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 * In `openssl speed`, changed the default hash function used with `hmac` from
36 * The build of exporters (such as `.pc` files for pkg-config) cleaned up to
37 be less hard coded in the build file templates, and to allow easier
38 addition of more exporters. With that, an exporter for CMake is also
43 * The BLAKE2s hash algorithm matches BLAKE2b's support
44 for configurable output length.
48 * New option `SSL_OP_PREFER_NO_DHE_KEX`, which allows configuring a TLS1.3
49 server to prefer session resumption using PSK-only key exchange over PSK
50 with DHE, if both are available.
52 *Markus Minichmayr, Tapkey GmbH*
57 ### Changes between 3.1 and 3.2 [xx XXX xxxx]
59 * The BLAKE2b hash algorithm supports a configurable output length
60 by setting the "size" parameter.
62 *Čestmír Kalina and Tomáš Mráz*
64 * Enable extra Arm64 optimization on Windows for GHASH, RAND and AES.
68 * Added a function to delete objects from store by URI - OSSL_STORE_delete()
69 and the corresponding provider-storemgmt API function
70 OSSL_FUNC_store_delete().
74 * Added OSSL_FUNC_store_open_ex() provider-storemgmt API function to pass
75 a passphrase callback when opening a store.
79 * Changed the default salt length used by PBES2 KDF's (PBKDF2 and scrypt)
80 from 8 bytes to 16 bytes.
81 The PKCS5 (RFC 8018) standard uses a 64 bit salt length for PBE, and
82 recommends a minimum of 64 bits for PBES2. For FIPS compliance PBKDF2
83 requires a salt length of 128 bits. This affects OpenSSL command line
84 applications such as "genrsa" and "pkcs8" and API's such as
85 PEM_write_bio_PrivateKey() that are reliant on the default value.
86 The additional commandline option 'saltlen' has been added to the
87 OpenSSL command line applications for "pkcs8" and "enc" to allow the
88 salt length to be set to a non default value.
92 * Changed the default value of the `ess_cert_id_alg` configuration
93 option which is used to calculate the TSA's public key certificate
94 identifier. The default algorithm is updated to be sha256 instead
99 * Added optimization for SM2 algorithm on aarch64. It uses a huge precomputed
100 table for point multiplication of the base point, which increases the size of
101 libcrypto from 4.4 MB to 4.9 MB. A new configure option `no-sm2-precomp` has
102 been added to disable the precomputed table.
106 * Added client side support for QUIC
108 *Hugo Landau, Matt Caswell, Paul Dale, Tomáš Mráz, Richard Levitte*
110 * Added multiple tutorials on the OpenSSL library and in particular
111 on writing various clients (using TLS and QUIC protocols) with libssl.
115 * Added secp384r1 implementation using Solinas' reduction to improve
116 speed of the NIST P-384 elliptic curve. To enable the implementation
117 the build option `enable-ec_nistp_64_gcc_128` must be used.
121 * Improved RFC7468 compliance of the asn1parse command.
123 *Matthias St. Pierre*
125 * Added SHA256/192 algorithm support.
129 * Added support for securely getting root CA certificate update in
134 * Improved contention on global write locks by using more read locks where
139 * Improved performance of OSSL_PARAM lookups in performance critical
144 * Added the SSL_get0_group_name() function to provide access to the
145 name of the group used for the TLS key exchange.
149 * Provide a new configure option `no-http` that can be used to disable the
150 HTTP support. Provide new configure options `no-apps` and `no-docs` to
151 disable building the openssl command line application and the documentation.
155 * Provide a new configure option `no-ecx` that can be used to disable the
156 X25519, X448, and EdDSA support.
160 * When multiple OSSL_KDF_PARAM_INFO parameters are passed to
161 the EVP_KDF_CTX_set_params() function they are now concatenated not just
162 for the HKDF algorithm but also for SSKDF and X9.63 KDF algorithms.
166 * Added OSSL_FUNC_keymgmt_im/export_types_ex() provider functions that get
167 the provider context as a parameter.
171 * TLS round-trip time calculation was added by a Brigham Young University
172 Capstone team partnering with Sandia National Laboratories. A new function
173 in ssl_lib titled SSL_get_handshake_rtt will calculate and retrieve this
178 * Added the "-quic" option to s_client to enable connectivity to QUIC servers.
179 QUIC requires the use of ALPN, so this must be specified via the "-alpn"
180 option. Use of the "advanced" s_client command command via the "-adv" option
185 * Added an "advanced" command mode to s_client. Use this with the "-adv"
186 option. The old "basic" command mode recognises certain letters that must
187 always appear at the start of a line and cannot be escaped. The advanced
188 command mode enables commands to be entered anywhere and there is an
189 escaping mechanism. After starting s_client with "-adv" type "{help}"
190 to show a list of available commands.
194 * Add Raw Public Key (RFC7250) support. Authentication is supported
195 by matching keys against either local policy (TLSA records synthesised
196 from the expected keys) or DANE (TLSA records obtained by the
197 application from DNS). TLSA records will also match the same key in
198 the server certificate, should RPK use not happen to be negotiated.
202 * Added support for modular exponentiation and CRT offloading for the
207 * Added further assembler code for the RISC-V architecture.
211 * Added EC_GROUP_to_params() which creates an OSSL_PARAM array
212 from a given EC_GROUP.
216 * Improved support for non-default library contexts and property queries
217 when parsing PKCS#12 files.
221 * Implemented support for all five instances of EdDSA from RFC8032:
222 Ed25519, Ed25519ctx, Ed25519ph, Ed448, and Ed448ph.
223 The streaming is not yet supported for the HashEdDSA variants
224 (Ed25519ph and Ed448ph).
228 * Added SM4 optimization for ARM processors using ASIMD and AES HW
233 * Implemented SM4-XTS support.
237 * Added platform-agnostic OSSL_sleep() function.
241 * Implemented deterministic ECDSA signatures (RFC6979) support.
245 * Implemented AES-GCM-SIV (RFC8452) support.
249 * Added support for pluggable (provider-based) TLS signature algorithms.
250 This enables TLS 1.3 authentication operations with algorithms embedded
251 in providers not included by default in OpenSSL. In combination with
252 the already available pluggable KEM and X.509 support, this enables
253 for example suitable providers to deliver post-quantum or quantum-safe
254 cryptography to OpenSSL users.
258 * Added support for pluggable (provider-based) CMS signature algorithms.
259 This enables CMS sign and verify operations with algorithms embedded
260 in providers not included by default in OpenSSL.
264 * Added support for Hybrid Public Key Encryption (HPKE) as defined
265 in RFC9180. HPKE is required for TLS Encrypted ClientHello (ECH),
266 Message Layer Security (MLS) and other IETF specifications.
267 HPKE can also be used by other applications that require
268 encrypting "to" an ECDH public key. External APIs are defined in
269 include/openssl/hpke.h and documented in doc/man3/OSSL_HPKE_CTX_new.pod
273 * Implemented HPKE DHKEM support in providers used by HPKE (RFC9180)
278 * Add support for certificate compression (RFC8879), including
279 library support for Brotli and Zstandard compression.
283 * Add the ability to add custom attributes to PKCS12 files. Add a new API
284 PKCS12_create_ex2, identical to the existing PKCS12_create_ex but allows
285 for a user specified callback and optional argument.
286 Added a new PKCS12_SAFEBAG_set0_attr, which allows for a new attr to be
287 added to the existing STACK_OF attrs.
291 * Major refactor of the libssl record layer.
295 * Add a mac salt length option for the pkcs12 command.
299 * Add more SRTP protection profiles from RFC8723 and RFC8269.
303 * Extended Kernel TLS (KTLS) to support TLS 1.3 receive offload.
305 *Daiki Ueno, John Baldwin and Dmitry Podgorny*
307 * Add support for TCP Fast Open (RFC7413) to macOS, Linux, and FreeBSD where
308 supported and enabled.
312 * Add ciphersuites based on DHE_PSK (RFC 4279) and ECDHE_PSK (RFC 5489)
313 to the list of ciphersuites providing Perfect Forward Secrecy as
314 required by SECLEVEL >= 3.
316 *Dmitry Belyavskiy, Nicola Tuveri*
318 * Add new SSL APIs to aid in efficiently implementing TLS/SSL fingerprinting.
319 The SSL_CTRL_GET_IANA_GROUPS control code, exposed as the
320 SSL_get0_iana_groups() function-like macro, retrieves the list of
321 supported groups sent by the peer.
322 The function SSL_client_hello_get_extension_order() populates
323 a caller-supplied array with the list of extension types present in the
324 ClientHello, in order of appearance.
328 * Fixed PEM_write_bio_PKCS8PrivateKey() and PEM_write_bio_PKCS8PrivateKey_nid()
329 to make it possible to use empty passphrase strings.
333 * The PKCS12_parse() function now supports MAC-less PKCS12 files.
337 * Added ASYNC_set_mem_functions() and ASYNC_get_mem_functions() calls to be able
338 to change functions used for allocating the memory of asynchronous call stack.
342 * Added support for signed BIGNUMs in the OSSL_PARAM APIs.
346 * A failure exit code is returned when using the openssl x509 command to check
347 certificate attributes and the checks fail.
351 * The default SSL/TLS security level has been changed from 1 to 2. RSA,
352 DSA and DH keys of 1024 bits and above and less than 2048 bits and ECC keys
353 of 160 bits and above and less than 224 bits were previously accepted by
354 default but are now no longer allowed. By default TLS compression was
355 already disabled in previous OpenSSL versions. At security level 2 it cannot
360 * The SSL_CTX_set_cipher_list family functions now accept ciphers using their
365 * The PVK key derivation function has been moved from b2i_PVK_bio_ex() into
366 the legacy crypto provider as an EVP_KDF. Applications requiring this KDF
367 will need to load the legacy crypto provider.
371 * CCM8 cipher suites in TLS have been downgraded to security level zero
372 because they use a short authentication tag which lowers their strength.
376 * Subject or issuer names in X.509 objects are now displayed as UTF-8 strings
381 * Add X.509 certificate codeSigning purpose and related checks on key usage and
382 extended key usage of the leaf certificate according to the CA/Browser Forum.
386 * The `x509`, `ca`, and `req` apps now produce X.509 v3 certificates.
387 The `-x509v1` option of `req` prefers generation of X.509 v1 certificates.
388 `X509_sign()` and `X509_sign_ctx()` make sure that the certificate has
389 X.509 version 3 if the certificate information includes X.509 extensions.
393 * Fix and extend certificate handling and the apps `x509`, `verify` etc.
394 such as adding a trace facility for debugging certificate chain building.
398 * Various fixes and extensions to the CMP+CRMF implementation and the `cmp` app
399 in particular supporting requests for central key generation, generalized
400 polling, and various types of genm/genp exchanges defined in CMP Updates.
404 * Fixes and extensions to the HTTP client and to the HTTP server in `apps/`
405 like correcting the TLS and proxy support and adding tracing for debugging.
409 * Extended the CMS API for handling `CMS_SignedData` and `CMS_EnvelopedData`.
413 * `CMS_add0_cert()` and `CMS_add1_cert()` no longer throw an error if
414 a certificate to be added is already present. `CMS_sign_ex()` and
415 `CMS_sign()` now ignore any duplicate certificates in their `certs` argument
416 and no longer throw an error for them.
420 * Fixed and extended `util/check-format.pl` for checking adherence to the
421 coding style <https://www.openssl.org/policies/technical/coding-style.html>.
422 The checks are meanwhile more complete and yield fewer false positives.
426 * Added BIO_s_dgram_pair() and BIO_s_dgram_mem() that provide memory-based
427 BIOs with datagram semantics and support for BIO_sendmmsg() and BIO_recvmmsg()
428 calls. They can be used as the transport BIOs for QUIC.
430 *Hugo Landau, Matt Caswell and Tomáš Mráz*
432 * Add new BIO_sendmmsg() and BIO_recvmmsg() BIO methods which allow
433 sending and receiving multiple messages in a single call. An implementation
434 is provided for BIO_dgram. For further details, see BIO_sendmmsg(3).
438 * Support for loading root certificates from the Windows certificate store
439 has been added. The support is in the form of a store which recognises the
440 URI string of `org.openssl.winstore://`. This URI scheme currently takes no
441 arguments. This store is built by default and can be disabled using the new
442 compile-time option `no-winstore`. This store is not currently used by
443 default and must be loaded explicitly using the above store URI. It is
444 expected to be loaded by default in the future.
448 * Enable KTLS with the TLS 1.3 CCM mode ciphersuites. Note that some linux
449 kernel versions that support KTLS have a known bug in CCM processing. That
450 has been fixed in stable releases starting from 5.4.164, 5.10.84, 5.15.7,
451 and all releases since 5.16. KTLS with CCM ciphersuites should be only used
456 * Added `-ktls` option to `s_server` and `s_client` commands to enable the
461 * Zerocopy KTLS sendfile() support on Linux.
465 * The OBJ_ calls are now thread safe using a global lock.
469 * New parameter `-digest` for openssl cms command allowing signing
470 pre-computed digests and new CMS API functions supporting that
475 * OPENSSL_malloc() and other allocation functions now raise errors on
476 allocation failures. The callers do not need to explicitly raise errors
477 unless they want to for tracing purposes.
481 * Added and enabled by default implicit rejection in RSA PKCS#1 v1.5
482 decryption as a protection against Bleichenbacher-like attacks.
483 The RSA decryption API will now return a randomly generated deterministic
484 message instead of an error in case it detects an error when checking
485 padding during PKCS#1 v1.5 decryption. This is a general protection against
486 issues like CVE-2020-25659 and CVE-2020-25657. This protection can be
488 `EVP_PKEY_CTX_ctrl_str(ctx, "rsa_pkcs1_implicit_rejection". "0")`
489 on the RSA decryption context.
493 * Added support for Brainpool curves in TLS-1.3.
495 *Bernd Edlinger and Matt Caswell*
497 * Added OpenBSD specific build targets.
501 * Support for Argon2d, Argon2i, Argon2id KDFs has been added along with
502 a basic thread pool implementation for select platforms.
509 ### Changes between 3.1.4 and 3.1.5 [xx XXX xxxx]
511 * Fix excessive time spent in DH check / generation with large Q parameter
514 Applications that use the functions DH_generate_key() to generate an
515 X9.42 DH key may experience long delays. Likewise, applications that use
516 DH_check_pub_key(), DH_check_pub_key_ex() or EVP_PKEY_public_check()
517 to check an X9.42 DH key or X9.42 DH parameters may experience long delays.
518 Where the key or parameters that are being checked have been obtained from
519 an untrusted source this may lead to a Denial of Service.
525 ### Changes between 3.1.3 and 3.1.4 [24 Oct 2023]
527 * Fix incorrect key and IV resizing issues when calling EVP_EncryptInit_ex2(),
528 EVP_DecryptInit_ex2() or EVP_CipherInit_ex2() with OSSL_PARAM parameters
529 that alter the key or IV length ([CVE-2023-5363]).
533 ### Changes between 3.1.2 and 3.1.3 [19 Sep 2023]
535 * Fix POLY1305 MAC implementation corrupting XMM registers on Windows.
537 The POLY1305 MAC (message authentication code) implementation in OpenSSL
538 does not save the contents of non-volatile XMM registers on Windows 64
539 platform when calculating the MAC of data larger than 64 bytes. Before
540 returning to the caller all the XMM registers are set to zero rather than
541 restoring their previous content. The vulnerable code is used only on newer
542 x86_64 processors supporting the AVX512-IFMA instructions.
544 The consequences of this kind of internal application state corruption can
545 be various - from no consequences, if the calling application does not
546 depend on the contents of non-volatile XMM registers at all, to the worst
547 consequences, where the attacker could get complete control of the
548 application process. However given the contents of the registers are just
549 zeroized so the attacker cannot put arbitrary values inside, the most likely
550 consequence, if any, would be an incorrect result of some application
551 dependent calculations or a crash leading to a denial of service.
557 ### Changes between 3.1.1 and 3.1.2 [1 Aug 2023]
559 * Fix excessive time spent checking DH q parameter value.
561 The function DH_check() performs various checks on DH parameters. After
562 fixing CVE-2023-3446 it was discovered that a large q parameter value can
563 also trigger an overly long computation during some of these checks.
564 A correct q value, if present, cannot be larger than the modulus p
565 parameter, thus it is unnecessary to perform these checks if q is larger
568 If DH_check() is called with such q parameter value,
569 DH_CHECK_INVALID_Q_VALUE return flag is set and the computationally
570 intensive checks are skipped.
576 * Fix DH_check() excessive time with over sized modulus.
578 The function DH_check() performs various checks on DH parameters. One of
579 those checks confirms that the modulus ("p" parameter) is not too large.
580 Trying to use a very large modulus is slow and OpenSSL will not normally use
581 a modulus which is over 10,000 bits in length.
583 However the DH_check() function checks numerous aspects of the key or
584 parameters that have been supplied. Some of those checks use the supplied
585 modulus value even if it has already been found to be too large.
587 A new limit has been added to DH_check of 32,768 bits. Supplying a
588 key/parameters with a modulus over this size will simply cause DH_check() to
595 * Do not ignore empty associated data entries with AES-SIV.
597 The AES-SIV algorithm allows for authentication of multiple associated
598 data entries along with the encryption. To authenticate empty data the
599 application has to call `EVP_EncryptUpdate()` (or `EVP_CipherUpdate()`)
600 with NULL pointer as the output buffer and 0 as the input buffer length.
601 The AES-SIV implementation in OpenSSL just returns success for such call
602 instead of performing the associated data authentication operation.
603 The empty data thus will not be authenticated. ([CVE-2023-2975])
605 Thanks to Juerg Wullschleger (Google) for discovering the issue.
607 The fix changes the authentication tag value and the ciphertext for
608 applications that use empty associated data entries with AES-SIV.
609 To decrypt data encrypted with previous versions of OpenSSL the application
610 has to skip calls to `EVP_DecryptUpdate()` for empty associated data
615 * When building with the `enable-fips` option and using the resulting
616 FIPS provider, TLS 1.2 will, by default, mandate the use of an extended
617 master secret (FIPS 140-3 IG G.Q) and the Hash and HMAC DRBGs will
618 not operate with truncated digests (FIPS 140-3 IG G.R).
622 ### Changes between 3.1.0 and 3.1.1 [30 May 2023]
624 * Mitigate for the time it takes for `OBJ_obj2txt` to translate gigantic
625 OBJECT IDENTIFIER sub-identifiers to canonical numeric text form.
627 OBJ_obj2txt() would translate any size OBJECT IDENTIFIER to canonical
628 numeric text form. For gigantic sub-identifiers, this would take a very
629 long time, the time complexity being O(n^2) where n is the size of that
630 sub-identifier. ([CVE-2023-2650])
632 To mitigitate this, `OBJ_obj2txt()` will only translate an OBJECT
633 IDENTIFIER to canonical numeric text form if the size of that OBJECT
634 IDENTIFIER is 586 bytes or less, and fail otherwise.
636 The basis for this restriction is [RFC 2578 (STD 58), section 3.5]. OBJECT
637 IDENTIFIER values, which stipulates that OBJECT IDENTIFIERS may have at
638 most 128 sub-identifiers, and that the maximum value that each sub-
639 identifier may have is 2^32-1 (4294967295 decimal).
641 For each byte of every sub-identifier, only the 7 lower bits are part of
642 the value, so the maximum amount of bytes that an OBJECT IDENTIFIER with
643 these restrictions may occupy is 32 * 128 / 7, which is approximately 586
648 * Multiple algorithm implementation fixes for ARM BE platforms.
652 * Added a -pedantic option to fipsinstall that adjusts the various
653 settings to ensure strict FIPS compliance rather than backwards
658 * Fixed buffer overread in AES-XTS decryption on ARM 64 bit platforms which
659 happens if the buffer size is 4 mod 5 in 16 byte AES blocks. This can
660 trigger a crash of an application using AES-XTS decryption if the memory
661 just after the buffer being decrypted is not mapped.
662 Thanks to Anton Romanov (Amazon) for discovering the issue.
667 * Reworked the Fix for the Timing Oracle in RSA Decryption ([CVE-2022-4304]).
668 The previous fix for this timing side channel turned out to cause
669 a severe 2-3x performance regression in the typical use case
670 compared to 3.0.7. The new fix uses existing constant time
671 code paths, and restores the previous performance level while
672 fully eliminating all existing timing side channels.
673 The fix was developed by Bernd Edlinger with testing support
678 * Add FIPS provider configuration option to disallow the use of
679 truncated digests with Hash and HMAC DRBGs (q.v. FIPS 140-3 IG D.R.).
680 The option '-no_drbg_truncated_digests' can optionally be
681 supplied to 'openssl fipsinstall'.
685 * Corrected documentation of X509_VERIFY_PARAM_add0_policy() to mention
686 that it does not enable policy checking. Thanks to David Benjamin for
687 discovering this issue.
692 * Fixed an issue where invalid certificate policies in leaf certificates are
693 silently ignored by OpenSSL and other certificate policy checks are skipped
694 for that certificate. A malicious CA could use this to deliberately assert
695 invalid certificate policies in order to circumvent policy checking on the
696 certificate altogether.
701 * Limited the number of nodes created in a policy tree to mitigate
702 against CVE-2023-0464. The default limit is set to 1000 nodes, which
703 should be sufficient for most installations. If required, the limit
704 can be adjusted by setting the OPENSSL_POLICY_TREE_NODES_MAX build
705 time define to a desired maximum number of nodes or zero to allow
711 ### Changes between 3.0 and 3.1.0 [14 Mar 2023]
713 * Add FIPS provider configuration option to enforce the
714 Extended Master Secret (EMS) check during the TLS1_PRF KDF.
715 The option '-ems-check' can optionally be supplied to
716 'openssl fipsinstall'.
720 * The FIPS provider includes a few non-approved algorithms for
721 backward compatibility purposes and the "fips=yes" property query
722 must be used for all algorithm fetches to ensure FIPS compliance.
724 The algorithms that are included but not approved are Triple DES ECB,
725 Triple DES CBC and EdDSA.
729 * Added support for KMAC in KBKDF.
733 * RNDR and RNDRRS support in provider functions to provide
734 random number generation for Arm CPUs (aarch64).
738 * s_client and s_server apps now explicitly say when the TLS version
739 does not include the renegotiation mechanism. This avoids confusion
740 between that scenario versus when the TLS version includes secure
741 renegotiation but the peer lacks support for it.
745 * AES-GCM enabled with AVX512 vAES and vPCLMULQDQ.
747 *Tomasz Kantecki, Andrey Matyukov*
749 * The various OBJ_* functions have been made thread safe.
753 * Parallel dual-prime 1536/2048-bit modular exponentiation for
754 AVX512_IFMA capable processors.
756 *Sergey Kirillov, Andrey Matyukov (Intel Corp)*
758 * The functions `OPENSSL_LH_stats`, `OPENSSL_LH_node_stats`,
759 `OPENSSL_LH_node_usage_stats`, `OPENSSL_LH_stats_bio`,
760 `OPENSSL_LH_node_stats_bio` and `OPENSSL_LH_node_usage_stats_bio` are now
761 marked deprecated from OpenSSL 3.1 onwards and can be disabled by defining
762 `OPENSSL_NO_DEPRECATED_3_1`.
764 The macro `DEFINE_LHASH_OF` is now deprecated in favour of the macro
765 `DEFINE_LHASH_OF_EX`, which omits the corresponding type-specific function
766 definitions for these functions regardless of whether
767 `OPENSSL_NO_DEPRECATED_3_1` is defined.
769 Users of `DEFINE_LHASH_OF` may start receiving deprecation warnings for these
770 functions regardless of whether they are using them. It is recommended that
771 users transition to the new macro, `DEFINE_LHASH_OF_EX`.
775 * When generating safe-prime DH parameters set the recommended private key
776 length equivalent to minimum key lengths as in RFC 7919.
780 * Change the default salt length for PKCS#1 RSASSA-PSS signatures to the
781 maximum size that is smaller or equal to the digest length to comply with
782 FIPS 186-4 section 5. This is implemented by a new option
783 `OSSL_PKEY_RSA_PSS_SALT_LEN_AUTO_DIGEST_MAX` ("auto-digestmax") for the
784 `rsa_pss_saltlen` parameter, which is now the default. Signature
785 verification is not affected by this change and continues to work as before.
792 For OpenSSL 3.0 a [Migration guide][] has been added, so the CHANGES entries
793 listed here are only a brief description.
794 The migration guide contains more detailed information related to new features,
795 breaking changes, and mappings for the large list of deprecated functions.
797 [Migration guide]: https://github.com/openssl/openssl/tree/master/doc/man7/migration_guide.pod
799 ### Changes between 3.0.7 and 3.0.8 [7 Feb 2023]
801 * Fixed NULL dereference during PKCS7 data verification.
803 A NULL pointer can be dereferenced when signatures are being
804 verified on PKCS7 signed or signedAndEnveloped data. In case the hash
805 algorithm used for the signature is known to the OpenSSL library but
806 the implementation of the hash algorithm is not available the digest
807 initialization will fail. There is a missing check for the return
808 value from the initialization function which later leads to invalid
809 usage of the digest API most likely leading to a crash.
812 PKCS7 data is processed by the SMIME library calls and also by the
813 time stamp (TS) library calls. The TLS implementation in OpenSSL does
814 not call these functions however third party applications would be
815 affected if they call these functions to verify signatures on untrusted
820 * Fixed X.400 address type confusion in X.509 GeneralName.
822 There is a type confusion vulnerability relating to X.400 address processing
823 inside an X.509 GeneralName. X.400 addresses were parsed as an ASN1_STRING
824 but the public structure definition for GENERAL_NAME incorrectly specified
825 the type of the x400Address field as ASN1_TYPE. This field is subsequently
826 interpreted by the OpenSSL function GENERAL_NAME_cmp as an ASN1_TYPE rather
829 When CRL checking is enabled (i.e. the application sets the
830 X509_V_FLAG_CRL_CHECK flag), this vulnerability may allow an attacker to
831 pass arbitrary pointers to a memcmp call, enabling them to read memory
832 contents or enact a denial of service.
837 * Fixed NULL dereference validating DSA public key.
839 An invalid pointer dereference on read can be triggered when an
840 application tries to check a malformed DSA public key by the
841 EVP_PKEY_public_check() function. This will most likely lead
842 to an application crash. This function can be called on public
843 keys supplied from untrusted sources which could allow an attacker
844 to cause a denial of service attack.
846 The TLS implementation in OpenSSL does not call this function
847 but applications might call the function if there are additional
848 security requirements imposed by standards such as FIPS 140-3.
851 *Shane Lontis, Tomáš Mráz*
853 * Fixed Invalid pointer dereference in d2i_PKCS7 functions.
855 An invalid pointer dereference on read can be triggered when an
856 application tries to load malformed PKCS7 data with the
857 d2i_PKCS7(), d2i_PKCS7_bio() or d2i_PKCS7_fp() functions.
859 The result of the dereference is an application crash which could
860 lead to a denial of service attack. The TLS implementation in OpenSSL
861 does not call this function however third party applications might
862 call these functions on untrusted data.
867 * Fixed Use-after-free following BIO_new_NDEF.
869 The public API function BIO_new_NDEF is a helper function used for
870 streaming ASN.1 data via a BIO. It is primarily used internally to OpenSSL
871 to support the SMIME, CMS and PKCS7 streaming capabilities, but may also
872 be called directly by end user applications.
874 The function receives a BIO from the caller, prepends a new BIO_f_asn1
875 filter BIO onto the front of it to form a BIO chain, and then returns
876 the new head of the BIO chain to the caller. Under certain conditions,
877 for example if a CMS recipient public key is invalid, the new filter BIO
878 is freed and the function returns a NULL result indicating a failure.
879 However, in this case, the BIO chain is not properly cleaned up and the
880 BIO passed by the caller still retains internal pointers to the previously
881 freed filter BIO. If the caller then goes on to call BIO_pop() on the BIO
882 then a use-after-free will occur. This will most likely result in a crash.
885 *Viktor Dukhovni, Matt Caswell*
887 * Fixed Double free after calling PEM_read_bio_ex.
889 The function PEM_read_bio_ex() reads a PEM file from a BIO and parses and
890 decodes the "name" (e.g. "CERTIFICATE"), any header data and the payload
891 data. If the function succeeds then the "name_out", "header" and "data"
892 arguments are populated with pointers to buffers containing the relevant
893 decoded data. The caller is responsible for freeing those buffers. It is
894 possible to construct a PEM file that results in 0 bytes of payload data.
895 In this case PEM_read_bio_ex() will return a failure code but will populate
896 the header argument with a pointer to a buffer that has already been freed.
897 If the caller also frees this buffer then a double free will occur. This
898 will most likely lead to a crash.
900 The functions PEM_read_bio() and PEM_read() are simple wrappers around
901 PEM_read_bio_ex() and therefore these functions are also directly affected.
903 These functions are also called indirectly by a number of other OpenSSL
904 functions including PEM_X509_INFO_read_bio_ex() and
905 SSL_CTX_use_serverinfo_file() which are also vulnerable. Some OpenSSL
906 internal uses of these functions are not vulnerable because the caller does
907 not free the header argument if PEM_read_bio_ex() returns a failure code.
910 *Kurt Roeckx, Matt Caswell*
912 * Fixed Timing Oracle in RSA Decryption.
914 A timing based side channel exists in the OpenSSL RSA Decryption
915 implementation which could be sufficient to recover a plaintext across
916 a network in a Bleichenbacher style attack. To achieve a successful
917 decryption an attacker would have to be able to send a very large number
918 of trial messages for decryption. The vulnerability affects all RSA padding
919 modes: PKCS#1 v1.5, RSA-OEAP and RSASVE.
922 *Dmitry Belyavsky, Hubert Kario*
924 * Fixed X.509 Name Constraints Read Buffer Overflow.
926 A read buffer overrun can be triggered in X.509 certificate verification,
927 specifically in name constraint checking. The read buffer overrun might
928 result in a crash which could lead to a denial of service attack.
929 In a TLS client, this can be triggered by connecting to a malicious
930 server. In a TLS server, this can be triggered if the server requests
931 client authentication and a malicious client connects.
936 * Fixed X.509 Policy Constraints Double Locking security issue.
938 If an X.509 certificate contains a malformed policy constraint and
939 policy processing is enabled, then a write lock will be taken twice
940 recursively. On some operating systems (most widely: Windows) this
941 results in a denial of service when the affected process hangs. Policy
942 processing being enabled on a publicly facing server is not considered
943 to be a common setup.
948 * Our provider implementations of `OSSL_FUNC_KEYMGMT_EXPORT` and
949 `OSSL_FUNC_KEYMGMT_GET_PARAMS` for EC and SM2 keys now honor
950 `OSSL_PKEY_PARAM_EC_POINT_CONVERSION_FORMAT` as set (and
951 default to `POINT_CONVERSION_UNCOMPRESSED`) when exporting
952 `OSSL_PKEY_PARAM_PUB_KEY`, instead of unconditionally using
953 `POINT_CONVERSION_COMPRESSED` as in previous 3.x releases.
954 For symmetry, our implementation of `EVP_PKEY_ASN1_METHOD->export_to`
955 for legacy EC and SM2 keys is also changed similarly to honor the
956 equivalent conversion format flag as specified in the underlying
957 `EC_KEY` object being exported to a provider, when this function is
958 called through `EVP_PKEY_export()`.
962 ### Changes between 3.0.6 and 3.0.7 [1 Nov 2022]
964 * Fixed two buffer overflows in punycode decoding functions.
966 A buffer overrun can be triggered in X.509 certificate verification,
967 specifically in name constraint checking. Note that this occurs after
968 certificate chain signature verification and requires either a CA to
969 have signed the malicious certificate or for the application to continue
970 certificate verification despite failure to construct a path to a trusted
973 In a TLS client, this can be triggered by connecting to a malicious
974 server. In a TLS server, this can be triggered if the server requests
975 client authentication and a malicious client connects.
977 An attacker can craft a malicious email address to overflow
978 an arbitrary number of bytes containing the `.` character (decimal 46)
979 on the stack. This buffer overflow could result in a crash (causing a
983 An attacker can craft a malicious email address to overflow four
984 attacker-controlled bytes on the stack. This buffer overflow could
985 result in a crash (causing a denial of service) or potentially remote code
986 execution depending on stack layout for any given platform/compiler.
991 * Removed all references to invalid OSSL_PKEY_PARAM_RSA names for CRT
992 parameters in OpenSSL code.
993 Applications should not use the names OSSL_PKEY_PARAM_RSA_FACTOR,
994 OSSL_PKEY_PARAM_RSA_EXPONENT and OSSL_PKEY_PARAM_RSA_COEFFICIENT.
995 Use the numbered names such as OSSL_PKEY_PARAM_RSA_FACTOR1 instead.
996 Using these invalid names may cause algorithms to use slower methods
997 that ignore the CRT parameters.
1001 * Fixed a regression introduced in 3.0.6 version raising errors on some stack
1006 * Fixed a regression introduced in 3.0.6 version not refreshing the certificate
1007 data to be signed before signing the certificate.
1011 * Added RIPEMD160 to the default provider.
1015 * Ensured that the key share group sent or accepted for the key exchange
1016 is allowed for the protocol version.
1020 ### Changes between 3.0.5 and 3.0.6 [11 Oct 2022]
1022 * OpenSSL supports creating a custom cipher via the legacy
1023 EVP_CIPHER_meth_new() function and associated function calls. This function
1024 was deprecated in OpenSSL 3.0 and application authors are instead encouraged
1025 to use the new provider mechanism in order to implement custom ciphers.
1027 OpenSSL versions 3.0.0 to 3.0.5 incorrectly handle legacy custom ciphers
1028 passed to the EVP_EncryptInit_ex2(), EVP_DecryptInit_ex2() and
1029 EVP_CipherInit_ex2() functions (as well as other similarly named encryption
1030 and decryption initialisation functions). Instead of using the custom cipher
1031 directly it incorrectly tries to fetch an equivalent cipher from the
1032 available providers. An equivalent cipher is found based on the NID passed to
1033 EVP_CIPHER_meth_new(). This NID is supposed to represent the unique NID for a
1034 given cipher. However it is possible for an application to incorrectly pass
1035 NID_undef as this value in the call to EVP_CIPHER_meth_new(). When NID_undef
1036 is used in this way the OpenSSL encryption/decryption initialisation function
1037 will match the NULL cipher as being equivalent and will fetch this from the
1038 available providers. This will succeed if the default provider has been
1039 loaded (or if a third party provider has been loaded that offers this
1040 cipher). Using the NULL cipher means that the plaintext is emitted as the
1043 Applications are only affected by this issue if they call
1044 EVP_CIPHER_meth_new() using NID_undef and subsequently use it in a call to an
1045 encryption/decryption initialisation function. Applications that only use
1046 SSL/TLS are not impacted by this issue.
1051 * Fix LLVM vs Apple LLVM version numbering confusion that caused build failures
1056 * Fixed the linux-mips64 Configure target which was missing the
1057 SIXTY_FOUR_BIT bn_ops flag. This was causing heap corruption on that
1062 * Fix handling of a ticket key callback that returns 0 in TLSv1.3 to not send a
1067 * Correctly handle a retransmitted ClientHello in DTLS
1071 * Fixed detection of ktls support in cross-compile environment on Linux
1075 * Fixed some regressions and test failures when running the 3.0.0 FIPS provider
1080 * Fixed SSL_pending() and SSL_has_pending() with DTLS which were failing to
1081 report correct results in some cases
1085 * Fix UWP builds by defining VirtualLock
1089 * For known safe primes use the minimum key length according to RFC 7919.
1090 Longer private key sizes unnecessarily raise the cycles needed to compute the
1091 shared secret without any increase of the real security. This fixes a
1092 regression from 1.1.1 where these shorter keys were generated for the known
1097 * Added the loongarch64 target
1101 * Fixed EC ASM flag passing. Flags for ASM implementations of EC curves were
1102 only passed to the FIPS provider and not to the default or legacy provider.
1106 * Fixed reported performance degradation on aarch64. Restored the
1107 implementation prior to commit 2621751 ("aes/asm/aesv8-armx.pl: avoid
1108 32-bit lane assignment in CTR mode") for 64bit targets only, since it is
1109 reportedly 2-17% slower and the silicon errata only affects 32bit targets.
1110 The new algorithm is still used for 32 bit targets.
1114 * Added a missing header for memcmp that caused compilation failure on some
1119 ### Changes between 3.0.4 and 3.0.5 [5 Jul 2022]
1121 * The OpenSSL 3.0.4 release introduced a serious bug in the RSA
1122 implementation for X86_64 CPUs supporting the AVX512IFMA instructions.
1123 This issue makes the RSA implementation with 2048 bit private keys
1124 incorrect on such machines and memory corruption will happen during
1125 the computation. As a consequence of the memory corruption an attacker
1126 may be able to trigger a remote code execution on the machine performing
1129 SSL/TLS servers or other servers using 2048 bit RSA private keys running
1130 on machines supporting AVX512IFMA instructions of the X86_64 architecture
1131 are affected by this issue.
1136 * AES OCB mode for 32-bit x86 platforms using the AES-NI assembly optimised
1137 implementation would not encrypt the entirety of the data under some
1138 circumstances. This could reveal sixteen bytes of data that was
1139 preexisting in the memory that wasn't written. In the special case of
1140 "in place" encryption, sixteen bytes of the plaintext would be revealed.
1142 Since OpenSSL does not support OCB based cipher suites for TLS and DTLS,
1143 they are both unaffected.
1146 *Alex Chernyakhovsky, David Benjamin, Alejandro Sedeño*
1148 ### Changes between 3.0.3 and 3.0.4 [21 Jun 2022]
1150 * In addition to the c_rehash shell command injection identified in
1151 CVE-2022-1292, further bugs where the c_rehash script does not
1152 properly sanitise shell metacharacters to prevent command injection have been
1155 When the CVE-2022-1292 was fixed it was not discovered that there
1156 are other places in the script where the file names of certificates
1157 being hashed were possibly passed to a command executed through the shell.
1159 This script is distributed by some operating systems in a manner where
1160 it is automatically executed. On such operating systems, an attacker
1161 could execute arbitrary commands with the privileges of the script.
1163 Use of the c_rehash script is considered obsolete and should be replaced
1164 by the OpenSSL rehash command line tool.
1167 *Daniel Fiala, Tomáš Mráz*
1169 * Case insensitive string comparison no longer uses locales. It has instead
1170 been directly implemented.
1174 ### Changes between 3.0.2 and 3.0.3 [3 May 2022]
1176 * Case insensitive string comparison is reimplemented via new locale-agnostic
1177 comparison functions OPENSSL_str[n]casecmp always using the POSIX locale for
1178 comparison. The previous implementation had problems when the Turkish locale
1183 * Fixed a bug in the c_rehash script which was not properly sanitising shell
1184 metacharacters to prevent command injection. This script is distributed by
1185 some operating systems in a manner where it is automatically executed. On
1186 such operating systems, an attacker could execute arbitrary commands with the
1187 privileges of the script.
1189 Use of the c_rehash script is considered obsolete and should be replaced
1190 by the OpenSSL rehash command line tool.
1195 * Fixed a bug in the function `OCSP_basic_verify` that verifies the signer
1196 certificate on an OCSP response. The bug caused the function in the case
1197 where the (non-default) flag OCSP_NOCHECKS is used to return a postivie
1198 response (meaning a successful verification) even in the case where the
1199 response signing certificate fails to verify.
1201 It is anticipated that most users of `OCSP_basic_verify` will not use the
1202 OCSP_NOCHECKS flag. In this case the `OCSP_basic_verify` function will return
1203 a negative value (indicating a fatal error) in the case of a certificate
1204 verification failure. The normal expected return value in this case would be
1207 This issue also impacts the command line OpenSSL "ocsp" application. When
1208 verifying an ocsp response with the "-no_cert_checks" option the command line
1209 application will report that the verification is successful even though it
1210 has in fact failed. In this case the incorrect successful response will also
1211 be accompanied by error messages showing the failure and contradicting the
1212 apparently successful result.
1217 * Fixed a bug where the RC4-MD5 ciphersuite incorrectly used the
1218 AAD data as the MAC key. This made the MAC key trivially predictable.
1220 An attacker could exploit this issue by performing a man-in-the-middle attack
1221 to modify data being sent from one endpoint to an OpenSSL 3.0 recipient such
1222 that the modified data would still pass the MAC integrity check.
1224 Note that data sent from an OpenSSL 3.0 endpoint to a non-OpenSSL 3.0
1225 endpoint will always be rejected by the recipient and the connection will
1226 fail at that point. Many application protocols require data to be sent from
1227 the client to the server first. Therefore, in such a case, only an OpenSSL
1228 3.0 server would be impacted when talking to a non-OpenSSL 3.0 client.
1230 If both endpoints are OpenSSL 3.0 then the attacker could modify data being
1231 sent in both directions. In this case both clients and servers could be
1232 affected, regardless of the application protocol.
1234 Note that in the absence of an attacker this bug means that an OpenSSL 3.0
1235 endpoint communicating with a non-OpenSSL 3.0 endpoint will fail to complete
1236 the handshake when using this ciphersuite.
1238 The confidentiality of data is not impacted by this issue, i.e. an attacker
1239 cannot decrypt data that has been encrypted using this ciphersuite - they can
1242 In order for this attack to work both endpoints must legitimately negotiate
1243 the RC4-MD5 ciphersuite. This ciphersuite is not compiled by default in
1244 OpenSSL 3.0, and is not available within the default provider or the default
1245 ciphersuite list. This ciphersuite will never be used if TLSv1.3 has been
1246 negotiated. In order for an OpenSSL 3.0 endpoint to use this ciphersuite the
1247 following must have occurred:
1249 1) OpenSSL must have been compiled with the (non-default) compile time option
1250 enable-weak-ssl-ciphers
1252 2) OpenSSL must have had the legacy provider explicitly loaded (either
1253 through application code or via configuration)
1255 3) The ciphersuite must have been explicitly added to the ciphersuite list
1257 4) The libssl security level must have been set to 0 (default is 1)
1259 5) A version of SSL/TLS below TLSv1.3 must have been negotiated
1261 6) Both endpoints must negotiate the RC4-MD5 ciphersuite in preference to any
1262 others that both endpoints have in common
1267 * Fix a bug in the OPENSSL_LH_flush() function that breaks reuse of the memory
1268 occupied by the removed hash table entries.
1270 This function is used when decoding certificates or keys. If a long lived
1271 process periodically decodes certificates or keys its memory usage will
1272 expand without bounds and the process might be terminated by the operating
1273 system causing a denial of service. Also traversing the empty hash table
1274 entries will take increasingly more time.
1276 Typically such long lived processes might be TLS clients or TLS servers
1277 configured to accept client certificate authentication.
1280 *Hugo Landau, Aliaksei Levin*
1282 * The functions `OPENSSL_LH_stats` and `OPENSSL_LH_stats_bio` now only report
1283 the `num_items`, `num_nodes` and `num_alloc_nodes` statistics. All other
1284 statistics are no longer supported. For compatibility, these statistics are
1285 still listed in the output but are now always reported as zero.
1289 ### Changes between 3.0.1 and 3.0.2 [15 Mar 2022]
1291 * Fixed a bug in the BN_mod_sqrt() function that can cause it to loop forever
1292 for non-prime moduli.
1294 Internally this function is used when parsing certificates that contain
1295 elliptic curve public keys in compressed form or explicit elliptic curve
1296 parameters with a base point encoded in compressed form.
1298 It is possible to trigger the infinite loop by crafting a certificate that
1299 has invalid explicit curve parameters.
1301 Since certificate parsing happens prior to verification of the certificate
1302 signature, any process that parses an externally supplied certificate may thus
1303 be subject to a denial of service attack. The infinite loop can also be
1304 reached when parsing crafted private keys as they can contain explicit
1305 elliptic curve parameters.
1307 Thus vulnerable situations include:
1309 - TLS clients consuming server certificates
1310 - TLS servers consuming client certificates
1311 - Hosting providers taking certificates or private keys from customers
1312 - Certificate authorities parsing certification requests from subscribers
1313 - Anything else which parses ASN.1 elliptic curve parameters
1315 Also any other applications that use the BN_mod_sqrt() where the attacker
1316 can control the parameter values are vulnerable to this DoS issue.
1321 * Add ciphersuites based on DHE_PSK (RFC 4279) and ECDHE_PSK (RFC 5489)
1322 to the list of ciphersuites providing Perfect Forward Secrecy as
1323 required by SECLEVEL >= 3.
1325 *Dmitry Belyavskiy, Nicola Tuveri*
1327 * Made the AES constant time code for no-asm configurations
1328 optional due to the resulting 95% performance degradation.
1329 The AES constant time code can be enabled, for no assembly
1330 builds, with: ./config no-asm -DOPENSSL_AES_CONST_TIME
1334 * Fixed PEM_write_bio_PKCS8PrivateKey() to make it possible to use empty
1339 * The negative return value handling of the certificate verification callback
1340 was reverted. The replacement is to set the verification retry state with
1341 the SSL_set_retry_verify() function.
1345 ### Changes between 3.0.0 and 3.0.1 [14 Dec 2021]
1347 * Fixed invalid handling of X509_verify_cert() internal errors in libssl
1348 Internally libssl in OpenSSL calls X509_verify_cert() on the client side to
1349 verify a certificate supplied by a server. That function may return a
1350 negative return value to indicate an internal error (for example out of
1351 memory). Such a negative return value is mishandled by OpenSSL and will cause
1352 an IO function (such as SSL_connect() or SSL_do_handshake()) to not indicate
1353 success and a subsequent call to SSL_get_error() to return the value
1354 SSL_ERROR_WANT_RETRY_VERIFY. This return value is only supposed to be
1355 returned by OpenSSL if the application has previously called
1356 SSL_CTX_set_cert_verify_callback(). Since most applications do not do this
1357 the SSL_ERROR_WANT_RETRY_VERIFY return value from SSL_get_error() will be
1358 totally unexpected and applications may not behave correctly as a result. The
1359 exact behaviour will depend on the application but it could result in
1360 crashes, infinite loops or other similar incorrect responses.
1362 This issue is made more serious in combination with a separate bug in OpenSSL
1363 3.0 that will cause X509_verify_cert() to indicate an internal error when
1364 processing a certificate chain. This will occur where a certificate does not
1365 include the Subject Alternative Name extension but where a Certificate
1366 Authority has enforced name constraints. This issue can occur even with valid
1372 * Corrected a few file name and file reference bugs in the build,
1373 installation and setup scripts, which lead to installation verification
1374 failures. Slightly enhanced the installation verification script.
1378 * Fixed EVP_PKEY_eq() to make it possible to use it with strictly private
1383 * Fixed PVK encoder to properly query for the passphrase.
1387 * Multiple fixes in the OSSL_HTTP API functions.
1391 * Allow sign extension in OSSL_PARAM_allocate_from_text() for the
1392 OSSL_PARAM_INTEGER data type and return error on negative numbers
1393 used with the OSSL_PARAM_UNSIGNED_INTEGER data type. Make
1394 OSSL_PARAM_BLD_push_BN{,_pad}() return an error on negative numbers.
1398 * Allow copying uninitialized digest contexts with EVP_MD_CTX_copy_ex.
1402 * Fixed detection of ARMv7 and ARM64 CPU features on FreeBSD.
1406 * Multiple threading fixes.
1410 * Added NULL digest implementation to keep compatibility with 1.1.1 version.
1414 * Allow fetching an operation from the provider that owns an unexportable key
1415 as a fallback if that is still allowed by the property query.
1419 ### Changes between 1.1.1 and 3.0.0 [7 Sep 2021]
1421 * TLS_MAX_VERSION, DTLS_MAX_VERSION and DTLS_MIN_VERSION constants are now
1426 * The `OPENSSL_s390xcap` environment variable can be used to set bits in the
1427 S390X capability vector to zero. This simplifies testing of different code
1428 paths on S390X architecture.
1432 * Encrypting more than 2^64 TLS records with AES-GCM is disallowed
1433 as per FIPS 140-2 IG A.5 "Key/IV Pair Uniqueness Requirements from
1434 SP 800-38D". The communication will fail at this point.
1438 * The EC_GROUP_clear_free() function is deprecated as there is nothing
1439 confidential in EC_GROUP data.
1443 * The byte order mark (BOM) character is ignored if encountered at the
1444 beginning of a PEM-formatted file.
1448 * Added CMS support for the Russian GOST algorithms.
1452 * Due to move of the implementation of cryptographic operations
1453 to the providers, validation of various operation parameters can
1454 be postponed until the actual operation is executed where previously
1455 it happened immediately when an operation parameter was set.
1457 For example when setting an unsupported curve with
1458 EVP_PKEY_CTX_set_ec_paramgen_curve_nid() this function call will not
1459 fail but later keygen operations with the EVP_PKEY_CTX will fail.
1461 *OpenSSL team members and many third party contributors*
1463 * The EVP_get_cipherbyname() function will return NULL for algorithms such as
1464 "AES-128-SIV", "AES-128-CBC-CTS" and "CAMELLIA-128-CBC-CTS" which were
1465 previously only accessible via low-level interfaces. Use EVP_CIPHER_fetch()
1466 instead to retrieve these algorithms from a provider.
1470 * On build targets where the multilib postfix is set in the build
1471 configuration the libdir directory was changing based on whether
1472 the lib directory with the multilib postfix exists on the system
1473 or not. This unpredictable behavior was removed and eventual
1474 multilib postfix is now always added to the default libdir. Use
1475 `--libdir=lib` to override the libdir if adding the postfix is
1480 * The triple DES key wrap functionality now conforms to RFC 3217 but is
1481 no longer interoperable with OpenSSL 1.1.1.
1485 * The ERR_GET_FUNC() function was removed. With the loss of meaningful
1486 function codes, this function can only cause problems for calling
1491 * Add a configurable flag to output date formats as ISO 8601. Does not
1492 change the default date format.
1496 * Version of MSVC earlier than 1300 could get link warnings, which could
1497 be suppressed if the undocumented -DI_CAN_LIVE_WITH_LNK4049 was set.
1498 Support for this flag has been removed.
1502 * Rework and make DEBUG macros consistent. Remove unused -DCONF_DEBUG,
1503 -DBN_CTX_DEBUG, and REF_PRINT. Add a new tracing category and use it for
1504 printing reference counts. Rename -DDEBUG_UNUSED to -DUNUSED_RESULT_DEBUG
1505 Fix BN_DEBUG_RAND so it compiles and, when set, force DEBUG_RAND to be set
1506 also. Rename engine_debug_ref to be ENGINE_REF_PRINT also for consistency.
1510 * The signatures of the functions to get and set options on SSL and
1511 SSL_CTX objects changed from "unsigned long" to "uint64_t" type.
1512 Some source code changes may be required.
1516 * The public definitions of conf_method_st and conf_st have been
1517 deprecated. They will be made opaque in a future release.
1519 *Rich Salz and Tomáš Mráz*
1521 * Client-initiated renegotiation is disabled by default. To allow it, use
1522 the -client_renegotiation option, the SSL_OP_ALLOW_CLIENT_RENEGOTIATION
1523 flag, or the "ClientRenegotiation" config parameter as appropriate.
1527 * Add "abspath" and "includedir" pragma's to config files, to prevent,
1528 or modify relative pathname inclusion.
1532 * OpenSSL includes a cryptographic module that is intended to be FIPS 140-2
1533 validated. Please consult the README-FIPS and
1534 README-PROVIDERS files, as well as the migration guide.
1536 *OpenSSL team members and many third party contributors*
1538 * For the key types DH and DHX the allowed settable parameters are now different.
1542 * The openssl commands that read keys, certificates, and CRLs now
1543 automatically detect the PEM or DER format of the input files.
1545 *David von Oheimb, Richard Levitte, and Tomáš Mráz*
1547 * Added enhanced PKCS#12 APIs which accept a library context.
1551 * The default manual page suffix ($MANSUFFIX) has been changed to "ossl"
1555 * Added support for Kernel TLS (KTLS).
1557 *Boris Pismenny, John Baldwin and Andrew Gallatin*
1559 * Support for RFC 5746 secure renegotiation is now required by default for
1560 SSL or TLS connections to succeed.
1564 * The signature of the `copy` functional parameter of the
1565 EVP_PKEY_meth_set_copy() function has changed so its `src` argument is
1566 now `const EVP_PKEY_CTX *` instead of `EVP_PKEY_CTX *`. Similarly
1567 the signature of the `pub_decode` functional parameter of the
1568 EVP_PKEY_asn1_set_public() function has changed so its `pub` argument is
1569 now `const X509_PUBKEY *` instead of `X509_PUBKEY *`.
1573 * The error return values from some control calls (ctrl) have changed.
1577 * A public key check is now performed during EVP_PKEY_derive_set_peer().
1581 * Many functions in the EVP_ namespace that are getters of values from
1582 implementations or contexts were renamed to include get or get0 in their
1583 names. Old names are provided as macro aliases for compatibility and
1588 * The EVP_PKEY_CTRL_PKCS7_ENCRYPT, EVP_PKEY_CTRL_PKCS7_DECRYPT,
1589 EVP_PKEY_CTRL_PKCS7_SIGN, EVP_PKEY_CTRL_CMS_ENCRYPT,
1590 EVP_PKEY_CTRL_CMS_DECRYPT, and EVP_PKEY_CTRL_CMS_SIGN control operations
1595 * The EVP_PKEY_public_check() and EVP_PKEY_param_check() functions now work for
1598 * The output from the command line applications may have minor
1603 * The output from numerous "printing" may have minor changes.
1607 * Windows thread synchronization uses read/write primitives (SRWLock) when
1608 supported by the OS, otherwise CriticalSection continues to be used.
1612 * Add filter BIO BIO_f_readbuffer() that allows BIO_tell() and BIO_seek() to
1613 work on read only BIO source/sinks that do not support these functions.
1614 This allows piping or redirection of a file BIO using stdin to be buffered
1615 into memory. This is used internally in OSSL_DECODER_from_bio().
1619 * OSSL_STORE_INFO_get_type() may now return an additional value. In 1.1.1
1620 this function would return one of the values OSSL_STORE_INFO_NAME,
1621 OSSL_STORE_INFO_PKEY, OSSL_STORE_INFO_PARAMS, OSSL_STORE_INFO_CERT or
1622 OSSL_STORE_INFO_CRL. Decoded public keys would previously have been reported
1623 as type OSSL_STORE_INFO_PKEY in 1.1.1. In 3.0 decoded public keys are now
1624 reported as having the new type OSSL_STORE_INFO_PUBKEY. Applications
1625 using this function should be amended to handle the changed return value.
1629 * Improved adherence to Enhanced Security Services (ESS, RFC 2634 and RFC 5035)
1630 for the TSP and CMS Advanced Electronic Signatures (CAdES) implementations.
1631 As required by RFC 5035 check both ESSCertID and ESSCertIDv2 if both present.
1632 Correct the semantics of checking the validation chain in case ESSCertID{,v2}
1633 contains more than one certificate identifier: This means that all
1634 certificates referenced there MUST be part of the validation chain.
1638 * The implementation of older EVP ciphers related to CAST, IDEA, SEED, RC2, RC4,
1639 RC5, DESX and DES have been moved to the legacy provider.
1643 * The implementation of the EVP digests MD2, MD4, MDC2, WHIRLPOOL and
1644 RIPEMD-160 have been moved to the legacy provider.
1648 * The deprecated function EVP_PKEY_get0() now returns NULL being called for a
1653 * The deprecated functions EVP_PKEY_get0_RSA(),
1654 EVP_PKEY_get0_DSA(), EVP_PKEY_get0_EC_KEY(), EVP_PKEY_get0_DH(),
1655 EVP_PKEY_get0_hmac(), EVP_PKEY_get0_poly1305() and EVP_PKEY_get0_siphash() as
1656 well as the similarly named "get1" functions behave differently in
1661 * A number of functions handling low-level keys or engines were deprecated
1662 including EVP_PKEY_set1_engine(), EVP_PKEY_get0_engine(), EVP_PKEY_assign(),
1663 EVP_PKEY_get0(), EVP_PKEY_get0_hmac(), EVP_PKEY_get0_poly1305() and
1664 EVP_PKEY_get0_siphash().
1668 * PKCS#5 PBKDF1 key derivation has been moved from PKCS5_PBE_keyivgen() into
1669 the legacy crypto provider as an EVP_KDF. Applications requiring this KDF
1670 will need to load the legacy crypto provider. This includes these PBE
1671 algorithms which use this KDF:
1672 - NID_pbeWithMD2AndDES_CBC
1673 - NID_pbeWithMD5AndDES_CBC
1674 - NID_pbeWithSHA1AndRC2_CBC
1675 - NID_pbeWithMD2AndRC2_CBC
1676 - NID_pbeWithMD5AndRC2_CBC
1677 - NID_pbeWithSHA1AndDES_CBC
1681 * Deprecated obsolete BIO_set_callback(), BIO_get_callback(), and
1682 BIO_debug_callback() functions.
1686 * Deprecated obsolete EVP_PKEY_CTX_get0_dh_kdf_ukm() and
1687 EVP_PKEY_CTX_get0_ecdh_kdf_ukm() functions.
1691 * The RAND_METHOD APIs have been deprecated.
1695 * The SRP APIs have been deprecated.
1699 * Add a compile time option to prevent the caching of provider fetched
1700 algorithms. This is enabled by including the no-cached-fetch option
1701 at configuration time.
1705 * pkcs12 now uses defaults of PBKDF2, AES and SHA-256, with a MAC iteration
1706 count of PKCS12_DEFAULT_ITER.
1708 *Tomáš Mráz and Sahana Prasad*
1710 * The openssl speed command does not use low-level API calls anymore.
1714 * Parallel dual-prime 1024-bit modular exponentiation for AVX512_IFMA
1717 *Ilya Albrekht, Sergey Kirillov, Andrey Matyukov (Intel Corp)*
1719 * Combining the Configure options no-ec and no-dh no longer disables TLSv1.3.
1723 * Implemented support for fully "pluggable" TLSv1.3 groups. This means that
1724 providers may supply their own group implementations (using either the "key
1725 exchange" or the "key encapsulation" methods) which will automatically be
1726 detected and used by libssl.
1728 *Matt Caswell, Nicola Tuveri*
1730 * The undocumented function X509_certificate_type() has been deprecated;
1734 * Deprecated the obsolete BN_pseudo_rand() and BN_pseudo_rand_range().
1738 * Removed RSA padding mode for SSLv23 (which was only used for
1739 SSLv2). This includes the functions RSA_padding_check_SSLv23() and
1740 RSA_padding_add_SSLv23() and the `-ssl` option in the deprecated
1745 * Deprecated the obsolete X9.31 RSA key generation related functions.
1747 * While a callback function set via `SSL_CTX_set_cert_verify_callback()`
1748 is not allowed to return a value > 1, this is no more taken as failure.
1750 *Viktor Dukhovni and David von Oheimb*
1752 * Deprecated the obsolete X9.31 RSA key generation related functions
1753 BN_X931_generate_Xpq(), BN_X931_derive_prime_ex(), and
1754 BN_X931_generate_prime_ex().
1758 * The default key generation method for the regular 2-prime RSA keys was
1759 changed to the FIPS 186-4 B.3.6 method.
1763 * Deprecated the BN_is_prime_ex() and BN_is_prime_fasttest_ex() functions.
1767 * Deprecated EVP_MD_CTX_set_update_fn() and EVP_MD_CTX_update_fn().
1771 * Deprecated the type OCSP_REQ_CTX and the functions OCSP_REQ_CTX_*() and
1772 replaced with OSSL_HTTP_REQ_CTX and the functions OSSL_HTTP_REQ_CTX_*().
1774 *Rich Salz, Richard Levitte, and David von Oheimb*
1776 * Deprecated `X509_http_nbio()` and `X509_CRL_http_nbio()`.
1780 * Deprecated `OCSP_parse_url()`.
1784 * Validation of SM2 keys has been separated from the validation of regular EC
1789 * Behavior of the `pkey` app is changed, when using the `-check` or `-pubcheck`
1790 switches: a validation failure triggers an early exit, returning a failure
1791 exit status to the parent process.
1795 * Changed behavior of SSL_CTX_set_ciphersuites() and SSL_set_ciphersuites()
1796 to ignore unknown ciphers.
1800 * The `-cipher-commands` and `-digest-commands` options
1801 of the command line utility `list` have been deprecated.
1802 Instead use the `-cipher-algorithms` and `-digest-algorithms` options.
1806 * Added convenience functions for generating asymmetric key pairs:
1807 The 'quick' one-shot (yet somewhat limited) function L<EVP_PKEY_Q_keygen(3)>
1808 and macros for the most common cases: <EVP_RSA_gen(3)> and L<EVP_EC_gen(3)>.
1812 * All of the low-level EC_KEY functions have been deprecated.
1814 *Shane Lontis, Paul Dale, Richard Levitte, and Tomáš Mráz*
1816 * Deprecated all the libcrypto and libssl error string loading
1821 * The functions SSL_CTX_set_tmp_dh_callback and SSL_set_tmp_dh_callback, as
1822 well as the macros SSL_CTX_set_tmp_dh() and SSL_set_tmp_dh() have been
1827 * The `-crypt` option to the `passwd` command line tool has been removed.
1831 * The -C option to the `x509`, `dhparam`, `dsaparam`, and `ecparam` commands
1836 * Add support for AES Key Wrap inverse ciphers to the EVP layer.
1840 * Deprecated EVP_PKEY_set1_tls_encodedpoint() and
1841 EVP_PKEY_get1_tls_encodedpoint().
1845 * The security callback, which can be customised by application code, supports
1846 the security operation SSL_SECOP_TMP_DH. One location of the "other" parameter
1847 was incorrectly passing a DH object. It now passed an EVP_PKEY in all cases.
1851 * Add PKCS7_get_octet_string() and PKCS7_type_is_other() to the public
1852 interface. Their functionality remains unchanged.
1856 * Added new option for 'openssl list', '-providers', which will display the
1857 list of loaded providers, their names, version and status. It optionally
1858 displays their gettable parameters.
1862 * Removed EVP_PKEY_set_alias_type().
1866 * Deprecated `EVP_PKEY_CTX_set_rsa_keygen_pubexp()` and introduced
1867 `EVP_PKEY_CTX_set1_rsa_keygen_pubexp()`, which is now preferred.
1871 * Changed all "STACK" functions to be macros instead of inline functions. Macro
1872 parameters are still checked for type safety at compile time via helper
1877 * Remove the RAND_DRBG API
1879 *Paul Dale and Matthias St. Pierre*
1881 * Allow `SSL_set1_host()` and `SSL_add1_host()` to take IP literal addresses
1882 as well as actual hostnames.
1886 * The 'MinProtocol' and 'MaxProtocol' configuration commands now silently
1887 ignore TLS protocol version bounds when configuring DTLS-based contexts, and
1888 conversely, silently ignore DTLS protocol version bounds when configuring
1889 TLS-based contexts. The commands can be repeated to set bounds of both
1890 types. The same applies with the corresponding "min_protocol" and
1891 "max_protocol" command-line switches, in case some application uses both TLS
1894 SSL_CTX instances that are created for a fixed protocol version (e.g.
1895 `TLSv1_server_method()`) also silently ignore version bounds. Previously
1896 attempts to apply bounds to these protocol versions would result in an
1897 error. Now only the "version-flexible" SSL_CTX instances are subject to
1898 limits in configuration files in command-line options.
1902 * Deprecated the `ENGINE` API. Engines should be replaced with providers
1907 * Reworked the recorded ERR codes to make better space for system errors.
1908 To distinguish them, the macro `ERR_SYSTEM_ERROR()` indicates if the
1909 given code is a system error (true) or an OpenSSL error (false).
1913 * Reworked the test perl framework to better allow parallel testing.
1915 *Nicola Tuveri and David von Oheimb*
1917 * Added ciphertext stealing algorithms AES-128-CBC-CTS, AES-192-CBC-CTS and
1918 AES-256-CBC-CTS to the providers. CS1, CS2 and CS3 variants are supported.
1922 * 'Configure' has been changed to figure out the configuration target if
1923 none is given on the command line. Consequently, the 'config' script is
1924 now only a mere wrapper. All documentation is changed to only mention
1927 *Rich Salz and Richard Levitte*
1929 * Added a library context `OSSL_LIB_CTX` that applications as well as
1930 other libraries can use to form a separate context within which
1931 libcrypto operations are performed.
1935 * Added various `_ex` functions to the OpenSSL API that support using
1936 a non-default `OSSL_LIB_CTX`.
1940 * Handshake now fails if Extended Master Secret extension is dropped
1945 * Dropped interactive mode from the `openssl` program.
1949 * Deprecated `EVP_PKEY_cmp()` and `EVP_PKEY_cmp_parameters()`.
1951 *David von Oheimb and Shane Lontis*
1953 * Deprecated `EC_METHOD_get_field_type()`.
1957 * Deprecated EC_GFp_simple_method(), EC_GFp_mont_method(),
1958 EC_GF2m_simple_method(), EC_GFp_nist_method(), EC_GFp_nistp224_method()
1959 EC_GFp_nistp256_method(), and EC_GFp_nistp521_method().
1963 * Deprecated EC_GROUP_new(), EC_GROUP_method_of(), and EC_POINT_method_of().
1967 * Add CAdES-BES signature verification support, mostly derived
1968 from ESSCertIDv2 TS (RFC 5816) contribution by Marek Klein.
1970 *Filipe Raimundo da Silva*
1972 * Add CAdES-BES signature scheme and attributes support (RFC 5126) to CMS API.
1976 * Added the AuthEnvelopedData content type structure (RFC 5083) with AES-GCM
1977 parameter (RFC 5084) for the Cryptographic Message Syntax (CMS).
1981 * Deprecated EC_POINT_make_affine() and EC_POINTs_make_affine().
1985 * Deprecated EC_GROUP_precompute_mult(), EC_GROUP_have_precompute_mult(), and
1986 EC_KEY_precompute_mult().
1990 * Deprecated EC_POINTs_mul().
1994 * Removed FIPS_mode() and FIPS_mode_set().
1998 * The SSL option SSL_OP_IGNORE_UNEXPECTED_EOF is introduced.
2002 * Deprecated EC_POINT_set_Jprojective_coordinates_GFp() and
2003 EC_POINT_get_Jprojective_coordinates_GFp().
2007 * Added OSSL_PARAM_BLD to the public interface. This allows OSSL_PARAM
2008 arrays to be more easily constructed via a series of utility functions.
2009 Create a parameter builder using OSSL_PARAM_BLD_new(), add parameters using
2010 the various push functions and finally convert to a passable OSSL_PARAM
2011 array using OSSL_PARAM_BLD_to_param().
2015 * The security strength of SHA1 and MD5 based signatures in TLS has been
2020 * Added EVP_PKEY_set_type_by_keymgmt(), to initialise an EVP_PKEY to
2021 contain a provider side internal key.
2025 * ASN1_verify(), ASN1_digest() and ASN1_sign() have been deprecated.
2029 * Project text documents not yet having a proper file name extension
2030 (`HACKING`, `LICENSE`, `NOTES*`, `README*`, `VERSION`) have been renamed to
2031 `*.md` as far as reasonable, else `*.txt`, for better use with file managers.
2035 * The main project documents (README, NEWS, CHANGES, INSTALL, SUPPORT)
2036 have been converted to Markdown with the goal to produce documents
2037 which not only look pretty when viewed online in the browser, but
2038 remain well readable inside a plain text editor.
2040 To achieve this goal, a 'minimalistic' Markdown style has been applied
2041 which avoids formatting elements that interfere too much with the
2042 reading flow in the text file. For example, it
2044 * avoids [ATX headings][] and uses [setext headings][] instead
2045 (which works for `<h1>` and `<h2>` headings only).
2046 * avoids [inline links][] and uses [reference links][] instead.
2047 * avoids [fenced code blocks][] and uses [indented code blocks][] instead.
2049 [ATX headings]: https://github.github.com/gfm/#atx-headings
2050 [setext headings]: https://github.github.com/gfm/#setext-headings
2051 [inline links]: https://github.github.com/gfm/#inline-link
2052 [reference links]: https://github.github.com/gfm/#reference-link
2053 [fenced code blocks]: https://github.github.com/gfm/#fenced-code-blocks
2054 [indented code blocks]: https://github.github.com/gfm/#indented-code-blocks
2056 *Matthias St. Pierre*
2058 * The test suite is changed to preserve results of each test recipe.
2059 A new directory test-runs/ with subdirectories named like the
2060 test recipes are created in the build tree for this purpose.
2064 * Added an implementation of CMP and CRMF (RFC 4210, RFC 4211 RFC 6712).
2065 This adds `crypto/cmp/`, `crpyto/crmf/`, `apps/cmp.c`, and `test/cmp_*`.
2066 See L<openssl-cmp(1)> and L<OSSL_CMP_exec_IR_ses(3)> as starting points.
2068 *David von Oheimb, Martin Peylo*
2070 * Generalized the HTTP client code from `crypto/ocsp/` into `crpyto/http/`.
2071 It supports arbitrary request and response content types, GET redirection,
2072 TLS, connections via HTTP(S) proxies, connections and exchange via
2073 user-defined BIOs (allowing implicit connections), persistent connections,
2074 and timeout checks. See L<OSSL_HTTP_transfer(3)> etc. for details.
2075 The legacy OCSP-focused (and only partly documented) API
2076 is retained for backward compatibility, while most of it is deprecated.
2080 * Added `util/check-format.pl`, a tool for checking adherence to the
2081 OpenSSL coding style <https://www.openssl.org/policies/codingstyle.html>.
2082 The checks performed are incomplete and yield some false positives.
2083 Still the tool should be useful for detecting most typical glitches.
2087 * `BIO_do_connect()` and `BIO_do_handshake()` have been extended:
2088 If domain name resolution yields multiple IP addresses all of them are tried
2089 after `connect()` failures.
2093 * All of the low-level RSA functions have been deprecated.
2097 * X509 certificates signed using SHA1 are no longer allowed at security
2102 * The command line utilities dhparam, dsa, gendsa and dsaparam have been
2103 modified to use PKEY APIs. These commands are now in maintenance mode
2104 and no new features will be added to them.
2108 * The command line utility rsautl has been deprecated.
2112 * The command line utilities genrsa and rsa have been modified to use PKEY
2113 APIs. They now write PKCS#8 keys by default. These commands are now in
2114 maintenance mode and no new features will be added to them.
2118 * All of the low-level DH functions have been deprecated.
2120 *Paul Dale and Matt Caswell*
2122 * All of the low-level DSA functions have been deprecated.
2126 * Reworked the treatment of EC EVP_PKEYs with the SM2 curve to
2127 automatically become EVP_PKEY_SM2 rather than EVP_PKEY_EC.
2131 * Deprecated low-level ECDH and ECDSA functions.
2135 * Deprecated EVP_PKEY_decrypt_old() and EVP_PKEY_encrypt_old().
2139 * Enhanced the documentation of EVP_PKEY_get_size(), EVP_PKEY_get_bits()
2140 and EVP_PKEY_get_security_bits(). Especially EVP_PKEY_get_size() needed
2141 a new formulation to include all the things it can be used for,
2142 as well as words of caution.
2146 * The SSL_CTX_set_tlsext_ticket_key_cb(3) function has been deprecated.
2150 * All of the low-level HMAC functions have been deprecated.
2152 *Paul Dale and David von Oheimb*
2154 * Over two thousand fixes were made to the documentation, including:
2155 - Common options (such as -rand/-writerand, TLS version control, etc)
2156 were refactored and point to newly-enhanced descriptions in openssl.pod.
2157 - Added style conformance for all options (with help from Richard Levitte),
2158 documented all reported missing options, added a CI build to check
2159 that all options are documented and that no unimplemented options
2161 - Documented some internals, such as all use of environment variables.
2162 - Addressed all internal broken L<> references.
2166 * All of the low-level CMAC functions have been deprecated.
2170 * The low-level MD2, MD4, MD5, MDC2, RIPEMD160 and Whirlpool digest
2171 functions have been deprecated.
2173 *Paul Dale and David von Oheimb*
2175 * Corrected the documentation of the return values from the `EVP_DigestSign*`
2176 set of functions. The documentation mentioned negative values for some
2177 errors, but this was never the case, so the mention of negative values
2180 Code that followed the documentation and thereby check with something
2181 like `EVP_DigestSignInit(...) <= 0` will continue to work undisturbed.
2185 * All of the low-level cipher functions have been deprecated.
2187 *Matt Caswell and Paul Dale*
2189 * Removed include/openssl/opensslconf.h.in and replaced it with
2190 include/openssl/configuration.h.in, which differs in not including
2191 <openssl/macros.h>. A short header include/openssl/opensslconf.h
2192 was added to include both.
2194 This allows internal hacks where one might need to modify the set
2195 of configured macros, for example this if deprecated symbols are
2196 still supposed to be available internally:
2198 #include <openssl/configuration.h>
2200 #undef OPENSSL_NO_DEPRECATED
2201 #define OPENSSL_SUPPRESS_DEPRECATED
2203 #include <openssl/macros.h>
2205 This should not be used by applications that use the exported
2206 symbols, as that will lead to linking errors.
2210 * Fixed an overflow bug in the x64_64 Montgomery squaring procedure
2211 used in exponentiation with 512-bit moduli. No EC algorithms are
2212 affected. Analysis suggests that attacks against 2-prime RSA1024,
2213 3-prime RSA1536, and DSA1024 as a result of this defect would be very
2214 difficult to perform and are not believed likely. Attacks against DH512
2215 are considered just feasible. However, for an attack the target would
2216 have to reuse the DH512 private key, which is not recommended anyway.
2217 Also applications directly using the low-level API BN_mod_exp may be
2218 affected if they use BN_FLG_CONSTTIME.
2223 * Most memory-debug features have been deprecated, and the functionality
2224 replaced with no-ops.
2228 * Added documentation for the STACK API.
2232 * Introduced a new method type and API, OSSL_ENCODER, to represent
2233 generic encoders. These do the same sort of job that PEM writers
2234 and d2i functions do, but with support for methods supplied by
2235 providers, and the possibility for providers to support other
2240 * Introduced a new method type and API, OSSL_DECODER, to represent
2241 generic decoders. These do the same sort of job that PEM readers
2242 and i2d functions do, but with support for methods supplied by
2243 providers, and the possibility for providers to support other
2248 * Added a .pragma directive to the syntax of configuration files, to
2249 allow varying behavior in a supported and predictable manner.
2250 Currently added pragma:
2254 This allows dollar signs to be a keyword character unless it's
2255 followed by a opening brace or parenthesis. This is useful for
2256 platforms where dollar signs are commonly used in names, such as
2257 volume names and system directory names on VMS.
2261 * Added functionality to create an EVP_PKEY from user data.
2265 * Change the interpretation of the '--api' configuration option to
2266 mean that this is a desired API compatibility level with no
2267 further meaning. The previous interpretation, that this would
2268 also mean to remove all deprecated symbols up to and including
2269 the given version, no requires that 'no-deprecated' is also used
2270 in the configuration.
2272 When building applications, the desired API compatibility level
2273 can be set with the OPENSSL_API_COMPAT macro like before. For
2274 API compatibility version below 3.0, the old style numerical
2275 value is valid as before, such as -DOPENSSL_API_COMPAT=0x10100000L.
2276 For version 3.0 and on, the value is expected to be the decimal
2277 value calculated from the major and minor version like this:
2279 MAJOR * 10000 + MINOR * 100
2283 -DOPENSSL_API_COMPAT=30000 For 3.0
2284 -DOPENSSL_API_COMPAT=30200 For 3.2
2286 To hide declarations that are deprecated up to and including the
2287 given API compatibility level, -DOPENSSL_NO_DEPRECATED must be
2288 given when building the application as well.
2292 * Added the X509_LOOKUP_METHOD called X509_LOOKUP_store, to allow
2293 access to certificate and CRL stores via URIs and OSSL_STORE
2296 This adds the following functions:
2298 - X509_LOOKUP_store()
2299 - X509_STORE_load_file()
2300 - X509_STORE_load_path()
2301 - X509_STORE_load_store()
2302 - SSL_add_store_cert_subjects_to_stack()
2303 - SSL_CTX_set_default_verify_store()
2304 - SSL_CTX_load_verify_file()
2305 - SSL_CTX_load_verify_dir()
2306 - SSL_CTX_load_verify_store()
2310 * Added a new method to gather entropy on VMS, based on SYS$GET_ENTROPY.
2311 The presence of this system service is determined at run-time.
2315 * Added functionality to create an EVP_PKEY context based on data
2316 for methods from providers. This takes an algorithm name and a
2317 property query string and simply stores them, with the intent
2318 that any operation that uses this context will use those strings
2319 to fetch the needed methods implicitly, thereby making the port
2320 of application written for pre-3.0 OpenSSL easier.
2324 * The undocumented function NCONF_WIN32() has been deprecated; for
2325 conversion details see the HISTORY section of doc/man5/config.pod
2329 * Introduced the new functions EVP_DigestSignInit_ex() and
2330 EVP_DigestVerifyInit_ex(). The macros EVP_DigestSignUpdate() and
2331 EVP_DigestVerifyUpdate() have been converted to functions. See the man
2332 pages for further details.
2336 * Over two thousand fixes were made to the documentation, including:
2337 adding missing command flags, better style conformance, documentation
2340 *Rich Salz, Richard Levitte*
2342 * s390x assembly pack: add hardware-support for P-256, P-384, P-521,
2343 X25519, X448, Ed25519 and Ed448.
2347 * Print all values for a PKCS#12 attribute with 'openssl pkcs12', not just
2352 * Deprecated the public definition of `ERR_STATE` as well as the function
2353 `ERR_get_state()`. This is done in preparation of making `ERR_STATE` an
2358 * Added ERR functionality to give callers access to the stored function
2359 names that have replaced the older function code based functions.
2361 New functions are ERR_peek_error_func(), ERR_peek_last_error_func(),
2362 ERR_peek_error_data(), ERR_peek_last_error_data(), ERR_get_error_all(),
2363 ERR_peek_error_all() and ERR_peek_last_error_all().
2365 Deprecate ERR functions ERR_get_error_line(), ERR_get_error_line_data(),
2366 ERR_peek_error_line_data(), ERR_peek_last_error_line_data() and
2367 ERR_func_error_string().
2371 * Extended testing to be verbose for failing tests only. The make variables
2372 VERBOSE_FAILURE or VF can be used to enable this:
2374 $ make VF=1 test # Unix
2375 $ mms /macro=(VF=1) test ! OpenVMS
2376 $ nmake VF=1 test # Windows
2380 * Added the `-copy_extensions` option to the `x509` command for use with
2381 `-req` and `-x509toreq`. When given with the `copy` or `copyall` argument,
2382 all extensions in the request are copied to the certificate or vice versa.
2384 *David von Oheimb*, *Kirill Stefanenkov <kirill_stefanenkov@rambler.ru>*
2386 * Added the `-copy_extensions` option to the `req` command for use with
2387 `-x509`. When given with the `copy` or `copyall` argument,
2388 all extensions in the certification request are copied to the certificate.
2392 * The `x509`, `req`, and `ca` commands now make sure that X.509v3 certificates
2393 they generate are by default RFC 5280 compliant in the following sense:
2394 There is a subjectKeyIdentifier extension with a hash value of the public key
2395 and for not self-signed certs there is an authorityKeyIdentifier extension
2396 with a keyIdentifier field or issuer information identifying the signing key.
2397 This is done unless some configuration overrides the new default behavior,
2398 such as `subjectKeyIdentifier = none` and `authorityKeyIdentifier = none`.
2402 * Added several checks to `X509_verify_cert()` according to requirements in
2403 RFC 5280 in case `X509_V_FLAG_X509_STRICT` is set
2404 (which may be done by using the CLI option `-x509_strict`):
2405 * The basicConstraints of CA certificates must be marked critical.
2406 * CA certificates must explicitly include the keyUsage extension.
2407 * If a pathlenConstraint is given the key usage keyCertSign must be allowed.
2408 * The issuer name of any certificate must not be empty.
2409 * The subject name of CA certs, certs with keyUsage crlSign,
2410 and certs without subjectAlternativeName must not be empty.
2411 * If a subjectAlternativeName extension is given it must not be empty.
2412 * The signatureAlgorithm field and the cert signature must be consistent.
2413 * Any given authorityKeyIdentifier and any given subjectKeyIdentifier
2414 must not be marked critical.
2415 * The authorityKeyIdentifier must be given for X.509v3 certs
2416 unless they are self-signed.
2417 * The subjectKeyIdentifier must be given for all X.509v3 CA certs.
2421 * Certificate verification using `X509_verify_cert()` meanwhile rejects EC keys
2422 with explicit curve parameters (specifiedCurve) as required by RFC 5480.
2426 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
2427 used even when parsing explicit parameters, when loading a encoded key
2428 or calling `EC_GROUP_new_from_ecpkparameters()`/
2429 `EC_GROUP_new_from_ecparameters()`.
2430 This prevents bypass of security hardening and performance gains,
2431 especially for curves with specialized EC_METHODs.
2432 By default, if a key encoded with explicit parameters is loaded and later
2433 encoded, the output is still encoded with explicit parameters, even if
2434 internally a "named" EC_GROUP is used for computation.
2438 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
2439 this change, EC_GROUP_set_generator would accept order and/or cofactor as
2440 NULL. After this change, only the cofactor parameter can be NULL. It also
2441 does some minimal sanity checks on the passed order.
2446 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
2447 An attack is simple, if the first CMS_recipientInfo is valid but the
2448 second CMS_recipientInfo is chosen ciphertext. If the second
2449 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
2450 encryption key will be replaced by garbage, and the message cannot be
2451 decoded, but if the RSA decryption fails, the correct encryption key is
2452 used and the recipient will not notice the attack.
2453 As a work around for this potential attack the length of the decrypted
2454 key must be equal to the cipher default key length, in case the
2455 certificate is not given and all recipientInfo are tried out.
2456 The old behaviour can be re-enabled in the CMS code by setting the
2457 CMS_DEBUG_DECRYPT flag.
2461 * Early start up entropy quality from the DEVRANDOM seed source has been
2462 improved for older Linux systems. The RAND subsystem will wait for
2463 /dev/random to be producing output before seeding from /dev/urandom.
2464 The seeded state is stored for future library initialisations using
2465 a system global shared memory segment. The shared memory identifier
2466 can be configured by defining OPENSSL_RAND_SEED_DEVRANDOM_SHM_ID to
2467 the desired value. The default identifier is 114.
2471 * Revised BN_generate_prime_ex to not avoid factors 2..17863 in p-1
2472 when primes for RSA keys are computed.
2473 Since we previously always generated primes == 2 (mod 3) for RSA keys,
2474 the 2-prime and 3-prime RSA modules were easy to distinguish, since
2475 `N = p*q = 1 (mod 3)`, but `N = p*q*r = 2 (mod 3)`. Therefore, fingerprinting
2476 2-prime vs. 3-prime RSA keys was possible by computing N mod 3.
2477 This avoids possible fingerprinting of newly generated RSA modules.
2481 * Correct the extended master secret constant on EBCDIC systems. Without this
2482 fix TLS connections between an EBCDIC system and a non-EBCDIC system that
2483 negotiate EMS will fail. Unfortunately this also means that TLS connections
2484 between EBCDIC systems with this fix, and EBCDIC systems without this
2485 fix will fail if they negotiate EMS.
2489 * Changed the library initialisation so that the config file is now loaded
2490 by default. This was already the case for libssl. It now occurs for both
2491 libcrypto and libssl. Use the OPENSSL_INIT_NO_LOAD_CONFIG option to
2492 `OPENSSL_init_crypto()` to suppress automatic loading of a config file.
2496 * Introduced new error raising macros, `ERR_raise()` and `ERR_raise_data()`,
2497 where the former acts as a replacement for `ERR_put_error()`, and the
2498 latter replaces the combination `ERR_put_error()` + `ERR_add_error_data()`.
2499 `ERR_raise_data()` adds more flexibility by taking a format string and
2500 an arbitrary number of arguments following it, to be processed with
2505 * Introduced a new function, `OSSL_PROVIDER_available()`, which can be used
2506 to check if a named provider is loaded and available. When called, it
2507 will also activate all fallback providers if such are still present.
2511 * Enforce a minimum DH modulus size of 512 bits.
2515 * Changed DH parameters to generate the order q subgroup instead of 2q.
2516 Previously generated DH parameters are still accepted by DH_check
2517 but DH_generate_key works around that by clearing bit 0 of the
2518 private key for those. This avoids leaking bit 0 of the private key.
2522 * Significantly reduce secure memory usage by the randomness pools.
2526 * `{CRYPTO,OPENSSL}_mem_debug_{push,pop}` are now no-ops and have been
2531 * A new type, EVP_KEYEXCH, has been introduced to represent key exchange
2532 algorithms. An implementation of a key exchange algorithm can be obtained
2533 by using the function EVP_KEYEXCH_fetch(). An EVP_KEYEXCH algorithm can be
2534 used in a call to EVP_PKEY_derive_init_ex() which works in a similar way to
2535 the older EVP_PKEY_derive_init() function. See the man pages for the new
2536 functions for further details.
2540 * The EVP_PKEY_CTX_set_dh_pad() macro has now been converted to a function.
2544 * Removed the function names from error messages and deprecated the
2549 * Removed NextStep support and the macro OPENSSL_UNISTD
2553 * Removed DES_check_key. Also removed OPENSSL_IMPLEMENT_GLOBAL,
2554 OPENSSL_GLOBAL_REF, OPENSSL_DECLARE_GLOBAL.
2555 Also removed "export var as function" capability; we do not export
2556 variables, only functions.
2560 * RC5_32_set_key has been changed to return an int type, with 0 indicating
2561 an error and 1 indicating success. In previous versions of OpenSSL this
2562 was a void type. If a key was set longer than the maximum possible this
2567 * Support SM2 signing and verification schemes with X509 certificate.
2571 * Use SHA256 as the default digest for TS query in the `ts` app.
2575 * Change PBKDF2 to conform to SP800-132 instead of the older PKCS5 RFC2898.
2579 * Default cipher lists/suites are now available via a function, the
2580 #defines are deprecated.
2584 * Add target VC-WIN32-UWP, VC-WIN64A-UWP, VC-WIN32-ARM-UWP and
2585 VC-WIN64-ARM-UWP in Windows OneCore target for making building libraries
2586 for Windows Store apps easier. Also, the "no-uplink" option has been added.
2590 * Join the directories crypto/x509 and crypto/x509v3
2594 * Added command 'openssl kdf' that uses the EVP_KDF API.
2598 * Added command 'openssl mac' that uses the EVP_MAC API.
2602 * Added OPENSSL_info() to get diverse built-in OpenSSL data, such
2603 as default directories. Also added the command 'openssl info'
2604 for scripting purposes.
2608 * The functions AES_ige_encrypt() and AES_bi_ige_encrypt() have been
2613 * Add prediction resistance to the DRBG reseeding process.
2617 * Limit the number of blocks in a data unit for AES-XTS to 2^20 as
2618 mandated by IEEE Std 1619-2018.
2622 * Added newline escaping functionality to a filename when using openssl dgst.
2623 This output format is to replicate the output format found in the `*sum`
2624 checksum programs. This aims to preserve backward compatibility.
2626 *Matt Eaton, Richard Levitte, and Paul Dale*
2628 * Removed the heartbeat message in DTLS feature, as it has very
2629 little usage and doesn't seem to fulfill a valuable purpose.
2630 The configuration option is now deprecated.
2634 * Changed the output of 'openssl {digestname} < file' to display the
2635 digest name in its output.
2639 * Added a new generic trace API which provides support for enabling
2640 instrumentation through trace output.
2642 *Richard Levitte & Matthias St. Pierre*
2644 * Added build tests for C++. These are generated files that only do one
2645 thing, to include one public OpenSSL head file each. This tests that
2646 the public header files can be usefully included in a C++ application.
2648 This test isn't enabled by default. It can be enabled with the option
2649 'enable-buildtest-c++'.
2653 * Added KB KDF (EVP_KDF_KB) to EVP_KDF.
2657 * Added SSH KDF (EVP_KDF_SSHKDF) and KRB5 KDF (EVP_KDF_KRB5KDF) to EVP_KDF.
2661 * Added Single Step KDF (EVP_KDF_SS), X963 KDF, and X942 KDF to EVP_KDF.
2665 * Added KMAC to EVP_MAC.
2669 * Added property based algorithm implementation selection framework to
2674 * Added SCA hardening for modular field inversion in EC_GROUP through
2675 a new dedicated field_inv() pointer in EC_METHOD.
2676 This also addresses a leakage affecting conversions from projective
2677 to affine coordinates.
2679 *Billy Bob Brumley, Nicola Tuveri*
2681 * Added EVP_KDF, an EVP layer KDF API, to simplify adding KDF and PRF
2682 implementations. This includes an EVP_PKEY to EVP_KDF bridge for
2683 those algorithms that were already supported through the EVP_PKEY API
2684 (scrypt, TLS1 PRF and HKDF). The low-level KDF functions for PBKDF2
2685 and scrypt are now wrappers that call EVP_KDF.
2689 * Build devcrypto engine as a dynamic engine.
2691 *Eneas U de Queiroz*
2693 * Add keyed BLAKE2 to EVP_MAC.
2697 * Fix a bug in the computation of the endpoint-pair shared secret used
2698 by DTLS over SCTP. This breaks interoperability with older versions
2699 of OpenSSL like OpenSSL 1.1.0 and OpenSSL 1.0.2. There is a runtime
2700 switch SSL_MODE_DTLS_SCTP_LABEL_LENGTH_BUG (off by default) enabling
2701 interoperability with such broken implementations. However, enabling
2702 this switch breaks interoperability with correct implementations.
2704 * Fix a use after free bug in d2i_X509_PUBKEY when overwriting a
2705 re-used X509_PUBKEY object if the second PUBKEY is malformed.
2709 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
2713 * Changed the license to the Apache License v2.0.
2717 * Switch to a new version scheme using three numbers MAJOR.MINOR.PATCH.
2719 - Major releases (indicated by incrementing the MAJOR release number)
2720 may introduce incompatible API/ABI changes.
2721 - Minor releases (indicated by incrementing the MINOR release number)
2722 may introduce new features but retain API/ABI compatibility.
2723 - Patch releases (indicated by incrementing the PATCH number)
2724 are intended for bug fixes and other improvements of existing
2725 features only (like improving performance or adding documentation)
2726 and retain API/ABI compatibility.
2730 * Add support for RFC5297 SIV mode (siv128), including AES-SIV.
2734 * Remove the 'dist' target and add a tarball building script. The
2735 'dist' target has fallen out of use, and it shouldn't be
2736 necessary to configure just to create a source distribution.
2740 * Recreate the OS390-Unix config target. It no longer relies on a
2741 special script like it did for OpenSSL pre-1.1.0.
2745 * Instead of having the source directories listed in Configure, add
2746 a 'build.info' keyword SUBDIRS to indicate what sub-directories to
2751 * Add GMAC to EVP_MAC.
2755 * Ported the HMAC, CMAC and SipHash EVP_PKEY_METHODs to EVP_MAC.
2759 * Added EVP_MAC, an EVP layer MAC API, to simplify adding MAC
2760 implementations. This includes a generic EVP_PKEY to EVP_MAC bridge,
2761 to facilitate the continued use of MACs through raw private keys in
2762 functionality such as `EVP_DigestSign*` and `EVP_DigestVerify*`.
2766 * Deprecate ECDH_KDF_X9_62().
2770 * Added EVP_PKEY_ECDH_KDF_X9_63 and ecdh_KDF_X9_63() as replacements for
2771 the EVP_PKEY_ECDH_KDF_X9_62 KDF type and ECDH_KDF_X9_62(). The old names
2772 are retained for backwards compatibility.
2776 * AES-XTS mode now enforces that its two keys are different to mitigate
2777 the attacked described in "Efficient Instantiations of Tweakable
2778 Blockciphers and Refinements to Modes OCB and PMAC" by Phillip Rogaway.
2779 Details of this attack can be obtained from:
2780 <http://web.cs.ucdavis.edu/%7Erogaway/papers/offsets.pdf>
2784 * Rename the object files, i.e. give them other names than in previous
2785 versions. Their names now include the name of the final product, as
2786 well as its type mnemonic (bin, lib, shlib).
2790 * Added new option for 'openssl list', '-objects', which will display the
2791 list of built in objects, i.e. OIDs with names.
2795 * Added the options `-crl_lastupdate` and `-crl_nextupdate` to `openssl ca`,
2796 allowing the `lastUpdate` and `nextUpdate` fields in the generated CRL to
2801 * Added support for Linux Kernel TLS data-path. The Linux Kernel data-path
2802 improves application performance by removing data copies and providing
2803 applications with zero-copy system calls such as sendfile and splice.
2807 * The SSL option SSL_OP_CLEANSE_PLAINTEXT is introduced.
2811 * `PKCS12_parse` now maintains the order of the parsed certificates
2812 when outputting them via `*ca` (rather than reversing it).
2816 * Deprecated pthread fork support methods.
2820 * Added support for FFDHE key exchange in TLS 1.3.
2824 * Added a new concept for OpenSSL plugability: providers. This
2825 functionality is designed to replace the ENGINE API and ENGINE
2826 implementations, and to be much more dynamic, allowing provider
2827 authors to introduce new algorithms among other things, as long as
2828 there's an API that supports the algorithm type.
2830 With this concept comes a new core API for interaction between
2831 libcrypto and provider implementations. Public libcrypto functions
2832 that want to use providers do so through this core API.
2834 The main documentation for this core API is found in
2835 doc/man7/provider.pod, doc/man7/provider-base.pod, and they in turn
2836 refer to other manuals describing the API specific for supported
2837 algorithm types (also called operations).
2844 ### Changes between 1.1.1m and 1.1.1n [xx XXX xxxx]
2846 ### Changes between 1.1.1l and 1.1.1m [14 Dec 2021]
2848 * Avoid loading of a dynamic engine twice.
2852 * Prioritise DANE TLSA issuer certs over peer certs
2856 * Fixed random API for MacOS prior to 10.12
2858 These MacOS versions don't support the CommonCrypto APIs
2862 ### Changes between 1.1.1k and 1.1.1l [24 Aug 2021]
2864 * Fixed an SM2 Decryption Buffer Overflow.
2866 In order to decrypt SM2 encrypted data an application is expected to
2867 call the API function EVP_PKEY_decrypt(). Typically an application will
2868 call this function twice. The first time, on entry, the "out" parameter
2869 can be NULL and, on exit, the "outlen" parameter is populated with the
2870 buffer size required to hold the decrypted plaintext. The application
2871 can then allocate a sufficiently sized buffer and call EVP_PKEY_decrypt()
2872 again, but this time passing a non-NULL value for the "out" parameter.
2874 A bug in the implementation of the SM2 decryption code means that the
2875 calculation of the buffer size required to hold the plaintext returned
2876 by the first call to EVP_PKEY_decrypt() can be smaller than the actual
2877 size required by the second call. This can lead to a buffer overflow
2878 when EVP_PKEY_decrypt() is called by the application a second time with
2879 a buffer that is too small.
2881 A malicious attacker who is able present SM2 content for decryption to
2882 an application could cause attacker chosen data to overflow the buffer
2883 by up to a maximum of 62 bytes altering the contents of other data held
2884 after the buffer, possibly changing application behaviour or causing
2885 the application to crash. The location of the buffer is application
2886 dependent but is typically heap allocated.
2891 * Fixed various read buffer overruns processing ASN.1 strings
2893 ASN.1 strings are represented internally within OpenSSL as an ASN1_STRING
2894 structure which contains a buffer holding the string data and a field
2895 holding the buffer length. This contrasts with normal C strings which
2896 are represented as a buffer for the string data which is terminated
2897 with a NUL (0) byte.
2899 Although not a strict requirement, ASN.1 strings that are parsed using
2900 OpenSSL's own "d2i" functions (and other similar parsing functions) as
2901 well as any string whose value has been set with the ASN1_STRING_set()
2902 function will additionally NUL terminate the byte array in the
2903 ASN1_STRING structure.
2905 However, it is possible for applications to directly construct valid
2906 ASN1_STRING structures which do not NUL terminate the byte array by
2907 directly setting the "data" and "length" fields in the ASN1_STRING
2908 array. This can also happen by using the ASN1_STRING_set0() function.
2910 Numerous OpenSSL functions that print ASN.1 data have been found to
2911 assume that the ASN1_STRING byte array will be NUL terminated, even
2912 though this is not guaranteed for strings that have been directly
2913 constructed. Where an application requests an ASN.1 structure to be
2914 printed, and where that ASN.1 structure contains ASN1_STRINGs that have
2915 been directly constructed by the application without NUL terminating
2916 the "data" field, then a read buffer overrun can occur.
2918 The same thing can also occur during name constraints processing
2919 of certificates (for example if a certificate has been directly
2920 constructed by the application instead of loading it via the OpenSSL
2921 parsing functions, and the certificate contains non NUL terminated
2922 ASN1_STRING structures). It can also occur in the X509_get1_email(),
2923 X509_REQ_get1_email() and X509_get1_ocsp() functions.
2925 If a malicious actor can cause an application to directly construct an
2926 ASN1_STRING and then process it through one of the affected OpenSSL
2927 functions then this issue could be hit. This might result in a crash
2928 (causing a Denial of Service attack). It could also result in the
2929 disclosure of private memory contents (such as private keys, or
2930 sensitive plaintext).
2935 ### Changes between 1.1.1j and 1.1.1k [25 Mar 2021]
2937 * Fixed a problem with verifying a certificate chain when using the
2938 X509_V_FLAG_X509_STRICT flag. This flag enables additional security checks of
2939 the certificates present in a certificate chain. It is not set by default.
2941 Starting from OpenSSL version 1.1.1h a check to disallow certificates in
2942 the chain that have explicitly encoded elliptic curve parameters was added
2943 as an additional strict check.
2945 An error in the implementation of this check meant that the result of a
2946 previous check to confirm that certificates in the chain are valid CA
2947 certificates was overwritten. This effectively bypasses the check
2948 that non-CA certificates must not be able to issue other certificates.
2950 If a "purpose" has been configured then there is a subsequent opportunity
2951 for checks that the certificate is a valid CA. All of the named "purpose"
2952 values implemented in libcrypto perform this check. Therefore, where
2953 a purpose is set the certificate chain will still be rejected even when the
2954 strict flag has been used. A purpose is set by default in libssl client and
2955 server certificate verification routines, but it can be overridden or
2956 removed by an application.
2958 In order to be affected, an application must explicitly set the
2959 X509_V_FLAG_X509_STRICT verification flag and either not set a purpose
2960 for the certificate verification or, in the case of TLS client or server
2961 applications, override the default purpose.
2966 * Fixed an issue where an OpenSSL TLS server may crash if sent a maliciously
2967 crafted renegotiation ClientHello message from a client. If a TLSv1.2
2968 renegotiation ClientHello omits the signature_algorithms extension (where it
2969 was present in the initial ClientHello), but includes a
2970 signature_algorithms_cert extension then a NULL pointer dereference will
2971 result, leading to a crash and a denial of service attack.
2973 A server is only vulnerable if it has TLSv1.2 and renegotiation enabled
2974 (which is the default configuration). OpenSSL TLS clients are not impacted by
2978 *Peter Kästle and Samuel Sapalski*
2980 ### Changes between 1.1.1i and 1.1.1j [16 Feb 2021]
2982 * Fixed the X509_issuer_and_serial_hash() function. It attempts to
2983 create a unique hash value based on the issuer and serial number data
2984 contained within an X509 certificate. However, it was failing to correctly
2985 handle any errors that may occur while parsing the issuer field (which might
2986 occur if the issuer field is maliciously constructed). This may subsequently
2987 result in a NULL pointer deref and a crash leading to a potential denial of
2993 * Fixed the RSA_padding_check_SSLv23() function and the RSA_SSLV23_PADDING
2994 padding mode to correctly check for rollback attacks. This is considered a
2995 bug in OpenSSL 1.1.1 because it does not support SSLv2. In 1.0.2 this is
3000 Fixed the EVP_CipherUpdate, EVP_EncryptUpdate and EVP_DecryptUpdate
3001 functions. Previously they could overflow the output length argument in some
3002 cases where the input length is close to the maximum permissible length for
3003 an integer on the platform. In such cases the return value from the function
3004 call would be 1 (indicating success), but the output length value would be
3005 negative. This could cause applications to behave incorrectly or crash.
3010 * Fixed SRP_Calc_client_key so that it runs in constant time. The previous
3011 implementation called BN_mod_exp without setting BN_FLG_CONSTTIME. This
3012 could be exploited in a side channel attack to recover the password. Since
3013 the attack is local host only this is outside of the current OpenSSL
3014 threat model and therefore no CVE is assigned.
3016 Thanks to Mohammed Sabt and Daniel De Almeida Braga for reporting this
3021 ### Changes between 1.1.1h and 1.1.1i [8 Dec 2020]
3023 * Fixed NULL pointer deref in the GENERAL_NAME_cmp function
3024 This function could crash if both GENERAL_NAMEs contain an EDIPARTYNAME.
3025 If an attacker can control both items being compared then this could lead
3026 to a possible denial of service attack. OpenSSL itself uses the
3027 GENERAL_NAME_cmp function for two purposes:
3028 1) Comparing CRL distribution point names between an available CRL and a
3029 CRL distribution point embedded in an X509 certificate
3030 2) When verifying that a timestamp response token signer matches the
3031 timestamp authority name (exposed via the API functions
3032 TS_RESP_verify_response and TS_RESP_verify_token)
3037 ### Changes between 1.1.1g and 1.1.1h [22 Sep 2020]
3039 * Certificates with explicit curve parameters are now disallowed in
3040 verification chains if the X509_V_FLAG_X509_STRICT flag is used.
3044 * The 'MinProtocol' and 'MaxProtocol' configuration commands now silently
3045 ignore TLS protocol version bounds when configuring DTLS-based contexts, and
3046 conversely, silently ignore DTLS protocol version bounds when configuring
3047 TLS-based contexts. The commands can be repeated to set bounds of both
3048 types. The same applies with the corresponding "min_protocol" and
3049 "max_protocol" command-line switches, in case some application uses both TLS
3052 SSL_CTX instances that are created for a fixed protocol version (e.g.
3053 TLSv1_server_method()) also silently ignore version bounds. Previously
3054 attempts to apply bounds to these protocol versions would result in an
3055 error. Now only the "version-flexible" SSL_CTX instances are subject to
3056 limits in configuration files in command-line options.
3060 * Handshake now fails if Extended Master Secret extension is dropped
3065 * The Oracle Developer Studio compiler will start reporting deprecated APIs
3067 ### Changes between 1.1.1f and 1.1.1g [21 Apr 2020]
3069 * Fixed segmentation fault in SSL_check_chain()
3070 Server or client applications that call the SSL_check_chain() function
3071 during or after a TLS 1.3 handshake may crash due to a NULL pointer
3072 dereference as a result of incorrect handling of the
3073 "signature_algorithms_cert" TLS extension. The crash occurs if an invalid
3074 or unrecognised signature algorithm is received from the peer. This could
3075 be exploited by a malicious peer in a Denial of Service attack.
3080 * Added AES consttime code for no-asm configurations
3081 an optional constant time support for AES was added
3082 when building openssl for no-asm.
3083 Enable with: ./config no-asm -DOPENSSL_AES_CONST_TIME
3084 Disable with: ./config no-asm -DOPENSSL_NO_AES_CONST_TIME
3085 At this time this feature is by default disabled.
3086 It will be enabled by default in 3.0.
3090 ### Changes between 1.1.1e and 1.1.1f [31 Mar 2020]
3092 * Revert the change of EOF detection while reading in libssl to avoid
3093 regressions in applications depending on the current way of reporting
3094 the EOF. As the existing method is not fully accurate the change to
3095 reporting the EOF via SSL_ERROR_SSL is kept on the current development
3096 branch and will be present in the 3.0 release.
3100 * Revised BN_generate_prime_ex to not avoid factors 3..17863 in p-1
3101 when primes for RSA keys are computed.
3102 Since we previously always generated primes == 2 (mod 3) for RSA keys,
3103 the 2-prime and 3-prime RSA modules were easy to distinguish, since
3104 N = p*q = 1 (mod 3), but N = p*q*r = 2 (mod 3). Therefore, fingerprinting
3105 2-prime vs. 3-prime RSA keys was possible by computing N mod 3.
3106 This avoids possible fingerprinting of newly generated RSA modules.
3110 ### Changes between 1.1.1d and 1.1.1e [17 Mar 2020]
3112 * Properly detect EOF while reading in libssl. Previously if we hit an EOF
3113 while reading in libssl then we would report an error back to the
3114 application (SSL_ERROR_SYSCALL) but errno would be 0. We now add
3115 an error to the stack (which means we instead return SSL_ERROR_SSL) and
3116 therefore give a hint as to what went wrong.
3120 * Check that ed25519 and ed448 are allowed by the security level. Previously
3121 signature algorithms not using an MD were not being checked that they were
3122 allowed by the security level.
3126 * Fixed SSL_get_servername() behaviour. The behaviour of SSL_get_servername()
3127 was not quite right. The behaviour was not consistent between resumption
3128 and normal handshakes, and also not quite consistent with historical
3129 behaviour. The behaviour in various scenarios has been clarified and
3130 it has been updated to make it match historical behaviour as closely as
3135 * *[VMS only]* The header files that the VMS compilers include automatically,
3136 `__DECC_INCLUDE_PROLOGUE.H` and `__DECC_INCLUDE_EPILOGUE.H`, use pragmas
3137 that the C++ compiler doesn't understand. This is a shortcoming in the
3138 compiler, but can be worked around with `__cplusplus` guards.
3140 C++ applications that use OpenSSL libraries must be compiled using the
3141 qualifier `/NAMES=(AS_IS,SHORTENED)` to be able to use all the OpenSSL
3142 functions. Otherwise, only functions with symbols of less than 31
3143 characters can be used, as the linker will not be able to successfully
3144 resolve symbols with longer names.
3148 * Added a new method to gather entropy on VMS, based on SYS$GET_ENTROPY.
3149 The presence of this system service is determined at run-time.
3153 * Print all values for a PKCS#12 attribute with 'openssl pkcs12', not just
3158 ### Changes between 1.1.1c and 1.1.1d [10 Sep 2019]
3160 * Fixed a fork protection issue. OpenSSL 1.1.1 introduced a rewritten random
3161 number generator (RNG). This was intended to include protection in the
3162 event of a fork() system call in order to ensure that the parent and child
3163 processes did not share the same RNG state. However, this protection was not
3164 being used in the default case.
3166 A partial mitigation for this issue is that the output from a high
3167 precision timer is mixed into the RNG state so the likelihood of a parent
3168 and child process sharing state is significantly reduced.
3170 If an application already calls OPENSSL_init_crypto() explicitly using
3171 OPENSSL_INIT_ATFORK then this problem does not occur at all.
3174 *Matthias St. Pierre*
3176 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
3177 used even when parsing explicit parameters, when loading a encoded key
3178 or calling `EC_GROUP_new_from_ecpkparameters()`/
3179 `EC_GROUP_new_from_ecparameters()`.
3180 This prevents bypass of security hardening and performance gains,
3181 especially for curves with specialized EC_METHODs.
3182 By default, if a key encoded with explicit parameters is loaded and later
3183 encoded, the output is still encoded with explicit parameters, even if
3184 internally a "named" EC_GROUP is used for computation.
3188 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
3189 this change, EC_GROUP_set_generator would accept order and/or cofactor as
3190 NULL. After this change, only the cofactor parameter can be NULL. It also
3191 does some minimal sanity checks on the passed order.
3196 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
3197 An attack is simple, if the first CMS_recipientInfo is valid but the
3198 second CMS_recipientInfo is chosen ciphertext. If the second
3199 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
3200 encryption key will be replaced by garbage, and the message cannot be
3201 decoded, but if the RSA decryption fails, the correct encryption key is
3202 used and the recipient will not notice the attack.
3203 As a work around for this potential attack the length of the decrypted
3204 key must be equal to the cipher default key length, in case the
3205 certificate is not given and all recipientInfo are tried out.
3206 The old behaviour can be re-enabled in the CMS code by setting the
3207 CMS_DEBUG_DECRYPT flag.
3212 * Early start up entropy quality from the DEVRANDOM seed source has been
3213 improved for older Linux systems. The RAND subsystem will wait for
3214 /dev/random to be producing output before seeding from /dev/urandom.
3215 The seeded state is stored for future library initialisations using
3216 a system global shared memory segment. The shared memory identifier
3217 can be configured by defining OPENSSL_RAND_SEED_DEVRANDOM_SHM_ID to
3218 the desired value. The default identifier is 114.
3222 * Correct the extended master secret constant on EBCDIC systems. Without this
3223 fix TLS connections between an EBCDIC system and a non-EBCDIC system that
3224 negotiate EMS will fail. Unfortunately this also means that TLS connections
3225 between EBCDIC systems with this fix, and EBCDIC systems without this
3226 fix will fail if they negotiate EMS.
3230 * Use Windows installation paths in the mingw builds
3232 Mingw isn't a POSIX environment per se, which means that Windows
3233 paths should be used for installation.
3238 * Changed DH_check to accept parameters with order q and 2q subgroups.
3239 With order 2q subgroups the bit 0 of the private key is not secret
3240 but DH_generate_key works around that by clearing bit 0 of the
3241 private key for those. This avoids leaking bit 0 of the private key.
3245 * Significantly reduce secure memory usage by the randomness pools.
3249 * Revert the DEVRANDOM_WAIT feature for Linux systems
3251 The DEVRANDOM_WAIT feature added a select() call to wait for the
3252 /dev/random device to become readable before reading from the
3253 /dev/urandom device.
3255 It turned out that this change had negative side effects on
3256 performance which were not acceptable. After some discussion it
3257 was decided to revert this feature and leave it up to the OS
3258 resp. the platform maintainer to ensure a proper initialization
3259 during early boot time.
3261 *Matthias St. Pierre*
3263 ### Changes between 1.1.1b and 1.1.1c [28 May 2019]
3265 * Add build tests for C++. These are generated files that only do one
3266 thing, to include one public OpenSSL head file each. This tests that
3267 the public header files can be usefully included in a C++ application.
3269 This test isn't enabled by default. It can be enabled with the option
3270 'enable-buildtest-c++'.
3274 * Enable SHA3 pre-hashing for ECDSA and DSA.
3278 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
3279 This changes the size when using the `genpkey` command when no size is given.
3280 It fixes an omission in earlier changes that changed all RSA, DSA and DH
3281 generation commands to use 2048 bits by default.
3285 * Reorganize the manual pages to consistently have RETURN VALUES,
3286 EXAMPLES, SEE ALSO and HISTORY come in that order, and adjust
3287 util/fix-doc-nits accordingly.
3289 *Paul Yang, Joshua Lock*
3291 * Add the missing accessor EVP_PKEY_get0_engine()
3295 * Have commands like `s_client` and `s_server` output the signature scheme
3296 along with other cipher suite parameters when debugging.
3300 * Make OPENSSL_config() error agnostic again.
3304 * Do the error handling in RSA decryption constant time.
3308 * Prevent over long nonces in ChaCha20-Poly1305.
3310 ChaCha20-Poly1305 is an AEAD cipher, and requires a unique nonce input
3311 for every encryption operation. RFC 7539 specifies that the nonce value
3312 (IV) should be 96 bits (12 bytes). OpenSSL allows a variable nonce length
3313 and front pads the nonce with 0 bytes if it is less than 12
3314 bytes. However it also incorrectly allows a nonce to be set of up to 16
3315 bytes. In this case only the last 12 bytes are significant and any
3316 additional leading bytes are ignored.
3318 It is a requirement of using this cipher that nonce values are
3319 unique. Messages encrypted using a reused nonce value are susceptible to
3320 serious confidentiality and integrity attacks. If an application changes
3321 the default nonce length to be longer than 12 bytes and then makes a
3322 change to the leading bytes of the nonce expecting the new value to be a
3323 new unique nonce then such an application could inadvertently encrypt
3324 messages with a reused nonce.
3326 Additionally the ignored bytes in a long nonce are not covered by the
3327 integrity guarantee of this cipher. Any application that relies on the
3328 integrity of these ignored leading bytes of a long nonce may be further
3329 affected. Any OpenSSL internal use of this cipher, including in SSL/TLS,
3330 is safe because no such use sets such a long nonce value. However user
3331 applications that use this cipher directly and set a non-default nonce
3332 length to be longer than 12 bytes may be vulnerable.
3334 This issue was reported to OpenSSL on 16th of March 2019 by Joran Dirk
3340 * Add DEVRANDOM_WAIT feature for Linux systems
3342 On older Linux systems where the getrandom() system call is not available,
3343 OpenSSL normally uses the /dev/urandom device for seeding its CSPRNG.
3344 Contrary to getrandom(), the /dev/urandom device will not block during
3345 early boot when the kernel CSPRNG has not been seeded yet.
3347 To mitigate this known weakness, use select() to wait for /dev/random to
3348 become readable before reading from /dev/urandom.
3350 * Ensure that SM2 only uses SM3 as digest algorithm
3354 ### Changes between 1.1.1a and 1.1.1b [26 Feb 2019]
3356 * Change the info callback signals for the start and end of a post-handshake
3357 message exchange in TLSv1.3. In 1.1.1/1.1.1a we used SSL_CB_HANDSHAKE_START
3358 and SSL_CB_HANDSHAKE_DONE. Experience has shown that many applications get
3359 confused by this and assume that a TLSv1.2 renegotiation has started. This
3360 can break KeyUpdate handling. Instead we no longer signal the start and end
3361 of a post handshake message exchange (although the messages themselves are
3362 still signalled). This could break some applications that were expecting
3363 the old signals. However without this KeyUpdate is not usable for many
3368 ### Changes between 1.1.1 and 1.1.1a [20 Nov 2018]
3370 * Timing vulnerability in DSA signature generation
3372 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
3373 timing side channel attack. An attacker could use variations in the signing
3374 algorithm to recover the private key.
3376 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
3381 * Timing vulnerability in ECDSA signature generation
3383 The OpenSSL ECDSA signature algorithm has been shown to be vulnerable to a
3384 timing side channel attack. An attacker could use variations in the signing
3385 algorithm to recover the private key.
3387 This issue was reported to OpenSSL on 25th October 2018 by Samuel Weiser.
3392 * Fixed the issue that RAND_add()/RAND_seed() silently discards random input
3393 if its length exceeds 4096 bytes. The limit has been raised to a buffer size
3394 of two gigabytes and the error handling improved.
3396 This issue was reported to OpenSSL by Dr. Falko Strenzke. It has been
3397 categorized as a normal bug, not a security issue, because the DRBG reseeds
3398 automatically and is fully functional even without additional randomness
3399 provided by the application.
3401 ### Changes between 1.1.0i and 1.1.1 [11 Sep 2018]
3403 * Add a new ClientHello callback. Provides a callback interface that gives
3404 the application the ability to adjust the nascent SSL object at the
3405 earliest stage of ClientHello processing, immediately after extensions have
3406 been collected but before they have been processed. In particular, this
3407 callback can adjust the supported TLS versions in response to the contents
3412 * Add SM2 base algorithm support.
3416 * s390x assembly pack: add (improved) hardware-support for the following
3417 cryptographic primitives: sha3, shake, aes-gcm, aes-ccm, aes-ctr, aes-ofb,
3418 aes-cfb/cfb8, aes-ecb.
3422 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
3423 parameter is no longer accepted, as it leads to a corrupt table. NULL
3424 pem_str is reserved for alias entries only.
3428 * Use the new ec_scalar_mul_ladder scaffold to implement a specialized ladder
3429 step for prime curves. The new implementation is based on formulae from
3430 differential addition-and-doubling in homogeneous projective coordinates
3431 from Izu-Takagi "A fast parallel elliptic curve multiplication resistant
3432 against side channel attacks" and Brier-Joye "Weierstrass Elliptic Curves
3433 and Side-Channel Attacks" Eq. (8) for y-coordinate recovery, modified
3434 to work in projective coordinates.
3436 *Billy Bob Brumley, Nicola Tuveri*
3438 * Change generating and checking of primes so that the error rate of not
3439 being prime depends on the intended use based on the size of the input.
3440 For larger primes this will result in more rounds of Miller-Rabin.
3441 The maximal error rate for primes with more than 1080 bits is lowered
3444 *Kurt Roeckx, Annie Yousar*
3446 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
3450 * The 'tsget' script is renamed to 'tsget.pl', to avoid confusion when
3451 moving between systems, and to avoid confusion when a Windows build is
3452 done with mingw vs with MSVC. For POSIX installs, there's still a
3453 symlink or copy named 'tsget' to avoid that confusion as well.
3457 * Revert blinding in ECDSA sign and instead make problematic addition
3458 length-invariant. Switch even to fixed-length Montgomery multiplication.
3462 * Use the new ec_scalar_mul_ladder scaffold to implement a specialized ladder
3463 step for binary curves. The new implementation is based on formulae from
3464 differential addition-and-doubling in mixed Lopez-Dahab projective
3465 coordinates, modified to independently blind the operands.
3467 *Billy Bob Brumley, Sohaib ul Hassan, Nicola Tuveri*
3469 * Add a scaffold to optionally enhance the Montgomery ladder implementation
3470 for `ec_scalar_mul_ladder` (formerly `ec_mul_consttime`) allowing
3471 EC_METHODs to implement their own specialized "ladder step", to take
3472 advantage of more favorable coordinate systems or more efficient
3473 differential addition-and-doubling algorithms.
3475 *Billy Bob Brumley, Sohaib ul Hassan, Nicola Tuveri*
3477 * Modified the random device based seed sources to keep the relevant
3478 file descriptors open rather than reopening them on each access.
3479 This allows such sources to operate in a chroot() jail without
3480 the associated device nodes being available. This behaviour can be
3481 controlled using RAND_keep_random_devices_open().
3485 * Numerous side-channel attack mitigations have been applied. This may have
3486 performance impacts for some algorithms for the benefit of improved
3487 security. Specific changes are noted in this change log by their respective
3492 * AIX shared library support overhaul. Switch to AIX "natural" way of
3493 handling shared libraries, which means collecting shared objects of
3494 different versions and bitnesses in one common archive. This allows to
3495 mitigate conflict between 1.0 and 1.1 side-by-side installations. It
3496 doesn't affect the way 3rd party applications are linked, only how
3497 multi-version installation is managed.
3501 * Make ec_group_do_inverse_ord() more robust and available to other
3502 EC cryptosystems, so that irrespective of BN_FLG_CONSTTIME, SCA
3503 mitigations are applied to the fallback BN_mod_inverse().
3504 When using this function rather than BN_mod_inverse() directly, new
3505 EC cryptosystem implementations are then safer-by-default.
3509 * Add coordinate blinding for EC_POINT and implement projective
3510 coordinate blinding for generic prime curves as a countermeasure to
3511 chosen point SCA attacks.
3513 *Sohaib ul Hassan, Nicola Tuveri, Billy Bob Brumley*
3515 * Add blinding to ECDSA and DSA signatures to protect against side channel
3516 attacks discovered by Keegan Ryan (NCC Group).
3520 * Enforce checking in the `pkeyutl` command to ensure that the input
3521 length does not exceed the maximum supported digest length when performing
3522 a sign, verify or verifyrecover operation.
3526 * SSL_MODE_AUTO_RETRY is enabled by default. Applications that use blocking
3527 I/O in combination with something like select() or poll() will hang. This
3528 can be turned off again using SSL_CTX_clear_mode().
3529 Many applications do not properly handle non-application data records, and
3530 TLS 1.3 sends more of such records. Setting SSL_MODE_AUTO_RETRY works
3531 around the problems in those applications, but can also break some.
3532 It's recommended to read the manpages about SSL_read(), SSL_write(),
3533 SSL_get_error(), SSL_shutdown(), SSL_CTX_set_mode() and
3534 SSL_CTX_set_read_ahead() again.
3538 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
3539 now allow empty (zero character) pass phrases.
3543 * Apply blinding to binary field modular inversion and remove patent
3544 pending (OPENSSL_SUN_GF2M_DIV) BN_GF2m_mod_div implementation.
3548 * Deprecate ec2_mult.c and unify scalar multiplication code paths for
3549 binary and prime elliptic curves.
3553 * Remove ECDSA nonce padding: EC_POINT_mul is now responsible for
3554 constant time fixed point multiplication.
3558 * Revise elliptic curve scalar multiplication with timing attack
3559 defenses: ec_wNAF_mul redirects to a constant time implementation
3560 when computing fixed point and variable point multiplication (which
3561 in OpenSSL are mostly used with secret scalars in keygen, sign,
3562 ECDH derive operations).
3563 *Billy Bob Brumley, Nicola Tuveri, Cesar Pereida García,
3566 * Updated CONTRIBUTING
3570 * Updated DRBG / RAND to request nonce and additional low entropy
3571 randomness from the system.
3573 *Matthias St. Pierre*
3575 * Updated 'openssl rehash' to use OpenSSL consistent default.
3579 * Moved the load of the ssl_conf module to libcrypto, which helps
3580 loading engines that libssl uses before libssl is initialised.
3584 * Added EVP_PKEY_sign() and EVP_PKEY_verify() for EdDSA
3588 * Fixed X509_NAME_ENTRY_set to get multi-valued RDNs right in all cases.
3590 *Ingo Schwarze, Rich Salz*
3592 * Added output of accepting IP address and port for 'openssl s_server'
3596 * Added a new API for TLSv1.3 ciphersuites:
3597 SSL_CTX_set_ciphersuites()
3598 SSL_set_ciphersuites()
3602 * Memory allocation failures consistently add an error to the error
3607 * Don't use OPENSSL_ENGINES and OPENSSL_CONF environment values
3608 in libcrypto when run as setuid/setgid.
3612 * Load any config file by default when libssl is used.
3616 * Added new public header file <openssl/rand_drbg.h> and documentation
3617 for the RAND_DRBG API. See manual page RAND_DRBG(7) for an overview.
3619 *Matthias St. Pierre*
3621 * QNX support removed (cannot find contributors to get their approval
3622 for the license change).
3626 * TLSv1.3 replay protection for early data has been implemented. See the
3627 SSL_read_early_data() man page for further details.
3631 * Separated TLSv1.3 ciphersuite configuration out from TLSv1.2 ciphersuite
3632 configuration. TLSv1.3 ciphersuites are not compatible with TLSv1.2 and
3633 below. Similarly TLSv1.2 ciphersuites are not compatible with TLSv1.3.
3634 In order to avoid issues where legacy TLSv1.2 ciphersuite configuration
3635 would otherwise inadvertently disable all TLSv1.3 ciphersuites the
3636 configuration has been separated out. See the ciphers man page or the
3637 SSL_CTX_set_ciphersuites() man page for more information.
3641 * On POSIX (BSD, Linux, ...) systems the ocsp(1) command running
3642 in responder mode now supports the new "-multi" option, which
3643 spawns the specified number of child processes to handle OCSP
3644 requests. The "-timeout" option now also limits the OCSP
3645 responder's patience to wait to receive the full client request
3646 on a newly accepted connection. Child processes are respawned
3647 as needed, and the CA index file is automatically reloaded
3648 when changed. This makes it possible to run the "ocsp" responder
3649 as a long-running service, making the OpenSSL CA somewhat more
3650 feature-complete. In this mode, most diagnostic messages logged
3651 after entering the event loop are logged via syslog(3) rather than
3656 * Added support for X448 and Ed448. Heavily based on original work by
3661 * Extend OSSL_STORE with capabilities to search and to narrow the set of
3662 objects loaded. This adds the functions OSSL_STORE_expect() and
3663 OSSL_STORE_find() as well as needed tools to construct searches and
3664 get the search data out of them.
3668 * Support for TLSv1.3 added. Note that users upgrading from an earlier
3669 version of OpenSSL should review their configuration settings to ensure
3670 that they are still appropriate for TLSv1.3. For further information see:
3671 <https://wiki.openssl.org/index.php/TLS1.3>
3675 * Grand redesign of the OpenSSL random generator
3677 The default RAND method now utilizes an AES-CTR DRBG according to
3678 NIST standard SP 800-90Ar1. The new random generator is essentially
3679 a port of the default random generator from the OpenSSL FIPS 2.0
3680 object module. It is a hybrid deterministic random bit generator
3681 using an AES-CTR bit stream and which seeds and reseeds itself
3682 automatically using trusted system entropy sources.
3684 Some of its new features are:
3685 - Support for multiple DRBG instances with seed chaining.
3686 - The default RAND method makes use of a DRBG.
3687 - There is a public and private DRBG instance.
3688 - The DRBG instances are fork-safe.
3689 - Keep all global DRBG instances on the secure heap if it is enabled.
3690 - The public and private DRBG instance are per thread for lock free
3693 *Paul Dale, Benjamin Kaduk, Kurt Roeckx, Rich Salz, Matthias St. Pierre*
3695 * Changed Configure so it only says what it does and doesn't dump
3696 so much data. Instead, ./configdata.pm should be used as a script
3697 to display all sorts of configuration data.
3701 * Added processing of "make variables" to Configure.
3705 * Added SHA512/224 and SHA512/256 algorithm support.
3709 * The last traces of Netware support, first removed in 1.1.0, have
3714 * Get rid of Makefile.shared, and in the process, make the processing
3715 of certain files (rc.obj, or the .def/.map/.opt files produced from
3716 the ordinal files) more visible and hopefully easier to trace and
3717 debug (or make silent).
3721 * Make it possible to have environment variable assignments as
3722 arguments to config / Configure.
3726 * Add multi-prime RSA (RFC 8017) support.
3730 * Add SM3 implemented according to GB/T 32905-2016
3731 *Jack Lloyd <jack.lloyd@ribose.com>,*
3732 *Ronald Tse <ronald.tse@ribose.com>,*
3733 *Erick Borsboom <erick.borsboom@ribose.com>*
3735 * Add 'Maximum Fragment Length' TLS extension negotiation and support
3736 as documented in RFC6066.
3737 Based on a patch from Tomasz Moń
3739 *Filipe Raimundo da Silva*
3741 * Add SM4 implemented according to GB/T 32907-2016.
3742 *Jack Lloyd <jack.lloyd@ribose.com>,*
3743 *Ronald Tse <ronald.tse@ribose.com>,*
3744 *Erick Borsboom <erick.borsboom@ribose.com>*
3746 * Reimplement -newreq-nodes and ERR_error_string_n; the
3747 original author does not agree with the license change.
3751 * Add ARIA AEAD TLS support.
3755 * Some macro definitions to support VS6 have been removed. Visual
3756 Studio 6 has not worked since 1.1.0
3760 * Add ERR_clear_last_mark(), to allow callers to clear the last mark
3761 without clearing the errors.
3765 * Add "atfork" functions. If building on a system that without
3766 pthreads, see doc/man3/OPENSSL_fork_prepare.pod for application
3767 requirements. The RAND facility now uses/requires this.
3775 * The UI API becomes a permanent and integral part of libcrypto, i.e.
3776 not possible to disable entirely. However, it's still possible to
3777 disable the console reading UI method, UI_OpenSSL() (use UI_null()
3780 To disable, configure with 'no-ui-console'. 'no-ui' is still
3781 possible to use as an alias. Check at compile time with the
3782 macro OPENSSL_NO_UI_CONSOLE. The macro OPENSSL_NO_UI is still
3783 possible to check and is an alias for OPENSSL_NO_UI_CONSOLE.
3787 * Add a STORE module, which implements a uniform and URI based reader of
3788 stores that can contain keys, certificates, CRLs and numerous other
3789 objects. The main API is loosely based on a few stdio functions,
3790 and includes OSSL_STORE_open, OSSL_STORE_load, OSSL_STORE_eof,
3791 OSSL_STORE_error and OSSL_STORE_close.
3792 The implementation uses backends called "loaders" to implement arbitrary
3793 URI schemes. There is one built in "loader" for the 'file' scheme.
3797 * Add devcrypto engine. This has been implemented against cryptodev-linux,
3798 then adjusted to work on FreeBSD 8.4 as well.
3799 Enable by configuring with 'enable-devcryptoeng'. This is done by default
3800 on BSD implementations, as cryptodev.h is assumed to exist on all of them.
3804 * Module names can prefixed with OSSL_ or OPENSSL_. This affects
3805 util/mkerr.pl, which is adapted to allow those prefixes, leading to
3806 error code calls like this:
3808 OSSL_FOOerr(OSSL_FOO_F_SOMETHING, OSSL_FOO_R_WHATEVER);
3810 With this change, we claim the namespaces OSSL and OPENSSL in a manner
3811 that can be encoded in C. For the foreseeable future, this will only
3814 *Richard Levitte and Tim Hudson*
3816 * Removed BSD cryptodev engine.
3820 * Add a build target 'build_all_generated', to build all generated files
3821 and only that. This can be used to prepare everything that requires
3822 things like perl for a system that lacks perl and then move everything
3823 to that system and do the rest of the build there.
3827 * In the UI interface, make it possible to duplicate the user data. This
3828 can be used by engines that need to retain the data for a longer time
3829 than just the call where this user data is passed.
3833 * Ignore the '-named_curve auto' value for compatibility of applications
3836 *Tomáš Mráz <tmraz@fedoraproject.org>*
3838 * Fragmented SSL/TLS alerts are no longer accepted. An alert message is 2
3839 bytes long. In theory it is permissible in SSLv3 - TLSv1.2 to fragment such
3840 alerts across multiple records (some of which could be empty). In practice
3841 it make no sense to send an empty alert record, or to fragment one. TLSv1.3
3842 prohibits this altogether and other libraries (BoringSSL, NSS) do not
3843 support this at all. Supporting it adds significant complexity to the
3844 record layer, and its removal is unlikely to cause interoperability
3849 * Add the ASN.1 types INT32, UINT32, INT64, UINT64 and variants prefixed
3850 with Z. These are meant to replace LONG and ZLONG and to be size safe.
3851 The use of LONG and ZLONG is discouraged and scheduled for deprecation
3856 * Add the 'z' and 'j' modifiers to BIO_printf() et al formatting string,
3857 'z' is to be used for [s]size_t, and 'j' - with [u]int64_t.
3859 *Richard Levitte, Andy Polyakov*
3861 * Add EC_KEY_get0_engine(), which does for EC_KEY what RSA_get0_engine()
3866 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
3867 platform rather than 'mingw'.
3871 * The functions X509_STORE_add_cert and X509_STORE_add_crl return
3872 success if they are asked to add an object which already exists
3873 in the store. This change cascades to other functions which load
3874 certificates and CRLs.
3878 * x86_64 assembly pack: annotate code with DWARF CFI directives to
3879 facilitate stack unwinding even from assembly subroutines.
3883 * Remove VAX C specific definitions of OPENSSL_EXPORT, OPENSSL_EXTERN.
3884 Also remove OPENSSL_GLOBAL entirely, as it became a no-op.
3888 * Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
3889 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
3890 which is the minimum version we support.
3894 * Certificate time validation (X509_cmp_time) enforces stricter
3895 compliance with RFC 5280. Fractional seconds and timezone offsets
3896 are no longer allowed.
3900 * Add support for ARIA
3904 * s_client will now send the Server Name Indication (SNI) extension by
3905 default unless the new "-noservername" option is used. The server name is
3906 based on the host provided to the "-connect" option unless overridden by
3907 using "-servername".
3911 * Add support for SipHash
3915 * OpenSSL now fails if it receives an unrecognised record type in TLS1.0
3916 or TLS1.1. Previously this only happened in SSLv3 and TLS1.2. This is to
3917 prevent issues where no progress is being made and the peer continually
3918 sends unrecognised record types, using up resources processing them.
3922 * 'openssl passwd' can now produce SHA256 and SHA512 based output,
3923 using the algorithm defined in
3924 <https://www.akkadia.org/drepper/SHA-crypt.txt>
3928 * Heartbeat support has been removed; the ABI is changed for now.
3930 *Richard Levitte, Rich Salz*
3932 * Support for SSL_OP_NO_ENCRYPT_THEN_MAC in SSL_CONF_cmd.
3936 * The RSA "null" method, which was partially supported to avoid patent
3937 issues, has been replaced to always returns NULL.
3944 ### Changes between 1.1.0k and 1.1.0l [10 Sep 2019]
3946 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
3947 used even when parsing explicit parameters, when loading a encoded key
3948 or calling `EC_GROUP_new_from_ecpkparameters()`/
3949 `EC_GROUP_new_from_ecparameters()`.
3950 This prevents bypass of security hardening and performance gains,
3951 especially for curves with specialized EC_METHODs.
3952 By default, if a key encoded with explicit parameters is loaded and later
3953 encoded, the output is still encoded with explicit parameters, even if
3954 internally a "named" EC_GROUP is used for computation.
3958 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
3959 this change, EC_GROUP_set_generator would accept order and/or cofactor as
3960 NULL. After this change, only the cofactor parameter can be NULL. It also
3961 does some minimal sanity checks on the passed order.
3966 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
3967 An attack is simple, if the first CMS_recipientInfo is valid but the
3968 second CMS_recipientInfo is chosen ciphertext. If the second
3969 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
3970 encryption key will be replaced by garbage, and the message cannot be
3971 decoded, but if the RSA decryption fails, the correct encryption key is
3972 used and the recipient will not notice the attack.
3973 As a work around for this potential attack the length of the decrypted
3974 key must be equal to the cipher default key length, in case the
3975 certificate is not given and all recipientInfo are tried out.
3976 The old behaviour can be re-enabled in the CMS code by setting the
3977 CMS_DEBUG_DECRYPT flag.
3982 * Use Windows installation paths in the mingw builds
3984 Mingw isn't a POSIX environment per se, which means that Windows
3985 paths should be used for installation.
3990 ### Changes between 1.1.0j and 1.1.0k [28 May 2019]
3992 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
3993 This changes the size when using the `genpkey` command when no size is given.
3994 It fixes an omission in earlier changes that changed all RSA, DSA and DH
3995 generation commands to use 2048 bits by default.
3999 * Prevent over long nonces in ChaCha20-Poly1305.
4001 ChaCha20-Poly1305 is an AEAD cipher, and requires a unique nonce input
4002 for every encryption operation. RFC 7539 specifies that the nonce value
4003 (IV) should be 96 bits (12 bytes). OpenSSL allows a variable nonce length
4004 and front pads the nonce with 0 bytes if it is less than 12
4005 bytes. However it also incorrectly allows a nonce to be set of up to 16
4006 bytes. In this case only the last 12 bytes are significant and any
4007 additional leading bytes are ignored.
4009 It is a requirement of using this cipher that nonce values are
4010 unique. Messages encrypted using a reused nonce value are susceptible to
4011 serious confidentiality and integrity attacks. If an application changes
4012 the default nonce length to be longer than 12 bytes and then makes a
4013 change to the leading bytes of the nonce expecting the new value to be a
4014 new unique nonce then such an application could inadvertently encrypt
4015 messages with a reused nonce.
4017 Additionally the ignored bytes in a long nonce are not covered by the
4018 integrity guarantee of this cipher. Any application that relies on the
4019 integrity of these ignored leading bytes of a long nonce may be further
4020 affected. Any OpenSSL internal use of this cipher, including in SSL/TLS,
4021 is safe because no such use sets such a long nonce value. However user
4022 applications that use this cipher directly and set a non-default nonce
4023 length to be longer than 12 bytes may be vulnerable.
4025 This issue was reported to OpenSSL on 16th of March 2019 by Joran Dirk
4031 * Added SCA hardening for modular field inversion in EC_GROUP through
4032 a new dedicated field_inv() pointer in EC_METHOD.
4033 This also addresses a leakage affecting conversions from projective
4034 to affine coordinates.
4036 *Billy Bob Brumley, Nicola Tuveri*
4038 * Fix a use after free bug in d2i_X509_PUBKEY when overwriting a
4039 re-used X509_PUBKEY object if the second PUBKEY is malformed.
4043 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
4047 * Remove the 'dist' target and add a tarball building script. The
4048 'dist' target has fallen out of use, and it shouldn't be
4049 necessary to configure just to create a source distribution.
4053 ### Changes between 1.1.0i and 1.1.0j [20 Nov 2018]
4055 * Timing vulnerability in DSA signature generation
4057 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
4058 timing side channel attack. An attacker could use variations in the signing
4059 algorithm to recover the private key.
4061 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
4066 * Timing vulnerability in ECDSA signature generation
4068 The OpenSSL ECDSA signature algorithm has been shown to be vulnerable to a
4069 timing side channel attack. An attacker could use variations in the signing
4070 algorithm to recover the private key.
4072 This issue was reported to OpenSSL on 25th October 2018 by Samuel Weiser.
4077 * Add coordinate blinding for EC_POINT and implement projective
4078 coordinate blinding for generic prime curves as a countermeasure to
4079 chosen point SCA attacks.
4081 *Sohaib ul Hassan, Nicola Tuveri, Billy Bob Brumley*
4083 ### Changes between 1.1.0h and 1.1.0i [14 Aug 2018]
4085 * Client DoS due to large DH parameter
4087 During key agreement in a TLS handshake using a DH(E) based ciphersuite a
4088 malicious server can send a very large prime value to the client. This will
4089 cause the client to spend an unreasonably long period of time generating a
4090 key for this prime resulting in a hang until the client has finished. This
4091 could be exploited in a Denial Of Service attack.
4093 This issue was reported to OpenSSL on 5th June 2018 by Guido Vranken
4098 * Cache timing vulnerability in RSA Key Generation
4100 The OpenSSL RSA Key generation algorithm has been shown to be vulnerable to
4101 a cache timing side channel attack. An attacker with sufficient access to
4102 mount cache timing attacks during the RSA key generation process could
4103 recover the private key.
4105 This issue was reported to OpenSSL on 4th April 2018 by Alejandro Cabrera
4106 Aldaya, Billy Brumley, Cesar Pereida Garcia and Luis Manuel Alvarez Tapia.
4111 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
4112 parameter is no longer accepted, as it leads to a corrupt table. NULL
4113 pem_str is reserved for alias entries only.
4117 * Revert blinding in ECDSA sign and instead make problematic addition
4118 length-invariant. Switch even to fixed-length Montgomery multiplication.
4122 * Change generating and checking of primes so that the error rate of not
4123 being prime depends on the intended use based on the size of the input.
4124 For larger primes this will result in more rounds of Miller-Rabin.
4125 The maximal error rate for primes with more than 1080 bits is lowered
4128 *Kurt Roeckx, Annie Yousar*
4130 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
4134 * Add blinding to ECDSA and DSA signatures to protect against side channel
4135 attacks discovered by Keegan Ryan (NCC Group).
4139 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
4140 now allow empty (zero character) pass phrases.
4144 * Certificate time validation (X509_cmp_time) enforces stricter
4145 compliance with RFC 5280. Fractional seconds and timezone offsets
4146 are no longer allowed.
4150 * Fixed a text canonicalisation bug in CMS
4152 Where a CMS detached signature is used with text content the text goes
4153 through a canonicalisation process first prior to signing or verifying a
4154 signature. This process strips trailing space at the end of lines, converts
4155 line terminators to CRLF and removes additional trailing line terminators
4156 at the end of a file. A bug in the canonicalisation process meant that
4157 some characters, such as form-feed, were incorrectly treated as whitespace
4158 and removed. This is contrary to the specification (RFC5485). This fix
4159 could mean that detached text data signed with an earlier version of
4160 OpenSSL 1.1.0 may fail to verify using the fixed version, or text data
4161 signed with a fixed OpenSSL may fail to verify with an earlier version of
4162 OpenSSL 1.1.0. A workaround is to only verify the canonicalised text data
4163 and use the "-binary" flag (for the "cms" command line application) or set
4164 the SMIME_BINARY/PKCS7_BINARY/CMS_BINARY flags (if using CMS_verify()).
4168 ### Changes between 1.1.0g and 1.1.0h [27 Mar 2018]
4170 * Constructed ASN.1 types with a recursive definition could exceed the stack
4172 Constructed ASN.1 types with a recursive definition (such as can be found
4173 in PKCS7) could eventually exceed the stack given malicious input with
4174 excessive recursion. This could result in a Denial Of Service attack. There
4175 are no such structures used within SSL/TLS that come from untrusted sources
4176 so this is considered safe.
4178 This issue was reported to OpenSSL on 4th January 2018 by the OSS-fuzz
4184 * Incorrect CRYPTO_memcmp on HP-UX PA-RISC
4186 Because of an implementation bug the PA-RISC CRYPTO_memcmp function is
4187 effectively reduced to only comparing the least significant bit of each
4188 byte. This allows an attacker to forge messages that would be considered as
4189 authenticated in an amount of tries lower than that guaranteed by the
4190 security claims of the scheme. The module can only be compiled by the
4191 HP-UX assembler, so that only HP-UX PA-RISC targets are affected.
4193 This issue was reported to OpenSSL on 2nd March 2018 by Peter Waltenberg
4199 * Add a build target 'build_all_generated', to build all generated files
4200 and only that. This can be used to prepare everything that requires
4201 things like perl for a system that lacks perl and then move everything
4202 to that system and do the rest of the build there.
4206 * Backport SSL_OP_NO_RENGOTIATION
4208 OpenSSL 1.0.2 and below had the ability to disable renegotiation using the
4209 (undocumented) SSL3_FLAGS_NO_RENEGOTIATE_CIPHERS flag. Due to the opacity
4210 changes this is no longer possible in 1.1.0. Therefore, the new
4211 SSL_OP_NO_RENEGOTIATION option from 1.1.1-dev has been backported to
4212 1.1.0 to provide equivalent functionality.
4214 Note that if an application built against 1.1.0h headers (or above) is run
4215 using an older version of 1.1.0 (prior to 1.1.0h) then the option will be
4216 accepted but nothing will happen, i.e. renegotiation will not be prevented.
4220 * Removed the OS390-Unix config target. It relied on a script that doesn't
4225 * rsaz_1024_mul_avx2 overflow bug on x86_64
4227 There is an overflow bug in the AVX2 Montgomery multiplication procedure
4228 used in exponentiation with 1024-bit moduli. No EC algorithms are affected.
4229 Analysis suggests that attacks against RSA and DSA as a result of this
4230 defect would be very difficult to perform and are not believed likely.
4231 Attacks against DH1024 are considered just feasible, because most of the
4232 work necessary to deduce information about a private key may be performed
4233 offline. The amount of resources required for such an attack would be
4234 significant. However, for an attack on TLS to be meaningful, the server
4235 would have to share the DH1024 private key among multiple clients, which is
4236 no longer an option since CVE-2016-0701.
4238 This only affects processors that support the AVX2 but not ADX extensions
4239 like Intel Haswell (4th generation).
4241 This issue was reported to OpenSSL by David Benjamin (Google). The issue
4242 was originally found via the OSS-Fuzz project.
4247 ### Changes between 1.1.0f and 1.1.0g [2 Nov 2017]
4249 * bn_sqrx8x_internal carry bug on x86_64
4251 There is a carry propagating bug in the x86_64 Montgomery squaring
4252 procedure. No EC algorithms are affected. Analysis suggests that attacks
4253 against RSA and DSA as a result of this defect would be very difficult to
4254 perform and are not believed likely. Attacks against DH are considered just
4255 feasible (although very difficult) because most of the work necessary to
4256 deduce information about a private key may be performed offline. The amount
4257 of resources required for such an attack would be very significant and
4258 likely only accessible to a limited number of attackers. An attacker would
4259 additionally need online access to an unpatched system using the target
4260 private key in a scenario with persistent DH parameters and a private
4261 key that is shared between multiple clients.
4263 This only affects processors that support the BMI1, BMI2 and ADX extensions
4264 like Intel Broadwell (5th generation) and later or AMD Ryzen.
4266 This issue was reported to OpenSSL by the OSS-Fuzz project.
4271 * Malformed X.509 IPAddressFamily could cause OOB read
4273 If an X.509 certificate has a malformed IPAddressFamily extension,
4274 OpenSSL could do a one-byte buffer overread. The most likely result
4275 would be an erroneous display of the certificate in text format.
4277 This issue was reported to OpenSSL by the OSS-Fuzz project.
4282 ### Changes between 1.1.0e and 1.1.0f [25 May 2017]
4284 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
4285 platform rather than 'mingw'.
4289 * Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
4290 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
4291 which is the minimum version we support.
4295 ### Changes between 1.1.0d and 1.1.0e [16 Feb 2017]
4297 * Encrypt-Then-Mac renegotiation crash
4299 During a renegotiation handshake if the Encrypt-Then-Mac extension is
4300 negotiated where it was not in the original handshake (or vice-versa) then
4301 this can cause OpenSSL to crash (dependent on ciphersuite). Both clients
4302 and servers are affected.
4304 This issue was reported to OpenSSL by Joe Orton (Red Hat).
4309 ### Changes between 1.1.0c and 1.1.0d [26 Jan 2017]
4311 * Truncated packet could crash via OOB read
4313 If one side of an SSL/TLS path is running on a 32-bit host and a specific
4314 cipher is being used, then a truncated packet can cause that host to
4315 perform an out-of-bounds read, usually resulting in a crash.
4317 This issue was reported to OpenSSL by Robert Święcki of Google.
4322 * Bad (EC)DHE parameters cause a client crash
4324 If a malicious server supplies bad parameters for a DHE or ECDHE key
4325 exchange then this can result in the client attempting to dereference a
4326 NULL pointer leading to a client crash. This could be exploited in a Denial
4329 This issue was reported to OpenSSL by Guido Vranken.
4334 * BN_mod_exp may produce incorrect results on x86_64
4336 There is a carry propagating bug in the x86_64 Montgomery squaring
4337 procedure. No EC algorithms are affected. Analysis suggests that attacks
4338 against RSA and DSA as a result of this defect would be very difficult to
4339 perform and are not believed likely. Attacks against DH are considered just
4340 feasible (although very difficult) because most of the work necessary to
4341 deduce information about a private key may be performed offline. The amount
4342 of resources required for such an attack would be very significant and
4343 likely only accessible to a limited number of attackers. An attacker would
4344 additionally need online access to an unpatched system using the target
4345 private key in a scenario with persistent DH parameters and a private
4346 key that is shared between multiple clients. For example this can occur by
4347 default in OpenSSL DHE based SSL/TLS ciphersuites. Note: This issue is very
4348 similar to CVE-2015-3193 but must be treated as a separate problem.
4350 This issue was reported to OpenSSL by the OSS-Fuzz project.
4355 ### Changes between 1.1.0b and 1.1.0c [10 Nov 2016]
4357 * ChaCha20/Poly1305 heap-buffer-overflow
4359 TLS connections using `*-CHACHA20-POLY1305` ciphersuites are susceptible to
4360 a DoS attack by corrupting larger payloads. This can result in an OpenSSL
4361 crash. This issue is not considered to be exploitable beyond a DoS.
4363 This issue was reported to OpenSSL by Robert Święcki (Google Security Team)
4368 * CMS Null dereference
4370 Applications parsing invalid CMS structures can crash with a NULL pointer
4371 dereference. This is caused by a bug in the handling of the ASN.1 CHOICE
4372 type in OpenSSL 1.1.0 which can result in a NULL value being passed to the
4373 structure callback if an attempt is made to free certain invalid encodings.
4374 Only CHOICE structures using a callback which do not handle NULL value are
4377 This issue was reported to OpenSSL by Tyler Nighswander of ForAllSecure.
4382 * Montgomery multiplication may produce incorrect results
4384 There is a carry propagating bug in the Broadwell-specific Montgomery
4385 multiplication procedure that handles input lengths divisible by, but
4386 longer than 256 bits. Analysis suggests that attacks against RSA, DSA
4387 and DH private keys are impossible. This is because the subroutine in
4388 question is not used in operations with the private key itself and an input
4389 of the attacker's direct choice. Otherwise the bug can manifest itself as
4390 transient authentication and key negotiation failures or reproducible
4391 erroneous outcome of public-key operations with specially crafted input.
4392 Among EC algorithms only Brainpool P-512 curves are affected and one
4393 presumably can attack ECDH key negotiation. Impact was not analyzed in
4394 detail, because pre-requisites for attack are considered unlikely. Namely
4395 multiple clients have to choose the curve in question and the server has to
4396 share the private key among them, neither of which is default behaviour.
4397 Even then only clients that chose the curve will be affected.
4399 This issue was publicly reported as transient failures and was not
4400 initially recognized as a security issue. Thanks to Richard Morgan for
4401 providing reproducible case.
4406 * Removed automatic addition of RPATH in shared libraries and executables,
4407 as this was a remainder from OpenSSL 1.0.x and isn't needed any more.
4411 ### Changes between 1.1.0a and 1.1.0b [26 Sep 2016]
4413 * Fix Use After Free for large message sizes
4415 The patch applied to address CVE-2016-6307 resulted in an issue where if a
4416 message larger than approx 16k is received then the underlying buffer to
4417 store the incoming message is reallocated and moved. Unfortunately a
4418 dangling pointer to the old location is left which results in an attempt to
4419 write to the previously freed location. This is likely to result in a
4420 crash, however it could potentially lead to execution of arbitrary code.
4422 This issue only affects OpenSSL 1.1.0a.
4424 This issue was reported to OpenSSL by Robert Święcki.
4429 ### Changes between 1.1.0 and 1.1.0a [22 Sep 2016]
4431 * OCSP Status Request extension unbounded memory growth
4433 A malicious client can send an excessively large OCSP Status Request
4434 extension. If that client continually requests renegotiation, sending a
4435 large OCSP Status Request extension each time, then there will be unbounded
4436 memory growth on the server. This will eventually lead to a Denial Of
4437 Service attack through memory exhaustion. Servers with a default
4438 configuration are vulnerable even if they do not support OCSP. Builds using
4439 the "no-ocsp" build time option are not affected.
4441 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4446 * SSL_peek() hang on empty record
4448 OpenSSL 1.1.0 SSL/TLS will hang during a call to SSL_peek() if the peer
4449 sends an empty record. This could be exploited by a malicious peer in a
4450 Denial Of Service attack.
4452 This issue was reported to OpenSSL by Alex Gaynor.
4457 * Excessive allocation of memory in tls_get_message_header() and
4458 dtls1_preprocess_fragment()
4460 A (D)TLS message includes 3 bytes for its length in the header for the
4461 message. This would allow for messages up to 16Mb in length. Messages of
4462 this length are excessive and OpenSSL includes a check to ensure that a
4463 peer is sending reasonably sized messages in order to avoid too much memory
4464 being consumed to service a connection. A flaw in the logic of version
4465 1.1.0 means that memory for the message is allocated too early, prior to
4466 the excessive message length check. Due to way memory is allocated in
4467 OpenSSL this could mean an attacker could force up to 21Mb to be allocated
4468 to service a connection. This could lead to a Denial of Service through
4469 memory exhaustion. However, the excessive message length check still takes
4470 place, and this would cause the connection to immediately fail. Assuming
4471 that the application calls SSL_free() on the failed connection in a timely
4472 manner then the 21Mb of allocated memory will then be immediately freed
4473 again. Therefore, the excessive memory allocation will be transitory in
4474 nature. This then means that there is only a security impact if:
4476 1) The application does not call SSL_free() in a timely manner in the event
4477 that the connection fails
4479 2) The application is working in a constrained environment where there is
4480 very little free memory
4482 3) The attacker initiates multiple connection attempts such that there are
4483 multiple connections in a state where memory has been allocated for the
4484 connection; SSL_free() has not yet been called; and there is insufficient
4485 memory to service the multiple requests.
4487 Except in the instance of (1) above any Denial Of Service is likely to be
4488 transitory because as soon as the connection fails the memory is
4489 subsequently freed again in the SSL_free() call. However there is an
4490 increased risk during this period of application crashes due to the lack of
4491 memory - which would then mean a more serious Denial of Service.
4493 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4494 (CVE-2016-6307 and CVE-2016-6308)
4498 * solaris-x86-cc, i.e. 32-bit configuration with vendor compiler,
4499 had to be removed. Primary reason is that vendor assembler can't
4500 assemble our modules with -KPIC flag. As result it, assembly
4501 support, was not even available as option. But its lack means
4502 lack of side-channel resistant code, which is incompatible with
4503 security by todays standards. Fortunately gcc is readily available
4504 prepackaged option, which we firmly point at...
4508 ### Changes between 1.0.2h and 1.1.0 [25 Aug 2016]
4510 * Windows command-line tool supports UTF-8 opt-in option for arguments
4511 and console input. Setting OPENSSL_WIN32_UTF8 environment variable
4512 (to any value) allows Windows user to access PKCS#12 file generated
4513 with Windows CryptoAPI and protected with non-ASCII password, as well
4514 as files generated under UTF-8 locale on Linux also protected with
4519 * To mitigate the SWEET32 attack ([CVE-2016-2183]), 3DES cipher suites
4520 have been disabled by default and removed from DEFAULT, just like RC4.
4521 See the RC4 item below to re-enable both.
4525 * The method for finding the storage location for the Windows RAND seed file
4526 has changed. First we check %RANDFILE%. If that is not set then we check
4527 the directories %HOME%, %USERPROFILE% and %SYSTEMROOT% in that order. If
4528 all else fails we fall back to C:\.
4532 * The EVP_EncryptUpdate() function has had its return type changed from void
4533 to int. A return of 0 indicates and error while a return of 1 indicates
4538 * The flags RSA_FLAG_NO_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME and
4539 DH_FLAG_NO_EXP_CONSTTIME which previously provided the ability to switch
4540 off the constant time implementation for RSA, DSA and DH have been made
4541 no-ops and deprecated.
4545 * Windows RAND implementation was simplified to only get entropy by
4546 calling CryptGenRandom(). Various other RAND-related tickets
4549 *Joseph Wylie Yandle, Rich Salz*
4551 * The stack and lhash API's were renamed to start with `OPENSSL_SK_`
4552 and `OPENSSL_LH_`, respectively. The old names are available
4553 with API compatibility. They new names are now completely documented.
4557 * Unify TYPE_up_ref(obj) methods signature.
4558 SSL_CTX_up_ref(), SSL_up_ref(), X509_up_ref(), EVP_PKEY_up_ref(),
4559 X509_CRL_up_ref(), X509_OBJECT_up_ref_count() methods are now returning an
4560 int (instead of void) like all others TYPE_up_ref() methods.
4561 So now these methods also check the return value of CRYPTO_atomic_add(),
4562 and the validity of object reference counter.
4564 *fdasilvayy@gmail.com*
4566 * With Windows Visual Studio builds, the .pdb files are installed
4567 alongside the installed libraries and executables. For a static
4568 library installation, ossl_static.pdb is the associate compiler
4569 generated .pdb file to be used when linking programs.
4573 * Remove openssl.spec. Packaging files belong with the packagers.
4577 * Automatic Darwin/OSX configuration has had a refresh, it will now
4578 recognise x86_64 architectures automatically. You can still decide
4579 to build for a different bitness with the environment variable
4580 KERNEL_BITS (can be 32 or 64), for example:
4582 KERNEL_BITS=32 ./config
4586 * Change default algorithms in pkcs8 utility to use PKCS#5 v2.0,
4587 256 bit AES and HMAC with SHA256.
4591 * Remove support for MIPS o32 ABI on IRIX (and IRIX only).
4595 * Triple-DES ciphers have been moved from HIGH to MEDIUM.
4599 * To enable users to have their own config files and build file templates,
4600 Configure looks in the directory indicated by the environment variable
4601 OPENSSL_LOCAL_CONFIG_DIR as well as the in-source Configurations/
4602 directory. On VMS, OPENSSL_LOCAL_CONFIG_DIR is expected to be a logical
4603 name and is used as is.
4607 * The following datatypes were made opaque: X509_OBJECT, X509_STORE_CTX,
4608 X509_STORE, X509_LOOKUP, and X509_LOOKUP_METHOD. The unused type
4609 X509_CERT_FILE_CTX was removed.
4613 * "shared" builds are now the default. To create only static libraries use
4614 the "no-shared" Configure option.
4618 * Remove the no-aes, no-hmac, no-rsa, no-sha and no-md5 Configure options.
4619 All of these option have not worked for some while and are fundamental
4624 * Make various cleanup routines no-ops and mark them as deprecated. Most
4625 global cleanup functions are no longer required because they are handled
4626 via auto-deinit (see OPENSSL_init_crypto and OPENSSL_init_ssl man pages).
4627 Explicitly de-initing can cause problems (e.g. where a library that uses
4628 OpenSSL de-inits, but an application is still using it). The affected
4629 functions are CONF_modules_free(), ENGINE_cleanup(), OBJ_cleanup(),
4630 EVP_cleanup(), BIO_sock_cleanup(), CRYPTO_cleanup_all_ex_data(),
4631 RAND_cleanup(), SSL_COMP_free_compression_methods(), ERR_free_strings() and
4632 COMP_zlib_cleanup().
4636 * --strict-warnings no longer enables runtime debugging options
4637 such as REF_DEBUG. Instead, debug options are automatically
4638 enabled with '--debug' builds.
4640 *Andy Polyakov, Emilia Käsper*
4642 * Made DH and DH_METHOD opaque. The structures for managing DH objects
4643 have been moved out of the public header files. New functions for managing
4644 these have been added.
4648 * Made RSA and RSA_METHOD opaque. The structures for managing RSA
4649 objects have been moved out of the public header files. New
4650 functions for managing these have been added.
4654 * Made DSA and DSA_METHOD opaque. The structures for managing DSA objects
4655 have been moved out of the public header files. New functions for managing
4656 these have been added.
4660 * Made BIO and BIO_METHOD opaque. The structures for managing BIOs have been
4661 moved out of the public header files. New functions for managing these
4666 * Removed no-rijndael as a config option. Rijndael is an old name for AES.
4670 * Removed the mk1mf build scripts.
4674 * Headers are now wrapped, if necessary, with OPENSSL_NO_xxx, so
4675 it is always safe to #include a header now.
4679 * Removed the aged BC-32 config and all its supporting scripts
4683 * Removed support for Ultrix, Netware, and OS/2.
4687 * Add support for HKDF.
4689 *Alessandro Ghedini*
4691 * Add support for blake2b and blake2s
4695 * Added support for "pipelining". Ciphers that have the
4696 EVP_CIPH_FLAG_PIPELINE flag set have a capability to process multiple
4697 encryptions/decryptions simultaneously. There are currently no built-in
4698 ciphers with this property but the expectation is that engines will be able
4699 to offer it to significantly improve throughput. Support has been extended
4700 into libssl so that multiple records for a single connection can be
4701 processed in one go (for >=TLS 1.1).
4705 * Added the AFALG engine. This is an async capable engine which is able to
4706 offload work to the Linux kernel. In this initial version it only supports
4707 AES128-CBC. The kernel must be version 4.1.0 or greater.
4711 * OpenSSL now uses a new threading API. It is no longer necessary to
4712 set locking callbacks to use OpenSSL in a multi-threaded environment. There
4713 are two supported threading models: pthreads and windows threads. It is
4714 also possible to configure OpenSSL at compile time for "no-threads". The
4715 old threading API should no longer be used. The functions have been
4716 replaced with "no-op" compatibility macros.
4718 *Alessandro Ghedini, Matt Caswell*
4720 * Modify behavior of ALPN to invoke callback after SNI/servername
4721 callback, such that updates to the SSL_CTX affect ALPN.
4725 * Add SSL_CIPHER queries for authentication and key-exchange.
4729 * Changes to the DEFAULT cipherlist:
4730 - Prefer (EC)DHE handshakes over plain RSA.
4731 - Prefer AEAD ciphers over legacy ciphers.
4732 - Prefer ECDSA over RSA when both certificates are available.
4733 - Prefer TLSv1.2 ciphers/PRF.
4734 - Remove DSS, SEED, IDEA, CAMELLIA, and AES-CCM from the
4739 * Change the ECC default curve list to be this, in order: x25519,
4740 secp256r1, secp521r1, secp384r1.
4744 * RC4 based libssl ciphersuites are now classed as "weak" ciphers and are
4745 disabled by default. They can be re-enabled using the
4746 enable-weak-ssl-ciphers option to Configure.
4750 * If the server has ALPN configured, but supports no protocols that the
4751 client advertises, send a fatal "no_application_protocol" alert.
4752 This behaviour is SHALL in RFC 7301, though it isn't universally
4753 implemented by other servers.
4757 * Add X25519 support.
4758 Add ASN.1 and EVP_PKEY methods for X25519. This includes support
4759 for public and private key encoding using the format documented in
4760 draft-ietf-curdle-pkix-02. The corresponding EVP_PKEY method supports
4761 key generation and key derivation.
4763 TLS support complies with draft-ietf-tls-rfc4492bis-08 and uses
4768 * Deprecate SRP_VBASE_get_by_user.
4769 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
4770 In order to fix an unavoidable memory leak ([CVE-2016-0798]),
4771 SRP_VBASE_get_by_user was changed to ignore the "fake user" SRP
4772 seed, even if the seed is configured.
4774 Users should use SRP_VBASE_get1_by_user instead. Note that in
4775 SRP_VBASE_get1_by_user, caller must free the returned value. Note
4776 also that even though configuring the SRP seed attempts to hide
4777 invalid usernames by continuing the handshake with fake
4778 credentials, this behaviour is not constant time and no strong
4779 guarantees are made that the handshake is indistinguishable from
4780 that of a valid user.
4784 * Configuration change; it's now possible to build dynamic engines
4785 without having to build shared libraries and vice versa. This
4786 only applies to the engines in `engines/`, those in `crypto/engine/`
4787 will always be built into libcrypto (i.e. "static").
4789 Building dynamic engines is enabled by default; to disable, use
4790 the configuration option "disable-dynamic-engine".
4792 The only requirements for building dynamic engines are the
4793 presence of the DSO module and building with position independent
4794 code, so they will also automatically be disabled if configuring
4795 with "disable-dso" or "disable-pic".
4797 The macros OPENSSL_NO_STATIC_ENGINE and OPENSSL_NO_DYNAMIC_ENGINE
4798 are also taken away from openssl/opensslconf.h, as they are
4803 * Configuration change; if there is a known flag to compile
4804 position independent code, it will always be applied on the
4805 libcrypto and libssl object files, and never on the application
4806 object files. This means other libraries that use routines from
4807 libcrypto / libssl can be made into shared libraries regardless
4808 of how OpenSSL was configured.
4810 If this isn't desirable, the configuration options "disable-pic"
4811 or "no-pic" can be used to disable the use of PIC. This will
4812 also disable building shared libraries and dynamic engines.
4816 * Removed JPAKE code. It was experimental and has no wide use.
4820 * The INSTALL_PREFIX Makefile variable has been renamed to
4821 DESTDIR. That makes for less confusion on what this variable
4822 is for. Also, the configuration option --install_prefix is
4827 * Heartbeat for TLS has been removed and is disabled by default
4828 for DTLS; configure with enable-heartbeats. Code that uses the
4829 old #define's might need to be updated.
4831 *Emilia Käsper, Rich Salz*
4833 * Rename REF_CHECK to REF_DEBUG.
4837 * New "unified" build system
4839 The "unified" build system is aimed to be a common system for all
4840 platforms we support. With it comes new support for VMS.
4842 This system builds supports building in a different directory tree
4843 than the source tree. It produces one Makefile (for unix family
4844 or lookalikes), or one descrip.mms (for VMS).
4846 The source of information to make the Makefile / descrip.mms is
4847 small files called 'build.info', holding the necessary
4848 information for each directory with source to compile, and a
4849 template in Configurations, like unix-Makefile.tmpl or
4852 With this change, the library names were also renamed on Windows
4853 and on VMS. They now have names that are closer to the standard
4854 on Unix, and include the major version number, and in certain
4855 cases, the architecture they are built for. See "Notes on shared
4856 libraries" in INSTALL.
4858 We rely heavily on the perl module Text::Template.
4862 * Added support for auto-initialisation and de-initialisation of the library.
4863 OpenSSL no longer requires explicit init or deinit routines to be called,
4864 except in certain circumstances. See the OPENSSL_init_crypto() and
4865 OPENSSL_init_ssl() man pages for further information.
4869 * The arguments to the DTLSv1_listen function have changed. Specifically the
4870 "peer" argument is now expected to be a BIO_ADDR object.
4872 * Rewrite of BIO networking library. The BIO library lacked consistent
4873 support of IPv6, and adding it required some more extensive
4874 modifications. This introduces the BIO_ADDR and BIO_ADDRINFO types,
4875 which hold all types of addresses and chains of address information.
4876 It also introduces a new API, with functions like BIO_socket,
4877 BIO_connect, BIO_listen, BIO_lookup and a rewrite of BIO_accept.
4878 The source/sink BIOs BIO_s_connect, BIO_s_accept and BIO_s_datagram
4879 have been adapted accordingly.
4883 * RSA_padding_check_PKCS1_type_1 now accepts inputs with and without
4888 * CRIME protection: disable compression by default, even if OpenSSL is
4889 compiled with zlib enabled. Applications can still enable compression
4890 by calling SSL_CTX_clear_options(ctx, SSL_OP_NO_COMPRESSION), or by
4891 using the SSL_CONF library to configure compression.
4895 * The signature of the session callback configured with
4896 SSL_CTX_sess_set_get_cb was changed. The read-only input buffer
4897 was explicitly marked as `const unsigned char*` instead of
4902 * Always DPURIFY. Remove the use of uninitialized memory in the
4903 RNG, and other conditional uses of DPURIFY. This makes -DPURIFY a no-op.
4907 * Removed many obsolete configuration items, including
4908 DES_PTR, DES_RISC1, DES_RISC2, DES_INT
4909 MD2_CHAR, MD2_INT, MD2_LONG
4911 IDEA_SHORT, IDEA_LONG
4912 RC2_SHORT, RC2_LONG, RC4_LONG, RC4_CHUNK, RC4_INDEX
4914 *Rich Salz, with advice from Andy Polyakov*
4916 * Many BN internals have been moved to an internal header file.
4918 *Rich Salz with help from Andy Polyakov*
4920 * Configuration and writing out the results from it has changed.
4921 Files such as Makefile include/openssl/opensslconf.h and are now
4922 produced through general templates, such as Makefile.in and
4923 crypto/opensslconf.h.in and some help from the perl module
4926 Also, the center of configuration information is no longer
4927 Makefile. Instead, Configure produces a perl module in
4928 configdata.pm which holds most of the config data (in the hash
4929 table %config), the target data that comes from the target
4930 configuration in one of the `Configurations/*.conf` files (in
4935 * To clarify their intended purposes, the Configure options
4936 --prefix and --openssldir change their semantics, and become more
4937 straightforward and less interdependent.
4939 --prefix shall be used exclusively to give the location INSTALLTOP
4940 where programs, scripts, libraries, include files and manuals are
4941 going to be installed. The default is now /usr/local.
4943 --openssldir shall be used exclusively to give the default
4944 location OPENSSLDIR where certificates, private keys, CRLs are
4945 managed. This is also where the default openssl.cnf gets
4947 If the directory given with this option is a relative path, the
4948 values of both the --prefix value and the --openssldir value will
4949 be combined to become OPENSSLDIR.
4950 The default for --openssldir is INSTALLTOP/ssl.
4952 Anyone who uses --openssldir to specify where OpenSSL is to be
4953 installed MUST change to use --prefix instead.
4957 * The GOST engine was out of date and therefore it has been removed. An up
4958 to date GOST engine is now being maintained in an external repository.
4959 See: <https://wiki.openssl.org/index.php/Binaries>. Libssl still retains
4960 support for GOST ciphersuites (these are only activated if a GOST engine
4965 * EGD is no longer supported by default; use enable-egd when
4968 *Ben Kaduk and Rich Salz*
4970 * The distribution now has Makefile.in files, which are used to
4971 create Makefile's when Configure is run. *Configure must be run
4972 before trying to build now.*
4976 * The return value for SSL_CIPHER_description() for error conditions
4981 * Support for RFC6698/RFC7671 DANE TLSA peer authentication.
4983 Obtaining and performing DNSSEC validation of TLSA records is
4984 the application's responsibility. The application provides
4985 the TLSA records of its choice to OpenSSL, and these are then
4986 used to authenticate the peer.
4988 The TLSA records need not even come from DNS. They can, for
4989 example, be used to implement local end-entity certificate or
4990 trust-anchor "pinning", where the "pin" data takes the form
4991 of TLSA records, which can augment or replace verification
4992 based on the usual WebPKI public certification authorities.
4996 * Revert default OPENSSL_NO_DEPRECATED setting. Instead OpenSSL
4997 continues to support deprecated interfaces in default builds.
4998 However, applications are strongly advised to compile their
4999 source files with -DOPENSSL_API_COMPAT=0x10100000L, which hides
5000 the declarations of all interfaces deprecated in 0.9.8, 1.0.0
5001 or the 1.1.0 releases.
5003 In environments in which all applications have been ported to
5004 not use any deprecated interfaces OpenSSL's Configure script
5005 should be used with the --api=1.1.0 option to entirely remove
5006 support for the deprecated features from the library and
5007 unconditionally disable them in the installed headers.
5008 Essentially the same effect can be achieved with the "no-deprecated"
5009 argument to Configure, except that this will always restrict
5010 the build to just the latest API, rather than a fixed API
5013 As applications are ported to future revisions of the API,
5014 they should update their compile-time OPENSSL_API_COMPAT define
5015 accordingly, but in most cases should be able to continue to
5016 compile with later releases.
5018 The OPENSSL_API_COMPAT versions for 1.0.0, and 0.9.8 are
5019 0x10000000L and 0x00908000L, respectively. However those
5020 versions did not support the OPENSSL_API_COMPAT feature, and
5021 so applications are not typically tested for explicit support
5022 of just the undeprecated features of either release.
5026 * Add support for setting the minimum and maximum supported protocol.
5027 It can bet set via the SSL_set_min_proto_version() and
5028 SSL_set_max_proto_version(), or via the SSL_CONF's MinProtocol and
5029 MaxProtocol. It's recommended to use the new APIs to disable
5030 protocols instead of disabling individual protocols using
5031 SSL_set_options() or SSL_CONF's Protocol. This change also
5032 removes support for disabling TLS 1.2 in the OpenSSL TLS
5033 client at compile time by defining OPENSSL_NO_TLS1_2_CLIENT.
5037 * Support for ChaCha20 and Poly1305 added to libcrypto and libssl.
5041 * New EC_KEY_METHOD, this replaces the older ECDSA_METHOD and ECDH_METHOD
5042 and integrates ECDSA and ECDH functionality into EC. Implementations can
5043 now redirect key generation and no longer need to convert to or from
5046 Note: the ecdsa.h and ecdh.h headers are now no longer needed and just
5047 include the ec.h header file instead.
5051 * Remove support for all 40 and 56 bit ciphers. This includes all the export
5052 ciphers who are no longer supported and drops support the ephemeral RSA key
5053 exchange. The LOW ciphers currently doesn't have any ciphers in it.
5057 * Made EVP_MD_CTX, EVP_MD, EVP_CIPHER_CTX, EVP_CIPHER and HMAC_CTX
5058 opaque. For HMAC_CTX, the following constructors and destructors
5061 HMAC_CTX *HMAC_CTX_new(void);
5062 void HMAC_CTX_free(HMAC_CTX *ctx);
5064 For EVP_MD and EVP_CIPHER, complete APIs to create, fill and
5065 destroy such methods has been added. See EVP_MD_meth_new(3) and
5066 EVP_CIPHER_meth_new(3) for documentation.
5069 1) `EVP_MD_CTX_cleanup()`, `EVP_CIPHER_CTX_cleanup()` and
5070 `HMAC_CTX_cleanup()` were removed. `HMAC_CTX_reset()` and
5071 `EVP_MD_CTX_reset()` should be called instead to reinitialise
5072 an already created structure.
5073 2) For consistency with the majority of our object creators and
5074 destructors, `EVP_MD_CTX_(create|destroy)` were renamed to
5075 `EVP_MD_CTX_(new|free)`. The old names are retained as macros
5076 for deprecated builds.
5080 * Added ASYNC support. Libcrypto now includes the async sub-library to enable
5081 cryptographic operations to be performed asynchronously as long as an
5082 asynchronous capable engine is used. See the ASYNC_start_job() man page for
5083 further details. Libssl has also had this capability integrated with the
5084 introduction of the new mode SSL_MODE_ASYNC and associated error
5085 SSL_ERROR_WANT_ASYNC. See the SSL_CTX_set_mode() and SSL_get_error() man
5086 pages. This work was developed in partnership with Intel Corp.
5090 * SSL_{CTX_}set_ecdh_auto() has been removed and ECDH is support is
5091 always enabled now. If you want to disable the support you should
5092 exclude it using the list of supported ciphers. This also means that the
5093 "-no_ecdhe" option has been removed from s_server.
5097 * SSL_{CTX}_set_tmp_ecdh() which can set 1 EC curve now internally calls
5098 SSL_{CTX_}set1_curves() which can set a list.
5102 * Remove support for SSL_{CTX_}set_tmp_ecdh_callback(). You should set the
5103 curve you want to support using SSL_{CTX_}set1_curves().
5107 * State machine rewrite. The state machine code has been significantly
5108 refactored in order to remove much duplication of code and solve issues
5109 with the old code (see [ssl/statem/README.md](ssl/statem/README.md) for
5110 further details). This change does have some associated API changes.
5111 Notably the SSL_state() function has been removed and replaced by
5112 SSL_get_state which now returns an "OSSL_HANDSHAKE_STATE" instead of an int.
5113 SSL_set_state() has been removed altogether. The previous handshake states
5114 defined in ssl.h and ssl3.h have also been removed.
5118 * All instances of the string "ssleay" in the public API were replaced
5119 with OpenSSL (case-matching; e.g., OPENSSL_VERSION for #define's)
5120 Some error codes related to internal RSA_eay API's were renamed.
5124 * The demo files in crypto/threads were moved to demo/threads.
5128 * Removed obsolete engines: 4758cca, aep, atalla, cswift, nuron, gmp,
5131 *Matt Caswell, Rich Salz*
5133 * New ASN.1 embed macro.
5135 New ASN.1 macro ASN1_EMBED. This is the same as ASN1_SIMPLE except the
5136 structure is not allocated: it is part of the parent. That is instead of
5144 This reduces memory fragmentation and make it impossible to accidentally
5145 set a mandatory field to NULL.
5147 This currently only works for some fields specifically a SEQUENCE, CHOICE,
5148 or ASN1_STRING type which is part of a parent SEQUENCE. Since it is
5149 equivalent to ASN1_SIMPLE it cannot be tagged, OPTIONAL, SET OF or
5154 * Remove EVP_CHECK_DES_KEY, a compile-time option that never compiled.
5158 * Removed DES and RC4 ciphersuites from DEFAULT. Also removed RC2 although
5159 in 1.0.2 EXPORT was already removed and the only RC2 ciphersuite is also
5160 an EXPORT one. COMPLEMENTOFDEFAULT has been updated accordingly to add
5161 DES and RC4 ciphersuites.
5165 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
5166 This changes the decoding behaviour for some invalid messages,
5167 though the change is mostly in the more lenient direction, and
5168 legacy behaviour is preserved as much as possible.
5172 * Fix no-stdio build.
5173 *David Woodhouse <David.Woodhouse@intel.com> and also*
5174 *Ivan Nestlerode <ivan.nestlerode@sonos.com>*
5176 * New testing framework
5177 The testing framework has been largely rewritten and is now using
5178 perl and the perl modules Test::Harness and an extended variant of
5179 Test::More called OpenSSL::Test to do its work. All test scripts in
5180 test/ have been rewritten into test recipes, and all direct calls to
5181 executables in test/Makefile have become individual recipes using the
5182 simplified testing OpenSSL::Test::Simple.
5184 For documentation on our testing modules, do:
5186 perldoc test/testlib/OpenSSL/Test/Simple.pm
5187 perldoc test/testlib/OpenSSL/Test.pm
5191 * Revamped memory debug; only -DCRYPTO_MDEBUG and -DCRYPTO_MDEBUG_ABORT
5192 are used; the latter aborts on memory leaks (usually checked on exit).
5193 Some undocumented "set malloc, etc., hooks" functions were removed
5194 and others were changed. All are now documented.
5198 * In DSA_generate_parameters_ex, if the provided seed is too short,
5201 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
5203 * Rewrite PSK to support ECDHE_PSK, DHE_PSK and RSA_PSK. Add ciphersuites
5204 from RFC4279, RFC4785, RFC5487, RFC5489.
5206 Thanks to Christian J. Dietrich and Giuseppe D'Angelo for the
5207 original RSA_PSK patch.
5211 * Dropped support for the SSL3_FLAGS_DELAY_CLIENT_FINISHED flag. This SSLeay
5212 era flag was never set throughout the codebase (only read). Also removed
5213 SSL3_FLAGS_POP_BUFFER which was only used if
5214 SSL3_FLAGS_DELAY_CLIENT_FINISHED was also set.
5218 * Changed the default name options in the "ca", "crl", "req" and "x509"
5219 to be "oneline" instead of "compat".
5223 * Remove SSL_OP_TLS_BLOCK_PADDING_BUG. This is SSLeay legacy, we're
5224 not aware of clients that still exhibit this bug, and the workaround
5225 hasn't been working properly for a while.
5229 * The return type of BIO_number_read() and BIO_number_written() as well as
5230 the corresponding num_read and num_write members in the BIO structure has
5231 changed from unsigned long to uint64_t. On platforms where an unsigned
5232 long is 32 bits (e.g. Windows) these counters could overflow if >4Gb is
5237 * Given the pervasive nature of TLS extensions it is inadvisable to run
5238 OpenSSL without support for them. It also means that maintaining
5239 the OPENSSL_NO_TLSEXT option within the code is very invasive (and probably
5240 not well tested). Therefore, the OPENSSL_NO_TLSEXT option has been removed.
5244 * Removed support for the two export grade static DH ciphersuites
5245 EXP-DH-RSA-DES-CBC-SHA and EXP-DH-DSS-DES-CBC-SHA. These two ciphersuites
5246 were newly added (along with a number of other static DH ciphersuites) to
5247 1.0.2. However the two export ones have *never* worked since they were
5248 introduced. It seems strange in any case to be adding new export
5249 ciphersuites, and given "logjam" it also does not seem correct to fix them.
5253 * Version negotiation has been rewritten. In particular SSLv23_method(),
5254 SSLv23_client_method() and SSLv23_server_method() have been deprecated,
5255 and turned into macros which simply call the new preferred function names
5256 TLS_method(), TLS_client_method() and TLS_server_method(). All new code
5257 should use the new names instead. Also as part of this change the ssl23.h
5258 header file has been removed.
5262 * Support for Kerberos ciphersuites in TLS (RFC2712) has been removed. This
5263 code and the associated standard is no longer considered fit-for-purpose.
5267 * RT2547 was closed. When generating a private key, try to make the
5268 output file readable only by the owner. This behavior change might
5269 be noticeable when interacting with other software.
5271 * Documented all exdata functions. Added CRYPTO_free_ex_index.
5276 * Added HTTP GET support to the ocsp command.
5280 * Changed default digest for the dgst and enc commands from MD5 to
5285 * RAND_pseudo_bytes has been deprecated. Users should use RAND_bytes instead.
5289 * Added support for TLS extended master secret from
5290 draft-ietf-tls-session-hash-03.txt. Thanks for Alfredo Pironti for an
5291 initial patch which was a great help during development.
5295 * All libssl internal structures have been removed from the public header
5296 files, and the OPENSSL_NO_SSL_INTERN option has been removed (since it is
5297 now redundant). Users should not attempt to access internal structures
5298 directly. Instead they should use the provided API functions.
5302 * config has been changed so that by default OPENSSL_NO_DEPRECATED is used.
5303 Access to deprecated functions can be re-enabled by running config with
5304 "enable-deprecated". In addition applications wishing to use deprecated
5305 functions must define OPENSSL_USE_DEPRECATED. Note that this new behaviour
5306 will, by default, disable some transitive includes that previously existed
5307 in the header files (e.g. ec.h will no longer, by default, include bn.h)
5311 * Added support for OCB mode. OpenSSL has been granted a patent license
5312 compatible with the OpenSSL license for use of OCB. Details are available
5313 at <https://www.openssl.org/source/OCB-patent-grant-OpenSSL.pdf>. Support
5314 for OCB can be removed by calling config with no-ocb.
5318 * SSLv2 support has been removed. It still supports receiving an SSLv2
5319 compatible client hello.
5323 * Increased the minimal RSA keysize from 256 to 512 bits [Rich Salz],
5324 done while fixing the error code for the key-too-small case.
5326 *Annie Yousar <a.yousar@informatik.hu-berlin.de>*
5328 * CA.sh has been removed; use CA.pl instead.
5332 * Removed old DES API.
5336 * Remove various unsupported platforms:
5342 Sinix/ReliantUNIX RM400
5347 16-bit platforms such as WIN16
5351 * Clean up OPENSSL_NO_xxx #define's
5352 - Use setbuf() and remove OPENSSL_NO_SETVBUF_IONBF
5353 - Rename OPENSSL_SYSNAME_xxx to OPENSSL_SYS_xxx
5354 - OPENSSL_NO_EC{DH,DSA} merged into OPENSSL_NO_EC
5355 - OPENSSL_NO_RIPEMD160, OPENSSL_NO_RIPEMD merged into OPENSSL_NO_RMD160
5356 - OPENSSL_NO_FP_API merged into OPENSSL_NO_STDIO
5357 - Remove OPENSSL_NO_BIO OPENSSL_NO_BUFFER OPENSSL_NO_CHAIN_VERIFY
5358 OPENSSL_NO_EVP OPENSSL_NO_FIPS_ERR OPENSSL_NO_HASH_COMP
5359 OPENSSL_NO_LHASH OPENSSL_NO_OBJECT OPENSSL_NO_SPEED OPENSSL_NO_STACK
5360 OPENSSL_NO_X509 OPENSSL_NO_X509_VERIFY
5361 - Remove MS_STATIC; it's a relic from platforms <32 bits.
5365 * Cleaned up dead code
5366 Remove all but one '#ifdef undef' which is to be looked at.
5370 * Clean up calling of xxx_free routines.
5371 Just like free(), fix most of the xxx_free routines to accept
5372 NULL. Remove the non-null checks from callers. Save much code.
5376 * Add secure heap for storage of private keys (when possible).
5377 Add BIO_s_secmem(), CBIGNUM, etc.
5378 Contributed by Akamai Technologies under our Corporate CLA.
5382 * Experimental support for a new, fast, unbiased prime candidate generator,
5383 bn_probable_prime_dh_coprime(). Not currently used by any prime generator.
5385 *Felix Laurie von Massenbach <felix@erbridge.co.uk>*
5387 * New output format NSS in the sess_id command line tool. This allows
5388 exporting the session id and the master key in NSS keylog format.
5390 *Martin Kaiser <martin@kaiser.cx>*
5392 * Harmonize version and its documentation. -f flag is used to display
5395 *mancha <mancha1@zoho.com>*
5397 * Fix eckey_priv_encode so it immediately returns an error upon a failure
5398 in i2d_ECPrivateKey. Thanks to Ted Unangst for feedback on this issue.
5400 *mancha <mancha1@zoho.com>*
5402 * Fix some double frees. These are not thought to be exploitable.
5404 *mancha <mancha1@zoho.com>*
5406 * A missing bounds check in the handling of the TLS heartbeat extension
5407 can be used to reveal up to 64k of memory to a connected client or
5410 Thanks for Neel Mehta of Google Security for discovering this bug and to
5411 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
5412 preparing the fix ([CVE-2014-0160])
5414 *Adam Langley, Bodo Moeller*
5416 * Fix for the attack described in the paper "Recovering OpenSSL
5417 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
5418 by Yuval Yarom and Naomi Benger. Details can be obtained from:
5419 <http://eprint.iacr.org/2014/140>
5421 Thanks to Yuval Yarom and Naomi Benger for discovering this
5422 flaw and to Yuval Yarom for supplying a fix ([CVE-2014-0076])
5424 *Yuval Yarom and Naomi Benger*
5426 * Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
5427 this fixes a limitation in previous versions of OpenSSL.
5431 * Experimental encrypt-then-mac support.
5433 Experimental support for encrypt then mac from
5434 draft-gutmann-tls-encrypt-then-mac-02.txt
5436 To enable it set the appropriate extension number (0x42 for the test
5437 server) using e.g. -DTLSEXT_TYPE_encrypt_then_mac=0x42
5439 For non-compliant peers (i.e. just about everything) this should have no
5442 WARNING: EXPERIMENTAL, SUBJECT TO CHANGE.
5446 * Add EVP support for key wrapping algorithms, to avoid problems with
5447 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
5448 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
5449 algorithms and include tests cases.
5453 * Extend CMS code to support RSA-PSS signatures and RSA-OAEP for
5458 * Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
5459 MGF1 digest and OAEP label.
5463 * Make openssl verify return errors.
5465 *Chris Palmer <palmer@google.com> and Ben Laurie*
5467 * New function ASN1_TIME_diff to calculate the difference between two
5468 ASN1_TIME structures or one structure and the current time.
5472 * Update fips_test_suite to support multiple command line options. New
5473 test to induce all self test errors in sequence and check expected
5478 * Add FIPS_{rsa,dsa,ecdsa}_{sign,verify} functions which digest and
5479 sign or verify all in one operation.
5483 * Add fips_algvs: a multicall fips utility incorporating all the algorithm
5484 test programs and fips_test_suite. Includes functionality to parse
5485 the minimal script output of fipsalgest.pl directly.
5489 * Add authorisation parameter to FIPS_module_mode_set().
5493 * Add FIPS selftest for ECDH algorithm using P-224 and B-233 curves.
5497 * Use separate DRBG fields for internal and external flags. New function
5498 FIPS_drbg_health_check() to perform on demand health checking. Add
5499 generation tests to fips_test_suite with reduced health check interval to
5500 demonstrate periodic health checking. Add "nodh" option to
5501 fips_test_suite to skip very slow DH test.
5505 * New function FIPS_get_cipherbynid() to lookup FIPS supported ciphers
5510 * More extensive health check for DRBG checking many more failure modes.
5511 New function FIPS_selftest_drbg_all() to handle every possible DRBG
5512 combination: call this in fips_test_suite.
5516 * Add support for canonical generation of DSA parameter 'g'. See
5519 * Add support for HMAC DRBG from SP800-90. Update DRBG algorithm test and
5520 POST to handle HMAC cases.
5524 * Add functions FIPS_module_version() and FIPS_module_version_text()
5525 to return numerical and string versions of the FIPS module number.
5529 * Rename FIPS_mode_set and FIPS_mode to FIPS_module_mode_set and
5530 FIPS_module_mode. FIPS_mode and FIPS_mode_set will be implemented
5531 outside the validated module in the FIPS capable OpenSSL.
5535 * Minor change to DRBG entropy callback semantics. In some cases
5536 there is no multiple of the block length between min_len and
5537 max_len. Allow the callback to return more than max_len bytes
5538 of entropy but discard any extra: it is the callback's responsibility
5539 to ensure that the extra data discarded does not impact the
5540 requested amount of entropy.
5544 * Add PRNG security strength checks to RSA, DSA and ECDSA using
5545 information in FIPS186-3, SP800-57 and SP800-131A.
5549 * CCM support via EVP. Interface is very similar to GCM case except we
5550 must supply all data in one chunk (i.e. no update, final) and the
5551 message length must be supplied if AAD is used. Add algorithm test
5556 * Initial version of POST overhaul. Add POST callback to allow the status
5557 of POST to be monitored and/or failures induced. Modify fips_test_suite
5558 to use callback. Always run all selftests even if one fails.
5562 * XTS support including algorithm test driver in the fips_gcmtest program.
5563 Note: this does increase the maximum key length from 32 to 64 bytes but
5564 there should be no binary compatibility issues as existing applications
5565 will never use XTS mode.
5569 * Extensive reorganisation of FIPS PRNG behaviour. Remove all dependencies
5570 to OpenSSL RAND code and replace with a tiny FIPS RAND API which also
5571 performs algorithm blocking for unapproved PRNG types. Also do not
5572 set PRNG type in FIPS_mode_set(): leave this to the application.
5573 Add default OpenSSL DRBG handling: sets up FIPS PRNG and seeds with
5574 the standard OpenSSL PRNG: set additional data to a date time vector.
5578 * Rename old X9.31 PRNG functions of the form `FIPS_rand*` to `FIPS_x931*`.
5579 This shouldn't present any incompatibility problems because applications
5580 shouldn't be using these directly and any that are will need to rethink
5581 anyway as the X9.31 PRNG is now deprecated by FIPS 140-2
5585 * Extensive self tests and health checking required by SP800-90 DRBG.
5586 Remove strength parameter from FIPS_drbg_instantiate and always
5587 instantiate at maximum supported strength.
5591 * Add ECDH code to fips module and fips_ecdhvs for primitives only testing.
5595 * New algorithm test program fips_dhvs to handle DH primitives only testing.
5599 * New function DH_compute_key_padded() to compute a DH key and pad with
5600 leading zeroes if needed: this complies with SP800-56A et al.
5604 * Initial implementation of SP800-90 DRBGs for Hash and CTR. Not used by
5605 anything, incomplete, subject to change and largely untested at present.
5609 * Modify fipscanisteronly build option to only build the necessary object
5610 files by filtering FIPS_EX_OBJ through a perl script in crypto/Makefile.
5614 * Add experimental option FIPSSYMS to give all symbols in
5615 fipscanister.o and FIPS or fips prefix. This will avoid
5616 conflicts with future versions of OpenSSL. Add perl script
5617 util/fipsas.pl to preprocess assembly language source files
5618 and rename any affected symbols.
5622 * Add selftest checks and algorithm block of non-fips algorithms in
5623 FIPS mode. Remove DES2 from selftests.
5627 * Add ECDSA code to fips module. Add tiny fips_ecdsa_check to just
5628 return internal method without any ENGINE dependencies. Add new
5629 tiny fips sign and verify functions.
5633 * New build option no-ec2m to disable characteristic 2 code.
5637 * New build option "fipscanisteronly". This only builds fipscanister.o
5638 and (currently) associated fips utilities. Uses the file Makefile.fips
5639 instead of Makefile.org as the prototype.
5643 * Add some FIPS mode restrictions to GCM. Add internal IV generator.
5644 Update fips_gcmtest to use IV generator.
5648 * Initial, experimental EVP support for AES-GCM. AAD can be input by
5649 setting output buffer to NULL. The `*Final` function must be
5650 called although it will not retrieve any additional data. The tag
5651 can be set or retrieved with a ctrl. The IV length is by default 12
5652 bytes (96 bits) but can be set to an alternative value. If the IV
5653 length exceeds the maximum IV length (currently 16 bytes) it cannot be
5658 * New flag in ciphers: EVP_CIPH_FLAG_CUSTOM_CIPHER. This means the
5659 underlying do_cipher function handles all cipher semantics itself
5660 including padding and finalisation. This is useful if (for example)
5661 an ENGINE cipher handles block padding itself. The behaviour of
5662 do_cipher is subtly changed if this flag is set: the return value
5663 is the number of characters written to the output buffer (zero is
5664 no longer an error code) or a negative error code. Also if the
5665 input buffer is NULL and length 0 finalisation should be performed.
5669 * If a candidate issuer certificate is already part of the constructed
5670 path ignore it: new debug notification X509_V_ERR_PATH_LOOP for this case.
5674 * Improve forward-security support: add functions
5676 void SSL_CTX_set_not_resumable_session_callback(
5677 SSL_CTX *ctx, int (*cb)(SSL *ssl, int is_forward_secure))
5678 void SSL_set_not_resumable_session_callback(
5679 SSL *ssl, int (*cb)(SSL *ssl, int is_forward_secure))
5681 for use by SSL/TLS servers; the callback function will be called whenever a
5682 new session is created, and gets to decide whether the session may be
5683 cached to make it resumable (return 0) or not (return 1). (As by the
5684 SSL/TLS protocol specifications, the session_id sent by the server will be
5685 empty to indicate that the session is not resumable; also, the server will
5686 not generate RFC 4507 (RFC 5077) session tickets.)
5688 A simple reasonable callback implementation is to return is_forward_secure.
5689 This parameter will be set to 1 or 0 depending on the ciphersuite selected
5690 by the SSL/TLS server library, indicating whether it can provide forward
5693 *Emilia Käsper <emilia.kasper@esat.kuleuven.be> (Google)*
5695 * New -verify_name option in command line utilities to set verification
5700 * Initial CMAC implementation. WARNING: EXPERIMENTAL, API MAY CHANGE.
5701 Add CMAC pkey methods.
5705 * Experimental renegotiation in s_server -www mode. If the client
5706 browses /reneg connection is renegotiated. If /renegcert it is
5707 renegotiated requesting a certificate.
5711 * Add an "external" session cache for debugging purposes to s_server. This
5712 should help trace issues which normally are only apparent in deployed
5713 multi-process servers.
5717 * Extensive audit of libcrypto with DEBUG_UNUSED. Fix many cases where
5718 return value is ignored. NB. The functions RAND_add(), RAND_seed(),
5719 BIO_set_cipher() and some obscure PEM functions were changed so they
5720 can now return an error. The RAND changes required a change to the
5721 RAND_METHOD structure.
5725 * New macro `__owur` for "OpenSSL Warn Unused Result". This makes use of
5726 a gcc attribute to warn if the result of a function is ignored. This
5727 is enable if DEBUG_UNUSED is set. Add to several functions in evp.h
5728 whose return value is often ignored.
5732 * New -noct, -requestct, -requirect and -ctlogfile options for s_client.
5733 These allow SCTs (signed certificate timestamps) to be requested and
5734 validated when establishing a connection.
5736 *Rob Percival <robpercival@google.com>*
5741 ### Changes between 1.0.2s and 1.0.2t [10 Sep 2019]
5743 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
5744 used even when parsing explicit parameters, when loading a encoded key
5745 or calling `EC_GROUP_new_from_ecpkparameters()`/
5746 `EC_GROUP_new_from_ecparameters()`.
5747 This prevents bypass of security hardening and performance gains,
5748 especially for curves with specialized EC_METHODs.
5749 By default, if a key encoded with explicit parameters is loaded and later
5750 encoded, the output is still encoded with explicit parameters, even if
5751 internally a "named" EC_GROUP is used for computation.
5755 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
5756 this change, EC_GROUP_set_generator would accept order and/or cofactor as
5757 NULL. After this change, only the cofactor parameter can be NULL. It also
5758 does some minimal sanity checks on the passed order.
5763 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
5764 An attack is simple, if the first CMS_recipientInfo is valid but the
5765 second CMS_recipientInfo is chosen ciphertext. If the second
5766 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
5767 encryption key will be replaced by garbage, and the message cannot be
5768 decoded, but if the RSA decryption fails, the correct encryption key is
5769 used and the recipient will not notice the attack.
5770 As a work around for this potential attack the length of the decrypted
5771 key must be equal to the cipher default key length, in case the
5772 certificate is not given and all recipientInfo are tried out.
5773 The old behaviour can be re-enabled in the CMS code by setting the
5774 CMS_DEBUG_DECRYPT flag.
5779 * Document issue with installation paths in diverse Windows builds
5781 '/usr/local/ssl' is an unsafe prefix for location to install OpenSSL
5782 binaries and run-time config file.
5787 ### Changes between 1.0.2r and 1.0.2s [28 May 2019]
5789 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
5790 This changes the size when using the `genpkey` command when no size is given.
5791 It fixes an omission in earlier changes that changed all RSA, DSA and DH
5792 generation commands to use 2048 bits by default.
5796 * Add FIPS support for Android Arm 64-bit
5798 Support for Android Arm 64-bit was added to the OpenSSL FIPS Object
5799 Module in Version 2.0.10. For some reason, the corresponding target
5800 'android64-aarch64' was missing OpenSSL 1.0.2, whence it could not be
5801 built with FIPS support on Android Arm 64-bit. This omission has been
5804 *Matthias St. Pierre*
5806 ### Changes between 1.0.2q and 1.0.2r [26 Feb 2019]
5808 * 0-byte record padding oracle
5810 If an application encounters a fatal protocol error and then calls
5811 SSL_shutdown() twice (once to send a close_notify, and once to receive one)
5812 then OpenSSL can respond differently to the calling application if a 0 byte
5813 record is received with invalid padding compared to if a 0 byte record is
5814 received with an invalid MAC. If the application then behaves differently
5815 based on that in a way that is detectable to the remote peer, then this
5816 amounts to a padding oracle that could be used to decrypt data.
5818 In order for this to be exploitable "non-stitched" ciphersuites must be in
5819 use. Stitched ciphersuites are optimised implementations of certain
5820 commonly used ciphersuites. Also the application must call SSL_shutdown()
5821 twice even if a protocol error has occurred (applications should not do
5822 this but some do anyway).
5824 This issue was discovered by Juraj Somorovsky, Robert Merget and Nimrod
5825 Aviram, with additional investigation by Steven Collison and Andrew
5826 Hourselt. It was reported to OpenSSL on 10th December 2018.
5831 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
5835 ### Changes between 1.0.2p and 1.0.2q [20 Nov 2018]
5837 * Microarchitecture timing vulnerability in ECC scalar multiplication
5839 OpenSSL ECC scalar multiplication, used in e.g. ECDSA and ECDH, has been
5840 shown to be vulnerable to a microarchitecture timing side channel attack.
5841 An attacker with sufficient access to mount local timing attacks during
5842 ECDSA signature generation could recover the private key.
5844 This issue was reported to OpenSSL on 26th October 2018 by Alejandro
5845 Cabrera Aldaya, Billy Brumley, Sohaib ul Hassan, Cesar Pereida Garcia and
5851 * Timing vulnerability in DSA signature generation
5853 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
5854 timing side channel attack. An attacker could use variations in the signing
5855 algorithm to recover the private key.
5857 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
5862 * Resolve a compatibility issue in EC_GROUP handling with the FIPS Object
5863 Module, accidentally introduced while backporting security fixes from the
5864 development branch and hindering the use of ECC in FIPS mode.
5868 ### Changes between 1.0.2o and 1.0.2p [14 Aug 2018]
5870 * Client DoS due to large DH parameter
5872 During key agreement in a TLS handshake using a DH(E) based ciphersuite a
5873 malicious server can send a very large prime value to the client. This will
5874 cause the client to spend an unreasonably long period of time generating a
5875 key for this prime resulting in a hang until the client has finished. This
5876 could be exploited in a Denial Of Service attack.
5878 This issue was reported to OpenSSL on 5th June 2018 by Guido Vranken
5883 * Cache timing vulnerability in RSA Key Generation
5885 The OpenSSL RSA Key generation algorithm has been shown to be vulnerable to
5886 a cache timing side channel attack. An attacker with sufficient access to
5887 mount cache timing attacks during the RSA key generation process could
5888 recover the private key.
5890 This issue was reported to OpenSSL on 4th April 2018 by Alejandro Cabrera
5891 Aldaya, Billy Brumley, Cesar Pereida Garcia and Luis Manuel Alvarez Tapia.
5896 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
5897 parameter is no longer accepted, as it leads to a corrupt table. NULL
5898 pem_str is reserved for alias entries only.
5902 * Revert blinding in ECDSA sign and instead make problematic addition
5903 length-invariant. Switch even to fixed-length Montgomery multiplication.
5907 * Change generating and checking of primes so that the error rate of not
5908 being prime depends on the intended use based on the size of the input.
5909 For larger primes this will result in more rounds of Miller-Rabin.
5910 The maximal error rate for primes with more than 1080 bits is lowered
5913 *Kurt Roeckx, Annie Yousar*
5915 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
5919 * Add blinding to ECDSA and DSA signatures to protect against side channel
5920 attacks discovered by Keegan Ryan (NCC Group).
5924 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
5925 now allow empty (zero character) pass phrases.
5929 * Certificate time validation (X509_cmp_time) enforces stricter
5930 compliance with RFC 5280. Fractional seconds and timezone offsets
5931 are no longer allowed.
5935 ### Changes between 1.0.2n and 1.0.2o [27 Mar 2018]
5937 * Constructed ASN.1 types with a recursive definition could exceed the stack
5939 Constructed ASN.1 types with a recursive definition (such as can be found
5940 in PKCS7) could eventually exceed the stack given malicious input with
5941 excessive recursion. This could result in a Denial Of Service attack. There
5942 are no such structures used within SSL/TLS that come from untrusted sources
5943 so this is considered safe.
5945 This issue was reported to OpenSSL on 4th January 2018 by the OSS-fuzz
5951 ### Changes between 1.0.2m and 1.0.2n [7 Dec 2017]
5953 * Read/write after SSL object in error state
5955 OpenSSL 1.0.2 (starting from version 1.0.2b) introduced an "error state"
5956 mechanism. The intent was that if a fatal error occurred during a handshake
5957 then OpenSSL would move into the error state and would immediately fail if
5958 you attempted to continue the handshake. This works as designed for the
5959 explicit handshake functions (SSL_do_handshake(), SSL_accept() and
5960 SSL_connect()), however due to a bug it does not work correctly if
5961 SSL_read() or SSL_write() is called directly. In that scenario, if the
5962 handshake fails then a fatal error will be returned in the initial function
5963 call. If SSL_read()/SSL_write() is subsequently called by the application
5964 for the same SSL object then it will succeed and the data is passed without
5965 being decrypted/encrypted directly from the SSL/TLS record layer.
5967 In order to exploit this issue an application bug would have to be present
5968 that resulted in a call to SSL_read()/SSL_write() being issued after having
5969 already received a fatal error.
5971 This issue was reported to OpenSSL by David Benjamin (Google).
5976 * rsaz_1024_mul_avx2 overflow bug on x86_64
5978 There is an overflow bug in the AVX2 Montgomery multiplication procedure
5979 used in exponentiation with 1024-bit moduli. No EC algorithms are affected.
5980 Analysis suggests that attacks against RSA and DSA as a result of this
5981 defect would be very difficult to perform and are not believed likely.
5982 Attacks against DH1024 are considered just feasible, because most of the
5983 work necessary to deduce information about a private key may be performed
5984 offline. The amount of resources required for such an attack would be
5985 significant. However, for an attack on TLS to be meaningful, the server
5986 would have to share the DH1024 private key among multiple clients, which is
5987 no longer an option since CVE-2016-0701.
5989 This only affects processors that support the AVX2 but not ADX extensions
5990 like Intel Haswell (4th generation).
5992 This issue was reported to OpenSSL by David Benjamin (Google). The issue
5993 was originally found via the OSS-Fuzz project.
5998 ### Changes between 1.0.2l and 1.0.2m [2 Nov 2017]
6000 * bn_sqrx8x_internal carry bug on x86_64
6002 There is a carry propagating bug in the x86_64 Montgomery squaring
6003 procedure. No EC algorithms are affected. Analysis suggests that attacks
6004 against RSA and DSA as a result of this defect would be very difficult to
6005 perform and are not believed likely. Attacks against DH are considered just
6006 feasible (although very difficult) because most of the work necessary to
6007 deduce information about a private key may be performed offline. The amount
6008 of resources required for such an attack would be very significant and
6009 likely only accessible to a limited number of attackers. An attacker would
6010 additionally need online access to an unpatched system using the target
6011 private key in a scenario with persistent DH parameters and a private
6012 key that is shared between multiple clients.
6014 This only affects processors that support the BMI1, BMI2 and ADX extensions
6015 like Intel Broadwell (5th generation) and later or AMD Ryzen.
6017 This issue was reported to OpenSSL by the OSS-Fuzz project.
6022 * Malformed X.509 IPAddressFamily could cause OOB read
6024 If an X.509 certificate has a malformed IPAddressFamily extension,
6025 OpenSSL could do a one-byte buffer overread. The most likely result
6026 would be an erroneous display of the certificate in text format.
6028 This issue was reported to OpenSSL by the OSS-Fuzz project.
6032 ### Changes between 1.0.2k and 1.0.2l [25 May 2017]
6034 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
6035 platform rather than 'mingw'.
6039 ### Changes between 1.0.2j and 1.0.2k [26 Jan 2017]
6041 * Truncated packet could crash via OOB read
6043 If one side of an SSL/TLS path is running on a 32-bit host and a specific
6044 cipher is being used, then a truncated packet can cause that host to
6045 perform an out-of-bounds read, usually resulting in a crash.
6047 This issue was reported to OpenSSL by Robert Święcki of Google.
6052 * BN_mod_exp may produce incorrect results on x86_64
6054 There is a carry propagating bug in the x86_64 Montgomery squaring
6055 procedure. No EC algorithms are affected. Analysis suggests that attacks
6056 against RSA and DSA as a result of this defect would be very difficult to
6057 perform and are not believed likely. Attacks against DH are considered just
6058 feasible (although very difficult) because most of the work necessary to
6059 deduce information about a private key may be performed offline. The amount
6060 of resources required for such an attack would be very significant and
6061 likely only accessible to a limited number of attackers. An attacker would
6062 additionally need online access to an unpatched system using the target
6063 private key in a scenario with persistent DH parameters and a private
6064 key that is shared between multiple clients. For example this can occur by
6065 default in OpenSSL DHE based SSL/TLS ciphersuites. Note: This issue is very
6066 similar to CVE-2015-3193 but must be treated as a separate problem.
6068 This issue was reported to OpenSSL by the OSS-Fuzz project.
6073 * Montgomery multiplication may produce incorrect results
6075 There is a carry propagating bug in the Broadwell-specific Montgomery
6076 multiplication procedure that handles input lengths divisible by, but
6077 longer than 256 bits. Analysis suggests that attacks against RSA, DSA
6078 and DH private keys are impossible. This is because the subroutine in
6079 question is not used in operations with the private key itself and an input
6080 of the attacker's direct choice. Otherwise the bug can manifest itself as
6081 transient authentication and key negotiation failures or reproducible
6082 erroneous outcome of public-key operations with specially crafted input.
6083 Among EC algorithms only Brainpool P-512 curves are affected and one
6084 presumably can attack ECDH key negotiation. Impact was not analyzed in
6085 detail, because pre-requisites for attack are considered unlikely. Namely
6086 multiple clients have to choose the curve in question and the server has to
6087 share the private key among them, neither of which is default behaviour.
6088 Even then only clients that chose the curve will be affected.
6090 This issue was publicly reported as transient failures and was not
6091 initially recognized as a security issue. Thanks to Richard Morgan for
6092 providing reproducible case.
6097 * OpenSSL now fails if it receives an unrecognised record type in TLS1.0
6098 or TLS1.1. Previously this only happened in SSLv3 and TLS1.2. This is to
6099 prevent issues where no progress is being made and the peer continually
6100 sends unrecognised record types, using up resources processing them.
6104 ### Changes between 1.0.2i and 1.0.2j [26 Sep 2016]
6106 * Missing CRL sanity check
6108 A bug fix which included a CRL sanity check was added to OpenSSL 1.1.0
6109 but was omitted from OpenSSL 1.0.2i. As a result any attempt to use
6110 CRLs in OpenSSL 1.0.2i will crash with a null pointer exception.
6112 This issue only affects the OpenSSL 1.0.2i
6117 ### Changes between 1.0.2h and 1.0.2i [22 Sep 2016]
6119 * OCSP Status Request extension unbounded memory growth
6121 A malicious client can send an excessively large OCSP Status Request
6122 extension. If that client continually requests renegotiation, sending a
6123 large OCSP Status Request extension each time, then there will be unbounded
6124 memory growth on the server. This will eventually lead to a Denial Of
6125 Service attack through memory exhaustion. Servers with a default
6126 configuration are vulnerable even if they do not support OCSP. Builds using
6127 the "no-ocsp" build time option are not affected.
6129 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6134 * In order to mitigate the SWEET32 attack, the DES ciphers were moved from
6137 This issue was reported to OpenSSL Karthikeyan Bhargavan and Gaetan
6143 * OOB write in MDC2_Update()
6145 An overflow can occur in MDC2_Update() either if called directly or
6146 through the EVP_DigestUpdate() function using MDC2. If an attacker
6147 is able to supply very large amounts of input data after a previous
6148 call to EVP_EncryptUpdate() with a partial block then a length check
6149 can overflow resulting in a heap corruption.
6151 The amount of data needed is comparable to SIZE_MAX which is impractical
6154 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6159 * Malformed SHA512 ticket DoS
6161 If a server uses SHA512 for TLS session ticket HMAC it is vulnerable to a
6162 DoS attack where a malformed ticket will result in an OOB read which will
6165 The use of SHA512 in TLS session tickets is comparatively rare as it requires
6166 a custom server callback and ticket lookup mechanism.
6168 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6173 * OOB write in BN_bn2dec()
6175 The function BN_bn2dec() does not check the return value of BN_div_word().
6176 This can cause an OOB write if an application uses this function with an
6177 overly large BIGNUM. This could be a problem if an overly large certificate
6178 or CRL is printed out from an untrusted source. TLS is not affected because
6179 record limits will reject an oversized certificate before it is parsed.
6181 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6186 * OOB read in TS_OBJ_print_bio()
6188 The function TS_OBJ_print_bio() misuses OBJ_obj2txt(): the return value is
6189 the total length the OID text representation would use and not the amount
6190 of data written. This will result in OOB reads when large OIDs are
6193 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6198 * Pointer arithmetic undefined behaviour
6200 Avoid some undefined pointer arithmetic
6202 A common idiom in the codebase is to check limits in the following manner:
6205 Where "p" points to some malloc'd data of SIZE bytes and
6208 "len" here could be from some externally supplied data (e.g. from a TLS
6211 The rules of C pointer arithmetic are such that "p + len" is only well
6212 defined where len <= SIZE. Therefore the above idiom is actually
6213 undefined behaviour.
6215 For example this could cause problems if some malloc implementation
6216 provides an address for "p" such that "p + len" actually overflows for
6217 values of len that are too big and therefore p + len < limit.
6219 This issue was reported to OpenSSL by Guido Vranken
6224 * Constant time flag not preserved in DSA signing
6226 Operations in the DSA signing algorithm should run in constant time in
6227 order to avoid side channel attacks. A flaw in the OpenSSL DSA
6228 implementation means that a non-constant time codepath is followed for
6229 certain operations. This has been demonstrated through a cache-timing
6230 attack to be sufficient for an attacker to recover the private DSA key.
6232 This issue was reported by César Pereida (Aalto University), Billy Brumley
6233 (Tampere University of Technology), and Yuval Yarom (The University of
6234 Adelaide and NICTA).
6239 * DTLS buffered message DoS
6241 In a DTLS connection where handshake messages are delivered out-of-order
6242 those messages that OpenSSL is not yet ready to process will be buffered
6243 for later use. Under certain circumstances, a flaw in the logic means that
6244 those messages do not get removed from the buffer even though the handshake
6245 has been completed. An attacker could force up to approx. 15 messages to
6246 remain in the buffer when they are no longer required. These messages will
6247 be cleared when the DTLS connection is closed. The default maximum size for
6248 a message is 100k. Therefore, the attacker could force an additional 1500k
6249 to be consumed per connection. By opening many simultaneous connections an
6250 attacker could cause a DoS attack through memory exhaustion.
6252 This issue was reported to OpenSSL by Quan Luo.
6257 * DTLS replay protection DoS
6259 A flaw in the DTLS replay attack protection mechanism means that records
6260 that arrive for future epochs update the replay protection "window" before
6261 the MAC for the record has been validated. This could be exploited by an
6262 attacker by sending a record for the next epoch (which does not have to
6263 decrypt or have a valid MAC), with a very large sequence number. This means
6264 that all subsequent legitimate packets are dropped causing a denial of
6265 service for a specific DTLS connection.
6267 This issue was reported to OpenSSL by the OCAP audit team.
6272 * Certificate message OOB reads
6274 In OpenSSL 1.0.2 and earlier some missing message length checks can result
6275 in OOB reads of up to 2 bytes beyond an allocated buffer. There is a
6276 theoretical DoS risk but this has not been observed in practice on common
6279 The messages affected are client certificate, client certificate request
6280 and server certificate. As a result the attack can only be performed
6281 against a client or a server which enables client authentication.
6283 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6288 ### Changes between 1.0.2g and 1.0.2h [3 May 2016]
6290 * Prevent padding oracle in AES-NI CBC MAC check
6292 A MITM attacker can use a padding oracle attack to decrypt traffic
6293 when the connection uses an AES CBC cipher and the server support
6296 This issue was introduced as part of the fix for Lucky 13 padding
6297 attack ([CVE-2013-0169]). The padding check was rewritten to be in
6298 constant time by making sure that always the same bytes are read and
6299 compared against either the MAC or padding bytes. But it no longer
6300 checked that there was enough data to have both the MAC and padding
6303 This issue was reported by Juraj Somorovsky using TLS-Attacker.
6307 * Fix EVP_EncodeUpdate overflow
6309 An overflow can occur in the EVP_EncodeUpdate() function which is used for
6310 Base64 encoding of binary data. If an attacker is able to supply very large
6311 amounts of input data then a length check can overflow resulting in a heap
6314 Internally to OpenSSL the EVP_EncodeUpdate() function is primarily used by
6315 the `PEM_write_bio*` family of functions. These are mainly used within the
6316 OpenSSL command line applications, so any application which processes data
6317 from an untrusted source and outputs it as a PEM file should be considered
6318 vulnerable to this issue. User applications that call these APIs directly
6319 with large amounts of untrusted data may also be vulnerable.
6321 This issue was reported by Guido Vranken.
6326 * Fix EVP_EncryptUpdate overflow
6328 An overflow can occur in the EVP_EncryptUpdate() function. If an attacker
6329 is able to supply very large amounts of input data after a previous call to
6330 EVP_EncryptUpdate() with a partial block then a length check can overflow
6331 resulting in a heap corruption. Following an analysis of all OpenSSL
6332 internal usage of the EVP_EncryptUpdate() function all usage is one of two
6333 forms. The first form is where the EVP_EncryptUpdate() call is known to be
6334 the first called function after an EVP_EncryptInit(), and therefore that
6335 specific call must be safe. The second form is where the length passed to
6336 EVP_EncryptUpdate() can be seen from the code to be some small value and
6337 therefore there is no possibility of an overflow. Since all instances are
6338 one of these two forms, it is believed that there can be no overflows in
6339 internal code due to this problem. It should be noted that
6340 EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths.
6341 Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances
6342 of these calls have also been analysed too and it is believed there are no
6343 instances in internal usage where an overflow could occur.
6345 This issue was reported by Guido Vranken.
6350 * Prevent ASN.1 BIO excessive memory allocation
6352 When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio()
6353 a short invalid encoding can cause allocation of large amounts of memory
6354 potentially consuming excessive resources or exhausting memory.
6356 Any application parsing untrusted data through d2i BIO functions is
6357 affected. The memory based functions such as d2i_X509() are *not* affected.
6358 Since the memory based functions are used by the TLS library, TLS
6359 applications are not affected.
6361 This issue was reported by Brian Carpenter.
6368 ASN1 Strings that are over 1024 bytes can cause an overread in applications
6369 using the X509_NAME_oneline() function on EBCDIC systems. This could result
6370 in arbitrary stack data being returned in the buffer.
6372 This issue was reported by Guido Vranken.
6377 * Modify behavior of ALPN to invoke callback after SNI/servername
6378 callback, such that updates to the SSL_CTX affect ALPN.
6382 * Remove LOW from the DEFAULT cipher list. This removes singles DES from the
6387 * Only remove the SSLv2 methods with the no-ssl2-method option. When the
6388 methods are enabled and ssl2 is disabled the methods return NULL.
6392 ### Changes between 1.0.2f and 1.0.2g [1 Mar 2016]
6394 * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL.
6395 Builds that are not configured with "enable-weak-ssl-ciphers" will not
6396 provide any "EXPORT" or "LOW" strength ciphers.
6400 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2
6401 is by default disabled at build-time. Builds that are not configured with
6402 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used,
6403 users who want to negotiate SSLv2 via the version-flexible SSLv23_method()
6404 will need to explicitly call either of:
6406 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2);
6408 SSL_clear_options(ssl, SSL_OP_NO_SSLv2);
6410 as appropriate. Even if either of those is used, or the application
6411 explicitly uses the version-specific SSLv2_method() or its client and
6412 server variants, SSLv2 ciphers vulnerable to exhaustive search key
6413 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT
6414 ciphers, and SSLv2 56-bit DES are no longer available.
6419 * Fix a double-free in DSA code
6421 A double free bug was discovered when OpenSSL parses malformed DSA private
6422 keys and could lead to a DoS attack or memory corruption for applications
6423 that receive DSA private keys from untrusted sources. This scenario is
6426 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using
6432 * Disable SRP fake user seed to address a server memory leak.
6434 Add a new method SRP_VBASE_get1_by_user that handles the seed properly.
6436 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
6437 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user
6438 was changed to ignore the "fake user" SRP seed, even if the seed
6441 Users should use SRP_VBASE_get1_by_user instead. Note that in
6442 SRP_VBASE_get1_by_user, caller must free the returned value. Note
6443 also that even though configuring the SRP seed attempts to hide
6444 invalid usernames by continuing the handshake with fake
6445 credentials, this behaviour is not constant time and no strong
6446 guarantees are made that the handshake is indistinguishable from
6447 that of a valid user.
6452 * Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption
6454 In the BN_hex2bn function the number of hex digits is calculated using an
6455 int value `i`. Later `bn_expand` is called with a value of `i * 4`. For
6456 large values of `i` this can result in `bn_expand` not allocating any
6457 memory because `i * 4` is negative. This can leave the internal BIGNUM data
6458 field as NULL leading to a subsequent NULL ptr deref. For very large values
6459 of `i`, the calculation `i * 4` could be a positive value smaller than `i`.
6460 In this case memory is allocated to the internal BIGNUM data field, but it
6461 is insufficiently sized leading to heap corruption. A similar issue exists
6462 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn
6463 is ever called by user applications with very large untrusted hex/dec data.
6464 This is anticipated to be a rare occurrence.
6466 All OpenSSL internal usage of these functions use data that is not expected
6467 to be untrusted, e.g. config file data or application command line
6468 arguments. If user developed applications generate config file data based
6469 on untrusted data then it is possible that this could also lead to security
6470 consequences. This is also anticipated to be rare.
6472 This issue was reported to OpenSSL by Guido Vranken.
6477 * Fix memory issues in `BIO_*printf` functions
6479 The internal `fmtstr` function used in processing a "%s" format string in
6480 the `BIO_*printf` functions could overflow while calculating the length of a
6481 string and cause an OOB read when printing very long strings.
6483 Additionally the internal `doapr_outch` function can attempt to write to an
6484 OOB memory location (at an offset from the NULL pointer) in the event of a
6485 memory allocation failure. In 1.0.2 and below this could be caused where
6486 the size of a buffer to be allocated is greater than INT_MAX. E.g. this
6487 could be in processing a very long "%s" format string. Memory leaks can
6490 The first issue may mask the second issue dependent on compiler behaviour.
6491 These problems could enable attacks where large amounts of untrusted data
6492 is passed to the `BIO_*printf` functions. If applications use these functions
6493 in this way then they could be vulnerable. OpenSSL itself uses these
6494 functions when printing out human-readable dumps of ASN.1 data. Therefore
6495 applications that print this data could be vulnerable if the data is from
6496 untrusted sources. OpenSSL command line applications could also be
6497 vulnerable where they print out ASN.1 data, or if untrusted data is passed
6498 as command line arguments.
6500 Libssl is not considered directly vulnerable. Additionally certificates etc
6501 received via remote connections via libssl are also unlikely to be able to
6502 trigger these issues because of message size limits enforced within libssl.
6504 This issue was reported to OpenSSL Guido Vranken.
6509 * Side channel attack on modular exponentiation
6511 A side-channel attack was found which makes use of cache-bank conflicts on
6512 the Intel Sandy-Bridge microarchitecture which could lead to the recovery
6513 of RSA keys. The ability to exploit this issue is limited as it relies on
6514 an attacker who has control of code in a thread running on the same
6515 hyper-threaded core as the victim thread which is performing decryptions.
6517 This issue was reported to OpenSSL by Yuval Yarom, The University of
6518 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and
6519 Nadia Heninger, University of Pennsylvania with more information at
6520 <http://cachebleed.info>.
6525 * Change the `req` command to generate a 2048-bit RSA/DSA key by default,
6526 if no keysize is specified with default_bits. This fixes an
6527 omission in an earlier change that changed all RSA/DSA key generation
6528 commands to use 2048 bits by default.
6532 ### Changes between 1.0.2e and 1.0.2f [28 Jan 2016]
6534 * DH small subgroups
6536 Historically OpenSSL only ever generated DH parameters based on "safe"
6537 primes. More recently (in version 1.0.2) support was provided for
6538 generating X9.42 style parameter files such as those required for RFC 5114
6539 support. The primes used in such files may not be "safe". Where an
6540 application is using DH configured with parameters based on primes that are
6541 not "safe" then an attacker could use this fact to find a peer's private
6542 DH exponent. This attack requires that the attacker complete multiple
6543 handshakes in which the peer uses the same private DH exponent. For example
6544 this could be used to discover a TLS server's private DH exponent if it's
6545 reusing the private DH exponent or it's using a static DH ciphersuite.
6547 OpenSSL provides the option SSL_OP_SINGLE_DH_USE for ephemeral DH (DHE) in
6548 TLS. It is not on by default. If the option is not set then the server
6549 reuses the same private DH exponent for the life of the server process and
6550 would be vulnerable to this attack. It is believed that many popular
6551 applications do set this option and would therefore not be at risk.
6553 The fix for this issue adds an additional check where a "q" parameter is
6554 available (as is the case in X9.42 based parameters). This detects the
6555 only known attack, and is the only possible defense for static DH
6556 ciphersuites. This could have some performance impact.
6558 Additionally the SSL_OP_SINGLE_DH_USE option has been switched on by
6559 default and cannot be disabled. This could have some performance impact.
6561 This issue was reported to OpenSSL by Antonio Sanso (Adobe).
6566 * SSLv2 doesn't block disabled ciphers
6568 A malicious client can negotiate SSLv2 ciphers that have been disabled on
6569 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
6570 been disabled, provided that the SSLv2 protocol was not also disabled via
6573 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
6574 and Sebastian Schinzel.
6579 ### Changes between 1.0.2d and 1.0.2e [3 Dec 2015]
6581 * BN_mod_exp may produce incorrect results on x86_64
6583 There is a carry propagating bug in the x86_64 Montgomery squaring
6584 procedure. No EC algorithms are affected. Analysis suggests that attacks
6585 against RSA and DSA as a result of this defect would be very difficult to
6586 perform and are not believed likely. Attacks against DH are considered just
6587 feasible (although very difficult) because most of the work necessary to
6588 deduce information about a private key may be performed offline. The amount
6589 of resources required for such an attack would be very significant and
6590 likely only accessible to a limited number of attackers. An attacker would
6591 additionally need online access to an unpatched system using the target
6592 private key in a scenario with persistent DH parameters and a private
6593 key that is shared between multiple clients. For example this can occur by
6594 default in OpenSSL DHE based SSL/TLS ciphersuites.
6596 This issue was reported to OpenSSL by Hanno Böck.
6601 * Certificate verify crash with missing PSS parameter
6603 The signature verification routines will crash with a NULL pointer
6604 dereference if presented with an ASN.1 signature using the RSA PSS
6605 algorithm and absent mask generation function parameter. Since these
6606 routines are used to verify certificate signature algorithms this can be
6607 used to crash any certificate verification operation and exploited in a
6608 DoS attack. Any application which performs certificate verification is
6609 vulnerable including OpenSSL clients and servers which enable client
6612 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
6617 * X509_ATTRIBUTE memory leak
6619 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
6620 memory. This structure is used by the PKCS#7 and CMS routines so any
6621 application which reads PKCS#7 or CMS data from untrusted sources is
6622 affected. SSL/TLS is not affected.
6624 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
6630 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
6631 This changes the decoding behaviour for some invalid messages,
6632 though the change is mostly in the more lenient direction, and
6633 legacy behaviour is preserved as much as possible.
6637 * In DSA_generate_parameters_ex, if the provided seed is too short,
6640 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
6642 ### Changes between 1.0.2c and 1.0.2d [9 Jul 2015]
6644 * Alternate chains certificate forgery
6646 During certificate verification, OpenSSL will attempt to find an
6647 alternative certificate chain if the first attempt to build such a chain
6648 fails. An error in the implementation of this logic can mean that an
6649 attacker could cause certain checks on untrusted certificates to be
6650 bypassed, such as the CA flag, enabling them to use a valid leaf
6651 certificate to act as a CA and "issue" an invalid certificate.
6653 This issue was reported to OpenSSL by Adam Langley/David Benjamin
6658 ### Changes between 1.0.2b and 1.0.2c [12 Jun 2015]
6660 * Fix HMAC ABI incompatibility. The previous version introduced an ABI
6661 incompatibility in the handling of HMAC. The previous ABI has now been
6666 ### Changes between 1.0.2a and 1.0.2b [11 Jun 2015]
6668 * Malformed ECParameters causes infinite loop
6670 When processing an ECParameters structure OpenSSL enters an infinite loop
6671 if the curve specified is over a specially malformed binary polynomial
6674 This can be used to perform denial of service against any
6675 system which processes public keys, certificate requests or
6676 certificates. This includes TLS clients and TLS servers with
6677 client authentication enabled.
6679 This issue was reported to OpenSSL by Joseph Barr-Pixton.
6684 * Exploitable out-of-bounds read in X509_cmp_time
6686 X509_cmp_time does not properly check the length of the ASN1_TIME
6687 string and can read a few bytes out of bounds. In addition,
6688 X509_cmp_time accepts an arbitrary number of fractional seconds in the
6691 An attacker can use this to craft malformed certificates and CRLs of
6692 various sizes and potentially cause a segmentation fault, resulting in
6693 a DoS on applications that verify certificates or CRLs. TLS clients
6694 that verify CRLs are affected. TLS clients and servers with client
6695 authentication enabled may be affected if they use custom verification
6698 This issue was reported to OpenSSL by Robert Swiecki (Google), and
6699 independently by Hanno Böck.
6704 * PKCS7 crash with missing EnvelopedContent
6706 The PKCS#7 parsing code does not handle missing inner EncryptedContent
6707 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
6708 with missing content and trigger a NULL pointer dereference on parsing.
6710 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
6711 structures from untrusted sources are affected. OpenSSL clients and
6712 servers are not affected.
6714 This issue was reported to OpenSSL by Michal Zalewski (Google).
6719 * CMS verify infinite loop with unknown hash function
6721 When verifying a signedData message the CMS code can enter an infinite loop
6722 if presented with an unknown hash function OID. This can be used to perform
6723 denial of service against any system which verifies signedData messages using
6725 This issue was reported to OpenSSL by Johannes Bauer.
6730 * Race condition handling NewSessionTicket
6732 If a NewSessionTicket is received by a multi-threaded client when attempting to
6733 reuse a previous ticket then a race condition can occur potentially leading to
6734 a double free of the ticket data.
6739 * Only support 256-bit or stronger elliptic curves with the
6740 'ecdh_auto' setting (server) or by default (client). Of supported
6741 curves, prefer P-256 (both).
6745 ### Changes between 1.0.2 and 1.0.2a [19 Mar 2015]
6747 * ClientHello sigalgs DoS fix
6749 If a client connects to an OpenSSL 1.0.2 server and renegotiates with an
6750 invalid signature algorithms extension a NULL pointer dereference will
6751 occur. This can be exploited in a DoS attack against the server.
6753 This issue was was reported to OpenSSL by David Ramos of Stanford
6757 *Stephen Henson and Matt Caswell*
6759 * Multiblock corrupted pointer fix
6761 OpenSSL 1.0.2 introduced the "multiblock" performance improvement. This
6762 feature only applies on 64 bit x86 architecture platforms that support AES
6763 NI instructions. A defect in the implementation of "multiblock" can cause
6764 OpenSSL's internal write buffer to become incorrectly set to NULL when
6765 using non-blocking IO. Typically, when the user application is using a
6766 socket BIO for writing, this will only result in a failed connection.
6767 However if some other BIO is used then it is likely that a segmentation
6768 fault will be triggered, thus enabling a potential DoS attack.
6770 This issue was reported to OpenSSL by Daniel Danner and Rainer Mueller.
6775 * Segmentation fault in DTLSv1_listen fix
6777 The DTLSv1_listen function is intended to be stateless and processes the
6778 initial ClientHello from many peers. It is common for user code to loop
6779 over the call to DTLSv1_listen until a valid ClientHello is received with
6780 an associated cookie. A defect in the implementation of DTLSv1_listen means
6781 that state is preserved in the SSL object from one invocation to the next
6782 that can lead to a segmentation fault. Errors processing the initial
6783 ClientHello can trigger this scenario. An example of such an error could be
6784 that a DTLS1.0 only client is attempting to connect to a DTLS1.2 only
6787 This issue was reported to OpenSSL by Per Allansson.
6792 * Segmentation fault in ASN1_TYPE_cmp fix
6794 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
6795 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
6796 certificate signature algorithm consistency this can be used to crash any
6797 certificate verification operation and exploited in a DoS attack. Any
6798 application which performs certificate verification is vulnerable including
6799 OpenSSL clients and servers which enable client authentication.
6804 * Segmentation fault for invalid PSS parameters fix
6806 The signature verification routines will crash with a NULL pointer
6807 dereference if presented with an ASN.1 signature using the RSA PSS
6808 algorithm and invalid parameters. Since these routines are used to verify
6809 certificate signature algorithms this can be used to crash any
6810 certificate verification operation and exploited in a DoS attack. Any
6811 application which performs certificate verification is vulnerable including
6812 OpenSSL clients and servers which enable client authentication.
6814 This issue was was reported to OpenSSL by Brian Carpenter.
6819 * ASN.1 structure reuse memory corruption fix
6821 Reusing a structure in ASN.1 parsing may allow an attacker to cause
6822 memory corruption via an invalid write. Such reuse is and has been
6823 strongly discouraged and is believed to be rare.
6825 Applications that parse structures containing CHOICE or ANY DEFINED BY
6826 components may be affected. Certificate parsing (d2i_X509 and related
6827 functions) are however not affected. OpenSSL clients and servers are
6833 * PKCS7 NULL pointer dereferences fix
6835 The PKCS#7 parsing code does not handle missing outer ContentInfo
6836 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
6837 missing content and trigger a NULL pointer dereference on parsing.
6839 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
6840 otherwise parse PKCS#7 structures from untrusted sources are
6841 affected. OpenSSL clients and servers are not affected.
6843 This issue was reported to OpenSSL by Michal Zalewski (Google).
6848 * DoS via reachable assert in SSLv2 servers fix
6850 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
6851 servers that both support SSLv2 and enable export cipher suites by sending
6852 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
6854 This issue was discovered by Sean Burford (Google) and Emilia Käsper
6855 (OpenSSL development team).
6860 * Empty CKE with client auth and DHE fix
6862 If client auth is used then a server can seg fault in the event of a DHE
6863 ciphersuite being selected and a zero length ClientKeyExchange message
6864 being sent by the client. This could be exploited in a DoS attack.
6869 * Handshake with unseeded PRNG fix
6871 Under certain conditions an OpenSSL 1.0.2 client can complete a handshake
6872 with an unseeded PRNG. The conditions are:
6873 - The client is on a platform where the PRNG has not been seeded
6874 automatically, and the user has not seeded manually
6875 - A protocol specific client method version has been used (i.e. not
6876 SSL_client_methodv23)
6877 - A ciphersuite is used that does not require additional random data from
6878 the PRNG beyond the initial ClientHello client random (e.g. PSK-RC4-SHA).
6880 If the handshake succeeds then the client random that has been used will
6881 have been generated from a PRNG with insufficient entropy and therefore the
6882 output may be predictable.
6884 For example using the following command with an unseeded openssl will
6885 succeed on an unpatched platform:
6887 openssl s_client -psk 1a2b3c4d -tls1_2 -cipher PSK-RC4-SHA
6892 * Use After Free following d2i_ECPrivatekey error fix
6894 A malformed EC private key file consumed via the d2i_ECPrivateKey function
6895 could cause a use after free condition. This, in turn, could cause a double
6896 free in several private key parsing functions (such as d2i_PrivateKey
6897 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
6898 for applications that receive EC private keys from untrusted
6899 sources. This scenario is considered rare.
6901 This issue was discovered by the BoringSSL project and fixed in their
6907 * X509_to_X509_REQ NULL pointer deref fix
6909 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
6910 the certificate key is invalid. This function is rarely used in practice.
6912 This issue was discovered by Brian Carpenter.
6917 * Removed the export ciphers from the DEFAULT ciphers
6921 ### Changes between 1.0.1l and 1.0.2 [22 Jan 2015]
6923 * Facilitate "universal" ARM builds targeting range of ARM ISAs, e.g.
6924 ARMv5 through ARMv8, as opposite to "locking" it to single one.
6925 So far those who have to target multiple platforms would compromise
6926 and argue that binary targeting say ARMv5 would still execute on
6927 ARMv8. "Universal" build resolves this compromise by providing
6928 near-optimal performance even on newer platforms.
6932 * Accelerated NIST P-256 elliptic curve implementation for x86_64
6933 (other platforms pending).
6935 *Shay Gueron & Vlad Krasnov (Intel Corp), Andy Polyakov*
6937 * Add support for the SignedCertificateTimestampList certificate and
6938 OCSP response extensions from RFC6962.
6942 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
6943 for corner cases. (Certain input points at infinity could lead to
6944 bogus results, with non-infinity inputs mapped to infinity too.)
6948 * Initial support for PowerISA 2.0.7, first implemented in POWER8.
6949 This covers AES, SHA256/512 and GHASH. "Initial" means that most
6950 common cases are optimized and there still is room for further
6951 improvements. Vector Permutation AES for Altivec is also added.
6955 * Add support for little-endian ppc64 Linux target.
6957 *Marcelo Cerri (IBM)*
6959 * Initial support for AMRv8 ISA crypto extensions. This covers AES,
6960 SHA1, SHA256 and GHASH. "Initial" means that most common cases
6961 are optimized and there still is room for further improvements.
6962 Both 32- and 64-bit modes are supported.
6964 *Andy Polyakov, Ard Biesheuvel (Linaro)*
6966 * Improved ARMv7 NEON support.
6970 * Support for SPARC Architecture 2011 crypto extensions, first
6971 implemented in SPARC T4. This covers AES, DES, Camellia, SHA1,
6972 SHA256/512, MD5, GHASH and modular exponentiation.
6974 *Andy Polyakov, David Miller*
6976 * Accelerated modular exponentiation for Intel processors, a.k.a.
6979 *Shay Gueron & Vlad Krasnov (Intel Corp)*
6981 * Support for new and upcoming Intel processors, including AVX2,
6982 BMI and SHA ISA extensions. This includes additional "stitched"
6983 implementations, AESNI-SHA256 and GCM, and multi-buffer support
6986 This work was sponsored by Intel Corp.
6990 * Support for DTLS 1.2. This adds two sets of DTLS methods: DTLS_*_method()
6991 supports both DTLS 1.2 and 1.0 and should use whatever version the peer
6992 supports and DTLSv1_2_*_method() which supports DTLS 1.2 only.
6996 * Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
6997 this fixes a limitation in previous versions of OpenSSL.
7001 * Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
7002 MGF1 digest and OAEP label.
7006 * Add EVP support for key wrapping algorithms, to avoid problems with
7007 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
7008 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
7009 algorithms and include tests cases.
7013 * Add functions to allocate and set the fields of an ECDSA_METHOD
7016 *Douglas E. Engert, Steve Henson*
7018 * New functions OPENSSL_gmtime_diff and ASN1_TIME_diff to find the
7019 difference in days and seconds between two tm or ASN1_TIME structures.
7023 * Add -rev test option to s_server to just reverse order of characters
7024 received by client and send back to server. Also prints an abbreviated
7025 summary of the connection parameters.
7029 * New option -brief for s_client and s_server to print out a brief summary
7030 of connection parameters.
7034 * Add callbacks for arbitrary TLS extensions.
7036 *Trevor Perrin <trevp@trevp.net> and Ben Laurie*
7038 * New option -crl_download in several openssl utilities to download CRLs
7039 from CRLDP extension in certificates.
7043 * New options -CRL and -CRLform for s_client and s_server for CRLs.
7047 * New function X509_CRL_diff to generate a delta CRL from the difference
7048 of two full CRLs. Add support to "crl" utility.
7052 * New functions to set lookup_crls function and to retrieve
7053 X509_STORE from X509_STORE_CTX.
7057 * Print out deprecated issuer and subject unique ID fields in
7062 * Extend OCSP I/O functions so they can be used for simple general purpose
7063 HTTP as well as OCSP. New wrapper function which can be used to download
7064 CRLs using the OCSP API.
7068 * Delegate command line handling in s_client/s_server to SSL_CONF APIs.
7072 * `SSL_CONF*` functions. These provide a common framework for application
7073 configuration using configuration files or command lines.
7077 * SSL/TLS tracing code. This parses out SSL/TLS records using the
7078 message callback and prints the results. Needs compile time option
7079 "enable-ssl-trace". New options to s_client and s_server to enable
7084 * New ctrl and macro to retrieve supported points extensions.
7085 Print out extension in s_server and s_client.
7089 * New functions to retrieve certificate signature and signature
7094 * Add functions to retrieve and manipulate the raw cipherlist sent by a
7099 * New Suite B modes for TLS code. These use and enforce the requirements
7100 of RFC6460: restrict ciphersuites, only permit Suite B algorithms and
7101 only use Suite B curves. The Suite B modes can be set by using the
7102 strings "SUITEB128", "SUITEB192" or "SUITEB128ONLY" for the cipherstring.
7106 * New chain verification flags for Suite B levels of security. Check
7107 algorithms are acceptable when flags are set in X509_verify_cert.
7111 * Make tls1_check_chain return a set of flags indicating checks passed
7112 by a certificate chain. Add additional tests to handle client
7113 certificates: checks for matching certificate type and issuer name
7118 * If an attempt is made to use a signature algorithm not in the peer
7119 preference list abort the handshake. If client has no suitable
7120 signature algorithms in response to a certificate request do not
7121 use the certificate.
7125 * If server EC tmp key is not in client preference list abort handshake.
7129 * Add support for certificate stores in CERT structure. This makes it
7130 possible to have different stores per SSL structure or one store in
7131 the parent SSL_CTX. Include distinct stores for certificate chain
7132 verification and chain building. New ctrl SSL_CTRL_BUILD_CERT_CHAIN
7133 to build and store a certificate chain in CERT structure: returning
7134 an error if the chain cannot be built: this will allow applications
7135 to test if a chain is correctly configured.
7137 Note: if the CERT based stores are not set then the parent SSL_CTX
7138 store is used to retain compatibility with existing behaviour.
7142 * New function ssl_set_client_disabled to set a ciphersuite disabled
7143 mask based on the current session, check mask when sending client
7144 hello and checking the requested ciphersuite.
7148 * New ctrls to retrieve and set certificate types in a certificate
7149 request message. Print out received values in s_client. If certificate
7150 types is not set with custom values set sensible values based on
7151 supported signature algorithms.
7155 * Support for distinct client and server supported signature algorithms.
7159 * Add certificate callback. If set this is called whenever a certificate
7160 is required by client or server. An application can decide which
7161 certificate chain to present based on arbitrary criteria: for example
7162 supported signature algorithms. Add very simple example to s_server.
7163 This fixes many of the problems and restrictions of the existing client
7164 certificate callback: for example you can now clear an existing
7165 certificate and specify the whole chain.
7169 * Add new "valid_flags" field to CERT_PKEY structure which determines what
7170 the certificate can be used for (if anything). Set valid_flags field
7171 in new tls1_check_chain function. Simplify ssl_set_cert_masks which used
7172 to have similar checks in it.
7174 Add new "cert_flags" field to CERT structure and include a "strict mode".
7175 This enforces some TLS certificate requirements (such as only permitting
7176 certificate signature algorithms contained in the supported algorithms
7177 extension) which some implementations ignore: this option should be used
7178 with caution as it could cause interoperability issues.
7182 * Update and tidy signature algorithm extension processing. Work out
7183 shared signature algorithms based on preferences and peer algorithms
7184 and print them out in s_client and s_server. Abort handshake if no
7185 shared signature algorithms.
7189 * Add new functions to allow customised supported signature algorithms
7190 for SSL and SSL_CTX structures. Add options to s_client and s_server
7195 * New function SSL_certs_clear() to delete all references to certificates
7196 from an SSL structure. Before this once a certificate had been added
7197 it couldn't be removed.
7201 * Integrate hostname, email address and IP address checking with certificate
7202 verification. New verify options supporting checking in openssl utility.
7206 * Fixes and wildcard matching support to hostname and email checking
7207 functions. Add manual page.
7209 *Florian Weimer (Red Hat Product Security Team)*
7211 * New functions to check a hostname email or IP address against a
7212 certificate. Add options x509 utility to print results of checks against
7217 * Fix OCSP checking.
7219 *Rob Stradling <rob.stradling@comodo.com> and Ben Laurie*
7221 * Initial experimental support for explicitly trusted non-root CAs.
7222 OpenSSL still tries to build a complete chain to a root but if an
7223 intermediate CA has a trust setting included that is used. The first
7224 setting is used: whether to trust (e.g., -addtrust option to the x509
7229 * Add -trusted_first option which attempts to find certificates in the
7230 trusted store even if an untrusted chain is also supplied.
7234 * MIPS assembly pack updates: support for MIPS32r2 and SmartMIPS ASE,
7235 platform support for Linux and Android.
7239 * Support for linux-x32, ILP32 environment in x86_64 framework.
7243 * Experimental multi-implementation support for FIPS capable OpenSSL.
7244 When in FIPS mode the approved implementations are used as normal,
7245 when not in FIPS mode the internal unapproved versions are used instead.
7246 This means that the FIPS capable OpenSSL isn't forced to use the
7247 (often lower performance) FIPS implementations outside FIPS mode.
7251 * Transparently support X9.42 DH parameters when calling
7252 PEM_read_bio_DHparameters. This means existing applications can handle
7253 the new parameter format automatically.
7257 * Initial experimental support for X9.42 DH parameter format: mainly
7258 to support use of 'q' parameter for RFC5114 parameters.
7262 * Add DH parameters from RFC5114 including test data to dhtest.
7266 * Support for automatic EC temporary key parameter selection. If enabled
7267 the most preferred EC parameters are automatically used instead of
7268 hardcoded fixed parameters. Now a server just has to call:
7269 SSL_CTX_set_ecdh_auto(ctx, 1) and the server will automatically
7270 support ECDH and use the most appropriate parameters.
7274 * Enhance and tidy EC curve and point format TLS extension code. Use
7275 static structures instead of allocation if default values are used.
7276 New ctrls to set curves we wish to support and to retrieve shared curves.
7277 Print out shared curves in s_server. New options to s_server and s_client
7278 to set list of supported curves.
7282 * New ctrls to retrieve supported signature algorithms and
7283 supported curve values as an array of NIDs. Extend openssl utility
7284 to print out received values.
7288 * Add new APIs EC_curve_nist2nid and EC_curve_nid2nist which convert
7289 between NIDs and the more common NIST names such as "P-256". Enhance
7290 ecparam utility and ECC method to recognise the NIST names for curves.
7294 * Enhance SSL/TLS certificate chain handling to support different
7295 chains for each certificate instead of one chain in the parent SSL_CTX.
7299 * Support for fixed DH ciphersuite client authentication: where both
7300 server and client use DH certificates with common parameters.
7304 * Support for fixed DH ciphersuites: those requiring DH server
7309 * New function i2d_re_X509_tbs for re-encoding the TBS portion of
7311 Note: Related 1.0.2-beta specific macros X509_get_cert_info,
7312 X509_CINF_set_modified, X509_CINF_get_issuer, X509_CINF_get_extensions and
7313 X509_CINF_get_signature were reverted post internal team review.
7318 ### Changes between 1.0.1t and 1.0.1u [22 Sep 2016]
7320 * OCSP Status Request extension unbounded memory growth
7322 A malicious client can send an excessively large OCSP Status Request
7323 extension. If that client continually requests renegotiation, sending a
7324 large OCSP Status Request extension each time, then there will be unbounded
7325 memory growth on the server. This will eventually lead to a Denial Of
7326 Service attack through memory exhaustion. Servers with a default
7327 configuration are vulnerable even if they do not support OCSP. Builds using
7328 the "no-ocsp" build time option are not affected.
7330 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
7335 * In order to mitigate the SWEET32 attack, the DES ciphers were moved from
7338 This issue was reported to OpenSSL Karthikeyan Bhargavan and Gaetan
7344 * OOB write in MDC2_Update()
7346 An overflow can occur in MDC2_Update() either if called directly or
7347 through the EVP_DigestUpdate() function using MDC2. If an attacker
7348 is able to supply very large amounts of input data after a previous
7349 call to EVP_EncryptUpdate() with a partial block then a length check
7350 can overflow resulting in a heap corruption.
7352 The amount of data needed is comparable to SIZE_MAX which is impractical
7355 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
7360 * Malformed SHA512 ticket DoS
7362 If a server uses SHA512 for TLS session ticket HMAC it is vulnerable to a
7363 DoS attack where a malformed ticket will result in an OOB read which will
7366 The use of SHA512 in TLS session tickets is comparatively rare as it requires
7367 a custom server callback and ticket lookup mechanism.
7369 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
7374 * OOB write in BN_bn2dec()
7376 The function BN_bn2dec() does not check the return value of BN_div_word().
7377 This can cause an OOB write if an application uses this function with an
7378 overly large BIGNUM. This could be a problem if an overly large certificate
7379 or CRL is printed out from an untrusted source. TLS is not affected because
7380 record limits will reject an oversized certificate before it is parsed.
7382 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
7387 * OOB read in TS_OBJ_print_bio()
7389 The function TS_OBJ_print_bio() misuses OBJ_obj2txt(): the return value is
7390 the total length the OID text representation would use and not the amount
7391 of data written. This will result in OOB reads when large OIDs are
7394 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
7399 * Pointer arithmetic undefined behaviour
7401 Avoid some undefined pointer arithmetic
7403 A common idiom in the codebase is to check limits in the following manner:
7406 Where "p" points to some malloc'd data of SIZE bytes and
7409 "len" here could be from some externally supplied data (e.g. from a TLS
7412 The rules of C pointer arithmetic are such that "p + len" is only well
7413 defined where len <= SIZE. Therefore, the above idiom is actually
7414 undefined behaviour.
7416 For example this could cause problems if some malloc implementation
7417 provides an address for "p" such that "p + len" actually overflows for
7418 values of len that are too big and therefore p + len < limit.
7420 This issue was reported to OpenSSL by Guido Vranken
7425 * Constant time flag not preserved in DSA signing
7427 Operations in the DSA signing algorithm should run in constant time in
7428 order to avoid side channel attacks. A flaw in the OpenSSL DSA
7429 implementation means that a non-constant time codepath is followed for
7430 certain operations. This has been demonstrated through a cache-timing
7431 attack to be sufficient for an attacker to recover the private DSA key.
7433 This issue was reported by César Pereida (Aalto University), Billy Brumley
7434 (Tampere University of Technology), and Yuval Yarom (The University of
7435 Adelaide and NICTA).
7440 * DTLS buffered message DoS
7442 In a DTLS connection where handshake messages are delivered out-of-order
7443 those messages that OpenSSL is not yet ready to process will be buffered
7444 for later use. Under certain circumstances, a flaw in the logic means that
7445 those messages do not get removed from the buffer even though the handshake
7446 has been completed. An attacker could force up to approx. 15 messages to
7447 remain in the buffer when they are no longer required. These messages will
7448 be cleared when the DTLS connection is closed. The default maximum size for
7449 a message is 100k. Therefore, the attacker could force an additional 1500k
7450 to be consumed per connection. By opening many simultaneous connections an
7451 attacker could cause a DoS attack through memory exhaustion.
7453 This issue was reported to OpenSSL by Quan Luo.
7458 * DTLS replay protection DoS
7460 A flaw in the DTLS replay attack protection mechanism means that records
7461 that arrive for future epochs update the replay protection "window" before
7462 the MAC for the record has been validated. This could be exploited by an
7463 attacker by sending a record for the next epoch (which does not have to
7464 decrypt or have a valid MAC), with a very large sequence number. This means
7465 that all subsequent legitimate packets are dropped causing a denial of
7466 service for a specific DTLS connection.
7468 This issue was reported to OpenSSL by the OCAP audit team.
7473 * Certificate message OOB reads
7475 In OpenSSL 1.0.2 and earlier some missing message length checks can result
7476 in OOB reads of up to 2 bytes beyond an allocated buffer. There is a
7477 theoretical DoS risk but this has not been observed in practice on common
7480 The messages affected are client certificate, client certificate request
7481 and server certificate. As a result the attack can only be performed
7482 against a client or a server which enables client authentication.
7484 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
7489 ### Changes between 1.0.1s and 1.0.1t [3 May 2016]
7491 * Prevent padding oracle in AES-NI CBC MAC check
7493 A MITM attacker can use a padding oracle attack to decrypt traffic
7494 when the connection uses an AES CBC cipher and the server support
7497 This issue was introduced as part of the fix for Lucky 13 padding
7498 attack ([CVE-2013-0169]). The padding check was rewritten to be in
7499 constant time by making sure that always the same bytes are read and
7500 compared against either the MAC or padding bytes. But it no longer
7501 checked that there was enough data to have both the MAC and padding
7504 This issue was reported by Juraj Somorovsky using TLS-Attacker.
7509 * Fix EVP_EncodeUpdate overflow
7511 An overflow can occur in the EVP_EncodeUpdate() function which is used for
7512 Base64 encoding of binary data. If an attacker is able to supply very large
7513 amounts of input data then a length check can overflow resulting in a heap
7516 Internally to OpenSSL the EVP_EncodeUpdate() function is primarily used by
7517 the `PEM_write_bio*` family of functions. These are mainly used within the
7518 OpenSSL command line applications, so any application which processes data
7519 from an untrusted source and outputs it as a PEM file should be considered
7520 vulnerable to this issue. User applications that call these APIs directly
7521 with large amounts of untrusted data may also be vulnerable.
7523 This issue was reported by Guido Vranken.
7528 * Fix EVP_EncryptUpdate overflow
7530 An overflow can occur in the EVP_EncryptUpdate() function. If an attacker
7531 is able to supply very large amounts of input data after a previous call to
7532 EVP_EncryptUpdate() with a partial block then a length check can overflow
7533 resulting in a heap corruption. Following an analysis of all OpenSSL
7534 internal usage of the EVP_EncryptUpdate() function all usage is one of two
7535 forms. The first form is where the EVP_EncryptUpdate() call is known to be
7536 the first called function after an EVP_EncryptInit(), and therefore that
7537 specific call must be safe. The second form is where the length passed to
7538 EVP_EncryptUpdate() can be seen from the code to be some small value and
7539 therefore there is no possibility of an overflow. Since all instances are
7540 one of these two forms, it is believed that there can be no overflows in
7541 internal code due to this problem. It should be noted that
7542 EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths.
7543 Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances
7544 of these calls have also been analysed too and it is believed there are no
7545 instances in internal usage where an overflow could occur.
7547 This issue was reported by Guido Vranken.
7552 * Prevent ASN.1 BIO excessive memory allocation
7554 When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio()
7555 a short invalid encoding can casuse allocation of large amounts of memory
7556 potentially consuming excessive resources or exhausting memory.
7558 Any application parsing untrusted data through d2i BIO functions is
7559 affected. The memory based functions such as d2i_X509() are *not* affected.
7560 Since the memory based functions are used by the TLS library, TLS
7561 applications are not affected.
7563 This issue was reported by Brian Carpenter.
7570 ASN1 Strings that are over 1024 bytes can cause an overread in applications
7571 using the X509_NAME_oneline() function on EBCDIC systems. This could result
7572 in arbitrary stack data being returned in the buffer.
7574 This issue was reported by Guido Vranken.
7579 * Modify behavior of ALPN to invoke callback after SNI/servername
7580 callback, such that updates to the SSL_CTX affect ALPN.
7584 * Remove LOW from the DEFAULT cipher list. This removes singles DES from the
7589 * Only remove the SSLv2 methods with the no-ssl2-method option. When the
7590 methods are enabled and ssl2 is disabled the methods return NULL.
7594 ### Changes between 1.0.1r and 1.0.1s [1 Mar 2016]
7596 * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL.
7597 Builds that are not configured with "enable-weak-ssl-ciphers" will not
7598 provide any "EXPORT" or "LOW" strength ciphers.
7602 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2
7603 is by default disabled at build-time. Builds that are not configured with
7604 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used,
7605 users who want to negotiate SSLv2 via the version-flexible SSLv23_method()
7606 will need to explicitly call either of:
7608 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2);
7610 SSL_clear_options(ssl, SSL_OP_NO_SSLv2);
7612 as appropriate. Even if either of those is used, or the application
7613 explicitly uses the version-specific SSLv2_method() or its client and
7614 server variants, SSLv2 ciphers vulnerable to exhaustive search key
7615 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT
7616 ciphers, and SSLv2 56-bit DES are no longer available.
7621 * Fix a double-free in DSA code
7623 A double free bug was discovered when OpenSSL parses malformed DSA private
7624 keys and could lead to a DoS attack or memory corruption for applications
7625 that receive DSA private keys from untrusted sources. This scenario is
7628 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using
7634 * Disable SRP fake user seed to address a server memory leak.
7636 Add a new method SRP_VBASE_get1_by_user that handles the seed properly.
7638 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
7639 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user
7640 was changed to ignore the "fake user" SRP seed, even if the seed
7643 Users should use SRP_VBASE_get1_by_user instead. Note that in
7644 SRP_VBASE_get1_by_user, caller must free the returned value. Note
7645 also that even though configuring the SRP seed attempts to hide
7646 invalid usernames by continuing the handshake with fake
7647 credentials, this behaviour is not constant time and no strong
7648 guarantees are made that the handshake is indistinguishable from
7649 that of a valid user.
7654 * Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption
7656 In the BN_hex2bn function the number of hex digits is calculated using an
7657 int value `i`. Later `bn_expand` is called with a value of `i * 4`. For
7658 large values of `i` this can result in `bn_expand` not allocating any
7659 memory because `i * 4` is negative. This can leave the internal BIGNUM data
7660 field as NULL leading to a subsequent NULL ptr deref. For very large values
7661 of `i`, the calculation `i * 4` could be a positive value smaller than `i`.
7662 In this case memory is allocated to the internal BIGNUM data field, but it
7663 is insufficiently sized leading to heap corruption. A similar issue exists
7664 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn
7665 is ever called by user applications with very large untrusted hex/dec data.
7666 This is anticipated to be a rare occurrence.
7668 All OpenSSL internal usage of these functions use data that is not expected
7669 to be untrusted, e.g. config file data or application command line
7670 arguments. If user developed applications generate config file data based
7671 on untrusted data then it is possible that this could also lead to security
7672 consequences. This is also anticipated to be rare.
7674 This issue was reported to OpenSSL by Guido Vranken.
7679 * Fix memory issues in `BIO_*printf` functions
7681 The internal `fmtstr` function used in processing a "%s" format string in
7682 the `BIO_*printf` functions could overflow while calculating the length of a
7683 string and cause an OOB read when printing very long strings.
7685 Additionally the internal `doapr_outch` function can attempt to write to an
7686 OOB memory location (at an offset from the NULL pointer) in the event of a
7687 memory allocation failure. In 1.0.2 and below this could be caused where
7688 the size of a buffer to be allocated is greater than INT_MAX. E.g. this
7689 could be in processing a very long "%s" format string. Memory leaks can
7692 The first issue may mask the second issue dependent on compiler behaviour.
7693 These problems could enable attacks where large amounts of untrusted data
7694 is passed to the `BIO_*printf` functions. If applications use these functions
7695 in this way then they could be vulnerable. OpenSSL itself uses these
7696 functions when printing out human-readable dumps of ASN.1 data. Therefore
7697 applications that print this data could be vulnerable if the data is from
7698 untrusted sources. OpenSSL command line applications could also be
7699 vulnerable where they print out ASN.1 data, or if untrusted data is passed
7700 as command line arguments.
7702 Libssl is not considered directly vulnerable. Additionally certificates etc
7703 received via remote connections via libssl are also unlikely to be able to
7704 trigger these issues because of message size limits enforced within libssl.
7706 This issue was reported to OpenSSL Guido Vranken.
7711 * Side channel attack on modular exponentiation
7713 A side-channel attack was found which makes use of cache-bank conflicts on
7714 the Intel Sandy-Bridge microarchitecture which could lead to the recovery
7715 of RSA keys. The ability to exploit this issue is limited as it relies on
7716 an attacker who has control of code in a thread running on the same
7717 hyper-threaded core as the victim thread which is performing decryptions.
7719 This issue was reported to OpenSSL by Yuval Yarom, The University of
7720 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and
7721 Nadia Heninger, University of Pennsylvania with more information at
7722 <http://cachebleed.info>.
7727 * Change the req command to generate a 2048-bit RSA/DSA key by default,
7728 if no keysize is specified with default_bits. This fixes an
7729 omission in an earlier change that changed all RSA/DSA key generation
7730 commands to use 2048 bits by default.
7734 ### Changes between 1.0.1q and 1.0.1r [28 Jan 2016]
7736 * Protection for DH small subgroup attacks
7738 As a precautionary measure the SSL_OP_SINGLE_DH_USE option has been
7739 switched on by default and cannot be disabled. This could have some
7744 * SSLv2 doesn't block disabled ciphers
7746 A malicious client can negotiate SSLv2 ciphers that have been disabled on
7747 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
7748 been disabled, provided that the SSLv2 protocol was not also disabled via
7751 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
7752 and Sebastian Schinzel.
7757 * Reject DH handshakes with parameters shorter than 1024 bits.
7761 ### Changes between 1.0.1p and 1.0.1q [3 Dec 2015]
7763 * Certificate verify crash with missing PSS parameter
7765 The signature verification routines will crash with a NULL pointer
7766 dereference if presented with an ASN.1 signature using the RSA PSS
7767 algorithm and absent mask generation function parameter. Since these
7768 routines are used to verify certificate signature algorithms this can be
7769 used to crash any certificate verification operation and exploited in a
7770 DoS attack. Any application which performs certificate verification is
7771 vulnerable including OpenSSL clients and servers which enable client
7774 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
7779 * X509_ATTRIBUTE memory leak
7781 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
7782 memory. This structure is used by the PKCS#7 and CMS routines so any
7783 application which reads PKCS#7 or CMS data from untrusted sources is
7784 affected. SSL/TLS is not affected.
7786 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
7792 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
7793 This changes the decoding behaviour for some invalid messages,
7794 though the change is mostly in the more lenient direction, and
7795 legacy behaviour is preserved as much as possible.
7799 * In DSA_generate_parameters_ex, if the provided seed is too short,
7800 use a random seed, as already documented.
7802 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
7804 ### Changes between 1.0.1o and 1.0.1p [9 Jul 2015]
7806 * Alternate chains certificate forgery
7808 During certificate verification, OpenSSL will attempt to find an
7809 alternative certificate chain if the first attempt to build such a chain
7810 fails. An error in the implementation of this logic can mean that an
7811 attacker could cause certain checks on untrusted certificates to be
7812 bypassed, such as the CA flag, enabling them to use a valid leaf
7813 certificate to act as a CA and "issue" an invalid certificate.
7815 This issue was reported to OpenSSL by Adam Langley/David Benjamin
7821 * Race condition handling PSK identify hint
7823 If PSK identity hints are received by a multi-threaded client then
7824 the values are wrongly updated in the parent SSL_CTX structure. This can
7825 result in a race condition potentially leading to a double free of the
7831 ### Changes between 1.0.1n and 1.0.1o [12 Jun 2015]
7833 * Fix HMAC ABI incompatibility. The previous version introduced an ABI
7834 incompatibility in the handling of HMAC. The previous ABI has now been
7837 ### Changes between 1.0.1m and 1.0.1n [11 Jun 2015]
7839 * Malformed ECParameters causes infinite loop
7841 When processing an ECParameters structure OpenSSL enters an infinite loop
7842 if the curve specified is over a specially malformed binary polynomial
7845 This can be used to perform denial of service against any
7846 system which processes public keys, certificate requests or
7847 certificates. This includes TLS clients and TLS servers with
7848 client authentication enabled.
7850 This issue was reported to OpenSSL by Joseph Barr-Pixton.
7855 * Exploitable out-of-bounds read in X509_cmp_time
7857 X509_cmp_time does not properly check the length of the ASN1_TIME
7858 string and can read a few bytes out of bounds. In addition,
7859 X509_cmp_time accepts an arbitrary number of fractional seconds in the
7862 An attacker can use this to craft malformed certificates and CRLs of
7863 various sizes and potentially cause a segmentation fault, resulting in
7864 a DoS on applications that verify certificates or CRLs. TLS clients
7865 that verify CRLs are affected. TLS clients and servers with client
7866 authentication enabled may be affected if they use custom verification
7869 This issue was reported to OpenSSL by Robert Swiecki (Google), and
7870 independently by Hanno Böck.
7875 * PKCS7 crash with missing EnvelopedContent
7877 The PKCS#7 parsing code does not handle missing inner EncryptedContent
7878 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
7879 with missing content and trigger a NULL pointer dereference on parsing.
7881 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
7882 structures from untrusted sources are affected. OpenSSL clients and
7883 servers are not affected.
7885 This issue was reported to OpenSSL by Michal Zalewski (Google).
7890 * CMS verify infinite loop with unknown hash function
7892 When verifying a signedData message the CMS code can enter an infinite loop
7893 if presented with an unknown hash function OID. This can be used to perform
7894 denial of service against any system which verifies signedData messages using
7896 This issue was reported to OpenSSL by Johannes Bauer.
7901 * Race condition handling NewSessionTicket
7903 If a NewSessionTicket is received by a multi-threaded client when attempting to
7904 reuse a previous ticket then a race condition can occur potentially leading to
7905 a double free of the ticket data.
7910 * Reject DH handshakes with parameters shorter than 768 bits.
7912 *Kurt Roeckx and Emilia Kasper*
7914 * dhparam: generate 2048-bit parameters by default.
7916 *Kurt Roeckx and Emilia Kasper*
7918 ### Changes between 1.0.1l and 1.0.1m [19 Mar 2015]
7920 * Segmentation fault in ASN1_TYPE_cmp fix
7922 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
7923 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
7924 certificate signature algorithm consistency this can be used to crash any
7925 certificate verification operation and exploited in a DoS attack. Any
7926 application which performs certificate verification is vulnerable including
7927 OpenSSL clients and servers which enable client authentication.
7932 * ASN.1 structure reuse memory corruption fix
7934 Reusing a structure in ASN.1 parsing may allow an attacker to cause
7935 memory corruption via an invalid write. Such reuse is and has been
7936 strongly discouraged and is believed to be rare.
7938 Applications that parse structures containing CHOICE or ANY DEFINED BY
7939 components may be affected. Certificate parsing (d2i_X509 and related
7940 functions) are however not affected. OpenSSL clients and servers are
7946 * PKCS7 NULL pointer dereferences fix
7948 The PKCS#7 parsing code does not handle missing outer ContentInfo
7949 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
7950 missing content and trigger a NULL pointer dereference on parsing.
7952 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
7953 otherwise parse PKCS#7 structures from untrusted sources are
7954 affected. OpenSSL clients and servers are not affected.
7956 This issue was reported to OpenSSL by Michal Zalewski (Google).
7961 * DoS via reachable assert in SSLv2 servers fix
7963 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
7964 servers that both support SSLv2 and enable export cipher suites by sending
7965 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
7967 This issue was discovered by Sean Burford (Google) and Emilia Käsper
7968 (OpenSSL development team).
7973 * Use After Free following d2i_ECPrivatekey error fix
7975 A malformed EC private key file consumed via the d2i_ECPrivateKey function
7976 could cause a use after free condition. This, in turn, could cause a double
7977 free in several private key parsing functions (such as d2i_PrivateKey
7978 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
7979 for applications that receive EC private keys from untrusted
7980 sources. This scenario is considered rare.
7982 This issue was discovered by the BoringSSL project and fixed in their
7988 * X509_to_X509_REQ NULL pointer deref fix
7990 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
7991 the certificate key is invalid. This function is rarely used in practice.
7993 This issue was discovered by Brian Carpenter.
7998 * Removed the export ciphers from the DEFAULT ciphers
8002 ### Changes between 1.0.1k and 1.0.1l [15 Jan 2015]
8004 * Build fixes for the Windows and OpenVMS platforms
8006 *Matt Caswell and Richard Levitte*
8008 ### Changes between 1.0.1j and 1.0.1k [8 Jan 2015]
8010 * Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
8011 message can cause a segmentation fault in OpenSSL due to a NULL pointer
8012 dereference. This could lead to a Denial Of Service attack. Thanks to
8013 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
8018 * Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
8019 dtls1_buffer_record function under certain conditions. In particular this
8020 could occur if an attacker sent repeated DTLS records with the same
8021 sequence number but for the next epoch. The memory leak could be exploited
8022 by an attacker in a Denial of Service attack through memory exhaustion.
8023 Thanks to Chris Mueller for reporting this issue.
8028 * Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
8029 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
8030 method would be set to NULL which could later result in a NULL pointer
8031 dereference. Thanks to Frank Schmirler for reporting this issue.
8036 * Abort handshake if server key exchange message is omitted for ephemeral
8039 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
8040 reporting this issue.
8045 * Remove non-export ephemeral RSA code on client and server. This code
8046 violated the TLS standard by allowing the use of temporary RSA keys in
8047 non-export ciphersuites and could be used by a server to effectively
8048 downgrade the RSA key length used to a value smaller than the server
8049 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
8050 INRIA or reporting this issue.
8055 * Fixed issue where DH client certificates are accepted without verification.
8056 An OpenSSL server will accept a DH certificate for client authentication
8057 without the certificate verify message. This effectively allows a client to
8058 authenticate without the use of a private key. This only affects servers
8059 which trust a client certificate authority which issues certificates
8060 containing DH keys: these are extremely rare and hardly ever encountered.
8061 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
8067 * Ensure that the session ID context of an SSL is updated when its
8068 SSL_CTX is updated via SSL_set_SSL_CTX.
8070 The session ID context is typically set from the parent SSL_CTX,
8071 and can vary with the CTX.
8075 * Fix various certificate fingerprint issues.
8077 By using non-DER or invalid encodings outside the signed portion of a
8078 certificate the fingerprint can be changed without breaking the signature.
8079 Although no details of the signed portion of the certificate can be changed
8080 this can cause problems with some applications: e.g. those using the
8081 certificate fingerprint for blacklists.
8083 1. Reject signatures with non zero unused bits.
8085 If the BIT STRING containing the signature has non zero unused bits reject
8086 the signature. All current signature algorithms require zero unused bits.
8088 2. Check certificate algorithm consistency.
8090 Check the AlgorithmIdentifier inside TBS matches the one in the
8091 certificate signature. NB: this will result in signature failure
8092 errors for some broken certificates.
8094 Thanks to Konrad Kraszewski from Google for reporting this issue.
8096 3. Check DSA/ECDSA signatures use DER.
8098 Re-encode DSA/ECDSA signatures and compare with the original received
8099 signature. Return an error if there is a mismatch.
8101 This will reject various cases including garbage after signature
8102 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
8103 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
8104 (negative or with leading zeroes).
8106 Further analysis was conducted and fixes were developed by Stephen Henson
8107 of the OpenSSL core team.
8113 * Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
8114 results on some platforms, including x86_64. This bug occurs at random
8115 with a very low probability, and is not known to be exploitable in any
8116 way, though its exact impact is difficult to determine. Thanks to Pieter
8117 Wuille (Blockstream) who reported this issue and also suggested an initial
8118 fix. Further analysis was conducted by the OpenSSL development team and
8119 Adam Langley of Google. The final fix was developed by Andy Polyakov of
8120 the OpenSSL core team.
8125 * Do not resume sessions on the server if the negotiated protocol
8126 version does not match the session's version. Resuming with a different
8127 version, while not strictly forbidden by the RFC, is of questionable
8128 sanity and breaks all known clients.
8130 *David Benjamin, Emilia Käsper*
8132 * Tighten handling of the ChangeCipherSpec (CCS) message: reject
8133 early CCS messages during renegotiation. (Note that because
8134 renegotiation is encrypted, this early CCS was not exploitable.)
8138 * Tighten client-side session ticket handling during renegotiation:
8139 ensure that the client only accepts a session ticket if the server sends
8140 the extension anew in the ServerHello. Previously, a TLS client would
8141 reuse the old extension state and thus accept a session ticket if one was
8142 announced in the initial ServerHello.
8144 Similarly, ensure that the client requires a session ticket if one
8145 was advertised in the ServerHello. Previously, a TLS client would
8146 ignore a missing NewSessionTicket message.
8150 ### Changes between 1.0.1i and 1.0.1j [15 Oct 2014]
8154 A flaw in the DTLS SRTP extension parsing code allows an attacker, who
8155 sends a carefully crafted handshake message, to cause OpenSSL to fail
8156 to free up to 64k of memory causing a memory leak. This could be
8157 exploited in a Denial Of Service attack. This issue affects OpenSSL
8158 1.0.1 server implementations for both SSL/TLS and DTLS regardless of
8159 whether SRTP is used or configured. Implementations of OpenSSL that
8160 have been compiled with OPENSSL_NO_SRTP defined are not affected.
8162 The fix was developed by the OpenSSL team.
8167 * Session Ticket Memory Leak.
8169 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
8170 integrity of that ticket is first verified. In the event of a session
8171 ticket integrity check failing, OpenSSL will fail to free memory
8172 causing a memory leak. By sending a large number of invalid session
8173 tickets an attacker could exploit this issue in a Denial Of Service
8179 * Build option no-ssl3 is incomplete.
8181 When OpenSSL is configured with "no-ssl3" as a build option, servers
8182 could accept and complete an SSL 3.0 handshake, and clients could be
8183 configured to send them.
8186 *Akamai and the OpenSSL team*
8188 * Add support for TLS_FALLBACK_SCSV.
8189 Client applications doing fallback retries should call
8190 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
8193 *Adam Langley, Bodo Moeller*
8195 * Add additional DigestInfo checks.
8197 Re-encode DigestInto in DER and check against the original when
8198 verifying RSA signature: this will reject any improperly encoded
8199 DigestInfo structures.
8201 Note: this is a precautionary measure and no attacks are currently known.
8205 ### Changes between 1.0.1h and 1.0.1i [6 Aug 2014]
8207 * Fix SRP buffer overrun vulnerability. Invalid parameters passed to the
8208 SRP code can be overrun an internal buffer. Add sanity check that
8209 g, A, B < N to SRP code.
8211 Thanks to Sean Devlin and Watson Ladd of Cryptography Services, NCC
8212 Group for discovering this issue.
8217 * A flaw in the OpenSSL SSL/TLS server code causes the server to negotiate
8218 TLS 1.0 instead of higher protocol versions when the ClientHello message
8219 is badly fragmented. This allows a man-in-the-middle attacker to force a
8220 downgrade to TLS 1.0 even if both the server and the client support a
8221 higher protocol version, by modifying the client's TLS records.
8223 Thanks to David Benjamin and Adam Langley (Google) for discovering and
8224 researching this issue.
8229 * OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
8230 to a denial of service attack. A malicious server can crash the client
8231 with a null pointer dereference (read) by specifying an anonymous (EC)DH
8232 ciphersuite and sending carefully crafted handshake messages.
8234 Thanks to Felix Gröbert (Google) for discovering and researching this
8240 * By sending carefully crafted DTLS packets an attacker could cause openssl
8241 to leak memory. This can be exploited through a Denial of Service attack.
8242 Thanks to Adam Langley for discovering and researching this issue.
8247 * An attacker can force openssl to consume large amounts of memory whilst
8248 processing DTLS handshake messages. This can be exploited through a
8249 Denial of Service attack.
8250 Thanks to Adam Langley for discovering and researching this issue.
8255 * An attacker can force an error condition which causes openssl to crash
8256 whilst processing DTLS packets due to memory being freed twice. This
8257 can be exploited through a Denial of Service attack.
8258 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
8264 * If a multithreaded client connects to a malicious server using a resumed
8265 session and the server sends an ec point format extension it could write
8266 up to 255 bytes to freed memory.
8268 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
8274 * A malicious server can crash an OpenSSL client with a null pointer
8275 dereference (read) by specifying an SRP ciphersuite even though it was not
8276 properly negotiated with the client. This can be exploited through a
8277 Denial of Service attack.
8279 Thanks to Joonas Kuorilehto and Riku Hietamäki (Codenomicon) for
8280 discovering and researching this issue.
8285 * A flaw in OBJ_obj2txt may cause pretty printing functions such as
8286 X509_name_oneline, X509_name_print_ex et al. to leak some information
8287 from the stack. Applications may be affected if they echo pretty printing
8288 output to the attacker.
8290 Thanks to Ivan Fratric (Google) for discovering this issue.
8293 *Emilia Käsper, and Steve Henson*
8295 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
8296 for corner cases. (Certain input points at infinity could lead to
8297 bogus results, with non-infinity inputs mapped to infinity too.)
8301 ### Changes between 1.0.1g and 1.0.1h [5 Jun 2014]
8303 * Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
8304 handshake can force the use of weak keying material in OpenSSL
8305 SSL/TLS clients and servers.
8307 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
8308 researching this issue. ([CVE-2014-0224])
8310 *KIKUCHI Masashi, Steve Henson*
8312 * Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
8313 OpenSSL DTLS client the code can be made to recurse eventually crashing
8316 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
8319 *Imre Rad, Steve Henson*
8321 * Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
8322 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
8323 client or server. This is potentially exploitable to run arbitrary
8324 code on a vulnerable client or server.
8326 Thanks to Jüri Aedla for reporting this issue. ([CVE-2014-0195])
8328 *Jüri Aedla, Steve Henson*
8330 * Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
8331 are subject to a denial of service attack.
8333 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
8334 this issue. ([CVE-2014-3470])
8336 *Felix Gröbert, Ivan Fratric, Steve Henson*
8338 * Harmonize version and its documentation. -f flag is used to display
8341 *mancha <mancha1@zoho.com>*
8343 * Fix eckey_priv_encode so it immediately returns an error upon a failure
8344 in i2d_ECPrivateKey.
8346 *mancha <mancha1@zoho.com>*
8348 * Fix some double frees. These are not thought to be exploitable.
8350 *mancha <mancha1@zoho.com>*
8352 ### Changes between 1.0.1f and 1.0.1g [7 Apr 2014]
8354 * A missing bounds check in the handling of the TLS heartbeat extension
8355 can be used to reveal up to 64k of memory to a connected client or
8358 Thanks for Neel Mehta of Google Security for discovering this bug and to
8359 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
8360 preparing the fix ([CVE-2014-0160])
8362 *Adam Langley, Bodo Moeller*
8364 * Fix for the attack described in the paper "Recovering OpenSSL
8365 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
8366 by Yuval Yarom and Naomi Benger. Details can be obtained from:
8367 <http://eprint.iacr.org/2014/140>
8369 Thanks to Yuval Yarom and Naomi Benger for discovering this
8370 flaw and to Yuval Yarom for supplying a fix ([CVE-2014-0076])
8372 *Yuval Yarom and Naomi Benger*
8374 * TLS pad extension: draft-agl-tls-padding-03
8376 Workaround for the "TLS hang bug" (see FAQ and PR#2771): if the
8377 TLS client Hello record length value would otherwise be > 255 and
8378 less that 512 pad with a dummy extension containing zeroes so it
8379 is at least 512 bytes long.
8381 *Adam Langley, Steve Henson*
8383 ### Changes between 1.0.1e and 1.0.1f [6 Jan 2014]
8385 * Fix for TLS record tampering bug. A carefully crafted invalid
8386 handshake could crash OpenSSL with a NULL pointer exception.
8387 Thanks to Anton Johansson for reporting this issues.
8390 * Keep original DTLS digest and encryption contexts in retransmission
8391 structures so we can use the previous session parameters if they need
8392 to be resent. ([CVE-2013-6450])
8396 * Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
8397 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
8398 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
8399 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
8400 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
8401 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
8403 *Rob Stradling, Adam Langley*
8405 ### Changes between 1.0.1d and 1.0.1e [11 Feb 2013]
8407 * Correct fix for CVE-2013-0169. The original didn't work on AES-NI
8408 supporting platforms or when small records were transferred.
8410 *Andy Polyakov, Steve Henson*
8412 ### Changes between 1.0.1c and 1.0.1d [5 Feb 2013]
8414 * Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
8416 This addresses the flaw in CBC record processing discovered by
8417 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
8418 at: <http://www.isg.rhul.ac.uk/tls/>
8420 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
8421 Security Group at Royal Holloway, University of London
8422 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
8423 Emilia Käsper for the initial patch.
8426 *Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson*
8428 * Fix flaw in AESNI handling of TLS 1.2 and 1.1 records for CBC mode
8429 ciphersuites which can be exploited in a denial of service attack.
8430 Thanks go to and to Adam Langley <agl@chromium.org> for discovering
8431 and detecting this bug and to Wolfgang Ettlinger
8432 <wolfgang.ettlinger@gmail.com> for independently discovering this issue.
8437 * Return an error when checking OCSP signatures when key is NULL.
8438 This fixes a DoS attack. ([CVE-2013-0166])
8442 * Make openssl verify return errors.
8444 *Chris Palmer <palmer@google.com> and Ben Laurie*
8446 * Call OCSP Stapling callback after ciphersuite has been chosen, so
8447 the right response is stapled. Also change SSL_get_certificate()
8448 so it returns the certificate actually sent.
8449 See <http://rt.openssl.org/Ticket/Display.html?id=2836>.
8451 *Rob Stradling <rob.stradling@comodo.com>*
8453 * Fix possible deadlock when decoding public keys.
8457 * Don't use TLS 1.0 record version number in initial client hello
8462 ### Changes between 1.0.1b and 1.0.1c [10 May 2012]
8464 * Sanity check record length before skipping explicit IV in TLS
8465 1.2, 1.1 and DTLS to fix DoS attack.
8467 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
8468 fuzzing as a service testing platform.
8473 * Initialise tkeylen properly when encrypting CMS messages.
8474 Thanks to Solar Designer of Openwall for reporting this issue.
8478 * In FIPS mode don't try to use composite ciphers as they are not
8483 ### Changes between 1.0.1a and 1.0.1b [26 Apr 2012]
8485 * OpenSSL 1.0.0 sets SSL_OP_ALL to 0x80000FFFL and OpenSSL 1.0.1 and
8486 1.0.1a set SSL_OP_NO_TLSv1_1 to 0x00000400L which would unfortunately
8487 mean any application compiled against OpenSSL 1.0.0 headers setting
8488 SSL_OP_ALL would also set SSL_OP_NO_TLSv1_1, unintentionally disabling
8489 TLS 1.1 also. Fix this by changing the value of SSL_OP_NO_TLSv1_1 to
8490 0x10000000L Any application which was previously compiled against
8491 OpenSSL 1.0.1 or 1.0.1a headers and which cares about SSL_OP_NO_TLSv1_1
8492 will need to be recompiled as a result. Letting be results in
8493 inability to disable specifically TLS 1.1 and in client context,
8494 in unlike event, limit maximum offered version to TLS 1.0 [see below].
8498 * In order to ensure interoperability SSL_OP_NO_protocolX does not
8499 disable just protocol X, but all protocols above X *if* there are
8500 protocols *below* X still enabled. In more practical terms it means
8501 that if application wants to disable TLS1.0 in favor of TLS1.1 and
8502 above, it's not sufficient to pass `SSL_OP_NO_TLSv1`, one has to pass
8503 `SSL_OP_NO_TLSv1|SSL_OP_NO_SSLv3|SSL_OP_NO_SSLv2`. This applies to
8508 ### Changes between 1.0.1 and 1.0.1a [19 Apr 2012]
8510 * Check for potentially exploitable overflows in asn1_d2i_read_bio
8511 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
8512 in CRYPTO_realloc_clean.
8514 Thanks to Tavis Ormandy, Google Security Team, for discovering this
8515 issue and to Adam Langley <agl@chromium.org> for fixing it.
8518 *Adam Langley (Google), Tavis Ormandy, Google Security Team*
8520 * Don't allow TLS 1.2 SHA-256 ciphersuites in TLS 1.0, 1.1 connections.
8524 * Workarounds for some broken servers that "hang" if a client hello
8525 record length exceeds 255 bytes.
8527 1. Do not use record version number > TLS 1.0 in initial client
8528 hello: some (but not all) hanging servers will now work.
8529 2. If we set OPENSSL_MAX_TLS1_2_CIPHER_LENGTH this will truncate
8530 the number of ciphers sent in the client hello. This should be
8531 set to an even number, such as 50, for example by passing:
8532 -DOPENSSL_MAX_TLS1_2_CIPHER_LENGTH=50 to config or Configure.
8533 Most broken servers should now work.
8534 3. If all else fails setting OPENSSL_NO_TLS1_2_CLIENT will disable
8535 TLS 1.2 client support entirely.
8539 * Fix SEGV in Vector Permutation AES module observed in OpenSSH.
8543 ### Changes between 1.0.0h and 1.0.1 [14 Mar 2012]
8545 * Add compatibility with old MDC2 signatures which use an ASN1 OCTET
8546 STRING form instead of a DigestInfo.
8550 * The format used for MDC2 RSA signatures is inconsistent between EVP
8551 and the RSA_sign/RSA_verify functions. This was made more apparent when
8552 OpenSSL used RSA_sign/RSA_verify for some RSA signatures in particular
8553 those which went through EVP_PKEY_METHOD in 1.0.0 and later. Detect
8554 the correct format in RSA_verify so both forms transparently work.
8558 * Some servers which support TLS 1.0 can choke if we initially indicate
8559 support for TLS 1.2 and later renegotiate using TLS 1.0 in the RSA
8560 encrypted premaster secret. As a workaround use the maximum permitted
8561 client version in client hello, this should keep such servers happy
8562 and still work with previous versions of OpenSSL.
8566 * Add support for TLS/DTLS heartbeats.
8568 *Robin Seggelmann <seggelmann@fh-muenster.de>*
8570 * Add support for SCTP.
8572 *Robin Seggelmann <seggelmann@fh-muenster.de>*
8574 * Improved PRNG seeding for VOS.
8576 *Paul Green <Paul.Green@stratus.com>*
8578 * Extensive assembler packs updates, most notably:
8580 - x86[_64]: AES-NI, PCLMULQDQ, RDRAND support;
8581 - x86[_64]: SSSE3 support (SHA1, vector-permutation AES);
8582 - x86_64: bit-sliced AES implementation;
8583 - ARM: NEON support, contemporary platforms optimizations;
8584 - s390x: z196 support;
8585 - `*`: GHASH and GF(2^m) multiplication implementations;
8589 * Make TLS-SRP code conformant with RFC 5054 API cleanup
8590 (removal of unnecessary code)
8592 *Peter Sylvester <peter.sylvester@edelweb.fr>*
8594 * Add TLS key material exporter from RFC 5705.
8598 * Add DTLS-SRTP negotiation from RFC 5764.
8602 * Add Next Protocol Negotiation,
8603 <http://tools.ietf.org/html/draft-agl-tls-nextprotoneg-00>. Can be
8604 disabled with a no-npn flag to config or Configure. Code donated
8607 *Adam Langley <agl@google.com> and Ben Laurie*
8609 * Add optional 64-bit optimized implementations of elliptic curves NIST-P224,
8610 NIST-P256, NIST-P521, with constant-time single point multiplication on
8611 typical inputs. Compiler support for the nonstandard type `__uint128_t` is
8612 required to use this (present in gcc 4.4 and later, for 64-bit builds).
8613 Code made available under Apache License version 2.0.
8615 Specify "enable-ec_nistp_64_gcc_128" on the Configure (or config) command
8616 line to include this in your build of OpenSSL, and run "make depend" (or
8617 "make update"). This enables the following EC_METHODs:
8619 EC_GFp_nistp224_method()
8620 EC_GFp_nistp256_method()
8621 EC_GFp_nistp521_method()
8623 EC_GROUP_new_by_curve_name() will automatically use these (while
8624 EC_GROUP_new_curve_GFp() currently prefers the more flexible
8627 *Emilia Käsper, Adam Langley, Bodo Moeller (Google)*
8629 * Use type ossl_ssize_t instead of ssize_t which isn't available on
8630 all platforms. Move ssize_t definition from e_os.h to the public
8631 header file e_os2.h as it now appears in public header file cms.h
8635 * New -sigopt option to the ca, req and x509 utilities. Additional
8636 signature parameters can be passed using this option and in
8641 * Add RSA PSS signing function. This will generate and set the
8642 appropriate AlgorithmIdentifiers for PSS based on those in the
8643 corresponding EVP_MD_CTX structure. No application support yet.
8647 * Support for companion algorithm specific ASN1 signing routines.
8648 New function ASN1_item_sign_ctx() signs a pre-initialised
8649 EVP_MD_CTX structure and sets AlgorithmIdentifiers based on
8650 the appropriate parameters.
8654 * Add new algorithm specific ASN1 verification initialisation function
8655 to EVP_PKEY_ASN1_METHOD: this is not in EVP_PKEY_METHOD since the ASN1
8656 handling will be the same no matter what EVP_PKEY_METHOD is used.
8657 Add a PSS handler to support verification of PSS signatures: checked
8658 against a number of sample certificates.
8662 * Add signature printing for PSS. Add PSS OIDs.
8664 *Steve Henson, Martin Kaiser <lists@kaiser.cx>*
8666 * Add algorithm specific signature printing. An individual ASN1 method
8667 can now print out signatures instead of the standard hex dump.
8669 More complex signatures (e.g. PSS) can print out more meaningful
8670 information. Include DSA version that prints out the signature
8675 * Password based recipient info support for CMS library: implementing
8680 * Split password based encryption into PBES2 and PBKDF2 functions. This
8681 neatly separates the code into cipher and PBE sections and is required
8682 for some algorithms that split PBES2 into separate pieces (such as
8683 password based CMS).
8687 * Session-handling fixes:
8688 - Fix handling of connections that are resuming with a session ID,
8689 but also support Session Tickets.
8690 - Fix a bug that suppressed issuing of a new ticket if the client
8691 presented a ticket with an expired session.
8692 - Try to set the ticket lifetime hint to something reasonable.
8693 - Make tickets shorter by excluding irrelevant information.
8694 - On the client side, don't ignore renewed tickets.
8696 *Adam Langley, Bodo Moeller (Google)*
8698 * Fix PSK session representation.
8702 * Add RC4-MD5 and AESNI-SHA1 "stitched" implementations.
8704 This work was sponsored by Intel.
8708 * Add GCM support to TLS library. Some custom code is needed to split
8709 the IV between the fixed (from PRF) and explicit (from TLS record)
8710 portions. This adds all GCM ciphersuites supported by RFC5288 and
8711 RFC5289. Generalise some `AES*` cipherstrings to include GCM and
8712 add a special AESGCM string for GCM only.
8716 * Expand range of ctrls for AES GCM. Permit setting invocation
8717 field on decrypt and retrieval of invocation field only on encrypt.
8721 * Add HMAC ECC ciphersuites from RFC5289. Include SHA384 PRF support.
8722 As required by RFC5289 these ciphersuites cannot be used if for
8723 versions of TLS earlier than 1.2.
8727 * For FIPS capable OpenSSL interpret a NULL default public key method
8728 as unset and return the appropriate default but do *not* set the default.
8729 This means we can return the appropriate method in applications that
8730 switch between FIPS and non-FIPS modes.
8734 * Redirect HMAC and CMAC operations to FIPS module in FIPS mode. If an
8735 ENGINE is used then we cannot handle that in the FIPS module so we
8736 keep original code iff non-FIPS operations are allowed.
8740 * Add -attime option to openssl utilities.
8742 *Peter Eckersley <pde@eff.org>, Ben Laurie and Steve Henson*
8744 * Redirect DSA and DH operations to FIPS module in FIPS mode.
8748 * Redirect ECDSA and ECDH operations to FIPS module in FIPS mode. Also use
8749 FIPS EC methods unconditionally for now.
8753 * New build option no-ec2m to disable characteristic 2 code.
8757 * Backport libcrypto audit of return value checking from 1.1.0-dev; not
8758 all cases can be covered as some introduce binary incompatibilities.
8762 * Redirect RSA operations to FIPS module including keygen,
8763 encrypt, decrypt, sign and verify. Block use of non FIPS RSA methods.
8767 * Add similar low-level API blocking to ciphers.
8771 * low-level digest APIs are not approved in FIPS mode: any attempt
8772 to use these will cause a fatal error. Applications that *really* want
8773 to use them can use the `private_*` version instead.
8777 * Redirect cipher operations to FIPS module for FIPS builds.
8781 * Redirect digest operations to FIPS module for FIPS builds.
8785 * Update build system to add "fips" flag which will link in fipscanister.o
8786 for static and shared library builds embedding a signature if needed.
8790 * Output TLS supported curves in preference order instead of numerical
8791 order. This is currently hardcoded for the highest order curves first.
8792 This should be configurable so applications can judge speed vs strength.
8796 * Add TLS v1.2 server support for client authentication.
8800 * Add support for FIPS mode in ssl library: disable SSLv3, non-FIPS ciphers
8805 * Functions FIPS_mode_set() and FIPS_mode() which call the underlying
8806 FIPS modules versions.
8810 * Add TLS v1.2 client side support for client authentication. Keep cache
8811 of handshake records longer as we don't know the hash algorithm to use
8812 until after the certificate request message is received.
8816 * Initial TLS v1.2 client support. Add a default signature algorithms
8817 extension including all the algorithms we support. Parse new signature
8818 format in client key exchange. Relax some ECC signing restrictions for
8819 TLS v1.2 as indicated in RFC5246.
8823 * Add server support for TLS v1.2 signature algorithms extension. Switch
8824 to new signature format when needed using client digest preference.
8825 All server ciphersuites should now work correctly in TLS v1.2. No client
8826 support yet and no support for client certificates.
8830 * Initial TLS v1.2 support. Add new SHA256 digest to ssl code, switch
8831 to SHA256 for PRF when using TLS v1.2 and later. Add new SHA256 based
8832 ciphersuites. At present only RSA key exchange ciphersuites work with
8833 TLS v1.2. Add new option for TLS v1.2 replacing the old and obsolete
8834 SSL_OP_PKCS1_CHECK flags with SSL_OP_NO_TLSv1_2. New TLSv1.2 methods
8835 and version checking.
8839 * New option OPENSSL_NO_SSL_INTERN. If an application can be compiled
8840 with this defined it will not be affected by any changes to ssl internal
8841 structures. Add several utility functions to allow openssl application
8842 to work with OPENSSL_NO_SSL_INTERN defined.
8846 * A long standing patch to add support for SRP from EdelWeb (Peter
8847 Sylvester and Christophe Renou) was integrated.
8848 *Christophe Renou <christophe.renou@edelweb.fr>, Peter Sylvester
8849 <peter.sylvester@edelweb.fr>, Tom Wu <tjw@cs.stanford.edu>, and
8852 * Add functions to copy EVP_PKEY_METHOD and retrieve flags and id.
8856 * Permit abbreviated handshakes when renegotiating using the function
8857 SSL_renegotiate_abbreviated().
8859 *Robin Seggelmann <seggelmann@fh-muenster.de>*
8861 * Add call to ENGINE_register_all_complete() to
8862 ENGINE_load_builtin_engines(), so some implementations get used
8863 automatically instead of needing explicit application support.
8867 * Add support for TLS key exporter as described in RFC5705.
8869 *Robin Seggelmann <seggelmann@fh-muenster.de>, Steve Henson*
8871 * Initial TLSv1.1 support. Since TLSv1.1 is very similar to TLS v1.0 only
8872 a few changes are required:
8874 Add SSL_OP_NO_TLSv1_1 flag.
8875 Add TLSv1_1 methods.
8876 Update version checking logic to handle version 1.1.
8877 Add explicit IV handling (ported from DTLS code).
8878 Add command line options to s_client/s_server.
8885 ### Changes between 1.0.0s and 1.0.0t [3 Dec 2015]
8887 * X509_ATTRIBUTE memory leak
8889 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
8890 memory. This structure is used by the PKCS#7 and CMS routines so any
8891 application which reads PKCS#7 or CMS data from untrusted sources is
8892 affected. SSL/TLS is not affected.
8894 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
8900 * Race condition handling PSK identify hint
8902 If PSK identity hints are received by a multi-threaded client then
8903 the values are wrongly updated in the parent SSL_CTX structure. This can
8904 result in a race condition potentially leading to a double free of the
8910 ### Changes between 1.0.0r and 1.0.0s [11 Jun 2015]
8912 * Malformed ECParameters causes infinite loop
8914 When processing an ECParameters structure OpenSSL enters an infinite loop
8915 if the curve specified is over a specially malformed binary polynomial
8918 This can be used to perform denial of service against any
8919 system which processes public keys, certificate requests or
8920 certificates. This includes TLS clients and TLS servers with
8921 client authentication enabled.
8923 This issue was reported to OpenSSL by Joseph Barr-Pixton.
8928 * Exploitable out-of-bounds read in X509_cmp_time
8930 X509_cmp_time does not properly check the length of the ASN1_TIME
8931 string and can read a few bytes out of bounds. In addition,
8932 X509_cmp_time accepts an arbitrary number of fractional seconds in the
8935 An attacker can use this to craft malformed certificates and CRLs of
8936 various sizes and potentially cause a segmentation fault, resulting in
8937 a DoS on applications that verify certificates or CRLs. TLS clients
8938 that verify CRLs are affected. TLS clients and servers with client
8939 authentication enabled may be affected if they use custom verification
8942 This issue was reported to OpenSSL by Robert Swiecki (Google), and
8943 independently by Hanno Böck.
8948 * PKCS7 crash with missing EnvelopedContent
8950 The PKCS#7 parsing code does not handle missing inner EncryptedContent
8951 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
8952 with missing content and trigger a NULL pointer dereference on parsing.
8954 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
8955 structures from untrusted sources are affected. OpenSSL clients and
8956 servers are not affected.
8958 This issue was reported to OpenSSL by Michal Zalewski (Google).
8963 * CMS verify infinite loop with unknown hash function
8965 When verifying a signedData message the CMS code can enter an infinite loop
8966 if presented with an unknown hash function OID. This can be used to perform
8967 denial of service against any system which verifies signedData messages using
8969 This issue was reported to OpenSSL by Johannes Bauer.
8974 * Race condition handling NewSessionTicket
8976 If a NewSessionTicket is received by a multi-threaded client when attempting to
8977 reuse a previous ticket then a race condition can occur potentially leading to
8978 a double free of the ticket data.
8983 ### Changes between 1.0.0q and 1.0.0r [19 Mar 2015]
8985 * Segmentation fault in ASN1_TYPE_cmp fix
8987 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
8988 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
8989 certificate signature algorithm consistency this can be used to crash any
8990 certificate verification operation and exploited in a DoS attack. Any
8991 application which performs certificate verification is vulnerable including
8992 OpenSSL clients and servers which enable client authentication.
8997 * ASN.1 structure reuse memory corruption fix
8999 Reusing a structure in ASN.1 parsing may allow an attacker to cause
9000 memory corruption via an invalid write. Such reuse is and has been
9001 strongly discouraged and is believed to be rare.
9003 Applications that parse structures containing CHOICE or ANY DEFINED BY
9004 components may be affected. Certificate parsing (d2i_X509 and related
9005 functions) are however not affected. OpenSSL clients and servers are
9011 * PKCS7 NULL pointer dereferences fix
9013 The PKCS#7 parsing code does not handle missing outer ContentInfo
9014 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
9015 missing content and trigger a NULL pointer dereference on parsing.
9017 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
9018 otherwise parse PKCS#7 structures from untrusted sources are
9019 affected. OpenSSL clients and servers are not affected.
9021 This issue was reported to OpenSSL by Michal Zalewski (Google).
9026 * DoS via reachable assert in SSLv2 servers fix
9028 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
9029 servers that both support SSLv2 and enable export cipher suites by sending
9030 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
9032 This issue was discovered by Sean Burford (Google) and Emilia Käsper
9033 (OpenSSL development team).
9038 * Use After Free following d2i_ECPrivatekey error fix
9040 A malformed EC private key file consumed via the d2i_ECPrivateKey function
9041 could cause a use after free condition. This, in turn, could cause a double
9042 free in several private key parsing functions (such as d2i_PrivateKey
9043 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
9044 for applications that receive EC private keys from untrusted
9045 sources. This scenario is considered rare.
9047 This issue was discovered by the BoringSSL project and fixed in their
9053 * X509_to_X509_REQ NULL pointer deref fix
9055 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
9056 the certificate key is invalid. This function is rarely used in practice.
9058 This issue was discovered by Brian Carpenter.
9063 * Removed the export ciphers from the DEFAULT ciphers
9067 ### Changes between 1.0.0p and 1.0.0q [15 Jan 2015]
9069 * Build fixes for the Windows and OpenVMS platforms
9071 *Matt Caswell and Richard Levitte*
9073 ### Changes between 1.0.0o and 1.0.0p [8 Jan 2015]
9075 * Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
9076 message can cause a segmentation fault in OpenSSL due to a NULL pointer
9077 dereference. This could lead to a Denial Of Service attack. Thanks to
9078 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
9083 * Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
9084 dtls1_buffer_record function under certain conditions. In particular this
9085 could occur if an attacker sent repeated DTLS records with the same
9086 sequence number but for the next epoch. The memory leak could be exploited
9087 by an attacker in a Denial of Service attack through memory exhaustion.
9088 Thanks to Chris Mueller for reporting this issue.
9093 * Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
9094 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
9095 method would be set to NULL which could later result in a NULL pointer
9096 dereference. Thanks to Frank Schmirler for reporting this issue.
9101 * Abort handshake if server key exchange message is omitted for ephemeral
9104 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
9105 reporting this issue.
9110 * Remove non-export ephemeral RSA code on client and server. This code
9111 violated the TLS standard by allowing the use of temporary RSA keys in
9112 non-export ciphersuites and could be used by a server to effectively
9113 downgrade the RSA key length used to a value smaller than the server
9114 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
9115 INRIA or reporting this issue.
9120 * Fixed issue where DH client certificates are accepted without verification.
9121 An OpenSSL server will accept a DH certificate for client authentication
9122 without the certificate verify message. This effectively allows a client to
9123 authenticate without the use of a private key. This only affects servers
9124 which trust a client certificate authority which issues certificates
9125 containing DH keys: these are extremely rare and hardly ever encountered.
9126 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
9132 * Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
9133 results on some platforms, including x86_64. This bug occurs at random
9134 with a very low probability, and is not known to be exploitable in any
9135 way, though its exact impact is difficult to determine. Thanks to Pieter
9136 Wuille (Blockstream) who reported this issue and also suggested an initial
9137 fix. Further analysis was conducted by the OpenSSL development team and
9138 Adam Langley of Google. The final fix was developed by Andy Polyakov of
9139 the OpenSSL core team.
9144 * Fix various certificate fingerprint issues.
9146 By using non-DER or invalid encodings outside the signed portion of a
9147 certificate the fingerprint can be changed without breaking the signature.
9148 Although no details of the signed portion of the certificate can be changed
9149 this can cause problems with some applications: e.g. those using the
9150 certificate fingerprint for blacklists.
9152 1. Reject signatures with non zero unused bits.
9154 If the BIT STRING containing the signature has non zero unused bits reject
9155 the signature. All current signature algorithms require zero unused bits.
9157 2. Check certificate algorithm consistency.
9159 Check the AlgorithmIdentifier inside TBS matches the one in the
9160 certificate signature. NB: this will result in signature failure
9161 errors for some broken certificates.
9163 Thanks to Konrad Kraszewski from Google for reporting this issue.
9165 3. Check DSA/ECDSA signatures use DER.
9167 Re-encode DSA/ECDSA signatures and compare with the original received
9168 signature. Return an error if there is a mismatch.
9170 This will reject various cases including garbage after signature
9171 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
9172 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
9173 (negative or with leading zeroes).
9175 Further analysis was conducted and fixes were developed by Stephen Henson
9176 of the OpenSSL core team.
9182 ### Changes between 1.0.0n and 1.0.0o [15 Oct 2014]
9184 * Session Ticket Memory Leak.
9186 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
9187 integrity of that ticket is first verified. In the event of a session
9188 ticket integrity check failing, OpenSSL will fail to free memory
9189 causing a memory leak. By sending a large number of invalid session
9190 tickets an attacker could exploit this issue in a Denial Of Service
9196 * Build option no-ssl3 is incomplete.
9198 When OpenSSL is configured with "no-ssl3" as a build option, servers
9199 could accept and complete an SSL 3.0 handshake, and clients could be
9200 configured to send them.
9203 *Akamai and the OpenSSL team*
9205 * Add support for TLS_FALLBACK_SCSV.
9206 Client applications doing fallback retries should call
9207 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
9210 *Adam Langley, Bodo Moeller*
9212 * Add additional DigestInfo checks.
9214 Re-encode DigestInto in DER and check against the original when
9215 verifying RSA signature: this will reject any improperly encoded
9216 DigestInfo structures.
9218 Note: this is a precautionary measure and no attacks are currently known.
9222 ### Changes between 1.0.0m and 1.0.0n [6 Aug 2014]
9224 * OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
9225 to a denial of service attack. A malicious server can crash the client
9226 with a null pointer dereference (read) by specifying an anonymous (EC)DH
9227 ciphersuite and sending carefully crafted handshake messages.
9229 Thanks to Felix Gröbert (Google) for discovering and researching this
9235 * By sending carefully crafted DTLS packets an attacker could cause openssl
9236 to leak memory. This can be exploited through a Denial of Service attack.
9237 Thanks to Adam Langley for discovering and researching this issue.
9242 * An attacker can force openssl to consume large amounts of memory whilst
9243 processing DTLS handshake messages. This can be exploited through a
9244 Denial of Service attack.
9245 Thanks to Adam Langley for discovering and researching this issue.
9250 * An attacker can force an error condition which causes openssl to crash
9251 whilst processing DTLS packets due to memory being freed twice. This
9252 can be exploited through a Denial of Service attack.
9253 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
9259 * If a multithreaded client connects to a malicious server using a resumed
9260 session and the server sends an ec point format extension it could write
9261 up to 255 bytes to freed memory.
9263 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
9269 * A flaw in OBJ_obj2txt may cause pretty printing functions such as
9270 X509_name_oneline, X509_name_print_ex et al. to leak some information
9271 from the stack. Applications may be affected if they echo pretty printing
9272 output to the attacker.
9274 Thanks to Ivan Fratric (Google) for discovering this issue.
9277 *Emilia Käsper, and Steve Henson*
9279 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
9280 for corner cases. (Certain input points at infinity could lead to
9281 bogus results, with non-infinity inputs mapped to infinity too.)
9285 ### Changes between 1.0.0l and 1.0.0m [5 Jun 2014]
9287 * Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
9288 handshake can force the use of weak keying material in OpenSSL
9289 SSL/TLS clients and servers.
9291 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
9292 researching this issue. ([CVE-2014-0224])
9294 *KIKUCHI Masashi, Steve Henson*
9296 * Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
9297 OpenSSL DTLS client the code can be made to recurse eventually crashing
9300 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
9303 *Imre Rad, Steve Henson*
9305 * Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
9306 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
9307 client or server. This is potentially exploitable to run arbitrary
9308 code on a vulnerable client or server.
9310 Thanks to Jüri Aedla for reporting this issue. ([CVE-2014-0195])
9312 *Jüri Aedla, Steve Henson*
9314 * Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
9315 are subject to a denial of service attack.
9317 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
9318 this issue. ([CVE-2014-3470])
9320 *Felix Gröbert, Ivan Fratric, Steve Henson*
9322 * Harmonize version and its documentation. -f flag is used to display
9325 *mancha <mancha1@zoho.com>*
9327 * Fix eckey_priv_encode so it immediately returns an error upon a failure
9328 in i2d_ECPrivateKey.
9330 *mancha <mancha1@zoho.com>*
9332 * Fix some double frees. These are not thought to be exploitable.
9334 *mancha <mancha1@zoho.com>*
9336 * Fix for the attack described in the paper "Recovering OpenSSL
9337 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
9338 by Yuval Yarom and Naomi Benger. Details can be obtained from:
9339 <http://eprint.iacr.org/2014/140>
9341 Thanks to Yuval Yarom and Naomi Benger for discovering this
9342 flaw and to Yuval Yarom for supplying a fix ([CVE-2014-0076])
9344 *Yuval Yarom and Naomi Benger*
9346 ### Changes between 1.0.0k and 1.0.0l [6 Jan 2014]
9348 * Keep original DTLS digest and encryption contexts in retransmission
9349 structures so we can use the previous session parameters if they need
9350 to be resent. ([CVE-2013-6450])
9354 * Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
9355 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
9356 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
9357 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
9358 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
9359 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
9361 *Rob Stradling, Adam Langley*
9363 ### Changes between 1.0.0j and 1.0.0k [5 Feb 2013]
9365 * Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
9367 This addresses the flaw in CBC record processing discovered by
9368 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
9369 at: <http://www.isg.rhul.ac.uk/tls/>
9371 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
9372 Security Group at Royal Holloway, University of London
9373 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
9374 Emilia Käsper for the initial patch.
9377 *Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson*
9379 * Return an error when checking OCSP signatures when key is NULL.
9380 This fixes a DoS attack. ([CVE-2013-0166])
9384 * Call OCSP Stapling callback after ciphersuite has been chosen, so
9385 the right response is stapled. Also change SSL_get_certificate()
9386 so it returns the certificate actually sent.
9387 See <http://rt.openssl.org/Ticket/Display.html?id=2836>.
9388 (This is a backport)
9390 *Rob Stradling <rob.stradling@comodo.com>*
9392 * Fix possible deadlock when decoding public keys.
9396 ### Changes between 1.0.0i and 1.0.0j [10 May 2012]
9398 [NB: OpenSSL 1.0.0i and later 1.0.0 patch levels were released after
9401 * Sanity check record length before skipping explicit IV in DTLS
9404 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
9405 fuzzing as a service testing platform.
9410 * Initialise tkeylen properly when encrypting CMS messages.
9411 Thanks to Solar Designer of Openwall for reporting this issue.
9415 ### Changes between 1.0.0h and 1.0.0i [19 Apr 2012]
9417 * Check for potentially exploitable overflows in asn1_d2i_read_bio
9418 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
9419 in CRYPTO_realloc_clean.
9421 Thanks to Tavis Ormandy, Google Security Team, for discovering this
9422 issue and to Adam Langley <agl@chromium.org> for fixing it.
9425 *Adam Langley (Google), Tavis Ormandy, Google Security Team*
9427 ### Changes between 1.0.0g and 1.0.0h [12 Mar 2012]
9429 * Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness
9430 in CMS and PKCS7 code. When RSA decryption fails use a random key for
9431 content decryption and always return the same error. Note: this attack
9432 needs on average 2^20 messages so it only affects automated senders. The
9433 old behaviour can be re-enabled in the CMS code by setting the
9434 CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where
9435 an MMA defence is not necessary.
9436 Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering
9437 this issue. ([CVE-2012-0884])
9441 * Fix CVE-2011-4619: make sure we really are receiving a
9442 client hello before rejecting multiple SGC restarts. Thanks to
9443 Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug.
9447 ### Changes between 1.0.0f and 1.0.0g [18 Jan 2012]
9449 * Fix for DTLS DoS issue introduced by fix for CVE-2011-4109.
9450 Thanks to Antonio Martin, Enterprise Secure Access Research and
9451 Development, Cisco Systems, Inc. for discovering this bug and
9452 preparing a fix. ([CVE-2012-0050])
9456 ### Changes between 1.0.0e and 1.0.0f [4 Jan 2012]
9458 * Nadhem Alfardan and Kenny Paterson have discovered an extension
9459 of the Vaudenay padding oracle attack on CBC mode encryption
9460 which enables an efficient plaintext recovery attack against
9461 the OpenSSL implementation of DTLS. Their attack exploits timing
9462 differences arising during decryption processing. A research
9463 paper describing this attack can be found at:
9464 <http://www.isg.rhul.ac.uk/~kp/dtls.pdf>
9465 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
9466 Security Group at Royal Holloway, University of London
9467 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
9468 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
9469 for preparing the fix. ([CVE-2011-4108])
9471 *Robin Seggelmann, Michael Tuexen*
9473 * Clear bytes used for block padding of SSL 3.0 records.
9476 *Adam Langley (Google)*
9478 * Only allow one SGC handshake restart for SSL/TLS. Thanks to George
9479 Kadianakis <desnacked@gmail.com> for discovering this issue and
9480 Adam Langley for preparing the fix. ([CVE-2011-4619])
9482 *Adam Langley (Google)*
9484 * Check parameters are not NULL in GOST ENGINE. ([CVE-2012-0027])
9486 *Andrey Kulikov <amdeich@gmail.com>*
9488 * Prevent malformed RFC3779 data triggering an assertion failure.
9489 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
9490 and Rob Austein <sra@hactrn.net> for fixing it. ([CVE-2011-4577])
9492 *Rob Austein <sra@hactrn.net>*
9494 * Improved PRNG seeding for VOS.
9496 *Paul Green <Paul.Green@stratus.com>*
9498 * Fix ssl_ciph.c set-up race.
9500 *Adam Langley (Google)*
9502 * Fix spurious failures in ecdsatest.c.
9504 *Emilia Käsper (Google)*
9506 * Fix the BIO_f_buffer() implementation (which was mixing different
9507 interpretations of the `..._len` fields).
9509 *Adam Langley (Google)*
9511 * Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
9512 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
9513 threads won't reuse the same blinding coefficients.
9515 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
9516 lock to call BN_BLINDING_invert_ex, and avoids one use of
9517 BN_BLINDING_update for each BN_BLINDING structure (previously,
9518 the last update always remained unused).
9520 *Emilia Käsper (Google)*
9522 * In ssl3_clear, preserve s3->init_extra along with s3->rbuf.
9524 *Bob Buckholz (Google)*
9526 ### Changes between 1.0.0d and 1.0.0e [6 Sep 2011]
9528 * Fix bug where CRLs with nextUpdate in the past are sometimes accepted
9529 by initialising X509_STORE_CTX properly. ([CVE-2011-3207])
9531 *Kaspar Brand <ossl@velox.ch>*
9533 * Fix SSL memory handling for (EC)DH ciphersuites, in particular
9534 for multi-threaded use of ECDH. ([CVE-2011-3210])
9536 *Adam Langley (Google)*
9538 * Fix x509_name_ex_d2i memory leak on bad inputs.
9542 * Remove hard coded ecdsaWithSHA1 signature tests in ssl code and check
9543 signature public key algorithm by using OID xref utilities instead.
9544 Before this you could only use some ECC ciphersuites with SHA1 only.
9548 * Add protection against ECDSA timing attacks as mentioned in the paper
9549 by Billy Bob Brumley and Nicola Tuveri, see:
9550 <http://eprint.iacr.org/2011/232.pdf>
9552 *Billy Bob Brumley and Nicola Tuveri*
9554 ### Changes between 1.0.0c and 1.0.0d [8 Feb 2011]
9556 * Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
9558 *Neel Mehta, Adam Langley, Bodo Moeller (Google)*
9560 * Fix bug in string printing code: if *any* escaping is enabled we must
9561 escape the escape character (backslash) or the resulting string is
9566 ### Changes between 1.0.0b and 1.0.0c [2 Dec 2010]
9568 * Disable code workaround for ancient and obsolete Netscape browsers
9569 and servers: an attacker can use it in a ciphersuite downgrade attack.
9570 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
9574 * Fixed J-PAKE implementation error, originally discovered by
9575 Sebastien Martini, further info and confirmation from Stefan
9576 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
9580 ### Changes between 1.0.0a and 1.0.0b [16 Nov 2010]
9582 * Fix extension code to avoid race conditions which can result in a buffer
9583 overrun vulnerability: resumed sessions must not be modified as they can
9584 be shared by multiple threads. CVE-2010-3864
9588 * Fix WIN32 build system to correctly link an ENGINE directory into
9593 ### Changes between 1.0.0 and 1.0.0a [01 Jun 2010]
9595 * Check return value of int_rsa_verify in pkey_rsa_verifyrecover
9598 *Steve Henson, Peter-Michael Hager <hager@dortmund.net>*
9600 ### Changes between 0.9.8n and 1.0.0 [29 Mar 2010]
9602 * Add "missing" function EVP_CIPHER_CTX_copy(). This copies a cipher
9603 context. The operation can be customised via the ctrl mechanism in
9604 case ENGINEs want to include additional functionality.
9608 * Tolerate yet another broken PKCS#8 key format: private key value negative.
9612 * Add new -subject_hash_old and -issuer_hash_old options to x509 utility to
9613 output hashes compatible with older versions of OpenSSL.
9615 *Willy Weisz <weisz@vcpc.univie.ac.at>*
9617 * Fix compression algorithm handling: if resuming a session use the
9618 compression algorithm of the resumed session instead of determining
9619 it from client hello again. Don't allow server to change algorithm.
9623 * Add load_crls() function to commands tidying load_certs() too. Add option
9624 to verify utility to allow additional CRLs to be included.
9628 * Update OCSP request code to permit adding custom headers to the request:
9629 some responders need this.
9633 * The function EVP_PKEY_sign() returns <=0 on error: check return code
9636 *Julia Lawall <julia@diku.dk>*
9638 * Update verify callback code in `apps/s_cb.c` and `apps/verify.c`, it
9639 needlessly dereferenced structures, used obsolete functions and
9640 didn't handle all updated verify codes correctly.
9644 * Disable MD2 in the default configuration.
9648 * In BIO_pop() and BIO_push() use the ctrl argument (which was NULL) to
9649 indicate the initial BIO being pushed or popped. This makes it possible
9650 to determine whether the BIO is the one explicitly called or as a result
9651 of the ctrl being passed down the chain. Fix BIO_pop() and SSL BIOs so
9652 it handles reference counts correctly and doesn't zero out the I/O bio
9653 when it is not being explicitly popped. WARNING: applications which
9654 included workarounds for the old buggy behaviour will need to be modified
9655 or they could free up already freed BIOs.
9659 * Extend the uni2asc/asc2uni => OPENSSL_uni2asc/OPENSSL_asc2uni
9660 renaming to all platforms (within the 0.9.8 branch, this was
9661 done conditionally on Netware platforms to avoid a name clash).
9663 *Guenter <lists@gknw.net>*
9665 * Add ECDHE and PSK support to DTLS.
9667 *Michael Tuexen <tuexen@fh-muenster.de>*
9669 * Add CHECKED_STACK_OF macro to safestack.h, otherwise safestack can't
9674 * Add "missing" function EVP_MD_flags() (without this the only way to
9675 retrieve a digest flags is by accessing the structure directly. Update
9676 `EVP_MD_do_all*()` and `EVP_CIPHER_do_all*()` to include the name a digest
9677 or cipher is registered as in the "from" argument. Print out all
9678 registered digests in the dgst usage message instead of manually
9679 attempting to work them out.
9683 * If no SSLv2 ciphers are used don't use an SSLv2 compatible client hello:
9684 this allows the use of compression and extensions. Change default cipher
9685 string to remove SSLv2 ciphersuites. This effectively avoids ancient SSLv2
9686 by default unless an application cipher string requests it.
9690 * Alter match criteria in PKCS12_parse(). It used to try to use local
9691 key ids to find matching certificates and keys but some PKCS#12 files
9692 don't follow the (somewhat unwritten) rules and this strategy fails.
9693 Now just gather all certificates together and the first private key
9694 then look for the first certificate that matches the key.
9698 * Support use of registered digest and cipher names for dgst and cipher
9699 commands instead of having to add each one as a special case. So now
9706 openssl dgst -sha256 foo
9708 and this works for ENGINE based algorithms too.
9712 * Update Gost ENGINE to support parameter files.
9714 *Victor B. Wagner <vitus@cryptocom.ru>*
9716 * Support GeneralizedTime in ca utility.
9718 *Oliver Martin <oliver@volatilevoid.net>, Steve Henson*
9720 * Enhance the hash format used for certificate directory links. The new
9721 form uses the canonical encoding (meaning equivalent names will work
9722 even if they aren't identical) and uses SHA1 instead of MD5. This form
9723 is incompatible with the older format and as a result c_rehash should
9724 be used to rebuild symbolic links.
9728 * Make PKCS#8 the default write format for private keys, replacing the
9729 traditional format. This form is standardised, more secure and doesn't
9730 include an implicit MD5 dependency.
9734 * Add a $gcc_devteam_warn option to Configure. The idea is that any code
9735 committed to OpenSSL should pass this lot as a minimum.
9739 * Add session ticket override functionality for use by EAP-FAST.
9741 *Jouni Malinen <j@w1.fi>*
9743 * Modify HMAC functions to return a value. Since these can be implemented
9744 in an ENGINE errors can occur.
9748 * Type-checked OBJ_bsearch_ex.
9752 * Type-checked OBJ_bsearch. Also some constification necessitated
9753 by type-checking. Still to come: TXT_DB, bsearch(?),
9754 OBJ_bsearch_ex, qsort, CRYPTO_EX_DATA, ASN1_VALUE, ASN1_STRING,
9759 * New function OPENSSL_gmtime_adj() to add a specific number of days and
9760 seconds to a tm structure directly, instead of going through OS
9761 specific date routines. This avoids any issues with OS routines such
9762 as the year 2038 bug. New `*_adj()` functions for ASN1 time structures
9763 and X509_time_adj_ex() to cover the extended range. The existing
9764 X509_time_adj() is still usable and will no longer have any date issues.
9768 * Delta CRL support. New use deltas option which will attempt to locate
9769 and search any appropriate delta CRLs available.
9771 This work was sponsored by Google.
9775 * Support for CRLs partitioned by reason code. Reorganise CRL processing
9776 code and add additional score elements. Validate alternate CRL paths
9777 as part of the CRL checking and indicate a new error "CRL path validation
9778 error" in this case. Applications wanting additional details can use
9779 the verify callback and check the new "parent" field. If this is not
9780 NULL CRL path validation is taking place. Existing applications won't
9781 see this because it requires extended CRL support which is off by
9784 This work was sponsored by Google.
9788 * Support for freshest CRL extension.
9790 This work was sponsored by Google.
9794 * Initial indirect CRL support. Currently only supported in the CRLs
9795 passed directly and not via lookup. Process certificate issuer
9796 CRL entry extension and lookup CRL entries by bother issuer name
9797 and serial number. Check and process CRL issuer entry in IDP extension.
9799 This work was sponsored by Google.
9803 * Add support for distinct certificate and CRL paths. The CRL issuer
9804 certificate is validated separately in this case. Only enabled if
9805 an extended CRL support flag is set: this flag will enable additional
9806 CRL functionality in future.
9808 This work was sponsored by Google.
9812 * Add support for policy mappings extension.
9814 This work was sponsored by Google.
9818 * Fixes to pathlength constraint, self issued certificate handling,
9819 policy processing to align with RFC3280 and PKITS tests.
9821 This work was sponsored by Google.
9825 * Support for name constraints certificate extension. DN, email, DNS
9826 and URI types are currently supported.
9828 This work was sponsored by Google.
9832 * To cater for systems that provide a pointer-based thread ID rather
9833 than numeric, deprecate the current numeric thread ID mechanism and
9834 replace it with a structure and associated callback type. This
9835 mechanism allows a numeric "hash" to be extracted from a thread ID in
9836 either case, and on platforms where pointers are larger than 'long',
9837 mixing is done to help ensure the numeric 'hash' is usable even if it
9838 can't be guaranteed unique. The default mechanism is to use "&errno"
9839 as a pointer-based thread ID to distinguish between threads.
9841 Applications that want to provide their own thread IDs should now use
9842 CRYPTO_THREADID_set_callback() to register a callback that will call
9843 either CRYPTO_THREADID_set_numeric() or CRYPTO_THREADID_set_pointer().
9845 Note that ERR_remove_state() is now deprecated, because it is tied
9846 to the assumption that thread IDs are numeric. ERR_remove_state(0)
9847 to free the current thread's error state should be replaced by
9848 ERR_remove_thread_state(NULL).
9850 (This new approach replaces the functions CRYPTO_set_idptr_callback(),
9851 CRYPTO_get_idptr_callback(), and CRYPTO_thread_idptr() that existed in
9852 OpenSSL 0.9.9-dev between June 2006 and August 2008. Also, if an
9853 application was previously providing a numeric thread callback that
9854 was inappropriate for distinguishing threads, then uniqueness might
9855 have been obtained with &errno that happened immediately in the
9856 intermediate development versions of OpenSSL; this is no longer the
9857 case, the numeric thread callback will now override the automatic use
9860 *Geoff Thorpe, with help from Bodo Moeller*
9862 * Initial support for different CRL issuing certificates. This covers a
9863 simple case where the self issued certificates in the chain exist and
9864 the real CRL issuer is higher in the existing chain.
9866 This work was sponsored by Google.
9870 * Removed effectively defunct crypto/store from the build.
9874 * Revamp of STACK to provide stronger type-checking. Still to come:
9875 TXT_DB, bsearch(?), OBJ_bsearch, qsort, CRYPTO_EX_DATA, ASN1_VALUE,
9876 ASN1_STRING, CONF_VALUE.
9880 * Add a new SSL_MODE_RELEASE_BUFFERS mode flag to release unused buffer
9881 RAM on SSL connections. This option can save about 34k per idle SSL.
9885 * Revamp of LHASH to provide stronger type-checking. Still to come:
9886 STACK, TXT_DB, bsearch, qsort.
9890 * Initial support for Cryptographic Message Syntax (aka CMS) based
9891 on RFC3850, RFC3851 and RFC3852. New cms directory and cms utility,
9892 support for data, signedData, compressedData, digestedData and
9893 encryptedData, envelopedData types included. Scripts to check against
9894 RFC4134 examples draft and interop and consistency checks of many
9895 content types and variants.
9899 * Add options to enc utility to support use of zlib compression BIO.
9903 * Extend mk1mf to support importing of options and assembly language
9904 files from Configure script, currently only included in VC-WIN32.
9905 The assembly language rules can now optionally generate the source
9906 files from the associated perl scripts.
9910 * Implement remaining functionality needed to support GOST ciphersuites.
9911 Interop testing has been performed using CryptoPro implementations.
9913 *Victor B. Wagner <vitus@cryptocom.ru>*
9915 * s390x assembler pack.
9919 * ARMv4 assembler pack. ARMv4 refers to v4 and later ISA, not CPU
9924 * Implement Opaque PRF Input TLS extension as specified in
9925 draft-rescorla-tls-opaque-prf-input-00.txt. Since this is not an
9926 official specification yet and no extension type assignment by
9927 IANA exists, this extension (for now) will have to be explicitly
9928 enabled when building OpenSSL by providing the extension number
9929 to use. For example, specify an option
9931 -DTLSEXT_TYPE_opaque_prf_input=0x9527
9933 to the "config" or "Configure" script to enable the extension,
9934 assuming extension number 0x9527 (which is a completely arbitrary
9935 and unofficial assignment based on the MD5 hash of the Internet
9936 Draft). Note that by doing so, you potentially lose
9937 interoperability with other TLS implementations since these might
9938 be using the same extension number for other purposes.
9940 SSL_set_tlsext_opaque_prf_input(ssl, src, len) is used to set the
9941 opaque PRF input value to use in the handshake. This will create
9942 an internal copy of the length-'len' string at 'src', and will
9943 return non-zero for success.
9945 To get more control and flexibility, provide a callback function
9948 SSL_CTX_set_tlsext_opaque_prf_input_callback(ctx, cb)
9949 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg(ctx, arg)
9953 int (*cb)(SSL *, void *peerinput, size_t len, void *arg);
9956 Callback function 'cb' will be called in handshakes, and is
9957 expected to use SSL_set_tlsext_opaque_prf_input() as appropriate.
9958 Argument 'arg' is for application purposes (the value as given to
9959 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg() will directly
9960 be provided to the callback function). The callback function
9961 has to return non-zero to report success: usually 1 to use opaque
9962 PRF input just if possible, or 2 to enforce use of the opaque PRF
9963 input. In the latter case, the library will abort the handshake
9964 if opaque PRF input is not successfully negotiated.
9966 Arguments 'peerinput' and 'len' given to the callback function
9967 will always be NULL and 0 in the case of a client. A server will
9968 see the client's opaque PRF input through these variables if
9969 available (NULL and 0 otherwise). Note that if the server
9970 provides an opaque PRF input, the length must be the same as the
9971 length of the client's opaque PRF input.
9973 Note that the callback function will only be called when creating
9974 a new session (session resumption can resume whatever was
9975 previously negotiated), and will not be called in SSL 2.0
9976 handshakes; thus, SSL_CTX_set_options(ctx, SSL_OP_NO_SSLv2) or
9977 SSL_set_options(ssl, SSL_OP_NO_SSLv2) is especially recommended
9978 for applications that need to enforce opaque PRF input.
9982 * Update ssl code to support digests other than SHA1+MD5 for handshake
9985 *Victor B. Wagner <vitus@cryptocom.ru>*
9987 * Add RFC4507 support to OpenSSL. This includes the corrections in
9988 RFC4507bis. The encrypted ticket format is an encrypted encoded
9989 SSL_SESSION structure, that way new session features are automatically
9992 If a client application caches session in an SSL_SESSION structure
9993 support is transparent because tickets are now stored in the encoded
9996 The SSL_CTX structure automatically generates keys for ticket
9997 protection in servers so again support should be possible
9998 with no application modification.
10000 If a client or server wishes to disable RFC4507 support then the option
10001 SSL_OP_NO_TICKET can be set.
10003 Add a TLS extension debugging callback to allow the contents of any client
10004 or server extensions to be examined.
10006 This work was sponsored by Google.
10010 * Final changes to avoid use of pointer pointer casts in OpenSSL.
10011 OpenSSL should now compile cleanly on gcc 4.2
10013 *Peter Hartley <pdh@utter.chaos.org.uk>, Steve Henson*
10015 * Update SSL library to use new EVP_PKEY MAC API. Include generic MAC
10016 support including streaming MAC support: this is required for GOST
10017 ciphersuite support.
10019 *Victor B. Wagner <vitus@cryptocom.ru>, Steve Henson*
10021 * Add option -stream to use PKCS#7 streaming in smime utility. New
10022 function i2d_PKCS7_bio_stream() and PEM_write_PKCS7_bio_stream()
10023 to output in BER and PEM format.
10027 * Experimental support for use of HMAC via EVP_PKEY interface. This
10028 allows HMAC to be handled via the `EVP_DigestSign*()` interface. The
10029 EVP_PKEY "key" in this case is the HMAC key, potentially allowing
10030 ENGINE support for HMAC keys which are unextractable. New -mac and
10031 -macopt options to dgst utility.
10035 * New option -sigopt to dgst utility. Update dgst to use
10036 `EVP_Digest{Sign,Verify}*`. These two changes make it possible to use
10037 alternative signing parameters such as X9.31 or PSS in the dgst
10042 * Change ssl_cipher_apply_rule(), the internal function that does
10043 the work each time a ciphersuite string requests enabling
10044 ("foo+bar"), moving ("+foo+bar"), disabling ("-foo+bar", or
10045 removing ("!foo+bar") a class of ciphersuites: Now it maintains
10046 the order of disabled ciphersuites such that those ciphersuites
10047 that most recently went from enabled to disabled not only stay
10048 in order with respect to each other, but also have higher priority
10049 than other disabled ciphersuites the next time ciphersuites are
10052 This means that you can now say, e.g., "PSK:-PSK:HIGH" to enable
10053 the same ciphersuites as with "HIGH" alone, but in a specific
10054 order where the PSK ciphersuites come first (since they are the
10055 most recently disabled ciphersuites when "HIGH" is parsed).
10057 Also, change ssl_create_cipher_list() (using this new
10058 functionality) such that between otherwise identical
10059 ciphersuites, ephemeral ECDH is preferred over ephemeral DH in
10064 * Change ssl_create_cipher_list() so that it automatically
10065 arranges the ciphersuites in reasonable order before starting
10066 to process the rule string. Thus, the definition for "DEFAULT"
10067 (SSL_DEFAULT_CIPHER_LIST) now is just "ALL:!aNULL:!eNULL", but
10068 remains equivalent to `"AES:ALL:!aNULL:!eNULL:+aECDH:+kRSA:+RC4:@STRENGTH"`.
10069 This makes it much easier to arrive at a reasonable default order
10070 in applications for which anonymous ciphers are OK (meaning
10071 that you can't actually use DEFAULT).
10073 *Bodo Moeller; suggested by Victor Duchovni*
10075 * Split the SSL/TLS algorithm mask (as used for ciphersuite string
10076 processing) into multiple integers instead of setting
10077 "SSL_MKEY_MASK" bits, "SSL_AUTH_MASK" bits, "SSL_ENC_MASK",
10078 "SSL_MAC_MASK", and "SSL_SSL_MASK" bits all in a single integer.
10079 (These masks as well as the individual bit definitions are hidden
10080 away into the non-exported interface ssl/ssl_locl.h, so this
10081 change to the definition of the SSL_CIPHER structure shouldn't
10082 affect applications.) This give us more bits for each of these
10083 categories, so there is no longer a need to coagulate AES128 and
10084 AES256 into a single algorithm bit, and to coagulate Camellia128
10085 and Camellia256 into a single algorithm bit, which has led to all
10088 Thus, among other things, the kludge introduced in 0.9.7m and
10089 0.9.8e for masking out AES256 independently of AES128 or masking
10090 out Camellia256 independently of AES256 is not needed here in 0.9.9.
10092 With the change, we also introduce new ciphersuite aliases that
10093 so far were missing: "AES128", "AES256", "CAMELLIA128", and
10098 * Add support for dsa-with-SHA224 and dsa-with-SHA256.
10099 Use the leftmost N bytes of the signature input if the input is
10100 larger than the prime q (with N being the size in bytes of q).
10104 * Very *very* experimental PKCS#7 streaming encoder support. Nothing uses
10105 it yet and it is largely untested.
10109 * Add support for the ecdsa-with-SHA224/256/384/512 signature types.
10113 * Initial incomplete changes to avoid need for function casts in OpenSSL
10114 some compilers (gcc 4.2 and later) reject their use. Safestack is
10115 reimplemented. Update ASN1 to avoid use of legacy functions.
10119 * Win32/64 targets are linked with Winsock2.
10123 * Add an X509_CRL_METHOD structure to allow CRL processing to be redirected
10124 to external functions. This can be used to increase CRL handling
10125 efficiency especially when CRLs are very large by (for example) storing
10126 the CRL revoked certificates in a database.
10130 * Overhaul of by_dir code. Add support for dynamic loading of CRLs so
10131 new CRLs added to a directory can be used. New command line option
10132 -verify_return_error to s_client and s_server. This causes real errors
10133 to be returned by the verify callback instead of carrying on no matter
10134 what. This reflects the way a "real world" verify callback would behave.
10138 * GOST engine, supporting several GOST algorithms and public key formats.
10139 Kindly donated by Cryptocom.
10143 * Partial support for Issuing Distribution Point CRL extension. CRLs
10144 partitioned by DP are handled but no indirect CRL or reason partitioning
10145 (yet). Complete overhaul of CRL handling: now the most suitable CRL is
10146 selected via a scoring technique which handles IDP and AKID in CRLs.
10150 * New X509_STORE_CTX callbacks lookup_crls() and lookup_certs() which
10151 will ultimately be used for all verify operations: this will remove the
10152 X509_STORE dependency on certificate verification and allow alternative
10153 lookup methods. X509_STORE based implementations of these two callbacks.
10157 * Allow multiple CRLs to exist in an X509_STORE with matching issuer names.
10158 Modify get_crl() to find a valid (unexpired) CRL if possible.
10162 * New function X509_CRL_match() to check if two CRLs are identical. Normally
10163 this would be called X509_CRL_cmp() but that name is already used by
10164 a function that just compares CRL issuer names. Cache several CRL
10165 extensions in X509_CRL structure and cache CRLDP in X509.
10169 * Store a "canonical" representation of X509_NAME structure (ASN1 Name)
10170 this maps equivalent X509_NAME structures into a consistent structure.
10171 Name comparison can then be performed rapidly using memcmp().
10175 * Non-blocking OCSP request processing. Add -timeout option to ocsp
10180 * Allow digests to supply their own micalg string for S/MIME type using
10181 the ctrl EVP_MD_CTRL_MICALG.
10185 * During PKCS7 signing pass the PKCS7 SignerInfo structure to the
10186 EVP_PKEY_METHOD before and after signing via the EVP_PKEY_CTRL_PKCS7_SIGN
10187 ctrl. It can then customise the structure before and/or after signing
10192 * New function OBJ_add_sigid() to allow application defined signature OIDs
10193 to be added to OpenSSLs internal tables. New function OBJ_sigid_free()
10194 to free up any added signature OIDs.
10198 * New functions EVP_CIPHER_do_all(), EVP_CIPHER_do_all_sorted(),
10199 EVP_MD_do_all() and EVP_MD_do_all_sorted() to enumerate internal
10200 digest and cipher tables. New options added to openssl utility:
10201 list-message-digest-algorithms and list-cipher-algorithms.
10205 * Change the array representation of binary polynomials: the list
10206 of degrees of non-zero coefficients is now terminated with -1.
10207 Previously it was terminated with 0, which was also part of the
10208 value; thus, the array representation was not applicable to
10209 polynomials where t^0 has coefficient zero. This change makes
10210 the array representation useful in a more general context.
10214 * Various modifications and fixes to SSL/TLS cipher string
10215 handling. For ECC, the code now distinguishes between fixed ECDH
10216 with RSA certificates on the one hand and with ECDSA certificates
10217 on the other hand, since these are separate ciphersuites. The
10218 unused code for Fortezza ciphersuites has been removed.
10220 For consistency with EDH, ephemeral ECDH is now called "EECDH"
10221 (not "ECDHE"). For consistency with the code for DH
10222 certificates, use of ECDH certificates is now considered ECDH
10223 authentication, not RSA or ECDSA authentication (the latter is
10224 merely the CA's signing algorithm and not actively used in the
10227 The temporary ciphersuite alias "ECCdraft" is no longer
10228 available, and ECC ciphersuites are no longer excluded from "ALL"
10229 and "DEFAULT". The following aliases now exist for RFC 4492
10230 ciphersuites, most of these by analogy with the DH case:
10232 kECDHr - ECDH cert, signed with RSA
10233 kECDHe - ECDH cert, signed with ECDSA
10234 kECDH - ECDH cert (signed with either RSA or ECDSA)
10235 kEECDH - ephemeral ECDH
10236 ECDH - ECDH cert or ephemeral ECDH
10239 aECDSA - ECDSA cert
10242 AECDH - anonymous ECDH
10243 EECDH - non-anonymous ephemeral ECDH (equivalent to "kEECDH:-AECDH")
10247 * Add additional S/MIME capabilities for AES and GOST ciphers if supported.
10248 Use correct micalg parameters depending on digest(s) in signed message.
10252 * Add engine support for EVP_PKEY_ASN1_METHOD. Add functions to process
10253 an ENGINE asn1 method. Support ENGINE lookups in the ASN1 code.
10257 * Initial engine support for EVP_PKEY_METHOD. New functions to permit
10258 an engine to register a method. Add ENGINE lookups for methods and
10259 functional reference processing.
10263 * New functions `EVP_Digest{Sign,Verify)*`. These are enhanced versions of
10264 `EVP_{Sign,Verify}*` which allow an application to customise the signature
10269 * New -resign option to smime utility. This adds one or more signers
10270 to an existing PKCS#7 signedData structure. Also -md option to use an
10271 alternative message digest algorithm for signing.
10275 * Tidy up PKCS#7 routines and add new functions to make it easier to
10276 create PKCS7 structures containing multiple signers. Update smime
10277 application to support multiple signers.
10281 * New -macalg option to pkcs12 utility to allow setting of an alternative
10286 * Initial support for PKCS#5 v2.0 PRFs other than default SHA1 HMAC.
10287 Reorganize PBE internals to lookup from a static table using NIDs,
10288 add support for HMAC PBE OID translation. Add a EVP_CIPHER ctrl:
10289 EVP_CTRL_PBE_PRF_NID this allows a cipher to specify an alternative
10290 PRF which will be automatically used with PBES2.
10294 * Replace the algorithm specific calls to generate keys in "req" with the
10299 * Update PKCS#7 enveloped data routines to use new API. This is now
10300 supported by any public key method supporting the encrypt operation. A
10301 ctrl is added to allow the public key algorithm to examine or modify
10302 the PKCS#7 RecipientInfo structure if it needs to: for RSA this is
10307 * Add a ctrl to asn1 method to allow a public key algorithm to express
10308 a default digest type to use. In most cases this will be SHA1 but some
10309 algorithms (such as GOST) need to specify an alternative digest. The
10310 return value indicates how strong the preference is 1 means optional and
10311 2 is mandatory (that is it is the only supported type). Modify
10312 ASN1_item_sign() to accept a NULL digest argument to indicate it should
10313 use the default md. Update openssl utilities to use the default digest
10314 type for signing if it is not explicitly indicated.
10318 * Use OID cross reference table in ASN1_sign() and ASN1_verify(). New
10319 EVP_MD flag EVP_MD_FLAG_PKEY_METHOD_SIGNATURE. This uses the relevant
10320 signing method from the key type. This effectively removes the link
10321 between digests and public key types.
10325 * Add an OID cross reference table and utility functions. Its purpose is to
10326 translate between signature OIDs such as SHA1WithrsaEncryption and SHA1,
10327 rsaEncryption. This will allow some of the algorithm specific hackery
10328 needed to use the correct OID to be removed.
10332 * Remove algorithm specific dependencies when setting PKCS7_SIGNER_INFO
10333 structures for PKCS7_sign(). They are now set up by the relevant public
10338 * Add provisional EC pkey method with support for ECDSA and ECDH.
10342 * Add support for key derivation (agreement) in the API, DH method and
10347 * Add DSA pkey method and DH pkey methods, extend DH ASN1 method to support
10348 public and private key formats. As a side effect these add additional
10349 command line functionality not previously available: DSA signatures can be
10350 generated and verified using pkeyutl and DH key support and generation in
10357 *Oliver Tappe <zooey@hirschkaefer.de>*
10359 * New make target "install_html_docs" installs HTML renditions of the
10362 *Oliver Tappe <zooey@hirschkaefer.de>*
10364 * New utility "genpkey" this is analogous to "genrsa" etc except it can
10365 generate keys for any algorithm. Extend and update EVP_PKEY_METHOD to
10366 support key and parameter generation and add initial key generation
10367 functionality for RSA.
10371 * Add functions for main EVP_PKEY_method operations. The undocumented
10372 functions `EVP_PKEY_{encrypt,decrypt}` have been renamed to
10373 `EVP_PKEY_{encrypt,decrypt}_old`.
10377 * Initial definitions for EVP_PKEY_METHOD. This will be a high level public
10378 key API, doesn't do much yet.
10382 * New function EVP_PKEY_asn1_get0_info() to retrieve information about
10383 public key algorithms. New option to openssl utility:
10384 "list-public-key-algorithms" to print out info.
10388 * Implement the Supported Elliptic Curves Extension for
10389 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
10393 * Don't free up OIDs in OBJ_cleanup() if they are in use by EVP_MD or
10394 EVP_CIPHER structures to avoid later problems in EVP_cleanup().
10398 * New utilities pkey and pkeyparam. These are similar to algorithm specific
10399 utilities such as rsa, dsa, dsaparam etc except they process any key
10404 * Transfer public key printing routines to EVP_PKEY_ASN1_METHOD. New
10405 functions EVP_PKEY_print_public(), EVP_PKEY_print_private(),
10406 EVP_PKEY_print_param() to print public key data from an EVP_PKEY
10411 * Initial support for pluggable public key ASN1.
10412 De-spaghettify the public key ASN1 handling. Move public and private
10413 key ASN1 handling to a new EVP_PKEY_ASN1_METHOD structure. Relocate
10414 algorithm specific handling to a single module within the relevant
10415 algorithm directory. Add functions to allow (near) opaque processing
10416 of public and private key structures.
10420 * Implement the Supported Point Formats Extension for
10421 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
10425 * Add initial support for RFC 4279 PSK TLS ciphersuites. Add members
10426 for the psk identity [hint] and the psk callback functions to the
10427 SSL_SESSION, SSL and SSL_CTX structure.
10430 PSK-RC4-SHA, PSK-3DES-EDE-CBC-SHA, PSK-AES128-CBC-SHA,
10434 SSL_CTX_use_psk_identity_hint
10435 SSL_get_psk_identity_hint
10436 SSL_get_psk_identity
10437 SSL_use_psk_identity_hint
10439 *Mika Kousa and Pasi Eronen of Nokia Corporation*
10441 * Add RFC 3161 compliant time stamp request creation, response generation
10442 and response verification functionality.
10444 *Zoltán Glózik <zglozik@opentsa.org>, The OpenTSA Project*
10446 * Add initial support for TLS extensions, specifically for the server_name
10447 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
10448 have new members for a hostname. The SSL data structure has an
10449 additional member `SSL_CTX *initial_ctx` so that new sessions can be
10450 stored in that context to allow for session resumption, even after the
10451 SSL has been switched to a new SSL_CTX in reaction to a client's
10452 server_name extension.
10454 New functions (subject to change):
10456 SSL_get_servername()
10457 SSL_get_servername_type()
10460 New CTRL codes and macros (subject to change):
10462 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
10463 - SSL_CTX_set_tlsext_servername_callback()
10464 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
10465 - SSL_CTX_set_tlsext_servername_arg()
10466 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
10468 openssl s_client has a new '-servername ...' option.
10470 openssl s_server has new options '-servername_host ...', '-cert2 ...',
10471 '-key2 ...', '-servername_fatal' (subject to change). This allows
10472 testing the HostName extension for a specific single hostname ('-cert'
10473 and '-key' remain fallbacks for handshakes without HostName
10474 negotiation). If the unrecognized_name alert has to be sent, this by
10475 default is a warning; it becomes fatal with the '-servername_fatal'
10478 *Peter Sylvester, Remy Allais, Christophe Renou*
10480 * Whirlpool hash implementation is added.
10484 * BIGNUM code on 64-bit SPARCv9 targets is switched from bn(64,64) to
10485 bn(64,32). Because of instruction set limitations it doesn't have
10486 any negative impact on performance. This was done mostly in order
10487 to make it possible to share assembler modules, such as bn_mul_mont
10488 implementations, between 32- and 64-bit builds without hassle.
10492 * Move code previously exiled into file crypto/ec/ec2_smpt.c
10493 to ec2_smpl.c, and no longer require the OPENSSL_EC_BIN_PT_COMP
10498 * New candidate for BIGNUM assembler implementation, bn_mul_mont,
10499 dedicated Montgomery multiplication procedure, is introduced.
10500 BN_MONT_CTX is modified to allow bn_mul_mont to reach for higher
10501 "64-bit" performance on certain 32-bit targets.
10505 * New option SSL_OP_NO_COMP to disable use of compression selectively
10506 in SSL structures. New SSL ctrl to set maximum send fragment size.
10507 Save memory by setting the I/O buffer sizes dynamically instead of
10508 using the maximum available value.
10512 * New option -V for 'openssl ciphers'. This prints the ciphersuite code
10513 in addition to the text details.
10517 * Very, very preliminary EXPERIMENTAL support for printing of general
10518 ASN1 structures. This currently produces rather ugly output and doesn't
10519 handle several customised structures at all.
10523 * Integrated support for PVK file format and some related formats such
10524 as MS PUBLICKEYBLOB and PRIVATEKEYBLOB. Command line switches to support
10525 these in the 'rsa' and 'dsa' utilities.
10529 * Support for PKCS#1 RSAPublicKey format on rsa utility command line.
10533 * Remove the ancient ASN1_METHOD code. This was only ever used in one
10534 place for the (very old) "NETSCAPE" format certificates which are now
10535 handled using new ASN1 code equivalents.
10539 * Let the TLSv1_method() etc. functions return a 'const' SSL_METHOD
10540 pointer and make the SSL_METHOD parameter in SSL_CTX_new,
10541 SSL_CTX_set_ssl_version and SSL_set_ssl_method 'const'.
10545 * Modify CRL distribution points extension code to print out previously
10546 unsupported fields. Enhance extension setting code to allow setting of
10551 * Add print and set support for Issuing Distribution Point CRL extension.
10555 * Change 'Configure' script to enable Camellia by default.
10562 ### Changes between 0.9.8m and 0.9.8n [24 Mar 2010]
10564 * When rejecting SSL/TLS records due to an incorrect version number, never
10565 update s->server with a new major version number. As of
10566 - OpenSSL 0.9.8m if 'short' is a 16-bit type,
10567 - OpenSSL 0.9.8f if 'short' is longer than 16 bits,
10568 the previous behavior could result in a read attempt at NULL when
10569 receiving specific incorrect SSL/TLS records once record payload
10570 protection is active. ([CVE-2010-0740])
10572 *Bodo Moeller, Adam Langley <agl@chromium.org>*
10574 * Fix for CVE-2010-0433 where some kerberos enabled versions of OpenSSL
10575 could be crashed if the relevant tables were not present (e.g. chrooted).
10577 *Tomas Hoger <thoger@redhat.com>*
10579 ### Changes between 0.9.8l and 0.9.8m [25 Feb 2010]
10581 * Always check bn_wexpand() return values for failure. ([CVE-2009-3245])
10583 *Martin Olsson, Neel Mehta*
10585 * Fix X509_STORE locking: Every 'objs' access requires a lock (to
10586 accommodate for stack sorting, always a write lock!).
10590 * On some versions of WIN32 Heap32Next is very slow. This can cause
10591 excessive delays in the RAND_poll(): over a minute. As a workaround
10592 include a time check in the inner Heap32Next loop too.
10596 * The code that handled flushing of data in SSL/TLS originally used the
10597 BIO_CTRL_INFO ctrl to see if any data was pending first. This caused
10598 the problem outlined in PR#1949. The fix suggested there however can
10599 trigger problems with buggy BIO_CTRL_WPENDING (e.g. some versions
10600 of Apache). So instead simplify the code to flush unconditionally.
10601 This should be fine since flushing with no data to flush is a no op.
10605 * Handle TLS versions 2.0 and later properly and correctly use the
10606 highest version of TLS/SSL supported. Although TLS >= 2.0 is some way
10607 off ancient servers have a habit of sticking around for a while...
10611 * Modify compression code so it frees up structures without using the
10612 ex_data callbacks. This works around a problem where some applications
10613 call CRYPTO_cleanup_all_ex_data() before application exit (e.g. when
10614 restarting) then use compression (e.g. SSL with compression) later.
10615 This results in significant per-connection memory leaks and
10616 has caused some security issues including CVE-2008-1678 and
10621 * Constify crypto/cast (i.e., <openssl/cast.h>): a CAST_KEY doesn't
10622 change when encrypting or decrypting.
10626 * Add option SSL_OP_LEGACY_SERVER_CONNECT which will allow clients to
10627 connect and renegotiate with servers which do not support RI.
10628 Until RI is more widely deployed this option is enabled by default.
10632 * Add "missing" ssl ctrls to clear options and mode.
10636 * If client attempts to renegotiate and doesn't support RI respond with
10637 a no_renegotiation alert as required by RFC5746. Some renegotiating
10638 TLS clients will continue a connection gracefully when they receive
10639 the alert. Unfortunately OpenSSL mishandled this alert and would hang
10640 waiting for a server hello which it will never receive. Now we treat a
10641 received no_renegotiation alert as a fatal error. This is because
10642 applications requesting a renegotiation might well expect it to succeed
10643 and would have no code in place to handle the server denying it so the
10644 only safe thing to do is to terminate the connection.
10648 * Add ctrl macro SSL_get_secure_renegotiation_support() which returns 1 if
10649 peer supports secure renegotiation and 0 otherwise. Print out peer
10650 renegotiation support in s_client/s_server.
10654 * Replace the highly broken and deprecated SPKAC certification method with
10655 the updated NID creation version. This should correctly handle UTF8.
10659 * Implement RFC5746. Re-enable renegotiation but require the extension
10660 as needed. Unfortunately, SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION
10661 turns out to be a bad idea. It has been replaced by
10662 SSL_OP_ALLOW_UNSAFE_LEGACY_RENEGOTIATION which can be set with
10663 SSL_CTX_set_options(). This is really not recommended unless you
10664 know what you are doing.
10666 *Eric Rescorla <ekr@networkresonance.com>, Ben Laurie, Steve Henson*
10668 * Fixes to stateless session resumption handling. Use initial_ctx when
10669 issuing and attempting to decrypt tickets in case it has changed during
10670 servername handling. Use a non-zero length session ID when attempting
10671 stateless session resumption: this makes it possible to determine if
10672 a resumption has occurred immediately after receiving server hello
10673 (several places in OpenSSL subtly assume this) instead of later in
10678 * The functions ENGINE_ctrl(), OPENSSL_isservice(),
10679 CMS_get1_RecipientRequest() and RAND_bytes() can return <=0 on error
10680 fixes for a few places where the return code is not checked
10683 *Julia Lawall <julia@diku.dk>*
10685 * Add --strict-warnings option to Configure script to include devteam
10686 warnings in other configurations.
10690 * Add support for --libdir option and LIBDIR variable in makefiles. This
10691 makes it possible to install openssl libraries in locations which
10692 have names other than "lib", for example "/usr/lib64" which some
10695 *Steve Henson, based on patch from Jeremy Utley*
10697 * Don't allow the use of leading 0x80 in OIDs. This is a violation of
10698 X690 8.9.12 and can produce some misleading textual output of OIDs.
10700 *Steve Henson, reported by Dan Kaminsky*
10702 * Delete MD2 from algorithm tables. This follows the recommendation in
10703 several standards that it is not used in new applications due to
10704 several cryptographic weaknesses. For binary compatibility reasons
10705 the MD2 API is still compiled in by default.
10709 * Add compression id to {d2i,i2d}_SSL_SESSION so it is correctly saved
10714 * Rename uni2asc and asc2uni functions to OPENSSL_uni2asc and
10715 OPENSSL_asc2uni conditionally on Netware platforms to avoid a name
10718 *Guenter <lists@gknw.net>*
10720 * Fix the server certificate chain building code to use X509_verify_cert(),
10721 it used to have an ad-hoc builder which was unable to cope with anything
10722 other than a simple chain.
10724 *David Woodhouse <dwmw2@infradead.org>, Steve Henson*
10726 * Don't check self signed certificate signatures in X509_verify_cert()
10727 by default (a flag can override this): it just wastes time without
10728 adding any security. As a useful side effect self signed root CAs
10729 with non-FIPS digests are now usable in FIPS mode.
10733 * In dtls1_process_out_of_seq_message() the check if the current message
10734 is already buffered was missing. For every new message was memory
10735 allocated, allowing an attacker to perform an denial of service attack
10736 with sending out of seq handshake messages until there is no memory
10737 left. Additionally every future message was buffered, even if the
10738 sequence number made no sense and would be part of another handshake.
10739 So only messages with sequence numbers less than 10 in advance will be
10740 buffered. ([CVE-2009-1378])
10742 *Robin Seggelmann, discovered by Daniel Mentz*
10744 * Records are buffered if they arrive with a future epoch to be
10745 processed after finishing the corresponding handshake. There is
10746 currently no limitation to this buffer allowing an attacker to perform
10747 a DOS attack with sending records with future epochs until there is no
10748 memory left. This patch adds the pqueue_size() function to determine
10749 the size of a buffer and limits the record buffer to 100 entries.
10752 *Robin Seggelmann, discovered by Daniel Mentz*
10754 * Keep a copy of frag->msg_header.frag_len so it can be used after the
10755 parent structure is freed. ([CVE-2009-1379])
10759 * Handle non-blocking I/O properly in SSL_shutdown() call.
10761 *Darryl Miles <darryl-mailinglists@netbauds.net>*
10763 * Add `2.5.4.*` OIDs
10765 *Ilya O. <vrghost@gmail.com>*
10767 ### Changes between 0.9.8k and 0.9.8l [5 Nov 2009]
10769 * Disable renegotiation completely - this fixes a severe security
10770 problem ([CVE-2009-3555]) at the cost of breaking all
10771 renegotiation. Renegotiation can be re-enabled by setting
10772 SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION in s3->flags at
10773 run-time. This is really not recommended unless you know what
10778 ### Changes between 0.9.8j and 0.9.8k [25 Mar 2009]
10780 * Don't set val to NULL when freeing up structures, it is freed up by
10781 underlying code. If `sizeof(void *) > sizeof(long)` this can result in
10782 zeroing past the valid field. ([CVE-2009-0789])
10784 *Paolo Ganci <Paolo.Ganci@AdNovum.CH>*
10786 * Fix bug where return value of CMS_SignerInfo_verify_content() was not
10787 checked correctly. This would allow some invalid signed attributes to
10788 appear to verify correctly. ([CVE-2009-0591])
10790 *Ivan Nestlerode <inestlerode@us.ibm.com>*
10792 * Reject UniversalString and BMPString types with invalid lengths. This
10793 prevents a crash in ASN1_STRING_print_ex() which assumes the strings have
10794 a legal length. ([CVE-2009-0590])
10798 * Set S/MIME signing as the default purpose rather than setting it
10799 unconditionally. This allows applications to override it at the store
10804 * Permit restricted recursion of ASN1 strings. This is needed in practice
10805 to handle some structures.
10809 * Improve efficiency of mem_gets: don't search whole buffer each time
10812 *Jeremy Shapiro <jnshapir@us.ibm.com>*
10814 * New -hex option for openssl rand.
10818 * Print out UTF8String and NumericString when parsing ASN1.
10822 * Support NumericString type for name components.
10826 * Allow CC in the environment to override the automatically chosen
10827 compiler. Note that nothing is done to ensure flags work with the
10832 ### Changes between 0.9.8i and 0.9.8j [07 Jan 2009]
10834 * Properly check EVP_VerifyFinal() and similar return values
10837 *Ben Laurie, Bodo Moeller, Google Security Team*
10839 * Enable TLS extensions by default.
10843 * Allow the CHIL engine to be loaded, whether the application is
10844 multithreaded or not. (This does not release the developer from the
10845 obligation to set up the dynamic locking callbacks.)
10847 *Sander Temme <sander@temme.net>*
10849 * Use correct exit code if there is an error in dgst command.
10851 *Steve Henson; problem pointed out by Roland Dirlewanger*
10853 * Tweak Configure so that you need to say "experimental-jpake" to enable
10854 JPAKE, and need to use -DOPENSSL_EXPERIMENTAL_JPAKE in applications.
10858 * Add experimental JPAKE support, including demo authentication in
10859 s_client and s_server.
10863 * Set the comparison function in v3_addr_canonize().
10865 *Rob Austein <sra@hactrn.net>*
10867 * Add support for XMPP STARTTLS in s_client.
10869 *Philip Paeps <philip@freebsd.org>*
10871 * Change the server-side SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG behavior
10872 to ensure that even with this option, only ciphersuites in the
10873 server's preference list will be accepted. (Note that the option
10874 applies only when resuming a session, so the earlier behavior was
10875 just about the algorithm choice for symmetric cryptography.)
10879 ### Changes between 0.9.8h and 0.9.8i [15 Sep 2008]
10881 * Fix NULL pointer dereference if a DTLS server received
10882 ChangeCipherSpec as first record ([CVE-2009-1386]).
10886 * Fix a state transition in s3_srvr.c and d1_srvr.c
10887 (was using SSL3_ST_CW_CLNT_HELLO_B, should be `..._ST_SW_SRVR_...`).
10889 *Nagendra Modadugu*
10891 * The fix in 0.9.8c that supposedly got rid of unsafe
10892 double-checked locking was incomplete for RSA blinding,
10893 addressing just one layer of what turns out to have been
10894 doubly unsafe triple-checked locking.
10896 So now fix this for real by retiring the MONT_HELPER macro
10897 in crypto/rsa/rsa_eay.c.
10899 *Bodo Moeller; problem pointed out by Marius Schilder*
10901 * Various precautionary measures:
10903 - Avoid size_t integer overflow in HASH_UPDATE (md32_common.h).
10905 - Avoid a buffer overflow in d2i_SSL_SESSION() (ssl_asn1.c).
10906 (NB: This would require knowledge of the secret session ticket key
10907 to exploit, in which case you'd be SOL either way.)
10909 - Change bn_nist.c so that it will properly handle input BIGNUMs
10910 outside the expected range.
10912 - Enforce the 'num' check in BN_div() (bn_div.c) for non-BN_DEBUG
10915 *Neel Mehta, Bodo Moeller*
10917 * Allow engines to be "soft loaded" - i.e. optionally don't die if
10918 the load fails. Useful for distros.
10920 *Ben Laurie and the FreeBSD team*
10922 * Add support for Local Machine Keyset attribute in PKCS#12 files.
10926 * Fix BN_GF2m_mod_arr() top-bit cleanup code.
10930 * Expand ENGINE to support engine supplied SSL client certificate functions.
10932 This work was sponsored by Logica.
10936 * Add CryptoAPI ENGINE to support use of RSA and DSA keys held in Windows
10937 keystores. Support for SSL/TLS client authentication too.
10938 Not compiled unless enable-capieng specified to Configure.
10940 This work was sponsored by Logica.
10944 * Fix bug in X509_ATTRIBUTE creation: don't set attribute using
10945 ASN1_TYPE_set1 if MBSTRING flag set. This bug would crash certain
10946 attribute creation routines such as certificate requests and PKCS#12
10951 ### Changes between 0.9.8g and 0.9.8h [28 May 2008]
10953 * Fix flaw if 'Server Key exchange message' is omitted from a TLS
10954 handshake which could lead to a client crash as found using the
10955 Codenomicon TLS test suite ([CVE-2008-1672])
10957 *Steve Henson, Mark Cox*
10959 * Fix double free in TLS server name extensions which could lead to
10960 a remote crash found by Codenomicon TLS test suite ([CVE-2008-0891])
10964 * Clear error queue in SSL_CTX_use_certificate_chain_file()
10966 Clear the error queue to ensure that error entries left from
10967 older function calls do not interfere with the correct operation.
10969 *Lutz Jaenicke, Erik de Castro Lopo*
10971 * Remove root CA certificates of commercial CAs:
10973 The OpenSSL project does not recommend any specific CA and does not
10974 have any policy with respect to including or excluding any CA.
10975 Therefore, it does not make any sense to ship an arbitrary selection
10976 of root CA certificates with the OpenSSL software.
10980 * RSA OAEP patches to fix two separate invalid memory reads.
10981 The first one involves inputs when 'lzero' is greater than
10982 'SHA_DIGEST_LENGTH' (it would read about SHA_DIGEST_LENGTH bytes
10983 before the beginning of from). The second one involves inputs where
10984 the 'db' section contains nothing but zeroes (there is a one-byte
10985 invalid read after the end of 'db').
10987 *Ivan Nestlerode <inestlerode@us.ibm.com>*
10989 * Partial backport from 0.9.9-dev:
10991 Introduce bn_mul_mont (dedicated Montgomery multiplication
10992 procedure) as a candidate for BIGNUM assembler implementation.
10993 While 0.9.9-dev uses assembler for various architectures, only
10994 x86_64 is available by default here in the 0.9.8 branch, and
10995 32-bit x86 is available through a compile-time setting.
10997 To try the 32-bit x86 assembler implementation, use Configure
10998 option "enable-montasm" (which exists only for this backport).
11000 As "enable-montasm" for 32-bit x86 disclaims code stability
11001 anyway, in this constellation we activate additional code
11002 backported from 0.9.9-dev for further performance improvements,
11003 namely BN_from_montgomery_word. (To enable this otherwise,
11004 e.g. x86_64, try `-DMONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD`.)
11006 *Andy Polyakov (backport partially by Bodo Moeller)*
11008 * Add TLS session ticket callback. This allows an application to set
11009 TLS ticket cipher and HMAC keys rather than relying on hardcoded fixed
11010 values. This is useful for key rollover for example where several key
11011 sets may exist with different names.
11015 * Reverse ENGINE-internal logic for caching default ENGINE handles.
11016 This was broken until now in 0.9.8 releases, such that the only way
11017 a registered ENGINE could be used (assuming it initialises
11018 successfully on the host) was to explicitly set it as the default
11019 for the relevant algorithms. This is in contradiction with 0.9.7
11020 behaviour and the documentation. With this fix, when an ENGINE is
11021 registered into a given algorithm's table of implementations, the
11022 'uptodate' flag is reset so that auto-discovery will be used next
11023 time a new context for that algorithm attempts to select an
11026 *Ian Lister (tweaked by Geoff Thorpe)*
11028 * Backport of CMS code to OpenSSL 0.9.8. This differs from the 0.9.9
11029 implementation in the following ways:
11031 Lack of EVP_PKEY_ASN1_METHOD means algorithm parameters have to be
11034 Lack of BER streaming support means one pass streaming processing is
11035 only supported if data is detached: setting the streaming flag is
11036 ignored for embedded content.
11038 CMS support is disabled by default and must be explicitly enabled
11039 with the enable-cms configuration option.
11043 * Update the GMP engine glue to do direct copies between BIGNUM and
11044 mpz_t when openssl and GMP use the same limb size. Otherwise the
11045 existing "conversion via a text string export" trick is still used.
11047 *Paul Sheer <paulsheer@gmail.com>*
11049 * Zlib compression BIO. This is a filter BIO which compressed and
11050 uncompresses any data passed through it.
11054 * Add AES_wrap_key() and AES_unwrap_key() functions to implement
11055 RFC3394 compatible AES key wrapping.
11059 * Add utility functions to handle ASN1 structures. ASN1_STRING_set0():
11060 sets string data without copying. X509_ALGOR_set0() and
11061 X509_ALGOR_get0(): set and retrieve X509_ALGOR (AlgorithmIdentifier)
11062 data. Attribute function X509at_get0_data_by_OBJ(): retrieves data
11063 from an X509_ATTRIBUTE structure optionally checking it occurs only
11064 once. ASN1_TYPE_set1(): set and ASN1_TYPE structure copying supplied
11069 * Fix BN flag handling in RSA_eay_mod_exp() and BN_MONT_CTX_set()
11070 to get the expected BN_FLG_CONSTTIME behavior.
11072 *Bodo Moeller (Google)*
11076 - fixed wrong usage of ioctlsocket() when build for LIBC BSD sockets
11077 - fixed do_tests.pl to run the test suite with CLIB builds too (CLIB_OPT)
11078 - added some more tests to do_tests.pl
11079 - fixed RunningProcess usage so that it works with newer LIBC NDKs too
11080 - removed usage of BN_LLONG for CLIB builds to avoid runtime dependency
11081 - added new Configure targets netware-clib-bsdsock, netware-clib-gcc,
11082 netware-clib-bsdsock-gcc, netware-libc-bsdsock-gcc
11083 - various changes to netware.pl to enable gcc-cross builds on Win32
11085 - changed crypto/bio/b_sock.c to work with macro functions (CLIB BSD)
11086 - various changes to fix missing prototype warnings
11087 - fixed x86nasm.pl to create correct asm files for NASM COFF output
11088 - added AES, WHIRLPOOL and CPUID assembler code to build files
11089 - added missing AES assembler make rules to mk1mf.pl
11090 - fixed order of includes in `apps/ocsp.c` so that `e_os.h` settings apply
11092 *Guenter Knauf <eflash@gmx.net>*
11094 * Implement certificate status request TLS extension defined in RFC3546.
11095 A client can set the appropriate parameters and receive the encoded
11096 OCSP response via a callback. A server can query the supplied parameters
11097 and set the encoded OCSP response in the callback. Add simplified examples
11098 to s_client and s_server.
11102 ### Changes between 0.9.8f and 0.9.8g [19 Oct 2007]
11104 * Fix various bugs:
11105 + Binary incompatibility of ssl_ctx_st structure
11106 + DTLS interoperation with non-compliant servers
11107 + Don't call get_session_cb() without proposed session
11108 + Fix ia64 assembler code
11110 *Andy Polyakov, Steve Henson*
11112 ### Changes between 0.9.8e and 0.9.8f [11 Oct 2007]
11114 * DTLS Handshake overhaul. There were longstanding issues with
11115 OpenSSL DTLS implementation, which were making it impossible for
11116 RFC 4347 compliant client to communicate with OpenSSL server.
11117 Unfortunately just fixing these incompatibilities would "cut off"
11118 pre-0.9.8f clients. To allow for hassle free upgrade post-0.9.8e
11119 server keeps tolerating non RFC compliant syntax. The opposite is
11120 not true, 0.9.8f client can not communicate with earlier server.
11121 This update even addresses CVE-2007-4995.
11125 * Changes to avoid need for function casts in OpenSSL: some compilers
11126 (gcc 4.2 and later) reject their use.
11127 *Kurt Roeckx <kurt@roeckx.be>, Peter Hartley <pdh@utter.chaos.org.uk>,
11130 * Add RFC4507 support to OpenSSL. This includes the corrections in
11131 RFC4507bis. The encrypted ticket format is an encrypted encoded
11132 SSL_SESSION structure, that way new session features are automatically
11135 If a client application caches session in an SSL_SESSION structure
11136 support is transparent because tickets are now stored in the encoded
11139 The SSL_CTX structure automatically generates keys for ticket
11140 protection in servers so again support should be possible
11141 with no application modification.
11143 If a client or server wishes to disable RFC4507 support then the option
11144 SSL_OP_NO_TICKET can be set.
11146 Add a TLS extension debugging callback to allow the contents of any client
11147 or server extensions to be examined.
11149 This work was sponsored by Google.
11153 * Add initial support for TLS extensions, specifically for the server_name
11154 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
11155 have new members for a hostname. The SSL data structure has an
11156 additional member `SSL_CTX *initial_ctx` so that new sessions can be
11157 stored in that context to allow for session resumption, even after the
11158 SSL has been switched to a new SSL_CTX in reaction to a client's
11159 server_name extension.
11161 New functions (subject to change):
11163 SSL_get_servername()
11164 SSL_get_servername_type()
11167 New CTRL codes and macros (subject to change):
11169 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
11170 - SSL_CTX_set_tlsext_servername_callback()
11171 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
11172 - SSL_CTX_set_tlsext_servername_arg()
11173 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
11175 openssl s_client has a new '-servername ...' option.
11177 openssl s_server has new options '-servername_host ...', '-cert2 ...',
11178 '-key2 ...', '-servername_fatal' (subject to change). This allows
11179 testing the HostName extension for a specific single hostname ('-cert'
11180 and '-key' remain fallbacks for handshakes without HostName
11181 negotiation). If the unrecognized_name alert has to be sent, this by
11182 default is a warning; it becomes fatal with the '-servername_fatal'
11185 *Peter Sylvester, Remy Allais, Christophe Renou, Steve Henson*
11187 * Add AES and SSE2 assembly language support to VC++ build.
11191 * Mitigate attack on final subtraction in Montgomery reduction.
11195 * Fix crypto/ec/ec_mult.c to work properly with scalars of value 0
11196 (which previously caused an internal error).
11200 * Squeeze another 10% out of IGE mode when in != out.
11204 * AES IGE mode speedup.
11206 *Dean Gaudet (Google)*
11208 * Add the Korean symmetric 128-bit cipher SEED (see
11209 <http://www.kisa.or.kr/kisa/seed/jsp/seed_eng.jsp>) and
11210 add SEED ciphersuites from RFC 4162:
11212 TLS_RSA_WITH_SEED_CBC_SHA = "SEED-SHA"
11213 TLS_DHE_DSS_WITH_SEED_CBC_SHA = "DHE-DSS-SEED-SHA"
11214 TLS_DHE_RSA_WITH_SEED_CBC_SHA = "DHE-RSA-SEED-SHA"
11215 TLS_DH_anon_WITH_SEED_CBC_SHA = "ADH-SEED-SHA"
11217 To minimize changes between patchlevels in the OpenSSL 0.9.8
11218 series, SEED remains excluded from compilation unless OpenSSL
11219 is configured with 'enable-seed'.
11221 *KISA, Bodo Moeller*
11223 * Mitigate branch prediction attacks, which can be practical if a
11224 single processor is shared, allowing a spy process to extract
11225 information. For detailed background information, see
11226 <http://eprint.iacr.org/2007/039> (O. Aciicmez, S. Gueron,
11227 J.-P. Seifert, "New Branch Prediction Vulnerabilities in OpenSSL
11228 and Necessary Software Countermeasures"). The core of the change
11229 are new versions BN_div_no_branch() and
11230 BN_mod_inverse_no_branch() of BN_div() and BN_mod_inverse(),
11231 respectively, which are slower, but avoid the security-relevant
11232 conditional branches. These are automatically called by BN_div()
11233 and BN_mod_inverse() if the flag BN_FLG_CONSTTIME is set for one
11234 of the input BIGNUMs. Also, BN_is_bit_set() has been changed to
11235 remove a conditional branch.
11237 BN_FLG_CONSTTIME is the new name for the previous
11238 BN_FLG_EXP_CONSTTIME flag, since it now affects more than just
11239 modular exponentiation. (Since OpenSSL 0.9.7h, setting this flag
11240 in the exponent causes BN_mod_exp_mont() to use the alternative
11241 implementation in BN_mod_exp_mont_consttime().) The old name
11242 remains as a deprecated alias.
11244 Similarly, RSA_FLAG_NO_EXP_CONSTTIME is replaced by a more general
11245 RSA_FLAG_NO_CONSTTIME flag since the RSA implementation now uses
11246 constant-time implementations for more than just exponentiation.
11247 Here too the old name is kept as a deprecated alias.
11249 BN_BLINDING_new() will now use BN_dup() for the modulus so that
11250 the BN_BLINDING structure gets an independent copy of the
11251 modulus. This means that the previous `BIGNUM *m` argument to
11252 BN_BLINDING_new() and to BN_BLINDING_create_param() now
11253 essentially becomes `const BIGNUM *m`, although we can't actually
11254 change this in the header file before 0.9.9. It allows
11255 RSA_setup_blinding() to use BN_with_flags() on the modulus to
11256 enable BN_FLG_CONSTTIME.
11258 *Matthew D Wood (Intel Corp)*
11260 * In the SSL/TLS server implementation, be strict about session ID
11261 context matching (which matters if an application uses a single
11262 external cache for different purposes). Previously,
11263 out-of-context reuse was forbidden only if SSL_VERIFY_PEER was
11264 set. This did ensure strict client verification, but meant that,
11265 with applications using a single external cache for quite
11266 different requirements, clients could circumvent ciphersuite
11267 restrictions for a given session ID context by starting a session
11268 in a different context.
11272 * Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
11273 a ciphersuite string such as "DEFAULT:RSA" cannot enable
11274 authentication-only ciphersuites.
11278 * Update the SSL_get_shared_ciphers() fix CVE-2006-3738 which was
11279 not complete and could lead to a possible single byte overflow
11280 ([CVE-2007-5135]) [Ben Laurie]
11282 ### Changes between 0.9.8d and 0.9.8e [23 Feb 2007]
11284 * Since AES128 and AES256 (and similarly Camellia128 and
11285 Camellia256) share a single mask bit in the logic of
11286 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
11287 kludge to work properly if AES128 is available and AES256 isn't
11288 (or if Camellia128 is available and Camellia256 isn't).
11292 * Fix the BIT STRING encoding generated by crypto/ec/ec_asn1.c
11293 (within i2d_ECPrivateKey, i2d_ECPKParameters, i2d_ECParameters):
11294 When a point or a seed is encoded in a BIT STRING, we need to
11295 prevent the removal of trailing zero bits to get the proper DER
11296 encoding. (By default, crypto/asn1/a_bitstr.c assumes the case
11297 of a NamedBitList, for which trailing 0 bits need to be removed.)
11301 * Have SSL/TLS server implementation tolerate "mismatched" record
11302 protocol version while receiving ClientHello even if the
11303 ClientHello is fragmented. (The server can't insist on the
11304 particular protocol version it has chosen before the ServerHello
11305 message has informed the client about his choice.)
11309 * Add RFC 3779 support.
11311 *Rob Austein for ARIN, Ben Laurie*
11313 * Load error codes if they are not already present instead of using a
11314 static variable. This allows them to be cleanly unloaded and reloaded.
11315 Improve header file function name parsing.
11319 * extend SMTP and IMAP protocol emulation in s_client to use EHLO
11320 or CAPABILITY handshake as required by RFCs.
11322 *Goetz Babin-Ebell*
11324 ### Changes between 0.9.8c and 0.9.8d [28 Sep 2006]
11326 * Introduce limits to prevent malicious keys being able to
11327 cause a denial of service. ([CVE-2006-2940])
11329 *Steve Henson, Bodo Moeller*
11331 * Fix ASN.1 parsing of certain invalid structures that can result
11332 in a denial of service. ([CVE-2006-2937]) [Steve Henson]
11334 * Fix buffer overflow in SSL_get_shared_ciphers() function.
11335 ([CVE-2006-3738]) [Tavis Ormandy and Will Drewry, Google Security Team]
11337 * Fix SSL client code which could crash if connecting to a
11338 malicious SSLv2 server. ([CVE-2006-4343])
11340 *Tavis Ormandy and Will Drewry, Google Security Team*
11342 * Since 0.9.8b, ciphersuite strings naming explicit ciphersuites
11343 match only those. Before that, "AES256-SHA" would be interpreted
11344 as a pattern and match "AES128-SHA" too (since AES128-SHA got
11345 the same strength classification in 0.9.7h) as we currently only
11346 have a single AES bit in the ciphersuite description bitmap.
11347 That change, however, also applied to ciphersuite strings such as
11348 "RC4-MD5" that intentionally matched multiple ciphersuites --
11349 namely, SSL 2.0 ciphersuites in addition to the more common ones
11350 from SSL 3.0/TLS 1.0.
11352 So we change the selection algorithm again: Naming an explicit
11353 ciphersuite selects this one ciphersuite, and any other similar
11354 ciphersuite (same bitmap) from *other* protocol versions.
11355 Thus, "RC4-MD5" again will properly select both the SSL 2.0
11356 ciphersuite and the SSL 3.0/TLS 1.0 ciphersuite.
11358 Since SSL 2.0 does not have any ciphersuites for which the
11359 128/256 bit distinction would be relevant, this works for now.
11360 The proper fix will be to use different bits for AES128 and
11361 AES256, which would have avoided the problems from the beginning;
11362 however, bits are scarce, so we can only do this in a new release
11363 (not just a patchlevel) when we can change the SSL_CIPHER
11364 definition to split the single 'unsigned long mask' bitmap into
11365 multiple values to extend the available space.
11369 ### Changes between 0.9.8b and 0.9.8c [05 Sep 2006]
11371 * Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
11372 ([CVE-2006-4339]) [Ben Laurie and Google Security Team]
11374 * Add AES IGE and biIGE modes.
11378 * Change the Unix randomness entropy gathering to use poll() when
11379 possible instead of select(), since the latter has some
11380 undesirable limitations.
11382 *Darryl Miles via Richard Levitte and Bodo Moeller*
11384 * Disable "ECCdraft" ciphersuites more thoroughly. Now special
11385 treatment in ssl/ssl_ciph.s makes sure that these ciphersuites
11386 cannot be implicitly activated as part of, e.g., the "AES" alias.
11387 However, please upgrade to OpenSSL 0.9.9[-dev] for
11388 non-experimental use of the ECC ciphersuites to get TLS extension
11389 support, which is required for curve and point format negotiation
11390 to avoid potential handshake problems.
11394 * Disable rogue ciphersuites:
11396 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
11397 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
11398 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
11400 The latter two were purportedly from
11401 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
11404 Also deactivate the remaining ciphersuites from
11405 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
11406 unofficial, and the ID has long expired.
11410 * Fix RSA blinding Heisenbug (problems sometimes occurred on
11411 dual-core machines) and other potential thread-safety issues.
11415 * Add the symmetric cipher Camellia (128-bit, 192-bit, 256-bit key
11416 versions), which is now available for royalty-free use
11417 (see <http://info.isl.ntt.co.jp/crypt/eng/info/chiteki.html>).
11418 Also, add Camellia TLS ciphersuites from RFC 4132.
11420 To minimize changes between patchlevels in the OpenSSL 0.9.8
11421 series, Camellia remains excluded from compilation unless OpenSSL
11422 is configured with 'enable-camellia'.
11426 * Disable the padding bug check when compression is in use. The padding
11427 bug check assumes the first packet is of even length, this is not
11428 necessarily true if compression is enabled and can result in false
11429 positives causing handshake failure. The actual bug test is ancient
11430 code so it is hoped that implementations will either have fixed it by
11431 now or any which still have the bug do not support compression.
11435 ### Changes between 0.9.8a and 0.9.8b [04 May 2006]
11437 * When applying a cipher rule check to see if string match is an explicit
11438 cipher suite and only match that one cipher suite if it is.
11442 * Link in manifests for VC++ if needed.
11444 *Austin Ziegler <halostatue@gmail.com>*
11446 * Update support for ECC-based TLS ciphersuites according to
11447 draft-ietf-tls-ecc-12.txt with proposed changes (but without
11448 TLS extensions, which are supported starting with the 0.9.9
11449 branch, not in the OpenSSL 0.9.8 branch).
11453 * New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support
11454 opaque EVP_CIPHER_CTX handling.
11458 * Fixes and enhancements to zlib compression code. We now only use
11459 "zlib1.dll" and use the default `__cdecl` calling convention on Win32
11460 to conform with the standards mentioned here:
11461 <http://www.zlib.net/DLL_FAQ.txt>
11462 Static zlib linking now works on Windows and the new --with-zlib-include
11463 --with-zlib-lib options to Configure can be used to supply the location
11464 of the headers and library. Gracefully handle case where zlib library
11469 * Several fixes and enhancements to the OID generation code. The old code
11470 sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't
11471 handle numbers larger than ULONG_MAX, truncated printing and had a
11472 non standard OBJ_obj2txt() behaviour.
11476 * Add support for building of engines under engine/ as shared libraries
11477 under VC++ build system.
11481 * Corrected the numerous bugs in the Win32 path splitter in DSO.
11482 Hopefully, we will not see any false combination of paths any more.
11486 ### Changes between 0.9.8 and 0.9.8a [11 Oct 2005]
11488 * Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
11489 (part of SSL_OP_ALL). This option used to disable the
11490 countermeasure against man-in-the-middle protocol-version
11491 rollback in the SSL 2.0 server implementation, which is a bad
11492 idea. ([CVE-2005-2969])
11494 *Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
11495 for Information Security, National Institute of Advanced Industrial
11496 Science and Technology [AIST], Japan)*
11498 * Add two function to clear and return the verify parameter flags.
11502 * Keep cipherlists sorted in the source instead of sorting them at
11503 runtime, thus removing the need for a lock.
11507 * Avoid some small subgroup attacks in Diffie-Hellman.
11509 *Nick Mathewson and Ben Laurie*
11511 * Add functions for well-known primes.
11515 * Extended Windows CE support.
11517 *Satoshi Nakamura and Andy Polyakov*
11519 * Initialize SSL_METHOD structures at compile time instead of during
11520 runtime, thus removing the need for a lock.
11524 * Make PKCS7_decrypt() work even if no certificate is supplied by
11525 attempting to decrypt each encrypted key in turn. Add support to
11530 ### Changes between 0.9.7h and 0.9.8 [05 Jul 2005]
11532 [NB: OpenSSL 0.9.7i and later 0.9.7 patch levels were released after
11535 * Add libcrypto.pc and libssl.pc for those who feel they need them.
11539 * Change CA.sh and CA.pl so they don't bundle the CSR and the private
11540 key into the same file any more.
11544 * Add initial support for Win64, both IA64 and AMD64/x64 flavors.
11548 * Add -utf8 command line and config file option to 'ca'.
11550 *Stefan <stf@udoma.org*
11552 * Removed the macro des_crypt(), as it seems to conflict with some
11553 libraries. Use DES_crypt().
11557 * Correct naming of the 'chil' and '4758cca' ENGINEs. This
11558 involves renaming the source and generated shared-libs for
11559 both. The engines will accept the corrected or legacy ids
11560 ('ncipher' and '4758_cca' respectively) when binding. NB,
11561 this only applies when building 'shared'.
11563 *Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe*
11565 * Add attribute functions to EVP_PKEY structure. Modify
11566 PKCS12_create() to recognize a CSP name attribute and
11567 use it. Make -CSP option work again in pkcs12 utility.
11571 * Add new functionality to the bn blinding code:
11572 - automatic re-creation of the BN_BLINDING parameters after
11573 a fixed number of uses (currently 32)
11574 - add new function for parameter creation
11575 - introduce flags to control the update behaviour of the
11576 BN_BLINDING parameters
11577 - hide BN_BLINDING structure
11578 Add a second BN_BLINDING slot to the RSA structure to improve
11579 performance when a single RSA object is shared among several
11584 * Add support for DTLS.
11586 *Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie*
11588 * Add support for DER encoded private keys (SSL_FILETYPE_ASN1)
11589 to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file()
11593 * Remove buggy and incomplete DH cert support from
11594 ssl/ssl_rsa.c and ssl/s3_both.c
11598 * Use SHA-1 instead of MD5 as the default digest algorithm for
11599 the `apps/openssl` commands.
11603 * Compile clean with "-Wall -Wmissing-prototypes
11604 -Wstrict-prototypes -Wmissing-declarations -Werror". Currently
11605 DEBUG_SAFESTACK must also be set.
11609 * Change ./Configure so that certain algorithms can be disabled by default.
11610 The new counterpiece to "no-xxx" is "enable-xxx".
11612 The patented RC5 and MDC2 algorithms will now be disabled unless
11613 "enable-rc5" and "enable-mdc2", respectively, are specified.
11615 (IDEA remains enabled despite being patented. This is because IDEA
11616 is frequently required for interoperability, and there is no license
11617 fee for non-commercial use. As before, "no-idea" can be used to
11618 avoid this algorithm.)
11622 * Add processing of proxy certificates (see RFC 3820). This work was
11623 sponsored by KTH (The Royal Institute of Technology in Stockholm) and
11624 EGEE (Enabling Grids for E-science in Europe).
11628 * RC4 performance overhaul on modern architectures/implementations, such
11629 as Intel P4, IA-64 and AMD64.
11633 * New utility extract-section.pl. This can be used specify an alternative
11634 section number in a pod file instead of having to treat each file as
11635 a separate case in Makefile. This can be done by adding two lines to the
11638 =for comment openssl_section:XXX
11640 The blank line is mandatory.
11644 * New arguments -certform, -keyform and -pass for s_client and s_server
11645 to allow alternative format key and certificate files and passphrase
11650 * New structure X509_VERIFY_PARAM which combines current verify parameters,
11651 update associated structures and add various utility functions.
11653 Add new policy related verify parameters, include policy checking in
11654 standard verify code. Enhance 'smime' application with extra parameters
11655 to support policy checking and print out.
11659 * Add a new engine to support VIA PadLock ACE extensions in the VIA C3
11660 Nehemiah processors. These extensions support AES encryption in hardware
11661 as well as RNG (though RNG support is currently disabled).
11663 *Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov*
11665 * Deprecate `BN_[get|set]_params()` functions (they were ignored internally).
11669 * New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented.
11671 *Andy Polyakov and a number of other people*
11673 * Improved PowerPC platform support. Most notably BIGNUM assembler
11674 implementation contributed by IBM.
11676 *Suresh Chari, Peter Waltenberg, Andy Polyakov*
11678 * The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public
11679 exponent rather than 'unsigned long'. There is a corresponding change to
11680 the new 'rsa_keygen' element of the RSA_METHOD structure.
11682 *Jelte Jansen, Geoff Thorpe*
11684 * Functionality for creating the initial serial number file is now
11685 moved from CA.pl to the 'ca' utility with a new option -create_serial.
11687 (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial
11688 number file to 1, which is bound to cause problems. To avoid
11689 the problems while respecting compatibility between different 0.9.7
11690 patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in
11691 CA.pl for serial number initialization. With the new release 0.9.8,
11692 we can fix the problem directly in the 'ca' utility.)
11696 * Reduced header interdependencies by declaring more opaque objects in
11697 ossl_typ.h. As a consequence, including some headers (eg. engine.h) will
11698 give fewer recursive includes, which could break lazy source code - so
11699 this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always,
11700 developers should define this symbol when building and using openssl to
11701 ensure they track the recommended behaviour, interfaces, [etc], but
11702 backwards-compatible behaviour prevails when this isn't defined.
11706 * New function X509_POLICY_NODE_print() which prints out policy nodes.
11710 * Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality.
11711 This will generate a random key of the appropriate length based on the
11712 cipher context. The EVP_CIPHER can provide its own random key generation
11713 routine to support keys of a specific form. This is used in the des and
11714 3des routines to generate a key of the correct parity. Update S/MIME
11715 code to use new functions and hence generate correct parity DES keys.
11716 Add EVP_CHECK_DES_KEY #define to return an error if the key is not
11717 valid (weak or incorrect parity).
11721 * Add a local set of CRLs that can be used by X509_verify_cert() as well
11722 as looking them up. This is useful when the verified structure may contain
11723 CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs
11724 present unless the new PKCS7_NO_CRL flag is asserted.
11728 * Extend ASN1 oid configuration module. It now additionally accepts the
11731 shortName = some long name, 1.2.3.4
11735 * Reimplemented the BN_CTX implementation. There is now no more static
11736 limitation on the number of variables it can handle nor the depth of the
11737 "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack
11738 information can now expand as required, and rather than having a single
11739 static array of bignums, BN_CTX now uses a linked-list of such arrays
11740 allowing it to expand on demand whilst maintaining the usefulness of
11741 BN_CTX's "bundling".
11745 * Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD
11746 to allow all RSA operations to function using a single BN_CTX.
11750 * Preliminary support for certificate policy evaluation and checking. This
11751 is initially intended to pass the tests outlined in "Conformance Testing
11752 of Relying Party Client Certificate Path Processing Logic" v1.07.
11756 * bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and
11757 remained unused and not that useful. A variety of other little bignum
11758 tweaks and fixes have also been made continuing on from the audit (see
11763 * Constify all or almost all d2i, c2i, s2i and r2i functions, along with
11764 associated ASN1, EVP and SSL functions and old ASN1 macros.
11768 * BN_zero() only needs to set 'top' and 'neg' to zero for correct results,
11769 and this should never fail. So the return value from the use of
11770 BN_set_word() (which can fail due to needless expansion) is now deprecated;
11771 if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro.
11775 * BN_CTX_get() should return zero-valued bignums, providing the same
11776 initialised value as BN_new().
11778 *Geoff Thorpe, suggested by Ulf Möller*
11780 * Support for inhibitAnyPolicy certificate extension.
11784 * An audit of the BIGNUM code is underway, for which debugging code is
11785 enabled when BN_DEBUG is defined. This makes stricter enforcements on what
11786 is considered valid when processing BIGNUMs, and causes execution to
11787 assert() when a problem is discovered. If BN_DEBUG_RAND is defined,
11788 further steps are taken to deliberately pollute unused data in BIGNUM
11789 structures to try and expose faulty code further on. For now, openssl will
11790 (in its default mode of operation) continue to tolerate the inconsistent
11791 forms that it has tolerated in the past, but authors and packagers should
11792 consider trying openssl and their own applications when compiled with
11793 these debugging symbols defined. It will help highlight potential bugs in
11794 their own code, and will improve the test coverage for OpenSSL itself. At
11795 some point, these tighter rules will become openssl's default to improve
11796 maintainability, though the assert()s and other overheads will remain only
11797 in debugging configurations. See bn.h for more details.
11799 *Geoff Thorpe, Nils Larsch, Ulf Möller*
11801 * BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure
11802 that can only be obtained through BN_CTX_new() (which implicitly
11803 initialises it). The presence of this function only made it possible
11804 to overwrite an existing structure (and cause memory leaks).
11808 * Because of the callback-based approach for implementing LHASH as a
11809 template type, lh_insert() adds opaque objects to hash-tables and
11810 lh_doall() or lh_doall_arg() are typically used with a destructor callback
11811 to clean up those corresponding objects before destroying the hash table
11812 (and losing the object pointers). So some over-zealous constifications in
11813 LHASH have been relaxed so that lh_insert() does not take (nor store) the
11814 objects as "const" and the `lh_doall[_arg]` callback wrappers are not
11815 prototyped to have "const" restrictions on the object pointers they are
11816 given (and so aren't required to cast them away any more).
11820 * The tmdiff.h API was so ugly and minimal that our own timing utility
11821 (speed) prefers to use its own implementation. The two implementations
11822 haven't been consolidated as yet (volunteers?) but the tmdiff API has had
11823 its object type properly exposed (MS_TM) instead of casting to/from
11824 `char *`. This may still change yet if someone realises MS_TM and
11826 aren't necessarily the greatest nomenclatures - but this is what was used
11827 internally to the implementation so I've used that for now.
11831 * Ensure that deprecated functions do not get compiled when
11832 OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of
11833 the self-tests were still using deprecated key-generation functions so
11834 these have been updated also.
11838 * Reorganise PKCS#7 code to separate the digest location functionality
11839 into PKCS7_find_digest(), digest addition into PKCS7_bio_add_digest().
11840 New function PKCS7_set_digest() to set the digest type for PKCS#7
11841 digestedData type. Add additional code to correctly generate the
11842 digestedData type and add support for this type in PKCS7 initialization
11847 * New function PKCS7_set0_type_other() this initializes a PKCS7
11848 structure of type "other".
11852 * Fix prime generation loop in crypto/bn/bn_prime.pl by making
11853 sure the loop does correctly stop and breaking ("division by zero")
11854 modulus operations are not performed. The (pre-generated) prime
11855 table crypto/bn/bn_prime.h was already correct, but it could not be
11856 re-generated on some platforms because of the "division by zero"
11857 situation in the script.
11859 *Ralf S. Engelschall*
11861 * Update support for ECC-based TLS ciphersuites according to
11862 draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with
11863 SHA-1 now is only used for "small" curves (where the
11864 representation of a field element takes up to 24 bytes); for
11865 larger curves, the field element resulting from ECDH is directly
11866 used as premaster secret.
11868 *Douglas Stebila (Sun Microsystems Laboratories)*
11870 * Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2
11871 curve secp160r1 to the tests.
11873 *Douglas Stebila (Sun Microsystems Laboratories)*
11875 * Add the possibility to load symbols globally with DSO.
11877 *Götz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte*
11879 * Add the functions ERR_set_mark() and ERR_pop_to_mark() for better
11880 control of the error stack.
11884 * Add support for STORE in ENGINE.
11888 * Add the STORE type. The intention is to provide a common interface
11889 to certificate and key stores, be they simple file-based stores, or
11890 HSM-type store, or LDAP stores, or...
11891 NOTE: The code is currently UNTESTED and isn't really used anywhere.
11895 * Add a generic structure called OPENSSL_ITEM. This can be used to
11896 pass a list of arguments to any function as well as provide a way
11897 for a function to pass data back to the caller.
11901 * Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup()
11902 works like BUF_strdup() but can be used to duplicate a portion of
11903 a string. The copy gets NUL-terminated. BUF_memdup() duplicates
11908 * Add the function sk_find_ex() which works like sk_find(), but will
11909 return an index to an element even if an exact match couldn't be
11910 found. The index is guaranteed to point at the element where the
11911 searched-for key would be inserted to preserve sorting order.
11915 * Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but
11916 takes an extra flags argument for optional functionality. Currently,
11917 the following flags are defined:
11919 OBJ_BSEARCH_VALUE_ON_NOMATCH
11920 This one gets OBJ_bsearch_ex() to return a pointer to the first
11921 element where the comparing function returns a negative or zero
11924 OBJ_BSEARCH_FIRST_VALUE_ON_MATCH
11925 This one gets OBJ_bsearch_ex() to return a pointer to the first
11926 element where the comparing function returns zero. This is useful
11927 if there are more than one element where the comparing function
11932 * Make it possible to create self-signed certificates with 'openssl ca'
11933 in such a way that the self-signed certificate becomes part of the
11934 CA database and uses the same mechanisms for serial number generation
11935 as all other certificate signing. The new flag '-selfsign' enables
11936 this functionality. Adapt CA.sh and CA.pl.in.
11940 * Add functionality to check the public key of a certificate request
11941 against a given private. This is useful to check that a certificate
11942 request can be signed by that key (self-signing).
11946 * Make it possible to have multiple active certificates with the same
11947 subject in the CA index file. This is done only if the keyword
11948 'unique_subject' is set to 'no' in the main CA section (default
11949 if 'CA_default') of the configuration file. The value is saved
11950 with the database itself in a separate index attribute file,
11951 named like the index file with '.attr' appended to the name.
11955 * Generate multi-valued AVAs using '+' notation in config files for
11960 * Support for nameConstraints certificate extension.
11964 * Support for policyConstraints certificate extension.
11968 * Support for policyMappings certificate extension.
11972 * Make sure the default DSA_METHOD implementation only uses its
11973 dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL,
11974 and change its own handlers to be NULL so as to remove unnecessary
11975 indirection. This lets alternative implementations fallback to the
11976 default implementation more easily.
11980 * Support for directoryName in GeneralName related extensions
11985 * Make it possible to link applications using Makefile.shared.
11986 Make that possible even when linking against static libraries!
11990 * Support for single pass processing for S/MIME signing. This now
11991 means that S/MIME signing can be done from a pipe, in addition
11992 cleartext signing (multipart/signed type) is effectively streaming
11993 and the signed data does not need to be all held in memory.
11995 This is done with a new flag PKCS7_STREAM. When this flag is set
11996 PKCS7_sign() only initializes the PKCS7 structure and the actual signing
11997 is done after the data is output (and digests calculated) in
11998 SMIME_write_PKCS7().
12002 * Add full support for -rpath/-R, both in shared libraries and
12003 applications, at least on the platforms where it's known how
12008 * In crypto/ec/ec_mult.c, implement fast point multiplication with
12009 precomputation, based on wNAF splitting: EC_GROUP_precompute_mult()
12010 will now compute a table of multiples of the generator that
12011 makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul()
12012 faster (notably in the case of a single point multiplication,
12013 scalar * generator).
12015 *Nils Larsch, Bodo Moeller*
12017 * IPv6 support for certificate extensions. The various extensions
12018 which use the IP:a.b.c.d can now take IPv6 addresses using the
12019 formats of RFC1884 2.2 . IPv6 addresses are now also displayed
12024 * Added an ENGINE that implements RSA by performing private key
12025 exponentiations with the GMP library. The conversions to and from
12026 GMP's mpz_t format aren't optimised nor are any montgomery forms
12027 cached, and on x86 it appears OpenSSL's own performance has caught up.
12028 However there are likely to be other architectures where GMP could
12029 provide a boost. This ENGINE is not built in by default, but it can be
12030 specified at Configure time and should be accompanied by the necessary
12031 linker additions, eg;
12032 ./config -DOPENSSL_USE_GMP -lgmp
12036 * "openssl engine" will not display ENGINE/DSO load failure errors when
12037 testing availability of engines with "-t" - the old behaviour is
12038 produced by increasing the feature's verbosity with "-tt".
12042 * ECDSA routines: under certain error conditions uninitialized BN objects
12043 could be freed. Solution: make sure initialization is performed early
12044 enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de>
12049 * Key-generation can now be implemented in RSA_METHOD, DSA_METHOD
12050 and DH_METHOD (eg. by ENGINE implementations) to override the normal
12051 software implementations. For DSA and DH, parameter generation can
12052 also be overridden by providing the appropriate method callbacks.
12056 * Change the "progress" mechanism used in key-generation and
12057 primality testing to functions that take a new BN_GENCB pointer in
12058 place of callback/argument pairs. The new API functions have `_ex`
12059 postfixes and the older functions are reimplemented as wrappers for
12060 the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide
12061 declarations of the old functions to help (graceful) attempts to
12062 migrate to the new functions. Also, the new key-generation API
12063 functions operate on a caller-supplied key-structure and return
12064 success/failure rather than returning a key or NULL - this is to
12065 help make "keygen" another member function of RSA_METHOD etc.
12067 Example for using the new callback interface:
12069 int (*my_callback)(int a, int b, BN_GENCB *cb) = ...;
12070 void *my_arg = ...;
12073 BN_GENCB_set(&my_cb, my_callback, my_arg);
12075 return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb);
12076 /* For the meaning of a, b in calls to my_callback(), see the
12077 * documentation of the function that calls the callback.
12078 * cb will point to my_cb; my_arg can be retrieved as cb->arg.
12079 * my_callback should return 1 if it wants BN_is_prime_ex()
12080 * to continue, or 0 to stop.
12085 * Change the ZLIB compression method to be stateful, and make it
12086 available to TLS with the number defined in
12087 draft-ietf-tls-compression-04.txt.
12091 * Add the ASN.1 structures and functions for CertificatePair, which
12092 is defined as follows (according to X.509_4thEditionDraftV6.pdf):
12094 CertificatePair ::= SEQUENCE {
12095 forward [0] Certificate OPTIONAL,
12096 reverse [1] Certificate OPTIONAL,
12097 -- at least one of the pair shall be present -- }
12099 Also implement the PEM functions to read and write certificate
12100 pairs, and defined the PEM tag as "CERTIFICATE PAIR".
12102 This needed to be defined, mostly for the sake of the LDAP
12103 attribute crossCertificatePair, but may prove useful elsewhere as
12108 * Make it possible to inhibit symlinking of shared libraries in
12109 Makefile.shared, for Cygwin's sake.
12113 * Extend the BIGNUM API by creating a function
12114 void BN_set_negative(BIGNUM *a, int neg);
12115 and a macro that behave like
12116 int BN_is_negative(const BIGNUM *a);
12118 to avoid the need to access 'a->neg' directly in applications.
12122 * Implement fast modular reduction for pseudo-Mersenne primes
12123 used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c).
12124 EC_GROUP_new_curve_GFp() will now automatically use this
12127 *Nils Larsch <nla@trustcenter.de>*
12129 * Add new lock type (CRYPTO_LOCK_BN).
12133 * Change the ENGINE framework to automatically load engines
12134 dynamically from specific directories unless they could be
12135 found to already be built in or loaded. Move all the
12136 current engines except for the cryptodev one to a new
12137 directory engines/.
12138 The engines in engines/ are built as shared libraries if
12139 the "shared" options was given to ./Configure or ./config.
12140 Otherwise, they are inserted in libcrypto.a.
12141 /usr/local/ssl/engines is the default directory for dynamic
12142 engines, but that can be overridden at configure time through
12143 the usual use of --prefix and/or --openssldir, and at run
12144 time with the environment variable OPENSSL_ENGINES.
12146 *Geoff Thorpe and Richard Levitte*
12148 * Add Makefile.shared, a helper makefile to build shared
12149 libraries. Adapt Makefile.org.
12153 * Add version info to Win32 DLLs.
12155 *Peter 'Luna' Runestig" <peter@runestig.com>*
12157 * Add new 'medium level' PKCS#12 API. Certificates and keys
12158 can be added using this API to created arbitrary PKCS#12
12159 files while avoiding the low-level API.
12161 New options to PKCS12_create(), key or cert can be NULL and
12162 will then be omitted from the output file. The encryption
12163 algorithm NIDs can be set to -1 for no encryption, the mac
12164 iteration count can be set to 0 to omit the mac.
12166 Enhance pkcs12 utility by making the -nokeys and -nocerts
12167 options work when creating a PKCS#12 file. New option -nomac
12168 to omit the mac, NONE can be set for an encryption algorithm.
12169 New code is modified to use the enhanced PKCS12_create()
12170 instead of the low-level API.
12174 * Extend ASN1 encoder to support indefinite length constructed
12175 encoding. This can output sequences tags and octet strings in
12176 this form. Modify pk7_asn1.c to support indefinite length
12177 encoding. This is experimental and needs additional code to
12178 be useful, such as an ASN1 bio and some enhanced streaming
12181 Extend template encode functionality so that tagging is passed
12182 down to the template encoder.
12186 * Let 'openssl req' fail if an argument to '-newkey' is not
12187 recognized instead of using RSA as a default.
12191 * Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt.
12192 As these are not official, they are not included in "ALL";
12193 the "ECCdraft" ciphersuite group alias can be used to select them.
12195 *Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)*
12197 * Add ECDH engine support.
12199 *Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)*
12201 * Add ECDH in new directory crypto/ecdh/.
12203 *Douglas Stebila (Sun Microsystems Laboratories)*
12205 * Let BN_rand_range() abort with an error after 100 iterations
12206 without success (which indicates a broken PRNG).
12210 * Change BN_mod_sqrt() so that it verifies that the input value
12211 is really the square of the return value. (Previously,
12212 BN_mod_sqrt would show GIGO behaviour.)
12216 * Add named elliptic curves over binary fields from X9.62, SECG,
12217 and WAP/WTLS; add OIDs that were still missing.
12219 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
12221 * Extend the EC library for elliptic curves over binary fields
12222 (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/).
12225 EC_GF2m_simple_method
12229 EC_GROUP_new_curve_GF2m
12230 EC_GROUP_set_curve_GF2m
12231 EC_GROUP_get_curve_GF2m
12232 EC_POINT_set_affine_coordinates_GF2m
12233 EC_POINT_get_affine_coordinates_GF2m
12234 EC_POINT_set_compressed_coordinates_GF2m
12236 Point compression for binary fields is disabled by default for
12237 patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to
12240 As binary polynomials are represented as BIGNUMs, various members
12241 of the EC_GROUP and EC_POINT data structures can be shared
12242 between the implementations for prime fields and binary fields;
12243 the above `..._GF2m functions` (except for EX_GROUP_new_curve_GF2m)
12244 are essentially identical to their `..._GFp` counterparts.
12245 (For simplicity, the `..._GFp` prefix has been dropped from
12246 various internal method names.)
12248 An internal 'field_div' method (similar to 'field_mul' and
12249 'field_sqr') has been added; this is used only for binary fields.
12251 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
12253 * Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult()
12254 through methods ('mul', 'precompute_mult').
12256 The generic implementations (now internally called 'ec_wNAF_mul'
12257 and 'ec_wNAF_precomputed_mult') remain the default if these
12258 methods are undefined.
12260 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
12262 * New function EC_GROUP_get_degree, which is defined through
12263 EC_METHOD. For curves over prime fields, this returns the bit
12264 length of the modulus.
12266 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
12268 * New functions EC_GROUP_dup, EC_POINT_dup.
12269 (These simply call ..._new and ..._copy).
12271 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
12273 * Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c.
12274 Polynomials are represented as BIGNUMs (where the sign bit is not
12275 used) in the following functions [macros]:
12278 BN_GF2m_sub [= BN_GF2m_add]
12279 BN_GF2m_mod [wrapper for BN_GF2m_mod_arr]
12280 BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr]
12281 BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr]
12283 BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr]
12284 BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr]
12285 BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr]
12286 BN_GF2m_cmp [= BN_ucmp]
12288 (Note that only the 'mod' functions are actually for fields GF(2^m).
12289 BN_GF2m_add() is misnomer, but this is for the sake of consistency.)
12291 For some functions, an the irreducible polynomial defining a
12292 field can be given as an 'unsigned int[]' with strictly
12293 decreasing elements giving the indices of those bits that are set;
12294 i.e., p[] represents the polynomial
12295 f(t) = t^p[0] + t^p[1] + ... + t^p[k]
12297 p[0] > p[1] > ... > p[k] = 0.
12298 This applies to the following functions:
12301 BN_GF2m_mod_mul_arr
12302 BN_GF2m_mod_sqr_arr
12303 BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv]
12304 BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div]
12305 BN_GF2m_mod_exp_arr
12306 BN_GF2m_mod_sqrt_arr
12307 BN_GF2m_mod_solve_quad_arr
12311 Conversion can be performed by the following functions:
12316 bntest.c has additional tests for binary polynomial arithmetic.
12318 Two implementations for BN_GF2m_mod_div() are available.
12319 The default algorithm simply uses BN_GF2m_mod_inv() and
12320 BN_GF2m_mod_mul(). The alternative algorithm is compiled in only
12321 if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the
12322 copyright notice in crypto/bn/bn_gf2m.c before enabling it).
12324 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
12326 * Add new error code 'ERR_R_DISABLED' that can be used when some
12327 functionality is disabled at compile-time.
12329 *Douglas Stebila <douglas.stebila@sun.com>*
12331 * Change default behaviour of 'openssl asn1parse' so that more
12332 information is visible when viewing, e.g., a certificate:
12334 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump'
12335 mode the content of non-printable OCTET STRINGs is output in a
12336 style similar to INTEGERs, but with '[HEX DUMP]' prepended to
12337 avoid the appearance of a printable string.
12339 *Nils Larsch <nla@trustcenter.de>*
12341 * Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access
12343 EC_GROUP_set_asn1_flag()
12344 EC_GROUP_get_asn1_flag()
12345 EC_GROUP_set_point_conversion_form()
12346 EC_GROUP_get_point_conversion_form()
12347 These control ASN1 encoding details:
12348 - Curves (i.e., groups) are encoded explicitly unless asn1_flag
12349 has been set to OPENSSL_EC_NAMED_CURVE.
12350 - Points are encoded in uncompressed form by default; options for
12351 asn1_for are as for point2oct, namely
12352 POINT_CONVERSION_COMPRESSED
12353 POINT_CONVERSION_UNCOMPRESSED
12354 POINT_CONVERSION_HYBRID
12356 Also add 'seed' and 'seed_len' members to EC_GROUP with access
12358 EC_GROUP_set_seed()
12359 EC_GROUP_get0_seed()
12360 EC_GROUP_get_seed_len()
12361 This is used only for ASN1 purposes (so far).
12363 *Nils Larsch <nla@trustcenter.de>*
12365 * Add 'field_type' member to EC_METHOD, which holds the NID
12366 of the appropriate field type OID. The new function
12367 EC_METHOD_get_field_type() returns this value.
12369 *Nils Larsch <nla@trustcenter.de>*
12372 EC_POINT_point2bn()
12373 EC_POINT_bn2point()
12374 EC_POINT_point2hex()
12375 EC_POINT_hex2point()
12376 providing useful interfaces to EC_POINT_point2oct() and
12377 EC_POINT_oct2point().
12379 *Nils Larsch <nla@trustcenter.de>*
12381 * Change internals of the EC library so that the functions
12382 EC_GROUP_set_generator()
12383 EC_GROUP_get_generator()
12384 EC_GROUP_get_order()
12385 EC_GROUP_get_cofactor()
12386 are implemented directly in crypto/ec/ec_lib.c and not dispatched
12387 to methods, which would lead to unnecessary code duplication when
12388 adding different types of curves.
12390 *Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller*
12392 * Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM
12393 arithmetic, and such that modified wNAFs are generated
12394 (which avoid length expansion in many cases).
12398 * Add a function EC_GROUP_check_discriminant() (defined via
12399 EC_METHOD) that verifies that the curve discriminant is non-zero.
12401 Add a function EC_GROUP_check() that makes some sanity tests
12402 on a EC_GROUP, its generator and order. This includes
12403 EC_GROUP_check_discriminant().
12405 *Nils Larsch <nla@trustcenter.de>*
12407 * Add ECDSA in new directory crypto/ecdsa/.
12409 Add applications 'openssl ecparam' and 'openssl ecdsa'
12410 (these are based on 'openssl dsaparam' and 'openssl dsa').
12412 ECDSA support is also included in various other files across the
12413 library. Most notably,
12414 - 'openssl req' now has a '-newkey ecdsa:file' option;
12415 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA;
12416 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and
12417 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make
12418 them suitable for ECDSA where domain parameters must be
12419 extracted before the specific public key;
12420 - ECDSA engine support has been added.
12422 *Nils Larsch <nla@trustcenter.de>*
12424 * Include some named elliptic curves, and add OIDs from X9.62,
12425 SECG, and WAP/WTLS. Each curve can be obtained from the new
12427 EC_GROUP_new_by_curve_name(),
12428 and the list of available named curves can be obtained with
12429 EC_get_builtin_curves().
12430 Also add a 'curve_name' member to EC_GROUP objects, which can be
12432 EC_GROUP_set_curve_name()
12433 EC_GROUP_get_curve_name()
12435 *Nils Larsch <larsch@trustcenter.de, Bodo Moeller*
12437 * Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
12438 was actually never needed) and in BN_mul(). The removal in BN_mul()
12439 required a small change in bn_mul_part_recursive() and the addition
12440 of the functions bn_cmp_part_words(), bn_sub_part_words() and
12441 bn_add_part_words(), which do the same thing as bn_cmp_words(),
12442 bn_sub_words() and bn_add_words() except they take arrays with
12447 ### Changes between 0.9.7l and 0.9.7m [23 Feb 2007]
12449 * Cleanse PEM buffers before freeing them since they may contain
12452 *Benjamin Bennett <ben@psc.edu>*
12454 * Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
12455 a ciphersuite string such as "DEFAULT:RSA" cannot enable
12456 authentication-only ciphersuites.
12460 * Since AES128 and AES256 share a single mask bit in the logic of
12461 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
12462 kludge to work properly if AES128 is available and AES256 isn't.
12466 * Expand security boundary to match 1.1.1 module.
12470 * Remove redundant features: hash file source, editing of test vectors
12471 modify fipsld to use external fips_premain.c signature.
12475 * New perl script mkfipsscr.pl to create shell scripts or batch files to
12476 run algorithm test programs.
12480 * Make algorithm test programs more tolerant of whitespace.
12484 * Have SSL/TLS server implementation tolerate "mismatched" record
12485 protocol version while receiving ClientHello even if the
12486 ClientHello is fragmented. (The server can't insist on the
12487 particular protocol version it has chosen before the ServerHello
12488 message has informed the client about his choice.)
12492 * Load error codes if they are not already present instead of using a
12493 static variable. This allows them to be cleanly unloaded and reloaded.
12497 ### Changes between 0.9.7k and 0.9.7l [28 Sep 2006]
12499 * Introduce limits to prevent malicious keys being able to
12500 cause a denial of service. ([CVE-2006-2940])
12502 *Steve Henson, Bodo Moeller*
12504 * Fix ASN.1 parsing of certain invalid structures that can result
12505 in a denial of service. ([CVE-2006-2937]) [Steve Henson]
12507 * Fix buffer overflow in SSL_get_shared_ciphers() function.
12508 ([CVE-2006-3738]) [Tavis Ormandy and Will Drewry, Google Security Team]
12510 * Fix SSL client code which could crash if connecting to a
12511 malicious SSLv2 server. ([CVE-2006-4343])
12513 *Tavis Ormandy and Will Drewry, Google Security Team*
12515 * Change ciphersuite string processing so that an explicit
12516 ciphersuite selects this one ciphersuite (so that "AES256-SHA"
12517 will no longer include "AES128-SHA"), and any other similar
12518 ciphersuite (same bitmap) from *other* protocol versions (so that
12519 "RC4-MD5" will still include both the SSL 2.0 ciphersuite and the
12520 SSL 3.0/TLS 1.0 ciphersuite). This is a backport combining
12521 changes from 0.9.8b and 0.9.8d.
12525 ### Changes between 0.9.7j and 0.9.7k [05 Sep 2006]
12527 * Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
12528 ([CVE-2006-4339]) [Ben Laurie and Google Security Team]
12530 * Change the Unix randomness entropy gathering to use poll() when
12531 possible instead of select(), since the latter has some
12532 undesirable limitations.
12534 *Darryl Miles via Richard Levitte and Bodo Moeller*
12536 * Disable rogue ciphersuites:
12538 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
12539 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
12540 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
12542 The latter two were purportedly from
12543 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
12546 Also deactivate the remaining ciphersuites from
12547 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
12548 unofficial, and the ID has long expired.
12552 * Fix RSA blinding Heisenbug (problems sometimes occurred on
12553 dual-core machines) and other potential thread-safety issues.
12557 ### Changes between 0.9.7i and 0.9.7j [04 May 2006]
12559 * Adapt fipsld and the build system to link against the validated FIPS
12560 module in FIPS mode.
12564 * Fixes for VC++ 2005 build under Windows.
12568 * Add new Windows build target VC-32-GMAKE for VC++. This uses GNU make
12569 from a Windows bash shell such as MSYS. It is autodetected from the
12570 "config" script when run from a VC++ environment. Modify standard VC++
12571 build to use fipscanister.o from the GNU make build.
12575 ### Changes between 0.9.7h and 0.9.7i [14 Oct 2005]
12577 * Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS.
12578 The value now differs depending on if you build for FIPS or not.
12579 BEWARE! A program linked with a shared FIPSed libcrypto can't be
12580 safely run with a non-FIPSed libcrypto, as it may crash because of
12581 the difference induced by this change.
12585 ### Changes between 0.9.7g and 0.9.7h [11 Oct 2005]
12587 * Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
12588 (part of SSL_OP_ALL). This option used to disable the
12589 countermeasure against man-in-the-middle protocol-version
12590 rollback in the SSL 2.0 server implementation, which is a bad
12591 idea. ([CVE-2005-2969])
12593 *Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
12594 for Information Security, National Institute of Advanced Industrial
12595 Science and Technology [AIST, Japan)]*
12597 * Minimal support for X9.31 signatures and PSS padding modes. This is
12598 mainly for FIPS compliance and not fully integrated at this stage.
12602 * For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform
12603 the exponentiation using a fixed-length exponent. (Otherwise,
12604 the information leaked through timing could expose the secret key
12605 after many signatures; cf. Bleichenbacher's attack on DSA with
12610 * Make a new fixed-window mod_exp implementation the default for
12611 RSA, DSA, and DH private-key operations so that the sequence of
12612 squares and multiplies and the memory access pattern are
12613 independent of the particular secret key. This will mitigate
12614 cache-timing and potential related attacks.
12616 BN_mod_exp_mont_consttime() is the new exponentiation implementation,
12617 and this is automatically used by BN_mod_exp_mont() if the new flag
12618 BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH
12619 will use this BN flag for private exponents unless the flag
12620 RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or
12621 DH_FLAG_NO_EXP_CONSTTIME, respectively, is set.
12623 *Matthew D Wood (Intel Corp), with some changes by Bodo Moeller*
12625 * Change the client implementation for SSLv23_method() and
12626 SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0
12627 Client Hello message format if the SSL_OP_NO_SSLv2 option is set.
12628 (Previously, the SSL 2.0 backwards compatible Client Hello
12629 message format would be used even with SSL_OP_NO_SSLv2.)
12633 * Add support for smime-type MIME parameter in S/MIME messages which some
12638 * New function BN_MONT_CTX_set_locked() to set montgomery parameters in
12639 a threadsafe manner. Modify rsa code to use new function and add calls
12640 to dsa and dh code (which had race conditions before).
12644 * Include the fixed error library code in the C error file definitions
12645 instead of fixing them up at runtime. This keeps the error code
12646 structures constant.
12650 ### Changes between 0.9.7f and 0.9.7g [11 Apr 2005]
12652 [NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after
12655 * Fixes for newer kerberos headers. NB: the casts are needed because
12656 the 'length' field is signed on one version and unsigned on another
12657 with no (?) obvious way to tell the difference, without these VC++
12658 complains. Also the "definition" of FAR (blank) is no longer included
12659 nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up
12660 some needed definitions.
12664 * Undo Cygwin change.
12668 * Added support for proxy certificates according to RFC 3820.
12669 Because they may be a security thread to unaware applications,
12670 they must be explicitly allowed in run-time. See
12671 docs/HOWTO/proxy_certificates.txt for further information.
12675 ### Changes between 0.9.7e and 0.9.7f [22 Mar 2005]
12677 * Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating
12678 server and client random values. Previously
12679 (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in
12680 less random data when sizeof(time_t) > 4 (some 64 bit platforms).
12682 This change has negligible security impact because:
12684 1. Server and client random values still have 24 bytes of pseudo random
12687 2. Server and client random values are sent in the clear in the initial
12690 3. The master secret is derived using the premaster secret (48 bytes in
12691 size for static RSA ciphersuites) as well as client server and random
12694 The OpenSSL team would like to thank the UK NISCC for bringing this issue
12697 *Stephen Henson, reported by UK NISCC*
12699 * Use Windows randomness collection on Cygwin.
12703 * Fix hang in EGD/PRNGD query when communication socket is closed
12704 prematurely by EGD/PRNGD.
12706 *Darren Tucker <dtucker@zip.com.au> via Lutz Jänicke, resolves #1014*
12708 * Prompt for pass phrases when appropriate for PKCS12 input format.
12712 * Back-port of selected performance improvements from development
12713 branch, as well as improved support for PowerPC platforms.
12717 * Add lots of checks for memory allocation failure, error codes to indicate
12718 failure and freeing up memory if a failure occurs.
12720 *Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson*
12722 * Add new -passin argument to dgst.
12726 * Perform some character comparisons of different types in X509_NAME_cmp:
12727 this is needed for some certificates that re-encode DNs into UTF8Strings
12728 (in violation of RFC3280) and can't or won't issue name rollover
12733 * Make an explicit check during certificate validation to see that
12734 the CA setting in each certificate on the chain is correct. As a
12735 side effect always do the following basic checks on extensions,
12736 not just when there's an associated purpose to the check:
12738 - if there is an unhandled critical extension (unless the user
12739 has chosen to ignore this fault)
12740 - if the path length has been exceeded (if one is set at all)
12741 - that certain extensions fit the associated purpose (if one has
12746 ### Changes between 0.9.7d and 0.9.7e [25 Oct 2004]
12748 * Avoid a race condition when CRLs are checked in a multi threaded
12749 environment. This would happen due to the reordering of the revoked
12750 entries during signature checking and serial number lookup. Now the
12751 encoding is cached and the serial number sort performed under a lock.
12752 Add new STACK function sk_is_sorted().
12756 * Add Delta CRL to the extension code.
12760 * Various fixes to s3_pkt.c so alerts are sent properly.
12762 *David Holmes <d.holmes@f5.com>*
12764 * Reduce the chances of duplicate issuer name and serial numbers (in
12765 violation of RFC3280) using the OpenSSL certificate creation utilities.
12766 This is done by creating a random 64 bit value for the initial serial
12767 number when a serial number file is created or when a self signed
12768 certificate is created using 'openssl req -x509'. The initial serial
12769 number file is created using 'openssl x509 -next_serial' in CA.pl
12770 rather than being initialized to 1.
12774 ### Changes between 0.9.7c and 0.9.7d [17 Mar 2004]
12776 * Fix null-pointer assignment in do_change_cipher_spec() revealed
12777 by using the Codenomicon TLS Test Tool ([CVE-2004-0079])
12779 *Joe Orton, Steve Henson*
12781 * Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites
12784 *Joe Orton, Steve Henson*
12786 * Make it possible to have multiple active certificates with the same
12787 subject in the CA index file. This is done only if the keyword
12788 'unique_subject' is set to 'no' in the main CA section (default
12789 if 'CA_default') of the configuration file. The value is saved
12790 with the database itself in a separate index attribute file,
12791 named like the index file with '.attr' appended to the name.
12795 * X509 verify fixes. Disable broken certificate workarounds when
12796 X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if
12797 keyUsage extension present. Don't accept CRLs with unhandled critical
12798 extensions: since verify currently doesn't process CRL extensions this
12799 rejects a CRL with *any* critical extensions. Add new verify error codes
12804 * When creating an OCSP nonce use an OCTET STRING inside the extnValue.
12805 A clarification of RFC2560 will require the use of OCTET STRINGs and
12806 some implementations cannot handle the current raw format. Since OpenSSL
12807 copies and compares OCSP nonces as opaque blobs without any attempt at
12808 parsing them this should not create any compatibility issues.
12812 * New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when
12813 calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without
12814 this HMAC (and other) operations are several times slower than OpenSSL
12819 * Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex().
12821 *Peter Sylvester <Peter.Sylvester@EdelWeb.fr>*
12823 * Use the correct content when signing type "other".
12827 ### Changes between 0.9.7b and 0.9.7c [30 Sep 2003]
12829 * Fix various bugs revealed by running the NISCC test suite:
12831 Stop out of bounds reads in the ASN1 code when presented with
12832 invalid tags (CVE-2003-0543 and CVE-2003-0544).
12834 Free up ASN1_TYPE correctly if ANY type is invalid ([CVE-2003-0545]).
12836 If verify callback ignores invalid public key errors don't try to check
12837 certificate signature with the NULL public key.
12841 * New -ignore_err option in ocsp application to stop the server
12842 exiting on the first error in a request.
12846 * In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
12847 if the server requested one: as stated in TLS 1.0 and SSL 3.0
12852 * In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
12853 extra data after the compression methods not only for TLS 1.0
12854 but also for SSL 3.0 (as required by the specification).
12856 *Bodo Moeller; problem pointed out by Matthias Loepfe*
12858 * Change X509_certificate_type() to mark the key as exported/exportable
12859 when it's 512 *bits* long, not 512 bytes.
12863 * Change AES_cbc_encrypt() so it outputs exact multiple of
12864 blocks during encryption.
12868 * Various fixes to base64 BIO and non blocking I/O. On write
12869 flushes were not handled properly if the BIO retried. On read
12870 data was not being buffered properly and had various logic bugs.
12871 This also affects blocking I/O when the data being decoded is a
12876 * Various S/MIME bugfixes and compatibility changes:
12877 output correct application/pkcs7 MIME type if
12878 PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures.
12879 Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening
12880 of files as .eml work). Correctly handle very long lines in MIME
12885 ### Changes between 0.9.7a and 0.9.7b [10 Apr 2003]
12887 * Countermeasure against the Klima-Pokorny-Rosa extension of
12888 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
12889 a protocol version number mismatch like a decryption error
12890 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
12894 * Turn on RSA blinding by default in the default implementation
12895 to avoid a timing attack. Applications that don't want it can call
12896 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
12897 They would be ill-advised to do so in most cases.
12899 *Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller*
12901 * Change RSA blinding code so that it works when the PRNG is not
12902 seeded (in this case, the secret RSA exponent is abused as
12903 an unpredictable seed -- if it is not unpredictable, there
12904 is no point in blinding anyway). Make RSA blinding thread-safe
12905 by remembering the creator's thread ID in rsa->blinding and
12906 having all other threads use local one-time blinding factors
12907 (this requires more computation than sharing rsa->blinding, but
12908 avoids excessive locking; and if an RSA object is not shared
12909 between threads, blinding will still be very fast).
12913 * Fixed a typo bug that would cause ENGINE_set_default() to set an
12914 ENGINE as defaults for all supported algorithms irrespective of
12915 the 'flags' parameter. 'flags' is now honoured, so applications
12916 should make sure they are passing it correctly.
12920 * Target "mingw" now allows native Windows code to be generated in
12921 the Cygwin environment as well as with the MinGW compiler.
12925 ### Changes between 0.9.7 and 0.9.7a [19 Feb 2003]
12927 * In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
12928 via timing by performing a MAC computation even if incorrect
12929 block cipher padding has been found. This is a countermeasure
12930 against active attacks where the attacker has to distinguish
12931 between bad padding and a MAC verification error. ([CVE-2003-0078])
12933 *Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
12934 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
12935 Martin Vuagnoux (EPFL, Ilion)*
12937 * Make the no-err option work as intended. The intention with no-err
12938 is not to have the whole error stack handling routines removed from
12939 libcrypto, it's only intended to remove all the function name and
12940 reason texts, thereby removing some of the footprint that may not
12941 be interesting if those errors aren't displayed anyway.
12943 NOTE: it's still possible for any application or module to have its
12944 own set of error texts inserted. The routines are there, just not
12945 used by default when no-err is given.
12949 * Add support for FreeBSD on IA64.
12951 *dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454*
12953 * Adjust DES_cbc_cksum() so it returns the same value as the MIT
12954 Kerberos function mit_des_cbc_cksum(). Before this change,
12955 the value returned by DES_cbc_cksum() was like the one from
12956 mit_des_cbc_cksum(), except the bytes were swapped.
12958 *Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte*
12960 * Allow an application to disable the automatic SSL chain building.
12961 Before this a rather primitive chain build was always performed in
12962 ssl3_output_cert_chain(): an application had no way to send the
12963 correct chain if the automatic operation produced an incorrect result.
12965 Now the chain builder is disabled if either:
12967 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert().
12969 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set.
12971 The reasoning behind this is that an application would not want the
12972 auto chain building to take place if extra chain certificates are
12973 present and it might also want a means of sending no additional
12974 certificates (for example the chain has two certificates and the
12979 * Add the possibility to build without the ENGINE framework.
12981 *Steven Reddie <smr@essemer.com.au> via Richard Levitte*
12983 * Under Win32 gmtime() can return NULL: check return value in
12984 OPENSSL_gmtime(). Add error code for case where gmtime() fails.
12988 * DSA routines: under certain error conditions uninitialized BN objects
12989 could be freed. Solution: make sure initialization is performed early
12990 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>,
12991 Nils Larsch <nla@trustcenter.de> via PR#459)
12995 * Another fix for SSLv2 session ID handling: the session ID was incorrectly
12996 checked on reconnect on the client side, therefore session resumption
12997 could still fail with a "ssl session id is different" error. This
12998 behaviour is masked when SSL_OP_ALL is used due to
12999 SSL_OP_MICROSOFT_SESS_ID_BUG being set.
13000 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
13001 followup to PR #377.
13005 * IA-32 assembler support enhancements: unified ELF targets, support
13006 for SCO/Caldera platforms, fix for Cygwin shared build.
13010 * Add support for FreeBSD on sparc64. As a consequence, support for
13011 FreeBSD on non-x86 processors is separate from x86 processors on
13012 the config script, much like the NetBSD support.
13014 *Richard Levitte & Kris Kennaway <kris@obsecurity.org>*
13016 ### Changes between 0.9.6h and 0.9.7 [31 Dec 2002]
13018 [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after
13021 * Fix session ID handling in SSLv2 client code: the SERVER FINISHED
13022 code (06) was taken as the first octet of the session ID and the last
13023 octet was ignored consequently. As a result SSLv2 client side session
13024 caching could not have worked due to the session ID mismatch between
13026 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
13031 * Change the declaration of needed Kerberos libraries to use EX_LIBS
13032 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is
13037 * The hw_ncipher.c engine requires dynamic locks. Unfortunately, it
13038 seems that in spite of existing for more than a year, many application
13039 author have done nothing to provide the necessary callbacks, which
13040 means that this particular engine will not work properly anywhere.
13041 This is a very unfortunate situation which forces us, in the name
13042 of usability, to give the hw_ncipher.c a static lock, which is part
13044 NOTE: This is for the 0.9.7 series ONLY. This hack will never
13045 appear in 0.9.8 or later. We EXPECT application authors to have
13046 dealt properly with this when 0.9.8 is released (unless we actually
13047 make such changes in the libcrypto locking code that changes will
13048 have to be made anyway).
13052 * In asn1_d2i_read_bio() repeatedly call BIO_read() until all content
13053 octets have been read, EOF or an error occurs. Without this change
13054 some truncated ASN1 structures will not produce an error.
13058 * Disable Heimdal support, since it hasn't been fully implemented.
13059 Still give the possibility to force the use of Heimdal, but with
13060 warnings and a request that patches get sent to openssl-dev.
13064 * Add the VC-CE target, introduce the WINCE sysname, and add
13065 INSTALL.WCE and appropriate conditionals to make it build.
13067 *Steven Reddie <smr@essemer.com.au> via Richard Levitte*
13069 * Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and
13070 cygssl-x.y.z.dll, where x, y and z are the major, minor and
13071 edit numbers of the version.
13073 *Corinna Vinschen <vinschen@redhat.com> and Richard Levitte*
13075 * Introduce safe string copy and catenation functions
13076 (BUF_strlcpy() and BUF_strlcat()).
13078 *Ben Laurie (CHATS) and Richard Levitte*
13080 * Avoid using fixed-size buffers for one-line DNs.
13082 *Ben Laurie (CHATS)*
13084 * Add BUF_MEM_grow_clean() to avoid information leakage when
13085 resizing buffers containing secrets, and use where appropriate.
13087 *Ben Laurie (CHATS)*
13089 * Avoid using fixed size buffers for configuration file location.
13091 *Ben Laurie (CHATS)*
13093 * Avoid filename truncation for various CA files.
13095 *Ben Laurie (CHATS)*
13097 * Use sizeof in preference to magic numbers.
13099 *Ben Laurie (CHATS)*
13101 * Avoid filename truncation in cert requests.
13103 *Ben Laurie (CHATS)*
13105 * Add assertions to check for (supposedly impossible) buffer
13108 *Ben Laurie (CHATS)*
13110 * Don't cache truncated DNS entries in the local cache (this could
13111 potentially lead to a spoofing attack).
13113 *Ben Laurie (CHATS)*
13115 * Fix various buffers to be large enough for hex/decimal
13116 representations in a platform independent manner.
13118 *Ben Laurie (CHATS)*
13120 * Add CRYPTO_realloc_clean() to avoid information leakage when
13121 resizing buffers containing secrets, and use where appropriate.
13123 *Ben Laurie (CHATS)*
13125 * Add BIO_indent() to avoid much slightly worrying code to do
13128 *Ben Laurie (CHATS)*
13130 * Convert sprintf()/BIO_puts() to BIO_printf().
13132 *Ben Laurie (CHATS)*
13134 * buffer_gets() could terminate with the buffer only half
13137 *Ben Laurie (CHATS)*
13139 * Add assertions to prevent user-supplied crypto functions from
13140 overflowing internal buffers by having large block sizes, etc.
13142 *Ben Laurie (CHATS)*
13144 * New OPENSSL_assert() macro (similar to assert(), but enabled
13147 *Ben Laurie (CHATS)*
13149 * Eliminate unused copy of key in RC4.
13151 *Ben Laurie (CHATS)*
13153 * Eliminate unused and incorrectly sized buffers for IV in pem.h.
13155 *Ben Laurie (CHATS)*
13157 * Fix off-by-one error in EGD path.
13159 *Ben Laurie (CHATS)*
13161 * If RANDFILE path is too long, ignore instead of truncating.
13163 *Ben Laurie (CHATS)*
13165 * Eliminate unused and incorrectly sized X.509 structure
13168 *Ben Laurie (CHATS)*
13170 * Eliminate unused and dangerous function knumber().
13172 *Ben Laurie (CHATS)*
13174 * Eliminate unused and dangerous structure, KSSL_ERR.
13176 *Ben Laurie (CHATS)*
13178 * Protect against overlong session ID context length in an encoded
13179 session object. Since these are local, this does not appear to be
13182 *Ben Laurie (CHATS)*
13184 * Change from security patch (see 0.9.6e below) that did not affect
13185 the 0.9.6 release series:
13187 Remote buffer overflow in SSL3 protocol - an attacker could
13188 supply an oversized master key in Kerberos-enabled versions.
13191 *Ben Laurie (CHATS)*
13193 * Change the SSL kerb5 codes to match RFC 2712.
13197 * Make -nameopt work fully for req and add -reqopt switch.
13199 *Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson*
13201 * The "block size" for block ciphers in CFB and OFB mode should be 1.
13203 *Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>*
13205 * Make sure tests can be performed even if the corresponding algorithms
13206 have been removed entirely. This was also the last step to make
13207 OpenSSL compilable with DJGPP under all reasonable conditions.
13209 *Richard Levitte, Doug Kaufman <dkaufman@rahul.net>*
13211 * Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT
13212 to allow version independent disabling of normally unselected ciphers,
13213 which may be activated as a side-effect of selecting a single cipher.
13215 (E.g., cipher list string "RSA" enables ciphersuites that are left
13216 out of "ALL" because they do not provide symmetric encryption.
13217 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.)
13219 *Lutz Jaenicke, Bodo Moeller*
13221 * Add appropriate support for separate platform-dependent build
13222 directories. The recommended way to make a platform-dependent
13223 build directory is the following (tested on Linux), maybe with
13226 # Place yourself outside of the OpenSSL source tree. In
13227 # this example, the environment variable OPENSSL_SOURCE
13228 # is assumed to contain the absolute OpenSSL source directory.
13229 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`"
13230 cd objtree/"`uname -s`-`uname -r`-`uname -m`"
13231 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do
13232 mkdir -p `dirname $F`
13233 ln -s $OPENSSL_SOURCE/$F $F
13236 To be absolutely sure not to disturb the source tree, a "make clean"
13237 is a good thing. If it isn't successful, don't worry about it,
13238 it probably means the source directory is very clean.
13242 * Make sure any ENGINE control commands make local copies of string
13243 pointers passed to them whenever necessary. Otherwise it is possible
13244 the caller may have overwritten (or deallocated) the original string
13245 data when a later ENGINE operation tries to use the stored values.
13247 *Götz Babin-Ebell <babinebell@trustcenter.de>*
13249 * Improve diagnostics in file reading and command-line digests.
13251 *Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>*
13253 * Add AES modes CFB and OFB to the object database. Correct an
13254 error in AES-CFB decryption.
13258 * Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this
13259 allows existing EVP_CIPHER_CTX structures to be reused after
13260 calling `EVP_*Final()`. This behaviour is used by encryption
13261 BIOs and some applications. This has the side effect that
13262 applications must explicitly clean up cipher contexts with
13263 EVP_CIPHER_CTX_cleanup() or they will leak memory.
13267 * Check the values of dna and dnb in bn_mul_recursive before calling
13268 bn_mul_comba (a non zero value means the a or b arrays do not contain
13269 n2 elements) and fallback to bn_mul_normal if either is not zero.
13273 * Fix escaping of non-ASCII characters when using the -subj option
13274 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>)
13278 * Make object definitions compliant to LDAP (RFC2256): SN is the short
13279 form for "surname", serialNumber has no short form.
13280 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798;
13281 therefore remove "mail" short name for "internet 7".
13282 The OID for unique identifiers in X509 certificates is
13283 x500UniqueIdentifier, not uniqueIdentifier.
13284 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>)
13288 * Add an "init" command to the ENGINE config module and auto initialize
13289 ENGINEs. Without any "init" command the ENGINE will be initialized
13290 after all ctrl commands have been executed on it. If init=1 the
13291 ENGINE is initialized at that point (ctrls before that point are run
13292 on the uninitialized ENGINE and after on the initialized one). If
13293 init=0 then the ENGINE will not be initialized at all.
13297 * Fix the 'app_verify_callback' interface so that the user-defined
13298 argument is actually passed to the callback: In the
13299 SSL_CTX_set_cert_verify_callback() prototype, the callback
13300 declaration has been changed from
13303 int (*cb)(X509_STORE_CTX *,void *);
13304 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
13305 i=s->ctx->app_verify_callback(&ctx)
13306 has been changed into
13307 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
13309 To update applications using SSL_CTX_set_cert_verify_callback(),
13310 a dummy argument can be added to their callback functions.
13312 *D. K. Smetters <smetters@parc.xerox.com>*
13314 * Added the '4758cca' ENGINE to support IBM 4758 cards.
13316 *Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe*
13318 * Add and OPENSSL_LOAD_CONF define which will cause
13319 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
13320 This allows older applications to transparently support certain
13321 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
13322 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
13323 load the config file and OPENSSL_add_all_algorithms_conf() which will
13324 always load it have also been added.
13328 * Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
13329 Adjust NIDs and EVP layer.
13331 *Stephen Sprunk <stephen@sprunk.org> and Richard Levitte*
13333 * Config modules support in openssl utility.
13335 Most commands now load modules from the config file,
13336 though in a few (such as version) this isn't done
13337 because it couldn't be used for anything.
13339 In the case of ca and req the config file used is
13340 the same as the utility itself: that is the -config
13341 command line option can be used to specify an
13346 * Move default behaviour from OPENSSL_config(). If appname is NULL
13347 use "openssl_conf" if filename is NULL use default openssl config file.
13351 * Add an argument to OPENSSL_config() to allow the use of an alternative
13352 config section name. Add a new flag to tolerate a missing config file
13353 and move code to CONF_modules_load_file().
13357 * Support for crypto accelerator cards from Accelerated Encryption
13358 Processing, www.aep.ie. (Use engine 'aep')
13359 The support was copied from 0.9.6c [engine] and adapted/corrected
13360 to work with the new engine framework.
13362 *AEP Inc. and Richard Levitte*
13364 * Support for SureWare crypto accelerator cards from Baltimore
13365 Technologies. (Use engine 'sureware')
13366 The support was copied from 0.9.6c [engine] and adapted
13367 to work with the new engine framework.
13371 * Have the CHIL engine fork-safe (as defined by nCipher) and actually
13372 make the newer ENGINE framework commands for the CHIL engine work.
13374 *Toomas Kiisk <vix@cyber.ee> and Richard Levitte*
13376 * Make it possible to produce shared libraries on ReliantUNIX.
13378 *Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte*
13380 * Add the configuration target debug-linux-ppro.
13381 Make 'openssl rsa' use the general key loading routines
13382 implemented in `apps.c`, and make those routines able to
13383 handle the key format FORMAT_NETSCAPE and the variant
13386 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
13388 * Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
13390 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
13392 * Add -keyform to rsautl, and document -engine.
13394 *Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>*
13396 * Change BIO_new_file (crypto/bio/bss_file.c) to use new
13397 BIO_R_NO_SUCH_FILE error code rather than the generic
13398 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
13402 * Add new functions
13403 ERR_peek_last_error
13404 ERR_peek_last_error_line
13405 ERR_peek_last_error_line_data.
13406 These are similar to
13408 ERR_peek_error_line
13409 ERR_peek_error_line_data,
13410 but report on the latest error recorded rather than the first one
13411 still in the error queue.
13413 *Ben Laurie, Bodo Moeller*
13415 * default_algorithms option in ENGINE config module. This allows things
13417 default_algorithms = ALL
13418 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
13422 * Preliminary ENGINE config module.
13426 * New experimental application configuration code.
13430 * Change the AES code to follow the same name structure as all other
13431 symmetric ciphers, and behave the same way. Move everything to
13432 the directory crypto/aes, thereby obsoleting crypto/rijndael.
13434 *Stephen Sprunk <stephen@sprunk.org> and Richard Levitte*
13436 * SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
13438 *Ben Laurie and Theo de Raadt*
13440 * Add option to output public keys in req command.
13442 *Massimiliano Pala madwolf@openca.org*
13444 * Use wNAFs in EC_POINTs_mul() for improved efficiency
13445 (up to about 10% better than before for P-192 and P-224).
13449 * New functions/macros
13451 SSL_CTX_set_msg_callback(ctx, cb)
13452 SSL_CTX_set_msg_callback_arg(ctx, arg)
13453 SSL_set_msg_callback(ssl, cb)
13454 SSL_set_msg_callback_arg(ssl, arg)
13456 to request calling a callback function
13458 void cb(int write_p, int version, int content_type,
13459 const void *buf, size_t len, SSL *ssl, void *arg)
13461 whenever a protocol message has been completely received
13462 (write_p == 0) or sent (write_p == 1). Here 'version' is the
13463 protocol version according to which the SSL library interprets
13464 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
13465 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
13466 the content type as defined in the SSL 3.0/TLS 1.0 protocol
13467 specification (change_cipher_spec(20), alert(21), handshake(22)).
13468 'buf' and 'len' point to the actual message, 'ssl' to the
13469 SSL object, and 'arg' is the application-defined value set by
13470 SSL[_CTX]_set_msg_callback_arg().
13472 'openssl s_client' and 'openssl s_server' have new '-msg' options
13473 to enable a callback that displays all protocol messages.
13477 * Change the shared library support so shared libraries are built as
13478 soon as the corresponding static library is finished, and thereby get
13479 openssl and the test programs linked against the shared library.
13480 This still only happens when the keyword "shard" has been given to
13481 the configuration scripts.
13483 NOTE: shared library support is still an experimental thing, and
13484 backward binary compatibility is still not guaranteed.
13486 *"Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte*
13488 * Add support for Subject Information Access extension.
13490 *Peter Sylvester <Peter.Sylvester@EdelWeb.fr>*
13492 * Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
13493 additional bytes when new memory had to be allocated, not just
13494 when reusing an existing buffer.
13498 * New command line and configuration option 'utf8' for the req command.
13499 This allows field values to be specified as UTF8 strings.
13503 * Add -multi and -mr options to "openssl speed" - giving multiple parallel
13504 runs for the former and machine-readable output for the latter.
13508 * Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
13509 of the e-mail address in the DN (i.e., it will go into a certificate
13510 extension only). The new configuration file option 'email_in_dn = no'
13511 has the same effect.
13513 *Massimiliano Pala madwolf@openca.org*
13515 * Change all functions with names starting with `des_` to be starting
13516 with `DES_` instead. Add wrappers that are compatible with libdes,
13517 but are named `_ossl_old_des_*`. Finally, add macros that map the
13518 `des_*` symbols to the corresponding `_ossl_old_des_*` if libdes
13519 compatibility is desired. If OpenSSL 0.9.6c compatibility is
13520 desired, the `des_*` symbols will be mapped to `DES_*`, with one
13523 Since we provide two compatibility mappings, the user needs to
13524 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
13525 compatibility is desired. The default (i.e., when that macro
13526 isn't defined) is OpenSSL 0.9.6c compatibility.
13528 There are also macros that enable and disable the support of old
13529 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
13530 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
13531 are defined, the default will apply: to support the old des routines.
13533 In either case, one must include openssl/des.h to get the correct
13534 definitions. Do not try to just include openssl/des_old.h, that
13537 NOTE: This is a major break of an old API into a new one. Software
13538 authors are encouraged to switch to the `DES_` style functions. Some
13539 time in the future, des_old.h and the libdes compatibility functions
13540 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
13541 default), and then completely removed.
13545 * Test for certificates which contain unsupported critical extensions.
13546 If such a certificate is found during a verify operation it is
13547 rejected by default: this behaviour can be overridden by either
13548 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
13549 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
13550 X509_supported_extension() has also been added which returns 1 if a
13551 particular extension is supported.
13555 * Modify the behaviour of EVP cipher functions in similar way to digests
13556 to retain compatibility with existing code.
13560 * Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
13561 compatibility with existing code. In particular the 'ctx' parameter does
13562 not have to be to be initialized before the call to EVP_DigestInit() and
13563 it is tidied up after a call to EVP_DigestFinal(). New function
13564 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
13565 EVP_MD_CTX_copy() changed to not require the destination to be
13566 initialized valid and new function EVP_MD_CTX_copy_ex() added which
13567 requires the destination to be valid.
13569 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
13570 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
13574 * Change ssl3_get_message (ssl/s3_both.c) and the functions using it
13575 so that complete 'Handshake' protocol structures are kept in memory
13576 instead of overwriting 'msg_type' and 'length' with 'body' data.
13580 * Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
13582 *Massimo Santin via Richard Levitte*
13584 * Major restructuring to the underlying ENGINE code. This includes
13585 reduction of linker bloat, separation of pure "ENGINE" manipulation
13586 (initialisation, etc) from functionality dealing with implementations
13587 of specific crypto interfaces. This change also introduces integrated
13588 support for symmetric ciphers and digest implementations - so ENGINEs
13589 can now accelerate these by providing EVP_CIPHER and EVP_MD
13590 implementations of their own. This is detailed in
13591 [crypto/engine/README.md](crypto/engine/README.md)
13592 as it couldn't be adequately described here. However, there are a few
13593 API changes worth noting - some RSA, DSA, DH, and RAND functions that
13594 were changed in the original introduction of ENGINE code have now
13595 reverted back - the hooking from this code to ENGINE is now a good
13596 deal more passive and at run-time, operations deal directly with
13597 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
13598 dereferencing through an ENGINE pointer any more. Also, the ENGINE
13599 functions dealing with `BN_MOD_EXP[_CRT]` handlers have been removed -
13600 they were not being used by the framework as there is no concept of a
13601 BIGNUM_METHOD and they could not be generalised to the new
13602 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
13603 ENGINE_cpy() has been removed as it cannot be consistently defined in
13608 * Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
13612 * Change mkdef.pl to sort symbols that get the same entry number,
13613 and make sure the automatically generated functions `ERR_load_*`
13614 become part of libeay.num as well.
13618 * New function SSL_renegotiate_pending(). This returns true once
13619 renegotiation has been requested (either SSL_renegotiate() call
13620 or HelloRequest/ClientHello received from the peer) and becomes
13621 false once a handshake has been completed.
13622 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
13623 sends a HelloRequest, but does not ensure that a handshake takes
13624 place. SSL_renegotiate_pending() is useful for checking if the
13625 client has followed the request.)
13629 * New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
13630 By default, clients may request session resumption even during
13631 renegotiation (if session ID contexts permit); with this option,
13632 session resumption is possible only in the first handshake.
13634 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes
13635 more bits available for options that should not be part of
13636 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION).
13640 * Add some demos for certificate and certificate request creation.
13644 * Make maximum certificate chain size accepted from the peer application
13645 settable (`SSL*_get/set_max_cert_list()`), as proposed by
13646 "Douglas E. Engert" <deengert@anl.gov>.
13650 * Add support for shared libraries for Unixware-7
13651 (Boyd Lynn Gerber <gerberb@zenez.com>).
13655 * Add a "destroy" handler to ENGINEs that allows structural cleanup to
13656 be done prior to destruction. Use this to unload error strings from
13657 ENGINEs that load their own error strings. NB: This adds two new API
13658 functions to "get" and "set" this destroy handler in an ENGINE.
13662 * Alter all existing ENGINE implementations (except "openssl" and
13663 "openbsd") to dynamically instantiate their own error strings. This
13664 makes them more flexible to be built both as statically-linked ENGINEs
13665 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
13666 Also, add stub code to each that makes building them as self-contained
13667 shared-libraries easier (see [README-Engine.md](README-Engine.md)).
13671 * Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
13672 implementations into applications that are completely implemented in
13673 self-contained shared-libraries. The "dynamic" ENGINE exposes control
13674 commands that can be used to configure what shared-library to load and
13675 to control aspects of the way it is handled. Also, made an update to
13676 the [README-Engine.md](README-Engine.md) file
13677 that brings its information up-to-date and
13678 provides some information and instructions on the "dynamic" ENGINE
13679 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
13683 * Make it possible to unload ranges of ERR strings with a new
13684 "ERR_unload_strings" function.
13688 * Add a copy() function to EVP_MD.
13692 * Make EVP_MD routines take a context pointer instead of just the
13693 md_data void pointer.
13697 * Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
13698 that the digest can only process a single chunk of data
13699 (typically because it is provided by a piece of
13700 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
13701 is only going to provide a single chunk of data, and hence the
13702 framework needn't accumulate the data for oneshot drivers.
13706 * As with "ERR", make it possible to replace the underlying "ex_data"
13707 functions. This change also alters the storage and management of global
13708 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
13709 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
13710 index counters. The API functions that use this state have been changed
13711 to take a "class_index" rather than pointers to the class's local STACK
13712 and counter, and there is now an API function to dynamically create new
13713 classes. This centralisation allows us to (a) plug a lot of the
13714 thread-safety problems that existed, and (b) makes it possible to clean
13715 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
13716 such data would previously have always leaked in application code and
13717 workarounds were in place to make the memory debugging turn a blind eye
13718 to it. Application code that doesn't use this new function will still
13719 leak as before, but their memory debugging output will announce it now
13720 rather than letting it slide.
13722 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
13723 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
13724 has a return value to indicate success or failure.
13728 * Make it possible to replace the underlying "ERR" functions such that the
13729 global state (2 LHASH tables and 2 locks) is only used by the "default"
13730 implementation. This change also adds two functions to "get" and "set"
13731 the implementation prior to it being automatically set the first time
13732 any other ERR function takes place. Ie. an application can call "get",
13733 pass the return value to a module it has just loaded, and that module
13734 can call its own "set" function using that value. This means the
13735 module's "ERR" operations will use (and modify) the error state in the
13736 application and not in its own statically linked copy of OpenSSL code.
13740 * Give DH, DSA, and RSA types their own `*_up_ref()` function to increment
13741 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
13742 the operation, and provides a more encapsulated way for external code
13743 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
13744 to use these functions rather than manually incrementing the counts.
13746 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
13750 * Add EVP test program.
13754 * Add symmetric cipher support to ENGINE. Expect the API to change!
13758 * New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
13759 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
13760 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
13761 These allow a CRL to be built without having to access X509_CRL fields
13762 directly. Modify 'ca' application to use new functions.
13766 * Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
13767 bug workarounds. Rollback attack detection is a security feature.
13768 The problem will only arise on OpenSSL servers when TLSv1 is not
13769 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
13770 Software authors not wanting to support TLSv1 will have special reasons
13771 for their choice and can explicitly enable this option.
13773 *Bodo Moeller, Lutz Jaenicke*
13775 * Rationalise EVP so it can be extended: don't include a union of
13776 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX
13777 (similar to those existing for EVP_CIPHER_CTX).
13782 EVP_MD_CTX_init(&md); /* new function call */
13783 EVP_DigestInit(&md, EVP_sha1());
13784 EVP_DigestUpdate(&md, in, len);
13785 EVP_DigestFinal(&md, out, NULL);
13786 EVP_MD_CTX_cleanup(&md); /* new function call */
13790 * Make DES key schedule conform to the usual scheme, as well as
13791 correcting its structure. This means that calls to DES functions
13792 now have to pass a pointer to a des_key_schedule instead of a
13793 plain des_key_schedule (which was actually always a pointer
13796 des_key_schedule ks;
13798 des_set_key_checked(..., &ks);
13799 des_ncbc_encrypt(..., &ks, ...);
13801 (Note that a later change renames 'des_...' into 'DES_...'.)
13805 * Initial reduction of linker bloat: the use of some functions, such as
13806 PEM causes large amounts of unused functions to be linked in due to
13807 poor organisation. For example pem_all.c contains every PEM function
13808 which has a knock on effect of linking in large amounts of (unused)
13809 ASN1 code. Grouping together similar functions and splitting unrelated
13810 functions prevents this.
13814 * Cleanup of EVP macros.
13818 * Change historical references to `{NID,SN,LN}_des_ede` and ede3 to add the
13819 correct `_ecb suffix`.
13823 * Add initial OCSP responder support to ocsp application. The
13824 revocation information is handled using the text based index
13825 use by the ca application. The responder can either handle
13826 requests generated internally, supplied in files (for example
13827 via a CGI script) or using an internal minimal server.
13831 * Add configuration choices to get zlib compression for TLS.
13835 * Changes to Kerberos SSL for RFC 2712 compliance:
13836 1. Implemented real KerberosWrapper, instead of just using
13837 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
13838 2. Implemented optional authenticator field of KerberosWrapper.
13840 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
13841 and authenticator structs; see crypto/krb5/.
13843 Generalized Kerberos calls to support multiple Kerberos libraries.
13844 *Vern Staats <staatsvr@asc.hpc.mil>, Jeffrey Altman <jaltman@columbia.edu>
13845 via Richard Levitte*
13847 * Cause 'openssl speed' to use fully hard-coded DSA keys as it
13848 already does with RSA. testdsa.h now has 'priv_key/pub_key'
13849 values for each of the key sizes rather than having just
13850 parameters (and 'speed' generating keys each time).
13854 * Speed up EVP routines.
13857 pe 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
13858 s-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
13859 s-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
13860 s-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
13862 s-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
13863 s-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
13864 s-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
13867 s-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
13869 s-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
13873 * Added the OS2-EMX target.
13875 *"Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte*
13877 * Rewrite commands to use `NCONF` routines instead of the old `CONF`.
13878 New functions to support `NCONF` routines in extension code.
13879 New function `CONF_set_nconf()`
13880 to allow functions which take an `NCONF` to also handle the old `LHASH`
13881 structure: this means that the old `CONF` compatible routines can be
13882 retained (in particular w.rt. extensions) without having to duplicate the
13883 code. New function `X509V3_add_ext_nconf_sk()` to add extensions to a stack.
13887 * Enhance the general user interface with mechanisms for inner control
13888 and with possibilities to have yes/no kind of prompts.
13892 * Change all calls to low-level digest routines in the library and
13893 applications to use EVP. Add missing calls to HMAC_cleanup() and
13894 don't assume HMAC_CTX can be copied using memcpy().
13896 *Verdon Walker <VWalker@novell.com>, Steve Henson*
13898 * Add the possibility to control engines through control names but with
13899 arbitrary arguments instead of just a string.
13900 Change the key loaders to take a UI_METHOD instead of a callback
13901 function pointer. NOTE: this breaks binary compatibility with earlier
13902 versions of OpenSSL [engine].
13903 Adapt the nCipher code for these new conditions and add a card insertion
13908 * Enhance the general user interface with mechanisms to better support
13909 dialog box interfaces, application-defined prompts, the possibility
13910 to use defaults (for example default passwords from somewhere else)
13911 and interrupts/cancellations.
13915 * Tidy up PKCS#12 attribute handling. Add support for the CSP name
13916 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
13920 * Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
13921 tidy up some unnecessarily weird code in 'sk_new()').
13923 *Geoff, reported by Diego Tartara <dtartara@novamens.com>*
13925 * Change the key loading routines for ENGINEs to use the same kind
13926 callback (pem_password_cb) as all other routines that need this
13931 * Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
13932 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
13933 than this minimum value is recommended.
13937 * New random seeder for OpenVMS, using the system process statistics
13938 that are easily reachable.
13942 * Windows apparently can't transparently handle global
13943 variables defined in DLLs. Initialisations such as:
13945 const ASN1_ITEM *it = &ASN1_INTEGER_it;
13947 won't compile. This is used by the any applications that need to
13948 declare their own ASN1 modules. This was fixed by adding the option
13949 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
13950 needed for static libraries under Win32.
13954 * New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
13955 setting of purpose and trust fields. New X509_STORE trust and
13956 purpose functions and tidy up setting in other SSL functions.
13960 * Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
13961 structure. These are inherited by X509_STORE_CTX when it is
13962 initialised. This allows various defaults to be set in the
13963 X509_STORE structure (such as flags for CRL checking and custom
13964 purpose or trust settings) for functions which only use X509_STORE_CTX
13965 internally such as S/MIME.
13967 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
13968 trust settings if they are not set in X509_STORE. This allows X509_STORE
13969 purposes and trust (in S/MIME for example) to override any set by default.
13971 Add command line options for CRL checking to smime, s_client and s_server
13976 * Initial CRL based revocation checking. If the CRL checking flag(s)
13977 are set then the CRL is looked up in the X509_STORE structure and
13978 its validity and signature checked, then if the certificate is found
13979 in the CRL the verify fails with a revoked error.
13981 Various new CRL related callbacks added to X509_STORE_CTX structure.
13983 Command line options added to 'verify' application to support this.
13985 This needs some additional work, such as being able to handle multiple
13986 CRLs with different times, extension based lookup (rather than just
13987 by subject name) and ultimately more complete V2 CRL extension
13992 * Add a general user interface API (crypto/ui/). This is designed
13993 to replace things like des_read_password and friends (backward
13994 compatibility functions using this new API are provided).
13995 The purpose is to remove prompting functions from the DES code
13996 section as well as provide for prompting through dialog boxes in
13997 a window system and the like.
14001 * Add "ex_data" support to ENGINE so implementations can add state at a
14002 per-structure level rather than having to store it globally.
14006 * Make it possible for ENGINE structures to be copied when retrieved by
14007 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
14008 This causes the "original" ENGINE structure to act like a template,
14009 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
14010 operational state can be localised to each ENGINE structure, despite the
14011 fact they all share the same "methods". New ENGINE structures returned in
14012 this case have no functional references and the return value is the single
14013 structural reference. This matches the single structural reference returned
14014 by ENGINE_by_id() normally, when it is incremented on the pre-existing
14019 * Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
14020 needs to match any other type at all we need to manually clear the
14025 * Changes to the "openssl engine" utility to include;
14026 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
14027 about an ENGINE's available control commands.
14028 - executing control commands from command line arguments using the
14029 '-pre' and '-post' switches. '-post' is only used if '-t' is
14030 specified and the ENGINE is successfully initialised. The syntax for
14031 the individual commands are colon-separated, for example;
14032 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
14036 * New dynamic control command support for ENGINEs. ENGINEs can now
14037 declare their own commands (numbers), names (strings), descriptions,
14038 and input types for run-time discovery by calling applications. A
14039 subset of these commands are implicitly classed as "executable"
14040 depending on their input type, and only these can be invoked through
14041 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
14042 can be based on user input, config files, etc). The distinction is
14043 that "executable" commands cannot return anything other than a boolean
14044 result and can only support numeric or string input, whereas some
14045 discoverable commands may only be for direct use through
14046 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
14047 pointers, or other custom uses. The "executable" commands are to
14048 support parameterisations of ENGINE behaviour that can be
14049 unambiguously defined by ENGINEs and used consistently across any
14050 OpenSSL-based application. Commands have been added to all the
14051 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
14052 control over shared-library paths without source code alterations.
14056 * Changed all ENGINE implementations to dynamically allocate their
14057 ENGINEs rather than declaring them statically. Apart from this being
14058 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
14059 this also allows the implementations to compile without using the
14060 internal engine_int.h header.
14064 * Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
14065 'const' value. Any code that should be able to modify a RAND_METHOD
14066 should already have non-const pointers to it (ie. they should only
14067 modify their own ones).
14071 * Made a variety of little tweaks to the ENGINE code.
14072 - "atalla" and "ubsec" string definitions were moved from header files
14073 to C code. "nuron" string definitions were placed in variables
14074 rather than hard-coded - allowing parameterisation of these values
14075 later on via ctrl() commands.
14076 - Removed unused "#if 0"'d code.
14077 - Fixed engine list iteration code so it uses ENGINE_free() to release
14078 structural references.
14079 - Constified the RAND_METHOD element of ENGINE structures.
14080 - Constified various get/set functions as appropriate and added
14081 missing functions (including a catch-all ENGINE_cpy that duplicates
14082 all ENGINE values onto a new ENGINE except reference counts/state).
14083 - Removed NULL parameter checks in get/set functions. Setting a method
14084 or function to NULL is a way of cancelling out a previously set
14085 value. Passing a NULL ENGINE parameter is just plain stupid anyway
14086 and doesn't justify the extra error symbols and code.
14087 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
14088 flags from engine_int.h to engine.h.
14089 - Changed prototypes for ENGINE handler functions (init(), finish(),
14090 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
14094 * Implement binary inversion algorithm for BN_mod_inverse in addition
14095 to the algorithm using long division. The binary algorithm can be
14096 used only if the modulus is odd. On 32-bit systems, it is faster
14097 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
14098 roughly 5-15% for 256-bit moduli), so we use it only for moduli
14099 up to 450 bits. In 64-bit environments, the binary algorithm
14100 appears to be advantageous for much longer moduli; here we use it
14101 for moduli up to 2048 bits.
14105 * Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
14106 could not support the combine flag in choice fields.
14110 * Add a 'copy_extensions' option to the 'ca' utility. This copies
14111 extensions from a certificate request to the certificate.
14115 * Allow multiple 'certopt' and 'nameopt' options to be separated
14116 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
14117 file: this allows the display of the certificate about to be
14118 signed to be customised, to allow certain fields to be included
14119 or excluded and extension details. The old system didn't display
14120 multicharacter strings properly, omitted fields not in the policy
14121 and couldn't display additional details such as extensions.
14125 * Function EC_POINTs_mul for multiple scalar multiplication
14126 of an arbitrary number of elliptic curve points
14127 \sum scalars[i]*points[i],
14128 optionally including the generator defined for the EC_GROUP:
14129 scalar*generator + \sum scalars[i]*points[i].
14131 EC_POINT_mul is a simple wrapper function for the typical case
14132 that the point list has just one item (besides the optional
14137 * First EC_METHODs for curves over GF(p):
14139 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
14140 operations and provides various method functions that can also
14141 operate with faster implementations of modular arithmetic.
14143 EC_GFp_mont_method() reuses most functions that are part of
14144 EC_GFp_simple_method, but uses Montgomery arithmetic.
14146 *Bodo Moeller; point addition and point doubling
14147 implementation directly derived from source code provided by
14148 Lenka Fibikova <fibikova@exp-math.uni-essen.de>*
14150 * Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
14151 crypto/ec/ec_lib.c):
14153 Curves are EC_GROUP objects (with an optional group generator)
14154 based on EC_METHODs that are built into the library.
14156 Points are EC_POINT objects based on EC_GROUP objects.
14158 Most of the framework would be able to handle curves over arbitrary
14159 finite fields, but as there are no obvious types for fields other
14160 than GF(p), some functions are limited to that for now.
14164 * Add the -HTTP option to s_server. It is similar to -WWW, but requires
14165 that the file contains a complete HTTP response.
14169 * Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
14170 change the def and num file printf format specifier from "%-40sXXX"
14171 to "%-39s XXX". The latter will always guarantee a space after the
14172 field while the former will cause them to run together if the field
14173 is 40 of more characters long.
14177 * Constify the cipher and digest 'method' functions and structures
14178 and modify related functions to take constant EVP_MD and EVP_CIPHER
14183 * Hide BN_CTX structure details in bn_lcl.h instead of publishing them
14184 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
14188 * Modify `EVP_Digest*()` routines so they now return values. Although the
14189 internal software routines can never fail additional hardware versions
14194 * Clean up crypto/err/err.h and change some error codes to avoid conflicts:
14196 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
14197 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
14200 ERR_R_NESTED_ASN1_ERROR
14202 ERR_R_MISSING_ASN1_EOS
14203 were 4 .. 9, conflicting with
14204 ERR_LIB_RSA (= ERR_R_RSA_LIB)
14206 ERR_LIB_PEM (= ERR_R_PEM_LIB).
14207 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
14209 Add new error code 'ERR_R_INTERNAL_ERROR'.
14213 * Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
14218 * New option '-subj arg' for 'openssl req' and 'openssl ca'. This
14219 sets the subject name for a new request or supersedes the
14220 subject name in a given request. Formats that can be parsed are
14221 'CN=Some Name, OU=myOU, C=IT'
14223 'CN=Some Name/OU=myOU/C=IT'.
14225 Add options '-batch' and '-verbose' to 'openssl req'.
14227 *Massimiliano Pala <madwolf@hackmasters.net>*
14229 * Introduce the possibility to access global variables through
14230 functions on platform were that's the best way to handle exporting
14231 global variables in shared libraries. To enable this functionality,
14232 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
14233 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
14234 is normally done by Configure or something similar).
14236 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
14237 in the source file (foo.c) like this:
14239 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
14240 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
14242 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
14243 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
14245 OPENSSL_DECLARE_GLOBAL(int,foo);
14246 #define foo OPENSSL_GLOBAL_REF(foo)
14247 OPENSSL_DECLARE_GLOBAL(double,bar);
14248 #define bar OPENSSL_GLOBAL_REF(bar)
14250 The #defines are very important, and therefore so is including the
14251 header file everywhere where the defined globals are used.
14253 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
14254 of ASN.1 items, but that structure is a bit different.
14256 The largest change is in util/mkdef.pl which has been enhanced with
14257 better and easier to understand logic to choose which symbols should
14258 go into the Windows .def files as well as a number of fixes and code
14259 cleanup (among others, algorithm keywords are now sorted
14260 lexicographically to avoid constant rewrites).
14264 * In BN_div() keep a copy of the sign of 'num' before writing the
14265 result to 'rm' because if rm==num the value will be overwritten
14266 and produce the wrong result if 'num' is negative: this caused
14267 problems with BN_mod() and BN_nnmod().
14271 * Function OCSP_request_verify(). This checks the signature on an
14272 OCSP request and verifies the signer certificate. The signer
14273 certificate is just checked for a generic purpose and OCSP request
14278 * Add OCSP_check_validity() function to check the validity of OCSP
14279 responses. OCSP responses are prepared in real time and may only
14280 be a few seconds old. Simply checking that the current time lies
14281 between thisUpdate and nextUpdate max reject otherwise valid responses
14282 caused by either OCSP responder or client clock inaccuracy. Instead
14283 we allow thisUpdate and nextUpdate to fall within a certain period of
14284 the current time. The age of the response can also optionally be
14285 checked. Two new options -validity_period and -status_age added to
14290 * If signature or public key algorithm is unrecognized print out its
14291 OID rather that just UNKNOWN.
14295 * Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
14296 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
14297 ID to be generated from the issuer certificate alone which can then be
14298 passed to OCSP_id_issuer_cmp().
14302 * New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
14303 ASN1 modules to export functions returning ASN1_ITEM pointers
14304 instead of the ASN1_ITEM structures themselves. This adds several
14305 new macros which allow the underlying ASN1 function/structure to
14306 be accessed transparently. As a result code should not use ASN1_ITEM
14307 references directly (such as &X509_it) but instead use the relevant
14308 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
14309 use of the new ASN1 code on platforms where exporting structures
14310 is problematical (for example in shared libraries) but exporting
14311 functions returning pointers to structures is not.
14315 * Add support for overriding the generation of SSL/TLS session IDs.
14316 These callbacks can be registered either in an SSL_CTX or per SSL.
14317 The purpose of this is to allow applications to control, if they wish,
14318 the arbitrary values chosen for use as session IDs, particularly as it
14319 can be useful for session caching in multiple-server environments. A
14320 command-line switch for testing this (and any client code that wishes
14321 to use such a feature) has been added to "s_server".
14323 *Geoff Thorpe, Lutz Jaenicke*
14325 * Modify mkdef.pl to recognise and parse preprocessor conditionals
14326 of the form `#if defined(...) || defined(...) || ...` and
14327 `#if !defined(...) && !defined(...) && ...`. This also avoids
14328 the growing number of special cases it was previously handling.
14332 * Make all configuration macros available for application by making
14333 sure they are available in opensslconf.h, by giving them names starting
14334 with `OPENSSL_` to avoid conflicts with other packages and by making
14335 sure e_os2.h will cover all platform-specific cases together with
14337 Additionally, it is now possible to define configuration/platform-
14338 specific names (called "system identities"). In the C code, these
14339 are prefixed with `OPENSSL_SYSNAME_`. e_os2.h will create another
14340 macro with the name beginning with `OPENSSL_SYS_`, which is determined
14341 from `OPENSSL_SYSNAME_*` or compiler-specific macros depending on
14346 * New option -set_serial to 'req' and 'x509' this allows the serial
14347 number to use to be specified on the command line. Previously self
14348 signed certificates were hard coded with serial number 0 and the
14349 CA options of 'x509' had to use a serial number in a file which was
14354 * New options to 'ca' utility to support V2 CRL entry extensions.
14355 Currently CRL reason, invalidity date and hold instruction are
14356 supported. Add new CRL extensions to V3 code and some new objects.
14360 * New function EVP_CIPHER_CTX_set_padding() this is used to
14361 disable standard block padding (aka PKCS#5 padding) in the EVP
14362 API, which was previously mandatory. This means that the data is
14363 not padded in any way and so the total length much be a multiple
14364 of the block size, otherwise an error occurs.
14368 * Initial (incomplete) OCSP SSL support.
14372 * New function OCSP_parse_url(). This splits up a URL into its host,
14373 port and path components: primarily to parse OCSP URLs. New -url
14374 option to ocsp utility.
14378 * New nonce behavior. The return value of OCSP_check_nonce() now
14379 reflects the various checks performed. Applications can decide
14380 whether to tolerate certain situations such as an absent nonce
14381 in a response when one was present in a request: the ocsp application
14382 just prints out a warning. New function OCSP_add1_basic_nonce()
14383 this is to allow responders to include a nonce in a response even if
14384 the request is nonce-less.
14388 * Disable stdin buffering in `load_cert()` (`apps/apps.c`) so that no certs are
14389 skipped when using openssl x509 multiple times on a single input file,
14390 e.g. `(openssl x509 -out cert1; openssl x509 -out cert2) <certs`.
14394 * Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
14395 set string type: to handle setting ASN1_TIME structures. Fix ca
14396 utility to correctly initialize revocation date of CRLs.
14400 * New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
14401 the clients preferred ciphersuites and rather use its own preferences.
14402 Should help to work around M$ SGC (Server Gated Cryptography) bug in
14403 Internet Explorer by ensuring unchanged hash method during stepup.
14404 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
14408 * Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
14409 to aes and add a new 'exist' option to print out symbols that don't
14414 * Additional options to ocsp utility to allow flags to be set and
14415 additional certificates supplied.
14419 * Add the option -VAfile to 'openssl ocsp', so the user can give the
14420 OCSP client a number of certificate to only verify the response
14425 * Update Rijndael code to version 3.0 and change EVP AES ciphers to
14426 handle the new API. Currently only ECB, CBC modes supported. Add new
14429 Add TLS AES ciphersuites as described in RFC3268, "Advanced
14430 Encryption Standard (AES) Ciphersuites for Transport Layer
14431 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were
14432 not enabled by default and were not part of the "ALL" ciphersuite
14433 alias because they were not yet official; they could be
14434 explicitly requested by specifying the "AESdraft" ciphersuite
14435 group alias. In the final release of OpenSSL 0.9.7, the group
14436 alias is called "AES" and is part of "ALL".)
14438 *Ben Laurie, Steve Henson, Bodo Moeller*
14440 * New function OCSP_copy_nonce() to copy nonce value (if present) from
14441 request to response.
14445 * Functions for OCSP responders. OCSP_request_onereq_count(),
14446 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
14447 extract information from a certificate request. OCSP_response_create()
14448 creates a response and optionally adds a basic response structure.
14449 OCSP_basic_add1_status() adds a complete single response to a basic
14450 response and returns the OCSP_SINGLERESP structure just added (to allow
14451 extensions to be included for example). OCSP_basic_add1_cert() adds a
14452 certificate to a basic response and OCSP_basic_sign() signs a basic
14453 response with various flags. New helper functions ASN1_TIME_check()
14454 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
14455 (converts ASN1_TIME to GeneralizedTime).
14459 * Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
14460 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
14461 structure from a certificate. X509_pubkey_digest() digests the public_key
14462 contents: this is used in various key identifiers.
14466 * Make sk_sort() tolerate a NULL argument.
14468 *Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>*
14470 * New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
14471 passed by the function are trusted implicitly. If any of them signed the
14472 response then it is assumed to be valid and is not verified.
14476 * In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
14477 to data. This was previously part of the PKCS7 ASN1 code. This
14478 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
14479 *Steve Henson, reported by Kenneth R. Robinette
14480 <support@securenetterm.com>*
14482 * Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
14483 routines: without these tracing memory leaks is very painful.
14484 Fix leaks in PKCS12 and PKCS7 routines.
14488 * Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
14489 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
14490 effectively meant GeneralizedTime would never be used. Now it
14491 is initialised to -1 but X509_time_adj() now has to check the value
14492 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
14493 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
14494 *Steve Henson, reported by Kenneth R. Robinette
14495 <support@securenetterm.com>*
14497 * Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
14498 result in a zero length in the ASN1_INTEGER structure which was
14499 not consistent with the structure when d2i_ASN1_INTEGER() was used
14500 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
14501 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
14502 where it did not print out a minus for negative ASN1_INTEGER.
14506 * Add summary printout to ocsp utility. The various functions which
14507 convert status values to strings have been renamed to:
14508 OCSP_response_status_str(), OCSP_cert_status_str() and
14509 OCSP_crl_reason_str() and are no longer static. New options
14510 to verify nonce values and to disable verification. OCSP response
14511 printout format cleaned up.
14515 * Add additional OCSP certificate checks. These are those specified
14516 in RFC2560. This consists of two separate checks: the CA of the
14517 certificate being checked must either be the OCSP signer certificate
14518 or the issuer of the OCSP signer certificate. In the latter case the
14519 OCSP signer certificate must contain the OCSP signing extended key
14520 usage. This check is performed by attempting to match the OCSP
14521 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
14522 in the OCSP_CERTID structures of the response.
14526 * Initial OCSP certificate verification added to OCSP_basic_verify()
14527 and related routines. This uses the standard OpenSSL certificate
14528 verify routines to perform initial checks (just CA validity) and
14529 to obtain the certificate chain. Then additional checks will be
14530 performed on the chain. Currently the root CA is checked to see
14531 if it is explicitly trusted for OCSP signing. This is used to set
14532 a root CA as a global signing root: that is any certificate that
14533 chains to that CA is an acceptable OCSP signing certificate.
14537 * New '-extfile ...' option to 'openssl ca' for reading X.509v3
14538 extensions from a separate configuration file.
14539 As when reading extensions from the main configuration file,
14540 the '-extensions ...' option may be used for specifying the
14543 *Massimiliano Pala <madwolf@comune.modena.it>*
14545 * New OCSP utility. Allows OCSP requests to be generated or
14546 read. The request can be sent to a responder and the output
14547 parsed, outputted or printed in text form. Not complete yet:
14548 still needs to check the OCSP response validity.
14552 * New subcommands for 'openssl ca':
14553 `openssl ca -status <serial>` prints the status of the cert with
14554 the given serial number (according to the index file).
14555 `openssl ca -updatedb` updates the expiry status of certificates
14558 *Massimiliano Pala <madwolf@comune.modena.it>*
14560 * New '-newreq-nodes' command option to CA.pl. This is like
14561 '-newreq', but calls 'openssl req' with the '-nodes' option
14562 so that the resulting key is not encrypted.
14564 *Damien Miller <djm@mindrot.org>*
14566 * New configuration for the GNU Hurd.
14568 *Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte*
14570 * Initial code to implement OCSP basic response verify. This
14571 is currently incomplete. Currently just finds the signer's
14572 certificate and verifies the signature on the response.
14576 * New SSLeay_version code SSLEAY_DIR to determine the compiled-in
14577 value of OPENSSLDIR. This is available via the new '-d' option
14578 to 'openssl version', and is also included in 'openssl version -a'.
14582 * Allowing defining memory allocation callbacks that will be given
14583 file name and line number information in additional arguments
14584 (a `const char*` and an int). The basic functionality remains, as
14585 well as the original possibility to just replace malloc(),
14586 realloc() and free() by functions that do not know about these
14587 additional arguments. To register and find out the current
14588 settings for extended allocation functions, the following
14589 functions are provided:
14591 CRYPTO_set_mem_ex_functions
14592 CRYPTO_set_locked_mem_ex_functions
14593 CRYPTO_get_mem_ex_functions
14594 CRYPTO_get_locked_mem_ex_functions
14596 These work the same way as CRYPTO_set_mem_functions and friends.
14597 `CRYPTO_get_[locked_]mem_functions` now writes 0 where such an
14598 extended allocation function is enabled.
14599 Similarly, `CRYPTO_get_[locked_]mem_ex_functions` writes 0 where
14600 a conventional allocation function is enabled.
14602 *Richard Levitte, Bodo Moeller*
14604 * Finish off removing the remaining LHASH function pointer casts.
14605 There should no longer be any prototype-casting required when using
14606 the LHASH abstraction, and any casts that remain are "bugs". See
14607 the callback types and macros at the head of lhash.h for details
14608 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
14612 * Add automatic query of EGD sockets in RAND_poll() for the unix variant.
14613 If /dev/[u]random devices are not available or do not return enough
14614 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
14616 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
14617 /etc/entropy will be queried once each in this sequence, querying stops
14618 when enough entropy was collected without querying more sockets.
14622 * Change the Unix RAND_poll() variant to be able to poll several
14623 random devices, as specified by DEVRANDOM, until a sufficient amount
14624 of data has been collected. We spend at most 10 ms on each file
14625 (select timeout) and read in non-blocking mode. DEVRANDOM now
14626 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
14627 (previously it was just the string "/dev/urandom"), so on typical
14628 platforms the 10 ms delay will never occur.
14629 Also separate out the Unix variant to its own file, rand_unix.c.
14630 For VMS, there's a currently-empty rand_vms.c.
14634 * Move OCSP client related routines to ocsp_cl.c. These
14635 provide utility functions which an application needing
14636 to issue a request to an OCSP responder and analyse the
14637 response will typically need: as opposed to those which an
14638 OCSP responder itself would need which will be added later.
14640 OCSP_request_sign() signs an OCSP request with an API similar
14641 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
14642 response. OCSP_response_get1_basic() extracts basic response
14643 from response. OCSP_resp_find_status(): finds and extracts status
14644 information from an OCSP_CERTID structure (which will be created
14645 when the request structure is built). These are built from lower
14646 level functions which work on OCSP_SINGLERESP structures but
14647 won't normally be used unless the application wishes to examine
14648 extensions in the OCSP response for example.
14650 Replace nonce routines with a pair of functions.
14651 OCSP_request_add1_nonce() adds a nonce value and optionally
14652 generates a random value. OCSP_check_nonce() checks the
14653 validity of the nonce in an OCSP response.
14657 * Change function OCSP_request_add() to OCSP_request_add0_id().
14658 This doesn't copy the supplied OCSP_CERTID and avoids the
14659 need to free up the newly created id. Change return type
14660 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
14661 This can then be used to add extensions to the request.
14662 Deleted OCSP_request_new(), since most of its functionality
14663 is now in OCSP_REQUEST_new() (and the case insensitive name
14664 clash) apart from the ability to set the request name which
14665 will be added elsewhere.
14669 * Update OCSP API. Remove obsolete extensions argument from
14670 various functions. Extensions are now handled using the new
14671 OCSP extension code. New simple OCSP HTTP function which
14672 can be used to send requests and parse the response.
14676 * Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
14677 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
14678 uses the special reorder version of SET OF to sort the attributes
14679 and reorder them to match the encoded order. This resolves a long
14680 standing problem: a verify on a PKCS7 structure just after signing
14681 it used to fail because the attribute order did not match the
14682 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
14683 it uses the received order. This is necessary to tolerate some broken
14684 software that does not order SET OF. This is handled by encoding
14685 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
14686 to produce the required SET OF.
14690 * Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
14691 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
14692 files to get correct declarations of the ASN.1 item variables.
14696 * Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
14697 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
14698 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
14699 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
14700 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
14701 ASN1_ITEM and no wrapper functions.
14705 * New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
14706 replace the old function pointer based I/O routines. Change most of
14707 the `*_d2i_bio()` and `*_d2i_fp()` functions to use these.
14711 * Enhance mkdef.pl to be more accepting about spacing in C preprocessor
14712 lines, recognize more "algorithms" that can be deselected, and make
14713 it complain about algorithm deselection that isn't recognised.
14717 * New ASN1 functions to handle dup, sign, verify, digest, pack and
14718 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
14719 to use new functions. Add NO_ASN1_OLD which can be set to remove
14720 some old style ASN1 functions: this can be used to determine if old
14721 code will still work when these eventually go away.
14725 * New extension functions for OCSP structures, these follow the
14726 same conventions as certificates and CRLs.
14730 * New function X509V3_add1_i2d(). This automatically encodes and
14731 adds an extension. Its behaviour can be customised with various
14732 flags to append, replace or delete. Various wrappers added for
14733 certificates and CRLs.
14737 * Fix to avoid calling the underlying ASN1 print routine when
14738 an extension cannot be parsed. Correct a typo in the
14739 OCSP_SERVICELOC extension. Tidy up print OCSP format.
14743 * Make mkdef.pl parse some of the ASN1 macros and add appropriate
14744 entries for variables.
14748 * Add functionality to `apps/openssl.c` for detecting locking
14749 problems: As the program is single-threaded, all we have
14750 to do is register a locking callback using an array for
14751 storing which locks are currently held by the program.
14755 * Use a lock around the call to CRYPTO_get_ex_new_index() in
14756 SSL_get_ex_data_X509_STORE_idx(), which is used in
14757 ssl_verify_cert_chain() and thus can be called at any time
14758 during TLS/SSL handshakes so that thread-safety is essential.
14759 Unfortunately, the ex_data design is not at all suited
14760 for multi-threaded use, so it probably should be abolished.
14764 * Added Broadcom "ubsec" ENGINE to OpenSSL.
14766 *Broadcom, tweaked and integrated by Geoff Thorpe*
14768 * Move common extension printing code to new function
14769 X509V3_print_extensions(). Reorganise OCSP print routines and
14770 implement some needed OCSP ASN1 functions. Add OCSP extensions.
14774 * New function X509_signature_print() to remove duplication in some
14779 * Add a special meaning when SET OF and SEQUENCE OF flags are both
14780 set (this was treated exactly the same as SET OF previously). This
14781 is used to reorder the STACK representing the structure to match the
14782 encoding. This will be used to get round a problem where a PKCS7
14783 structure which was signed could not be verified because the STACK
14784 order did not reflect the encoded order.
14788 * Reimplement the OCSP ASN1 module using the new code.
14792 * Update the X509V3 code to permit the use of an ASN1_ITEM structure
14793 for its ASN1 operations. The old style function pointers still exist
14794 for now but they will eventually go away.
14798 * Merge in replacement ASN1 code from the ASN1 branch. This almost
14799 completely replaces the old ASN1 functionality with a table driven
14800 encoder and decoder which interprets an ASN1_ITEM structure describing
14801 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
14802 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
14803 has also been converted to the new form.
14807 * Change BN_mod_exp_recp so that negative moduli are tolerated
14808 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
14809 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
14810 for negative moduli.
14814 * Fix BN_uadd and BN_usub: Always return non-negative results instead
14815 of not touching the result's sign bit.
14819 * BN_div bugfix: If the result is 0, the sign (res->neg) must not be
14824 * Changed the LHASH code to use prototypes for callbacks, and created
14825 macros to declare and implement thin (optionally static) functions
14826 that provide type-safety and avoid function pointer casting for the
14827 type-specific callbacks.
14831 * Added Kerberos Cipher Suites to be used with TLS, as written in
14833 *Veers Staats <staatsvr@asc.hpc.mil>,
14834 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte*
14836 * Reformat the FAQ so the different questions and answers can be divided
14837 in sections depending on the subject.
14841 * Have the zlib compression code load ZLIB.DLL dynamically under
14846 * New function BN_mod_sqrt for computing square roots modulo a prime
14847 (using the probabilistic Tonelli-Shanks algorithm unless
14848 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
14849 be handled deterministically).
14851 *Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller*
14853 * Make BN_mod_inverse faster by explicitly handling small quotients
14854 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
14855 512 bits], about 30% for larger ones [1024 or 2048 bits].)
14859 * New function BN_kronecker.
14863 * Fix BN_gcd so that it works on negative inputs; the result is
14864 positive unless both parameters are zero.
14865 Previously something reasonably close to an infinite loop was
14866 possible because numbers could be growing instead of shrinking
14867 in the implementation of Euclid's algorithm.
14871 * Fix BN_is_word() and BN_is_one() macros to take into account the
14872 sign of the number in question.
14874 Fix BN_is_word(a,w) to work correctly for w == 0.
14876 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
14877 because its test if the absolute value of 'a' equals 'w'.
14878 Note that BN_abs_is_word does *not* handle w == 0 reliably;
14879 it exists mostly for use in the implementations of BN_is_zero(),
14880 BN_is_one(), and BN_is_word().
14884 * New function BN_swap.
14888 * Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
14889 the exponentiation functions are more likely to produce reasonable
14890 results on negative inputs.
14894 * Change BN_mod_mul so that the result is always non-negative.
14895 Previously, it could be negative if one of the factors was negative;
14896 I don't think anyone really wanted that behaviour.
14900 * Move `BN_mod_...` functions into new file `crypto/bn/bn_mod.c`
14901 (except for exponentiation, which stays in `crypto/bn/bn_exp.c`,
14902 and `BN_mod_mul_reciprocal`, which stays in `crypto/bn/bn_recp.c`)
14903 and add new functions:
14912 BN_mod_lshift1_quick
14914 BN_mod_lshift_quick
14916 These functions always generate non-negative results.
14918 `BN_nnmod` otherwise is `like BN_mod` (if `BN_mod` computes a remainder `r`
14919 such that `|m| < r < 0`, `BN_nnmod` will output `rem + |m|` instead).
14921 `BN_mod_XXX_quick(r, a, [b,] m)` generates the same result as
14922 `BN_mod_XXX(r, a, [b,] m, ctx)`, but requires that `a` [and `b`]
14923 be reduced modulo `m`.
14925 *Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller*
14928 The following entry accidentally appeared in the CHANGES file
14929 distributed with OpenSSL 0.9.7. The modifications described in
14930 it do *not* apply to OpenSSL 0.9.7.
14932 * Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
14933 was actually never needed) and in BN_mul(). The removal in BN_mul()
14934 required a small change in bn_mul_part_recursive() and the addition
14935 of the functions bn_cmp_part_words(), bn_sub_part_words() and
14936 bn_add_part_words(), which do the same thing as bn_cmp_words(),
14937 bn_sub_words() and bn_add_words() except they take arrays with
14943 * In 'openssl passwd', verify passwords read from the terminal
14944 unless the '-salt' option is used (which usually means that
14945 verification would just waste user's time since the resulting
14946 hash is going to be compared with some given password hash)
14947 or the new '-noverify' option is used.
14949 This is an incompatible change, but it does not affect
14950 non-interactive use of 'openssl passwd' (passwords on the command
14951 line, '-stdin' option, '-in ...' option) and thus should not
14952 cause any problems.
14956 * Remove all references to RSAref, since there's no more need for it.
14960 * Make DSO load along a path given through an environment variable
14961 (SHLIB_PATH) with shl_load().
14965 * Constify the ENGINE code as a result of BIGNUM constification.
14966 Also constify the RSA code and most things related to it. In a
14967 few places, most notable in the depth of the ASN.1 code, ugly
14968 casts back to non-const were required (to be solved at a later
14973 * Make it so the openssl application has all engines loaded by default.
14977 * Constify the BIGNUM routines a little more.
14981 * Add the following functions:
14983 ENGINE_load_cswift()
14985 ENGINE_load_atalla()
14986 ENGINE_load_nuron()
14987 ENGINE_load_builtin_engines()
14989 That way, an application can itself choose if external engines that
14990 are built-in in OpenSSL shall ever be used or not. The benefit is
14991 that applications won't have to be linked with libdl or other dso
14992 libraries unless it's really needed.
14994 Changed 'openssl engine' to load all engines on demand.
14995 Changed the engine header files to avoid the duplication of some
14996 declarations (they differed!).
15000 * 'openssl engine' can now list capabilities.
15004 * Better error reporting in 'openssl engine'.
15008 * Never call load_dh_param(NULL) in s_server.
15012 * Add engine application. It can currently list engines by name and
15013 identity, and test if they are actually available.
15017 * Improve RPM specification file by forcing symbolic linking and making
15018 sure the installed documentation is also owned by root.root.
15020 *Damien Miller <djm@mindrot.org>*
15022 * Give the OpenSSL applications more possibilities to make use of
15023 keys (public as well as private) handled by engines.
15027 * Add OCSP code that comes from CertCo.
15031 * Add VMS support for the Rijndael code.
15035 * Added untested support for Nuron crypto accelerator.
15039 * Add support for external cryptographic devices. This code was
15040 previously distributed separately as the "engine" branch.
15042 *Geoff Thorpe, Richard Levitte*
15044 * Rework the filename-translation in the DSO code. It is now possible to
15045 have far greater control over how a "name" is turned into a filename
15046 depending on the operating environment and any oddities about the
15047 different shared library filenames on each system.
15051 * Support threads on FreeBSD-elf in Configure.
15055 * Fix for SHA1 assembly problem with MASM: it produces
15056 warnings about corrupt line number information when assembling
15057 with debugging information. This is caused by the overlapping
15060 *Bernd Matthes <mainbug@celocom.de>, Steve Henson*
15063 NCONF_get_number() has no error checking at all. As a replacement,
15064 NCONF_get_number_e() is defined (`_e` for "error checking") and is
15065 promoted strongly. The old NCONF_get_number is kept around for
15066 binary backward compatibility.
15067 Make it possible for methods to load from something other than a BIO,
15068 by providing a function pointer that is given a name instead of a BIO.
15069 For example, this could be used to load configuration data from an
15074 * Fix for non blocking accept BIOs. Added new I/O special reason
15075 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
15076 with non blocking I/O was not possible because no retry code was
15077 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
15082 * Added the beginnings of Rijndael support.
15086 * Fix for bug in DirectoryString mask setting. Add support for
15087 X509_NAME_print_ex() in 'req' and X509_print_ex() function
15088 to allow certificate printing to more controllable, additional
15089 'certopt' option to 'x509' to allow new printing options to be
15094 * Clean old EAY MD5 hack from e_os.h.
15098 ### Changes between 0.9.6l and 0.9.6m [17 Mar 2004]
15100 * Fix null-pointer assignment in do_change_cipher_spec() revealed
15101 by using the Codenomicon TLS Test Tool ([CVE-2004-0079])
15103 *Joe Orton, Steve Henson*
15105 ### Changes between 0.9.6k and 0.9.6l [04 Nov 2003]
15107 * Fix additional bug revealed by the NISCC test suite:
15109 Stop bug triggering large recursion when presented with
15110 certain ASN.1 tags ([CVE-2003-0851])
15114 ### Changes between 0.9.6j and 0.9.6k [30 Sep 2003]
15116 * Fix various bugs revealed by running the NISCC test suite:
15118 Stop out of bounds reads in the ASN1 code when presented with
15119 invalid tags (CVE-2003-0543 and CVE-2003-0544).
15121 If verify callback ignores invalid public key errors don't try to check
15122 certificate signature with the NULL public key.
15126 * In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
15127 if the server requested one: as stated in TLS 1.0 and SSL 3.0
15132 * In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
15133 extra data after the compression methods not only for TLS 1.0
15134 but also for SSL 3.0 (as required by the specification).
15136 *Bodo Moeller; problem pointed out by Matthias Loepfe*
15138 * Change X509_certificate_type() to mark the key as exported/exportable
15139 when it's 512 *bits* long, not 512 bytes.
15143 ### Changes between 0.9.6i and 0.9.6j [10 Apr 2003]
15145 * Countermeasure against the Klima-Pokorny-Rosa extension of
15146 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
15147 a protocol version number mismatch like a decryption error
15148 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
15152 * Turn on RSA blinding by default in the default implementation
15153 to avoid a timing attack. Applications that don't want it can call
15154 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
15155 They would be ill-advised to do so in most cases.
15157 *Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller*
15159 * Change RSA blinding code so that it works when the PRNG is not
15160 seeded (in this case, the secret RSA exponent is abused as
15161 an unpredictable seed -- if it is not unpredictable, there
15162 is no point in blinding anyway). Make RSA blinding thread-safe
15163 by remembering the creator's thread ID in rsa->blinding and
15164 having all other threads use local one-time blinding factors
15165 (this requires more computation than sharing rsa->blinding, but
15166 avoids excessive locking; and if an RSA object is not shared
15167 between threads, blinding will still be very fast).
15171 ### Changes between 0.9.6h and 0.9.6i [19 Feb 2003]
15173 * In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
15174 via timing by performing a MAC computation even if incorrect
15175 block cipher padding has been found. This is a countermeasure
15176 against active attacks where the attacker has to distinguish
15177 between bad padding and a MAC verification error. ([CVE-2003-0078])
15179 *Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
15180 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
15181 Martin Vuagnoux (EPFL, Ilion)*
15183 ### Changes between 0.9.6g and 0.9.6h [5 Dec 2002]
15185 * New function OPENSSL_cleanse(), which is used to cleanse a section of
15186 memory from its contents. This is done with a counter that will
15187 place alternating values in each byte. This can be used to solve
15188 two issues: 1) the removal of calls to memset() by highly optimizing
15189 compilers, and 2) cleansing with other values than 0, since those can
15190 be read through on certain media, for example a swap space on disk.
15194 * Bugfix: client side session caching did not work with external caching,
15195 because the session->cipher setting was not restored when reloading
15196 from the external cache. This problem was masked, when
15197 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set.
15198 (Found by Steve Haslam <steve@araqnid.ddts.net>.)
15202 * Fix client_certificate (ssl/s2_clnt.c): The permissible total
15203 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33.
15205 *Zeev Lieber <zeev-l@yahoo.com>*
15207 * Undo an undocumented change introduced in 0.9.6e which caused
15208 repeated calls to OpenSSL_add_all_ciphers() and
15209 OpenSSL_add_all_digests() to be ignored, even after calling
15214 * Change the default configuration reader to deal with last line not
15215 being properly terminated.
15219 * Change X509_NAME_cmp() so it applies the special rules on handling
15220 DN values that are of type PrintableString, as well as RDNs of type
15221 emailAddress where the value has the type ia5String.
15223 *stefank@valicert.com via Richard Levitte*
15225 * Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half
15226 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently
15227 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be
15228 the bitwise-OR of the two for use by the majority of applications
15229 wanting this behaviour, and update the docs. The documented
15230 behaviour and actual behaviour were inconsistent and had been
15231 changing anyway, so this is more a bug-fix than a behavioural
15234 *Geoff Thorpe, diagnosed by Nadav Har'El*
15236 * Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c
15237 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes).
15241 * Fix initialization code race conditions in
15242 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(),
15243 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(),
15244 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(),
15245 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(),
15246 ssl2_get_cipher_by_char(),
15247 ssl3_get_cipher_by_char().
15249 *Patrick McCormick <patrick@tellme.com>, Bodo Moeller*
15251 * Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after
15252 the cached sessions are flushed, as the remove_cb() might use ex_data
15253 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com>
15254 (see [openssl.org #212]).
15256 *Geoff Thorpe, Lutz Jaenicke*
15258 * Fix typo in OBJ_txt2obj which incorrectly passed the content
15259 length, instead of the encoding length to d2i_ASN1_OBJECT.
15263 ### Changes between 0.9.6f and 0.9.6g [9 Aug 2002]
15265 * [In 0.9.6g-engine release:]
15266 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use `_stdcall`).
15268 *Lynn Gazis <lgazis@rainbow.com>*
15270 ### Changes between 0.9.6e and 0.9.6f [8 Aug 2002]
15272 * Fix ASN1 checks. Check for overflow by comparing with LONG_MAX
15273 and get fix the header length calculation.
15274 *Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>,
15275 Alon Kantor <alonk@checkpoint.com> (and others), Steve Henson*
15277 * Use proper error handling instead of 'assertions' in buffer
15278 overflow checks added in 0.9.6e. This prevents DoS (the
15279 assertions could call abort()).
15281 *Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller*
15283 ### Changes between 0.9.6d and 0.9.6e [30 Jul 2002]
15285 * Add various sanity checks to asn1_get_length() to reject
15286 the ASN1 length bytes if they exceed sizeof(long), will appear
15287 negative or the content length exceeds the length of the
15290 *Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>*
15292 * Fix cipher selection routines: ciphers without encryption had no flags
15293 for the cipher strength set and where therefore not handled correctly
15294 by the selection routines (PR #130).
15298 * Fix EVP_dsa_sha macro.
15303 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS
15304 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure
15305 that was added in OpenSSL 0.9.6d.
15307 As the countermeasure turned out to be incompatible with some
15308 broken SSL implementations, the new option is part of SSL_OP_ALL.
15309 SSL_OP_ALL is usually employed when compatibility with weird SSL
15310 implementations is desired (e.g. '-bugs' option to 's_client' and
15311 's_server'), so the new option is automatically set in many
15316 * Changes in security patch:
15318 Changes marked "(CHATS)" were sponsored by the Defense Advanced
15319 Research Projects Agency (DARPA) and Air Force Research Laboratory,
15320 Air Force Materiel Command, USAF, under agreement number
15323 * Add various sanity checks to asn1_get_length() to reject
15324 the ASN1 length bytes if they exceed sizeof(long), will appear
15325 negative or the content length exceeds the length of the
15326 supplied buffer. ([CVE-2002-0659])
15328 *Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>*
15330 * Assertions for various potential buffer overflows, not known to
15331 happen in practice.
15333 *Ben Laurie (CHATS)*
15335 * Various temporary buffers to hold ASCII versions of integers were
15336 too small for 64 bit platforms. ([CVE-2002-0655])
15337 *Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>*
15339 * Remote buffer overflow in SSL3 protocol - an attacker could
15340 supply an oversized session ID to a client. ([CVE-2002-0656])
15342 *Ben Laurie (CHATS)*
15344 * Remote buffer overflow in SSL2 protocol - an attacker could
15345 supply an oversized client master key. ([CVE-2002-0656])
15347 *Ben Laurie (CHATS)*
15349 ### Changes between 0.9.6c and 0.9.6d [9 May 2002]
15351 * Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
15352 encoded as NULL) with id-dsa-with-sha1.
15354 *Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller*
15356 * Check various `X509_...()` return values in `apps/req.c`.
15358 *Nils Larsch <nla@trustcenter.de>*
15360 * Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
15361 an end-of-file condition would erroneously be flagged, when the CRLF
15362 was just at the end of a processed block. The bug was discovered when
15363 processing data through a buffering memory BIO handing the data to a
15364 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
15365 <ptsekov@syntrex.com> and Nedelcho Stanev.
15369 * Implement a countermeasure against a vulnerability recently found
15370 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
15371 before application data chunks to avoid the use of known IVs
15372 with data potentially chosen by the attacker.
15376 * Fix length checks in ssl3_get_client_hello().
15380 * TLS/SSL library bugfix: use s->s3->in_read_app_data differently
15381 to prevent ssl3_read_internal() from incorrectly assuming that
15382 ssl3_read_bytes() found application data while handshake
15383 processing was enabled when in fact s->s3->in_read_app_data was
15384 merely automatically cleared during the initial handshake.
15386 *Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>*
15388 * Fix object definitions for Private and Enterprise: they were not
15389 recognized in their shortname (=lowercase) representation. Extend
15390 obj_dat.pl to issue an error when using undefined keywords instead
15391 of silently ignoring the problem (Svenning Sorensen
15392 <sss@sss.dnsalias.net>).
15396 * Fix DH_generate_parameters() so that it works for 'non-standard'
15397 generators, i.e. generators other than 2 and 5. (Previously, the
15398 code did not properly initialise the 'add' and 'rem' values to
15399 BN_generate_prime().)
15401 In the new general case, we do not insist that 'generator' is
15402 actually a primitive root: This requirement is rather pointless;
15403 a generator of the order-q subgroup is just as good, if not
15408 * Map new X509 verification errors to alerts. Discovered and submitted by
15409 Tom Wu <tom@arcot.com>.
15413 * Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
15414 returning non-zero before the data has been completely received
15415 when using non-blocking I/O.
15417 *Bodo Moeller; problem pointed out by John Hughes*
15419 * Some of the ciphers missed the strength entry (SSL_LOW etc).
15421 *Ben Laurie, Lutz Jaenicke*
15423 * Fix bug in SSL_clear(): bad sessions were not removed (found by
15424 Yoram Zahavi <YoramZ@gilian.com>).
15428 * Add information about CygWin 1.3 and on, and preserve proper
15429 configuration for the versions before that.
15431 *Corinna Vinschen <vinschen@redhat.com> and Richard Levitte*
15433 * Make removal from session cache (SSL_CTX_remove_session()) more robust:
15434 check whether we deal with a copy of a session and do not delete from
15435 the cache in this case. Problem reported by "Izhar Shoshani Levi"
15436 <izhar@checkpoint.com>.
15440 * Do not store session data into the internal session cache, if it
15441 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
15442 flag is set). Proposed by Aslam <aslam@funk.com>.
15446 * Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
15451 * [In 0.9.6d-engine release:]
15452 Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
15454 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
15456 * Add the configuration target linux-s390x.
15458 *Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte*
15460 * The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
15461 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
15462 variable as an indication that a ClientHello message has been
15463 received. As the flag value will be lost between multiple
15464 invocations of ssl3_accept when using non-blocking I/O, the
15465 function may not be aware that a handshake has actually taken
15466 place, thus preventing a new session from being added to the
15469 To avoid this problem, we now set s->new_session to 2 instead of
15470 using a local variable.
15472 *Lutz Jaenicke, Bodo Moeller*
15474 * Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
15475 if the SSL_R_LENGTH_MISMATCH error is detected.
15477 *Geoff Thorpe, Bodo Moeller*
15479 * New 'shared_ldflag' column in Configure platform table.
15483 * Fix EVP_CIPHER_mode macro.
15485 *"Dan S. Camper" <dan@bti.net>*
15487 * Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
15488 type, we must throw them away by setting rr->length to 0.
15490 *D P Chang <dpc@qualys.com>*
15492 ### Changes between 0.9.6b and 0.9.6c [21 dec 2001]
15494 * Fix BN_rand_range bug pointed out by Dominikus Scherkl
15495 <Dominikus.Scherkl@biodata.com>. (The previous implementation
15496 worked incorrectly for those cases where range = `10..._2` and
15497 `3*range` is two bits longer than range.)
15501 * Only add signing time to PKCS7 structures if it is not already
15506 * Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
15507 OBJ_ld_ce should be OBJ_id_ce.
15508 Also some ip-pda OIDs in crypto/objects/objects.txt were
15509 incorrect (cf. RFC 3039).
15511 *Matt Cooper, Frederic Giudicelli, Bodo Moeller*
15513 * Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
15514 returns early because it has nothing to do.
15516 *Andy Schneider <andy.schneider@bjss.co.uk>*
15518 * [In 0.9.6c-engine release:]
15519 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
15521 *Andy Schneider <andy.schneider@bjss.co.uk>*
15523 * [In 0.9.6c-engine release:]
15524 Add support for Cryptographic Appliance's keyserver technology.
15525 (Use engine 'keyclient')
15527 *Cryptographic Appliances and Geoff Thorpe*
15529 * Add a configuration entry for OS/390 Unix. The C compiler 'c89'
15530 is called via tools/c89.sh because arguments have to be
15531 rearranged (all '-L' options must appear before the first object
15534 *Richard Shapiro <rshapiro@abinitio.com>*
15536 * [In 0.9.6c-engine release:]
15537 Add support for Broadcom crypto accelerator cards, backported
15540 *Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox*
15542 * [In 0.9.6c-engine release:]
15543 Add support for SureWare crypto accelerator cards from
15544 Baltimore Technologies. (Use engine 'sureware')
15546 *Baltimore Technologies and Mark Cox*
15548 * [In 0.9.6c-engine release:]
15549 Add support for crypto accelerator cards from Accelerated
15550 Encryption Processing, www.aep.ie. (Use engine 'aep')
15552 *AEP Inc. and Mark Cox*
15554 * Add a configuration entry for gcc on UnixWare.
15556 *Gary Benson <gbenson@redhat.com>*
15558 * Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
15559 messages are stored in a single piece (fixed-length part and
15560 variable-length part combined) and fix various bugs found on the way.
15564 * Disable caching in BIO_gethostbyname(), directly use gethostbyname()
15565 instead. BIO_gethostbyname() does not know what timeouts are
15566 appropriate, so entries would stay in cache even when they have
15568 *Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>*
15570 * Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
15571 faced with a pathologically small ClientHello fragment that does
15572 not contain client_version: Instead of aborting with an error,
15573 simply choose the highest available protocol version (i.e.,
15574 TLS 1.0 unless it is disabled). In practice, ClientHello
15575 messages are never sent like this, but this change gives us
15576 strictly correct behaviour at least for TLS.
15580 * Fix SSL handshake functions and SSL_clear() such that SSL_clear()
15581 never resets s->method to s->ctx->method when called from within
15582 one of the SSL handshake functions.
15584 *Bodo Moeller; problem pointed out by Niko Baric*
15586 * In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
15587 (sent using the client's version number) if client_version is
15588 smaller than the protocol version in use. Also change
15589 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
15590 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
15591 the client will at least see that alert.
15595 * Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
15600 * Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
15601 client receives HelloRequest while in a handshake.
15603 *Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>*
15605 * Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
15606 should end in 'break', not 'goto end' which circumvents various
15607 cleanups done in state SSL_ST_OK. But session related stuff
15608 must be disabled for SSL_ST_OK in the case that we just sent a
15611 Also avoid some overhead by not calling ssl_init_wbio_buffer()
15612 before just sending a HelloRequest.
15614 *Bodo Moeller, Eric Rescorla <ekr@rtfm.com>*
15616 * Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
15617 reveal whether illegal block cipher padding was found or a MAC
15618 verification error occurred. (Neither SSLerr() codes nor alerts
15619 are directly visible to potential attackers, but the information
15620 may leak via logfiles.)
15622 Similar changes are not required for the SSL 2.0 implementation
15623 because the number of padding bytes is sent in clear for SSL 2.0,
15624 and the extra bytes are just ignored. However ssl/s2_pkt.c
15625 failed to verify that the purported number of padding bytes is in
15630 * Add OpenUNIX-8 support including shared libraries
15631 (Boyd Lynn Gerber <gerberb@zenez.com>).
15635 * Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
15636 'wristwatch attack' using huge encoding parameters (cf.
15637 James H. Manger's CRYPTO 2001 paper). Note that the
15638 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
15639 encoding parameters and hence was not vulnerable.
15643 * BN_sqr() bug fix.
15645 *Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>*
15647 * Rabin-Miller test analyses assume uniformly distributed witnesses,
15648 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
15649 followed by modular reduction.
15651 *Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>*
15653 * Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
15654 equivalent based on BN_pseudo_rand() instead of BN_rand().
15658 * s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
15659 This function was broken, as the check for a new client hello message
15660 to handle SGC did not allow these large messages.
15661 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
15665 * Add alert descriptions for TLSv1 to `SSL_alert_desc_string[_long]()`.
15669 * Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
15670 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
15674 * Rework the configuration and shared library support for Tru64 Unix.
15675 The configuration part makes use of modern compiler features and
15676 still retains old compiler behavior for those that run older versions
15677 of the OS. The shared library support part includes a variant that
15678 uses the RPATH feature, and is available through the special
15679 configuration target "alpha-cc-rpath", which will never be selected
15682 *Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte*
15684 * In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
15685 with the same message size as in ssl3_get_certificate_request().
15686 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
15687 messages might inadvertently be reject as too long.
15689 *Petr Lampa <lampa@fee.vutbr.cz>*
15691 * Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
15695 * Modified SSL library such that the verify_callback that has been set
15696 specifically for an SSL object with SSL_set_verify() is actually being
15697 used. Before the change, a verify_callback set with this function was
15698 ignored and the verify_callback() set in the SSL_CTX at the time of
15699 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
15700 to allow the necessary settings.
15704 * Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
15705 explicitly to NULL, as at least on Solaris 8 this seems not always to be
15706 done automatically (in contradiction to the requirements of the C
15707 standard). This made problems when used from OpenSSH.
15711 * In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
15712 dh->length and always used
15714 BN_rand_range(priv_key, dh->p).
15716 BN_rand_range() is not necessary for Diffie-Hellman, and this
15717 specific range makes Diffie-Hellman unnecessarily inefficient if
15718 dh->length (recommended exponent length) is much smaller than the
15719 length of dh->p. We could use BN_rand_range() if the order of
15720 the subgroup was stored in the DH structure, but we only have
15725 BN_rand(priv_key, l, ...)
15727 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
15734 RSA_eay_public_encrypt
15735 RSA_eay_private_decrypt
15736 RSA_eay_private_encrypt (signing)
15737 RSA_eay_public_decrypt (signature verification)
15739 (default implementations for RSA_public_encrypt,
15740 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
15741 always reject numbers >= n.
15745 * In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
15746 to synchronize access to 'locking_thread'. This is necessary on
15747 systems where access to 'locking_thread' (an 'unsigned long'
15748 variable) is not atomic.
15752 * In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
15753 *before* setting the 'crypto_lock_rand' flag. The previous code had
15754 a race condition if 0 is a valid thread ID.
15756 *Travis Vitek <vitek@roguewave.com>*
15758 * Add support for shared libraries under Irix.
15760 *Albert Chin-A-Young <china@thewrittenword.com>*
15762 * Add configuration option to build on Linux on both big-endian and
15763 little-endian MIPS.
15765 *Ralf Baechle <ralf@uni-koblenz.de>*
15767 * Add the possibility to create shared libraries on HP-UX.
15771 ### Changes between 0.9.6a and 0.9.6b [9 Jul 2001]
15773 * Change ssleay_rand_bytes (crypto/rand/md_rand.c)
15774 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
15775 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
15776 PRNG state recovery was possible based on the output of
15777 one PRNG request appropriately sized to gain knowledge on
15778 'md' followed by enough consecutive 1-byte PRNG requests
15779 to traverse all of 'state'.
15781 1. When updating 'md_local' (the current thread's copy of 'md')
15782 during PRNG output generation, hash all of the previous
15783 'md_local' value, not just the half used for PRNG output.
15785 2. Make the number of bytes from 'state' included into the hash
15786 independent from the number of PRNG bytes requested.
15788 The first measure alone would be sufficient to avoid
15789 Markku-Juhani's attack. (Actually it had never occurred
15790 to me that the half of 'md_local' used for chaining was the
15791 half from which PRNG output bytes were taken -- I had always
15792 assumed that the secret half would be used.) The second
15793 measure makes sure that additional data from 'state' is never
15794 mixed into 'md_local' in small portions; this heuristically
15795 further strengthens the PRNG.
15799 * Fix crypto/bn/asm/mips3.s.
15803 * When only the key is given to "enc", the IV is undefined. Print out
15804 an error message in this case.
15808 * Handle special case when X509_NAME is empty in X509 printing routines.
15812 * In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
15813 positive and less than q.
15817 * Don't change `*pointer` in CRYPTO_add_lock() is add_lock_callback is
15818 used: it isn't thread safe and the add_lock_callback should handle
15821 *Paul Rose <Paul.Rose@bridge.com>*
15823 * Verify that incoming data obeys the block size in
15824 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
15830 *Ulf Möller, Bodo Möller*
15832 * The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
15833 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
15834 when fixing the server behaviour for backwards-compatible 'client
15835 hello' messages. (Note that the attack is impractical against
15836 SSL 3.0 and TLS 1.0 anyway because length and version checking
15837 means that the probability of guessing a valid ciphertext is
15838 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
15841 Before 0.9.5, the countermeasure (hide the error by generating a
15842 random 'decryption result') did not work properly because
15843 ERR_clear_error() was missing, meaning that SSL_get_error() would
15844 detect the supposedly ignored error.
15846 Both problems are now fixed.
15850 * In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
15851 (previously it was 1024).
15855 * Fix for compatibility mode trust settings: ignore trust settings
15856 unless some valid trust or reject settings are present.
15860 * Fix for blowfish EVP: its a variable length cipher.
15864 * Fix various bugs related to DSA S/MIME verification. Handle missing
15865 parameters in DSA public key structures and return an error in the
15866 DSA routines if parameters are absent.
15870 * In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
15871 in the current directory if neither $RANDFILE nor $HOME was set.
15872 RAND_file_name() in 0.9.6a returned NULL in this case. This has
15873 caused some confusion to Windows users who haven't defined $HOME.
15874 Thus RAND_file_name() is changed again: e_os.h can define a
15875 DEFAULT_HOME, which will be used if $HOME is not set.
15876 For Windows, we use "C:"; on other platforms, we still require
15877 environment variables.
15879 * Move 'if (!initialized) RAND_poll()' into regions protected by
15880 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
15881 having multiple threads call RAND_poll() concurrently.
15885 * In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
15886 combination of a flag and a thread ID variable.
15887 Otherwise while one thread is in ssleay_rand_bytes (which sets the
15888 flag), *other* threads can enter ssleay_add_bytes without obeying
15889 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
15890 that they do not hold after the first thread unsets add_do_not_lock).
15894 * Change bctest again: '-x' expressions are not available in all
15895 versions of 'test'.
15899 ### Changes between 0.9.6 and 0.9.6a [5 Apr 2001]
15901 * Fix a couple of memory leaks in PKCS7_dataDecode()
15903 *Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>*
15905 * Change Configure and Makefiles to provide EXE_EXT, which will contain
15906 the default extension for executables, if any. Also, make the perl
15907 scripts that use symlink() to test if it really exists and use "cp"
15908 if it doesn't. All this made OpenSSL compilable and installable in
15913 * Fix for asn1_GetSequence() for indefinite length constructed data.
15914 If SEQUENCE is length is indefinite just set c->slen to the total
15915 amount of data available.
15917 *Steve Henson, reported by shige@FreeBSD.org*
15919 *This change does not apply to 0.9.7.*
15921 * Change bctest to avoid here-documents inside command substitution
15922 (workaround for FreeBSD /bin/sh bug).
15923 For compatibility with Ultrix, avoid shell functions (introduced
15924 in the bctest version that searches along $PATH).
15928 * Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
15929 with des_encrypt() defined on some operating systems, like Solaris
15934 * Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
15935 On the Importance of Eliminating Errors in Cryptographic
15936 Computations, J. Cryptology 14 (2001) 2, 101-119,
15937 <http://theory.stanford.edu/~dabo/papers/faults.ps.gz>).
15941 * MIPS assembler BIGNUM division bug fix.
15945 * Disabled incorrect Alpha assembler code.
15949 * Fix PKCS#7 decode routines so they correctly update the length
15950 after reading an EOC for the EXPLICIT tag.
15954 *This change does not apply to 0.9.7.*
15956 * Fix bug in PKCS#12 key generation routines. This was triggered
15957 if a 3DES key was generated with a 0 initial byte. Include
15958 PKCS12_BROKEN_KEYGEN compilation option to retain the old
15959 (but broken) behaviour.
15963 * Enhance bctest to search for a working bc along $PATH and print
15966 *Tim Rice <tim@multitalents.net> via Richard Levitte*
15968 * Fix memory leaks in err.c: free err_data string if necessary;
15969 don't write to the wrong index in ERR_set_error_data.
15973 * Implement ssl23_peek (analogous to ssl23_read), which previously
15978 * Replace rdtsc with `_emit` statements for VC++ version 5.
15980 *Jeremy Cooper <jeremy@baymoo.org>*
15982 * Make it possible to reuse SSLv2 sessions.
15986 * In copy_email() check for >= 0 as a return value for
15987 X509_NAME_get_index_by_NID() since 0 is a valid index.
15989 *Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>*
15991 * Avoid coredump with unsupported or invalid public keys by checking if
15992 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
15993 PKCS7_verify() fails with non detached data.
15997 * Don't use getenv in library functions when run as setuid/setgid.
15998 New function OPENSSL_issetugid().
16002 * Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
16003 due to incorrect handling of multi-threading:
16005 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
16007 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
16009 3. Count how many times MemCheck_off() has been called so that
16010 nested use can be treated correctly. This also avoids
16011 inband-signalling in the previous code (which relied on the
16012 assumption that thread ID 0 is impossible).
16016 * Add "-rand" option also to s_client and s_server.
16020 * Fix CPU detection on Irix 6.x.
16021 *Kurt Hockenbury <khockenb@stevens-tech.edu> and
16022 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>*
16024 * Fix X509_NAME bug which produced incorrect encoding if X509_NAME
16029 *This change does not apply to 0.9.7.*
16031 * Use the cached encoding of an X509_NAME structure rather than
16032 copying it. This is apparently the reason for the libsafe "errors"
16033 but the code is actually correct.
16037 * Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
16038 Bleichenbacher's DSA attack.
16039 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
16040 to be set and top=0 forces the highest bit to be set; top=-1 is new
16041 and leaves the highest bit random.
16043 *Ulf Moeller, Bodo Moeller*
16045 * In the `NCONF_...`-based implementations for `CONF_...` queries
16046 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
16047 a temporary CONF structure with the data component set to NULL
16048 (which gives segmentation faults in lh_retrieve).
16049 Instead, use NULL for the CONF pointer in CONF_get_string and
16050 CONF_get_number (which may use environment variables) and directly
16051 return NULL from CONF_get_section.
16055 * Fix potential buffer overrun for EBCDIC.
16059 * Tolerate nonRepudiation as being valid for S/MIME signing and certSign
16060 keyUsage if basicConstraints absent for a CA.
16064 * Make SMIME_write_PKCS7() write mail header values with a format that
16065 is more generally accepted (no spaces before the semicolon), since
16066 some programs can't parse those values properly otherwise. Also make
16067 sure BIO's that break lines after each write do not create invalid
16072 * Make the CRL encoding routines work with empty SEQUENCE OF. The
16073 macros previously used would not encode an empty SEQUENCE OF
16074 and break the signature.
16078 *This change does not apply to 0.9.7.*
16080 * Zero the premaster secret after deriving the master secret in
16085 * Add some EVP_add_digest_alias registrations (as found in
16086 OpenSSL_add_all_digests()) to SSL_library_init()
16087 aka OpenSSL_add_ssl_algorithms(). This provides improved
16088 compatibility with peers using X.509 certificates
16089 with unconventional AlgorithmIdentifier OIDs.
16093 * Fix for Irix with NO_ASM.
16095 *"Bruce W. Forsberg" <bruce.forsberg@baesystems.com>*
16097 * ./config script fixes.
16099 *Ulf Moeller, Richard Levitte*
16101 * Fix 'openssl passwd -1'.
16105 * Change PKCS12_key_gen_asc() so it can cope with non null
16106 terminated strings whose length is passed in the passlen
16107 parameter, for example from PEM callbacks. This was done
16108 by adding an extra length parameter to asc2uni().
16110 *Steve Henson, reported by <oddissey@samsung.co.kr>*
16112 * Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
16113 call failed, free the DSA structure.
16117 * Fix to uni2asc() to cope with zero length Unicode strings.
16118 These are present in some PKCS#12 files.
16122 * Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
16123 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
16124 when writing a 32767 byte record.
16126 *Bodo Moeller; problem reported by Eric Day <eday@concentric.net>*
16128 * In `RSA_eay_public_{en,ed}crypt` and RSA_eay_mod_exp (rsa_eay.c),
16129 obtain lock CRYPTO_LOCK_RSA before setting `rsa->_method_mod_{n,p,q}`.
16131 (RSA objects have a reference count access to which is protected
16132 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
16133 so they are meant to be shared between threads.)
16134 *Bodo Moeller, Geoff Thorpe; original patch submitted by
16135 "Reddie, Steven" <Steven.Reddie@ca.com>*
16137 * Fix a deadlock in CRYPTO_mem_leaks().
16141 * Use better test patterns in bntest.
16145 * rand_win.c fix for Borland C.
16149 * BN_rshift bugfix for n == 0.
16153 * Add a 'bctest' script that checks for some known 'bc' bugs
16154 so that 'make test' does not abort just because 'bc' is broken.
16158 * Store verify_result within SSL_SESSION also for client side to
16159 avoid potential security hole. (Re-used sessions on the client side
16160 always resulted in verify_result==X509_V_OK, not using the original
16161 result of the server certificate verification.)
16165 * Fix ssl3_pending: If the record in s->s3->rrec is not of type
16166 SSL3_RT_APPLICATION_DATA, return 0.
16167 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
16172 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
16173 releases, have been re-implemented by renaming the previous
16174 implementations of ssl2_read and ssl3_read to ssl2_read_internal
16175 and ssl3_read_internal, respectively, and adding 'peek' parameters
16176 to them. The new ssl[23]_{read,peek} functions are calls to
16177 ssl[23]_read_internal with the 'peek' flag set appropriately.
16178 A 'peek' parameter has also been added to ssl3_read_bytes, which
16179 does the actual work for ssl3_read_internal.
16183 * Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
16184 the method-specific "init()" handler. Also clean up ex_data after
16185 calling the method-specific "finish()" handler. Previously, this was
16186 happening the other way round.
16190 * Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
16191 The previous value, 12, was not always sufficient for BN_mod_exp().
16195 * Make sure that shared libraries get the internal name engine with
16196 the full version number and not just 0. This should mark the
16197 shared libraries as not backward compatible. Of course, this should
16198 be changed again when we can guarantee backward binary compatibility.
16202 * Fix typo in get_cert_by_subject() in by_dir.c
16204 *Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>*
16206 * Rework the system to generate shared libraries:
16208 - Make note of the expected extension for the shared libraries and
16209 if there is a need for symbolic links from for example libcrypto.so.0
16210 to libcrypto.so.0.9.7. There is extended info in Configure for
16213 - Make as few rebuilds of the shared libraries as possible.
16215 - Still avoid linking the OpenSSL programs with the shared libraries.
16217 - When installing, install the shared libraries separately from the
16222 * Fix SSL_CTX_set_read_ahead macro to actually use its argument.
16224 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
16225 and not in SSL_clear because the latter is also used by the
16226 accept/connect functions; previously, the settings made by
16227 SSL_set_read_ahead would be lost during the handshake.
16229 *Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>*
16231 * Correct util/mkdef.pl to be selective about disabled algorithms.
16232 Previously, it would create entries for disabled algorithms no
16237 * Added several new manual pages for SSL_* function.
16241 ### Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
16243 * In ssl23_get_client_hello, generate an error message when faced
16244 with an initial SSL 3.0/TLS record that is too small to contain the
16245 first two bytes of the ClientHello message, i.e. client_version.
16246 (Note that this is a pathologic case that probably has never happened
16247 in real life.) The previous approach was to use the version number
16248 from the record header as a substitute; but our protocol choice
16249 should not depend on that one because it is not authenticated
16250 by the Finished messages.
16254 * More robust randomness gathering functions for Windows.
16256 *Jeffrey Altman <jaltman@columbia.edu>*
16258 * For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
16259 not set then we don't setup the error code for issuer check errors
16260 to avoid possibly overwriting other errors which the callback does
16261 handle. If an application does set the flag then we assume it knows
16262 what it is doing and can handle the new informational codes
16267 * Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
16268 a general "ANY" type, as such it should be able to decode anything
16269 including tagged types. However it didn't check the class so it would
16270 wrongly interpret tagged types in the same way as their universal
16271 counterpart and unknown types were just rejected. Changed so that the
16272 tagged and unknown types are handled in the same way as a SEQUENCE:
16273 that is the encoding is stored intact. There is also a new type
16274 "V_ASN1_OTHER" which is used when the class is not universal, in this
16275 case we have no idea what the actual type is so we just lump them all
16280 * On VMS, stdout may very well lead to a file that is written to
16281 in a record-oriented fashion. That means that every write() will
16282 write a separate record, which will be read separately by the
16283 programs trying to read from it. This can be very confusing.
16285 The solution is to put a BIO filter in the way that will buffer
16286 text until a linefeed is reached, and then write everything a
16287 line at a time, so every record written will be an actual line,
16288 not chunks of lines and not (usually doesn't happen, but I've
16289 seen it once) several lines in one record. BIO_f_linebuffer() is
16292 Currently, it's a VMS-only method, because that's where it has
16293 been tested well enough.
16297 * Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
16298 it can return incorrect results.
16299 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
16300 but it was in 0.9.6-beta[12].)
16304 * Disable the check for content being present when verifying detached
16305 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
16306 include zero length content when signing messages.
16310 * New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
16311 BIO_ctrl (for BIO pairs).
16315 * Add DSO method for VMS.
16319 * Bug fix: Montgomery multiplication could produce results with the
16324 * Add RPM specification openssl.spec and modify it to build three
16325 packages. The default package contains applications, application
16326 documentation and run-time libraries. The devel package contains
16327 include files, static libraries and function documentation. The
16328 doc package contains the contents of the doc directory. The original
16329 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
16333 * Add a large number of documentation files for many SSL routines.
16335 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>*
16337 * Add a configuration entry for Sony News 4.
16339 *NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>*
16341 * Don't set the two most significant bits to one when generating a
16342 random number < q in the DSA library.
16346 * New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
16347 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
16348 the underlying transport is blocking) if a handshake took place.
16349 (The default behaviour is needed by applications such as s_client
16350 and s_server that use select() to determine when to use SSL_read;
16351 but for applications that know in advance when to expect data, it
16352 just makes things more complicated.)
16356 * Add RAND_egd_bytes(), which gives control over the number of bytes read
16361 * Add a few more EBCDIC conditionals that make `req` and `x509`
16362 work better on such systems.
16364 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
16366 * Add two demo programs for PKCS12_parse() and PKCS12_create().
16367 Update PKCS12_parse() so it copies the friendlyName and the
16368 keyid to the certificates aux info.
16372 * Fix bug in PKCS7_verify() which caused an infinite loop
16373 if there was more than one signature.
16375 *Sven Uszpelkat <su@celocom.de>*
16377 * Major change in util/mkdef.pl to include extra information
16378 about each symbol, as well as presenting variables as well
16379 as functions. This change means that there's n more need
16380 to rebuild the .num files when some algorithms are excluded.
16384 * Allow the verify time to be set by an application,
16385 rather than always using the current time.
16389 * Phase 2 verify code reorganisation. The certificate
16390 verify code now looks up an issuer certificate by a
16391 number of criteria: subject name, authority key id
16392 and key usage. It also verifies self signed certificates
16393 by the same criteria. The main comparison function is
16394 X509_check_issued() which performs these checks.
16396 Lot of changes were necessary in order to support this
16397 without completely rewriting the lookup code.
16399 Authority and subject key identifier are now cached.
16401 The LHASH 'certs' is X509_STORE has now been replaced
16402 by a STACK_OF(X509_OBJECT). This is mainly because an
16403 LHASH can't store or retrieve multiple objects with
16404 the same hash value.
16406 As a result various functions (which were all internal
16407 use only) have changed to handle the new X509_STORE
16408 structure. This will break anything that messed round
16409 with X509_STORE internally.
16411 The functions X509_STORE_add_cert() now checks for an
16412 exact match, rather than just subject name.
16414 The X509_STORE API doesn't directly support the retrieval
16415 of multiple certificates matching a given criteria, however
16416 this can be worked round by performing a lookup first
16417 (which will fill the cache with candidate certificates)
16418 and then examining the cache for matches. This is probably
16419 the best we can do without throwing out X509_LOOKUP
16420 entirely (maybe later...).
16422 The X509_VERIFY_CTX structure has been enhanced considerably.
16424 All certificate lookup operations now go via a get_issuer()
16425 callback. Although this currently uses an X509_STORE it
16426 can be replaced by custom lookups. This is a simple way
16427 to bypass the X509_STORE hackery necessary to make this
16428 work and makes it possible to use more efficient techniques
16429 in future. A very simple version which uses a simple
16430 STACK for its trusted certificate store is also provided
16431 using X509_STORE_CTX_trusted_stack().
16433 The verify_cb() and verify() callbacks now have equivalents
16434 in the X509_STORE_CTX structure.
16436 X509_STORE_CTX also has a 'flags' field which can be used
16437 to customise the verify behaviour.
16441 * Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
16442 excludes S/MIME capabilities.
16446 * When a certificate request is read in keep a copy of the
16447 original encoding of the signed data and use it when outputting
16448 again. Signatures then use the original encoding rather than
16449 a decoded, encoded version which may cause problems if the
16450 request is improperly encoded.
16454 * For consistency with other BIO_puts implementations, call
16455 buffer_write(b, ...) directly in buffer_puts instead of calling
16458 In BIO_puts, increment b->num_write as in BIO_write.
16460 *Peter.Sylvester@EdelWeb.fr*
16462 * Fix BN_mul_word for the case where the word is 0. (We have to use
16463 BN_zero, we may not return a BIGNUM with an array consisting of
16464 words set to zero.)
16468 * Avoid calling abort() from within the library when problems are
16469 detected, except if preprocessor symbols have been defined
16470 (such as REF_CHECK, BN_DEBUG etc.).
16474 * New openssl application 'rsautl'. This utility can be
16475 used for low-level RSA operations. DER public key
16476 BIO/fp routines also added.
16480 * New Configure entry and patches for compiling on QNX 4.
16482 *Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>*
16484 * A demo state-machine implementation was sponsored by
16485 Nuron (<http://www.nuron.com/>) and is now available in
16486 demos/state_machine.
16490 * New options added to the 'dgst' utility for signature
16491 generation and verification.
16495 * Unrecognized PKCS#7 content types are now handled via a
16496 catch all ASN1_TYPE structure. This allows unsupported
16497 types to be stored as a "blob" and an application can
16498 encode and decode it manually.
16502 * Fix various signed/unsigned issues to make a_strex.c
16503 compile under VC++.
16505 *Oscar Jacobsson <oscar.jacobsson@celocom.com>*
16507 * ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
16508 length if passed a buffer. ASN1_INTEGER_to_BN failed
16509 if passed a NULL BN and its argument was negative.
16511 *Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>*
16513 * Modification to PKCS#7 encoding routines to output definite
16514 length encoding. Since currently the whole structures are in
16515 memory there's not real point in using indefinite length
16516 constructed encoding. However if OpenSSL is compiled with
16517 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
16521 * Added BIO_vprintf() and BIO_vsnprintf().
16525 * Added more prefixes to parse for in the strings written
16526 through a logging bio, to cover all the levels that are available
16527 through syslog. The prefixes are now:
16529 PANIC, EMERG, EMR => LOG_EMERG
16530 ALERT, ALR => LOG_ALERT
16531 CRIT, CRI => LOG_CRIT
16532 ERROR, ERR => LOG_ERR
16533 WARNING, WARN, WAR => LOG_WARNING
16534 NOTICE, NOTE, NOT => LOG_NOTICE
16535 INFO, INF => LOG_INFO
16536 DEBUG, DBG => LOG_DEBUG
16538 and as before, if none of those prefixes are present at the
16539 beginning of the string, LOG_ERR is chosen.
16541 On Win32, the `LOG_*` levels are mapped according to this:
16543 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
16544 LOG_WARNING => EVENTLOG_WARNING_TYPE
16545 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
16549 * Made it possible to reconfigure with just the configuration
16550 argument "reconf" or "reconfigure". The command line arguments
16551 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
16552 and are retrieved from there when reconfiguring.
16558 *Assar Westerlund <assar@sics.se>, Richard Levitte*
16560 * Add the arguments -CAfile and -CApath to the pkcs12 utility.
16564 * The obj_dat.pl script was messing up the sorting of object
16565 names. The reason was that it compared the quoted version
16566 of strings as a result "OCSP" > "OCSP Signing" because
16567 " > SPACE. Changed script to store unquoted versions of
16568 names and add quotes on output. It was also omitting some
16569 names from the lookup table if they were given a default
16570 value (that is if SN is missing it is given the same
16571 value as LN and vice versa), these are now added on the
16572 grounds that if an object has a name we should be able to
16573 look it up. Finally added warning output when duplicate
16574 short or long names are found.
16578 * Changes needed for Tandem NSK.
16580 *Scott Uroff <scott@xypro.com>*
16582 * Fix SSL 2.0 rollback checking: Due to an off-by-one error in
16583 RSA_padding_check_SSLv23(), special padding was never detected
16584 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
16585 version rollback attacks was not effective.
16587 In s23_clnt.c, don't use special rollback-attack detection padding
16588 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
16589 client; similarly, in s23_srvr.c, don't do the rollback check if
16590 SSL 2.0 is the only protocol enabled in the server.
16594 * Make it possible to get hexdumps of unprintable data with 'openssl
16595 asn1parse'. By implication, the functions ASN1_parse_dump() and
16596 BIO_dump_indent() are added.
16600 * New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
16601 these print out strings and name structures based on various
16602 flags including RFC2253 support and proper handling of
16603 multibyte characters. Added options to the 'x509' utility
16604 to allow the various flags to be set.
16608 * Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
16609 Also change the functions X509_cmp_current_time() and
16610 X509_gmtime_adj() work with an ASN1_TIME structure,
16611 this will enable certificates using GeneralizedTime in validity
16612 dates to be checked.
16616 * Make the NEG_PUBKEY_BUG code (which tolerates invalid
16617 negative public key encodings) on by default,
16618 NO_NEG_PUBKEY_BUG can be set to disable it.
16622 * New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
16623 content octets. An i2c_ASN1_OBJECT is unnecessary because
16624 the encoding can be trivially obtained from the structure.
16628 * crypto/err.c locking bugfix: Use write locks (`CRYPTO_w_[un]lock`),
16629 not read locks (`CRYPTO_r_[un]lock`).
16633 * A first attempt at creating official support for shared
16634 libraries through configuration. I've kept it so the
16635 default is static libraries only, and the OpenSSL programs
16636 are always statically linked for now, but there are
16637 preparations for dynamic linking in place.
16638 This has been tested on Linux and Tru64.
16642 * Randomness polling function for Win9x, as described in:
16643 Peter Gutmann, Software Generation of Practically Strong
16648 * Fix so PRNG is seeded in req if using an already existing
16653 * New options to smime application. -inform and -outform
16654 allow alternative formats for the S/MIME message including
16655 PEM and DER. The -content option allows the content to be
16656 specified separately. This should allow things like Netscape
16657 form signing output easier to verify.
16661 * Fix the ASN1 encoding of tags using the 'long form'.
16665 * New ASN1 functions, `i2c_*` and `c2i_*` for INTEGER and BIT
16666 STRING types. These convert content octets to and from the
16667 underlying type. The actual tag and length octets are
16668 already assumed to have been read in and checked. These
16669 are needed because all other string types have virtually
16670 identical handling apart from the tag. By having versions
16671 of the ASN1 functions that just operate on content octets
16672 IMPLICIT tagging can be handled properly. It also allows
16673 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
16674 and ASN1_INTEGER are identical apart from the tag.
16678 * Change the handling of OID objects as follows:
16680 - New object identifiers are inserted in objects.txt, following
16681 the syntax given in [crypto/objects/README.md](crypto/objects/README.md).
16682 - objects.pl is used to process obj_mac.num and create a new
16684 - obj_dat.pl is used to create a new obj_dat.h, using the data in
16687 This is currently kind of a hack, and the perl code in objects.pl
16688 isn't very elegant, but it works as I intended. The simplest way
16689 to check that it worked correctly is to look in obj_dat.h and
16690 check the array nid_objs and make sure the objects haven't moved
16691 around (this is important!). Additions are OK, as well as
16692 consistent name changes.
16696 * Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
16700 * Addition of the command line parameter '-rand file' to 'openssl req'.
16701 The given file adds to whatever has already been seeded into the
16702 random pool through the RANDFILE configuration file option or
16703 environment variable, or the default random state file.
16707 * mkstack.pl now sorts each macro group into lexical order.
16708 Previously the output order depended on the order the files
16709 appeared in the directory, resulting in needless rewriting
16714 * Patches to make OpenSSL compile under Win32 again. Mostly
16715 work arounds for the VC++ problem that it treats func() as
16716 func(void). Also stripped out the parts of mkdef.pl that
16717 added extra typesafe functions: these no longer exist.
16721 * Reorganisation of the stack code. The macros are now all
16722 collected in safestack.h . Each macro is defined in terms of
16723 a "stack macro" of the form `SKM_<name>(type, a, b)`. The
16724 DEBUG_SAFESTACK is now handled in terms of function casts,
16725 this has the advantage of retaining type safety without the
16726 use of additional functions. If DEBUG_SAFESTACK is not defined
16727 then the non typesafe macros are used instead. Also modified the
16728 mkstack.pl script to handle the new form. Needs testing to see
16729 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
16730 the default if no major problems. Similar behaviour for ASN1_SET_OF
16731 and PKCS12_STACK_OF.
16735 * When some versions of IIS use the 'NET' form of private key the
16736 key derivation algorithm is different. Normally MD5(password) is
16737 used as a 128 bit RC4 key. In the modified case
16738 MD5(MD5(password) + "SGCKEYSALT") is used instead. Added some
16739 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
16740 as the old Netscape_RSA functions except they have an additional
16741 'sgckey' parameter which uses the modified algorithm. Also added
16742 an -sgckey command line option to the rsa utility. Thanks to
16743 Adrian Peck <bertie@ncipher.com> for posting details of the modified
16744 algorithm to openssl-dev.
16748 * The evp_local.h macros were using 'c.##kname' which resulted in
16749 invalid expansion on some systems (SCO 5.0.5 for example).
16750 Corrected to 'c.kname'.
16752 *Phillip Porch <root@theporch.com>*
16754 * New X509_get1_email() and X509_REQ_get1_email() functions that return
16755 a STACK of email addresses from a certificate or request, these look
16756 in the subject name and the subject alternative name extensions and
16757 omit any duplicate addresses.
16761 * Re-implement BN_mod_exp2_mont using independent (and larger) windows.
16762 This makes DSA verification about 2 % faster.
16766 * Increase maximum window size in `BN_mod_exp_...` to 6 bits instead of 5
16767 (meaning that now 2^5 values will be precomputed, which is only 4 KB
16768 plus overhead for 1024 bit moduli).
16769 This makes exponentiations about 0.5 % faster for 1024 bit
16770 exponents (as measured by "openssl speed rsa2048").
16774 * Rename memory handling macros to avoid conflicts with other
16776 Malloc => OPENSSL_malloc
16777 Malloc_locked => OPENSSL_malloc_locked
16778 Realloc => OPENSSL_realloc
16779 Free => OPENSSL_free
16783 * New function BN_mod_exp_mont_word for small bases (roughly 15%
16784 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
16788 * CygWin32 support.
16790 *John Jarvie <jjarvie@newsguy.com>*
16792 * The type-safe stack code has been rejigged. It is now only compiled
16793 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
16794 by default all type-specific stack functions are "#define"d back to
16795 standard stack functions. This results in more streamlined output
16796 but retains the type-safety checking possibilities of the original
16801 * The STACK code has been cleaned up, and certain type declarations
16802 that didn't make a lot of sense have been brought in line. This has
16803 also involved a cleanup of sorts in safestack.h to more correctly
16804 map type-safe stack functions onto their plain stack counterparts.
16805 This work has also resulted in a variety of "const"ifications of
16806 lots of the code, especially `_cmp` operations which should normally
16807 be prototyped with "const" parameters anyway.
16811 * When generating bytes for the first time in md_rand.c, 'stir the pool'
16812 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
16813 (The PRNG state consists of two parts, the large pool 'state' and 'md',
16814 where all of 'md' is used each time the PRNG is used, but 'state'
16815 is used only indexed by a cyclic counter. As entropy may not be
16816 well distributed from the beginning, 'md' is important as a
16817 chaining variable. However, the output function chains only half
16818 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
16819 all of 'md', and seeding with STATE_SIZE dummy bytes will result
16820 in all of 'state' being rewritten, with the new values depending
16821 on virtually all of 'md'. This overcomes the 80 bit limitation.)
16825 * In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
16826 the handshake is continued after ssl_verify_cert_chain();
16827 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
16828 can lead to 'unexplainable' connection aborts later.
16830 *Bodo Moeller; problem tracked down by Lutz Jaenicke*
16832 * Major EVP API cipher revision.
16833 Add hooks for extra EVP features. This allows various cipher
16834 parameters to be set in the EVP interface. Support added for variable
16835 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
16836 setting of RC2 and RC5 parameters.
16838 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
16841 Remove lots of duplicated code from the EVP library. For example *every*
16842 cipher init() function handles the 'iv' in the same way according to the
16843 cipher mode. They also all do nothing if the 'key' parameter is NULL and
16844 for CFB and OFB modes they zero ctx->num.
16846 New functionality allows removal of S/MIME code RC2 hack.
16848 Most of the routines have the same form and so can be declared in terms
16851 By shifting this to the top level EVP_CipherInit() it can be removed from
16852 all individual ciphers. If the cipher wants to handle IVs or keys
16853 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
16856 Change lots of functions like EVP_EncryptUpdate() to now return a
16857 value: although software versions of the algorithms cannot fail
16858 any installed hardware versions can.
16862 * Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
16863 this option is set, tolerate broken clients that send the negotiated
16864 protocol version number instead of the requested protocol version
16869 * Call dh_tmp_cb (set by `..._TMP_DH_CB`) with correct 'is_export' flag;
16870 i.e. non-zero for export ciphersuites, zero otherwise.
16871 Previous versions had this flag inverted, inconsistent with
16872 rsa_tmp_cb (..._TMP_RSA_CB).
16874 *Bodo Moeller; problem reported by Amit Chopra*
16876 * Add missing DSA library text string. Work around for some IIS
16877 key files with invalid SEQUENCE encoding.
16881 * Add a document (doc/standards.txt) that list all kinds of standards
16882 and so on that are implemented in OpenSSL.
16886 * Enhance c_rehash script. Old version would mishandle certificates
16887 with the same subject name hash and wouldn't handle CRLs at all.
16888 Added -fingerprint option to crl utility, to support new c_rehash
16893 * Eliminate non-ANSI declarations in crypto.h and stack.h.
16897 * Fix for SSL server purpose checking. Server checking was
16898 rejecting certificates which had extended key usage present
16899 but no ssl client purpose.
16901 *Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>*
16903 * Make PKCS#12 code work with no password. The PKCS#12 spec
16904 is a little unclear about how a blank password is handled.
16905 Since the password in encoded as a BMPString with terminating
16906 double NULL a zero length password would end up as just the
16907 double NULL. However no password at all is different and is
16908 handled differently in the PKCS#12 key generation code. NS
16909 treats a blank password as zero length. MSIE treats it as no
16910 password on export: but it will try both on import. We now do
16911 the same: PKCS12_parse() tries zero length and no password if
16912 the password is set to "" or NULL (NULL is now a valid password:
16913 it wasn't before) as does the pkcs12 application.
16917 * Bugfixes in `apps/x509.c`: Avoid a memory leak; and don't use
16918 perror when PEM_read_bio_X509_REQ fails, the error message must
16919 be obtained from the error queue.
16923 * Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
16924 it in ERR_remove_state if appropriate, and change ERR_get_state
16925 accordingly to avoid race conditions (this is necessary because
16926 thread_hash is no longer constant once set).
16930 * Bugfix for linux-elf makefile.one.
16934 * RSA_get_default_method() will now cause a default
16935 RSA_METHOD to be chosen if one doesn't exist already.
16936 Previously this was only set during a call to RSA_new()
16937 or RSA_new_method(NULL) meaning it was possible for
16938 RSA_get_default_method() to return NULL.
16942 * Added native name translation to the existing DSO code
16943 that will convert (if the flag to do so is set) filenames
16944 that are sufficiently small and have no path information
16945 into a canonical native form. Eg. "blah" converted to
16946 "libblah.so" or "blah.dll" etc.
16950 * New function ERR_error_string_n(e, buf, len) which is like
16951 ERR_error_string(e, buf), but writes at most 'len' bytes
16952 including the 0 terminator. For ERR_error_string_n, 'buf'
16955 *Damien Miller <djm@mindrot.org>, Bodo Moeller*
16957 * CONF library reworked to become more general. A new CONF
16958 configuration file reader "class" is implemented as well as a
16959 new functions (`NCONF_*`, for "New CONF") to handle it. The now
16960 old `CONF_*` functions are still there, but are reimplemented to
16961 work in terms of the new functions. Also, a set of functions
16962 to handle the internal storage of the configuration data is
16963 provided to make it easier to write new configuration file
16964 reader "classes" (I can definitely see something reading a
16965 configuration file in XML format, for example), called `_CONF_*`,
16966 or "the configuration storage API"...
16968 The new configuration file reading functions are:
16970 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
16971 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
16973 NCONF_default, NCONF_WIN32
16975 NCONF_dump_fp, NCONF_dump_bio
16977 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
16978 NCONF_new creates a new CONF object. This works in the same way
16979 as other interfaces in OpenSSL, like the BIO interface.
16980 `NCONF_dump_*` dump the internal storage of the configuration file,
16981 which is useful for debugging. All other functions take the same
16982 arguments as the old `CONF_*` functions with the exception of the
16983 first that must be a `CONF *` instead of a `LHASH *`.
16985 To make it easier to use the new classes with the old `CONF_*` functions,
16986 the function CONF_set_default_method is provided.
16990 * Add '-tls1' option to 'openssl ciphers', which was already
16991 mentioned in the documentation but had not been implemented.
16992 (This option is not yet really useful because even the additional
16993 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
16997 * Initial DSO code added into libcrypto for letting OpenSSL (and
16998 OpenSSL-based applications) load shared libraries and bind to
16999 them in a portable way.
17001 *Geoff Thorpe, with contributions from Richard Levitte*
17003 ### Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
17005 * Make sure _lrotl and _lrotr are only used with MSVC.
17007 * Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
17008 (the default implementation of RAND_status).
17010 * Rename openssl x509 option '-crlext', which was added in 0.9.5,
17011 to '-clrext' (= clear extensions), as intended and documented.
17012 *Bodo Moeller; inconsistency pointed out by Michael Attili
17013 <attili@amaxo.com>*
17015 * Fix for HMAC. It wasn't zeroing the rest of the block if the key length
17016 was larger than the MD block size.
17018 *Steve Henson, pointed out by Yost William <YostW@tce.com>*
17020 * Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
17021 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
17022 using the passed key: if the passed key was a private key the result
17023 of X509_print(), for example, would be to print out all the private key
17028 * des_quad_cksum() byte order bug fix.
17029 *Ulf Möller, using the problem description in krb4-0.9.7, where
17030 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>*
17032 * Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
17035 *Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>*
17037 * For easily testing in shell scripts whether some command
17038 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
17039 returns with exit code 0 iff no command of the given name is available.
17040 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
17041 the output goes to stdout and nothing is printed to stderr.
17042 Additional arguments are always ignored.
17044 Since for each cipher there is a command of the same name,
17045 the 'no-cipher' compilation switches can be tested this way.
17047 ('openssl no-XXX' is not able to detect pseudo-commands such
17048 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
17052 * Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
17056 * For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
17057 is set; it will be thrown away anyway because each handshake creates
17059 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
17060 to parameters -- in previous versions (since OpenSSL 0.9.3) the
17061 'default key' from SSL_CTX_set_tmp_dh would always be lost, meaning
17062 you effectively got SSL_OP_SINGLE_DH_USE when using this macro.
17066 * New s_client option -ign_eof: EOF at stdin is ignored, and
17067 'Q' and 'R' lose their special meanings (quit/renegotiate).
17068 This is part of what -quiet does; unlike -quiet, -ign_eof
17069 does not suppress any output.
17073 * Add compatibility options to the purpose and trust code. The
17074 purpose X509_PURPOSE_ANY is "any purpose" which automatically
17075 accepts a certificate or CA, this was the previous behaviour,
17076 with all the associated security issues.
17078 X509_TRUST_COMPAT is the old trust behaviour: only and
17079 automatically trust self signed roots in certificate store. A
17080 new trust setting X509_TRUST_DEFAULT is used to specify that
17081 a purpose has no associated trust setting and it should instead
17082 use the value in the default purpose.
17086 * Fix the PKCS#8 DSA private key code so it decodes keys again
17087 and fix a memory leak.
17091 * In util/mkerr.pl (which implements 'make errors'), preserve
17092 reason strings from the previous version of the .c file, as
17093 the default to have only downcase letters (and digits) in
17094 automatically generated reasons codes is not always appropriate.
17098 * In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
17099 using strerror. Previously, ERR_reason_error_string() returned
17100 library names as reason strings for SYSerr; but SYSerr is a special
17101 case where small numbers are errno values, not library numbers.
17105 * Add '-dsaparam' option to 'openssl dhparam' application. This
17106 converts DSA parameters into DH parameters. (When creating parameters,
17107 DSA_generate_parameters is used.)
17111 * Include 'length' (recommended exponent length) in C code generated
17112 by 'openssl dhparam -C'.
17116 * The second argument to set_label in perlasm was already being used
17117 so couldn't be used as a "file scope" flag. Moved to third argument
17122 * In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
17123 instead of RAND_bytes for encryption IVs and salts.
17127 * Include RAND_status() into RAND_METHOD instead of implementing
17128 it only for md_rand.c Otherwise replacing the PRNG by calling
17129 RAND_set_rand_method would be impossible.
17133 * Don't let DSA_generate_key() enter an infinite loop if the random
17134 number generation fails.
17138 * New 'rand' application for creating pseudo-random output.
17142 * Added configuration support for Linux/IA64
17144 *Rolf Haberrecker <rolf@suse.de>*
17146 * Assembler module support for Mingw32.
17150 * Shared library support for HPUX (in shlib/).
17152 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous*
17154 * Shared library support for Solaris gcc.
17156 *Lutz Behnke <behnke@trustcenter.de>*
17158 ### Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
17160 * PKCS7_encrypt() was adding text MIME headers twice because they
17161 were added manually and by SMIME_crlf_copy().
17165 * In bntest.c don't call BN_rand with zero bits argument.
17167 *Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>*
17169 * BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
17170 case was implemented. This caused BN_div_recp() to fail occasionally.
17174 * Add an optional second argument to the set_label() in the perl
17175 assembly language builder. If this argument exists and is set
17176 to 1 it signals that the assembler should use a symbol whose
17177 scope is the entire file, not just the current function. This
17178 is needed with MASM which uses the format label:: for this scope.
17180 *Steve Henson, pointed out by Peter Runestig <peter@runestig.com>*
17182 * Change the ASN1 types so they are typedefs by default. Before
17183 almost all types were #define'd to ASN1_STRING which was causing
17184 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
17189 * Change names of new functions to the new get1/get0 naming
17190 convention: After 'get1', the caller owns a reference count
17191 and has to call `..._free`; 'get0' returns a pointer to some
17192 data structure without incrementing reference counters.
17193 (Some of the existing 'get' functions increment a reference
17194 counter, some don't.)
17195 Similarly, 'set1' and 'add1' functions increase reference
17196 counters or duplicate objects.
17200 * Allow for the possibility of temp RSA key generation failure:
17201 the code used to assume it always worked and crashed on failure.
17205 * Fix potential buffer overrun problem in BIO_printf().
17206 *Ulf Möller, using public domain code by Patrick Powell; problem
17207 pointed out by David Sacerdote <das33@cornell.edu>*
17209 * Support EGD <http://www.lothar.com/tech/crypto/>. New functions
17210 RAND_egd() and RAND_status(). In the command line application,
17211 the EGD socket can be specified like a seed file using RANDFILE
17216 * Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
17217 Some CAs (e.g. Verisign) distribute certificates in this form.
17221 * Remove the SSL_ALLOW_ADH compile option and set the default cipher
17222 list to exclude them. This means that no special compilation option
17223 is needed to use anonymous DH: it just needs to be included in the
17228 * Change the EVP_MD_CTX_type macro so its meaning consistent with
17229 EVP_MD_type. The old functionality is available in a new macro called
17230 EVP_MD_md(). Change code that uses it and update docs.
17234 * `..._ctrl` functions now have corresponding `..._callback_ctrl` functions
17235 where the `void *` argument is replaced by a function pointer argument.
17236 Previously `void *` was abused to point to functions, which works on
17237 many platforms, but is not correct. As these functions are usually
17238 called by macros defined in OpenSSL header files, most source code
17239 should work without changes.
17243 * `<openssl/opensslconf.h>` (which is created by Configure) now contains
17244 sections with information on -D... compiler switches used for
17245 compiling the library so that applications can see them. To enable
17246 one of these sections, a pre-processor symbol `OPENSSL_..._DEFINES`
17247 must be defined. E.g.,
17248 #define OPENSSL_ALGORITHM_DEFINES
17249 #include <openssl/opensslconf.h>
17250 defines all pertinent `NO_<algo>` symbols, such as NO_IDEA, NO_RSA, etc.
17252 *Richard Levitte, Ulf and Bodo Möller*
17254 * Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
17259 * Change the 'other' type in certificate aux info to a STACK_OF
17260 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
17261 the required ASN1 format: arbitrary types determined by an OID.
17265 * Add some PEM_write_X509_REQ_NEW() functions and a command line
17266 argument to 'req'. This is not because the function is newer or
17267 better than others it just uses the work 'NEW' in the certificate
17268 request header lines. Some software needs this.
17272 * Reorganise password command line arguments: now passwords can be
17273 obtained from various sources. Delete the PEM_cb function and make
17274 it the default behaviour: i.e. if the callback is NULL and the
17275 usrdata argument is not NULL interpret it as a null terminated pass
17276 phrase. If usrdata and the callback are NULL then the pass phrase
17277 is prompted for as usual.
17281 * Add support for the Compaq Atalla crypto accelerator. If it is installed,
17282 the support is automatically enabled. The resulting binaries will
17283 autodetect the card and use it if present.
17285 *Ben Laurie and Compaq Inc.*
17287 * Work around for Netscape hang bug. This sends certificate request
17288 and server done in one record. Since this is perfectly legal in the
17289 SSL/TLS protocol it isn't a "bug" option and is on by default. See
17290 the bugs/SSLv3 entry for more info.
17294 * HP-UX tune-up: new unified configs, HP C compiler bug workaround.
17298 * Add -rand argument to smime and pkcs12 applications and read/write
17303 * New 'passwd' tool for crypt(3) and apr1 password hashes.
17307 * Add command line password options to the remaining applications.
17311 * Bug fix for BN_div_recp() for numerators with an even number of
17316 * More tests in bntest.c, and changed test_bn output.
17320 * ./config recognizes MacOS X now.
17324 * Bug fix for BN_div() when the first words of num and divisor are
17325 equal (it gave wrong results if `(rem=(n1-q*d0)&BN_MASK2) < d0)`.
17329 * Add support for various broken PKCS#8 formats, and command line
17330 options to produce them.
17334 * New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
17335 get temporary BIGNUMs from a BN_CTX.
17339 * Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
17344 * Change the `SSLeay_add_all_*()` functions to `OpenSSL_add_all_*()` and
17345 include a #define from the old name to the new. The original intent
17346 was that statically linked binaries could for example just call
17347 SSLeay_add_all_ciphers() to just add ciphers to the table and not
17348 link with digests. This never worked because SSLeay_add_all_digests()
17349 and SSLeay_add_all_ciphers() were in the same source file so calling
17350 one would link with the other. They are now in separate source files.
17354 * Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
17358 * Use a less unusual form of the Miller-Rabin primality test (it used
17359 a binary algorithm for exponentiation integrated into the Miller-Rabin
17360 loop, our standard modexp algorithms are faster).
17364 * Support for the EBCDIC character set completed.
17366 *Martin Kraemer <Martin.Kraemer@Mch.SNI.De>*
17368 * Source code cleanups: use const where appropriate, eliminate casts,
17369 use `void *` instead of `char *` in lhash.
17373 * Bugfix: ssl3_send_server_key_exchange was not restartable
17374 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
17375 this the server could overwrite ephemeral keys that the client
17380 * Turn DSA_is_prime into a macro that calls BN_is_prime,
17381 using 50 iterations of the Rabin-Miller test.
17383 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
17384 iterations of the Rabin-Miller test as required by the appendix
17385 to FIPS PUB 186[-1]) instead of DSA_is_prime.
17386 As BN_is_prime_fasttest includes trial division, DSA parameter
17387 generation becomes much faster.
17389 This implies a change for the callback functions in DSA_is_prime
17390 and DSA_generate_parameters: The callback function is called once
17391 for each positive witness in the Rabin-Miller test, not just
17392 occasionally in the inner loop; and the parameters to the
17393 callback function now provide an iteration count for the outer
17394 loop rather than for the current invocation of the inner loop.
17395 DSA_generate_parameters additionally can call the callback
17396 function with an 'iteration count' of -1, meaning that a
17397 candidate has passed the trial division test (when q is generated
17398 from an application-provided seed, trial division is skipped).
17402 * New function BN_is_prime_fasttest that optionally does trial
17403 division before starting the Rabin-Miller test and has
17404 an additional BN_CTX * argument (whereas BN_is_prime always
17405 has to allocate at least one BN_CTX).
17406 'callback(1, -1, cb_arg)' is called when a number has passed the
17407 trial division stage.
17411 * Fix for bug in CRL encoding. The validity dates weren't being handled
17416 * New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
17420 * New function BN_pseudo_rand().
17424 * Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
17425 bignum version of BN_from_montgomery() with the working code from
17426 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
17431 * Avoid a race condition in s2_clnt.c (function get_server_hello) that
17432 made it impossible to use the same SSL_SESSION data structure in
17433 SSL2 clients in multiple threads.
17437 * The return value of RAND_load_file() no longer counts bytes obtained
17438 by stat(). RAND_load_file(..., -1) is new and uses the complete file
17439 to seed the PRNG (previously an explicit byte count was required).
17441 *Ulf Möller, Bodo Möller*
17443 * Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
17444 used `char *` instead of `void *` and had casts all over the place.
17448 * Make BN_generate_prime() return NULL on error if ret!=NULL.
17452 * Retain source code compatibility for BN_prime_checks macro:
17453 BN_is_prime(..., BN_prime_checks, ...) now uses
17454 BN_prime_checks_for_size to determine the appropriate number of
17455 Rabin-Miller iterations.
17459 * Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
17460 DH_CHECK_P_NOT_SAFE_PRIME.
17461 (Check if this is true? OpenPGP calls them "strong".)
17465 * Merge the functionality of "dh" and "gendh" programs into a new program
17466 "dhparam". The old programs are retained for now but will handle DH keys
17467 (instead of parameters) in future.
17471 * Make the ciphers, s_server and s_client programs check the return values
17472 when a new cipher list is set.
17476 * Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
17477 ciphers. Before when the 56bit ciphers were enabled the sorting was
17480 The syntax for the cipher sorting has been extended to support sorting by
17481 cipher-strength (using the strength_bits hard coded in the tables).
17482 The new command is `@STRENGTH` (see also `doc/apps/ciphers.pod`).
17484 Fix a bug in the cipher-command parser: when supplying a cipher command
17485 string with an "undefined" symbol (neither command nor alphanumeric
17486 *A-Za-z0-9*, ssl_set_cipher_list used to hang in an endless loop. Now
17487 an error is flagged.
17489 Due to the strength-sorting extension, the code of the
17490 ssl_create_cipher_list() function was completely rearranged. I hope that
17491 the readability was also increased :-)
17493 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>*
17495 * Minor change to 'x509' utility. The -CAcreateserial option now uses 1
17496 for the first serial number and places 2 in the serial number file. This
17497 avoids problems when the root CA is created with serial number zero and
17498 the first user certificate has the same issuer name and serial number
17503 * Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
17504 the new code. Add documentation for this stuff.
17508 * Changes to X509_ATTRIBUTE utilities. These have been renamed from
17509 `X509_*()` to `X509at_*()` on the grounds that they don't handle X509
17510 structures and behave in an analogous way to the X509v3 functions:
17511 they shouldn't be called directly but wrapper functions should be used
17514 So we also now have some wrapper functions that call the X509at functions
17515 when passed certificate requests. (TO DO: similar things can be done with
17516 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
17517 things. Some of these need some d2i or i2d and print functionality
17518 because they handle more complex structures.)
17522 * Add missing #ifndefs that caused missing symbols when building libssl
17523 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
17524 NO_RSA in `ssl/s2*.c`.
17526 *Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller*
17528 * Precautions against using the PRNG uninitialized: RAND_bytes() now
17529 has a return value which indicates the quality of the random data
17530 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
17531 error queue. New function RAND_pseudo_bytes() generates output that is
17532 guaranteed to be unique but not unpredictable. RAND_add is like
17533 RAND_seed, but takes an extra argument for an entropy estimate
17534 (RAND_seed always assumes full entropy).
17538 * Do more iterations of Rabin-Miller probable prime test (specifically,
17539 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
17540 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
17541 in crypto/bn/bn_prime.c for the complete table). This guarantees a
17542 false-positive rate of at most 2^-80 for random input.
17546 * Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
17550 * New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
17551 in the 0.9.5 release), this returns the chain
17552 from an X509_CTX structure with a dup of the stack and all
17553 the X509 reference counts upped: so the stack will exist
17554 after X509_CTX_cleanup() has been called. Modify pkcs12.c
17557 Also make SSL_SESSION_print() print out the verify return
17562 * Add manpage for the pkcs12 command. Also change the default
17563 behaviour so MAC iteration counts are used unless the new
17564 -nomaciter option is used. This improves file security and
17565 only older versions of MSIE (4.0 for example) need it.
17569 * Honor the no-xxx Configure options when creating .DEF files.
17573 * Add PKCS#10 attributes to field table: challengePassword,
17574 unstructuredName and unstructuredAddress. These are taken from
17575 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
17576 international characters are used.
17578 More changes to X509_ATTRIBUTE code: allow the setting of types
17579 based on strings. Remove the 'loc' parameter when adding
17580 attributes because these will be a SET OF encoding which is sorted
17585 * Initial changes to the 'req' utility to allow request generation
17586 automation. This will allow an application to just generate a template
17587 file containing all the field values and have req construct the
17590 Initial support for X509_ATTRIBUTE handling. Stacks of these are
17591 used all over the place including certificate requests and PKCS#7
17592 structures. They are currently handled manually where necessary with
17593 some primitive wrappers for PKCS#7. The new functions behave in a
17594 manner analogous to the X509 extension functions: they allow
17595 attributes to be looked up by NID and added.
17597 Later something similar to the X509V3 code would be desirable to
17598 automatically handle the encoding, decoding and printing of the
17599 more complex types. The string types like challengePassword can
17600 be handled by the string table functions.
17602 Also modified the multi byte string table handling. Now there is
17603 a 'global mask' which masks out certain types. The table itself
17604 can use the flag STABLE_NO_MASK to ignore the mask setting: this
17605 is useful when for example there is only one permissible type
17606 (as in countryName) and using the mask might result in no valid
17611 * Clean up 'Finished' handling, and add functions SSL_get_finished and
17612 SSL_get_peer_finished to allow applications to obtain the latest
17613 Finished messages sent to the peer or expected from the peer,
17614 respectively. (SSL_get_peer_finished is usually the Finished message
17615 actually received from the peer, otherwise the protocol will be aborted.)
17617 As the Finished message are message digests of the complete handshake
17618 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
17619 be used for external authentication procedures when the authentication
17620 provided by SSL/TLS is not desired or is not enough.
17624 * Enhanced support for Alpha Linux is added. Now ./config checks if
17625 the host supports BWX extension and if Compaq C is present on the
17626 $PATH. Just exploiting of the BWX extension results in 20-30%
17627 performance kick for some algorithms, e.g. DES and RC4 to mention
17628 a couple. Compaq C in turn generates ~20% faster code for MD5 and
17633 * Add support for MS "fast SGC". This is arguably a violation of the
17634 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
17635 weak crypto and after checking the certificate is SGC a second one
17636 with strong crypto. MS SGC stops the first handshake after receiving
17637 the server certificate message and sends a second client hello. Since
17638 a server will typically do all the time consuming operations before
17639 expecting any further messages from the client (server key exchange
17640 is the most expensive) there is little difference between the two.
17642 To get OpenSSL to support MS SGC we have to permit a second client
17643 hello message after we have sent server done. In addition we have to
17644 reset the MAC if we do get this second client hello.
17648 * Add a function 'd2i_AutoPrivateKey()' this will automatically decide
17649 if a DER encoded private key is RSA or DSA traditional format. Changed
17650 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
17651 format DER encoded private key. Newer code should use PKCS#8 format which
17652 has the key type encoded in the ASN1 structure. Added DER private key
17653 support to pkcs8 application.
17657 * SSL 3/TLS 1 servers now don't request certificates when an anonymous
17658 ciphersuites has been selected (as required by the SSL 3/TLS 1
17659 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
17660 is set, we interpret this as a request to violate the specification
17661 (the worst that can happen is a handshake failure, and 'correct'
17662 behaviour would result in a handshake failure anyway).
17666 * In SSL_CTX_add_session, take into account that there might be multiple
17667 SSL_SESSION structures with the same session ID (e.g. when two threads
17668 concurrently obtain them from an external cache).
17669 The internal cache can handle only one SSL_SESSION with a given ID,
17670 so if there's a conflict, we now throw out the old one to achieve
17675 * Add OIDs for idea and blowfish in CBC mode. This will allow both
17676 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
17677 some routines that use cipher OIDs: some ciphers do not have OIDs
17678 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
17683 * Simplify the trust setting structure and code. Now we just have
17684 two sequences of OIDs for trusted and rejected settings. These will
17685 typically have values the same as the extended key usage extension
17686 and any application specific purposes.
17688 The trust checking code now has a default behaviour: it will just
17689 check for an object with the same NID as the passed id. Functions can
17690 be provided to override either the default behaviour or the behaviour
17691 for a given id. SSL client, server and email already have functions
17692 in place for compatibility: they check the NID and also return "trusted"
17693 if the certificate is self signed.
17697 * Add d2i,i2d bio/fp functions for PrivateKey: these convert the
17698 traditional format into an EVP_PKEY structure.
17702 * Add a password callback function PEM_cb() which either prompts for
17703 a password if usr_data is NULL or otherwise assumes it is a null
17704 terminated password. Allow passwords to be passed on command line
17705 environment or config files in a few more utilities.
17709 * Add a bunch of DER and PEM functions to handle PKCS#8 format private
17710 keys. Add some short names for PKCS#8 PBE algorithms and allow them
17711 to be specified on the command line for the pkcs8 and pkcs12 utilities.
17712 Update documentation.
17716 * Support for ASN1 "NULL" type. This could be handled before by using
17717 ASN1_TYPE but there wasn't any function that would try to read a NULL
17718 and produce an error if it couldn't. For compatibility we also have
17719 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
17720 don't allocate anything because they don't need to.
17724 * Initial support for MacOS is now provided. Examine INSTALL.MacOS
17727 *Andy Polyakov, Roy Woods <roy@centicsystems.ca>*
17729 * Rebuild of the memory allocation routines used by OpenSSL code and
17730 possibly others as well. The purpose is to make an interface that
17731 provide hooks so anyone can build a separate set of allocation and
17732 deallocation routines to be used by OpenSSL, for example memory
17733 pool implementations, or something else, which was previously hard
17734 since Malloc(), Realloc() and Free() were defined as macros having
17735 the values malloc, realloc and free, respectively (except for Win32
17736 compilations). The same is provided for memory debugging code.
17737 OpenSSL already comes with functionality to find memory leaks, but
17738 this gives people a chance to debug other memory problems.
17740 With these changes, a new set of functions and macros have appeared:
17742 CRYPTO_set_mem_debug_functions() [F]
17743 CRYPTO_get_mem_debug_functions() [F]
17744 CRYPTO_dbg_set_options() [F]
17745 CRYPTO_dbg_get_options() [F]
17746 CRYPTO_malloc_debug_init() [M]
17748 The memory debug functions are NULL by default, unless the library
17749 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
17750 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
17751 gives the standard debugging functions that come with OpenSSL) or
17752 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
17753 provided by the library user) must be used. When the standard
17754 debugging functions are used, CRYPTO_dbg_set_options can be used to
17755 request additional information:
17756 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
17757 the CRYPTO_MDEBUG_xxx macro when compiling the library.
17759 Also, things like CRYPTO_set_mem_functions will always give the
17760 expected result (the new set of functions is used for allocation
17761 and deallocation) at all times, regardless of platform and compiler
17764 To finish it up, some functions that were never use in any other
17765 way than through macros have a new API and new semantic:
17767 CRYPTO_dbg_malloc()
17768 CRYPTO_dbg_realloc()
17771 All macros of value have retained their old syntax.
17773 *Richard Levitte and Bodo Moeller*
17775 * Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
17776 ordering of SMIMECapabilities wasn't in "strength order" and there
17777 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
17782 * Some ASN1 types with illegal zero length encoding (INTEGER,
17783 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
17785 *Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson*
17787 * Merge in my S/MIME library for OpenSSL. This provides a simple
17788 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
17789 functionality to handle multipart/signed properly) and a utility
17790 called 'smime' to call all this stuff. This is based on code I
17791 originally wrote for Celo who have kindly allowed it to be
17792 included in OpenSSL.
17796 * Add variants des_set_key_checked and des_set_key_unchecked of
17797 des_set_key (aka des_key_sched). Global variable des_check_key
17798 decides which of these is called by des_set_key; this way
17799 des_check_key behaves as it always did, but applications and
17800 the library itself, which was buggy for des_check_key == 1,
17801 have a cleaner way to pick the version they need.
17805 * New function PKCS12_newpass() which changes the password of a
17810 * Modify X509_TRUST and X509_PURPOSE so it also uses a static and
17811 dynamic mix. In both cases the ids can be used as an index into the
17812 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
17813 functions so they accept a list of the field values and the
17814 application doesn't need to directly manipulate the X509_TRUST
17819 * Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
17824 * Modify the way the V3 extension code looks up extensions. This now
17825 works in a similar way to the object code: we have some "standard"
17826 extensions in a static table which is searched with OBJ_bsearch()
17827 and the application can add dynamic ones if needed. The file
17828 crypto/x509v3/ext_dat.h now has the info: this file needs to be
17829 updated whenever a new extension is added to the core code and kept
17830 in ext_nid order. There is a simple program 'tabtest.c' which checks
17831 this. New extensions are not added too often so this file can readily
17832 be maintained manually.
17834 There are two big advantages in doing things this way. The extensions
17835 can be looked up immediately and no longer need to be "added" using
17836 X509V3_add_standard_extensions(): this function now does nothing.
17837 Side note: I get *lots* of email saying the extension code doesn't
17838 work because people forget to call this function.
17839 Also no dynamic allocation is done unless new extensions are added:
17840 so if we don't add custom extensions there is no need to call
17841 X509V3_EXT_cleanup().
17845 * Modify enc utility's salting as follows: make salting the default. Add a
17846 magic header, so unsalted files fail gracefully instead of just decrypting
17847 to garbage. This is because not salting is a big security hole, so people
17848 should be discouraged from doing it.
17852 * Fixes and enhancements to the 'x509' utility. It allowed a message
17853 digest to be passed on the command line but it only used this
17854 parameter when signing a certificate. Modified so all relevant
17855 operations are affected by the digest parameter including the
17856 -fingerprint and -x509toreq options. Also -x509toreq choked if a
17857 DSA key was used because it didn't fix the digest.
17861 * Initial certificate chain verify code. Currently tests the untrusted
17862 certificates for consistency with the verify purpose (which is set
17863 when the X509_STORE_CTX structure is set up) and checks the pathlength.
17865 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
17866 this is because it will reject chains with invalid extensions whereas
17867 every previous version of OpenSSL and SSLeay made no checks at all.
17869 Trust code: checks the root CA for the relevant trust settings. Trust
17870 settings have an initial value consistent with the verify purpose: e.g.
17871 if the verify purpose is for SSL client use it expects the CA to be
17872 trusted for SSL client use. However the default value can be changed to
17873 permit custom trust settings: one example of this would be to only trust
17874 certificates from a specific "secure" set of CAs.
17876 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
17877 which should be used for version portability: especially since the
17878 verify structure is likely to change more often now.
17880 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
17881 to set them. If not set then assume SSL clients will verify SSL servers
17884 Two new options to the verify program: -untrusted allows a set of
17885 untrusted certificates to be passed in and -purpose which sets the
17886 intended purpose of the certificate. If a purpose is set then the
17887 new chain verify code is used to check extension consistency.
17891 * Support for the authority information access extension.
17895 * Modify RSA and DSA PEM read routines to transparently handle
17896 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
17897 public keys in a format compatible with certificate
17898 SubjectPublicKeyInfo structures. Unfortunately there were already
17899 functions called *_PublicKey_* which used various odd formats so
17900 these are retained for compatibility: however the DSA variants were
17901 never in a public release so they have been deleted. Changed dsa/rsa
17902 utilities to handle the new format: note no releases ever handled public
17903 keys so we should be OK.
17905 The primary motivation for this change is to avoid the same fiasco
17906 that dogs private keys: there are several incompatible private key
17907 formats some of which are standard and some OpenSSL specific and
17908 require various evil hacks to allow partial transparent handling and
17909 even then it doesn't work with DER formats. Given the option anything
17910 other than PKCS#8 should be dumped: but the other formats have to
17911 stay in the name of compatibility.
17913 With public keys and the benefit of hindsight one standard format
17914 is used which works with EVP_PKEY, RSA or DSA structures: though
17915 it clearly returns an error if you try to read the wrong kind of key.
17917 Added a -pubkey option to the 'x509' utility to output the public key.
17918 Also rename the `EVP_PKEY_get_*()` to `EVP_PKEY_rget_*()`
17919 (renamed to `EVP_PKEY_get1_*()` in the OpenSSL 0.9.5 release) and add
17920 `EVP_PKEY_rset_*()` functions (renamed to `EVP_PKEY_set1_*()`)
17921 that do the same as the `EVP_PKEY_assign_*()` except they up the
17922 reference count of the added key (they don't "swallow" the
17927 * Fixes to crypto/x509/by_file.c the code to read in certificates and
17928 CRLs would fail if the file contained no certificates or no CRLs:
17929 added a new function to read in both types and return the number
17930 read: this means that if none are read it will be an error. The
17931 DER versions of the certificate and CRL reader would always fail
17932 because it isn't possible to mix certificates and CRLs in DER format
17933 without choking one or the other routine. Changed this to just read
17934 a certificate: this is the best we can do. Also modified the code
17935 in `apps/verify.c` to take notice of return codes: it was previously
17936 attempting to read in certificates from NULL pointers and ignoring
17937 any errors: this is one reason why the cert and CRL reader seemed
17938 to work. It doesn't check return codes from the default certificate
17939 routines: these may well fail if the certificates aren't installed.
17943 * Code to support otherName option in GeneralName.
17947 * First update to verify code. Change the verify utility
17948 so it warns if it is passed a self signed certificate:
17949 for consistency with the normal behaviour. X509_verify
17950 has been modified to it will now verify a self signed
17951 certificate if *exactly* the same certificate appears
17952 in the store: it was previously impossible to trust a
17953 single self signed certificate. This means that:
17954 openssl verify ss.pem
17955 now gives a warning about a self signed certificate but
17956 openssl verify -CAfile ss.pem ss.pem
17961 * For servers, store verify_result in SSL_SESSION data structure
17962 (and add it to external session representation).
17963 This is needed when client certificate verifications fails,
17964 but an application-provided verification callback (set by
17965 SSL_CTX_set_cert_verify_callback) allows accepting the session
17966 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
17967 but returns 1): When the session is reused, we have to set
17968 ssl->verify_result to the appropriate error code to avoid
17971 *Bodo Moeller, problem pointed out by Lutz Jaenicke*
17973 * Fix a bug in the new PKCS#7 code: it didn't consider the
17974 case in PKCS7_dataInit() where the signed PKCS7 structure
17975 didn't contain any existing data because it was being created.
17977 *Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson*
17979 * Add a salt to the key derivation routines in enc.c. This
17980 forms the first 8 bytes of the encrypted file. Also add a
17981 -S option to allow a salt to be input on the command line.
17985 * New function X509_cmp(). Oddly enough there wasn't a function
17986 to compare two certificates. We do this by working out the SHA1
17987 hash and comparing that. X509_cmp() will be needed by the trust
17992 * SSL_get1_session() is like SSL_get_session(), but increments
17993 the reference count in the SSL_SESSION returned.
17995 *Geoff Thorpe <geoff@eu.c2.net>*
17997 * Fix for 'req': it was adding a null to request attributes.
17998 Also change the X509_LOOKUP and X509_INFO code to handle
17999 certificate auxiliary information.
18003 * Add support for 40 and 64 bit RC2 and RC4 algorithms: document
18008 * Add the possibility to add extra information to the memory leak
18009 detecting output, to form tracebacks, showing from where each
18010 allocation was originated: CRYPTO_push_info("constant string") adds
18011 the string plus current file name and line number to a per-thread
18012 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
18013 is like calling CYRPTO_pop_info() until the stack is empty.
18014 Also updated memory leak detection code to be multi-thread-safe.
18018 * Add options -text and -noout to pkcs7 utility and delete the
18019 encryption options which never did anything. Update docs.
18023 * Add options to some of the utilities to allow the pass phrase
18024 to be included on either the command line (not recommended on
18025 OSes like Unix) or read from the environment. Update the
18026 manpages and fix a few bugs.
18030 * Add a few manpages for some of the openssl commands.
18034 * Fix the -revoke option in ca. It was freeing up memory twice,
18035 leaking and not finding already revoked certificates.
18039 * Extensive changes to support certificate auxiliary information.
18040 This involves the use of X509_CERT_AUX structure and X509_AUX
18041 functions. An X509_AUX function such as PEM_read_X509_AUX()
18042 can still read in a certificate file in the usual way but it
18043 will also read in any additional "auxiliary information". By
18044 doing things this way a fair degree of compatibility can be
18045 retained: existing certificates can have this information added
18046 using the new 'x509' options.
18048 Current auxiliary information includes an "alias" and some trust
18049 settings. The trust settings will ultimately be used in enhanced
18050 certificate chain verification routines: currently a certificate
18051 can only be trusted if it is self signed and then it is trusted
18056 * Fix assembler for Alpha (tested only on DEC OSF not Linux or `*BSD`).
18057 The problem was that one of the replacement routines had not been working
18058 since SSLeay releases. For now the offending routine has been replaced
18059 with non-optimised assembler. Even so, this now gives around 95%
18060 performance improvement for 1024 bit RSA signs.
18064 * Hack to fix PKCS#7 decryption when used with some unorthodox RC2
18065 handling. Most clients have the effective key size in bits equal to
18066 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
18067 A few however don't do this and instead use the size of the decrypted key
18068 to determine the RC2 key length and the AlgorithmIdentifier to determine
18069 the effective key length. In this case the effective key length can still
18070 be 40 bits but the key length can be 168 bits for example. This is fixed
18071 by manually forcing an RC2 key into the EVP_PKEY structure because the
18072 EVP code can't currently handle unusual RC2 key sizes: it always assumes
18073 the key length and effective key length are equal.
18077 * Add a bunch of functions that should simplify the creation of
18078 X509_NAME structures. Now you should be able to do:
18079 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
18080 and have it automatically work out the correct field type and fill in
18081 the structures. The more adventurous can try:
18082 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
18083 and it will (hopefully) work out the correct multibyte encoding.
18087 * Change the 'req' utility to use the new field handling and multibyte
18088 copy routines. Before the DN field creation was handled in an ad hoc
18089 way in req, ca, and x509 which was rather broken and didn't support
18090 BMPStrings or UTF8Strings. Since some software doesn't implement
18091 BMPStrings or UTF8Strings yet, they can be enabled using the config file
18092 using the dirstring_type option. See the new comment in the default
18093 openssl.cnf for more info.
18097 * Make crypto/rand/md_rand.c more robust:
18098 - Assure unique random numbers after fork().
18099 - Make sure that concurrent threads access the global counter and
18100 md serializably so that we never lose entropy in them
18101 or use exactly the same state in multiple threads.
18102 Access to the large state is not always serializable because
18103 the additional locking could be a performance killer, and
18104 md should be large enough anyway.
18108 * New file `apps/app_rand.c` with commonly needed functionality
18109 for handling the random seed file.
18111 Use the random seed file in some applications that previously did not:
18113 dsaparam -genkey (which also ignored its '-rand' option),
18116 x509 (when signing).
18117 Except on systems with /dev/urandom, it is crucial to have a random
18118 seed file at least for key creation, DSA signing, and for DH exchanges;
18119 for RSA signatures we could do without one.
18121 gendh and gendsa (unlike genrsa) used to read only the first byte
18122 of each file listed in the '-rand' option. The function as previously
18123 found in genrsa is now in app_rand.c and is used by all programs
18124 that support '-rand'.
18128 * In RAND_write_file, use mode 0600 for creating files;
18129 don't just chmod when it may be too late.
18133 * Report an error from X509_STORE_load_locations
18134 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
18138 * New function ASN1_mbstring_copy() this copies a string in either
18139 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
18140 into an ASN1_STRING type. A mask of permissible types is passed
18141 and it chooses the "minimal" type to use or an error if not type
18146 * Add function equivalents to the various macros in asn1.h. The old
18147 macros are retained with an `M_` prefix. Code inside the library can
18148 use the `M_` macros. External code (including the openssl utility)
18149 should *NOT* in order to be "shared library friendly".
18153 * Add various functions that can check a certificate's extensions
18154 to see if it usable for various purposes such as SSL client,
18155 server or S/MIME and CAs of these types. This is currently
18156 VERY EXPERIMENTAL but will ultimately be used for certificate chain
18157 verification. Also added a -purpose flag to x509 utility to
18158 print out all the purposes.
18162 * Add a CRYPTO_EX_DATA to X509 certificate structure and associated
18167 * New `X509V3_{X509,CRL,REVOKED}_get_d2i()` functions. These will search
18168 for, obtain and decode and extension and obtain its critical flag.
18169 This allows all the necessary extension code to be handled in a
18170 single function call.
18174 * RC4 tune-up featuring 30-40% performance improvement on most RISC
18175 platforms. See crypto/rc4/rc4_enc.c for further details.
18179 * New -noout option to asn1parse. This causes no output to be produced
18180 its main use is when combined with -strparse and -out to extract data
18181 from a file (which may not be in ASN.1 format).
18185 * Fix for pkcs12 program. It was hashing an invalid certificate pointer
18186 when producing the local key id.
18188 *Richard Levitte <levitte@stacken.kth.se>*
18190 * New option -dhparam in s_server. This allows a DH parameter file to be
18191 stated explicitly. If it is not stated then it tries the first server
18192 certificate file. The previous behaviour hard coded the filename
18197 * Add -pubin and -pubout options to the rsa and dsa commands. These allow
18198 a public key to be input or output. For example:
18199 openssl rsa -in key.pem -pubout -out pubkey.pem
18200 Also added necessary DSA public key functions to handle this.
18204 * Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
18205 in the message. This was handled by allowing
18206 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
18208 *Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>*
18210 * Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
18211 to the end of the strings whereas this didn't. This would cause problems
18212 if strings read with d2i_ASN1_bytes() were later modified.
18214 *Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>*
18216 * Fix for base64 decode bug. When a base64 bio reads only one line of
18217 data and it contains EOF it will end up returning an error. This is
18218 caused by input 46 bytes long. The cause is due to the way base64
18219 BIOs find the start of base64 encoded data. They do this by trying a
18220 trial decode on each line until they find one that works. When they
18221 do a flag is set and it starts again knowing it can pass all the
18222 data directly through the decoder. Unfortunately it doesn't reset
18223 the context it uses. This means that if EOF is reached an attempt
18224 is made to pass two EOFs through the context and this causes the
18225 resulting error. This can also cause other problems as well. As is
18226 usual with these problems it takes *ages* to find and the fix is
18227 trivial: move one line.
18229 *Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer)*
18231 * Ugly workaround to get s_client and s_server working under Windows. The
18232 old code wouldn't work because it needed to select() on sockets and the
18233 tty (for keypresses and to see if data could be written). Win32 only
18234 supports select() on sockets so we select() with a 1s timeout on the
18235 sockets and then see if any characters are waiting to be read, if none
18236 are present then we retry, we also assume we can always write data to
18237 the tty. This isn't nice because the code then blocks until we've
18238 received a complete line of data and it is effectively polling the
18239 keyboard at 1s intervals: however it's quite a bit better than not
18240 working at all :-) A dedicated Windows application might handle this
18241 with an event loop for example.
18245 * Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
18246 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
18247 will be called when RSA_sign() and RSA_verify() are used. This is useful
18248 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
18249 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
18250 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
18251 This necessitated the support of an extra signature type NID_md5_sha1
18252 for SSL signatures and modifications to the SSL library to use it instead
18253 of calling RSA_public_decrypt() and RSA_private_encrypt().
18257 * Add new -verify -CAfile and -CApath options to the crl program, these
18258 will lookup a CRL issuers certificate and verify the signature in a
18259 similar way to the verify program. Tidy up the crl program so it
18260 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
18261 less strict. It will now permit CRL extensions even if it is not
18262 a V2 CRL: this will allow it to tolerate some broken CRLs.
18266 * Initialize all non-automatic variables each time one of the openssl
18267 sub-programs is started (this is necessary as they may be started
18268 multiple times from the "OpenSSL>" prompt).
18270 *Lennart Bang, Bodo Moeller*
18272 * Preliminary compilation option RSA_NULL which disables RSA crypto without
18273 removing all other RSA functionality (this is what NO_RSA does). This
18274 is so (for example) those in the US can disable those operations covered
18275 by the RSA patent while allowing storage and parsing of RSA keys and RSA
18280 * Non-copying interface to BIO pairs.
18281 (still largely untested)
18285 * New function ASN1_tag2str() to convert an ASN1 tag to a descriptive
18286 ASCII string. This was handled independently in various places before.
18290 * New functions UTF8_getc() and UTF8_putc() that parse and generate
18291 UTF8 strings a character at a time.
18295 * Use client_version from client hello to select the protocol
18296 (s23_srvr.c) and for RSA client key exchange verification
18297 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
18301 * Add various utility functions to handle SPKACs, these were previously
18302 handled by poking round in the structure internals. Added new function
18303 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
18304 print, verify and generate SPKACs. Based on an original idea from
18305 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
18309 * RIPEMD160 is operational on all platforms and is back in 'make test'.
18313 * Allow the config file extension section to be overwritten on the
18314 command line. Based on an original idea from Massimiliano Pala
18315 <madwolf@comune.modena.it>. The new option is called -extensions
18316 and can be applied to ca, req and x509. Also -reqexts to override
18317 the request extensions in req and -crlexts to override the crl extensions
18322 * Add new feature to the SPKAC handling in ca. Now you can include
18323 the same field multiple times by preceding it by "XXXX." for example:
18326 this is the same syntax as used in the req config file.
18330 * Allow certificate extensions to be added to certificate requests. These
18331 are specified in a 'req_extensions' option of the req section of the
18332 config file. They can be printed out with the -text option to req but
18333 are otherwise ignored at present.
18337 * Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
18338 data read consists of only the final block it would not decrypted because
18339 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
18340 A misplaced 'break' also meant the decrypted final block might not be
18341 copied until the next read.
18345 * Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
18346 a few extra parameters to the DH structure: these will be useful if
18347 for example we want the value of 'q' or implement X9.42 DH.
18351 * Initial support for DSA_METHOD. This is based on the RSA_METHOD and
18352 provides hooks that allow the default DSA functions or functions on a
18353 "per key" basis to be replaced. This allows hardware acceleration and
18354 hardware key storage to be handled without major modification to the
18355 library. Also added low-level modexp hooks and CRYPTO_EX structure and
18356 associated functions.
18360 * Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
18361 as "read only": it can't be written to and the buffer it points to will
18362 not be freed. Reading from a read only BIO is much more efficient than
18363 a normal memory BIO. This was added because there are several times when
18364 an area of memory needs to be read from a BIO. The previous method was
18365 to create a memory BIO and write the data to it, this results in two
18366 copies of the data and an O(n^2) reading algorithm. There is a new
18367 function BIO_new_mem_buf() which creates a read only memory BIO from
18368 an area of memory. Also modified the PKCS#7 routines to use read only
18373 * Bugfix: ssl23_get_client_hello did not work properly when called in
18374 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
18375 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
18376 but a retry condition occurred while trying to read the rest.
18380 * The PKCS7_ENC_CONTENT_new() function was setting the content type as
18381 NID_pkcs7_encrypted by default: this was wrong since this should almost
18382 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
18383 the encrypted data type: this is a more sensible place to put it and it
18384 allows the PKCS#12 code to be tidied up that duplicated this
18389 * Changed obj_dat.pl script so it takes its input and output files on
18390 the command line. This should avoid shell escape redirection problems
18395 * Initial support for certificate extension requests, these are included
18396 in things like Xenroll certificate requests. Included functions to allow
18397 extensions to be obtained and added.
18401 * -crlf option to s_client and s_server for sending newlines as
18402 CRLF (as required by many protocols).
18406 ### Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
18408 * Install libRSAglue.a when OpenSSL is built with RSAref.
18410 *Ralf S. Engelschall*
18412 * A few more `#ifndef NO_FP_API / #endif` pairs for consistency.
18414 *Andrija Antonijevic <TheAntony2@bigfoot.com>*
18416 * Fix -startdate and -enddate (which was missing) arguments to 'ca'
18421 * New function DSA_dup_DH, which duplicates DSA parameters/keys as
18422 DH parameters/keys (q is lost during that conversion, but the resulting
18423 DH parameters contain its length).
18425 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
18426 much faster than DH_generate_parameters (which creates parameters
18427 where `p = 2*q + 1`), and also the smaller q makes DH computations
18428 much more efficient (160-bit exponentiation instead of 1024-bit
18429 exponentiation); so this provides a convenient way to support DHE
18430 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
18431 utter importance to use
18432 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
18434 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
18435 when such DH parameters are used, because otherwise small subgroup
18436 attacks may become possible!
18440 * Avoid memory leak in i2d_DHparams.
18444 * Allow the -k option to be used more than once in the enc program:
18445 this allows the same encrypted message to be read by multiple recipients.
18449 * New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
18450 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
18451 it will always use the numerical form of the OID, even if it has a short
18456 * Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
18457 method only got called if p,q,dmp1,dmq1,iqmp components were present,
18458 otherwise bn_mod_exp was called. In the case of hardware keys for example
18459 no private key components need be present and it might store extra data
18460 in the RSA structure, which cannot be accessed from bn_mod_exp.
18461 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
18462 private key operations.
18466 * Added support for SPARC Linux.
18470 * pem_password_cb function type incompatibly changed from
18471 typedef int pem_password_cb(char *buf, int size, int rwflag);
18473 ....(char *buf, int size, int rwflag, void *userdata);
18474 so that applications can pass data to their callbacks:
18475 The `PEM[_ASN1]_{read,write}...` functions and macros now take an
18476 additional void * argument, which is just handed through whenever
18477 the password callback is called.
18479 *Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller*
18481 New function SSL_CTX_set_default_passwd_cb_userdata.
18483 Compatibility note: As many C implementations push function arguments
18484 onto the stack in reverse order, the new library version is likely to
18485 interoperate with programs that have been compiled with the old
18486 pem_password_cb definition (PEM_whatever takes some data that
18487 happens to be on the stack as its last argument, and the callback
18488 just ignores this garbage); but there is no guarantee whatsoever that
18491 * The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
18492 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
18493 problems not only on Windows, but also on some Unix platforms.
18494 To avoid problematic command lines, these definitions are now in an
18495 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
18496 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
18500 * MIPS III/IV assembler module is reimplemented.
18504 * More DES library cleanups: remove references to srand/rand and
18505 delete an unused file.
18509 * Add support for the free Netwide assembler (NASM) under Win32,
18510 since not many people have MASM (ml) and it can be hard to obtain.
18511 This is currently experimental but it seems to work OK and pass all
18512 the tests. Check out INSTALL.W32 for info.
18516 * Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
18517 without temporary keys kept an extra copy of the server key,
18518 and connections with temporary keys did not free everything in case
18523 * New function RSA_check_key and new openssl rsa option -check
18524 for verifying the consistency of RSA keys.
18526 *Ulf Moeller, Bodo Moeller*
18528 * Various changes to make Win32 compile work:
18529 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
18530 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
18531 comparison" warnings.
18532 3. Add `sk_<TYPE>_sort` to DEF file generator and do make update.
18536 * Add a debugging option to PKCS#5 v2 key generation function: when
18537 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
18538 derived keys are printed to stderr.
18542 * Copy the flags in ASN1_STRING_dup().
18544 *Roman E. Pavlov <pre@mo.msk.ru>*
18546 * The x509 application mishandled signing requests containing DSA
18547 keys when the signing key was also DSA and the parameters didn't match.
18549 It was supposed to omit the parameters when they matched the signing key:
18550 the verifying software was then supposed to automatically use the CA's
18551 parameters if they were absent from the end user certificate.
18553 Omitting parameters is no longer recommended. The test was also
18554 the wrong way round! This was probably due to unusual behaviour in
18555 EVP_cmp_parameters() which returns 1 if the parameters match.
18556 This meant that parameters were omitted when they *didn't* match and
18557 the certificate was useless. Certificates signed with 'ca' didn't have
18560 *Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>*
18562 * Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
18563 The interface is as follows:
18564 Applications can use
18565 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
18566 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
18567 "off" is now the default.
18568 The library internally uses
18569 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
18570 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
18571 to disable memory-checking temporarily.
18573 Some inconsistent states that previously were possible (and were
18574 even the default) are now avoided.
18576 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
18577 with each memory chunk allocated; this is occasionally more helpful
18578 than just having a counter.
18580 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
18582 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
18587 * Introduce "mode" for SSL structures (with defaults in SSL_CTX),
18588 which largely parallels "options", but is for changing API behaviour,
18589 whereas "options" are about protocol behaviour.
18590 Initial "mode" flags are:
18592 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
18593 a single record has been written.
18594 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
18595 retries use the same buffer location.
18596 (But all of the contents must be
18601 * Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
18604 * Fix problems with no-hmac etc.
18606 *Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>*
18608 * New functions RSA_get_default_method(), RSA_set_method() and
18609 RSA_get_method(). These allows replacement of RSA_METHODs without having
18610 to mess around with the internals of an RSA structure.
18614 * Fix memory leaks in DSA_do_sign and DSA_is_prime.
18615 Also really enable memory leak checks in openssl.c and in some
18618 *Chad C. Mulligan, Bodo Moeller*
18620 * Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
18621 up the length of negative integers. This has now been simplified to just
18622 store the length when it is first determined and use it later, rather
18623 than trying to keep track of where data is copied and updating it to
18625 *Steve Henson, reported by Brien Wheeler <bwheeler@authentica-security.com>*
18627 * Add a new function PKCS7_signatureVerify. This allows the verification
18628 of a PKCS#7 signature but with the signing certificate passed to the
18629 function itself. This contrasts with PKCS7_dataVerify which assumes the
18630 certificate is present in the PKCS#7 structure. This isn't always the
18631 case: certificates can be omitted from a PKCS#7 structure and be
18632 distributed by "out of band" means (such as a certificate database).
18636 * Complete the `PEM_*` macros with DECLARE_PEM versions to replace the
18637 function prototypes in pem.h, also change util/mkdef.pl to add the
18638 necessary function names.
18642 * mk1mf.pl (used by Windows builds) did not properly read the
18643 options set by Configure in the top level Makefile, and Configure
18644 was not even able to write more than one option correctly.
18645 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
18649 * New functions CONF_load_bio() and CONF_load_fp() to allow a config
18650 file to be loaded from a BIO or FILE pointer. The BIO version will
18651 for example allow memory BIOs to contain config info.
18655 * New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
18656 Whoever hopes to achieve shared-library compatibility across versions
18657 must use this, not the compile-time macro.
18658 (Exercise 0.9.4: Which is the minimum library version required by
18660 Note: All this applies only to multi-threaded programs, others don't
18665 * Add missing case to s3_clnt.c state machine -- one of the new SSL tests
18666 through a BIO pair triggered the default case, i.e.
18667 SSLerr(...,SSL_R_UNKNOWN_STATE).
18671 * New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
18672 can use the SSL library even if none of the specific BIOs is
18677 * Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
18678 for the encoded length.
18680 *Jeon KyoungHo <khjeon@sds.samsung.co.kr>*
18682 * Add initial documentation of the X509V3 functions.
18686 * Add a new pair of functions PEM_write_PKCS8PrivateKey() and
18687 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
18688 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
18689 secure PKCS#8 private key format with a high iteration count.
18693 * Fix determination of Perl interpreter: A perl or perl5
18694 *directory* in $PATH was also accepted as the interpreter.
18696 *Ralf S. Engelschall*
18698 * Fix demos/sign/sign.c: well there wasn't anything strictly speaking
18699 wrong with it but it was very old and did things like calling
18700 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
18701 unusual formatting.
18705 * Fix demos/selfsign.c: it used obsolete and deleted functions, changed
18706 to use the new extension code.
18710 * Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
18711 with macros. This should make it easier to change their form, add extra
18712 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
18717 * Add to configuration table a new entry that can specify an alternative
18718 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
18719 according to Mark Crispin <MRC@Panda.COM>.
18723 * DES CBC did not update the IV. Weird.
18727 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
18728 Changing the behaviour of the former might break existing programs --
18729 where IV updating is needed, des_ncbc_encrypt can be used.
18732 * When bntest is run from "make test" it drives bc to check its
18733 calculations, as well as internally checking them. If an internal check
18734 fails, it needs to cause bc to give a non-zero result or make test carries
18735 on without noticing the failure. Fixed.
18739 * DES library cleanups.
18743 * Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
18744 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
18745 ciphers. NOTE: although the key derivation function has been verified
18746 against some published test vectors it has not been extensively tested
18747 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
18752 * Instead of "mkdir -p", which is not fully portable, use new
18753 Perl script "util/mkdir-p.pl".
18757 * Rewrite the way password based encryption (PBE) is handled. It used to
18758 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
18759 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
18760 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
18761 the 'parameter' field of the AlgorithmIdentifier is passed to the
18762 underlying key generation function so it must do its own ASN1 parsing.
18763 This has also changed the EVP_PBE_CipherInit() function which now has a
18764 'parameter' argument instead of literal salt and iteration count values
18765 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
18769 * Support for PKCS#5 v1.5 compatible password based encryption algorithms
18770 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
18771 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
18772 KEY" because this clashed with PKCS#8 unencrypted string. Since this
18773 value was just used as a "magic string" and not used directly its
18774 value doesn't matter.
18778 * Introduce some semblance of const correctness to BN. Shame C doesn't
18783 * "linux-sparc64" configuration (ultrapenguin).
18785 *Ray Miller <ray.miller@oucs.ox.ac.uk>*
18786 "linux-sparc" configuration.
18788 *Christian Forster <fo@hawo.stw.uni-erlangen.de>*
18790 * config now generates no-xxx options for missing ciphers.
18794 * Support the EBCDIC character set (work in progress).
18795 File ebcdic.c not yet included because it has a different license.
18797 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
18799 * Support BS2000/OSD-POSIX.
18801 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
18803 * Make callbacks for key generation use `void *` instead of `char *`.
18807 * Make S/MIME samples compile (not yet tested).
18811 * Additional typesafe stacks.
18815 * New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
18819 ### Changes between 0.9.3 and 0.9.3a [29 May 1999]
18821 * New configuration variant "sco5-gcc".
18823 * Updated some demos.
18825 *Sean O Riordain, Wade Scholine*
18827 * Add missing BIO_free at exit of pkcs12 application.
18831 * Fix memory leak in conf.c.
18835 * Updates for Win32 to assembler version of MD5.
18839 * Set #! path to perl in `apps/der_chop` to where we found it
18840 instead of using a fixed path.
18844 * SHA library changes for irix64-mips4-cc.
18848 * Improvements for VMS support.
18852 ### Changes between 0.9.2b and 0.9.3 [24 May 1999]
18854 * Bignum library bug fix. IRIX 6 passes "make test" now!
18855 This also avoids the problems with SC4.2 and unpatched SC5.
18857 *Andy Polyakov <appro@fy.chalmers.se>*
18859 * New functions sk_num, sk_value and sk_set to replace the previous macros.
18860 These are required because of the typesafe stack would otherwise break
18861 existing code. If old code used a structure member which used to be STACK
18862 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
18863 sk_num or sk_value it would produce an error because the num, data members
18864 are not present in STACK_OF. Now it just produces a warning. sk_set
18865 replaces the old method of assigning a value to sk_value
18866 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
18867 that does this will no longer work (and should use sk_set instead) but
18868 this could be regarded as a "questionable" behaviour anyway.
18872 * Fix most of the other PKCS#7 bugs. The "experimental" code can now
18873 correctly handle encrypted S/MIME data.
18877 * Change type of various DES function arguments from des_cblock
18878 (which means, in function argument declarations, pointer to char)
18879 to des_cblock * (meaning pointer to array with 8 char elements),
18880 which allows the compiler to do more typechecking; it was like
18881 that back in SSLeay, but with lots of ugly casts.
18883 Introduce new type const_des_cblock.
18887 * Reorganise the PKCS#7 library and get rid of some of the more obvious
18888 problems: find RecipientInfo structure that matches recipient certificate
18889 and initialise the ASN1 structures properly based on passed cipher.
18893 * Belatedly make the BN tests actually check the results.
18897 * Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
18898 to and from BNs: it was completely broken. New compilation option
18899 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
18900 key elements as negative integers.
18904 * Reorganize and speed up MD5.
18906 *Andy Polyakov <appro@fy.chalmers.se>*
18910 *Richard Levitte <richard@levitte.org>*
18912 * New option -out to asn1parse to allow the parsed structure to be
18913 output to a file. This is most useful when combined with the -strparse
18914 option to examine the output of things like OCTET STRINGS.
18918 * Make SSL library a little more fool-proof by not requiring any longer
18919 that `SSL_set_{accept,connect}_state` be called before
18920 `SSL_{accept,connect}` may be used (`SSL_set_..._state` is omitted
18921 in many applications because usually everything *appeared* to work as
18922 intended anyway -- now it really works as intended).
18926 * Move openssl.cnf out of lib/.
18930 * Fix various things to let OpenSSL even pass "egcc -pipe -O2 -Wall
18931 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
18932 -Wmissing-declarations -Wnested-externs -Winline" with EGCS 1.1.2+
18934 *Ralf S. Engelschall*
18936 * Various fixes to the EVP and PKCS#7 code. It may now be able to
18937 handle PKCS#7 enveloped data properly.
18939 *Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve*
18941 * Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
18942 copying pointers. The cert_st handling is changed by this in
18943 various ways (and thus what used to be known as ctx->default_cert
18944 is now called ctx->cert, since we don't resort to `s->ctx->[default_]cert`
18945 any longer when s->cert does not give us what we need).
18946 ssl_cert_instantiate becomes obsolete by this change.
18947 As soon as we've got the new code right (possibly it already is?),
18948 we have solved a couple of bugs of the earlier code where s->cert
18949 was used as if it could not have been shared with other SSL structures.
18951 Note that using the SSL API in certain dirty ways now will result
18952 in different behaviour than observed with earlier library versions:
18953 Changing settings for an `SSL_CTX *ctx` after having done s = SSL_new(ctx)
18954 does not influence s as it used to.
18956 In order to clean up things more thoroughly, inside SSL_SESSION
18957 we don't use CERT any longer, but a new structure SESS_CERT
18958 that holds per-session data (if available); currently, this is
18959 the peer's certificate chain and, for clients, the server's certificate
18960 and temporary key. CERT holds only those values that can have
18961 meaningful defaults in an SSL_CTX.
18965 * New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
18966 from the internal representation. Various PKCS#7 fixes: remove some
18967 evil casts and set the enc_dig_alg field properly based on the signing
18972 * Allow PKCS#12 password to be set from the command line or the
18973 environment. Let 'ca' get its config file name from the environment
18974 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
18979 * Allow certificate policies extension to use an IA5STRING for the
18980 organization field. This is contrary to the PKIX definition but
18981 VeriSign uses it and IE5 only recognises this form. Document 'x509'
18986 * Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
18987 without disallowing inline assembler and the like for non-pedantic builds.
18991 * Support Borland C++ builder.
18993 *Janez Jere <jj@void.si>, modified by Ulf Möller*
18999 * SHA-1 cleanups and performance enhancements.
19001 *Andy Polyakov <appro@fy.chalmers.se>*
19003 * Sparc v8plus assembler for the bignum library.
19005 *Andy Polyakov <appro@fy.chalmers.se>*
19007 * Accept any -xxx and +xxx compiler options in Configure.
19011 * Update HPUX configuration.
19015 * Add missing `sk_<type>_unshift()` function to safestack.h
19017 *Ralf S. Engelschall*
19019 * New function SSL_CTX_use_certificate_chain_file that sets the
19020 "extra_cert"s in addition to the certificate. (This makes sense
19021 only for "PEM" format files, as chains as a whole are not
19026 * Support verify_depth from the SSL API.
19027 x509_vfy.c had what can be considered an off-by-one-error:
19028 Its depth (which was not part of the external interface)
19029 was actually counting the number of certificates in a chain;
19030 now it really counts the depth.
19034 * Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
19035 instead of X509err, which often resulted in confusing error
19036 messages since the error codes are not globally unique
19037 (e.g. an alleged error in ssl3_accept when a certificate
19038 didn't match the private key).
19040 * New function SSL_CTX_set_session_id_context that allows to set a default
19041 value (so that you don't need SSL_set_session_id_context for each
19042 connection using the SSL_CTX).
19046 * OAEP decoding bug fix.
19050 * Support INSTALL_PREFIX for package builders, as proposed by
19055 * New Configure options "threads" and "no-threads". For systems
19056 where the proper compiler options are known (currently Solaris
19057 and Linux), "threads" is the default.
19061 * New script util/mklink.pl as a faster substitute for util/mklink.sh.
19065 * Install various scripts to $(OPENSSLDIR)/misc, not to
19066 $(INSTALLTOP)/bin -- they shouldn't clutter directories
19067 such as /usr/local/bin.
19071 * "make linux-shared" to build shared libraries.
19073 *Niels Poppe <niels@netbox.org>*
19075 * New Configure option `no-<cipher>` (rsa, idea, rc5, ...).
19079 * Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
19080 extension adding in x509 utility.
19084 * Remove NOPROTO sections and error code comments.
19088 * Partial rewrite of the DEF file generator to now parse the ANSI
19093 * New Configure options --prefix=DIR and --openssldir=DIR.
19097 * Complete rewrite of the error code script(s). It is all now handled
19098 by one script at the top level which handles error code gathering,
19099 header rewriting and C source file generation. It should be much better
19100 than the old method: it now uses a modified version of Ulf's parser to
19101 read the ANSI prototypes in all header files (thus the old K&R definitions
19102 aren't needed for error creation any more) and do a better job of
19103 translating function codes into names. The old 'ASN1 error code embedded
19104 in a comment' is no longer necessary and it doesn't use .err files which
19105 have now been deleted. Also the error code call doesn't have to appear all
19106 on one line (which resulted in some large lines...).
19110 * Change #include filenames from `<foo.h>` to `<openssl/foo.h>`.
19114 * Change behaviour of ssl2_read when facing length-0 packets: Don't return
19115 0 (which usually indicates a closed connection), but continue reading.
19119 * Fix some race conditions.
19123 * Add support for CRL distribution points extension. Add Certificate
19124 Policies and CRL distribution points documentation.
19128 * Move the autogenerated header file parts to crypto/opensslconf.h.
19132 * Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
19133 8 of keying material. Merlin has also confirmed interop with this fix
19134 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
19136 *Merlin Hughes <merlin@baltimore.ie>*
19138 * Fix lots of warnings.
19140 *Richard Levitte <levitte@stacken.kth.se>*
19142 * In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
19143 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
19145 *Richard Levitte <levitte@stacken.kth.se>*
19147 * Fix problems with sizeof(long) == 8.
19149 *Andy Polyakov <appro@fy.chalmers.se>*
19151 * Change functions to ANSI C.
19155 * Fix typos in error codes.
19157 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller*
19159 * Remove defunct assembler files from Configure.
19163 * SPARC v8 assembler BIGNUM implementation.
19165 *Andy Polyakov <appro@fy.chalmers.se>*
19167 * Support for Certificate Policies extension: both print and set.
19168 Various additions to support the r2i method this uses.
19172 * A lot of constification, and fix a bug in X509_NAME_oneline() that could
19173 return a const string when you are expecting an allocated buffer.
19177 * Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
19178 types DirectoryString and DisplayText.
19182 * Add code to allow r2i extensions to access the configuration database,
19183 add an LHASH database driver and add several ctx helper functions.
19187 * Fix an evil bug in bn_expand2() which caused various BN functions to
19188 fail when they extended the size of a BIGNUM.
19192 * Various utility functions to handle SXNet extension. Modify mkdef.pl to
19193 support typesafe stack.
19197 * Fix typo in SSL_[gs]et_options().
19199 *Nils Frostberg <nils@medcom.se>*
19201 * Delete various functions and files that belonged to the (now obsolete)
19202 old X509V3 handling code.
19206 * New Configure option "rsaref".
19210 * Don't auto-generate pem.h.
19214 * Introduce type-safe ASN.1 SETs.
19218 * Convert various additional casted stacks to type-safe STACK_OF() variants.
19220 *Ben Laurie, Ralf S. Engelschall, Steve Henson*
19222 * Introduce type-safe STACKs. This will almost certainly break lots of code
19223 that links with OpenSSL (well at least cause lots of warnings), but fear
19224 not: the conversion is trivial, and it eliminates loads of evil casts. A
19225 few STACKed things have been converted already. Feel free to convert more.
19226 In the fullness of time, I'll do away with the STACK type altogether.
19230 * Add `openssl ca -revoke <certfile>` facility which revokes a certificate
19231 specified in `<certfile>` by updating the entry in the index.txt file.
19232 This way one no longer has to edit the index.txt file manually for
19233 revoking a certificate. The -revoke option does the gory details now.
19235 *Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall*
19237 * Fix `openssl crl -noout -text` combination where `-noout` killed the
19238 `-text` option at all and this way the `-noout -text` combination was
19239 inconsistent in `openssl crl` with the friends in `openssl x509|rsa|dsa`.
19241 *Ralf S. Engelschall*
19243 * Make sure a corresponding plain text error message exists for the
19244 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
19245 verify callback function determined that a certificate was revoked.
19247 *Ralf S. Engelschall*
19249 * Bugfix: In test/testenc, don't test `openssl <cipher>` for
19250 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
19251 all available ciphers including rc5, which was forgotten until now.
19252 In order to let the testing shell script know which algorithms
19253 are available, a new (up to now undocumented) command
19254 `openssl list-cipher-commands` is used.
19258 * Bugfix: s_client occasionally would sleep in select() when
19259 it should have checked SSL_pending() first.
19263 * New functions DSA_do_sign and DSA_do_verify to provide access to
19264 the raw DSA values prior to ASN.1 encoding.
19268 * Tweaks to Configure
19270 *Niels Poppe <niels@netbox.org>*
19272 * Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
19277 * New variables $(RANLIB) and $(PERL) in the Makefiles.
19281 * New config option to avoid instructions that are illegal on the 80386.
19282 The default code is faster, but requires at least a 486.
19286 * Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
19287 SSL2_SERVER_VERSION (not used at all) macros, which are now the
19288 same as SSL2_VERSION anyway.
19292 * New "-showcerts" option for s_client.
19296 * Still more PKCS#12 integration. Add pkcs12 application to openssl
19297 application. Various cleanups and fixes.
19301 * More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
19302 modify error routines to work internally. Add error codes and PBE init
19303 to library startup routines.
19307 * Further PKCS#12 integration. Added password based encryption, PKCS#8 and
19308 packing functions to asn1 and evp. Changed function names and error
19309 codes along the way.
19313 * PKCS12 integration: and so it begins... First of several patches to
19314 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
19315 objects to objects.h
19319 * Add a new 'indent' option to some X509V3 extension code. Initial ASN1
19320 and display support for Thawte strong extranet extension.
19324 * Add LinuxPPC support.
19326 *Jeff Dubrule <igor@pobox.org>*
19328 * Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
19329 bn_div_words in alpha.s.
19331 *Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie*
19333 * Make sure the RSA OAEP test is skipped under -DRSAref because
19334 OAEP isn't supported when OpenSSL is built with RSAref.
19336 *Ulf Moeller <ulf@fitug.de>*
19338 * Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
19339 so they no longer are missing under -DNOPROTO.
19341 *Soren S. Jorvang <soren@t.dk>*
19343 ### Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
19345 * Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
19346 doesn't work when the session is reused. Coming soon!
19350 * Fix a security hole, that allows sessions to be reused in the wrong
19351 context thus bypassing client cert protection! All software that uses
19352 client certs and session caches in multiple contexts NEEDS PATCHING to
19353 allow session reuse! A fuller solution is in the works.
19355 *Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)*
19357 * Some more source tree cleanups (removed obsolete files
19358 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
19359 permission on "config" script to be executable) and a fix for the INSTALL
19362 *Ulf Moeller <ulf@fitug.de>*
19364 * Remove some legacy and erroneous uses of malloc, free instead of
19367 *Lennart Bang <lob@netstream.se>, with minor changes by Steve*
19369 * Make rsa_oaep_test return non-zero on error.
19371 *Ulf Moeller <ulf@fitug.de>*
19373 * Add support for native Solaris shared libraries. Configure
19374 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
19375 if someone would make that last step automatic.
19377 *Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>*
19379 * ctx_size was not built with the right compiler during "make links". Fixed.
19383 * Change the meaning of 'ALL' in the cipher list. It now means "everything
19384 except NULL ciphers". This means the default cipher list will no longer
19385 enable NULL ciphers. They need to be specifically enabled e.g. with
19386 the string "DEFAULT:eNULL".
19390 * Fix to RSA private encryption routines: if p < q then it would
19391 occasionally produce an invalid result. This will only happen with
19392 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
19396 * Be less restrictive and allow also `perl util/perlpath.pl
19397 /path/to/bin/perl` in addition to `perl util/perlpath.pl /path/to/bin`,
19398 because this way one can also use an interpreter named `perl5` (which is
19399 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
19400 installed as `perl`).
19402 *Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
19404 * Let util/clean-depend.pl work also with older Perl 5.00x versions.
19406 *Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
19408 * Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
19409 advapi32.lib to Win32 build and change the pem test comparison
19410 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
19411 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
19412 and crypto/des/ede_cbcm_enc.c.
19416 * DES quad checksum was broken on big-endian architectures. Fixed.
19420 * Comment out two functions in bio.h that aren't implemented. Fix up the
19421 Win32 test batch file so it (might) work again. The Win32 test batch file
19422 is horrible: I feel ill....
19426 * Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
19427 in e_os.h. Audit of header files to check ANSI and non ANSI
19428 sections: 10 functions were absent from non ANSI section and not exported
19429 from Windows DLLs. Fixed up libeay.num for new functions.
19433 * Make `openssl version` output lines consistent.
19435 *Ralf S. Engelschall*
19437 * Fix Win32 symbol export lists for BIO functions: Added
19438 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
19439 to ms/libeay{16,32}.def.
19441 *Ralf S. Engelschall*
19443 * Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
19444 fine under Unix and passes some trivial tests I've now added. But the
19445 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
19446 added to make sure no one expects that this stuff really works in the
19447 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
19448 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
19451 *Ralf S. Engelschall*
19453 * Fix the generation of two part addresses in perl.
19455 *Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie*
19457 * Add config entry for Linux on MIPS.
19459 *John Tobey <jtobey@channel1.com>*
19461 * Make links whenever Configure is run, unless we are on Windoze.
19465 * Permit extensions to be added to CRLs using crl_section in openssl.cnf.
19466 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
19471 * Add a useful kludge to allow package maintainers to specify compiler and
19472 other platforms details on the command line without having to patch the
19473 Configure script every time: One now can use
19474 `perl Configure <id>:<details>`,
19475 i.e. platform ids are allowed to have details appended
19476 to them (separated by colons). This is treated as there would be a static
19477 pre-configured entry in Configure's %table under key `<id>` with value
19478 `<details>` and `perl Configure <id>` is called. So, when you want to
19479 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
19480 assembler stuff you can use `perl Configure "FreeBSD-elf:pgcc:-O6:::"`
19481 now, which overrides the FreeBSD-elf entry on-the-fly.
19483 *Ralf S. Engelschall*
19485 * Disable new TLS1 ciphersuites by default: they aren't official yet.
19489 * Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
19490 on the `perl Configure ...` command line. This way one can compile
19491 OpenSSL libraries with Position Independent Code (PIC) which is needed
19492 for linking it into DSOs.
19494 *Ralf S. Engelschall*
19496 * Remarkably, export ciphers were totally broken and no-one had noticed!
19501 * Cleaned up the LICENSE document: The official contact for any license
19502 questions now is the OpenSSL core team under openssl-core@openssl.org.
19503 And add a paragraph about the dual-license situation to make sure people
19504 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
19505 to the OpenSSL toolkit.
19507 *Ralf S. Engelschall*
19509 * General source tree makefile cleanups: Made `making xxx in yyy...`
19510 display consistent in the source tree and replaced `/bin/rm` by `rm`.
19511 Additionally cleaned up the `make links` target: Remove unnecessary
19512 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
19513 to speed processing and no longer clutter the display with confusing
19514 stuff. Instead only the actually done links are displayed.
19516 *Ralf S. Engelschall*
19518 * Permit null encryption ciphersuites, used for authentication only. It used
19519 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
19520 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
19525 * Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
19526 signed attributes when verifying signatures (this would break them),
19527 the detached data encoding was wrong and public keys obtained using
19528 X509_get_pubkey() weren't freed.
19532 * Add text documentation for the BUFFER functions. Also added a work around
19533 to a Win95 console bug. This was triggered by the password read stuff: the
19534 last character typed gets carried over to the next fread(). If you were
19535 generating a new cert request using 'req' for example then the last
19536 character of the passphrase would be CR which would then enter the first
19541 * Added the new 'Includes OpenSSL Cryptography Software' button as
19542 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
19543 button and can be used by applications based on OpenSSL to show the
19544 relationship to the OpenSSL project.
19546 *Ralf S. Engelschall*
19548 * Remove confusing variables in function signatures in files
19549 ssl/ssl_lib.c and ssl/ssl.h.
19551 *Lennart Bong <lob@kulthea.stacken.kth.se>*
19553 * Don't install bss_file.c under PREFIX/include/
19555 *Lennart Bong <lob@kulthea.stacken.kth.se>*
19557 * Get the Win32 compile working again. Modify mkdef.pl so it can handle
19558 functions that return function pointers and has support for NT specific
19559 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
19560 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
19561 unsigned to signed types: this was killing the Win32 compile.
19565 * Add new certificate file to stack functions,
19566 SSL_add_dir_cert_subjects_to_stack() and
19567 SSL_add_file_cert_subjects_to_stack(). These largely supplant
19568 SSL_load_client_CA_file(), and can be used to add multiple certs easily
19569 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
19570 This means that Apache-SSL and similar packages don't have to mess around
19571 to add as many CAs as they want to the preferred list.
19575 * Experiment with doxygen documentation. Currently only partially applied to
19577 See <http://www.stack.nl/~dimitri/doxygen/index.html>, and run doxygen with
19578 openssl.doxy as the configuration file.
19582 * Get rid of remaining C++-style comments which strict C compilers hate.
19584 *Ralf S. Engelschall, pointed out by Carlos Amengual*
19586 * Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
19587 compiled in by default: it has problems with large keys.
19591 * Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
19592 DH private keys and/or callback functions which directly correspond to
19593 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
19594 is needed for applications which have to configure certificates on a
19595 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
19597 For the RSA certificate situation is makes no difference, but
19598 for the DSA certificate situation this fixes the "no shared cipher"
19599 problem where the OpenSSL cipher selection procedure failed because the
19600 temporary keys were not overtaken from the context and the API provided
19601 no way to reconfigure them.
19602 The new functions now let applications reconfigure the stuff and they
19603 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
19604 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
19605 non-public-API function ssl_cert_instantiate() is used as a helper
19606 function and also to reduce code redundancy inside ssl_rsa.c.
19608 *Ralf S. Engelschall*
19610 * Move s_server -dcert and -dkey options out of the undocumented feature
19611 area because they are useful for the DSA situation and should be
19612 recognized by the users.
19614 *Ralf S. Engelschall*
19616 * Fix the cipher decision scheme for export ciphers: the export bits are
19617 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
19618 SSL_EXP_MASK. So, the original variable has to be used instead of the
19619 already masked variable.
19621 *Richard Levitte <levitte@stacken.kth.se>*
19623 * Fix `port` variable from `int` to `unsigned int` in crypto/bio/b_sock.c
19625 *Richard Levitte <levitte@stacken.kth.se>*
19627 * Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
19628 from `int` to `unsigned int` because it is a length and initialized by
19629 EVP_DigestFinal() which expects an `unsigned int *`.
19631 *Richard Levitte <levitte@stacken.kth.se>*
19633 * Don't hard-code path to Perl interpreter on shebang line of Configure
19634 script. Instead use the usual Shell->Perl transition trick.
19636 *Ralf S. Engelschall*
19638 * Make `openssl x509 -noout -modulus`' functional also for DSA certificates
19639 (in addition to RSA certificates) to match the behaviour of `openssl dsa
19640 -noout -modulus` as it's already the case for `openssl rsa -noout
19641 -modulus`. For RSA the -modulus is the real "modulus" while for DSA
19642 currently the public key is printed (a decision which was already done by
19643 `openssl dsa -modulus` in the past) which serves a similar purpose.
19644 Additionally the NO_RSA no longer completely removes the whole -modulus
19645 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
19648 *Ralf S. Engelschall*
19650 * Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
19651 BIO. See the source (crypto/evp/bio_ok.c) for more info.
19653 *Arne Ansper <arne@ats.cyber.ee>*
19655 * Dump the old yucky req code that tried (and failed) to allow raw OIDs
19656 to be added. Now both 'req' and 'ca' can use new objects defined in the
19661 * Add cool BIO that does syslog (or event log on NT).
19663 *Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie*
19665 * Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
19666 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
19667 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
19668 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
19672 * Add preliminary config info for new extension code.
19676 * Make RSA_NO_PADDING really use no padding.
19678 *Ulf Moeller <ulf@fitug.de>*
19680 * Generate errors when private/public key check is done.
19684 * Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
19685 for some CRL extensions and new objects added.
19689 * Really fix the ASN1 IMPLICIT bug this time... Partial support for private
19690 key usage extension and fuller support for authority key id.
19694 * Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
19695 padding method for RSA, which is recommended for new applications in PKCS
19696 #1 v2.0 (RFC 2437, October 1998).
19697 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
19698 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
19699 against Bleichbacher's attack on RSA.
19700 *Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
19703 * Updates to the new SSL compression code
19705 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
19707 * Fix so that the version number in the master secret, when passed
19708 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
19709 (because the server will not accept higher), that the version number
19710 is 0x03,0x01, not 0x03,0x00
19712 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
19714 * Run extensive memory leak checks on SSL commands. Fixed *lots* of memory
19715 leaks in `ssl/` relating to new `X509_get_pubkey()` behaviour. Also fixes
19716 in `apps/` and an unrelated leak in `crypto/dsa/dsa_vrf.c`.
19720 * Support for RAW extensions where an arbitrary extension can be
19721 created by including its DER encoding. See `apps/openssl.cnf` for
19726 * Make sure latest Perl versions don't interpret some generated C array
19727 code as Perl array code in the crypto/err/err_genc.pl script.
19729 *Lars Weber <3weber@informatik.uni-hamburg.de>*
19731 * Modify ms/do_ms.bat to not generate assembly language makefiles since
19732 not many people have the assembler. Various Win32 compilation fixes and
19733 update to the INSTALL.W32 file with (hopefully) more accurate Win32
19734 build instructions.
19738 * Modify configure script 'Configure' to automatically create crypto/date.h
19739 file under Win32 and also build pem.h from pem.org. New script
19740 util/mkfiles.pl to create the MINFO file on environments that can't do a
19741 'make files': perl util/mkfiles.pl >MINFO should work.
19745 * Major rework of DES function declarations, in the pursuit of correctness
19746 and purity. As a result, many evil casts evaporated, and some weirdness,
19747 too. You may find this causes warnings in your code. Zapping your evil
19748 casts will probably fix them. Mostly.
19752 * Fix for a typo in asn1.h. Bug fix to object creation script
19753 obj_dat.pl. It considered a zero in an object definition to mean
19754 "end of object": none of the objects in objects.h have any zeros
19755 so it wasn't spotted.
19757 *Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>*
19759 * Add support for Triple DES Cipher Block Chaining with Output Feedback
19760 Masking (CBCM). In the absence of test vectors, the best I have been able
19761 to do is check that the decrypt undoes the encrypt, so far. Send me test
19762 vectors if you have them.
19766 * Correct calculation of key length for export ciphers (too much space was
19767 allocated for null ciphers). This has not been tested!
19771 * Modifications to the mkdef.pl for Win32 DEF file creation. The usage
19772 message is now correct (it understands "crypto" and "ssl" on its
19773 command line). There is also now an "update" option. This will update
19774 the util/ssleay.num and util/libeay.num files with any new functions.
19776 perl util/mkdef.pl crypto ssl update
19777 it will update them.
19781 * Overhauled the Perl interface:
19782 - ported BN stuff to OpenSSL's different BN library
19783 - made the perl/ source tree CVS-aware
19784 - renamed the package from SSLeay to OpenSSL (the files still contain
19785 their history because I've copied them in the repository)
19786 - removed obsolete files (the test scripts will be replaced
19787 by better Test::Harness variants in the future)
19789 *Ralf S. Engelschall*
19791 * First cut for a very conservative source tree cleanup:
19792 1. merge various obsolete readme texts into doc/ssleay.txt
19793 where we collect the old documents and readme texts.
19794 2. remove the first part of files where I'm already sure that we no
19795 longer need them because of three reasons: either they are just temporary
19796 files which were left by Eric or they are preserved original files where
19797 I've verified that the diff is also available in the CVS via "cvs diff
19798 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
19799 the crypto/md/ stuff).
19801 *Ralf S. Engelschall*
19803 * More extension code. Incomplete support for subject and issuer alt
19804 name, issuer and authority key id. Change the i2v function parameters
19805 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
19806 what that's for :-) Fix to ASN1 macro which messed up
19807 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
19811 * Preliminary support for ENUMERATED type. This is largely copied from the
19816 * Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
19818 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
19820 * Make sure `make rehash` target really finds the `openssl` program.
19822 *Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
19824 * Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
19825 like to hear about it if this slows down other processors.
19829 * Add CygWin32 platform information to Configure script.
19831 *Alan Batie <batie@aahz.jf.intel.com>*
19833 * Fixed ms/32all.bat script: `no_asm` -> `no-asm`
19835 *Rainer W. Gerling <gerling@mpg-gv.mpg.de>*
19837 * New program nseq to manipulate netscape certificate sequences
19841 * Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
19846 * Fixes to BN code. Previously the default was to define BN_RECURSION
19847 but the BN code had some problems that would cause failures when
19848 doing certificate verification and some other functions.
19850 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
19852 * Add ASN1 and PEM code to support netscape certificate sequences.
19856 * Add ASN1 and PEM code to support netscape certificate sequences.
19860 * Add several PKIX and private extended key usage OIDs.
19864 * Modify the 'ca' program to handle the new extension code. Modify
19865 openssl.cnf for new extension format, add comments.
19869 * More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
19870 and add a sample to openssl.cnf so req -x509 now adds appropriate
19875 * Continued X509 V3 changes. Add to other makefiles, integrate with the
19876 error code, add initial support to X509_print() and x509 application.
19880 * Takes a deep breath and start adding X509 V3 extension support code. Add
19881 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
19882 stuff is currently isolated and isn't even compiled yet.
19886 * Continuing patches for GeneralizedTime. Fix up certificate and CRL
19887 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
19888 Removed the versions check from X509 routines when loading extensions:
19889 this allows certain broken certificates that don't set the version
19890 properly to be processed.
19894 * Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
19895 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
19896 can still be regenerated with "make depend".
19900 * Spelling mistake in C version of CAST-128.
19902 *Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>*
19904 * Changes to the error generation code. The perl script err-code.pl
19905 now reads in the old error codes and retains the old numbers, only
19906 adding new ones if necessary. It also only changes the .err files if new
19907 codes are added. The makefiles have been modified to only insert errors
19908 when needed (to avoid needlessly modifying header files). This is done
19909 by only inserting errors if the .err file is newer than the auto generated
19910 C file. To rebuild all the error codes from scratch (the old behaviour)
19911 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
19912 or delete all the .err files.
19916 * CAST-128 was incorrectly implemented for short keys. The C version has
19917 been fixed, but is untested. The assembler versions are also fixed, but
19918 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
19919 to regenerate it if needed.
19920 *Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
19921 Hagino <itojun@kame.net>*
19923 * File was opened incorrectly in randfile.c.
19925 *Ulf Möller <ulf@fitug.de>*
19927 * Beginning of support for GeneralizedTime. d2i, i2d, check and print
19928 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
19929 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
19930 al: it's just almost always a UTCTime. Note this patch adds new error
19931 codes so do a "make errors" if there are problems.
19935 * Correct Linux 1 recognition in config.
19937 *Ulf Möller <ulf@fitug.de>*
19939 * Remove pointless MD5 hash when using DSA keys in ca.
19941 *Anonymous <nobody@replay.com>*
19943 * Generate an error if given an empty string as a cert directory. Also
19944 generate an error if handed NULL (previously returned 0 to indicate an
19945 error, but didn't set one).
19947 *Ben Laurie, reported by Anonymous <nobody@replay.com>*
19949 * Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
19953 * Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
19954 parameters. This was causing a warning which killed off the Win32 compile.
19958 * Remove C++ style comments from crypto/bn/bn_local.h.
19960 *Neil Costigan <neil.costigan@celocom.com>*
19962 * The function OBJ_txt2nid was broken. It was supposed to return a nid
19963 based on a text string, looking up short and long names and finally
19964 "dot" format. The "dot" format stuff didn't work. Added new function
19965 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
19966 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
19967 OID is not part of the table.
19971 * Add prototypes to X509 lookup/verify methods, fixing a bug in
19972 X509_LOOKUP_by_alias().
19976 * Sort openssl functions by name.
19980 * Get the `gendsa` command working and add it to the `list` command. Remove
19981 encryption from sample DSA keys (in case anyone is interested the password
19986 * Make *all* `*_free` functions accept a NULL pointer.
19988 *Frans Heymans <fheymans@isaserver.be>*
19990 * If a DH key is generated in s3_srvr.c, don't blow it by trying to use
19993 *Anonymous <nobody@replay.com>*
19995 * s_server should send the CAfile as acceptable CAs, not its own cert.
19997 *Bodo Moeller <3moeller@informatik.uni-hamburg.de>*
19999 * Don't blow it for numeric `-newkey` arguments to `apps/req`.
20001 *Bodo Moeller <3moeller@informatik.uni-hamburg.de>*
20003 * Temp key "for export" tests were wrong in s3_srvr.c.
20005 *Anonymous <nobody@replay.com>*
20007 * Add prototype for temp key callback functions
20008 SSL_CTX_set_tmp_{rsa,dh}_callback().
20012 * Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
20013 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
20017 * X509_name_add_entry() freed the wrong thing after an error.
20019 *Arne Ansper <arne@ats.cyber.ee>*
20021 * rsa_eay.c would attempt to free a NULL context.
20023 *Arne Ansper <arne@ats.cyber.ee>*
20025 * BIO_s_socket() had a broken should_retry() on Windoze.
20027 *Arne Ansper <arne@ats.cyber.ee>*
20029 * BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
20031 *Arne Ansper <arne@ats.cyber.ee>*
20033 * Make sure the already existing X509_STORE->depth variable is initialized
20034 in X509_STORE_new(), but document the fact that this variable is still
20035 unused in the certificate verification process.
20037 *Ralf S. Engelschall*
20039 * Fix the various library and `apps/` files to free up pkeys obtained from
20040 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
20044 * Fix reference counting in X509_PUBKEY_get(). This makes
20045 demos/maurice/example2.c work, amongst others, probably.
20047 *Steve Henson and Ben Laurie*
20049 * First cut of a cleanup for `apps/`. First the `ssleay` program is now named
20050 `openssl` and second, the shortcut symlinks for the `openssl <command>`
20051 are no longer created. This way we have a single and consistent command
20052 line interface `openssl <command>`, similar to `cvs <command>`.
20054 *Ralf S. Engelschall, Paul Sutton and Ben Laurie*
20056 * ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
20057 BIT STRING wrapper always have zero unused bits.
20061 * Add CA.pl, perl version of CA.sh, add extended key usage OID.
20065 * Make the top-level INSTALL documentation easier to understand.
20069 * Makefiles updated to exit if an error occurs in a sub-directory
20070 make (including if user presses ^C) [Paul Sutton]
20072 * Make Montgomery context stuff explicit in RSA data structure.
20076 * Fix build order of pem and err to allow for generated pem.h.
20080 * Fix renumbering bug in X509_NAME_delete_entry().
20084 * Enhanced the err-ins.pl script so it makes the error library number
20085 global and can add a library name. This is needed for external ASN1 and
20086 other error libraries.
20090 * Fixed sk_insert which never worked properly.
20094 * Fix ASN1 macros so they can handle indefinite length constructed
20095 EXPLICIT tags. Some non standard certificates use these: they can now
20100 * Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
20101 into a single doc/ssleay.txt bundle. This way the information is still
20102 preserved but no longer messes up this directory. Now it's new room for
20103 the new set of documentation files.
20105 *Ralf S. Engelschall*
20107 * SETs were incorrectly DER encoded. This was a major pain, because they
20108 shared code with SEQUENCEs, which aren't coded the same. This means that
20109 almost everything to do with SETs or SEQUENCEs has either changed name or
20110 number of arguments.
20112 *Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>*
20114 * Fix test data to work with the above.
20118 * Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
20119 was already fixed by Eric for 0.9.1 it seems.
20121 *Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>*
20123 * Autodetect FreeBSD3.
20127 * Fix various bugs in Configure. This affects the following platforms:
20131 unixware-2.0-pentium
20136 * Eliminate generated files from CVS. Reorder tests to regenerate files
20137 before they are needed.
20141 * Generate Makefile.ssl from Makefile.org (to keep CVS happy).
20145 ### Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
20147 * Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
20148 changed SSLeay to OpenSSL in version strings.
20150 *Ralf S. Engelschall*
20152 * Some fixups to the top-level documents.
20156 * Fixed the nasty bug where rsaref.h was not found under compile-time
20157 because the symlink to include/ was missing.
20159 *Ralf S. Engelschall*
20161 * Incorporated the popular no-RSA/DSA-only patches
20162 which allow to compile an RSA-free SSLeay.
20164 *Andrew Cooke / Interrader Ldt., Ralf S. Engelschall*
20166 * Fixed nasty rehash problem under `make -f Makefile.ssl links`
20167 when "ssleay" is still not found.
20169 *Ralf S. Engelschall*
20171 * Added more platforms to Configure: Cray T3E, HPUX 11,
20173 *Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>*
20175 * Updated the README file.
20177 *Ralf S. Engelschall*
20179 * Added various .cvsignore files in the CVS repository subdirs
20180 to make a "cvs update" really silent.
20182 *Ralf S. Engelschall*
20184 * Recompiled the error-definition header files and added
20185 missing symbols to the Win32 linker tables.
20187 *Ralf S. Engelschall*
20189 * Cleaned up the top-level documents;
20190 o new files: CHANGES and LICENSE
20191 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
20192 o merged COPYRIGHT into LICENSE
20193 o removed obsolete TODO file
20194 o renamed MICROSOFT to INSTALL.W32
20196 *Ralf S. Engelschall*
20198 * Removed dummy files from the 0.9.1b source tree:
20199 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
20200 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
20201 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
20202 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
20203 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
20205 *Ralf S. Engelschall*
20207 * Added various platform portability fixes.
20211 * The Genesis of the OpenSSL rpject:
20212 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
20213 Young and Tim J. Hudson created while they were working for C2Net until
20216 *The OpenSSL Project*
20218 ### Changes between 0.9.0b and 0.9.1b [not released]
20220 * Updated a few CA certificates under certs/
20224 * Changed some BIGNUM api stuff.
20228 * Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
20229 DGUX x86, Linux Alpha, etc.
20233 * New COMP library [crypto/comp/] for SSL Record Layer Compression:
20234 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
20239 * Add -strparse option to asn1pars program which parses nested
20242 *Dr Stephen Henson <shenson@bigfoot.com>*
20244 * Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
20248 * DSA fix for "ca" program.
20252 * Added "-genkey" option to "dsaparam" program.
20256 * Added RIPE MD160 (rmd160) message digest.
20260 * Added -a (all) option to "ssleay version" command.
20264 * Added PLATFORM define which is the id given to Configure.
20268 * Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
20272 * Extended the ASN.1 parser routines.
20276 * Extended BIO routines to support REUSEADDR, seek, tell, etc.
20280 * Added a BN_CTX to the BN library.
20284 * Fixed the weak key values in DES library
20288 * Changed API in EVP library for cipher aliases.
20292 * Added support for RC2/64bit cipher.
20296 * Converted the lhash library to the crypto/mem.c functions.
20300 * Added more recognized ASN.1 object ids.
20304 * Added more RSA padding checks for SSL/TLS.
20308 * Added BIO proxy/filter functionality.
20312 * Added extra_certs to SSL_CTX which can be used
20313 send extra CA certificates to the client in the CA cert chain sending
20314 process. It can be configured with SSL_CTX_add_extra_chain_cert().
20318 * Now Fortezza is denied in the authentication phase because
20319 this is key exchange mechanism is not supported by SSLeay at all.
20323 * Additional PKCS1 checks.
20327 * Support the string "TLSv1" for all TLS v1 ciphers.
20331 * Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
20332 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
20336 * Fixed a few memory leaks.
20340 * Fixed various code and comment typos.
20344 * A minor bug in ssl/s3_clnt.c where there would always be 4 0
20345 bytes sent in the client random.
20347 *Edward Bishop <ebishop@spyglass.com>*
20351 [CVE-2023-5678]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-5678
20352 [CVE-2023-5363]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-5363
20353 [CVE-2023-4807]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-4807
20354 [CVE-2023-3817]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-3817
20355 [CVE-2023-3446]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-3446
20356 [CVE-2023-2975]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-2975
20357 [RFC 2578 (STD 58), section 3.5]: https://datatracker.ietf.org/doc/html/rfc2578#section-3.5
20358 [CVE-2023-2650]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-2650
20359 [CVE-2023-1255]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-1255
20360 [CVE-2023-0466]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-0466
20361 [CVE-2023-0465]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-0465
20362 [CVE-2023-0464]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-0464
20363 [CVE-2023-0401]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-0401
20364 [CVE-2023-0286]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-0286
20365 [CVE-2023-0217]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-0217
20366 [CVE-2023-0216]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-0216
20367 [CVE-2023-0215]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-0215
20368 [CVE-2022-4450]: https://www.openssl.org/news/vulnerabilities.html#CVE-2022-4450
20369 [CVE-2022-4304]: https://www.openssl.org/news/vulnerabilities.html#CVE-2022-4304
20370 [CVE-2022-4203]: https://www.openssl.org/news/vulnerabilities.html#CVE-2022-4203
20371 [CVE-2022-3996]: https://www.openssl.org/news/vulnerabilities.html#CVE-2022-3996
20372 [CVE-2022-2274]: https://www.openssl.org/news/vulnerabilities.html#CVE-2022-2274
20373 [CVE-2022-2097]: https://www.openssl.org/news/vulnerabilities.html#CVE-2022-2097
20374 [CVE-2020-1971]: https://www.openssl.org/news/vulnerabilities.html#CVE-2020-1971
20375 [CVE-2020-1967]: https://www.openssl.org/news/vulnerabilities.html#CVE-2020-1967
20376 [CVE-2019-1563]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1563
20377 [CVE-2019-1559]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1559
20378 [CVE-2019-1552]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1552
20379 [CVE-2019-1551]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1551
20380 [CVE-2019-1549]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1549
20381 [CVE-2019-1547]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1547
20382 [CVE-2019-1543]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1543
20383 [CVE-2018-5407]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-5407
20384 [CVE-2018-0739]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0739
20385 [CVE-2018-0737]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0737
20386 [CVE-2018-0735]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0735
20387 [CVE-2018-0734]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0734
20388 [CVE-2018-0733]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0733
20389 [CVE-2018-0732]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0732
20390 [CVE-2017-3738]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3738
20391 [CVE-2017-3737]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3737
20392 [CVE-2017-3736]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3736
20393 [CVE-2017-3735]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3735
20394 [CVE-2017-3733]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3733
20395 [CVE-2017-3732]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3732
20396 [CVE-2017-3731]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3731
20397 [CVE-2017-3730]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3730
20398 [CVE-2016-7055]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7055
20399 [CVE-2016-7054]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7054
20400 [CVE-2016-7053]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7053
20401 [CVE-2016-7052]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7052
20402 [CVE-2016-6309]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6309
20403 [CVE-2016-6308]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6308
20404 [CVE-2016-6307]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6307
20405 [CVE-2016-6306]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6306
20406 [CVE-2016-6305]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6305
20407 [CVE-2016-6304]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6304
20408 [CVE-2016-6303]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6303
20409 [CVE-2016-6302]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6302
20410 [CVE-2016-2183]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2183
20411 [CVE-2016-2182]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2182
20412 [CVE-2016-2181]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2181
20413 [CVE-2016-2180]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2180
20414 [CVE-2016-2179]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2179
20415 [CVE-2016-2178]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2178
20416 [CVE-2016-2177]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2177
20417 [CVE-2016-2176]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2176
20418 [CVE-2016-2109]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2109
20419 [CVE-2016-2107]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2107
20420 [CVE-2016-2106]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2106
20421 [CVE-2016-2105]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2105
20422 [CVE-2016-0800]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0800
20423 [CVE-2016-0799]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0799
20424 [CVE-2016-0798]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0798
20425 [CVE-2016-0797]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0797
20426 [CVE-2016-0705]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0705
20427 [CVE-2016-0702]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0702
20428 [CVE-2016-0701]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0701
20429 [CVE-2015-3197]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3197
20430 [CVE-2015-3196]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3196
20431 [CVE-2015-3195]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3195
20432 [CVE-2015-3194]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3194
20433 [CVE-2015-3193]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3193
20434 [CVE-2015-1793]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1793
20435 [CVE-2015-1792]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1792
20436 [CVE-2015-1791]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1791
20437 [CVE-2015-1790]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1790
20438 [CVE-2015-1789]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1789
20439 [CVE-2015-1788]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1788
20440 [CVE-2015-1787]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1787
20441 [CVE-2015-0293]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0293
20442 [CVE-2015-0291]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0291
20443 [CVE-2015-0290]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0290
20444 [CVE-2015-0289]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0289
20445 [CVE-2015-0288]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0288
20446 [CVE-2015-0287]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0287
20447 [CVE-2015-0286]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0286
20448 [CVE-2015-0285]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0285
20449 [CVE-2015-0209]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0209
20450 [CVE-2015-0208]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0208
20451 [CVE-2015-0207]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0207
20452 [CVE-2015-0206]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0206
20453 [CVE-2015-0205]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0205
20454 [CVE-2015-0204]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0204
20455 [CVE-2014-8275]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-8275
20456 [CVE-2014-5139]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-5139
20457 [CVE-2014-3572]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3572
20458 [CVE-2014-3571]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3571
20459 [CVE-2014-3570]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3570
20460 [CVE-2014-3569]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3569
20461 [CVE-2014-3568]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3568
20462 [CVE-2014-3567]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3567
20463 [CVE-2014-3566]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3566
20464 [CVE-2014-3513]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3513
20465 [CVE-2014-3512]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3512
20466 [CVE-2014-3511]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3511
20467 [CVE-2014-3510]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3510
20468 [CVE-2014-3509]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3509
20469 [CVE-2014-3508]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3508
20470 [CVE-2014-3507]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3507
20471 [CVE-2014-3506]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3506
20472 [CVE-2014-3505]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3505
20473 [CVE-2014-3470]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3470
20474 [CVE-2014-0224]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0224
20475 [CVE-2014-0221]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0221
20476 [CVE-2014-0195]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0195
20477 [CVE-2014-0160]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0160
20478 [CVE-2014-0076]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0076
20479 [CVE-2013-6450]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-6450
20480 [CVE-2013-4353]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-4353
20481 [CVE-2013-0169]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-0169
20482 [CVE-2013-0166]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-0166
20483 [CVE-2012-2686]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2686
20484 [CVE-2012-2333]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2333
20485 [CVE-2012-2110]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2110
20486 [CVE-2012-0884]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0884
20487 [CVE-2012-0050]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0050
20488 [CVE-2012-0027]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0027
20489 [CVE-2011-4619]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4619
20490 [CVE-2011-4577]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4577
20491 [CVE-2011-4576]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4576
20492 [CVE-2011-4109]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4109
20493 [CVE-2011-4108]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4108
20494 [CVE-2011-3210]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-3210
20495 [CVE-2011-3207]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-3207
20496 [CVE-2011-0014]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-0014
20497 [CVE-2010-4252]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-4252
20498 [CVE-2010-4180]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-4180
20499 [CVE-2010-3864]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-3864
20500 [CVE-2010-1633]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-1633
20501 [CVE-2010-0740]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-0740
20502 [CVE-2010-0433]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-0433
20503 [CVE-2009-4355]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-4355
20504 [CVE-2009-3555]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-3555
20505 [CVE-2009-3245]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-3245
20506 [CVE-2009-1386]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1386
20507 [CVE-2009-1379]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1379
20508 [CVE-2009-1378]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1378
20509 [CVE-2009-1377]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1377
20510 [CVE-2009-0789]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0789
20511 [CVE-2009-0591]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0591
20512 [CVE-2009-0590]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0590
20513 [CVE-2008-5077]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-5077
20514 [CVE-2008-1678]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-1678
20515 [CVE-2008-1672]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-1672
20516 [CVE-2008-0891]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-0891
20517 [CVE-2007-5135]: https://www.openssl.org/news/vulnerabilities.html#CVE-2007-5135
20518 [CVE-2007-4995]: https://www.openssl.org/news/vulnerabilities.html#CVE-2007-4995
20519 [CVE-2006-4343]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-4343
20520 [CVE-2006-4339]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-4339
20521 [CVE-2006-3738]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-3738
20522 [CVE-2006-2940]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-2940
20523 [CVE-2006-2937]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-2937
20524 [CVE-2005-2969]: https://www.openssl.org/news/vulnerabilities.html#CVE-2005-2969
20525 [CVE-2004-0112]: https://www.openssl.org/news/vulnerabilities.html#CVE-2004-0112
20526 [CVE-2004-0079]: https://www.openssl.org/news/vulnerabilities.html#CVE-2004-0079
20527 [CVE-2003-0851]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0851
20528 [CVE-2003-0545]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0545
20529 [CVE-2003-0544]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0544
20530 [CVE-2003-0543]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0543
20531 [CVE-2003-0078]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0078
20532 [CVE-2002-0659]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0659
20533 [CVE-2002-0657]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0657
20534 [CVE-2002-0656]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0656
20535 [CVE-2002-0655]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0655