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 EVP_PKEY_fromdata function has been augmented to allow for the derivation
32 of CRT (Chinese Remainder Theorem) parameters when requested. See the
33 OSSL_PKEY_PARAM_DERIVE_FROM_PQ param in the EVP_PKEY-RSA documentation.
37 * The activate and soft_load configuration settings for providers in
38 openssl.cnf have been updated to require a value of [1|yes|true|on]
39 (in lower or UPPER case) to enable the setting. Conversely a value
40 of [0|no|false|off] will disable the setting. All other values, or the
41 omission of a value for these settings will result in an error.
45 * Added `-set_issuer` and `-set_subject` options to `openssl x509` to
46 override the Issuer and Subject when creating a certificate. The `-subj`
47 option now is an alias for `-set_subject`.
49 *Job Snijders, George Michaelson*
51 * OPENSSL_sk_push() and sk_<TYPE>_push() functions now return 0 instead of -1
52 if called with a NULL stack argument.
56 * In `openssl speed`, changed the default hash function used with `hmac` from
61 * Added several new features of CMPv3 defined in RFC 9480 and RFC 9483:
62 - `certProfile` request message header and respective `-profile` CLI option
63 - support for delayed delivery of all types of response messages
67 * The build of exporters (such as `.pc` files for pkg-config) cleaned up to
68 be less hard coded in the build file templates, and to allow easier
69 addition of more exporters. With that, an exporter for CMake is also
74 * The BLAKE2s hash algorithm matches BLAKE2b's support
75 for configurable output length.
79 * New option `SSL_OP_PREFER_NO_DHE_KEX`, which allows configuring a TLS1.3
80 server to prefer session resumption using PSK-only key exchange over PSK
81 with DHE, if both are available.
83 *Markus Minichmayr, Tapkey GmbH*
85 * New API `SSL_write_ex2`, which can be used to send an end-of-stream (FIN)
86 condition in an optimised way when using QUIC.
93 ### Changes between 3.2.0 and 3.2.1 [xx XXX xxxx]
95 * A file in PKCS12 format can contain certificates and keys and may come from
96 an untrusted source. The PKCS12 specification allows certain fields to be
97 NULL, but OpenSSL did not correctly check for this case. A fix has been
98 applied to prevent a NULL pointer dereference that results in OpenSSL
99 crashing. If an application processes PKCS12 files from an untrusted source
100 using the OpenSSL APIs then that application will be vulnerable to this
101 issue prior to this fix.
103 OpenSSL APIs that were vulnerable to this are: PKCS12_parse(),
104 PKCS12_unpack_p7data(), PKCS12_unpack_p7encdata(), PKCS12_unpack_authsafes()
105 and PKCS12_newpass().
107 We have also fixed a similar issue in SMIME_write_PKCS7(). However since this
108 function is related to writing data we do not consider it security
115 * When function EVP_PKEY_public_check() is called on RSA public keys,
116 a computation is done to confirm that the RSA modulus, n, is composite.
117 For valid RSA keys, n is a product of two or more large primes and this
118 computation completes quickly. However, if n is an overly large prime,
119 then this computation would take a long time.
121 An application that calls EVP_PKEY_public_check() and supplies an RSA key
122 obtained from an untrusted source could be vulnerable to a Denial of Service
125 The function EVP_PKEY_public_check() is not called from other OpenSSL
126 functions however it is called from the OpenSSL pkey command line
127 application. For that reason that application is also vulnerable if used
128 with the "-pubin" and "-check" options on untrusted data.
130 To resolve this issue RSA keys larger than OPENSSL_RSA_MAX_MODULUS_BITS will
131 now fail the check immediately with an RSA_R_MODULUS_TOO_LARGE error reason.
137 * Restore the encoding of SM2 PrivateKeyInfo and SubjectPublicKeyInfo to
138 have the contained AlgorithmIdentifier.algorithm set to id-ecPublicKey
143 * The POLY1305 MAC (message authentication code) implementation in OpenSSL
144 for PowerPC CPUs saves the contents of vector registers in different
145 order than they are restored. Thus the contents of some of these vector
146 registers is corrupted when returning to the caller. The vulnerable code is
147 used only on newer PowerPC processors supporting the PowerISA 2.07
150 The consequences of this kind of internal application state corruption can
151 be various - from no consequences, if the calling application does not
152 depend on the contents of non-volatile XMM registers at all, to the worst
153 consequences, where the attacker could get complete control of the
154 application process. However unless the compiler uses the vector registers
155 for storing pointers, the most likely consequence, if any, would be an
156 incorrect result of some application dependent calculations or a crash
157 leading to a denial of service.
163 * Fix excessive time spent in DH check / generation with large Q parameter
166 Applications that use the functions DH_generate_key() to generate an
167 X9.42 DH key may experience long delays. Likewise, applications that use
168 DH_check_pub_key(), DH_check_pub_key_ex() or EVP_PKEY_public_check()
169 to check an X9.42 DH key or X9.42 DH parameters may experience long delays.
170 Where the key or parameters that are being checked have been obtained from
171 an untrusted source this may lead to a Denial of Service.
177 * Disable building QUIC server utility when OpenSSL is configured with
182 ### Changes between 3.1 and 3.2.0 [23 Nov 2023]
184 * The BLAKE2b hash algorithm supports a configurable output length
185 by setting the "size" parameter.
187 *Čestmír Kalina and Tomáš Mráz*
189 * Enable extra Arm64 optimization on Windows for GHASH, RAND and AES.
193 * Added a function to delete objects from store by URI - OSSL_STORE_delete()
194 and the corresponding provider-storemgmt API function
195 OSSL_FUNC_store_delete().
199 * Added OSSL_FUNC_store_open_ex() provider-storemgmt API function to pass
200 a passphrase callback when opening a store.
204 * Changed the default salt length used by PBES2 KDF's (PBKDF2 and scrypt)
205 from 8 bytes to 16 bytes.
206 The PKCS5 (RFC 8018) standard uses a 64 bit salt length for PBE, and
207 recommends a minimum of 64 bits for PBES2. For FIPS compliance PBKDF2
208 requires a salt length of 128 bits. This affects OpenSSL command line
209 applications such as "genrsa" and "pkcs8" and API's such as
210 PEM_write_bio_PrivateKey() that are reliant on the default value.
211 The additional commandline option 'saltlen' has been added to the
212 OpenSSL command line applications for "pkcs8" and "enc" to allow the
213 salt length to be set to a non default value.
217 * Changed the default value of the `ess_cert_id_alg` configuration
218 option which is used to calculate the TSA's public key certificate
219 identifier. The default algorithm is updated to be sha256 instead
222 *Małgorzata Olszówka*
224 * Added optimization for SM2 algorithm on aarch64. It uses a huge precomputed
225 table for point multiplication of the base point, which increases the size of
226 libcrypto from 4.4 MB to 4.9 MB. A new configure option `no-sm2-precomp` has
227 been added to disable the precomputed table.
231 * Added client side support for QUIC
233 *Hugo Landau, Matt Caswell, Paul Dale, Tomáš Mráz, Richard Levitte*
235 * Added multiple tutorials on the OpenSSL library and in particular
236 on writing various clients (using TLS and QUIC protocols) with libssl.
240 * Added secp384r1 implementation using Solinas' reduction to improve
241 speed of the NIST P-384 elliptic curve. To enable the implementation
242 the build option `enable-ec_nistp_64_gcc_128` must be used.
246 * Improved RFC7468 compliance of the asn1parse command.
248 *Matthias St. Pierre*
250 * Added SHA256/192 algorithm support.
254 * Added support for securely getting root CA certificate update in
259 * Improved contention on global write locks by using more read locks where
264 * Improved performance of OSSL_PARAM lookups in performance critical
269 * Added the SSL_get0_group_name() function to provide access to the
270 name of the group used for the TLS key exchange.
274 * Provide a new configure option `no-http` that can be used to disable the
275 HTTP support. Provide new configure options `no-apps` and `no-docs` to
276 disable building the openssl command line application and the documentation.
280 * Provide a new configure option `no-ecx` that can be used to disable the
281 X25519, X448, and EdDSA support.
285 * When multiple OSSL_KDF_PARAM_INFO parameters are passed to
286 the EVP_KDF_CTX_set_params() function they are now concatenated not just
287 for the HKDF algorithm but also for SSKDF and X9.63 KDF algorithms.
291 * Added OSSL_FUNC_keymgmt_im/export_types_ex() provider functions that get
292 the provider context as a parameter.
296 * TLS round-trip time calculation was added by a Brigham Young University
297 Capstone team partnering with Sandia National Laboratories. A new function
298 in ssl_lib titled SSL_get_handshake_rtt will calculate and retrieve this
303 * Added the "-quic" option to s_client to enable connectivity to QUIC servers.
304 QUIC requires the use of ALPN, so this must be specified via the "-alpn"
305 option. Use of the "advanced" s_client command command via the "-adv" option
310 * Added an "advanced" command mode to s_client. Use this with the "-adv"
311 option. The old "basic" command mode recognises certain letters that must
312 always appear at the start of a line and cannot be escaped. The advanced
313 command mode enables commands to be entered anywhere and there is an
314 escaping mechanism. After starting s_client with "-adv" type "{help}"
315 to show a list of available commands.
319 * Add Raw Public Key (RFC7250) support. Authentication is supported
320 by matching keys against either local policy (TLSA records synthesised
321 from the expected keys) or DANE (TLSA records obtained by the
322 application from DNS). TLSA records will also match the same key in
323 the server certificate, should RPK use not happen to be negotiated.
327 * Added support for modular exponentiation and CRT offloading for the
332 * Added further assembler code for the RISC-V architecture.
336 * Added EC_GROUP_to_params() which creates an OSSL_PARAM array
337 from a given EC_GROUP.
341 * Improved support for non-default library contexts and property queries
342 when parsing PKCS#12 files.
346 * Implemented support for all five instances of EdDSA from RFC8032:
347 Ed25519, Ed25519ctx, Ed25519ph, Ed448, and Ed448ph.
348 The streaming is not yet supported for the HashEdDSA variants
349 (Ed25519ph and Ed448ph).
353 * Added SM4 optimization for ARM processors using ASIMD and AES HW
358 * Implemented SM4-XTS support.
362 * Added platform-agnostic OSSL_sleep() function.
366 * Implemented deterministic ECDSA signatures (RFC6979) support.
370 * Implemented AES-GCM-SIV (RFC8452) support.
374 * Added support for pluggable (provider-based) TLS signature algorithms.
375 This enables TLS 1.3 authentication operations with algorithms embedded
376 in providers not included by default in OpenSSL. In combination with
377 the already available pluggable KEM and X.509 support, this enables
378 for example suitable providers to deliver post-quantum or quantum-safe
379 cryptography to OpenSSL users.
383 * Added support for pluggable (provider-based) CMS signature algorithms.
384 This enables CMS sign and verify operations with algorithms embedded
385 in providers not included by default in OpenSSL.
389 * Added support for Hybrid Public Key Encryption (HPKE) as defined
390 in RFC9180. HPKE is required for TLS Encrypted ClientHello (ECH),
391 Message Layer Security (MLS) and other IETF specifications.
392 HPKE can also be used by other applications that require
393 encrypting "to" an ECDH public key. External APIs are defined in
394 include/openssl/hpke.h and documented in doc/man3/OSSL_HPKE_CTX_new.pod
398 * Implemented HPKE DHKEM support in providers used by HPKE (RFC9180)
403 * Add support for certificate compression (RFC8879), including
404 library support for Brotli and Zstandard compression.
408 * Add the ability to add custom attributes to PKCS12 files. Add a new API
409 PKCS12_create_ex2, identical to the existing PKCS12_create_ex but allows
410 for a user specified callback and optional argument.
411 Added a new PKCS12_SAFEBAG_set0_attr, which allows for a new attr to be
412 added to the existing STACK_OF attrs.
416 * Major refactor of the libssl record layer.
420 * Add a mac salt length option for the pkcs12 command.
424 * Add more SRTP protection profiles from RFC8723 and RFC8269.
428 * Extended Kernel TLS (KTLS) to support TLS 1.3 receive offload.
430 *Daiki Ueno, John Baldwin and Dmitry Podgorny*
432 * Add support for TCP Fast Open (RFC7413) to macOS, Linux, and FreeBSD where
433 supported and enabled.
437 * Add ciphersuites based on DHE_PSK (RFC 4279) and ECDHE_PSK (RFC 5489)
438 to the list of ciphersuites providing Perfect Forward Secrecy as
439 required by SECLEVEL >= 3.
441 *Dmitry Belyavskiy, Nicola Tuveri*
443 * Add new SSL APIs to aid in efficiently implementing TLS/SSL fingerprinting.
444 The SSL_CTRL_GET_IANA_GROUPS control code, exposed as the
445 SSL_get0_iana_groups() function-like macro, retrieves the list of
446 supported groups sent by the peer.
447 The function SSL_client_hello_get_extension_order() populates
448 a caller-supplied array with the list of extension types present in the
449 ClientHello, in order of appearance.
453 * Fixed PEM_write_bio_PKCS8PrivateKey() and PEM_write_bio_PKCS8PrivateKey_nid()
454 to make it possible to use empty passphrase strings.
458 * The PKCS12_parse() function now supports MAC-less PKCS12 files.
462 * Added ASYNC_set_mem_functions() and ASYNC_get_mem_functions() calls to be able
463 to change functions used for allocating the memory of asynchronous call stack.
467 * Added support for signed BIGNUMs in the OSSL_PARAM APIs.
471 * A failure exit code is returned when using the openssl x509 command to check
472 certificate attributes and the checks fail.
476 * The default SSL/TLS security level has been changed from 1 to 2. RSA,
477 DSA and DH keys of 1024 bits and above and less than 2048 bits and ECC keys
478 of 160 bits and above and less than 224 bits were previously accepted by
479 default but are now no longer allowed. By default TLS compression was
480 already disabled in previous OpenSSL versions. At security level 2 it cannot
485 * The SSL_CTX_set_cipher_list family functions now accept ciphers using their
490 * The PVK key derivation function has been moved from b2i_PVK_bio_ex() into
491 the legacy crypto provider as an EVP_KDF. Applications requiring this KDF
492 will need to load the legacy crypto provider.
496 * CCM8 cipher suites in TLS have been downgraded to security level zero
497 because they use a short authentication tag which lowers their strength.
501 * Subject or issuer names in X.509 objects are now displayed as UTF-8 strings
506 * Add X.509 certificate codeSigning purpose and related checks on key usage and
507 extended key usage of the leaf certificate according to the CA/Browser Forum.
511 * The `x509`, `ca`, and `req` apps now produce X.509 v3 certificates.
512 The `-x509v1` option of `req` prefers generation of X.509 v1 certificates.
513 `X509_sign()` and `X509_sign_ctx()` make sure that the certificate has
514 X.509 version 3 if the certificate information includes X.509 extensions.
518 * Fix and extend certificate handling and the apps `x509`, `verify` etc.
519 such as adding a trace facility for debugging certificate chain building.
523 * Various fixes and extensions to the CMP+CRMF implementation and the `cmp` app
524 in particular supporting requests for central key generation, generalized
525 polling, and various types of genm/genp exchanges defined in CMP Updates.
529 * Fixes and extensions to the HTTP client and to the HTTP server in `apps/`
530 like correcting the TLS and proxy support and adding tracing for debugging.
534 * Extended the CMS API for handling `CMS_SignedData` and `CMS_EnvelopedData`.
538 * `CMS_add0_cert()` and `CMS_add1_cert()` no longer throw an error if
539 a certificate to be added is already present. `CMS_sign_ex()` and
540 `CMS_sign()` now ignore any duplicate certificates in their `certs` argument
541 and no longer throw an error for them.
545 * Fixed and extended `util/check-format.pl` for checking adherence to the
546 coding style <https://www.openssl.org/policies/technical/coding-style.html>.
547 The checks are meanwhile more complete and yield fewer false positives.
551 * Added BIO_s_dgram_pair() and BIO_s_dgram_mem() that provide memory-based
552 BIOs with datagram semantics and support for BIO_sendmmsg() and BIO_recvmmsg()
553 calls. They can be used as the transport BIOs for QUIC.
555 *Hugo Landau, Matt Caswell and Tomáš Mráz*
557 * Add new BIO_sendmmsg() and BIO_recvmmsg() BIO methods which allow
558 sending and receiving multiple messages in a single call. An implementation
559 is provided for BIO_dgram. For further details, see BIO_sendmmsg(3).
563 * Support for loading root certificates from the Windows certificate store
564 has been added. The support is in the form of a store which recognises the
565 URI string of `org.openssl.winstore://`. This URI scheme currently takes no
566 arguments. This store is built by default and can be disabled using the new
567 compile-time option `no-winstore`. This store is not currently used by
568 default and must be loaded explicitly using the above store URI. It is
569 expected to be loaded by default in the future.
573 * Enable KTLS with the TLS 1.3 CCM mode ciphersuites. Note that some linux
574 kernel versions that support KTLS have a known bug in CCM processing. That
575 has been fixed in stable releases starting from 5.4.164, 5.10.84, 5.15.7,
576 and all releases since 5.16. KTLS with CCM ciphersuites should be only used
581 * Added `-ktls` option to `s_server` and `s_client` commands to enable the
586 * Zerocopy KTLS sendfile() support on Linux.
590 * The OBJ_ calls are now thread safe using a global lock.
594 * New parameter `-digest` for openssl cms command allowing signing
595 pre-computed digests and new CMS API functions supporting that
600 * OPENSSL_malloc() and other allocation functions now raise errors on
601 allocation failures. The callers do not need to explicitly raise errors
602 unless they want to for tracing purposes.
606 * Added and enabled by default implicit rejection in RSA PKCS#1 v1.5
607 decryption as a protection against Bleichenbacher-like attacks.
608 The RSA decryption API will now return a randomly generated deterministic
609 message instead of an error in case it detects an error when checking
610 padding during PKCS#1 v1.5 decryption. This is a general protection against
611 issues like CVE-2020-25659 and CVE-2020-25657. This protection can be
613 `EVP_PKEY_CTX_ctrl_str(ctx, "rsa_pkcs1_implicit_rejection". "0")`
614 on the RSA decryption context.
618 * Added support for Brainpool curves in TLS-1.3.
620 *Bernd Edlinger and Matt Caswell*
622 * Added OpenBSD specific build targets.
626 * Support for Argon2d, Argon2i, Argon2id KDFs has been added along with
627 a basic thread pool implementation for select platforms.
634 ### Changes between 3.1.3 and 3.1.4 [24 Oct 2023]
636 * Fix incorrect key and IV resizing issues when calling EVP_EncryptInit_ex2(),
637 EVP_DecryptInit_ex2() or EVP_CipherInit_ex2() with OSSL_PARAM parameters
638 that alter the key or IV length ([CVE-2023-5363]).
642 ### Changes between 3.1.2 and 3.1.3 [19 Sep 2023]
644 * Fix POLY1305 MAC implementation corrupting XMM registers on Windows.
646 The POLY1305 MAC (message authentication code) implementation in OpenSSL
647 does not save the contents of non-volatile XMM registers on Windows 64
648 platform when calculating the MAC of data larger than 64 bytes. Before
649 returning to the caller all the XMM registers are set to zero rather than
650 restoring their previous content. The vulnerable code is used only on newer
651 x86_64 processors supporting the AVX512-IFMA instructions.
653 The consequences of this kind of internal application state corruption can
654 be various - from no consequences, if the calling application does not
655 depend on the contents of non-volatile XMM registers at all, to the worst
656 consequences, where the attacker could get complete control of the
657 application process. However given the contents of the registers are just
658 zeroized so the attacker cannot put arbitrary values inside, the most likely
659 consequence, if any, would be an incorrect result of some application
660 dependent calculations or a crash leading to a denial of service.
666 ### Changes between 3.1.1 and 3.1.2 [1 Aug 2023]
668 * Fix excessive time spent checking DH q parameter value.
670 The function DH_check() performs various checks on DH parameters. After
671 fixing CVE-2023-3446 it was discovered that a large q parameter value can
672 also trigger an overly long computation during some of these checks.
673 A correct q value, if present, cannot be larger than the modulus p
674 parameter, thus it is unnecessary to perform these checks if q is larger
677 If DH_check() is called with such q parameter value,
678 DH_CHECK_INVALID_Q_VALUE return flag is set and the computationally
679 intensive checks are skipped.
685 * Fix DH_check() excessive time with over sized modulus.
687 The function DH_check() performs various checks on DH parameters. One of
688 those checks confirms that the modulus ("p" parameter) is not too large.
689 Trying to use a very large modulus is slow and OpenSSL will not normally use
690 a modulus which is over 10,000 bits in length.
692 However the DH_check() function checks numerous aspects of the key or
693 parameters that have been supplied. Some of those checks use the supplied
694 modulus value even if it has already been found to be too large.
696 A new limit has been added to DH_check of 32,768 bits. Supplying a
697 key/parameters with a modulus over this size will simply cause DH_check() to
704 * Do not ignore empty associated data entries with AES-SIV.
706 The AES-SIV algorithm allows for authentication of multiple associated
707 data entries along with the encryption. To authenticate empty data the
708 application has to call `EVP_EncryptUpdate()` (or `EVP_CipherUpdate()`)
709 with NULL pointer as the output buffer and 0 as the input buffer length.
710 The AES-SIV implementation in OpenSSL just returns success for such call
711 instead of performing the associated data authentication operation.
712 The empty data thus will not be authenticated. ([CVE-2023-2975])
714 Thanks to Juerg Wullschleger (Google) for discovering the issue.
716 The fix changes the authentication tag value and the ciphertext for
717 applications that use empty associated data entries with AES-SIV.
718 To decrypt data encrypted with previous versions of OpenSSL the application
719 has to skip calls to `EVP_DecryptUpdate()` for empty associated data
724 * When building with the `enable-fips` option and using the resulting
725 FIPS provider, TLS 1.2 will, by default, mandate the use of an extended
726 master secret (FIPS 140-3 IG G.Q) and the Hash and HMAC DRBGs will
727 not operate with truncated digests (FIPS 140-3 IG G.R).
731 ### Changes between 3.1.0 and 3.1.1 [30 May 2023]
733 * Mitigate for the time it takes for `OBJ_obj2txt` to translate gigantic
734 OBJECT IDENTIFIER sub-identifiers to canonical numeric text form.
736 OBJ_obj2txt() would translate any size OBJECT IDENTIFIER to canonical
737 numeric text form. For gigantic sub-identifiers, this would take a very
738 long time, the time complexity being O(n^2) where n is the size of that
739 sub-identifier. ([CVE-2023-2650])
741 To mitigitate this, `OBJ_obj2txt()` will only translate an OBJECT
742 IDENTIFIER to canonical numeric text form if the size of that OBJECT
743 IDENTIFIER is 586 bytes or less, and fail otherwise.
745 The basis for this restriction is [RFC 2578 (STD 58), section 3.5]. OBJECT
746 IDENTIFIER values, which stipulates that OBJECT IDENTIFIERS may have at
747 most 128 sub-identifiers, and that the maximum value that each sub-
748 identifier may have is 2^32-1 (4294967295 decimal).
750 For each byte of every sub-identifier, only the 7 lower bits are part of
751 the value, so the maximum amount of bytes that an OBJECT IDENTIFIER with
752 these restrictions may occupy is 32 * 128 / 7, which is approximately 586
757 * Multiple algorithm implementation fixes for ARM BE platforms.
761 * Added a -pedantic option to fipsinstall that adjusts the various
762 settings to ensure strict FIPS compliance rather than backwards
767 * Fixed buffer overread in AES-XTS decryption on ARM 64 bit platforms which
768 happens if the buffer size is 4 mod 5 in 16 byte AES blocks. This can
769 trigger a crash of an application using AES-XTS decryption if the memory
770 just after the buffer being decrypted is not mapped.
771 Thanks to Anton Romanov (Amazon) for discovering the issue.
776 * Reworked the Fix for the Timing Oracle in RSA Decryption ([CVE-2022-4304]).
777 The previous fix for this timing side channel turned out to cause
778 a severe 2-3x performance regression in the typical use case
779 compared to 3.0.7. The new fix uses existing constant time
780 code paths, and restores the previous performance level while
781 fully eliminating all existing timing side channels.
782 The fix was developed by Bernd Edlinger with testing support
787 * Add FIPS provider configuration option to disallow the use of
788 truncated digests with Hash and HMAC DRBGs (q.v. FIPS 140-3 IG D.R.).
789 The option '-no_drbg_truncated_digests' can optionally be
790 supplied to 'openssl fipsinstall'.
794 * Corrected documentation of X509_VERIFY_PARAM_add0_policy() to mention
795 that it does not enable policy checking. Thanks to David Benjamin for
796 discovering this issue.
801 * Fixed an issue where invalid certificate policies in leaf certificates are
802 silently ignored by OpenSSL and other certificate policy checks are skipped
803 for that certificate. A malicious CA could use this to deliberately assert
804 invalid certificate policies in order to circumvent policy checking on the
805 certificate altogether.
810 * Limited the number of nodes created in a policy tree to mitigate
811 against CVE-2023-0464. The default limit is set to 1000 nodes, which
812 should be sufficient for most installations. If required, the limit
813 can be adjusted by setting the OPENSSL_POLICY_TREE_NODES_MAX build
814 time define to a desired maximum number of nodes or zero to allow
820 ### Changes between 3.0 and 3.1.0 [14 Mar 2023]
822 * Add FIPS provider configuration option to enforce the
823 Extended Master Secret (EMS) check during the TLS1_PRF KDF.
824 The option '-ems-check' can optionally be supplied to
825 'openssl fipsinstall'.
829 * The FIPS provider includes a few non-approved algorithms for
830 backward compatibility purposes and the "fips=yes" property query
831 must be used for all algorithm fetches to ensure FIPS compliance.
833 The algorithms that are included but not approved are Triple DES ECB,
834 Triple DES CBC and EdDSA.
838 * Added support for KMAC in KBKDF.
842 * RNDR and RNDRRS support in provider functions to provide
843 random number generation for Arm CPUs (aarch64).
847 * s_client and s_server apps now explicitly say when the TLS version
848 does not include the renegotiation mechanism. This avoids confusion
849 between that scenario versus when the TLS version includes secure
850 renegotiation but the peer lacks support for it.
854 * AES-GCM enabled with AVX512 vAES and vPCLMULQDQ.
856 *Tomasz Kantecki, Andrey Matyukov*
858 * The various OBJ_* functions have been made thread safe.
862 * Parallel dual-prime 1536/2048-bit modular exponentiation for
863 AVX512_IFMA capable processors.
865 *Sergey Kirillov, Andrey Matyukov (Intel Corp)*
867 * The functions `OPENSSL_LH_stats`, `OPENSSL_LH_node_stats`,
868 `OPENSSL_LH_node_usage_stats`, `OPENSSL_LH_stats_bio`,
869 `OPENSSL_LH_node_stats_bio` and `OPENSSL_LH_node_usage_stats_bio` are now
870 marked deprecated from OpenSSL 3.1 onwards and can be disabled by defining
871 `OPENSSL_NO_DEPRECATED_3_1`.
873 The macro `DEFINE_LHASH_OF` is now deprecated in favour of the macro
874 `DEFINE_LHASH_OF_EX`, which omits the corresponding type-specific function
875 definitions for these functions regardless of whether
876 `OPENSSL_NO_DEPRECATED_3_1` is defined.
878 Users of `DEFINE_LHASH_OF` may start receiving deprecation warnings for these
879 functions regardless of whether they are using them. It is recommended that
880 users transition to the new macro, `DEFINE_LHASH_OF_EX`.
884 * When generating safe-prime DH parameters set the recommended private key
885 length equivalent to minimum key lengths as in RFC 7919.
889 * Change the default salt length for PKCS#1 RSASSA-PSS signatures to the
890 maximum size that is smaller or equal to the digest length to comply with
891 FIPS 186-4 section 5. This is implemented by a new option
892 `OSSL_PKEY_RSA_PSS_SALT_LEN_AUTO_DIGEST_MAX` ("auto-digestmax") for the
893 `rsa_pss_saltlen` parameter, which is now the default. Signature
894 verification is not affected by this change and continues to work as before.
901 For OpenSSL 3.0 a [Migration guide][] has been added, so the CHANGES entries
902 listed here are only a brief description.
903 The migration guide contains more detailed information related to new features,
904 breaking changes, and mappings for the large list of deprecated functions.
906 [Migration guide]: https://github.com/openssl/openssl/tree/master/doc/man7/migration_guide.pod
908 ### Changes between 3.0.7 and 3.0.8 [7 Feb 2023]
910 * Fixed NULL dereference during PKCS7 data verification.
912 A NULL pointer can be dereferenced when signatures are being
913 verified on PKCS7 signed or signedAndEnveloped data. In case the hash
914 algorithm used for the signature is known to the OpenSSL library but
915 the implementation of the hash algorithm is not available the digest
916 initialization will fail. There is a missing check for the return
917 value from the initialization function which later leads to invalid
918 usage of the digest API most likely leading to a crash.
921 PKCS7 data is processed by the SMIME library calls and also by the
922 time stamp (TS) library calls. The TLS implementation in OpenSSL does
923 not call these functions however third party applications would be
924 affected if they call these functions to verify signatures on untrusted
929 * Fixed X.400 address type confusion in X.509 GeneralName.
931 There is a type confusion vulnerability relating to X.400 address processing
932 inside an X.509 GeneralName. X.400 addresses were parsed as an ASN1_STRING
933 but the public structure definition for GENERAL_NAME incorrectly specified
934 the type of the x400Address field as ASN1_TYPE. This field is subsequently
935 interpreted by the OpenSSL function GENERAL_NAME_cmp as an ASN1_TYPE rather
938 When CRL checking is enabled (i.e. the application sets the
939 X509_V_FLAG_CRL_CHECK flag), this vulnerability may allow an attacker to
940 pass arbitrary pointers to a memcmp call, enabling them to read memory
941 contents or enact a denial of service.
946 * Fixed NULL dereference validating DSA public key.
948 An invalid pointer dereference on read can be triggered when an
949 application tries to check a malformed DSA public key by the
950 EVP_PKEY_public_check() function. This will most likely lead
951 to an application crash. This function can be called on public
952 keys supplied from untrusted sources which could allow an attacker
953 to cause a denial of service attack.
955 The TLS implementation in OpenSSL does not call this function
956 but applications might call the function if there are additional
957 security requirements imposed by standards such as FIPS 140-3.
960 *Shane Lontis, Tomáš Mráz*
962 * Fixed Invalid pointer dereference in d2i_PKCS7 functions.
964 An invalid pointer dereference on read can be triggered when an
965 application tries to load malformed PKCS7 data with the
966 d2i_PKCS7(), d2i_PKCS7_bio() or d2i_PKCS7_fp() functions.
968 The result of the dereference is an application crash which could
969 lead to a denial of service attack. The TLS implementation in OpenSSL
970 does not call this function however third party applications might
971 call these functions on untrusted data.
976 * Fixed Use-after-free following BIO_new_NDEF.
978 The public API function BIO_new_NDEF is a helper function used for
979 streaming ASN.1 data via a BIO. It is primarily used internally to OpenSSL
980 to support the SMIME, CMS and PKCS7 streaming capabilities, but may also
981 be called directly by end user applications.
983 The function receives a BIO from the caller, prepends a new BIO_f_asn1
984 filter BIO onto the front of it to form a BIO chain, and then returns
985 the new head of the BIO chain to the caller. Under certain conditions,
986 for example if a CMS recipient public key is invalid, the new filter BIO
987 is freed and the function returns a NULL result indicating a failure.
988 However, in this case, the BIO chain is not properly cleaned up and the
989 BIO passed by the caller still retains internal pointers to the previously
990 freed filter BIO. If the caller then goes on to call BIO_pop() on the BIO
991 then a use-after-free will occur. This will most likely result in a crash.
994 *Viktor Dukhovni, Matt Caswell*
996 * Fixed Double free after calling PEM_read_bio_ex.
998 The function PEM_read_bio_ex() reads a PEM file from a BIO and parses and
999 decodes the "name" (e.g. "CERTIFICATE"), any header data and the payload
1000 data. If the function succeeds then the "name_out", "header" and "data"
1001 arguments are populated with pointers to buffers containing the relevant
1002 decoded data. The caller is responsible for freeing those buffers. It is
1003 possible to construct a PEM file that results in 0 bytes of payload data.
1004 In this case PEM_read_bio_ex() will return a failure code but will populate
1005 the header argument with a pointer to a buffer that has already been freed.
1006 If the caller also frees this buffer then a double free will occur. This
1007 will most likely lead to a crash.
1009 The functions PEM_read_bio() and PEM_read() are simple wrappers around
1010 PEM_read_bio_ex() and therefore these functions are also directly affected.
1012 These functions are also called indirectly by a number of other OpenSSL
1013 functions including PEM_X509_INFO_read_bio_ex() and
1014 SSL_CTX_use_serverinfo_file() which are also vulnerable. Some OpenSSL
1015 internal uses of these functions are not vulnerable because the caller does
1016 not free the header argument if PEM_read_bio_ex() returns a failure code.
1019 *Kurt Roeckx, Matt Caswell*
1021 * Fixed Timing Oracle in RSA Decryption.
1023 A timing based side channel exists in the OpenSSL RSA Decryption
1024 implementation which could be sufficient to recover a plaintext across
1025 a network in a Bleichenbacher style attack. To achieve a successful
1026 decryption an attacker would have to be able to send a very large number
1027 of trial messages for decryption. The vulnerability affects all RSA padding
1028 modes: PKCS#1 v1.5, RSA-OEAP and RSASVE.
1031 *Dmitry Belyavsky, Hubert Kario*
1033 * Fixed X.509 Name Constraints Read Buffer Overflow.
1035 A read buffer overrun can be triggered in X.509 certificate verification,
1036 specifically in name constraint checking. The read buffer overrun might
1037 result in a crash which could lead to a denial of service attack.
1038 In a TLS client, this can be triggered by connecting to a malicious
1039 server. In a TLS server, this can be triggered if the server requests
1040 client authentication and a malicious client connects.
1045 * Fixed X.509 Policy Constraints Double Locking security issue.
1047 If an X.509 certificate contains a malformed policy constraint and
1048 policy processing is enabled, then a write lock will be taken twice
1049 recursively. On some operating systems (most widely: Windows) this
1050 results in a denial of service when the affected process hangs. Policy
1051 processing being enabled on a publicly facing server is not considered
1052 to be a common setup.
1057 * Our provider implementations of `OSSL_FUNC_KEYMGMT_EXPORT` and
1058 `OSSL_FUNC_KEYMGMT_GET_PARAMS` for EC and SM2 keys now honor
1059 `OSSL_PKEY_PARAM_EC_POINT_CONVERSION_FORMAT` as set (and
1060 default to `POINT_CONVERSION_UNCOMPRESSED`) when exporting
1061 `OSSL_PKEY_PARAM_PUB_KEY`, instead of unconditionally using
1062 `POINT_CONVERSION_COMPRESSED` as in previous 3.x releases.
1063 For symmetry, our implementation of `EVP_PKEY_ASN1_METHOD->export_to`
1064 for legacy EC and SM2 keys is also changed similarly to honor the
1065 equivalent conversion format flag as specified in the underlying
1066 `EC_KEY` object being exported to a provider, when this function is
1067 called through `EVP_PKEY_export()`.
1071 ### Changes between 3.0.6 and 3.0.7 [1 Nov 2022]
1073 * Fixed two buffer overflows in punycode decoding functions.
1075 A buffer overrun can be triggered in X.509 certificate verification,
1076 specifically in name constraint checking. Note that this occurs after
1077 certificate chain signature verification and requires either a CA to
1078 have signed the malicious certificate or for the application to continue
1079 certificate verification despite failure to construct a path to a trusted
1082 In a TLS client, this can be triggered by connecting to a malicious
1083 server. In a TLS server, this can be triggered if the server requests
1084 client authentication and a malicious client connects.
1086 An attacker can craft a malicious email address to overflow
1087 an arbitrary number of bytes containing the `.` character (decimal 46)
1088 on the stack. This buffer overflow could result in a crash (causing a
1092 An attacker can craft a malicious email address to overflow four
1093 attacker-controlled bytes on the stack. This buffer overflow could
1094 result in a crash (causing a denial of service) or potentially remote code
1095 execution depending on stack layout for any given platform/compiler.
1100 * Removed all references to invalid OSSL_PKEY_PARAM_RSA names for CRT
1101 parameters in OpenSSL code.
1102 Applications should not use the names OSSL_PKEY_PARAM_RSA_FACTOR,
1103 OSSL_PKEY_PARAM_RSA_EXPONENT and OSSL_PKEY_PARAM_RSA_COEFFICIENT.
1104 Use the numbered names such as OSSL_PKEY_PARAM_RSA_FACTOR1 instead.
1105 Using these invalid names may cause algorithms to use slower methods
1106 that ignore the CRT parameters.
1110 * Fixed a regression introduced in 3.0.6 version raising errors on some stack
1115 * Fixed a regression introduced in 3.0.6 version not refreshing the certificate
1116 data to be signed before signing the certificate.
1120 * Added RIPEMD160 to the default provider.
1124 * Ensured that the key share group sent or accepted for the key exchange
1125 is allowed for the protocol version.
1129 ### Changes between 3.0.5 and 3.0.6 [11 Oct 2022]
1131 * OpenSSL supports creating a custom cipher via the legacy
1132 EVP_CIPHER_meth_new() function and associated function calls. This function
1133 was deprecated in OpenSSL 3.0 and application authors are instead encouraged
1134 to use the new provider mechanism in order to implement custom ciphers.
1136 OpenSSL versions 3.0.0 to 3.0.5 incorrectly handle legacy custom ciphers
1137 passed to the EVP_EncryptInit_ex2(), EVP_DecryptInit_ex2() and
1138 EVP_CipherInit_ex2() functions (as well as other similarly named encryption
1139 and decryption initialisation functions). Instead of using the custom cipher
1140 directly it incorrectly tries to fetch an equivalent cipher from the
1141 available providers. An equivalent cipher is found based on the NID passed to
1142 EVP_CIPHER_meth_new(). This NID is supposed to represent the unique NID for a
1143 given cipher. However it is possible for an application to incorrectly pass
1144 NID_undef as this value in the call to EVP_CIPHER_meth_new(). When NID_undef
1145 is used in this way the OpenSSL encryption/decryption initialisation function
1146 will match the NULL cipher as being equivalent and will fetch this from the
1147 available providers. This will succeed if the default provider has been
1148 loaded (or if a third party provider has been loaded that offers this
1149 cipher). Using the NULL cipher means that the plaintext is emitted as the
1152 Applications are only affected by this issue if they call
1153 EVP_CIPHER_meth_new() using NID_undef and subsequently use it in a call to an
1154 encryption/decryption initialisation function. Applications that only use
1155 SSL/TLS are not impacted by this issue.
1160 * Fix LLVM vs Apple LLVM version numbering confusion that caused build failures
1165 * Fixed the linux-mips64 Configure target which was missing the
1166 SIXTY_FOUR_BIT bn_ops flag. This was causing heap corruption on that
1171 * Fix handling of a ticket key callback that returns 0 in TLSv1.3 to not send a
1176 * Correctly handle a retransmitted ClientHello in DTLS
1180 * Fixed detection of ktls support in cross-compile environment on Linux
1184 * Fixed some regressions and test failures when running the 3.0.0 FIPS provider
1189 * Fixed SSL_pending() and SSL_has_pending() with DTLS which were failing to
1190 report correct results in some cases
1194 * Fix UWP builds by defining VirtualLock
1198 * For known safe primes use the minimum key length according to RFC 7919.
1199 Longer private key sizes unnecessarily raise the cycles needed to compute the
1200 shared secret without any increase of the real security. This fixes a
1201 regression from 1.1.1 where these shorter keys were generated for the known
1206 * Added the loongarch64 target
1210 * Fixed EC ASM flag passing. Flags for ASM implementations of EC curves were
1211 only passed to the FIPS provider and not to the default or legacy provider.
1215 * Fixed reported performance degradation on aarch64. Restored the
1216 implementation prior to commit 2621751 ("aes/asm/aesv8-armx.pl: avoid
1217 32-bit lane assignment in CTR mode") for 64bit targets only, since it is
1218 reportedly 2-17% slower and the silicon errata only affects 32bit targets.
1219 The new algorithm is still used for 32 bit targets.
1223 * Added a missing header for memcmp that caused compilation failure on some
1228 ### Changes between 3.0.4 and 3.0.5 [5 Jul 2022]
1230 * The OpenSSL 3.0.4 release introduced a serious bug in the RSA
1231 implementation for X86_64 CPUs supporting the AVX512IFMA instructions.
1232 This issue makes the RSA implementation with 2048 bit private keys
1233 incorrect on such machines and memory corruption will happen during
1234 the computation. As a consequence of the memory corruption an attacker
1235 may be able to trigger a remote code execution on the machine performing
1238 SSL/TLS servers or other servers using 2048 bit RSA private keys running
1239 on machines supporting AVX512IFMA instructions of the X86_64 architecture
1240 are affected by this issue.
1245 * AES OCB mode for 32-bit x86 platforms using the AES-NI assembly optimised
1246 implementation would not encrypt the entirety of the data under some
1247 circumstances. This could reveal sixteen bytes of data that was
1248 preexisting in the memory that wasn't written. In the special case of
1249 "in place" encryption, sixteen bytes of the plaintext would be revealed.
1251 Since OpenSSL does not support OCB based cipher suites for TLS and DTLS,
1252 they are both unaffected.
1255 *Alex Chernyakhovsky, David Benjamin, Alejandro Sedeño*
1257 ### Changes between 3.0.3 and 3.0.4 [21 Jun 2022]
1259 * In addition to the c_rehash shell command injection identified in
1260 CVE-2022-1292, further bugs where the c_rehash script does not
1261 properly sanitise shell metacharacters to prevent command injection have been
1264 When the CVE-2022-1292 was fixed it was not discovered that there
1265 are other places in the script where the file names of certificates
1266 being hashed were possibly passed to a command executed through the shell.
1268 This script is distributed by some operating systems in a manner where
1269 it is automatically executed. On such operating systems, an attacker
1270 could execute arbitrary commands with the privileges of the script.
1272 Use of the c_rehash script is considered obsolete and should be replaced
1273 by the OpenSSL rehash command line tool.
1276 *Daniel Fiala, Tomáš Mráz*
1278 * Case insensitive string comparison no longer uses locales. It has instead
1279 been directly implemented.
1283 ### Changes between 3.0.2 and 3.0.3 [3 May 2022]
1285 * Case insensitive string comparison is reimplemented via new locale-agnostic
1286 comparison functions OPENSSL_str[n]casecmp always using the POSIX locale for
1287 comparison. The previous implementation had problems when the Turkish locale
1292 * Fixed a bug in the c_rehash script which was not properly sanitising shell
1293 metacharacters to prevent command injection. This script is distributed by
1294 some operating systems in a manner where it is automatically executed. On
1295 such operating systems, an attacker could execute arbitrary commands with the
1296 privileges of the script.
1298 Use of the c_rehash script is considered obsolete and should be replaced
1299 by the OpenSSL rehash command line tool.
1304 * Fixed a bug in the function `OCSP_basic_verify` that verifies the signer
1305 certificate on an OCSP response. The bug caused the function in the case
1306 where the (non-default) flag OCSP_NOCHECKS is used to return a postivie
1307 response (meaning a successful verification) even in the case where the
1308 response signing certificate fails to verify.
1310 It is anticipated that most users of `OCSP_basic_verify` will not use the
1311 OCSP_NOCHECKS flag. In this case the `OCSP_basic_verify` function will return
1312 a negative value (indicating a fatal error) in the case of a certificate
1313 verification failure. The normal expected return value in this case would be
1316 This issue also impacts the command line OpenSSL "ocsp" application. When
1317 verifying an ocsp response with the "-no_cert_checks" option the command line
1318 application will report that the verification is successful even though it
1319 has in fact failed. In this case the incorrect successful response will also
1320 be accompanied by error messages showing the failure and contradicting the
1321 apparently successful result.
1326 * Fixed a bug where the RC4-MD5 ciphersuite incorrectly used the
1327 AAD data as the MAC key. This made the MAC key trivially predictable.
1329 An attacker could exploit this issue by performing a man-in-the-middle attack
1330 to modify data being sent from one endpoint to an OpenSSL 3.0 recipient such
1331 that the modified data would still pass the MAC integrity check.
1333 Note that data sent from an OpenSSL 3.0 endpoint to a non-OpenSSL 3.0
1334 endpoint will always be rejected by the recipient and the connection will
1335 fail at that point. Many application protocols require data to be sent from
1336 the client to the server first. Therefore, in such a case, only an OpenSSL
1337 3.0 server would be impacted when talking to a non-OpenSSL 3.0 client.
1339 If both endpoints are OpenSSL 3.0 then the attacker could modify data being
1340 sent in both directions. In this case both clients and servers could be
1341 affected, regardless of the application protocol.
1343 Note that in the absence of an attacker this bug means that an OpenSSL 3.0
1344 endpoint communicating with a non-OpenSSL 3.0 endpoint will fail to complete
1345 the handshake when using this ciphersuite.
1347 The confidentiality of data is not impacted by this issue, i.e. an attacker
1348 cannot decrypt data that has been encrypted using this ciphersuite - they can
1351 In order for this attack to work both endpoints must legitimately negotiate
1352 the RC4-MD5 ciphersuite. This ciphersuite is not compiled by default in
1353 OpenSSL 3.0, and is not available within the default provider or the default
1354 ciphersuite list. This ciphersuite will never be used if TLSv1.3 has been
1355 negotiated. In order for an OpenSSL 3.0 endpoint to use this ciphersuite the
1356 following must have occurred:
1358 1) OpenSSL must have been compiled with the (non-default) compile time option
1359 enable-weak-ssl-ciphers
1361 2) OpenSSL must have had the legacy provider explicitly loaded (either
1362 through application code or via configuration)
1364 3) The ciphersuite must have been explicitly added to the ciphersuite list
1366 4) The libssl security level must have been set to 0 (default is 1)
1368 5) A version of SSL/TLS below TLSv1.3 must have been negotiated
1370 6) Both endpoints must negotiate the RC4-MD5 ciphersuite in preference to any
1371 others that both endpoints have in common
1376 * Fix a bug in the OPENSSL_LH_flush() function that breaks reuse of the memory
1377 occupied by the removed hash table entries.
1379 This function is used when decoding certificates or keys. If a long lived
1380 process periodically decodes certificates or keys its memory usage will
1381 expand without bounds and the process might be terminated by the operating
1382 system causing a denial of service. Also traversing the empty hash table
1383 entries will take increasingly more time.
1385 Typically such long lived processes might be TLS clients or TLS servers
1386 configured to accept client certificate authentication.
1389 *Hugo Landau, Aliaksei Levin*
1391 * The functions `OPENSSL_LH_stats` and `OPENSSL_LH_stats_bio` now only report
1392 the `num_items`, `num_nodes` and `num_alloc_nodes` statistics. All other
1393 statistics are no longer supported. For compatibility, these statistics are
1394 still listed in the output but are now always reported as zero.
1398 ### Changes between 3.0.1 and 3.0.2 [15 Mar 2022]
1400 * Fixed a bug in the BN_mod_sqrt() function that can cause it to loop forever
1401 for non-prime moduli.
1403 Internally this function is used when parsing certificates that contain
1404 elliptic curve public keys in compressed form or explicit elliptic curve
1405 parameters with a base point encoded in compressed form.
1407 It is possible to trigger the infinite loop by crafting a certificate that
1408 has invalid explicit curve parameters.
1410 Since certificate parsing happens prior to verification of the certificate
1411 signature, any process that parses an externally supplied certificate may thus
1412 be subject to a denial of service attack. The infinite loop can also be
1413 reached when parsing crafted private keys as they can contain explicit
1414 elliptic curve parameters.
1416 Thus vulnerable situations include:
1418 - TLS clients consuming server certificates
1419 - TLS servers consuming client certificates
1420 - Hosting providers taking certificates or private keys from customers
1421 - Certificate authorities parsing certification requests from subscribers
1422 - Anything else which parses ASN.1 elliptic curve parameters
1424 Also any other applications that use the BN_mod_sqrt() where the attacker
1425 can control the parameter values are vulnerable to this DoS issue.
1430 * Add ciphersuites based on DHE_PSK (RFC 4279) and ECDHE_PSK (RFC 5489)
1431 to the list of ciphersuites providing Perfect Forward Secrecy as
1432 required by SECLEVEL >= 3.
1434 *Dmitry Belyavskiy, Nicola Tuveri*
1436 * Made the AES constant time code for no-asm configurations
1437 optional due to the resulting 95% performance degradation.
1438 The AES constant time code can be enabled, for no assembly
1439 builds, with: ./config no-asm -DOPENSSL_AES_CONST_TIME
1443 * Fixed PEM_write_bio_PKCS8PrivateKey() to make it possible to use empty
1448 * The negative return value handling of the certificate verification callback
1449 was reverted. The replacement is to set the verification retry state with
1450 the SSL_set_retry_verify() function.
1454 ### Changes between 3.0.0 and 3.0.1 [14 Dec 2021]
1456 * Fixed invalid handling of X509_verify_cert() internal errors in libssl
1457 Internally libssl in OpenSSL calls X509_verify_cert() on the client side to
1458 verify a certificate supplied by a server. That function may return a
1459 negative return value to indicate an internal error (for example out of
1460 memory). Such a negative return value is mishandled by OpenSSL and will cause
1461 an IO function (such as SSL_connect() or SSL_do_handshake()) to not indicate
1462 success and a subsequent call to SSL_get_error() to return the value
1463 SSL_ERROR_WANT_RETRY_VERIFY. This return value is only supposed to be
1464 returned by OpenSSL if the application has previously called
1465 SSL_CTX_set_cert_verify_callback(). Since most applications do not do this
1466 the SSL_ERROR_WANT_RETRY_VERIFY return value from SSL_get_error() will be
1467 totally unexpected and applications may not behave correctly as a result. The
1468 exact behaviour will depend on the application but it could result in
1469 crashes, infinite loops or other similar incorrect responses.
1471 This issue is made more serious in combination with a separate bug in OpenSSL
1472 3.0 that will cause X509_verify_cert() to indicate an internal error when
1473 processing a certificate chain. This will occur where a certificate does not
1474 include the Subject Alternative Name extension but where a Certificate
1475 Authority has enforced name constraints. This issue can occur even with valid
1481 * Corrected a few file name and file reference bugs in the build,
1482 installation and setup scripts, which lead to installation verification
1483 failures. Slightly enhanced the installation verification script.
1487 * Fixed EVP_PKEY_eq() to make it possible to use it with strictly private
1492 * Fixed PVK encoder to properly query for the passphrase.
1496 * Multiple fixes in the OSSL_HTTP API functions.
1500 * Allow sign extension in OSSL_PARAM_allocate_from_text() for the
1501 OSSL_PARAM_INTEGER data type and return error on negative numbers
1502 used with the OSSL_PARAM_UNSIGNED_INTEGER data type. Make
1503 OSSL_PARAM_BLD_push_BN{,_pad}() return an error on negative numbers.
1507 * Allow copying uninitialized digest contexts with EVP_MD_CTX_copy_ex.
1511 * Fixed detection of ARMv7 and ARM64 CPU features on FreeBSD.
1515 * Multiple threading fixes.
1519 * Added NULL digest implementation to keep compatibility with 1.1.1 version.
1523 * Allow fetching an operation from the provider that owns an unexportable key
1524 as a fallback if that is still allowed by the property query.
1528 ### Changes between 1.1.1 and 3.0.0 [7 Sep 2021]
1530 * TLS_MAX_VERSION, DTLS_MAX_VERSION and DTLS_MIN_VERSION constants are now
1535 * The `OPENSSL_s390xcap` environment variable can be used to set bits in the
1536 S390X capability vector to zero. This simplifies testing of different code
1537 paths on S390X architecture.
1541 * Encrypting more than 2^64 TLS records with AES-GCM is disallowed
1542 as per FIPS 140-2 IG A.5 "Key/IV Pair Uniqueness Requirements from
1543 SP 800-38D". The communication will fail at this point.
1547 * The EC_GROUP_clear_free() function is deprecated as there is nothing
1548 confidential in EC_GROUP data.
1552 * The byte order mark (BOM) character is ignored if encountered at the
1553 beginning of a PEM-formatted file.
1557 * Added CMS support for the Russian GOST algorithms.
1561 * Due to move of the implementation of cryptographic operations
1562 to the providers, validation of various operation parameters can
1563 be postponed until the actual operation is executed where previously
1564 it happened immediately when an operation parameter was set.
1566 For example when setting an unsupported curve with
1567 EVP_PKEY_CTX_set_ec_paramgen_curve_nid() this function call will not
1568 fail but later keygen operations with the EVP_PKEY_CTX will fail.
1570 *OpenSSL team members and many third party contributors*
1572 * The EVP_get_cipherbyname() function will return NULL for algorithms such as
1573 "AES-128-SIV", "AES-128-CBC-CTS" and "CAMELLIA-128-CBC-CTS" which were
1574 previously only accessible via low-level interfaces. Use EVP_CIPHER_fetch()
1575 instead to retrieve these algorithms from a provider.
1579 * On build targets where the multilib postfix is set in the build
1580 configuration the libdir directory was changing based on whether
1581 the lib directory with the multilib postfix exists on the system
1582 or not. This unpredictable behavior was removed and eventual
1583 multilib postfix is now always added to the default libdir. Use
1584 `--libdir=lib` to override the libdir if adding the postfix is
1589 * The triple DES key wrap functionality now conforms to RFC 3217 but is
1590 no longer interoperable with OpenSSL 1.1.1.
1594 * The ERR_GET_FUNC() function was removed. With the loss of meaningful
1595 function codes, this function can only cause problems for calling
1600 * Add a configurable flag to output date formats as ISO 8601. Does not
1601 change the default date format.
1605 * Version of MSVC earlier than 1300 could get link warnings, which could
1606 be suppressed if the undocumented -DI_CAN_LIVE_WITH_LNK4049 was set.
1607 Support for this flag has been removed.
1611 * Rework and make DEBUG macros consistent. Remove unused -DCONF_DEBUG,
1612 -DBN_CTX_DEBUG, and REF_PRINT. Add a new tracing category and use it for
1613 printing reference counts. Rename -DDEBUG_UNUSED to -DUNUSED_RESULT_DEBUG
1614 Fix BN_DEBUG_RAND so it compiles and, when set, force DEBUG_RAND to be set
1615 also. Rename engine_debug_ref to be ENGINE_REF_PRINT also for consistency.
1619 * The signatures of the functions to get and set options on SSL and
1620 SSL_CTX objects changed from "unsigned long" to "uint64_t" type.
1621 Some source code changes may be required.
1625 * The public definitions of conf_method_st and conf_st have been
1626 deprecated. They will be made opaque in a future release.
1628 *Rich Salz and Tomáš Mráz*
1630 * Client-initiated renegotiation is disabled by default. To allow it, use
1631 the -client_renegotiation option, the SSL_OP_ALLOW_CLIENT_RENEGOTIATION
1632 flag, or the "ClientRenegotiation" config parameter as appropriate.
1636 * Add "abspath" and "includedir" pragma's to config files, to prevent,
1637 or modify relative pathname inclusion.
1641 * OpenSSL includes a cryptographic module that is intended to be FIPS 140-2
1642 validated. Please consult the README-FIPS and
1643 README-PROVIDERS files, as well as the migration guide.
1645 *OpenSSL team members and many third party contributors*
1647 * For the key types DH and DHX the allowed settable parameters are now different.
1651 * The openssl commands that read keys, certificates, and CRLs now
1652 automatically detect the PEM or DER format of the input files.
1654 *David von Oheimb, Richard Levitte, and Tomáš Mráz*
1656 * Added enhanced PKCS#12 APIs which accept a library context.
1660 * The default manual page suffix ($MANSUFFIX) has been changed to "ossl"
1664 * Added support for Kernel TLS (KTLS).
1666 *Boris Pismenny, John Baldwin and Andrew Gallatin*
1668 * Support for RFC 5746 secure renegotiation is now required by default for
1669 SSL or TLS connections to succeed.
1673 * The signature of the `copy` functional parameter of the
1674 EVP_PKEY_meth_set_copy() function has changed so its `src` argument is
1675 now `const EVP_PKEY_CTX *` instead of `EVP_PKEY_CTX *`. Similarly
1676 the signature of the `pub_decode` functional parameter of the
1677 EVP_PKEY_asn1_set_public() function has changed so its `pub` argument is
1678 now `const X509_PUBKEY *` instead of `X509_PUBKEY *`.
1682 * The error return values from some control calls (ctrl) have changed.
1686 * A public key check is now performed during EVP_PKEY_derive_set_peer().
1690 * Many functions in the EVP_ namespace that are getters of values from
1691 implementations or contexts were renamed to include get or get0 in their
1692 names. Old names are provided as macro aliases for compatibility and
1697 * The EVP_PKEY_CTRL_PKCS7_ENCRYPT, EVP_PKEY_CTRL_PKCS7_DECRYPT,
1698 EVP_PKEY_CTRL_PKCS7_SIGN, EVP_PKEY_CTRL_CMS_ENCRYPT,
1699 EVP_PKEY_CTRL_CMS_DECRYPT, and EVP_PKEY_CTRL_CMS_SIGN control operations
1704 * The EVP_PKEY_public_check() and EVP_PKEY_param_check() functions now work for
1707 * The output from the command line applications may have minor
1712 * The output from numerous "printing" may have minor changes.
1716 * Windows thread synchronization uses read/write primitives (SRWLock) when
1717 supported by the OS, otherwise CriticalSection continues to be used.
1721 * Add filter BIO BIO_f_readbuffer() that allows BIO_tell() and BIO_seek() to
1722 work on read only BIO source/sinks that do not support these functions.
1723 This allows piping or redirection of a file BIO using stdin to be buffered
1724 into memory. This is used internally in OSSL_DECODER_from_bio().
1728 * OSSL_STORE_INFO_get_type() may now return an additional value. In 1.1.1
1729 this function would return one of the values OSSL_STORE_INFO_NAME,
1730 OSSL_STORE_INFO_PKEY, OSSL_STORE_INFO_PARAMS, OSSL_STORE_INFO_CERT or
1731 OSSL_STORE_INFO_CRL. Decoded public keys would previously have been reported
1732 as type OSSL_STORE_INFO_PKEY in 1.1.1. In 3.0 decoded public keys are now
1733 reported as having the new type OSSL_STORE_INFO_PUBKEY. Applications
1734 using this function should be amended to handle the changed return value.
1738 * Improved adherence to Enhanced Security Services (ESS, RFC 2634 and RFC 5035)
1739 for the TSP and CMS Advanced Electronic Signatures (CAdES) implementations.
1740 As required by RFC 5035 check both ESSCertID and ESSCertIDv2 if both present.
1741 Correct the semantics of checking the validation chain in case ESSCertID{,v2}
1742 contains more than one certificate identifier: This means that all
1743 certificates referenced there MUST be part of the validation chain.
1747 * The implementation of older EVP ciphers related to CAST, IDEA, SEED, RC2, RC4,
1748 RC5, DESX and DES have been moved to the legacy provider.
1752 * The implementation of the EVP digests MD2, MD4, MDC2, WHIRLPOOL and
1753 RIPEMD-160 have been moved to the legacy provider.
1757 * The deprecated function EVP_PKEY_get0() now returns NULL being called for a
1762 * The deprecated functions EVP_PKEY_get0_RSA(),
1763 EVP_PKEY_get0_DSA(), EVP_PKEY_get0_EC_KEY(), EVP_PKEY_get0_DH(),
1764 EVP_PKEY_get0_hmac(), EVP_PKEY_get0_poly1305() and EVP_PKEY_get0_siphash() as
1765 well as the similarly named "get1" functions behave differently in
1770 * A number of functions handling low-level keys or engines were deprecated
1771 including EVP_PKEY_set1_engine(), EVP_PKEY_get0_engine(), EVP_PKEY_assign(),
1772 EVP_PKEY_get0(), EVP_PKEY_get0_hmac(), EVP_PKEY_get0_poly1305() and
1773 EVP_PKEY_get0_siphash().
1777 * PKCS#5 PBKDF1 key derivation has been moved from PKCS5_PBE_keyivgen() into
1778 the legacy crypto provider as an EVP_KDF. Applications requiring this KDF
1779 will need to load the legacy crypto provider. This includes these PBE
1780 algorithms which use this KDF:
1781 - NID_pbeWithMD2AndDES_CBC
1782 - NID_pbeWithMD5AndDES_CBC
1783 - NID_pbeWithSHA1AndRC2_CBC
1784 - NID_pbeWithMD2AndRC2_CBC
1785 - NID_pbeWithMD5AndRC2_CBC
1786 - NID_pbeWithSHA1AndDES_CBC
1790 * Deprecated obsolete BIO_set_callback(), BIO_get_callback(), and
1791 BIO_debug_callback() functions.
1795 * Deprecated obsolete EVP_PKEY_CTX_get0_dh_kdf_ukm() and
1796 EVP_PKEY_CTX_get0_ecdh_kdf_ukm() functions.
1800 * The RAND_METHOD APIs have been deprecated.
1804 * The SRP APIs have been deprecated.
1808 * Add a compile time option to prevent the caching of provider fetched
1809 algorithms. This is enabled by including the no-cached-fetch option
1810 at configuration time.
1814 * pkcs12 now uses defaults of PBKDF2, AES and SHA-256, with a MAC iteration
1815 count of PKCS12_DEFAULT_ITER.
1817 *Tomáš Mráz and Sahana Prasad*
1819 * The openssl speed command does not use low-level API calls anymore.
1823 * Parallel dual-prime 1024-bit modular exponentiation for AVX512_IFMA
1826 *Ilya Albrekht, Sergey Kirillov, Andrey Matyukov (Intel Corp)*
1828 * Combining the Configure options no-ec and no-dh no longer disables TLSv1.3.
1832 * Implemented support for fully "pluggable" TLSv1.3 groups. This means that
1833 providers may supply their own group implementations (using either the "key
1834 exchange" or the "key encapsulation" methods) which will automatically be
1835 detected and used by libssl.
1837 *Matt Caswell, Nicola Tuveri*
1839 * The undocumented function X509_certificate_type() has been deprecated;
1843 * Deprecated the obsolete BN_pseudo_rand() and BN_pseudo_rand_range().
1847 * Removed RSA padding mode for SSLv23 (which was only used for
1848 SSLv2). This includes the functions RSA_padding_check_SSLv23() and
1849 RSA_padding_add_SSLv23() and the `-ssl` option in the deprecated
1854 * Deprecated the obsolete X9.31 RSA key generation related functions.
1856 * While a callback function set via `SSL_CTX_set_cert_verify_callback()`
1857 is not allowed to return a value > 1, this is no more taken as failure.
1859 *Viktor Dukhovni and David von Oheimb*
1861 * Deprecated the obsolete X9.31 RSA key generation related functions
1862 BN_X931_generate_Xpq(), BN_X931_derive_prime_ex(), and
1863 BN_X931_generate_prime_ex().
1867 * The default key generation method for the regular 2-prime RSA keys was
1868 changed to the FIPS 186-4 B.3.6 method.
1872 * Deprecated the BN_is_prime_ex() and BN_is_prime_fasttest_ex() functions.
1876 * Deprecated EVP_MD_CTX_set_update_fn() and EVP_MD_CTX_update_fn().
1880 * Deprecated the type OCSP_REQ_CTX and the functions OCSP_REQ_CTX_*() and
1881 replaced with OSSL_HTTP_REQ_CTX and the functions OSSL_HTTP_REQ_CTX_*().
1883 *Rich Salz, Richard Levitte, and David von Oheimb*
1885 * Deprecated `X509_http_nbio()` and `X509_CRL_http_nbio()`.
1889 * Deprecated `OCSP_parse_url()`.
1893 * Validation of SM2 keys has been separated from the validation of regular EC
1898 * Behavior of the `pkey` app is changed, when using the `-check` or `-pubcheck`
1899 switches: a validation failure triggers an early exit, returning a failure
1900 exit status to the parent process.
1904 * Changed behavior of SSL_CTX_set_ciphersuites() and SSL_set_ciphersuites()
1905 to ignore unknown ciphers.
1909 * The `-cipher-commands` and `-digest-commands` options
1910 of the command line utility `list` have been deprecated.
1911 Instead use the `-cipher-algorithms` and `-digest-algorithms` options.
1915 * Added convenience functions for generating asymmetric key pairs:
1916 The 'quick' one-shot (yet somewhat limited) function L<EVP_PKEY_Q_keygen(3)>
1917 and macros for the most common cases: <EVP_RSA_gen(3)> and L<EVP_EC_gen(3)>.
1921 * All of the low-level EC_KEY functions have been deprecated.
1923 *Shane Lontis, Paul Dale, Richard Levitte, and Tomáš Mráz*
1925 * Deprecated all the libcrypto and libssl error string loading
1930 * The functions SSL_CTX_set_tmp_dh_callback and SSL_set_tmp_dh_callback, as
1931 well as the macros SSL_CTX_set_tmp_dh() and SSL_set_tmp_dh() have been
1936 * The `-crypt` option to the `passwd` command line tool has been removed.
1940 * The -C option to the `x509`, `dhparam`, `dsaparam`, and `ecparam` commands
1945 * Add support for AES Key Wrap inverse ciphers to the EVP layer.
1949 * Deprecated EVP_PKEY_set1_tls_encodedpoint() and
1950 EVP_PKEY_get1_tls_encodedpoint().
1954 * The security callback, which can be customised by application code, supports
1955 the security operation SSL_SECOP_TMP_DH. One location of the "other" parameter
1956 was incorrectly passing a DH object. It now passed an EVP_PKEY in all cases.
1960 * Add PKCS7_get_octet_string() and PKCS7_type_is_other() to the public
1961 interface. Their functionality remains unchanged.
1965 * Added new option for 'openssl list', '-providers', which will display the
1966 list of loaded providers, their names, version and status. It optionally
1967 displays their gettable parameters.
1971 * Removed EVP_PKEY_set_alias_type().
1975 * Deprecated `EVP_PKEY_CTX_set_rsa_keygen_pubexp()` and introduced
1976 `EVP_PKEY_CTX_set1_rsa_keygen_pubexp()`, which is now preferred.
1980 * Changed all "STACK" functions to be macros instead of inline functions. Macro
1981 parameters are still checked for type safety at compile time via helper
1986 * Remove the RAND_DRBG API
1988 *Paul Dale and Matthias St. Pierre*
1990 * Allow `SSL_set1_host()` and `SSL_add1_host()` to take IP literal addresses
1991 as well as actual hostnames.
1995 * The 'MinProtocol' and 'MaxProtocol' configuration commands now silently
1996 ignore TLS protocol version bounds when configuring DTLS-based contexts, and
1997 conversely, silently ignore DTLS protocol version bounds when configuring
1998 TLS-based contexts. The commands can be repeated to set bounds of both
1999 types. The same applies with the corresponding "min_protocol" and
2000 "max_protocol" command-line switches, in case some application uses both TLS
2003 SSL_CTX instances that are created for a fixed protocol version (e.g.
2004 `TLSv1_server_method()`) also silently ignore version bounds. Previously
2005 attempts to apply bounds to these protocol versions would result in an
2006 error. Now only the "version-flexible" SSL_CTX instances are subject to
2007 limits in configuration files in command-line options.
2011 * Deprecated the `ENGINE` API. Engines should be replaced with providers
2016 * Reworked the recorded ERR codes to make better space for system errors.
2017 To distinguish them, the macro `ERR_SYSTEM_ERROR()` indicates if the
2018 given code is a system error (true) or an OpenSSL error (false).
2022 * Reworked the test perl framework to better allow parallel testing.
2024 *Nicola Tuveri and David von Oheimb*
2026 * Added ciphertext stealing algorithms AES-128-CBC-CTS, AES-192-CBC-CTS and
2027 AES-256-CBC-CTS to the providers. CS1, CS2 and CS3 variants are supported.
2031 * 'Configure' has been changed to figure out the configuration target if
2032 none is given on the command line. Consequently, the 'config' script is
2033 now only a mere wrapper. All documentation is changed to only mention
2036 *Rich Salz and Richard Levitte*
2038 * Added a library context `OSSL_LIB_CTX` that applications as well as
2039 other libraries can use to form a separate context within which
2040 libcrypto operations are performed.
2044 * Added various `_ex` functions to the OpenSSL API that support using
2045 a non-default `OSSL_LIB_CTX`.
2049 * Handshake now fails if Extended Master Secret extension is dropped
2054 * Dropped interactive mode from the `openssl` program.
2058 * Deprecated `EVP_PKEY_cmp()` and `EVP_PKEY_cmp_parameters()`.
2060 *David von Oheimb and Shane Lontis*
2062 * Deprecated `EC_METHOD_get_field_type()`.
2066 * Deprecated EC_GFp_simple_method(), EC_GFp_mont_method(),
2067 EC_GF2m_simple_method(), EC_GFp_nist_method(), EC_GFp_nistp224_method()
2068 EC_GFp_nistp256_method(), and EC_GFp_nistp521_method().
2072 * Deprecated EC_GROUP_new(), EC_GROUP_method_of(), and EC_POINT_method_of().
2076 * Add CAdES-BES signature verification support, mostly derived
2077 from ESSCertIDv2 TS (RFC 5816) contribution by Marek Klein.
2079 *Filipe Raimundo da Silva*
2081 * Add CAdES-BES signature scheme and attributes support (RFC 5126) to CMS API.
2085 * Added the AuthEnvelopedData content type structure (RFC 5083) with AES-GCM
2086 parameter (RFC 5084) for the Cryptographic Message Syntax (CMS).
2090 * Deprecated EC_POINT_make_affine() and EC_POINTs_make_affine().
2094 * Deprecated EC_GROUP_precompute_mult(), EC_GROUP_have_precompute_mult(), and
2095 EC_KEY_precompute_mult().
2099 * Deprecated EC_POINTs_mul().
2103 * Removed FIPS_mode() and FIPS_mode_set().
2107 * The SSL option SSL_OP_IGNORE_UNEXPECTED_EOF is introduced.
2111 * Deprecated EC_POINT_set_Jprojective_coordinates_GFp() and
2112 EC_POINT_get_Jprojective_coordinates_GFp().
2116 * Added OSSL_PARAM_BLD to the public interface. This allows OSSL_PARAM
2117 arrays to be more easily constructed via a series of utility functions.
2118 Create a parameter builder using OSSL_PARAM_BLD_new(), add parameters using
2119 the various push functions and finally convert to a passable OSSL_PARAM
2120 array using OSSL_PARAM_BLD_to_param().
2124 * The security strength of SHA1 and MD5 based signatures in TLS has been
2129 * Added EVP_PKEY_set_type_by_keymgmt(), to initialise an EVP_PKEY to
2130 contain a provider side internal key.
2134 * ASN1_verify(), ASN1_digest() and ASN1_sign() have been deprecated.
2138 * Project text documents not yet having a proper file name extension
2139 (`HACKING`, `LICENSE`, `NOTES*`, `README*`, `VERSION`) have been renamed to
2140 `*.md` as far as reasonable, else `*.txt`, for better use with file managers.
2144 * The main project documents (README, NEWS, CHANGES, INSTALL, SUPPORT)
2145 have been converted to Markdown with the goal to produce documents
2146 which not only look pretty when viewed online in the browser, but
2147 remain well readable inside a plain text editor.
2149 To achieve this goal, a 'minimalistic' Markdown style has been applied
2150 which avoids formatting elements that interfere too much with the
2151 reading flow in the text file. For example, it
2153 * avoids [ATX headings][] and uses [setext headings][] instead
2154 (which works for `<h1>` and `<h2>` headings only).
2155 * avoids [inline links][] and uses [reference links][] instead.
2156 * avoids [fenced code blocks][] and uses [indented code blocks][] instead.
2158 [ATX headings]: https://github.github.com/gfm/#atx-headings
2159 [setext headings]: https://github.github.com/gfm/#setext-headings
2160 [inline links]: https://github.github.com/gfm/#inline-link
2161 [reference links]: https://github.github.com/gfm/#reference-link
2162 [fenced code blocks]: https://github.github.com/gfm/#fenced-code-blocks
2163 [indented code blocks]: https://github.github.com/gfm/#indented-code-blocks
2165 *Matthias St. Pierre*
2167 * The test suite is changed to preserve results of each test recipe.
2168 A new directory test-runs/ with subdirectories named like the
2169 test recipes are created in the build tree for this purpose.
2173 * Added an implementation of CMP and CRMF (RFC 4210, RFC 4211 RFC 6712).
2174 This adds `crypto/cmp/`, `crpyto/crmf/`, `apps/cmp.c`, and `test/cmp_*`.
2175 See L<openssl-cmp(1)> and L<OSSL_CMP_exec_IR_ses(3)> as starting points.
2177 *David von Oheimb, Martin Peylo*
2179 * Generalized the HTTP client code from `crypto/ocsp/` into `crpyto/http/`.
2180 It supports arbitrary request and response content types, GET redirection,
2181 TLS, connections via HTTP(S) proxies, connections and exchange via
2182 user-defined BIOs (allowing implicit connections), persistent connections,
2183 and timeout checks. See L<OSSL_HTTP_transfer(3)> etc. for details.
2184 The legacy OCSP-focused (and only partly documented) API
2185 is retained for backward compatibility, while most of it is deprecated.
2189 * Added `util/check-format.pl`, a tool for checking adherence to the
2190 OpenSSL coding style <https://www.openssl.org/policies/codingstyle.html>.
2191 The checks performed are incomplete and yield some false positives.
2192 Still the tool should be useful for detecting most typical glitches.
2196 * `BIO_do_connect()` and `BIO_do_handshake()` have been extended:
2197 If domain name resolution yields multiple IP addresses all of them are tried
2198 after `connect()` failures.
2202 * All of the low-level RSA functions have been deprecated.
2206 * X509 certificates signed using SHA1 are no longer allowed at security
2211 * The command line utilities dhparam, dsa, gendsa and dsaparam have been
2212 modified to use PKEY APIs. These commands are now in maintenance mode
2213 and no new features will be added to them.
2217 * The command line utility rsautl has been deprecated.
2221 * The command line utilities genrsa and rsa have been modified to use PKEY
2222 APIs. They now write PKCS#8 keys by default. These commands are now in
2223 maintenance mode and no new features will be added to them.
2227 * All of the low-level DH functions have been deprecated.
2229 *Paul Dale and Matt Caswell*
2231 * All of the low-level DSA functions have been deprecated.
2235 * Reworked the treatment of EC EVP_PKEYs with the SM2 curve to
2236 automatically become EVP_PKEY_SM2 rather than EVP_PKEY_EC.
2240 * Deprecated low-level ECDH and ECDSA functions.
2244 * Deprecated EVP_PKEY_decrypt_old() and EVP_PKEY_encrypt_old().
2248 * Enhanced the documentation of EVP_PKEY_get_size(), EVP_PKEY_get_bits()
2249 and EVP_PKEY_get_security_bits(). Especially EVP_PKEY_get_size() needed
2250 a new formulation to include all the things it can be used for,
2251 as well as words of caution.
2255 * The SSL_CTX_set_tlsext_ticket_key_cb(3) function has been deprecated.
2259 * All of the low-level HMAC functions have been deprecated.
2261 *Paul Dale and David von Oheimb*
2263 * Over two thousand fixes were made to the documentation, including:
2264 - Common options (such as -rand/-writerand, TLS version control, etc)
2265 were refactored and point to newly-enhanced descriptions in openssl.pod.
2266 - Added style conformance for all options (with help from Richard Levitte),
2267 documented all reported missing options, added a CI build to check
2268 that all options are documented and that no unimplemented options
2270 - Documented some internals, such as all use of environment variables.
2271 - Addressed all internal broken L<> references.
2275 * All of the low-level CMAC functions have been deprecated.
2279 * The low-level MD2, MD4, MD5, MDC2, RIPEMD160 and Whirlpool digest
2280 functions have been deprecated.
2282 *Paul Dale and David von Oheimb*
2284 * Corrected the documentation of the return values from the `EVP_DigestSign*`
2285 set of functions. The documentation mentioned negative values for some
2286 errors, but this was never the case, so the mention of negative values
2289 Code that followed the documentation and thereby check with something
2290 like `EVP_DigestSignInit(...) <= 0` will continue to work undisturbed.
2294 * All of the low-level cipher functions have been deprecated.
2296 *Matt Caswell and Paul Dale*
2298 * Removed include/openssl/opensslconf.h.in and replaced it with
2299 include/openssl/configuration.h.in, which differs in not including
2300 <openssl/macros.h>. A short header include/openssl/opensslconf.h
2301 was added to include both.
2303 This allows internal hacks where one might need to modify the set
2304 of configured macros, for example this if deprecated symbols are
2305 still supposed to be available internally:
2307 #include <openssl/configuration.h>
2309 #undef OPENSSL_NO_DEPRECATED
2310 #define OPENSSL_SUPPRESS_DEPRECATED
2312 #include <openssl/macros.h>
2314 This should not be used by applications that use the exported
2315 symbols, as that will lead to linking errors.
2319 * Fixed an overflow bug in the x64_64 Montgomery squaring procedure
2320 used in exponentiation with 512-bit moduli. No EC algorithms are
2321 affected. Analysis suggests that attacks against 2-prime RSA1024,
2322 3-prime RSA1536, and DSA1024 as a result of this defect would be very
2323 difficult to perform and are not believed likely. Attacks against DH512
2324 are considered just feasible. However, for an attack the target would
2325 have to reuse the DH512 private key, which is not recommended anyway.
2326 Also applications directly using the low-level API BN_mod_exp may be
2327 affected if they use BN_FLG_CONSTTIME.
2332 * Most memory-debug features have been deprecated, and the functionality
2333 replaced with no-ops.
2337 * Added documentation for the STACK API.
2341 * Introduced a new method type and API, OSSL_ENCODER, to represent
2342 generic encoders. These do the same sort of job that PEM writers
2343 and d2i functions do, but with support for methods supplied by
2344 providers, and the possibility for providers to support other
2349 * Introduced a new method type and API, OSSL_DECODER, to represent
2350 generic decoders. These do the same sort of job that PEM readers
2351 and i2d functions do, but with support for methods supplied by
2352 providers, and the possibility for providers to support other
2357 * Added a .pragma directive to the syntax of configuration files, to
2358 allow varying behavior in a supported and predictable manner.
2359 Currently added pragma:
2363 This allows dollar signs to be a keyword character unless it's
2364 followed by a opening brace or parenthesis. This is useful for
2365 platforms where dollar signs are commonly used in names, such as
2366 volume names and system directory names on VMS.
2370 * Added functionality to create an EVP_PKEY from user data.
2374 * Change the interpretation of the '--api' configuration option to
2375 mean that this is a desired API compatibility level with no
2376 further meaning. The previous interpretation, that this would
2377 also mean to remove all deprecated symbols up to and including
2378 the given version, no requires that 'no-deprecated' is also used
2379 in the configuration.
2381 When building applications, the desired API compatibility level
2382 can be set with the OPENSSL_API_COMPAT macro like before. For
2383 API compatibility version below 3.0, the old style numerical
2384 value is valid as before, such as -DOPENSSL_API_COMPAT=0x10100000L.
2385 For version 3.0 and on, the value is expected to be the decimal
2386 value calculated from the major and minor version like this:
2388 MAJOR * 10000 + MINOR * 100
2392 -DOPENSSL_API_COMPAT=30000 For 3.0
2393 -DOPENSSL_API_COMPAT=30200 For 3.2
2395 To hide declarations that are deprecated up to and including the
2396 given API compatibility level, -DOPENSSL_NO_DEPRECATED must be
2397 given when building the application as well.
2401 * Added the X509_LOOKUP_METHOD called X509_LOOKUP_store, to allow
2402 access to certificate and CRL stores via URIs and OSSL_STORE
2405 This adds the following functions:
2407 - X509_LOOKUP_store()
2408 - X509_STORE_load_file()
2409 - X509_STORE_load_path()
2410 - X509_STORE_load_store()
2411 - SSL_add_store_cert_subjects_to_stack()
2412 - SSL_CTX_set_default_verify_store()
2413 - SSL_CTX_load_verify_file()
2414 - SSL_CTX_load_verify_dir()
2415 - SSL_CTX_load_verify_store()
2419 * Added a new method to gather entropy on VMS, based on SYS$GET_ENTROPY.
2420 The presence of this system service is determined at run-time.
2424 * Added functionality to create an EVP_PKEY context based on data
2425 for methods from providers. This takes an algorithm name and a
2426 property query string and simply stores them, with the intent
2427 that any operation that uses this context will use those strings
2428 to fetch the needed methods implicitly, thereby making the port
2429 of application written for pre-3.0 OpenSSL easier.
2433 * The undocumented function NCONF_WIN32() has been deprecated; for
2434 conversion details see the HISTORY section of doc/man5/config.pod
2438 * Introduced the new functions EVP_DigestSignInit_ex() and
2439 EVP_DigestVerifyInit_ex(). The macros EVP_DigestSignUpdate() and
2440 EVP_DigestVerifyUpdate() have been converted to functions. See the man
2441 pages for further details.
2445 * Over two thousand fixes were made to the documentation, including:
2446 adding missing command flags, better style conformance, documentation
2449 *Rich Salz, Richard Levitte*
2451 * s390x assembly pack: add hardware-support for P-256, P-384, P-521,
2452 X25519, X448, Ed25519 and Ed448.
2456 * Print all values for a PKCS#12 attribute with 'openssl pkcs12', not just
2461 * Deprecated the public definition of `ERR_STATE` as well as the function
2462 `ERR_get_state()`. This is done in preparation of making `ERR_STATE` an
2467 * Added ERR functionality to give callers access to the stored function
2468 names that have replaced the older function code based functions.
2470 New functions are ERR_peek_error_func(), ERR_peek_last_error_func(),
2471 ERR_peek_error_data(), ERR_peek_last_error_data(), ERR_get_error_all(),
2472 ERR_peek_error_all() and ERR_peek_last_error_all().
2474 Deprecate ERR functions ERR_get_error_line(), ERR_get_error_line_data(),
2475 ERR_peek_error_line_data(), ERR_peek_last_error_line_data() and
2476 ERR_func_error_string().
2480 * Extended testing to be verbose for failing tests only. The make variables
2481 VERBOSE_FAILURE or VF can be used to enable this:
2483 $ make VF=1 test # Unix
2484 $ mms /macro=(VF=1) test ! OpenVMS
2485 $ nmake VF=1 test # Windows
2489 * Added the `-copy_extensions` option to the `x509` command for use with
2490 `-req` and `-x509toreq`. When given with the `copy` or `copyall` argument,
2491 all extensions in the request are copied to the certificate or vice versa.
2493 *David von Oheimb*, *Kirill Stefanenkov <kirill_stefanenkov@rambler.ru>*
2495 * Added the `-copy_extensions` option to the `req` command for use with
2496 `-x509`. When given with the `copy` or `copyall` argument,
2497 all extensions in the certification request are copied to the certificate.
2501 * The `x509`, `req`, and `ca` commands now make sure that X.509v3 certificates
2502 they generate are by default RFC 5280 compliant in the following sense:
2503 There is a subjectKeyIdentifier extension with a hash value of the public key
2504 and for not self-signed certs there is an authorityKeyIdentifier extension
2505 with a keyIdentifier field or issuer information identifying the signing key.
2506 This is done unless some configuration overrides the new default behavior,
2507 such as `subjectKeyIdentifier = none` and `authorityKeyIdentifier = none`.
2511 * Added several checks to `X509_verify_cert()` according to requirements in
2512 RFC 5280 in case `X509_V_FLAG_X509_STRICT` is set
2513 (which may be done by using the CLI option `-x509_strict`):
2514 * The basicConstraints of CA certificates must be marked critical.
2515 * CA certificates must explicitly include the keyUsage extension.
2516 * If a pathlenConstraint is given the key usage keyCertSign must be allowed.
2517 * The issuer name of any certificate must not be empty.
2518 * The subject name of CA certs, certs with keyUsage crlSign,
2519 and certs without subjectAlternativeName must not be empty.
2520 * If a subjectAlternativeName extension is given it must not be empty.
2521 * The signatureAlgorithm field and the cert signature must be consistent.
2522 * Any given authorityKeyIdentifier and any given subjectKeyIdentifier
2523 must not be marked critical.
2524 * The authorityKeyIdentifier must be given for X.509v3 certs
2525 unless they are self-signed.
2526 * The subjectKeyIdentifier must be given for all X.509v3 CA certs.
2530 * Certificate verification using `X509_verify_cert()` meanwhile rejects EC keys
2531 with explicit curve parameters (specifiedCurve) as required by RFC 5480.
2535 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
2536 used even when parsing explicit parameters, when loading a encoded key
2537 or calling `EC_GROUP_new_from_ecpkparameters()`/
2538 `EC_GROUP_new_from_ecparameters()`.
2539 This prevents bypass of security hardening and performance gains,
2540 especially for curves with specialized EC_METHODs.
2541 By default, if a key encoded with explicit parameters is loaded and later
2542 encoded, the output is still encoded with explicit parameters, even if
2543 internally a "named" EC_GROUP is used for computation.
2547 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
2548 this change, EC_GROUP_set_generator would accept order and/or cofactor as
2549 NULL. After this change, only the cofactor parameter can be NULL. It also
2550 does some minimal sanity checks on the passed order.
2555 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
2556 An attack is simple, if the first CMS_recipientInfo is valid but the
2557 second CMS_recipientInfo is chosen ciphertext. If the second
2558 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
2559 encryption key will be replaced by garbage, and the message cannot be
2560 decoded, but if the RSA decryption fails, the correct encryption key is
2561 used and the recipient will not notice the attack.
2562 As a work around for this potential attack the length of the decrypted
2563 key must be equal to the cipher default key length, in case the
2564 certificate is not given and all recipientInfo are tried out.
2565 The old behaviour can be re-enabled in the CMS code by setting the
2566 CMS_DEBUG_DECRYPT flag.
2570 * Early start up entropy quality from the DEVRANDOM seed source has been
2571 improved for older Linux systems. The RAND subsystem will wait for
2572 /dev/random to be producing output before seeding from /dev/urandom.
2573 The seeded state is stored for future library initialisations using
2574 a system global shared memory segment. The shared memory identifier
2575 can be configured by defining OPENSSL_RAND_SEED_DEVRANDOM_SHM_ID to
2576 the desired value. The default identifier is 114.
2580 * Revised BN_generate_prime_ex to not avoid factors 2..17863 in p-1
2581 when primes for RSA keys are computed.
2582 Since we previously always generated primes == 2 (mod 3) for RSA keys,
2583 the 2-prime and 3-prime RSA modules were easy to distinguish, since
2584 `N = p*q = 1 (mod 3)`, but `N = p*q*r = 2 (mod 3)`. Therefore, fingerprinting
2585 2-prime vs. 3-prime RSA keys was possible by computing N mod 3.
2586 This avoids possible fingerprinting of newly generated RSA modules.
2590 * Correct the extended master secret constant on EBCDIC systems. Without this
2591 fix TLS connections between an EBCDIC system and a non-EBCDIC system that
2592 negotiate EMS will fail. Unfortunately this also means that TLS connections
2593 between EBCDIC systems with this fix, and EBCDIC systems without this
2594 fix will fail if they negotiate EMS.
2598 * Changed the library initialisation so that the config file is now loaded
2599 by default. This was already the case for libssl. It now occurs for both
2600 libcrypto and libssl. Use the OPENSSL_INIT_NO_LOAD_CONFIG option to
2601 `OPENSSL_init_crypto()` to suppress automatic loading of a config file.
2605 * Introduced new error raising macros, `ERR_raise()` and `ERR_raise_data()`,
2606 where the former acts as a replacement for `ERR_put_error()`, and the
2607 latter replaces the combination `ERR_put_error()` + `ERR_add_error_data()`.
2608 `ERR_raise_data()` adds more flexibility by taking a format string and
2609 an arbitrary number of arguments following it, to be processed with
2614 * Introduced a new function, `OSSL_PROVIDER_available()`, which can be used
2615 to check if a named provider is loaded and available. When called, it
2616 will also activate all fallback providers if such are still present.
2620 * Enforce a minimum DH modulus size of 512 bits.
2624 * Changed DH parameters to generate the order q subgroup instead of 2q.
2625 Previously generated DH parameters are still accepted by DH_check
2626 but DH_generate_key works around that by clearing bit 0 of the
2627 private key for those. This avoids leaking bit 0 of the private key.
2631 * Significantly reduce secure memory usage by the randomness pools.
2635 * `{CRYPTO,OPENSSL}_mem_debug_{push,pop}` are now no-ops and have been
2640 * A new type, EVP_KEYEXCH, has been introduced to represent key exchange
2641 algorithms. An implementation of a key exchange algorithm can be obtained
2642 by using the function EVP_KEYEXCH_fetch(). An EVP_KEYEXCH algorithm can be
2643 used in a call to EVP_PKEY_derive_init_ex() which works in a similar way to
2644 the older EVP_PKEY_derive_init() function. See the man pages for the new
2645 functions for further details.
2649 * The EVP_PKEY_CTX_set_dh_pad() macro has now been converted to a function.
2653 * Removed the function names from error messages and deprecated the
2658 * Removed NextStep support and the macro OPENSSL_UNISTD
2662 * Removed DES_check_key. Also removed OPENSSL_IMPLEMENT_GLOBAL,
2663 OPENSSL_GLOBAL_REF, OPENSSL_DECLARE_GLOBAL.
2664 Also removed "export var as function" capability; we do not export
2665 variables, only functions.
2669 * RC5_32_set_key has been changed to return an int type, with 0 indicating
2670 an error and 1 indicating success. In previous versions of OpenSSL this
2671 was a void type. If a key was set longer than the maximum possible this
2676 * Support SM2 signing and verification schemes with X509 certificate.
2680 * Use SHA256 as the default digest for TS query in the `ts` app.
2684 * Change PBKDF2 to conform to SP800-132 instead of the older PKCS5 RFC2898.
2688 * Default cipher lists/suites are now available via a function, the
2689 #defines are deprecated.
2693 * Add target VC-WIN32-UWP, VC-WIN64A-UWP, VC-WIN32-ARM-UWP and
2694 VC-WIN64-ARM-UWP in Windows OneCore target for making building libraries
2695 for Windows Store apps easier. Also, the "no-uplink" option has been added.
2699 * Join the directories crypto/x509 and crypto/x509v3
2703 * Added command 'openssl kdf' that uses the EVP_KDF API.
2707 * Added command 'openssl mac' that uses the EVP_MAC API.
2711 * Added OPENSSL_info() to get diverse built-in OpenSSL data, such
2712 as default directories. Also added the command 'openssl info'
2713 for scripting purposes.
2717 * The functions AES_ige_encrypt() and AES_bi_ige_encrypt() have been
2722 * Add prediction resistance to the DRBG reseeding process.
2726 * Limit the number of blocks in a data unit for AES-XTS to 2^20 as
2727 mandated by IEEE Std 1619-2018.
2731 * Added newline escaping functionality to a filename when using openssl dgst.
2732 This output format is to replicate the output format found in the `*sum`
2733 checksum programs. This aims to preserve backward compatibility.
2735 *Matt Eaton, Richard Levitte, and Paul Dale*
2737 * Removed the heartbeat message in DTLS feature, as it has very
2738 little usage and doesn't seem to fulfill a valuable purpose.
2739 The configuration option is now deprecated.
2743 * Changed the output of 'openssl {digestname} < file' to display the
2744 digest name in its output.
2748 * Added a new generic trace API which provides support for enabling
2749 instrumentation through trace output.
2751 *Richard Levitte & Matthias St. Pierre*
2753 * Added build tests for C++. These are generated files that only do one
2754 thing, to include one public OpenSSL head file each. This tests that
2755 the public header files can be usefully included in a C++ application.
2757 This test isn't enabled by default. It can be enabled with the option
2758 'enable-buildtest-c++'.
2762 * Added KB KDF (EVP_KDF_KB) to EVP_KDF.
2766 * Added SSH KDF (EVP_KDF_SSHKDF) and KRB5 KDF (EVP_KDF_KRB5KDF) to EVP_KDF.
2770 * Added Single Step KDF (EVP_KDF_SS), X963 KDF, and X942 KDF to EVP_KDF.
2774 * Added KMAC to EVP_MAC.
2778 * Added property based algorithm implementation selection framework to
2783 * Added SCA hardening for modular field inversion in EC_GROUP through
2784 a new dedicated field_inv() pointer in EC_METHOD.
2785 This also addresses a leakage affecting conversions from projective
2786 to affine coordinates.
2788 *Billy Bob Brumley, Nicola Tuveri*
2790 * Added EVP_KDF, an EVP layer KDF API, to simplify adding KDF and PRF
2791 implementations. This includes an EVP_PKEY to EVP_KDF bridge for
2792 those algorithms that were already supported through the EVP_PKEY API
2793 (scrypt, TLS1 PRF and HKDF). The low-level KDF functions for PBKDF2
2794 and scrypt are now wrappers that call EVP_KDF.
2798 * Build devcrypto engine as a dynamic engine.
2800 *Eneas U de Queiroz*
2802 * Add keyed BLAKE2 to EVP_MAC.
2806 * Fix a bug in the computation of the endpoint-pair shared secret used
2807 by DTLS over SCTP. This breaks interoperability with older versions
2808 of OpenSSL like OpenSSL 1.1.0 and OpenSSL 1.0.2. There is a runtime
2809 switch SSL_MODE_DTLS_SCTP_LABEL_LENGTH_BUG (off by default) enabling
2810 interoperability with such broken implementations. However, enabling
2811 this switch breaks interoperability with correct implementations.
2813 * Fix a use after free bug in d2i_X509_PUBKEY when overwriting a
2814 re-used X509_PUBKEY object if the second PUBKEY is malformed.
2818 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
2822 * Changed the license to the Apache License v2.0.
2826 * Switch to a new version scheme using three numbers MAJOR.MINOR.PATCH.
2828 - Major releases (indicated by incrementing the MAJOR release number)
2829 may introduce incompatible API/ABI changes.
2830 - Minor releases (indicated by incrementing the MINOR release number)
2831 may introduce new features but retain API/ABI compatibility.
2832 - Patch releases (indicated by incrementing the PATCH number)
2833 are intended for bug fixes and other improvements of existing
2834 features only (like improving performance or adding documentation)
2835 and retain API/ABI compatibility.
2839 * Add support for RFC5297 SIV mode (siv128), including AES-SIV.
2843 * Remove the 'dist' target and add a tarball building script. The
2844 'dist' target has fallen out of use, and it shouldn't be
2845 necessary to configure just to create a source distribution.
2849 * Recreate the OS390-Unix config target. It no longer relies on a
2850 special script like it did for OpenSSL pre-1.1.0.
2854 * Instead of having the source directories listed in Configure, add
2855 a 'build.info' keyword SUBDIRS to indicate what sub-directories to
2860 * Add GMAC to EVP_MAC.
2864 * Ported the HMAC, CMAC and SipHash EVP_PKEY_METHODs to EVP_MAC.
2868 * Added EVP_MAC, an EVP layer MAC API, to simplify adding MAC
2869 implementations. This includes a generic EVP_PKEY to EVP_MAC bridge,
2870 to facilitate the continued use of MACs through raw private keys in
2871 functionality such as `EVP_DigestSign*` and `EVP_DigestVerify*`.
2875 * Deprecate ECDH_KDF_X9_62().
2879 * Added EVP_PKEY_ECDH_KDF_X9_63 and ecdh_KDF_X9_63() as replacements for
2880 the EVP_PKEY_ECDH_KDF_X9_62 KDF type and ECDH_KDF_X9_62(). The old names
2881 are retained for backwards compatibility.
2885 * AES-XTS mode now enforces that its two keys are different to mitigate
2886 the attacked described in "Efficient Instantiations of Tweakable
2887 Blockciphers and Refinements to Modes OCB and PMAC" by Phillip Rogaway.
2888 Details of this attack can be obtained from:
2889 <http://web.cs.ucdavis.edu/%7Erogaway/papers/offsets.pdf>
2893 * Rename the object files, i.e. give them other names than in previous
2894 versions. Their names now include the name of the final product, as
2895 well as its type mnemonic (bin, lib, shlib).
2899 * Added new option for 'openssl list', '-objects', which will display the
2900 list of built in objects, i.e. OIDs with names.
2904 * Added the options `-crl_lastupdate` and `-crl_nextupdate` to `openssl ca`,
2905 allowing the `lastUpdate` and `nextUpdate` fields in the generated CRL to
2910 * Added support for Linux Kernel TLS data-path. The Linux Kernel data-path
2911 improves application performance by removing data copies and providing
2912 applications with zero-copy system calls such as sendfile and splice.
2916 * The SSL option SSL_OP_CLEANSE_PLAINTEXT is introduced.
2920 * `PKCS12_parse` now maintains the order of the parsed certificates
2921 when outputting them via `*ca` (rather than reversing it).
2925 * Deprecated pthread fork support methods.
2929 * Added support for FFDHE key exchange in TLS 1.3.
2933 * Added a new concept for OpenSSL plugability: providers. This
2934 functionality is designed to replace the ENGINE API and ENGINE
2935 implementations, and to be much more dynamic, allowing provider
2936 authors to introduce new algorithms among other things, as long as
2937 there's an API that supports the algorithm type.
2939 With this concept comes a new core API for interaction between
2940 libcrypto and provider implementations. Public libcrypto functions
2941 that want to use providers do so through this core API.
2943 The main documentation for this core API is found in
2944 doc/man7/provider.pod, doc/man7/provider-base.pod, and they in turn
2945 refer to other manuals describing the API specific for supported
2946 algorithm types (also called operations).
2953 ### Changes between 1.1.1m and 1.1.1n [xx XXX xxxx]
2955 ### Changes between 1.1.1l and 1.1.1m [14 Dec 2021]
2957 * Avoid loading of a dynamic engine twice.
2961 * Prioritise DANE TLSA issuer certs over peer certs
2965 * Fixed random API for MacOS prior to 10.12
2967 These MacOS versions don't support the CommonCrypto APIs
2971 ### Changes between 1.1.1k and 1.1.1l [24 Aug 2021]
2973 * Fixed an SM2 Decryption Buffer Overflow.
2975 In order to decrypt SM2 encrypted data an application is expected to
2976 call the API function EVP_PKEY_decrypt(). Typically an application will
2977 call this function twice. The first time, on entry, the "out" parameter
2978 can be NULL and, on exit, the "outlen" parameter is populated with the
2979 buffer size required to hold the decrypted plaintext. The application
2980 can then allocate a sufficiently sized buffer and call EVP_PKEY_decrypt()
2981 again, but this time passing a non-NULL value for the "out" parameter.
2983 A bug in the implementation of the SM2 decryption code means that the
2984 calculation of the buffer size required to hold the plaintext returned
2985 by the first call to EVP_PKEY_decrypt() can be smaller than the actual
2986 size required by the second call. This can lead to a buffer overflow
2987 when EVP_PKEY_decrypt() is called by the application a second time with
2988 a buffer that is too small.
2990 A malicious attacker who is able present SM2 content for decryption to
2991 an application could cause attacker chosen data to overflow the buffer
2992 by up to a maximum of 62 bytes altering the contents of other data held
2993 after the buffer, possibly changing application behaviour or causing
2994 the application to crash. The location of the buffer is application
2995 dependent but is typically heap allocated.
3000 * Fixed various read buffer overruns processing ASN.1 strings
3002 ASN.1 strings are represented internally within OpenSSL as an ASN1_STRING
3003 structure which contains a buffer holding the string data and a field
3004 holding the buffer length. This contrasts with normal C strings which
3005 are represented as a buffer for the string data which is terminated
3006 with a NUL (0) byte.
3008 Although not a strict requirement, ASN.1 strings that are parsed using
3009 OpenSSL's own "d2i" functions (and other similar parsing functions) as
3010 well as any string whose value has been set with the ASN1_STRING_set()
3011 function will additionally NUL terminate the byte array in the
3012 ASN1_STRING structure.
3014 However, it is possible for applications to directly construct valid
3015 ASN1_STRING structures which do not NUL terminate the byte array by
3016 directly setting the "data" and "length" fields in the ASN1_STRING
3017 array. This can also happen by using the ASN1_STRING_set0() function.
3019 Numerous OpenSSL functions that print ASN.1 data have been found to
3020 assume that the ASN1_STRING byte array will be NUL terminated, even
3021 though this is not guaranteed for strings that have been directly
3022 constructed. Where an application requests an ASN.1 structure to be
3023 printed, and where that ASN.1 structure contains ASN1_STRINGs that have
3024 been directly constructed by the application without NUL terminating
3025 the "data" field, then a read buffer overrun can occur.
3027 The same thing can also occur during name constraints processing
3028 of certificates (for example if a certificate has been directly
3029 constructed by the application instead of loading it via the OpenSSL
3030 parsing functions, and the certificate contains non NUL terminated
3031 ASN1_STRING structures). It can also occur in the X509_get1_email(),
3032 X509_REQ_get1_email() and X509_get1_ocsp() functions.
3034 If a malicious actor can cause an application to directly construct an
3035 ASN1_STRING and then process it through one of the affected OpenSSL
3036 functions then this issue could be hit. This might result in a crash
3037 (causing a Denial of Service attack). It could also result in the
3038 disclosure of private memory contents (such as private keys, or
3039 sensitive plaintext).
3044 ### Changes between 1.1.1j and 1.1.1k [25 Mar 2021]
3046 * Fixed a problem with verifying a certificate chain when using the
3047 X509_V_FLAG_X509_STRICT flag. This flag enables additional security checks of
3048 the certificates present in a certificate chain. It is not set by default.
3050 Starting from OpenSSL version 1.1.1h a check to disallow certificates in
3051 the chain that have explicitly encoded elliptic curve parameters was added
3052 as an additional strict check.
3054 An error in the implementation of this check meant that the result of a
3055 previous check to confirm that certificates in the chain are valid CA
3056 certificates was overwritten. This effectively bypasses the check
3057 that non-CA certificates must not be able to issue other certificates.
3059 If a "purpose" has been configured then there is a subsequent opportunity
3060 for checks that the certificate is a valid CA. All of the named "purpose"
3061 values implemented in libcrypto perform this check. Therefore, where
3062 a purpose is set the certificate chain will still be rejected even when the
3063 strict flag has been used. A purpose is set by default in libssl client and
3064 server certificate verification routines, but it can be overridden or
3065 removed by an application.
3067 In order to be affected, an application must explicitly set the
3068 X509_V_FLAG_X509_STRICT verification flag and either not set a purpose
3069 for the certificate verification or, in the case of TLS client or server
3070 applications, override the default purpose.
3075 * Fixed an issue where an OpenSSL TLS server may crash if sent a maliciously
3076 crafted renegotiation ClientHello message from a client. If a TLSv1.2
3077 renegotiation ClientHello omits the signature_algorithms extension (where it
3078 was present in the initial ClientHello), but includes a
3079 signature_algorithms_cert extension then a NULL pointer dereference will
3080 result, leading to a crash and a denial of service attack.
3082 A server is only vulnerable if it has TLSv1.2 and renegotiation enabled
3083 (which is the default configuration). OpenSSL TLS clients are not impacted by
3087 *Peter Kästle and Samuel Sapalski*
3089 ### Changes between 1.1.1i and 1.1.1j [16 Feb 2021]
3091 * Fixed the X509_issuer_and_serial_hash() function. It attempts to
3092 create a unique hash value based on the issuer and serial number data
3093 contained within an X509 certificate. However, it was failing to correctly
3094 handle any errors that may occur while parsing the issuer field (which might
3095 occur if the issuer field is maliciously constructed). This may subsequently
3096 result in a NULL pointer deref and a crash leading to a potential denial of
3102 * Fixed the RSA_padding_check_SSLv23() function and the RSA_SSLV23_PADDING
3103 padding mode to correctly check for rollback attacks. This is considered a
3104 bug in OpenSSL 1.1.1 because it does not support SSLv2. In 1.0.2 this is
3109 Fixed the EVP_CipherUpdate, EVP_EncryptUpdate and EVP_DecryptUpdate
3110 functions. Previously they could overflow the output length argument in some
3111 cases where the input length is close to the maximum permissible length for
3112 an integer on the platform. In such cases the return value from the function
3113 call would be 1 (indicating success), but the output length value would be
3114 negative. This could cause applications to behave incorrectly or crash.
3119 * Fixed SRP_Calc_client_key so that it runs in constant time. The previous
3120 implementation called BN_mod_exp without setting BN_FLG_CONSTTIME. This
3121 could be exploited in a side channel attack to recover the password. Since
3122 the attack is local host only this is outside of the current OpenSSL
3123 threat model and therefore no CVE is assigned.
3125 Thanks to Mohammed Sabt and Daniel De Almeida Braga for reporting this
3130 ### Changes between 1.1.1h and 1.1.1i [8 Dec 2020]
3132 * Fixed NULL pointer deref in the GENERAL_NAME_cmp function
3133 This function could crash if both GENERAL_NAMEs contain an EDIPARTYNAME.
3134 If an attacker can control both items being compared then this could lead
3135 to a possible denial of service attack. OpenSSL itself uses the
3136 GENERAL_NAME_cmp function for two purposes:
3137 1) Comparing CRL distribution point names between an available CRL and a
3138 CRL distribution point embedded in an X509 certificate
3139 2) When verifying that a timestamp response token signer matches the
3140 timestamp authority name (exposed via the API functions
3141 TS_RESP_verify_response and TS_RESP_verify_token)
3146 ### Changes between 1.1.1g and 1.1.1h [22 Sep 2020]
3148 * Certificates with explicit curve parameters are now disallowed in
3149 verification chains if the X509_V_FLAG_X509_STRICT flag is used.
3153 * The 'MinProtocol' and 'MaxProtocol' configuration commands now silently
3154 ignore TLS protocol version bounds when configuring DTLS-based contexts, and
3155 conversely, silently ignore DTLS protocol version bounds when configuring
3156 TLS-based contexts. The commands can be repeated to set bounds of both
3157 types. The same applies with the corresponding "min_protocol" and
3158 "max_protocol" command-line switches, in case some application uses both TLS
3161 SSL_CTX instances that are created for a fixed protocol version (e.g.
3162 TLSv1_server_method()) also silently ignore version bounds. Previously
3163 attempts to apply bounds to these protocol versions would result in an
3164 error. Now only the "version-flexible" SSL_CTX instances are subject to
3165 limits in configuration files in command-line options.
3169 * Handshake now fails if Extended Master Secret extension is dropped
3174 * The Oracle Developer Studio compiler will start reporting deprecated APIs
3176 ### Changes between 1.1.1f and 1.1.1g [21 Apr 2020]
3178 * Fixed segmentation fault in SSL_check_chain()
3179 Server or client applications that call the SSL_check_chain() function
3180 during or after a TLS 1.3 handshake may crash due to a NULL pointer
3181 dereference as a result of incorrect handling of the
3182 "signature_algorithms_cert" TLS extension. The crash occurs if an invalid
3183 or unrecognised signature algorithm is received from the peer. This could
3184 be exploited by a malicious peer in a Denial of Service attack.
3189 * Added AES consttime code for no-asm configurations
3190 an optional constant time support for AES was added
3191 when building openssl for no-asm.
3192 Enable with: ./config no-asm -DOPENSSL_AES_CONST_TIME
3193 Disable with: ./config no-asm -DOPENSSL_NO_AES_CONST_TIME
3194 At this time this feature is by default disabled.
3195 It will be enabled by default in 3.0.
3199 ### Changes between 1.1.1e and 1.1.1f [31 Mar 2020]
3201 * Revert the change of EOF detection while reading in libssl to avoid
3202 regressions in applications depending on the current way of reporting
3203 the EOF. As the existing method is not fully accurate the change to
3204 reporting the EOF via SSL_ERROR_SSL is kept on the current development
3205 branch and will be present in the 3.0 release.
3209 * Revised BN_generate_prime_ex to not avoid factors 3..17863 in p-1
3210 when primes for RSA keys are computed.
3211 Since we previously always generated primes == 2 (mod 3) for RSA keys,
3212 the 2-prime and 3-prime RSA modules were easy to distinguish, since
3213 N = p*q = 1 (mod 3), but N = p*q*r = 2 (mod 3). Therefore, fingerprinting
3214 2-prime vs. 3-prime RSA keys was possible by computing N mod 3.
3215 This avoids possible fingerprinting of newly generated RSA modules.
3219 ### Changes between 1.1.1d and 1.1.1e [17 Mar 2020]
3221 * Properly detect EOF while reading in libssl. Previously if we hit an EOF
3222 while reading in libssl then we would report an error back to the
3223 application (SSL_ERROR_SYSCALL) but errno would be 0. We now add
3224 an error to the stack (which means we instead return SSL_ERROR_SSL) and
3225 therefore give a hint as to what went wrong.
3229 * Check that ed25519 and ed448 are allowed by the security level. Previously
3230 signature algorithms not using an MD were not being checked that they were
3231 allowed by the security level.
3235 * Fixed SSL_get_servername() behaviour. The behaviour of SSL_get_servername()
3236 was not quite right. The behaviour was not consistent between resumption
3237 and normal handshakes, and also not quite consistent with historical
3238 behaviour. The behaviour in various scenarios has been clarified and
3239 it has been updated to make it match historical behaviour as closely as
3244 * *[VMS only]* The header files that the VMS compilers include automatically,
3245 `__DECC_INCLUDE_PROLOGUE.H` and `__DECC_INCLUDE_EPILOGUE.H`, use pragmas
3246 that the C++ compiler doesn't understand. This is a shortcoming in the
3247 compiler, but can be worked around with `__cplusplus` guards.
3249 C++ applications that use OpenSSL libraries must be compiled using the
3250 qualifier `/NAMES=(AS_IS,SHORTENED)` to be able to use all the OpenSSL
3251 functions. Otherwise, only functions with symbols of less than 31
3252 characters can be used, as the linker will not be able to successfully
3253 resolve symbols with longer names.
3257 * Added a new method to gather entropy on VMS, based on SYS$GET_ENTROPY.
3258 The presence of this system service is determined at run-time.
3262 * Print all values for a PKCS#12 attribute with 'openssl pkcs12', not just
3267 ### Changes between 1.1.1c and 1.1.1d [10 Sep 2019]
3269 * Fixed a fork protection issue. OpenSSL 1.1.1 introduced a rewritten random
3270 number generator (RNG). This was intended to include protection in the
3271 event of a fork() system call in order to ensure that the parent and child
3272 processes did not share the same RNG state. However, this protection was not
3273 being used in the default case.
3275 A partial mitigation for this issue is that the output from a high
3276 precision timer is mixed into the RNG state so the likelihood of a parent
3277 and child process sharing state is significantly reduced.
3279 If an application already calls OPENSSL_init_crypto() explicitly using
3280 OPENSSL_INIT_ATFORK then this problem does not occur at all.
3283 *Matthias St. Pierre*
3285 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
3286 used even when parsing explicit parameters, when loading a encoded key
3287 or calling `EC_GROUP_new_from_ecpkparameters()`/
3288 `EC_GROUP_new_from_ecparameters()`.
3289 This prevents bypass of security hardening and performance gains,
3290 especially for curves with specialized EC_METHODs.
3291 By default, if a key encoded with explicit parameters is loaded and later
3292 encoded, the output is still encoded with explicit parameters, even if
3293 internally a "named" EC_GROUP is used for computation.
3297 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
3298 this change, EC_GROUP_set_generator would accept order and/or cofactor as
3299 NULL. After this change, only the cofactor parameter can be NULL. It also
3300 does some minimal sanity checks on the passed order.
3305 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
3306 An attack is simple, if the first CMS_recipientInfo is valid but the
3307 second CMS_recipientInfo is chosen ciphertext. If the second
3308 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
3309 encryption key will be replaced by garbage, and the message cannot be
3310 decoded, but if the RSA decryption fails, the correct encryption key is
3311 used and the recipient will not notice the attack.
3312 As a work around for this potential attack the length of the decrypted
3313 key must be equal to the cipher default key length, in case the
3314 certificate is not given and all recipientInfo are tried out.
3315 The old behaviour can be re-enabled in the CMS code by setting the
3316 CMS_DEBUG_DECRYPT flag.
3321 * Early start up entropy quality from the DEVRANDOM seed source has been
3322 improved for older Linux systems. The RAND subsystem will wait for
3323 /dev/random to be producing output before seeding from /dev/urandom.
3324 The seeded state is stored for future library initialisations using
3325 a system global shared memory segment. The shared memory identifier
3326 can be configured by defining OPENSSL_RAND_SEED_DEVRANDOM_SHM_ID to
3327 the desired value. The default identifier is 114.
3331 * Correct the extended master secret constant on EBCDIC systems. Without this
3332 fix TLS connections between an EBCDIC system and a non-EBCDIC system that
3333 negotiate EMS will fail. Unfortunately this also means that TLS connections
3334 between EBCDIC systems with this fix, and EBCDIC systems without this
3335 fix will fail if they negotiate EMS.
3339 * Use Windows installation paths in the mingw builds
3341 Mingw isn't a POSIX environment per se, which means that Windows
3342 paths should be used for installation.
3347 * Changed DH_check to accept parameters with order q and 2q subgroups.
3348 With order 2q subgroups the bit 0 of the private key is not secret
3349 but DH_generate_key works around that by clearing bit 0 of the
3350 private key for those. This avoids leaking bit 0 of the private key.
3354 * Significantly reduce secure memory usage by the randomness pools.
3358 * Revert the DEVRANDOM_WAIT feature for Linux systems
3360 The DEVRANDOM_WAIT feature added a select() call to wait for the
3361 /dev/random device to become readable before reading from the
3362 /dev/urandom device.
3364 It turned out that this change had negative side effects on
3365 performance which were not acceptable. After some discussion it
3366 was decided to revert this feature and leave it up to the OS
3367 resp. the platform maintainer to ensure a proper initialization
3368 during early boot time.
3370 *Matthias St. Pierre*
3372 ### Changes between 1.1.1b and 1.1.1c [28 May 2019]
3374 * Add build tests for C++. These are generated files that only do one
3375 thing, to include one public OpenSSL head file each. This tests that
3376 the public header files can be usefully included in a C++ application.
3378 This test isn't enabled by default. It can be enabled with the option
3379 'enable-buildtest-c++'.
3383 * Enable SHA3 pre-hashing for ECDSA and DSA.
3387 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
3388 This changes the size when using the `genpkey` command when no size is given.
3389 It fixes an omission in earlier changes that changed all RSA, DSA and DH
3390 generation commands to use 2048 bits by default.
3394 * Reorganize the manual pages to consistently have RETURN VALUES,
3395 EXAMPLES, SEE ALSO and HISTORY come in that order, and adjust
3396 util/fix-doc-nits accordingly.
3398 *Paul Yang, Joshua Lock*
3400 * Add the missing accessor EVP_PKEY_get0_engine()
3404 * Have commands like `s_client` and `s_server` output the signature scheme
3405 along with other cipher suite parameters when debugging.
3409 * Make OPENSSL_config() error agnostic again.
3413 * Do the error handling in RSA decryption constant time.
3417 * Prevent over long nonces in ChaCha20-Poly1305.
3419 ChaCha20-Poly1305 is an AEAD cipher, and requires a unique nonce input
3420 for every encryption operation. RFC 7539 specifies that the nonce value
3421 (IV) should be 96 bits (12 bytes). OpenSSL allows a variable nonce length
3422 and front pads the nonce with 0 bytes if it is less than 12
3423 bytes. However it also incorrectly allows a nonce to be set of up to 16
3424 bytes. In this case only the last 12 bytes are significant and any
3425 additional leading bytes are ignored.
3427 It is a requirement of using this cipher that nonce values are
3428 unique. Messages encrypted using a reused nonce value are susceptible to
3429 serious confidentiality and integrity attacks. If an application changes
3430 the default nonce length to be longer than 12 bytes and then makes a
3431 change to the leading bytes of the nonce expecting the new value to be a
3432 new unique nonce then such an application could inadvertently encrypt
3433 messages with a reused nonce.
3435 Additionally the ignored bytes in a long nonce are not covered by the
3436 integrity guarantee of this cipher. Any application that relies on the
3437 integrity of these ignored leading bytes of a long nonce may be further
3438 affected. Any OpenSSL internal use of this cipher, including in SSL/TLS,
3439 is safe because no such use sets such a long nonce value. However user
3440 applications that use this cipher directly and set a non-default nonce
3441 length to be longer than 12 bytes may be vulnerable.
3443 This issue was reported to OpenSSL on 16th of March 2019 by Joran Dirk
3449 * Add DEVRANDOM_WAIT feature for Linux systems
3451 On older Linux systems where the getrandom() system call is not available,
3452 OpenSSL normally uses the /dev/urandom device for seeding its CSPRNG.
3453 Contrary to getrandom(), the /dev/urandom device will not block during
3454 early boot when the kernel CSPRNG has not been seeded yet.
3456 To mitigate this known weakness, use select() to wait for /dev/random to
3457 become readable before reading from /dev/urandom.
3459 * Ensure that SM2 only uses SM3 as digest algorithm
3463 ### Changes between 1.1.1a and 1.1.1b [26 Feb 2019]
3465 * Change the info callback signals for the start and end of a post-handshake
3466 message exchange in TLSv1.3. In 1.1.1/1.1.1a we used SSL_CB_HANDSHAKE_START
3467 and SSL_CB_HANDSHAKE_DONE. Experience has shown that many applications get
3468 confused by this and assume that a TLSv1.2 renegotiation has started. This
3469 can break KeyUpdate handling. Instead we no longer signal the start and end
3470 of a post handshake message exchange (although the messages themselves are
3471 still signalled). This could break some applications that were expecting
3472 the old signals. However without this KeyUpdate is not usable for many
3477 ### Changes between 1.1.1 and 1.1.1a [20 Nov 2018]
3479 * Timing vulnerability in DSA signature generation
3481 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
3482 timing side channel attack. An attacker could use variations in the signing
3483 algorithm to recover the private key.
3485 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
3490 * Timing vulnerability in ECDSA signature generation
3492 The OpenSSL ECDSA signature algorithm has been shown to be vulnerable to a
3493 timing side channel attack. An attacker could use variations in the signing
3494 algorithm to recover the private key.
3496 This issue was reported to OpenSSL on 25th October 2018 by Samuel Weiser.
3501 * Fixed the issue that RAND_add()/RAND_seed() silently discards random input
3502 if its length exceeds 4096 bytes. The limit has been raised to a buffer size
3503 of two gigabytes and the error handling improved.
3505 This issue was reported to OpenSSL by Dr. Falko Strenzke. It has been
3506 categorized as a normal bug, not a security issue, because the DRBG reseeds
3507 automatically and is fully functional even without additional randomness
3508 provided by the application.
3510 ### Changes between 1.1.0i and 1.1.1 [11 Sep 2018]
3512 * Add a new ClientHello callback. Provides a callback interface that gives
3513 the application the ability to adjust the nascent SSL object at the
3514 earliest stage of ClientHello processing, immediately after extensions have
3515 been collected but before they have been processed. In particular, this
3516 callback can adjust the supported TLS versions in response to the contents
3521 * Add SM2 base algorithm support.
3525 * s390x assembly pack: add (improved) hardware-support for the following
3526 cryptographic primitives: sha3, shake, aes-gcm, aes-ccm, aes-ctr, aes-ofb,
3527 aes-cfb/cfb8, aes-ecb.
3531 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
3532 parameter is no longer accepted, as it leads to a corrupt table. NULL
3533 pem_str is reserved for alias entries only.
3537 * Use the new ec_scalar_mul_ladder scaffold to implement a specialized ladder
3538 step for prime curves. The new implementation is based on formulae from
3539 differential addition-and-doubling in homogeneous projective coordinates
3540 from Izu-Takagi "A fast parallel elliptic curve multiplication resistant
3541 against side channel attacks" and Brier-Joye "Weierstrass Elliptic Curves
3542 and Side-Channel Attacks" Eq. (8) for y-coordinate recovery, modified
3543 to work in projective coordinates.
3545 *Billy Bob Brumley, Nicola Tuveri*
3547 * Change generating and checking of primes so that the error rate of not
3548 being prime depends on the intended use based on the size of the input.
3549 For larger primes this will result in more rounds of Miller-Rabin.
3550 The maximal error rate for primes with more than 1080 bits is lowered
3553 *Kurt Roeckx, Annie Yousar*
3555 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
3559 * The 'tsget' script is renamed to 'tsget.pl', to avoid confusion when
3560 moving between systems, and to avoid confusion when a Windows build is
3561 done with mingw vs with MSVC. For POSIX installs, there's still a
3562 symlink or copy named 'tsget' to avoid that confusion as well.
3566 * Revert blinding in ECDSA sign and instead make problematic addition
3567 length-invariant. Switch even to fixed-length Montgomery multiplication.
3571 * Use the new ec_scalar_mul_ladder scaffold to implement a specialized ladder
3572 step for binary curves. The new implementation is based on formulae from
3573 differential addition-and-doubling in mixed Lopez-Dahab projective
3574 coordinates, modified to independently blind the operands.
3576 *Billy Bob Brumley, Sohaib ul Hassan, Nicola Tuveri*
3578 * Add a scaffold to optionally enhance the Montgomery ladder implementation
3579 for `ec_scalar_mul_ladder` (formerly `ec_mul_consttime`) allowing
3580 EC_METHODs to implement their own specialized "ladder step", to take
3581 advantage of more favorable coordinate systems or more efficient
3582 differential addition-and-doubling algorithms.
3584 *Billy Bob Brumley, Sohaib ul Hassan, Nicola Tuveri*
3586 * Modified the random device based seed sources to keep the relevant
3587 file descriptors open rather than reopening them on each access.
3588 This allows such sources to operate in a chroot() jail without
3589 the associated device nodes being available. This behaviour can be
3590 controlled using RAND_keep_random_devices_open().
3594 * Numerous side-channel attack mitigations have been applied. This may have
3595 performance impacts for some algorithms for the benefit of improved
3596 security. Specific changes are noted in this change log by their respective
3601 * AIX shared library support overhaul. Switch to AIX "natural" way of
3602 handling shared libraries, which means collecting shared objects of
3603 different versions and bitnesses in one common archive. This allows to
3604 mitigate conflict between 1.0 and 1.1 side-by-side installations. It
3605 doesn't affect the way 3rd party applications are linked, only how
3606 multi-version installation is managed.
3610 * Make ec_group_do_inverse_ord() more robust and available to other
3611 EC cryptosystems, so that irrespective of BN_FLG_CONSTTIME, SCA
3612 mitigations are applied to the fallback BN_mod_inverse().
3613 When using this function rather than BN_mod_inverse() directly, new
3614 EC cryptosystem implementations are then safer-by-default.
3618 * Add coordinate blinding for EC_POINT and implement projective
3619 coordinate blinding for generic prime curves as a countermeasure to
3620 chosen point SCA attacks.
3622 *Sohaib ul Hassan, Nicola Tuveri, Billy Bob Brumley*
3624 * Add blinding to ECDSA and DSA signatures to protect against side channel
3625 attacks discovered by Keegan Ryan (NCC Group).
3629 * Enforce checking in the `pkeyutl` command to ensure that the input
3630 length does not exceed the maximum supported digest length when performing
3631 a sign, verify or verifyrecover operation.
3635 * SSL_MODE_AUTO_RETRY is enabled by default. Applications that use blocking
3636 I/O in combination with something like select() or poll() will hang. This
3637 can be turned off again using SSL_CTX_clear_mode().
3638 Many applications do not properly handle non-application data records, and
3639 TLS 1.3 sends more of such records. Setting SSL_MODE_AUTO_RETRY works
3640 around the problems in those applications, but can also break some.
3641 It's recommended to read the manpages about SSL_read(), SSL_write(),
3642 SSL_get_error(), SSL_shutdown(), SSL_CTX_set_mode() and
3643 SSL_CTX_set_read_ahead() again.
3647 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
3648 now allow empty (zero character) pass phrases.
3652 * Apply blinding to binary field modular inversion and remove patent
3653 pending (OPENSSL_SUN_GF2M_DIV) BN_GF2m_mod_div implementation.
3657 * Deprecate ec2_mult.c and unify scalar multiplication code paths for
3658 binary and prime elliptic curves.
3662 * Remove ECDSA nonce padding: EC_POINT_mul is now responsible for
3663 constant time fixed point multiplication.
3667 * Revise elliptic curve scalar multiplication with timing attack
3668 defenses: ec_wNAF_mul redirects to a constant time implementation
3669 when computing fixed point and variable point multiplication (which
3670 in OpenSSL are mostly used with secret scalars in keygen, sign,
3671 ECDH derive operations).
3672 *Billy Bob Brumley, Nicola Tuveri, Cesar Pereida García,
3675 * Updated CONTRIBUTING
3679 * Updated DRBG / RAND to request nonce and additional low entropy
3680 randomness from the system.
3682 *Matthias St. Pierre*
3684 * Updated 'openssl rehash' to use OpenSSL consistent default.
3688 * Moved the load of the ssl_conf module to libcrypto, which helps
3689 loading engines that libssl uses before libssl is initialised.
3693 * Added EVP_PKEY_sign() and EVP_PKEY_verify() for EdDSA
3697 * Fixed X509_NAME_ENTRY_set to get multi-valued RDNs right in all cases.
3699 *Ingo Schwarze, Rich Salz*
3701 * Added output of accepting IP address and port for 'openssl s_server'
3705 * Added a new API for TLSv1.3 ciphersuites:
3706 SSL_CTX_set_ciphersuites()
3707 SSL_set_ciphersuites()
3711 * Memory allocation failures consistently add an error to the error
3716 * Don't use OPENSSL_ENGINES and OPENSSL_CONF environment values
3717 in libcrypto when run as setuid/setgid.
3721 * Load any config file by default when libssl is used.
3725 * Added new public header file <openssl/rand_drbg.h> and documentation
3726 for the RAND_DRBG API. See manual page RAND_DRBG(7) for an overview.
3728 *Matthias St. Pierre*
3730 * QNX support removed (cannot find contributors to get their approval
3731 for the license change).
3735 * TLSv1.3 replay protection for early data has been implemented. See the
3736 SSL_read_early_data() man page for further details.
3740 * Separated TLSv1.3 ciphersuite configuration out from TLSv1.2 ciphersuite
3741 configuration. TLSv1.3 ciphersuites are not compatible with TLSv1.2 and
3742 below. Similarly TLSv1.2 ciphersuites are not compatible with TLSv1.3.
3743 In order to avoid issues where legacy TLSv1.2 ciphersuite configuration
3744 would otherwise inadvertently disable all TLSv1.3 ciphersuites the
3745 configuration has been separated out. See the ciphers man page or the
3746 SSL_CTX_set_ciphersuites() man page for more information.
3750 * On POSIX (BSD, Linux, ...) systems the ocsp(1) command running
3751 in responder mode now supports the new "-multi" option, which
3752 spawns the specified number of child processes to handle OCSP
3753 requests. The "-timeout" option now also limits the OCSP
3754 responder's patience to wait to receive the full client request
3755 on a newly accepted connection. Child processes are respawned
3756 as needed, and the CA index file is automatically reloaded
3757 when changed. This makes it possible to run the "ocsp" responder
3758 as a long-running service, making the OpenSSL CA somewhat more
3759 feature-complete. In this mode, most diagnostic messages logged
3760 after entering the event loop are logged via syslog(3) rather than
3765 * Added support for X448 and Ed448. Heavily based on original work by
3770 * Extend OSSL_STORE with capabilities to search and to narrow the set of
3771 objects loaded. This adds the functions OSSL_STORE_expect() and
3772 OSSL_STORE_find() as well as needed tools to construct searches and
3773 get the search data out of them.
3777 * Support for TLSv1.3 added. Note that users upgrading from an earlier
3778 version of OpenSSL should review their configuration settings to ensure
3779 that they are still appropriate for TLSv1.3. For further information see:
3780 <https://wiki.openssl.org/index.php/TLS1.3>
3784 * Grand redesign of the OpenSSL random generator
3786 The default RAND method now utilizes an AES-CTR DRBG according to
3787 NIST standard SP 800-90Ar1. The new random generator is essentially
3788 a port of the default random generator from the OpenSSL FIPS 2.0
3789 object module. It is a hybrid deterministic random bit generator
3790 using an AES-CTR bit stream and which seeds and reseeds itself
3791 automatically using trusted system entropy sources.
3793 Some of its new features are:
3794 - Support for multiple DRBG instances with seed chaining.
3795 - The default RAND method makes use of a DRBG.
3796 - There is a public and private DRBG instance.
3797 - The DRBG instances are fork-safe.
3798 - Keep all global DRBG instances on the secure heap if it is enabled.
3799 - The public and private DRBG instance are per thread for lock free
3802 *Paul Dale, Benjamin Kaduk, Kurt Roeckx, Rich Salz, Matthias St. Pierre*
3804 * Changed Configure so it only says what it does and doesn't dump
3805 so much data. Instead, ./configdata.pm should be used as a script
3806 to display all sorts of configuration data.
3810 * Added processing of "make variables" to Configure.
3814 * Added SHA512/224 and SHA512/256 algorithm support.
3818 * The last traces of Netware support, first removed in 1.1.0, have
3823 * Get rid of Makefile.shared, and in the process, make the processing
3824 of certain files (rc.obj, or the .def/.map/.opt files produced from
3825 the ordinal files) more visible and hopefully easier to trace and
3826 debug (or make silent).
3830 * Make it possible to have environment variable assignments as
3831 arguments to config / Configure.
3835 * Add multi-prime RSA (RFC 8017) support.
3839 * Add SM3 implemented according to GB/T 32905-2016
3840 *Jack Lloyd <jack.lloyd@ribose.com>,*
3841 *Ronald Tse <ronald.tse@ribose.com>,*
3842 *Erick Borsboom <erick.borsboom@ribose.com>*
3844 * Add 'Maximum Fragment Length' TLS extension negotiation and support
3845 as documented in RFC6066.
3846 Based on a patch from Tomasz Moń
3848 *Filipe Raimundo da Silva*
3850 * Add SM4 implemented according to GB/T 32907-2016.
3851 *Jack Lloyd <jack.lloyd@ribose.com>,*
3852 *Ronald Tse <ronald.tse@ribose.com>,*
3853 *Erick Borsboom <erick.borsboom@ribose.com>*
3855 * Reimplement -newreq-nodes and ERR_error_string_n; the
3856 original author does not agree with the license change.
3860 * Add ARIA AEAD TLS support.
3864 * Some macro definitions to support VS6 have been removed. Visual
3865 Studio 6 has not worked since 1.1.0
3869 * Add ERR_clear_last_mark(), to allow callers to clear the last mark
3870 without clearing the errors.
3874 * Add "atfork" functions. If building on a system that without
3875 pthreads, see doc/man3/OPENSSL_fork_prepare.pod for application
3876 requirements. The RAND facility now uses/requires this.
3884 * The UI API becomes a permanent and integral part of libcrypto, i.e.
3885 not possible to disable entirely. However, it's still possible to
3886 disable the console reading UI method, UI_OpenSSL() (use UI_null()
3889 To disable, configure with 'no-ui-console'. 'no-ui' is still
3890 possible to use as an alias. Check at compile time with the
3891 macro OPENSSL_NO_UI_CONSOLE. The macro OPENSSL_NO_UI is still
3892 possible to check and is an alias for OPENSSL_NO_UI_CONSOLE.
3896 * Add a STORE module, which implements a uniform and URI based reader of
3897 stores that can contain keys, certificates, CRLs and numerous other
3898 objects. The main API is loosely based on a few stdio functions,
3899 and includes OSSL_STORE_open, OSSL_STORE_load, OSSL_STORE_eof,
3900 OSSL_STORE_error and OSSL_STORE_close.
3901 The implementation uses backends called "loaders" to implement arbitrary
3902 URI schemes. There is one built in "loader" for the 'file' scheme.
3906 * Add devcrypto engine. This has been implemented against cryptodev-linux,
3907 then adjusted to work on FreeBSD 8.4 as well.
3908 Enable by configuring with 'enable-devcryptoeng'. This is done by default
3909 on BSD implementations, as cryptodev.h is assumed to exist on all of them.
3913 * Module names can prefixed with OSSL_ or OPENSSL_. This affects
3914 util/mkerr.pl, which is adapted to allow those prefixes, leading to
3915 error code calls like this:
3917 OSSL_FOOerr(OSSL_FOO_F_SOMETHING, OSSL_FOO_R_WHATEVER);
3919 With this change, we claim the namespaces OSSL and OPENSSL in a manner
3920 that can be encoded in C. For the foreseeable future, this will only
3923 *Richard Levitte and Tim Hudson*
3925 * Removed BSD cryptodev engine.
3929 * Add a build target 'build_all_generated', to build all generated files
3930 and only that. This can be used to prepare everything that requires
3931 things like perl for a system that lacks perl and then move everything
3932 to that system and do the rest of the build there.
3936 * In the UI interface, make it possible to duplicate the user data. This
3937 can be used by engines that need to retain the data for a longer time
3938 than just the call where this user data is passed.
3942 * Ignore the '-named_curve auto' value for compatibility of applications
3945 *Tomáš Mráz <tmraz@fedoraproject.org>*
3947 * Fragmented SSL/TLS alerts are no longer accepted. An alert message is 2
3948 bytes long. In theory it is permissible in SSLv3 - TLSv1.2 to fragment such
3949 alerts across multiple records (some of which could be empty). In practice
3950 it make no sense to send an empty alert record, or to fragment one. TLSv1.3
3951 prohibits this altogether and other libraries (BoringSSL, NSS) do not
3952 support this at all. Supporting it adds significant complexity to the
3953 record layer, and its removal is unlikely to cause interoperability
3958 * Add the ASN.1 types INT32, UINT32, INT64, UINT64 and variants prefixed
3959 with Z. These are meant to replace LONG and ZLONG and to be size safe.
3960 The use of LONG and ZLONG is discouraged and scheduled for deprecation
3965 * Add the 'z' and 'j' modifiers to BIO_printf() et al formatting string,
3966 'z' is to be used for [s]size_t, and 'j' - with [u]int64_t.
3968 *Richard Levitte, Andy Polyakov*
3970 * Add EC_KEY_get0_engine(), which does for EC_KEY what RSA_get0_engine()
3975 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
3976 platform rather than 'mingw'.
3980 * The functions X509_STORE_add_cert and X509_STORE_add_crl return
3981 success if they are asked to add an object which already exists
3982 in the store. This change cascades to other functions which load
3983 certificates and CRLs.
3987 * x86_64 assembly pack: annotate code with DWARF CFI directives to
3988 facilitate stack unwinding even from assembly subroutines.
3992 * Remove VAX C specific definitions of OPENSSL_EXPORT, OPENSSL_EXTERN.
3993 Also remove OPENSSL_GLOBAL entirely, as it became a no-op.
3997 * Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
3998 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
3999 which is the minimum version we support.
4003 * Certificate time validation (X509_cmp_time) enforces stricter
4004 compliance with RFC 5280. Fractional seconds and timezone offsets
4005 are no longer allowed.
4009 * Add support for ARIA
4013 * s_client will now send the Server Name Indication (SNI) extension by
4014 default unless the new "-noservername" option is used. The server name is
4015 based on the host provided to the "-connect" option unless overridden by
4016 using "-servername".
4020 * Add support for SipHash
4024 * OpenSSL now fails if it receives an unrecognised record type in TLS1.0
4025 or TLS1.1. Previously this only happened in SSLv3 and TLS1.2. This is to
4026 prevent issues where no progress is being made and the peer continually
4027 sends unrecognised record types, using up resources processing them.
4031 * 'openssl passwd' can now produce SHA256 and SHA512 based output,
4032 using the algorithm defined in
4033 <https://www.akkadia.org/drepper/SHA-crypt.txt>
4037 * Heartbeat support has been removed; the ABI is changed for now.
4039 *Richard Levitte, Rich Salz*
4041 * Support for SSL_OP_NO_ENCRYPT_THEN_MAC in SSL_CONF_cmd.
4045 * The RSA "null" method, which was partially supported to avoid patent
4046 issues, has been replaced to always returns NULL.
4053 ### Changes between 1.1.0k and 1.1.0l [10 Sep 2019]
4055 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
4056 used even when parsing explicit parameters, when loading a encoded key
4057 or calling `EC_GROUP_new_from_ecpkparameters()`/
4058 `EC_GROUP_new_from_ecparameters()`.
4059 This prevents bypass of security hardening and performance gains,
4060 especially for curves with specialized EC_METHODs.
4061 By default, if a key encoded with explicit parameters is loaded and later
4062 encoded, the output is still encoded with explicit parameters, even if
4063 internally a "named" EC_GROUP is used for computation.
4067 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
4068 this change, EC_GROUP_set_generator would accept order and/or cofactor as
4069 NULL. After this change, only the cofactor parameter can be NULL. It also
4070 does some minimal sanity checks on the passed order.
4075 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
4076 An attack is simple, if the first CMS_recipientInfo is valid but the
4077 second CMS_recipientInfo is chosen ciphertext. If the second
4078 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
4079 encryption key will be replaced by garbage, and the message cannot be
4080 decoded, but if the RSA decryption fails, the correct encryption key is
4081 used and the recipient will not notice the attack.
4082 As a work around for this potential attack the length of the decrypted
4083 key must be equal to the cipher default key length, in case the
4084 certificate is not given and all recipientInfo are tried out.
4085 The old behaviour can be re-enabled in the CMS code by setting the
4086 CMS_DEBUG_DECRYPT flag.
4091 * Use Windows installation paths in the mingw builds
4093 Mingw isn't a POSIX environment per se, which means that Windows
4094 paths should be used for installation.
4099 ### Changes between 1.1.0j and 1.1.0k [28 May 2019]
4101 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
4102 This changes the size when using the `genpkey` command when no size is given.
4103 It fixes an omission in earlier changes that changed all RSA, DSA and DH
4104 generation commands to use 2048 bits by default.
4108 * Prevent over long nonces in ChaCha20-Poly1305.
4110 ChaCha20-Poly1305 is an AEAD cipher, and requires a unique nonce input
4111 for every encryption operation. RFC 7539 specifies that the nonce value
4112 (IV) should be 96 bits (12 bytes). OpenSSL allows a variable nonce length
4113 and front pads the nonce with 0 bytes if it is less than 12
4114 bytes. However it also incorrectly allows a nonce to be set of up to 16
4115 bytes. In this case only the last 12 bytes are significant and any
4116 additional leading bytes are ignored.
4118 It is a requirement of using this cipher that nonce values are
4119 unique. Messages encrypted using a reused nonce value are susceptible to
4120 serious confidentiality and integrity attacks. If an application changes
4121 the default nonce length to be longer than 12 bytes and then makes a
4122 change to the leading bytes of the nonce expecting the new value to be a
4123 new unique nonce then such an application could inadvertently encrypt
4124 messages with a reused nonce.
4126 Additionally the ignored bytes in a long nonce are not covered by the
4127 integrity guarantee of this cipher. Any application that relies on the
4128 integrity of these ignored leading bytes of a long nonce may be further
4129 affected. Any OpenSSL internal use of this cipher, including in SSL/TLS,
4130 is safe because no such use sets such a long nonce value. However user
4131 applications that use this cipher directly and set a non-default nonce
4132 length to be longer than 12 bytes may be vulnerable.
4134 This issue was reported to OpenSSL on 16th of March 2019 by Joran Dirk
4140 * Added SCA hardening for modular field inversion in EC_GROUP through
4141 a new dedicated field_inv() pointer in EC_METHOD.
4142 This also addresses a leakage affecting conversions from projective
4143 to affine coordinates.
4145 *Billy Bob Brumley, Nicola Tuveri*
4147 * Fix a use after free bug in d2i_X509_PUBKEY when overwriting a
4148 re-used X509_PUBKEY object if the second PUBKEY is malformed.
4152 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
4156 * Remove the 'dist' target and add a tarball building script. The
4157 'dist' target has fallen out of use, and it shouldn't be
4158 necessary to configure just to create a source distribution.
4162 ### Changes between 1.1.0i and 1.1.0j [20 Nov 2018]
4164 * Timing vulnerability in DSA signature generation
4166 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
4167 timing side channel attack. An attacker could use variations in the signing
4168 algorithm to recover the private key.
4170 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
4175 * Timing vulnerability in ECDSA signature generation
4177 The OpenSSL ECDSA signature algorithm has been shown to be vulnerable to a
4178 timing side channel attack. An attacker could use variations in the signing
4179 algorithm to recover the private key.
4181 This issue was reported to OpenSSL on 25th October 2018 by Samuel Weiser.
4186 * Add coordinate blinding for EC_POINT and implement projective
4187 coordinate blinding for generic prime curves as a countermeasure to
4188 chosen point SCA attacks.
4190 *Sohaib ul Hassan, Nicola Tuveri, Billy Bob Brumley*
4192 ### Changes between 1.1.0h and 1.1.0i [14 Aug 2018]
4194 * Client DoS due to large DH parameter
4196 During key agreement in a TLS handshake using a DH(E) based ciphersuite a
4197 malicious server can send a very large prime value to the client. This will
4198 cause the client to spend an unreasonably long period of time generating a
4199 key for this prime resulting in a hang until the client has finished. This
4200 could be exploited in a Denial Of Service attack.
4202 This issue was reported to OpenSSL on 5th June 2018 by Guido Vranken
4207 * Cache timing vulnerability in RSA Key Generation
4209 The OpenSSL RSA Key generation algorithm has been shown to be vulnerable to
4210 a cache timing side channel attack. An attacker with sufficient access to
4211 mount cache timing attacks during the RSA key generation process could
4212 recover the private key.
4214 This issue was reported to OpenSSL on 4th April 2018 by Alejandro Cabrera
4215 Aldaya, Billy Brumley, Cesar Pereida Garcia and Luis Manuel Alvarez Tapia.
4220 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
4221 parameter is no longer accepted, as it leads to a corrupt table. NULL
4222 pem_str is reserved for alias entries only.
4226 * Revert blinding in ECDSA sign and instead make problematic addition
4227 length-invariant. Switch even to fixed-length Montgomery multiplication.
4231 * Change generating and checking of primes so that the error rate of not
4232 being prime depends on the intended use based on the size of the input.
4233 For larger primes this will result in more rounds of Miller-Rabin.
4234 The maximal error rate for primes with more than 1080 bits is lowered
4237 *Kurt Roeckx, Annie Yousar*
4239 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
4243 * Add blinding to ECDSA and DSA signatures to protect against side channel
4244 attacks discovered by Keegan Ryan (NCC Group).
4248 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
4249 now allow empty (zero character) pass phrases.
4253 * Certificate time validation (X509_cmp_time) enforces stricter
4254 compliance with RFC 5280. Fractional seconds and timezone offsets
4255 are no longer allowed.
4259 * Fixed a text canonicalisation bug in CMS
4261 Where a CMS detached signature is used with text content the text goes
4262 through a canonicalisation process first prior to signing or verifying a
4263 signature. This process strips trailing space at the end of lines, converts
4264 line terminators to CRLF and removes additional trailing line terminators
4265 at the end of a file. A bug in the canonicalisation process meant that
4266 some characters, such as form-feed, were incorrectly treated as whitespace
4267 and removed. This is contrary to the specification (RFC5485). This fix
4268 could mean that detached text data signed with an earlier version of
4269 OpenSSL 1.1.0 may fail to verify using the fixed version, or text data
4270 signed with a fixed OpenSSL may fail to verify with an earlier version of
4271 OpenSSL 1.1.0. A workaround is to only verify the canonicalised text data
4272 and use the "-binary" flag (for the "cms" command line application) or set
4273 the SMIME_BINARY/PKCS7_BINARY/CMS_BINARY flags (if using CMS_verify()).
4277 ### Changes between 1.1.0g and 1.1.0h [27 Mar 2018]
4279 * Constructed ASN.1 types with a recursive definition could exceed the stack
4281 Constructed ASN.1 types with a recursive definition (such as can be found
4282 in PKCS7) could eventually exceed the stack given malicious input with
4283 excessive recursion. This could result in a Denial Of Service attack. There
4284 are no such structures used within SSL/TLS that come from untrusted sources
4285 so this is considered safe.
4287 This issue was reported to OpenSSL on 4th January 2018 by the OSS-fuzz
4293 * Incorrect CRYPTO_memcmp on HP-UX PA-RISC
4295 Because of an implementation bug the PA-RISC CRYPTO_memcmp function is
4296 effectively reduced to only comparing the least significant bit of each
4297 byte. This allows an attacker to forge messages that would be considered as
4298 authenticated in an amount of tries lower than that guaranteed by the
4299 security claims of the scheme. The module can only be compiled by the
4300 HP-UX assembler, so that only HP-UX PA-RISC targets are affected.
4302 This issue was reported to OpenSSL on 2nd March 2018 by Peter Waltenberg
4308 * Add a build target 'build_all_generated', to build all generated files
4309 and only that. This can be used to prepare everything that requires
4310 things like perl for a system that lacks perl and then move everything
4311 to that system and do the rest of the build there.
4315 * Backport SSL_OP_NO_RENGOTIATION
4317 OpenSSL 1.0.2 and below had the ability to disable renegotiation using the
4318 (undocumented) SSL3_FLAGS_NO_RENEGOTIATE_CIPHERS flag. Due to the opacity
4319 changes this is no longer possible in 1.1.0. Therefore, the new
4320 SSL_OP_NO_RENEGOTIATION option from 1.1.1-dev has been backported to
4321 1.1.0 to provide equivalent functionality.
4323 Note that if an application built against 1.1.0h headers (or above) is run
4324 using an older version of 1.1.0 (prior to 1.1.0h) then the option will be
4325 accepted but nothing will happen, i.e. renegotiation will not be prevented.
4329 * Removed the OS390-Unix config target. It relied on a script that doesn't
4334 * rsaz_1024_mul_avx2 overflow bug on x86_64
4336 There is an overflow bug in the AVX2 Montgomery multiplication procedure
4337 used in exponentiation with 1024-bit moduli. No EC algorithms are affected.
4338 Analysis suggests that attacks against RSA and DSA as a result of this
4339 defect would be very difficult to perform and are not believed likely.
4340 Attacks against DH1024 are considered just feasible, because most of the
4341 work necessary to deduce information about a private key may be performed
4342 offline. The amount of resources required for such an attack would be
4343 significant. However, for an attack on TLS to be meaningful, the server
4344 would have to share the DH1024 private key among multiple clients, which is
4345 no longer an option since CVE-2016-0701.
4347 This only affects processors that support the AVX2 but not ADX extensions
4348 like Intel Haswell (4th generation).
4350 This issue was reported to OpenSSL by David Benjamin (Google). The issue
4351 was originally found via the OSS-Fuzz project.
4356 ### Changes between 1.1.0f and 1.1.0g [2 Nov 2017]
4358 * bn_sqrx8x_internal carry bug on x86_64
4360 There is a carry propagating bug in the x86_64 Montgomery squaring
4361 procedure. No EC algorithms are affected. Analysis suggests that attacks
4362 against RSA and DSA as a result of this defect would be very difficult to
4363 perform and are not believed likely. Attacks against DH are considered just
4364 feasible (although very difficult) because most of the work necessary to
4365 deduce information about a private key may be performed offline. The amount
4366 of resources required for such an attack would be very significant and
4367 likely only accessible to a limited number of attackers. An attacker would
4368 additionally need online access to an unpatched system using the target
4369 private key in a scenario with persistent DH parameters and a private
4370 key that is shared between multiple clients.
4372 This only affects processors that support the BMI1, BMI2 and ADX extensions
4373 like Intel Broadwell (5th generation) and later or AMD Ryzen.
4375 This issue was reported to OpenSSL by the OSS-Fuzz project.
4380 * Malformed X.509 IPAddressFamily could cause OOB read
4382 If an X.509 certificate has a malformed IPAddressFamily extension,
4383 OpenSSL could do a one-byte buffer overread. The most likely result
4384 would be an erroneous display of the certificate in text format.
4386 This issue was reported to OpenSSL by the OSS-Fuzz project.
4391 ### Changes between 1.1.0e and 1.1.0f [25 May 2017]
4393 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
4394 platform rather than 'mingw'.
4398 * Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
4399 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
4400 which is the minimum version we support.
4404 ### Changes between 1.1.0d and 1.1.0e [16 Feb 2017]
4406 * Encrypt-Then-Mac renegotiation crash
4408 During a renegotiation handshake if the Encrypt-Then-Mac extension is
4409 negotiated where it was not in the original handshake (or vice-versa) then
4410 this can cause OpenSSL to crash (dependent on ciphersuite). Both clients
4411 and servers are affected.
4413 This issue was reported to OpenSSL by Joe Orton (Red Hat).
4418 ### Changes between 1.1.0c and 1.1.0d [26 Jan 2017]
4420 * Truncated packet could crash via OOB read
4422 If one side of an SSL/TLS path is running on a 32-bit host and a specific
4423 cipher is being used, then a truncated packet can cause that host to
4424 perform an out-of-bounds read, usually resulting in a crash.
4426 This issue was reported to OpenSSL by Robert Święcki of Google.
4431 * Bad (EC)DHE parameters cause a client crash
4433 If a malicious server supplies bad parameters for a DHE or ECDHE key
4434 exchange then this can result in the client attempting to dereference a
4435 NULL pointer leading to a client crash. This could be exploited in a Denial
4438 This issue was reported to OpenSSL by Guido Vranken.
4443 * BN_mod_exp may produce incorrect results on x86_64
4445 There is a carry propagating bug in the x86_64 Montgomery squaring
4446 procedure. No EC algorithms are affected. Analysis suggests that attacks
4447 against RSA and DSA as a result of this defect would be very difficult to
4448 perform and are not believed likely. Attacks against DH are considered just
4449 feasible (although very difficult) because most of the work necessary to
4450 deduce information about a private key may be performed offline. The amount
4451 of resources required for such an attack would be very significant and
4452 likely only accessible to a limited number of attackers. An attacker would
4453 additionally need online access to an unpatched system using the target
4454 private key in a scenario with persistent DH parameters and a private
4455 key that is shared between multiple clients. For example this can occur by
4456 default in OpenSSL DHE based SSL/TLS ciphersuites. Note: This issue is very
4457 similar to CVE-2015-3193 but must be treated as a separate problem.
4459 This issue was reported to OpenSSL by the OSS-Fuzz project.
4464 ### Changes between 1.1.0b and 1.1.0c [10 Nov 2016]
4466 * ChaCha20/Poly1305 heap-buffer-overflow
4468 TLS connections using `*-CHACHA20-POLY1305` ciphersuites are susceptible to
4469 a DoS attack by corrupting larger payloads. This can result in an OpenSSL
4470 crash. This issue is not considered to be exploitable beyond a DoS.
4472 This issue was reported to OpenSSL by Robert Święcki (Google Security Team)
4477 * CMS Null dereference
4479 Applications parsing invalid CMS structures can crash with a NULL pointer
4480 dereference. This is caused by a bug in the handling of the ASN.1 CHOICE
4481 type in OpenSSL 1.1.0 which can result in a NULL value being passed to the
4482 structure callback if an attempt is made to free certain invalid encodings.
4483 Only CHOICE structures using a callback which do not handle NULL value are
4486 This issue was reported to OpenSSL by Tyler Nighswander of ForAllSecure.
4491 * Montgomery multiplication may produce incorrect results
4493 There is a carry propagating bug in the Broadwell-specific Montgomery
4494 multiplication procedure that handles input lengths divisible by, but
4495 longer than 256 bits. Analysis suggests that attacks against RSA, DSA
4496 and DH private keys are impossible. This is because the subroutine in
4497 question is not used in operations with the private key itself and an input
4498 of the attacker's direct choice. Otherwise the bug can manifest itself as
4499 transient authentication and key negotiation failures or reproducible
4500 erroneous outcome of public-key operations with specially crafted input.
4501 Among EC algorithms only Brainpool P-512 curves are affected and one
4502 presumably can attack ECDH key negotiation. Impact was not analyzed in
4503 detail, because pre-requisites for attack are considered unlikely. Namely
4504 multiple clients have to choose the curve in question and the server has to
4505 share the private key among them, neither of which is default behaviour.
4506 Even then only clients that chose the curve will be affected.
4508 This issue was publicly reported as transient failures and was not
4509 initially recognized as a security issue. Thanks to Richard Morgan for
4510 providing reproducible case.
4515 * Removed automatic addition of RPATH in shared libraries and executables,
4516 as this was a remainder from OpenSSL 1.0.x and isn't needed any more.
4520 ### Changes between 1.1.0a and 1.1.0b [26 Sep 2016]
4522 * Fix Use After Free for large message sizes
4524 The patch applied to address CVE-2016-6307 resulted in an issue where if a
4525 message larger than approx 16k is received then the underlying buffer to
4526 store the incoming message is reallocated and moved. Unfortunately a
4527 dangling pointer to the old location is left which results in an attempt to
4528 write to the previously freed location. This is likely to result in a
4529 crash, however it could potentially lead to execution of arbitrary code.
4531 This issue only affects OpenSSL 1.1.0a.
4533 This issue was reported to OpenSSL by Robert Święcki.
4538 ### Changes between 1.1.0 and 1.1.0a [22 Sep 2016]
4540 * OCSP Status Request extension unbounded memory growth
4542 A malicious client can send an excessively large OCSP Status Request
4543 extension. If that client continually requests renegotiation, sending a
4544 large OCSP Status Request extension each time, then there will be unbounded
4545 memory growth on the server. This will eventually lead to a Denial Of
4546 Service attack through memory exhaustion. Servers with a default
4547 configuration are vulnerable even if they do not support OCSP. Builds using
4548 the "no-ocsp" build time option are not affected.
4550 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4555 * SSL_peek() hang on empty record
4557 OpenSSL 1.1.0 SSL/TLS will hang during a call to SSL_peek() if the peer
4558 sends an empty record. This could be exploited by a malicious peer in a
4559 Denial Of Service attack.
4561 This issue was reported to OpenSSL by Alex Gaynor.
4566 * Excessive allocation of memory in tls_get_message_header() and
4567 dtls1_preprocess_fragment()
4569 A (D)TLS message includes 3 bytes for its length in the header for the
4570 message. This would allow for messages up to 16Mb in length. Messages of
4571 this length are excessive and OpenSSL includes a check to ensure that a
4572 peer is sending reasonably sized messages in order to avoid too much memory
4573 being consumed to service a connection. A flaw in the logic of version
4574 1.1.0 means that memory for the message is allocated too early, prior to
4575 the excessive message length check. Due to way memory is allocated in
4576 OpenSSL this could mean an attacker could force up to 21Mb to be allocated
4577 to service a connection. This could lead to a Denial of Service through
4578 memory exhaustion. However, the excessive message length check still takes
4579 place, and this would cause the connection to immediately fail. Assuming
4580 that the application calls SSL_free() on the failed connection in a timely
4581 manner then the 21Mb of allocated memory will then be immediately freed
4582 again. Therefore, the excessive memory allocation will be transitory in
4583 nature. This then means that there is only a security impact if:
4585 1) The application does not call SSL_free() in a timely manner in the event
4586 that the connection fails
4588 2) The application is working in a constrained environment where there is
4589 very little free memory
4591 3) The attacker initiates multiple connection attempts such that there are
4592 multiple connections in a state where memory has been allocated for the
4593 connection; SSL_free() has not yet been called; and there is insufficient
4594 memory to service the multiple requests.
4596 Except in the instance of (1) above any Denial Of Service is likely to be
4597 transitory because as soon as the connection fails the memory is
4598 subsequently freed again in the SSL_free() call. However there is an
4599 increased risk during this period of application crashes due to the lack of
4600 memory - which would then mean a more serious Denial of Service.
4602 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4603 (CVE-2016-6307 and CVE-2016-6308)
4607 * solaris-x86-cc, i.e. 32-bit configuration with vendor compiler,
4608 had to be removed. Primary reason is that vendor assembler can't
4609 assemble our modules with -KPIC flag. As result it, assembly
4610 support, was not even available as option. But its lack means
4611 lack of side-channel resistant code, which is incompatible with
4612 security by todays standards. Fortunately gcc is readily available
4613 prepackaged option, which we firmly point at...
4617 ### Changes between 1.0.2h and 1.1.0 [25 Aug 2016]
4619 * Windows command-line tool supports UTF-8 opt-in option for arguments
4620 and console input. Setting OPENSSL_WIN32_UTF8 environment variable
4621 (to any value) allows Windows user to access PKCS#12 file generated
4622 with Windows CryptoAPI and protected with non-ASCII password, as well
4623 as files generated under UTF-8 locale on Linux also protected with
4628 * To mitigate the SWEET32 attack ([CVE-2016-2183]), 3DES cipher suites
4629 have been disabled by default and removed from DEFAULT, just like RC4.
4630 See the RC4 item below to re-enable both.
4634 * The method for finding the storage location for the Windows RAND seed file
4635 has changed. First we check %RANDFILE%. If that is not set then we check
4636 the directories %HOME%, %USERPROFILE% and %SYSTEMROOT% in that order. If
4637 all else fails we fall back to C:\.
4641 * The EVP_EncryptUpdate() function has had its return type changed from void
4642 to int. A return of 0 indicates and error while a return of 1 indicates
4647 * The flags RSA_FLAG_NO_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME and
4648 DH_FLAG_NO_EXP_CONSTTIME which previously provided the ability to switch
4649 off the constant time implementation for RSA, DSA and DH have been made
4650 no-ops and deprecated.
4654 * Windows RAND implementation was simplified to only get entropy by
4655 calling CryptGenRandom(). Various other RAND-related tickets
4658 *Joseph Wylie Yandle, Rich Salz*
4660 * The stack and lhash API's were renamed to start with `OPENSSL_SK_`
4661 and `OPENSSL_LH_`, respectively. The old names are available
4662 with API compatibility. They new names are now completely documented.
4666 * Unify TYPE_up_ref(obj) methods signature.
4667 SSL_CTX_up_ref(), SSL_up_ref(), X509_up_ref(), EVP_PKEY_up_ref(),
4668 X509_CRL_up_ref(), X509_OBJECT_up_ref_count() methods are now returning an
4669 int (instead of void) like all others TYPE_up_ref() methods.
4670 So now these methods also check the return value of CRYPTO_atomic_add(),
4671 and the validity of object reference counter.
4673 *fdasilvayy@gmail.com*
4675 * With Windows Visual Studio builds, the .pdb files are installed
4676 alongside the installed libraries and executables. For a static
4677 library installation, ossl_static.pdb is the associate compiler
4678 generated .pdb file to be used when linking programs.
4682 * Remove openssl.spec. Packaging files belong with the packagers.
4686 * Automatic Darwin/OSX configuration has had a refresh, it will now
4687 recognise x86_64 architectures automatically. You can still decide
4688 to build for a different bitness with the environment variable
4689 KERNEL_BITS (can be 32 or 64), for example:
4691 KERNEL_BITS=32 ./config
4695 * Change default algorithms in pkcs8 utility to use PKCS#5 v2.0,
4696 256 bit AES and HMAC with SHA256.
4700 * Remove support for MIPS o32 ABI on IRIX (and IRIX only).
4704 * Triple-DES ciphers have been moved from HIGH to MEDIUM.
4708 * To enable users to have their own config files and build file templates,
4709 Configure looks in the directory indicated by the environment variable
4710 OPENSSL_LOCAL_CONFIG_DIR as well as the in-source Configurations/
4711 directory. On VMS, OPENSSL_LOCAL_CONFIG_DIR is expected to be a logical
4712 name and is used as is.
4716 * The following datatypes were made opaque: X509_OBJECT, X509_STORE_CTX,
4717 X509_STORE, X509_LOOKUP, and X509_LOOKUP_METHOD. The unused type
4718 X509_CERT_FILE_CTX was removed.
4722 * "shared" builds are now the default. To create only static libraries use
4723 the "no-shared" Configure option.
4727 * Remove the no-aes, no-hmac, no-rsa, no-sha and no-md5 Configure options.
4728 All of these option have not worked for some while and are fundamental
4733 * Make various cleanup routines no-ops and mark them as deprecated. Most
4734 global cleanup functions are no longer required because they are handled
4735 via auto-deinit (see OPENSSL_init_crypto and OPENSSL_init_ssl man pages).
4736 Explicitly de-initing can cause problems (e.g. where a library that uses
4737 OpenSSL de-inits, but an application is still using it). The affected
4738 functions are CONF_modules_free(), ENGINE_cleanup(), OBJ_cleanup(),
4739 EVP_cleanup(), BIO_sock_cleanup(), CRYPTO_cleanup_all_ex_data(),
4740 RAND_cleanup(), SSL_COMP_free_compression_methods(), ERR_free_strings() and
4741 COMP_zlib_cleanup().
4745 * --strict-warnings no longer enables runtime debugging options
4746 such as REF_DEBUG. Instead, debug options are automatically
4747 enabled with '--debug' builds.
4749 *Andy Polyakov, Emilia Käsper*
4751 * Made DH and DH_METHOD opaque. The structures for managing DH objects
4752 have been moved out of the public header files. New functions for managing
4753 these have been added.
4757 * Made RSA and RSA_METHOD opaque. The structures for managing RSA
4758 objects have been moved out of the public header files. New
4759 functions for managing these have been added.
4763 * Made DSA and DSA_METHOD opaque. The structures for managing DSA objects
4764 have been moved out of the public header files. New functions for managing
4765 these have been added.
4769 * Made BIO and BIO_METHOD opaque. The structures for managing BIOs have been
4770 moved out of the public header files. New functions for managing these
4775 * Removed no-rijndael as a config option. Rijndael is an old name for AES.
4779 * Removed the mk1mf build scripts.
4783 * Headers are now wrapped, if necessary, with OPENSSL_NO_xxx, so
4784 it is always safe to #include a header now.
4788 * Removed the aged BC-32 config and all its supporting scripts
4792 * Removed support for Ultrix, Netware, and OS/2.
4796 * Add support for HKDF.
4798 *Alessandro Ghedini*
4800 * Add support for blake2b and blake2s
4804 * Added support for "pipelining". Ciphers that have the
4805 EVP_CIPH_FLAG_PIPELINE flag set have a capability to process multiple
4806 encryptions/decryptions simultaneously. There are currently no built-in
4807 ciphers with this property but the expectation is that engines will be able
4808 to offer it to significantly improve throughput. Support has been extended
4809 into libssl so that multiple records for a single connection can be
4810 processed in one go (for >=TLS 1.1).
4814 * Added the AFALG engine. This is an async capable engine which is able to
4815 offload work to the Linux kernel. In this initial version it only supports
4816 AES128-CBC. The kernel must be version 4.1.0 or greater.
4820 * OpenSSL now uses a new threading API. It is no longer necessary to
4821 set locking callbacks to use OpenSSL in a multi-threaded environment. There
4822 are two supported threading models: pthreads and windows threads. It is
4823 also possible to configure OpenSSL at compile time for "no-threads". The
4824 old threading API should no longer be used. The functions have been
4825 replaced with "no-op" compatibility macros.
4827 *Alessandro Ghedini, Matt Caswell*
4829 * Modify behavior of ALPN to invoke callback after SNI/servername
4830 callback, such that updates to the SSL_CTX affect ALPN.
4834 * Add SSL_CIPHER queries for authentication and key-exchange.
4838 * Changes to the DEFAULT cipherlist:
4839 - Prefer (EC)DHE handshakes over plain RSA.
4840 - Prefer AEAD ciphers over legacy ciphers.
4841 - Prefer ECDSA over RSA when both certificates are available.
4842 - Prefer TLSv1.2 ciphers/PRF.
4843 - Remove DSS, SEED, IDEA, CAMELLIA, and AES-CCM from the
4848 * Change the ECC default curve list to be this, in order: x25519,
4849 secp256r1, secp521r1, secp384r1.
4853 * RC4 based libssl ciphersuites are now classed as "weak" ciphers and are
4854 disabled by default. They can be re-enabled using the
4855 enable-weak-ssl-ciphers option to Configure.
4859 * If the server has ALPN configured, but supports no protocols that the
4860 client advertises, send a fatal "no_application_protocol" alert.
4861 This behaviour is SHALL in RFC 7301, though it isn't universally
4862 implemented by other servers.
4866 * Add X25519 support.
4867 Add ASN.1 and EVP_PKEY methods for X25519. This includes support
4868 for public and private key encoding using the format documented in
4869 draft-ietf-curdle-pkix-02. The corresponding EVP_PKEY method supports
4870 key generation and key derivation.
4872 TLS support complies with draft-ietf-tls-rfc4492bis-08 and uses
4877 * Deprecate SRP_VBASE_get_by_user.
4878 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
4879 In order to fix an unavoidable memory leak ([CVE-2016-0798]),
4880 SRP_VBASE_get_by_user was changed to ignore the "fake user" SRP
4881 seed, even if the seed is configured.
4883 Users should use SRP_VBASE_get1_by_user instead. Note that in
4884 SRP_VBASE_get1_by_user, caller must free the returned value. Note
4885 also that even though configuring the SRP seed attempts to hide
4886 invalid usernames by continuing the handshake with fake
4887 credentials, this behaviour is not constant time and no strong
4888 guarantees are made that the handshake is indistinguishable from
4889 that of a valid user.
4893 * Configuration change; it's now possible to build dynamic engines
4894 without having to build shared libraries and vice versa. This
4895 only applies to the engines in `engines/`, those in `crypto/engine/`
4896 will always be built into libcrypto (i.e. "static").
4898 Building dynamic engines is enabled by default; to disable, use
4899 the configuration option "disable-dynamic-engine".
4901 The only requirements for building dynamic engines are the
4902 presence of the DSO module and building with position independent
4903 code, so they will also automatically be disabled if configuring
4904 with "disable-dso" or "disable-pic".
4906 The macros OPENSSL_NO_STATIC_ENGINE and OPENSSL_NO_DYNAMIC_ENGINE
4907 are also taken away from openssl/opensslconf.h, as they are
4912 * Configuration change; if there is a known flag to compile
4913 position independent code, it will always be applied on the
4914 libcrypto and libssl object files, and never on the application
4915 object files. This means other libraries that use routines from
4916 libcrypto / libssl can be made into shared libraries regardless
4917 of how OpenSSL was configured.
4919 If this isn't desirable, the configuration options "disable-pic"
4920 or "no-pic" can be used to disable the use of PIC. This will
4921 also disable building shared libraries and dynamic engines.
4925 * Removed JPAKE code. It was experimental and has no wide use.
4929 * The INSTALL_PREFIX Makefile variable has been renamed to
4930 DESTDIR. That makes for less confusion on what this variable
4931 is for. Also, the configuration option --install_prefix is
4936 * Heartbeat for TLS has been removed and is disabled by default
4937 for DTLS; configure with enable-heartbeats. Code that uses the
4938 old #define's might need to be updated.
4940 *Emilia Käsper, Rich Salz*
4942 * Rename REF_CHECK to REF_DEBUG.
4946 * New "unified" build system
4948 The "unified" build system is aimed to be a common system for all
4949 platforms we support. With it comes new support for VMS.
4951 This system builds supports building in a different directory tree
4952 than the source tree. It produces one Makefile (for unix family
4953 or lookalikes), or one descrip.mms (for VMS).
4955 The source of information to make the Makefile / descrip.mms is
4956 small files called 'build.info', holding the necessary
4957 information for each directory with source to compile, and a
4958 template in Configurations, like unix-Makefile.tmpl or
4961 With this change, the library names were also renamed on Windows
4962 and on VMS. They now have names that are closer to the standard
4963 on Unix, and include the major version number, and in certain
4964 cases, the architecture they are built for. See "Notes on shared
4965 libraries" in INSTALL.
4967 We rely heavily on the perl module Text::Template.
4971 * Added support for auto-initialisation and de-initialisation of the library.
4972 OpenSSL no longer requires explicit init or deinit routines to be called,
4973 except in certain circumstances. See the OPENSSL_init_crypto() and
4974 OPENSSL_init_ssl() man pages for further information.
4978 * The arguments to the DTLSv1_listen function have changed. Specifically the
4979 "peer" argument is now expected to be a BIO_ADDR object.
4981 * Rewrite of BIO networking library. The BIO library lacked consistent
4982 support of IPv6, and adding it required some more extensive
4983 modifications. This introduces the BIO_ADDR and BIO_ADDRINFO types,
4984 which hold all types of addresses and chains of address information.
4985 It also introduces a new API, with functions like BIO_socket,
4986 BIO_connect, BIO_listen, BIO_lookup and a rewrite of BIO_accept.
4987 The source/sink BIOs BIO_s_connect, BIO_s_accept and BIO_s_datagram
4988 have been adapted accordingly.
4992 * RSA_padding_check_PKCS1_type_1 now accepts inputs with and without
4997 * CRIME protection: disable compression by default, even if OpenSSL is
4998 compiled with zlib enabled. Applications can still enable compression
4999 by calling SSL_CTX_clear_options(ctx, SSL_OP_NO_COMPRESSION), or by
5000 using the SSL_CONF library to configure compression.
5004 * The signature of the session callback configured with
5005 SSL_CTX_sess_set_get_cb was changed. The read-only input buffer
5006 was explicitly marked as `const unsigned char*` instead of
5011 * Always DPURIFY. Remove the use of uninitialized memory in the
5012 RNG, and other conditional uses of DPURIFY. This makes -DPURIFY a no-op.
5016 * Removed many obsolete configuration items, including
5017 DES_PTR, DES_RISC1, DES_RISC2, DES_INT
5018 MD2_CHAR, MD2_INT, MD2_LONG
5020 IDEA_SHORT, IDEA_LONG
5021 RC2_SHORT, RC2_LONG, RC4_LONG, RC4_CHUNK, RC4_INDEX
5023 *Rich Salz, with advice from Andy Polyakov*
5025 * Many BN internals have been moved to an internal header file.
5027 *Rich Salz with help from Andy Polyakov*
5029 * Configuration and writing out the results from it has changed.
5030 Files such as Makefile include/openssl/opensslconf.h and are now
5031 produced through general templates, such as Makefile.in and
5032 crypto/opensslconf.h.in and some help from the perl module
5035 Also, the center of configuration information is no longer
5036 Makefile. Instead, Configure produces a perl module in
5037 configdata.pm which holds most of the config data (in the hash
5038 table %config), the target data that comes from the target
5039 configuration in one of the `Configurations/*.conf` files (in
5044 * To clarify their intended purposes, the Configure options
5045 --prefix and --openssldir change their semantics, and become more
5046 straightforward and less interdependent.
5048 --prefix shall be used exclusively to give the location INSTALLTOP
5049 where programs, scripts, libraries, include files and manuals are
5050 going to be installed. The default is now /usr/local.
5052 --openssldir shall be used exclusively to give the default
5053 location OPENSSLDIR where certificates, private keys, CRLs are
5054 managed. This is also where the default openssl.cnf gets
5056 If the directory given with this option is a relative path, the
5057 values of both the --prefix value and the --openssldir value will
5058 be combined to become OPENSSLDIR.
5059 The default for --openssldir is INSTALLTOP/ssl.
5061 Anyone who uses --openssldir to specify where OpenSSL is to be
5062 installed MUST change to use --prefix instead.
5066 * The GOST engine was out of date and therefore it has been removed. An up
5067 to date GOST engine is now being maintained in an external repository.
5068 See: <https://wiki.openssl.org/index.php/Binaries>. Libssl still retains
5069 support for GOST ciphersuites (these are only activated if a GOST engine
5074 * EGD is no longer supported by default; use enable-egd when
5077 *Ben Kaduk and Rich Salz*
5079 * The distribution now has Makefile.in files, which are used to
5080 create Makefile's when Configure is run. *Configure must be run
5081 before trying to build now.*
5085 * The return value for SSL_CIPHER_description() for error conditions
5090 * Support for RFC6698/RFC7671 DANE TLSA peer authentication.
5092 Obtaining and performing DNSSEC validation of TLSA records is
5093 the application's responsibility. The application provides
5094 the TLSA records of its choice to OpenSSL, and these are then
5095 used to authenticate the peer.
5097 The TLSA records need not even come from DNS. They can, for
5098 example, be used to implement local end-entity certificate or
5099 trust-anchor "pinning", where the "pin" data takes the form
5100 of TLSA records, which can augment or replace verification
5101 based on the usual WebPKI public certification authorities.
5105 * Revert default OPENSSL_NO_DEPRECATED setting. Instead OpenSSL
5106 continues to support deprecated interfaces in default builds.
5107 However, applications are strongly advised to compile their
5108 source files with -DOPENSSL_API_COMPAT=0x10100000L, which hides
5109 the declarations of all interfaces deprecated in 0.9.8, 1.0.0
5110 or the 1.1.0 releases.
5112 In environments in which all applications have been ported to
5113 not use any deprecated interfaces OpenSSL's Configure script
5114 should be used with the --api=1.1.0 option to entirely remove
5115 support for the deprecated features from the library and
5116 unconditionally disable them in the installed headers.
5117 Essentially the same effect can be achieved with the "no-deprecated"
5118 argument to Configure, except that this will always restrict
5119 the build to just the latest API, rather than a fixed API
5122 As applications are ported to future revisions of the API,
5123 they should update their compile-time OPENSSL_API_COMPAT define
5124 accordingly, but in most cases should be able to continue to
5125 compile with later releases.
5127 The OPENSSL_API_COMPAT versions for 1.0.0, and 0.9.8 are
5128 0x10000000L and 0x00908000L, respectively. However those
5129 versions did not support the OPENSSL_API_COMPAT feature, and
5130 so applications are not typically tested for explicit support
5131 of just the undeprecated features of either release.
5135 * Add support for setting the minimum and maximum supported protocol.
5136 It can bet set via the SSL_set_min_proto_version() and
5137 SSL_set_max_proto_version(), or via the SSL_CONF's MinProtocol and
5138 MaxProtocol. It's recommended to use the new APIs to disable
5139 protocols instead of disabling individual protocols using
5140 SSL_set_options() or SSL_CONF's Protocol. This change also
5141 removes support for disabling TLS 1.2 in the OpenSSL TLS
5142 client at compile time by defining OPENSSL_NO_TLS1_2_CLIENT.
5146 * Support for ChaCha20 and Poly1305 added to libcrypto and libssl.
5150 * New EC_KEY_METHOD, this replaces the older ECDSA_METHOD and ECDH_METHOD
5151 and integrates ECDSA and ECDH functionality into EC. Implementations can
5152 now redirect key generation and no longer need to convert to or from
5155 Note: the ecdsa.h and ecdh.h headers are now no longer needed and just
5156 include the ec.h header file instead.
5160 * Remove support for all 40 and 56 bit ciphers. This includes all the export
5161 ciphers who are no longer supported and drops support the ephemeral RSA key
5162 exchange. The LOW ciphers currently doesn't have any ciphers in it.
5166 * Made EVP_MD_CTX, EVP_MD, EVP_CIPHER_CTX, EVP_CIPHER and HMAC_CTX
5167 opaque. For HMAC_CTX, the following constructors and destructors
5170 HMAC_CTX *HMAC_CTX_new(void);
5171 void HMAC_CTX_free(HMAC_CTX *ctx);
5173 For EVP_MD and EVP_CIPHER, complete APIs to create, fill and
5174 destroy such methods has been added. See EVP_MD_meth_new(3) and
5175 EVP_CIPHER_meth_new(3) for documentation.
5178 1) `EVP_MD_CTX_cleanup()`, `EVP_CIPHER_CTX_cleanup()` and
5179 `HMAC_CTX_cleanup()` were removed. `HMAC_CTX_reset()` and
5180 `EVP_MD_CTX_reset()` should be called instead to reinitialise
5181 an already created structure.
5182 2) For consistency with the majority of our object creators and
5183 destructors, `EVP_MD_CTX_(create|destroy)` were renamed to
5184 `EVP_MD_CTX_(new|free)`. The old names are retained as macros
5185 for deprecated builds.
5189 * Added ASYNC support. Libcrypto now includes the async sub-library to enable
5190 cryptographic operations to be performed asynchronously as long as an
5191 asynchronous capable engine is used. See the ASYNC_start_job() man page for
5192 further details. Libssl has also had this capability integrated with the
5193 introduction of the new mode SSL_MODE_ASYNC and associated error
5194 SSL_ERROR_WANT_ASYNC. See the SSL_CTX_set_mode() and SSL_get_error() man
5195 pages. This work was developed in partnership with Intel Corp.
5199 * SSL_{CTX_}set_ecdh_auto() has been removed and ECDH is support is
5200 always enabled now. If you want to disable the support you should
5201 exclude it using the list of supported ciphers. This also means that the
5202 "-no_ecdhe" option has been removed from s_server.
5206 * SSL_{CTX}_set_tmp_ecdh() which can set 1 EC curve now internally calls
5207 SSL_{CTX_}set1_curves() which can set a list.
5211 * Remove support for SSL_{CTX_}set_tmp_ecdh_callback(). You should set the
5212 curve you want to support using SSL_{CTX_}set1_curves().
5216 * State machine rewrite. The state machine code has been significantly
5217 refactored in order to remove much duplication of code and solve issues
5218 with the old code (see [ssl/statem/README.md](ssl/statem/README.md) for
5219 further details). This change does have some associated API changes.
5220 Notably the SSL_state() function has been removed and replaced by
5221 SSL_get_state which now returns an "OSSL_HANDSHAKE_STATE" instead of an int.
5222 SSL_set_state() has been removed altogether. The previous handshake states
5223 defined in ssl.h and ssl3.h have also been removed.
5227 * All instances of the string "ssleay" in the public API were replaced
5228 with OpenSSL (case-matching; e.g., OPENSSL_VERSION for #define's)
5229 Some error codes related to internal RSA_eay API's were renamed.
5233 * The demo files in crypto/threads were moved to demo/threads.
5237 * Removed obsolete engines: 4758cca, aep, atalla, cswift, nuron, gmp,
5240 *Matt Caswell, Rich Salz*
5242 * New ASN.1 embed macro.
5244 New ASN.1 macro ASN1_EMBED. This is the same as ASN1_SIMPLE except the
5245 structure is not allocated: it is part of the parent. That is instead of
5253 This reduces memory fragmentation and make it impossible to accidentally
5254 set a mandatory field to NULL.
5256 This currently only works for some fields specifically a SEQUENCE, CHOICE,
5257 or ASN1_STRING type which is part of a parent SEQUENCE. Since it is
5258 equivalent to ASN1_SIMPLE it cannot be tagged, OPTIONAL, SET OF or
5263 * Remove EVP_CHECK_DES_KEY, a compile-time option that never compiled.
5267 * Removed DES and RC4 ciphersuites from DEFAULT. Also removed RC2 although
5268 in 1.0.2 EXPORT was already removed and the only RC2 ciphersuite is also
5269 an EXPORT one. COMPLEMENTOFDEFAULT has been updated accordingly to add
5270 DES and RC4 ciphersuites.
5274 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
5275 This changes the decoding behaviour for some invalid messages,
5276 though the change is mostly in the more lenient direction, and
5277 legacy behaviour is preserved as much as possible.
5281 * Fix no-stdio build.
5282 *David Woodhouse <David.Woodhouse@intel.com> and also*
5283 *Ivan Nestlerode <ivan.nestlerode@sonos.com>*
5285 * New testing framework
5286 The testing framework has been largely rewritten and is now using
5287 perl and the perl modules Test::Harness and an extended variant of
5288 Test::More called OpenSSL::Test to do its work. All test scripts in
5289 test/ have been rewritten into test recipes, and all direct calls to
5290 executables in test/Makefile have become individual recipes using the
5291 simplified testing OpenSSL::Test::Simple.
5293 For documentation on our testing modules, do:
5295 perldoc test/testlib/OpenSSL/Test/Simple.pm
5296 perldoc test/testlib/OpenSSL/Test.pm
5300 * Revamped memory debug; only -DCRYPTO_MDEBUG and -DCRYPTO_MDEBUG_ABORT
5301 are used; the latter aborts on memory leaks (usually checked on exit).
5302 Some undocumented "set malloc, etc., hooks" functions were removed
5303 and others were changed. All are now documented.
5307 * In DSA_generate_parameters_ex, if the provided seed is too short,
5310 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
5312 * Rewrite PSK to support ECDHE_PSK, DHE_PSK and RSA_PSK. Add ciphersuites
5313 from RFC4279, RFC4785, RFC5487, RFC5489.
5315 Thanks to Christian J. Dietrich and Giuseppe D'Angelo for the
5316 original RSA_PSK patch.
5320 * Dropped support for the SSL3_FLAGS_DELAY_CLIENT_FINISHED flag. This SSLeay
5321 era flag was never set throughout the codebase (only read). Also removed
5322 SSL3_FLAGS_POP_BUFFER which was only used if
5323 SSL3_FLAGS_DELAY_CLIENT_FINISHED was also set.
5327 * Changed the default name options in the "ca", "crl", "req" and "x509"
5328 to be "oneline" instead of "compat".
5332 * Remove SSL_OP_TLS_BLOCK_PADDING_BUG. This is SSLeay legacy, we're
5333 not aware of clients that still exhibit this bug, and the workaround
5334 hasn't been working properly for a while.
5338 * The return type of BIO_number_read() and BIO_number_written() as well as
5339 the corresponding num_read and num_write members in the BIO structure has
5340 changed from unsigned long to uint64_t. On platforms where an unsigned
5341 long is 32 bits (e.g. Windows) these counters could overflow if >4Gb is
5346 * Given the pervasive nature of TLS extensions it is inadvisable to run
5347 OpenSSL without support for them. It also means that maintaining
5348 the OPENSSL_NO_TLSEXT option within the code is very invasive (and probably
5349 not well tested). Therefore, the OPENSSL_NO_TLSEXT option has been removed.
5353 * Removed support for the two export grade static DH ciphersuites
5354 EXP-DH-RSA-DES-CBC-SHA and EXP-DH-DSS-DES-CBC-SHA. These two ciphersuites
5355 were newly added (along with a number of other static DH ciphersuites) to
5356 1.0.2. However the two export ones have *never* worked since they were
5357 introduced. It seems strange in any case to be adding new export
5358 ciphersuites, and given "logjam" it also does not seem correct to fix them.
5362 * Version negotiation has been rewritten. In particular SSLv23_method(),
5363 SSLv23_client_method() and SSLv23_server_method() have been deprecated,
5364 and turned into macros which simply call the new preferred function names
5365 TLS_method(), TLS_client_method() and TLS_server_method(). All new code
5366 should use the new names instead. Also as part of this change the ssl23.h
5367 header file has been removed.
5371 * Support for Kerberos ciphersuites in TLS (RFC2712) has been removed. This
5372 code and the associated standard is no longer considered fit-for-purpose.
5376 * RT2547 was closed. When generating a private key, try to make the
5377 output file readable only by the owner. This behavior change might
5378 be noticeable when interacting with other software.
5380 * Documented all exdata functions. Added CRYPTO_free_ex_index.
5385 * Added HTTP GET support to the ocsp command.
5389 * Changed default digest for the dgst and enc commands from MD5 to
5394 * RAND_pseudo_bytes has been deprecated. Users should use RAND_bytes instead.
5398 * Added support for TLS extended master secret from
5399 draft-ietf-tls-session-hash-03.txt. Thanks for Alfredo Pironti for an
5400 initial patch which was a great help during development.
5404 * All libssl internal structures have been removed from the public header
5405 files, and the OPENSSL_NO_SSL_INTERN option has been removed (since it is
5406 now redundant). Users should not attempt to access internal structures
5407 directly. Instead they should use the provided API functions.
5411 * config has been changed so that by default OPENSSL_NO_DEPRECATED is used.
5412 Access to deprecated functions can be re-enabled by running config with
5413 "enable-deprecated". In addition applications wishing to use deprecated
5414 functions must define OPENSSL_USE_DEPRECATED. Note that this new behaviour
5415 will, by default, disable some transitive includes that previously existed
5416 in the header files (e.g. ec.h will no longer, by default, include bn.h)
5420 * Added support for OCB mode. OpenSSL has been granted a patent license
5421 compatible with the OpenSSL license for use of OCB. Details are available
5422 at <https://www.openssl.org/source/OCB-patent-grant-OpenSSL.pdf>. Support
5423 for OCB can be removed by calling config with no-ocb.
5427 * SSLv2 support has been removed. It still supports receiving an SSLv2
5428 compatible client hello.
5432 * Increased the minimal RSA keysize from 256 to 512 bits [Rich Salz],
5433 done while fixing the error code for the key-too-small case.
5435 *Annie Yousar <a.yousar@informatik.hu-berlin.de>*
5437 * CA.sh has been removed; use CA.pl instead.
5441 * Removed old DES API.
5445 * Remove various unsupported platforms:
5451 Sinix/ReliantUNIX RM400
5456 16-bit platforms such as WIN16
5460 * Clean up OPENSSL_NO_xxx #define's
5461 - Use setbuf() and remove OPENSSL_NO_SETVBUF_IONBF
5462 - Rename OPENSSL_SYSNAME_xxx to OPENSSL_SYS_xxx
5463 - OPENSSL_NO_EC{DH,DSA} merged into OPENSSL_NO_EC
5464 - OPENSSL_NO_RIPEMD160, OPENSSL_NO_RIPEMD merged into OPENSSL_NO_RMD160
5465 - OPENSSL_NO_FP_API merged into OPENSSL_NO_STDIO
5466 - Remove OPENSSL_NO_BIO OPENSSL_NO_BUFFER OPENSSL_NO_CHAIN_VERIFY
5467 OPENSSL_NO_EVP OPENSSL_NO_FIPS_ERR OPENSSL_NO_HASH_COMP
5468 OPENSSL_NO_LHASH OPENSSL_NO_OBJECT OPENSSL_NO_SPEED OPENSSL_NO_STACK
5469 OPENSSL_NO_X509 OPENSSL_NO_X509_VERIFY
5470 - Remove MS_STATIC; it's a relic from platforms <32 bits.
5474 * Cleaned up dead code
5475 Remove all but one '#ifdef undef' which is to be looked at.
5479 * Clean up calling of xxx_free routines.
5480 Just like free(), fix most of the xxx_free routines to accept
5481 NULL. Remove the non-null checks from callers. Save much code.
5485 * Add secure heap for storage of private keys (when possible).
5486 Add BIO_s_secmem(), CBIGNUM, etc.
5487 Contributed by Akamai Technologies under our Corporate CLA.
5491 * Experimental support for a new, fast, unbiased prime candidate generator,
5492 bn_probable_prime_dh_coprime(). Not currently used by any prime generator.
5494 *Felix Laurie von Massenbach <felix@erbridge.co.uk>*
5496 * New output format NSS in the sess_id command line tool. This allows
5497 exporting the session id and the master key in NSS keylog format.
5499 *Martin Kaiser <martin@kaiser.cx>*
5501 * Harmonize version and its documentation. -f flag is used to display
5504 *mancha <mancha1@zoho.com>*
5506 * Fix eckey_priv_encode so it immediately returns an error upon a failure
5507 in i2d_ECPrivateKey. Thanks to Ted Unangst for feedback on this issue.
5509 *mancha <mancha1@zoho.com>*
5511 * Fix some double frees. These are not thought to be exploitable.
5513 *mancha <mancha1@zoho.com>*
5515 * A missing bounds check in the handling of the TLS heartbeat extension
5516 can be used to reveal up to 64k of memory to a connected client or
5519 Thanks for Neel Mehta of Google Security for discovering this bug and to
5520 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
5521 preparing the fix ([CVE-2014-0160])
5523 *Adam Langley, Bodo Moeller*
5525 * Fix for the attack described in the paper "Recovering OpenSSL
5526 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
5527 by Yuval Yarom and Naomi Benger. Details can be obtained from:
5528 <http://eprint.iacr.org/2014/140>
5530 Thanks to Yuval Yarom and Naomi Benger for discovering this
5531 flaw and to Yuval Yarom for supplying a fix ([CVE-2014-0076])
5533 *Yuval Yarom and Naomi Benger*
5535 * Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
5536 this fixes a limitation in previous versions of OpenSSL.
5540 * Experimental encrypt-then-mac support.
5542 Experimental support for encrypt then mac from
5543 draft-gutmann-tls-encrypt-then-mac-02.txt
5545 To enable it set the appropriate extension number (0x42 for the test
5546 server) using e.g. -DTLSEXT_TYPE_encrypt_then_mac=0x42
5548 For non-compliant peers (i.e. just about everything) this should have no
5551 WARNING: EXPERIMENTAL, SUBJECT TO CHANGE.
5555 * Add EVP support for key wrapping algorithms, to avoid problems with
5556 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
5557 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
5558 algorithms and include tests cases.
5562 * Extend CMS code to support RSA-PSS signatures and RSA-OAEP for
5567 * Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
5568 MGF1 digest and OAEP label.
5572 * Make openssl verify return errors.
5574 *Chris Palmer <palmer@google.com> and Ben Laurie*
5576 * New function ASN1_TIME_diff to calculate the difference between two
5577 ASN1_TIME structures or one structure and the current time.
5581 * Update fips_test_suite to support multiple command line options. New
5582 test to induce all self test errors in sequence and check expected
5587 * Add FIPS_{rsa,dsa,ecdsa}_{sign,verify} functions which digest and
5588 sign or verify all in one operation.
5592 * Add fips_algvs: a multicall fips utility incorporating all the algorithm
5593 test programs and fips_test_suite. Includes functionality to parse
5594 the minimal script output of fipsalgest.pl directly.
5598 * Add authorisation parameter to FIPS_module_mode_set().
5602 * Add FIPS selftest for ECDH algorithm using P-224 and B-233 curves.
5606 * Use separate DRBG fields for internal and external flags. New function
5607 FIPS_drbg_health_check() to perform on demand health checking. Add
5608 generation tests to fips_test_suite with reduced health check interval to
5609 demonstrate periodic health checking. Add "nodh" option to
5610 fips_test_suite to skip very slow DH test.
5614 * New function FIPS_get_cipherbynid() to lookup FIPS supported ciphers
5619 * More extensive health check for DRBG checking many more failure modes.
5620 New function FIPS_selftest_drbg_all() to handle every possible DRBG
5621 combination: call this in fips_test_suite.
5625 * Add support for canonical generation of DSA parameter 'g'. See
5628 * Add support for HMAC DRBG from SP800-90. Update DRBG algorithm test and
5629 POST to handle HMAC cases.
5633 * Add functions FIPS_module_version() and FIPS_module_version_text()
5634 to return numerical and string versions of the FIPS module number.
5638 * Rename FIPS_mode_set and FIPS_mode to FIPS_module_mode_set and
5639 FIPS_module_mode. FIPS_mode and FIPS_mode_set will be implemented
5640 outside the validated module in the FIPS capable OpenSSL.
5644 * Minor change to DRBG entropy callback semantics. In some cases
5645 there is no multiple of the block length between min_len and
5646 max_len. Allow the callback to return more than max_len bytes
5647 of entropy but discard any extra: it is the callback's responsibility
5648 to ensure that the extra data discarded does not impact the
5649 requested amount of entropy.
5653 * Add PRNG security strength checks to RSA, DSA and ECDSA using
5654 information in FIPS186-3, SP800-57 and SP800-131A.
5658 * CCM support via EVP. Interface is very similar to GCM case except we
5659 must supply all data in one chunk (i.e. no update, final) and the
5660 message length must be supplied if AAD is used. Add algorithm test
5665 * Initial version of POST overhaul. Add POST callback to allow the status
5666 of POST to be monitored and/or failures induced. Modify fips_test_suite
5667 to use callback. Always run all selftests even if one fails.
5671 * XTS support including algorithm test driver in the fips_gcmtest program.
5672 Note: this does increase the maximum key length from 32 to 64 bytes but
5673 there should be no binary compatibility issues as existing applications
5674 will never use XTS mode.
5678 * Extensive reorganisation of FIPS PRNG behaviour. Remove all dependencies
5679 to OpenSSL RAND code and replace with a tiny FIPS RAND API which also
5680 performs algorithm blocking for unapproved PRNG types. Also do not
5681 set PRNG type in FIPS_mode_set(): leave this to the application.
5682 Add default OpenSSL DRBG handling: sets up FIPS PRNG and seeds with
5683 the standard OpenSSL PRNG: set additional data to a date time vector.
5687 * Rename old X9.31 PRNG functions of the form `FIPS_rand*` to `FIPS_x931*`.
5688 This shouldn't present any incompatibility problems because applications
5689 shouldn't be using these directly and any that are will need to rethink
5690 anyway as the X9.31 PRNG is now deprecated by FIPS 140-2
5694 * Extensive self tests and health checking required by SP800-90 DRBG.
5695 Remove strength parameter from FIPS_drbg_instantiate and always
5696 instantiate at maximum supported strength.
5700 * Add ECDH code to fips module and fips_ecdhvs for primitives only testing.
5704 * New algorithm test program fips_dhvs to handle DH primitives only testing.
5708 * New function DH_compute_key_padded() to compute a DH key and pad with
5709 leading zeroes if needed: this complies with SP800-56A et al.
5713 * Initial implementation of SP800-90 DRBGs for Hash and CTR. Not used by
5714 anything, incomplete, subject to change and largely untested at present.
5718 * Modify fipscanisteronly build option to only build the necessary object
5719 files by filtering FIPS_EX_OBJ through a perl script in crypto/Makefile.
5723 * Add experimental option FIPSSYMS to give all symbols in
5724 fipscanister.o and FIPS or fips prefix. This will avoid
5725 conflicts with future versions of OpenSSL. Add perl script
5726 util/fipsas.pl to preprocess assembly language source files
5727 and rename any affected symbols.
5731 * Add selftest checks and algorithm block of non-fips algorithms in
5732 FIPS mode. Remove DES2 from selftests.
5736 * Add ECDSA code to fips module. Add tiny fips_ecdsa_check to just
5737 return internal method without any ENGINE dependencies. Add new
5738 tiny fips sign and verify functions.
5742 * New build option no-ec2m to disable characteristic 2 code.
5746 * New build option "fipscanisteronly". This only builds fipscanister.o
5747 and (currently) associated fips utilities. Uses the file Makefile.fips
5748 instead of Makefile.org as the prototype.
5752 * Add some FIPS mode restrictions to GCM. Add internal IV generator.
5753 Update fips_gcmtest to use IV generator.
5757 * Initial, experimental EVP support for AES-GCM. AAD can be input by
5758 setting output buffer to NULL. The `*Final` function must be
5759 called although it will not retrieve any additional data. The tag
5760 can be set or retrieved with a ctrl. The IV length is by default 12
5761 bytes (96 bits) but can be set to an alternative value. If the IV
5762 length exceeds the maximum IV length (currently 16 bytes) it cannot be
5767 * New flag in ciphers: EVP_CIPH_FLAG_CUSTOM_CIPHER. This means the
5768 underlying do_cipher function handles all cipher semantics itself
5769 including padding and finalisation. This is useful if (for example)
5770 an ENGINE cipher handles block padding itself. The behaviour of
5771 do_cipher is subtly changed if this flag is set: the return value
5772 is the number of characters written to the output buffer (zero is
5773 no longer an error code) or a negative error code. Also if the
5774 input buffer is NULL and length 0 finalisation should be performed.
5778 * If a candidate issuer certificate is already part of the constructed
5779 path ignore it: new debug notification X509_V_ERR_PATH_LOOP for this case.
5783 * Improve forward-security support: add functions
5785 void SSL_CTX_set_not_resumable_session_callback(
5786 SSL_CTX *ctx, int (*cb)(SSL *ssl, int is_forward_secure))
5787 void SSL_set_not_resumable_session_callback(
5788 SSL *ssl, int (*cb)(SSL *ssl, int is_forward_secure))
5790 for use by SSL/TLS servers; the callback function will be called whenever a
5791 new session is created, and gets to decide whether the session may be
5792 cached to make it resumable (return 0) or not (return 1). (As by the
5793 SSL/TLS protocol specifications, the session_id sent by the server will be
5794 empty to indicate that the session is not resumable; also, the server will
5795 not generate RFC 4507 (RFC 5077) session tickets.)
5797 A simple reasonable callback implementation is to return is_forward_secure.
5798 This parameter will be set to 1 or 0 depending on the ciphersuite selected
5799 by the SSL/TLS server library, indicating whether it can provide forward
5802 *Emilia Käsper <emilia.kasper@esat.kuleuven.be> (Google)*
5804 * New -verify_name option in command line utilities to set verification
5809 * Initial CMAC implementation. WARNING: EXPERIMENTAL, API MAY CHANGE.
5810 Add CMAC pkey methods.
5814 * Experimental renegotiation in s_server -www mode. If the client
5815 browses /reneg connection is renegotiated. If /renegcert it is
5816 renegotiated requesting a certificate.
5820 * Add an "external" session cache for debugging purposes to s_server. This
5821 should help trace issues which normally are only apparent in deployed
5822 multi-process servers.
5826 * Extensive audit of libcrypto with DEBUG_UNUSED. Fix many cases where
5827 return value is ignored. NB. The functions RAND_add(), RAND_seed(),
5828 BIO_set_cipher() and some obscure PEM functions were changed so they
5829 can now return an error. The RAND changes required a change to the
5830 RAND_METHOD structure.
5834 * New macro `__owur` for "OpenSSL Warn Unused Result". This makes use of
5835 a gcc attribute to warn if the result of a function is ignored. This
5836 is enable if DEBUG_UNUSED is set. Add to several functions in evp.h
5837 whose return value is often ignored.
5841 * New -noct, -requestct, -requirect and -ctlogfile options for s_client.
5842 These allow SCTs (signed certificate timestamps) to be requested and
5843 validated when establishing a connection.
5845 *Rob Percival <robpercival@google.com>*
5850 ### Changes between 1.0.2s and 1.0.2t [10 Sep 2019]
5852 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
5853 used even when parsing explicit parameters, when loading a encoded key
5854 or calling `EC_GROUP_new_from_ecpkparameters()`/
5855 `EC_GROUP_new_from_ecparameters()`.
5856 This prevents bypass of security hardening and performance gains,
5857 especially for curves with specialized EC_METHODs.
5858 By default, if a key encoded with explicit parameters is loaded and later
5859 encoded, the output is still encoded with explicit parameters, even if
5860 internally a "named" EC_GROUP is used for computation.
5864 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
5865 this change, EC_GROUP_set_generator would accept order and/or cofactor as
5866 NULL. After this change, only the cofactor parameter can be NULL. It also
5867 does some minimal sanity checks on the passed order.
5872 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
5873 An attack is simple, if the first CMS_recipientInfo is valid but the
5874 second CMS_recipientInfo is chosen ciphertext. If the second
5875 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
5876 encryption key will be replaced by garbage, and the message cannot be
5877 decoded, but if the RSA decryption fails, the correct encryption key is
5878 used and the recipient will not notice the attack.
5879 As a work around for this potential attack the length of the decrypted
5880 key must be equal to the cipher default key length, in case the
5881 certificate is not given and all recipientInfo are tried out.
5882 The old behaviour can be re-enabled in the CMS code by setting the
5883 CMS_DEBUG_DECRYPT flag.
5888 * Document issue with installation paths in diverse Windows builds
5890 '/usr/local/ssl' is an unsafe prefix for location to install OpenSSL
5891 binaries and run-time config file.
5896 ### Changes between 1.0.2r and 1.0.2s [28 May 2019]
5898 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
5899 This changes the size when using the `genpkey` command when no size is given.
5900 It fixes an omission in earlier changes that changed all RSA, DSA and DH
5901 generation commands to use 2048 bits by default.
5905 * Add FIPS support for Android Arm 64-bit
5907 Support for Android Arm 64-bit was added to the OpenSSL FIPS Object
5908 Module in Version 2.0.10. For some reason, the corresponding target
5909 'android64-aarch64' was missing OpenSSL 1.0.2, whence it could not be
5910 built with FIPS support on Android Arm 64-bit. This omission has been
5913 *Matthias St. Pierre*
5915 ### Changes between 1.0.2q and 1.0.2r [26 Feb 2019]
5917 * 0-byte record padding oracle
5919 If an application encounters a fatal protocol error and then calls
5920 SSL_shutdown() twice (once to send a close_notify, and once to receive one)
5921 then OpenSSL can respond differently to the calling application if a 0 byte
5922 record is received with invalid padding compared to if a 0 byte record is
5923 received with an invalid MAC. If the application then behaves differently
5924 based on that in a way that is detectable to the remote peer, then this
5925 amounts to a padding oracle that could be used to decrypt data.
5927 In order for this to be exploitable "non-stitched" ciphersuites must be in
5928 use. Stitched ciphersuites are optimised implementations of certain
5929 commonly used ciphersuites. Also the application must call SSL_shutdown()
5930 twice even if a protocol error has occurred (applications should not do
5931 this but some do anyway).
5933 This issue was discovered by Juraj Somorovsky, Robert Merget and Nimrod
5934 Aviram, with additional investigation by Steven Collison and Andrew
5935 Hourselt. It was reported to OpenSSL on 10th December 2018.
5940 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
5944 ### Changes between 1.0.2p and 1.0.2q [20 Nov 2018]
5946 * Microarchitecture timing vulnerability in ECC scalar multiplication
5948 OpenSSL ECC scalar multiplication, used in e.g. ECDSA and ECDH, has been
5949 shown to be vulnerable to a microarchitecture timing side channel attack.
5950 An attacker with sufficient access to mount local timing attacks during
5951 ECDSA signature generation could recover the private key.
5953 This issue was reported to OpenSSL on 26th October 2018 by Alejandro
5954 Cabrera Aldaya, Billy Brumley, Sohaib ul Hassan, Cesar Pereida Garcia and
5960 * Timing vulnerability in DSA signature generation
5962 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
5963 timing side channel attack. An attacker could use variations in the signing
5964 algorithm to recover the private key.
5966 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
5971 * Resolve a compatibility issue in EC_GROUP handling with the FIPS Object
5972 Module, accidentally introduced while backporting security fixes from the
5973 development branch and hindering the use of ECC in FIPS mode.
5977 ### Changes between 1.0.2o and 1.0.2p [14 Aug 2018]
5979 * Client DoS due to large DH parameter
5981 During key agreement in a TLS handshake using a DH(E) based ciphersuite a
5982 malicious server can send a very large prime value to the client. This will
5983 cause the client to spend an unreasonably long period of time generating a
5984 key for this prime resulting in a hang until the client has finished. This
5985 could be exploited in a Denial Of Service attack.
5987 This issue was reported to OpenSSL on 5th June 2018 by Guido Vranken
5992 * Cache timing vulnerability in RSA Key Generation
5994 The OpenSSL RSA Key generation algorithm has been shown to be vulnerable to
5995 a cache timing side channel attack. An attacker with sufficient access to
5996 mount cache timing attacks during the RSA key generation process could
5997 recover the private key.
5999 This issue was reported to OpenSSL on 4th April 2018 by Alejandro Cabrera
6000 Aldaya, Billy Brumley, Cesar Pereida Garcia and Luis Manuel Alvarez Tapia.
6005 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
6006 parameter is no longer accepted, as it leads to a corrupt table. NULL
6007 pem_str is reserved for alias entries only.
6011 * Revert blinding in ECDSA sign and instead make problematic addition
6012 length-invariant. Switch even to fixed-length Montgomery multiplication.
6016 * Change generating and checking of primes so that the error rate of not
6017 being prime depends on the intended use based on the size of the input.
6018 For larger primes this will result in more rounds of Miller-Rabin.
6019 The maximal error rate for primes with more than 1080 bits is lowered
6022 *Kurt Roeckx, Annie Yousar*
6024 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
6028 * Add blinding to ECDSA and DSA signatures to protect against side channel
6029 attacks discovered by Keegan Ryan (NCC Group).
6033 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
6034 now allow empty (zero character) pass phrases.
6038 * Certificate time validation (X509_cmp_time) enforces stricter
6039 compliance with RFC 5280. Fractional seconds and timezone offsets
6040 are no longer allowed.
6044 ### Changes between 1.0.2n and 1.0.2o [27 Mar 2018]
6046 * Constructed ASN.1 types with a recursive definition could exceed the stack
6048 Constructed ASN.1 types with a recursive definition (such as can be found
6049 in PKCS7) could eventually exceed the stack given malicious input with
6050 excessive recursion. This could result in a Denial Of Service attack. There
6051 are no such structures used within SSL/TLS that come from untrusted sources
6052 so this is considered safe.
6054 This issue was reported to OpenSSL on 4th January 2018 by the OSS-fuzz
6060 ### Changes between 1.0.2m and 1.0.2n [7 Dec 2017]
6062 * Read/write after SSL object in error state
6064 OpenSSL 1.0.2 (starting from version 1.0.2b) introduced an "error state"
6065 mechanism. The intent was that if a fatal error occurred during a handshake
6066 then OpenSSL would move into the error state and would immediately fail if
6067 you attempted to continue the handshake. This works as designed for the
6068 explicit handshake functions (SSL_do_handshake(), SSL_accept() and
6069 SSL_connect()), however due to a bug it does not work correctly if
6070 SSL_read() or SSL_write() is called directly. In that scenario, if the
6071 handshake fails then a fatal error will be returned in the initial function
6072 call. If SSL_read()/SSL_write() is subsequently called by the application
6073 for the same SSL object then it will succeed and the data is passed without
6074 being decrypted/encrypted directly from the SSL/TLS record layer.
6076 In order to exploit this issue an application bug would have to be present
6077 that resulted in a call to SSL_read()/SSL_write() being issued after having
6078 already received a fatal error.
6080 This issue was reported to OpenSSL by David Benjamin (Google).
6085 * rsaz_1024_mul_avx2 overflow bug on x86_64
6087 There is an overflow bug in the AVX2 Montgomery multiplication procedure
6088 used in exponentiation with 1024-bit moduli. No EC algorithms are affected.
6089 Analysis suggests that attacks against RSA and DSA as a result of this
6090 defect would be very difficult to perform and are not believed likely.
6091 Attacks against DH1024 are considered just feasible, because most of the
6092 work necessary to deduce information about a private key may be performed
6093 offline. The amount of resources required for such an attack would be
6094 significant. However, for an attack on TLS to be meaningful, the server
6095 would have to share the DH1024 private key among multiple clients, which is
6096 no longer an option since CVE-2016-0701.
6098 This only affects processors that support the AVX2 but not ADX extensions
6099 like Intel Haswell (4th generation).
6101 This issue was reported to OpenSSL by David Benjamin (Google). The issue
6102 was originally found via the OSS-Fuzz project.
6107 ### Changes between 1.0.2l and 1.0.2m [2 Nov 2017]
6109 * bn_sqrx8x_internal carry bug on x86_64
6111 There is a carry propagating bug in the x86_64 Montgomery squaring
6112 procedure. No EC algorithms are affected. Analysis suggests that attacks
6113 against RSA and DSA as a result of this defect would be very difficult to
6114 perform and are not believed likely. Attacks against DH are considered just
6115 feasible (although very difficult) because most of the work necessary to
6116 deduce information about a private key may be performed offline. The amount
6117 of resources required for such an attack would be very significant and
6118 likely only accessible to a limited number of attackers. An attacker would
6119 additionally need online access to an unpatched system using the target
6120 private key in a scenario with persistent DH parameters and a private
6121 key that is shared between multiple clients.
6123 This only affects processors that support the BMI1, BMI2 and ADX extensions
6124 like Intel Broadwell (5th generation) and later or AMD Ryzen.
6126 This issue was reported to OpenSSL by the OSS-Fuzz project.
6131 * Malformed X.509 IPAddressFamily could cause OOB read
6133 If an X.509 certificate has a malformed IPAddressFamily extension,
6134 OpenSSL could do a one-byte buffer overread. The most likely result
6135 would be an erroneous display of the certificate in text format.
6137 This issue was reported to OpenSSL by the OSS-Fuzz project.
6141 ### Changes between 1.0.2k and 1.0.2l [25 May 2017]
6143 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
6144 platform rather than 'mingw'.
6148 ### Changes between 1.0.2j and 1.0.2k [26 Jan 2017]
6150 * Truncated packet could crash via OOB read
6152 If one side of an SSL/TLS path is running on a 32-bit host and a specific
6153 cipher is being used, then a truncated packet can cause that host to
6154 perform an out-of-bounds read, usually resulting in a crash.
6156 This issue was reported to OpenSSL by Robert Święcki of Google.
6161 * BN_mod_exp may produce incorrect results on x86_64
6163 There is a carry propagating bug in the x86_64 Montgomery squaring
6164 procedure. No EC algorithms are affected. Analysis suggests that attacks
6165 against RSA and DSA as a result of this defect would be very difficult to
6166 perform and are not believed likely. Attacks against DH are considered just
6167 feasible (although very difficult) because most of the work necessary to
6168 deduce information about a private key may be performed offline. The amount
6169 of resources required for such an attack would be very significant and
6170 likely only accessible to a limited number of attackers. An attacker would
6171 additionally need online access to an unpatched system using the target
6172 private key in a scenario with persistent DH parameters and a private
6173 key that is shared between multiple clients. For example this can occur by
6174 default in OpenSSL DHE based SSL/TLS ciphersuites. Note: This issue is very
6175 similar to CVE-2015-3193 but must be treated as a separate problem.
6177 This issue was reported to OpenSSL by the OSS-Fuzz project.
6182 * Montgomery multiplication may produce incorrect results
6184 There is a carry propagating bug in the Broadwell-specific Montgomery
6185 multiplication procedure that handles input lengths divisible by, but
6186 longer than 256 bits. Analysis suggests that attacks against RSA, DSA
6187 and DH private keys are impossible. This is because the subroutine in
6188 question is not used in operations with the private key itself and an input
6189 of the attacker's direct choice. Otherwise the bug can manifest itself as
6190 transient authentication and key negotiation failures or reproducible
6191 erroneous outcome of public-key operations with specially crafted input.
6192 Among EC algorithms only Brainpool P-512 curves are affected and one
6193 presumably can attack ECDH key negotiation. Impact was not analyzed in
6194 detail, because pre-requisites for attack are considered unlikely. Namely
6195 multiple clients have to choose the curve in question and the server has to
6196 share the private key among them, neither of which is default behaviour.
6197 Even then only clients that chose the curve will be affected.
6199 This issue was publicly reported as transient failures and was not
6200 initially recognized as a security issue. Thanks to Richard Morgan for
6201 providing reproducible case.
6206 * OpenSSL now fails if it receives an unrecognised record type in TLS1.0
6207 or TLS1.1. Previously this only happened in SSLv3 and TLS1.2. This is to
6208 prevent issues where no progress is being made and the peer continually
6209 sends unrecognised record types, using up resources processing them.
6213 ### Changes between 1.0.2i and 1.0.2j [26 Sep 2016]
6215 * Missing CRL sanity check
6217 A bug fix which included a CRL sanity check was added to OpenSSL 1.1.0
6218 but was omitted from OpenSSL 1.0.2i. As a result any attempt to use
6219 CRLs in OpenSSL 1.0.2i will crash with a null pointer exception.
6221 This issue only affects the OpenSSL 1.0.2i
6226 ### Changes between 1.0.2h and 1.0.2i [22 Sep 2016]
6228 * OCSP Status Request extension unbounded memory growth
6230 A malicious client can send an excessively large OCSP Status Request
6231 extension. If that client continually requests renegotiation, sending a
6232 large OCSP Status Request extension each time, then there will be unbounded
6233 memory growth on the server. This will eventually lead to a Denial Of
6234 Service attack through memory exhaustion. Servers with a default
6235 configuration are vulnerable even if they do not support OCSP. Builds using
6236 the "no-ocsp" build time option are not affected.
6238 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6243 * In order to mitigate the SWEET32 attack, the DES ciphers were moved from
6246 This issue was reported to OpenSSL Karthikeyan Bhargavan and Gaetan
6252 * OOB write in MDC2_Update()
6254 An overflow can occur in MDC2_Update() either if called directly or
6255 through the EVP_DigestUpdate() function using MDC2. If an attacker
6256 is able to supply very large amounts of input data after a previous
6257 call to EVP_EncryptUpdate() with a partial block then a length check
6258 can overflow resulting in a heap corruption.
6260 The amount of data needed is comparable to SIZE_MAX which is impractical
6263 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6268 * Malformed SHA512 ticket DoS
6270 If a server uses SHA512 for TLS session ticket HMAC it is vulnerable to a
6271 DoS attack where a malformed ticket will result in an OOB read which will
6274 The use of SHA512 in TLS session tickets is comparatively rare as it requires
6275 a custom server callback and ticket lookup mechanism.
6277 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6282 * OOB write in BN_bn2dec()
6284 The function BN_bn2dec() does not check the return value of BN_div_word().
6285 This can cause an OOB write if an application uses this function with an
6286 overly large BIGNUM. This could be a problem if an overly large certificate
6287 or CRL is printed out from an untrusted source. TLS is not affected because
6288 record limits will reject an oversized certificate before it is parsed.
6290 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6295 * OOB read in TS_OBJ_print_bio()
6297 The function TS_OBJ_print_bio() misuses OBJ_obj2txt(): the return value is
6298 the total length the OID text representation would use and not the amount
6299 of data written. This will result in OOB reads when large OIDs are
6302 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6307 * Pointer arithmetic undefined behaviour
6309 Avoid some undefined pointer arithmetic
6311 A common idiom in the codebase is to check limits in the following manner:
6314 Where "p" points to some malloc'd data of SIZE bytes and
6317 "len" here could be from some externally supplied data (e.g. from a TLS
6320 The rules of C pointer arithmetic are such that "p + len" is only well
6321 defined where len <= SIZE. Therefore the above idiom is actually
6322 undefined behaviour.
6324 For example this could cause problems if some malloc implementation
6325 provides an address for "p" such that "p + len" actually overflows for
6326 values of len that are too big and therefore p + len < limit.
6328 This issue was reported to OpenSSL by Guido Vranken
6333 * Constant time flag not preserved in DSA signing
6335 Operations in the DSA signing algorithm should run in constant time in
6336 order to avoid side channel attacks. A flaw in the OpenSSL DSA
6337 implementation means that a non-constant time codepath is followed for
6338 certain operations. This has been demonstrated through a cache-timing
6339 attack to be sufficient for an attacker to recover the private DSA key.
6341 This issue was reported by César Pereida (Aalto University), Billy Brumley
6342 (Tampere University of Technology), and Yuval Yarom (The University of
6343 Adelaide and NICTA).
6348 * DTLS buffered message DoS
6350 In a DTLS connection where handshake messages are delivered out-of-order
6351 those messages that OpenSSL is not yet ready to process will be buffered
6352 for later use. Under certain circumstances, a flaw in the logic means that
6353 those messages do not get removed from the buffer even though the handshake
6354 has been completed. An attacker could force up to approx. 15 messages to
6355 remain in the buffer when they are no longer required. These messages will
6356 be cleared when the DTLS connection is closed. The default maximum size for
6357 a message is 100k. Therefore, the attacker could force an additional 1500k
6358 to be consumed per connection. By opening many simultaneous connections an
6359 attacker could cause a DoS attack through memory exhaustion.
6361 This issue was reported to OpenSSL by Quan Luo.
6366 * DTLS replay protection DoS
6368 A flaw in the DTLS replay attack protection mechanism means that records
6369 that arrive for future epochs update the replay protection "window" before
6370 the MAC for the record has been validated. This could be exploited by an
6371 attacker by sending a record for the next epoch (which does not have to
6372 decrypt or have a valid MAC), with a very large sequence number. This means
6373 that all subsequent legitimate packets are dropped causing a denial of
6374 service for a specific DTLS connection.
6376 This issue was reported to OpenSSL by the OCAP audit team.
6381 * Certificate message OOB reads
6383 In OpenSSL 1.0.2 and earlier some missing message length checks can result
6384 in OOB reads of up to 2 bytes beyond an allocated buffer. There is a
6385 theoretical DoS risk but this has not been observed in practice on common
6388 The messages affected are client certificate, client certificate request
6389 and server certificate. As a result the attack can only be performed
6390 against a client or a server which enables client authentication.
6392 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6397 ### Changes between 1.0.2g and 1.0.2h [3 May 2016]
6399 * Prevent padding oracle in AES-NI CBC MAC check
6401 A MITM attacker can use a padding oracle attack to decrypt traffic
6402 when the connection uses an AES CBC cipher and the server support
6405 This issue was introduced as part of the fix for Lucky 13 padding
6406 attack ([CVE-2013-0169]). The padding check was rewritten to be in
6407 constant time by making sure that always the same bytes are read and
6408 compared against either the MAC or padding bytes. But it no longer
6409 checked that there was enough data to have both the MAC and padding
6412 This issue was reported by Juraj Somorovsky using TLS-Attacker.
6416 * Fix EVP_EncodeUpdate overflow
6418 An overflow can occur in the EVP_EncodeUpdate() function which is used for
6419 Base64 encoding of binary data. If an attacker is able to supply very large
6420 amounts of input data then a length check can overflow resulting in a heap
6423 Internally to OpenSSL the EVP_EncodeUpdate() function is primarily used by
6424 the `PEM_write_bio*` family of functions. These are mainly used within the
6425 OpenSSL command line applications, so any application which processes data
6426 from an untrusted source and outputs it as a PEM file should be considered
6427 vulnerable to this issue. User applications that call these APIs directly
6428 with large amounts of untrusted data may also be vulnerable.
6430 This issue was reported by Guido Vranken.
6435 * Fix EVP_EncryptUpdate overflow
6437 An overflow can occur in the EVP_EncryptUpdate() function. If an attacker
6438 is able to supply very large amounts of input data after a previous call to
6439 EVP_EncryptUpdate() with a partial block then a length check can overflow
6440 resulting in a heap corruption. Following an analysis of all OpenSSL
6441 internal usage of the EVP_EncryptUpdate() function all usage is one of two
6442 forms. The first form is where the EVP_EncryptUpdate() call is known to be
6443 the first called function after an EVP_EncryptInit(), and therefore that
6444 specific call must be safe. The second form is where the length passed to
6445 EVP_EncryptUpdate() can be seen from the code to be some small value and
6446 therefore there is no possibility of an overflow. Since all instances are
6447 one of these two forms, it is believed that there can be no overflows in
6448 internal code due to this problem. It should be noted that
6449 EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths.
6450 Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances
6451 of these calls have also been analysed too and it is believed there are no
6452 instances in internal usage where an overflow could occur.
6454 This issue was reported by Guido Vranken.
6459 * Prevent ASN.1 BIO excessive memory allocation
6461 When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio()
6462 a short invalid encoding can cause allocation of large amounts of memory
6463 potentially consuming excessive resources or exhausting memory.
6465 Any application parsing untrusted data through d2i BIO functions is
6466 affected. The memory based functions such as d2i_X509() are *not* affected.
6467 Since the memory based functions are used by the TLS library, TLS
6468 applications are not affected.
6470 This issue was reported by Brian Carpenter.
6477 ASN1 Strings that are over 1024 bytes can cause an overread in applications
6478 using the X509_NAME_oneline() function on EBCDIC systems. This could result
6479 in arbitrary stack data being returned in the buffer.
6481 This issue was reported by Guido Vranken.
6486 * Modify behavior of ALPN to invoke callback after SNI/servername
6487 callback, such that updates to the SSL_CTX affect ALPN.
6491 * Remove LOW from the DEFAULT cipher list. This removes singles DES from the
6496 * Only remove the SSLv2 methods with the no-ssl2-method option. When the
6497 methods are enabled and ssl2 is disabled the methods return NULL.
6501 ### Changes between 1.0.2f and 1.0.2g [1 Mar 2016]
6503 * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL.
6504 Builds that are not configured with "enable-weak-ssl-ciphers" will not
6505 provide any "EXPORT" or "LOW" strength ciphers.
6509 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2
6510 is by default disabled at build-time. Builds that are not configured with
6511 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used,
6512 users who want to negotiate SSLv2 via the version-flexible SSLv23_method()
6513 will need to explicitly call either of:
6515 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2);
6517 SSL_clear_options(ssl, SSL_OP_NO_SSLv2);
6519 as appropriate. Even if either of those is used, or the application
6520 explicitly uses the version-specific SSLv2_method() or its client and
6521 server variants, SSLv2 ciphers vulnerable to exhaustive search key
6522 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT
6523 ciphers, and SSLv2 56-bit DES are no longer available.
6528 * Fix a double-free in DSA code
6530 A double free bug was discovered when OpenSSL parses malformed DSA private
6531 keys and could lead to a DoS attack or memory corruption for applications
6532 that receive DSA private keys from untrusted sources. This scenario is
6535 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using
6541 * Disable SRP fake user seed to address a server memory leak.
6543 Add a new method SRP_VBASE_get1_by_user that handles the seed properly.
6545 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
6546 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user
6547 was changed to ignore the "fake user" SRP seed, even if the seed
6550 Users should use SRP_VBASE_get1_by_user instead. Note that in
6551 SRP_VBASE_get1_by_user, caller must free the returned value. Note
6552 also that even though configuring the SRP seed attempts to hide
6553 invalid usernames by continuing the handshake with fake
6554 credentials, this behaviour is not constant time and no strong
6555 guarantees are made that the handshake is indistinguishable from
6556 that of a valid user.
6561 * Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption
6563 In the BN_hex2bn function the number of hex digits is calculated using an
6564 int value `i`. Later `bn_expand` is called with a value of `i * 4`. For
6565 large values of `i` this can result in `bn_expand` not allocating any
6566 memory because `i * 4` is negative. This can leave the internal BIGNUM data
6567 field as NULL leading to a subsequent NULL ptr deref. For very large values
6568 of `i`, the calculation `i * 4` could be a positive value smaller than `i`.
6569 In this case memory is allocated to the internal BIGNUM data field, but it
6570 is insufficiently sized leading to heap corruption. A similar issue exists
6571 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn
6572 is ever called by user applications with very large untrusted hex/dec data.
6573 This is anticipated to be a rare occurrence.
6575 All OpenSSL internal usage of these functions use data that is not expected
6576 to be untrusted, e.g. config file data or application command line
6577 arguments. If user developed applications generate config file data based
6578 on untrusted data then it is possible that this could also lead to security
6579 consequences. This is also anticipated to be rare.
6581 This issue was reported to OpenSSL by Guido Vranken.
6586 * Fix memory issues in `BIO_*printf` functions
6588 The internal `fmtstr` function used in processing a "%s" format string in
6589 the `BIO_*printf` functions could overflow while calculating the length of a
6590 string and cause an OOB read when printing very long strings.
6592 Additionally the internal `doapr_outch` function can attempt to write to an
6593 OOB memory location (at an offset from the NULL pointer) in the event of a
6594 memory allocation failure. In 1.0.2 and below this could be caused where
6595 the size of a buffer to be allocated is greater than INT_MAX. E.g. this
6596 could be in processing a very long "%s" format string. Memory leaks can
6599 The first issue may mask the second issue dependent on compiler behaviour.
6600 These problems could enable attacks where large amounts of untrusted data
6601 is passed to the `BIO_*printf` functions. If applications use these functions
6602 in this way then they could be vulnerable. OpenSSL itself uses these
6603 functions when printing out human-readable dumps of ASN.1 data. Therefore
6604 applications that print this data could be vulnerable if the data is from
6605 untrusted sources. OpenSSL command line applications could also be
6606 vulnerable where they print out ASN.1 data, or if untrusted data is passed
6607 as command line arguments.
6609 Libssl is not considered directly vulnerable. Additionally certificates etc
6610 received via remote connections via libssl are also unlikely to be able to
6611 trigger these issues because of message size limits enforced within libssl.
6613 This issue was reported to OpenSSL Guido Vranken.
6618 * Side channel attack on modular exponentiation
6620 A side-channel attack was found which makes use of cache-bank conflicts on
6621 the Intel Sandy-Bridge microarchitecture which could lead to the recovery
6622 of RSA keys. The ability to exploit this issue is limited as it relies on
6623 an attacker who has control of code in a thread running on the same
6624 hyper-threaded core as the victim thread which is performing decryptions.
6626 This issue was reported to OpenSSL by Yuval Yarom, The University of
6627 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and
6628 Nadia Heninger, University of Pennsylvania with more information at
6629 <http://cachebleed.info>.
6634 * Change the `req` command to generate a 2048-bit RSA/DSA key by default,
6635 if no keysize is specified with default_bits. This fixes an
6636 omission in an earlier change that changed all RSA/DSA key generation
6637 commands to use 2048 bits by default.
6641 ### Changes between 1.0.2e and 1.0.2f [28 Jan 2016]
6643 * DH small subgroups
6645 Historically OpenSSL only ever generated DH parameters based on "safe"
6646 primes. More recently (in version 1.0.2) support was provided for
6647 generating X9.42 style parameter files such as those required for RFC 5114
6648 support. The primes used in such files may not be "safe". Where an
6649 application is using DH configured with parameters based on primes that are
6650 not "safe" then an attacker could use this fact to find a peer's private
6651 DH exponent. This attack requires that the attacker complete multiple
6652 handshakes in which the peer uses the same private DH exponent. For example
6653 this could be used to discover a TLS server's private DH exponent if it's
6654 reusing the private DH exponent or it's using a static DH ciphersuite.
6656 OpenSSL provides the option SSL_OP_SINGLE_DH_USE for ephemeral DH (DHE) in
6657 TLS. It is not on by default. If the option is not set then the server
6658 reuses the same private DH exponent for the life of the server process and
6659 would be vulnerable to this attack. It is believed that many popular
6660 applications do set this option and would therefore not be at risk.
6662 The fix for this issue adds an additional check where a "q" parameter is
6663 available (as is the case in X9.42 based parameters). This detects the
6664 only known attack, and is the only possible defense for static DH
6665 ciphersuites. This could have some performance impact.
6667 Additionally the SSL_OP_SINGLE_DH_USE option has been switched on by
6668 default and cannot be disabled. This could have some performance impact.
6670 This issue was reported to OpenSSL by Antonio Sanso (Adobe).
6675 * SSLv2 doesn't block disabled ciphers
6677 A malicious client can negotiate SSLv2 ciphers that have been disabled on
6678 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
6679 been disabled, provided that the SSLv2 protocol was not also disabled via
6682 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
6683 and Sebastian Schinzel.
6688 ### Changes between 1.0.2d and 1.0.2e [3 Dec 2015]
6690 * BN_mod_exp may produce incorrect results on x86_64
6692 There is a carry propagating bug in the x86_64 Montgomery squaring
6693 procedure. No EC algorithms are affected. Analysis suggests that attacks
6694 against RSA and DSA as a result of this defect would be very difficult to
6695 perform and are not believed likely. Attacks against DH are considered just
6696 feasible (although very difficult) because most of the work necessary to
6697 deduce information about a private key may be performed offline. The amount
6698 of resources required for such an attack would be very significant and
6699 likely only accessible to a limited number of attackers. An attacker would
6700 additionally need online access to an unpatched system using the target
6701 private key in a scenario with persistent DH parameters and a private
6702 key that is shared between multiple clients. For example this can occur by
6703 default in OpenSSL DHE based SSL/TLS ciphersuites.
6705 This issue was reported to OpenSSL by Hanno Böck.
6710 * Certificate verify crash with missing PSS parameter
6712 The signature verification routines will crash with a NULL pointer
6713 dereference if presented with an ASN.1 signature using the RSA PSS
6714 algorithm and absent mask generation function parameter. Since these
6715 routines are used to verify certificate signature algorithms this can be
6716 used to crash any certificate verification operation and exploited in a
6717 DoS attack. Any application which performs certificate verification is
6718 vulnerable including OpenSSL clients and servers which enable client
6721 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
6726 * X509_ATTRIBUTE memory leak
6728 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
6729 memory. This structure is used by the PKCS#7 and CMS routines so any
6730 application which reads PKCS#7 or CMS data from untrusted sources is
6731 affected. SSL/TLS is not affected.
6733 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
6739 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
6740 This changes the decoding behaviour for some invalid messages,
6741 though the change is mostly in the more lenient direction, and
6742 legacy behaviour is preserved as much as possible.
6746 * In DSA_generate_parameters_ex, if the provided seed is too short,
6749 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
6751 ### Changes between 1.0.2c and 1.0.2d [9 Jul 2015]
6753 * Alternate chains certificate forgery
6755 During certificate verification, OpenSSL will attempt to find an
6756 alternative certificate chain if the first attempt to build such a chain
6757 fails. An error in the implementation of this logic can mean that an
6758 attacker could cause certain checks on untrusted certificates to be
6759 bypassed, such as the CA flag, enabling them to use a valid leaf
6760 certificate to act as a CA and "issue" an invalid certificate.
6762 This issue was reported to OpenSSL by Adam Langley/David Benjamin
6767 ### Changes between 1.0.2b and 1.0.2c [12 Jun 2015]
6769 * Fix HMAC ABI incompatibility. The previous version introduced an ABI
6770 incompatibility in the handling of HMAC. The previous ABI has now been
6775 ### Changes between 1.0.2a and 1.0.2b [11 Jun 2015]
6777 * Malformed ECParameters causes infinite loop
6779 When processing an ECParameters structure OpenSSL enters an infinite loop
6780 if the curve specified is over a specially malformed binary polynomial
6783 This can be used to perform denial of service against any
6784 system which processes public keys, certificate requests or
6785 certificates. This includes TLS clients and TLS servers with
6786 client authentication enabled.
6788 This issue was reported to OpenSSL by Joseph Barr-Pixton.
6793 * Exploitable out-of-bounds read in X509_cmp_time
6795 X509_cmp_time does not properly check the length of the ASN1_TIME
6796 string and can read a few bytes out of bounds. In addition,
6797 X509_cmp_time accepts an arbitrary number of fractional seconds in the
6800 An attacker can use this to craft malformed certificates and CRLs of
6801 various sizes and potentially cause a segmentation fault, resulting in
6802 a DoS on applications that verify certificates or CRLs. TLS clients
6803 that verify CRLs are affected. TLS clients and servers with client
6804 authentication enabled may be affected if they use custom verification
6807 This issue was reported to OpenSSL by Robert Swiecki (Google), and
6808 independently by Hanno Böck.
6813 * PKCS7 crash with missing EnvelopedContent
6815 The PKCS#7 parsing code does not handle missing inner EncryptedContent
6816 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
6817 with missing content and trigger a NULL pointer dereference on parsing.
6819 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
6820 structures from untrusted sources are affected. OpenSSL clients and
6821 servers are not affected.
6823 This issue was reported to OpenSSL by Michal Zalewski (Google).
6828 * CMS verify infinite loop with unknown hash function
6830 When verifying a signedData message the CMS code can enter an infinite loop
6831 if presented with an unknown hash function OID. This can be used to perform
6832 denial of service against any system which verifies signedData messages using
6834 This issue was reported to OpenSSL by Johannes Bauer.
6839 * Race condition handling NewSessionTicket
6841 If a NewSessionTicket is received by a multi-threaded client when attempting to
6842 reuse a previous ticket then a race condition can occur potentially leading to
6843 a double free of the ticket data.
6848 * Only support 256-bit or stronger elliptic curves with the
6849 'ecdh_auto' setting (server) or by default (client). Of supported
6850 curves, prefer P-256 (both).
6854 ### Changes between 1.0.2 and 1.0.2a [19 Mar 2015]
6856 * ClientHello sigalgs DoS fix
6858 If a client connects to an OpenSSL 1.0.2 server and renegotiates with an
6859 invalid signature algorithms extension a NULL pointer dereference will
6860 occur. This can be exploited in a DoS attack against the server.
6862 This issue was was reported to OpenSSL by David Ramos of Stanford
6866 *Stephen Henson and Matt Caswell*
6868 * Multiblock corrupted pointer fix
6870 OpenSSL 1.0.2 introduced the "multiblock" performance improvement. This
6871 feature only applies on 64 bit x86 architecture platforms that support AES
6872 NI instructions. A defect in the implementation of "multiblock" can cause
6873 OpenSSL's internal write buffer to become incorrectly set to NULL when
6874 using non-blocking IO. Typically, when the user application is using a
6875 socket BIO for writing, this will only result in a failed connection.
6876 However if some other BIO is used then it is likely that a segmentation
6877 fault will be triggered, thus enabling a potential DoS attack.
6879 This issue was reported to OpenSSL by Daniel Danner and Rainer Mueller.
6884 * Segmentation fault in DTLSv1_listen fix
6886 The DTLSv1_listen function is intended to be stateless and processes the
6887 initial ClientHello from many peers. It is common for user code to loop
6888 over the call to DTLSv1_listen until a valid ClientHello is received with
6889 an associated cookie. A defect in the implementation of DTLSv1_listen means
6890 that state is preserved in the SSL object from one invocation to the next
6891 that can lead to a segmentation fault. Errors processing the initial
6892 ClientHello can trigger this scenario. An example of such an error could be
6893 that a DTLS1.0 only client is attempting to connect to a DTLS1.2 only
6896 This issue was reported to OpenSSL by Per Allansson.
6901 * Segmentation fault in ASN1_TYPE_cmp fix
6903 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
6904 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
6905 certificate signature algorithm consistency this can be used to crash any
6906 certificate verification operation and exploited in a DoS attack. Any
6907 application which performs certificate verification is vulnerable including
6908 OpenSSL clients and servers which enable client authentication.
6913 * Segmentation fault for invalid PSS parameters fix
6915 The signature verification routines will crash with a NULL pointer
6916 dereference if presented with an ASN.1 signature using the RSA PSS
6917 algorithm and invalid parameters. Since these routines are used to verify
6918 certificate signature algorithms this can be used to crash any
6919 certificate verification operation and exploited in a DoS attack. Any
6920 application which performs certificate verification is vulnerable including
6921 OpenSSL clients and servers which enable client authentication.
6923 This issue was was reported to OpenSSL by Brian Carpenter.
6928 * ASN.1 structure reuse memory corruption fix
6930 Reusing a structure in ASN.1 parsing may allow an attacker to cause
6931 memory corruption via an invalid write. Such reuse is and has been
6932 strongly discouraged and is believed to be rare.
6934 Applications that parse structures containing CHOICE or ANY DEFINED BY
6935 components may be affected. Certificate parsing (d2i_X509 and related
6936 functions) are however not affected. OpenSSL clients and servers are
6942 * PKCS7 NULL pointer dereferences fix
6944 The PKCS#7 parsing code does not handle missing outer ContentInfo
6945 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
6946 missing content and trigger a NULL pointer dereference on parsing.
6948 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
6949 otherwise parse PKCS#7 structures from untrusted sources are
6950 affected. OpenSSL clients and servers are not affected.
6952 This issue was reported to OpenSSL by Michal Zalewski (Google).
6957 * DoS via reachable assert in SSLv2 servers fix
6959 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
6960 servers that both support SSLv2 and enable export cipher suites by sending
6961 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
6963 This issue was discovered by Sean Burford (Google) and Emilia Käsper
6964 (OpenSSL development team).
6969 * Empty CKE with client auth and DHE fix
6971 If client auth is used then a server can seg fault in the event of a DHE
6972 ciphersuite being selected and a zero length ClientKeyExchange message
6973 being sent by the client. This could be exploited in a DoS attack.
6978 * Handshake with unseeded PRNG fix
6980 Under certain conditions an OpenSSL 1.0.2 client can complete a handshake
6981 with an unseeded PRNG. The conditions are:
6982 - The client is on a platform where the PRNG has not been seeded
6983 automatically, and the user has not seeded manually
6984 - A protocol specific client method version has been used (i.e. not
6985 SSL_client_methodv23)
6986 - A ciphersuite is used that does not require additional random data from
6987 the PRNG beyond the initial ClientHello client random (e.g. PSK-RC4-SHA).
6989 If the handshake succeeds then the client random that has been used will
6990 have been generated from a PRNG with insufficient entropy and therefore the
6991 output may be predictable.
6993 For example using the following command with an unseeded openssl will
6994 succeed on an unpatched platform:
6996 openssl s_client -psk 1a2b3c4d -tls1_2 -cipher PSK-RC4-SHA
7001 * Use After Free following d2i_ECPrivatekey error fix
7003 A malformed EC private key file consumed via the d2i_ECPrivateKey function
7004 could cause a use after free condition. This, in turn, could cause a double
7005 free in several private key parsing functions (such as d2i_PrivateKey
7006 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
7007 for applications that receive EC private keys from untrusted
7008 sources. This scenario is considered rare.
7010 This issue was discovered by the BoringSSL project and fixed in their
7016 * X509_to_X509_REQ NULL pointer deref fix
7018 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
7019 the certificate key is invalid. This function is rarely used in practice.
7021 This issue was discovered by Brian Carpenter.
7026 * Removed the export ciphers from the DEFAULT ciphers
7030 ### Changes between 1.0.1l and 1.0.2 [22 Jan 2015]
7032 * Facilitate "universal" ARM builds targeting range of ARM ISAs, e.g.
7033 ARMv5 through ARMv8, as opposite to "locking" it to single one.
7034 So far those who have to target multiple platforms would compromise
7035 and argue that binary targeting say ARMv5 would still execute on
7036 ARMv8. "Universal" build resolves this compromise by providing
7037 near-optimal performance even on newer platforms.
7041 * Accelerated NIST P-256 elliptic curve implementation for x86_64
7042 (other platforms pending).
7044 *Shay Gueron & Vlad Krasnov (Intel Corp), Andy Polyakov*
7046 * Add support for the SignedCertificateTimestampList certificate and
7047 OCSP response extensions from RFC6962.
7051 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
7052 for corner cases. (Certain input points at infinity could lead to
7053 bogus results, with non-infinity inputs mapped to infinity too.)
7057 * Initial support for PowerISA 2.0.7, first implemented in POWER8.
7058 This covers AES, SHA256/512 and GHASH. "Initial" means that most
7059 common cases are optimized and there still is room for further
7060 improvements. Vector Permutation AES for Altivec is also added.
7064 * Add support for little-endian ppc64 Linux target.
7066 *Marcelo Cerri (IBM)*
7068 * Initial support for AMRv8 ISA crypto extensions. This covers AES,
7069 SHA1, SHA256 and GHASH. "Initial" means that most common cases
7070 are optimized and there still is room for further improvements.
7071 Both 32- and 64-bit modes are supported.
7073 *Andy Polyakov, Ard Biesheuvel (Linaro)*
7075 * Improved ARMv7 NEON support.
7079 * Support for SPARC Architecture 2011 crypto extensions, first
7080 implemented in SPARC T4. This covers AES, DES, Camellia, SHA1,
7081 SHA256/512, MD5, GHASH and modular exponentiation.
7083 *Andy Polyakov, David Miller*
7085 * Accelerated modular exponentiation for Intel processors, a.k.a.
7088 *Shay Gueron & Vlad Krasnov (Intel Corp)*
7090 * Support for new and upcoming Intel processors, including AVX2,
7091 BMI and SHA ISA extensions. This includes additional "stitched"
7092 implementations, AESNI-SHA256 and GCM, and multi-buffer support
7095 This work was sponsored by Intel Corp.
7099 * Support for DTLS 1.2. This adds two sets of DTLS methods: DTLS_*_method()
7100 supports both DTLS 1.2 and 1.0 and should use whatever version the peer
7101 supports and DTLSv1_2_*_method() which supports DTLS 1.2 only.
7105 * Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
7106 this fixes a limitation in previous versions of OpenSSL.
7110 * Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
7111 MGF1 digest and OAEP label.
7115 * Add EVP support for key wrapping algorithms, to avoid problems with
7116 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
7117 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
7118 algorithms and include tests cases.
7122 * Add functions to allocate and set the fields of an ECDSA_METHOD
7125 *Douglas E. Engert, Steve Henson*
7127 * New functions OPENSSL_gmtime_diff and ASN1_TIME_diff to find the
7128 difference in days and seconds between two tm or ASN1_TIME structures.
7132 * Add -rev test option to s_server to just reverse order of characters
7133 received by client and send back to server. Also prints an abbreviated
7134 summary of the connection parameters.
7138 * New option -brief for s_client and s_server to print out a brief summary
7139 of connection parameters.
7143 * Add callbacks for arbitrary TLS extensions.
7145 *Trevor Perrin <trevp@trevp.net> and Ben Laurie*
7147 * New option -crl_download in several openssl utilities to download CRLs
7148 from CRLDP extension in certificates.
7152 * New options -CRL and -CRLform for s_client and s_server for CRLs.
7156 * New function X509_CRL_diff to generate a delta CRL from the difference
7157 of two full CRLs. Add support to "crl" utility.
7161 * New functions to set lookup_crls function and to retrieve
7162 X509_STORE from X509_STORE_CTX.
7166 * Print out deprecated issuer and subject unique ID fields in
7171 * Extend OCSP I/O functions so they can be used for simple general purpose
7172 HTTP as well as OCSP. New wrapper function which can be used to download
7173 CRLs using the OCSP API.
7177 * Delegate command line handling in s_client/s_server to SSL_CONF APIs.
7181 * `SSL_CONF*` functions. These provide a common framework for application
7182 configuration using configuration files or command lines.
7186 * SSL/TLS tracing code. This parses out SSL/TLS records using the
7187 message callback and prints the results. Needs compile time option
7188 "enable-ssl-trace". New options to s_client and s_server to enable
7193 * New ctrl and macro to retrieve supported points extensions.
7194 Print out extension in s_server and s_client.
7198 * New functions to retrieve certificate signature and signature
7203 * Add functions to retrieve and manipulate the raw cipherlist sent by a
7208 * New Suite B modes for TLS code. These use and enforce the requirements
7209 of RFC6460: restrict ciphersuites, only permit Suite B algorithms and
7210 only use Suite B curves. The Suite B modes can be set by using the
7211 strings "SUITEB128", "SUITEB192" or "SUITEB128ONLY" for the cipherstring.
7215 * New chain verification flags for Suite B levels of security. Check
7216 algorithms are acceptable when flags are set in X509_verify_cert.
7220 * Make tls1_check_chain return a set of flags indicating checks passed
7221 by a certificate chain. Add additional tests to handle client
7222 certificates: checks for matching certificate type and issuer name
7227 * If an attempt is made to use a signature algorithm not in the peer
7228 preference list abort the handshake. If client has no suitable
7229 signature algorithms in response to a certificate request do not
7230 use the certificate.
7234 * If server EC tmp key is not in client preference list abort handshake.
7238 * Add support for certificate stores in CERT structure. This makes it
7239 possible to have different stores per SSL structure or one store in
7240 the parent SSL_CTX. Include distinct stores for certificate chain
7241 verification and chain building. New ctrl SSL_CTRL_BUILD_CERT_CHAIN
7242 to build and store a certificate chain in CERT structure: returning
7243 an error if the chain cannot be built: this will allow applications
7244 to test if a chain is correctly configured.
7246 Note: if the CERT based stores are not set then the parent SSL_CTX
7247 store is used to retain compatibility with existing behaviour.
7251 * New function ssl_set_client_disabled to set a ciphersuite disabled
7252 mask based on the current session, check mask when sending client
7253 hello and checking the requested ciphersuite.
7257 * New ctrls to retrieve and set certificate types in a certificate
7258 request message. Print out received values in s_client. If certificate
7259 types is not set with custom values set sensible values based on
7260 supported signature algorithms.
7264 * Support for distinct client and server supported signature algorithms.
7268 * Add certificate callback. If set this is called whenever a certificate
7269 is required by client or server. An application can decide which
7270 certificate chain to present based on arbitrary criteria: for example
7271 supported signature algorithms. Add very simple example to s_server.
7272 This fixes many of the problems and restrictions of the existing client
7273 certificate callback: for example you can now clear an existing
7274 certificate and specify the whole chain.
7278 * Add new "valid_flags" field to CERT_PKEY structure which determines what
7279 the certificate can be used for (if anything). Set valid_flags field
7280 in new tls1_check_chain function. Simplify ssl_set_cert_masks which used
7281 to have similar checks in it.
7283 Add new "cert_flags" field to CERT structure and include a "strict mode".
7284 This enforces some TLS certificate requirements (such as only permitting
7285 certificate signature algorithms contained in the supported algorithms
7286 extension) which some implementations ignore: this option should be used
7287 with caution as it could cause interoperability issues.
7291 * Update and tidy signature algorithm extension processing. Work out
7292 shared signature algorithms based on preferences and peer algorithms
7293 and print them out in s_client and s_server. Abort handshake if no
7294 shared signature algorithms.
7298 * Add new functions to allow customised supported signature algorithms
7299 for SSL and SSL_CTX structures. Add options to s_client and s_server
7304 * New function SSL_certs_clear() to delete all references to certificates
7305 from an SSL structure. Before this once a certificate had been added
7306 it couldn't be removed.
7310 * Integrate hostname, email address and IP address checking with certificate
7311 verification. New verify options supporting checking in openssl utility.
7315 * Fixes and wildcard matching support to hostname and email checking
7316 functions. Add manual page.
7318 *Florian Weimer (Red Hat Product Security Team)*
7320 * New functions to check a hostname email or IP address against a
7321 certificate. Add options x509 utility to print results of checks against
7326 * Fix OCSP checking.
7328 *Rob Stradling <rob.stradling@comodo.com> and Ben Laurie*
7330 * Initial experimental support for explicitly trusted non-root CAs.
7331 OpenSSL still tries to build a complete chain to a root but if an
7332 intermediate CA has a trust setting included that is used. The first
7333 setting is used: whether to trust (e.g., -addtrust option to the x509
7338 * Add -trusted_first option which attempts to find certificates in the
7339 trusted store even if an untrusted chain is also supplied.
7343 * MIPS assembly pack updates: support for MIPS32r2 and SmartMIPS ASE,
7344 platform support for Linux and Android.
7348 * Support for linux-x32, ILP32 environment in x86_64 framework.
7352 * Experimental multi-implementation support for FIPS capable OpenSSL.
7353 When in FIPS mode the approved implementations are used as normal,
7354 when not in FIPS mode the internal unapproved versions are used instead.
7355 This means that the FIPS capable OpenSSL isn't forced to use the
7356 (often lower performance) FIPS implementations outside FIPS mode.
7360 * Transparently support X9.42 DH parameters when calling
7361 PEM_read_bio_DHparameters. This means existing applications can handle
7362 the new parameter format automatically.
7366 * Initial experimental support for X9.42 DH parameter format: mainly
7367 to support use of 'q' parameter for RFC5114 parameters.
7371 * Add DH parameters from RFC5114 including test data to dhtest.
7375 * Support for automatic EC temporary key parameter selection. If enabled
7376 the most preferred EC parameters are automatically used instead of
7377 hardcoded fixed parameters. Now a server just has to call:
7378 SSL_CTX_set_ecdh_auto(ctx, 1) and the server will automatically
7379 support ECDH and use the most appropriate parameters.
7383 * Enhance and tidy EC curve and point format TLS extension code. Use
7384 static structures instead of allocation if default values are used.
7385 New ctrls to set curves we wish to support and to retrieve shared curves.
7386 Print out shared curves in s_server. New options to s_server and s_client
7387 to set list of supported curves.
7391 * New ctrls to retrieve supported signature algorithms and
7392 supported curve values as an array of NIDs. Extend openssl utility
7393 to print out received values.
7397 * Add new APIs EC_curve_nist2nid and EC_curve_nid2nist which convert
7398 between NIDs and the more common NIST names such as "P-256". Enhance
7399 ecparam utility and ECC method to recognise the NIST names for curves.
7403 * Enhance SSL/TLS certificate chain handling to support different
7404 chains for each certificate instead of one chain in the parent SSL_CTX.
7408 * Support for fixed DH ciphersuite client authentication: where both
7409 server and client use DH certificates with common parameters.
7413 * Support for fixed DH ciphersuites: those requiring DH server
7418 * New function i2d_re_X509_tbs for re-encoding the TBS portion of
7420 Note: Related 1.0.2-beta specific macros X509_get_cert_info,
7421 X509_CINF_set_modified, X509_CINF_get_issuer, X509_CINF_get_extensions and
7422 X509_CINF_get_signature were reverted post internal team review.
7427 ### Changes between 1.0.1t and 1.0.1u [22 Sep 2016]
7429 * OCSP Status Request extension unbounded memory growth
7431 A malicious client can send an excessively large OCSP Status Request
7432 extension. If that client continually requests renegotiation, sending a
7433 large OCSP Status Request extension each time, then there will be unbounded
7434 memory growth on the server. This will eventually lead to a Denial Of
7435 Service attack through memory exhaustion. Servers with a default
7436 configuration are vulnerable even if they do not support OCSP. Builds using
7437 the "no-ocsp" build time option are not affected.
7439 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
7444 * In order to mitigate the SWEET32 attack, the DES ciphers were moved from
7447 This issue was reported to OpenSSL Karthikeyan Bhargavan and Gaetan
7453 * OOB write in MDC2_Update()
7455 An overflow can occur in MDC2_Update() either if called directly or
7456 through the EVP_DigestUpdate() function using MDC2. If an attacker
7457 is able to supply very large amounts of input data after a previous
7458 call to EVP_EncryptUpdate() with a partial block then a length check
7459 can overflow resulting in a heap corruption.
7461 The amount of data needed is comparable to SIZE_MAX which is impractical
7464 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
7469 * Malformed SHA512 ticket DoS
7471 If a server uses SHA512 for TLS session ticket HMAC it is vulnerable to a
7472 DoS attack where a malformed ticket will result in an OOB read which will
7475 The use of SHA512 in TLS session tickets is comparatively rare as it requires
7476 a custom server callback and ticket lookup mechanism.
7478 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
7483 * OOB write in BN_bn2dec()
7485 The function BN_bn2dec() does not check the return value of BN_div_word().
7486 This can cause an OOB write if an application uses this function with an
7487 overly large BIGNUM. This could be a problem if an overly large certificate
7488 or CRL is printed out from an untrusted source. TLS is not affected because
7489 record limits will reject an oversized certificate before it is parsed.
7491 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
7496 * OOB read in TS_OBJ_print_bio()
7498 The function TS_OBJ_print_bio() misuses OBJ_obj2txt(): the return value is
7499 the total length the OID text representation would use and not the amount
7500 of data written. This will result in OOB reads when large OIDs are
7503 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
7508 * Pointer arithmetic undefined behaviour
7510 Avoid some undefined pointer arithmetic
7512 A common idiom in the codebase is to check limits in the following manner:
7515 Where "p" points to some malloc'd data of SIZE bytes and
7518 "len" here could be from some externally supplied data (e.g. from a TLS
7521 The rules of C pointer arithmetic are such that "p + len" is only well
7522 defined where len <= SIZE. Therefore, the above idiom is actually
7523 undefined behaviour.
7525 For example this could cause problems if some malloc implementation
7526 provides an address for "p" such that "p + len" actually overflows for
7527 values of len that are too big and therefore p + len < limit.
7529 This issue was reported to OpenSSL by Guido Vranken
7534 * Constant time flag not preserved in DSA signing
7536 Operations in the DSA signing algorithm should run in constant time in
7537 order to avoid side channel attacks. A flaw in the OpenSSL DSA
7538 implementation means that a non-constant time codepath is followed for
7539 certain operations. This has been demonstrated through a cache-timing
7540 attack to be sufficient for an attacker to recover the private DSA key.
7542 This issue was reported by César Pereida (Aalto University), Billy Brumley
7543 (Tampere University of Technology), and Yuval Yarom (The University of
7544 Adelaide and NICTA).
7549 * DTLS buffered message DoS
7551 In a DTLS connection where handshake messages are delivered out-of-order
7552 those messages that OpenSSL is not yet ready to process will be buffered
7553 for later use. Under certain circumstances, a flaw in the logic means that
7554 those messages do not get removed from the buffer even though the handshake
7555 has been completed. An attacker could force up to approx. 15 messages to
7556 remain in the buffer when they are no longer required. These messages will
7557 be cleared when the DTLS connection is closed. The default maximum size for
7558 a message is 100k. Therefore, the attacker could force an additional 1500k
7559 to be consumed per connection. By opening many simultaneous connections an
7560 attacker could cause a DoS attack through memory exhaustion.
7562 This issue was reported to OpenSSL by Quan Luo.
7567 * DTLS replay protection DoS
7569 A flaw in the DTLS replay attack protection mechanism means that records
7570 that arrive for future epochs update the replay protection "window" before
7571 the MAC for the record has been validated. This could be exploited by an
7572 attacker by sending a record for the next epoch (which does not have to
7573 decrypt or have a valid MAC), with a very large sequence number. This means
7574 that all subsequent legitimate packets are dropped causing a denial of
7575 service for a specific DTLS connection.
7577 This issue was reported to OpenSSL by the OCAP audit team.
7582 * Certificate message OOB reads
7584 In OpenSSL 1.0.2 and earlier some missing message length checks can result
7585 in OOB reads of up to 2 bytes beyond an allocated buffer. There is a
7586 theoretical DoS risk but this has not been observed in practice on common
7589 The messages affected are client certificate, client certificate request
7590 and server certificate. As a result the attack can only be performed
7591 against a client or a server which enables client authentication.
7593 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
7598 ### Changes between 1.0.1s and 1.0.1t [3 May 2016]
7600 * Prevent padding oracle in AES-NI CBC MAC check
7602 A MITM attacker can use a padding oracle attack to decrypt traffic
7603 when the connection uses an AES CBC cipher and the server support
7606 This issue was introduced as part of the fix for Lucky 13 padding
7607 attack ([CVE-2013-0169]). The padding check was rewritten to be in
7608 constant time by making sure that always the same bytes are read and
7609 compared against either the MAC or padding bytes. But it no longer
7610 checked that there was enough data to have both the MAC and padding
7613 This issue was reported by Juraj Somorovsky using TLS-Attacker.
7618 * Fix EVP_EncodeUpdate overflow
7620 An overflow can occur in the EVP_EncodeUpdate() function which is used for
7621 Base64 encoding of binary data. If an attacker is able to supply very large
7622 amounts of input data then a length check can overflow resulting in a heap
7625 Internally to OpenSSL the EVP_EncodeUpdate() function is primarily used by
7626 the `PEM_write_bio*` family of functions. These are mainly used within the
7627 OpenSSL command line applications, so any application which processes data
7628 from an untrusted source and outputs it as a PEM file should be considered
7629 vulnerable to this issue. User applications that call these APIs directly
7630 with large amounts of untrusted data may also be vulnerable.
7632 This issue was reported by Guido Vranken.
7637 * Fix EVP_EncryptUpdate overflow
7639 An overflow can occur in the EVP_EncryptUpdate() function. If an attacker
7640 is able to supply very large amounts of input data after a previous call to
7641 EVP_EncryptUpdate() with a partial block then a length check can overflow
7642 resulting in a heap corruption. Following an analysis of all OpenSSL
7643 internal usage of the EVP_EncryptUpdate() function all usage is one of two
7644 forms. The first form is where the EVP_EncryptUpdate() call is known to be
7645 the first called function after an EVP_EncryptInit(), and therefore that
7646 specific call must be safe. The second form is where the length passed to
7647 EVP_EncryptUpdate() can be seen from the code to be some small value and
7648 therefore there is no possibility of an overflow. Since all instances are
7649 one of these two forms, it is believed that there can be no overflows in
7650 internal code due to this problem. It should be noted that
7651 EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths.
7652 Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances
7653 of these calls have also been analysed too and it is believed there are no
7654 instances in internal usage where an overflow could occur.
7656 This issue was reported by Guido Vranken.
7661 * Prevent ASN.1 BIO excessive memory allocation
7663 When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio()
7664 a short invalid encoding can casuse allocation of large amounts of memory
7665 potentially consuming excessive resources or exhausting memory.
7667 Any application parsing untrusted data through d2i BIO functions is
7668 affected. The memory based functions such as d2i_X509() are *not* affected.
7669 Since the memory based functions are used by the TLS library, TLS
7670 applications are not affected.
7672 This issue was reported by Brian Carpenter.
7679 ASN1 Strings that are over 1024 bytes can cause an overread in applications
7680 using the X509_NAME_oneline() function on EBCDIC systems. This could result
7681 in arbitrary stack data being returned in the buffer.
7683 This issue was reported by Guido Vranken.
7688 * Modify behavior of ALPN to invoke callback after SNI/servername
7689 callback, such that updates to the SSL_CTX affect ALPN.
7693 * Remove LOW from the DEFAULT cipher list. This removes singles DES from the
7698 * Only remove the SSLv2 methods with the no-ssl2-method option. When the
7699 methods are enabled and ssl2 is disabled the methods return NULL.
7703 ### Changes between 1.0.1r and 1.0.1s [1 Mar 2016]
7705 * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL.
7706 Builds that are not configured with "enable-weak-ssl-ciphers" will not
7707 provide any "EXPORT" or "LOW" strength ciphers.
7711 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2
7712 is by default disabled at build-time. Builds that are not configured with
7713 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used,
7714 users who want to negotiate SSLv2 via the version-flexible SSLv23_method()
7715 will need to explicitly call either of:
7717 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2);
7719 SSL_clear_options(ssl, SSL_OP_NO_SSLv2);
7721 as appropriate. Even if either of those is used, or the application
7722 explicitly uses the version-specific SSLv2_method() or its client and
7723 server variants, SSLv2 ciphers vulnerable to exhaustive search key
7724 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT
7725 ciphers, and SSLv2 56-bit DES are no longer available.
7730 * Fix a double-free in DSA code
7732 A double free bug was discovered when OpenSSL parses malformed DSA private
7733 keys and could lead to a DoS attack or memory corruption for applications
7734 that receive DSA private keys from untrusted sources. This scenario is
7737 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using
7743 * Disable SRP fake user seed to address a server memory leak.
7745 Add a new method SRP_VBASE_get1_by_user that handles the seed properly.
7747 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
7748 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user
7749 was changed to ignore the "fake user" SRP seed, even if the seed
7752 Users should use SRP_VBASE_get1_by_user instead. Note that in
7753 SRP_VBASE_get1_by_user, caller must free the returned value. Note
7754 also that even though configuring the SRP seed attempts to hide
7755 invalid usernames by continuing the handshake with fake
7756 credentials, this behaviour is not constant time and no strong
7757 guarantees are made that the handshake is indistinguishable from
7758 that of a valid user.
7763 * Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption
7765 In the BN_hex2bn function the number of hex digits is calculated using an
7766 int value `i`. Later `bn_expand` is called with a value of `i * 4`. For
7767 large values of `i` this can result in `bn_expand` not allocating any
7768 memory because `i * 4` is negative. This can leave the internal BIGNUM data
7769 field as NULL leading to a subsequent NULL ptr deref. For very large values
7770 of `i`, the calculation `i * 4` could be a positive value smaller than `i`.
7771 In this case memory is allocated to the internal BIGNUM data field, but it
7772 is insufficiently sized leading to heap corruption. A similar issue exists
7773 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn
7774 is ever called by user applications with very large untrusted hex/dec data.
7775 This is anticipated to be a rare occurrence.
7777 All OpenSSL internal usage of these functions use data that is not expected
7778 to be untrusted, e.g. config file data or application command line
7779 arguments. If user developed applications generate config file data based
7780 on untrusted data then it is possible that this could also lead to security
7781 consequences. This is also anticipated to be rare.
7783 This issue was reported to OpenSSL by Guido Vranken.
7788 * Fix memory issues in `BIO_*printf` functions
7790 The internal `fmtstr` function used in processing a "%s" format string in
7791 the `BIO_*printf` functions could overflow while calculating the length of a
7792 string and cause an OOB read when printing very long strings.
7794 Additionally the internal `doapr_outch` function can attempt to write to an
7795 OOB memory location (at an offset from the NULL pointer) in the event of a
7796 memory allocation failure. In 1.0.2 and below this could be caused where
7797 the size of a buffer to be allocated is greater than INT_MAX. E.g. this
7798 could be in processing a very long "%s" format string. Memory leaks can
7801 The first issue may mask the second issue dependent on compiler behaviour.
7802 These problems could enable attacks where large amounts of untrusted data
7803 is passed to the `BIO_*printf` functions. If applications use these functions
7804 in this way then they could be vulnerable. OpenSSL itself uses these
7805 functions when printing out human-readable dumps of ASN.1 data. Therefore
7806 applications that print this data could be vulnerable if the data is from
7807 untrusted sources. OpenSSL command line applications could also be
7808 vulnerable where they print out ASN.1 data, or if untrusted data is passed
7809 as command line arguments.
7811 Libssl is not considered directly vulnerable. Additionally certificates etc
7812 received via remote connections via libssl are also unlikely to be able to
7813 trigger these issues because of message size limits enforced within libssl.
7815 This issue was reported to OpenSSL Guido Vranken.
7820 * Side channel attack on modular exponentiation
7822 A side-channel attack was found which makes use of cache-bank conflicts on
7823 the Intel Sandy-Bridge microarchitecture which could lead to the recovery
7824 of RSA keys. The ability to exploit this issue is limited as it relies on
7825 an attacker who has control of code in a thread running on the same
7826 hyper-threaded core as the victim thread which is performing decryptions.
7828 This issue was reported to OpenSSL by Yuval Yarom, The University of
7829 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and
7830 Nadia Heninger, University of Pennsylvania with more information at
7831 <http://cachebleed.info>.
7836 * Change the req command to generate a 2048-bit RSA/DSA key by default,
7837 if no keysize is specified with default_bits. This fixes an
7838 omission in an earlier change that changed all RSA/DSA key generation
7839 commands to use 2048 bits by default.
7843 ### Changes between 1.0.1q and 1.0.1r [28 Jan 2016]
7845 * Protection for DH small subgroup attacks
7847 As a precautionary measure the SSL_OP_SINGLE_DH_USE option has been
7848 switched on by default and cannot be disabled. This could have some
7853 * SSLv2 doesn't block disabled ciphers
7855 A malicious client can negotiate SSLv2 ciphers that have been disabled on
7856 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
7857 been disabled, provided that the SSLv2 protocol was not also disabled via
7860 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
7861 and Sebastian Schinzel.
7866 * Reject DH handshakes with parameters shorter than 1024 bits.
7870 ### Changes between 1.0.1p and 1.0.1q [3 Dec 2015]
7872 * Certificate verify crash with missing PSS parameter
7874 The signature verification routines will crash with a NULL pointer
7875 dereference if presented with an ASN.1 signature using the RSA PSS
7876 algorithm and absent mask generation function parameter. Since these
7877 routines are used to verify certificate signature algorithms this can be
7878 used to crash any certificate verification operation and exploited in a
7879 DoS attack. Any application which performs certificate verification is
7880 vulnerable including OpenSSL clients and servers which enable client
7883 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
7888 * X509_ATTRIBUTE memory leak
7890 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
7891 memory. This structure is used by the PKCS#7 and CMS routines so any
7892 application which reads PKCS#7 or CMS data from untrusted sources is
7893 affected. SSL/TLS is not affected.
7895 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
7901 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
7902 This changes the decoding behaviour for some invalid messages,
7903 though the change is mostly in the more lenient direction, and
7904 legacy behaviour is preserved as much as possible.
7908 * In DSA_generate_parameters_ex, if the provided seed is too short,
7909 use a random seed, as already documented.
7911 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
7913 ### Changes between 1.0.1o and 1.0.1p [9 Jul 2015]
7915 * Alternate chains certificate forgery
7917 During certificate verification, OpenSSL will attempt to find an
7918 alternative certificate chain if the first attempt to build such a chain
7919 fails. An error in the implementation of this logic can mean that an
7920 attacker could cause certain checks on untrusted certificates to be
7921 bypassed, such as the CA flag, enabling them to use a valid leaf
7922 certificate to act as a CA and "issue" an invalid certificate.
7924 This issue was reported to OpenSSL by Adam Langley/David Benjamin
7930 * Race condition handling PSK identify hint
7932 If PSK identity hints are received by a multi-threaded client then
7933 the values are wrongly updated in the parent SSL_CTX structure. This can
7934 result in a race condition potentially leading to a double free of the
7940 ### Changes between 1.0.1n and 1.0.1o [12 Jun 2015]
7942 * Fix HMAC ABI incompatibility. The previous version introduced an ABI
7943 incompatibility in the handling of HMAC. The previous ABI has now been
7946 ### Changes between 1.0.1m and 1.0.1n [11 Jun 2015]
7948 * Malformed ECParameters causes infinite loop
7950 When processing an ECParameters structure OpenSSL enters an infinite loop
7951 if the curve specified is over a specially malformed binary polynomial
7954 This can be used to perform denial of service against any
7955 system which processes public keys, certificate requests or
7956 certificates. This includes TLS clients and TLS servers with
7957 client authentication enabled.
7959 This issue was reported to OpenSSL by Joseph Barr-Pixton.
7964 * Exploitable out-of-bounds read in X509_cmp_time
7966 X509_cmp_time does not properly check the length of the ASN1_TIME
7967 string and can read a few bytes out of bounds. In addition,
7968 X509_cmp_time accepts an arbitrary number of fractional seconds in the
7971 An attacker can use this to craft malformed certificates and CRLs of
7972 various sizes and potentially cause a segmentation fault, resulting in
7973 a DoS on applications that verify certificates or CRLs. TLS clients
7974 that verify CRLs are affected. TLS clients and servers with client
7975 authentication enabled may be affected if they use custom verification
7978 This issue was reported to OpenSSL by Robert Swiecki (Google), and
7979 independently by Hanno Böck.
7984 * PKCS7 crash with missing EnvelopedContent
7986 The PKCS#7 parsing code does not handle missing inner EncryptedContent
7987 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
7988 with missing content and trigger a NULL pointer dereference on parsing.
7990 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
7991 structures from untrusted sources are affected. OpenSSL clients and
7992 servers are not affected.
7994 This issue was reported to OpenSSL by Michal Zalewski (Google).
7999 * CMS verify infinite loop with unknown hash function
8001 When verifying a signedData message the CMS code can enter an infinite loop
8002 if presented with an unknown hash function OID. This can be used to perform
8003 denial of service against any system which verifies signedData messages using
8005 This issue was reported to OpenSSL by Johannes Bauer.
8010 * Race condition handling NewSessionTicket
8012 If a NewSessionTicket is received by a multi-threaded client when attempting to
8013 reuse a previous ticket then a race condition can occur potentially leading to
8014 a double free of the ticket data.
8019 * Reject DH handshakes with parameters shorter than 768 bits.
8021 *Kurt Roeckx and Emilia Kasper*
8023 * dhparam: generate 2048-bit parameters by default.
8025 *Kurt Roeckx and Emilia Kasper*
8027 ### Changes between 1.0.1l and 1.0.1m [19 Mar 2015]
8029 * Segmentation fault in ASN1_TYPE_cmp fix
8031 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
8032 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
8033 certificate signature algorithm consistency this can be used to crash any
8034 certificate verification operation and exploited in a DoS attack. Any
8035 application which performs certificate verification is vulnerable including
8036 OpenSSL clients and servers which enable client authentication.
8041 * ASN.1 structure reuse memory corruption fix
8043 Reusing a structure in ASN.1 parsing may allow an attacker to cause
8044 memory corruption via an invalid write. Such reuse is and has been
8045 strongly discouraged and is believed to be rare.
8047 Applications that parse structures containing CHOICE or ANY DEFINED BY
8048 components may be affected. Certificate parsing (d2i_X509 and related
8049 functions) are however not affected. OpenSSL clients and servers are
8055 * PKCS7 NULL pointer dereferences fix
8057 The PKCS#7 parsing code does not handle missing outer ContentInfo
8058 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
8059 missing content and trigger a NULL pointer dereference on parsing.
8061 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
8062 otherwise parse PKCS#7 structures from untrusted sources are
8063 affected. OpenSSL clients and servers are not affected.
8065 This issue was reported to OpenSSL by Michal Zalewski (Google).
8070 * DoS via reachable assert in SSLv2 servers fix
8072 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
8073 servers that both support SSLv2 and enable export cipher suites by sending
8074 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
8076 This issue was discovered by Sean Burford (Google) and Emilia Käsper
8077 (OpenSSL development team).
8082 * Use After Free following d2i_ECPrivatekey error fix
8084 A malformed EC private key file consumed via the d2i_ECPrivateKey function
8085 could cause a use after free condition. This, in turn, could cause a double
8086 free in several private key parsing functions (such as d2i_PrivateKey
8087 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
8088 for applications that receive EC private keys from untrusted
8089 sources. This scenario is considered rare.
8091 This issue was discovered by the BoringSSL project and fixed in their
8097 * X509_to_X509_REQ NULL pointer deref fix
8099 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
8100 the certificate key is invalid. This function is rarely used in practice.
8102 This issue was discovered by Brian Carpenter.
8107 * Removed the export ciphers from the DEFAULT ciphers
8111 ### Changes between 1.0.1k and 1.0.1l [15 Jan 2015]
8113 * Build fixes for the Windows and OpenVMS platforms
8115 *Matt Caswell and Richard Levitte*
8117 ### Changes between 1.0.1j and 1.0.1k [8 Jan 2015]
8119 * Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
8120 message can cause a segmentation fault in OpenSSL due to a NULL pointer
8121 dereference. This could lead to a Denial Of Service attack. Thanks to
8122 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
8127 * Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
8128 dtls1_buffer_record function under certain conditions. In particular this
8129 could occur if an attacker sent repeated DTLS records with the same
8130 sequence number but for the next epoch. The memory leak could be exploited
8131 by an attacker in a Denial of Service attack through memory exhaustion.
8132 Thanks to Chris Mueller for reporting this issue.
8137 * Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
8138 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
8139 method would be set to NULL which could later result in a NULL pointer
8140 dereference. Thanks to Frank Schmirler for reporting this issue.
8145 * Abort handshake if server key exchange message is omitted for ephemeral
8148 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
8149 reporting this issue.
8154 * Remove non-export ephemeral RSA code on client and server. This code
8155 violated the TLS standard by allowing the use of temporary RSA keys in
8156 non-export ciphersuites and could be used by a server to effectively
8157 downgrade the RSA key length used to a value smaller than the server
8158 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
8159 INRIA or reporting this issue.
8164 * Fixed issue where DH client certificates are accepted without verification.
8165 An OpenSSL server will accept a DH certificate for client authentication
8166 without the certificate verify message. This effectively allows a client to
8167 authenticate without the use of a private key. This only affects servers
8168 which trust a client certificate authority which issues certificates
8169 containing DH keys: these are extremely rare and hardly ever encountered.
8170 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
8176 * Ensure that the session ID context of an SSL is updated when its
8177 SSL_CTX is updated via SSL_set_SSL_CTX.
8179 The session ID context is typically set from the parent SSL_CTX,
8180 and can vary with the CTX.
8184 * Fix various certificate fingerprint issues.
8186 By using non-DER or invalid encodings outside the signed portion of a
8187 certificate the fingerprint can be changed without breaking the signature.
8188 Although no details of the signed portion of the certificate can be changed
8189 this can cause problems with some applications: e.g. those using the
8190 certificate fingerprint for blacklists.
8192 1. Reject signatures with non zero unused bits.
8194 If the BIT STRING containing the signature has non zero unused bits reject
8195 the signature. All current signature algorithms require zero unused bits.
8197 2. Check certificate algorithm consistency.
8199 Check the AlgorithmIdentifier inside TBS matches the one in the
8200 certificate signature. NB: this will result in signature failure
8201 errors for some broken certificates.
8203 Thanks to Konrad Kraszewski from Google for reporting this issue.
8205 3. Check DSA/ECDSA signatures use DER.
8207 Re-encode DSA/ECDSA signatures and compare with the original received
8208 signature. Return an error if there is a mismatch.
8210 This will reject various cases including garbage after signature
8211 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
8212 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
8213 (negative or with leading zeroes).
8215 Further analysis was conducted and fixes were developed by Stephen Henson
8216 of the OpenSSL core team.
8222 * Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
8223 results on some platforms, including x86_64. This bug occurs at random
8224 with a very low probability, and is not known to be exploitable in any
8225 way, though its exact impact is difficult to determine. Thanks to Pieter
8226 Wuille (Blockstream) who reported this issue and also suggested an initial
8227 fix. Further analysis was conducted by the OpenSSL development team and
8228 Adam Langley of Google. The final fix was developed by Andy Polyakov of
8229 the OpenSSL core team.
8234 * Do not resume sessions on the server if the negotiated protocol
8235 version does not match the session's version. Resuming with a different
8236 version, while not strictly forbidden by the RFC, is of questionable
8237 sanity and breaks all known clients.
8239 *David Benjamin, Emilia Käsper*
8241 * Tighten handling of the ChangeCipherSpec (CCS) message: reject
8242 early CCS messages during renegotiation. (Note that because
8243 renegotiation is encrypted, this early CCS was not exploitable.)
8247 * Tighten client-side session ticket handling during renegotiation:
8248 ensure that the client only accepts a session ticket if the server sends
8249 the extension anew in the ServerHello. Previously, a TLS client would
8250 reuse the old extension state and thus accept a session ticket if one was
8251 announced in the initial ServerHello.
8253 Similarly, ensure that the client requires a session ticket if one
8254 was advertised in the ServerHello. Previously, a TLS client would
8255 ignore a missing NewSessionTicket message.
8259 ### Changes between 1.0.1i and 1.0.1j [15 Oct 2014]
8263 A flaw in the DTLS SRTP extension parsing code allows an attacker, who
8264 sends a carefully crafted handshake message, to cause OpenSSL to fail
8265 to free up to 64k of memory causing a memory leak. This could be
8266 exploited in a Denial Of Service attack. This issue affects OpenSSL
8267 1.0.1 server implementations for both SSL/TLS and DTLS regardless of
8268 whether SRTP is used or configured. Implementations of OpenSSL that
8269 have been compiled with OPENSSL_NO_SRTP defined are not affected.
8271 The fix was developed by the OpenSSL team.
8276 * Session Ticket Memory Leak.
8278 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
8279 integrity of that ticket is first verified. In the event of a session
8280 ticket integrity check failing, OpenSSL will fail to free memory
8281 causing a memory leak. By sending a large number of invalid session
8282 tickets an attacker could exploit this issue in a Denial Of Service
8288 * Build option no-ssl3 is incomplete.
8290 When OpenSSL is configured with "no-ssl3" as a build option, servers
8291 could accept and complete an SSL 3.0 handshake, and clients could be
8292 configured to send them.
8295 *Akamai and the OpenSSL team*
8297 * Add support for TLS_FALLBACK_SCSV.
8298 Client applications doing fallback retries should call
8299 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
8302 *Adam Langley, Bodo Moeller*
8304 * Add additional DigestInfo checks.
8306 Re-encode DigestInto in DER and check against the original when
8307 verifying RSA signature: this will reject any improperly encoded
8308 DigestInfo structures.
8310 Note: this is a precautionary measure and no attacks are currently known.
8314 ### Changes between 1.0.1h and 1.0.1i [6 Aug 2014]
8316 * Fix SRP buffer overrun vulnerability. Invalid parameters passed to the
8317 SRP code can be overrun an internal buffer. Add sanity check that
8318 g, A, B < N to SRP code.
8320 Thanks to Sean Devlin and Watson Ladd of Cryptography Services, NCC
8321 Group for discovering this issue.
8326 * A flaw in the OpenSSL SSL/TLS server code causes the server to negotiate
8327 TLS 1.0 instead of higher protocol versions when the ClientHello message
8328 is badly fragmented. This allows a man-in-the-middle attacker to force a
8329 downgrade to TLS 1.0 even if both the server and the client support a
8330 higher protocol version, by modifying the client's TLS records.
8332 Thanks to David Benjamin and Adam Langley (Google) for discovering and
8333 researching this issue.
8338 * OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
8339 to a denial of service attack. A malicious server can crash the client
8340 with a null pointer dereference (read) by specifying an anonymous (EC)DH
8341 ciphersuite and sending carefully crafted handshake messages.
8343 Thanks to Felix Gröbert (Google) for discovering and researching this
8349 * By sending carefully crafted DTLS packets an attacker could cause openssl
8350 to leak memory. This can be exploited through a Denial of Service attack.
8351 Thanks to Adam Langley for discovering and researching this issue.
8356 * An attacker can force openssl to consume large amounts of memory whilst
8357 processing DTLS handshake messages. This can be exploited through a
8358 Denial of Service attack.
8359 Thanks to Adam Langley for discovering and researching this issue.
8364 * An attacker can force an error condition which causes openssl to crash
8365 whilst processing DTLS packets due to memory being freed twice. This
8366 can be exploited through a Denial of Service attack.
8367 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
8373 * If a multithreaded client connects to a malicious server using a resumed
8374 session and the server sends an ec point format extension it could write
8375 up to 255 bytes to freed memory.
8377 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
8383 * A malicious server can crash an OpenSSL client with a null pointer
8384 dereference (read) by specifying an SRP ciphersuite even though it was not
8385 properly negotiated with the client. This can be exploited through a
8386 Denial of Service attack.
8388 Thanks to Joonas Kuorilehto and Riku Hietamäki (Codenomicon) for
8389 discovering and researching this issue.
8394 * A flaw in OBJ_obj2txt may cause pretty printing functions such as
8395 X509_name_oneline, X509_name_print_ex et al. to leak some information
8396 from the stack. Applications may be affected if they echo pretty printing
8397 output to the attacker.
8399 Thanks to Ivan Fratric (Google) for discovering this issue.
8402 *Emilia Käsper, and Steve Henson*
8404 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
8405 for corner cases. (Certain input points at infinity could lead to
8406 bogus results, with non-infinity inputs mapped to infinity too.)
8410 ### Changes between 1.0.1g and 1.0.1h [5 Jun 2014]
8412 * Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
8413 handshake can force the use of weak keying material in OpenSSL
8414 SSL/TLS clients and servers.
8416 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
8417 researching this issue. ([CVE-2014-0224])
8419 *KIKUCHI Masashi, Steve Henson*
8421 * Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
8422 OpenSSL DTLS client the code can be made to recurse eventually crashing
8425 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
8428 *Imre Rad, Steve Henson*
8430 * Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
8431 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
8432 client or server. This is potentially exploitable to run arbitrary
8433 code on a vulnerable client or server.
8435 Thanks to Jüri Aedla for reporting this issue. ([CVE-2014-0195])
8437 *Jüri Aedla, Steve Henson*
8439 * Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
8440 are subject to a denial of service attack.
8442 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
8443 this issue. ([CVE-2014-3470])
8445 *Felix Gröbert, Ivan Fratric, Steve Henson*
8447 * Harmonize version and its documentation. -f flag is used to display
8450 *mancha <mancha1@zoho.com>*
8452 * Fix eckey_priv_encode so it immediately returns an error upon a failure
8453 in i2d_ECPrivateKey.
8455 *mancha <mancha1@zoho.com>*
8457 * Fix some double frees. These are not thought to be exploitable.
8459 *mancha <mancha1@zoho.com>*
8461 ### Changes between 1.0.1f and 1.0.1g [7 Apr 2014]
8463 * A missing bounds check in the handling of the TLS heartbeat extension
8464 can be used to reveal up to 64k of memory to a connected client or
8467 Thanks for Neel Mehta of Google Security for discovering this bug and to
8468 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
8469 preparing the fix ([CVE-2014-0160])
8471 *Adam Langley, Bodo Moeller*
8473 * Fix for the attack described in the paper "Recovering OpenSSL
8474 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
8475 by Yuval Yarom and Naomi Benger. Details can be obtained from:
8476 <http://eprint.iacr.org/2014/140>
8478 Thanks to Yuval Yarom and Naomi Benger for discovering this
8479 flaw and to Yuval Yarom for supplying a fix ([CVE-2014-0076])
8481 *Yuval Yarom and Naomi Benger*
8483 * TLS pad extension: draft-agl-tls-padding-03
8485 Workaround for the "TLS hang bug" (see FAQ and PR#2771): if the
8486 TLS client Hello record length value would otherwise be > 255 and
8487 less that 512 pad with a dummy extension containing zeroes so it
8488 is at least 512 bytes long.
8490 *Adam Langley, Steve Henson*
8492 ### Changes between 1.0.1e and 1.0.1f [6 Jan 2014]
8494 * Fix for TLS record tampering bug. A carefully crafted invalid
8495 handshake could crash OpenSSL with a NULL pointer exception.
8496 Thanks to Anton Johansson for reporting this issues.
8499 * Keep original DTLS digest and encryption contexts in retransmission
8500 structures so we can use the previous session parameters if they need
8501 to be resent. ([CVE-2013-6450])
8505 * Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
8506 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
8507 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
8508 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
8509 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
8510 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
8512 *Rob Stradling, Adam Langley*
8514 ### Changes between 1.0.1d and 1.0.1e [11 Feb 2013]
8516 * Correct fix for CVE-2013-0169. The original didn't work on AES-NI
8517 supporting platforms or when small records were transferred.
8519 *Andy Polyakov, Steve Henson*
8521 ### Changes between 1.0.1c and 1.0.1d [5 Feb 2013]
8523 * Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
8525 This addresses the flaw in CBC record processing discovered by
8526 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
8527 at: <http://www.isg.rhul.ac.uk/tls/>
8529 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
8530 Security Group at Royal Holloway, University of London
8531 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
8532 Emilia Käsper for the initial patch.
8535 *Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson*
8537 * Fix flaw in AESNI handling of TLS 1.2 and 1.1 records for CBC mode
8538 ciphersuites which can be exploited in a denial of service attack.
8539 Thanks go to and to Adam Langley <agl@chromium.org> for discovering
8540 and detecting this bug and to Wolfgang Ettlinger
8541 <wolfgang.ettlinger@gmail.com> for independently discovering this issue.
8546 * Return an error when checking OCSP signatures when key is NULL.
8547 This fixes a DoS attack. ([CVE-2013-0166])
8551 * Make openssl verify return errors.
8553 *Chris Palmer <palmer@google.com> and Ben Laurie*
8555 * Call OCSP Stapling callback after ciphersuite has been chosen, so
8556 the right response is stapled. Also change SSL_get_certificate()
8557 so it returns the certificate actually sent.
8558 See <http://rt.openssl.org/Ticket/Display.html?id=2836>.
8560 *Rob Stradling <rob.stradling@comodo.com>*
8562 * Fix possible deadlock when decoding public keys.
8566 * Don't use TLS 1.0 record version number in initial client hello
8571 ### Changes between 1.0.1b and 1.0.1c [10 May 2012]
8573 * Sanity check record length before skipping explicit IV in TLS
8574 1.2, 1.1 and DTLS to fix DoS attack.
8576 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
8577 fuzzing as a service testing platform.
8582 * Initialise tkeylen properly when encrypting CMS messages.
8583 Thanks to Solar Designer of Openwall for reporting this issue.
8587 * In FIPS mode don't try to use composite ciphers as they are not
8592 ### Changes between 1.0.1a and 1.0.1b [26 Apr 2012]
8594 * OpenSSL 1.0.0 sets SSL_OP_ALL to 0x80000FFFL and OpenSSL 1.0.1 and
8595 1.0.1a set SSL_OP_NO_TLSv1_1 to 0x00000400L which would unfortunately
8596 mean any application compiled against OpenSSL 1.0.0 headers setting
8597 SSL_OP_ALL would also set SSL_OP_NO_TLSv1_1, unintentionally disabling
8598 TLS 1.1 also. Fix this by changing the value of SSL_OP_NO_TLSv1_1 to
8599 0x10000000L Any application which was previously compiled against
8600 OpenSSL 1.0.1 or 1.0.1a headers and which cares about SSL_OP_NO_TLSv1_1
8601 will need to be recompiled as a result. Letting be results in
8602 inability to disable specifically TLS 1.1 and in client context,
8603 in unlike event, limit maximum offered version to TLS 1.0 [see below].
8607 * In order to ensure interoperability SSL_OP_NO_protocolX does not
8608 disable just protocol X, but all protocols above X *if* there are
8609 protocols *below* X still enabled. In more practical terms it means
8610 that if application wants to disable TLS1.0 in favor of TLS1.1 and
8611 above, it's not sufficient to pass `SSL_OP_NO_TLSv1`, one has to pass
8612 `SSL_OP_NO_TLSv1|SSL_OP_NO_SSLv3|SSL_OP_NO_SSLv2`. This applies to
8617 ### Changes between 1.0.1 and 1.0.1a [19 Apr 2012]
8619 * Check for potentially exploitable overflows in asn1_d2i_read_bio
8620 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
8621 in CRYPTO_realloc_clean.
8623 Thanks to Tavis Ormandy, Google Security Team, for discovering this
8624 issue and to Adam Langley <agl@chromium.org> for fixing it.
8627 *Adam Langley (Google), Tavis Ormandy, Google Security Team*
8629 * Don't allow TLS 1.2 SHA-256 ciphersuites in TLS 1.0, 1.1 connections.
8633 * Workarounds for some broken servers that "hang" if a client hello
8634 record length exceeds 255 bytes.
8636 1. Do not use record version number > TLS 1.0 in initial client
8637 hello: some (but not all) hanging servers will now work.
8638 2. If we set OPENSSL_MAX_TLS1_2_CIPHER_LENGTH this will truncate
8639 the number of ciphers sent in the client hello. This should be
8640 set to an even number, such as 50, for example by passing:
8641 -DOPENSSL_MAX_TLS1_2_CIPHER_LENGTH=50 to config or Configure.
8642 Most broken servers should now work.
8643 3. If all else fails setting OPENSSL_NO_TLS1_2_CLIENT will disable
8644 TLS 1.2 client support entirely.
8648 * Fix SEGV in Vector Permutation AES module observed in OpenSSH.
8652 ### Changes between 1.0.0h and 1.0.1 [14 Mar 2012]
8654 * Add compatibility with old MDC2 signatures which use an ASN1 OCTET
8655 STRING form instead of a DigestInfo.
8659 * The format used for MDC2 RSA signatures is inconsistent between EVP
8660 and the RSA_sign/RSA_verify functions. This was made more apparent when
8661 OpenSSL used RSA_sign/RSA_verify for some RSA signatures in particular
8662 those which went through EVP_PKEY_METHOD in 1.0.0 and later. Detect
8663 the correct format in RSA_verify so both forms transparently work.
8667 * Some servers which support TLS 1.0 can choke if we initially indicate
8668 support for TLS 1.2 and later renegotiate using TLS 1.0 in the RSA
8669 encrypted premaster secret. As a workaround use the maximum permitted
8670 client version in client hello, this should keep such servers happy
8671 and still work with previous versions of OpenSSL.
8675 * Add support for TLS/DTLS heartbeats.
8677 *Robin Seggelmann <seggelmann@fh-muenster.de>*
8679 * Add support for SCTP.
8681 *Robin Seggelmann <seggelmann@fh-muenster.de>*
8683 * Improved PRNG seeding for VOS.
8685 *Paul Green <Paul.Green@stratus.com>*
8687 * Extensive assembler packs updates, most notably:
8689 - x86[_64]: AES-NI, PCLMULQDQ, RDRAND support;
8690 - x86[_64]: SSSE3 support (SHA1, vector-permutation AES);
8691 - x86_64: bit-sliced AES implementation;
8692 - ARM: NEON support, contemporary platforms optimizations;
8693 - s390x: z196 support;
8694 - `*`: GHASH and GF(2^m) multiplication implementations;
8698 * Make TLS-SRP code conformant with RFC 5054 API cleanup
8699 (removal of unnecessary code)
8701 *Peter Sylvester <peter.sylvester@edelweb.fr>*
8703 * Add TLS key material exporter from RFC 5705.
8707 * Add DTLS-SRTP negotiation from RFC 5764.
8711 * Add Next Protocol Negotiation,
8712 <http://tools.ietf.org/html/draft-agl-tls-nextprotoneg-00>. Can be
8713 disabled with a no-npn flag to config or Configure. Code donated
8716 *Adam Langley <agl@google.com> and Ben Laurie*
8718 * Add optional 64-bit optimized implementations of elliptic curves NIST-P224,
8719 NIST-P256, NIST-P521, with constant-time single point multiplication on
8720 typical inputs. Compiler support for the nonstandard type `__uint128_t` is
8721 required to use this (present in gcc 4.4 and later, for 64-bit builds).
8722 Code made available under Apache License version 2.0.
8724 Specify "enable-ec_nistp_64_gcc_128" on the Configure (or config) command
8725 line to include this in your build of OpenSSL, and run "make depend" (or
8726 "make update"). This enables the following EC_METHODs:
8728 EC_GFp_nistp224_method()
8729 EC_GFp_nistp256_method()
8730 EC_GFp_nistp521_method()
8732 EC_GROUP_new_by_curve_name() will automatically use these (while
8733 EC_GROUP_new_curve_GFp() currently prefers the more flexible
8736 *Emilia Käsper, Adam Langley, Bodo Moeller (Google)*
8738 * Use type ossl_ssize_t instead of ssize_t which isn't available on
8739 all platforms. Move ssize_t definition from e_os.h to the public
8740 header file e_os2.h as it now appears in public header file cms.h
8744 * New -sigopt option to the ca, req and x509 utilities. Additional
8745 signature parameters can be passed using this option and in
8750 * Add RSA PSS signing function. This will generate and set the
8751 appropriate AlgorithmIdentifiers for PSS based on those in the
8752 corresponding EVP_MD_CTX structure. No application support yet.
8756 * Support for companion algorithm specific ASN1 signing routines.
8757 New function ASN1_item_sign_ctx() signs a pre-initialised
8758 EVP_MD_CTX structure and sets AlgorithmIdentifiers based on
8759 the appropriate parameters.
8763 * Add new algorithm specific ASN1 verification initialisation function
8764 to EVP_PKEY_ASN1_METHOD: this is not in EVP_PKEY_METHOD since the ASN1
8765 handling will be the same no matter what EVP_PKEY_METHOD is used.
8766 Add a PSS handler to support verification of PSS signatures: checked
8767 against a number of sample certificates.
8771 * Add signature printing for PSS. Add PSS OIDs.
8773 *Steve Henson, Martin Kaiser <lists@kaiser.cx>*
8775 * Add algorithm specific signature printing. An individual ASN1 method
8776 can now print out signatures instead of the standard hex dump.
8778 More complex signatures (e.g. PSS) can print out more meaningful
8779 information. Include DSA version that prints out the signature
8784 * Password based recipient info support for CMS library: implementing
8789 * Split password based encryption into PBES2 and PBKDF2 functions. This
8790 neatly separates the code into cipher and PBE sections and is required
8791 for some algorithms that split PBES2 into separate pieces (such as
8792 password based CMS).
8796 * Session-handling fixes:
8797 - Fix handling of connections that are resuming with a session ID,
8798 but also support Session Tickets.
8799 - Fix a bug that suppressed issuing of a new ticket if the client
8800 presented a ticket with an expired session.
8801 - Try to set the ticket lifetime hint to something reasonable.
8802 - Make tickets shorter by excluding irrelevant information.
8803 - On the client side, don't ignore renewed tickets.
8805 *Adam Langley, Bodo Moeller (Google)*
8807 * Fix PSK session representation.
8811 * Add RC4-MD5 and AESNI-SHA1 "stitched" implementations.
8813 This work was sponsored by Intel.
8817 * Add GCM support to TLS library. Some custom code is needed to split
8818 the IV between the fixed (from PRF) and explicit (from TLS record)
8819 portions. This adds all GCM ciphersuites supported by RFC5288 and
8820 RFC5289. Generalise some `AES*` cipherstrings to include GCM and
8821 add a special AESGCM string for GCM only.
8825 * Expand range of ctrls for AES GCM. Permit setting invocation
8826 field on decrypt and retrieval of invocation field only on encrypt.
8830 * Add HMAC ECC ciphersuites from RFC5289. Include SHA384 PRF support.
8831 As required by RFC5289 these ciphersuites cannot be used if for
8832 versions of TLS earlier than 1.2.
8836 * For FIPS capable OpenSSL interpret a NULL default public key method
8837 as unset and return the appropriate default but do *not* set the default.
8838 This means we can return the appropriate method in applications that
8839 switch between FIPS and non-FIPS modes.
8843 * Redirect HMAC and CMAC operations to FIPS module in FIPS mode. If an
8844 ENGINE is used then we cannot handle that in the FIPS module so we
8845 keep original code iff non-FIPS operations are allowed.
8849 * Add -attime option to openssl utilities.
8851 *Peter Eckersley <pde@eff.org>, Ben Laurie and Steve Henson*
8853 * Redirect DSA and DH operations to FIPS module in FIPS mode.
8857 * Redirect ECDSA and ECDH operations to FIPS module in FIPS mode. Also use
8858 FIPS EC methods unconditionally for now.
8862 * New build option no-ec2m to disable characteristic 2 code.
8866 * Backport libcrypto audit of return value checking from 1.1.0-dev; not
8867 all cases can be covered as some introduce binary incompatibilities.
8871 * Redirect RSA operations to FIPS module including keygen,
8872 encrypt, decrypt, sign and verify. Block use of non FIPS RSA methods.
8876 * Add similar low-level API blocking to ciphers.
8880 * low-level digest APIs are not approved in FIPS mode: any attempt
8881 to use these will cause a fatal error. Applications that *really* want
8882 to use them can use the `private_*` version instead.
8886 * Redirect cipher operations to FIPS module for FIPS builds.
8890 * Redirect digest operations to FIPS module for FIPS builds.
8894 * Update build system to add "fips" flag which will link in fipscanister.o
8895 for static and shared library builds embedding a signature if needed.
8899 * Output TLS supported curves in preference order instead of numerical
8900 order. This is currently hardcoded for the highest order curves first.
8901 This should be configurable so applications can judge speed vs strength.
8905 * Add TLS v1.2 server support for client authentication.
8909 * Add support for FIPS mode in ssl library: disable SSLv3, non-FIPS ciphers
8914 * Functions FIPS_mode_set() and FIPS_mode() which call the underlying
8915 FIPS modules versions.
8919 * Add TLS v1.2 client side support for client authentication. Keep cache
8920 of handshake records longer as we don't know the hash algorithm to use
8921 until after the certificate request message is received.
8925 * Initial TLS v1.2 client support. Add a default signature algorithms
8926 extension including all the algorithms we support. Parse new signature
8927 format in client key exchange. Relax some ECC signing restrictions for
8928 TLS v1.2 as indicated in RFC5246.
8932 * Add server support for TLS v1.2 signature algorithms extension. Switch
8933 to new signature format when needed using client digest preference.
8934 All server ciphersuites should now work correctly in TLS v1.2. No client
8935 support yet and no support for client certificates.
8939 * Initial TLS v1.2 support. Add new SHA256 digest to ssl code, switch
8940 to SHA256 for PRF when using TLS v1.2 and later. Add new SHA256 based
8941 ciphersuites. At present only RSA key exchange ciphersuites work with
8942 TLS v1.2. Add new option for TLS v1.2 replacing the old and obsolete
8943 SSL_OP_PKCS1_CHECK flags with SSL_OP_NO_TLSv1_2. New TLSv1.2 methods
8944 and version checking.
8948 * New option OPENSSL_NO_SSL_INTERN. If an application can be compiled
8949 with this defined it will not be affected by any changes to ssl internal
8950 structures. Add several utility functions to allow openssl application
8951 to work with OPENSSL_NO_SSL_INTERN defined.
8955 * A long standing patch to add support for SRP from EdelWeb (Peter
8956 Sylvester and Christophe Renou) was integrated.
8957 *Christophe Renou <christophe.renou@edelweb.fr>, Peter Sylvester
8958 <peter.sylvester@edelweb.fr>, Tom Wu <tjw@cs.stanford.edu>, and
8961 * Add functions to copy EVP_PKEY_METHOD and retrieve flags and id.
8965 * Permit abbreviated handshakes when renegotiating using the function
8966 SSL_renegotiate_abbreviated().
8968 *Robin Seggelmann <seggelmann@fh-muenster.de>*
8970 * Add call to ENGINE_register_all_complete() to
8971 ENGINE_load_builtin_engines(), so some implementations get used
8972 automatically instead of needing explicit application support.
8976 * Add support for TLS key exporter as described in RFC5705.
8978 *Robin Seggelmann <seggelmann@fh-muenster.de>, Steve Henson*
8980 * Initial TLSv1.1 support. Since TLSv1.1 is very similar to TLS v1.0 only
8981 a few changes are required:
8983 Add SSL_OP_NO_TLSv1_1 flag.
8984 Add TLSv1_1 methods.
8985 Update version checking logic to handle version 1.1.
8986 Add explicit IV handling (ported from DTLS code).
8987 Add command line options to s_client/s_server.
8994 ### Changes between 1.0.0s and 1.0.0t [3 Dec 2015]
8996 * X509_ATTRIBUTE memory leak
8998 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
8999 memory. This structure is used by the PKCS#7 and CMS routines so any
9000 application which reads PKCS#7 or CMS data from untrusted sources is
9001 affected. SSL/TLS is not affected.
9003 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
9009 * Race condition handling PSK identify hint
9011 If PSK identity hints are received by a multi-threaded client then
9012 the values are wrongly updated in the parent SSL_CTX structure. This can
9013 result in a race condition potentially leading to a double free of the
9019 ### Changes between 1.0.0r and 1.0.0s [11 Jun 2015]
9021 * Malformed ECParameters causes infinite loop
9023 When processing an ECParameters structure OpenSSL enters an infinite loop
9024 if the curve specified is over a specially malformed binary polynomial
9027 This can be used to perform denial of service against any
9028 system which processes public keys, certificate requests or
9029 certificates. This includes TLS clients and TLS servers with
9030 client authentication enabled.
9032 This issue was reported to OpenSSL by Joseph Barr-Pixton.
9037 * Exploitable out-of-bounds read in X509_cmp_time
9039 X509_cmp_time does not properly check the length of the ASN1_TIME
9040 string and can read a few bytes out of bounds. In addition,
9041 X509_cmp_time accepts an arbitrary number of fractional seconds in the
9044 An attacker can use this to craft malformed certificates and CRLs of
9045 various sizes and potentially cause a segmentation fault, resulting in
9046 a DoS on applications that verify certificates or CRLs. TLS clients
9047 that verify CRLs are affected. TLS clients and servers with client
9048 authentication enabled may be affected if they use custom verification
9051 This issue was reported to OpenSSL by Robert Swiecki (Google), and
9052 independently by Hanno Böck.
9057 * PKCS7 crash with missing EnvelopedContent
9059 The PKCS#7 parsing code does not handle missing inner EncryptedContent
9060 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
9061 with missing content and trigger a NULL pointer dereference on parsing.
9063 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
9064 structures from untrusted sources are affected. OpenSSL clients and
9065 servers are not affected.
9067 This issue was reported to OpenSSL by Michal Zalewski (Google).
9072 * CMS verify infinite loop with unknown hash function
9074 When verifying a signedData message the CMS code can enter an infinite loop
9075 if presented with an unknown hash function OID. This can be used to perform
9076 denial of service against any system which verifies signedData messages using
9078 This issue was reported to OpenSSL by Johannes Bauer.
9083 * Race condition handling NewSessionTicket
9085 If a NewSessionTicket is received by a multi-threaded client when attempting to
9086 reuse a previous ticket then a race condition can occur potentially leading to
9087 a double free of the ticket data.
9092 ### Changes between 1.0.0q and 1.0.0r [19 Mar 2015]
9094 * Segmentation fault in ASN1_TYPE_cmp fix
9096 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
9097 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
9098 certificate signature algorithm consistency this can be used to crash any
9099 certificate verification operation and exploited in a DoS attack. Any
9100 application which performs certificate verification is vulnerable including
9101 OpenSSL clients and servers which enable client authentication.
9106 * ASN.1 structure reuse memory corruption fix
9108 Reusing a structure in ASN.1 parsing may allow an attacker to cause
9109 memory corruption via an invalid write. Such reuse is and has been
9110 strongly discouraged and is believed to be rare.
9112 Applications that parse structures containing CHOICE or ANY DEFINED BY
9113 components may be affected. Certificate parsing (d2i_X509 and related
9114 functions) are however not affected. OpenSSL clients and servers are
9120 * PKCS7 NULL pointer dereferences fix
9122 The PKCS#7 parsing code does not handle missing outer ContentInfo
9123 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
9124 missing content and trigger a NULL pointer dereference on parsing.
9126 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
9127 otherwise parse PKCS#7 structures from untrusted sources are
9128 affected. OpenSSL clients and servers are not affected.
9130 This issue was reported to OpenSSL by Michal Zalewski (Google).
9135 * DoS via reachable assert in SSLv2 servers fix
9137 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
9138 servers that both support SSLv2 and enable export cipher suites by sending
9139 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
9141 This issue was discovered by Sean Burford (Google) and Emilia Käsper
9142 (OpenSSL development team).
9147 * Use After Free following d2i_ECPrivatekey error fix
9149 A malformed EC private key file consumed via the d2i_ECPrivateKey function
9150 could cause a use after free condition. This, in turn, could cause a double
9151 free in several private key parsing functions (such as d2i_PrivateKey
9152 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
9153 for applications that receive EC private keys from untrusted
9154 sources. This scenario is considered rare.
9156 This issue was discovered by the BoringSSL project and fixed in their
9162 * X509_to_X509_REQ NULL pointer deref fix
9164 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
9165 the certificate key is invalid. This function is rarely used in practice.
9167 This issue was discovered by Brian Carpenter.
9172 * Removed the export ciphers from the DEFAULT ciphers
9176 ### Changes between 1.0.0p and 1.0.0q [15 Jan 2015]
9178 * Build fixes for the Windows and OpenVMS platforms
9180 *Matt Caswell and Richard Levitte*
9182 ### Changes between 1.0.0o and 1.0.0p [8 Jan 2015]
9184 * Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
9185 message can cause a segmentation fault in OpenSSL due to a NULL pointer
9186 dereference. This could lead to a Denial Of Service attack. Thanks to
9187 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
9192 * Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
9193 dtls1_buffer_record function under certain conditions. In particular this
9194 could occur if an attacker sent repeated DTLS records with the same
9195 sequence number but for the next epoch. The memory leak could be exploited
9196 by an attacker in a Denial of Service attack through memory exhaustion.
9197 Thanks to Chris Mueller for reporting this issue.
9202 * Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
9203 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
9204 method would be set to NULL which could later result in a NULL pointer
9205 dereference. Thanks to Frank Schmirler for reporting this issue.
9210 * Abort handshake if server key exchange message is omitted for ephemeral
9213 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
9214 reporting this issue.
9219 * Remove non-export ephemeral RSA code on client and server. This code
9220 violated the TLS standard by allowing the use of temporary RSA keys in
9221 non-export ciphersuites and could be used by a server to effectively
9222 downgrade the RSA key length used to a value smaller than the server
9223 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
9224 INRIA or reporting this issue.
9229 * Fixed issue where DH client certificates are accepted without verification.
9230 An OpenSSL server will accept a DH certificate for client authentication
9231 without the certificate verify message. This effectively allows a client to
9232 authenticate without the use of a private key. This only affects servers
9233 which trust a client certificate authority which issues certificates
9234 containing DH keys: these are extremely rare and hardly ever encountered.
9235 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
9241 * Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
9242 results on some platforms, including x86_64. This bug occurs at random
9243 with a very low probability, and is not known to be exploitable in any
9244 way, though its exact impact is difficult to determine. Thanks to Pieter
9245 Wuille (Blockstream) who reported this issue and also suggested an initial
9246 fix. Further analysis was conducted by the OpenSSL development team and
9247 Adam Langley of Google. The final fix was developed by Andy Polyakov of
9248 the OpenSSL core team.
9253 * Fix various certificate fingerprint issues.
9255 By using non-DER or invalid encodings outside the signed portion of a
9256 certificate the fingerprint can be changed without breaking the signature.
9257 Although no details of the signed portion of the certificate can be changed
9258 this can cause problems with some applications: e.g. those using the
9259 certificate fingerprint for blacklists.
9261 1. Reject signatures with non zero unused bits.
9263 If the BIT STRING containing the signature has non zero unused bits reject
9264 the signature. All current signature algorithms require zero unused bits.
9266 2. Check certificate algorithm consistency.
9268 Check the AlgorithmIdentifier inside TBS matches the one in the
9269 certificate signature. NB: this will result in signature failure
9270 errors for some broken certificates.
9272 Thanks to Konrad Kraszewski from Google for reporting this issue.
9274 3. Check DSA/ECDSA signatures use DER.
9276 Re-encode DSA/ECDSA signatures and compare with the original received
9277 signature. Return an error if there is a mismatch.
9279 This will reject various cases including garbage after signature
9280 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
9281 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
9282 (negative or with leading zeroes).
9284 Further analysis was conducted and fixes were developed by Stephen Henson
9285 of the OpenSSL core team.
9291 ### Changes between 1.0.0n and 1.0.0o [15 Oct 2014]
9293 * Session Ticket Memory Leak.
9295 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
9296 integrity of that ticket is first verified. In the event of a session
9297 ticket integrity check failing, OpenSSL will fail to free memory
9298 causing a memory leak. By sending a large number of invalid session
9299 tickets an attacker could exploit this issue in a Denial Of Service
9305 * Build option no-ssl3 is incomplete.
9307 When OpenSSL is configured with "no-ssl3" as a build option, servers
9308 could accept and complete an SSL 3.0 handshake, and clients could be
9309 configured to send them.
9312 *Akamai and the OpenSSL team*
9314 * Add support for TLS_FALLBACK_SCSV.
9315 Client applications doing fallback retries should call
9316 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
9319 *Adam Langley, Bodo Moeller*
9321 * Add additional DigestInfo checks.
9323 Re-encode DigestInto in DER and check against the original when
9324 verifying RSA signature: this will reject any improperly encoded
9325 DigestInfo structures.
9327 Note: this is a precautionary measure and no attacks are currently known.
9331 ### Changes between 1.0.0m and 1.0.0n [6 Aug 2014]
9333 * OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
9334 to a denial of service attack. A malicious server can crash the client
9335 with a null pointer dereference (read) by specifying an anonymous (EC)DH
9336 ciphersuite and sending carefully crafted handshake messages.
9338 Thanks to Felix Gröbert (Google) for discovering and researching this
9344 * By sending carefully crafted DTLS packets an attacker could cause openssl
9345 to leak memory. This can be exploited through a Denial of Service attack.
9346 Thanks to Adam Langley for discovering and researching this issue.
9351 * An attacker can force openssl to consume large amounts of memory whilst
9352 processing DTLS handshake messages. This can be exploited through a
9353 Denial of Service attack.
9354 Thanks to Adam Langley for discovering and researching this issue.
9359 * An attacker can force an error condition which causes openssl to crash
9360 whilst processing DTLS packets due to memory being freed twice. This
9361 can be exploited through a Denial of Service attack.
9362 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
9368 * If a multithreaded client connects to a malicious server using a resumed
9369 session and the server sends an ec point format extension it could write
9370 up to 255 bytes to freed memory.
9372 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
9378 * A flaw in OBJ_obj2txt may cause pretty printing functions such as
9379 X509_name_oneline, X509_name_print_ex et al. to leak some information
9380 from the stack. Applications may be affected if they echo pretty printing
9381 output to the attacker.
9383 Thanks to Ivan Fratric (Google) for discovering this issue.
9386 *Emilia Käsper, and Steve Henson*
9388 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
9389 for corner cases. (Certain input points at infinity could lead to
9390 bogus results, with non-infinity inputs mapped to infinity too.)
9394 ### Changes between 1.0.0l and 1.0.0m [5 Jun 2014]
9396 * Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
9397 handshake can force the use of weak keying material in OpenSSL
9398 SSL/TLS clients and servers.
9400 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
9401 researching this issue. ([CVE-2014-0224])
9403 *KIKUCHI Masashi, Steve Henson*
9405 * Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
9406 OpenSSL DTLS client the code can be made to recurse eventually crashing
9409 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
9412 *Imre Rad, Steve Henson*
9414 * Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
9415 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
9416 client or server. This is potentially exploitable to run arbitrary
9417 code on a vulnerable client or server.
9419 Thanks to Jüri Aedla for reporting this issue. ([CVE-2014-0195])
9421 *Jüri Aedla, Steve Henson*
9423 * Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
9424 are subject to a denial of service attack.
9426 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
9427 this issue. ([CVE-2014-3470])
9429 *Felix Gröbert, Ivan Fratric, Steve Henson*
9431 * Harmonize version and its documentation. -f flag is used to display
9434 *mancha <mancha1@zoho.com>*
9436 * Fix eckey_priv_encode so it immediately returns an error upon a failure
9437 in i2d_ECPrivateKey.
9439 *mancha <mancha1@zoho.com>*
9441 * Fix some double frees. These are not thought to be exploitable.
9443 *mancha <mancha1@zoho.com>*
9445 * Fix for the attack described in the paper "Recovering OpenSSL
9446 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
9447 by Yuval Yarom and Naomi Benger. Details can be obtained from:
9448 <http://eprint.iacr.org/2014/140>
9450 Thanks to Yuval Yarom and Naomi Benger for discovering this
9451 flaw and to Yuval Yarom for supplying a fix ([CVE-2014-0076])
9453 *Yuval Yarom and Naomi Benger*
9455 ### Changes between 1.0.0k and 1.0.0l [6 Jan 2014]
9457 * Keep original DTLS digest and encryption contexts in retransmission
9458 structures so we can use the previous session parameters if they need
9459 to be resent. ([CVE-2013-6450])
9463 * Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
9464 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
9465 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
9466 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
9467 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
9468 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
9470 *Rob Stradling, Adam Langley*
9472 ### Changes between 1.0.0j and 1.0.0k [5 Feb 2013]
9474 * Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
9476 This addresses the flaw in CBC record processing discovered by
9477 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
9478 at: <http://www.isg.rhul.ac.uk/tls/>
9480 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
9481 Security Group at Royal Holloway, University of London
9482 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
9483 Emilia Käsper for the initial patch.
9486 *Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson*
9488 * Return an error when checking OCSP signatures when key is NULL.
9489 This fixes a DoS attack. ([CVE-2013-0166])
9493 * Call OCSP Stapling callback after ciphersuite has been chosen, so
9494 the right response is stapled. Also change SSL_get_certificate()
9495 so it returns the certificate actually sent.
9496 See <http://rt.openssl.org/Ticket/Display.html?id=2836>.
9497 (This is a backport)
9499 *Rob Stradling <rob.stradling@comodo.com>*
9501 * Fix possible deadlock when decoding public keys.
9505 ### Changes between 1.0.0i and 1.0.0j [10 May 2012]
9507 [NB: OpenSSL 1.0.0i and later 1.0.0 patch levels were released after
9510 * Sanity check record length before skipping explicit IV in DTLS
9513 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
9514 fuzzing as a service testing platform.
9519 * Initialise tkeylen properly when encrypting CMS messages.
9520 Thanks to Solar Designer of Openwall for reporting this issue.
9524 ### Changes between 1.0.0h and 1.0.0i [19 Apr 2012]
9526 * Check for potentially exploitable overflows in asn1_d2i_read_bio
9527 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
9528 in CRYPTO_realloc_clean.
9530 Thanks to Tavis Ormandy, Google Security Team, for discovering this
9531 issue and to Adam Langley <agl@chromium.org> for fixing it.
9534 *Adam Langley (Google), Tavis Ormandy, Google Security Team*
9536 ### Changes between 1.0.0g and 1.0.0h [12 Mar 2012]
9538 * Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness
9539 in CMS and PKCS7 code. When RSA decryption fails use a random key for
9540 content decryption and always return the same error. Note: this attack
9541 needs on average 2^20 messages so it only affects automated senders. The
9542 old behaviour can be re-enabled in the CMS code by setting the
9543 CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where
9544 an MMA defence is not necessary.
9545 Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering
9546 this issue. ([CVE-2012-0884])
9550 * Fix CVE-2011-4619: make sure we really are receiving a
9551 client hello before rejecting multiple SGC restarts. Thanks to
9552 Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug.
9556 ### Changes between 1.0.0f and 1.0.0g [18 Jan 2012]
9558 * Fix for DTLS DoS issue introduced by fix for CVE-2011-4109.
9559 Thanks to Antonio Martin, Enterprise Secure Access Research and
9560 Development, Cisco Systems, Inc. for discovering this bug and
9561 preparing a fix. ([CVE-2012-0050])
9565 ### Changes between 1.0.0e and 1.0.0f [4 Jan 2012]
9567 * Nadhem Alfardan and Kenny Paterson have discovered an extension
9568 of the Vaudenay padding oracle attack on CBC mode encryption
9569 which enables an efficient plaintext recovery attack against
9570 the OpenSSL implementation of DTLS. Their attack exploits timing
9571 differences arising during decryption processing. A research
9572 paper describing this attack can be found at:
9573 <http://www.isg.rhul.ac.uk/~kp/dtls.pdf>
9574 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
9575 Security Group at Royal Holloway, University of London
9576 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
9577 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
9578 for preparing the fix. ([CVE-2011-4108])
9580 *Robin Seggelmann, Michael Tuexen*
9582 * Clear bytes used for block padding of SSL 3.0 records.
9585 *Adam Langley (Google)*
9587 * Only allow one SGC handshake restart for SSL/TLS. Thanks to George
9588 Kadianakis <desnacked@gmail.com> for discovering this issue and
9589 Adam Langley for preparing the fix. ([CVE-2011-4619])
9591 *Adam Langley (Google)*
9593 * Check parameters are not NULL in GOST ENGINE. ([CVE-2012-0027])
9595 *Andrey Kulikov <amdeich@gmail.com>*
9597 * Prevent malformed RFC3779 data triggering an assertion failure.
9598 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
9599 and Rob Austein <sra@hactrn.net> for fixing it. ([CVE-2011-4577])
9601 *Rob Austein <sra@hactrn.net>*
9603 * Improved PRNG seeding for VOS.
9605 *Paul Green <Paul.Green@stratus.com>*
9607 * Fix ssl_ciph.c set-up race.
9609 *Adam Langley (Google)*
9611 * Fix spurious failures in ecdsatest.c.
9613 *Emilia Käsper (Google)*
9615 * Fix the BIO_f_buffer() implementation (which was mixing different
9616 interpretations of the `..._len` fields).
9618 *Adam Langley (Google)*
9620 * Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
9621 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
9622 threads won't reuse the same blinding coefficients.
9624 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
9625 lock to call BN_BLINDING_invert_ex, and avoids one use of
9626 BN_BLINDING_update for each BN_BLINDING structure (previously,
9627 the last update always remained unused).
9629 *Emilia Käsper (Google)*
9631 * In ssl3_clear, preserve s3->init_extra along with s3->rbuf.
9633 *Bob Buckholz (Google)*
9635 ### Changes between 1.0.0d and 1.0.0e [6 Sep 2011]
9637 * Fix bug where CRLs with nextUpdate in the past are sometimes accepted
9638 by initialising X509_STORE_CTX properly. ([CVE-2011-3207])
9640 *Kaspar Brand <ossl@velox.ch>*
9642 * Fix SSL memory handling for (EC)DH ciphersuites, in particular
9643 for multi-threaded use of ECDH. ([CVE-2011-3210])
9645 *Adam Langley (Google)*
9647 * Fix x509_name_ex_d2i memory leak on bad inputs.
9651 * Remove hard coded ecdsaWithSHA1 signature tests in ssl code and check
9652 signature public key algorithm by using OID xref utilities instead.
9653 Before this you could only use some ECC ciphersuites with SHA1 only.
9657 * Add protection against ECDSA timing attacks as mentioned in the paper
9658 by Billy Bob Brumley and Nicola Tuveri, see:
9659 <http://eprint.iacr.org/2011/232.pdf>
9661 *Billy Bob Brumley and Nicola Tuveri*
9663 ### Changes between 1.0.0c and 1.0.0d [8 Feb 2011]
9665 * Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
9667 *Neel Mehta, Adam Langley, Bodo Moeller (Google)*
9669 * Fix bug in string printing code: if *any* escaping is enabled we must
9670 escape the escape character (backslash) or the resulting string is
9675 ### Changes between 1.0.0b and 1.0.0c [2 Dec 2010]
9677 * Disable code workaround for ancient and obsolete Netscape browsers
9678 and servers: an attacker can use it in a ciphersuite downgrade attack.
9679 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
9683 * Fixed J-PAKE implementation error, originally discovered by
9684 Sebastien Martini, further info and confirmation from Stefan
9685 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
9689 ### Changes between 1.0.0a and 1.0.0b [16 Nov 2010]
9691 * Fix extension code to avoid race conditions which can result in a buffer
9692 overrun vulnerability: resumed sessions must not be modified as they can
9693 be shared by multiple threads. CVE-2010-3864
9697 * Fix WIN32 build system to correctly link an ENGINE directory into
9702 ### Changes between 1.0.0 and 1.0.0a [01 Jun 2010]
9704 * Check return value of int_rsa_verify in pkey_rsa_verifyrecover
9707 *Steve Henson, Peter-Michael Hager <hager@dortmund.net>*
9709 ### Changes between 0.9.8n and 1.0.0 [29 Mar 2010]
9711 * Add "missing" function EVP_CIPHER_CTX_copy(). This copies a cipher
9712 context. The operation can be customised via the ctrl mechanism in
9713 case ENGINEs want to include additional functionality.
9717 * Tolerate yet another broken PKCS#8 key format: private key value negative.
9721 * Add new -subject_hash_old and -issuer_hash_old options to x509 utility to
9722 output hashes compatible with older versions of OpenSSL.
9724 *Willy Weisz <weisz@vcpc.univie.ac.at>*
9726 * Fix compression algorithm handling: if resuming a session use the
9727 compression algorithm of the resumed session instead of determining
9728 it from client hello again. Don't allow server to change algorithm.
9732 * Add load_crls() function to commands tidying load_certs() too. Add option
9733 to verify utility to allow additional CRLs to be included.
9737 * Update OCSP request code to permit adding custom headers to the request:
9738 some responders need this.
9742 * The function EVP_PKEY_sign() returns <=0 on error: check return code
9745 *Julia Lawall <julia@diku.dk>*
9747 * Update verify callback code in `apps/s_cb.c` and `apps/verify.c`, it
9748 needlessly dereferenced structures, used obsolete functions and
9749 didn't handle all updated verify codes correctly.
9753 * Disable MD2 in the default configuration.
9757 * In BIO_pop() and BIO_push() use the ctrl argument (which was NULL) to
9758 indicate the initial BIO being pushed or popped. This makes it possible
9759 to determine whether the BIO is the one explicitly called or as a result
9760 of the ctrl being passed down the chain. Fix BIO_pop() and SSL BIOs so
9761 it handles reference counts correctly and doesn't zero out the I/O bio
9762 when it is not being explicitly popped. WARNING: applications which
9763 included workarounds for the old buggy behaviour will need to be modified
9764 or they could free up already freed BIOs.
9768 * Extend the uni2asc/asc2uni => OPENSSL_uni2asc/OPENSSL_asc2uni
9769 renaming to all platforms (within the 0.9.8 branch, this was
9770 done conditionally on Netware platforms to avoid a name clash).
9772 *Guenter <lists@gknw.net>*
9774 * Add ECDHE and PSK support to DTLS.
9776 *Michael Tuexen <tuexen@fh-muenster.de>*
9778 * Add CHECKED_STACK_OF macro to safestack.h, otherwise safestack can't
9783 * Add "missing" function EVP_MD_flags() (without this the only way to
9784 retrieve a digest flags is by accessing the structure directly. Update
9785 `EVP_MD_do_all*()` and `EVP_CIPHER_do_all*()` to include the name a digest
9786 or cipher is registered as in the "from" argument. Print out all
9787 registered digests in the dgst usage message instead of manually
9788 attempting to work them out.
9792 * If no SSLv2 ciphers are used don't use an SSLv2 compatible client hello:
9793 this allows the use of compression and extensions. Change default cipher
9794 string to remove SSLv2 ciphersuites. This effectively avoids ancient SSLv2
9795 by default unless an application cipher string requests it.
9799 * Alter match criteria in PKCS12_parse(). It used to try to use local
9800 key ids to find matching certificates and keys but some PKCS#12 files
9801 don't follow the (somewhat unwritten) rules and this strategy fails.
9802 Now just gather all certificates together and the first private key
9803 then look for the first certificate that matches the key.
9807 * Support use of registered digest and cipher names for dgst and cipher
9808 commands instead of having to add each one as a special case. So now
9815 openssl dgst -sha256 foo
9817 and this works for ENGINE based algorithms too.
9821 * Update Gost ENGINE to support parameter files.
9823 *Victor B. Wagner <vitus@cryptocom.ru>*
9825 * Support GeneralizedTime in ca utility.
9827 *Oliver Martin <oliver@volatilevoid.net>, Steve Henson*
9829 * Enhance the hash format used for certificate directory links. The new
9830 form uses the canonical encoding (meaning equivalent names will work
9831 even if they aren't identical) and uses SHA1 instead of MD5. This form
9832 is incompatible with the older format and as a result c_rehash should
9833 be used to rebuild symbolic links.
9837 * Make PKCS#8 the default write format for private keys, replacing the
9838 traditional format. This form is standardised, more secure and doesn't
9839 include an implicit MD5 dependency.
9843 * Add a $gcc_devteam_warn option to Configure. The idea is that any code
9844 committed to OpenSSL should pass this lot as a minimum.
9848 * Add session ticket override functionality for use by EAP-FAST.
9850 *Jouni Malinen <j@w1.fi>*
9852 * Modify HMAC functions to return a value. Since these can be implemented
9853 in an ENGINE errors can occur.
9857 * Type-checked OBJ_bsearch_ex.
9861 * Type-checked OBJ_bsearch. Also some constification necessitated
9862 by type-checking. Still to come: TXT_DB, bsearch(?),
9863 OBJ_bsearch_ex, qsort, CRYPTO_EX_DATA, ASN1_VALUE, ASN1_STRING,
9868 * New function OPENSSL_gmtime_adj() to add a specific number of days and
9869 seconds to a tm structure directly, instead of going through OS
9870 specific date routines. This avoids any issues with OS routines such
9871 as the year 2038 bug. New `*_adj()` functions for ASN1 time structures
9872 and X509_time_adj_ex() to cover the extended range. The existing
9873 X509_time_adj() is still usable and will no longer have any date issues.
9877 * Delta CRL support. New use deltas option which will attempt to locate
9878 and search any appropriate delta CRLs available.
9880 This work was sponsored by Google.
9884 * Support for CRLs partitioned by reason code. Reorganise CRL processing
9885 code and add additional score elements. Validate alternate CRL paths
9886 as part of the CRL checking and indicate a new error "CRL path validation
9887 error" in this case. Applications wanting additional details can use
9888 the verify callback and check the new "parent" field. If this is not
9889 NULL CRL path validation is taking place. Existing applications won't
9890 see this because it requires extended CRL support which is off by
9893 This work was sponsored by Google.
9897 * Support for freshest CRL extension.
9899 This work was sponsored by Google.
9903 * Initial indirect CRL support. Currently only supported in the CRLs
9904 passed directly and not via lookup. Process certificate issuer
9905 CRL entry extension and lookup CRL entries by bother issuer name
9906 and serial number. Check and process CRL issuer entry in IDP extension.
9908 This work was sponsored by Google.
9912 * Add support for distinct certificate and CRL paths. The CRL issuer
9913 certificate is validated separately in this case. Only enabled if
9914 an extended CRL support flag is set: this flag will enable additional
9915 CRL functionality in future.
9917 This work was sponsored by Google.
9921 * Add support for policy mappings extension.
9923 This work was sponsored by Google.
9927 * Fixes to pathlength constraint, self issued certificate handling,
9928 policy processing to align with RFC3280 and PKITS tests.
9930 This work was sponsored by Google.
9934 * Support for name constraints certificate extension. DN, email, DNS
9935 and URI types are currently supported.
9937 This work was sponsored by Google.
9941 * To cater for systems that provide a pointer-based thread ID rather
9942 than numeric, deprecate the current numeric thread ID mechanism and
9943 replace it with a structure and associated callback type. This
9944 mechanism allows a numeric "hash" to be extracted from a thread ID in
9945 either case, and on platforms where pointers are larger than 'long',
9946 mixing is done to help ensure the numeric 'hash' is usable even if it
9947 can't be guaranteed unique. The default mechanism is to use "&errno"
9948 as a pointer-based thread ID to distinguish between threads.
9950 Applications that want to provide their own thread IDs should now use
9951 CRYPTO_THREADID_set_callback() to register a callback that will call
9952 either CRYPTO_THREADID_set_numeric() or CRYPTO_THREADID_set_pointer().
9954 Note that ERR_remove_state() is now deprecated, because it is tied
9955 to the assumption that thread IDs are numeric. ERR_remove_state(0)
9956 to free the current thread's error state should be replaced by
9957 ERR_remove_thread_state(NULL).
9959 (This new approach replaces the functions CRYPTO_set_idptr_callback(),
9960 CRYPTO_get_idptr_callback(), and CRYPTO_thread_idptr() that existed in
9961 OpenSSL 0.9.9-dev between June 2006 and August 2008. Also, if an
9962 application was previously providing a numeric thread callback that
9963 was inappropriate for distinguishing threads, then uniqueness might
9964 have been obtained with &errno that happened immediately in the
9965 intermediate development versions of OpenSSL; this is no longer the
9966 case, the numeric thread callback will now override the automatic use
9969 *Geoff Thorpe, with help from Bodo Moeller*
9971 * Initial support for different CRL issuing certificates. This covers a
9972 simple case where the self issued certificates in the chain exist and
9973 the real CRL issuer is higher in the existing chain.
9975 This work was sponsored by Google.
9979 * Removed effectively defunct crypto/store from the build.
9983 * Revamp of STACK to provide stronger type-checking. Still to come:
9984 TXT_DB, bsearch(?), OBJ_bsearch, qsort, CRYPTO_EX_DATA, ASN1_VALUE,
9985 ASN1_STRING, CONF_VALUE.
9989 * Add a new SSL_MODE_RELEASE_BUFFERS mode flag to release unused buffer
9990 RAM on SSL connections. This option can save about 34k per idle SSL.
9994 * Revamp of LHASH to provide stronger type-checking. Still to come:
9995 STACK, TXT_DB, bsearch, qsort.
9999 * Initial support for Cryptographic Message Syntax (aka CMS) based
10000 on RFC3850, RFC3851 and RFC3852. New cms directory and cms utility,
10001 support for data, signedData, compressedData, digestedData and
10002 encryptedData, envelopedData types included. Scripts to check against
10003 RFC4134 examples draft and interop and consistency checks of many
10004 content types and variants.
10008 * Add options to enc utility to support use of zlib compression BIO.
10012 * Extend mk1mf to support importing of options and assembly language
10013 files from Configure script, currently only included in VC-WIN32.
10014 The assembly language rules can now optionally generate the source
10015 files from the associated perl scripts.
10019 * Implement remaining functionality needed to support GOST ciphersuites.
10020 Interop testing has been performed using CryptoPro implementations.
10022 *Victor B. Wagner <vitus@cryptocom.ru>*
10024 * s390x assembler pack.
10028 * ARMv4 assembler pack. ARMv4 refers to v4 and later ISA, not CPU
10033 * Implement Opaque PRF Input TLS extension as specified in
10034 draft-rescorla-tls-opaque-prf-input-00.txt. Since this is not an
10035 official specification yet and no extension type assignment by
10036 IANA exists, this extension (for now) will have to be explicitly
10037 enabled when building OpenSSL by providing the extension number
10038 to use. For example, specify an option
10040 -DTLSEXT_TYPE_opaque_prf_input=0x9527
10042 to the "config" or "Configure" script to enable the extension,
10043 assuming extension number 0x9527 (which is a completely arbitrary
10044 and unofficial assignment based on the MD5 hash of the Internet
10045 Draft). Note that by doing so, you potentially lose
10046 interoperability with other TLS implementations since these might
10047 be using the same extension number for other purposes.
10049 SSL_set_tlsext_opaque_prf_input(ssl, src, len) is used to set the
10050 opaque PRF input value to use in the handshake. This will create
10051 an internal copy of the length-'len' string at 'src', and will
10052 return non-zero for success.
10054 To get more control and flexibility, provide a callback function
10057 SSL_CTX_set_tlsext_opaque_prf_input_callback(ctx, cb)
10058 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg(ctx, arg)
10062 int (*cb)(SSL *, void *peerinput, size_t len, void *arg);
10065 Callback function 'cb' will be called in handshakes, and is
10066 expected to use SSL_set_tlsext_opaque_prf_input() as appropriate.
10067 Argument 'arg' is for application purposes (the value as given to
10068 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg() will directly
10069 be provided to the callback function). The callback function
10070 has to return non-zero to report success: usually 1 to use opaque
10071 PRF input just if possible, or 2 to enforce use of the opaque PRF
10072 input. In the latter case, the library will abort the handshake
10073 if opaque PRF input is not successfully negotiated.
10075 Arguments 'peerinput' and 'len' given to the callback function
10076 will always be NULL and 0 in the case of a client. A server will
10077 see the client's opaque PRF input through these variables if
10078 available (NULL and 0 otherwise). Note that if the server
10079 provides an opaque PRF input, the length must be the same as the
10080 length of the client's opaque PRF input.
10082 Note that the callback function will only be called when creating
10083 a new session (session resumption can resume whatever was
10084 previously negotiated), and will not be called in SSL 2.0
10085 handshakes; thus, SSL_CTX_set_options(ctx, SSL_OP_NO_SSLv2) or
10086 SSL_set_options(ssl, SSL_OP_NO_SSLv2) is especially recommended
10087 for applications that need to enforce opaque PRF input.
10091 * Update ssl code to support digests other than SHA1+MD5 for handshake
10094 *Victor B. Wagner <vitus@cryptocom.ru>*
10096 * Add RFC4507 support to OpenSSL. This includes the corrections in
10097 RFC4507bis. The encrypted ticket format is an encrypted encoded
10098 SSL_SESSION structure, that way new session features are automatically
10101 If a client application caches session in an SSL_SESSION structure
10102 support is transparent because tickets are now stored in the encoded
10105 The SSL_CTX structure automatically generates keys for ticket
10106 protection in servers so again support should be possible
10107 with no application modification.
10109 If a client or server wishes to disable RFC4507 support then the option
10110 SSL_OP_NO_TICKET can be set.
10112 Add a TLS extension debugging callback to allow the contents of any client
10113 or server extensions to be examined.
10115 This work was sponsored by Google.
10119 * Final changes to avoid use of pointer pointer casts in OpenSSL.
10120 OpenSSL should now compile cleanly on gcc 4.2
10122 *Peter Hartley <pdh@utter.chaos.org.uk>, Steve Henson*
10124 * Update SSL library to use new EVP_PKEY MAC API. Include generic MAC
10125 support including streaming MAC support: this is required for GOST
10126 ciphersuite support.
10128 *Victor B. Wagner <vitus@cryptocom.ru>, Steve Henson*
10130 * Add option -stream to use PKCS#7 streaming in smime utility. New
10131 function i2d_PKCS7_bio_stream() and PEM_write_PKCS7_bio_stream()
10132 to output in BER and PEM format.
10136 * Experimental support for use of HMAC via EVP_PKEY interface. This
10137 allows HMAC to be handled via the `EVP_DigestSign*()` interface. The
10138 EVP_PKEY "key" in this case is the HMAC key, potentially allowing
10139 ENGINE support for HMAC keys which are unextractable. New -mac and
10140 -macopt options to dgst utility.
10144 * New option -sigopt to dgst utility. Update dgst to use
10145 `EVP_Digest{Sign,Verify}*`. These two changes make it possible to use
10146 alternative signing parameters such as X9.31 or PSS in the dgst
10151 * Change ssl_cipher_apply_rule(), the internal function that does
10152 the work each time a ciphersuite string requests enabling
10153 ("foo+bar"), moving ("+foo+bar"), disabling ("-foo+bar", or
10154 removing ("!foo+bar") a class of ciphersuites: Now it maintains
10155 the order of disabled ciphersuites such that those ciphersuites
10156 that most recently went from enabled to disabled not only stay
10157 in order with respect to each other, but also have higher priority
10158 than other disabled ciphersuites the next time ciphersuites are
10161 This means that you can now say, e.g., "PSK:-PSK:HIGH" to enable
10162 the same ciphersuites as with "HIGH" alone, but in a specific
10163 order where the PSK ciphersuites come first (since they are the
10164 most recently disabled ciphersuites when "HIGH" is parsed).
10166 Also, change ssl_create_cipher_list() (using this new
10167 functionality) such that between otherwise identical
10168 ciphersuites, ephemeral ECDH is preferred over ephemeral DH in
10173 * Change ssl_create_cipher_list() so that it automatically
10174 arranges the ciphersuites in reasonable order before starting
10175 to process the rule string. Thus, the definition for "DEFAULT"
10176 (SSL_DEFAULT_CIPHER_LIST) now is just "ALL:!aNULL:!eNULL", but
10177 remains equivalent to `"AES:ALL:!aNULL:!eNULL:+aECDH:+kRSA:+RC4:@STRENGTH"`.
10178 This makes it much easier to arrive at a reasonable default order
10179 in applications for which anonymous ciphers are OK (meaning
10180 that you can't actually use DEFAULT).
10182 *Bodo Moeller; suggested by Victor Duchovni*
10184 * Split the SSL/TLS algorithm mask (as used for ciphersuite string
10185 processing) into multiple integers instead of setting
10186 "SSL_MKEY_MASK" bits, "SSL_AUTH_MASK" bits, "SSL_ENC_MASK",
10187 "SSL_MAC_MASK", and "SSL_SSL_MASK" bits all in a single integer.
10188 (These masks as well as the individual bit definitions are hidden
10189 away into the non-exported interface ssl/ssl_locl.h, so this
10190 change to the definition of the SSL_CIPHER structure shouldn't
10191 affect applications.) This give us more bits for each of these
10192 categories, so there is no longer a need to coagulate AES128 and
10193 AES256 into a single algorithm bit, and to coagulate Camellia128
10194 and Camellia256 into a single algorithm bit, which has led to all
10197 Thus, among other things, the kludge introduced in 0.9.7m and
10198 0.9.8e for masking out AES256 independently of AES128 or masking
10199 out Camellia256 independently of AES256 is not needed here in 0.9.9.
10201 With the change, we also introduce new ciphersuite aliases that
10202 so far were missing: "AES128", "AES256", "CAMELLIA128", and
10207 * Add support for dsa-with-SHA224 and dsa-with-SHA256.
10208 Use the leftmost N bytes of the signature input if the input is
10209 larger than the prime q (with N being the size in bytes of q).
10213 * Very *very* experimental PKCS#7 streaming encoder support. Nothing uses
10214 it yet and it is largely untested.
10218 * Add support for the ecdsa-with-SHA224/256/384/512 signature types.
10222 * Initial incomplete changes to avoid need for function casts in OpenSSL
10223 some compilers (gcc 4.2 and later) reject their use. Safestack is
10224 reimplemented. Update ASN1 to avoid use of legacy functions.
10228 * Win32/64 targets are linked with Winsock2.
10232 * Add an X509_CRL_METHOD structure to allow CRL processing to be redirected
10233 to external functions. This can be used to increase CRL handling
10234 efficiency especially when CRLs are very large by (for example) storing
10235 the CRL revoked certificates in a database.
10239 * Overhaul of by_dir code. Add support for dynamic loading of CRLs so
10240 new CRLs added to a directory can be used. New command line option
10241 -verify_return_error to s_client and s_server. This causes real errors
10242 to be returned by the verify callback instead of carrying on no matter
10243 what. This reflects the way a "real world" verify callback would behave.
10247 * GOST engine, supporting several GOST algorithms and public key formats.
10248 Kindly donated by Cryptocom.
10252 * Partial support for Issuing Distribution Point CRL extension. CRLs
10253 partitioned by DP are handled but no indirect CRL or reason partitioning
10254 (yet). Complete overhaul of CRL handling: now the most suitable CRL is
10255 selected via a scoring technique which handles IDP and AKID in CRLs.
10259 * New X509_STORE_CTX callbacks lookup_crls() and lookup_certs() which
10260 will ultimately be used for all verify operations: this will remove the
10261 X509_STORE dependency on certificate verification and allow alternative
10262 lookup methods. X509_STORE based implementations of these two callbacks.
10266 * Allow multiple CRLs to exist in an X509_STORE with matching issuer names.
10267 Modify get_crl() to find a valid (unexpired) CRL if possible.
10271 * New function X509_CRL_match() to check if two CRLs are identical. Normally
10272 this would be called X509_CRL_cmp() but that name is already used by
10273 a function that just compares CRL issuer names. Cache several CRL
10274 extensions in X509_CRL structure and cache CRLDP in X509.
10278 * Store a "canonical" representation of X509_NAME structure (ASN1 Name)
10279 this maps equivalent X509_NAME structures into a consistent structure.
10280 Name comparison can then be performed rapidly using memcmp().
10284 * Non-blocking OCSP request processing. Add -timeout option to ocsp
10289 * Allow digests to supply their own micalg string for S/MIME type using
10290 the ctrl EVP_MD_CTRL_MICALG.
10294 * During PKCS7 signing pass the PKCS7 SignerInfo structure to the
10295 EVP_PKEY_METHOD before and after signing via the EVP_PKEY_CTRL_PKCS7_SIGN
10296 ctrl. It can then customise the structure before and/or after signing
10301 * New function OBJ_add_sigid() to allow application defined signature OIDs
10302 to be added to OpenSSLs internal tables. New function OBJ_sigid_free()
10303 to free up any added signature OIDs.
10307 * New functions EVP_CIPHER_do_all(), EVP_CIPHER_do_all_sorted(),
10308 EVP_MD_do_all() and EVP_MD_do_all_sorted() to enumerate internal
10309 digest and cipher tables. New options added to openssl utility:
10310 list-message-digest-algorithms and list-cipher-algorithms.
10314 * Change the array representation of binary polynomials: the list
10315 of degrees of non-zero coefficients is now terminated with -1.
10316 Previously it was terminated with 0, which was also part of the
10317 value; thus, the array representation was not applicable to
10318 polynomials where t^0 has coefficient zero. This change makes
10319 the array representation useful in a more general context.
10323 * Various modifications and fixes to SSL/TLS cipher string
10324 handling. For ECC, the code now distinguishes between fixed ECDH
10325 with RSA certificates on the one hand and with ECDSA certificates
10326 on the other hand, since these are separate ciphersuites. The
10327 unused code for Fortezza ciphersuites has been removed.
10329 For consistency with EDH, ephemeral ECDH is now called "EECDH"
10330 (not "ECDHE"). For consistency with the code for DH
10331 certificates, use of ECDH certificates is now considered ECDH
10332 authentication, not RSA or ECDSA authentication (the latter is
10333 merely the CA's signing algorithm and not actively used in the
10336 The temporary ciphersuite alias "ECCdraft" is no longer
10337 available, and ECC ciphersuites are no longer excluded from "ALL"
10338 and "DEFAULT". The following aliases now exist for RFC 4492
10339 ciphersuites, most of these by analogy with the DH case:
10341 kECDHr - ECDH cert, signed with RSA
10342 kECDHe - ECDH cert, signed with ECDSA
10343 kECDH - ECDH cert (signed with either RSA or ECDSA)
10344 kEECDH - ephemeral ECDH
10345 ECDH - ECDH cert or ephemeral ECDH
10348 aECDSA - ECDSA cert
10351 AECDH - anonymous ECDH
10352 EECDH - non-anonymous ephemeral ECDH (equivalent to "kEECDH:-AECDH")
10356 * Add additional S/MIME capabilities for AES and GOST ciphers if supported.
10357 Use correct micalg parameters depending on digest(s) in signed message.
10361 * Add engine support for EVP_PKEY_ASN1_METHOD. Add functions to process
10362 an ENGINE asn1 method. Support ENGINE lookups in the ASN1 code.
10366 * Initial engine support for EVP_PKEY_METHOD. New functions to permit
10367 an engine to register a method. Add ENGINE lookups for methods and
10368 functional reference processing.
10372 * New functions `EVP_Digest{Sign,Verify)*`. These are enhanced versions of
10373 `EVP_{Sign,Verify}*` which allow an application to customise the signature
10378 * New -resign option to smime utility. This adds one or more signers
10379 to an existing PKCS#7 signedData structure. Also -md option to use an
10380 alternative message digest algorithm for signing.
10384 * Tidy up PKCS#7 routines and add new functions to make it easier to
10385 create PKCS7 structures containing multiple signers. Update smime
10386 application to support multiple signers.
10390 * New -macalg option to pkcs12 utility to allow setting of an alternative
10395 * Initial support for PKCS#5 v2.0 PRFs other than default SHA1 HMAC.
10396 Reorganize PBE internals to lookup from a static table using NIDs,
10397 add support for HMAC PBE OID translation. Add a EVP_CIPHER ctrl:
10398 EVP_CTRL_PBE_PRF_NID this allows a cipher to specify an alternative
10399 PRF which will be automatically used with PBES2.
10403 * Replace the algorithm specific calls to generate keys in "req" with the
10408 * Update PKCS#7 enveloped data routines to use new API. This is now
10409 supported by any public key method supporting the encrypt operation. A
10410 ctrl is added to allow the public key algorithm to examine or modify
10411 the PKCS#7 RecipientInfo structure if it needs to: for RSA this is
10416 * Add a ctrl to asn1 method to allow a public key algorithm to express
10417 a default digest type to use. In most cases this will be SHA1 but some
10418 algorithms (such as GOST) need to specify an alternative digest. The
10419 return value indicates how strong the preference is 1 means optional and
10420 2 is mandatory (that is it is the only supported type). Modify
10421 ASN1_item_sign() to accept a NULL digest argument to indicate it should
10422 use the default md. Update openssl utilities to use the default digest
10423 type for signing if it is not explicitly indicated.
10427 * Use OID cross reference table in ASN1_sign() and ASN1_verify(). New
10428 EVP_MD flag EVP_MD_FLAG_PKEY_METHOD_SIGNATURE. This uses the relevant
10429 signing method from the key type. This effectively removes the link
10430 between digests and public key types.
10434 * Add an OID cross reference table and utility functions. Its purpose is to
10435 translate between signature OIDs such as SHA1WithrsaEncryption and SHA1,
10436 rsaEncryption. This will allow some of the algorithm specific hackery
10437 needed to use the correct OID to be removed.
10441 * Remove algorithm specific dependencies when setting PKCS7_SIGNER_INFO
10442 structures for PKCS7_sign(). They are now set up by the relevant public
10447 * Add provisional EC pkey method with support for ECDSA and ECDH.
10451 * Add support for key derivation (agreement) in the API, DH method and
10456 * Add DSA pkey method and DH pkey methods, extend DH ASN1 method to support
10457 public and private key formats. As a side effect these add additional
10458 command line functionality not previously available: DSA signatures can be
10459 generated and verified using pkeyutl and DH key support and generation in
10466 *Oliver Tappe <zooey@hirschkaefer.de>*
10468 * New make target "install_html_docs" installs HTML renditions of the
10471 *Oliver Tappe <zooey@hirschkaefer.de>*
10473 * New utility "genpkey" this is analogous to "genrsa" etc except it can
10474 generate keys for any algorithm. Extend and update EVP_PKEY_METHOD to
10475 support key and parameter generation and add initial key generation
10476 functionality for RSA.
10480 * Add functions for main EVP_PKEY_method operations. The undocumented
10481 functions `EVP_PKEY_{encrypt,decrypt}` have been renamed to
10482 `EVP_PKEY_{encrypt,decrypt}_old`.
10486 * Initial definitions for EVP_PKEY_METHOD. This will be a high level public
10487 key API, doesn't do much yet.
10491 * New function EVP_PKEY_asn1_get0_info() to retrieve information about
10492 public key algorithms. New option to openssl utility:
10493 "list-public-key-algorithms" to print out info.
10497 * Implement the Supported Elliptic Curves Extension for
10498 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
10502 * Don't free up OIDs in OBJ_cleanup() if they are in use by EVP_MD or
10503 EVP_CIPHER structures to avoid later problems in EVP_cleanup().
10507 * New utilities pkey and pkeyparam. These are similar to algorithm specific
10508 utilities such as rsa, dsa, dsaparam etc except they process any key
10513 * Transfer public key printing routines to EVP_PKEY_ASN1_METHOD. New
10514 functions EVP_PKEY_print_public(), EVP_PKEY_print_private(),
10515 EVP_PKEY_print_param() to print public key data from an EVP_PKEY
10520 * Initial support for pluggable public key ASN1.
10521 De-spaghettify the public key ASN1 handling. Move public and private
10522 key ASN1 handling to a new EVP_PKEY_ASN1_METHOD structure. Relocate
10523 algorithm specific handling to a single module within the relevant
10524 algorithm directory. Add functions to allow (near) opaque processing
10525 of public and private key structures.
10529 * Implement the Supported Point Formats Extension for
10530 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
10534 * Add initial support for RFC 4279 PSK TLS ciphersuites. Add members
10535 for the psk identity [hint] and the psk callback functions to the
10536 SSL_SESSION, SSL and SSL_CTX structure.
10539 PSK-RC4-SHA, PSK-3DES-EDE-CBC-SHA, PSK-AES128-CBC-SHA,
10543 SSL_CTX_use_psk_identity_hint
10544 SSL_get_psk_identity_hint
10545 SSL_get_psk_identity
10546 SSL_use_psk_identity_hint
10548 *Mika Kousa and Pasi Eronen of Nokia Corporation*
10550 * Add RFC 3161 compliant time stamp request creation, response generation
10551 and response verification functionality.
10553 *Zoltán Glózik <zglozik@opentsa.org>, The OpenTSA Project*
10555 * Add initial support for TLS extensions, specifically for the server_name
10556 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
10557 have new members for a hostname. The SSL data structure has an
10558 additional member `SSL_CTX *initial_ctx` so that new sessions can be
10559 stored in that context to allow for session resumption, even after the
10560 SSL has been switched to a new SSL_CTX in reaction to a client's
10561 server_name extension.
10563 New functions (subject to change):
10565 SSL_get_servername()
10566 SSL_get_servername_type()
10569 New CTRL codes and macros (subject to change):
10571 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
10572 - SSL_CTX_set_tlsext_servername_callback()
10573 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
10574 - SSL_CTX_set_tlsext_servername_arg()
10575 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
10577 openssl s_client has a new '-servername ...' option.
10579 openssl s_server has new options '-servername_host ...', '-cert2 ...',
10580 '-key2 ...', '-servername_fatal' (subject to change). This allows
10581 testing the HostName extension for a specific single hostname ('-cert'
10582 and '-key' remain fallbacks for handshakes without HostName
10583 negotiation). If the unrecognized_name alert has to be sent, this by
10584 default is a warning; it becomes fatal with the '-servername_fatal'
10587 *Peter Sylvester, Remy Allais, Christophe Renou*
10589 * Whirlpool hash implementation is added.
10593 * BIGNUM code on 64-bit SPARCv9 targets is switched from bn(64,64) to
10594 bn(64,32). Because of instruction set limitations it doesn't have
10595 any negative impact on performance. This was done mostly in order
10596 to make it possible to share assembler modules, such as bn_mul_mont
10597 implementations, between 32- and 64-bit builds without hassle.
10601 * Move code previously exiled into file crypto/ec/ec2_smpt.c
10602 to ec2_smpl.c, and no longer require the OPENSSL_EC_BIN_PT_COMP
10607 * New candidate for BIGNUM assembler implementation, bn_mul_mont,
10608 dedicated Montgomery multiplication procedure, is introduced.
10609 BN_MONT_CTX is modified to allow bn_mul_mont to reach for higher
10610 "64-bit" performance on certain 32-bit targets.
10614 * New option SSL_OP_NO_COMP to disable use of compression selectively
10615 in SSL structures. New SSL ctrl to set maximum send fragment size.
10616 Save memory by setting the I/O buffer sizes dynamically instead of
10617 using the maximum available value.
10621 * New option -V for 'openssl ciphers'. This prints the ciphersuite code
10622 in addition to the text details.
10626 * Very, very preliminary EXPERIMENTAL support for printing of general
10627 ASN1 structures. This currently produces rather ugly output and doesn't
10628 handle several customised structures at all.
10632 * Integrated support for PVK file format and some related formats such
10633 as MS PUBLICKEYBLOB and PRIVATEKEYBLOB. Command line switches to support
10634 these in the 'rsa' and 'dsa' utilities.
10638 * Support for PKCS#1 RSAPublicKey format on rsa utility command line.
10642 * Remove the ancient ASN1_METHOD code. This was only ever used in one
10643 place for the (very old) "NETSCAPE" format certificates which are now
10644 handled using new ASN1 code equivalents.
10648 * Let the TLSv1_method() etc. functions return a 'const' SSL_METHOD
10649 pointer and make the SSL_METHOD parameter in SSL_CTX_new,
10650 SSL_CTX_set_ssl_version and SSL_set_ssl_method 'const'.
10654 * Modify CRL distribution points extension code to print out previously
10655 unsupported fields. Enhance extension setting code to allow setting of
10660 * Add print and set support for Issuing Distribution Point CRL extension.
10664 * Change 'Configure' script to enable Camellia by default.
10671 ### Changes between 0.9.8m and 0.9.8n [24 Mar 2010]
10673 * When rejecting SSL/TLS records due to an incorrect version number, never
10674 update s->server with a new major version number. As of
10675 - OpenSSL 0.9.8m if 'short' is a 16-bit type,
10676 - OpenSSL 0.9.8f if 'short' is longer than 16 bits,
10677 the previous behavior could result in a read attempt at NULL when
10678 receiving specific incorrect SSL/TLS records once record payload
10679 protection is active. ([CVE-2010-0740])
10681 *Bodo Moeller, Adam Langley <agl@chromium.org>*
10683 * Fix for CVE-2010-0433 where some kerberos enabled versions of OpenSSL
10684 could be crashed if the relevant tables were not present (e.g. chrooted).
10686 *Tomas Hoger <thoger@redhat.com>*
10688 ### Changes between 0.9.8l and 0.9.8m [25 Feb 2010]
10690 * Always check bn_wexpand() return values for failure. ([CVE-2009-3245])
10692 *Martin Olsson, Neel Mehta*
10694 * Fix X509_STORE locking: Every 'objs' access requires a lock (to
10695 accommodate for stack sorting, always a write lock!).
10699 * On some versions of WIN32 Heap32Next is very slow. This can cause
10700 excessive delays in the RAND_poll(): over a minute. As a workaround
10701 include a time check in the inner Heap32Next loop too.
10705 * The code that handled flushing of data in SSL/TLS originally used the
10706 BIO_CTRL_INFO ctrl to see if any data was pending first. This caused
10707 the problem outlined in PR#1949. The fix suggested there however can
10708 trigger problems with buggy BIO_CTRL_WPENDING (e.g. some versions
10709 of Apache). So instead simplify the code to flush unconditionally.
10710 This should be fine since flushing with no data to flush is a no op.
10714 * Handle TLS versions 2.0 and later properly and correctly use the
10715 highest version of TLS/SSL supported. Although TLS >= 2.0 is some way
10716 off ancient servers have a habit of sticking around for a while...
10720 * Modify compression code so it frees up structures without using the
10721 ex_data callbacks. This works around a problem where some applications
10722 call CRYPTO_cleanup_all_ex_data() before application exit (e.g. when
10723 restarting) then use compression (e.g. SSL with compression) later.
10724 This results in significant per-connection memory leaks and
10725 has caused some security issues including CVE-2008-1678 and
10730 * Constify crypto/cast (i.e., <openssl/cast.h>): a CAST_KEY doesn't
10731 change when encrypting or decrypting.
10735 * Add option SSL_OP_LEGACY_SERVER_CONNECT which will allow clients to
10736 connect and renegotiate with servers which do not support RI.
10737 Until RI is more widely deployed this option is enabled by default.
10741 * Add "missing" ssl ctrls to clear options and mode.
10745 * If client attempts to renegotiate and doesn't support RI respond with
10746 a no_renegotiation alert as required by RFC5746. Some renegotiating
10747 TLS clients will continue a connection gracefully when they receive
10748 the alert. Unfortunately OpenSSL mishandled this alert and would hang
10749 waiting for a server hello which it will never receive. Now we treat a
10750 received no_renegotiation alert as a fatal error. This is because
10751 applications requesting a renegotiation might well expect it to succeed
10752 and would have no code in place to handle the server denying it so the
10753 only safe thing to do is to terminate the connection.
10757 * Add ctrl macro SSL_get_secure_renegotiation_support() which returns 1 if
10758 peer supports secure renegotiation and 0 otherwise. Print out peer
10759 renegotiation support in s_client/s_server.
10763 * Replace the highly broken and deprecated SPKAC certification method with
10764 the updated NID creation version. This should correctly handle UTF8.
10768 * Implement RFC5746. Re-enable renegotiation but require the extension
10769 as needed. Unfortunately, SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION
10770 turns out to be a bad idea. It has been replaced by
10771 SSL_OP_ALLOW_UNSAFE_LEGACY_RENEGOTIATION which can be set with
10772 SSL_CTX_set_options(). This is really not recommended unless you
10773 know what you are doing.
10775 *Eric Rescorla <ekr@networkresonance.com>, Ben Laurie, Steve Henson*
10777 * Fixes to stateless session resumption handling. Use initial_ctx when
10778 issuing and attempting to decrypt tickets in case it has changed during
10779 servername handling. Use a non-zero length session ID when attempting
10780 stateless session resumption: this makes it possible to determine if
10781 a resumption has occurred immediately after receiving server hello
10782 (several places in OpenSSL subtly assume this) instead of later in
10787 * The functions ENGINE_ctrl(), OPENSSL_isservice(),
10788 CMS_get1_RecipientRequest() and RAND_bytes() can return <=0 on error
10789 fixes for a few places where the return code is not checked
10792 *Julia Lawall <julia@diku.dk>*
10794 * Add --strict-warnings option to Configure script to include devteam
10795 warnings in other configurations.
10799 * Add support for --libdir option and LIBDIR variable in makefiles. This
10800 makes it possible to install openssl libraries in locations which
10801 have names other than "lib", for example "/usr/lib64" which some
10804 *Steve Henson, based on patch from Jeremy Utley*
10806 * Don't allow the use of leading 0x80 in OIDs. This is a violation of
10807 X690 8.9.12 and can produce some misleading textual output of OIDs.
10809 *Steve Henson, reported by Dan Kaminsky*
10811 * Delete MD2 from algorithm tables. This follows the recommendation in
10812 several standards that it is not used in new applications due to
10813 several cryptographic weaknesses. For binary compatibility reasons
10814 the MD2 API is still compiled in by default.
10818 * Add compression id to {d2i,i2d}_SSL_SESSION so it is correctly saved
10823 * Rename uni2asc and asc2uni functions to OPENSSL_uni2asc and
10824 OPENSSL_asc2uni conditionally on Netware platforms to avoid a name
10827 *Guenter <lists@gknw.net>*
10829 * Fix the server certificate chain building code to use X509_verify_cert(),
10830 it used to have an ad-hoc builder which was unable to cope with anything
10831 other than a simple chain.
10833 *David Woodhouse <dwmw2@infradead.org>, Steve Henson*
10835 * Don't check self signed certificate signatures in X509_verify_cert()
10836 by default (a flag can override this): it just wastes time without
10837 adding any security. As a useful side effect self signed root CAs
10838 with non-FIPS digests are now usable in FIPS mode.
10842 * In dtls1_process_out_of_seq_message() the check if the current message
10843 is already buffered was missing. For every new message was memory
10844 allocated, allowing an attacker to perform an denial of service attack
10845 with sending out of seq handshake messages until there is no memory
10846 left. Additionally every future message was buffered, even if the
10847 sequence number made no sense and would be part of another handshake.
10848 So only messages with sequence numbers less than 10 in advance will be
10849 buffered. ([CVE-2009-1378])
10851 *Robin Seggelmann, discovered by Daniel Mentz*
10853 * Records are buffered if they arrive with a future epoch to be
10854 processed after finishing the corresponding handshake. There is
10855 currently no limitation to this buffer allowing an attacker to perform
10856 a DOS attack with sending records with future epochs until there is no
10857 memory left. This patch adds the pqueue_size() function to determine
10858 the size of a buffer and limits the record buffer to 100 entries.
10861 *Robin Seggelmann, discovered by Daniel Mentz*
10863 * Keep a copy of frag->msg_header.frag_len so it can be used after the
10864 parent structure is freed. ([CVE-2009-1379])
10868 * Handle non-blocking I/O properly in SSL_shutdown() call.
10870 *Darryl Miles <darryl-mailinglists@netbauds.net>*
10872 * Add `2.5.4.*` OIDs
10874 *Ilya O. <vrghost@gmail.com>*
10876 ### Changes between 0.9.8k and 0.9.8l [5 Nov 2009]
10878 * Disable renegotiation completely - this fixes a severe security
10879 problem ([CVE-2009-3555]) at the cost of breaking all
10880 renegotiation. Renegotiation can be re-enabled by setting
10881 SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION in s3->flags at
10882 run-time. This is really not recommended unless you know what
10887 ### Changes between 0.9.8j and 0.9.8k [25 Mar 2009]
10889 * Don't set val to NULL when freeing up structures, it is freed up by
10890 underlying code. If `sizeof(void *) > sizeof(long)` this can result in
10891 zeroing past the valid field. ([CVE-2009-0789])
10893 *Paolo Ganci <Paolo.Ganci@AdNovum.CH>*
10895 * Fix bug where return value of CMS_SignerInfo_verify_content() was not
10896 checked correctly. This would allow some invalid signed attributes to
10897 appear to verify correctly. ([CVE-2009-0591])
10899 *Ivan Nestlerode <inestlerode@us.ibm.com>*
10901 * Reject UniversalString and BMPString types with invalid lengths. This
10902 prevents a crash in ASN1_STRING_print_ex() which assumes the strings have
10903 a legal length. ([CVE-2009-0590])
10907 * Set S/MIME signing as the default purpose rather than setting it
10908 unconditionally. This allows applications to override it at the store
10913 * Permit restricted recursion of ASN1 strings. This is needed in practice
10914 to handle some structures.
10918 * Improve efficiency of mem_gets: don't search whole buffer each time
10921 *Jeremy Shapiro <jnshapir@us.ibm.com>*
10923 * New -hex option for openssl rand.
10927 * Print out UTF8String and NumericString when parsing ASN1.
10931 * Support NumericString type for name components.
10935 * Allow CC in the environment to override the automatically chosen
10936 compiler. Note that nothing is done to ensure flags work with the
10941 ### Changes between 0.9.8i and 0.9.8j [07 Jan 2009]
10943 * Properly check EVP_VerifyFinal() and similar return values
10946 *Ben Laurie, Bodo Moeller, Google Security Team*
10948 * Enable TLS extensions by default.
10952 * Allow the CHIL engine to be loaded, whether the application is
10953 multithreaded or not. (This does not release the developer from the
10954 obligation to set up the dynamic locking callbacks.)
10956 *Sander Temme <sander@temme.net>*
10958 * Use correct exit code if there is an error in dgst command.
10960 *Steve Henson; problem pointed out by Roland Dirlewanger*
10962 * Tweak Configure so that you need to say "experimental-jpake" to enable
10963 JPAKE, and need to use -DOPENSSL_EXPERIMENTAL_JPAKE in applications.
10967 * Add experimental JPAKE support, including demo authentication in
10968 s_client and s_server.
10972 * Set the comparison function in v3_addr_canonize().
10974 *Rob Austein <sra@hactrn.net>*
10976 * Add support for XMPP STARTTLS in s_client.
10978 *Philip Paeps <philip@freebsd.org>*
10980 * Change the server-side SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG behavior
10981 to ensure that even with this option, only ciphersuites in the
10982 server's preference list will be accepted. (Note that the option
10983 applies only when resuming a session, so the earlier behavior was
10984 just about the algorithm choice for symmetric cryptography.)
10988 ### Changes between 0.9.8h and 0.9.8i [15 Sep 2008]
10990 * Fix NULL pointer dereference if a DTLS server received
10991 ChangeCipherSpec as first record ([CVE-2009-1386]).
10995 * Fix a state transition in s3_srvr.c and d1_srvr.c
10996 (was using SSL3_ST_CW_CLNT_HELLO_B, should be `..._ST_SW_SRVR_...`).
10998 *Nagendra Modadugu*
11000 * The fix in 0.9.8c that supposedly got rid of unsafe
11001 double-checked locking was incomplete for RSA blinding,
11002 addressing just one layer of what turns out to have been
11003 doubly unsafe triple-checked locking.
11005 So now fix this for real by retiring the MONT_HELPER macro
11006 in crypto/rsa/rsa_eay.c.
11008 *Bodo Moeller; problem pointed out by Marius Schilder*
11010 * Various precautionary measures:
11012 - Avoid size_t integer overflow in HASH_UPDATE (md32_common.h).
11014 - Avoid a buffer overflow in d2i_SSL_SESSION() (ssl_asn1.c).
11015 (NB: This would require knowledge of the secret session ticket key
11016 to exploit, in which case you'd be SOL either way.)
11018 - Change bn_nist.c so that it will properly handle input BIGNUMs
11019 outside the expected range.
11021 - Enforce the 'num' check in BN_div() (bn_div.c) for non-BN_DEBUG
11024 *Neel Mehta, Bodo Moeller*
11026 * Allow engines to be "soft loaded" - i.e. optionally don't die if
11027 the load fails. Useful for distros.
11029 *Ben Laurie and the FreeBSD team*
11031 * Add support for Local Machine Keyset attribute in PKCS#12 files.
11035 * Fix BN_GF2m_mod_arr() top-bit cleanup code.
11039 * Expand ENGINE to support engine supplied SSL client certificate functions.
11041 This work was sponsored by Logica.
11045 * Add CryptoAPI ENGINE to support use of RSA and DSA keys held in Windows
11046 keystores. Support for SSL/TLS client authentication too.
11047 Not compiled unless enable-capieng specified to Configure.
11049 This work was sponsored by Logica.
11053 * Fix bug in X509_ATTRIBUTE creation: don't set attribute using
11054 ASN1_TYPE_set1 if MBSTRING flag set. This bug would crash certain
11055 attribute creation routines such as certificate requests and PKCS#12
11060 ### Changes between 0.9.8g and 0.9.8h [28 May 2008]
11062 * Fix flaw if 'Server Key exchange message' is omitted from a TLS
11063 handshake which could lead to a client crash as found using the
11064 Codenomicon TLS test suite ([CVE-2008-1672])
11066 *Steve Henson, Mark Cox*
11068 * Fix double free in TLS server name extensions which could lead to
11069 a remote crash found by Codenomicon TLS test suite ([CVE-2008-0891])
11073 * Clear error queue in SSL_CTX_use_certificate_chain_file()
11075 Clear the error queue to ensure that error entries left from
11076 older function calls do not interfere with the correct operation.
11078 *Lutz Jaenicke, Erik de Castro Lopo*
11080 * Remove root CA certificates of commercial CAs:
11082 The OpenSSL project does not recommend any specific CA and does not
11083 have any policy with respect to including or excluding any CA.
11084 Therefore, it does not make any sense to ship an arbitrary selection
11085 of root CA certificates with the OpenSSL software.
11089 * RSA OAEP patches to fix two separate invalid memory reads.
11090 The first one involves inputs when 'lzero' is greater than
11091 'SHA_DIGEST_LENGTH' (it would read about SHA_DIGEST_LENGTH bytes
11092 before the beginning of from). The second one involves inputs where
11093 the 'db' section contains nothing but zeroes (there is a one-byte
11094 invalid read after the end of 'db').
11096 *Ivan Nestlerode <inestlerode@us.ibm.com>*
11098 * Partial backport from 0.9.9-dev:
11100 Introduce bn_mul_mont (dedicated Montgomery multiplication
11101 procedure) as a candidate for BIGNUM assembler implementation.
11102 While 0.9.9-dev uses assembler for various architectures, only
11103 x86_64 is available by default here in the 0.9.8 branch, and
11104 32-bit x86 is available through a compile-time setting.
11106 To try the 32-bit x86 assembler implementation, use Configure
11107 option "enable-montasm" (which exists only for this backport).
11109 As "enable-montasm" for 32-bit x86 disclaims code stability
11110 anyway, in this constellation we activate additional code
11111 backported from 0.9.9-dev for further performance improvements,
11112 namely BN_from_montgomery_word. (To enable this otherwise,
11113 e.g. x86_64, try `-DMONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD`.)
11115 *Andy Polyakov (backport partially by Bodo Moeller)*
11117 * Add TLS session ticket callback. This allows an application to set
11118 TLS ticket cipher and HMAC keys rather than relying on hardcoded fixed
11119 values. This is useful for key rollover for example where several key
11120 sets may exist with different names.
11124 * Reverse ENGINE-internal logic for caching default ENGINE handles.
11125 This was broken until now in 0.9.8 releases, such that the only way
11126 a registered ENGINE could be used (assuming it initialises
11127 successfully on the host) was to explicitly set it as the default
11128 for the relevant algorithms. This is in contradiction with 0.9.7
11129 behaviour and the documentation. With this fix, when an ENGINE is
11130 registered into a given algorithm's table of implementations, the
11131 'uptodate' flag is reset so that auto-discovery will be used next
11132 time a new context for that algorithm attempts to select an
11135 *Ian Lister (tweaked by Geoff Thorpe)*
11137 * Backport of CMS code to OpenSSL 0.9.8. This differs from the 0.9.9
11138 implementation in the following ways:
11140 Lack of EVP_PKEY_ASN1_METHOD means algorithm parameters have to be
11143 Lack of BER streaming support means one pass streaming processing is
11144 only supported if data is detached: setting the streaming flag is
11145 ignored for embedded content.
11147 CMS support is disabled by default and must be explicitly enabled
11148 with the enable-cms configuration option.
11152 * Update the GMP engine glue to do direct copies between BIGNUM and
11153 mpz_t when openssl and GMP use the same limb size. Otherwise the
11154 existing "conversion via a text string export" trick is still used.
11156 *Paul Sheer <paulsheer@gmail.com>*
11158 * Zlib compression BIO. This is a filter BIO which compressed and
11159 uncompresses any data passed through it.
11163 * Add AES_wrap_key() and AES_unwrap_key() functions to implement
11164 RFC3394 compatible AES key wrapping.
11168 * Add utility functions to handle ASN1 structures. ASN1_STRING_set0():
11169 sets string data without copying. X509_ALGOR_set0() and
11170 X509_ALGOR_get0(): set and retrieve X509_ALGOR (AlgorithmIdentifier)
11171 data. Attribute function X509at_get0_data_by_OBJ(): retrieves data
11172 from an X509_ATTRIBUTE structure optionally checking it occurs only
11173 once. ASN1_TYPE_set1(): set and ASN1_TYPE structure copying supplied
11178 * Fix BN flag handling in RSA_eay_mod_exp() and BN_MONT_CTX_set()
11179 to get the expected BN_FLG_CONSTTIME behavior.
11181 *Bodo Moeller (Google)*
11185 - fixed wrong usage of ioctlsocket() when build for LIBC BSD sockets
11186 - fixed do_tests.pl to run the test suite with CLIB builds too (CLIB_OPT)
11187 - added some more tests to do_tests.pl
11188 - fixed RunningProcess usage so that it works with newer LIBC NDKs too
11189 - removed usage of BN_LLONG for CLIB builds to avoid runtime dependency
11190 - added new Configure targets netware-clib-bsdsock, netware-clib-gcc,
11191 netware-clib-bsdsock-gcc, netware-libc-bsdsock-gcc
11192 - various changes to netware.pl to enable gcc-cross builds on Win32
11194 - changed crypto/bio/b_sock.c to work with macro functions (CLIB BSD)
11195 - various changes to fix missing prototype warnings
11196 - fixed x86nasm.pl to create correct asm files for NASM COFF output
11197 - added AES, WHIRLPOOL and CPUID assembler code to build files
11198 - added missing AES assembler make rules to mk1mf.pl
11199 - fixed order of includes in `apps/ocsp.c` so that `e_os.h` settings apply
11201 *Guenter Knauf <eflash@gmx.net>*
11203 * Implement certificate status request TLS extension defined in RFC3546.
11204 A client can set the appropriate parameters and receive the encoded
11205 OCSP response via a callback. A server can query the supplied parameters
11206 and set the encoded OCSP response in the callback. Add simplified examples
11207 to s_client and s_server.
11211 ### Changes between 0.9.8f and 0.9.8g [19 Oct 2007]
11213 * Fix various bugs:
11214 + Binary incompatibility of ssl_ctx_st structure
11215 + DTLS interoperation with non-compliant servers
11216 + Don't call get_session_cb() without proposed session
11217 + Fix ia64 assembler code
11219 *Andy Polyakov, Steve Henson*
11221 ### Changes between 0.9.8e and 0.9.8f [11 Oct 2007]
11223 * DTLS Handshake overhaul. There were longstanding issues with
11224 OpenSSL DTLS implementation, which were making it impossible for
11225 RFC 4347 compliant client to communicate with OpenSSL server.
11226 Unfortunately just fixing these incompatibilities would "cut off"
11227 pre-0.9.8f clients. To allow for hassle free upgrade post-0.9.8e
11228 server keeps tolerating non RFC compliant syntax. The opposite is
11229 not true, 0.9.8f client can not communicate with earlier server.
11230 This update even addresses CVE-2007-4995.
11234 * Changes to avoid need for function casts in OpenSSL: some compilers
11235 (gcc 4.2 and later) reject their use.
11236 *Kurt Roeckx <kurt@roeckx.be>, Peter Hartley <pdh@utter.chaos.org.uk>,
11239 * Add RFC4507 support to OpenSSL. This includes the corrections in
11240 RFC4507bis. The encrypted ticket format is an encrypted encoded
11241 SSL_SESSION structure, that way new session features are automatically
11244 If a client application caches session in an SSL_SESSION structure
11245 support is transparent because tickets are now stored in the encoded
11248 The SSL_CTX structure automatically generates keys for ticket
11249 protection in servers so again support should be possible
11250 with no application modification.
11252 If a client or server wishes to disable RFC4507 support then the option
11253 SSL_OP_NO_TICKET can be set.
11255 Add a TLS extension debugging callback to allow the contents of any client
11256 or server extensions to be examined.
11258 This work was sponsored by Google.
11262 * Add initial support for TLS extensions, specifically for the server_name
11263 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
11264 have new members for a hostname. The SSL data structure has an
11265 additional member `SSL_CTX *initial_ctx` so that new sessions can be
11266 stored in that context to allow for session resumption, even after the
11267 SSL has been switched to a new SSL_CTX in reaction to a client's
11268 server_name extension.
11270 New functions (subject to change):
11272 SSL_get_servername()
11273 SSL_get_servername_type()
11276 New CTRL codes and macros (subject to change):
11278 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
11279 - SSL_CTX_set_tlsext_servername_callback()
11280 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
11281 - SSL_CTX_set_tlsext_servername_arg()
11282 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
11284 openssl s_client has a new '-servername ...' option.
11286 openssl s_server has new options '-servername_host ...', '-cert2 ...',
11287 '-key2 ...', '-servername_fatal' (subject to change). This allows
11288 testing the HostName extension for a specific single hostname ('-cert'
11289 and '-key' remain fallbacks for handshakes without HostName
11290 negotiation). If the unrecognized_name alert has to be sent, this by
11291 default is a warning; it becomes fatal with the '-servername_fatal'
11294 *Peter Sylvester, Remy Allais, Christophe Renou, Steve Henson*
11296 * Add AES and SSE2 assembly language support to VC++ build.
11300 * Mitigate attack on final subtraction in Montgomery reduction.
11304 * Fix crypto/ec/ec_mult.c to work properly with scalars of value 0
11305 (which previously caused an internal error).
11309 * Squeeze another 10% out of IGE mode when in != out.
11313 * AES IGE mode speedup.
11315 *Dean Gaudet (Google)*
11317 * Add the Korean symmetric 128-bit cipher SEED (see
11318 <http://www.kisa.or.kr/kisa/seed/jsp/seed_eng.jsp>) and
11319 add SEED ciphersuites from RFC 4162:
11321 TLS_RSA_WITH_SEED_CBC_SHA = "SEED-SHA"
11322 TLS_DHE_DSS_WITH_SEED_CBC_SHA = "DHE-DSS-SEED-SHA"
11323 TLS_DHE_RSA_WITH_SEED_CBC_SHA = "DHE-RSA-SEED-SHA"
11324 TLS_DH_anon_WITH_SEED_CBC_SHA = "ADH-SEED-SHA"
11326 To minimize changes between patchlevels in the OpenSSL 0.9.8
11327 series, SEED remains excluded from compilation unless OpenSSL
11328 is configured with 'enable-seed'.
11330 *KISA, Bodo Moeller*
11332 * Mitigate branch prediction attacks, which can be practical if a
11333 single processor is shared, allowing a spy process to extract
11334 information. For detailed background information, see
11335 <http://eprint.iacr.org/2007/039> (O. Aciicmez, S. Gueron,
11336 J.-P. Seifert, "New Branch Prediction Vulnerabilities in OpenSSL
11337 and Necessary Software Countermeasures"). The core of the change
11338 are new versions BN_div_no_branch() and
11339 BN_mod_inverse_no_branch() of BN_div() and BN_mod_inverse(),
11340 respectively, which are slower, but avoid the security-relevant
11341 conditional branches. These are automatically called by BN_div()
11342 and BN_mod_inverse() if the flag BN_FLG_CONSTTIME is set for one
11343 of the input BIGNUMs. Also, BN_is_bit_set() has been changed to
11344 remove a conditional branch.
11346 BN_FLG_CONSTTIME is the new name for the previous
11347 BN_FLG_EXP_CONSTTIME flag, since it now affects more than just
11348 modular exponentiation. (Since OpenSSL 0.9.7h, setting this flag
11349 in the exponent causes BN_mod_exp_mont() to use the alternative
11350 implementation in BN_mod_exp_mont_consttime().) The old name
11351 remains as a deprecated alias.
11353 Similarly, RSA_FLAG_NO_EXP_CONSTTIME is replaced by a more general
11354 RSA_FLAG_NO_CONSTTIME flag since the RSA implementation now uses
11355 constant-time implementations for more than just exponentiation.
11356 Here too the old name is kept as a deprecated alias.
11358 BN_BLINDING_new() will now use BN_dup() for the modulus so that
11359 the BN_BLINDING structure gets an independent copy of the
11360 modulus. This means that the previous `BIGNUM *m` argument to
11361 BN_BLINDING_new() and to BN_BLINDING_create_param() now
11362 essentially becomes `const BIGNUM *m`, although we can't actually
11363 change this in the header file before 0.9.9. It allows
11364 RSA_setup_blinding() to use BN_with_flags() on the modulus to
11365 enable BN_FLG_CONSTTIME.
11367 *Matthew D Wood (Intel Corp)*
11369 * In the SSL/TLS server implementation, be strict about session ID
11370 context matching (which matters if an application uses a single
11371 external cache for different purposes). Previously,
11372 out-of-context reuse was forbidden only if SSL_VERIFY_PEER was
11373 set. This did ensure strict client verification, but meant that,
11374 with applications using a single external cache for quite
11375 different requirements, clients could circumvent ciphersuite
11376 restrictions for a given session ID context by starting a session
11377 in a different context.
11381 * Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
11382 a ciphersuite string such as "DEFAULT:RSA" cannot enable
11383 authentication-only ciphersuites.
11387 * Update the SSL_get_shared_ciphers() fix CVE-2006-3738 which was
11388 not complete and could lead to a possible single byte overflow
11389 ([CVE-2007-5135]) [Ben Laurie]
11391 ### Changes between 0.9.8d and 0.9.8e [23 Feb 2007]
11393 * Since AES128 and AES256 (and similarly Camellia128 and
11394 Camellia256) share a single mask bit in the logic of
11395 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
11396 kludge to work properly if AES128 is available and AES256 isn't
11397 (or if Camellia128 is available and Camellia256 isn't).
11401 * Fix the BIT STRING encoding generated by crypto/ec/ec_asn1.c
11402 (within i2d_ECPrivateKey, i2d_ECPKParameters, i2d_ECParameters):
11403 When a point or a seed is encoded in a BIT STRING, we need to
11404 prevent the removal of trailing zero bits to get the proper DER
11405 encoding. (By default, crypto/asn1/a_bitstr.c assumes the case
11406 of a NamedBitList, for which trailing 0 bits need to be removed.)
11410 * Have SSL/TLS server implementation tolerate "mismatched" record
11411 protocol version while receiving ClientHello even if the
11412 ClientHello is fragmented. (The server can't insist on the
11413 particular protocol version it has chosen before the ServerHello
11414 message has informed the client about his choice.)
11418 * Add RFC 3779 support.
11420 *Rob Austein for ARIN, Ben Laurie*
11422 * Load error codes if they are not already present instead of using a
11423 static variable. This allows them to be cleanly unloaded and reloaded.
11424 Improve header file function name parsing.
11428 * extend SMTP and IMAP protocol emulation in s_client to use EHLO
11429 or CAPABILITY handshake as required by RFCs.
11431 *Goetz Babin-Ebell*
11433 ### Changes between 0.9.8c and 0.9.8d [28 Sep 2006]
11435 * Introduce limits to prevent malicious keys being able to
11436 cause a denial of service. ([CVE-2006-2940])
11438 *Steve Henson, Bodo Moeller*
11440 * Fix ASN.1 parsing of certain invalid structures that can result
11441 in a denial of service. ([CVE-2006-2937]) [Steve Henson]
11443 * Fix buffer overflow in SSL_get_shared_ciphers() function.
11444 ([CVE-2006-3738]) [Tavis Ormandy and Will Drewry, Google Security Team]
11446 * Fix SSL client code which could crash if connecting to a
11447 malicious SSLv2 server. ([CVE-2006-4343])
11449 *Tavis Ormandy and Will Drewry, Google Security Team*
11451 * Since 0.9.8b, ciphersuite strings naming explicit ciphersuites
11452 match only those. Before that, "AES256-SHA" would be interpreted
11453 as a pattern and match "AES128-SHA" too (since AES128-SHA got
11454 the same strength classification in 0.9.7h) as we currently only
11455 have a single AES bit in the ciphersuite description bitmap.
11456 That change, however, also applied to ciphersuite strings such as
11457 "RC4-MD5" that intentionally matched multiple ciphersuites --
11458 namely, SSL 2.0 ciphersuites in addition to the more common ones
11459 from SSL 3.0/TLS 1.0.
11461 So we change the selection algorithm again: Naming an explicit
11462 ciphersuite selects this one ciphersuite, and any other similar
11463 ciphersuite (same bitmap) from *other* protocol versions.
11464 Thus, "RC4-MD5" again will properly select both the SSL 2.0
11465 ciphersuite and the SSL 3.0/TLS 1.0 ciphersuite.
11467 Since SSL 2.0 does not have any ciphersuites for which the
11468 128/256 bit distinction would be relevant, this works for now.
11469 The proper fix will be to use different bits for AES128 and
11470 AES256, which would have avoided the problems from the beginning;
11471 however, bits are scarce, so we can only do this in a new release
11472 (not just a patchlevel) when we can change the SSL_CIPHER
11473 definition to split the single 'unsigned long mask' bitmap into
11474 multiple values to extend the available space.
11478 ### Changes between 0.9.8b and 0.9.8c [05 Sep 2006]
11480 * Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
11481 ([CVE-2006-4339]) [Ben Laurie and Google Security Team]
11483 * Add AES IGE and biIGE modes.
11487 * Change the Unix randomness entropy gathering to use poll() when
11488 possible instead of select(), since the latter has some
11489 undesirable limitations.
11491 *Darryl Miles via Richard Levitte and Bodo Moeller*
11493 * Disable "ECCdraft" ciphersuites more thoroughly. Now special
11494 treatment in ssl/ssl_ciph.s makes sure that these ciphersuites
11495 cannot be implicitly activated as part of, e.g., the "AES" alias.
11496 However, please upgrade to OpenSSL 0.9.9[-dev] for
11497 non-experimental use of the ECC ciphersuites to get TLS extension
11498 support, which is required for curve and point format negotiation
11499 to avoid potential handshake problems.
11503 * Disable rogue ciphersuites:
11505 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
11506 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
11507 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
11509 The latter two were purportedly from
11510 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
11513 Also deactivate the remaining ciphersuites from
11514 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
11515 unofficial, and the ID has long expired.
11519 * Fix RSA blinding Heisenbug (problems sometimes occurred on
11520 dual-core machines) and other potential thread-safety issues.
11524 * Add the symmetric cipher Camellia (128-bit, 192-bit, 256-bit key
11525 versions), which is now available for royalty-free use
11526 (see <http://info.isl.ntt.co.jp/crypt/eng/info/chiteki.html>).
11527 Also, add Camellia TLS ciphersuites from RFC 4132.
11529 To minimize changes between patchlevels in the OpenSSL 0.9.8
11530 series, Camellia remains excluded from compilation unless OpenSSL
11531 is configured with 'enable-camellia'.
11535 * Disable the padding bug check when compression is in use. The padding
11536 bug check assumes the first packet is of even length, this is not
11537 necessarily true if compression is enabled and can result in false
11538 positives causing handshake failure. The actual bug test is ancient
11539 code so it is hoped that implementations will either have fixed it by
11540 now or any which still have the bug do not support compression.
11544 ### Changes between 0.9.8a and 0.9.8b [04 May 2006]
11546 * When applying a cipher rule check to see if string match is an explicit
11547 cipher suite and only match that one cipher suite if it is.
11551 * Link in manifests for VC++ if needed.
11553 *Austin Ziegler <halostatue@gmail.com>*
11555 * Update support for ECC-based TLS ciphersuites according to
11556 draft-ietf-tls-ecc-12.txt with proposed changes (but without
11557 TLS extensions, which are supported starting with the 0.9.9
11558 branch, not in the OpenSSL 0.9.8 branch).
11562 * New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support
11563 opaque EVP_CIPHER_CTX handling.
11567 * Fixes and enhancements to zlib compression code. We now only use
11568 "zlib1.dll" and use the default `__cdecl` calling convention on Win32
11569 to conform with the standards mentioned here:
11570 <http://www.zlib.net/DLL_FAQ.txt>
11571 Static zlib linking now works on Windows and the new --with-zlib-include
11572 --with-zlib-lib options to Configure can be used to supply the location
11573 of the headers and library. Gracefully handle case where zlib library
11578 * Several fixes and enhancements to the OID generation code. The old code
11579 sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't
11580 handle numbers larger than ULONG_MAX, truncated printing and had a
11581 non standard OBJ_obj2txt() behaviour.
11585 * Add support for building of engines under engine/ as shared libraries
11586 under VC++ build system.
11590 * Corrected the numerous bugs in the Win32 path splitter in DSO.
11591 Hopefully, we will not see any false combination of paths any more.
11595 ### Changes between 0.9.8 and 0.9.8a [11 Oct 2005]
11597 * Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
11598 (part of SSL_OP_ALL). This option used to disable the
11599 countermeasure against man-in-the-middle protocol-version
11600 rollback in the SSL 2.0 server implementation, which is a bad
11601 idea. ([CVE-2005-2969])
11603 *Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
11604 for Information Security, National Institute of Advanced Industrial
11605 Science and Technology [AIST], Japan)*
11607 * Add two function to clear and return the verify parameter flags.
11611 * Keep cipherlists sorted in the source instead of sorting them at
11612 runtime, thus removing the need for a lock.
11616 * Avoid some small subgroup attacks in Diffie-Hellman.
11618 *Nick Mathewson and Ben Laurie*
11620 * Add functions for well-known primes.
11624 * Extended Windows CE support.
11626 *Satoshi Nakamura and Andy Polyakov*
11628 * Initialize SSL_METHOD structures at compile time instead of during
11629 runtime, thus removing the need for a lock.
11633 * Make PKCS7_decrypt() work even if no certificate is supplied by
11634 attempting to decrypt each encrypted key in turn. Add support to
11639 ### Changes between 0.9.7h and 0.9.8 [05 Jul 2005]
11641 [NB: OpenSSL 0.9.7i and later 0.9.7 patch levels were released after
11644 * Add libcrypto.pc and libssl.pc for those who feel they need them.
11648 * Change CA.sh and CA.pl so they don't bundle the CSR and the private
11649 key into the same file any more.
11653 * Add initial support for Win64, both IA64 and AMD64/x64 flavors.
11657 * Add -utf8 command line and config file option to 'ca'.
11659 *Stefan <stf@udoma.org*
11661 * Removed the macro des_crypt(), as it seems to conflict with some
11662 libraries. Use DES_crypt().
11666 * Correct naming of the 'chil' and '4758cca' ENGINEs. This
11667 involves renaming the source and generated shared-libs for
11668 both. The engines will accept the corrected or legacy ids
11669 ('ncipher' and '4758_cca' respectively) when binding. NB,
11670 this only applies when building 'shared'.
11672 *Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe*
11674 * Add attribute functions to EVP_PKEY structure. Modify
11675 PKCS12_create() to recognize a CSP name attribute and
11676 use it. Make -CSP option work again in pkcs12 utility.
11680 * Add new functionality to the bn blinding code:
11681 - automatic re-creation of the BN_BLINDING parameters after
11682 a fixed number of uses (currently 32)
11683 - add new function for parameter creation
11684 - introduce flags to control the update behaviour of the
11685 BN_BLINDING parameters
11686 - hide BN_BLINDING structure
11687 Add a second BN_BLINDING slot to the RSA structure to improve
11688 performance when a single RSA object is shared among several
11693 * Add support for DTLS.
11695 *Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie*
11697 * Add support for DER encoded private keys (SSL_FILETYPE_ASN1)
11698 to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file()
11702 * Remove buggy and incomplete DH cert support from
11703 ssl/ssl_rsa.c and ssl/s3_both.c
11707 * Use SHA-1 instead of MD5 as the default digest algorithm for
11708 the `apps/openssl` commands.
11712 * Compile clean with "-Wall -Wmissing-prototypes
11713 -Wstrict-prototypes -Wmissing-declarations -Werror". Currently
11714 DEBUG_SAFESTACK must also be set.
11718 * Change ./Configure so that certain algorithms can be disabled by default.
11719 The new counterpiece to "no-xxx" is "enable-xxx".
11721 The patented RC5 and MDC2 algorithms will now be disabled unless
11722 "enable-rc5" and "enable-mdc2", respectively, are specified.
11724 (IDEA remains enabled despite being patented. This is because IDEA
11725 is frequently required for interoperability, and there is no license
11726 fee for non-commercial use. As before, "no-idea" can be used to
11727 avoid this algorithm.)
11731 * Add processing of proxy certificates (see RFC 3820). This work was
11732 sponsored by KTH (The Royal Institute of Technology in Stockholm) and
11733 EGEE (Enabling Grids for E-science in Europe).
11737 * RC4 performance overhaul on modern architectures/implementations, such
11738 as Intel P4, IA-64 and AMD64.
11742 * New utility extract-section.pl. This can be used specify an alternative
11743 section number in a pod file instead of having to treat each file as
11744 a separate case in Makefile. This can be done by adding two lines to the
11747 =for comment openssl_section:XXX
11749 The blank line is mandatory.
11753 * New arguments -certform, -keyform and -pass for s_client and s_server
11754 to allow alternative format key and certificate files and passphrase
11759 * New structure X509_VERIFY_PARAM which combines current verify parameters,
11760 update associated structures and add various utility functions.
11762 Add new policy related verify parameters, include policy checking in
11763 standard verify code. Enhance 'smime' application with extra parameters
11764 to support policy checking and print out.
11768 * Add a new engine to support VIA PadLock ACE extensions in the VIA C3
11769 Nehemiah processors. These extensions support AES encryption in hardware
11770 as well as RNG (though RNG support is currently disabled).
11772 *Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov*
11774 * Deprecate `BN_[get|set]_params()` functions (they were ignored internally).
11778 * New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented.
11780 *Andy Polyakov and a number of other people*
11782 * Improved PowerPC platform support. Most notably BIGNUM assembler
11783 implementation contributed by IBM.
11785 *Suresh Chari, Peter Waltenberg, Andy Polyakov*
11787 * The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public
11788 exponent rather than 'unsigned long'. There is a corresponding change to
11789 the new 'rsa_keygen' element of the RSA_METHOD structure.
11791 *Jelte Jansen, Geoff Thorpe*
11793 * Functionality for creating the initial serial number file is now
11794 moved from CA.pl to the 'ca' utility with a new option -create_serial.
11796 (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial
11797 number file to 1, which is bound to cause problems. To avoid
11798 the problems while respecting compatibility between different 0.9.7
11799 patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in
11800 CA.pl for serial number initialization. With the new release 0.9.8,
11801 we can fix the problem directly in the 'ca' utility.)
11805 * Reduced header interdependencies by declaring more opaque objects in
11806 ossl_typ.h. As a consequence, including some headers (eg. engine.h) will
11807 give fewer recursive includes, which could break lazy source code - so
11808 this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always,
11809 developers should define this symbol when building and using openssl to
11810 ensure they track the recommended behaviour, interfaces, [etc], but
11811 backwards-compatible behaviour prevails when this isn't defined.
11815 * New function X509_POLICY_NODE_print() which prints out policy nodes.
11819 * Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality.
11820 This will generate a random key of the appropriate length based on the
11821 cipher context. The EVP_CIPHER can provide its own random key generation
11822 routine to support keys of a specific form. This is used in the des and
11823 3des routines to generate a key of the correct parity. Update S/MIME
11824 code to use new functions and hence generate correct parity DES keys.
11825 Add EVP_CHECK_DES_KEY #define to return an error if the key is not
11826 valid (weak or incorrect parity).
11830 * Add a local set of CRLs that can be used by X509_verify_cert() as well
11831 as looking them up. This is useful when the verified structure may contain
11832 CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs
11833 present unless the new PKCS7_NO_CRL flag is asserted.
11837 * Extend ASN1 oid configuration module. It now additionally accepts the
11840 shortName = some long name, 1.2.3.4
11844 * Reimplemented the BN_CTX implementation. There is now no more static
11845 limitation on the number of variables it can handle nor the depth of the
11846 "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack
11847 information can now expand as required, and rather than having a single
11848 static array of bignums, BN_CTX now uses a linked-list of such arrays
11849 allowing it to expand on demand whilst maintaining the usefulness of
11850 BN_CTX's "bundling".
11854 * Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD
11855 to allow all RSA operations to function using a single BN_CTX.
11859 * Preliminary support for certificate policy evaluation and checking. This
11860 is initially intended to pass the tests outlined in "Conformance Testing
11861 of Relying Party Client Certificate Path Processing Logic" v1.07.
11865 * bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and
11866 remained unused and not that useful. A variety of other little bignum
11867 tweaks and fixes have also been made continuing on from the audit (see
11872 * Constify all or almost all d2i, c2i, s2i and r2i functions, along with
11873 associated ASN1, EVP and SSL functions and old ASN1 macros.
11877 * BN_zero() only needs to set 'top' and 'neg' to zero for correct results,
11878 and this should never fail. So the return value from the use of
11879 BN_set_word() (which can fail due to needless expansion) is now deprecated;
11880 if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro.
11884 * BN_CTX_get() should return zero-valued bignums, providing the same
11885 initialised value as BN_new().
11887 *Geoff Thorpe, suggested by Ulf Möller*
11889 * Support for inhibitAnyPolicy certificate extension.
11893 * An audit of the BIGNUM code is underway, for which debugging code is
11894 enabled when BN_DEBUG is defined. This makes stricter enforcements on what
11895 is considered valid when processing BIGNUMs, and causes execution to
11896 assert() when a problem is discovered. If BN_DEBUG_RAND is defined,
11897 further steps are taken to deliberately pollute unused data in BIGNUM
11898 structures to try and expose faulty code further on. For now, openssl will
11899 (in its default mode of operation) continue to tolerate the inconsistent
11900 forms that it has tolerated in the past, but authors and packagers should
11901 consider trying openssl and their own applications when compiled with
11902 these debugging symbols defined. It will help highlight potential bugs in
11903 their own code, and will improve the test coverage for OpenSSL itself. At
11904 some point, these tighter rules will become openssl's default to improve
11905 maintainability, though the assert()s and other overheads will remain only
11906 in debugging configurations. See bn.h for more details.
11908 *Geoff Thorpe, Nils Larsch, Ulf Möller*
11910 * BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure
11911 that can only be obtained through BN_CTX_new() (which implicitly
11912 initialises it). The presence of this function only made it possible
11913 to overwrite an existing structure (and cause memory leaks).
11917 * Because of the callback-based approach for implementing LHASH as a
11918 template type, lh_insert() adds opaque objects to hash-tables and
11919 lh_doall() or lh_doall_arg() are typically used with a destructor callback
11920 to clean up those corresponding objects before destroying the hash table
11921 (and losing the object pointers). So some over-zealous constifications in
11922 LHASH have been relaxed so that lh_insert() does not take (nor store) the
11923 objects as "const" and the `lh_doall[_arg]` callback wrappers are not
11924 prototyped to have "const" restrictions on the object pointers they are
11925 given (and so aren't required to cast them away any more).
11929 * The tmdiff.h API was so ugly and minimal that our own timing utility
11930 (speed) prefers to use its own implementation. The two implementations
11931 haven't been consolidated as yet (volunteers?) but the tmdiff API has had
11932 its object type properly exposed (MS_TM) instead of casting to/from
11933 `char *`. This may still change yet if someone realises MS_TM and
11935 aren't necessarily the greatest nomenclatures - but this is what was used
11936 internally to the implementation so I've used that for now.
11940 * Ensure that deprecated functions do not get compiled when
11941 OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of
11942 the self-tests were still using deprecated key-generation functions so
11943 these have been updated also.
11947 * Reorganise PKCS#7 code to separate the digest location functionality
11948 into PKCS7_find_digest(), digest addition into PKCS7_bio_add_digest().
11949 New function PKCS7_set_digest() to set the digest type for PKCS#7
11950 digestedData type. Add additional code to correctly generate the
11951 digestedData type and add support for this type in PKCS7 initialization
11956 * New function PKCS7_set0_type_other() this initializes a PKCS7
11957 structure of type "other".
11961 * Fix prime generation loop in crypto/bn/bn_prime.pl by making
11962 sure the loop does correctly stop and breaking ("division by zero")
11963 modulus operations are not performed. The (pre-generated) prime
11964 table crypto/bn/bn_prime.h was already correct, but it could not be
11965 re-generated on some platforms because of the "division by zero"
11966 situation in the script.
11968 *Ralf S. Engelschall*
11970 * Update support for ECC-based TLS ciphersuites according to
11971 draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with
11972 SHA-1 now is only used for "small" curves (where the
11973 representation of a field element takes up to 24 bytes); for
11974 larger curves, the field element resulting from ECDH is directly
11975 used as premaster secret.
11977 *Douglas Stebila (Sun Microsystems Laboratories)*
11979 * Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2
11980 curve secp160r1 to the tests.
11982 *Douglas Stebila (Sun Microsystems Laboratories)*
11984 * Add the possibility to load symbols globally with DSO.
11986 *Götz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte*
11988 * Add the functions ERR_set_mark() and ERR_pop_to_mark() for better
11989 control of the error stack.
11993 * Add support for STORE in ENGINE.
11997 * Add the STORE type. The intention is to provide a common interface
11998 to certificate and key stores, be they simple file-based stores, or
11999 HSM-type store, or LDAP stores, or...
12000 NOTE: The code is currently UNTESTED and isn't really used anywhere.
12004 * Add a generic structure called OPENSSL_ITEM. This can be used to
12005 pass a list of arguments to any function as well as provide a way
12006 for a function to pass data back to the caller.
12010 * Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup()
12011 works like BUF_strdup() but can be used to duplicate a portion of
12012 a string. The copy gets NUL-terminated. BUF_memdup() duplicates
12017 * Add the function sk_find_ex() which works like sk_find(), but will
12018 return an index to an element even if an exact match couldn't be
12019 found. The index is guaranteed to point at the element where the
12020 searched-for key would be inserted to preserve sorting order.
12024 * Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but
12025 takes an extra flags argument for optional functionality. Currently,
12026 the following flags are defined:
12028 OBJ_BSEARCH_VALUE_ON_NOMATCH
12029 This one gets OBJ_bsearch_ex() to return a pointer to the first
12030 element where the comparing function returns a negative or zero
12033 OBJ_BSEARCH_FIRST_VALUE_ON_MATCH
12034 This one gets OBJ_bsearch_ex() to return a pointer to the first
12035 element where the comparing function returns zero. This is useful
12036 if there are more than one element where the comparing function
12041 * Make it possible to create self-signed certificates with 'openssl ca'
12042 in such a way that the self-signed certificate becomes part of the
12043 CA database and uses the same mechanisms for serial number generation
12044 as all other certificate signing. The new flag '-selfsign' enables
12045 this functionality. Adapt CA.sh and CA.pl.in.
12049 * Add functionality to check the public key of a certificate request
12050 against a given private. This is useful to check that a certificate
12051 request can be signed by that key (self-signing).
12055 * Make it possible to have multiple active certificates with the same
12056 subject in the CA index file. This is done only if the keyword
12057 'unique_subject' is set to 'no' in the main CA section (default
12058 if 'CA_default') of the configuration file. The value is saved
12059 with the database itself in a separate index attribute file,
12060 named like the index file with '.attr' appended to the name.
12064 * Generate multi-valued AVAs using '+' notation in config files for
12069 * Support for nameConstraints certificate extension.
12073 * Support for policyConstraints certificate extension.
12077 * Support for policyMappings certificate extension.
12081 * Make sure the default DSA_METHOD implementation only uses its
12082 dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL,
12083 and change its own handlers to be NULL so as to remove unnecessary
12084 indirection. This lets alternative implementations fallback to the
12085 default implementation more easily.
12089 * Support for directoryName in GeneralName related extensions
12094 * Make it possible to link applications using Makefile.shared.
12095 Make that possible even when linking against static libraries!
12099 * Support for single pass processing for S/MIME signing. This now
12100 means that S/MIME signing can be done from a pipe, in addition
12101 cleartext signing (multipart/signed type) is effectively streaming
12102 and the signed data does not need to be all held in memory.
12104 This is done with a new flag PKCS7_STREAM. When this flag is set
12105 PKCS7_sign() only initializes the PKCS7 structure and the actual signing
12106 is done after the data is output (and digests calculated) in
12107 SMIME_write_PKCS7().
12111 * Add full support for -rpath/-R, both in shared libraries and
12112 applications, at least on the platforms where it's known how
12117 * In crypto/ec/ec_mult.c, implement fast point multiplication with
12118 precomputation, based on wNAF splitting: EC_GROUP_precompute_mult()
12119 will now compute a table of multiples of the generator that
12120 makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul()
12121 faster (notably in the case of a single point multiplication,
12122 scalar * generator).
12124 *Nils Larsch, Bodo Moeller*
12126 * IPv6 support for certificate extensions. The various extensions
12127 which use the IP:a.b.c.d can now take IPv6 addresses using the
12128 formats of RFC1884 2.2 . IPv6 addresses are now also displayed
12133 * Added an ENGINE that implements RSA by performing private key
12134 exponentiations with the GMP library. The conversions to and from
12135 GMP's mpz_t format aren't optimised nor are any montgomery forms
12136 cached, and on x86 it appears OpenSSL's own performance has caught up.
12137 However there are likely to be other architectures where GMP could
12138 provide a boost. This ENGINE is not built in by default, but it can be
12139 specified at Configure time and should be accompanied by the necessary
12140 linker additions, eg;
12141 ./config -DOPENSSL_USE_GMP -lgmp
12145 * "openssl engine" will not display ENGINE/DSO load failure errors when
12146 testing availability of engines with "-t" - the old behaviour is
12147 produced by increasing the feature's verbosity with "-tt".
12151 * ECDSA routines: under certain error conditions uninitialized BN objects
12152 could be freed. Solution: make sure initialization is performed early
12153 enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de>
12158 * Key-generation can now be implemented in RSA_METHOD, DSA_METHOD
12159 and DH_METHOD (eg. by ENGINE implementations) to override the normal
12160 software implementations. For DSA and DH, parameter generation can
12161 also be overridden by providing the appropriate method callbacks.
12165 * Change the "progress" mechanism used in key-generation and
12166 primality testing to functions that take a new BN_GENCB pointer in
12167 place of callback/argument pairs. The new API functions have `_ex`
12168 postfixes and the older functions are reimplemented as wrappers for
12169 the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide
12170 declarations of the old functions to help (graceful) attempts to
12171 migrate to the new functions. Also, the new key-generation API
12172 functions operate on a caller-supplied key-structure and return
12173 success/failure rather than returning a key or NULL - this is to
12174 help make "keygen" another member function of RSA_METHOD etc.
12176 Example for using the new callback interface:
12178 int (*my_callback)(int a, int b, BN_GENCB *cb) = ...;
12179 void *my_arg = ...;
12182 BN_GENCB_set(&my_cb, my_callback, my_arg);
12184 return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb);
12185 /* For the meaning of a, b in calls to my_callback(), see the
12186 * documentation of the function that calls the callback.
12187 * cb will point to my_cb; my_arg can be retrieved as cb->arg.
12188 * my_callback should return 1 if it wants BN_is_prime_ex()
12189 * to continue, or 0 to stop.
12194 * Change the ZLIB compression method to be stateful, and make it
12195 available to TLS with the number defined in
12196 draft-ietf-tls-compression-04.txt.
12200 * Add the ASN.1 structures and functions for CertificatePair, which
12201 is defined as follows (according to X.509_4thEditionDraftV6.pdf):
12203 CertificatePair ::= SEQUENCE {
12204 forward [0] Certificate OPTIONAL,
12205 reverse [1] Certificate OPTIONAL,
12206 -- at least one of the pair shall be present -- }
12208 Also implement the PEM functions to read and write certificate
12209 pairs, and defined the PEM tag as "CERTIFICATE PAIR".
12211 This needed to be defined, mostly for the sake of the LDAP
12212 attribute crossCertificatePair, but may prove useful elsewhere as
12217 * Make it possible to inhibit symlinking of shared libraries in
12218 Makefile.shared, for Cygwin's sake.
12222 * Extend the BIGNUM API by creating a function
12223 void BN_set_negative(BIGNUM *a, int neg);
12224 and a macro that behave like
12225 int BN_is_negative(const BIGNUM *a);
12227 to avoid the need to access 'a->neg' directly in applications.
12231 * Implement fast modular reduction for pseudo-Mersenne primes
12232 used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c).
12233 EC_GROUP_new_curve_GFp() will now automatically use this
12236 *Nils Larsch <nla@trustcenter.de>*
12238 * Add new lock type (CRYPTO_LOCK_BN).
12242 * Change the ENGINE framework to automatically load engines
12243 dynamically from specific directories unless they could be
12244 found to already be built in or loaded. Move all the
12245 current engines except for the cryptodev one to a new
12246 directory engines/.
12247 The engines in engines/ are built as shared libraries if
12248 the "shared" options was given to ./Configure or ./config.
12249 Otherwise, they are inserted in libcrypto.a.
12250 /usr/local/ssl/engines is the default directory for dynamic
12251 engines, but that can be overridden at configure time through
12252 the usual use of --prefix and/or --openssldir, and at run
12253 time with the environment variable OPENSSL_ENGINES.
12255 *Geoff Thorpe and Richard Levitte*
12257 * Add Makefile.shared, a helper makefile to build shared
12258 libraries. Adapt Makefile.org.
12262 * Add version info to Win32 DLLs.
12264 *Peter 'Luna' Runestig" <peter@runestig.com>*
12266 * Add new 'medium level' PKCS#12 API. Certificates and keys
12267 can be added using this API to created arbitrary PKCS#12
12268 files while avoiding the low-level API.
12270 New options to PKCS12_create(), key or cert can be NULL and
12271 will then be omitted from the output file. The encryption
12272 algorithm NIDs can be set to -1 for no encryption, the mac
12273 iteration count can be set to 0 to omit the mac.
12275 Enhance pkcs12 utility by making the -nokeys and -nocerts
12276 options work when creating a PKCS#12 file. New option -nomac
12277 to omit the mac, NONE can be set for an encryption algorithm.
12278 New code is modified to use the enhanced PKCS12_create()
12279 instead of the low-level API.
12283 * Extend ASN1 encoder to support indefinite length constructed
12284 encoding. This can output sequences tags and octet strings in
12285 this form. Modify pk7_asn1.c to support indefinite length
12286 encoding. This is experimental and needs additional code to
12287 be useful, such as an ASN1 bio and some enhanced streaming
12290 Extend template encode functionality so that tagging is passed
12291 down to the template encoder.
12295 * Let 'openssl req' fail if an argument to '-newkey' is not
12296 recognized instead of using RSA as a default.
12300 * Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt.
12301 As these are not official, they are not included in "ALL";
12302 the "ECCdraft" ciphersuite group alias can be used to select them.
12304 *Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)*
12306 * Add ECDH engine support.
12308 *Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)*
12310 * Add ECDH in new directory crypto/ecdh/.
12312 *Douglas Stebila (Sun Microsystems Laboratories)*
12314 * Let BN_rand_range() abort with an error after 100 iterations
12315 without success (which indicates a broken PRNG).
12319 * Change BN_mod_sqrt() so that it verifies that the input value
12320 is really the square of the return value. (Previously,
12321 BN_mod_sqrt would show GIGO behaviour.)
12325 * Add named elliptic curves over binary fields from X9.62, SECG,
12326 and WAP/WTLS; add OIDs that were still missing.
12328 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
12330 * Extend the EC library for elliptic curves over binary fields
12331 (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/).
12334 EC_GF2m_simple_method
12338 EC_GROUP_new_curve_GF2m
12339 EC_GROUP_set_curve_GF2m
12340 EC_GROUP_get_curve_GF2m
12341 EC_POINT_set_affine_coordinates_GF2m
12342 EC_POINT_get_affine_coordinates_GF2m
12343 EC_POINT_set_compressed_coordinates_GF2m
12345 Point compression for binary fields is disabled by default for
12346 patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to
12349 As binary polynomials are represented as BIGNUMs, various members
12350 of the EC_GROUP and EC_POINT data structures can be shared
12351 between the implementations for prime fields and binary fields;
12352 the above `..._GF2m functions` (except for EX_GROUP_new_curve_GF2m)
12353 are essentially identical to their `..._GFp` counterparts.
12354 (For simplicity, the `..._GFp` prefix has been dropped from
12355 various internal method names.)
12357 An internal 'field_div' method (similar to 'field_mul' and
12358 'field_sqr') has been added; this is used only for binary fields.
12360 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
12362 * Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult()
12363 through methods ('mul', 'precompute_mult').
12365 The generic implementations (now internally called 'ec_wNAF_mul'
12366 and 'ec_wNAF_precomputed_mult') remain the default if these
12367 methods are undefined.
12369 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
12371 * New function EC_GROUP_get_degree, which is defined through
12372 EC_METHOD. For curves over prime fields, this returns the bit
12373 length of the modulus.
12375 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
12377 * New functions EC_GROUP_dup, EC_POINT_dup.
12378 (These simply call ..._new and ..._copy).
12380 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
12382 * Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c.
12383 Polynomials are represented as BIGNUMs (where the sign bit is not
12384 used) in the following functions [macros]:
12387 BN_GF2m_sub [= BN_GF2m_add]
12388 BN_GF2m_mod [wrapper for BN_GF2m_mod_arr]
12389 BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr]
12390 BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr]
12392 BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr]
12393 BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr]
12394 BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr]
12395 BN_GF2m_cmp [= BN_ucmp]
12397 (Note that only the 'mod' functions are actually for fields GF(2^m).
12398 BN_GF2m_add() is misnomer, but this is for the sake of consistency.)
12400 For some functions, an the irreducible polynomial defining a
12401 field can be given as an 'unsigned int[]' with strictly
12402 decreasing elements giving the indices of those bits that are set;
12403 i.e., p[] represents the polynomial
12404 f(t) = t^p[0] + t^p[1] + ... + t^p[k]
12406 p[0] > p[1] > ... > p[k] = 0.
12407 This applies to the following functions:
12410 BN_GF2m_mod_mul_arr
12411 BN_GF2m_mod_sqr_arr
12412 BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv]
12413 BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div]
12414 BN_GF2m_mod_exp_arr
12415 BN_GF2m_mod_sqrt_arr
12416 BN_GF2m_mod_solve_quad_arr
12420 Conversion can be performed by the following functions:
12425 bntest.c has additional tests for binary polynomial arithmetic.
12427 Two implementations for BN_GF2m_mod_div() are available.
12428 The default algorithm simply uses BN_GF2m_mod_inv() and
12429 BN_GF2m_mod_mul(). The alternative algorithm is compiled in only
12430 if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the
12431 copyright notice in crypto/bn/bn_gf2m.c before enabling it).
12433 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
12435 * Add new error code 'ERR_R_DISABLED' that can be used when some
12436 functionality is disabled at compile-time.
12438 *Douglas Stebila <douglas.stebila@sun.com>*
12440 * Change default behaviour of 'openssl asn1parse' so that more
12441 information is visible when viewing, e.g., a certificate:
12443 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump'
12444 mode the content of non-printable OCTET STRINGs is output in a
12445 style similar to INTEGERs, but with '[HEX DUMP]' prepended to
12446 avoid the appearance of a printable string.
12448 *Nils Larsch <nla@trustcenter.de>*
12450 * Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access
12452 EC_GROUP_set_asn1_flag()
12453 EC_GROUP_get_asn1_flag()
12454 EC_GROUP_set_point_conversion_form()
12455 EC_GROUP_get_point_conversion_form()
12456 These control ASN1 encoding details:
12457 - Curves (i.e., groups) are encoded explicitly unless asn1_flag
12458 has been set to OPENSSL_EC_NAMED_CURVE.
12459 - Points are encoded in uncompressed form by default; options for
12460 asn1_for are as for point2oct, namely
12461 POINT_CONVERSION_COMPRESSED
12462 POINT_CONVERSION_UNCOMPRESSED
12463 POINT_CONVERSION_HYBRID
12465 Also add 'seed' and 'seed_len' members to EC_GROUP with access
12467 EC_GROUP_set_seed()
12468 EC_GROUP_get0_seed()
12469 EC_GROUP_get_seed_len()
12470 This is used only for ASN1 purposes (so far).
12472 *Nils Larsch <nla@trustcenter.de>*
12474 * Add 'field_type' member to EC_METHOD, which holds the NID
12475 of the appropriate field type OID. The new function
12476 EC_METHOD_get_field_type() returns this value.
12478 *Nils Larsch <nla@trustcenter.de>*
12481 EC_POINT_point2bn()
12482 EC_POINT_bn2point()
12483 EC_POINT_point2hex()
12484 EC_POINT_hex2point()
12485 providing useful interfaces to EC_POINT_point2oct() and
12486 EC_POINT_oct2point().
12488 *Nils Larsch <nla@trustcenter.de>*
12490 * Change internals of the EC library so that the functions
12491 EC_GROUP_set_generator()
12492 EC_GROUP_get_generator()
12493 EC_GROUP_get_order()
12494 EC_GROUP_get_cofactor()
12495 are implemented directly in crypto/ec/ec_lib.c and not dispatched
12496 to methods, which would lead to unnecessary code duplication when
12497 adding different types of curves.
12499 *Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller*
12501 * Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM
12502 arithmetic, and such that modified wNAFs are generated
12503 (which avoid length expansion in many cases).
12507 * Add a function EC_GROUP_check_discriminant() (defined via
12508 EC_METHOD) that verifies that the curve discriminant is non-zero.
12510 Add a function EC_GROUP_check() that makes some sanity tests
12511 on a EC_GROUP, its generator and order. This includes
12512 EC_GROUP_check_discriminant().
12514 *Nils Larsch <nla@trustcenter.de>*
12516 * Add ECDSA in new directory crypto/ecdsa/.
12518 Add applications 'openssl ecparam' and 'openssl ecdsa'
12519 (these are based on 'openssl dsaparam' and 'openssl dsa').
12521 ECDSA support is also included in various other files across the
12522 library. Most notably,
12523 - 'openssl req' now has a '-newkey ecdsa:file' option;
12524 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA;
12525 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and
12526 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make
12527 them suitable for ECDSA where domain parameters must be
12528 extracted before the specific public key;
12529 - ECDSA engine support has been added.
12531 *Nils Larsch <nla@trustcenter.de>*
12533 * Include some named elliptic curves, and add OIDs from X9.62,
12534 SECG, and WAP/WTLS. Each curve can be obtained from the new
12536 EC_GROUP_new_by_curve_name(),
12537 and the list of available named curves can be obtained with
12538 EC_get_builtin_curves().
12539 Also add a 'curve_name' member to EC_GROUP objects, which can be
12541 EC_GROUP_set_curve_name()
12542 EC_GROUP_get_curve_name()
12544 *Nils Larsch <larsch@trustcenter.de, Bodo Moeller*
12546 * Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
12547 was actually never needed) and in BN_mul(). The removal in BN_mul()
12548 required a small change in bn_mul_part_recursive() and the addition
12549 of the functions bn_cmp_part_words(), bn_sub_part_words() and
12550 bn_add_part_words(), which do the same thing as bn_cmp_words(),
12551 bn_sub_words() and bn_add_words() except they take arrays with
12556 ### Changes between 0.9.7l and 0.9.7m [23 Feb 2007]
12558 * Cleanse PEM buffers before freeing them since they may contain
12561 *Benjamin Bennett <ben@psc.edu>*
12563 * Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
12564 a ciphersuite string such as "DEFAULT:RSA" cannot enable
12565 authentication-only ciphersuites.
12569 * Since AES128 and AES256 share a single mask bit in the logic of
12570 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
12571 kludge to work properly if AES128 is available and AES256 isn't.
12575 * Expand security boundary to match 1.1.1 module.
12579 * Remove redundant features: hash file source, editing of test vectors
12580 modify fipsld to use external fips_premain.c signature.
12584 * New perl script mkfipsscr.pl to create shell scripts or batch files to
12585 run algorithm test programs.
12589 * Make algorithm test programs more tolerant of whitespace.
12593 * Have SSL/TLS server implementation tolerate "mismatched" record
12594 protocol version while receiving ClientHello even if the
12595 ClientHello is fragmented. (The server can't insist on the
12596 particular protocol version it has chosen before the ServerHello
12597 message has informed the client about his choice.)
12601 * Load error codes if they are not already present instead of using a
12602 static variable. This allows them to be cleanly unloaded and reloaded.
12606 ### Changes between 0.9.7k and 0.9.7l [28 Sep 2006]
12608 * Introduce limits to prevent malicious keys being able to
12609 cause a denial of service. ([CVE-2006-2940])
12611 *Steve Henson, Bodo Moeller*
12613 * Fix ASN.1 parsing of certain invalid structures that can result
12614 in a denial of service. ([CVE-2006-2937]) [Steve Henson]
12616 * Fix buffer overflow in SSL_get_shared_ciphers() function.
12617 ([CVE-2006-3738]) [Tavis Ormandy and Will Drewry, Google Security Team]
12619 * Fix SSL client code which could crash if connecting to a
12620 malicious SSLv2 server. ([CVE-2006-4343])
12622 *Tavis Ormandy and Will Drewry, Google Security Team*
12624 * Change ciphersuite string processing so that an explicit
12625 ciphersuite selects this one ciphersuite (so that "AES256-SHA"
12626 will no longer include "AES128-SHA"), and any other similar
12627 ciphersuite (same bitmap) from *other* protocol versions (so that
12628 "RC4-MD5" will still include both the SSL 2.0 ciphersuite and the
12629 SSL 3.0/TLS 1.0 ciphersuite). This is a backport combining
12630 changes from 0.9.8b and 0.9.8d.
12634 ### Changes between 0.9.7j and 0.9.7k [05 Sep 2006]
12636 * Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
12637 ([CVE-2006-4339]) [Ben Laurie and Google Security Team]
12639 * Change the Unix randomness entropy gathering to use poll() when
12640 possible instead of select(), since the latter has some
12641 undesirable limitations.
12643 *Darryl Miles via Richard Levitte and Bodo Moeller*
12645 * Disable rogue ciphersuites:
12647 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
12648 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
12649 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
12651 The latter two were purportedly from
12652 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
12655 Also deactivate the remaining ciphersuites from
12656 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
12657 unofficial, and the ID has long expired.
12661 * Fix RSA blinding Heisenbug (problems sometimes occurred on
12662 dual-core machines) and other potential thread-safety issues.
12666 ### Changes between 0.9.7i and 0.9.7j [04 May 2006]
12668 * Adapt fipsld and the build system to link against the validated FIPS
12669 module in FIPS mode.
12673 * Fixes for VC++ 2005 build under Windows.
12677 * Add new Windows build target VC-32-GMAKE for VC++. This uses GNU make
12678 from a Windows bash shell such as MSYS. It is autodetected from the
12679 "config" script when run from a VC++ environment. Modify standard VC++
12680 build to use fipscanister.o from the GNU make build.
12684 ### Changes between 0.9.7h and 0.9.7i [14 Oct 2005]
12686 * Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS.
12687 The value now differs depending on if you build for FIPS or not.
12688 BEWARE! A program linked with a shared FIPSed libcrypto can't be
12689 safely run with a non-FIPSed libcrypto, as it may crash because of
12690 the difference induced by this change.
12694 ### Changes between 0.9.7g and 0.9.7h [11 Oct 2005]
12696 * Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
12697 (part of SSL_OP_ALL). This option used to disable the
12698 countermeasure against man-in-the-middle protocol-version
12699 rollback in the SSL 2.0 server implementation, which is a bad
12700 idea. ([CVE-2005-2969])
12702 *Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
12703 for Information Security, National Institute of Advanced Industrial
12704 Science and Technology [AIST, Japan)]*
12706 * Minimal support for X9.31 signatures and PSS padding modes. This is
12707 mainly for FIPS compliance and not fully integrated at this stage.
12711 * For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform
12712 the exponentiation using a fixed-length exponent. (Otherwise,
12713 the information leaked through timing could expose the secret key
12714 after many signatures; cf. Bleichenbacher's attack on DSA with
12719 * Make a new fixed-window mod_exp implementation the default for
12720 RSA, DSA, and DH private-key operations so that the sequence of
12721 squares and multiplies and the memory access pattern are
12722 independent of the particular secret key. This will mitigate
12723 cache-timing and potential related attacks.
12725 BN_mod_exp_mont_consttime() is the new exponentiation implementation,
12726 and this is automatically used by BN_mod_exp_mont() if the new flag
12727 BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH
12728 will use this BN flag for private exponents unless the flag
12729 RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or
12730 DH_FLAG_NO_EXP_CONSTTIME, respectively, is set.
12732 *Matthew D Wood (Intel Corp), with some changes by Bodo Moeller*
12734 * Change the client implementation for SSLv23_method() and
12735 SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0
12736 Client Hello message format if the SSL_OP_NO_SSLv2 option is set.
12737 (Previously, the SSL 2.0 backwards compatible Client Hello
12738 message format would be used even with SSL_OP_NO_SSLv2.)
12742 * Add support for smime-type MIME parameter in S/MIME messages which some
12747 * New function BN_MONT_CTX_set_locked() to set montgomery parameters in
12748 a threadsafe manner. Modify rsa code to use new function and add calls
12749 to dsa and dh code (which had race conditions before).
12753 * Include the fixed error library code in the C error file definitions
12754 instead of fixing them up at runtime. This keeps the error code
12755 structures constant.
12759 ### Changes between 0.9.7f and 0.9.7g [11 Apr 2005]
12761 [NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after
12764 * Fixes for newer kerberos headers. NB: the casts are needed because
12765 the 'length' field is signed on one version and unsigned on another
12766 with no (?) obvious way to tell the difference, without these VC++
12767 complains. Also the "definition" of FAR (blank) is no longer included
12768 nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up
12769 some needed definitions.
12773 * Undo Cygwin change.
12777 * Added support for proxy certificates according to RFC 3820.
12778 Because they may be a security thread to unaware applications,
12779 they must be explicitly allowed in run-time. See
12780 docs/HOWTO/proxy_certificates.txt for further information.
12784 ### Changes between 0.9.7e and 0.9.7f [22 Mar 2005]
12786 * Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating
12787 server and client random values. Previously
12788 (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in
12789 less random data when sizeof(time_t) > 4 (some 64 bit platforms).
12791 This change has negligible security impact because:
12793 1. Server and client random values still have 24 bytes of pseudo random
12796 2. Server and client random values are sent in the clear in the initial
12799 3. The master secret is derived using the premaster secret (48 bytes in
12800 size for static RSA ciphersuites) as well as client server and random
12803 The OpenSSL team would like to thank the UK NISCC for bringing this issue
12806 *Stephen Henson, reported by UK NISCC*
12808 * Use Windows randomness collection on Cygwin.
12812 * Fix hang in EGD/PRNGD query when communication socket is closed
12813 prematurely by EGD/PRNGD.
12815 *Darren Tucker <dtucker@zip.com.au> via Lutz Jänicke, resolves #1014*
12817 * Prompt for pass phrases when appropriate for PKCS12 input format.
12821 * Back-port of selected performance improvements from development
12822 branch, as well as improved support for PowerPC platforms.
12826 * Add lots of checks for memory allocation failure, error codes to indicate
12827 failure and freeing up memory if a failure occurs.
12829 *Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson*
12831 * Add new -passin argument to dgst.
12835 * Perform some character comparisons of different types in X509_NAME_cmp:
12836 this is needed for some certificates that re-encode DNs into UTF8Strings
12837 (in violation of RFC3280) and can't or won't issue name rollover
12842 * Make an explicit check during certificate validation to see that
12843 the CA setting in each certificate on the chain is correct. As a
12844 side effect always do the following basic checks on extensions,
12845 not just when there's an associated purpose to the check:
12847 - if there is an unhandled critical extension (unless the user
12848 has chosen to ignore this fault)
12849 - if the path length has been exceeded (if one is set at all)
12850 - that certain extensions fit the associated purpose (if one has
12855 ### Changes between 0.9.7d and 0.9.7e [25 Oct 2004]
12857 * Avoid a race condition when CRLs are checked in a multi threaded
12858 environment. This would happen due to the reordering of the revoked
12859 entries during signature checking and serial number lookup. Now the
12860 encoding is cached and the serial number sort performed under a lock.
12861 Add new STACK function sk_is_sorted().
12865 * Add Delta CRL to the extension code.
12869 * Various fixes to s3_pkt.c so alerts are sent properly.
12871 *David Holmes <d.holmes@f5.com>*
12873 * Reduce the chances of duplicate issuer name and serial numbers (in
12874 violation of RFC3280) using the OpenSSL certificate creation utilities.
12875 This is done by creating a random 64 bit value for the initial serial
12876 number when a serial number file is created or when a self signed
12877 certificate is created using 'openssl req -x509'. The initial serial
12878 number file is created using 'openssl x509 -next_serial' in CA.pl
12879 rather than being initialized to 1.
12883 ### Changes between 0.9.7c and 0.9.7d [17 Mar 2004]
12885 * Fix null-pointer assignment in do_change_cipher_spec() revealed
12886 by using the Codenomicon TLS Test Tool ([CVE-2004-0079])
12888 *Joe Orton, Steve Henson*
12890 * Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites
12893 *Joe Orton, Steve Henson*
12895 * Make it possible to have multiple active certificates with the same
12896 subject in the CA index file. This is done only if the keyword
12897 'unique_subject' is set to 'no' in the main CA section (default
12898 if 'CA_default') of the configuration file. The value is saved
12899 with the database itself in a separate index attribute file,
12900 named like the index file with '.attr' appended to the name.
12904 * X509 verify fixes. Disable broken certificate workarounds when
12905 X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if
12906 keyUsage extension present. Don't accept CRLs with unhandled critical
12907 extensions: since verify currently doesn't process CRL extensions this
12908 rejects a CRL with *any* critical extensions. Add new verify error codes
12913 * When creating an OCSP nonce use an OCTET STRING inside the extnValue.
12914 A clarification of RFC2560 will require the use of OCTET STRINGs and
12915 some implementations cannot handle the current raw format. Since OpenSSL
12916 copies and compares OCSP nonces as opaque blobs without any attempt at
12917 parsing them this should not create any compatibility issues.
12921 * New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when
12922 calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without
12923 this HMAC (and other) operations are several times slower than OpenSSL
12928 * Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex().
12930 *Peter Sylvester <Peter.Sylvester@EdelWeb.fr>*
12932 * Use the correct content when signing type "other".
12936 ### Changes between 0.9.7b and 0.9.7c [30 Sep 2003]
12938 * Fix various bugs revealed by running the NISCC test suite:
12940 Stop out of bounds reads in the ASN1 code when presented with
12941 invalid tags (CVE-2003-0543 and CVE-2003-0544).
12943 Free up ASN1_TYPE correctly if ANY type is invalid ([CVE-2003-0545]).
12945 If verify callback ignores invalid public key errors don't try to check
12946 certificate signature with the NULL public key.
12950 * New -ignore_err option in ocsp application to stop the server
12951 exiting on the first error in a request.
12955 * In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
12956 if the server requested one: as stated in TLS 1.0 and SSL 3.0
12961 * In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
12962 extra data after the compression methods not only for TLS 1.0
12963 but also for SSL 3.0 (as required by the specification).
12965 *Bodo Moeller; problem pointed out by Matthias Loepfe*
12967 * Change X509_certificate_type() to mark the key as exported/exportable
12968 when it's 512 *bits* long, not 512 bytes.
12972 * Change AES_cbc_encrypt() so it outputs exact multiple of
12973 blocks during encryption.
12977 * Various fixes to base64 BIO and non blocking I/O. On write
12978 flushes were not handled properly if the BIO retried. On read
12979 data was not being buffered properly and had various logic bugs.
12980 This also affects blocking I/O when the data being decoded is a
12985 * Various S/MIME bugfixes and compatibility changes:
12986 output correct application/pkcs7 MIME type if
12987 PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures.
12988 Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening
12989 of files as .eml work). Correctly handle very long lines in MIME
12994 ### Changes between 0.9.7a and 0.9.7b [10 Apr 2003]
12996 * Countermeasure against the Klima-Pokorny-Rosa extension of
12997 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
12998 a protocol version number mismatch like a decryption error
12999 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
13003 * Turn on RSA blinding by default in the default implementation
13004 to avoid a timing attack. Applications that don't want it can call
13005 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
13006 They would be ill-advised to do so in most cases.
13008 *Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller*
13010 * Change RSA blinding code so that it works when the PRNG is not
13011 seeded (in this case, the secret RSA exponent is abused as
13012 an unpredictable seed -- if it is not unpredictable, there
13013 is no point in blinding anyway). Make RSA blinding thread-safe
13014 by remembering the creator's thread ID in rsa->blinding and
13015 having all other threads use local one-time blinding factors
13016 (this requires more computation than sharing rsa->blinding, but
13017 avoids excessive locking; and if an RSA object is not shared
13018 between threads, blinding will still be very fast).
13022 * Fixed a typo bug that would cause ENGINE_set_default() to set an
13023 ENGINE as defaults for all supported algorithms irrespective of
13024 the 'flags' parameter. 'flags' is now honoured, so applications
13025 should make sure they are passing it correctly.
13029 * Target "mingw" now allows native Windows code to be generated in
13030 the Cygwin environment as well as with the MinGW compiler.
13034 ### Changes between 0.9.7 and 0.9.7a [19 Feb 2003]
13036 * In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
13037 via timing by performing a MAC computation even if incorrect
13038 block cipher padding has been found. This is a countermeasure
13039 against active attacks where the attacker has to distinguish
13040 between bad padding and a MAC verification error. ([CVE-2003-0078])
13042 *Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
13043 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
13044 Martin Vuagnoux (EPFL, Ilion)*
13046 * Make the no-err option work as intended. The intention with no-err
13047 is not to have the whole error stack handling routines removed from
13048 libcrypto, it's only intended to remove all the function name and
13049 reason texts, thereby removing some of the footprint that may not
13050 be interesting if those errors aren't displayed anyway.
13052 NOTE: it's still possible for any application or module to have its
13053 own set of error texts inserted. The routines are there, just not
13054 used by default when no-err is given.
13058 * Add support for FreeBSD on IA64.
13060 *dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454*
13062 * Adjust DES_cbc_cksum() so it returns the same value as the MIT
13063 Kerberos function mit_des_cbc_cksum(). Before this change,
13064 the value returned by DES_cbc_cksum() was like the one from
13065 mit_des_cbc_cksum(), except the bytes were swapped.
13067 *Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte*
13069 * Allow an application to disable the automatic SSL chain building.
13070 Before this a rather primitive chain build was always performed in
13071 ssl3_output_cert_chain(): an application had no way to send the
13072 correct chain if the automatic operation produced an incorrect result.
13074 Now the chain builder is disabled if either:
13076 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert().
13078 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set.
13080 The reasoning behind this is that an application would not want the
13081 auto chain building to take place if extra chain certificates are
13082 present and it might also want a means of sending no additional
13083 certificates (for example the chain has two certificates and the
13088 * Add the possibility to build without the ENGINE framework.
13090 *Steven Reddie <smr@essemer.com.au> via Richard Levitte*
13092 * Under Win32 gmtime() can return NULL: check return value in
13093 OPENSSL_gmtime(). Add error code for case where gmtime() fails.
13097 * DSA routines: under certain error conditions uninitialized BN objects
13098 could be freed. Solution: make sure initialization is performed early
13099 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>,
13100 Nils Larsch <nla@trustcenter.de> via PR#459)
13104 * Another fix for SSLv2 session ID handling: the session ID was incorrectly
13105 checked on reconnect on the client side, therefore session resumption
13106 could still fail with a "ssl session id is different" error. This
13107 behaviour is masked when SSL_OP_ALL is used due to
13108 SSL_OP_MICROSOFT_SESS_ID_BUG being set.
13109 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
13110 followup to PR #377.
13114 * IA-32 assembler support enhancements: unified ELF targets, support
13115 for SCO/Caldera platforms, fix for Cygwin shared build.
13119 * Add support for FreeBSD on sparc64. As a consequence, support for
13120 FreeBSD on non-x86 processors is separate from x86 processors on
13121 the config script, much like the NetBSD support.
13123 *Richard Levitte & Kris Kennaway <kris@obsecurity.org>*
13125 ### Changes between 0.9.6h and 0.9.7 [31 Dec 2002]
13127 [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after
13130 * Fix session ID handling in SSLv2 client code: the SERVER FINISHED
13131 code (06) was taken as the first octet of the session ID and the last
13132 octet was ignored consequently. As a result SSLv2 client side session
13133 caching could not have worked due to the session ID mismatch between
13135 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
13140 * Change the declaration of needed Kerberos libraries to use EX_LIBS
13141 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is
13146 * The hw_ncipher.c engine requires dynamic locks. Unfortunately, it
13147 seems that in spite of existing for more than a year, many application
13148 author have done nothing to provide the necessary callbacks, which
13149 means that this particular engine will not work properly anywhere.
13150 This is a very unfortunate situation which forces us, in the name
13151 of usability, to give the hw_ncipher.c a static lock, which is part
13153 NOTE: This is for the 0.9.7 series ONLY. This hack will never
13154 appear in 0.9.8 or later. We EXPECT application authors to have
13155 dealt properly with this when 0.9.8 is released (unless we actually
13156 make such changes in the libcrypto locking code that changes will
13157 have to be made anyway).
13161 * In asn1_d2i_read_bio() repeatedly call BIO_read() until all content
13162 octets have been read, EOF or an error occurs. Without this change
13163 some truncated ASN1 structures will not produce an error.
13167 * Disable Heimdal support, since it hasn't been fully implemented.
13168 Still give the possibility to force the use of Heimdal, but with
13169 warnings and a request that patches get sent to openssl-dev.
13173 * Add the VC-CE target, introduce the WINCE sysname, and add
13174 INSTALL.WCE and appropriate conditionals to make it build.
13176 *Steven Reddie <smr@essemer.com.au> via Richard Levitte*
13178 * Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and
13179 cygssl-x.y.z.dll, where x, y and z are the major, minor and
13180 edit numbers of the version.
13182 *Corinna Vinschen <vinschen@redhat.com> and Richard Levitte*
13184 * Introduce safe string copy and catenation functions
13185 (BUF_strlcpy() and BUF_strlcat()).
13187 *Ben Laurie (CHATS) and Richard Levitte*
13189 * Avoid using fixed-size buffers for one-line DNs.
13191 *Ben Laurie (CHATS)*
13193 * Add BUF_MEM_grow_clean() to avoid information leakage when
13194 resizing buffers containing secrets, and use where appropriate.
13196 *Ben Laurie (CHATS)*
13198 * Avoid using fixed size buffers for configuration file location.
13200 *Ben Laurie (CHATS)*
13202 * Avoid filename truncation for various CA files.
13204 *Ben Laurie (CHATS)*
13206 * Use sizeof in preference to magic numbers.
13208 *Ben Laurie (CHATS)*
13210 * Avoid filename truncation in cert requests.
13212 *Ben Laurie (CHATS)*
13214 * Add assertions to check for (supposedly impossible) buffer
13217 *Ben Laurie (CHATS)*
13219 * Don't cache truncated DNS entries in the local cache (this could
13220 potentially lead to a spoofing attack).
13222 *Ben Laurie (CHATS)*
13224 * Fix various buffers to be large enough for hex/decimal
13225 representations in a platform independent manner.
13227 *Ben Laurie (CHATS)*
13229 * Add CRYPTO_realloc_clean() to avoid information leakage when
13230 resizing buffers containing secrets, and use where appropriate.
13232 *Ben Laurie (CHATS)*
13234 * Add BIO_indent() to avoid much slightly worrying code to do
13237 *Ben Laurie (CHATS)*
13239 * Convert sprintf()/BIO_puts() to BIO_printf().
13241 *Ben Laurie (CHATS)*
13243 * buffer_gets() could terminate with the buffer only half
13246 *Ben Laurie (CHATS)*
13248 * Add assertions to prevent user-supplied crypto functions from
13249 overflowing internal buffers by having large block sizes, etc.
13251 *Ben Laurie (CHATS)*
13253 * New OPENSSL_assert() macro (similar to assert(), but enabled
13256 *Ben Laurie (CHATS)*
13258 * Eliminate unused copy of key in RC4.
13260 *Ben Laurie (CHATS)*
13262 * Eliminate unused and incorrectly sized buffers for IV in pem.h.
13264 *Ben Laurie (CHATS)*
13266 * Fix off-by-one error in EGD path.
13268 *Ben Laurie (CHATS)*
13270 * If RANDFILE path is too long, ignore instead of truncating.
13272 *Ben Laurie (CHATS)*
13274 * Eliminate unused and incorrectly sized X.509 structure
13277 *Ben Laurie (CHATS)*
13279 * Eliminate unused and dangerous function knumber().
13281 *Ben Laurie (CHATS)*
13283 * Eliminate unused and dangerous structure, KSSL_ERR.
13285 *Ben Laurie (CHATS)*
13287 * Protect against overlong session ID context length in an encoded
13288 session object. Since these are local, this does not appear to be
13291 *Ben Laurie (CHATS)*
13293 * Change from security patch (see 0.9.6e below) that did not affect
13294 the 0.9.6 release series:
13296 Remote buffer overflow in SSL3 protocol - an attacker could
13297 supply an oversized master key in Kerberos-enabled versions.
13300 *Ben Laurie (CHATS)*
13302 * Change the SSL kerb5 codes to match RFC 2712.
13306 * Make -nameopt work fully for req and add -reqopt switch.
13308 *Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson*
13310 * The "block size" for block ciphers in CFB and OFB mode should be 1.
13312 *Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>*
13314 * Make sure tests can be performed even if the corresponding algorithms
13315 have been removed entirely. This was also the last step to make
13316 OpenSSL compilable with DJGPP under all reasonable conditions.
13318 *Richard Levitte, Doug Kaufman <dkaufman@rahul.net>*
13320 * Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT
13321 to allow version independent disabling of normally unselected ciphers,
13322 which may be activated as a side-effect of selecting a single cipher.
13324 (E.g., cipher list string "RSA" enables ciphersuites that are left
13325 out of "ALL" because they do not provide symmetric encryption.
13326 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.)
13328 *Lutz Jaenicke, Bodo Moeller*
13330 * Add appropriate support for separate platform-dependent build
13331 directories. The recommended way to make a platform-dependent
13332 build directory is the following (tested on Linux), maybe with
13335 # Place yourself outside of the OpenSSL source tree. In
13336 # this example, the environment variable OPENSSL_SOURCE
13337 # is assumed to contain the absolute OpenSSL source directory.
13338 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`"
13339 cd objtree/"`uname -s`-`uname -r`-`uname -m`"
13340 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do
13341 mkdir -p `dirname $F`
13342 ln -s $OPENSSL_SOURCE/$F $F
13345 To be absolutely sure not to disturb the source tree, a "make clean"
13346 is a good thing. If it isn't successful, don't worry about it,
13347 it probably means the source directory is very clean.
13351 * Make sure any ENGINE control commands make local copies of string
13352 pointers passed to them whenever necessary. Otherwise it is possible
13353 the caller may have overwritten (or deallocated) the original string
13354 data when a later ENGINE operation tries to use the stored values.
13356 *Götz Babin-Ebell <babinebell@trustcenter.de>*
13358 * Improve diagnostics in file reading and command-line digests.
13360 *Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>*
13362 * Add AES modes CFB and OFB to the object database. Correct an
13363 error in AES-CFB decryption.
13367 * Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this
13368 allows existing EVP_CIPHER_CTX structures to be reused after
13369 calling `EVP_*Final()`. This behaviour is used by encryption
13370 BIOs and some applications. This has the side effect that
13371 applications must explicitly clean up cipher contexts with
13372 EVP_CIPHER_CTX_cleanup() or they will leak memory.
13376 * Check the values of dna and dnb in bn_mul_recursive before calling
13377 bn_mul_comba (a non zero value means the a or b arrays do not contain
13378 n2 elements) and fallback to bn_mul_normal if either is not zero.
13382 * Fix escaping of non-ASCII characters when using the -subj option
13383 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>)
13387 * Make object definitions compliant to LDAP (RFC2256): SN is the short
13388 form for "surname", serialNumber has no short form.
13389 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798;
13390 therefore remove "mail" short name for "internet 7".
13391 The OID for unique identifiers in X509 certificates is
13392 x500UniqueIdentifier, not uniqueIdentifier.
13393 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>)
13397 * Add an "init" command to the ENGINE config module and auto initialize
13398 ENGINEs. Without any "init" command the ENGINE will be initialized
13399 after all ctrl commands have been executed on it. If init=1 the
13400 ENGINE is initialized at that point (ctrls before that point are run
13401 on the uninitialized ENGINE and after on the initialized one). If
13402 init=0 then the ENGINE will not be initialized at all.
13406 * Fix the 'app_verify_callback' interface so that the user-defined
13407 argument is actually passed to the callback: In the
13408 SSL_CTX_set_cert_verify_callback() prototype, the callback
13409 declaration has been changed from
13412 int (*cb)(X509_STORE_CTX *,void *);
13413 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
13414 i=s->ctx->app_verify_callback(&ctx)
13415 has been changed into
13416 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
13418 To update applications using SSL_CTX_set_cert_verify_callback(),
13419 a dummy argument can be added to their callback functions.
13421 *D. K. Smetters <smetters@parc.xerox.com>*
13423 * Added the '4758cca' ENGINE to support IBM 4758 cards.
13425 *Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe*
13427 * Add and OPENSSL_LOAD_CONF define which will cause
13428 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
13429 This allows older applications to transparently support certain
13430 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
13431 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
13432 load the config file and OPENSSL_add_all_algorithms_conf() which will
13433 always load it have also been added.
13437 * Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
13438 Adjust NIDs and EVP layer.
13440 *Stephen Sprunk <stephen@sprunk.org> and Richard Levitte*
13442 * Config modules support in openssl utility.
13444 Most commands now load modules from the config file,
13445 though in a few (such as version) this isn't done
13446 because it couldn't be used for anything.
13448 In the case of ca and req the config file used is
13449 the same as the utility itself: that is the -config
13450 command line option can be used to specify an
13455 * Move default behaviour from OPENSSL_config(). If appname is NULL
13456 use "openssl_conf" if filename is NULL use default openssl config file.
13460 * Add an argument to OPENSSL_config() to allow the use of an alternative
13461 config section name. Add a new flag to tolerate a missing config file
13462 and move code to CONF_modules_load_file().
13466 * Support for crypto accelerator cards from Accelerated Encryption
13467 Processing, www.aep.ie. (Use engine 'aep')
13468 The support was copied from 0.9.6c [engine] and adapted/corrected
13469 to work with the new engine framework.
13471 *AEP Inc. and Richard Levitte*
13473 * Support for SureWare crypto accelerator cards from Baltimore
13474 Technologies. (Use engine 'sureware')
13475 The support was copied from 0.9.6c [engine] and adapted
13476 to work with the new engine framework.
13480 * Have the CHIL engine fork-safe (as defined by nCipher) and actually
13481 make the newer ENGINE framework commands for the CHIL engine work.
13483 *Toomas Kiisk <vix@cyber.ee> and Richard Levitte*
13485 * Make it possible to produce shared libraries on ReliantUNIX.
13487 *Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte*
13489 * Add the configuration target debug-linux-ppro.
13490 Make 'openssl rsa' use the general key loading routines
13491 implemented in `apps.c`, and make those routines able to
13492 handle the key format FORMAT_NETSCAPE and the variant
13495 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
13497 * Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
13499 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
13501 * Add -keyform to rsautl, and document -engine.
13503 *Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>*
13505 * Change BIO_new_file (crypto/bio/bss_file.c) to use new
13506 BIO_R_NO_SUCH_FILE error code rather than the generic
13507 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
13511 * Add new functions
13512 ERR_peek_last_error
13513 ERR_peek_last_error_line
13514 ERR_peek_last_error_line_data.
13515 These are similar to
13517 ERR_peek_error_line
13518 ERR_peek_error_line_data,
13519 but report on the latest error recorded rather than the first one
13520 still in the error queue.
13522 *Ben Laurie, Bodo Moeller*
13524 * default_algorithms option in ENGINE config module. This allows things
13526 default_algorithms = ALL
13527 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
13531 * Preliminary ENGINE config module.
13535 * New experimental application configuration code.
13539 * Change the AES code to follow the same name structure as all other
13540 symmetric ciphers, and behave the same way. Move everything to
13541 the directory crypto/aes, thereby obsoleting crypto/rijndael.
13543 *Stephen Sprunk <stephen@sprunk.org> and Richard Levitte*
13545 * SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
13547 *Ben Laurie and Theo de Raadt*
13549 * Add option to output public keys in req command.
13551 *Massimiliano Pala madwolf@openca.org*
13553 * Use wNAFs in EC_POINTs_mul() for improved efficiency
13554 (up to about 10% better than before for P-192 and P-224).
13558 * New functions/macros
13560 SSL_CTX_set_msg_callback(ctx, cb)
13561 SSL_CTX_set_msg_callback_arg(ctx, arg)
13562 SSL_set_msg_callback(ssl, cb)
13563 SSL_set_msg_callback_arg(ssl, arg)
13565 to request calling a callback function
13567 void cb(int write_p, int version, int content_type,
13568 const void *buf, size_t len, SSL *ssl, void *arg)
13570 whenever a protocol message has been completely received
13571 (write_p == 0) or sent (write_p == 1). Here 'version' is the
13572 protocol version according to which the SSL library interprets
13573 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
13574 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
13575 the content type as defined in the SSL 3.0/TLS 1.0 protocol
13576 specification (change_cipher_spec(20), alert(21), handshake(22)).
13577 'buf' and 'len' point to the actual message, 'ssl' to the
13578 SSL object, and 'arg' is the application-defined value set by
13579 SSL[_CTX]_set_msg_callback_arg().
13581 'openssl s_client' and 'openssl s_server' have new '-msg' options
13582 to enable a callback that displays all protocol messages.
13586 * Change the shared library support so shared libraries are built as
13587 soon as the corresponding static library is finished, and thereby get
13588 openssl and the test programs linked against the shared library.
13589 This still only happens when the keyword "shard" has been given to
13590 the configuration scripts.
13592 NOTE: shared library support is still an experimental thing, and
13593 backward binary compatibility is still not guaranteed.
13595 *"Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte*
13597 * Add support for Subject Information Access extension.
13599 *Peter Sylvester <Peter.Sylvester@EdelWeb.fr>*
13601 * Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
13602 additional bytes when new memory had to be allocated, not just
13603 when reusing an existing buffer.
13607 * New command line and configuration option 'utf8' for the req command.
13608 This allows field values to be specified as UTF8 strings.
13612 * Add -multi and -mr options to "openssl speed" - giving multiple parallel
13613 runs for the former and machine-readable output for the latter.
13617 * Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
13618 of the e-mail address in the DN (i.e., it will go into a certificate
13619 extension only). The new configuration file option 'email_in_dn = no'
13620 has the same effect.
13622 *Massimiliano Pala madwolf@openca.org*
13624 * Change all functions with names starting with `des_` to be starting
13625 with `DES_` instead. Add wrappers that are compatible with libdes,
13626 but are named `_ossl_old_des_*`. Finally, add macros that map the
13627 `des_*` symbols to the corresponding `_ossl_old_des_*` if libdes
13628 compatibility is desired. If OpenSSL 0.9.6c compatibility is
13629 desired, the `des_*` symbols will be mapped to `DES_*`, with one
13632 Since we provide two compatibility mappings, the user needs to
13633 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
13634 compatibility is desired. The default (i.e., when that macro
13635 isn't defined) is OpenSSL 0.9.6c compatibility.
13637 There are also macros that enable and disable the support of old
13638 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
13639 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
13640 are defined, the default will apply: to support the old des routines.
13642 In either case, one must include openssl/des.h to get the correct
13643 definitions. Do not try to just include openssl/des_old.h, that
13646 NOTE: This is a major break of an old API into a new one. Software
13647 authors are encouraged to switch to the `DES_` style functions. Some
13648 time in the future, des_old.h and the libdes compatibility functions
13649 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
13650 default), and then completely removed.
13654 * Test for certificates which contain unsupported critical extensions.
13655 If such a certificate is found during a verify operation it is
13656 rejected by default: this behaviour can be overridden by either
13657 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
13658 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
13659 X509_supported_extension() has also been added which returns 1 if a
13660 particular extension is supported.
13664 * Modify the behaviour of EVP cipher functions in similar way to digests
13665 to retain compatibility with existing code.
13669 * Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
13670 compatibility with existing code. In particular the 'ctx' parameter does
13671 not have to be to be initialized before the call to EVP_DigestInit() and
13672 it is tidied up after a call to EVP_DigestFinal(). New function
13673 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
13674 EVP_MD_CTX_copy() changed to not require the destination to be
13675 initialized valid and new function EVP_MD_CTX_copy_ex() added which
13676 requires the destination to be valid.
13678 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
13679 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
13683 * Change ssl3_get_message (ssl/s3_both.c) and the functions using it
13684 so that complete 'Handshake' protocol structures are kept in memory
13685 instead of overwriting 'msg_type' and 'length' with 'body' data.
13689 * Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
13691 *Massimo Santin via Richard Levitte*
13693 * Major restructuring to the underlying ENGINE code. This includes
13694 reduction of linker bloat, separation of pure "ENGINE" manipulation
13695 (initialisation, etc) from functionality dealing with implementations
13696 of specific crypto interfaces. This change also introduces integrated
13697 support for symmetric ciphers and digest implementations - so ENGINEs
13698 can now accelerate these by providing EVP_CIPHER and EVP_MD
13699 implementations of their own. This is detailed in
13700 [crypto/engine/README.md](crypto/engine/README.md)
13701 as it couldn't be adequately described here. However, there are a few
13702 API changes worth noting - some RSA, DSA, DH, and RAND functions that
13703 were changed in the original introduction of ENGINE code have now
13704 reverted back - the hooking from this code to ENGINE is now a good
13705 deal more passive and at run-time, operations deal directly with
13706 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
13707 dereferencing through an ENGINE pointer any more. Also, the ENGINE
13708 functions dealing with `BN_MOD_EXP[_CRT]` handlers have been removed -
13709 they were not being used by the framework as there is no concept of a
13710 BIGNUM_METHOD and they could not be generalised to the new
13711 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
13712 ENGINE_cpy() has been removed as it cannot be consistently defined in
13717 * Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
13721 * Change mkdef.pl to sort symbols that get the same entry number,
13722 and make sure the automatically generated functions `ERR_load_*`
13723 become part of libeay.num as well.
13727 * New function SSL_renegotiate_pending(). This returns true once
13728 renegotiation has been requested (either SSL_renegotiate() call
13729 or HelloRequest/ClientHello received from the peer) and becomes
13730 false once a handshake has been completed.
13731 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
13732 sends a HelloRequest, but does not ensure that a handshake takes
13733 place. SSL_renegotiate_pending() is useful for checking if the
13734 client has followed the request.)
13738 * New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
13739 By default, clients may request session resumption even during
13740 renegotiation (if session ID contexts permit); with this option,
13741 session resumption is possible only in the first handshake.
13743 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes
13744 more bits available for options that should not be part of
13745 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION).
13749 * Add some demos for certificate and certificate request creation.
13753 * Make maximum certificate chain size accepted from the peer application
13754 settable (`SSL*_get/set_max_cert_list()`), as proposed by
13755 "Douglas E. Engert" <deengert@anl.gov>.
13759 * Add support for shared libraries for Unixware-7
13760 (Boyd Lynn Gerber <gerberb@zenez.com>).
13764 * Add a "destroy" handler to ENGINEs that allows structural cleanup to
13765 be done prior to destruction. Use this to unload error strings from
13766 ENGINEs that load their own error strings. NB: This adds two new API
13767 functions to "get" and "set" this destroy handler in an ENGINE.
13771 * Alter all existing ENGINE implementations (except "openssl" and
13772 "openbsd") to dynamically instantiate their own error strings. This
13773 makes them more flexible to be built both as statically-linked ENGINEs
13774 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
13775 Also, add stub code to each that makes building them as self-contained
13776 shared-libraries easier (see [README-Engine.md](README-Engine.md)).
13780 * Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
13781 implementations into applications that are completely implemented in
13782 self-contained shared-libraries. The "dynamic" ENGINE exposes control
13783 commands that can be used to configure what shared-library to load and
13784 to control aspects of the way it is handled. Also, made an update to
13785 the [README-Engine.md](README-Engine.md) file
13786 that brings its information up-to-date and
13787 provides some information and instructions on the "dynamic" ENGINE
13788 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
13792 * Make it possible to unload ranges of ERR strings with a new
13793 "ERR_unload_strings" function.
13797 * Add a copy() function to EVP_MD.
13801 * Make EVP_MD routines take a context pointer instead of just the
13802 md_data void pointer.
13806 * Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
13807 that the digest can only process a single chunk of data
13808 (typically because it is provided by a piece of
13809 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
13810 is only going to provide a single chunk of data, and hence the
13811 framework needn't accumulate the data for oneshot drivers.
13815 * As with "ERR", make it possible to replace the underlying "ex_data"
13816 functions. This change also alters the storage and management of global
13817 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
13818 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
13819 index counters. The API functions that use this state have been changed
13820 to take a "class_index" rather than pointers to the class's local STACK
13821 and counter, and there is now an API function to dynamically create new
13822 classes. This centralisation allows us to (a) plug a lot of the
13823 thread-safety problems that existed, and (b) makes it possible to clean
13824 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
13825 such data would previously have always leaked in application code and
13826 workarounds were in place to make the memory debugging turn a blind eye
13827 to it. Application code that doesn't use this new function will still
13828 leak as before, but their memory debugging output will announce it now
13829 rather than letting it slide.
13831 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
13832 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
13833 has a return value to indicate success or failure.
13837 * Make it possible to replace the underlying "ERR" functions such that the
13838 global state (2 LHASH tables and 2 locks) is only used by the "default"
13839 implementation. This change also adds two functions to "get" and "set"
13840 the implementation prior to it being automatically set the first time
13841 any other ERR function takes place. Ie. an application can call "get",
13842 pass the return value to a module it has just loaded, and that module
13843 can call its own "set" function using that value. This means the
13844 module's "ERR" operations will use (and modify) the error state in the
13845 application and not in its own statically linked copy of OpenSSL code.
13849 * Give DH, DSA, and RSA types their own `*_up_ref()` function to increment
13850 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
13851 the operation, and provides a more encapsulated way for external code
13852 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
13853 to use these functions rather than manually incrementing the counts.
13855 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
13859 * Add EVP test program.
13863 * Add symmetric cipher support to ENGINE. Expect the API to change!
13867 * New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
13868 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
13869 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
13870 These allow a CRL to be built without having to access X509_CRL fields
13871 directly. Modify 'ca' application to use new functions.
13875 * Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
13876 bug workarounds. Rollback attack detection is a security feature.
13877 The problem will only arise on OpenSSL servers when TLSv1 is not
13878 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
13879 Software authors not wanting to support TLSv1 will have special reasons
13880 for their choice and can explicitly enable this option.
13882 *Bodo Moeller, Lutz Jaenicke*
13884 * Rationalise EVP so it can be extended: don't include a union of
13885 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX
13886 (similar to those existing for EVP_CIPHER_CTX).
13891 EVP_MD_CTX_init(&md); /* new function call */
13892 EVP_DigestInit(&md, EVP_sha1());
13893 EVP_DigestUpdate(&md, in, len);
13894 EVP_DigestFinal(&md, out, NULL);
13895 EVP_MD_CTX_cleanup(&md); /* new function call */
13899 * Make DES key schedule conform to the usual scheme, as well as
13900 correcting its structure. This means that calls to DES functions
13901 now have to pass a pointer to a des_key_schedule instead of a
13902 plain des_key_schedule (which was actually always a pointer
13905 des_key_schedule ks;
13907 des_set_key_checked(..., &ks);
13908 des_ncbc_encrypt(..., &ks, ...);
13910 (Note that a later change renames 'des_...' into 'DES_...'.)
13914 * Initial reduction of linker bloat: the use of some functions, such as
13915 PEM causes large amounts of unused functions to be linked in due to
13916 poor organisation. For example pem_all.c contains every PEM function
13917 which has a knock on effect of linking in large amounts of (unused)
13918 ASN1 code. Grouping together similar functions and splitting unrelated
13919 functions prevents this.
13923 * Cleanup of EVP macros.
13927 * Change historical references to `{NID,SN,LN}_des_ede` and ede3 to add the
13928 correct `_ecb suffix`.
13932 * Add initial OCSP responder support to ocsp application. The
13933 revocation information is handled using the text based index
13934 use by the ca application. The responder can either handle
13935 requests generated internally, supplied in files (for example
13936 via a CGI script) or using an internal minimal server.
13940 * Add configuration choices to get zlib compression for TLS.
13944 * Changes to Kerberos SSL for RFC 2712 compliance:
13945 1. Implemented real KerberosWrapper, instead of just using
13946 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
13947 2. Implemented optional authenticator field of KerberosWrapper.
13949 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
13950 and authenticator structs; see crypto/krb5/.
13952 Generalized Kerberos calls to support multiple Kerberos libraries.
13953 *Vern Staats <staatsvr@asc.hpc.mil>, Jeffrey Altman <jaltman@columbia.edu>
13954 via Richard Levitte*
13956 * Cause 'openssl speed' to use fully hard-coded DSA keys as it
13957 already does with RSA. testdsa.h now has 'priv_key/pub_key'
13958 values for each of the key sizes rather than having just
13959 parameters (and 'speed' generating keys each time).
13963 * Speed up EVP routines.
13966 pe 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
13967 s-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
13968 s-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
13969 s-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
13971 s-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
13972 s-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
13973 s-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
13976 s-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
13978 s-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
13982 * Added the OS2-EMX target.
13984 *"Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte*
13986 * Rewrite commands to use `NCONF` routines instead of the old `CONF`.
13987 New functions to support `NCONF` routines in extension code.
13988 New function `CONF_set_nconf()`
13989 to allow functions which take an `NCONF` to also handle the old `LHASH`
13990 structure: this means that the old `CONF` compatible routines can be
13991 retained (in particular w.rt. extensions) without having to duplicate the
13992 code. New function `X509V3_add_ext_nconf_sk()` to add extensions to a stack.
13996 * Enhance the general user interface with mechanisms for inner control
13997 and with possibilities to have yes/no kind of prompts.
14001 * Change all calls to low-level digest routines in the library and
14002 applications to use EVP. Add missing calls to HMAC_cleanup() and
14003 don't assume HMAC_CTX can be copied using memcpy().
14005 *Verdon Walker <VWalker@novell.com>, Steve Henson*
14007 * Add the possibility to control engines through control names but with
14008 arbitrary arguments instead of just a string.
14009 Change the key loaders to take a UI_METHOD instead of a callback
14010 function pointer. NOTE: this breaks binary compatibility with earlier
14011 versions of OpenSSL [engine].
14012 Adapt the nCipher code for these new conditions and add a card insertion
14017 * Enhance the general user interface with mechanisms to better support
14018 dialog box interfaces, application-defined prompts, the possibility
14019 to use defaults (for example default passwords from somewhere else)
14020 and interrupts/cancellations.
14024 * Tidy up PKCS#12 attribute handling. Add support for the CSP name
14025 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
14029 * Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
14030 tidy up some unnecessarily weird code in 'sk_new()').
14032 *Geoff, reported by Diego Tartara <dtartara@novamens.com>*
14034 * Change the key loading routines for ENGINEs to use the same kind
14035 callback (pem_password_cb) as all other routines that need this
14040 * Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
14041 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
14042 than this minimum value is recommended.
14046 * New random seeder for OpenVMS, using the system process statistics
14047 that are easily reachable.
14051 * Windows apparently can't transparently handle global
14052 variables defined in DLLs. Initialisations such as:
14054 const ASN1_ITEM *it = &ASN1_INTEGER_it;
14056 won't compile. This is used by the any applications that need to
14057 declare their own ASN1 modules. This was fixed by adding the option
14058 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
14059 needed for static libraries under Win32.
14063 * New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
14064 setting of purpose and trust fields. New X509_STORE trust and
14065 purpose functions and tidy up setting in other SSL functions.
14069 * Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
14070 structure. These are inherited by X509_STORE_CTX when it is
14071 initialised. This allows various defaults to be set in the
14072 X509_STORE structure (such as flags for CRL checking and custom
14073 purpose or trust settings) for functions which only use X509_STORE_CTX
14074 internally such as S/MIME.
14076 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
14077 trust settings if they are not set in X509_STORE. This allows X509_STORE
14078 purposes and trust (in S/MIME for example) to override any set by default.
14080 Add command line options for CRL checking to smime, s_client and s_server
14085 * Initial CRL based revocation checking. If the CRL checking flag(s)
14086 are set then the CRL is looked up in the X509_STORE structure and
14087 its validity and signature checked, then if the certificate is found
14088 in the CRL the verify fails with a revoked error.
14090 Various new CRL related callbacks added to X509_STORE_CTX structure.
14092 Command line options added to 'verify' application to support this.
14094 This needs some additional work, such as being able to handle multiple
14095 CRLs with different times, extension based lookup (rather than just
14096 by subject name) and ultimately more complete V2 CRL extension
14101 * Add a general user interface API (crypto/ui/). This is designed
14102 to replace things like des_read_password and friends (backward
14103 compatibility functions using this new API are provided).
14104 The purpose is to remove prompting functions from the DES code
14105 section as well as provide for prompting through dialog boxes in
14106 a window system and the like.
14110 * Add "ex_data" support to ENGINE so implementations can add state at a
14111 per-structure level rather than having to store it globally.
14115 * Make it possible for ENGINE structures to be copied when retrieved by
14116 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
14117 This causes the "original" ENGINE structure to act like a template,
14118 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
14119 operational state can be localised to each ENGINE structure, despite the
14120 fact they all share the same "methods". New ENGINE structures returned in
14121 this case have no functional references and the return value is the single
14122 structural reference. This matches the single structural reference returned
14123 by ENGINE_by_id() normally, when it is incremented on the pre-existing
14128 * Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
14129 needs to match any other type at all we need to manually clear the
14134 * Changes to the "openssl engine" utility to include;
14135 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
14136 about an ENGINE's available control commands.
14137 - executing control commands from command line arguments using the
14138 '-pre' and '-post' switches. '-post' is only used if '-t' is
14139 specified and the ENGINE is successfully initialised. The syntax for
14140 the individual commands are colon-separated, for example;
14141 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
14145 * New dynamic control command support for ENGINEs. ENGINEs can now
14146 declare their own commands (numbers), names (strings), descriptions,
14147 and input types for run-time discovery by calling applications. A
14148 subset of these commands are implicitly classed as "executable"
14149 depending on their input type, and only these can be invoked through
14150 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
14151 can be based on user input, config files, etc). The distinction is
14152 that "executable" commands cannot return anything other than a boolean
14153 result and can only support numeric or string input, whereas some
14154 discoverable commands may only be for direct use through
14155 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
14156 pointers, or other custom uses. The "executable" commands are to
14157 support parameterisations of ENGINE behaviour that can be
14158 unambiguously defined by ENGINEs and used consistently across any
14159 OpenSSL-based application. Commands have been added to all the
14160 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
14161 control over shared-library paths without source code alterations.
14165 * Changed all ENGINE implementations to dynamically allocate their
14166 ENGINEs rather than declaring them statically. Apart from this being
14167 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
14168 this also allows the implementations to compile without using the
14169 internal engine_int.h header.
14173 * Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
14174 'const' value. Any code that should be able to modify a RAND_METHOD
14175 should already have non-const pointers to it (ie. they should only
14176 modify their own ones).
14180 * Made a variety of little tweaks to the ENGINE code.
14181 - "atalla" and "ubsec" string definitions were moved from header files
14182 to C code. "nuron" string definitions were placed in variables
14183 rather than hard-coded - allowing parameterisation of these values
14184 later on via ctrl() commands.
14185 - Removed unused "#if 0"'d code.
14186 - Fixed engine list iteration code so it uses ENGINE_free() to release
14187 structural references.
14188 - Constified the RAND_METHOD element of ENGINE structures.
14189 - Constified various get/set functions as appropriate and added
14190 missing functions (including a catch-all ENGINE_cpy that duplicates
14191 all ENGINE values onto a new ENGINE except reference counts/state).
14192 - Removed NULL parameter checks in get/set functions. Setting a method
14193 or function to NULL is a way of cancelling out a previously set
14194 value. Passing a NULL ENGINE parameter is just plain stupid anyway
14195 and doesn't justify the extra error symbols and code.
14196 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
14197 flags from engine_int.h to engine.h.
14198 - Changed prototypes for ENGINE handler functions (init(), finish(),
14199 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
14203 * Implement binary inversion algorithm for BN_mod_inverse in addition
14204 to the algorithm using long division. The binary algorithm can be
14205 used only if the modulus is odd. On 32-bit systems, it is faster
14206 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
14207 roughly 5-15% for 256-bit moduli), so we use it only for moduli
14208 up to 450 bits. In 64-bit environments, the binary algorithm
14209 appears to be advantageous for much longer moduli; here we use it
14210 for moduli up to 2048 bits.
14214 * Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
14215 could not support the combine flag in choice fields.
14219 * Add a 'copy_extensions' option to the 'ca' utility. This copies
14220 extensions from a certificate request to the certificate.
14224 * Allow multiple 'certopt' and 'nameopt' options to be separated
14225 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
14226 file: this allows the display of the certificate about to be
14227 signed to be customised, to allow certain fields to be included
14228 or excluded and extension details. The old system didn't display
14229 multicharacter strings properly, omitted fields not in the policy
14230 and couldn't display additional details such as extensions.
14234 * Function EC_POINTs_mul for multiple scalar multiplication
14235 of an arbitrary number of elliptic curve points
14236 \sum scalars[i]*points[i],
14237 optionally including the generator defined for the EC_GROUP:
14238 scalar*generator + \sum scalars[i]*points[i].
14240 EC_POINT_mul is a simple wrapper function for the typical case
14241 that the point list has just one item (besides the optional
14246 * First EC_METHODs for curves over GF(p):
14248 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
14249 operations and provides various method functions that can also
14250 operate with faster implementations of modular arithmetic.
14252 EC_GFp_mont_method() reuses most functions that are part of
14253 EC_GFp_simple_method, but uses Montgomery arithmetic.
14255 *Bodo Moeller; point addition and point doubling
14256 implementation directly derived from source code provided by
14257 Lenka Fibikova <fibikova@exp-math.uni-essen.de>*
14259 * Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
14260 crypto/ec/ec_lib.c):
14262 Curves are EC_GROUP objects (with an optional group generator)
14263 based on EC_METHODs that are built into the library.
14265 Points are EC_POINT objects based on EC_GROUP objects.
14267 Most of the framework would be able to handle curves over arbitrary
14268 finite fields, but as there are no obvious types for fields other
14269 than GF(p), some functions are limited to that for now.
14273 * Add the -HTTP option to s_server. It is similar to -WWW, but requires
14274 that the file contains a complete HTTP response.
14278 * Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
14279 change the def and num file printf format specifier from "%-40sXXX"
14280 to "%-39s XXX". The latter will always guarantee a space after the
14281 field while the former will cause them to run together if the field
14282 is 40 of more characters long.
14286 * Constify the cipher and digest 'method' functions and structures
14287 and modify related functions to take constant EVP_MD and EVP_CIPHER
14292 * Hide BN_CTX structure details in bn_lcl.h instead of publishing them
14293 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
14297 * Modify `EVP_Digest*()` routines so they now return values. Although the
14298 internal software routines can never fail additional hardware versions
14303 * Clean up crypto/err/err.h and change some error codes to avoid conflicts:
14305 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
14306 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
14309 ERR_R_NESTED_ASN1_ERROR
14311 ERR_R_MISSING_ASN1_EOS
14312 were 4 .. 9, conflicting with
14313 ERR_LIB_RSA (= ERR_R_RSA_LIB)
14315 ERR_LIB_PEM (= ERR_R_PEM_LIB).
14316 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
14318 Add new error code 'ERR_R_INTERNAL_ERROR'.
14322 * Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
14327 * New option '-subj arg' for 'openssl req' and 'openssl ca'. This
14328 sets the subject name for a new request or supersedes the
14329 subject name in a given request. Formats that can be parsed are
14330 'CN=Some Name, OU=myOU, C=IT'
14332 'CN=Some Name/OU=myOU/C=IT'.
14334 Add options '-batch' and '-verbose' to 'openssl req'.
14336 *Massimiliano Pala <madwolf@hackmasters.net>*
14338 * Introduce the possibility to access global variables through
14339 functions on platform were that's the best way to handle exporting
14340 global variables in shared libraries. To enable this functionality,
14341 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
14342 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
14343 is normally done by Configure or something similar).
14345 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
14346 in the source file (foo.c) like this:
14348 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
14349 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
14351 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
14352 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
14354 OPENSSL_DECLARE_GLOBAL(int,foo);
14355 #define foo OPENSSL_GLOBAL_REF(foo)
14356 OPENSSL_DECLARE_GLOBAL(double,bar);
14357 #define bar OPENSSL_GLOBAL_REF(bar)
14359 The #defines are very important, and therefore so is including the
14360 header file everywhere where the defined globals are used.
14362 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
14363 of ASN.1 items, but that structure is a bit different.
14365 The largest change is in util/mkdef.pl which has been enhanced with
14366 better and easier to understand logic to choose which symbols should
14367 go into the Windows .def files as well as a number of fixes and code
14368 cleanup (among others, algorithm keywords are now sorted
14369 lexicographically to avoid constant rewrites).
14373 * In BN_div() keep a copy of the sign of 'num' before writing the
14374 result to 'rm' because if rm==num the value will be overwritten
14375 and produce the wrong result if 'num' is negative: this caused
14376 problems with BN_mod() and BN_nnmod().
14380 * Function OCSP_request_verify(). This checks the signature on an
14381 OCSP request and verifies the signer certificate. The signer
14382 certificate is just checked for a generic purpose and OCSP request
14387 * Add OCSP_check_validity() function to check the validity of OCSP
14388 responses. OCSP responses are prepared in real time and may only
14389 be a few seconds old. Simply checking that the current time lies
14390 between thisUpdate and nextUpdate max reject otherwise valid responses
14391 caused by either OCSP responder or client clock inaccuracy. Instead
14392 we allow thisUpdate and nextUpdate to fall within a certain period of
14393 the current time. The age of the response can also optionally be
14394 checked. Two new options -validity_period and -status_age added to
14399 * If signature or public key algorithm is unrecognized print out its
14400 OID rather that just UNKNOWN.
14404 * Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
14405 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
14406 ID to be generated from the issuer certificate alone which can then be
14407 passed to OCSP_id_issuer_cmp().
14411 * New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
14412 ASN1 modules to export functions returning ASN1_ITEM pointers
14413 instead of the ASN1_ITEM structures themselves. This adds several
14414 new macros which allow the underlying ASN1 function/structure to
14415 be accessed transparently. As a result code should not use ASN1_ITEM
14416 references directly (such as &X509_it) but instead use the relevant
14417 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
14418 use of the new ASN1 code on platforms where exporting structures
14419 is problematical (for example in shared libraries) but exporting
14420 functions returning pointers to structures is not.
14424 * Add support for overriding the generation of SSL/TLS session IDs.
14425 These callbacks can be registered either in an SSL_CTX or per SSL.
14426 The purpose of this is to allow applications to control, if they wish,
14427 the arbitrary values chosen for use as session IDs, particularly as it
14428 can be useful for session caching in multiple-server environments. A
14429 command-line switch for testing this (and any client code that wishes
14430 to use such a feature) has been added to "s_server".
14432 *Geoff Thorpe, Lutz Jaenicke*
14434 * Modify mkdef.pl to recognise and parse preprocessor conditionals
14435 of the form `#if defined(...) || defined(...) || ...` and
14436 `#if !defined(...) && !defined(...) && ...`. This also avoids
14437 the growing number of special cases it was previously handling.
14441 * Make all configuration macros available for application by making
14442 sure they are available in opensslconf.h, by giving them names starting
14443 with `OPENSSL_` to avoid conflicts with other packages and by making
14444 sure e_os2.h will cover all platform-specific cases together with
14446 Additionally, it is now possible to define configuration/platform-
14447 specific names (called "system identities"). In the C code, these
14448 are prefixed with `OPENSSL_SYSNAME_`. e_os2.h will create another
14449 macro with the name beginning with `OPENSSL_SYS_`, which is determined
14450 from `OPENSSL_SYSNAME_*` or compiler-specific macros depending on
14455 * New option -set_serial to 'req' and 'x509' this allows the serial
14456 number to use to be specified on the command line. Previously self
14457 signed certificates were hard coded with serial number 0 and the
14458 CA options of 'x509' had to use a serial number in a file which was
14463 * New options to 'ca' utility to support V2 CRL entry extensions.
14464 Currently CRL reason, invalidity date and hold instruction are
14465 supported. Add new CRL extensions to V3 code and some new objects.
14469 * New function EVP_CIPHER_CTX_set_padding() this is used to
14470 disable standard block padding (aka PKCS#5 padding) in the EVP
14471 API, which was previously mandatory. This means that the data is
14472 not padded in any way and so the total length much be a multiple
14473 of the block size, otherwise an error occurs.
14477 * Initial (incomplete) OCSP SSL support.
14481 * New function OCSP_parse_url(). This splits up a URL into its host,
14482 port and path components: primarily to parse OCSP URLs. New -url
14483 option to ocsp utility.
14487 * New nonce behavior. The return value of OCSP_check_nonce() now
14488 reflects the various checks performed. Applications can decide
14489 whether to tolerate certain situations such as an absent nonce
14490 in a response when one was present in a request: the ocsp application
14491 just prints out a warning. New function OCSP_add1_basic_nonce()
14492 this is to allow responders to include a nonce in a response even if
14493 the request is nonce-less.
14497 * Disable stdin buffering in `load_cert()` (`apps/apps.c`) so that no certs are
14498 skipped when using openssl x509 multiple times on a single input file,
14499 e.g. `(openssl x509 -out cert1; openssl x509 -out cert2) <certs`.
14503 * Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
14504 set string type: to handle setting ASN1_TIME structures. Fix ca
14505 utility to correctly initialize revocation date of CRLs.
14509 * New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
14510 the clients preferred ciphersuites and rather use its own preferences.
14511 Should help to work around M$ SGC (Server Gated Cryptography) bug in
14512 Internet Explorer by ensuring unchanged hash method during stepup.
14513 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
14517 * Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
14518 to aes and add a new 'exist' option to print out symbols that don't
14523 * Additional options to ocsp utility to allow flags to be set and
14524 additional certificates supplied.
14528 * Add the option -VAfile to 'openssl ocsp', so the user can give the
14529 OCSP client a number of certificate to only verify the response
14534 * Update Rijndael code to version 3.0 and change EVP AES ciphers to
14535 handle the new API. Currently only ECB, CBC modes supported. Add new
14538 Add TLS AES ciphersuites as described in RFC3268, "Advanced
14539 Encryption Standard (AES) Ciphersuites for Transport Layer
14540 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were
14541 not enabled by default and were not part of the "ALL" ciphersuite
14542 alias because they were not yet official; they could be
14543 explicitly requested by specifying the "AESdraft" ciphersuite
14544 group alias. In the final release of OpenSSL 0.9.7, the group
14545 alias is called "AES" and is part of "ALL".)
14547 *Ben Laurie, Steve Henson, Bodo Moeller*
14549 * New function OCSP_copy_nonce() to copy nonce value (if present) from
14550 request to response.
14554 * Functions for OCSP responders. OCSP_request_onereq_count(),
14555 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
14556 extract information from a certificate request. OCSP_response_create()
14557 creates a response and optionally adds a basic response structure.
14558 OCSP_basic_add1_status() adds a complete single response to a basic
14559 response and returns the OCSP_SINGLERESP structure just added (to allow
14560 extensions to be included for example). OCSP_basic_add1_cert() adds a
14561 certificate to a basic response and OCSP_basic_sign() signs a basic
14562 response with various flags. New helper functions ASN1_TIME_check()
14563 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
14564 (converts ASN1_TIME to GeneralizedTime).
14568 * Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
14569 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
14570 structure from a certificate. X509_pubkey_digest() digests the public_key
14571 contents: this is used in various key identifiers.
14575 * Make sk_sort() tolerate a NULL argument.
14577 *Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>*
14579 * New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
14580 passed by the function are trusted implicitly. If any of them signed the
14581 response then it is assumed to be valid and is not verified.
14585 * In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
14586 to data. This was previously part of the PKCS7 ASN1 code. This
14587 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
14588 *Steve Henson, reported by Kenneth R. Robinette
14589 <support@securenetterm.com>*
14591 * Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
14592 routines: without these tracing memory leaks is very painful.
14593 Fix leaks in PKCS12 and PKCS7 routines.
14597 * Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
14598 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
14599 effectively meant GeneralizedTime would never be used. Now it
14600 is initialised to -1 but X509_time_adj() now has to check the value
14601 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
14602 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
14603 *Steve Henson, reported by Kenneth R. Robinette
14604 <support@securenetterm.com>*
14606 * Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
14607 result in a zero length in the ASN1_INTEGER structure which was
14608 not consistent with the structure when d2i_ASN1_INTEGER() was used
14609 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
14610 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
14611 where it did not print out a minus for negative ASN1_INTEGER.
14615 * Add summary printout to ocsp utility. The various functions which
14616 convert status values to strings have been renamed to:
14617 OCSP_response_status_str(), OCSP_cert_status_str() and
14618 OCSP_crl_reason_str() and are no longer static. New options
14619 to verify nonce values and to disable verification. OCSP response
14620 printout format cleaned up.
14624 * Add additional OCSP certificate checks. These are those specified
14625 in RFC2560. This consists of two separate checks: the CA of the
14626 certificate being checked must either be the OCSP signer certificate
14627 or the issuer of the OCSP signer certificate. In the latter case the
14628 OCSP signer certificate must contain the OCSP signing extended key
14629 usage. This check is performed by attempting to match the OCSP
14630 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
14631 in the OCSP_CERTID structures of the response.
14635 * Initial OCSP certificate verification added to OCSP_basic_verify()
14636 and related routines. This uses the standard OpenSSL certificate
14637 verify routines to perform initial checks (just CA validity) and
14638 to obtain the certificate chain. Then additional checks will be
14639 performed on the chain. Currently the root CA is checked to see
14640 if it is explicitly trusted for OCSP signing. This is used to set
14641 a root CA as a global signing root: that is any certificate that
14642 chains to that CA is an acceptable OCSP signing certificate.
14646 * New '-extfile ...' option to 'openssl ca' for reading X.509v3
14647 extensions from a separate configuration file.
14648 As when reading extensions from the main configuration file,
14649 the '-extensions ...' option may be used for specifying the
14652 *Massimiliano Pala <madwolf@comune.modena.it>*
14654 * New OCSP utility. Allows OCSP requests to be generated or
14655 read. The request can be sent to a responder and the output
14656 parsed, outputted or printed in text form. Not complete yet:
14657 still needs to check the OCSP response validity.
14661 * New subcommands for 'openssl ca':
14662 `openssl ca -status <serial>` prints the status of the cert with
14663 the given serial number (according to the index file).
14664 `openssl ca -updatedb` updates the expiry status of certificates
14667 *Massimiliano Pala <madwolf@comune.modena.it>*
14669 * New '-newreq-nodes' command option to CA.pl. This is like
14670 '-newreq', but calls 'openssl req' with the '-nodes' option
14671 so that the resulting key is not encrypted.
14673 *Damien Miller <djm@mindrot.org>*
14675 * New configuration for the GNU Hurd.
14677 *Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte*
14679 * Initial code to implement OCSP basic response verify. This
14680 is currently incomplete. Currently just finds the signer's
14681 certificate and verifies the signature on the response.
14685 * New SSLeay_version code SSLEAY_DIR to determine the compiled-in
14686 value of OPENSSLDIR. This is available via the new '-d' option
14687 to 'openssl version', and is also included in 'openssl version -a'.
14691 * Allowing defining memory allocation callbacks that will be given
14692 file name and line number information in additional arguments
14693 (a `const char*` and an int). The basic functionality remains, as
14694 well as the original possibility to just replace malloc(),
14695 realloc() and free() by functions that do not know about these
14696 additional arguments. To register and find out the current
14697 settings for extended allocation functions, the following
14698 functions are provided:
14700 CRYPTO_set_mem_ex_functions
14701 CRYPTO_set_locked_mem_ex_functions
14702 CRYPTO_get_mem_ex_functions
14703 CRYPTO_get_locked_mem_ex_functions
14705 These work the same way as CRYPTO_set_mem_functions and friends.
14706 `CRYPTO_get_[locked_]mem_functions` now writes 0 where such an
14707 extended allocation function is enabled.
14708 Similarly, `CRYPTO_get_[locked_]mem_ex_functions` writes 0 where
14709 a conventional allocation function is enabled.
14711 *Richard Levitte, Bodo Moeller*
14713 * Finish off removing the remaining LHASH function pointer casts.
14714 There should no longer be any prototype-casting required when using
14715 the LHASH abstraction, and any casts that remain are "bugs". See
14716 the callback types and macros at the head of lhash.h for details
14717 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
14721 * Add automatic query of EGD sockets in RAND_poll() for the unix variant.
14722 If /dev/[u]random devices are not available or do not return enough
14723 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
14725 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
14726 /etc/entropy will be queried once each in this sequence, querying stops
14727 when enough entropy was collected without querying more sockets.
14731 * Change the Unix RAND_poll() variant to be able to poll several
14732 random devices, as specified by DEVRANDOM, until a sufficient amount
14733 of data has been collected. We spend at most 10 ms on each file
14734 (select timeout) and read in non-blocking mode. DEVRANDOM now
14735 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
14736 (previously it was just the string "/dev/urandom"), so on typical
14737 platforms the 10 ms delay will never occur.
14738 Also separate out the Unix variant to its own file, rand_unix.c.
14739 For VMS, there's a currently-empty rand_vms.c.
14743 * Move OCSP client related routines to ocsp_cl.c. These
14744 provide utility functions which an application needing
14745 to issue a request to an OCSP responder and analyse the
14746 response will typically need: as opposed to those which an
14747 OCSP responder itself would need which will be added later.
14749 OCSP_request_sign() signs an OCSP request with an API similar
14750 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
14751 response. OCSP_response_get1_basic() extracts basic response
14752 from response. OCSP_resp_find_status(): finds and extracts status
14753 information from an OCSP_CERTID structure (which will be created
14754 when the request structure is built). These are built from lower
14755 level functions which work on OCSP_SINGLERESP structures but
14756 won't normally be used unless the application wishes to examine
14757 extensions in the OCSP response for example.
14759 Replace nonce routines with a pair of functions.
14760 OCSP_request_add1_nonce() adds a nonce value and optionally
14761 generates a random value. OCSP_check_nonce() checks the
14762 validity of the nonce in an OCSP response.
14766 * Change function OCSP_request_add() to OCSP_request_add0_id().
14767 This doesn't copy the supplied OCSP_CERTID and avoids the
14768 need to free up the newly created id. Change return type
14769 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
14770 This can then be used to add extensions to the request.
14771 Deleted OCSP_request_new(), since most of its functionality
14772 is now in OCSP_REQUEST_new() (and the case insensitive name
14773 clash) apart from the ability to set the request name which
14774 will be added elsewhere.
14778 * Update OCSP API. Remove obsolete extensions argument from
14779 various functions. Extensions are now handled using the new
14780 OCSP extension code. New simple OCSP HTTP function which
14781 can be used to send requests and parse the response.
14785 * Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
14786 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
14787 uses the special reorder version of SET OF to sort the attributes
14788 and reorder them to match the encoded order. This resolves a long
14789 standing problem: a verify on a PKCS7 structure just after signing
14790 it used to fail because the attribute order did not match the
14791 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
14792 it uses the received order. This is necessary to tolerate some broken
14793 software that does not order SET OF. This is handled by encoding
14794 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
14795 to produce the required SET OF.
14799 * Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
14800 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
14801 files to get correct declarations of the ASN.1 item variables.
14805 * Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
14806 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
14807 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
14808 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
14809 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
14810 ASN1_ITEM and no wrapper functions.
14814 * New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
14815 replace the old function pointer based I/O routines. Change most of
14816 the `*_d2i_bio()` and `*_d2i_fp()` functions to use these.
14820 * Enhance mkdef.pl to be more accepting about spacing in C preprocessor
14821 lines, recognize more "algorithms" that can be deselected, and make
14822 it complain about algorithm deselection that isn't recognised.
14826 * New ASN1 functions to handle dup, sign, verify, digest, pack and
14827 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
14828 to use new functions. Add NO_ASN1_OLD which can be set to remove
14829 some old style ASN1 functions: this can be used to determine if old
14830 code will still work when these eventually go away.
14834 * New extension functions for OCSP structures, these follow the
14835 same conventions as certificates and CRLs.
14839 * New function X509V3_add1_i2d(). This automatically encodes and
14840 adds an extension. Its behaviour can be customised with various
14841 flags to append, replace or delete. Various wrappers added for
14842 certificates and CRLs.
14846 * Fix to avoid calling the underlying ASN1 print routine when
14847 an extension cannot be parsed. Correct a typo in the
14848 OCSP_SERVICELOC extension. Tidy up print OCSP format.
14852 * Make mkdef.pl parse some of the ASN1 macros and add appropriate
14853 entries for variables.
14857 * Add functionality to `apps/openssl.c` for detecting locking
14858 problems: As the program is single-threaded, all we have
14859 to do is register a locking callback using an array for
14860 storing which locks are currently held by the program.
14864 * Use a lock around the call to CRYPTO_get_ex_new_index() in
14865 SSL_get_ex_data_X509_STORE_idx(), which is used in
14866 ssl_verify_cert_chain() and thus can be called at any time
14867 during TLS/SSL handshakes so that thread-safety is essential.
14868 Unfortunately, the ex_data design is not at all suited
14869 for multi-threaded use, so it probably should be abolished.
14873 * Added Broadcom "ubsec" ENGINE to OpenSSL.
14875 *Broadcom, tweaked and integrated by Geoff Thorpe*
14877 * Move common extension printing code to new function
14878 X509V3_print_extensions(). Reorganise OCSP print routines and
14879 implement some needed OCSP ASN1 functions. Add OCSP extensions.
14883 * New function X509_signature_print() to remove duplication in some
14888 * Add a special meaning when SET OF and SEQUENCE OF flags are both
14889 set (this was treated exactly the same as SET OF previously). This
14890 is used to reorder the STACK representing the structure to match the
14891 encoding. This will be used to get round a problem where a PKCS7
14892 structure which was signed could not be verified because the STACK
14893 order did not reflect the encoded order.
14897 * Reimplement the OCSP ASN1 module using the new code.
14901 * Update the X509V3 code to permit the use of an ASN1_ITEM structure
14902 for its ASN1 operations. The old style function pointers still exist
14903 for now but they will eventually go away.
14907 * Merge in replacement ASN1 code from the ASN1 branch. This almost
14908 completely replaces the old ASN1 functionality with a table driven
14909 encoder and decoder which interprets an ASN1_ITEM structure describing
14910 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
14911 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
14912 has also been converted to the new form.
14916 * Change BN_mod_exp_recp so that negative moduli are tolerated
14917 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
14918 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
14919 for negative moduli.
14923 * Fix BN_uadd and BN_usub: Always return non-negative results instead
14924 of not touching the result's sign bit.
14928 * BN_div bugfix: If the result is 0, the sign (res->neg) must not be
14933 * Changed the LHASH code to use prototypes for callbacks, and created
14934 macros to declare and implement thin (optionally static) functions
14935 that provide type-safety and avoid function pointer casting for the
14936 type-specific callbacks.
14940 * Added Kerberos Cipher Suites to be used with TLS, as written in
14942 *Veers Staats <staatsvr@asc.hpc.mil>,
14943 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte*
14945 * Reformat the FAQ so the different questions and answers can be divided
14946 in sections depending on the subject.
14950 * Have the zlib compression code load ZLIB.DLL dynamically under
14955 * New function BN_mod_sqrt for computing square roots modulo a prime
14956 (using the probabilistic Tonelli-Shanks algorithm unless
14957 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
14958 be handled deterministically).
14960 *Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller*
14962 * Make BN_mod_inverse faster by explicitly handling small quotients
14963 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
14964 512 bits], about 30% for larger ones [1024 or 2048 bits].)
14968 * New function BN_kronecker.
14972 * Fix BN_gcd so that it works on negative inputs; the result is
14973 positive unless both parameters are zero.
14974 Previously something reasonably close to an infinite loop was
14975 possible because numbers could be growing instead of shrinking
14976 in the implementation of Euclid's algorithm.
14980 * Fix BN_is_word() and BN_is_one() macros to take into account the
14981 sign of the number in question.
14983 Fix BN_is_word(a,w) to work correctly for w == 0.
14985 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
14986 because its test if the absolute value of 'a' equals 'w'.
14987 Note that BN_abs_is_word does *not* handle w == 0 reliably;
14988 it exists mostly for use in the implementations of BN_is_zero(),
14989 BN_is_one(), and BN_is_word().
14993 * New function BN_swap.
14997 * Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
14998 the exponentiation functions are more likely to produce reasonable
14999 results on negative inputs.
15003 * Change BN_mod_mul so that the result is always non-negative.
15004 Previously, it could be negative if one of the factors was negative;
15005 I don't think anyone really wanted that behaviour.
15009 * Move `BN_mod_...` functions into new file `crypto/bn/bn_mod.c`
15010 (except for exponentiation, which stays in `crypto/bn/bn_exp.c`,
15011 and `BN_mod_mul_reciprocal`, which stays in `crypto/bn/bn_recp.c`)
15012 and add new functions:
15021 BN_mod_lshift1_quick
15023 BN_mod_lshift_quick
15025 These functions always generate non-negative results.
15027 `BN_nnmod` otherwise is `like BN_mod` (if `BN_mod` computes a remainder `r`
15028 such that `|m| < r < 0`, `BN_nnmod` will output `rem + |m|` instead).
15030 `BN_mod_XXX_quick(r, a, [b,] m)` generates the same result as
15031 `BN_mod_XXX(r, a, [b,] m, ctx)`, but requires that `a` [and `b`]
15032 be reduced modulo `m`.
15034 *Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller*
15037 The following entry accidentally appeared in the CHANGES file
15038 distributed with OpenSSL 0.9.7. The modifications described in
15039 it do *not* apply to OpenSSL 0.9.7.
15041 * Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
15042 was actually never needed) and in BN_mul(). The removal in BN_mul()
15043 required a small change in bn_mul_part_recursive() and the addition
15044 of the functions bn_cmp_part_words(), bn_sub_part_words() and
15045 bn_add_part_words(), which do the same thing as bn_cmp_words(),
15046 bn_sub_words() and bn_add_words() except they take arrays with
15052 * In 'openssl passwd', verify passwords read from the terminal
15053 unless the '-salt' option is used (which usually means that
15054 verification would just waste user's time since the resulting
15055 hash is going to be compared with some given password hash)
15056 or the new '-noverify' option is used.
15058 This is an incompatible change, but it does not affect
15059 non-interactive use of 'openssl passwd' (passwords on the command
15060 line, '-stdin' option, '-in ...' option) and thus should not
15061 cause any problems.
15065 * Remove all references to RSAref, since there's no more need for it.
15069 * Make DSO load along a path given through an environment variable
15070 (SHLIB_PATH) with shl_load().
15074 * Constify the ENGINE code as a result of BIGNUM constification.
15075 Also constify the RSA code and most things related to it. In a
15076 few places, most notable in the depth of the ASN.1 code, ugly
15077 casts back to non-const were required (to be solved at a later
15082 * Make it so the openssl application has all engines loaded by default.
15086 * Constify the BIGNUM routines a little more.
15090 * Add the following functions:
15092 ENGINE_load_cswift()
15094 ENGINE_load_atalla()
15095 ENGINE_load_nuron()
15096 ENGINE_load_builtin_engines()
15098 That way, an application can itself choose if external engines that
15099 are built-in in OpenSSL shall ever be used or not. The benefit is
15100 that applications won't have to be linked with libdl or other dso
15101 libraries unless it's really needed.
15103 Changed 'openssl engine' to load all engines on demand.
15104 Changed the engine header files to avoid the duplication of some
15105 declarations (they differed!).
15109 * 'openssl engine' can now list capabilities.
15113 * Better error reporting in 'openssl engine'.
15117 * Never call load_dh_param(NULL) in s_server.
15121 * Add engine application. It can currently list engines by name and
15122 identity, and test if they are actually available.
15126 * Improve RPM specification file by forcing symbolic linking and making
15127 sure the installed documentation is also owned by root.root.
15129 *Damien Miller <djm@mindrot.org>*
15131 * Give the OpenSSL applications more possibilities to make use of
15132 keys (public as well as private) handled by engines.
15136 * Add OCSP code that comes from CertCo.
15140 * Add VMS support for the Rijndael code.
15144 * Added untested support for Nuron crypto accelerator.
15148 * Add support for external cryptographic devices. This code was
15149 previously distributed separately as the "engine" branch.
15151 *Geoff Thorpe, Richard Levitte*
15153 * Rework the filename-translation in the DSO code. It is now possible to
15154 have far greater control over how a "name" is turned into a filename
15155 depending on the operating environment and any oddities about the
15156 different shared library filenames on each system.
15160 * Support threads on FreeBSD-elf in Configure.
15164 * Fix for SHA1 assembly problem with MASM: it produces
15165 warnings about corrupt line number information when assembling
15166 with debugging information. This is caused by the overlapping
15169 *Bernd Matthes <mainbug@celocom.de>, Steve Henson*
15172 NCONF_get_number() has no error checking at all. As a replacement,
15173 NCONF_get_number_e() is defined (`_e` for "error checking") and is
15174 promoted strongly. The old NCONF_get_number is kept around for
15175 binary backward compatibility.
15176 Make it possible for methods to load from something other than a BIO,
15177 by providing a function pointer that is given a name instead of a BIO.
15178 For example, this could be used to load configuration data from an
15183 * Fix for non blocking accept BIOs. Added new I/O special reason
15184 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
15185 with non blocking I/O was not possible because no retry code was
15186 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
15191 * Added the beginnings of Rijndael support.
15195 * Fix for bug in DirectoryString mask setting. Add support for
15196 X509_NAME_print_ex() in 'req' and X509_print_ex() function
15197 to allow certificate printing to more controllable, additional
15198 'certopt' option to 'x509' to allow new printing options to be
15203 * Clean old EAY MD5 hack from e_os.h.
15207 ### Changes between 0.9.6l and 0.9.6m [17 Mar 2004]
15209 * Fix null-pointer assignment in do_change_cipher_spec() revealed
15210 by using the Codenomicon TLS Test Tool ([CVE-2004-0079])
15212 *Joe Orton, Steve Henson*
15214 ### Changes between 0.9.6k and 0.9.6l [04 Nov 2003]
15216 * Fix additional bug revealed by the NISCC test suite:
15218 Stop bug triggering large recursion when presented with
15219 certain ASN.1 tags ([CVE-2003-0851])
15223 ### Changes between 0.9.6j and 0.9.6k [30 Sep 2003]
15225 * Fix various bugs revealed by running the NISCC test suite:
15227 Stop out of bounds reads in the ASN1 code when presented with
15228 invalid tags (CVE-2003-0543 and CVE-2003-0544).
15230 If verify callback ignores invalid public key errors don't try to check
15231 certificate signature with the NULL public key.
15235 * In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
15236 if the server requested one: as stated in TLS 1.0 and SSL 3.0
15241 * In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
15242 extra data after the compression methods not only for TLS 1.0
15243 but also for SSL 3.0 (as required by the specification).
15245 *Bodo Moeller; problem pointed out by Matthias Loepfe*
15247 * Change X509_certificate_type() to mark the key as exported/exportable
15248 when it's 512 *bits* long, not 512 bytes.
15252 ### Changes between 0.9.6i and 0.9.6j [10 Apr 2003]
15254 * Countermeasure against the Klima-Pokorny-Rosa extension of
15255 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
15256 a protocol version number mismatch like a decryption error
15257 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
15261 * Turn on RSA blinding by default in the default implementation
15262 to avoid a timing attack. Applications that don't want it can call
15263 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
15264 They would be ill-advised to do so in most cases.
15266 *Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller*
15268 * Change RSA blinding code so that it works when the PRNG is not
15269 seeded (in this case, the secret RSA exponent is abused as
15270 an unpredictable seed -- if it is not unpredictable, there
15271 is no point in blinding anyway). Make RSA blinding thread-safe
15272 by remembering the creator's thread ID in rsa->blinding and
15273 having all other threads use local one-time blinding factors
15274 (this requires more computation than sharing rsa->blinding, but
15275 avoids excessive locking; and if an RSA object is not shared
15276 between threads, blinding will still be very fast).
15280 ### Changes between 0.9.6h and 0.9.6i [19 Feb 2003]
15282 * In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
15283 via timing by performing a MAC computation even if incorrect
15284 block cipher padding has been found. This is a countermeasure
15285 against active attacks where the attacker has to distinguish
15286 between bad padding and a MAC verification error. ([CVE-2003-0078])
15288 *Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
15289 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
15290 Martin Vuagnoux (EPFL, Ilion)*
15292 ### Changes between 0.9.6g and 0.9.6h [5 Dec 2002]
15294 * New function OPENSSL_cleanse(), which is used to cleanse a section of
15295 memory from its contents. This is done with a counter that will
15296 place alternating values in each byte. This can be used to solve
15297 two issues: 1) the removal of calls to memset() by highly optimizing
15298 compilers, and 2) cleansing with other values than 0, since those can
15299 be read through on certain media, for example a swap space on disk.
15303 * Bugfix: client side session caching did not work with external caching,
15304 because the session->cipher setting was not restored when reloading
15305 from the external cache. This problem was masked, when
15306 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set.
15307 (Found by Steve Haslam <steve@araqnid.ddts.net>.)
15311 * Fix client_certificate (ssl/s2_clnt.c): The permissible total
15312 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33.
15314 *Zeev Lieber <zeev-l@yahoo.com>*
15316 * Undo an undocumented change introduced in 0.9.6e which caused
15317 repeated calls to OpenSSL_add_all_ciphers() and
15318 OpenSSL_add_all_digests() to be ignored, even after calling
15323 * Change the default configuration reader to deal with last line not
15324 being properly terminated.
15328 * Change X509_NAME_cmp() so it applies the special rules on handling
15329 DN values that are of type PrintableString, as well as RDNs of type
15330 emailAddress where the value has the type ia5String.
15332 *stefank@valicert.com via Richard Levitte*
15334 * Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half
15335 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently
15336 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be
15337 the bitwise-OR of the two for use by the majority of applications
15338 wanting this behaviour, and update the docs. The documented
15339 behaviour and actual behaviour were inconsistent and had been
15340 changing anyway, so this is more a bug-fix than a behavioural
15343 *Geoff Thorpe, diagnosed by Nadav Har'El*
15345 * Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c
15346 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes).
15350 * Fix initialization code race conditions in
15351 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(),
15352 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(),
15353 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(),
15354 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(),
15355 ssl2_get_cipher_by_char(),
15356 ssl3_get_cipher_by_char().
15358 *Patrick McCormick <patrick@tellme.com>, Bodo Moeller*
15360 * Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after
15361 the cached sessions are flushed, as the remove_cb() might use ex_data
15362 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com>
15363 (see [openssl.org #212]).
15365 *Geoff Thorpe, Lutz Jaenicke*
15367 * Fix typo in OBJ_txt2obj which incorrectly passed the content
15368 length, instead of the encoding length to d2i_ASN1_OBJECT.
15372 ### Changes between 0.9.6f and 0.9.6g [9 Aug 2002]
15374 * [In 0.9.6g-engine release:]
15375 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use `_stdcall`).
15377 *Lynn Gazis <lgazis@rainbow.com>*
15379 ### Changes between 0.9.6e and 0.9.6f [8 Aug 2002]
15381 * Fix ASN1 checks. Check for overflow by comparing with LONG_MAX
15382 and get fix the header length calculation.
15383 *Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>,
15384 Alon Kantor <alonk@checkpoint.com> (and others), Steve Henson*
15386 * Use proper error handling instead of 'assertions' in buffer
15387 overflow checks added in 0.9.6e. This prevents DoS (the
15388 assertions could call abort()).
15390 *Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller*
15392 ### Changes between 0.9.6d and 0.9.6e [30 Jul 2002]
15394 * Add various sanity checks to asn1_get_length() to reject
15395 the ASN1 length bytes if they exceed sizeof(long), will appear
15396 negative or the content length exceeds the length of the
15399 *Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>*
15401 * Fix cipher selection routines: ciphers without encryption had no flags
15402 for the cipher strength set and where therefore not handled correctly
15403 by the selection routines (PR #130).
15407 * Fix EVP_dsa_sha macro.
15412 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS
15413 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure
15414 that was added in OpenSSL 0.9.6d.
15416 As the countermeasure turned out to be incompatible with some
15417 broken SSL implementations, the new option is part of SSL_OP_ALL.
15418 SSL_OP_ALL is usually employed when compatibility with weird SSL
15419 implementations is desired (e.g. '-bugs' option to 's_client' and
15420 's_server'), so the new option is automatically set in many
15425 * Changes in security patch:
15427 Changes marked "(CHATS)" were sponsored by the Defense Advanced
15428 Research Projects Agency (DARPA) and Air Force Research Laboratory,
15429 Air Force Materiel Command, USAF, under agreement number
15432 * Add various sanity checks to asn1_get_length() to reject
15433 the ASN1 length bytes if they exceed sizeof(long), will appear
15434 negative or the content length exceeds the length of the
15435 supplied buffer. ([CVE-2002-0659])
15437 *Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>*
15439 * Assertions for various potential buffer overflows, not known to
15440 happen in practice.
15442 *Ben Laurie (CHATS)*
15444 * Various temporary buffers to hold ASCII versions of integers were
15445 too small for 64 bit platforms. ([CVE-2002-0655])
15446 *Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>*
15448 * Remote buffer overflow in SSL3 protocol - an attacker could
15449 supply an oversized session ID to a client. ([CVE-2002-0656])
15451 *Ben Laurie (CHATS)*
15453 * Remote buffer overflow in SSL2 protocol - an attacker could
15454 supply an oversized client master key. ([CVE-2002-0656])
15456 *Ben Laurie (CHATS)*
15458 ### Changes between 0.9.6c and 0.9.6d [9 May 2002]
15460 * Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
15461 encoded as NULL) with id-dsa-with-sha1.
15463 *Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller*
15465 * Check various `X509_...()` return values in `apps/req.c`.
15467 *Nils Larsch <nla@trustcenter.de>*
15469 * Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
15470 an end-of-file condition would erroneously be flagged, when the CRLF
15471 was just at the end of a processed block. The bug was discovered when
15472 processing data through a buffering memory BIO handing the data to a
15473 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
15474 <ptsekov@syntrex.com> and Nedelcho Stanev.
15478 * Implement a countermeasure against a vulnerability recently found
15479 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
15480 before application data chunks to avoid the use of known IVs
15481 with data potentially chosen by the attacker.
15485 * Fix length checks in ssl3_get_client_hello().
15489 * TLS/SSL library bugfix: use s->s3->in_read_app_data differently
15490 to prevent ssl3_read_internal() from incorrectly assuming that
15491 ssl3_read_bytes() found application data while handshake
15492 processing was enabled when in fact s->s3->in_read_app_data was
15493 merely automatically cleared during the initial handshake.
15495 *Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>*
15497 * Fix object definitions for Private and Enterprise: they were not
15498 recognized in their shortname (=lowercase) representation. Extend
15499 obj_dat.pl to issue an error when using undefined keywords instead
15500 of silently ignoring the problem (Svenning Sorensen
15501 <sss@sss.dnsalias.net>).
15505 * Fix DH_generate_parameters() so that it works for 'non-standard'
15506 generators, i.e. generators other than 2 and 5. (Previously, the
15507 code did not properly initialise the 'add' and 'rem' values to
15508 BN_generate_prime().)
15510 In the new general case, we do not insist that 'generator' is
15511 actually a primitive root: This requirement is rather pointless;
15512 a generator of the order-q subgroup is just as good, if not
15517 * Map new X509 verification errors to alerts. Discovered and submitted by
15518 Tom Wu <tom@arcot.com>.
15522 * Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
15523 returning non-zero before the data has been completely received
15524 when using non-blocking I/O.
15526 *Bodo Moeller; problem pointed out by John Hughes*
15528 * Some of the ciphers missed the strength entry (SSL_LOW etc).
15530 *Ben Laurie, Lutz Jaenicke*
15532 * Fix bug in SSL_clear(): bad sessions were not removed (found by
15533 Yoram Zahavi <YoramZ@gilian.com>).
15537 * Add information about CygWin 1.3 and on, and preserve proper
15538 configuration for the versions before that.
15540 *Corinna Vinschen <vinschen@redhat.com> and Richard Levitte*
15542 * Make removal from session cache (SSL_CTX_remove_session()) more robust:
15543 check whether we deal with a copy of a session and do not delete from
15544 the cache in this case. Problem reported by "Izhar Shoshani Levi"
15545 <izhar@checkpoint.com>.
15549 * Do not store session data into the internal session cache, if it
15550 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
15551 flag is set). Proposed by Aslam <aslam@funk.com>.
15555 * Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
15560 * [In 0.9.6d-engine release:]
15561 Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
15563 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
15565 * Add the configuration target linux-s390x.
15567 *Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte*
15569 * The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
15570 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
15571 variable as an indication that a ClientHello message has been
15572 received. As the flag value will be lost between multiple
15573 invocations of ssl3_accept when using non-blocking I/O, the
15574 function may not be aware that a handshake has actually taken
15575 place, thus preventing a new session from being added to the
15578 To avoid this problem, we now set s->new_session to 2 instead of
15579 using a local variable.
15581 *Lutz Jaenicke, Bodo Moeller*
15583 * Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
15584 if the SSL_R_LENGTH_MISMATCH error is detected.
15586 *Geoff Thorpe, Bodo Moeller*
15588 * New 'shared_ldflag' column in Configure platform table.
15592 * Fix EVP_CIPHER_mode macro.
15594 *"Dan S. Camper" <dan@bti.net>*
15596 * Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
15597 type, we must throw them away by setting rr->length to 0.
15599 *D P Chang <dpc@qualys.com>*
15601 ### Changes between 0.9.6b and 0.9.6c [21 dec 2001]
15603 * Fix BN_rand_range bug pointed out by Dominikus Scherkl
15604 <Dominikus.Scherkl@biodata.com>. (The previous implementation
15605 worked incorrectly for those cases where range = `10..._2` and
15606 `3*range` is two bits longer than range.)
15610 * Only add signing time to PKCS7 structures if it is not already
15615 * Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
15616 OBJ_ld_ce should be OBJ_id_ce.
15617 Also some ip-pda OIDs in crypto/objects/objects.txt were
15618 incorrect (cf. RFC 3039).
15620 *Matt Cooper, Frederic Giudicelli, Bodo Moeller*
15622 * Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
15623 returns early because it has nothing to do.
15625 *Andy Schneider <andy.schneider@bjss.co.uk>*
15627 * [In 0.9.6c-engine release:]
15628 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
15630 *Andy Schneider <andy.schneider@bjss.co.uk>*
15632 * [In 0.9.6c-engine release:]
15633 Add support for Cryptographic Appliance's keyserver technology.
15634 (Use engine 'keyclient')
15636 *Cryptographic Appliances and Geoff Thorpe*
15638 * Add a configuration entry for OS/390 Unix. The C compiler 'c89'
15639 is called via tools/c89.sh because arguments have to be
15640 rearranged (all '-L' options must appear before the first object
15643 *Richard Shapiro <rshapiro@abinitio.com>*
15645 * [In 0.9.6c-engine release:]
15646 Add support for Broadcom crypto accelerator cards, backported
15649 *Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox*
15651 * [In 0.9.6c-engine release:]
15652 Add support for SureWare crypto accelerator cards from
15653 Baltimore Technologies. (Use engine 'sureware')
15655 *Baltimore Technologies and Mark Cox*
15657 * [In 0.9.6c-engine release:]
15658 Add support for crypto accelerator cards from Accelerated
15659 Encryption Processing, www.aep.ie. (Use engine 'aep')
15661 *AEP Inc. and Mark Cox*
15663 * Add a configuration entry for gcc on UnixWare.
15665 *Gary Benson <gbenson@redhat.com>*
15667 * Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
15668 messages are stored in a single piece (fixed-length part and
15669 variable-length part combined) and fix various bugs found on the way.
15673 * Disable caching in BIO_gethostbyname(), directly use gethostbyname()
15674 instead. BIO_gethostbyname() does not know what timeouts are
15675 appropriate, so entries would stay in cache even when they have
15677 *Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>*
15679 * Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
15680 faced with a pathologically small ClientHello fragment that does
15681 not contain client_version: Instead of aborting with an error,
15682 simply choose the highest available protocol version (i.e.,
15683 TLS 1.0 unless it is disabled). In practice, ClientHello
15684 messages are never sent like this, but this change gives us
15685 strictly correct behaviour at least for TLS.
15689 * Fix SSL handshake functions and SSL_clear() such that SSL_clear()
15690 never resets s->method to s->ctx->method when called from within
15691 one of the SSL handshake functions.
15693 *Bodo Moeller; problem pointed out by Niko Baric*
15695 * In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
15696 (sent using the client's version number) if client_version is
15697 smaller than the protocol version in use. Also change
15698 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
15699 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
15700 the client will at least see that alert.
15704 * Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
15709 * Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
15710 client receives HelloRequest while in a handshake.
15712 *Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>*
15714 * Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
15715 should end in 'break', not 'goto end' which circumvents various
15716 cleanups done in state SSL_ST_OK. But session related stuff
15717 must be disabled for SSL_ST_OK in the case that we just sent a
15720 Also avoid some overhead by not calling ssl_init_wbio_buffer()
15721 before just sending a HelloRequest.
15723 *Bodo Moeller, Eric Rescorla <ekr@rtfm.com>*
15725 * Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
15726 reveal whether illegal block cipher padding was found or a MAC
15727 verification error occurred. (Neither SSLerr() codes nor alerts
15728 are directly visible to potential attackers, but the information
15729 may leak via logfiles.)
15731 Similar changes are not required for the SSL 2.0 implementation
15732 because the number of padding bytes is sent in clear for SSL 2.0,
15733 and the extra bytes are just ignored. However ssl/s2_pkt.c
15734 failed to verify that the purported number of padding bytes is in
15739 * Add OpenUNIX-8 support including shared libraries
15740 (Boyd Lynn Gerber <gerberb@zenez.com>).
15744 * Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
15745 'wristwatch attack' using huge encoding parameters (cf.
15746 James H. Manger's CRYPTO 2001 paper). Note that the
15747 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
15748 encoding parameters and hence was not vulnerable.
15752 * BN_sqr() bug fix.
15754 *Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>*
15756 * Rabin-Miller test analyses assume uniformly distributed witnesses,
15757 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
15758 followed by modular reduction.
15760 *Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>*
15762 * Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
15763 equivalent based on BN_pseudo_rand() instead of BN_rand().
15767 * s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
15768 This function was broken, as the check for a new client hello message
15769 to handle SGC did not allow these large messages.
15770 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
15774 * Add alert descriptions for TLSv1 to `SSL_alert_desc_string[_long]()`.
15778 * Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
15779 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
15783 * Rework the configuration and shared library support for Tru64 Unix.
15784 The configuration part makes use of modern compiler features and
15785 still retains old compiler behavior for those that run older versions
15786 of the OS. The shared library support part includes a variant that
15787 uses the RPATH feature, and is available through the special
15788 configuration target "alpha-cc-rpath", which will never be selected
15791 *Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte*
15793 * In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
15794 with the same message size as in ssl3_get_certificate_request().
15795 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
15796 messages might inadvertently be reject as too long.
15798 *Petr Lampa <lampa@fee.vutbr.cz>*
15800 * Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
15804 * Modified SSL library such that the verify_callback that has been set
15805 specifically for an SSL object with SSL_set_verify() is actually being
15806 used. Before the change, a verify_callback set with this function was
15807 ignored and the verify_callback() set in the SSL_CTX at the time of
15808 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
15809 to allow the necessary settings.
15813 * Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
15814 explicitly to NULL, as at least on Solaris 8 this seems not always to be
15815 done automatically (in contradiction to the requirements of the C
15816 standard). This made problems when used from OpenSSH.
15820 * In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
15821 dh->length and always used
15823 BN_rand_range(priv_key, dh->p).
15825 BN_rand_range() is not necessary for Diffie-Hellman, and this
15826 specific range makes Diffie-Hellman unnecessarily inefficient if
15827 dh->length (recommended exponent length) is much smaller than the
15828 length of dh->p. We could use BN_rand_range() if the order of
15829 the subgroup was stored in the DH structure, but we only have
15834 BN_rand(priv_key, l, ...)
15836 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
15843 RSA_eay_public_encrypt
15844 RSA_eay_private_decrypt
15845 RSA_eay_private_encrypt (signing)
15846 RSA_eay_public_decrypt (signature verification)
15848 (default implementations for RSA_public_encrypt,
15849 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
15850 always reject numbers >= n.
15854 * In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
15855 to synchronize access to 'locking_thread'. This is necessary on
15856 systems where access to 'locking_thread' (an 'unsigned long'
15857 variable) is not atomic.
15861 * In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
15862 *before* setting the 'crypto_lock_rand' flag. The previous code had
15863 a race condition if 0 is a valid thread ID.
15865 *Travis Vitek <vitek@roguewave.com>*
15867 * Add support for shared libraries under Irix.
15869 *Albert Chin-A-Young <china@thewrittenword.com>*
15871 * Add configuration option to build on Linux on both big-endian and
15872 little-endian MIPS.
15874 *Ralf Baechle <ralf@uni-koblenz.de>*
15876 * Add the possibility to create shared libraries on HP-UX.
15880 ### Changes between 0.9.6a and 0.9.6b [9 Jul 2001]
15882 * Change ssleay_rand_bytes (crypto/rand/md_rand.c)
15883 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
15884 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
15885 PRNG state recovery was possible based on the output of
15886 one PRNG request appropriately sized to gain knowledge on
15887 'md' followed by enough consecutive 1-byte PRNG requests
15888 to traverse all of 'state'.
15890 1. When updating 'md_local' (the current thread's copy of 'md')
15891 during PRNG output generation, hash all of the previous
15892 'md_local' value, not just the half used for PRNG output.
15894 2. Make the number of bytes from 'state' included into the hash
15895 independent from the number of PRNG bytes requested.
15897 The first measure alone would be sufficient to avoid
15898 Markku-Juhani's attack. (Actually it had never occurred
15899 to me that the half of 'md_local' used for chaining was the
15900 half from which PRNG output bytes were taken -- I had always
15901 assumed that the secret half would be used.) The second
15902 measure makes sure that additional data from 'state' is never
15903 mixed into 'md_local' in small portions; this heuristically
15904 further strengthens the PRNG.
15908 * Fix crypto/bn/asm/mips3.s.
15912 * When only the key is given to "enc", the IV is undefined. Print out
15913 an error message in this case.
15917 * Handle special case when X509_NAME is empty in X509 printing routines.
15921 * In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
15922 positive and less than q.
15926 * Don't change `*pointer` in CRYPTO_add_lock() is add_lock_callback is
15927 used: it isn't thread safe and the add_lock_callback should handle
15930 *Paul Rose <Paul.Rose@bridge.com>*
15932 * Verify that incoming data obeys the block size in
15933 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
15939 *Ulf Möller, Bodo Möller*
15941 * The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
15942 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
15943 when fixing the server behaviour for backwards-compatible 'client
15944 hello' messages. (Note that the attack is impractical against
15945 SSL 3.0 and TLS 1.0 anyway because length and version checking
15946 means that the probability of guessing a valid ciphertext is
15947 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
15950 Before 0.9.5, the countermeasure (hide the error by generating a
15951 random 'decryption result') did not work properly because
15952 ERR_clear_error() was missing, meaning that SSL_get_error() would
15953 detect the supposedly ignored error.
15955 Both problems are now fixed.
15959 * In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
15960 (previously it was 1024).
15964 * Fix for compatibility mode trust settings: ignore trust settings
15965 unless some valid trust or reject settings are present.
15969 * Fix for blowfish EVP: its a variable length cipher.
15973 * Fix various bugs related to DSA S/MIME verification. Handle missing
15974 parameters in DSA public key structures and return an error in the
15975 DSA routines if parameters are absent.
15979 * In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
15980 in the current directory if neither $RANDFILE nor $HOME was set.
15981 RAND_file_name() in 0.9.6a returned NULL in this case. This has
15982 caused some confusion to Windows users who haven't defined $HOME.
15983 Thus RAND_file_name() is changed again: e_os.h can define a
15984 DEFAULT_HOME, which will be used if $HOME is not set.
15985 For Windows, we use "C:"; on other platforms, we still require
15986 environment variables.
15988 * Move 'if (!initialized) RAND_poll()' into regions protected by
15989 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
15990 having multiple threads call RAND_poll() concurrently.
15994 * In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
15995 combination of a flag and a thread ID variable.
15996 Otherwise while one thread is in ssleay_rand_bytes (which sets the
15997 flag), *other* threads can enter ssleay_add_bytes without obeying
15998 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
15999 that they do not hold after the first thread unsets add_do_not_lock).
16003 * Change bctest again: '-x' expressions are not available in all
16004 versions of 'test'.
16008 ### Changes between 0.9.6 and 0.9.6a [5 Apr 2001]
16010 * Fix a couple of memory leaks in PKCS7_dataDecode()
16012 *Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>*
16014 * Change Configure and Makefiles to provide EXE_EXT, which will contain
16015 the default extension for executables, if any. Also, make the perl
16016 scripts that use symlink() to test if it really exists and use "cp"
16017 if it doesn't. All this made OpenSSL compilable and installable in
16022 * Fix for asn1_GetSequence() for indefinite length constructed data.
16023 If SEQUENCE is length is indefinite just set c->slen to the total
16024 amount of data available.
16026 *Steve Henson, reported by shige@FreeBSD.org*
16028 *This change does not apply to 0.9.7.*
16030 * Change bctest to avoid here-documents inside command substitution
16031 (workaround for FreeBSD /bin/sh bug).
16032 For compatibility with Ultrix, avoid shell functions (introduced
16033 in the bctest version that searches along $PATH).
16037 * Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
16038 with des_encrypt() defined on some operating systems, like Solaris
16043 * Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
16044 On the Importance of Eliminating Errors in Cryptographic
16045 Computations, J. Cryptology 14 (2001) 2, 101-119,
16046 <http://theory.stanford.edu/~dabo/papers/faults.ps.gz>).
16050 * MIPS assembler BIGNUM division bug fix.
16054 * Disabled incorrect Alpha assembler code.
16058 * Fix PKCS#7 decode routines so they correctly update the length
16059 after reading an EOC for the EXPLICIT tag.
16063 *This change does not apply to 0.9.7.*
16065 * Fix bug in PKCS#12 key generation routines. This was triggered
16066 if a 3DES key was generated with a 0 initial byte. Include
16067 PKCS12_BROKEN_KEYGEN compilation option to retain the old
16068 (but broken) behaviour.
16072 * Enhance bctest to search for a working bc along $PATH and print
16075 *Tim Rice <tim@multitalents.net> via Richard Levitte*
16077 * Fix memory leaks in err.c: free err_data string if necessary;
16078 don't write to the wrong index in ERR_set_error_data.
16082 * Implement ssl23_peek (analogous to ssl23_read), which previously
16087 * Replace rdtsc with `_emit` statements for VC++ version 5.
16089 *Jeremy Cooper <jeremy@baymoo.org>*
16091 * Make it possible to reuse SSLv2 sessions.
16095 * In copy_email() check for >= 0 as a return value for
16096 X509_NAME_get_index_by_NID() since 0 is a valid index.
16098 *Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>*
16100 * Avoid coredump with unsupported or invalid public keys by checking if
16101 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
16102 PKCS7_verify() fails with non detached data.
16106 * Don't use getenv in library functions when run as setuid/setgid.
16107 New function OPENSSL_issetugid().
16111 * Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
16112 due to incorrect handling of multi-threading:
16114 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
16116 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
16118 3. Count how many times MemCheck_off() has been called so that
16119 nested use can be treated correctly. This also avoids
16120 inband-signalling in the previous code (which relied on the
16121 assumption that thread ID 0 is impossible).
16125 * Add "-rand" option also to s_client and s_server.
16129 * Fix CPU detection on Irix 6.x.
16130 *Kurt Hockenbury <khockenb@stevens-tech.edu> and
16131 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>*
16133 * Fix X509_NAME bug which produced incorrect encoding if X509_NAME
16138 *This change does not apply to 0.9.7.*
16140 * Use the cached encoding of an X509_NAME structure rather than
16141 copying it. This is apparently the reason for the libsafe "errors"
16142 but the code is actually correct.
16146 * Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
16147 Bleichenbacher's DSA attack.
16148 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
16149 to be set and top=0 forces the highest bit to be set; top=-1 is new
16150 and leaves the highest bit random.
16152 *Ulf Moeller, Bodo Moeller*
16154 * In the `NCONF_...`-based implementations for `CONF_...` queries
16155 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
16156 a temporary CONF structure with the data component set to NULL
16157 (which gives segmentation faults in lh_retrieve).
16158 Instead, use NULL for the CONF pointer in CONF_get_string and
16159 CONF_get_number (which may use environment variables) and directly
16160 return NULL from CONF_get_section.
16164 * Fix potential buffer overrun for EBCDIC.
16168 * Tolerate nonRepudiation as being valid for S/MIME signing and certSign
16169 keyUsage if basicConstraints absent for a CA.
16173 * Make SMIME_write_PKCS7() write mail header values with a format that
16174 is more generally accepted (no spaces before the semicolon), since
16175 some programs can't parse those values properly otherwise. Also make
16176 sure BIO's that break lines after each write do not create invalid
16181 * Make the CRL encoding routines work with empty SEQUENCE OF. The
16182 macros previously used would not encode an empty SEQUENCE OF
16183 and break the signature.
16187 *This change does not apply to 0.9.7.*
16189 * Zero the premaster secret after deriving the master secret in
16194 * Add some EVP_add_digest_alias registrations (as found in
16195 OpenSSL_add_all_digests()) to SSL_library_init()
16196 aka OpenSSL_add_ssl_algorithms(). This provides improved
16197 compatibility with peers using X.509 certificates
16198 with unconventional AlgorithmIdentifier OIDs.
16202 * Fix for Irix with NO_ASM.
16204 *"Bruce W. Forsberg" <bruce.forsberg@baesystems.com>*
16206 * ./config script fixes.
16208 *Ulf Moeller, Richard Levitte*
16210 * Fix 'openssl passwd -1'.
16214 * Change PKCS12_key_gen_asc() so it can cope with non null
16215 terminated strings whose length is passed in the passlen
16216 parameter, for example from PEM callbacks. This was done
16217 by adding an extra length parameter to asc2uni().
16219 *Steve Henson, reported by <oddissey@samsung.co.kr>*
16221 * Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
16222 call failed, free the DSA structure.
16226 * Fix to uni2asc() to cope with zero length Unicode strings.
16227 These are present in some PKCS#12 files.
16231 * Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
16232 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
16233 when writing a 32767 byte record.
16235 *Bodo Moeller; problem reported by Eric Day <eday@concentric.net>*
16237 * In `RSA_eay_public_{en,ed}crypt` and RSA_eay_mod_exp (rsa_eay.c),
16238 obtain lock CRYPTO_LOCK_RSA before setting `rsa->_method_mod_{n,p,q}`.
16240 (RSA objects have a reference count access to which is protected
16241 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
16242 so they are meant to be shared between threads.)
16243 *Bodo Moeller, Geoff Thorpe; original patch submitted by
16244 "Reddie, Steven" <Steven.Reddie@ca.com>*
16246 * Fix a deadlock in CRYPTO_mem_leaks().
16250 * Use better test patterns in bntest.
16254 * rand_win.c fix for Borland C.
16258 * BN_rshift bugfix for n == 0.
16262 * Add a 'bctest' script that checks for some known 'bc' bugs
16263 so that 'make test' does not abort just because 'bc' is broken.
16267 * Store verify_result within SSL_SESSION also for client side to
16268 avoid potential security hole. (Re-used sessions on the client side
16269 always resulted in verify_result==X509_V_OK, not using the original
16270 result of the server certificate verification.)
16274 * Fix ssl3_pending: If the record in s->s3->rrec is not of type
16275 SSL3_RT_APPLICATION_DATA, return 0.
16276 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
16281 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
16282 releases, have been re-implemented by renaming the previous
16283 implementations of ssl2_read and ssl3_read to ssl2_read_internal
16284 and ssl3_read_internal, respectively, and adding 'peek' parameters
16285 to them. The new ssl[23]_{read,peek} functions are calls to
16286 ssl[23]_read_internal with the 'peek' flag set appropriately.
16287 A 'peek' parameter has also been added to ssl3_read_bytes, which
16288 does the actual work for ssl3_read_internal.
16292 * Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
16293 the method-specific "init()" handler. Also clean up ex_data after
16294 calling the method-specific "finish()" handler. Previously, this was
16295 happening the other way round.
16299 * Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
16300 The previous value, 12, was not always sufficient for BN_mod_exp().
16304 * Make sure that shared libraries get the internal name engine with
16305 the full version number and not just 0. This should mark the
16306 shared libraries as not backward compatible. Of course, this should
16307 be changed again when we can guarantee backward binary compatibility.
16311 * Fix typo in get_cert_by_subject() in by_dir.c
16313 *Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>*
16315 * Rework the system to generate shared libraries:
16317 - Make note of the expected extension for the shared libraries and
16318 if there is a need for symbolic links from for example libcrypto.so.0
16319 to libcrypto.so.0.9.7. There is extended info in Configure for
16322 - Make as few rebuilds of the shared libraries as possible.
16324 - Still avoid linking the OpenSSL programs with the shared libraries.
16326 - When installing, install the shared libraries separately from the
16331 * Fix SSL_CTX_set_read_ahead macro to actually use its argument.
16333 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
16334 and not in SSL_clear because the latter is also used by the
16335 accept/connect functions; previously, the settings made by
16336 SSL_set_read_ahead would be lost during the handshake.
16338 *Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>*
16340 * Correct util/mkdef.pl to be selective about disabled algorithms.
16341 Previously, it would create entries for disabled algorithms no
16346 * Added several new manual pages for SSL_* function.
16350 ### Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
16352 * In ssl23_get_client_hello, generate an error message when faced
16353 with an initial SSL 3.0/TLS record that is too small to contain the
16354 first two bytes of the ClientHello message, i.e. client_version.
16355 (Note that this is a pathologic case that probably has never happened
16356 in real life.) The previous approach was to use the version number
16357 from the record header as a substitute; but our protocol choice
16358 should not depend on that one because it is not authenticated
16359 by the Finished messages.
16363 * More robust randomness gathering functions for Windows.
16365 *Jeffrey Altman <jaltman@columbia.edu>*
16367 * For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
16368 not set then we don't setup the error code for issuer check errors
16369 to avoid possibly overwriting other errors which the callback does
16370 handle. If an application does set the flag then we assume it knows
16371 what it is doing and can handle the new informational codes
16376 * Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
16377 a general "ANY" type, as such it should be able to decode anything
16378 including tagged types. However it didn't check the class so it would
16379 wrongly interpret tagged types in the same way as their universal
16380 counterpart and unknown types were just rejected. Changed so that the
16381 tagged and unknown types are handled in the same way as a SEQUENCE:
16382 that is the encoding is stored intact. There is also a new type
16383 "V_ASN1_OTHER" which is used when the class is not universal, in this
16384 case we have no idea what the actual type is so we just lump them all
16389 * On VMS, stdout may very well lead to a file that is written to
16390 in a record-oriented fashion. That means that every write() will
16391 write a separate record, which will be read separately by the
16392 programs trying to read from it. This can be very confusing.
16394 The solution is to put a BIO filter in the way that will buffer
16395 text until a linefeed is reached, and then write everything a
16396 line at a time, so every record written will be an actual line,
16397 not chunks of lines and not (usually doesn't happen, but I've
16398 seen it once) several lines in one record. BIO_f_linebuffer() is
16401 Currently, it's a VMS-only method, because that's where it has
16402 been tested well enough.
16406 * Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
16407 it can return incorrect results.
16408 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
16409 but it was in 0.9.6-beta[12].)
16413 * Disable the check for content being present when verifying detached
16414 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
16415 include zero length content when signing messages.
16419 * New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
16420 BIO_ctrl (for BIO pairs).
16424 * Add DSO method for VMS.
16428 * Bug fix: Montgomery multiplication could produce results with the
16433 * Add RPM specification openssl.spec and modify it to build three
16434 packages. The default package contains applications, application
16435 documentation and run-time libraries. The devel package contains
16436 include files, static libraries and function documentation. The
16437 doc package contains the contents of the doc directory. The original
16438 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
16442 * Add a large number of documentation files for many SSL routines.
16444 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>*
16446 * Add a configuration entry for Sony News 4.
16448 *NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>*
16450 * Don't set the two most significant bits to one when generating a
16451 random number < q in the DSA library.
16455 * New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
16456 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
16457 the underlying transport is blocking) if a handshake took place.
16458 (The default behaviour is needed by applications such as s_client
16459 and s_server that use select() to determine when to use SSL_read;
16460 but for applications that know in advance when to expect data, it
16461 just makes things more complicated.)
16465 * Add RAND_egd_bytes(), which gives control over the number of bytes read
16470 * Add a few more EBCDIC conditionals that make `req` and `x509`
16471 work better on such systems.
16473 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
16475 * Add two demo programs for PKCS12_parse() and PKCS12_create().
16476 Update PKCS12_parse() so it copies the friendlyName and the
16477 keyid to the certificates aux info.
16481 * Fix bug in PKCS7_verify() which caused an infinite loop
16482 if there was more than one signature.
16484 *Sven Uszpelkat <su@celocom.de>*
16486 * Major change in util/mkdef.pl to include extra information
16487 about each symbol, as well as presenting variables as well
16488 as functions. This change means that there's n more need
16489 to rebuild the .num files when some algorithms are excluded.
16493 * Allow the verify time to be set by an application,
16494 rather than always using the current time.
16498 * Phase 2 verify code reorganisation. The certificate
16499 verify code now looks up an issuer certificate by a
16500 number of criteria: subject name, authority key id
16501 and key usage. It also verifies self signed certificates
16502 by the same criteria. The main comparison function is
16503 X509_check_issued() which performs these checks.
16505 Lot of changes were necessary in order to support this
16506 without completely rewriting the lookup code.
16508 Authority and subject key identifier are now cached.
16510 The LHASH 'certs' is X509_STORE has now been replaced
16511 by a STACK_OF(X509_OBJECT). This is mainly because an
16512 LHASH can't store or retrieve multiple objects with
16513 the same hash value.
16515 As a result various functions (which were all internal
16516 use only) have changed to handle the new X509_STORE
16517 structure. This will break anything that messed round
16518 with X509_STORE internally.
16520 The functions X509_STORE_add_cert() now checks for an
16521 exact match, rather than just subject name.
16523 The X509_STORE API doesn't directly support the retrieval
16524 of multiple certificates matching a given criteria, however
16525 this can be worked round by performing a lookup first
16526 (which will fill the cache with candidate certificates)
16527 and then examining the cache for matches. This is probably
16528 the best we can do without throwing out X509_LOOKUP
16529 entirely (maybe later...).
16531 The X509_VERIFY_CTX structure has been enhanced considerably.
16533 All certificate lookup operations now go via a get_issuer()
16534 callback. Although this currently uses an X509_STORE it
16535 can be replaced by custom lookups. This is a simple way
16536 to bypass the X509_STORE hackery necessary to make this
16537 work and makes it possible to use more efficient techniques
16538 in future. A very simple version which uses a simple
16539 STACK for its trusted certificate store is also provided
16540 using X509_STORE_CTX_trusted_stack().
16542 The verify_cb() and verify() callbacks now have equivalents
16543 in the X509_STORE_CTX structure.
16545 X509_STORE_CTX also has a 'flags' field which can be used
16546 to customise the verify behaviour.
16550 * Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
16551 excludes S/MIME capabilities.
16555 * When a certificate request is read in keep a copy of the
16556 original encoding of the signed data and use it when outputting
16557 again. Signatures then use the original encoding rather than
16558 a decoded, encoded version which may cause problems if the
16559 request is improperly encoded.
16563 * For consistency with other BIO_puts implementations, call
16564 buffer_write(b, ...) directly in buffer_puts instead of calling
16567 In BIO_puts, increment b->num_write as in BIO_write.
16569 *Peter.Sylvester@EdelWeb.fr*
16571 * Fix BN_mul_word for the case where the word is 0. (We have to use
16572 BN_zero, we may not return a BIGNUM with an array consisting of
16573 words set to zero.)
16577 * Avoid calling abort() from within the library when problems are
16578 detected, except if preprocessor symbols have been defined
16579 (such as REF_CHECK, BN_DEBUG etc.).
16583 * New openssl application 'rsautl'. This utility can be
16584 used for low-level RSA operations. DER public key
16585 BIO/fp routines also added.
16589 * New Configure entry and patches for compiling on QNX 4.
16591 *Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>*
16593 * A demo state-machine implementation was sponsored by
16594 Nuron (<http://www.nuron.com/>) and is now available in
16595 demos/state_machine.
16599 * New options added to the 'dgst' utility for signature
16600 generation and verification.
16604 * Unrecognized PKCS#7 content types are now handled via a
16605 catch all ASN1_TYPE structure. This allows unsupported
16606 types to be stored as a "blob" and an application can
16607 encode and decode it manually.
16611 * Fix various signed/unsigned issues to make a_strex.c
16612 compile under VC++.
16614 *Oscar Jacobsson <oscar.jacobsson@celocom.com>*
16616 * ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
16617 length if passed a buffer. ASN1_INTEGER_to_BN failed
16618 if passed a NULL BN and its argument was negative.
16620 *Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>*
16622 * Modification to PKCS#7 encoding routines to output definite
16623 length encoding. Since currently the whole structures are in
16624 memory there's not real point in using indefinite length
16625 constructed encoding. However if OpenSSL is compiled with
16626 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
16630 * Added BIO_vprintf() and BIO_vsnprintf().
16634 * Added more prefixes to parse for in the strings written
16635 through a logging bio, to cover all the levels that are available
16636 through syslog. The prefixes are now:
16638 PANIC, EMERG, EMR => LOG_EMERG
16639 ALERT, ALR => LOG_ALERT
16640 CRIT, CRI => LOG_CRIT
16641 ERROR, ERR => LOG_ERR
16642 WARNING, WARN, WAR => LOG_WARNING
16643 NOTICE, NOTE, NOT => LOG_NOTICE
16644 INFO, INF => LOG_INFO
16645 DEBUG, DBG => LOG_DEBUG
16647 and as before, if none of those prefixes are present at the
16648 beginning of the string, LOG_ERR is chosen.
16650 On Win32, the `LOG_*` levels are mapped according to this:
16652 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
16653 LOG_WARNING => EVENTLOG_WARNING_TYPE
16654 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
16658 * Made it possible to reconfigure with just the configuration
16659 argument "reconf" or "reconfigure". The command line arguments
16660 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
16661 and are retrieved from there when reconfiguring.
16667 *Assar Westerlund <assar@sics.se>, Richard Levitte*
16669 * Add the arguments -CAfile and -CApath to the pkcs12 utility.
16673 * The obj_dat.pl script was messing up the sorting of object
16674 names. The reason was that it compared the quoted version
16675 of strings as a result "OCSP" > "OCSP Signing" because
16676 " > SPACE. Changed script to store unquoted versions of
16677 names and add quotes on output. It was also omitting some
16678 names from the lookup table if they were given a default
16679 value (that is if SN is missing it is given the same
16680 value as LN and vice versa), these are now added on the
16681 grounds that if an object has a name we should be able to
16682 look it up. Finally added warning output when duplicate
16683 short or long names are found.
16687 * Changes needed for Tandem NSK.
16689 *Scott Uroff <scott@xypro.com>*
16691 * Fix SSL 2.0 rollback checking: Due to an off-by-one error in
16692 RSA_padding_check_SSLv23(), special padding was never detected
16693 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
16694 version rollback attacks was not effective.
16696 In s23_clnt.c, don't use special rollback-attack detection padding
16697 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
16698 client; similarly, in s23_srvr.c, don't do the rollback check if
16699 SSL 2.0 is the only protocol enabled in the server.
16703 * Make it possible to get hexdumps of unprintable data with 'openssl
16704 asn1parse'. By implication, the functions ASN1_parse_dump() and
16705 BIO_dump_indent() are added.
16709 * New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
16710 these print out strings and name structures based on various
16711 flags including RFC2253 support and proper handling of
16712 multibyte characters. Added options to the 'x509' utility
16713 to allow the various flags to be set.
16717 * Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
16718 Also change the functions X509_cmp_current_time() and
16719 X509_gmtime_adj() work with an ASN1_TIME structure,
16720 this will enable certificates using GeneralizedTime in validity
16721 dates to be checked.
16725 * Make the NEG_PUBKEY_BUG code (which tolerates invalid
16726 negative public key encodings) on by default,
16727 NO_NEG_PUBKEY_BUG can be set to disable it.
16731 * New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
16732 content octets. An i2c_ASN1_OBJECT is unnecessary because
16733 the encoding can be trivially obtained from the structure.
16737 * crypto/err.c locking bugfix: Use write locks (`CRYPTO_w_[un]lock`),
16738 not read locks (`CRYPTO_r_[un]lock`).
16742 * A first attempt at creating official support for shared
16743 libraries through configuration. I've kept it so the
16744 default is static libraries only, and the OpenSSL programs
16745 are always statically linked for now, but there are
16746 preparations for dynamic linking in place.
16747 This has been tested on Linux and Tru64.
16751 * Randomness polling function for Win9x, as described in:
16752 Peter Gutmann, Software Generation of Practically Strong
16757 * Fix so PRNG is seeded in req if using an already existing
16762 * New options to smime application. -inform and -outform
16763 allow alternative formats for the S/MIME message including
16764 PEM and DER. The -content option allows the content to be
16765 specified separately. This should allow things like Netscape
16766 form signing output easier to verify.
16770 * Fix the ASN1 encoding of tags using the 'long form'.
16774 * New ASN1 functions, `i2c_*` and `c2i_*` for INTEGER and BIT
16775 STRING types. These convert content octets to and from the
16776 underlying type. The actual tag and length octets are
16777 already assumed to have been read in and checked. These
16778 are needed because all other string types have virtually
16779 identical handling apart from the tag. By having versions
16780 of the ASN1 functions that just operate on content octets
16781 IMPLICIT tagging can be handled properly. It also allows
16782 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
16783 and ASN1_INTEGER are identical apart from the tag.
16787 * Change the handling of OID objects as follows:
16789 - New object identifiers are inserted in objects.txt, following
16790 the syntax given in [crypto/objects/README.md](crypto/objects/README.md).
16791 - objects.pl is used to process obj_mac.num and create a new
16793 - obj_dat.pl is used to create a new obj_dat.h, using the data in
16796 This is currently kind of a hack, and the perl code in objects.pl
16797 isn't very elegant, but it works as I intended. The simplest way
16798 to check that it worked correctly is to look in obj_dat.h and
16799 check the array nid_objs and make sure the objects haven't moved
16800 around (this is important!). Additions are OK, as well as
16801 consistent name changes.
16805 * Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
16809 * Addition of the command line parameter '-rand file' to 'openssl req'.
16810 The given file adds to whatever has already been seeded into the
16811 random pool through the RANDFILE configuration file option or
16812 environment variable, or the default random state file.
16816 * mkstack.pl now sorts each macro group into lexical order.
16817 Previously the output order depended on the order the files
16818 appeared in the directory, resulting in needless rewriting
16823 * Patches to make OpenSSL compile under Win32 again. Mostly
16824 work arounds for the VC++ problem that it treats func() as
16825 func(void). Also stripped out the parts of mkdef.pl that
16826 added extra typesafe functions: these no longer exist.
16830 * Reorganisation of the stack code. The macros are now all
16831 collected in safestack.h . Each macro is defined in terms of
16832 a "stack macro" of the form `SKM_<name>(type, a, b)`. The
16833 DEBUG_SAFESTACK is now handled in terms of function casts,
16834 this has the advantage of retaining type safety without the
16835 use of additional functions. If DEBUG_SAFESTACK is not defined
16836 then the non typesafe macros are used instead. Also modified the
16837 mkstack.pl script to handle the new form. Needs testing to see
16838 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
16839 the default if no major problems. Similar behaviour for ASN1_SET_OF
16840 and PKCS12_STACK_OF.
16844 * When some versions of IIS use the 'NET' form of private key the
16845 key derivation algorithm is different. Normally MD5(password) is
16846 used as a 128 bit RC4 key. In the modified case
16847 MD5(MD5(password) + "SGCKEYSALT") is used instead. Added some
16848 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
16849 as the old Netscape_RSA functions except they have an additional
16850 'sgckey' parameter which uses the modified algorithm. Also added
16851 an -sgckey command line option to the rsa utility. Thanks to
16852 Adrian Peck <bertie@ncipher.com> for posting details of the modified
16853 algorithm to openssl-dev.
16857 * The evp_local.h macros were using 'c.##kname' which resulted in
16858 invalid expansion on some systems (SCO 5.0.5 for example).
16859 Corrected to 'c.kname'.
16861 *Phillip Porch <root@theporch.com>*
16863 * New X509_get1_email() and X509_REQ_get1_email() functions that return
16864 a STACK of email addresses from a certificate or request, these look
16865 in the subject name and the subject alternative name extensions and
16866 omit any duplicate addresses.
16870 * Re-implement BN_mod_exp2_mont using independent (and larger) windows.
16871 This makes DSA verification about 2 % faster.
16875 * Increase maximum window size in `BN_mod_exp_...` to 6 bits instead of 5
16876 (meaning that now 2^5 values will be precomputed, which is only 4 KB
16877 plus overhead for 1024 bit moduli).
16878 This makes exponentiations about 0.5 % faster for 1024 bit
16879 exponents (as measured by "openssl speed rsa2048").
16883 * Rename memory handling macros to avoid conflicts with other
16885 Malloc => OPENSSL_malloc
16886 Malloc_locked => OPENSSL_malloc_locked
16887 Realloc => OPENSSL_realloc
16888 Free => OPENSSL_free
16892 * New function BN_mod_exp_mont_word for small bases (roughly 15%
16893 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
16897 * CygWin32 support.
16899 *John Jarvie <jjarvie@newsguy.com>*
16901 * The type-safe stack code has been rejigged. It is now only compiled
16902 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
16903 by default all type-specific stack functions are "#define"d back to
16904 standard stack functions. This results in more streamlined output
16905 but retains the type-safety checking possibilities of the original
16910 * The STACK code has been cleaned up, and certain type declarations
16911 that didn't make a lot of sense have been brought in line. This has
16912 also involved a cleanup of sorts in safestack.h to more correctly
16913 map type-safe stack functions onto their plain stack counterparts.
16914 This work has also resulted in a variety of "const"ifications of
16915 lots of the code, especially `_cmp` operations which should normally
16916 be prototyped with "const" parameters anyway.
16920 * When generating bytes for the first time in md_rand.c, 'stir the pool'
16921 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
16922 (The PRNG state consists of two parts, the large pool 'state' and 'md',
16923 where all of 'md' is used each time the PRNG is used, but 'state'
16924 is used only indexed by a cyclic counter. As entropy may not be
16925 well distributed from the beginning, 'md' is important as a
16926 chaining variable. However, the output function chains only half
16927 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
16928 all of 'md', and seeding with STATE_SIZE dummy bytes will result
16929 in all of 'state' being rewritten, with the new values depending
16930 on virtually all of 'md'. This overcomes the 80 bit limitation.)
16934 * In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
16935 the handshake is continued after ssl_verify_cert_chain();
16936 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
16937 can lead to 'unexplainable' connection aborts later.
16939 *Bodo Moeller; problem tracked down by Lutz Jaenicke*
16941 * Major EVP API cipher revision.
16942 Add hooks for extra EVP features. This allows various cipher
16943 parameters to be set in the EVP interface. Support added for variable
16944 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
16945 setting of RC2 and RC5 parameters.
16947 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
16950 Remove lots of duplicated code from the EVP library. For example *every*
16951 cipher init() function handles the 'iv' in the same way according to the
16952 cipher mode. They also all do nothing if the 'key' parameter is NULL and
16953 for CFB and OFB modes they zero ctx->num.
16955 New functionality allows removal of S/MIME code RC2 hack.
16957 Most of the routines have the same form and so can be declared in terms
16960 By shifting this to the top level EVP_CipherInit() it can be removed from
16961 all individual ciphers. If the cipher wants to handle IVs or keys
16962 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
16965 Change lots of functions like EVP_EncryptUpdate() to now return a
16966 value: although software versions of the algorithms cannot fail
16967 any installed hardware versions can.
16971 * Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
16972 this option is set, tolerate broken clients that send the negotiated
16973 protocol version number instead of the requested protocol version
16978 * Call dh_tmp_cb (set by `..._TMP_DH_CB`) with correct 'is_export' flag;
16979 i.e. non-zero for export ciphersuites, zero otherwise.
16980 Previous versions had this flag inverted, inconsistent with
16981 rsa_tmp_cb (..._TMP_RSA_CB).
16983 *Bodo Moeller; problem reported by Amit Chopra*
16985 * Add missing DSA library text string. Work around for some IIS
16986 key files with invalid SEQUENCE encoding.
16990 * Add a document (doc/standards.txt) that list all kinds of standards
16991 and so on that are implemented in OpenSSL.
16995 * Enhance c_rehash script. Old version would mishandle certificates
16996 with the same subject name hash and wouldn't handle CRLs at all.
16997 Added -fingerprint option to crl utility, to support new c_rehash
17002 * Eliminate non-ANSI declarations in crypto.h and stack.h.
17006 * Fix for SSL server purpose checking. Server checking was
17007 rejecting certificates which had extended key usage present
17008 but no ssl client purpose.
17010 *Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>*
17012 * Make PKCS#12 code work with no password. The PKCS#12 spec
17013 is a little unclear about how a blank password is handled.
17014 Since the password in encoded as a BMPString with terminating
17015 double NULL a zero length password would end up as just the
17016 double NULL. However no password at all is different and is
17017 handled differently in the PKCS#12 key generation code. NS
17018 treats a blank password as zero length. MSIE treats it as no
17019 password on export: but it will try both on import. We now do
17020 the same: PKCS12_parse() tries zero length and no password if
17021 the password is set to "" or NULL (NULL is now a valid password:
17022 it wasn't before) as does the pkcs12 application.
17026 * Bugfixes in `apps/x509.c`: Avoid a memory leak; and don't use
17027 perror when PEM_read_bio_X509_REQ fails, the error message must
17028 be obtained from the error queue.
17032 * Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
17033 it in ERR_remove_state if appropriate, and change ERR_get_state
17034 accordingly to avoid race conditions (this is necessary because
17035 thread_hash is no longer constant once set).
17039 * Bugfix for linux-elf makefile.one.
17043 * RSA_get_default_method() will now cause a default
17044 RSA_METHOD to be chosen if one doesn't exist already.
17045 Previously this was only set during a call to RSA_new()
17046 or RSA_new_method(NULL) meaning it was possible for
17047 RSA_get_default_method() to return NULL.
17051 * Added native name translation to the existing DSO code
17052 that will convert (if the flag to do so is set) filenames
17053 that are sufficiently small and have no path information
17054 into a canonical native form. Eg. "blah" converted to
17055 "libblah.so" or "blah.dll" etc.
17059 * New function ERR_error_string_n(e, buf, len) which is like
17060 ERR_error_string(e, buf), but writes at most 'len' bytes
17061 including the 0 terminator. For ERR_error_string_n, 'buf'
17064 *Damien Miller <djm@mindrot.org>, Bodo Moeller*
17066 * CONF library reworked to become more general. A new CONF
17067 configuration file reader "class" is implemented as well as a
17068 new functions (`NCONF_*`, for "New CONF") to handle it. The now
17069 old `CONF_*` functions are still there, but are reimplemented to
17070 work in terms of the new functions. Also, a set of functions
17071 to handle the internal storage of the configuration data is
17072 provided to make it easier to write new configuration file
17073 reader "classes" (I can definitely see something reading a
17074 configuration file in XML format, for example), called `_CONF_*`,
17075 or "the configuration storage API"...
17077 The new configuration file reading functions are:
17079 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
17080 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
17082 NCONF_default, NCONF_WIN32
17084 NCONF_dump_fp, NCONF_dump_bio
17086 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
17087 NCONF_new creates a new CONF object. This works in the same way
17088 as other interfaces in OpenSSL, like the BIO interface.
17089 `NCONF_dump_*` dump the internal storage of the configuration file,
17090 which is useful for debugging. All other functions take the same
17091 arguments as the old `CONF_*` functions with the exception of the
17092 first that must be a `CONF *` instead of a `LHASH *`.
17094 To make it easier to use the new classes with the old `CONF_*` functions,
17095 the function CONF_set_default_method is provided.
17099 * Add '-tls1' option to 'openssl ciphers', which was already
17100 mentioned in the documentation but had not been implemented.
17101 (This option is not yet really useful because even the additional
17102 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
17106 * Initial DSO code added into libcrypto for letting OpenSSL (and
17107 OpenSSL-based applications) load shared libraries and bind to
17108 them in a portable way.
17110 *Geoff Thorpe, with contributions from Richard Levitte*
17112 ### Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
17114 * Make sure _lrotl and _lrotr are only used with MSVC.
17116 * Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
17117 (the default implementation of RAND_status).
17119 * Rename openssl x509 option '-crlext', which was added in 0.9.5,
17120 to '-clrext' (= clear extensions), as intended and documented.
17121 *Bodo Moeller; inconsistency pointed out by Michael Attili
17122 <attili@amaxo.com>*
17124 * Fix for HMAC. It wasn't zeroing the rest of the block if the key length
17125 was larger than the MD block size.
17127 *Steve Henson, pointed out by Yost William <YostW@tce.com>*
17129 * Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
17130 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
17131 using the passed key: if the passed key was a private key the result
17132 of X509_print(), for example, would be to print out all the private key
17137 * des_quad_cksum() byte order bug fix.
17138 *Ulf Möller, using the problem description in krb4-0.9.7, where
17139 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>*
17141 * Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
17144 *Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>*
17146 * For easily testing in shell scripts whether some command
17147 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
17148 returns with exit code 0 iff no command of the given name is available.
17149 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
17150 the output goes to stdout and nothing is printed to stderr.
17151 Additional arguments are always ignored.
17153 Since for each cipher there is a command of the same name,
17154 the 'no-cipher' compilation switches can be tested this way.
17156 ('openssl no-XXX' is not able to detect pseudo-commands such
17157 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
17161 * Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
17165 * For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
17166 is set; it will be thrown away anyway because each handshake creates
17168 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
17169 to parameters -- in previous versions (since OpenSSL 0.9.3) the
17170 'default key' from SSL_CTX_set_tmp_dh would always be lost, meaning
17171 you effectively got SSL_OP_SINGLE_DH_USE when using this macro.
17175 * New s_client option -ign_eof: EOF at stdin is ignored, and
17176 'Q' and 'R' lose their special meanings (quit/renegotiate).
17177 This is part of what -quiet does; unlike -quiet, -ign_eof
17178 does not suppress any output.
17182 * Add compatibility options to the purpose and trust code. The
17183 purpose X509_PURPOSE_ANY is "any purpose" which automatically
17184 accepts a certificate or CA, this was the previous behaviour,
17185 with all the associated security issues.
17187 X509_TRUST_COMPAT is the old trust behaviour: only and
17188 automatically trust self signed roots in certificate store. A
17189 new trust setting X509_TRUST_DEFAULT is used to specify that
17190 a purpose has no associated trust setting and it should instead
17191 use the value in the default purpose.
17195 * Fix the PKCS#8 DSA private key code so it decodes keys again
17196 and fix a memory leak.
17200 * In util/mkerr.pl (which implements 'make errors'), preserve
17201 reason strings from the previous version of the .c file, as
17202 the default to have only downcase letters (and digits) in
17203 automatically generated reasons codes is not always appropriate.
17207 * In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
17208 using strerror. Previously, ERR_reason_error_string() returned
17209 library names as reason strings for SYSerr; but SYSerr is a special
17210 case where small numbers are errno values, not library numbers.
17214 * Add '-dsaparam' option to 'openssl dhparam' application. This
17215 converts DSA parameters into DH parameters. (When creating parameters,
17216 DSA_generate_parameters is used.)
17220 * Include 'length' (recommended exponent length) in C code generated
17221 by 'openssl dhparam -C'.
17225 * The second argument to set_label in perlasm was already being used
17226 so couldn't be used as a "file scope" flag. Moved to third argument
17231 * In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
17232 instead of RAND_bytes for encryption IVs and salts.
17236 * Include RAND_status() into RAND_METHOD instead of implementing
17237 it only for md_rand.c Otherwise replacing the PRNG by calling
17238 RAND_set_rand_method would be impossible.
17242 * Don't let DSA_generate_key() enter an infinite loop if the random
17243 number generation fails.
17247 * New 'rand' application for creating pseudo-random output.
17251 * Added configuration support for Linux/IA64
17253 *Rolf Haberrecker <rolf@suse.de>*
17255 * Assembler module support for Mingw32.
17259 * Shared library support for HPUX (in shlib/).
17261 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous*
17263 * Shared library support for Solaris gcc.
17265 *Lutz Behnke <behnke@trustcenter.de>*
17267 ### Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
17269 * PKCS7_encrypt() was adding text MIME headers twice because they
17270 were added manually and by SMIME_crlf_copy().
17274 * In bntest.c don't call BN_rand with zero bits argument.
17276 *Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>*
17278 * BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
17279 case was implemented. This caused BN_div_recp() to fail occasionally.
17283 * Add an optional second argument to the set_label() in the perl
17284 assembly language builder. If this argument exists and is set
17285 to 1 it signals that the assembler should use a symbol whose
17286 scope is the entire file, not just the current function. This
17287 is needed with MASM which uses the format label:: for this scope.
17289 *Steve Henson, pointed out by Peter Runestig <peter@runestig.com>*
17291 * Change the ASN1 types so they are typedefs by default. Before
17292 almost all types were #define'd to ASN1_STRING which was causing
17293 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
17298 * Change names of new functions to the new get1/get0 naming
17299 convention: After 'get1', the caller owns a reference count
17300 and has to call `..._free`; 'get0' returns a pointer to some
17301 data structure without incrementing reference counters.
17302 (Some of the existing 'get' functions increment a reference
17303 counter, some don't.)
17304 Similarly, 'set1' and 'add1' functions increase reference
17305 counters or duplicate objects.
17309 * Allow for the possibility of temp RSA key generation failure:
17310 the code used to assume it always worked and crashed on failure.
17314 * Fix potential buffer overrun problem in BIO_printf().
17315 *Ulf Möller, using public domain code by Patrick Powell; problem
17316 pointed out by David Sacerdote <das33@cornell.edu>*
17318 * Support EGD <http://www.lothar.com/tech/crypto/>. New functions
17319 RAND_egd() and RAND_status(). In the command line application,
17320 the EGD socket can be specified like a seed file using RANDFILE
17325 * Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
17326 Some CAs (e.g. Verisign) distribute certificates in this form.
17330 * Remove the SSL_ALLOW_ADH compile option and set the default cipher
17331 list to exclude them. This means that no special compilation option
17332 is needed to use anonymous DH: it just needs to be included in the
17337 * Change the EVP_MD_CTX_type macro so its meaning consistent with
17338 EVP_MD_type. The old functionality is available in a new macro called
17339 EVP_MD_md(). Change code that uses it and update docs.
17343 * `..._ctrl` functions now have corresponding `..._callback_ctrl` functions
17344 where the `void *` argument is replaced by a function pointer argument.
17345 Previously `void *` was abused to point to functions, which works on
17346 many platforms, but is not correct. As these functions are usually
17347 called by macros defined in OpenSSL header files, most source code
17348 should work without changes.
17352 * `<openssl/opensslconf.h>` (which is created by Configure) now contains
17353 sections with information on -D... compiler switches used for
17354 compiling the library so that applications can see them. To enable
17355 one of these sections, a pre-processor symbol `OPENSSL_..._DEFINES`
17356 must be defined. E.g.,
17357 #define OPENSSL_ALGORITHM_DEFINES
17358 #include <openssl/opensslconf.h>
17359 defines all pertinent `NO_<algo>` symbols, such as NO_IDEA, NO_RSA, etc.
17361 *Richard Levitte, Ulf and Bodo Möller*
17363 * Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
17368 * Change the 'other' type in certificate aux info to a STACK_OF
17369 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
17370 the required ASN1 format: arbitrary types determined by an OID.
17374 * Add some PEM_write_X509_REQ_NEW() functions and a command line
17375 argument to 'req'. This is not because the function is newer or
17376 better than others it just uses the work 'NEW' in the certificate
17377 request header lines. Some software needs this.
17381 * Reorganise password command line arguments: now passwords can be
17382 obtained from various sources. Delete the PEM_cb function and make
17383 it the default behaviour: i.e. if the callback is NULL and the
17384 usrdata argument is not NULL interpret it as a null terminated pass
17385 phrase. If usrdata and the callback are NULL then the pass phrase
17386 is prompted for as usual.
17390 * Add support for the Compaq Atalla crypto accelerator. If it is installed,
17391 the support is automatically enabled. The resulting binaries will
17392 autodetect the card and use it if present.
17394 *Ben Laurie and Compaq Inc.*
17396 * Work around for Netscape hang bug. This sends certificate request
17397 and server done in one record. Since this is perfectly legal in the
17398 SSL/TLS protocol it isn't a "bug" option and is on by default. See
17399 the bugs/SSLv3 entry for more info.
17403 * HP-UX tune-up: new unified configs, HP C compiler bug workaround.
17407 * Add -rand argument to smime and pkcs12 applications and read/write
17412 * New 'passwd' tool for crypt(3) and apr1 password hashes.
17416 * Add command line password options to the remaining applications.
17420 * Bug fix for BN_div_recp() for numerators with an even number of
17425 * More tests in bntest.c, and changed test_bn output.
17429 * ./config recognizes MacOS X now.
17433 * Bug fix for BN_div() when the first words of num and divisor are
17434 equal (it gave wrong results if `(rem=(n1-q*d0)&BN_MASK2) < d0)`.
17438 * Add support for various broken PKCS#8 formats, and command line
17439 options to produce them.
17443 * New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
17444 get temporary BIGNUMs from a BN_CTX.
17448 * Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
17453 * Change the `SSLeay_add_all_*()` functions to `OpenSSL_add_all_*()` and
17454 include a #define from the old name to the new. The original intent
17455 was that statically linked binaries could for example just call
17456 SSLeay_add_all_ciphers() to just add ciphers to the table and not
17457 link with digests. This never worked because SSLeay_add_all_digests()
17458 and SSLeay_add_all_ciphers() were in the same source file so calling
17459 one would link with the other. They are now in separate source files.
17463 * Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
17467 * Use a less unusual form of the Miller-Rabin primality test (it used
17468 a binary algorithm for exponentiation integrated into the Miller-Rabin
17469 loop, our standard modexp algorithms are faster).
17473 * Support for the EBCDIC character set completed.
17475 *Martin Kraemer <Martin.Kraemer@Mch.SNI.De>*
17477 * Source code cleanups: use const where appropriate, eliminate casts,
17478 use `void *` instead of `char *` in lhash.
17482 * Bugfix: ssl3_send_server_key_exchange was not restartable
17483 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
17484 this the server could overwrite ephemeral keys that the client
17489 * Turn DSA_is_prime into a macro that calls BN_is_prime,
17490 using 50 iterations of the Rabin-Miller test.
17492 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
17493 iterations of the Rabin-Miller test as required by the appendix
17494 to FIPS PUB 186[-1]) instead of DSA_is_prime.
17495 As BN_is_prime_fasttest includes trial division, DSA parameter
17496 generation becomes much faster.
17498 This implies a change for the callback functions in DSA_is_prime
17499 and DSA_generate_parameters: The callback function is called once
17500 for each positive witness in the Rabin-Miller test, not just
17501 occasionally in the inner loop; and the parameters to the
17502 callback function now provide an iteration count for the outer
17503 loop rather than for the current invocation of the inner loop.
17504 DSA_generate_parameters additionally can call the callback
17505 function with an 'iteration count' of -1, meaning that a
17506 candidate has passed the trial division test (when q is generated
17507 from an application-provided seed, trial division is skipped).
17511 * New function BN_is_prime_fasttest that optionally does trial
17512 division before starting the Rabin-Miller test and has
17513 an additional BN_CTX * argument (whereas BN_is_prime always
17514 has to allocate at least one BN_CTX).
17515 'callback(1, -1, cb_arg)' is called when a number has passed the
17516 trial division stage.
17520 * Fix for bug in CRL encoding. The validity dates weren't being handled
17525 * New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
17529 * New function BN_pseudo_rand().
17533 * Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
17534 bignum version of BN_from_montgomery() with the working code from
17535 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
17540 * Avoid a race condition in s2_clnt.c (function get_server_hello) that
17541 made it impossible to use the same SSL_SESSION data structure in
17542 SSL2 clients in multiple threads.
17546 * The return value of RAND_load_file() no longer counts bytes obtained
17547 by stat(). RAND_load_file(..., -1) is new and uses the complete file
17548 to seed the PRNG (previously an explicit byte count was required).
17550 *Ulf Möller, Bodo Möller*
17552 * Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
17553 used `char *` instead of `void *` and had casts all over the place.
17557 * Make BN_generate_prime() return NULL on error if ret!=NULL.
17561 * Retain source code compatibility for BN_prime_checks macro:
17562 BN_is_prime(..., BN_prime_checks, ...) now uses
17563 BN_prime_checks_for_size to determine the appropriate number of
17564 Rabin-Miller iterations.
17568 * Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
17569 DH_CHECK_P_NOT_SAFE_PRIME.
17570 (Check if this is true? OpenPGP calls them "strong".)
17574 * Merge the functionality of "dh" and "gendh" programs into a new program
17575 "dhparam". The old programs are retained for now but will handle DH keys
17576 (instead of parameters) in future.
17580 * Make the ciphers, s_server and s_client programs check the return values
17581 when a new cipher list is set.
17585 * Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
17586 ciphers. Before when the 56bit ciphers were enabled the sorting was
17589 The syntax for the cipher sorting has been extended to support sorting by
17590 cipher-strength (using the strength_bits hard coded in the tables).
17591 The new command is `@STRENGTH` (see also `doc/apps/ciphers.pod`).
17593 Fix a bug in the cipher-command parser: when supplying a cipher command
17594 string with an "undefined" symbol (neither command nor alphanumeric
17595 *A-Za-z0-9*, ssl_set_cipher_list used to hang in an endless loop. Now
17596 an error is flagged.
17598 Due to the strength-sorting extension, the code of the
17599 ssl_create_cipher_list() function was completely rearranged. I hope that
17600 the readability was also increased :-)
17602 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>*
17604 * Minor change to 'x509' utility. The -CAcreateserial option now uses 1
17605 for the first serial number and places 2 in the serial number file. This
17606 avoids problems when the root CA is created with serial number zero and
17607 the first user certificate has the same issuer name and serial number
17612 * Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
17613 the new code. Add documentation for this stuff.
17617 * Changes to X509_ATTRIBUTE utilities. These have been renamed from
17618 `X509_*()` to `X509at_*()` on the grounds that they don't handle X509
17619 structures and behave in an analogous way to the X509v3 functions:
17620 they shouldn't be called directly but wrapper functions should be used
17623 So we also now have some wrapper functions that call the X509at functions
17624 when passed certificate requests. (TO DO: similar things can be done with
17625 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
17626 things. Some of these need some d2i or i2d and print functionality
17627 because they handle more complex structures.)
17631 * Add missing #ifndefs that caused missing symbols when building libssl
17632 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
17633 NO_RSA in `ssl/s2*.c`.
17635 *Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller*
17637 * Precautions against using the PRNG uninitialized: RAND_bytes() now
17638 has a return value which indicates the quality of the random data
17639 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
17640 error queue. New function RAND_pseudo_bytes() generates output that is
17641 guaranteed to be unique but not unpredictable. RAND_add is like
17642 RAND_seed, but takes an extra argument for an entropy estimate
17643 (RAND_seed always assumes full entropy).
17647 * Do more iterations of Rabin-Miller probable prime test (specifically,
17648 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
17649 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
17650 in crypto/bn/bn_prime.c for the complete table). This guarantees a
17651 false-positive rate of at most 2^-80 for random input.
17655 * Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
17659 * New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
17660 in the 0.9.5 release), this returns the chain
17661 from an X509_CTX structure with a dup of the stack and all
17662 the X509 reference counts upped: so the stack will exist
17663 after X509_CTX_cleanup() has been called. Modify pkcs12.c
17666 Also make SSL_SESSION_print() print out the verify return
17671 * Add manpage for the pkcs12 command. Also change the default
17672 behaviour so MAC iteration counts are used unless the new
17673 -nomaciter option is used. This improves file security and
17674 only older versions of MSIE (4.0 for example) need it.
17678 * Honor the no-xxx Configure options when creating .DEF files.
17682 * Add PKCS#10 attributes to field table: challengePassword,
17683 unstructuredName and unstructuredAddress. These are taken from
17684 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
17685 international characters are used.
17687 More changes to X509_ATTRIBUTE code: allow the setting of types
17688 based on strings. Remove the 'loc' parameter when adding
17689 attributes because these will be a SET OF encoding which is sorted
17694 * Initial changes to the 'req' utility to allow request generation
17695 automation. This will allow an application to just generate a template
17696 file containing all the field values and have req construct the
17699 Initial support for X509_ATTRIBUTE handling. Stacks of these are
17700 used all over the place including certificate requests and PKCS#7
17701 structures. They are currently handled manually where necessary with
17702 some primitive wrappers for PKCS#7. The new functions behave in a
17703 manner analogous to the X509 extension functions: they allow
17704 attributes to be looked up by NID and added.
17706 Later something similar to the X509V3 code would be desirable to
17707 automatically handle the encoding, decoding and printing of the
17708 more complex types. The string types like challengePassword can
17709 be handled by the string table functions.
17711 Also modified the multi byte string table handling. Now there is
17712 a 'global mask' which masks out certain types. The table itself
17713 can use the flag STABLE_NO_MASK to ignore the mask setting: this
17714 is useful when for example there is only one permissible type
17715 (as in countryName) and using the mask might result in no valid
17720 * Clean up 'Finished' handling, and add functions SSL_get_finished and
17721 SSL_get_peer_finished to allow applications to obtain the latest
17722 Finished messages sent to the peer or expected from the peer,
17723 respectively. (SSL_get_peer_finished is usually the Finished message
17724 actually received from the peer, otherwise the protocol will be aborted.)
17726 As the Finished message are message digests of the complete handshake
17727 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
17728 be used for external authentication procedures when the authentication
17729 provided by SSL/TLS is not desired or is not enough.
17733 * Enhanced support for Alpha Linux is added. Now ./config checks if
17734 the host supports BWX extension and if Compaq C is present on the
17735 $PATH. Just exploiting of the BWX extension results in 20-30%
17736 performance kick for some algorithms, e.g. DES and RC4 to mention
17737 a couple. Compaq C in turn generates ~20% faster code for MD5 and
17742 * Add support for MS "fast SGC". This is arguably a violation of the
17743 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
17744 weak crypto and after checking the certificate is SGC a second one
17745 with strong crypto. MS SGC stops the first handshake after receiving
17746 the server certificate message and sends a second client hello. Since
17747 a server will typically do all the time consuming operations before
17748 expecting any further messages from the client (server key exchange
17749 is the most expensive) there is little difference between the two.
17751 To get OpenSSL to support MS SGC we have to permit a second client
17752 hello message after we have sent server done. In addition we have to
17753 reset the MAC if we do get this second client hello.
17757 * Add a function 'd2i_AutoPrivateKey()' this will automatically decide
17758 if a DER encoded private key is RSA or DSA traditional format. Changed
17759 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
17760 format DER encoded private key. Newer code should use PKCS#8 format which
17761 has the key type encoded in the ASN1 structure. Added DER private key
17762 support to pkcs8 application.
17766 * SSL 3/TLS 1 servers now don't request certificates when an anonymous
17767 ciphersuites has been selected (as required by the SSL 3/TLS 1
17768 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
17769 is set, we interpret this as a request to violate the specification
17770 (the worst that can happen is a handshake failure, and 'correct'
17771 behaviour would result in a handshake failure anyway).
17775 * In SSL_CTX_add_session, take into account that there might be multiple
17776 SSL_SESSION structures with the same session ID (e.g. when two threads
17777 concurrently obtain them from an external cache).
17778 The internal cache can handle only one SSL_SESSION with a given ID,
17779 so if there's a conflict, we now throw out the old one to achieve
17784 * Add OIDs for idea and blowfish in CBC mode. This will allow both
17785 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
17786 some routines that use cipher OIDs: some ciphers do not have OIDs
17787 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
17792 * Simplify the trust setting structure and code. Now we just have
17793 two sequences of OIDs for trusted and rejected settings. These will
17794 typically have values the same as the extended key usage extension
17795 and any application specific purposes.
17797 The trust checking code now has a default behaviour: it will just
17798 check for an object with the same NID as the passed id. Functions can
17799 be provided to override either the default behaviour or the behaviour
17800 for a given id. SSL client, server and email already have functions
17801 in place for compatibility: they check the NID and also return "trusted"
17802 if the certificate is self signed.
17806 * Add d2i,i2d bio/fp functions for PrivateKey: these convert the
17807 traditional format into an EVP_PKEY structure.
17811 * Add a password callback function PEM_cb() which either prompts for
17812 a password if usr_data is NULL or otherwise assumes it is a null
17813 terminated password. Allow passwords to be passed on command line
17814 environment or config files in a few more utilities.
17818 * Add a bunch of DER and PEM functions to handle PKCS#8 format private
17819 keys. Add some short names for PKCS#8 PBE algorithms and allow them
17820 to be specified on the command line for the pkcs8 and pkcs12 utilities.
17821 Update documentation.
17825 * Support for ASN1 "NULL" type. This could be handled before by using
17826 ASN1_TYPE but there wasn't any function that would try to read a NULL
17827 and produce an error if it couldn't. For compatibility we also have
17828 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
17829 don't allocate anything because they don't need to.
17833 * Initial support for MacOS is now provided. Examine INSTALL.MacOS
17836 *Andy Polyakov, Roy Woods <roy@centicsystems.ca>*
17838 * Rebuild of the memory allocation routines used by OpenSSL code and
17839 possibly others as well. The purpose is to make an interface that
17840 provide hooks so anyone can build a separate set of allocation and
17841 deallocation routines to be used by OpenSSL, for example memory
17842 pool implementations, or something else, which was previously hard
17843 since Malloc(), Realloc() and Free() were defined as macros having
17844 the values malloc, realloc and free, respectively (except for Win32
17845 compilations). The same is provided for memory debugging code.
17846 OpenSSL already comes with functionality to find memory leaks, but
17847 this gives people a chance to debug other memory problems.
17849 With these changes, a new set of functions and macros have appeared:
17851 CRYPTO_set_mem_debug_functions() [F]
17852 CRYPTO_get_mem_debug_functions() [F]
17853 CRYPTO_dbg_set_options() [F]
17854 CRYPTO_dbg_get_options() [F]
17855 CRYPTO_malloc_debug_init() [M]
17857 The memory debug functions are NULL by default, unless the library
17858 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
17859 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
17860 gives the standard debugging functions that come with OpenSSL) or
17861 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
17862 provided by the library user) must be used. When the standard
17863 debugging functions are used, CRYPTO_dbg_set_options can be used to
17864 request additional information:
17865 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
17866 the CRYPTO_MDEBUG_xxx macro when compiling the library.
17868 Also, things like CRYPTO_set_mem_functions will always give the
17869 expected result (the new set of functions is used for allocation
17870 and deallocation) at all times, regardless of platform and compiler
17873 To finish it up, some functions that were never use in any other
17874 way than through macros have a new API and new semantic:
17876 CRYPTO_dbg_malloc()
17877 CRYPTO_dbg_realloc()
17880 All macros of value have retained their old syntax.
17882 *Richard Levitte and Bodo Moeller*
17884 * Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
17885 ordering of SMIMECapabilities wasn't in "strength order" and there
17886 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
17891 * Some ASN1 types with illegal zero length encoding (INTEGER,
17892 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
17894 *Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson*
17896 * Merge in my S/MIME library for OpenSSL. This provides a simple
17897 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
17898 functionality to handle multipart/signed properly) and a utility
17899 called 'smime' to call all this stuff. This is based on code I
17900 originally wrote for Celo who have kindly allowed it to be
17901 included in OpenSSL.
17905 * Add variants des_set_key_checked and des_set_key_unchecked of
17906 des_set_key (aka des_key_sched). Global variable des_check_key
17907 decides which of these is called by des_set_key; this way
17908 des_check_key behaves as it always did, but applications and
17909 the library itself, which was buggy for des_check_key == 1,
17910 have a cleaner way to pick the version they need.
17914 * New function PKCS12_newpass() which changes the password of a
17919 * Modify X509_TRUST and X509_PURPOSE so it also uses a static and
17920 dynamic mix. In both cases the ids can be used as an index into the
17921 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
17922 functions so they accept a list of the field values and the
17923 application doesn't need to directly manipulate the X509_TRUST
17928 * Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
17933 * Modify the way the V3 extension code looks up extensions. This now
17934 works in a similar way to the object code: we have some "standard"
17935 extensions in a static table which is searched with OBJ_bsearch()
17936 and the application can add dynamic ones if needed. The file
17937 crypto/x509v3/ext_dat.h now has the info: this file needs to be
17938 updated whenever a new extension is added to the core code and kept
17939 in ext_nid order. There is a simple program 'tabtest.c' which checks
17940 this. New extensions are not added too often so this file can readily
17941 be maintained manually.
17943 There are two big advantages in doing things this way. The extensions
17944 can be looked up immediately and no longer need to be "added" using
17945 X509V3_add_standard_extensions(): this function now does nothing.
17946 Side note: I get *lots* of email saying the extension code doesn't
17947 work because people forget to call this function.
17948 Also no dynamic allocation is done unless new extensions are added:
17949 so if we don't add custom extensions there is no need to call
17950 X509V3_EXT_cleanup().
17954 * Modify enc utility's salting as follows: make salting the default. Add a
17955 magic header, so unsalted files fail gracefully instead of just decrypting
17956 to garbage. This is because not salting is a big security hole, so people
17957 should be discouraged from doing it.
17961 * Fixes and enhancements to the 'x509' utility. It allowed a message
17962 digest to be passed on the command line but it only used this
17963 parameter when signing a certificate. Modified so all relevant
17964 operations are affected by the digest parameter including the
17965 -fingerprint and -x509toreq options. Also -x509toreq choked if a
17966 DSA key was used because it didn't fix the digest.
17970 * Initial certificate chain verify code. Currently tests the untrusted
17971 certificates for consistency with the verify purpose (which is set
17972 when the X509_STORE_CTX structure is set up) and checks the pathlength.
17974 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
17975 this is because it will reject chains with invalid extensions whereas
17976 every previous version of OpenSSL and SSLeay made no checks at all.
17978 Trust code: checks the root CA for the relevant trust settings. Trust
17979 settings have an initial value consistent with the verify purpose: e.g.
17980 if the verify purpose is for SSL client use it expects the CA to be
17981 trusted for SSL client use. However the default value can be changed to
17982 permit custom trust settings: one example of this would be to only trust
17983 certificates from a specific "secure" set of CAs.
17985 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
17986 which should be used for version portability: especially since the
17987 verify structure is likely to change more often now.
17989 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
17990 to set them. If not set then assume SSL clients will verify SSL servers
17993 Two new options to the verify program: -untrusted allows a set of
17994 untrusted certificates to be passed in and -purpose which sets the
17995 intended purpose of the certificate. If a purpose is set then the
17996 new chain verify code is used to check extension consistency.
18000 * Support for the authority information access extension.
18004 * Modify RSA and DSA PEM read routines to transparently handle
18005 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
18006 public keys in a format compatible with certificate
18007 SubjectPublicKeyInfo structures. Unfortunately there were already
18008 functions called *_PublicKey_* which used various odd formats so
18009 these are retained for compatibility: however the DSA variants were
18010 never in a public release so they have been deleted. Changed dsa/rsa
18011 utilities to handle the new format: note no releases ever handled public
18012 keys so we should be OK.
18014 The primary motivation for this change is to avoid the same fiasco
18015 that dogs private keys: there are several incompatible private key
18016 formats some of which are standard and some OpenSSL specific and
18017 require various evil hacks to allow partial transparent handling and
18018 even then it doesn't work with DER formats. Given the option anything
18019 other than PKCS#8 should be dumped: but the other formats have to
18020 stay in the name of compatibility.
18022 With public keys and the benefit of hindsight one standard format
18023 is used which works with EVP_PKEY, RSA or DSA structures: though
18024 it clearly returns an error if you try to read the wrong kind of key.
18026 Added a -pubkey option to the 'x509' utility to output the public key.
18027 Also rename the `EVP_PKEY_get_*()` to `EVP_PKEY_rget_*()`
18028 (renamed to `EVP_PKEY_get1_*()` in the OpenSSL 0.9.5 release) and add
18029 `EVP_PKEY_rset_*()` functions (renamed to `EVP_PKEY_set1_*()`)
18030 that do the same as the `EVP_PKEY_assign_*()` except they up the
18031 reference count of the added key (they don't "swallow" the
18036 * Fixes to crypto/x509/by_file.c the code to read in certificates and
18037 CRLs would fail if the file contained no certificates or no CRLs:
18038 added a new function to read in both types and return the number
18039 read: this means that if none are read it will be an error. The
18040 DER versions of the certificate and CRL reader would always fail
18041 because it isn't possible to mix certificates and CRLs in DER format
18042 without choking one or the other routine. Changed this to just read
18043 a certificate: this is the best we can do. Also modified the code
18044 in `apps/verify.c` to take notice of return codes: it was previously
18045 attempting to read in certificates from NULL pointers and ignoring
18046 any errors: this is one reason why the cert and CRL reader seemed
18047 to work. It doesn't check return codes from the default certificate
18048 routines: these may well fail if the certificates aren't installed.
18052 * Code to support otherName option in GeneralName.
18056 * First update to verify code. Change the verify utility
18057 so it warns if it is passed a self signed certificate:
18058 for consistency with the normal behaviour. X509_verify
18059 has been modified to it will now verify a self signed
18060 certificate if *exactly* the same certificate appears
18061 in the store: it was previously impossible to trust a
18062 single self signed certificate. This means that:
18063 openssl verify ss.pem
18064 now gives a warning about a self signed certificate but
18065 openssl verify -CAfile ss.pem ss.pem
18070 * For servers, store verify_result in SSL_SESSION data structure
18071 (and add it to external session representation).
18072 This is needed when client certificate verifications fails,
18073 but an application-provided verification callback (set by
18074 SSL_CTX_set_cert_verify_callback) allows accepting the session
18075 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
18076 but returns 1): When the session is reused, we have to set
18077 ssl->verify_result to the appropriate error code to avoid
18080 *Bodo Moeller, problem pointed out by Lutz Jaenicke*
18082 * Fix a bug in the new PKCS#7 code: it didn't consider the
18083 case in PKCS7_dataInit() where the signed PKCS7 structure
18084 didn't contain any existing data because it was being created.
18086 *Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson*
18088 * Add a salt to the key derivation routines in enc.c. This
18089 forms the first 8 bytes of the encrypted file. Also add a
18090 -S option to allow a salt to be input on the command line.
18094 * New function X509_cmp(). Oddly enough there wasn't a function
18095 to compare two certificates. We do this by working out the SHA1
18096 hash and comparing that. X509_cmp() will be needed by the trust
18101 * SSL_get1_session() is like SSL_get_session(), but increments
18102 the reference count in the SSL_SESSION returned.
18104 *Geoff Thorpe <geoff@eu.c2.net>*
18106 * Fix for 'req': it was adding a null to request attributes.
18107 Also change the X509_LOOKUP and X509_INFO code to handle
18108 certificate auxiliary information.
18112 * Add support for 40 and 64 bit RC2 and RC4 algorithms: document
18117 * Add the possibility to add extra information to the memory leak
18118 detecting output, to form tracebacks, showing from where each
18119 allocation was originated: CRYPTO_push_info("constant string") adds
18120 the string plus current file name and line number to a per-thread
18121 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
18122 is like calling CYRPTO_pop_info() until the stack is empty.
18123 Also updated memory leak detection code to be multi-thread-safe.
18127 * Add options -text and -noout to pkcs7 utility and delete the
18128 encryption options which never did anything. Update docs.
18132 * Add options to some of the utilities to allow the pass phrase
18133 to be included on either the command line (not recommended on
18134 OSes like Unix) or read from the environment. Update the
18135 manpages and fix a few bugs.
18139 * Add a few manpages for some of the openssl commands.
18143 * Fix the -revoke option in ca. It was freeing up memory twice,
18144 leaking and not finding already revoked certificates.
18148 * Extensive changes to support certificate auxiliary information.
18149 This involves the use of X509_CERT_AUX structure and X509_AUX
18150 functions. An X509_AUX function such as PEM_read_X509_AUX()
18151 can still read in a certificate file in the usual way but it
18152 will also read in any additional "auxiliary information". By
18153 doing things this way a fair degree of compatibility can be
18154 retained: existing certificates can have this information added
18155 using the new 'x509' options.
18157 Current auxiliary information includes an "alias" and some trust
18158 settings. The trust settings will ultimately be used in enhanced
18159 certificate chain verification routines: currently a certificate
18160 can only be trusted if it is self signed and then it is trusted
18165 * Fix assembler for Alpha (tested only on DEC OSF not Linux or `*BSD`).
18166 The problem was that one of the replacement routines had not been working
18167 since SSLeay releases. For now the offending routine has been replaced
18168 with non-optimised assembler. Even so, this now gives around 95%
18169 performance improvement for 1024 bit RSA signs.
18173 * Hack to fix PKCS#7 decryption when used with some unorthodox RC2
18174 handling. Most clients have the effective key size in bits equal to
18175 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
18176 A few however don't do this and instead use the size of the decrypted key
18177 to determine the RC2 key length and the AlgorithmIdentifier to determine
18178 the effective key length. In this case the effective key length can still
18179 be 40 bits but the key length can be 168 bits for example. This is fixed
18180 by manually forcing an RC2 key into the EVP_PKEY structure because the
18181 EVP code can't currently handle unusual RC2 key sizes: it always assumes
18182 the key length and effective key length are equal.
18186 * Add a bunch of functions that should simplify the creation of
18187 X509_NAME structures. Now you should be able to do:
18188 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
18189 and have it automatically work out the correct field type and fill in
18190 the structures. The more adventurous can try:
18191 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
18192 and it will (hopefully) work out the correct multibyte encoding.
18196 * Change the 'req' utility to use the new field handling and multibyte
18197 copy routines. Before the DN field creation was handled in an ad hoc
18198 way in req, ca, and x509 which was rather broken and didn't support
18199 BMPStrings or UTF8Strings. Since some software doesn't implement
18200 BMPStrings or UTF8Strings yet, they can be enabled using the config file
18201 using the dirstring_type option. See the new comment in the default
18202 openssl.cnf for more info.
18206 * Make crypto/rand/md_rand.c more robust:
18207 - Assure unique random numbers after fork().
18208 - Make sure that concurrent threads access the global counter and
18209 md serializably so that we never lose entropy in them
18210 or use exactly the same state in multiple threads.
18211 Access to the large state is not always serializable because
18212 the additional locking could be a performance killer, and
18213 md should be large enough anyway.
18217 * New file `apps/app_rand.c` with commonly needed functionality
18218 for handling the random seed file.
18220 Use the random seed file in some applications that previously did not:
18222 dsaparam -genkey (which also ignored its '-rand' option),
18225 x509 (when signing).
18226 Except on systems with /dev/urandom, it is crucial to have a random
18227 seed file at least for key creation, DSA signing, and for DH exchanges;
18228 for RSA signatures we could do without one.
18230 gendh and gendsa (unlike genrsa) used to read only the first byte
18231 of each file listed in the '-rand' option. The function as previously
18232 found in genrsa is now in app_rand.c and is used by all programs
18233 that support '-rand'.
18237 * In RAND_write_file, use mode 0600 for creating files;
18238 don't just chmod when it may be too late.
18242 * Report an error from X509_STORE_load_locations
18243 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
18247 * New function ASN1_mbstring_copy() this copies a string in either
18248 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
18249 into an ASN1_STRING type. A mask of permissible types is passed
18250 and it chooses the "minimal" type to use or an error if not type
18255 * Add function equivalents to the various macros in asn1.h. The old
18256 macros are retained with an `M_` prefix. Code inside the library can
18257 use the `M_` macros. External code (including the openssl utility)
18258 should *NOT* in order to be "shared library friendly".
18262 * Add various functions that can check a certificate's extensions
18263 to see if it usable for various purposes such as SSL client,
18264 server or S/MIME and CAs of these types. This is currently
18265 VERY EXPERIMENTAL but will ultimately be used for certificate chain
18266 verification. Also added a -purpose flag to x509 utility to
18267 print out all the purposes.
18271 * Add a CRYPTO_EX_DATA to X509 certificate structure and associated
18276 * New `X509V3_{X509,CRL,REVOKED}_get_d2i()` functions. These will search
18277 for, obtain and decode and extension and obtain its critical flag.
18278 This allows all the necessary extension code to be handled in a
18279 single function call.
18283 * RC4 tune-up featuring 30-40% performance improvement on most RISC
18284 platforms. See crypto/rc4/rc4_enc.c for further details.
18288 * New -noout option to asn1parse. This causes no output to be produced
18289 its main use is when combined with -strparse and -out to extract data
18290 from a file (which may not be in ASN.1 format).
18294 * Fix for pkcs12 program. It was hashing an invalid certificate pointer
18295 when producing the local key id.
18297 *Richard Levitte <levitte@stacken.kth.se>*
18299 * New option -dhparam in s_server. This allows a DH parameter file to be
18300 stated explicitly. If it is not stated then it tries the first server
18301 certificate file. The previous behaviour hard coded the filename
18306 * Add -pubin and -pubout options to the rsa and dsa commands. These allow
18307 a public key to be input or output. For example:
18308 openssl rsa -in key.pem -pubout -out pubkey.pem
18309 Also added necessary DSA public key functions to handle this.
18313 * Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
18314 in the message. This was handled by allowing
18315 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
18317 *Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>*
18319 * Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
18320 to the end of the strings whereas this didn't. This would cause problems
18321 if strings read with d2i_ASN1_bytes() were later modified.
18323 *Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>*
18325 * Fix for base64 decode bug. When a base64 bio reads only one line of
18326 data and it contains EOF it will end up returning an error. This is
18327 caused by input 46 bytes long. The cause is due to the way base64
18328 BIOs find the start of base64 encoded data. They do this by trying a
18329 trial decode on each line until they find one that works. When they
18330 do a flag is set and it starts again knowing it can pass all the
18331 data directly through the decoder. Unfortunately it doesn't reset
18332 the context it uses. This means that if EOF is reached an attempt
18333 is made to pass two EOFs through the context and this causes the
18334 resulting error. This can also cause other problems as well. As is
18335 usual with these problems it takes *ages* to find and the fix is
18336 trivial: move one line.
18338 *Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer)*
18340 * Ugly workaround to get s_client and s_server working under Windows. The
18341 old code wouldn't work because it needed to select() on sockets and the
18342 tty (for keypresses and to see if data could be written). Win32 only
18343 supports select() on sockets so we select() with a 1s timeout on the
18344 sockets and then see if any characters are waiting to be read, if none
18345 are present then we retry, we also assume we can always write data to
18346 the tty. This isn't nice because the code then blocks until we've
18347 received a complete line of data and it is effectively polling the
18348 keyboard at 1s intervals: however it's quite a bit better than not
18349 working at all :-) A dedicated Windows application might handle this
18350 with an event loop for example.
18354 * Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
18355 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
18356 will be called when RSA_sign() and RSA_verify() are used. This is useful
18357 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
18358 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
18359 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
18360 This necessitated the support of an extra signature type NID_md5_sha1
18361 for SSL signatures and modifications to the SSL library to use it instead
18362 of calling RSA_public_decrypt() and RSA_private_encrypt().
18366 * Add new -verify -CAfile and -CApath options to the crl program, these
18367 will lookup a CRL issuers certificate and verify the signature in a
18368 similar way to the verify program. Tidy up the crl program so it
18369 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
18370 less strict. It will now permit CRL extensions even if it is not
18371 a V2 CRL: this will allow it to tolerate some broken CRLs.
18375 * Initialize all non-automatic variables each time one of the openssl
18376 sub-programs is started (this is necessary as they may be started
18377 multiple times from the "OpenSSL>" prompt).
18379 *Lennart Bang, Bodo Moeller*
18381 * Preliminary compilation option RSA_NULL which disables RSA crypto without
18382 removing all other RSA functionality (this is what NO_RSA does). This
18383 is so (for example) those in the US can disable those operations covered
18384 by the RSA patent while allowing storage and parsing of RSA keys and RSA
18389 * Non-copying interface to BIO pairs.
18390 (still largely untested)
18394 * New function ASN1_tag2str() to convert an ASN1 tag to a descriptive
18395 ASCII string. This was handled independently in various places before.
18399 * New functions UTF8_getc() and UTF8_putc() that parse and generate
18400 UTF8 strings a character at a time.
18404 * Use client_version from client hello to select the protocol
18405 (s23_srvr.c) and for RSA client key exchange verification
18406 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
18410 * Add various utility functions to handle SPKACs, these were previously
18411 handled by poking round in the structure internals. Added new function
18412 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
18413 print, verify and generate SPKACs. Based on an original idea from
18414 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
18418 * RIPEMD160 is operational on all platforms and is back in 'make test'.
18422 * Allow the config file extension section to be overwritten on the
18423 command line. Based on an original idea from Massimiliano Pala
18424 <madwolf@comune.modena.it>. The new option is called -extensions
18425 and can be applied to ca, req and x509. Also -reqexts to override
18426 the request extensions in req and -crlexts to override the crl extensions
18431 * Add new feature to the SPKAC handling in ca. Now you can include
18432 the same field multiple times by preceding it by "XXXX." for example:
18435 this is the same syntax as used in the req config file.
18439 * Allow certificate extensions to be added to certificate requests. These
18440 are specified in a 'req_extensions' option of the req section of the
18441 config file. They can be printed out with the -text option to req but
18442 are otherwise ignored at present.
18446 * Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
18447 data read consists of only the final block it would not decrypted because
18448 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
18449 A misplaced 'break' also meant the decrypted final block might not be
18450 copied until the next read.
18454 * Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
18455 a few extra parameters to the DH structure: these will be useful if
18456 for example we want the value of 'q' or implement X9.42 DH.
18460 * Initial support for DSA_METHOD. This is based on the RSA_METHOD and
18461 provides hooks that allow the default DSA functions or functions on a
18462 "per key" basis to be replaced. This allows hardware acceleration and
18463 hardware key storage to be handled without major modification to the
18464 library. Also added low-level modexp hooks and CRYPTO_EX structure and
18465 associated functions.
18469 * Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
18470 as "read only": it can't be written to and the buffer it points to will
18471 not be freed. Reading from a read only BIO is much more efficient than
18472 a normal memory BIO. This was added because there are several times when
18473 an area of memory needs to be read from a BIO. The previous method was
18474 to create a memory BIO and write the data to it, this results in two
18475 copies of the data and an O(n^2) reading algorithm. There is a new
18476 function BIO_new_mem_buf() which creates a read only memory BIO from
18477 an area of memory. Also modified the PKCS#7 routines to use read only
18482 * Bugfix: ssl23_get_client_hello did not work properly when called in
18483 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
18484 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
18485 but a retry condition occurred while trying to read the rest.
18489 * The PKCS7_ENC_CONTENT_new() function was setting the content type as
18490 NID_pkcs7_encrypted by default: this was wrong since this should almost
18491 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
18492 the encrypted data type: this is a more sensible place to put it and it
18493 allows the PKCS#12 code to be tidied up that duplicated this
18498 * Changed obj_dat.pl script so it takes its input and output files on
18499 the command line. This should avoid shell escape redirection problems
18504 * Initial support for certificate extension requests, these are included
18505 in things like Xenroll certificate requests. Included functions to allow
18506 extensions to be obtained and added.
18510 * -crlf option to s_client and s_server for sending newlines as
18511 CRLF (as required by many protocols).
18515 ### Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
18517 * Install libRSAglue.a when OpenSSL is built with RSAref.
18519 *Ralf S. Engelschall*
18521 * A few more `#ifndef NO_FP_API / #endif` pairs for consistency.
18523 *Andrija Antonijevic <TheAntony2@bigfoot.com>*
18525 * Fix -startdate and -enddate (which was missing) arguments to 'ca'
18530 * New function DSA_dup_DH, which duplicates DSA parameters/keys as
18531 DH parameters/keys (q is lost during that conversion, but the resulting
18532 DH parameters contain its length).
18534 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
18535 much faster than DH_generate_parameters (which creates parameters
18536 where `p = 2*q + 1`), and also the smaller q makes DH computations
18537 much more efficient (160-bit exponentiation instead of 1024-bit
18538 exponentiation); so this provides a convenient way to support DHE
18539 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
18540 utter importance to use
18541 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
18543 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
18544 when such DH parameters are used, because otherwise small subgroup
18545 attacks may become possible!
18549 * Avoid memory leak in i2d_DHparams.
18553 * Allow the -k option to be used more than once in the enc program:
18554 this allows the same encrypted message to be read by multiple recipients.
18558 * New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
18559 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
18560 it will always use the numerical form of the OID, even if it has a short
18565 * Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
18566 method only got called if p,q,dmp1,dmq1,iqmp components were present,
18567 otherwise bn_mod_exp was called. In the case of hardware keys for example
18568 no private key components need be present and it might store extra data
18569 in the RSA structure, which cannot be accessed from bn_mod_exp.
18570 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
18571 private key operations.
18575 * Added support for SPARC Linux.
18579 * pem_password_cb function type incompatibly changed from
18580 typedef int pem_password_cb(char *buf, int size, int rwflag);
18582 ....(char *buf, int size, int rwflag, void *userdata);
18583 so that applications can pass data to their callbacks:
18584 The `PEM[_ASN1]_{read,write}...` functions and macros now take an
18585 additional void * argument, which is just handed through whenever
18586 the password callback is called.
18588 *Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller*
18590 New function SSL_CTX_set_default_passwd_cb_userdata.
18592 Compatibility note: As many C implementations push function arguments
18593 onto the stack in reverse order, the new library version is likely to
18594 interoperate with programs that have been compiled with the old
18595 pem_password_cb definition (PEM_whatever takes some data that
18596 happens to be on the stack as its last argument, and the callback
18597 just ignores this garbage); but there is no guarantee whatsoever that
18600 * The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
18601 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
18602 problems not only on Windows, but also on some Unix platforms.
18603 To avoid problematic command lines, these definitions are now in an
18604 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
18605 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
18609 * MIPS III/IV assembler module is reimplemented.
18613 * More DES library cleanups: remove references to srand/rand and
18614 delete an unused file.
18618 * Add support for the free Netwide assembler (NASM) under Win32,
18619 since not many people have MASM (ml) and it can be hard to obtain.
18620 This is currently experimental but it seems to work OK and pass all
18621 the tests. Check out INSTALL.W32 for info.
18625 * Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
18626 without temporary keys kept an extra copy of the server key,
18627 and connections with temporary keys did not free everything in case
18632 * New function RSA_check_key and new openssl rsa option -check
18633 for verifying the consistency of RSA keys.
18635 *Ulf Moeller, Bodo Moeller*
18637 * Various changes to make Win32 compile work:
18638 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
18639 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
18640 comparison" warnings.
18641 3. Add `sk_<TYPE>_sort` to DEF file generator and do make update.
18645 * Add a debugging option to PKCS#5 v2 key generation function: when
18646 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
18647 derived keys are printed to stderr.
18651 * Copy the flags in ASN1_STRING_dup().
18653 *Roman E. Pavlov <pre@mo.msk.ru>*
18655 * The x509 application mishandled signing requests containing DSA
18656 keys when the signing key was also DSA and the parameters didn't match.
18658 It was supposed to omit the parameters when they matched the signing key:
18659 the verifying software was then supposed to automatically use the CA's
18660 parameters if they were absent from the end user certificate.
18662 Omitting parameters is no longer recommended. The test was also
18663 the wrong way round! This was probably due to unusual behaviour in
18664 EVP_cmp_parameters() which returns 1 if the parameters match.
18665 This meant that parameters were omitted when they *didn't* match and
18666 the certificate was useless. Certificates signed with 'ca' didn't have
18669 *Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>*
18671 * Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
18672 The interface is as follows:
18673 Applications can use
18674 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
18675 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
18676 "off" is now the default.
18677 The library internally uses
18678 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
18679 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
18680 to disable memory-checking temporarily.
18682 Some inconsistent states that previously were possible (and were
18683 even the default) are now avoided.
18685 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
18686 with each memory chunk allocated; this is occasionally more helpful
18687 than just having a counter.
18689 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
18691 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
18696 * Introduce "mode" for SSL structures (with defaults in SSL_CTX),
18697 which largely parallels "options", but is for changing API behaviour,
18698 whereas "options" are about protocol behaviour.
18699 Initial "mode" flags are:
18701 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
18702 a single record has been written.
18703 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
18704 retries use the same buffer location.
18705 (But all of the contents must be
18710 * Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
18713 * Fix problems with no-hmac etc.
18715 *Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>*
18717 * New functions RSA_get_default_method(), RSA_set_method() and
18718 RSA_get_method(). These allows replacement of RSA_METHODs without having
18719 to mess around with the internals of an RSA structure.
18723 * Fix memory leaks in DSA_do_sign and DSA_is_prime.
18724 Also really enable memory leak checks in openssl.c and in some
18727 *Chad C. Mulligan, Bodo Moeller*
18729 * Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
18730 up the length of negative integers. This has now been simplified to just
18731 store the length when it is first determined and use it later, rather
18732 than trying to keep track of where data is copied and updating it to
18734 *Steve Henson, reported by Brien Wheeler <bwheeler@authentica-security.com>*
18736 * Add a new function PKCS7_signatureVerify. This allows the verification
18737 of a PKCS#7 signature but with the signing certificate passed to the
18738 function itself. This contrasts with PKCS7_dataVerify which assumes the
18739 certificate is present in the PKCS#7 structure. This isn't always the
18740 case: certificates can be omitted from a PKCS#7 structure and be
18741 distributed by "out of band" means (such as a certificate database).
18745 * Complete the `PEM_*` macros with DECLARE_PEM versions to replace the
18746 function prototypes in pem.h, also change util/mkdef.pl to add the
18747 necessary function names.
18751 * mk1mf.pl (used by Windows builds) did not properly read the
18752 options set by Configure in the top level Makefile, and Configure
18753 was not even able to write more than one option correctly.
18754 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
18758 * New functions CONF_load_bio() and CONF_load_fp() to allow a config
18759 file to be loaded from a BIO or FILE pointer. The BIO version will
18760 for example allow memory BIOs to contain config info.
18764 * New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
18765 Whoever hopes to achieve shared-library compatibility across versions
18766 must use this, not the compile-time macro.
18767 (Exercise 0.9.4: Which is the minimum library version required by
18769 Note: All this applies only to multi-threaded programs, others don't
18774 * Add missing case to s3_clnt.c state machine -- one of the new SSL tests
18775 through a BIO pair triggered the default case, i.e.
18776 SSLerr(...,SSL_R_UNKNOWN_STATE).
18780 * New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
18781 can use the SSL library even if none of the specific BIOs is
18786 * Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
18787 for the encoded length.
18789 *Jeon KyoungHo <khjeon@sds.samsung.co.kr>*
18791 * Add initial documentation of the X509V3 functions.
18795 * Add a new pair of functions PEM_write_PKCS8PrivateKey() and
18796 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
18797 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
18798 secure PKCS#8 private key format with a high iteration count.
18802 * Fix determination of Perl interpreter: A perl or perl5
18803 *directory* in $PATH was also accepted as the interpreter.
18805 *Ralf S. Engelschall*
18807 * Fix demos/sign/sign.c: well there wasn't anything strictly speaking
18808 wrong with it but it was very old and did things like calling
18809 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
18810 unusual formatting.
18814 * Fix demos/selfsign.c: it used obsolete and deleted functions, changed
18815 to use the new extension code.
18819 * Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
18820 with macros. This should make it easier to change their form, add extra
18821 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
18826 * Add to configuration table a new entry that can specify an alternative
18827 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
18828 according to Mark Crispin <MRC@Panda.COM>.
18832 * DES CBC did not update the IV. Weird.
18836 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
18837 Changing the behaviour of the former might break existing programs --
18838 where IV updating is needed, des_ncbc_encrypt can be used.
18841 * When bntest is run from "make test" it drives bc to check its
18842 calculations, as well as internally checking them. If an internal check
18843 fails, it needs to cause bc to give a non-zero result or make test carries
18844 on without noticing the failure. Fixed.
18848 * DES library cleanups.
18852 * Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
18853 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
18854 ciphers. NOTE: although the key derivation function has been verified
18855 against some published test vectors it has not been extensively tested
18856 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
18861 * Instead of "mkdir -p", which is not fully portable, use new
18862 Perl script "util/mkdir-p.pl".
18866 * Rewrite the way password based encryption (PBE) is handled. It used to
18867 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
18868 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
18869 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
18870 the 'parameter' field of the AlgorithmIdentifier is passed to the
18871 underlying key generation function so it must do its own ASN1 parsing.
18872 This has also changed the EVP_PBE_CipherInit() function which now has a
18873 'parameter' argument instead of literal salt and iteration count values
18874 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
18878 * Support for PKCS#5 v1.5 compatible password based encryption algorithms
18879 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
18880 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
18881 KEY" because this clashed with PKCS#8 unencrypted string. Since this
18882 value was just used as a "magic string" and not used directly its
18883 value doesn't matter.
18887 * Introduce some semblance of const correctness to BN. Shame C doesn't
18892 * "linux-sparc64" configuration (ultrapenguin).
18894 *Ray Miller <ray.miller@oucs.ox.ac.uk>*
18895 "linux-sparc" configuration.
18897 *Christian Forster <fo@hawo.stw.uni-erlangen.de>*
18899 * config now generates no-xxx options for missing ciphers.
18903 * Support the EBCDIC character set (work in progress).
18904 File ebcdic.c not yet included because it has a different license.
18906 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
18908 * Support BS2000/OSD-POSIX.
18910 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
18912 * Make callbacks for key generation use `void *` instead of `char *`.
18916 * Make S/MIME samples compile (not yet tested).
18920 * Additional typesafe stacks.
18924 * New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
18928 ### Changes between 0.9.3 and 0.9.3a [29 May 1999]
18930 * New configuration variant "sco5-gcc".
18932 * Updated some demos.
18934 *Sean O Riordain, Wade Scholine*
18936 * Add missing BIO_free at exit of pkcs12 application.
18940 * Fix memory leak in conf.c.
18944 * Updates for Win32 to assembler version of MD5.
18948 * Set #! path to perl in `apps/der_chop` to where we found it
18949 instead of using a fixed path.
18953 * SHA library changes for irix64-mips4-cc.
18957 * Improvements for VMS support.
18961 ### Changes between 0.9.2b and 0.9.3 [24 May 1999]
18963 * Bignum library bug fix. IRIX 6 passes "make test" now!
18964 This also avoids the problems with SC4.2 and unpatched SC5.
18966 *Andy Polyakov <appro@fy.chalmers.se>*
18968 * New functions sk_num, sk_value and sk_set to replace the previous macros.
18969 These are required because of the typesafe stack would otherwise break
18970 existing code. If old code used a structure member which used to be STACK
18971 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
18972 sk_num or sk_value it would produce an error because the num, data members
18973 are not present in STACK_OF. Now it just produces a warning. sk_set
18974 replaces the old method of assigning a value to sk_value
18975 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
18976 that does this will no longer work (and should use sk_set instead) but
18977 this could be regarded as a "questionable" behaviour anyway.
18981 * Fix most of the other PKCS#7 bugs. The "experimental" code can now
18982 correctly handle encrypted S/MIME data.
18986 * Change type of various DES function arguments from des_cblock
18987 (which means, in function argument declarations, pointer to char)
18988 to des_cblock * (meaning pointer to array with 8 char elements),
18989 which allows the compiler to do more typechecking; it was like
18990 that back in SSLeay, but with lots of ugly casts.
18992 Introduce new type const_des_cblock.
18996 * Reorganise the PKCS#7 library and get rid of some of the more obvious
18997 problems: find RecipientInfo structure that matches recipient certificate
18998 and initialise the ASN1 structures properly based on passed cipher.
19002 * Belatedly make the BN tests actually check the results.
19006 * Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
19007 to and from BNs: it was completely broken. New compilation option
19008 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
19009 key elements as negative integers.
19013 * Reorganize and speed up MD5.
19015 *Andy Polyakov <appro@fy.chalmers.se>*
19019 *Richard Levitte <richard@levitte.org>*
19021 * New option -out to asn1parse to allow the parsed structure to be
19022 output to a file. This is most useful when combined with the -strparse
19023 option to examine the output of things like OCTET STRINGS.
19027 * Make SSL library a little more fool-proof by not requiring any longer
19028 that `SSL_set_{accept,connect}_state` be called before
19029 `SSL_{accept,connect}` may be used (`SSL_set_..._state` is omitted
19030 in many applications because usually everything *appeared* to work as
19031 intended anyway -- now it really works as intended).
19035 * Move openssl.cnf out of lib/.
19039 * Fix various things to let OpenSSL even pass "egcc -pipe -O2 -Wall
19040 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
19041 -Wmissing-declarations -Wnested-externs -Winline" with EGCS 1.1.2+
19043 *Ralf S. Engelschall*
19045 * Various fixes to the EVP and PKCS#7 code. It may now be able to
19046 handle PKCS#7 enveloped data properly.
19048 *Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve*
19050 * Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
19051 copying pointers. The cert_st handling is changed by this in
19052 various ways (and thus what used to be known as ctx->default_cert
19053 is now called ctx->cert, since we don't resort to `s->ctx->[default_]cert`
19054 any longer when s->cert does not give us what we need).
19055 ssl_cert_instantiate becomes obsolete by this change.
19056 As soon as we've got the new code right (possibly it already is?),
19057 we have solved a couple of bugs of the earlier code where s->cert
19058 was used as if it could not have been shared with other SSL structures.
19060 Note that using the SSL API in certain dirty ways now will result
19061 in different behaviour than observed with earlier library versions:
19062 Changing settings for an `SSL_CTX *ctx` after having done s = SSL_new(ctx)
19063 does not influence s as it used to.
19065 In order to clean up things more thoroughly, inside SSL_SESSION
19066 we don't use CERT any longer, but a new structure SESS_CERT
19067 that holds per-session data (if available); currently, this is
19068 the peer's certificate chain and, for clients, the server's certificate
19069 and temporary key. CERT holds only those values that can have
19070 meaningful defaults in an SSL_CTX.
19074 * New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
19075 from the internal representation. Various PKCS#7 fixes: remove some
19076 evil casts and set the enc_dig_alg field properly based on the signing
19081 * Allow PKCS#12 password to be set from the command line or the
19082 environment. Let 'ca' get its config file name from the environment
19083 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
19088 * Allow certificate policies extension to use an IA5STRING for the
19089 organization field. This is contrary to the PKIX definition but
19090 VeriSign uses it and IE5 only recognises this form. Document 'x509'
19095 * Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
19096 without disallowing inline assembler and the like for non-pedantic builds.
19100 * Support Borland C++ builder.
19102 *Janez Jere <jj@void.si>, modified by Ulf Möller*
19108 * SHA-1 cleanups and performance enhancements.
19110 *Andy Polyakov <appro@fy.chalmers.se>*
19112 * Sparc v8plus assembler for the bignum library.
19114 *Andy Polyakov <appro@fy.chalmers.se>*
19116 * Accept any -xxx and +xxx compiler options in Configure.
19120 * Update HPUX configuration.
19124 * Add missing `sk_<type>_unshift()` function to safestack.h
19126 *Ralf S. Engelschall*
19128 * New function SSL_CTX_use_certificate_chain_file that sets the
19129 "extra_cert"s in addition to the certificate. (This makes sense
19130 only for "PEM" format files, as chains as a whole are not
19135 * Support verify_depth from the SSL API.
19136 x509_vfy.c had what can be considered an off-by-one-error:
19137 Its depth (which was not part of the external interface)
19138 was actually counting the number of certificates in a chain;
19139 now it really counts the depth.
19143 * Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
19144 instead of X509err, which often resulted in confusing error
19145 messages since the error codes are not globally unique
19146 (e.g. an alleged error in ssl3_accept when a certificate
19147 didn't match the private key).
19149 * New function SSL_CTX_set_session_id_context that allows to set a default
19150 value (so that you don't need SSL_set_session_id_context for each
19151 connection using the SSL_CTX).
19155 * OAEP decoding bug fix.
19159 * Support INSTALL_PREFIX for package builders, as proposed by
19164 * New Configure options "threads" and "no-threads". For systems
19165 where the proper compiler options are known (currently Solaris
19166 and Linux), "threads" is the default.
19170 * New script util/mklink.pl as a faster substitute for util/mklink.sh.
19174 * Install various scripts to $(OPENSSLDIR)/misc, not to
19175 $(INSTALLTOP)/bin -- they shouldn't clutter directories
19176 such as /usr/local/bin.
19180 * "make linux-shared" to build shared libraries.
19182 *Niels Poppe <niels@netbox.org>*
19184 * New Configure option `no-<cipher>` (rsa, idea, rc5, ...).
19188 * Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
19189 extension adding in x509 utility.
19193 * Remove NOPROTO sections and error code comments.
19197 * Partial rewrite of the DEF file generator to now parse the ANSI
19202 * New Configure options --prefix=DIR and --openssldir=DIR.
19206 * Complete rewrite of the error code script(s). It is all now handled
19207 by one script at the top level which handles error code gathering,
19208 header rewriting and C source file generation. It should be much better
19209 than the old method: it now uses a modified version of Ulf's parser to
19210 read the ANSI prototypes in all header files (thus the old K&R definitions
19211 aren't needed for error creation any more) and do a better job of
19212 translating function codes into names. The old 'ASN1 error code embedded
19213 in a comment' is no longer necessary and it doesn't use .err files which
19214 have now been deleted. Also the error code call doesn't have to appear all
19215 on one line (which resulted in some large lines...).
19219 * Change #include filenames from `<foo.h>` to `<openssl/foo.h>`.
19223 * Change behaviour of ssl2_read when facing length-0 packets: Don't return
19224 0 (which usually indicates a closed connection), but continue reading.
19228 * Fix some race conditions.
19232 * Add support for CRL distribution points extension. Add Certificate
19233 Policies and CRL distribution points documentation.
19237 * Move the autogenerated header file parts to crypto/opensslconf.h.
19241 * Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
19242 8 of keying material. Merlin has also confirmed interop with this fix
19243 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
19245 *Merlin Hughes <merlin@baltimore.ie>*
19247 * Fix lots of warnings.
19249 *Richard Levitte <levitte@stacken.kth.se>*
19251 * In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
19252 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
19254 *Richard Levitte <levitte@stacken.kth.se>*
19256 * Fix problems with sizeof(long) == 8.
19258 *Andy Polyakov <appro@fy.chalmers.se>*
19260 * Change functions to ANSI C.
19264 * Fix typos in error codes.
19266 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller*
19268 * Remove defunct assembler files from Configure.
19272 * SPARC v8 assembler BIGNUM implementation.
19274 *Andy Polyakov <appro@fy.chalmers.se>*
19276 * Support for Certificate Policies extension: both print and set.
19277 Various additions to support the r2i method this uses.
19281 * A lot of constification, and fix a bug in X509_NAME_oneline() that could
19282 return a const string when you are expecting an allocated buffer.
19286 * Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
19287 types DirectoryString and DisplayText.
19291 * Add code to allow r2i extensions to access the configuration database,
19292 add an LHASH database driver and add several ctx helper functions.
19296 * Fix an evil bug in bn_expand2() which caused various BN functions to
19297 fail when they extended the size of a BIGNUM.
19301 * Various utility functions to handle SXNet extension. Modify mkdef.pl to
19302 support typesafe stack.
19306 * Fix typo in SSL_[gs]et_options().
19308 *Nils Frostberg <nils@medcom.se>*
19310 * Delete various functions and files that belonged to the (now obsolete)
19311 old X509V3 handling code.
19315 * New Configure option "rsaref".
19319 * Don't auto-generate pem.h.
19323 * Introduce type-safe ASN.1 SETs.
19327 * Convert various additional casted stacks to type-safe STACK_OF() variants.
19329 *Ben Laurie, Ralf S. Engelschall, Steve Henson*
19331 * Introduce type-safe STACKs. This will almost certainly break lots of code
19332 that links with OpenSSL (well at least cause lots of warnings), but fear
19333 not: the conversion is trivial, and it eliminates loads of evil casts. A
19334 few STACKed things have been converted already. Feel free to convert more.
19335 In the fullness of time, I'll do away with the STACK type altogether.
19339 * Add `openssl ca -revoke <certfile>` facility which revokes a certificate
19340 specified in `<certfile>` by updating the entry in the index.txt file.
19341 This way one no longer has to edit the index.txt file manually for
19342 revoking a certificate. The -revoke option does the gory details now.
19344 *Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall*
19346 * Fix `openssl crl -noout -text` combination where `-noout` killed the
19347 `-text` option at all and this way the `-noout -text` combination was
19348 inconsistent in `openssl crl` with the friends in `openssl x509|rsa|dsa`.
19350 *Ralf S. Engelschall*
19352 * Make sure a corresponding plain text error message exists for the
19353 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
19354 verify callback function determined that a certificate was revoked.
19356 *Ralf S. Engelschall*
19358 * Bugfix: In test/testenc, don't test `openssl <cipher>` for
19359 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
19360 all available ciphers including rc5, which was forgotten until now.
19361 In order to let the testing shell script know which algorithms
19362 are available, a new (up to now undocumented) command
19363 `openssl list-cipher-commands` is used.
19367 * Bugfix: s_client occasionally would sleep in select() when
19368 it should have checked SSL_pending() first.
19372 * New functions DSA_do_sign and DSA_do_verify to provide access to
19373 the raw DSA values prior to ASN.1 encoding.
19377 * Tweaks to Configure
19379 *Niels Poppe <niels@netbox.org>*
19381 * Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
19386 * New variables $(RANLIB) and $(PERL) in the Makefiles.
19390 * New config option to avoid instructions that are illegal on the 80386.
19391 The default code is faster, but requires at least a 486.
19395 * Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
19396 SSL2_SERVER_VERSION (not used at all) macros, which are now the
19397 same as SSL2_VERSION anyway.
19401 * New "-showcerts" option for s_client.
19405 * Still more PKCS#12 integration. Add pkcs12 application to openssl
19406 application. Various cleanups and fixes.
19410 * More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
19411 modify error routines to work internally. Add error codes and PBE init
19412 to library startup routines.
19416 * Further PKCS#12 integration. Added password based encryption, PKCS#8 and
19417 packing functions to asn1 and evp. Changed function names and error
19418 codes along the way.
19422 * PKCS12 integration: and so it begins... First of several patches to
19423 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
19424 objects to objects.h
19428 * Add a new 'indent' option to some X509V3 extension code. Initial ASN1
19429 and display support for Thawte strong extranet extension.
19433 * Add LinuxPPC support.
19435 *Jeff Dubrule <igor@pobox.org>*
19437 * Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
19438 bn_div_words in alpha.s.
19440 *Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie*
19442 * Make sure the RSA OAEP test is skipped under -DRSAref because
19443 OAEP isn't supported when OpenSSL is built with RSAref.
19445 *Ulf Moeller <ulf@fitug.de>*
19447 * Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
19448 so they no longer are missing under -DNOPROTO.
19450 *Soren S. Jorvang <soren@t.dk>*
19452 ### Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
19454 * Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
19455 doesn't work when the session is reused. Coming soon!
19459 * Fix a security hole, that allows sessions to be reused in the wrong
19460 context thus bypassing client cert protection! All software that uses
19461 client certs and session caches in multiple contexts NEEDS PATCHING to
19462 allow session reuse! A fuller solution is in the works.
19464 *Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)*
19466 * Some more source tree cleanups (removed obsolete files
19467 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
19468 permission on "config" script to be executable) and a fix for the INSTALL
19471 *Ulf Moeller <ulf@fitug.de>*
19473 * Remove some legacy and erroneous uses of malloc, free instead of
19476 *Lennart Bang <lob@netstream.se>, with minor changes by Steve*
19478 * Make rsa_oaep_test return non-zero on error.
19480 *Ulf Moeller <ulf@fitug.de>*
19482 * Add support for native Solaris shared libraries. Configure
19483 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
19484 if someone would make that last step automatic.
19486 *Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>*
19488 * ctx_size was not built with the right compiler during "make links". Fixed.
19492 * Change the meaning of 'ALL' in the cipher list. It now means "everything
19493 except NULL ciphers". This means the default cipher list will no longer
19494 enable NULL ciphers. They need to be specifically enabled e.g. with
19495 the string "DEFAULT:eNULL".
19499 * Fix to RSA private encryption routines: if p < q then it would
19500 occasionally produce an invalid result. This will only happen with
19501 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
19505 * Be less restrictive and allow also `perl util/perlpath.pl
19506 /path/to/bin/perl` in addition to `perl util/perlpath.pl /path/to/bin`,
19507 because this way one can also use an interpreter named `perl5` (which is
19508 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
19509 installed as `perl`).
19511 *Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
19513 * Let util/clean-depend.pl work also with older Perl 5.00x versions.
19515 *Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
19517 * Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
19518 advapi32.lib to Win32 build and change the pem test comparison
19519 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
19520 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
19521 and crypto/des/ede_cbcm_enc.c.
19525 * DES quad checksum was broken on big-endian architectures. Fixed.
19529 * Comment out two functions in bio.h that aren't implemented. Fix up the
19530 Win32 test batch file so it (might) work again. The Win32 test batch file
19531 is horrible: I feel ill....
19535 * Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
19536 in e_os.h. Audit of header files to check ANSI and non ANSI
19537 sections: 10 functions were absent from non ANSI section and not exported
19538 from Windows DLLs. Fixed up libeay.num for new functions.
19542 * Make `openssl version` output lines consistent.
19544 *Ralf S. Engelschall*
19546 * Fix Win32 symbol export lists for BIO functions: Added
19547 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
19548 to ms/libeay{16,32}.def.
19550 *Ralf S. Engelschall*
19552 * Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
19553 fine under Unix and passes some trivial tests I've now added. But the
19554 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
19555 added to make sure no one expects that this stuff really works in the
19556 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
19557 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
19560 *Ralf S. Engelschall*
19562 * Fix the generation of two part addresses in perl.
19564 *Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie*
19566 * Add config entry for Linux on MIPS.
19568 *John Tobey <jtobey@channel1.com>*
19570 * Make links whenever Configure is run, unless we are on Windoze.
19574 * Permit extensions to be added to CRLs using crl_section in openssl.cnf.
19575 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
19580 * Add a useful kludge to allow package maintainers to specify compiler and
19581 other platforms details on the command line without having to patch the
19582 Configure script every time: One now can use
19583 `perl Configure <id>:<details>`,
19584 i.e. platform ids are allowed to have details appended
19585 to them (separated by colons). This is treated as there would be a static
19586 pre-configured entry in Configure's %table under key `<id>` with value
19587 `<details>` and `perl Configure <id>` is called. So, when you want to
19588 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
19589 assembler stuff you can use `perl Configure "FreeBSD-elf:pgcc:-O6:::"`
19590 now, which overrides the FreeBSD-elf entry on-the-fly.
19592 *Ralf S. Engelschall*
19594 * Disable new TLS1 ciphersuites by default: they aren't official yet.
19598 * Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
19599 on the `perl Configure ...` command line. This way one can compile
19600 OpenSSL libraries with Position Independent Code (PIC) which is needed
19601 for linking it into DSOs.
19603 *Ralf S. Engelschall*
19605 * Remarkably, export ciphers were totally broken and no-one had noticed!
19610 * Cleaned up the LICENSE document: The official contact for any license
19611 questions now is the OpenSSL core team under openssl-core@openssl.org.
19612 And add a paragraph about the dual-license situation to make sure people
19613 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
19614 to the OpenSSL toolkit.
19616 *Ralf S. Engelschall*
19618 * General source tree makefile cleanups: Made `making xxx in yyy...`
19619 display consistent in the source tree and replaced `/bin/rm` by `rm`.
19620 Additionally cleaned up the `make links` target: Remove unnecessary
19621 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
19622 to speed processing and no longer clutter the display with confusing
19623 stuff. Instead only the actually done links are displayed.
19625 *Ralf S. Engelschall*
19627 * Permit null encryption ciphersuites, used for authentication only. It used
19628 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
19629 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
19634 * Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
19635 signed attributes when verifying signatures (this would break them),
19636 the detached data encoding was wrong and public keys obtained using
19637 X509_get_pubkey() weren't freed.
19641 * Add text documentation for the BUFFER functions. Also added a work around
19642 to a Win95 console bug. This was triggered by the password read stuff: the
19643 last character typed gets carried over to the next fread(). If you were
19644 generating a new cert request using 'req' for example then the last
19645 character of the passphrase would be CR which would then enter the first
19650 * Added the new 'Includes OpenSSL Cryptography Software' button as
19651 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
19652 button and can be used by applications based on OpenSSL to show the
19653 relationship to the OpenSSL project.
19655 *Ralf S. Engelschall*
19657 * Remove confusing variables in function signatures in files
19658 ssl/ssl_lib.c and ssl/ssl.h.
19660 *Lennart Bong <lob@kulthea.stacken.kth.se>*
19662 * Don't install bss_file.c under PREFIX/include/
19664 *Lennart Bong <lob@kulthea.stacken.kth.se>*
19666 * Get the Win32 compile working again. Modify mkdef.pl so it can handle
19667 functions that return function pointers and has support for NT specific
19668 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
19669 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
19670 unsigned to signed types: this was killing the Win32 compile.
19674 * Add new certificate file to stack functions,
19675 SSL_add_dir_cert_subjects_to_stack() and
19676 SSL_add_file_cert_subjects_to_stack(). These largely supplant
19677 SSL_load_client_CA_file(), and can be used to add multiple certs easily
19678 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
19679 This means that Apache-SSL and similar packages don't have to mess around
19680 to add as many CAs as they want to the preferred list.
19684 * Experiment with doxygen documentation. Currently only partially applied to
19686 See <http://www.stack.nl/~dimitri/doxygen/index.html>, and run doxygen with
19687 openssl.doxy as the configuration file.
19691 * Get rid of remaining C++-style comments which strict C compilers hate.
19693 *Ralf S. Engelschall, pointed out by Carlos Amengual*
19695 * Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
19696 compiled in by default: it has problems with large keys.
19700 * Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
19701 DH private keys and/or callback functions which directly correspond to
19702 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
19703 is needed for applications which have to configure certificates on a
19704 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
19706 For the RSA certificate situation is makes no difference, but
19707 for the DSA certificate situation this fixes the "no shared cipher"
19708 problem where the OpenSSL cipher selection procedure failed because the
19709 temporary keys were not overtaken from the context and the API provided
19710 no way to reconfigure them.
19711 The new functions now let applications reconfigure the stuff and they
19712 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
19713 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
19714 non-public-API function ssl_cert_instantiate() is used as a helper
19715 function and also to reduce code redundancy inside ssl_rsa.c.
19717 *Ralf S. Engelschall*
19719 * Move s_server -dcert and -dkey options out of the undocumented feature
19720 area because they are useful for the DSA situation and should be
19721 recognized by the users.
19723 *Ralf S. Engelschall*
19725 * Fix the cipher decision scheme for export ciphers: the export bits are
19726 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
19727 SSL_EXP_MASK. So, the original variable has to be used instead of the
19728 already masked variable.
19730 *Richard Levitte <levitte@stacken.kth.se>*
19732 * Fix `port` variable from `int` to `unsigned int` in crypto/bio/b_sock.c
19734 *Richard Levitte <levitte@stacken.kth.se>*
19736 * Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
19737 from `int` to `unsigned int` because it is a length and initialized by
19738 EVP_DigestFinal() which expects an `unsigned int *`.
19740 *Richard Levitte <levitte@stacken.kth.se>*
19742 * Don't hard-code path to Perl interpreter on shebang line of Configure
19743 script. Instead use the usual Shell->Perl transition trick.
19745 *Ralf S. Engelschall*
19747 * Make `openssl x509 -noout -modulus`' functional also for DSA certificates
19748 (in addition to RSA certificates) to match the behaviour of `openssl dsa
19749 -noout -modulus` as it's already the case for `openssl rsa -noout
19750 -modulus`. For RSA the -modulus is the real "modulus" while for DSA
19751 currently the public key is printed (a decision which was already done by
19752 `openssl dsa -modulus` in the past) which serves a similar purpose.
19753 Additionally the NO_RSA no longer completely removes the whole -modulus
19754 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
19757 *Ralf S. Engelschall*
19759 * Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
19760 BIO. See the source (crypto/evp/bio_ok.c) for more info.
19762 *Arne Ansper <arne@ats.cyber.ee>*
19764 * Dump the old yucky req code that tried (and failed) to allow raw OIDs
19765 to be added. Now both 'req' and 'ca' can use new objects defined in the
19770 * Add cool BIO that does syslog (or event log on NT).
19772 *Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie*
19774 * Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
19775 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
19776 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
19777 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
19781 * Add preliminary config info for new extension code.
19785 * Make RSA_NO_PADDING really use no padding.
19787 *Ulf Moeller <ulf@fitug.de>*
19789 * Generate errors when private/public key check is done.
19793 * Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
19794 for some CRL extensions and new objects added.
19798 * Really fix the ASN1 IMPLICIT bug this time... Partial support for private
19799 key usage extension and fuller support for authority key id.
19803 * Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
19804 padding method for RSA, which is recommended for new applications in PKCS
19805 #1 v2.0 (RFC 2437, October 1998).
19806 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
19807 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
19808 against Bleichbacher's attack on RSA.
19809 *Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
19812 * Updates to the new SSL compression code
19814 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
19816 * Fix so that the version number in the master secret, when passed
19817 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
19818 (because the server will not accept higher), that the version number
19819 is 0x03,0x01, not 0x03,0x00
19821 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
19823 * Run extensive memory leak checks on SSL commands. Fixed *lots* of memory
19824 leaks in `ssl/` relating to new `X509_get_pubkey()` behaviour. Also fixes
19825 in `apps/` and an unrelated leak in `crypto/dsa/dsa_vrf.c`.
19829 * Support for RAW extensions where an arbitrary extension can be
19830 created by including its DER encoding. See `apps/openssl.cnf` for
19835 * Make sure latest Perl versions don't interpret some generated C array
19836 code as Perl array code in the crypto/err/err_genc.pl script.
19838 *Lars Weber <3weber@informatik.uni-hamburg.de>*
19840 * Modify ms/do_ms.bat to not generate assembly language makefiles since
19841 not many people have the assembler. Various Win32 compilation fixes and
19842 update to the INSTALL.W32 file with (hopefully) more accurate Win32
19843 build instructions.
19847 * Modify configure script 'Configure' to automatically create crypto/date.h
19848 file under Win32 and also build pem.h from pem.org. New script
19849 util/mkfiles.pl to create the MINFO file on environments that can't do a
19850 'make files': perl util/mkfiles.pl >MINFO should work.
19854 * Major rework of DES function declarations, in the pursuit of correctness
19855 and purity. As a result, many evil casts evaporated, and some weirdness,
19856 too. You may find this causes warnings in your code. Zapping your evil
19857 casts will probably fix them. Mostly.
19861 * Fix for a typo in asn1.h. Bug fix to object creation script
19862 obj_dat.pl. It considered a zero in an object definition to mean
19863 "end of object": none of the objects in objects.h have any zeros
19864 so it wasn't spotted.
19866 *Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>*
19868 * Add support for Triple DES Cipher Block Chaining with Output Feedback
19869 Masking (CBCM). In the absence of test vectors, the best I have been able
19870 to do is check that the decrypt undoes the encrypt, so far. Send me test
19871 vectors if you have them.
19875 * Correct calculation of key length for export ciphers (too much space was
19876 allocated for null ciphers). This has not been tested!
19880 * Modifications to the mkdef.pl for Win32 DEF file creation. The usage
19881 message is now correct (it understands "crypto" and "ssl" on its
19882 command line). There is also now an "update" option. This will update
19883 the util/ssleay.num and util/libeay.num files with any new functions.
19885 perl util/mkdef.pl crypto ssl update
19886 it will update them.
19890 * Overhauled the Perl interface:
19891 - ported BN stuff to OpenSSL's different BN library
19892 - made the perl/ source tree CVS-aware
19893 - renamed the package from SSLeay to OpenSSL (the files still contain
19894 their history because I've copied them in the repository)
19895 - removed obsolete files (the test scripts will be replaced
19896 by better Test::Harness variants in the future)
19898 *Ralf S. Engelschall*
19900 * First cut for a very conservative source tree cleanup:
19901 1. merge various obsolete readme texts into doc/ssleay.txt
19902 where we collect the old documents and readme texts.
19903 2. remove the first part of files where I'm already sure that we no
19904 longer need them because of three reasons: either they are just temporary
19905 files which were left by Eric or they are preserved original files where
19906 I've verified that the diff is also available in the CVS via "cvs diff
19907 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
19908 the crypto/md/ stuff).
19910 *Ralf S. Engelschall*
19912 * More extension code. Incomplete support for subject and issuer alt
19913 name, issuer and authority key id. Change the i2v function parameters
19914 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
19915 what that's for :-) Fix to ASN1 macro which messed up
19916 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
19920 * Preliminary support for ENUMERATED type. This is largely copied from the
19925 * Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
19927 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
19929 * Make sure `make rehash` target really finds the `openssl` program.
19931 *Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
19933 * Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
19934 like to hear about it if this slows down other processors.
19938 * Add CygWin32 platform information to Configure script.
19940 *Alan Batie <batie@aahz.jf.intel.com>*
19942 * Fixed ms/32all.bat script: `no_asm` -> `no-asm`
19944 *Rainer W. Gerling <gerling@mpg-gv.mpg.de>*
19946 * New program nseq to manipulate netscape certificate sequences
19950 * Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
19955 * Fixes to BN code. Previously the default was to define BN_RECURSION
19956 but the BN code had some problems that would cause failures when
19957 doing certificate verification and some other functions.
19959 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
19961 * Add ASN1 and PEM code to support netscape certificate sequences.
19965 * Add ASN1 and PEM code to support netscape certificate sequences.
19969 * Add several PKIX and private extended key usage OIDs.
19973 * Modify the 'ca' program to handle the new extension code. Modify
19974 openssl.cnf for new extension format, add comments.
19978 * More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
19979 and add a sample to openssl.cnf so req -x509 now adds appropriate
19984 * Continued X509 V3 changes. Add to other makefiles, integrate with the
19985 error code, add initial support to X509_print() and x509 application.
19989 * Takes a deep breath and start adding X509 V3 extension support code. Add
19990 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
19991 stuff is currently isolated and isn't even compiled yet.
19995 * Continuing patches for GeneralizedTime. Fix up certificate and CRL
19996 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
19997 Removed the versions check from X509 routines when loading extensions:
19998 this allows certain broken certificates that don't set the version
19999 properly to be processed.
20003 * Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
20004 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
20005 can still be regenerated with "make depend".
20009 * Spelling mistake in C version of CAST-128.
20011 *Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>*
20013 * Changes to the error generation code. The perl script err-code.pl
20014 now reads in the old error codes and retains the old numbers, only
20015 adding new ones if necessary. It also only changes the .err files if new
20016 codes are added. The makefiles have been modified to only insert errors
20017 when needed (to avoid needlessly modifying header files). This is done
20018 by only inserting errors if the .err file is newer than the auto generated
20019 C file. To rebuild all the error codes from scratch (the old behaviour)
20020 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
20021 or delete all the .err files.
20025 * CAST-128 was incorrectly implemented for short keys. The C version has
20026 been fixed, but is untested. The assembler versions are also fixed, but
20027 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
20028 to regenerate it if needed.
20029 *Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
20030 Hagino <itojun@kame.net>*
20032 * File was opened incorrectly in randfile.c.
20034 *Ulf Möller <ulf@fitug.de>*
20036 * Beginning of support for GeneralizedTime. d2i, i2d, check and print
20037 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
20038 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
20039 al: it's just almost always a UTCTime. Note this patch adds new error
20040 codes so do a "make errors" if there are problems.
20044 * Correct Linux 1 recognition in config.
20046 *Ulf Möller <ulf@fitug.de>*
20048 * Remove pointless MD5 hash when using DSA keys in ca.
20050 *Anonymous <nobody@replay.com>*
20052 * Generate an error if given an empty string as a cert directory. Also
20053 generate an error if handed NULL (previously returned 0 to indicate an
20054 error, but didn't set one).
20056 *Ben Laurie, reported by Anonymous <nobody@replay.com>*
20058 * Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
20062 * Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
20063 parameters. This was causing a warning which killed off the Win32 compile.
20067 * Remove C++ style comments from crypto/bn/bn_local.h.
20069 *Neil Costigan <neil.costigan@celocom.com>*
20071 * The function OBJ_txt2nid was broken. It was supposed to return a nid
20072 based on a text string, looking up short and long names and finally
20073 "dot" format. The "dot" format stuff didn't work. Added new function
20074 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
20075 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
20076 OID is not part of the table.
20080 * Add prototypes to X509 lookup/verify methods, fixing a bug in
20081 X509_LOOKUP_by_alias().
20085 * Sort openssl functions by name.
20089 * Get the `gendsa` command working and add it to the `list` command. Remove
20090 encryption from sample DSA keys (in case anyone is interested the password
20095 * Make *all* `*_free` functions accept a NULL pointer.
20097 *Frans Heymans <fheymans@isaserver.be>*
20099 * If a DH key is generated in s3_srvr.c, don't blow it by trying to use
20102 *Anonymous <nobody@replay.com>*
20104 * s_server should send the CAfile as acceptable CAs, not its own cert.
20106 *Bodo Moeller <3moeller@informatik.uni-hamburg.de>*
20108 * Don't blow it for numeric `-newkey` arguments to `apps/req`.
20110 *Bodo Moeller <3moeller@informatik.uni-hamburg.de>*
20112 * Temp key "for export" tests were wrong in s3_srvr.c.
20114 *Anonymous <nobody@replay.com>*
20116 * Add prototype for temp key callback functions
20117 SSL_CTX_set_tmp_{rsa,dh}_callback().
20121 * Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
20122 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
20126 * X509_name_add_entry() freed the wrong thing after an error.
20128 *Arne Ansper <arne@ats.cyber.ee>*
20130 * rsa_eay.c would attempt to free a NULL context.
20132 *Arne Ansper <arne@ats.cyber.ee>*
20134 * BIO_s_socket() had a broken should_retry() on Windoze.
20136 *Arne Ansper <arne@ats.cyber.ee>*
20138 * BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
20140 *Arne Ansper <arne@ats.cyber.ee>*
20142 * Make sure the already existing X509_STORE->depth variable is initialized
20143 in X509_STORE_new(), but document the fact that this variable is still
20144 unused in the certificate verification process.
20146 *Ralf S. Engelschall*
20148 * Fix the various library and `apps/` files to free up pkeys obtained from
20149 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
20153 * Fix reference counting in X509_PUBKEY_get(). This makes
20154 demos/maurice/example2.c work, amongst others, probably.
20156 *Steve Henson and Ben Laurie*
20158 * First cut of a cleanup for `apps/`. First the `ssleay` program is now named
20159 `openssl` and second, the shortcut symlinks for the `openssl <command>`
20160 are no longer created. This way we have a single and consistent command
20161 line interface `openssl <command>`, similar to `cvs <command>`.
20163 *Ralf S. Engelschall, Paul Sutton and Ben Laurie*
20165 * ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
20166 BIT STRING wrapper always have zero unused bits.
20170 * Add CA.pl, perl version of CA.sh, add extended key usage OID.
20174 * Make the top-level INSTALL documentation easier to understand.
20178 * Makefiles updated to exit if an error occurs in a sub-directory
20179 make (including if user presses ^C) [Paul Sutton]
20181 * Make Montgomery context stuff explicit in RSA data structure.
20185 * Fix build order of pem and err to allow for generated pem.h.
20189 * Fix renumbering bug in X509_NAME_delete_entry().
20193 * Enhanced the err-ins.pl script so it makes the error library number
20194 global and can add a library name. This is needed for external ASN1 and
20195 other error libraries.
20199 * Fixed sk_insert which never worked properly.
20203 * Fix ASN1 macros so they can handle indefinite length constructed
20204 EXPLICIT tags. Some non standard certificates use these: they can now
20209 * Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
20210 into a single doc/ssleay.txt bundle. This way the information is still
20211 preserved but no longer messes up this directory. Now it's new room for
20212 the new set of documentation files.
20214 *Ralf S. Engelschall*
20216 * SETs were incorrectly DER encoded. This was a major pain, because they
20217 shared code with SEQUENCEs, which aren't coded the same. This means that
20218 almost everything to do with SETs or SEQUENCEs has either changed name or
20219 number of arguments.
20221 *Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>*
20223 * Fix test data to work with the above.
20227 * Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
20228 was already fixed by Eric for 0.9.1 it seems.
20230 *Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>*
20232 * Autodetect FreeBSD3.
20236 * Fix various bugs in Configure. This affects the following platforms:
20240 unixware-2.0-pentium
20245 * Eliminate generated files from CVS. Reorder tests to regenerate files
20246 before they are needed.
20250 * Generate Makefile.ssl from Makefile.org (to keep CVS happy).
20254 ### Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
20256 * Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
20257 changed SSLeay to OpenSSL in version strings.
20259 *Ralf S. Engelschall*
20261 * Some fixups to the top-level documents.
20265 * Fixed the nasty bug where rsaref.h was not found under compile-time
20266 because the symlink to include/ was missing.
20268 *Ralf S. Engelschall*
20270 * Incorporated the popular no-RSA/DSA-only patches
20271 which allow to compile an RSA-free SSLeay.
20273 *Andrew Cooke / Interrader Ldt., Ralf S. Engelschall*
20275 * Fixed nasty rehash problem under `make -f Makefile.ssl links`
20276 when "ssleay" is still not found.
20278 *Ralf S. Engelschall*
20280 * Added more platforms to Configure: Cray T3E, HPUX 11,
20282 *Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>*
20284 * Updated the README file.
20286 *Ralf S. Engelschall*
20288 * Added various .cvsignore files in the CVS repository subdirs
20289 to make a "cvs update" really silent.
20291 *Ralf S. Engelschall*
20293 * Recompiled the error-definition header files and added
20294 missing symbols to the Win32 linker tables.
20296 *Ralf S. Engelschall*
20298 * Cleaned up the top-level documents;
20299 o new files: CHANGES and LICENSE
20300 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
20301 o merged COPYRIGHT into LICENSE
20302 o removed obsolete TODO file
20303 o renamed MICROSOFT to INSTALL.W32
20305 *Ralf S. Engelschall*
20307 * Removed dummy files from the 0.9.1b source tree:
20308 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
20309 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
20310 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
20311 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
20312 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
20314 *Ralf S. Engelschall*
20316 * Added various platform portability fixes.
20320 * The Genesis of the OpenSSL rpject:
20321 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
20322 Young and Tim J. Hudson created while they were working for C2Net until
20325 *The OpenSSL Project*
20327 ### Changes between 0.9.0b and 0.9.1b [not released]
20329 * Updated a few CA certificates under certs/
20333 * Changed some BIGNUM api stuff.
20337 * Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
20338 DGUX x86, Linux Alpha, etc.
20342 * New COMP library [crypto/comp/] for SSL Record Layer Compression:
20343 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
20348 * Add -strparse option to asn1pars program which parses nested
20351 *Dr Stephen Henson <shenson@bigfoot.com>*
20353 * Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
20357 * DSA fix for "ca" program.
20361 * Added "-genkey" option to "dsaparam" program.
20365 * Added RIPE MD160 (rmd160) message digest.
20369 * Added -a (all) option to "ssleay version" command.
20373 * Added PLATFORM define which is the id given to Configure.
20377 * Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
20381 * Extended the ASN.1 parser routines.
20385 * Extended BIO routines to support REUSEADDR, seek, tell, etc.
20389 * Added a BN_CTX to the BN library.
20393 * Fixed the weak key values in DES library
20397 * Changed API in EVP library for cipher aliases.
20401 * Added support for RC2/64bit cipher.
20405 * Converted the lhash library to the crypto/mem.c functions.
20409 * Added more recognized ASN.1 object ids.
20413 * Added more RSA padding checks for SSL/TLS.
20417 * Added BIO proxy/filter functionality.
20421 * Added extra_certs to SSL_CTX which can be used
20422 send extra CA certificates to the client in the CA cert chain sending
20423 process. It can be configured with SSL_CTX_add_extra_chain_cert().
20427 * Now Fortezza is denied in the authentication phase because
20428 this is key exchange mechanism is not supported by SSLeay at all.
20432 * Additional PKCS1 checks.
20436 * Support the string "TLSv1" for all TLS v1 ciphers.
20440 * Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
20441 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
20445 * Fixed a few memory leaks.
20449 * Fixed various code and comment typos.
20453 * A minor bug in ssl/s3_clnt.c where there would always be 4 0
20454 bytes sent in the client random.
20456 *Edward Bishop <ebishop@spyglass.com>*
20460 [CVE-2024-0727]: https://www.openssl.org/news/vulnerabilities.html#CVE-2024-0727
20461 [CVE-2023-6237]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-6237
20462 [CVE-2023-6129]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-6129
20463 [CVE-2023-5678]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-5678
20464 [CVE-2023-5363]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-5363
20465 [CVE-2023-4807]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-4807
20466 [CVE-2023-3817]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-3817
20467 [CVE-2023-3446]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-3446
20468 [CVE-2023-2975]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-2975
20469 [RFC 2578 (STD 58), section 3.5]: https://datatracker.ietf.org/doc/html/rfc2578#section-3.5
20470 [CVE-2023-2650]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-2650
20471 [CVE-2023-1255]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-1255
20472 [CVE-2023-0466]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-0466
20473 [CVE-2023-0465]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-0465
20474 [CVE-2023-0464]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-0464
20475 [CVE-2023-0401]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-0401
20476 [CVE-2023-0286]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-0286
20477 [CVE-2023-0217]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-0217
20478 [CVE-2023-0216]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-0216
20479 [CVE-2023-0215]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-0215
20480 [CVE-2022-4450]: https://www.openssl.org/news/vulnerabilities.html#CVE-2022-4450
20481 [CVE-2022-4304]: https://www.openssl.org/news/vulnerabilities.html#CVE-2022-4304
20482 [CVE-2022-4203]: https://www.openssl.org/news/vulnerabilities.html#CVE-2022-4203
20483 [CVE-2022-3996]: https://www.openssl.org/news/vulnerabilities.html#CVE-2022-3996
20484 [CVE-2022-2274]: https://www.openssl.org/news/vulnerabilities.html#CVE-2022-2274
20485 [CVE-2022-2097]: https://www.openssl.org/news/vulnerabilities.html#CVE-2022-2097
20486 [CVE-2020-1971]: https://www.openssl.org/news/vulnerabilities.html#CVE-2020-1971
20487 [CVE-2020-1967]: https://www.openssl.org/news/vulnerabilities.html#CVE-2020-1967
20488 [CVE-2019-1563]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1563
20489 [CVE-2019-1559]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1559
20490 [CVE-2019-1552]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1552
20491 [CVE-2019-1551]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1551
20492 [CVE-2019-1549]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1549
20493 [CVE-2019-1547]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1547
20494 [CVE-2019-1543]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1543
20495 [CVE-2018-5407]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-5407
20496 [CVE-2018-0739]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0739
20497 [CVE-2018-0737]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0737
20498 [CVE-2018-0735]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0735
20499 [CVE-2018-0734]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0734
20500 [CVE-2018-0733]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0733
20501 [CVE-2018-0732]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0732
20502 [CVE-2017-3738]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3738
20503 [CVE-2017-3737]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3737
20504 [CVE-2017-3736]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3736
20505 [CVE-2017-3735]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3735
20506 [CVE-2017-3733]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3733
20507 [CVE-2017-3732]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3732
20508 [CVE-2017-3731]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3731
20509 [CVE-2017-3730]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3730
20510 [CVE-2016-7055]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7055
20511 [CVE-2016-7054]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7054
20512 [CVE-2016-7053]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7053
20513 [CVE-2016-7052]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7052
20514 [CVE-2016-6309]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6309
20515 [CVE-2016-6308]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6308
20516 [CVE-2016-6307]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6307
20517 [CVE-2016-6306]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6306
20518 [CVE-2016-6305]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6305
20519 [CVE-2016-6304]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6304
20520 [CVE-2016-6303]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6303
20521 [CVE-2016-6302]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6302
20522 [CVE-2016-2183]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2183
20523 [CVE-2016-2182]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2182
20524 [CVE-2016-2181]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2181
20525 [CVE-2016-2180]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2180
20526 [CVE-2016-2179]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2179
20527 [CVE-2016-2178]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2178
20528 [CVE-2016-2177]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2177
20529 [CVE-2016-2176]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2176
20530 [CVE-2016-2109]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2109
20531 [CVE-2016-2107]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2107
20532 [CVE-2016-2106]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2106
20533 [CVE-2016-2105]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2105
20534 [CVE-2016-0800]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0800
20535 [CVE-2016-0799]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0799
20536 [CVE-2016-0798]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0798
20537 [CVE-2016-0797]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0797
20538 [CVE-2016-0705]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0705
20539 [CVE-2016-0702]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0702
20540 [CVE-2016-0701]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0701
20541 [CVE-2015-3197]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3197
20542 [CVE-2015-3196]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3196
20543 [CVE-2015-3195]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3195
20544 [CVE-2015-3194]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3194
20545 [CVE-2015-3193]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3193
20546 [CVE-2015-1793]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1793
20547 [CVE-2015-1792]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1792
20548 [CVE-2015-1791]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1791
20549 [CVE-2015-1790]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1790
20550 [CVE-2015-1789]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1789
20551 [CVE-2015-1788]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1788
20552 [CVE-2015-1787]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1787
20553 [CVE-2015-0293]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0293
20554 [CVE-2015-0291]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0291
20555 [CVE-2015-0290]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0290
20556 [CVE-2015-0289]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0289
20557 [CVE-2015-0288]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0288
20558 [CVE-2015-0287]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0287
20559 [CVE-2015-0286]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0286
20560 [CVE-2015-0285]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0285
20561 [CVE-2015-0209]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0209
20562 [CVE-2015-0208]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0208
20563 [CVE-2015-0207]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0207
20564 [CVE-2015-0206]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0206
20565 [CVE-2015-0205]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0205
20566 [CVE-2015-0204]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0204
20567 [CVE-2014-8275]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-8275
20568 [CVE-2014-5139]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-5139
20569 [CVE-2014-3572]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3572
20570 [CVE-2014-3571]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3571
20571 [CVE-2014-3570]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3570
20572 [CVE-2014-3569]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3569
20573 [CVE-2014-3568]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3568
20574 [CVE-2014-3567]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3567
20575 [CVE-2014-3566]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3566
20576 [CVE-2014-3513]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3513
20577 [CVE-2014-3512]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3512
20578 [CVE-2014-3511]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3511
20579 [CVE-2014-3510]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3510
20580 [CVE-2014-3509]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3509
20581 [CVE-2014-3508]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3508
20582 [CVE-2014-3507]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3507
20583 [CVE-2014-3506]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3506
20584 [CVE-2014-3505]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3505
20585 [CVE-2014-3470]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3470
20586 [CVE-2014-0224]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0224
20587 [CVE-2014-0221]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0221
20588 [CVE-2014-0195]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0195
20589 [CVE-2014-0160]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0160
20590 [CVE-2014-0076]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0076
20591 [CVE-2013-6450]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-6450
20592 [CVE-2013-4353]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-4353
20593 [CVE-2013-0169]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-0169
20594 [CVE-2013-0166]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-0166
20595 [CVE-2012-2686]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2686
20596 [CVE-2012-2333]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2333
20597 [CVE-2012-2110]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2110
20598 [CVE-2012-0884]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0884
20599 [CVE-2012-0050]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0050
20600 [CVE-2012-0027]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0027
20601 [CVE-2011-4619]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4619
20602 [CVE-2011-4577]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4577
20603 [CVE-2011-4576]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4576
20604 [CVE-2011-4109]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4109
20605 [CVE-2011-4108]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4108
20606 [CVE-2011-3210]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-3210
20607 [CVE-2011-3207]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-3207
20608 [CVE-2011-0014]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-0014
20609 [CVE-2010-4252]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-4252
20610 [CVE-2010-4180]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-4180
20611 [CVE-2010-3864]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-3864
20612 [CVE-2010-1633]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-1633
20613 [CVE-2010-0740]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-0740
20614 [CVE-2010-0433]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-0433
20615 [CVE-2009-4355]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-4355
20616 [CVE-2009-3555]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-3555
20617 [CVE-2009-3245]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-3245
20618 [CVE-2009-1386]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1386
20619 [CVE-2009-1379]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1379
20620 [CVE-2009-1378]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1378
20621 [CVE-2009-1377]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1377
20622 [CVE-2009-0789]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0789
20623 [CVE-2009-0591]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0591
20624 [CVE-2009-0590]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0590
20625 [CVE-2008-5077]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-5077
20626 [CVE-2008-1678]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-1678
20627 [CVE-2008-1672]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-1672
20628 [CVE-2008-0891]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-0891
20629 [CVE-2007-5135]: https://www.openssl.org/news/vulnerabilities.html#CVE-2007-5135
20630 [CVE-2007-4995]: https://www.openssl.org/news/vulnerabilities.html#CVE-2007-4995
20631 [CVE-2006-4343]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-4343
20632 [CVE-2006-4339]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-4339
20633 [CVE-2006-3738]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-3738
20634 [CVE-2006-2940]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-2940
20635 [CVE-2006-2937]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-2937
20636 [CVE-2005-2969]: https://www.openssl.org/news/vulnerabilities.html#CVE-2005-2969
20637 [CVE-2004-0112]: https://www.openssl.org/news/vulnerabilities.html#CVE-2004-0112
20638 [CVE-2004-0079]: https://www.openssl.org/news/vulnerabilities.html#CVE-2004-0079
20639 [CVE-2003-0851]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0851
20640 [CVE-2003-0545]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0545
20641 [CVE-2003-0544]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0544
20642 [CVE-2003-0543]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0543
20643 [CVE-2003-0078]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0078
20644 [CVE-2002-0659]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0659
20645 [CVE-2002-0657]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0657
20646 [CVE-2002-0656]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0656
20647 [CVE-2002-0655]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0655