4 This is a high-level summary of the most important changes.
5 For a full list of changes, see the [git commit log][log] and
6 pick the appropriate release branch.
8 [log]: https://github.com/openssl/openssl/commits/
13 - [OpenSSL 3.2](#openssl-32)
14 - [OpenSSL 3.1](#openssl-31)
15 - [OpenSSL 3.0](#openssl-30)
16 - [OpenSSL 1.1.1](#openssl-111)
17 - [OpenSSL 1.1.0](#openssl-110)
18 - [OpenSSL 1.0.2](#openssl-102)
19 - [OpenSSL 1.0.1](#openssl-101)
20 - [OpenSSL 1.0.0](#openssl-100)
21 - [OpenSSL 0.9.x](#openssl-09x)
26 ### Changes between 3.1 and 3.2 [xx XXX xxxx]
28 * Added client side support for QUIC
30 *Hugo Landau, Matt Caswell, Paul Dale, Tomáš Mráz, Richard Levitte*
32 * Added multiple tutorials on the OpenSSL library and in particular
33 on writing various clients (using TLS and QUIC protocols) with libssl.
37 * Added secp384r1 implementation using Solinas' reduction to improve
38 speed of the NIST P-384 elliptic curve. To enable the implementation
39 the build option `enable-ec_nistp_64_gcc_128` must be used.
43 * Improved RFC7468 compliance of the asn1parse command.
47 * Added SHA256/192 algorithm support.
51 * Added support for securely getting root CA certificate update in
56 * Improved contention on global write locks by using more read locks where
61 * Improved performance of OSSL_PARAM lookups in performance critical
66 * Added the SSL_get0_group_name() function to provide access to the
67 name of the group used for the TLS key exchange.
71 * Provide a new configure option `no-http` that can be used to disable the
72 HTTP support. Provide new configure options `no-apps` and `no-docs` to
73 disable building the openssl command line application and the documentation.
77 * Provide a new configure option `no-ecx` that can be used to disable the
78 X25519, X448, and EdDSA support.
82 * When multiple OSSL_KDF_PARAM_INFO parameters are passed to
83 the EVP_KDF_CTX_set_params() function they are now concatenated not just
84 for the HKDF algorithm but also for SSKDF and X9.63 KDF algorithms.
88 * Added OSSL_FUNC_keymgmt_im/export_types_ex() provider functions that get
89 the provider context as a parameter.
93 * TLS round-trip time calculation was added by a Brigham Young University
94 Capstone team partnering with Sandia National Laboratories. A new function
95 in ssl_lib titled SSL_get_handshake_rtt will calculate and retrieve this
100 * Added the "-quic" option to s_client to enable connectivity to QUIC servers.
101 QUIC requires the use of ALPN, so this must be specified via the "-alpn"
102 option. Use of the "advanced" s_client command command via the "-adv" option
107 * Added an "advanced" command mode to s_client. Use this with the "-adv"
108 option. The old "basic" command mode recognises certain letters that must
109 always appear at the start of a line and cannot be escaped. The advanced
110 command mode enables commands to be entered anywhere and there is an
111 escaping mechanism. After starting s_client with "-adv" type "{help}"
112 to show a list of available commands.
116 * Add Raw Public Key (RFC7250) support. Authentication is supported
117 by matching keys against either local policy (TLSA records synthesised
118 from the expected keys) or DANE (TLSA records obtained by the
119 application from DNS). TLSA records will also match the same key in
120 the server certificate, should RPK use not happen to be negotiated.
124 * Added support for modular exponentiation and CRT offloading for the
129 * Added further assembler code for the RISC-V architecture.
133 * Added EC_GROUP_to_params() which creates an OSSL_PARAM array
134 from a given EC_GROUP.
138 * Improved support for non-default library contexts and property queries
139 when parsing PKCS#12 files.
143 * Implemented support for all five instances of EdDSA from RFC8032:
144 Ed25519, Ed25519ctx, Ed25519ph, Ed448, and Ed448ph.
145 The streaming is not yet supported for the HashEdDSA variants
146 (Ed25519ph and Ed448ph).
150 * Added SM4 optimization for ARM processors using ASIMD and AES HW
155 * Implemented SM4-XTS support.
159 * Added platform-agnostic OSSL_sleep() function.
163 * Implemented deterministic ECDSA signatures (RFC6979) support.
167 * Implemented AES-GCM-SIV (RFC8452) support.
171 * Added support for pluggable (provider-based) TLS signature algorithms.
172 This enables TLS 1.3 authentication operations with algorithms embedded
173 in providers not included by default in OpenSSL. In combination with
174 the already available pluggable KEM and X.509 support, this enables
175 for example suitable providers to deliver post-quantum or quantum-safe
176 cryptography to OpenSSL users.
180 * Added support for pluggable (provider-based) CMS signature algorithms.
181 This enables CMS sign and verify operations with algorithms embedded
182 in providers not included by default in OpenSSL.
186 * Added support for Hybrid Public Key Encryption (HPKE) as defined
187 in RFC9180. HPKE is required for TLS Encrypted ClientHello (ECH),
188 Message Layer Security (MLS) and other IETF specifications.
189 HPKE can also be used by other applications that require
190 encrypting "to" an ECDH public key. External APIs are defined in
191 include/openssl/hpke.h and documented in doc/man3/OSSL_HPKE_CTX_new.pod
195 * Implemented HPKE DHKEM support in providers used by HPKE (RFC9180)
200 * Add support for certificate compression (RFC8879), including
201 library support for Brotli and Zstandard compression.
205 * Add the ability to add custom attributes to PKCS12 files. Add a new API
206 PKCS12_create_ex2, identical to the existing PKCS12_create_ex but allows
207 for a user specified callback and optional argument.
208 Added a new PKCS12_SAFEBAG_set0_attr, which allows for a new attr to be
209 added to the existing STACK_OF attrs.
213 * Major refactor of the libssl record layer.
217 * Add a mac salt length option for the pkcs12 command.
221 * Add more SRTP protection profiles from RFC8723 and RFC8269.
225 * Extended Kernel TLS (KTLS) to support TLS 1.3 receive offload.
227 *Daiki Ueno, John Baldwin and Dmitry Podgorny*
229 * Add support for TCP Fast Open (RFC7413) to macOS, Linux, and FreeBSD where
230 supported and enabled.
234 * Add ciphersuites based on DHE_PSK (RFC 4279) and ECDHE_PSK (RFC 5489)
235 to the list of ciphersuites providing Perfect Forward Secrecy as
236 required by SECLEVEL >= 3.
238 *Dmitry Belyavskiy, Nicola Tuveri*
240 * Add new SSL APIs to aid in efficiently implementing TLS/SSL fingerprinting.
241 The SSL_CTRL_GET_IANA_GROUPS control code, exposed as the
242 SSL_get0_iana_groups() function-like macro, retrieves the list of
243 supported groups sent by the peer.
244 The function SSL_client_hello_get_extension_order() populates
245 a caller-supplied array with the list of extension types present in the
246 ClientHello, in order of appearance.
250 * Fixed PEM_write_bio_PKCS8PrivateKey() and PEM_write_bio_PKCS8PrivateKey_nid()
251 to make it possible to use empty passphrase strings.
255 * The PKCS12_parse() function now supports MAC-less PKCS12 files.
259 * Added ASYNC_set_mem_functions() and ASYNC_get_mem_functions() calls to be able
260 to change functions used for allocating the memory of asynchronous call stack.
264 * Added support for signed BIGNUMs in the OSSL_PARAM APIs.
268 * A failure exit code is returned when using the openssl x509 command to check
269 certificate attributes and the checks fail.
273 * The default SSL/TLS security level has been changed from 1 to 2. RSA,
274 DSA and DH keys of 1024 bits and above and less than 2048 bits and ECC keys
275 of 160 bits and above and less than 224 bits were previously accepted by
276 default but are now no longer allowed. By default TLS compression was
277 already disabled in previous OpenSSL versions. At security level 2 it cannot
282 * The SSL_CTX_set_cipher_list family functions now accept ciphers using their
287 * The PVK key derivation function has been moved from b2i_PVK_bio_ex() into
288 the legacy crypto provider as an EVP_KDF. Applications requiring this KDF
289 will need to load the legacy crypto provider.
293 * CCM8 cipher suites in TLS have been downgraded to security level zero
294 because they use a short authentication tag which lowers their strength.
298 * Subject or issuer names in X.509 objects are now displayed as UTF-8 strings
303 * Add X.509 certificate codeSigning purpose and related checks on key usage and
304 extended key usage of the leaf certificate according to the CA/Browser Forum.
308 * The `x509`, `ca`, and `req` apps now produce X.509 v3 certificates.
309 The `-x509v1` option of `req` prefers generation of X.509 v1 certificates.
310 `X509_sign()` and `X509_sign_ctx()` make sure that the certificate has
311 X.509 version 3 if the certificate information includes X.509 extensions.
315 * Fix and extend certificate handling and the apps `x509`, `verify` etc.
316 such as adding a trace facility for debugging certificate chain building.
320 * Various fixes and extensions to the CMP+CRMF implementation and the `cmp` app
321 in particular supporting requests for central key generation, generalized
322 polling, and various types of genm/genp exchanges defined in CMP Updates.
326 * Fixes and extensions to the HTTP client and to the HTTP server in `apps/`
327 like correcting the TLS and proxy support and adding tracing for debugging.
331 * Extended the CMS API for handling `CMS_SignedData` and `CMS_EnvelopedData`.
335 * `CMS_add0_cert()` and `CMS_add1_cert()` no longer throw an error if
336 a certificate to be added is already present. `CMS_sign_ex()` and
337 `CMS_sign()` now ignore any duplicate certificates in their `certs` argument
338 and no longer throw an error for them.
342 * Fixed and extended `util/check-format.pl` for checking adherence to the
343 coding style <https://www.openssl.org/policies/technical/coding-style.html>.
344 The checks are meanwhile more complete and yield fewer false positives.
348 * Added BIO_s_dgram_pair() and BIO_s_dgram_mem() that provide memory-based
349 BIOs with datagram semantics and support for BIO_sendmmsg() and BIO_recvmmsg()
350 calls. They can be used as the transport BIOs for QUIC.
352 *Hugo Landau, Matt Caswell and Tomáš Mráz*
354 * Add new BIO_sendmmsg() and BIO_recvmmsg() BIO methods which allow
355 sending and receiving multiple messages in a single call. An implementation
356 is provided for BIO_dgram. For further details, see BIO_sendmmsg(3).
360 * Support for loading root certificates from the Windows certificate store
361 has been added. The support is in the form of a store which recognises the
362 URI string of `org.openssl.winstore://`. This URI scheme currently takes no
363 arguments. This store is built by default and can be disabled using the new
364 compile-time option `no-winstore`. This store is not currently used by
365 default and must be loaded explicitly using the above store URI. It is
366 expected to be loaded by default in the future.
370 * Enable KTLS with the TLS 1.3 CCM mode ciphersuites. Note that some linux
371 kernel versions that support KTLS have a known bug in CCM processing. That
372 has been fixed in stable releases starting from 5.4.164, 5.10.84, 5.15.7,
373 and all releases since 5.16. KTLS with CCM ciphersuites should be only used
378 * Added `-ktls` option to `s_server` and `s_client` commands to enable the
383 * Zerocopy KTLS sendfile() support on Linux.
387 * The OBJ_ calls are now thread safe using a global lock.
391 * New parameter `-digest` for openssl cms command allowing signing
392 pre-computed digests and new CMS API functions supporting that
397 * OPENSSL_malloc() and other allocation functions now raise errors on
398 allocation failures. The callers do not need to explicitly raise errors
399 unless they want to for tracing purposes.
403 * Added and enabled by default implicit rejection in RSA PKCS#1 v1.5
404 decryption as a protection against Bleichenbacher-like attacks.
405 The RSA decryption API will now return a randomly generated deterministic
406 message instead of an error in case it detects an error when checking
407 padding during PKCS#1 v1.5 decryption. This is a general protection against
408 issues like CVE-2020-25659 and CVE-2020-25657. This protection can be
410 `EVP_PKEY_CTX_ctrl_str(ctx, "rsa_pkcs1_implicit_rejection". "0")`
411 on the RSA decryption context.
415 * Added support for Brainpool curves in TLS-1.3.
417 *Bernd Edlinger and Matt Caswell*
419 * Added OpenBSD specific build targets.
423 * Support for Argon2d, Argon2i, Argon2id KDFs has been added along with
424 a basic thread pool implementation for select platforms.
431 ### Changes between 3.1.1 and 3.1.2 [1 Aug 2023]
433 * Fix excessive time spent checking DH q parameter value.
435 The function DH_check() performs various checks on DH parameters. After
436 fixing CVE-2023-3446 it was discovered that a large q parameter value can
437 also trigger an overly long computation during some of these checks.
438 A correct q value, if present, cannot be larger than the modulus p
439 parameter, thus it is unnecessary to perform these checks if q is larger
442 If DH_check() is called with such q parameter value,
443 DH_CHECK_INVALID_Q_VALUE return flag is set and the computationally
444 intensive checks are skipped.
450 * Fix DH_check() excessive time with over sized modulus.
452 The function DH_check() performs various checks on DH parameters. One of
453 those checks confirms that the modulus ("p" parameter) is not too large.
454 Trying to use a very large modulus is slow and OpenSSL will not normally use
455 a modulus which is over 10,000 bits in length.
457 However the DH_check() function checks numerous aspects of the key or
458 parameters that have been supplied. Some of those checks use the supplied
459 modulus value even if it has already been found to be too large.
461 A new limit has been added to DH_check of 32,768 bits. Supplying a
462 key/parameters with a modulus over this size will simply cause DH_check() to
469 * Do not ignore empty associated data entries with AES-SIV.
471 The AES-SIV algorithm allows for authentication of multiple associated
472 data entries along with the encryption. To authenticate empty data the
473 application has to call `EVP_EncryptUpdate()` (or `EVP_CipherUpdate()`)
474 with NULL pointer as the output buffer and 0 as the input buffer length.
475 The AES-SIV implementation in OpenSSL just returns success for such call
476 instead of performing the associated data authentication operation.
477 The empty data thus will not be authenticated. ([CVE-2023-2975])
479 Thanks to Juerg Wullschleger (Google) for discovering the issue.
481 The fix changes the authentication tag value and the ciphertext for
482 applications that use empty associated data entries with AES-SIV.
483 To decrypt data encrypted with previous versions of OpenSSL the application
484 has to skip calls to `EVP_DecryptUpdate()` for empty associated data
489 * When building with the `enable-fips` option and using the resulting
490 FIPS provider, TLS 1.2 will, by default, mandate the use of an extended
491 master secret (FIPS 140-3 IG G.Q) and the Hash and HMAC DRBGs will
492 not operate with truncated digests (FIPS 140-3 IG G.R).
496 ### Changes between 3.1.0 and 3.1.1 [30 May 2023]
498 * Mitigate for the time it takes for `OBJ_obj2txt` to translate gigantic
499 OBJECT IDENTIFIER sub-identifiers to canonical numeric text form.
501 OBJ_obj2txt() would translate any size OBJECT IDENTIFIER to canonical
502 numeric text form. For gigantic sub-identifiers, this would take a very
503 long time, the time complexity being O(n^2) where n is the size of that
504 sub-identifier. ([CVE-2023-2650])
506 To mitigitate this, `OBJ_obj2txt()` will only translate an OBJECT
507 IDENTIFIER to canonical numeric text form if the size of that OBJECT
508 IDENTIFIER is 586 bytes or less, and fail otherwise.
510 The basis for this restriction is [RFC 2578 (STD 58), section 3.5]. OBJECT
511 IDENTIFIER values, which stipulates that OBJECT IDENTIFIERS may have at
512 most 128 sub-identifiers, and that the maximum value that each sub-
513 identifier may have is 2^32-1 (4294967295 decimal).
515 For each byte of every sub-identifier, only the 7 lower bits are part of
516 the value, so the maximum amount of bytes that an OBJECT IDENTIFIER with
517 these restrictions may occupy is 32 * 128 / 7, which is approximately 586
522 * Multiple algorithm implementation fixes for ARM BE platforms.
526 * Added a -pedantic option to fipsinstall that adjusts the various
527 settings to ensure strict FIPS compliance rather than backwards
532 * Fixed buffer overread in AES-XTS decryption on ARM 64 bit platforms which
533 happens if the buffer size is 4 mod 5 in 16 byte AES blocks. This can
534 trigger a crash of an application using AES-XTS decryption if the memory
535 just after the buffer being decrypted is not mapped.
536 Thanks to Anton Romanov (Amazon) for discovering the issue.
541 * Reworked the Fix for the Timing Oracle in RSA Decryption ([CVE-2022-4304]).
542 The previous fix for this timing side channel turned out to cause
543 a severe 2-3x performance regression in the typical use case
544 compared to 3.0.7. The new fix uses existing constant time
545 code paths, and restores the previous performance level while
546 fully eliminating all existing timing side channels.
547 The fix was developed by Bernd Edlinger with testing support
552 * Add FIPS provider configuration option to disallow the use of
553 truncated digests with Hash and HMAC DRBGs (q.v. FIPS 140-3 IG D.R.).
554 The option '-no_drbg_truncated_digests' can optionally be
555 supplied to 'openssl fipsinstall'.
559 * Corrected documentation of X509_VERIFY_PARAM_add0_policy() to mention
560 that it does not enable policy checking. Thanks to David Benjamin for
561 discovering this issue.
566 * Fixed an issue where invalid certificate policies in leaf certificates are
567 silently ignored by OpenSSL and other certificate policy checks are skipped
568 for that certificate. A malicious CA could use this to deliberately assert
569 invalid certificate policies in order to circumvent policy checking on the
570 certificate altogether.
575 * Limited the number of nodes created in a policy tree to mitigate
576 against CVE-2023-0464. The default limit is set to 1000 nodes, which
577 should be sufficient for most installations. If required, the limit
578 can be adjusted by setting the OPENSSL_POLICY_TREE_NODES_MAX build
579 time define to a desired maximum number of nodes or zero to allow
585 ### Changes between 3.0 and 3.1.0 [14 Mar 2023]
587 * Add FIPS provider configuration option to enforce the
588 Extended Master Secret (EMS) check during the TLS1_PRF KDF.
589 The option '-ems-check' can optionally be supplied to
590 'openssl fipsinstall'.
594 * The FIPS provider includes a few non-approved algorithms for
595 backward compatibility purposes and the "fips=yes" property query
596 must be used for all algorithm fetches to ensure FIPS compliance.
598 The algorithms that are included but not approved are Triple DES ECB,
599 Triple DES CBC and EdDSA.
603 * Added support for KMAC in KBKDF.
607 * RNDR and RNDRRS support in provider functions to provide
608 random number generation for Arm CPUs (aarch64).
612 * s_client and s_server apps now explicitly say when the TLS version
613 does not include the renegotiation mechanism. This avoids confusion
614 between that scenario versus when the TLS version includes secure
615 renegotiation but the peer lacks support for it.
619 * AES-GCM enabled with AVX512 vAES and vPCLMULQDQ.
621 *Tomasz Kantecki, Andrey Matyukov*
623 * The various OBJ_* functions have been made thread safe.
627 * Parallel dual-prime 1536/2048-bit modular exponentiation for
628 AVX512_IFMA capable processors.
630 *Sergey Kirillov, Andrey Matyukov (Intel Corp)*
632 * The functions `OPENSSL_LH_stats`, `OPENSSL_LH_node_stats`,
633 `OPENSSL_LH_node_usage_stats`, `OPENSSL_LH_stats_bio`,
634 `OPENSSL_LH_node_stats_bio` and `OPENSSL_LH_node_usage_stats_bio` are now
635 marked deprecated from OpenSSL 3.1 onwards and can be disabled by defining
636 `OPENSSL_NO_DEPRECATED_3_1`.
638 The macro `DEFINE_LHASH_OF` is now deprecated in favour of the macro
639 `DEFINE_LHASH_OF_EX`, which omits the corresponding type-specific function
640 definitions for these functions regardless of whether
641 `OPENSSL_NO_DEPRECATED_3_1` is defined.
643 Users of `DEFINE_LHASH_OF` may start receiving deprecation warnings for these
644 functions regardless of whether they are using them. It is recommended that
645 users transition to the new macro, `DEFINE_LHASH_OF_EX`.
649 * When generating safe-prime DH parameters set the recommended private key
650 length equivalent to minimum key lengths as in RFC 7919.
654 * Change the default salt length for PKCS#1 RSASSA-PSS signatures to the
655 maximum size that is smaller or equal to the digest length to comply with
656 FIPS 186-4 section 5. This is implemented by a new option
657 `OSSL_PKEY_RSA_PSS_SALT_LEN_AUTO_DIGEST_MAX` ("auto-digestmax") for the
658 `rsa_pss_saltlen` parameter, which is now the default. Signature
659 verification is not affected by this change and continues to work as before.
666 For OpenSSL 3.0 a [Migration guide][] has been added, so the CHANGES entries
667 listed here are only a brief description.
668 The migration guide contains more detailed information related to new features,
669 breaking changes, and mappings for the large list of deprecated functions.
671 [Migration guide]: https://github.com/openssl/openssl/tree/master/doc/man7/migration_guide.pod
673 ### Changes between 3.0.7 and 3.0.8 [7 Feb 2023]
675 * Fixed NULL dereference during PKCS7 data verification.
677 A NULL pointer can be dereferenced when signatures are being
678 verified on PKCS7 signed or signedAndEnveloped data. In case the hash
679 algorithm used for the signature is known to the OpenSSL library but
680 the implementation of the hash algorithm is not available the digest
681 initialization will fail. There is a missing check for the return
682 value from the initialization function which later leads to invalid
683 usage of the digest API most likely leading to a crash.
686 PKCS7 data is processed by the SMIME library calls and also by the
687 time stamp (TS) library calls. The TLS implementation in OpenSSL does
688 not call these functions however third party applications would be
689 affected if they call these functions to verify signatures on untrusted
694 * Fixed X.400 address type confusion in X.509 GeneralName.
696 There is a type confusion vulnerability relating to X.400 address processing
697 inside an X.509 GeneralName. X.400 addresses were parsed as an ASN1_STRING
698 but the public structure definition for GENERAL_NAME incorrectly specified
699 the type of the x400Address field as ASN1_TYPE. This field is subsequently
700 interpreted by the OpenSSL function GENERAL_NAME_cmp as an ASN1_TYPE rather
703 When CRL checking is enabled (i.e. the application sets the
704 X509_V_FLAG_CRL_CHECK flag), this vulnerability may allow an attacker to
705 pass arbitrary pointers to a memcmp call, enabling them to read memory
706 contents or enact a denial of service.
711 * Fixed NULL dereference validating DSA public key.
713 An invalid pointer dereference on read can be triggered when an
714 application tries to check a malformed DSA public key by the
715 EVP_PKEY_public_check() function. This will most likely lead
716 to an application crash. This function can be called on public
717 keys supplied from untrusted sources which could allow an attacker
718 to cause a denial of service attack.
720 The TLS implementation in OpenSSL does not call this function
721 but applications might call the function if there are additional
722 security requirements imposed by standards such as FIPS 140-3.
725 *Shane Lontis, Tomáš Mráz*
727 * Fixed Invalid pointer dereference in d2i_PKCS7 functions.
729 An invalid pointer dereference on read can be triggered when an
730 application tries to load malformed PKCS7 data with the
731 d2i_PKCS7(), d2i_PKCS7_bio() or d2i_PKCS7_fp() functions.
733 The result of the dereference is an application crash which could
734 lead to a denial of service attack. The TLS implementation in OpenSSL
735 does not call this function however third party applications might
736 call these functions on untrusted data.
741 * Fixed Use-after-free following BIO_new_NDEF.
743 The public API function BIO_new_NDEF is a helper function used for
744 streaming ASN.1 data via a BIO. It is primarily used internally to OpenSSL
745 to support the SMIME, CMS and PKCS7 streaming capabilities, but may also
746 be called directly by end user applications.
748 The function receives a BIO from the caller, prepends a new BIO_f_asn1
749 filter BIO onto the front of it to form a BIO chain, and then returns
750 the new head of the BIO chain to the caller. Under certain conditions,
751 for example if a CMS recipient public key is invalid, the new filter BIO
752 is freed and the function returns a NULL result indicating a failure.
753 However, in this case, the BIO chain is not properly cleaned up and the
754 BIO passed by the caller still retains internal pointers to the previously
755 freed filter BIO. If the caller then goes on to call BIO_pop() on the BIO
756 then a use-after-free will occur. This will most likely result in a crash.
759 *Viktor Dukhovni, Matt Caswell*
761 * Fixed Double free after calling PEM_read_bio_ex.
763 The function PEM_read_bio_ex() reads a PEM file from a BIO and parses and
764 decodes the "name" (e.g. "CERTIFICATE"), any header data and the payload
765 data. If the function succeeds then the "name_out", "header" and "data"
766 arguments are populated with pointers to buffers containing the relevant
767 decoded data. The caller is responsible for freeing those buffers. It is
768 possible to construct a PEM file that results in 0 bytes of payload data.
769 In this case PEM_read_bio_ex() will return a failure code but will populate
770 the header argument with a pointer to a buffer that has already been freed.
771 If the caller also frees this buffer then a double free will occur. This
772 will most likely lead to a crash.
774 The functions PEM_read_bio() and PEM_read() are simple wrappers around
775 PEM_read_bio_ex() and therefore these functions are also directly affected.
777 These functions are also called indirectly by a number of other OpenSSL
778 functions including PEM_X509_INFO_read_bio_ex() and
779 SSL_CTX_use_serverinfo_file() which are also vulnerable. Some OpenSSL
780 internal uses of these functions are not vulnerable because the caller does
781 not free the header argument if PEM_read_bio_ex() returns a failure code.
784 *Kurt Roeckx, Matt Caswell*
786 * Fixed Timing Oracle in RSA Decryption.
788 A timing based side channel exists in the OpenSSL RSA Decryption
789 implementation which could be sufficient to recover a plaintext across
790 a network in a Bleichenbacher style attack. To achieve a successful
791 decryption an attacker would have to be able to send a very large number
792 of trial messages for decryption. The vulnerability affects all RSA padding
793 modes: PKCS#1 v1.5, RSA-OEAP and RSASVE.
796 *Dmitry Belyavsky, Hubert Kario*
798 * Fixed X.509 Name Constraints Read Buffer Overflow.
800 A read buffer overrun can be triggered in X.509 certificate verification,
801 specifically in name constraint checking. The read buffer overrun might
802 result in a crash which could lead to a denial of service attack.
803 In a TLS client, this can be triggered by connecting to a malicious
804 server. In a TLS server, this can be triggered if the server requests
805 client authentication and a malicious client connects.
810 * Fixed X.509 Policy Constraints Double Locking security issue.
812 If an X.509 certificate contains a malformed policy constraint and
813 policy processing is enabled, then a write lock will be taken twice
814 recursively. On some operating systems (most widely: Windows) this
815 results in a denial of service when the affected process hangs. Policy
816 processing being enabled on a publicly facing server is not considered
817 to be a common setup.
822 * Our provider implementations of `OSSL_FUNC_KEYMGMT_EXPORT` and
823 `OSSL_FUNC_KEYMGMT_GET_PARAMS` for EC and SM2 keys now honor
824 `OSSL_PKEY_PARAM_EC_POINT_CONVERSION_FORMAT` as set (and
825 default to `POINT_CONVERSION_UNCOMPRESSED`) when exporting
826 `OSSL_PKEY_PARAM_PUB_KEY`, instead of unconditionally using
827 `POINT_CONVERSION_COMPRESSED` as in previous 3.x releases.
828 For symmetry, our implementation of `EVP_PKEY_ASN1_METHOD->export_to`
829 for legacy EC and SM2 keys is also changed similarly to honor the
830 equivalent conversion format flag as specified in the underlying
831 `EC_KEY` object being exported to a provider, when this function is
832 called through `EVP_PKEY_export()`.
836 ### Changes between 3.0.6 and 3.0.7 [1 Nov 2022]
838 * Fixed two buffer overflows in punycode decoding functions.
840 A buffer overrun can be triggered in X.509 certificate verification,
841 specifically in name constraint checking. Note that this occurs after
842 certificate chain signature verification and requires either a CA to
843 have signed the malicious certificate or for the application to continue
844 certificate verification despite failure to construct a path to a trusted
847 In a TLS client, this can be triggered by connecting to a malicious
848 server. In a TLS server, this can be triggered if the server requests
849 client authentication and a malicious client connects.
851 An attacker can craft a malicious email address to overflow
852 an arbitrary number of bytes containing the `.` character (decimal 46)
853 on the stack. This buffer overflow could result in a crash (causing a
857 An attacker can craft a malicious email address to overflow four
858 attacker-controlled bytes on the stack. This buffer overflow could
859 result in a crash (causing a denial of service) or potentially remote code
860 execution depending on stack layout for any given platform/compiler.
865 * Removed all references to invalid OSSL_PKEY_PARAM_RSA names for CRT
866 parameters in OpenSSL code.
867 Applications should not use the names OSSL_PKEY_PARAM_RSA_FACTOR,
868 OSSL_PKEY_PARAM_RSA_EXPONENT and OSSL_PKEY_PARAM_RSA_COEFFICIENT.
869 Use the numbered names such as OSSL_PKEY_PARAM_RSA_FACTOR1 instead.
870 Using these invalid names may cause algorithms to use slower methods
871 that ignore the CRT parameters.
875 * Fixed a regression introduced in 3.0.6 version raising errors on some stack
880 * Fixed a regression introduced in 3.0.6 version not refreshing the certificate
881 data to be signed before signing the certificate.
885 * Added RIPEMD160 to the default provider.
889 * Ensured that the key share group sent or accepted for the key exchange
890 is allowed for the protocol version.
894 ### Changes between 3.0.5 and 3.0.6 [11 Oct 2022]
896 * OpenSSL supports creating a custom cipher via the legacy
897 EVP_CIPHER_meth_new() function and associated function calls. This function
898 was deprecated in OpenSSL 3.0 and application authors are instead encouraged
899 to use the new provider mechanism in order to implement custom ciphers.
901 OpenSSL versions 3.0.0 to 3.0.5 incorrectly handle legacy custom ciphers
902 passed to the EVP_EncryptInit_ex2(), EVP_DecryptInit_ex2() and
903 EVP_CipherInit_ex2() functions (as well as other similarly named encryption
904 and decryption initialisation functions). Instead of using the custom cipher
905 directly it incorrectly tries to fetch an equivalent cipher from the
906 available providers. An equivalent cipher is found based on the NID passed to
907 EVP_CIPHER_meth_new(). This NID is supposed to represent the unique NID for a
908 given cipher. However it is possible for an application to incorrectly pass
909 NID_undef as this value in the call to EVP_CIPHER_meth_new(). When NID_undef
910 is used in this way the OpenSSL encryption/decryption initialisation function
911 will match the NULL cipher as being equivalent and will fetch this from the
912 available providers. This will succeed if the default provider has been
913 loaded (or if a third party provider has been loaded that offers this
914 cipher). Using the NULL cipher means that the plaintext is emitted as the
917 Applications are only affected by this issue if they call
918 EVP_CIPHER_meth_new() using NID_undef and subsequently use it in a call to an
919 encryption/decryption initialisation function. Applications that only use
920 SSL/TLS are not impacted by this issue.
925 * Fix LLVM vs Apple LLVM version numbering confusion that caused build failures
930 * Fixed the linux-mips64 Configure target which was missing the
931 SIXTY_FOUR_BIT bn_ops flag. This was causing heap corruption on that
936 * Fix handling of a ticket key callback that returns 0 in TLSv1.3 to not send a
941 * Correctly handle a retransmitted ClientHello in DTLS
945 * Fixed detection of ktls support in cross-compile environment on Linux
949 * Fixed some regressions and test failures when running the 3.0.0 FIPS provider
954 * Fixed SSL_pending() and SSL_has_pending() with DTLS which were failing to
955 report correct results in some cases
959 * Fix UWP builds by defining VirtualLock
963 * For known safe primes use the minimum key length according to RFC 7919.
964 Longer private key sizes unnecessarily raise the cycles needed to compute the
965 shared secret without any increase of the real security. This fixes a
966 regression from 1.1.1 where these shorter keys were generated for the known
971 * Added the loongarch64 target
975 * Fixed EC ASM flag passing. Flags for ASM implementations of EC curves were
976 only passed to the FIPS provider and not to the default or legacy provider.
980 * Fixed reported performance degradation on aarch64. Restored the
981 implementation prior to commit 2621751 ("aes/asm/aesv8-armx.pl: avoid
982 32-bit lane assignment in CTR mode") for 64bit targets only, since it is
983 reportedly 2-17% slower and the silicon errata only affects 32bit targets.
984 The new algorithm is still used for 32 bit targets.
988 * Added a missing header for memcmp that caused compilation failure on some
993 ### Changes between 3.0.4 and 3.0.5 [5 Jul 2022]
995 * The OpenSSL 3.0.4 release introduced a serious bug in the RSA
996 implementation for X86_64 CPUs supporting the AVX512IFMA instructions.
997 This issue makes the RSA implementation with 2048 bit private keys
998 incorrect on such machines and memory corruption will happen during
999 the computation. As a consequence of the memory corruption an attacker
1000 may be able to trigger a remote code execution on the machine performing
1003 SSL/TLS servers or other servers using 2048 bit RSA private keys running
1004 on machines supporting AVX512IFMA instructions of the X86_64 architecture
1005 are affected by this issue.
1010 * AES OCB mode for 32-bit x86 platforms using the AES-NI assembly optimised
1011 implementation would not encrypt the entirety of the data under some
1012 circumstances. This could reveal sixteen bytes of data that was
1013 preexisting in the memory that wasn't written. In the special case of
1014 "in place" encryption, sixteen bytes of the plaintext would be revealed.
1016 Since OpenSSL does not support OCB based cipher suites for TLS and DTLS,
1017 they are both unaffected.
1020 *Alex Chernyakhovsky, David Benjamin, Alejandro Sedeño*
1022 ### Changes between 3.0.3 and 3.0.4 [21 Jun 2022]
1024 * In addition to the c_rehash shell command injection identified in
1025 CVE-2022-1292, further bugs where the c_rehash script does not
1026 properly sanitise shell metacharacters to prevent command injection have been
1029 When the CVE-2022-1292 was fixed it was not discovered that there
1030 are other places in the script where the file names of certificates
1031 being hashed were possibly passed to a command executed through the shell.
1033 This script is distributed by some operating systems in a manner where
1034 it is automatically executed. On such operating systems, an attacker
1035 could execute arbitrary commands with the privileges of the script.
1037 Use of the c_rehash script is considered obsolete and should be replaced
1038 by the OpenSSL rehash command line tool.
1041 *Daniel Fiala, Tomáš Mráz*
1043 * Case insensitive string comparison no longer uses locales. It has instead
1044 been directly implemented.
1048 ### Changes between 3.0.2 and 3.0.3 [3 May 2022]
1050 * Case insensitive string comparison is reimplemented via new locale-agnostic
1051 comparison functions OPENSSL_str[n]casecmp always using the POSIX locale for
1052 comparison. The previous implementation had problems when the Turkish locale
1057 * Fixed a bug in the c_rehash script which was not properly sanitising shell
1058 metacharacters to prevent command injection. This script is distributed by
1059 some operating systems in a manner where it is automatically executed. On
1060 such operating systems, an attacker could execute arbitrary commands with the
1061 privileges of the script.
1063 Use of the c_rehash script is considered obsolete and should be replaced
1064 by the OpenSSL rehash command line tool.
1069 * Fixed a bug in the function `OCSP_basic_verify` that verifies the signer
1070 certificate on an OCSP response. The bug caused the function in the case
1071 where the (non-default) flag OCSP_NOCHECKS is used to return a postivie
1072 response (meaning a successful verification) even in the case where the
1073 response signing certificate fails to verify.
1075 It is anticipated that most users of `OCSP_basic_verify` will not use the
1076 OCSP_NOCHECKS flag. In this case the `OCSP_basic_verify` function will return
1077 a negative value (indicating a fatal error) in the case of a certificate
1078 verification failure. The normal expected return value in this case would be
1081 This issue also impacts the command line OpenSSL "ocsp" application. When
1082 verifying an ocsp response with the "-no_cert_checks" option the command line
1083 application will report that the verification is successful even though it
1084 has in fact failed. In this case the incorrect successful response will also
1085 be accompanied by error messages showing the failure and contradicting the
1086 apparently successful result.
1091 * Fixed a bug where the RC4-MD5 ciphersuite incorrectly used the
1092 AAD data as the MAC key. This made the MAC key trivially predictable.
1094 An attacker could exploit this issue by performing a man-in-the-middle attack
1095 to modify data being sent from one endpoint to an OpenSSL 3.0 recipient such
1096 that the modified data would still pass the MAC integrity check.
1098 Note that data sent from an OpenSSL 3.0 endpoint to a non-OpenSSL 3.0
1099 endpoint will always be rejected by the recipient and the connection will
1100 fail at that point. Many application protocols require data to be sent from
1101 the client to the server first. Therefore, in such a case, only an OpenSSL
1102 3.0 server would be impacted when talking to a non-OpenSSL 3.0 client.
1104 If both endpoints are OpenSSL 3.0 then the attacker could modify data being
1105 sent in both directions. In this case both clients and servers could be
1106 affected, regardless of the application protocol.
1108 Note that in the absence of an attacker this bug means that an OpenSSL 3.0
1109 endpoint communicating with a non-OpenSSL 3.0 endpoint will fail to complete
1110 the handshake when using this ciphersuite.
1112 The confidentiality of data is not impacted by this issue, i.e. an attacker
1113 cannot decrypt data that has been encrypted using this ciphersuite - they can
1116 In order for this attack to work both endpoints must legitimately negotiate
1117 the RC4-MD5 ciphersuite. This ciphersuite is not compiled by default in
1118 OpenSSL 3.0, and is not available within the default provider or the default
1119 ciphersuite list. This ciphersuite will never be used if TLSv1.3 has been
1120 negotiated. In order for an OpenSSL 3.0 endpoint to use this ciphersuite the
1121 following must have occurred:
1123 1) OpenSSL must have been compiled with the (non-default) compile time option
1124 enable-weak-ssl-ciphers
1126 2) OpenSSL must have had the legacy provider explicitly loaded (either
1127 through application code or via configuration)
1129 3) The ciphersuite must have been explicitly added to the ciphersuite list
1131 4) The libssl security level must have been set to 0 (default is 1)
1133 5) A version of SSL/TLS below TLSv1.3 must have been negotiated
1135 6) Both endpoints must negotiate the RC4-MD5 ciphersuite in preference to any
1136 others that both endpoints have in common
1141 * Fix a bug in the OPENSSL_LH_flush() function that breaks reuse of the memory
1142 occupied by the removed hash table entries.
1144 This function is used when decoding certificates or keys. If a long lived
1145 process periodically decodes certificates or keys its memory usage will
1146 expand without bounds and the process might be terminated by the operating
1147 system causing a denial of service. Also traversing the empty hash table
1148 entries will take increasingly more time.
1150 Typically such long lived processes might be TLS clients or TLS servers
1151 configured to accept client certificate authentication.
1154 *Hugo Landau, Aliaksei Levin*
1156 * The functions `OPENSSL_LH_stats` and `OPENSSL_LH_stats_bio` now only report
1157 the `num_items`, `num_nodes` and `num_alloc_nodes` statistics. All other
1158 statistics are no longer supported. For compatibility, these statistics are
1159 still listed in the output but are now always reported as zero.
1163 ### Changes between 3.0.1 and 3.0.2 [15 Mar 2022]
1165 * Fixed a bug in the BN_mod_sqrt() function that can cause it to loop forever
1166 for non-prime moduli.
1168 Internally this function is used when parsing certificates that contain
1169 elliptic curve public keys in compressed form or explicit elliptic curve
1170 parameters with a base point encoded in compressed form.
1172 It is possible to trigger the infinite loop by crafting a certificate that
1173 has invalid explicit curve parameters.
1175 Since certificate parsing happens prior to verification of the certificate
1176 signature, any process that parses an externally supplied certificate may thus
1177 be subject to a denial of service attack. The infinite loop can also be
1178 reached when parsing crafted private keys as they can contain explicit
1179 elliptic curve parameters.
1181 Thus vulnerable situations include:
1183 - TLS clients consuming server certificates
1184 - TLS servers consuming client certificates
1185 - Hosting providers taking certificates or private keys from customers
1186 - Certificate authorities parsing certification requests from subscribers
1187 - Anything else which parses ASN.1 elliptic curve parameters
1189 Also any other applications that use the BN_mod_sqrt() where the attacker
1190 can control the parameter values are vulnerable to this DoS issue.
1195 * Add ciphersuites based on DHE_PSK (RFC 4279) and ECDHE_PSK (RFC 5489)
1196 to the list of ciphersuites providing Perfect Forward Secrecy as
1197 required by SECLEVEL >= 3.
1199 *Dmitry Belyavskiy, Nicola Tuveri*
1201 * Made the AES constant time code for no-asm configurations
1202 optional due to the resulting 95% performance degradation.
1203 The AES constant time code can be enabled, for no assembly
1204 builds, with: ./config no-asm -DOPENSSL_AES_CONST_TIME
1208 * Fixed PEM_write_bio_PKCS8PrivateKey() to make it possible to use empty
1213 * The negative return value handling of the certificate verification callback
1214 was reverted. The replacement is to set the verification retry state with
1215 the SSL_set_retry_verify() function.
1219 ### Changes between 3.0.0 and 3.0.1 [14 Dec 2021]
1221 * Fixed invalid handling of X509_verify_cert() internal errors in libssl
1222 Internally libssl in OpenSSL calls X509_verify_cert() on the client side to
1223 verify a certificate supplied by a server. That function may return a
1224 negative return value to indicate an internal error (for example out of
1225 memory). Such a negative return value is mishandled by OpenSSL and will cause
1226 an IO function (such as SSL_connect() or SSL_do_handshake()) to not indicate
1227 success and a subsequent call to SSL_get_error() to return the value
1228 SSL_ERROR_WANT_RETRY_VERIFY. This return value is only supposed to be
1229 returned by OpenSSL if the application has previously called
1230 SSL_CTX_set_cert_verify_callback(). Since most applications do not do this
1231 the SSL_ERROR_WANT_RETRY_VERIFY return value from SSL_get_error() will be
1232 totally unexpected and applications may not behave correctly as a result. The
1233 exact behaviour will depend on the application but it could result in
1234 crashes, infinite loops or other similar incorrect responses.
1236 This issue is made more serious in combination with a separate bug in OpenSSL
1237 3.0 that will cause X509_verify_cert() to indicate an internal error when
1238 processing a certificate chain. This will occur where a certificate does not
1239 include the Subject Alternative Name extension but where a Certificate
1240 Authority has enforced name constraints. This issue can occur even with valid
1246 * Corrected a few file name and file reference bugs in the build,
1247 installation and setup scripts, which lead to installation verification
1248 failures. Slightly enhanced the installation verification script.
1252 * Fixed EVP_PKEY_eq() to make it possible to use it with strictly private
1257 * Fixed PVK encoder to properly query for the passphrase.
1261 * Multiple fixes in the OSSL_HTTP API functions.
1265 * Allow sign extension in OSSL_PARAM_allocate_from_text() for the
1266 OSSL_PARAM_INTEGER data type and return error on negative numbers
1267 used with the OSSL_PARAM_UNSIGNED_INTEGER data type. Make
1268 OSSL_PARAM_BLD_push_BN{,_pad}() return an error on negative numbers.
1272 * Allow copying uninitialized digest contexts with EVP_MD_CTX_copy_ex.
1276 * Fixed detection of ARMv7 and ARM64 CPU features on FreeBSD.
1280 * Multiple threading fixes.
1284 * Added NULL digest implementation to keep compatibility with 1.1.1 version.
1288 * Allow fetching an operation from the provider that owns an unexportable key
1289 as a fallback if that is still allowed by the property query.
1293 ### Changes between 1.1.1 and 3.0.0 [7 Sep 2021]
1295 * TLS_MAX_VERSION, DTLS_MAX_VERSION and DTLS_MIN_VERSION constants are now
1300 * The `OPENSSL_s390xcap` environment variable can be used to set bits in the
1301 S390X capability vector to zero. This simplifies testing of different code
1302 paths on S390X architecture.
1306 * Encrypting more than 2^64 TLS records with AES-GCM is disallowed
1307 as per FIPS 140-2 IG A.5 "Key/IV Pair Uniqueness Requirements from
1308 SP 800-38D". The communication will fail at this point.
1312 * The EC_GROUP_clear_free() function is deprecated as there is nothing
1313 confidential in EC_GROUP data.
1317 * The byte order mark (BOM) character is ignored if encountered at the
1318 beginning of a PEM-formatted file.
1322 * Added CMS support for the Russian GOST algorithms.
1326 * Due to move of the implementation of cryptographic operations
1327 to the providers, validation of various operation parameters can
1328 be postponed until the actual operation is executed where previously
1329 it happened immediately when an operation parameter was set.
1331 For example when setting an unsupported curve with
1332 EVP_PKEY_CTX_set_ec_paramgen_curve_nid() this function call will not
1333 fail but later keygen operations with the EVP_PKEY_CTX will fail.
1335 *OpenSSL team members and many third party contributors*
1337 * The EVP_get_cipherbyname() function will return NULL for algorithms such as
1338 "AES-128-SIV", "AES-128-CBC-CTS" and "CAMELLIA-128-CBC-CTS" which were
1339 previously only accessible via low-level interfaces. Use EVP_CIPHER_fetch()
1340 instead to retrieve these algorithms from a provider.
1344 * On build targets where the multilib postfix is set in the build
1345 configuration the libdir directory was changing based on whether
1346 the lib directory with the multilib postfix exists on the system
1347 or not. This unpredictable behavior was removed and eventual
1348 multilib postfix is now always added to the default libdir. Use
1349 `--libdir=lib` to override the libdir if adding the postfix is
1354 * The triple DES key wrap functionality now conforms to RFC 3217 but is
1355 no longer interoperable with OpenSSL 1.1.1.
1359 * The ERR_GET_FUNC() function was removed. With the loss of meaningful
1360 function codes, this function can only cause problems for calling
1365 * Add a configurable flag to output date formats as ISO 8601. Does not
1366 change the default date format.
1370 * Version of MSVC earlier than 1300 could get link warnings, which could
1371 be suppressed if the undocumented -DI_CAN_LIVE_WITH_LNK4049 was set.
1372 Support for this flag has been removed.
1376 * Rework and make DEBUG macros consistent. Remove unused -DCONF_DEBUG,
1377 -DBN_CTX_DEBUG, and REF_PRINT. Add a new tracing category and use it for
1378 printing reference counts. Rename -DDEBUG_UNUSED to -DUNUSED_RESULT_DEBUG
1379 Fix BN_DEBUG_RAND so it compiles and, when set, force DEBUG_RAND to be set
1380 also. Rename engine_debug_ref to be ENGINE_REF_PRINT also for consistency.
1384 * The signatures of the functions to get and set options on SSL and
1385 SSL_CTX objects changed from "unsigned long" to "uint64_t" type.
1386 Some source code changes may be required.
1390 * The public definitions of conf_method_st and conf_st have been
1391 deprecated. They will be made opaque in a future release.
1393 *Rich Salz and Tomáš Mráz*
1395 * Client-initiated renegotiation is disabled by default. To allow it, use
1396 the -client_renegotiation option, the SSL_OP_ALLOW_CLIENT_RENEGOTIATION
1397 flag, or the "ClientRenegotiation" config parameter as appropriate.
1401 * Add "abspath" and "includedir" pragma's to config files, to prevent,
1402 or modify relative pathname inclusion.
1406 * OpenSSL includes a cryptographic module that is intended to be FIPS 140-2
1407 validated. Please consult the README-FIPS and
1408 README-PROVIDERS files, as well as the migration guide.
1410 *OpenSSL team members and many third party contributors*
1412 * For the key types DH and DHX the allowed settable parameters are now different.
1416 * The openssl commands that read keys, certificates, and CRLs now
1417 automatically detect the PEM or DER format of the input files.
1419 *David von Oheimb, Richard Levitte, and Tomáš Mráz*
1421 * Added enhanced PKCS#12 APIs which accept a library context.
1425 * The default manual page suffix ($MANSUFFIX) has been changed to "ossl"
1429 * Added support for Kernel TLS (KTLS).
1431 *Boris Pismenny, John Baldwin and Andrew Gallatin*
1433 * Support for RFC 5746 secure renegotiation is now required by default for
1434 SSL or TLS connections to succeed.
1438 * The signature of the `copy` functional parameter of the
1439 EVP_PKEY_meth_set_copy() function has changed so its `src` argument is
1440 now `const EVP_PKEY_CTX *` instead of `EVP_PKEY_CTX *`. Similarly
1441 the signature of the `pub_decode` functional parameter of the
1442 EVP_PKEY_asn1_set_public() function has changed so its `pub` argument is
1443 now `const X509_PUBKEY *` instead of `X509_PUBKEY *`.
1447 * The error return values from some control calls (ctrl) have changed.
1451 * A public key check is now performed during EVP_PKEY_derive_set_peer().
1455 * Many functions in the EVP_ namespace that are getters of values from
1456 implementations or contexts were renamed to include get or get0 in their
1457 names. Old names are provided as macro aliases for compatibility and
1462 * The EVP_PKEY_CTRL_PKCS7_ENCRYPT, EVP_PKEY_CTRL_PKCS7_DECRYPT,
1463 EVP_PKEY_CTRL_PKCS7_SIGN, EVP_PKEY_CTRL_CMS_ENCRYPT,
1464 EVP_PKEY_CTRL_CMS_DECRYPT, and EVP_PKEY_CTRL_CMS_SIGN control operations
1469 * The EVP_PKEY_public_check() and EVP_PKEY_param_check() functions now work for
1472 * The output from the command line applications may have minor
1477 * The output from numerous "printing" may have minor changes.
1481 * Windows thread synchronization uses read/write primitives (SRWLock) when
1482 supported by the OS, otherwise CriticalSection continues to be used.
1486 * Add filter BIO BIO_f_readbuffer() that allows BIO_tell() and BIO_seek() to
1487 work on read only BIO source/sinks that do not support these functions.
1488 This allows piping or redirection of a file BIO using stdin to be buffered
1489 into memory. This is used internally in OSSL_DECODER_from_bio().
1493 * OSSL_STORE_INFO_get_type() may now return an additional value. In 1.1.1
1494 this function would return one of the values OSSL_STORE_INFO_NAME,
1495 OSSL_STORE_INFO_PKEY, OSSL_STORE_INFO_PARAMS, OSSL_STORE_INFO_CERT or
1496 OSSL_STORE_INFO_CRL. Decoded public keys would previously have been reported
1497 as type OSSL_STORE_INFO_PKEY in 1.1.1. In 3.0 decoded public keys are now
1498 reported as having the new type OSSL_STORE_INFO_PUBKEY. Applications
1499 using this function should be amended to handle the changed return value.
1503 * Improved adherence to Enhanced Security Services (ESS, RFC 2634 and RFC 5035)
1504 for the TSP and CMS Advanced Electronic Signatures (CAdES) implementations.
1505 As required by RFC 5035 check both ESSCertID and ESSCertIDv2 if both present.
1506 Correct the semantics of checking the validation chain in case ESSCertID{,v2}
1507 contains more than one certificate identifier: This means that all
1508 certificates referenced there MUST be part of the validation chain.
1512 * The implementation of older EVP ciphers related to CAST, IDEA, SEED, RC2, RC4,
1513 RC5, DESX and DES have been moved to the legacy provider.
1517 * The implementation of the EVP digests MD2, MD4, MDC2, WHIRLPOOL and
1518 RIPEMD-160 have been moved to the legacy provider.
1522 * The deprecated function EVP_PKEY_get0() now returns NULL being called for a
1527 * The deprecated functions EVP_PKEY_get0_RSA(),
1528 EVP_PKEY_get0_DSA(), EVP_PKEY_get0_EC_KEY(), EVP_PKEY_get0_DH(),
1529 EVP_PKEY_get0_hmac(), EVP_PKEY_get0_poly1305() and EVP_PKEY_get0_siphash() as
1530 well as the similarly named "get1" functions behave differently in
1535 * A number of functions handling low-level keys or engines were deprecated
1536 including EVP_PKEY_set1_engine(), EVP_PKEY_get0_engine(), EVP_PKEY_assign(),
1537 EVP_PKEY_get0(), EVP_PKEY_get0_hmac(), EVP_PKEY_get0_poly1305() and
1538 EVP_PKEY_get0_siphash().
1542 * PKCS#5 PBKDF1 key derivation has been moved from PKCS5_PBE_keyivgen() into
1543 the legacy crypto provider as an EVP_KDF. Applications requiring this KDF
1544 will need to load the legacy crypto provider. This includes these PBE
1545 algorithms which use this KDF:
1546 - NID_pbeWithMD2AndDES_CBC
1547 - NID_pbeWithMD5AndDES_CBC
1548 - NID_pbeWithSHA1AndRC2_CBC
1549 - NID_pbeWithMD2AndRC2_CBC
1550 - NID_pbeWithMD5AndRC2_CBC
1551 - NID_pbeWithSHA1AndDES_CBC
1555 * Deprecated obsolete BIO_set_callback(), BIO_get_callback(), and
1556 BIO_debug_callback() functions.
1560 * Deprecated obsolete EVP_PKEY_CTX_get0_dh_kdf_ukm() and
1561 EVP_PKEY_CTX_get0_ecdh_kdf_ukm() functions.
1565 * The RAND_METHOD APIs have been deprecated.
1569 * The SRP APIs have been deprecated.
1573 * Add a compile time option to prevent the caching of provider fetched
1574 algorithms. This is enabled by including the no-cached-fetch option
1575 at configuration time.
1579 * pkcs12 now uses defaults of PBKDF2, AES and SHA-256, with a MAC iteration
1580 count of PKCS12_DEFAULT_ITER.
1582 *Tomáš Mráz and Sahana Prasad*
1584 * The openssl speed command does not use low-level API calls anymore.
1588 * Parallel dual-prime 1024-bit modular exponentiation for AVX512_IFMA
1591 *Ilya Albrekht, Sergey Kirillov, Andrey Matyukov (Intel Corp)*
1593 * Combining the Configure options no-ec and no-dh no longer disables TLSv1.3.
1597 * Implemented support for fully "pluggable" TLSv1.3 groups. This means that
1598 providers may supply their own group implementations (using either the "key
1599 exchange" or the "key encapsulation" methods) which will automatically be
1600 detected and used by libssl.
1602 *Matt Caswell, Nicola Tuveri*
1604 * The undocumented function X509_certificate_type() has been deprecated;
1608 * Deprecated the obsolete BN_pseudo_rand() and BN_pseudo_rand_range().
1612 * Removed RSA padding mode for SSLv23 (which was only used for
1613 SSLv2). This includes the functions RSA_padding_check_SSLv23() and
1614 RSA_padding_add_SSLv23() and the `-ssl` option in the deprecated
1619 * Deprecated the obsolete X9.31 RSA key generation related functions.
1621 * While a callback function set via `SSL_CTX_set_cert_verify_callback()`
1622 is not allowed to return a value > 1, this is no more taken as failure.
1624 *Viktor Dukhovni and David von Oheimb*
1626 * Deprecated the obsolete X9.31 RSA key generation related functions
1627 BN_X931_generate_Xpq(), BN_X931_derive_prime_ex(), and
1628 BN_X931_generate_prime_ex().
1632 * The default key generation method for the regular 2-prime RSA keys was
1633 changed to the FIPS 186-4 B.3.6 method.
1637 * Deprecated the BN_is_prime_ex() and BN_is_prime_fasttest_ex() functions.
1641 * Deprecated EVP_MD_CTX_set_update_fn() and EVP_MD_CTX_update_fn().
1645 * Deprecated the type OCSP_REQ_CTX and the functions OCSP_REQ_CTX_*() and
1646 replaced with OSSL_HTTP_REQ_CTX and the functions OSSL_HTTP_REQ_CTX_*().
1648 *Rich Salz, Richard Levitte, and David von Oheimb*
1650 * Deprecated `X509_http_nbio()` and `X509_CRL_http_nbio()`.
1654 * Deprecated `OCSP_parse_url()`.
1658 * Validation of SM2 keys has been separated from the validation of regular EC
1663 * Behavior of the `pkey` app is changed, when using the `-check` or `-pubcheck`
1664 switches: a validation failure triggers an early exit, returning a failure
1665 exit status to the parent process.
1669 * Changed behavior of SSL_CTX_set_ciphersuites() and SSL_set_ciphersuites()
1670 to ignore unknown ciphers.
1674 * The `-cipher-commands` and `-digest-commands` options
1675 of the command line utility `list` have been deprecated.
1676 Instead use the `-cipher-algorithms` and `-digest-algorithms` options.
1680 * Added convenience functions for generating asymmetric key pairs:
1681 The 'quick' one-shot (yet somewhat limited) function L<EVP_PKEY_Q_keygen(3)>
1682 and macros for the most common cases: <EVP_RSA_gen(3)> and L<EVP_EC_gen(3)>.
1686 * All of the low-level EC_KEY functions have been deprecated.
1688 *Shane Lontis, Paul Dale, Richard Levitte, and Tomáš Mráz*
1690 * Deprecated all the libcrypto and libssl error string loading
1695 * The functions SSL_CTX_set_tmp_dh_callback and SSL_set_tmp_dh_callback, as
1696 well as the macros SSL_CTX_set_tmp_dh() and SSL_set_tmp_dh() have been
1701 * The `-crypt` option to the `passwd` command line tool has been removed.
1705 * The -C option to the `x509`, `dhparam`, `dsaparam`, and `ecparam` commands
1710 * Add support for AES Key Wrap inverse ciphers to the EVP layer.
1714 * Deprecated EVP_PKEY_set1_tls_encodedpoint() and
1715 EVP_PKEY_get1_tls_encodedpoint().
1719 * The security callback, which can be customised by application code, supports
1720 the security operation SSL_SECOP_TMP_DH. One location of the "other" parameter
1721 was incorrectly passing a DH object. It now passed an EVP_PKEY in all cases.
1725 * Add PKCS7_get_octet_string() and PKCS7_type_is_other() to the public
1726 interface. Their functionality remains unchanged.
1730 * Added new option for 'openssl list', '-providers', which will display the
1731 list of loaded providers, their names, version and status. It optionally
1732 displays their gettable parameters.
1736 * Removed EVP_PKEY_set_alias_type().
1740 * Deprecated `EVP_PKEY_CTX_set_rsa_keygen_pubexp()` and introduced
1741 `EVP_PKEY_CTX_set1_rsa_keygen_pubexp()`, which is now preferred.
1745 * Changed all "STACK" functions to be macros instead of inline functions. Macro
1746 parameters are still checked for type safety at compile time via helper
1751 * Remove the RAND_DRBG API
1753 *Paul Dale and Matthias St. Pierre*
1755 * Allow `SSL_set1_host()` and `SSL_add1_host()` to take IP literal addresses
1756 as well as actual hostnames.
1760 * The 'MinProtocol' and 'MaxProtocol' configuration commands now silently
1761 ignore TLS protocol version bounds when configuring DTLS-based contexts, and
1762 conversely, silently ignore DTLS protocol version bounds when configuring
1763 TLS-based contexts. The commands can be repeated to set bounds of both
1764 types. The same applies with the corresponding "min_protocol" and
1765 "max_protocol" command-line switches, in case some application uses both TLS
1768 SSL_CTX instances that are created for a fixed protocol version (e.g.
1769 `TLSv1_server_method()`) also silently ignore version bounds. Previously
1770 attempts to apply bounds to these protocol versions would result in an
1771 error. Now only the "version-flexible" SSL_CTX instances are subject to
1772 limits in configuration files in command-line options.
1776 * Deprecated the `ENGINE` API. Engines should be replaced with providers
1781 * Reworked the recorded ERR codes to make better space for system errors.
1782 To distinguish them, the macro `ERR_SYSTEM_ERROR()` indicates if the
1783 given code is a system error (true) or an OpenSSL error (false).
1787 * Reworked the test perl framework to better allow parallel testing.
1789 *Nicola Tuveri and David von Oheimb*
1791 * Added ciphertext stealing algorithms AES-128-CBC-CTS, AES-192-CBC-CTS and
1792 AES-256-CBC-CTS to the providers. CS1, CS2 and CS3 variants are supported.
1796 * 'Configure' has been changed to figure out the configuration target if
1797 none is given on the command line. Consequently, the 'config' script is
1798 now only a mere wrapper. All documentation is changed to only mention
1801 *Rich Salz and Richard Levitte*
1803 * Added a library context `OSSL_LIB_CTX` that applications as well as
1804 other libraries can use to form a separate context within which
1805 libcrypto operations are performed.
1809 * Added various `_ex` functions to the OpenSSL API that support using
1810 a non-default `OSSL_LIB_CTX`.
1814 * Handshake now fails if Extended Master Secret extension is dropped
1819 * Dropped interactive mode from the `openssl` program.
1823 * Deprecated `EVP_PKEY_cmp()` and `EVP_PKEY_cmp_parameters()`.
1825 *David von Oheimb and Shane Lontis*
1827 * Deprecated `EC_METHOD_get_field_type()`.
1831 * Deprecated EC_GFp_simple_method(), EC_GFp_mont_method(),
1832 EC_GF2m_simple_method(), EC_GFp_nist_method(), EC_GFp_nistp224_method()
1833 EC_GFp_nistp256_method(), and EC_GFp_nistp521_method().
1837 * Deprecated EC_GROUP_new(), EC_GROUP_method_of(), and EC_POINT_method_of().
1841 * Add CAdES-BES signature verification support, mostly derived
1842 from ESSCertIDv2 TS (RFC 5816) contribution by Marek Klein.
1844 *Filipe Raimundo da Silva*
1846 * Add CAdES-BES signature scheme and attributes support (RFC 5126) to CMS API.
1850 * Added the AuthEnvelopedData content type structure (RFC 5083) with AES-GCM
1851 parameter (RFC 5084) for the Cryptographic Message Syntax (CMS).
1855 * Deprecated EC_POINT_make_affine() and EC_POINTs_make_affine().
1859 * Deprecated EC_GROUP_precompute_mult(), EC_GROUP_have_precompute_mult(), and
1860 EC_KEY_precompute_mult().
1864 * Deprecated EC_POINTs_mul().
1868 * Removed FIPS_mode() and FIPS_mode_set().
1872 * The SSL option SSL_OP_IGNORE_UNEXPECTED_EOF is introduced.
1876 * Deprecated EC_POINT_set_Jprojective_coordinates_GFp() and
1877 EC_POINT_get_Jprojective_coordinates_GFp().
1881 * Added OSSL_PARAM_BLD to the public interface. This allows OSSL_PARAM
1882 arrays to be more easily constructed via a series of utility functions.
1883 Create a parameter builder using OSSL_PARAM_BLD_new(), add parameters using
1884 the various push functions and finally convert to a passable OSSL_PARAM
1885 array using OSSL_PARAM_BLD_to_param().
1889 * The security strength of SHA1 and MD5 based signatures in TLS has been
1894 * Added EVP_PKEY_set_type_by_keymgmt(), to initialise an EVP_PKEY to
1895 contain a provider side internal key.
1899 * ASN1_verify(), ASN1_digest() and ASN1_sign() have been deprecated.
1903 * Project text documents not yet having a proper file name extension
1904 (`HACKING`, `LICENSE`, `NOTES*`, `README*`, `VERSION`) have been renamed to
1905 `*.md` as far as reasonable, else `*.txt`, for better use with file managers.
1909 * The main project documents (README, NEWS, CHANGES, INSTALL, SUPPORT)
1910 have been converted to Markdown with the goal to produce documents
1911 which not only look pretty when viewed online in the browser, but
1912 remain well readable inside a plain text editor.
1914 To achieve this goal, a 'minimalistic' Markdown style has been applied
1915 which avoids formatting elements that interfere too much with the
1916 reading flow in the text file. For example, it
1918 * avoids [ATX headings][] and uses [setext headings][] instead
1919 (which works for `<h1>` and `<h2>` headings only).
1920 * avoids [inline links][] and uses [reference links][] instead.
1921 * avoids [fenced code blocks][] and uses [indented code blocks][] instead.
1923 [ATX headings]: https://github.github.com/gfm/#atx-headings
1924 [setext headings]: https://github.github.com/gfm/#setext-headings
1925 [inline links]: https://github.github.com/gfm/#inline-link
1926 [reference links]: https://github.github.com/gfm/#reference-link
1927 [fenced code blocks]: https://github.github.com/gfm/#fenced-code-blocks
1928 [indented code blocks]: https://github.github.com/gfm/#indented-code-blocks
1930 *Matthias St. Pierre*
1932 * The test suite is changed to preserve results of each test recipe.
1933 A new directory test-runs/ with subdirectories named like the
1934 test recipes are created in the build tree for this purpose.
1938 * Added an implementation of CMP and CRMF (RFC 4210, RFC 4211 RFC 6712).
1939 This adds `crypto/cmp/`, `crpyto/crmf/`, `apps/cmp.c`, and `test/cmp_*`.
1940 See L<openssl-cmp(1)> and L<OSSL_CMP_exec_IR_ses(3)> as starting points.
1942 *David von Oheimb, Martin Peylo*
1944 * Generalized the HTTP client code from `crypto/ocsp/` into `crpyto/http/`.
1945 It supports arbitrary request and response content types, GET redirection,
1946 TLS, connections via HTTP(S) proxies, connections and exchange via
1947 user-defined BIOs (allowing implicit connections), persistent connections,
1948 and timeout checks. See L<OSSL_HTTP_transfer(3)> etc. for details.
1949 The legacy OCSP-focused (and only partly documented) API
1950 is retained for backward compatibility, while most of it is deprecated.
1954 * Added `util/check-format.pl`, a tool for checking adherence to the
1955 OpenSSL coding style <https://www.openssl.org/policies/codingstyle.html>.
1956 The checks performed are incomplete and yield some false positives.
1957 Still the tool should be useful for detecting most typical glitches.
1961 * `BIO_do_connect()` and `BIO_do_handshake()` have been extended:
1962 If domain name resolution yields multiple IP addresses all of them are tried
1963 after `connect()` failures.
1967 * All of the low-level RSA functions have been deprecated.
1971 * X509 certificates signed using SHA1 are no longer allowed at security
1976 * The command line utilities dhparam, dsa, gendsa and dsaparam have been
1977 modified to use PKEY APIs. These commands are now in maintenance mode
1978 and no new features will be added to them.
1982 * The command line utility rsautl has been deprecated.
1986 * The command line utilities genrsa and rsa have been modified to use PKEY
1987 APIs. They now write PKCS#8 keys by default. These commands are now in
1988 maintenance mode and no new features will be added to them.
1992 * All of the low-level DH functions have been deprecated.
1994 *Paul Dale and Matt Caswell*
1996 * All of the low-level DSA functions have been deprecated.
2000 * Reworked the treatment of EC EVP_PKEYs with the SM2 curve to
2001 automatically become EVP_PKEY_SM2 rather than EVP_PKEY_EC.
2005 * Deprecated low-level ECDH and ECDSA functions.
2009 * Deprecated EVP_PKEY_decrypt_old() and EVP_PKEY_encrypt_old().
2013 * Enhanced the documentation of EVP_PKEY_get_size(), EVP_PKEY_get_bits()
2014 and EVP_PKEY_get_security_bits(). Especially EVP_PKEY_get_size() needed
2015 a new formulation to include all the things it can be used for,
2016 as well as words of caution.
2020 * The SSL_CTX_set_tlsext_ticket_key_cb(3) function has been deprecated.
2024 * All of the low-level HMAC functions have been deprecated.
2026 *Paul Dale and David von Oheimb*
2028 * Over two thousand fixes were made to the documentation, including:
2029 - Common options (such as -rand/-writerand, TLS version control, etc)
2030 were refactored and point to newly-enhanced descriptions in openssl.pod.
2031 - Added style conformance for all options (with help from Richard Levitte),
2032 documented all reported missing options, added a CI build to check
2033 that all options are documented and that no unimplemented options
2035 - Documented some internals, such as all use of environment variables.
2036 - Addressed all internal broken L<> references.
2040 * All of the low-level CMAC functions have been deprecated.
2044 * The low-level MD2, MD4, MD5, MDC2, RIPEMD160 and Whirlpool digest
2045 functions have been deprecated.
2047 *Paul Dale and David von Oheimb*
2049 * Corrected the documentation of the return values from the `EVP_DigestSign*`
2050 set of functions. The documentation mentioned negative values for some
2051 errors, but this was never the case, so the mention of negative values
2054 Code that followed the documentation and thereby check with something
2055 like `EVP_DigestSignInit(...) <= 0` will continue to work undisturbed.
2059 * All of the low-level cipher functions have been deprecated.
2061 *Matt Caswell and Paul Dale*
2063 * Removed include/openssl/opensslconf.h.in and replaced it with
2064 include/openssl/configuration.h.in, which differs in not including
2065 <openssl/macros.h>. A short header include/openssl/opensslconf.h
2066 was added to include both.
2068 This allows internal hacks where one might need to modify the set
2069 of configured macros, for example this if deprecated symbols are
2070 still supposed to be available internally:
2072 #include <openssl/configuration.h>
2074 #undef OPENSSL_NO_DEPRECATED
2075 #define OPENSSL_SUPPRESS_DEPRECATED
2077 #include <openssl/macros.h>
2079 This should not be used by applications that use the exported
2080 symbols, as that will lead to linking errors.
2084 * Fixed an overflow bug in the x64_64 Montgomery squaring procedure
2085 used in exponentiation with 512-bit moduli. No EC algorithms are
2086 affected. Analysis suggests that attacks against 2-prime RSA1024,
2087 3-prime RSA1536, and DSA1024 as a result of this defect would be very
2088 difficult to perform and are not believed likely. Attacks against DH512
2089 are considered just feasible. However, for an attack the target would
2090 have to reuse the DH512 private key, which is not recommended anyway.
2091 Also applications directly using the low-level API BN_mod_exp may be
2092 affected if they use BN_FLG_CONSTTIME.
2097 * Most memory-debug features have been deprecated, and the functionality
2098 replaced with no-ops.
2102 * Added documentation for the STACK API.
2106 * Introduced a new method type and API, OSSL_ENCODER, to represent
2107 generic encoders. These do the same sort of job that PEM writers
2108 and d2i functions do, but with support for methods supplied by
2109 providers, and the possibility for providers to support other
2114 * Introduced a new method type and API, OSSL_DECODER, to represent
2115 generic decoders. These do the same sort of job that PEM readers
2116 and i2d functions do, but with support for methods supplied by
2117 providers, and the possibility for providers to support other
2122 * Added a .pragma directive to the syntax of configuration files, to
2123 allow varying behavior in a supported and predictable manner.
2124 Currently added pragma:
2128 This allows dollar signs to be a keyword character unless it's
2129 followed by a opening brace or parenthesis. This is useful for
2130 platforms where dollar signs are commonly used in names, such as
2131 volume names and system directory names on VMS.
2135 * Added functionality to create an EVP_PKEY from user data.
2139 * Change the interpretation of the '--api' configuration option to
2140 mean that this is a desired API compatibility level with no
2141 further meaning. The previous interpretation, that this would
2142 also mean to remove all deprecated symbols up to and including
2143 the given version, no requires that 'no-deprecated' is also used
2144 in the configuration.
2146 When building applications, the desired API compatibility level
2147 can be set with the OPENSSL_API_COMPAT macro like before. For
2148 API compatibility version below 3.0, the old style numerical
2149 value is valid as before, such as -DOPENSSL_API_COMPAT=0x10100000L.
2150 For version 3.0 and on, the value is expected to be the decimal
2151 value calculated from the major and minor version like this:
2153 MAJOR * 10000 + MINOR * 100
2157 -DOPENSSL_API_COMPAT=30000 For 3.0
2158 -DOPENSSL_API_COMPAT=30200 For 3.2
2160 To hide declarations that are deprecated up to and including the
2161 given API compatibility level, -DOPENSSL_NO_DEPRECATED must be
2162 given when building the application as well.
2166 * Added the X509_LOOKUP_METHOD called X509_LOOKUP_store, to allow
2167 access to certificate and CRL stores via URIs and OSSL_STORE
2170 This adds the following functions:
2172 - X509_LOOKUP_store()
2173 - X509_STORE_load_file()
2174 - X509_STORE_load_path()
2175 - X509_STORE_load_store()
2176 - SSL_add_store_cert_subjects_to_stack()
2177 - SSL_CTX_set_default_verify_store()
2178 - SSL_CTX_load_verify_file()
2179 - SSL_CTX_load_verify_dir()
2180 - SSL_CTX_load_verify_store()
2184 * Added a new method to gather entropy on VMS, based on SYS$GET_ENTROPY.
2185 The presence of this system service is determined at run-time.
2189 * Added functionality to create an EVP_PKEY context based on data
2190 for methods from providers. This takes an algorithm name and a
2191 property query string and simply stores them, with the intent
2192 that any operation that uses this context will use those strings
2193 to fetch the needed methods implicitly, thereby making the port
2194 of application written for pre-3.0 OpenSSL easier.
2198 * The undocumented function NCONF_WIN32() has been deprecated; for
2199 conversion details see the HISTORY section of doc/man5/config.pod
2203 * Introduced the new functions EVP_DigestSignInit_ex() and
2204 EVP_DigestVerifyInit_ex(). The macros EVP_DigestSignUpdate() and
2205 EVP_DigestVerifyUpdate() have been converted to functions. See the man
2206 pages for further details.
2210 * Over two thousand fixes were made to the documentation, including:
2211 adding missing command flags, better style conformance, documentation
2214 *Rich Salz, Richard Levitte*
2216 * s390x assembly pack: add hardware-support for P-256, P-384, P-521,
2217 X25519, X448, Ed25519 and Ed448.
2221 * Print all values for a PKCS#12 attribute with 'openssl pkcs12', not just
2226 * Deprecated the public definition of `ERR_STATE` as well as the function
2227 `ERR_get_state()`. This is done in preparation of making `ERR_STATE` an
2232 * Added ERR functionality to give callers access to the stored function
2233 names that have replaced the older function code based functions.
2235 New functions are ERR_peek_error_func(), ERR_peek_last_error_func(),
2236 ERR_peek_error_data(), ERR_peek_last_error_data(), ERR_get_error_all(),
2237 ERR_peek_error_all() and ERR_peek_last_error_all().
2239 Deprecate ERR functions ERR_get_error_line(), ERR_get_error_line_data(),
2240 ERR_peek_error_line_data(), ERR_peek_last_error_line_data() and
2241 ERR_func_error_string().
2245 * Extended testing to be verbose for failing tests only. The make variables
2246 VERBOSE_FAILURE or VF can be used to enable this:
2248 $ make VF=1 test # Unix
2249 $ mms /macro=(VF=1) test ! OpenVMS
2250 $ nmake VF=1 test # Windows
2254 * Added the `-copy_extensions` option to the `x509` command for use with
2255 `-req` and `-x509toreq`. When given with the `copy` or `copyall` argument,
2256 all extensions in the request are copied to the certificate or vice versa.
2258 *David von Oheimb*, *Kirill Stefanenkov <kirill_stefanenkov@rambler.ru>*
2260 * Added the `-copy_extensions` option to the `req` command for use with
2261 `-x509`. When given with the `copy` or `copyall` argument,
2262 all extensions in the certification request are copied to the certificate.
2266 * The `x509`, `req`, and `ca` commands now make sure that X.509v3 certificates
2267 they generate are by default RFC 5280 compliant in the following sense:
2268 There is a subjectKeyIdentifier extension with a hash value of the public key
2269 and for not self-signed certs there is an authorityKeyIdentifier extension
2270 with a keyIdentifier field or issuer information identifying the signing key.
2271 This is done unless some configuration overrides the new default behavior,
2272 such as `subjectKeyIdentifier = none` and `authorityKeyIdentifier = none`.
2276 * Added several checks to `X509_verify_cert()` according to requirements in
2277 RFC 5280 in case `X509_V_FLAG_X509_STRICT` is set
2278 (which may be done by using the CLI option `-x509_strict`):
2279 * The basicConstraints of CA certificates must be marked critical.
2280 * CA certificates must explicitly include the keyUsage extension.
2281 * If a pathlenConstraint is given the key usage keyCertSign must be allowed.
2282 * The issuer name of any certificate must not be empty.
2283 * The subject name of CA certs, certs with keyUsage crlSign,
2284 and certs without subjectAlternativeName must not be empty.
2285 * If a subjectAlternativeName extension is given it must not be empty.
2286 * The signatureAlgorithm field and the cert signature must be consistent.
2287 * Any given authorityKeyIdentifier and any given subjectKeyIdentifier
2288 must not be marked critical.
2289 * The authorityKeyIdentifier must be given for X.509v3 certs
2290 unless they are self-signed.
2291 * The subjectKeyIdentifier must be given for all X.509v3 CA certs.
2295 * Certificate verification using `X509_verify_cert()` meanwhile rejects EC keys
2296 with explicit curve parameters (specifiedCurve) as required by RFC 5480.
2300 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
2301 used even when parsing explicit parameters, when loading a encoded key
2302 or calling `EC_GROUP_new_from_ecpkparameters()`/
2303 `EC_GROUP_new_from_ecparameters()`.
2304 This prevents bypass of security hardening and performance gains,
2305 especially for curves with specialized EC_METHODs.
2306 By default, if a key encoded with explicit parameters is loaded and later
2307 encoded, the output is still encoded with explicit parameters, even if
2308 internally a "named" EC_GROUP is used for computation.
2312 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
2313 this change, EC_GROUP_set_generator would accept order and/or cofactor as
2314 NULL. After this change, only the cofactor parameter can be NULL. It also
2315 does some minimal sanity checks on the passed order.
2320 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
2321 An attack is simple, if the first CMS_recipientInfo is valid but the
2322 second CMS_recipientInfo is chosen ciphertext. If the second
2323 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
2324 encryption key will be replaced by garbage, and the message cannot be
2325 decoded, but if the RSA decryption fails, the correct encryption key is
2326 used and the recipient will not notice the attack.
2327 As a work around for this potential attack the length of the decrypted
2328 key must be equal to the cipher default key length, in case the
2329 certificate is not given and all recipientInfo are tried out.
2330 The old behaviour can be re-enabled in the CMS code by setting the
2331 CMS_DEBUG_DECRYPT flag.
2335 * Early start up entropy quality from the DEVRANDOM seed source has been
2336 improved for older Linux systems. The RAND subsystem will wait for
2337 /dev/random to be producing output before seeding from /dev/urandom.
2338 The seeded state is stored for future library initialisations using
2339 a system global shared memory segment. The shared memory identifier
2340 can be configured by defining OPENSSL_RAND_SEED_DEVRANDOM_SHM_ID to
2341 the desired value. The default identifier is 114.
2345 * Revised BN_generate_prime_ex to not avoid factors 2..17863 in p-1
2346 when primes for RSA keys are computed.
2347 Since we previously always generated primes == 2 (mod 3) for RSA keys,
2348 the 2-prime and 3-prime RSA modules were easy to distinguish, since
2349 `N = p*q = 1 (mod 3)`, but `N = p*q*r = 2 (mod 3)`. Therefore, fingerprinting
2350 2-prime vs. 3-prime RSA keys was possible by computing N mod 3.
2351 This avoids possible fingerprinting of newly generated RSA modules.
2355 * Correct the extended master secret constant on EBCDIC systems. Without this
2356 fix TLS connections between an EBCDIC system and a non-EBCDIC system that
2357 negotiate EMS will fail. Unfortunately this also means that TLS connections
2358 between EBCDIC systems with this fix, and EBCDIC systems without this
2359 fix will fail if they negotiate EMS.
2363 * Changed the library initialisation so that the config file is now loaded
2364 by default. This was already the case for libssl. It now occurs for both
2365 libcrypto and libssl. Use the OPENSSL_INIT_NO_LOAD_CONFIG option to
2366 `OPENSSL_init_crypto()` to suppress automatic loading of a config file.
2370 * Introduced new error raising macros, `ERR_raise()` and `ERR_raise_data()`,
2371 where the former acts as a replacement for `ERR_put_error()`, and the
2372 latter replaces the combination `ERR_put_error()` + `ERR_add_error_data()`.
2373 `ERR_raise_data()` adds more flexibility by taking a format string and
2374 an arbitrary number of arguments following it, to be processed with
2379 * Introduced a new function, `OSSL_PROVIDER_available()`, which can be used
2380 to check if a named provider is loaded and available. When called, it
2381 will also activate all fallback providers if such are still present.
2385 * Enforce a minimum DH modulus size of 512 bits.
2389 * Changed DH parameters to generate the order q subgroup instead of 2q.
2390 Previously generated DH parameters are still accepted by DH_check
2391 but DH_generate_key works around that by clearing bit 0 of the
2392 private key for those. This avoids leaking bit 0 of the private key.
2396 * Significantly reduce secure memory usage by the randomness pools.
2400 * `{CRYPTO,OPENSSL}_mem_debug_{push,pop}` are now no-ops and have been
2405 * A new type, EVP_KEYEXCH, has been introduced to represent key exchange
2406 algorithms. An implementation of a key exchange algorithm can be obtained
2407 by using the function EVP_KEYEXCH_fetch(). An EVP_KEYEXCH algorithm can be
2408 used in a call to EVP_PKEY_derive_init_ex() which works in a similar way to
2409 the older EVP_PKEY_derive_init() function. See the man pages for the new
2410 functions for further details.
2414 * The EVP_PKEY_CTX_set_dh_pad() macro has now been converted to a function.
2418 * Removed the function names from error messages and deprecated the
2423 * Removed NextStep support and the macro OPENSSL_UNISTD
2427 * Removed DES_check_key. Also removed OPENSSL_IMPLEMENT_GLOBAL,
2428 OPENSSL_GLOBAL_REF, OPENSSL_DECLARE_GLOBAL.
2429 Also removed "export var as function" capability; we do not export
2430 variables, only functions.
2434 * RC5_32_set_key has been changed to return an int type, with 0 indicating
2435 an error and 1 indicating success. In previous versions of OpenSSL this
2436 was a void type. If a key was set longer than the maximum possible this
2441 * Support SM2 signing and verification schemes with X509 certificate.
2445 * Use SHA256 as the default digest for TS query in the `ts` app.
2449 * Change PBKDF2 to conform to SP800-132 instead of the older PKCS5 RFC2898.
2453 * Default cipher lists/suites are now available via a function, the
2454 #defines are deprecated.
2458 * Add target VC-WIN32-UWP, VC-WIN64A-UWP, VC-WIN32-ARM-UWP and
2459 VC-WIN64-ARM-UWP in Windows OneCore target for making building libraries
2460 for Windows Store apps easier. Also, the "no-uplink" option has been added.
2464 * Join the directories crypto/x509 and crypto/x509v3
2468 * Added command 'openssl kdf' that uses the EVP_KDF API.
2472 * Added command 'openssl mac' that uses the EVP_MAC API.
2476 * Added OPENSSL_info() to get diverse built-in OpenSSL data, such
2477 as default directories. Also added the command 'openssl info'
2478 for scripting purposes.
2482 * The functions AES_ige_encrypt() and AES_bi_ige_encrypt() have been
2487 * Add prediction resistance to the DRBG reseeding process.
2491 * Limit the number of blocks in a data unit for AES-XTS to 2^20 as
2492 mandated by IEEE Std 1619-2018.
2496 * Added newline escaping functionality to a filename when using openssl dgst.
2497 This output format is to replicate the output format found in the `*sum`
2498 checksum programs. This aims to preserve backward compatibility.
2500 *Matt Eaton, Richard Levitte, and Paul Dale*
2502 * Removed the heartbeat message in DTLS feature, as it has very
2503 little usage and doesn't seem to fulfill a valuable purpose.
2504 The configuration option is now deprecated.
2508 * Changed the output of 'openssl {digestname} < file' to display the
2509 digest name in its output.
2513 * Added a new generic trace API which provides support for enabling
2514 instrumentation through trace output.
2516 *Richard Levitte & Matthias St. Pierre*
2518 * Added build tests for C++. These are generated files that only do one
2519 thing, to include one public OpenSSL head file each. This tests that
2520 the public header files can be usefully included in a C++ application.
2522 This test isn't enabled by default. It can be enabled with the option
2523 'enable-buildtest-c++'.
2527 * Added KB KDF (EVP_KDF_KB) to EVP_KDF.
2531 * Added SSH KDF (EVP_KDF_SSHKDF) and KRB5 KDF (EVP_KDF_KRB5KDF) to EVP_KDF.
2535 * Added Single Step KDF (EVP_KDF_SS), X963 KDF, and X942 KDF to EVP_KDF.
2539 * Added KMAC to EVP_MAC.
2543 * Added property based algorithm implementation selection framework to
2548 * Added SCA hardening for modular field inversion in EC_GROUP through
2549 a new dedicated field_inv() pointer in EC_METHOD.
2550 This also addresses a leakage affecting conversions from projective
2551 to affine coordinates.
2553 *Billy Bob Brumley, Nicola Tuveri*
2555 * Added EVP_KDF, an EVP layer KDF API, to simplify adding KDF and PRF
2556 implementations. This includes an EVP_PKEY to EVP_KDF bridge for
2557 those algorithms that were already supported through the EVP_PKEY API
2558 (scrypt, TLS1 PRF and HKDF). The low-level KDF functions for PBKDF2
2559 and scrypt are now wrappers that call EVP_KDF.
2563 * Build devcrypto engine as a dynamic engine.
2565 *Eneas U de Queiroz*
2567 * Add keyed BLAKE2 to EVP_MAC.
2571 * Fix a bug in the computation of the endpoint-pair shared secret used
2572 by DTLS over SCTP. This breaks interoperability with older versions
2573 of OpenSSL like OpenSSL 1.1.0 and OpenSSL 1.0.2. There is a runtime
2574 switch SSL_MODE_DTLS_SCTP_LABEL_LENGTH_BUG (off by default) enabling
2575 interoperability with such broken implementations. However, enabling
2576 this switch breaks interoperability with correct implementations.
2578 * Fix a use after free bug in d2i_X509_PUBKEY when overwriting a
2579 re-used X509_PUBKEY object if the second PUBKEY is malformed.
2583 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
2587 * Changed the license to the Apache License v2.0.
2591 * Switch to a new version scheme using three numbers MAJOR.MINOR.PATCH.
2593 - Major releases (indicated by incrementing the MAJOR release number)
2594 may introduce incompatible API/ABI changes.
2595 - Minor releases (indicated by incrementing the MINOR release number)
2596 may introduce new features but retain API/ABI compatibility.
2597 - Patch releases (indicated by incrementing the PATCH number)
2598 are intended for bug fixes and other improvements of existing
2599 features only (like improving performance or adding documentation)
2600 and retain API/ABI compatibility.
2604 * Add support for RFC5297 SIV mode (siv128), including AES-SIV.
2608 * Remove the 'dist' target and add a tarball building script. The
2609 'dist' target has fallen out of use, and it shouldn't be
2610 necessary to configure just to create a source distribution.
2614 * Recreate the OS390-Unix config target. It no longer relies on a
2615 special script like it did for OpenSSL pre-1.1.0.
2619 * Instead of having the source directories listed in Configure, add
2620 a 'build.info' keyword SUBDIRS to indicate what sub-directories to
2625 * Add GMAC to EVP_MAC.
2629 * Ported the HMAC, CMAC and SipHash EVP_PKEY_METHODs to EVP_MAC.
2633 * Added EVP_MAC, an EVP layer MAC API, to simplify adding MAC
2634 implementations. This includes a generic EVP_PKEY to EVP_MAC bridge,
2635 to facilitate the continued use of MACs through raw private keys in
2636 functionality such as `EVP_DigestSign*` and `EVP_DigestVerify*`.
2640 * Deprecate ECDH_KDF_X9_62().
2644 * Added EVP_PKEY_ECDH_KDF_X9_63 and ecdh_KDF_X9_63() as replacements for
2645 the EVP_PKEY_ECDH_KDF_X9_62 KDF type and ECDH_KDF_X9_62(). The old names
2646 are retained for backwards compatibility.
2650 * AES-XTS mode now enforces that its two keys are different to mitigate
2651 the attacked described in "Efficient Instantiations of Tweakable
2652 Blockciphers and Refinements to Modes OCB and PMAC" by Phillip Rogaway.
2653 Details of this attack can be obtained from:
2654 <http://web.cs.ucdavis.edu/%7Erogaway/papers/offsets.pdf>
2658 * Rename the object files, i.e. give them other names than in previous
2659 versions. Their names now include the name of the final product, as
2660 well as its type mnemonic (bin, lib, shlib).
2664 * Added new option for 'openssl list', '-objects', which will display the
2665 list of built in objects, i.e. OIDs with names.
2669 * Added the options `-crl_lastupdate` and `-crl_nextupdate` to `openssl ca`,
2670 allowing the `lastUpdate` and `nextUpdate` fields in the generated CRL to
2675 * Added support for Linux Kernel TLS data-path. The Linux Kernel data-path
2676 improves application performance by removing data copies and providing
2677 applications with zero-copy system calls such as sendfile and splice.
2681 * The SSL option SSL_OP_CLEANSE_PLAINTEXT is introduced.
2685 * `PKCS12_parse` now maintains the order of the parsed certificates
2686 when outputting them via `*ca` (rather than reversing it).
2690 * Deprecated pthread fork support methods.
2694 * Added support for FFDHE key exchange in TLS 1.3.
2698 * Added a new concept for OpenSSL plugability: providers. This
2699 functionality is designed to replace the ENGINE API and ENGINE
2700 implementations, and to be much more dynamic, allowing provider
2701 authors to introduce new algorithms among other things, as long as
2702 there's an API that supports the algorithm type.
2704 With this concept comes a new core API for interaction between
2705 libcrypto and provider implementations. Public libcrypto functions
2706 that want to use providers do so through this core API.
2708 The main documentation for this core API is found in
2709 doc/man7/provider.pod, doc/man7/provider-base.pod, and they in turn
2710 refer to other manuals describing the API specific for supported
2711 algorithm types (also called operations).
2718 ### Changes between 1.1.1m and 1.1.1n [xx XXX xxxx]
2720 ### Changes between 1.1.1l and 1.1.1m [14 Dec 2021]
2722 * Avoid loading of a dynamic engine twice.
2726 * Prioritise DANE TLSA issuer certs over peer certs
2730 * Fixed random API for MacOS prior to 10.12
2732 These MacOS versions don't support the CommonCrypto APIs
2736 ### Changes between 1.1.1k and 1.1.1l [24 Aug 2021]
2738 * Fixed an SM2 Decryption Buffer Overflow.
2740 In order to decrypt SM2 encrypted data an application is expected to
2741 call the API function EVP_PKEY_decrypt(). Typically an application will
2742 call this function twice. The first time, on entry, the "out" parameter
2743 can be NULL and, on exit, the "outlen" parameter is populated with the
2744 buffer size required to hold the decrypted plaintext. The application
2745 can then allocate a sufficiently sized buffer and call EVP_PKEY_decrypt()
2746 again, but this time passing a non-NULL value for the "out" parameter.
2748 A bug in the implementation of the SM2 decryption code means that the
2749 calculation of the buffer size required to hold the plaintext returned
2750 by the first call to EVP_PKEY_decrypt() can be smaller than the actual
2751 size required by the second call. This can lead to a buffer overflow
2752 when EVP_PKEY_decrypt() is called by the application a second time with
2753 a buffer that is too small.
2755 A malicious attacker who is able present SM2 content for decryption to
2756 an application could cause attacker chosen data to overflow the buffer
2757 by up to a maximum of 62 bytes altering the contents of other data held
2758 after the buffer, possibly changing application behaviour or causing
2759 the application to crash. The location of the buffer is application
2760 dependent but is typically heap allocated.
2765 * Fixed various read buffer overruns processing ASN.1 strings
2767 ASN.1 strings are represented internally within OpenSSL as an ASN1_STRING
2768 structure which contains a buffer holding the string data and a field
2769 holding the buffer length. This contrasts with normal C strings which
2770 are represented as a buffer for the string data which is terminated
2771 with a NUL (0) byte.
2773 Although not a strict requirement, ASN.1 strings that are parsed using
2774 OpenSSL's own "d2i" functions (and other similar parsing functions) as
2775 well as any string whose value has been set with the ASN1_STRING_set()
2776 function will additionally NUL terminate the byte array in the
2777 ASN1_STRING structure.
2779 However, it is possible for applications to directly construct valid
2780 ASN1_STRING structures which do not NUL terminate the byte array by
2781 directly setting the "data" and "length" fields in the ASN1_STRING
2782 array. This can also happen by using the ASN1_STRING_set0() function.
2784 Numerous OpenSSL functions that print ASN.1 data have been found to
2785 assume that the ASN1_STRING byte array will be NUL terminated, even
2786 though this is not guaranteed for strings that have been directly
2787 constructed. Where an application requests an ASN.1 structure to be
2788 printed, and where that ASN.1 structure contains ASN1_STRINGs that have
2789 been directly constructed by the application without NUL terminating
2790 the "data" field, then a read buffer overrun can occur.
2792 The same thing can also occur during name constraints processing
2793 of certificates (for example if a certificate has been directly
2794 constructed by the application instead of loading it via the OpenSSL
2795 parsing functions, and the certificate contains non NUL terminated
2796 ASN1_STRING structures). It can also occur in the X509_get1_email(),
2797 X509_REQ_get1_email() and X509_get1_ocsp() functions.
2799 If a malicious actor can cause an application to directly construct an
2800 ASN1_STRING and then process it through one of the affected OpenSSL
2801 functions then this issue could be hit. This might result in a crash
2802 (causing a Denial of Service attack). It could also result in the
2803 disclosure of private memory contents (such as private keys, or
2804 sensitive plaintext).
2809 ### Changes between 1.1.1j and 1.1.1k [25 Mar 2021]
2811 * Fixed a problem with verifying a certificate chain when using the
2812 X509_V_FLAG_X509_STRICT flag. This flag enables additional security checks of
2813 the certificates present in a certificate chain. It is not set by default.
2815 Starting from OpenSSL version 1.1.1h a check to disallow certificates in
2816 the chain that have explicitly encoded elliptic curve parameters was added
2817 as an additional strict check.
2819 An error in the implementation of this check meant that the result of a
2820 previous check to confirm that certificates in the chain are valid CA
2821 certificates was overwritten. This effectively bypasses the check
2822 that non-CA certificates must not be able to issue other certificates.
2824 If a "purpose" has been configured then there is a subsequent opportunity
2825 for checks that the certificate is a valid CA. All of the named "purpose"
2826 values implemented in libcrypto perform this check. Therefore, where
2827 a purpose is set the certificate chain will still be rejected even when the
2828 strict flag has been used. A purpose is set by default in libssl client and
2829 server certificate verification routines, but it can be overridden or
2830 removed by an application.
2832 In order to be affected, an application must explicitly set the
2833 X509_V_FLAG_X509_STRICT verification flag and either not set a purpose
2834 for the certificate verification or, in the case of TLS client or server
2835 applications, override the default purpose.
2840 * Fixed an issue where an OpenSSL TLS server may crash if sent a maliciously
2841 crafted renegotiation ClientHello message from a client. If a TLSv1.2
2842 renegotiation ClientHello omits the signature_algorithms extension (where it
2843 was present in the initial ClientHello), but includes a
2844 signature_algorithms_cert extension then a NULL pointer dereference will
2845 result, leading to a crash and a denial of service attack.
2847 A server is only vulnerable if it has TLSv1.2 and renegotiation enabled
2848 (which is the default configuration). OpenSSL TLS clients are not impacted by
2852 *Peter Kästle and Samuel Sapalski*
2854 ### Changes between 1.1.1i and 1.1.1j [16 Feb 2021]
2856 * Fixed the X509_issuer_and_serial_hash() function. It attempts to
2857 create a unique hash value based on the issuer and serial number data
2858 contained within an X509 certificate. However, it was failing to correctly
2859 handle any errors that may occur while parsing the issuer field (which might
2860 occur if the issuer field is maliciously constructed). This may subsequently
2861 result in a NULL pointer deref and a crash leading to a potential denial of
2867 * Fixed the RSA_padding_check_SSLv23() function and the RSA_SSLV23_PADDING
2868 padding mode to correctly check for rollback attacks. This is considered a
2869 bug in OpenSSL 1.1.1 because it does not support SSLv2. In 1.0.2 this is
2874 Fixed the EVP_CipherUpdate, EVP_EncryptUpdate and EVP_DecryptUpdate
2875 functions. Previously they could overflow the output length argument in some
2876 cases where the input length is close to the maximum permissible length for
2877 an integer on the platform. In such cases the return value from the function
2878 call would be 1 (indicating success), but the output length value would be
2879 negative. This could cause applications to behave incorrectly or crash.
2884 * Fixed SRP_Calc_client_key so that it runs in constant time. The previous
2885 implementation called BN_mod_exp without setting BN_FLG_CONSTTIME. This
2886 could be exploited in a side channel attack to recover the password. Since
2887 the attack is local host only this is outside of the current OpenSSL
2888 threat model and therefore no CVE is assigned.
2890 Thanks to Mohammed Sabt and Daniel De Almeida Braga for reporting this
2895 ### Changes between 1.1.1h and 1.1.1i [8 Dec 2020]
2897 * Fixed NULL pointer deref in the GENERAL_NAME_cmp function
2898 This function could crash if both GENERAL_NAMEs contain an EDIPARTYNAME.
2899 If an attacker can control both items being compared then this could lead
2900 to a possible denial of service attack. OpenSSL itself uses the
2901 GENERAL_NAME_cmp function for two purposes:
2902 1) Comparing CRL distribution point names between an available CRL and a
2903 CRL distribution point embedded in an X509 certificate
2904 2) When verifying that a timestamp response token signer matches the
2905 timestamp authority name (exposed via the API functions
2906 TS_RESP_verify_response and TS_RESP_verify_token)
2911 ### Changes between 1.1.1g and 1.1.1h [22 Sep 2020]
2913 * Certificates with explicit curve parameters are now disallowed in
2914 verification chains if the X509_V_FLAG_X509_STRICT flag is used.
2918 * The 'MinProtocol' and 'MaxProtocol' configuration commands now silently
2919 ignore TLS protocol version bounds when configuring DTLS-based contexts, and
2920 conversely, silently ignore DTLS protocol version bounds when configuring
2921 TLS-based contexts. The commands can be repeated to set bounds of both
2922 types. The same applies with the corresponding "min_protocol" and
2923 "max_protocol" command-line switches, in case some application uses both TLS
2926 SSL_CTX instances that are created for a fixed protocol version (e.g.
2927 TLSv1_server_method()) also silently ignore version bounds. Previously
2928 attempts to apply bounds to these protocol versions would result in an
2929 error. Now only the "version-flexible" SSL_CTX instances are subject to
2930 limits in configuration files in command-line options.
2934 * Handshake now fails if Extended Master Secret extension is dropped
2939 * The Oracle Developer Studio compiler will start reporting deprecated APIs
2941 ### Changes between 1.1.1f and 1.1.1g [21 Apr 2020]
2943 * Fixed segmentation fault in SSL_check_chain()
2944 Server or client applications that call the SSL_check_chain() function
2945 during or after a TLS 1.3 handshake may crash due to a NULL pointer
2946 dereference as a result of incorrect handling of the
2947 "signature_algorithms_cert" TLS extension. The crash occurs if an invalid
2948 or unrecognised signature algorithm is received from the peer. This could
2949 be exploited by a malicious peer in a Denial of Service attack.
2954 * Added AES consttime code for no-asm configurations
2955 an optional constant time support for AES was added
2956 when building openssl for no-asm.
2957 Enable with: ./config no-asm -DOPENSSL_AES_CONST_TIME
2958 Disable with: ./config no-asm -DOPENSSL_NO_AES_CONST_TIME
2959 At this time this feature is by default disabled.
2960 It will be enabled by default in 3.0.
2964 ### Changes between 1.1.1e and 1.1.1f [31 Mar 2020]
2966 * Revert the change of EOF detection while reading in libssl to avoid
2967 regressions in applications depending on the current way of reporting
2968 the EOF. As the existing method is not fully accurate the change to
2969 reporting the EOF via SSL_ERROR_SSL is kept on the current development
2970 branch and will be present in the 3.0 release.
2974 * Revised BN_generate_prime_ex to not avoid factors 3..17863 in p-1
2975 when primes for RSA keys are computed.
2976 Since we previously always generated primes == 2 (mod 3) for RSA keys,
2977 the 2-prime and 3-prime RSA modules were easy to distinguish, since
2978 N = p*q = 1 (mod 3), but N = p*q*r = 2 (mod 3). Therefore, fingerprinting
2979 2-prime vs. 3-prime RSA keys was possible by computing N mod 3.
2980 This avoids possible fingerprinting of newly generated RSA modules.
2984 ### Changes between 1.1.1d and 1.1.1e [17 Mar 2020]
2986 * Properly detect EOF while reading in libssl. Previously if we hit an EOF
2987 while reading in libssl then we would report an error back to the
2988 application (SSL_ERROR_SYSCALL) but errno would be 0. We now add
2989 an error to the stack (which means we instead return SSL_ERROR_SSL) and
2990 therefore give a hint as to what went wrong.
2994 * Check that ed25519 and ed448 are allowed by the security level. Previously
2995 signature algorithms not using an MD were not being checked that they were
2996 allowed by the security level.
3000 * Fixed SSL_get_servername() behaviour. The behaviour of SSL_get_servername()
3001 was not quite right. The behaviour was not consistent between resumption
3002 and normal handshakes, and also not quite consistent with historical
3003 behaviour. The behaviour in various scenarios has been clarified and
3004 it has been updated to make it match historical behaviour as closely as
3009 * *[VMS only]* The header files that the VMS compilers include automatically,
3010 `__DECC_INCLUDE_PROLOGUE.H` and `__DECC_INCLUDE_EPILOGUE.H`, use pragmas
3011 that the C++ compiler doesn't understand. This is a shortcoming in the
3012 compiler, but can be worked around with `__cplusplus` guards.
3014 C++ applications that use OpenSSL libraries must be compiled using the
3015 qualifier `/NAMES=(AS_IS,SHORTENED)` to be able to use all the OpenSSL
3016 functions. Otherwise, only functions with symbols of less than 31
3017 characters can be used, as the linker will not be able to successfully
3018 resolve symbols with longer names.
3022 * Added a new method to gather entropy on VMS, based on SYS$GET_ENTROPY.
3023 The presence of this system service is determined at run-time.
3027 * Print all values for a PKCS#12 attribute with 'openssl pkcs12', not just
3032 ### Changes between 1.1.1c and 1.1.1d [10 Sep 2019]
3034 * Fixed a fork protection issue. OpenSSL 1.1.1 introduced a rewritten random
3035 number generator (RNG). This was intended to include protection in the
3036 event of a fork() system call in order to ensure that the parent and child
3037 processes did not share the same RNG state. However, this protection was not
3038 being used in the default case.
3040 A partial mitigation for this issue is that the output from a high
3041 precision timer is mixed into the RNG state so the likelihood of a parent
3042 and child process sharing state is significantly reduced.
3044 If an application already calls OPENSSL_init_crypto() explicitly using
3045 OPENSSL_INIT_ATFORK then this problem does not occur at all.
3048 *Matthias St. Pierre*
3050 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
3051 used even when parsing explicit parameters, when loading a encoded key
3052 or calling `EC_GROUP_new_from_ecpkparameters()`/
3053 `EC_GROUP_new_from_ecparameters()`.
3054 This prevents bypass of security hardening and performance gains,
3055 especially for curves with specialized EC_METHODs.
3056 By default, if a key encoded with explicit parameters is loaded and later
3057 encoded, the output is still encoded with explicit parameters, even if
3058 internally a "named" EC_GROUP is used for computation.
3062 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
3063 this change, EC_GROUP_set_generator would accept order and/or cofactor as
3064 NULL. After this change, only the cofactor parameter can be NULL. It also
3065 does some minimal sanity checks on the passed order.
3070 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
3071 An attack is simple, if the first CMS_recipientInfo is valid but the
3072 second CMS_recipientInfo is chosen ciphertext. If the second
3073 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
3074 encryption key will be replaced by garbage, and the message cannot be
3075 decoded, but if the RSA decryption fails, the correct encryption key is
3076 used and the recipient will not notice the attack.
3077 As a work around for this potential attack the length of the decrypted
3078 key must be equal to the cipher default key length, in case the
3079 certificate is not given and all recipientInfo are tried out.
3080 The old behaviour can be re-enabled in the CMS code by setting the
3081 CMS_DEBUG_DECRYPT flag.
3086 * Early start up entropy quality from the DEVRANDOM seed source has been
3087 improved for older Linux systems. The RAND subsystem will wait for
3088 /dev/random to be producing output before seeding from /dev/urandom.
3089 The seeded state is stored for future library initialisations using
3090 a system global shared memory segment. The shared memory identifier
3091 can be configured by defining OPENSSL_RAND_SEED_DEVRANDOM_SHM_ID to
3092 the desired value. The default identifier is 114.
3096 * Correct the extended master secret constant on EBCDIC systems. Without this
3097 fix TLS connections between an EBCDIC system and a non-EBCDIC system that
3098 negotiate EMS will fail. Unfortunately this also means that TLS connections
3099 between EBCDIC systems with this fix, and EBCDIC systems without this
3100 fix will fail if they negotiate EMS.
3104 * Use Windows installation paths in the mingw builds
3106 Mingw isn't a POSIX environment per se, which means that Windows
3107 paths should be used for installation.
3112 * Changed DH_check to accept parameters with order q and 2q subgroups.
3113 With order 2q subgroups the bit 0 of the private key is not secret
3114 but DH_generate_key works around that by clearing bit 0 of the
3115 private key for those. This avoids leaking bit 0 of the private key.
3119 * Significantly reduce secure memory usage by the randomness pools.
3123 * Revert the DEVRANDOM_WAIT feature for Linux systems
3125 The DEVRANDOM_WAIT feature added a select() call to wait for the
3126 /dev/random device to become readable before reading from the
3127 /dev/urandom device.
3129 It turned out that this change had negative side effects on
3130 performance which were not acceptable. After some discussion it
3131 was decided to revert this feature and leave it up to the OS
3132 resp. the platform maintainer to ensure a proper initialization
3133 during early boot time.
3135 *Matthias St. Pierre*
3137 ### Changes between 1.1.1b and 1.1.1c [28 May 2019]
3139 * Add build tests for C++. These are generated files that only do one
3140 thing, to include one public OpenSSL head file each. This tests that
3141 the public header files can be usefully included in a C++ application.
3143 This test isn't enabled by default. It can be enabled with the option
3144 'enable-buildtest-c++'.
3148 * Enable SHA3 pre-hashing for ECDSA and DSA.
3152 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
3153 This changes the size when using the `genpkey` command when no size is given.
3154 It fixes an omission in earlier changes that changed all RSA, DSA and DH
3155 generation commands to use 2048 bits by default.
3159 * Reorganize the manual pages to consistently have RETURN VALUES,
3160 EXAMPLES, SEE ALSO and HISTORY come in that order, and adjust
3161 util/fix-doc-nits accordingly.
3163 *Paul Yang, Joshua Lock*
3165 * Add the missing accessor EVP_PKEY_get0_engine()
3169 * Have commands like `s_client` and `s_server` output the signature scheme
3170 along with other cipher suite parameters when debugging.
3174 * Make OPENSSL_config() error agnostic again.
3178 * Do the error handling in RSA decryption constant time.
3182 * Prevent over long nonces in ChaCha20-Poly1305.
3184 ChaCha20-Poly1305 is an AEAD cipher, and requires a unique nonce input
3185 for every encryption operation. RFC 7539 specifies that the nonce value
3186 (IV) should be 96 bits (12 bytes). OpenSSL allows a variable nonce length
3187 and front pads the nonce with 0 bytes if it is less than 12
3188 bytes. However it also incorrectly allows a nonce to be set of up to 16
3189 bytes. In this case only the last 12 bytes are significant and any
3190 additional leading bytes are ignored.
3192 It is a requirement of using this cipher that nonce values are
3193 unique. Messages encrypted using a reused nonce value are susceptible to
3194 serious confidentiality and integrity attacks. If an application changes
3195 the default nonce length to be longer than 12 bytes and then makes a
3196 change to the leading bytes of the nonce expecting the new value to be a
3197 new unique nonce then such an application could inadvertently encrypt
3198 messages with a reused nonce.
3200 Additionally the ignored bytes in a long nonce are not covered by the
3201 integrity guarantee of this cipher. Any application that relies on the
3202 integrity of these ignored leading bytes of a long nonce may be further
3203 affected. Any OpenSSL internal use of this cipher, including in SSL/TLS,
3204 is safe because no such use sets such a long nonce value. However user
3205 applications that use this cipher directly and set a non-default nonce
3206 length to be longer than 12 bytes may be vulnerable.
3208 This issue was reported to OpenSSL on 16th of March 2019 by Joran Dirk
3214 * Add DEVRANDOM_WAIT feature for Linux systems
3216 On older Linux systems where the getrandom() system call is not available,
3217 OpenSSL normally uses the /dev/urandom device for seeding its CSPRNG.
3218 Contrary to getrandom(), the /dev/urandom device will not block during
3219 early boot when the kernel CSPRNG has not been seeded yet.
3221 To mitigate this known weakness, use select() to wait for /dev/random to
3222 become readable before reading from /dev/urandom.
3224 * Ensure that SM2 only uses SM3 as digest algorithm
3228 ### Changes between 1.1.1a and 1.1.1b [26 Feb 2019]
3230 * Change the info callback signals for the start and end of a post-handshake
3231 message exchange in TLSv1.3. In 1.1.1/1.1.1a we used SSL_CB_HANDSHAKE_START
3232 and SSL_CB_HANDSHAKE_DONE. Experience has shown that many applications get
3233 confused by this and assume that a TLSv1.2 renegotiation has started. This
3234 can break KeyUpdate handling. Instead we no longer signal the start and end
3235 of a post handshake message exchange (although the messages themselves are
3236 still signalled). This could break some applications that were expecting
3237 the old signals. However without this KeyUpdate is not usable for many
3242 ### Changes between 1.1.1 and 1.1.1a [20 Nov 2018]
3244 * Timing vulnerability in DSA signature generation
3246 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
3247 timing side channel attack. An attacker could use variations in the signing
3248 algorithm to recover the private key.
3250 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
3255 * Timing vulnerability in ECDSA signature generation
3257 The OpenSSL ECDSA signature algorithm has been shown to be vulnerable to a
3258 timing side channel attack. An attacker could use variations in the signing
3259 algorithm to recover the private key.
3261 This issue was reported to OpenSSL on 25th October 2018 by Samuel Weiser.
3266 * Fixed the issue that RAND_add()/RAND_seed() silently discards random input
3267 if its length exceeds 4096 bytes. The limit has been raised to a buffer size
3268 of two gigabytes and the error handling improved.
3270 This issue was reported to OpenSSL by Dr. Falko Strenzke. It has been
3271 categorized as a normal bug, not a security issue, because the DRBG reseeds
3272 automatically and is fully functional even without additional randomness
3273 provided by the application.
3275 ### Changes between 1.1.0i and 1.1.1 [11 Sep 2018]
3277 * Add a new ClientHello callback. Provides a callback interface that gives
3278 the application the ability to adjust the nascent SSL object at the
3279 earliest stage of ClientHello processing, immediately after extensions have
3280 been collected but before they have been processed. In particular, this
3281 callback can adjust the supported TLS versions in response to the contents
3286 * Add SM2 base algorithm support.
3290 * s390x assembly pack: add (improved) hardware-support for the following
3291 cryptographic primitives: sha3, shake, aes-gcm, aes-ccm, aes-ctr, aes-ofb,
3292 aes-cfb/cfb8, aes-ecb.
3296 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
3297 parameter is no longer accepted, as it leads to a corrupt table. NULL
3298 pem_str is reserved for alias entries only.
3302 * Use the new ec_scalar_mul_ladder scaffold to implement a specialized ladder
3303 step for prime curves. The new implementation is based on formulae from
3304 differential addition-and-doubling in homogeneous projective coordinates
3305 from Izu-Takagi "A fast parallel elliptic curve multiplication resistant
3306 against side channel attacks" and Brier-Joye "Weierstrass Elliptic Curves
3307 and Side-Channel Attacks" Eq. (8) for y-coordinate recovery, modified
3308 to work in projective coordinates.
3310 *Billy Bob Brumley, Nicola Tuveri*
3312 * Change generating and checking of primes so that the error rate of not
3313 being prime depends on the intended use based on the size of the input.
3314 For larger primes this will result in more rounds of Miller-Rabin.
3315 The maximal error rate for primes with more than 1080 bits is lowered
3318 *Kurt Roeckx, Annie Yousar*
3320 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
3324 * The 'tsget' script is renamed to 'tsget.pl', to avoid confusion when
3325 moving between systems, and to avoid confusion when a Windows build is
3326 done with mingw vs with MSVC. For POSIX installs, there's still a
3327 symlink or copy named 'tsget' to avoid that confusion as well.
3331 * Revert blinding in ECDSA sign and instead make problematic addition
3332 length-invariant. Switch even to fixed-length Montgomery multiplication.
3336 * Use the new ec_scalar_mul_ladder scaffold to implement a specialized ladder
3337 step for binary curves. The new implementation is based on formulae from
3338 differential addition-and-doubling in mixed Lopez-Dahab projective
3339 coordinates, modified to independently blind the operands.
3341 *Billy Bob Brumley, Sohaib ul Hassan, Nicola Tuveri*
3343 * Add a scaffold to optionally enhance the Montgomery ladder implementation
3344 for `ec_scalar_mul_ladder` (formerly `ec_mul_consttime`) allowing
3345 EC_METHODs to implement their own specialized "ladder step", to take
3346 advantage of more favorable coordinate systems or more efficient
3347 differential addition-and-doubling algorithms.
3349 *Billy Bob Brumley, Sohaib ul Hassan, Nicola Tuveri*
3351 * Modified the random device based seed sources to keep the relevant
3352 file descriptors open rather than reopening them on each access.
3353 This allows such sources to operate in a chroot() jail without
3354 the associated device nodes being available. This behaviour can be
3355 controlled using RAND_keep_random_devices_open().
3359 * Numerous side-channel attack mitigations have been applied. This may have
3360 performance impacts for some algorithms for the benefit of improved
3361 security. Specific changes are noted in this change log by their respective
3366 * AIX shared library support overhaul. Switch to AIX "natural" way of
3367 handling shared libraries, which means collecting shared objects of
3368 different versions and bitnesses in one common archive. This allows to
3369 mitigate conflict between 1.0 and 1.1 side-by-side installations. It
3370 doesn't affect the way 3rd party applications are linked, only how
3371 multi-version installation is managed.
3375 * Make ec_group_do_inverse_ord() more robust and available to other
3376 EC cryptosystems, so that irrespective of BN_FLG_CONSTTIME, SCA
3377 mitigations are applied to the fallback BN_mod_inverse().
3378 When using this function rather than BN_mod_inverse() directly, new
3379 EC cryptosystem implementations are then safer-by-default.
3383 * Add coordinate blinding for EC_POINT and implement projective
3384 coordinate blinding for generic prime curves as a countermeasure to
3385 chosen point SCA attacks.
3387 *Sohaib ul Hassan, Nicola Tuveri, Billy Bob Brumley*
3389 * Add blinding to ECDSA and DSA signatures to protect against side channel
3390 attacks discovered by Keegan Ryan (NCC Group).
3394 * Enforce checking in the `pkeyutl` command to ensure that the input
3395 length does not exceed the maximum supported digest length when performing
3396 a sign, verify or verifyrecover operation.
3400 * SSL_MODE_AUTO_RETRY is enabled by default. Applications that use blocking
3401 I/O in combination with something like select() or poll() will hang. This
3402 can be turned off again using SSL_CTX_clear_mode().
3403 Many applications do not properly handle non-application data records, and
3404 TLS 1.3 sends more of such records. Setting SSL_MODE_AUTO_RETRY works
3405 around the problems in those applications, but can also break some.
3406 It's recommended to read the manpages about SSL_read(), SSL_write(),
3407 SSL_get_error(), SSL_shutdown(), SSL_CTX_set_mode() and
3408 SSL_CTX_set_read_ahead() again.
3412 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
3413 now allow empty (zero character) pass phrases.
3417 * Apply blinding to binary field modular inversion and remove patent
3418 pending (OPENSSL_SUN_GF2M_DIV) BN_GF2m_mod_div implementation.
3422 * Deprecate ec2_mult.c and unify scalar multiplication code paths for
3423 binary and prime elliptic curves.
3427 * Remove ECDSA nonce padding: EC_POINT_mul is now responsible for
3428 constant time fixed point multiplication.
3432 * Revise elliptic curve scalar multiplication with timing attack
3433 defenses: ec_wNAF_mul redirects to a constant time implementation
3434 when computing fixed point and variable point multiplication (which
3435 in OpenSSL are mostly used with secret scalars in keygen, sign,
3436 ECDH derive operations).
3437 *Billy Bob Brumley, Nicola Tuveri, Cesar Pereida García,
3440 * Updated CONTRIBUTING
3444 * Updated DRBG / RAND to request nonce and additional low entropy
3445 randomness from the system.
3447 *Matthias St. Pierre*
3449 * Updated 'openssl rehash' to use OpenSSL consistent default.
3453 * Moved the load of the ssl_conf module to libcrypto, which helps
3454 loading engines that libssl uses before libssl is initialised.
3458 * Added EVP_PKEY_sign() and EVP_PKEY_verify() for EdDSA
3462 * Fixed X509_NAME_ENTRY_set to get multi-valued RDNs right in all cases.
3464 *Ingo Schwarze, Rich Salz*
3466 * Added output of accepting IP address and port for 'openssl s_server'
3470 * Added a new API for TLSv1.3 ciphersuites:
3471 SSL_CTX_set_ciphersuites()
3472 SSL_set_ciphersuites()
3476 * Memory allocation failures consistently add an error to the error
3481 * Don't use OPENSSL_ENGINES and OPENSSL_CONF environment values
3482 in libcrypto when run as setuid/setgid.
3486 * Load any config file by default when libssl is used.
3490 * Added new public header file <openssl/rand_drbg.h> and documentation
3491 for the RAND_DRBG API. See manual page RAND_DRBG(7) for an overview.
3493 *Matthias St. Pierre*
3495 * QNX support removed (cannot find contributors to get their approval
3496 for the license change).
3500 * TLSv1.3 replay protection for early data has been implemented. See the
3501 SSL_read_early_data() man page for further details.
3505 * Separated TLSv1.3 ciphersuite configuration out from TLSv1.2 ciphersuite
3506 configuration. TLSv1.3 ciphersuites are not compatible with TLSv1.2 and
3507 below. Similarly TLSv1.2 ciphersuites are not compatible with TLSv1.3.
3508 In order to avoid issues where legacy TLSv1.2 ciphersuite configuration
3509 would otherwise inadvertently disable all TLSv1.3 ciphersuites the
3510 configuration has been separated out. See the ciphers man page or the
3511 SSL_CTX_set_ciphersuites() man page for more information.
3515 * On POSIX (BSD, Linux, ...) systems the ocsp(1) command running
3516 in responder mode now supports the new "-multi" option, which
3517 spawns the specified number of child processes to handle OCSP
3518 requests. The "-timeout" option now also limits the OCSP
3519 responder's patience to wait to receive the full client request
3520 on a newly accepted connection. Child processes are respawned
3521 as needed, and the CA index file is automatically reloaded
3522 when changed. This makes it possible to run the "ocsp" responder
3523 as a long-running service, making the OpenSSL CA somewhat more
3524 feature-complete. In this mode, most diagnostic messages logged
3525 after entering the event loop are logged via syslog(3) rather than
3530 * Added support for X448 and Ed448. Heavily based on original work by
3535 * Extend OSSL_STORE with capabilities to search and to narrow the set of
3536 objects loaded. This adds the functions OSSL_STORE_expect() and
3537 OSSL_STORE_find() as well as needed tools to construct searches and
3538 get the search data out of them.
3542 * Support for TLSv1.3 added. Note that users upgrading from an earlier
3543 version of OpenSSL should review their configuration settings to ensure
3544 that they are still appropriate for TLSv1.3. For further information see:
3545 <https://wiki.openssl.org/index.php/TLS1.3>
3549 * Grand redesign of the OpenSSL random generator
3551 The default RAND method now utilizes an AES-CTR DRBG according to
3552 NIST standard SP 800-90Ar1. The new random generator is essentially
3553 a port of the default random generator from the OpenSSL FIPS 2.0
3554 object module. It is a hybrid deterministic random bit generator
3555 using an AES-CTR bit stream and which seeds and reseeds itself
3556 automatically using trusted system entropy sources.
3558 Some of its new features are:
3559 - Support for multiple DRBG instances with seed chaining.
3560 - The default RAND method makes use of a DRBG.
3561 - There is a public and private DRBG instance.
3562 - The DRBG instances are fork-safe.
3563 - Keep all global DRBG instances on the secure heap if it is enabled.
3564 - The public and private DRBG instance are per thread for lock free
3567 *Paul Dale, Benjamin Kaduk, Kurt Roeckx, Rich Salz, Matthias St. Pierre*
3569 * Changed Configure so it only says what it does and doesn't dump
3570 so much data. Instead, ./configdata.pm should be used as a script
3571 to display all sorts of configuration data.
3575 * Added processing of "make variables" to Configure.
3579 * Added SHA512/224 and SHA512/256 algorithm support.
3583 * The last traces of Netware support, first removed in 1.1.0, have
3588 * Get rid of Makefile.shared, and in the process, make the processing
3589 of certain files (rc.obj, or the .def/.map/.opt files produced from
3590 the ordinal files) more visible and hopefully easier to trace and
3591 debug (or make silent).
3595 * Make it possible to have environment variable assignments as
3596 arguments to config / Configure.
3600 * Add multi-prime RSA (RFC 8017) support.
3604 * Add SM3 implemented according to GB/T 32905-2016
3605 *Jack Lloyd <jack.lloyd@ribose.com>,*
3606 *Ronald Tse <ronald.tse@ribose.com>,*
3607 *Erick Borsboom <erick.borsboom@ribose.com>*
3609 * Add 'Maximum Fragment Length' TLS extension negotiation and support
3610 as documented in RFC6066.
3611 Based on a patch from Tomasz Moń
3613 *Filipe Raimundo da Silva*
3615 * Add SM4 implemented according to GB/T 32907-2016.
3616 *Jack Lloyd <jack.lloyd@ribose.com>,*
3617 *Ronald Tse <ronald.tse@ribose.com>,*
3618 *Erick Borsboom <erick.borsboom@ribose.com>*
3620 * Reimplement -newreq-nodes and ERR_error_string_n; the
3621 original author does not agree with the license change.
3625 * Add ARIA AEAD TLS support.
3629 * Some macro definitions to support VS6 have been removed. Visual
3630 Studio 6 has not worked since 1.1.0
3634 * Add ERR_clear_last_mark(), to allow callers to clear the last mark
3635 without clearing the errors.
3639 * Add "atfork" functions. If building on a system that without
3640 pthreads, see doc/man3/OPENSSL_fork_prepare.pod for application
3641 requirements. The RAND facility now uses/requires this.
3649 * The UI API becomes a permanent and integral part of libcrypto, i.e.
3650 not possible to disable entirely. However, it's still possible to
3651 disable the console reading UI method, UI_OpenSSL() (use UI_null()
3654 To disable, configure with 'no-ui-console'. 'no-ui' is still
3655 possible to use as an alias. Check at compile time with the
3656 macro OPENSSL_NO_UI_CONSOLE. The macro OPENSSL_NO_UI is still
3657 possible to check and is an alias for OPENSSL_NO_UI_CONSOLE.
3661 * Add a STORE module, which implements a uniform and URI based reader of
3662 stores that can contain keys, certificates, CRLs and numerous other
3663 objects. The main API is loosely based on a few stdio functions,
3664 and includes OSSL_STORE_open, OSSL_STORE_load, OSSL_STORE_eof,
3665 OSSL_STORE_error and OSSL_STORE_close.
3666 The implementation uses backends called "loaders" to implement arbitrary
3667 URI schemes. There is one built in "loader" for the 'file' scheme.
3671 * Add devcrypto engine. This has been implemented against cryptodev-linux,
3672 then adjusted to work on FreeBSD 8.4 as well.
3673 Enable by configuring with 'enable-devcryptoeng'. This is done by default
3674 on BSD implementations, as cryptodev.h is assumed to exist on all of them.
3678 * Module names can prefixed with OSSL_ or OPENSSL_. This affects
3679 util/mkerr.pl, which is adapted to allow those prefixes, leading to
3680 error code calls like this:
3682 OSSL_FOOerr(OSSL_FOO_F_SOMETHING, OSSL_FOO_R_WHATEVER);
3684 With this change, we claim the namespaces OSSL and OPENSSL in a manner
3685 that can be encoded in C. For the foreseeable future, this will only
3688 *Richard Levitte and Tim Hudson*
3690 * Removed BSD cryptodev engine.
3694 * Add a build target 'build_all_generated', to build all generated files
3695 and only that. This can be used to prepare everything that requires
3696 things like perl for a system that lacks perl and then move everything
3697 to that system and do the rest of the build there.
3701 * In the UI interface, make it possible to duplicate the user data. This
3702 can be used by engines that need to retain the data for a longer time
3703 than just the call where this user data is passed.
3707 * Ignore the '-named_curve auto' value for compatibility of applications
3710 *Tomáš Mráz <tmraz@fedoraproject.org>*
3712 * Fragmented SSL/TLS alerts are no longer accepted. An alert message is 2
3713 bytes long. In theory it is permissible in SSLv3 - TLSv1.2 to fragment such
3714 alerts across multiple records (some of which could be empty). In practice
3715 it make no sense to send an empty alert record, or to fragment one. TLSv1.3
3716 prohibits this altogether and other libraries (BoringSSL, NSS) do not
3717 support this at all. Supporting it adds significant complexity to the
3718 record layer, and its removal is unlikely to cause interoperability
3723 * Add the ASN.1 types INT32, UINT32, INT64, UINT64 and variants prefixed
3724 with Z. These are meant to replace LONG and ZLONG and to be size safe.
3725 The use of LONG and ZLONG is discouraged and scheduled for deprecation
3730 * Add the 'z' and 'j' modifiers to BIO_printf() et al formatting string,
3731 'z' is to be used for [s]size_t, and 'j' - with [u]int64_t.
3733 *Richard Levitte, Andy Polyakov*
3735 * Add EC_KEY_get0_engine(), which does for EC_KEY what RSA_get0_engine()
3740 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
3741 platform rather than 'mingw'.
3745 * The functions X509_STORE_add_cert and X509_STORE_add_crl return
3746 success if they are asked to add an object which already exists
3747 in the store. This change cascades to other functions which load
3748 certificates and CRLs.
3752 * x86_64 assembly pack: annotate code with DWARF CFI directives to
3753 facilitate stack unwinding even from assembly subroutines.
3757 * Remove VAX C specific definitions of OPENSSL_EXPORT, OPENSSL_EXTERN.
3758 Also remove OPENSSL_GLOBAL entirely, as it became a no-op.
3762 * Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
3763 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
3764 which is the minimum version we support.
3768 * Certificate time validation (X509_cmp_time) enforces stricter
3769 compliance with RFC 5280. Fractional seconds and timezone offsets
3770 are no longer allowed.
3774 * Add support for ARIA
3778 * s_client will now send the Server Name Indication (SNI) extension by
3779 default unless the new "-noservername" option is used. The server name is
3780 based on the host provided to the "-connect" option unless overridden by
3781 using "-servername".
3785 * Add support for SipHash
3789 * OpenSSL now fails if it receives an unrecognised record type in TLS1.0
3790 or TLS1.1. Previously this only happened in SSLv3 and TLS1.2. This is to
3791 prevent issues where no progress is being made and the peer continually
3792 sends unrecognised record types, using up resources processing them.
3796 * 'openssl passwd' can now produce SHA256 and SHA512 based output,
3797 using the algorithm defined in
3798 <https://www.akkadia.org/drepper/SHA-crypt.txt>
3802 * Heartbeat support has been removed; the ABI is changed for now.
3804 *Richard Levitte, Rich Salz*
3806 * Support for SSL_OP_NO_ENCRYPT_THEN_MAC in SSL_CONF_cmd.
3810 * The RSA "null" method, which was partially supported to avoid patent
3811 issues, has been replaced to always returns NULL.
3818 ### Changes between 1.1.0k and 1.1.0l [10 Sep 2019]
3820 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
3821 used even when parsing explicit parameters, when loading a encoded key
3822 or calling `EC_GROUP_new_from_ecpkparameters()`/
3823 `EC_GROUP_new_from_ecparameters()`.
3824 This prevents bypass of security hardening and performance gains,
3825 especially for curves with specialized EC_METHODs.
3826 By default, if a key encoded with explicit parameters is loaded and later
3827 encoded, the output is still encoded with explicit parameters, even if
3828 internally a "named" EC_GROUP is used for computation.
3832 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
3833 this change, EC_GROUP_set_generator would accept order and/or cofactor as
3834 NULL. After this change, only the cofactor parameter can be NULL. It also
3835 does some minimal sanity checks on the passed order.
3840 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
3841 An attack is simple, if the first CMS_recipientInfo is valid but the
3842 second CMS_recipientInfo is chosen ciphertext. If the second
3843 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
3844 encryption key will be replaced by garbage, and the message cannot be
3845 decoded, but if the RSA decryption fails, the correct encryption key is
3846 used and the recipient will not notice the attack.
3847 As a work around for this potential attack the length of the decrypted
3848 key must be equal to the cipher default key length, in case the
3849 certificate is not given and all recipientInfo are tried out.
3850 The old behaviour can be re-enabled in the CMS code by setting the
3851 CMS_DEBUG_DECRYPT flag.
3856 * Use Windows installation paths in the mingw builds
3858 Mingw isn't a POSIX environment per se, which means that Windows
3859 paths should be used for installation.
3864 ### Changes between 1.1.0j and 1.1.0k [28 May 2019]
3866 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
3867 This changes the size when using the `genpkey` command when no size is given.
3868 It fixes an omission in earlier changes that changed all RSA, DSA and DH
3869 generation commands to use 2048 bits by default.
3873 * Prevent over long nonces in ChaCha20-Poly1305.
3875 ChaCha20-Poly1305 is an AEAD cipher, and requires a unique nonce input
3876 for every encryption operation. RFC 7539 specifies that the nonce value
3877 (IV) should be 96 bits (12 bytes). OpenSSL allows a variable nonce length
3878 and front pads the nonce with 0 bytes if it is less than 12
3879 bytes. However it also incorrectly allows a nonce to be set of up to 16
3880 bytes. In this case only the last 12 bytes are significant and any
3881 additional leading bytes are ignored.
3883 It is a requirement of using this cipher that nonce values are
3884 unique. Messages encrypted using a reused nonce value are susceptible to
3885 serious confidentiality and integrity attacks. If an application changes
3886 the default nonce length to be longer than 12 bytes and then makes a
3887 change to the leading bytes of the nonce expecting the new value to be a
3888 new unique nonce then such an application could inadvertently encrypt
3889 messages with a reused nonce.
3891 Additionally the ignored bytes in a long nonce are not covered by the
3892 integrity guarantee of this cipher. Any application that relies on the
3893 integrity of these ignored leading bytes of a long nonce may be further
3894 affected. Any OpenSSL internal use of this cipher, including in SSL/TLS,
3895 is safe because no such use sets such a long nonce value. However user
3896 applications that use this cipher directly and set a non-default nonce
3897 length to be longer than 12 bytes may be vulnerable.
3899 This issue was reported to OpenSSL on 16th of March 2019 by Joran Dirk
3905 * Added SCA hardening for modular field inversion in EC_GROUP through
3906 a new dedicated field_inv() pointer in EC_METHOD.
3907 This also addresses a leakage affecting conversions from projective
3908 to affine coordinates.
3910 *Billy Bob Brumley, Nicola Tuveri*
3912 * Fix a use after free bug in d2i_X509_PUBKEY when overwriting a
3913 re-used X509_PUBKEY object if the second PUBKEY is malformed.
3917 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
3921 * Remove the 'dist' target and add a tarball building script. The
3922 'dist' target has fallen out of use, and it shouldn't be
3923 necessary to configure just to create a source distribution.
3927 ### Changes between 1.1.0i and 1.1.0j [20 Nov 2018]
3929 * Timing vulnerability in DSA signature generation
3931 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
3932 timing side channel attack. An attacker could use variations in the signing
3933 algorithm to recover the private key.
3935 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
3940 * Timing vulnerability in ECDSA signature generation
3942 The OpenSSL ECDSA signature algorithm has been shown to be vulnerable to a
3943 timing side channel attack. An attacker could use variations in the signing
3944 algorithm to recover the private key.
3946 This issue was reported to OpenSSL on 25th October 2018 by Samuel Weiser.
3951 * Add coordinate blinding for EC_POINT and implement projective
3952 coordinate blinding for generic prime curves as a countermeasure to
3953 chosen point SCA attacks.
3955 *Sohaib ul Hassan, Nicola Tuveri, Billy Bob Brumley*
3957 ### Changes between 1.1.0h and 1.1.0i [14 Aug 2018]
3959 * Client DoS due to large DH parameter
3961 During key agreement in a TLS handshake using a DH(E) based ciphersuite a
3962 malicious server can send a very large prime value to the client. This will
3963 cause the client to spend an unreasonably long period of time generating a
3964 key for this prime resulting in a hang until the client has finished. This
3965 could be exploited in a Denial Of Service attack.
3967 This issue was reported to OpenSSL on 5th June 2018 by Guido Vranken
3972 * Cache timing vulnerability in RSA Key Generation
3974 The OpenSSL RSA Key generation algorithm has been shown to be vulnerable to
3975 a cache timing side channel attack. An attacker with sufficient access to
3976 mount cache timing attacks during the RSA key generation process could
3977 recover the private key.
3979 This issue was reported to OpenSSL on 4th April 2018 by Alejandro Cabrera
3980 Aldaya, Billy Brumley, Cesar Pereida Garcia and Luis Manuel Alvarez Tapia.
3985 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
3986 parameter is no longer accepted, as it leads to a corrupt table. NULL
3987 pem_str is reserved for alias entries only.
3991 * Revert blinding in ECDSA sign and instead make problematic addition
3992 length-invariant. Switch even to fixed-length Montgomery multiplication.
3996 * Change generating and checking of primes so that the error rate of not
3997 being prime depends on the intended use based on the size of the input.
3998 For larger primes this will result in more rounds of Miller-Rabin.
3999 The maximal error rate for primes with more than 1080 bits is lowered
4002 *Kurt Roeckx, Annie Yousar*
4004 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
4008 * Add blinding to ECDSA and DSA signatures to protect against side channel
4009 attacks discovered by Keegan Ryan (NCC Group).
4013 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
4014 now allow empty (zero character) pass phrases.
4018 * Certificate time validation (X509_cmp_time) enforces stricter
4019 compliance with RFC 5280. Fractional seconds and timezone offsets
4020 are no longer allowed.
4024 * Fixed a text canonicalisation bug in CMS
4026 Where a CMS detached signature is used with text content the text goes
4027 through a canonicalisation process first prior to signing or verifying a
4028 signature. This process strips trailing space at the end of lines, converts
4029 line terminators to CRLF and removes additional trailing line terminators
4030 at the end of a file. A bug in the canonicalisation process meant that
4031 some characters, such as form-feed, were incorrectly treated as whitespace
4032 and removed. This is contrary to the specification (RFC5485). This fix
4033 could mean that detached text data signed with an earlier version of
4034 OpenSSL 1.1.0 may fail to verify using the fixed version, or text data
4035 signed with a fixed OpenSSL may fail to verify with an earlier version of
4036 OpenSSL 1.1.0. A workaround is to only verify the canonicalised text data
4037 and use the "-binary" flag (for the "cms" command line application) or set
4038 the SMIME_BINARY/PKCS7_BINARY/CMS_BINARY flags (if using CMS_verify()).
4042 ### Changes between 1.1.0g and 1.1.0h [27 Mar 2018]
4044 * Constructed ASN.1 types with a recursive definition could exceed the stack
4046 Constructed ASN.1 types with a recursive definition (such as can be found
4047 in PKCS7) could eventually exceed the stack given malicious input with
4048 excessive recursion. This could result in a Denial Of Service attack. There
4049 are no such structures used within SSL/TLS that come from untrusted sources
4050 so this is considered safe.
4052 This issue was reported to OpenSSL on 4th January 2018 by the OSS-fuzz
4058 * Incorrect CRYPTO_memcmp on HP-UX PA-RISC
4060 Because of an implementation bug the PA-RISC CRYPTO_memcmp function is
4061 effectively reduced to only comparing the least significant bit of each
4062 byte. This allows an attacker to forge messages that would be considered as
4063 authenticated in an amount of tries lower than that guaranteed by the
4064 security claims of the scheme. The module can only be compiled by the
4065 HP-UX assembler, so that only HP-UX PA-RISC targets are affected.
4067 This issue was reported to OpenSSL on 2nd March 2018 by Peter Waltenberg
4073 * Add a build target 'build_all_generated', to build all generated files
4074 and only that. This can be used to prepare everything that requires
4075 things like perl for a system that lacks perl and then move everything
4076 to that system and do the rest of the build there.
4080 * Backport SSL_OP_NO_RENGOTIATION
4082 OpenSSL 1.0.2 and below had the ability to disable renegotiation using the
4083 (undocumented) SSL3_FLAGS_NO_RENEGOTIATE_CIPHERS flag. Due to the opacity
4084 changes this is no longer possible in 1.1.0. Therefore, the new
4085 SSL_OP_NO_RENEGOTIATION option from 1.1.1-dev has been backported to
4086 1.1.0 to provide equivalent functionality.
4088 Note that if an application built against 1.1.0h headers (or above) is run
4089 using an older version of 1.1.0 (prior to 1.1.0h) then the option will be
4090 accepted but nothing will happen, i.e. renegotiation will not be prevented.
4094 * Removed the OS390-Unix config target. It relied on a script that doesn't
4099 * rsaz_1024_mul_avx2 overflow bug on x86_64
4101 There is an overflow bug in the AVX2 Montgomery multiplication procedure
4102 used in exponentiation with 1024-bit moduli. No EC algorithms are affected.
4103 Analysis suggests that attacks against RSA and DSA as a result of this
4104 defect would be very difficult to perform and are not believed likely.
4105 Attacks against DH1024 are considered just feasible, because most of the
4106 work necessary to deduce information about a private key may be performed
4107 offline. The amount of resources required for such an attack would be
4108 significant. However, for an attack on TLS to be meaningful, the server
4109 would have to share the DH1024 private key among multiple clients, which is
4110 no longer an option since CVE-2016-0701.
4112 This only affects processors that support the AVX2 but not ADX extensions
4113 like Intel Haswell (4th generation).
4115 This issue was reported to OpenSSL by David Benjamin (Google). The issue
4116 was originally found via the OSS-Fuzz project.
4121 ### Changes between 1.1.0f and 1.1.0g [2 Nov 2017]
4123 * bn_sqrx8x_internal carry bug on x86_64
4125 There is a carry propagating bug in the x86_64 Montgomery squaring
4126 procedure. No EC algorithms are affected. Analysis suggests that attacks
4127 against RSA and DSA as a result of this defect would be very difficult to
4128 perform and are not believed likely. Attacks against DH are considered just
4129 feasible (although very difficult) because most of the work necessary to
4130 deduce information about a private key may be performed offline. The amount
4131 of resources required for such an attack would be very significant and
4132 likely only accessible to a limited number of attackers. An attacker would
4133 additionally need online access to an unpatched system using the target
4134 private key in a scenario with persistent DH parameters and a private
4135 key that is shared between multiple clients.
4137 This only affects processors that support the BMI1, BMI2 and ADX extensions
4138 like Intel Broadwell (5th generation) and later or AMD Ryzen.
4140 This issue was reported to OpenSSL by the OSS-Fuzz project.
4145 * Malformed X.509 IPAddressFamily could cause OOB read
4147 If an X.509 certificate has a malformed IPAddressFamily extension,
4148 OpenSSL could do a one-byte buffer overread. The most likely result
4149 would be an erroneous display of the certificate in text format.
4151 This issue was reported to OpenSSL by the OSS-Fuzz project.
4156 ### Changes between 1.1.0e and 1.1.0f [25 May 2017]
4158 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
4159 platform rather than 'mingw'.
4163 * Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
4164 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
4165 which is the minimum version we support.
4169 ### Changes between 1.1.0d and 1.1.0e [16 Feb 2017]
4171 * Encrypt-Then-Mac renegotiation crash
4173 During a renegotiation handshake if the Encrypt-Then-Mac extension is
4174 negotiated where it was not in the original handshake (or vice-versa) then
4175 this can cause OpenSSL to crash (dependent on ciphersuite). Both clients
4176 and servers are affected.
4178 This issue was reported to OpenSSL by Joe Orton (Red Hat).
4183 ### Changes between 1.1.0c and 1.1.0d [26 Jan 2017]
4185 * Truncated packet could crash via OOB read
4187 If one side of an SSL/TLS path is running on a 32-bit host and a specific
4188 cipher is being used, then a truncated packet can cause that host to
4189 perform an out-of-bounds read, usually resulting in a crash.
4191 This issue was reported to OpenSSL by Robert Święcki of Google.
4196 * Bad (EC)DHE parameters cause a client crash
4198 If a malicious server supplies bad parameters for a DHE or ECDHE key
4199 exchange then this can result in the client attempting to dereference a
4200 NULL pointer leading to a client crash. This could be exploited in a Denial
4203 This issue was reported to OpenSSL by Guido Vranken.
4208 * BN_mod_exp may produce incorrect results on x86_64
4210 There is a carry propagating bug in the x86_64 Montgomery squaring
4211 procedure. No EC algorithms are affected. Analysis suggests that attacks
4212 against RSA and DSA as a result of this defect would be very difficult to
4213 perform and are not believed likely. Attacks against DH are considered just
4214 feasible (although very difficult) because most of the work necessary to
4215 deduce information about a private key may be performed offline. The amount
4216 of resources required for such an attack would be very significant and
4217 likely only accessible to a limited number of attackers. An attacker would
4218 additionally need online access to an unpatched system using the target
4219 private key in a scenario with persistent DH parameters and a private
4220 key that is shared between multiple clients. For example this can occur by
4221 default in OpenSSL DHE based SSL/TLS ciphersuites. Note: This issue is very
4222 similar to CVE-2015-3193 but must be treated as a separate problem.
4224 This issue was reported to OpenSSL by the OSS-Fuzz project.
4229 ### Changes between 1.1.0b and 1.1.0c [10 Nov 2016]
4231 * ChaCha20/Poly1305 heap-buffer-overflow
4233 TLS connections using `*-CHACHA20-POLY1305` ciphersuites are susceptible to
4234 a DoS attack by corrupting larger payloads. This can result in an OpenSSL
4235 crash. This issue is not considered to be exploitable beyond a DoS.
4237 This issue was reported to OpenSSL by Robert Święcki (Google Security Team)
4242 * CMS Null dereference
4244 Applications parsing invalid CMS structures can crash with a NULL pointer
4245 dereference. This is caused by a bug in the handling of the ASN.1 CHOICE
4246 type in OpenSSL 1.1.0 which can result in a NULL value being passed to the
4247 structure callback if an attempt is made to free certain invalid encodings.
4248 Only CHOICE structures using a callback which do not handle NULL value are
4251 This issue was reported to OpenSSL by Tyler Nighswander of ForAllSecure.
4256 * Montgomery multiplication may produce incorrect results
4258 There is a carry propagating bug in the Broadwell-specific Montgomery
4259 multiplication procedure that handles input lengths divisible by, but
4260 longer than 256 bits. Analysis suggests that attacks against RSA, DSA
4261 and DH private keys are impossible. This is because the subroutine in
4262 question is not used in operations with the private key itself and an input
4263 of the attacker's direct choice. Otherwise the bug can manifest itself as
4264 transient authentication and key negotiation failures or reproducible
4265 erroneous outcome of public-key operations with specially crafted input.
4266 Among EC algorithms only Brainpool P-512 curves are affected and one
4267 presumably can attack ECDH key negotiation. Impact was not analyzed in
4268 detail, because pre-requisites for attack are considered unlikely. Namely
4269 multiple clients have to choose the curve in question and the server has to
4270 share the private key among them, neither of which is default behaviour.
4271 Even then only clients that chose the curve will be affected.
4273 This issue was publicly reported as transient failures and was not
4274 initially recognized as a security issue. Thanks to Richard Morgan for
4275 providing reproducible case.
4280 * Removed automatic addition of RPATH in shared libraries and executables,
4281 as this was a remainder from OpenSSL 1.0.x and isn't needed any more.
4285 ### Changes between 1.1.0a and 1.1.0b [26 Sep 2016]
4287 * Fix Use After Free for large message sizes
4289 The patch applied to address CVE-2016-6307 resulted in an issue where if a
4290 message larger than approx 16k is received then the underlying buffer to
4291 store the incoming message is reallocated and moved. Unfortunately a
4292 dangling pointer to the old location is left which results in an attempt to
4293 write to the previously freed location. This is likely to result in a
4294 crash, however it could potentially lead to execution of arbitrary code.
4296 This issue only affects OpenSSL 1.1.0a.
4298 This issue was reported to OpenSSL by Robert Święcki.
4303 ### Changes between 1.1.0 and 1.1.0a [22 Sep 2016]
4305 * OCSP Status Request extension unbounded memory growth
4307 A malicious client can send an excessively large OCSP Status Request
4308 extension. If that client continually requests renegotiation, sending a
4309 large OCSP Status Request extension each time, then there will be unbounded
4310 memory growth on the server. This will eventually lead to a Denial Of
4311 Service attack through memory exhaustion. Servers with a default
4312 configuration are vulnerable even if they do not support OCSP. Builds using
4313 the "no-ocsp" build time option are not affected.
4315 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4320 * SSL_peek() hang on empty record
4322 OpenSSL 1.1.0 SSL/TLS will hang during a call to SSL_peek() if the peer
4323 sends an empty record. This could be exploited by a malicious peer in a
4324 Denial Of Service attack.
4326 This issue was reported to OpenSSL by Alex Gaynor.
4331 * Excessive allocation of memory in tls_get_message_header() and
4332 dtls1_preprocess_fragment()
4334 A (D)TLS message includes 3 bytes for its length in the header for the
4335 message. This would allow for messages up to 16Mb in length. Messages of
4336 this length are excessive and OpenSSL includes a check to ensure that a
4337 peer is sending reasonably sized messages in order to avoid too much memory
4338 being consumed to service a connection. A flaw in the logic of version
4339 1.1.0 means that memory for the message is allocated too early, prior to
4340 the excessive message length check. Due to way memory is allocated in
4341 OpenSSL this could mean an attacker could force up to 21Mb to be allocated
4342 to service a connection. This could lead to a Denial of Service through
4343 memory exhaustion. However, the excessive message length check still takes
4344 place, and this would cause the connection to immediately fail. Assuming
4345 that the application calls SSL_free() on the failed connection in a timely
4346 manner then the 21Mb of allocated memory will then be immediately freed
4347 again. Therefore, the excessive memory allocation will be transitory in
4348 nature. This then means that there is only a security impact if:
4350 1) The application does not call SSL_free() in a timely manner in the event
4351 that the connection fails
4353 2) The application is working in a constrained environment where there is
4354 very little free memory
4356 3) The attacker initiates multiple connection attempts such that there are
4357 multiple connections in a state where memory has been allocated for the
4358 connection; SSL_free() has not yet been called; and there is insufficient
4359 memory to service the multiple requests.
4361 Except in the instance of (1) above any Denial Of Service is likely to be
4362 transitory because as soon as the connection fails the memory is
4363 subsequently freed again in the SSL_free() call. However there is an
4364 increased risk during this period of application crashes due to the lack of
4365 memory - which would then mean a more serious Denial of Service.
4367 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4368 (CVE-2016-6307 and CVE-2016-6308)
4372 * solaris-x86-cc, i.e. 32-bit configuration with vendor compiler,
4373 had to be removed. Primary reason is that vendor assembler can't
4374 assemble our modules with -KPIC flag. As result it, assembly
4375 support, was not even available as option. But its lack means
4376 lack of side-channel resistant code, which is incompatible with
4377 security by todays standards. Fortunately gcc is readily available
4378 prepackaged option, which we firmly point at...
4382 ### Changes between 1.0.2h and 1.1.0 [25 Aug 2016]
4384 * Windows command-line tool supports UTF-8 opt-in option for arguments
4385 and console input. Setting OPENSSL_WIN32_UTF8 environment variable
4386 (to any value) allows Windows user to access PKCS#12 file generated
4387 with Windows CryptoAPI and protected with non-ASCII password, as well
4388 as files generated under UTF-8 locale on Linux also protected with
4393 * To mitigate the SWEET32 attack ([CVE-2016-2183]), 3DES cipher suites
4394 have been disabled by default and removed from DEFAULT, just like RC4.
4395 See the RC4 item below to re-enable both.
4399 * The method for finding the storage location for the Windows RAND seed file
4400 has changed. First we check %RANDFILE%. If that is not set then we check
4401 the directories %HOME%, %USERPROFILE% and %SYSTEMROOT% in that order. If
4402 all else fails we fall back to C:\.
4406 * The EVP_EncryptUpdate() function has had its return type changed from void
4407 to int. A return of 0 indicates and error while a return of 1 indicates
4412 * The flags RSA_FLAG_NO_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME and
4413 DH_FLAG_NO_EXP_CONSTTIME which previously provided the ability to switch
4414 off the constant time implementation for RSA, DSA and DH have been made
4415 no-ops and deprecated.
4419 * Windows RAND implementation was simplified to only get entropy by
4420 calling CryptGenRandom(). Various other RAND-related tickets
4423 *Joseph Wylie Yandle, Rich Salz*
4425 * The stack and lhash API's were renamed to start with `OPENSSL_SK_`
4426 and `OPENSSL_LH_`, respectively. The old names are available
4427 with API compatibility. They new names are now completely documented.
4431 * Unify TYPE_up_ref(obj) methods signature.
4432 SSL_CTX_up_ref(), SSL_up_ref(), X509_up_ref(), EVP_PKEY_up_ref(),
4433 X509_CRL_up_ref(), X509_OBJECT_up_ref_count() methods are now returning an
4434 int (instead of void) like all others TYPE_up_ref() methods.
4435 So now these methods also check the return value of CRYPTO_atomic_add(),
4436 and the validity of object reference counter.
4438 *fdasilvayy@gmail.com*
4440 * With Windows Visual Studio builds, the .pdb files are installed
4441 alongside the installed libraries and executables. For a static
4442 library installation, ossl_static.pdb is the associate compiler
4443 generated .pdb file to be used when linking programs.
4447 * Remove openssl.spec. Packaging files belong with the packagers.
4451 * Automatic Darwin/OSX configuration has had a refresh, it will now
4452 recognise x86_64 architectures automatically. You can still decide
4453 to build for a different bitness with the environment variable
4454 KERNEL_BITS (can be 32 or 64), for example:
4456 KERNEL_BITS=32 ./config
4460 * Change default algorithms in pkcs8 utility to use PKCS#5 v2.0,
4461 256 bit AES and HMAC with SHA256.
4465 * Remove support for MIPS o32 ABI on IRIX (and IRIX only).
4469 * Triple-DES ciphers have been moved from HIGH to MEDIUM.
4473 * To enable users to have their own config files and build file templates,
4474 Configure looks in the directory indicated by the environment variable
4475 OPENSSL_LOCAL_CONFIG_DIR as well as the in-source Configurations/
4476 directory. On VMS, OPENSSL_LOCAL_CONFIG_DIR is expected to be a logical
4477 name and is used as is.
4481 * The following datatypes were made opaque: X509_OBJECT, X509_STORE_CTX,
4482 X509_STORE, X509_LOOKUP, and X509_LOOKUP_METHOD. The unused type
4483 X509_CERT_FILE_CTX was removed.
4487 * "shared" builds are now the default. To create only static libraries use
4488 the "no-shared" Configure option.
4492 * Remove the no-aes, no-hmac, no-rsa, no-sha and no-md5 Configure options.
4493 All of these option have not worked for some while and are fundamental
4498 * Make various cleanup routines no-ops and mark them as deprecated. Most
4499 global cleanup functions are no longer required because they are handled
4500 via auto-deinit (see OPENSSL_init_crypto and OPENSSL_init_ssl man pages).
4501 Explicitly de-initing can cause problems (e.g. where a library that uses
4502 OpenSSL de-inits, but an application is still using it). The affected
4503 functions are CONF_modules_free(), ENGINE_cleanup(), OBJ_cleanup(),
4504 EVP_cleanup(), BIO_sock_cleanup(), CRYPTO_cleanup_all_ex_data(),
4505 RAND_cleanup(), SSL_COMP_free_compression_methods(), ERR_free_strings() and
4506 COMP_zlib_cleanup().
4510 * --strict-warnings no longer enables runtime debugging options
4511 such as REF_DEBUG. Instead, debug options are automatically
4512 enabled with '--debug' builds.
4514 *Andy Polyakov, Emilia Käsper*
4516 * Made DH and DH_METHOD opaque. The structures for managing DH objects
4517 have been moved out of the public header files. New functions for managing
4518 these have been added.
4522 * Made RSA and RSA_METHOD opaque. The structures for managing RSA
4523 objects have been moved out of the public header files. New
4524 functions for managing these have been added.
4528 * Made DSA and DSA_METHOD opaque. The structures for managing DSA objects
4529 have been moved out of the public header files. New functions for managing
4530 these have been added.
4534 * Made BIO and BIO_METHOD opaque. The structures for managing BIOs have been
4535 moved out of the public header files. New functions for managing these
4540 * Removed no-rijndael as a config option. Rijndael is an old name for AES.
4544 * Removed the mk1mf build scripts.
4548 * Headers are now wrapped, if necessary, with OPENSSL_NO_xxx, so
4549 it is always safe to #include a header now.
4553 * Removed the aged BC-32 config and all its supporting scripts
4557 * Removed support for Ultrix, Netware, and OS/2.
4561 * Add support for HKDF.
4563 *Alessandro Ghedini*
4565 * Add support for blake2b and blake2s
4569 * Added support for "pipelining". Ciphers that have the
4570 EVP_CIPH_FLAG_PIPELINE flag set have a capability to process multiple
4571 encryptions/decryptions simultaneously. There are currently no built-in
4572 ciphers with this property but the expectation is that engines will be able
4573 to offer it to significantly improve throughput. Support has been extended
4574 into libssl so that multiple records for a single connection can be
4575 processed in one go (for >=TLS 1.1).
4579 * Added the AFALG engine. This is an async capable engine which is able to
4580 offload work to the Linux kernel. In this initial version it only supports
4581 AES128-CBC. The kernel must be version 4.1.0 or greater.
4585 * OpenSSL now uses a new threading API. It is no longer necessary to
4586 set locking callbacks to use OpenSSL in a multi-threaded environment. There
4587 are two supported threading models: pthreads and windows threads. It is
4588 also possible to configure OpenSSL at compile time for "no-threads". The
4589 old threading API should no longer be used. The functions have been
4590 replaced with "no-op" compatibility macros.
4592 *Alessandro Ghedini, Matt Caswell*
4594 * Modify behavior of ALPN to invoke callback after SNI/servername
4595 callback, such that updates to the SSL_CTX affect ALPN.
4599 * Add SSL_CIPHER queries for authentication and key-exchange.
4603 * Changes to the DEFAULT cipherlist:
4604 - Prefer (EC)DHE handshakes over plain RSA.
4605 - Prefer AEAD ciphers over legacy ciphers.
4606 - Prefer ECDSA over RSA when both certificates are available.
4607 - Prefer TLSv1.2 ciphers/PRF.
4608 - Remove DSS, SEED, IDEA, CAMELLIA, and AES-CCM from the
4613 * Change the ECC default curve list to be this, in order: x25519,
4614 secp256r1, secp521r1, secp384r1.
4618 * RC4 based libssl ciphersuites are now classed as "weak" ciphers and are
4619 disabled by default. They can be re-enabled using the
4620 enable-weak-ssl-ciphers option to Configure.
4624 * If the server has ALPN configured, but supports no protocols that the
4625 client advertises, send a fatal "no_application_protocol" alert.
4626 This behaviour is SHALL in RFC 7301, though it isn't universally
4627 implemented by other servers.
4631 * Add X25519 support.
4632 Add ASN.1 and EVP_PKEY methods for X25519. This includes support
4633 for public and private key encoding using the format documented in
4634 draft-ietf-curdle-pkix-02. The corresponding EVP_PKEY method supports
4635 key generation and key derivation.
4637 TLS support complies with draft-ietf-tls-rfc4492bis-08 and uses
4642 * Deprecate SRP_VBASE_get_by_user.
4643 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
4644 In order to fix an unavoidable memory leak ([CVE-2016-0798]),
4645 SRP_VBASE_get_by_user was changed to ignore the "fake user" SRP
4646 seed, even if the seed is configured.
4648 Users should use SRP_VBASE_get1_by_user instead. Note that in
4649 SRP_VBASE_get1_by_user, caller must free the returned value. Note
4650 also that even though configuring the SRP seed attempts to hide
4651 invalid usernames by continuing the handshake with fake
4652 credentials, this behaviour is not constant time and no strong
4653 guarantees are made that the handshake is indistinguishable from
4654 that of a valid user.
4658 * Configuration change; it's now possible to build dynamic engines
4659 without having to build shared libraries and vice versa. This
4660 only applies to the engines in `engines/`, those in `crypto/engine/`
4661 will always be built into libcrypto (i.e. "static").
4663 Building dynamic engines is enabled by default; to disable, use
4664 the configuration option "disable-dynamic-engine".
4666 The only requirements for building dynamic engines are the
4667 presence of the DSO module and building with position independent
4668 code, so they will also automatically be disabled if configuring
4669 with "disable-dso" or "disable-pic".
4671 The macros OPENSSL_NO_STATIC_ENGINE and OPENSSL_NO_DYNAMIC_ENGINE
4672 are also taken away from openssl/opensslconf.h, as they are
4677 * Configuration change; if there is a known flag to compile
4678 position independent code, it will always be applied on the
4679 libcrypto and libssl object files, and never on the application
4680 object files. This means other libraries that use routines from
4681 libcrypto / libssl can be made into shared libraries regardless
4682 of how OpenSSL was configured.
4684 If this isn't desirable, the configuration options "disable-pic"
4685 or "no-pic" can be used to disable the use of PIC. This will
4686 also disable building shared libraries and dynamic engines.
4690 * Removed JPAKE code. It was experimental and has no wide use.
4694 * The INSTALL_PREFIX Makefile variable has been renamed to
4695 DESTDIR. That makes for less confusion on what this variable
4696 is for. Also, the configuration option --install_prefix is
4701 * Heartbeat for TLS has been removed and is disabled by default
4702 for DTLS; configure with enable-heartbeats. Code that uses the
4703 old #define's might need to be updated.
4705 *Emilia Käsper, Rich Salz*
4707 * Rename REF_CHECK to REF_DEBUG.
4711 * New "unified" build system
4713 The "unified" build system is aimed to be a common system for all
4714 platforms we support. With it comes new support for VMS.
4716 This system builds supports building in a different directory tree
4717 than the source tree. It produces one Makefile (for unix family
4718 or lookalikes), or one descrip.mms (for VMS).
4720 The source of information to make the Makefile / descrip.mms is
4721 small files called 'build.info', holding the necessary
4722 information for each directory with source to compile, and a
4723 template in Configurations, like unix-Makefile.tmpl or
4726 With this change, the library names were also renamed on Windows
4727 and on VMS. They now have names that are closer to the standard
4728 on Unix, and include the major version number, and in certain
4729 cases, the architecture they are built for. See "Notes on shared
4730 libraries" in INSTALL.
4732 We rely heavily on the perl module Text::Template.
4736 * Added support for auto-initialisation and de-initialisation of the library.
4737 OpenSSL no longer requires explicit init or deinit routines to be called,
4738 except in certain circumstances. See the OPENSSL_init_crypto() and
4739 OPENSSL_init_ssl() man pages for further information.
4743 * The arguments to the DTLSv1_listen function have changed. Specifically the
4744 "peer" argument is now expected to be a BIO_ADDR object.
4746 * Rewrite of BIO networking library. The BIO library lacked consistent
4747 support of IPv6, and adding it required some more extensive
4748 modifications. This introduces the BIO_ADDR and BIO_ADDRINFO types,
4749 which hold all types of addresses and chains of address information.
4750 It also introduces a new API, with functions like BIO_socket,
4751 BIO_connect, BIO_listen, BIO_lookup and a rewrite of BIO_accept.
4752 The source/sink BIOs BIO_s_connect, BIO_s_accept and BIO_s_datagram
4753 have been adapted accordingly.
4757 * RSA_padding_check_PKCS1_type_1 now accepts inputs with and without
4762 * CRIME protection: disable compression by default, even if OpenSSL is
4763 compiled with zlib enabled. Applications can still enable compression
4764 by calling SSL_CTX_clear_options(ctx, SSL_OP_NO_COMPRESSION), or by
4765 using the SSL_CONF library to configure compression.
4769 * The signature of the session callback configured with
4770 SSL_CTX_sess_set_get_cb was changed. The read-only input buffer
4771 was explicitly marked as `const unsigned char*` instead of
4776 * Always DPURIFY. Remove the use of uninitialized memory in the
4777 RNG, and other conditional uses of DPURIFY. This makes -DPURIFY a no-op.
4781 * Removed many obsolete configuration items, including
4782 DES_PTR, DES_RISC1, DES_RISC2, DES_INT
4783 MD2_CHAR, MD2_INT, MD2_LONG
4785 IDEA_SHORT, IDEA_LONG
4786 RC2_SHORT, RC2_LONG, RC4_LONG, RC4_CHUNK, RC4_INDEX
4788 *Rich Salz, with advice from Andy Polyakov*
4790 * Many BN internals have been moved to an internal header file.
4792 *Rich Salz with help from Andy Polyakov*
4794 * Configuration and writing out the results from it has changed.
4795 Files such as Makefile include/openssl/opensslconf.h and are now
4796 produced through general templates, such as Makefile.in and
4797 crypto/opensslconf.h.in and some help from the perl module
4800 Also, the center of configuration information is no longer
4801 Makefile. Instead, Configure produces a perl module in
4802 configdata.pm which holds most of the config data (in the hash
4803 table %config), the target data that comes from the target
4804 configuration in one of the `Configurations/*.conf` files (in
4809 * To clarify their intended purposes, the Configure options
4810 --prefix and --openssldir change their semantics, and become more
4811 straightforward and less interdependent.
4813 --prefix shall be used exclusively to give the location INSTALLTOP
4814 where programs, scripts, libraries, include files and manuals are
4815 going to be installed. The default is now /usr/local.
4817 --openssldir shall be used exclusively to give the default
4818 location OPENSSLDIR where certificates, private keys, CRLs are
4819 managed. This is also where the default openssl.cnf gets
4821 If the directory given with this option is a relative path, the
4822 values of both the --prefix value and the --openssldir value will
4823 be combined to become OPENSSLDIR.
4824 The default for --openssldir is INSTALLTOP/ssl.
4826 Anyone who uses --openssldir to specify where OpenSSL is to be
4827 installed MUST change to use --prefix instead.
4831 * The GOST engine was out of date and therefore it has been removed. An up
4832 to date GOST engine is now being maintained in an external repository.
4833 See: <https://wiki.openssl.org/index.php/Binaries>. Libssl still retains
4834 support for GOST ciphersuites (these are only activated if a GOST engine
4839 * EGD is no longer supported by default; use enable-egd when
4842 *Ben Kaduk and Rich Salz*
4844 * The distribution now has Makefile.in files, which are used to
4845 create Makefile's when Configure is run. *Configure must be run
4846 before trying to build now.*
4850 * The return value for SSL_CIPHER_description() for error conditions
4855 * Support for RFC6698/RFC7671 DANE TLSA peer authentication.
4857 Obtaining and performing DNSSEC validation of TLSA records is
4858 the application's responsibility. The application provides
4859 the TLSA records of its choice to OpenSSL, and these are then
4860 used to authenticate the peer.
4862 The TLSA records need not even come from DNS. They can, for
4863 example, be used to implement local end-entity certificate or
4864 trust-anchor "pinning", where the "pin" data takes the form
4865 of TLSA records, which can augment or replace verification
4866 based on the usual WebPKI public certification authorities.
4870 * Revert default OPENSSL_NO_DEPRECATED setting. Instead OpenSSL
4871 continues to support deprecated interfaces in default builds.
4872 However, applications are strongly advised to compile their
4873 source files with -DOPENSSL_API_COMPAT=0x10100000L, which hides
4874 the declarations of all interfaces deprecated in 0.9.8, 1.0.0
4875 or the 1.1.0 releases.
4877 In environments in which all applications have been ported to
4878 not use any deprecated interfaces OpenSSL's Configure script
4879 should be used with the --api=1.1.0 option to entirely remove
4880 support for the deprecated features from the library and
4881 unconditionally disable them in the installed headers.
4882 Essentially the same effect can be achieved with the "no-deprecated"
4883 argument to Configure, except that this will always restrict
4884 the build to just the latest API, rather than a fixed API
4887 As applications are ported to future revisions of the API,
4888 they should update their compile-time OPENSSL_API_COMPAT define
4889 accordingly, but in most cases should be able to continue to
4890 compile with later releases.
4892 The OPENSSL_API_COMPAT versions for 1.0.0, and 0.9.8 are
4893 0x10000000L and 0x00908000L, respectively. However those
4894 versions did not support the OPENSSL_API_COMPAT feature, and
4895 so applications are not typically tested for explicit support
4896 of just the undeprecated features of either release.
4900 * Add support for setting the minimum and maximum supported protocol.
4901 It can bet set via the SSL_set_min_proto_version() and
4902 SSL_set_max_proto_version(), or via the SSL_CONF's MinProtocol and
4903 MaxProtocol. It's recommended to use the new APIs to disable
4904 protocols instead of disabling individual protocols using
4905 SSL_set_options() or SSL_CONF's Protocol. This change also
4906 removes support for disabling TLS 1.2 in the OpenSSL TLS
4907 client at compile time by defining OPENSSL_NO_TLS1_2_CLIENT.
4911 * Support for ChaCha20 and Poly1305 added to libcrypto and libssl.
4915 * New EC_KEY_METHOD, this replaces the older ECDSA_METHOD and ECDH_METHOD
4916 and integrates ECDSA and ECDH functionality into EC. Implementations can
4917 now redirect key generation and no longer need to convert to or from
4920 Note: the ecdsa.h and ecdh.h headers are now no longer needed and just
4921 include the ec.h header file instead.
4925 * Remove support for all 40 and 56 bit ciphers. This includes all the export
4926 ciphers who are no longer supported and drops support the ephemeral RSA key
4927 exchange. The LOW ciphers currently doesn't have any ciphers in it.
4931 * Made EVP_MD_CTX, EVP_MD, EVP_CIPHER_CTX, EVP_CIPHER and HMAC_CTX
4932 opaque. For HMAC_CTX, the following constructors and destructors
4935 HMAC_CTX *HMAC_CTX_new(void);
4936 void HMAC_CTX_free(HMAC_CTX *ctx);
4938 For EVP_MD and EVP_CIPHER, complete APIs to create, fill and
4939 destroy such methods has been added. See EVP_MD_meth_new(3) and
4940 EVP_CIPHER_meth_new(3) for documentation.
4943 1) `EVP_MD_CTX_cleanup()`, `EVP_CIPHER_CTX_cleanup()` and
4944 `HMAC_CTX_cleanup()` were removed. `HMAC_CTX_reset()` and
4945 `EVP_MD_CTX_reset()` should be called instead to reinitialise
4946 an already created structure.
4947 2) For consistency with the majority of our object creators and
4948 destructors, `EVP_MD_CTX_(create|destroy)` were renamed to
4949 `EVP_MD_CTX_(new|free)`. The old names are retained as macros
4950 for deprecated builds.
4954 * Added ASYNC support. Libcrypto now includes the async sub-library to enable
4955 cryptographic operations to be performed asynchronously as long as an
4956 asynchronous capable engine is used. See the ASYNC_start_job() man page for
4957 further details. Libssl has also had this capability integrated with the
4958 introduction of the new mode SSL_MODE_ASYNC and associated error
4959 SSL_ERROR_WANT_ASYNC. See the SSL_CTX_set_mode() and SSL_get_error() man
4960 pages. This work was developed in partnership with Intel Corp.
4964 * SSL_{CTX_}set_ecdh_auto() has been removed and ECDH is support is
4965 always enabled now. If you want to disable the support you should
4966 exclude it using the list of supported ciphers. This also means that the
4967 "-no_ecdhe" option has been removed from s_server.
4971 * SSL_{CTX}_set_tmp_ecdh() which can set 1 EC curve now internally calls
4972 SSL_{CTX_}set1_curves() which can set a list.
4976 * Remove support for SSL_{CTX_}set_tmp_ecdh_callback(). You should set the
4977 curve you want to support using SSL_{CTX_}set1_curves().
4981 * State machine rewrite. The state machine code has been significantly
4982 refactored in order to remove much duplication of code and solve issues
4983 with the old code (see [ssl/statem/README.md](ssl/statem/README.md) for
4984 further details). This change does have some associated API changes.
4985 Notably the SSL_state() function has been removed and replaced by
4986 SSL_get_state which now returns an "OSSL_HANDSHAKE_STATE" instead of an int.
4987 SSL_set_state() has been removed altogether. The previous handshake states
4988 defined in ssl.h and ssl3.h have also been removed.
4992 * All instances of the string "ssleay" in the public API were replaced
4993 with OpenSSL (case-matching; e.g., OPENSSL_VERSION for #define's)
4994 Some error codes related to internal RSA_eay API's were renamed.
4998 * The demo files in crypto/threads were moved to demo/threads.
5002 * Removed obsolete engines: 4758cca, aep, atalla, cswift, nuron, gmp,
5005 *Matt Caswell, Rich Salz*
5007 * New ASN.1 embed macro.
5009 New ASN.1 macro ASN1_EMBED. This is the same as ASN1_SIMPLE except the
5010 structure is not allocated: it is part of the parent. That is instead of
5018 This reduces memory fragmentation and make it impossible to accidentally
5019 set a mandatory field to NULL.
5021 This currently only works for some fields specifically a SEQUENCE, CHOICE,
5022 or ASN1_STRING type which is part of a parent SEQUENCE. Since it is
5023 equivalent to ASN1_SIMPLE it cannot be tagged, OPTIONAL, SET OF or
5028 * Remove EVP_CHECK_DES_KEY, a compile-time option that never compiled.
5032 * Removed DES and RC4 ciphersuites from DEFAULT. Also removed RC2 although
5033 in 1.0.2 EXPORT was already removed and the only RC2 ciphersuite is also
5034 an EXPORT one. COMPLEMENTOFDEFAULT has been updated accordingly to add
5035 DES and RC4 ciphersuites.
5039 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
5040 This changes the decoding behaviour for some invalid messages,
5041 though the change is mostly in the more lenient direction, and
5042 legacy behaviour is preserved as much as possible.
5046 * Fix no-stdio build.
5047 *David Woodhouse <David.Woodhouse@intel.com> and also*
5048 *Ivan Nestlerode <ivan.nestlerode@sonos.com>*
5050 * New testing framework
5051 The testing framework has been largely rewritten and is now using
5052 perl and the perl modules Test::Harness and an extended variant of
5053 Test::More called OpenSSL::Test to do its work. All test scripts in
5054 test/ have been rewritten into test recipes, and all direct calls to
5055 executables in test/Makefile have become individual recipes using the
5056 simplified testing OpenSSL::Test::Simple.
5058 For documentation on our testing modules, do:
5060 perldoc test/testlib/OpenSSL/Test/Simple.pm
5061 perldoc test/testlib/OpenSSL/Test.pm
5065 * Revamped memory debug; only -DCRYPTO_MDEBUG and -DCRYPTO_MDEBUG_ABORT
5066 are used; the latter aborts on memory leaks (usually checked on exit).
5067 Some undocumented "set malloc, etc., hooks" functions were removed
5068 and others were changed. All are now documented.
5072 * In DSA_generate_parameters_ex, if the provided seed is too short,
5075 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
5077 * Rewrite PSK to support ECDHE_PSK, DHE_PSK and RSA_PSK. Add ciphersuites
5078 from RFC4279, RFC4785, RFC5487, RFC5489.
5080 Thanks to Christian J. Dietrich and Giuseppe D'Angelo for the
5081 original RSA_PSK patch.
5085 * Dropped support for the SSL3_FLAGS_DELAY_CLIENT_FINISHED flag. This SSLeay
5086 era flag was never set throughout the codebase (only read). Also removed
5087 SSL3_FLAGS_POP_BUFFER which was only used if
5088 SSL3_FLAGS_DELAY_CLIENT_FINISHED was also set.
5092 * Changed the default name options in the "ca", "crl", "req" and "x509"
5093 to be "oneline" instead of "compat".
5097 * Remove SSL_OP_TLS_BLOCK_PADDING_BUG. This is SSLeay legacy, we're
5098 not aware of clients that still exhibit this bug, and the workaround
5099 hasn't been working properly for a while.
5103 * The return type of BIO_number_read() and BIO_number_written() as well as
5104 the corresponding num_read and num_write members in the BIO structure has
5105 changed from unsigned long to uint64_t. On platforms where an unsigned
5106 long is 32 bits (e.g. Windows) these counters could overflow if >4Gb is
5111 * Given the pervasive nature of TLS extensions it is inadvisable to run
5112 OpenSSL without support for them. It also means that maintaining
5113 the OPENSSL_NO_TLSEXT option within the code is very invasive (and probably
5114 not well tested). Therefore, the OPENSSL_NO_TLSEXT option has been removed.
5118 * Removed support for the two export grade static DH ciphersuites
5119 EXP-DH-RSA-DES-CBC-SHA and EXP-DH-DSS-DES-CBC-SHA. These two ciphersuites
5120 were newly added (along with a number of other static DH ciphersuites) to
5121 1.0.2. However the two export ones have *never* worked since they were
5122 introduced. It seems strange in any case to be adding new export
5123 ciphersuites, and given "logjam" it also does not seem correct to fix them.
5127 * Version negotiation has been rewritten. In particular SSLv23_method(),
5128 SSLv23_client_method() and SSLv23_server_method() have been deprecated,
5129 and turned into macros which simply call the new preferred function names
5130 TLS_method(), TLS_client_method() and TLS_server_method(). All new code
5131 should use the new names instead. Also as part of this change the ssl23.h
5132 header file has been removed.
5136 * Support for Kerberos ciphersuites in TLS (RFC2712) has been removed. This
5137 code and the associated standard is no longer considered fit-for-purpose.
5141 * RT2547 was closed. When generating a private key, try to make the
5142 output file readable only by the owner. This behavior change might
5143 be noticeable when interacting with other software.
5145 * Documented all exdata functions. Added CRYPTO_free_ex_index.
5150 * Added HTTP GET support to the ocsp command.
5154 * Changed default digest for the dgst and enc commands from MD5 to
5159 * RAND_pseudo_bytes has been deprecated. Users should use RAND_bytes instead.
5163 * Added support for TLS extended master secret from
5164 draft-ietf-tls-session-hash-03.txt. Thanks for Alfredo Pironti for an
5165 initial patch which was a great help during development.
5169 * All libssl internal structures have been removed from the public header
5170 files, and the OPENSSL_NO_SSL_INTERN option has been removed (since it is
5171 now redundant). Users should not attempt to access internal structures
5172 directly. Instead they should use the provided API functions.
5176 * config has been changed so that by default OPENSSL_NO_DEPRECATED is used.
5177 Access to deprecated functions can be re-enabled by running config with
5178 "enable-deprecated". In addition applications wishing to use deprecated
5179 functions must define OPENSSL_USE_DEPRECATED. Note that this new behaviour
5180 will, by default, disable some transitive includes that previously existed
5181 in the header files (e.g. ec.h will no longer, by default, include bn.h)
5185 * Added support for OCB mode. OpenSSL has been granted a patent license
5186 compatible with the OpenSSL license for use of OCB. Details are available
5187 at <https://www.openssl.org/source/OCB-patent-grant-OpenSSL.pdf>. Support
5188 for OCB can be removed by calling config with no-ocb.
5192 * SSLv2 support has been removed. It still supports receiving an SSLv2
5193 compatible client hello.
5197 * Increased the minimal RSA keysize from 256 to 512 bits [Rich Salz],
5198 done while fixing the error code for the key-too-small case.
5200 *Annie Yousar <a.yousar@informatik.hu-berlin.de>*
5202 * CA.sh has been removed; use CA.pl instead.
5206 * Removed old DES API.
5210 * Remove various unsupported platforms:
5216 Sinix/ReliantUNIX RM400
5221 16-bit platforms such as WIN16
5225 * Clean up OPENSSL_NO_xxx #define's
5226 - Use setbuf() and remove OPENSSL_NO_SETVBUF_IONBF
5227 - Rename OPENSSL_SYSNAME_xxx to OPENSSL_SYS_xxx
5228 - OPENSSL_NO_EC{DH,DSA} merged into OPENSSL_NO_EC
5229 - OPENSSL_NO_RIPEMD160, OPENSSL_NO_RIPEMD merged into OPENSSL_NO_RMD160
5230 - OPENSSL_NO_FP_API merged into OPENSSL_NO_STDIO
5231 - Remove OPENSSL_NO_BIO OPENSSL_NO_BUFFER OPENSSL_NO_CHAIN_VERIFY
5232 OPENSSL_NO_EVP OPENSSL_NO_FIPS_ERR OPENSSL_NO_HASH_COMP
5233 OPENSSL_NO_LHASH OPENSSL_NO_OBJECT OPENSSL_NO_SPEED OPENSSL_NO_STACK
5234 OPENSSL_NO_X509 OPENSSL_NO_X509_VERIFY
5235 - Remove MS_STATIC; it's a relic from platforms <32 bits.
5239 * Cleaned up dead code
5240 Remove all but one '#ifdef undef' which is to be looked at.
5244 * Clean up calling of xxx_free routines.
5245 Just like free(), fix most of the xxx_free routines to accept
5246 NULL. Remove the non-null checks from callers. Save much code.
5250 * Add secure heap for storage of private keys (when possible).
5251 Add BIO_s_secmem(), CBIGNUM, etc.
5252 Contributed by Akamai Technologies under our Corporate CLA.
5256 * Experimental support for a new, fast, unbiased prime candidate generator,
5257 bn_probable_prime_dh_coprime(). Not currently used by any prime generator.
5259 *Felix Laurie von Massenbach <felix@erbridge.co.uk>*
5261 * New output format NSS in the sess_id command line tool. This allows
5262 exporting the session id and the master key in NSS keylog format.
5264 *Martin Kaiser <martin@kaiser.cx>*
5266 * Harmonize version and its documentation. -f flag is used to display
5269 *mancha <mancha1@zoho.com>*
5271 * Fix eckey_priv_encode so it immediately returns an error upon a failure
5272 in i2d_ECPrivateKey. Thanks to Ted Unangst for feedback on this issue.
5274 *mancha <mancha1@zoho.com>*
5276 * Fix some double frees. These are not thought to be exploitable.
5278 *mancha <mancha1@zoho.com>*
5280 * A missing bounds check in the handling of the TLS heartbeat extension
5281 can be used to reveal up to 64k of memory to a connected client or
5284 Thanks for Neel Mehta of Google Security for discovering this bug and to
5285 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
5286 preparing the fix ([CVE-2014-0160])
5288 *Adam Langley, Bodo Moeller*
5290 * Fix for the attack described in the paper "Recovering OpenSSL
5291 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
5292 by Yuval Yarom and Naomi Benger. Details can be obtained from:
5293 <http://eprint.iacr.org/2014/140>
5295 Thanks to Yuval Yarom and Naomi Benger for discovering this
5296 flaw and to Yuval Yarom for supplying a fix ([CVE-2014-0076])
5298 *Yuval Yarom and Naomi Benger*
5300 * Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
5301 this fixes a limitation in previous versions of OpenSSL.
5305 * Experimental encrypt-then-mac support.
5307 Experimental support for encrypt then mac from
5308 draft-gutmann-tls-encrypt-then-mac-02.txt
5310 To enable it set the appropriate extension number (0x42 for the test
5311 server) using e.g. -DTLSEXT_TYPE_encrypt_then_mac=0x42
5313 For non-compliant peers (i.e. just about everything) this should have no
5316 WARNING: EXPERIMENTAL, SUBJECT TO CHANGE.
5320 * Add EVP support for key wrapping algorithms, to avoid problems with
5321 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
5322 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
5323 algorithms and include tests cases.
5327 * Extend CMS code to support RSA-PSS signatures and RSA-OAEP for
5332 * Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
5333 MGF1 digest and OAEP label.
5337 * Make openssl verify return errors.
5339 *Chris Palmer <palmer@google.com> and Ben Laurie*
5341 * New function ASN1_TIME_diff to calculate the difference between two
5342 ASN1_TIME structures or one structure and the current time.
5346 * Update fips_test_suite to support multiple command line options. New
5347 test to induce all self test errors in sequence and check expected
5352 * Add FIPS_{rsa,dsa,ecdsa}_{sign,verify} functions which digest and
5353 sign or verify all in one operation.
5357 * Add fips_algvs: a multicall fips utility incorporating all the algorithm
5358 test programs and fips_test_suite. Includes functionality to parse
5359 the minimal script output of fipsalgest.pl directly.
5363 * Add authorisation parameter to FIPS_module_mode_set().
5367 * Add FIPS selftest for ECDH algorithm using P-224 and B-233 curves.
5371 * Use separate DRBG fields for internal and external flags. New function
5372 FIPS_drbg_health_check() to perform on demand health checking. Add
5373 generation tests to fips_test_suite with reduced health check interval to
5374 demonstrate periodic health checking. Add "nodh" option to
5375 fips_test_suite to skip very slow DH test.
5379 * New function FIPS_get_cipherbynid() to lookup FIPS supported ciphers
5384 * More extensive health check for DRBG checking many more failure modes.
5385 New function FIPS_selftest_drbg_all() to handle every possible DRBG
5386 combination: call this in fips_test_suite.
5390 * Add support for canonical generation of DSA parameter 'g'. See
5393 * Add support for HMAC DRBG from SP800-90. Update DRBG algorithm test and
5394 POST to handle HMAC cases.
5398 * Add functions FIPS_module_version() and FIPS_module_version_text()
5399 to return numerical and string versions of the FIPS module number.
5403 * Rename FIPS_mode_set and FIPS_mode to FIPS_module_mode_set and
5404 FIPS_module_mode. FIPS_mode and FIPS_mode_set will be implemented
5405 outside the validated module in the FIPS capable OpenSSL.
5409 * Minor change to DRBG entropy callback semantics. In some cases
5410 there is no multiple of the block length between min_len and
5411 max_len. Allow the callback to return more than max_len bytes
5412 of entropy but discard any extra: it is the callback's responsibility
5413 to ensure that the extra data discarded does not impact the
5414 requested amount of entropy.
5418 * Add PRNG security strength checks to RSA, DSA and ECDSA using
5419 information in FIPS186-3, SP800-57 and SP800-131A.
5423 * CCM support via EVP. Interface is very similar to GCM case except we
5424 must supply all data in one chunk (i.e. no update, final) and the
5425 message length must be supplied if AAD is used. Add algorithm test
5430 * Initial version of POST overhaul. Add POST callback to allow the status
5431 of POST to be monitored and/or failures induced. Modify fips_test_suite
5432 to use callback. Always run all selftests even if one fails.
5436 * XTS support including algorithm test driver in the fips_gcmtest program.
5437 Note: this does increase the maximum key length from 32 to 64 bytes but
5438 there should be no binary compatibility issues as existing applications
5439 will never use XTS mode.
5443 * Extensive reorganisation of FIPS PRNG behaviour. Remove all dependencies
5444 to OpenSSL RAND code and replace with a tiny FIPS RAND API which also
5445 performs algorithm blocking for unapproved PRNG types. Also do not
5446 set PRNG type in FIPS_mode_set(): leave this to the application.
5447 Add default OpenSSL DRBG handling: sets up FIPS PRNG and seeds with
5448 the standard OpenSSL PRNG: set additional data to a date time vector.
5452 * Rename old X9.31 PRNG functions of the form `FIPS_rand*` to `FIPS_x931*`.
5453 This shouldn't present any incompatibility problems because applications
5454 shouldn't be using these directly and any that are will need to rethink
5455 anyway as the X9.31 PRNG is now deprecated by FIPS 140-2
5459 * Extensive self tests and health checking required by SP800-90 DRBG.
5460 Remove strength parameter from FIPS_drbg_instantiate and always
5461 instantiate at maximum supported strength.
5465 * Add ECDH code to fips module and fips_ecdhvs for primitives only testing.
5469 * New algorithm test program fips_dhvs to handle DH primitives only testing.
5473 * New function DH_compute_key_padded() to compute a DH key and pad with
5474 leading zeroes if needed: this complies with SP800-56A et al.
5478 * Initial implementation of SP800-90 DRBGs for Hash and CTR. Not used by
5479 anything, incomplete, subject to change and largely untested at present.
5483 * Modify fipscanisteronly build option to only build the necessary object
5484 files by filtering FIPS_EX_OBJ through a perl script in crypto/Makefile.
5488 * Add experimental option FIPSSYMS to give all symbols in
5489 fipscanister.o and FIPS or fips prefix. This will avoid
5490 conflicts with future versions of OpenSSL. Add perl script
5491 util/fipsas.pl to preprocess assembly language source files
5492 and rename any affected symbols.
5496 * Add selftest checks and algorithm block of non-fips algorithms in
5497 FIPS mode. Remove DES2 from selftests.
5501 * Add ECDSA code to fips module. Add tiny fips_ecdsa_check to just
5502 return internal method without any ENGINE dependencies. Add new
5503 tiny fips sign and verify functions.
5507 * New build option no-ec2m to disable characteristic 2 code.
5511 * New build option "fipscanisteronly". This only builds fipscanister.o
5512 and (currently) associated fips utilities. Uses the file Makefile.fips
5513 instead of Makefile.org as the prototype.
5517 * Add some FIPS mode restrictions to GCM. Add internal IV generator.
5518 Update fips_gcmtest to use IV generator.
5522 * Initial, experimental EVP support for AES-GCM. AAD can be input by
5523 setting output buffer to NULL. The `*Final` function must be
5524 called although it will not retrieve any additional data. The tag
5525 can be set or retrieved with a ctrl. The IV length is by default 12
5526 bytes (96 bits) but can be set to an alternative value. If the IV
5527 length exceeds the maximum IV length (currently 16 bytes) it cannot be
5532 * New flag in ciphers: EVP_CIPH_FLAG_CUSTOM_CIPHER. This means the
5533 underlying do_cipher function handles all cipher semantics itself
5534 including padding and finalisation. This is useful if (for example)
5535 an ENGINE cipher handles block padding itself. The behaviour of
5536 do_cipher is subtly changed if this flag is set: the return value
5537 is the number of characters written to the output buffer (zero is
5538 no longer an error code) or a negative error code. Also if the
5539 input buffer is NULL and length 0 finalisation should be performed.
5543 * If a candidate issuer certificate is already part of the constructed
5544 path ignore it: new debug notification X509_V_ERR_PATH_LOOP for this case.
5548 * Improve forward-security support: add functions
5550 void SSL_CTX_set_not_resumable_session_callback(
5551 SSL_CTX *ctx, int (*cb)(SSL *ssl, int is_forward_secure))
5552 void SSL_set_not_resumable_session_callback(
5553 SSL *ssl, int (*cb)(SSL *ssl, int is_forward_secure))
5555 for use by SSL/TLS servers; the callback function will be called whenever a
5556 new session is created, and gets to decide whether the session may be
5557 cached to make it resumable (return 0) or not (return 1). (As by the
5558 SSL/TLS protocol specifications, the session_id sent by the server will be
5559 empty to indicate that the session is not resumable; also, the server will
5560 not generate RFC 4507 (RFC 5077) session tickets.)
5562 A simple reasonable callback implementation is to return is_forward_secure.
5563 This parameter will be set to 1 or 0 depending on the ciphersuite selected
5564 by the SSL/TLS server library, indicating whether it can provide forward
5567 *Emilia Käsper <emilia.kasper@esat.kuleuven.be> (Google)*
5569 * New -verify_name option in command line utilities to set verification
5574 * Initial CMAC implementation. WARNING: EXPERIMENTAL, API MAY CHANGE.
5575 Add CMAC pkey methods.
5579 * Experimental renegotiation in s_server -www mode. If the client
5580 browses /reneg connection is renegotiated. If /renegcert it is
5581 renegotiated requesting a certificate.
5585 * Add an "external" session cache for debugging purposes to s_server. This
5586 should help trace issues which normally are only apparent in deployed
5587 multi-process servers.
5591 * Extensive audit of libcrypto with DEBUG_UNUSED. Fix many cases where
5592 return value is ignored. NB. The functions RAND_add(), RAND_seed(),
5593 BIO_set_cipher() and some obscure PEM functions were changed so they
5594 can now return an error. The RAND changes required a change to the
5595 RAND_METHOD structure.
5599 * New macro `__owur` for "OpenSSL Warn Unused Result". This makes use of
5600 a gcc attribute to warn if the result of a function is ignored. This
5601 is enable if DEBUG_UNUSED is set. Add to several functions in evp.h
5602 whose return value is often ignored.
5606 * New -noct, -requestct, -requirect and -ctlogfile options for s_client.
5607 These allow SCTs (signed certificate timestamps) to be requested and
5608 validated when establishing a connection.
5610 *Rob Percival <robpercival@google.com>*
5615 ### Changes between 1.0.2s and 1.0.2t [10 Sep 2019]
5617 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
5618 used even when parsing explicit parameters, when loading a encoded key
5619 or calling `EC_GROUP_new_from_ecpkparameters()`/
5620 `EC_GROUP_new_from_ecparameters()`.
5621 This prevents bypass of security hardening and performance gains,
5622 especially for curves with specialized EC_METHODs.
5623 By default, if a key encoded with explicit parameters is loaded and later
5624 encoded, the output is still encoded with explicit parameters, even if
5625 internally a "named" EC_GROUP is used for computation.
5629 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
5630 this change, EC_GROUP_set_generator would accept order and/or cofactor as
5631 NULL. After this change, only the cofactor parameter can be NULL. It also
5632 does some minimal sanity checks on the passed order.
5637 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
5638 An attack is simple, if the first CMS_recipientInfo is valid but the
5639 second CMS_recipientInfo is chosen ciphertext. If the second
5640 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
5641 encryption key will be replaced by garbage, and the message cannot be
5642 decoded, but if the RSA decryption fails, the correct encryption key is
5643 used and the recipient will not notice the attack.
5644 As a work around for this potential attack the length of the decrypted
5645 key must be equal to the cipher default key length, in case the
5646 certificate is not given and all recipientInfo are tried out.
5647 The old behaviour can be re-enabled in the CMS code by setting the
5648 CMS_DEBUG_DECRYPT flag.
5653 * Document issue with installation paths in diverse Windows builds
5655 '/usr/local/ssl' is an unsafe prefix for location to install OpenSSL
5656 binaries and run-time config file.
5661 ### Changes between 1.0.2r and 1.0.2s [28 May 2019]
5663 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
5664 This changes the size when using the `genpkey` command when no size is given.
5665 It fixes an omission in earlier changes that changed all RSA, DSA and DH
5666 generation commands to use 2048 bits by default.
5670 * Add FIPS support for Android Arm 64-bit
5672 Support for Android Arm 64-bit was added to the OpenSSL FIPS Object
5673 Module in Version 2.0.10. For some reason, the corresponding target
5674 'android64-aarch64' was missing OpenSSL 1.0.2, whence it could not be
5675 built with FIPS support on Android Arm 64-bit. This omission has been
5678 *Matthias St. Pierre*
5680 ### Changes between 1.0.2q and 1.0.2r [26 Feb 2019]
5682 * 0-byte record padding oracle
5684 If an application encounters a fatal protocol error and then calls
5685 SSL_shutdown() twice (once to send a close_notify, and once to receive one)
5686 then OpenSSL can respond differently to the calling application if a 0 byte
5687 record is received with invalid padding compared to if a 0 byte record is
5688 received with an invalid MAC. If the application then behaves differently
5689 based on that in a way that is detectable to the remote peer, then this
5690 amounts to a padding oracle that could be used to decrypt data.
5692 In order for this to be exploitable "non-stitched" ciphersuites must be in
5693 use. Stitched ciphersuites are optimised implementations of certain
5694 commonly used ciphersuites. Also the application must call SSL_shutdown()
5695 twice even if a protocol error has occurred (applications should not do
5696 this but some do anyway).
5698 This issue was discovered by Juraj Somorovsky, Robert Merget and Nimrod
5699 Aviram, with additional investigation by Steven Collison and Andrew
5700 Hourselt. It was reported to OpenSSL on 10th December 2018.
5705 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
5709 ### Changes between 1.0.2p and 1.0.2q [20 Nov 2018]
5711 * Microarchitecture timing vulnerability in ECC scalar multiplication
5713 OpenSSL ECC scalar multiplication, used in e.g. ECDSA and ECDH, has been
5714 shown to be vulnerable to a microarchitecture timing side channel attack.
5715 An attacker with sufficient access to mount local timing attacks during
5716 ECDSA signature generation could recover the private key.
5718 This issue was reported to OpenSSL on 26th October 2018 by Alejandro
5719 Cabrera Aldaya, Billy Brumley, Sohaib ul Hassan, Cesar Pereida Garcia and
5725 * Timing vulnerability in DSA signature generation
5727 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
5728 timing side channel attack. An attacker could use variations in the signing
5729 algorithm to recover the private key.
5731 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
5736 * Resolve a compatibility issue in EC_GROUP handling with the FIPS Object
5737 Module, accidentally introduced while backporting security fixes from the
5738 development branch and hindering the use of ECC in FIPS mode.
5742 ### Changes between 1.0.2o and 1.0.2p [14 Aug 2018]
5744 * Client DoS due to large DH parameter
5746 During key agreement in a TLS handshake using a DH(E) based ciphersuite a
5747 malicious server can send a very large prime value to the client. This will
5748 cause the client to spend an unreasonably long period of time generating a
5749 key for this prime resulting in a hang until the client has finished. This
5750 could be exploited in a Denial Of Service attack.
5752 This issue was reported to OpenSSL on 5th June 2018 by Guido Vranken
5757 * Cache timing vulnerability in RSA Key Generation
5759 The OpenSSL RSA Key generation algorithm has been shown to be vulnerable to
5760 a cache timing side channel attack. An attacker with sufficient access to
5761 mount cache timing attacks during the RSA key generation process could
5762 recover the private key.
5764 This issue was reported to OpenSSL on 4th April 2018 by Alejandro Cabrera
5765 Aldaya, Billy Brumley, Cesar Pereida Garcia and Luis Manuel Alvarez Tapia.
5770 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
5771 parameter is no longer accepted, as it leads to a corrupt table. NULL
5772 pem_str is reserved for alias entries only.
5776 * Revert blinding in ECDSA sign and instead make problematic addition
5777 length-invariant. Switch even to fixed-length Montgomery multiplication.
5781 * Change generating and checking of primes so that the error rate of not
5782 being prime depends on the intended use based on the size of the input.
5783 For larger primes this will result in more rounds of Miller-Rabin.
5784 The maximal error rate for primes with more than 1080 bits is lowered
5787 *Kurt Roeckx, Annie Yousar*
5789 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
5793 * Add blinding to ECDSA and DSA signatures to protect against side channel
5794 attacks discovered by Keegan Ryan (NCC Group).
5798 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
5799 now allow empty (zero character) pass phrases.
5803 * Certificate time validation (X509_cmp_time) enforces stricter
5804 compliance with RFC 5280. Fractional seconds and timezone offsets
5805 are no longer allowed.
5809 ### Changes between 1.0.2n and 1.0.2o [27 Mar 2018]
5811 * Constructed ASN.1 types with a recursive definition could exceed the stack
5813 Constructed ASN.1 types with a recursive definition (such as can be found
5814 in PKCS7) could eventually exceed the stack given malicious input with
5815 excessive recursion. This could result in a Denial Of Service attack. There
5816 are no such structures used within SSL/TLS that come from untrusted sources
5817 so this is considered safe.
5819 This issue was reported to OpenSSL on 4th January 2018 by the OSS-fuzz
5825 ### Changes between 1.0.2m and 1.0.2n [7 Dec 2017]
5827 * Read/write after SSL object in error state
5829 OpenSSL 1.0.2 (starting from version 1.0.2b) introduced an "error state"
5830 mechanism. The intent was that if a fatal error occurred during a handshake
5831 then OpenSSL would move into the error state and would immediately fail if
5832 you attempted to continue the handshake. This works as designed for the
5833 explicit handshake functions (SSL_do_handshake(), SSL_accept() and
5834 SSL_connect()), however due to a bug it does not work correctly if
5835 SSL_read() or SSL_write() is called directly. In that scenario, if the
5836 handshake fails then a fatal error will be returned in the initial function
5837 call. If SSL_read()/SSL_write() is subsequently called by the application
5838 for the same SSL object then it will succeed and the data is passed without
5839 being decrypted/encrypted directly from the SSL/TLS record layer.
5841 In order to exploit this issue an application bug would have to be present
5842 that resulted in a call to SSL_read()/SSL_write() being issued after having
5843 already received a fatal error.
5845 This issue was reported to OpenSSL by David Benjamin (Google).
5850 * rsaz_1024_mul_avx2 overflow bug on x86_64
5852 There is an overflow bug in the AVX2 Montgomery multiplication procedure
5853 used in exponentiation with 1024-bit moduli. No EC algorithms are affected.
5854 Analysis suggests that attacks against RSA and DSA as a result of this
5855 defect would be very difficult to perform and are not believed likely.
5856 Attacks against DH1024 are considered just feasible, because most of the
5857 work necessary to deduce information about a private key may be performed
5858 offline. The amount of resources required for such an attack would be
5859 significant. However, for an attack on TLS to be meaningful, the server
5860 would have to share the DH1024 private key among multiple clients, which is
5861 no longer an option since CVE-2016-0701.
5863 This only affects processors that support the AVX2 but not ADX extensions
5864 like Intel Haswell (4th generation).
5866 This issue was reported to OpenSSL by David Benjamin (Google). The issue
5867 was originally found via the OSS-Fuzz project.
5872 ### Changes between 1.0.2l and 1.0.2m [2 Nov 2017]
5874 * bn_sqrx8x_internal carry bug on x86_64
5876 There is a carry propagating bug in the x86_64 Montgomery squaring
5877 procedure. No EC algorithms are affected. Analysis suggests that attacks
5878 against RSA and DSA as a result of this defect would be very difficult to
5879 perform and are not believed likely. Attacks against DH are considered just
5880 feasible (although very difficult) because most of the work necessary to
5881 deduce information about a private key may be performed offline. The amount
5882 of resources required for such an attack would be very significant and
5883 likely only accessible to a limited number of attackers. An attacker would
5884 additionally need online access to an unpatched system using the target
5885 private key in a scenario with persistent DH parameters and a private
5886 key that is shared between multiple clients.
5888 This only affects processors that support the BMI1, BMI2 and ADX extensions
5889 like Intel Broadwell (5th generation) and later or AMD Ryzen.
5891 This issue was reported to OpenSSL by the OSS-Fuzz project.
5896 * Malformed X.509 IPAddressFamily could cause OOB read
5898 If an X.509 certificate has a malformed IPAddressFamily extension,
5899 OpenSSL could do a one-byte buffer overread. The most likely result
5900 would be an erroneous display of the certificate in text format.
5902 This issue was reported to OpenSSL by the OSS-Fuzz project.
5906 ### Changes between 1.0.2k and 1.0.2l [25 May 2017]
5908 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
5909 platform rather than 'mingw'.
5913 ### Changes between 1.0.2j and 1.0.2k [26 Jan 2017]
5915 * Truncated packet could crash via OOB read
5917 If one side of an SSL/TLS path is running on a 32-bit host and a specific
5918 cipher is being used, then a truncated packet can cause that host to
5919 perform an out-of-bounds read, usually resulting in a crash.
5921 This issue was reported to OpenSSL by Robert Święcki of Google.
5926 * BN_mod_exp may produce incorrect results on x86_64
5928 There is a carry propagating bug in the x86_64 Montgomery squaring
5929 procedure. No EC algorithms are affected. Analysis suggests that attacks
5930 against RSA and DSA as a result of this defect would be very difficult to
5931 perform and are not believed likely. Attacks against DH are considered just
5932 feasible (although very difficult) because most of the work necessary to
5933 deduce information about a private key may be performed offline. The amount
5934 of resources required for such an attack would be very significant and
5935 likely only accessible to a limited number of attackers. An attacker would
5936 additionally need online access to an unpatched system using the target
5937 private key in a scenario with persistent DH parameters and a private
5938 key that is shared between multiple clients. For example this can occur by
5939 default in OpenSSL DHE based SSL/TLS ciphersuites. Note: This issue is very
5940 similar to CVE-2015-3193 but must be treated as a separate problem.
5942 This issue was reported to OpenSSL by the OSS-Fuzz project.
5947 * Montgomery multiplication may produce incorrect results
5949 There is a carry propagating bug in the Broadwell-specific Montgomery
5950 multiplication procedure that handles input lengths divisible by, but
5951 longer than 256 bits. Analysis suggests that attacks against RSA, DSA
5952 and DH private keys are impossible. This is because the subroutine in
5953 question is not used in operations with the private key itself and an input
5954 of the attacker's direct choice. Otherwise the bug can manifest itself as
5955 transient authentication and key negotiation failures or reproducible
5956 erroneous outcome of public-key operations with specially crafted input.
5957 Among EC algorithms only Brainpool P-512 curves are affected and one
5958 presumably can attack ECDH key negotiation. Impact was not analyzed in
5959 detail, because pre-requisites for attack are considered unlikely. Namely
5960 multiple clients have to choose the curve in question and the server has to
5961 share the private key among them, neither of which is default behaviour.
5962 Even then only clients that chose the curve will be affected.
5964 This issue was publicly reported as transient failures and was not
5965 initially recognized as a security issue. Thanks to Richard Morgan for
5966 providing reproducible case.
5971 * OpenSSL now fails if it receives an unrecognised record type in TLS1.0
5972 or TLS1.1. Previously this only happened in SSLv3 and TLS1.2. This is to
5973 prevent issues where no progress is being made and the peer continually
5974 sends unrecognised record types, using up resources processing them.
5978 ### Changes between 1.0.2i and 1.0.2j [26 Sep 2016]
5980 * Missing CRL sanity check
5982 A bug fix which included a CRL sanity check was added to OpenSSL 1.1.0
5983 but was omitted from OpenSSL 1.0.2i. As a result any attempt to use
5984 CRLs in OpenSSL 1.0.2i will crash with a null pointer exception.
5986 This issue only affects the OpenSSL 1.0.2i
5991 ### Changes between 1.0.2h and 1.0.2i [22 Sep 2016]
5993 * OCSP Status Request extension unbounded memory growth
5995 A malicious client can send an excessively large OCSP Status Request
5996 extension. If that client continually requests renegotiation, sending a
5997 large OCSP Status Request extension each time, then there will be unbounded
5998 memory growth on the server. This will eventually lead to a Denial Of
5999 Service attack through memory exhaustion. Servers with a default
6000 configuration are vulnerable even if they do not support OCSP. Builds using
6001 the "no-ocsp" build time option are not affected.
6003 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6008 * In order to mitigate the SWEET32 attack, the DES ciphers were moved from
6011 This issue was reported to OpenSSL Karthikeyan Bhargavan and Gaetan
6017 * OOB write in MDC2_Update()
6019 An overflow can occur in MDC2_Update() either if called directly or
6020 through the EVP_DigestUpdate() function using MDC2. If an attacker
6021 is able to supply very large amounts of input data after a previous
6022 call to EVP_EncryptUpdate() with a partial block then a length check
6023 can overflow resulting in a heap corruption.
6025 The amount of data needed is comparable to SIZE_MAX which is impractical
6028 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6033 * Malformed SHA512 ticket DoS
6035 If a server uses SHA512 for TLS session ticket HMAC it is vulnerable to a
6036 DoS attack where a malformed ticket will result in an OOB read which will
6039 The use of SHA512 in TLS session tickets is comparatively rare as it requires
6040 a custom server callback and ticket lookup mechanism.
6042 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6047 * OOB write in BN_bn2dec()
6049 The function BN_bn2dec() does not check the return value of BN_div_word().
6050 This can cause an OOB write if an application uses this function with an
6051 overly large BIGNUM. This could be a problem if an overly large certificate
6052 or CRL is printed out from an untrusted source. TLS is not affected because
6053 record limits will reject an oversized certificate before it is parsed.
6055 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6060 * OOB read in TS_OBJ_print_bio()
6062 The function TS_OBJ_print_bio() misuses OBJ_obj2txt(): the return value is
6063 the total length the OID text representation would use and not the amount
6064 of data written. This will result in OOB reads when large OIDs are
6067 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6072 * Pointer arithmetic undefined behaviour
6074 Avoid some undefined pointer arithmetic
6076 A common idiom in the codebase is to check limits in the following manner:
6079 Where "p" points to some malloc'd data of SIZE bytes and
6082 "len" here could be from some externally supplied data (e.g. from a TLS
6085 The rules of C pointer arithmetic are such that "p + len" is only well
6086 defined where len <= SIZE. Therefore the above idiom is actually
6087 undefined behaviour.
6089 For example this could cause problems if some malloc implementation
6090 provides an address for "p" such that "p + len" actually overflows for
6091 values of len that are too big and therefore p + len < limit.
6093 This issue was reported to OpenSSL by Guido Vranken
6098 * Constant time flag not preserved in DSA signing
6100 Operations in the DSA signing algorithm should run in constant time in
6101 order to avoid side channel attacks. A flaw in the OpenSSL DSA
6102 implementation means that a non-constant time codepath is followed for
6103 certain operations. This has been demonstrated through a cache-timing
6104 attack to be sufficient for an attacker to recover the private DSA key.
6106 This issue was reported by César Pereida (Aalto University), Billy Brumley
6107 (Tampere University of Technology), and Yuval Yarom (The University of
6108 Adelaide and NICTA).
6113 * DTLS buffered message DoS
6115 In a DTLS connection where handshake messages are delivered out-of-order
6116 those messages that OpenSSL is not yet ready to process will be buffered
6117 for later use. Under certain circumstances, a flaw in the logic means that
6118 those messages do not get removed from the buffer even though the handshake
6119 has been completed. An attacker could force up to approx. 15 messages to
6120 remain in the buffer when they are no longer required. These messages will
6121 be cleared when the DTLS connection is closed. The default maximum size for
6122 a message is 100k. Therefore, the attacker could force an additional 1500k
6123 to be consumed per connection. By opening many simultaneous connections an
6124 attacker could cause a DoS attack through memory exhaustion.
6126 This issue was reported to OpenSSL by Quan Luo.
6131 * DTLS replay protection DoS
6133 A flaw in the DTLS replay attack protection mechanism means that records
6134 that arrive for future epochs update the replay protection "window" before
6135 the MAC for the record has been validated. This could be exploited by an
6136 attacker by sending a record for the next epoch (which does not have to
6137 decrypt or have a valid MAC), with a very large sequence number. This means
6138 that all subsequent legitimate packets are dropped causing a denial of
6139 service for a specific DTLS connection.
6141 This issue was reported to OpenSSL by the OCAP audit team.
6146 * Certificate message OOB reads
6148 In OpenSSL 1.0.2 and earlier some missing message length checks can result
6149 in OOB reads of up to 2 bytes beyond an allocated buffer. There is a
6150 theoretical DoS risk but this has not been observed in practice on common
6153 The messages affected are client certificate, client certificate request
6154 and server certificate. As a result the attack can only be performed
6155 against a client or a server which enables client authentication.
6157 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6162 ### Changes between 1.0.2g and 1.0.2h [3 May 2016]
6164 * Prevent padding oracle in AES-NI CBC MAC check
6166 A MITM attacker can use a padding oracle attack to decrypt traffic
6167 when the connection uses an AES CBC cipher and the server support
6170 This issue was introduced as part of the fix for Lucky 13 padding
6171 attack ([CVE-2013-0169]). The padding check was rewritten to be in
6172 constant time by making sure that always the same bytes are read and
6173 compared against either the MAC or padding bytes. But it no longer
6174 checked that there was enough data to have both the MAC and padding
6177 This issue was reported by Juraj Somorovsky using TLS-Attacker.
6181 * Fix EVP_EncodeUpdate overflow
6183 An overflow can occur in the EVP_EncodeUpdate() function which is used for
6184 Base64 encoding of binary data. If an attacker is able to supply very large
6185 amounts of input data then a length check can overflow resulting in a heap
6188 Internally to OpenSSL the EVP_EncodeUpdate() function is primarily used by
6189 the `PEM_write_bio*` family of functions. These are mainly used within the
6190 OpenSSL command line applications, so any application which processes data
6191 from an untrusted source and outputs it as a PEM file should be considered
6192 vulnerable to this issue. User applications that call these APIs directly
6193 with large amounts of untrusted data may also be vulnerable.
6195 This issue was reported by Guido Vranken.
6200 * Fix EVP_EncryptUpdate overflow
6202 An overflow can occur in the EVP_EncryptUpdate() function. If an attacker
6203 is able to supply very large amounts of input data after a previous call to
6204 EVP_EncryptUpdate() with a partial block then a length check can overflow
6205 resulting in a heap corruption. Following an analysis of all OpenSSL
6206 internal usage of the EVP_EncryptUpdate() function all usage is one of two
6207 forms. The first form is where the EVP_EncryptUpdate() call is known to be
6208 the first called function after an EVP_EncryptInit(), and therefore that
6209 specific call must be safe. The second form is where the length passed to
6210 EVP_EncryptUpdate() can be seen from the code to be some small value and
6211 therefore there is no possibility of an overflow. Since all instances are
6212 one of these two forms, it is believed that there can be no overflows in
6213 internal code due to this problem. It should be noted that
6214 EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths.
6215 Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances
6216 of these calls have also been analysed too and it is believed there are no
6217 instances in internal usage where an overflow could occur.
6219 This issue was reported by Guido Vranken.
6224 * Prevent ASN.1 BIO excessive memory allocation
6226 When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio()
6227 a short invalid encoding can cause allocation of large amounts of memory
6228 potentially consuming excessive resources or exhausting memory.
6230 Any application parsing untrusted data through d2i BIO functions is
6231 affected. The memory based functions such as d2i_X509() are *not* affected.
6232 Since the memory based functions are used by the TLS library, TLS
6233 applications are not affected.
6235 This issue was reported by Brian Carpenter.
6242 ASN1 Strings that are over 1024 bytes can cause an overread in applications
6243 using the X509_NAME_oneline() function on EBCDIC systems. This could result
6244 in arbitrary stack data being returned in the buffer.
6246 This issue was reported by Guido Vranken.
6251 * Modify behavior of ALPN to invoke callback after SNI/servername
6252 callback, such that updates to the SSL_CTX affect ALPN.
6256 * Remove LOW from the DEFAULT cipher list. This removes singles DES from the
6261 * Only remove the SSLv2 methods with the no-ssl2-method option. When the
6262 methods are enabled and ssl2 is disabled the methods return NULL.
6266 ### Changes between 1.0.2f and 1.0.2g [1 Mar 2016]
6268 * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL.
6269 Builds that are not configured with "enable-weak-ssl-ciphers" will not
6270 provide any "EXPORT" or "LOW" strength ciphers.
6274 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2
6275 is by default disabled at build-time. Builds that are not configured with
6276 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used,
6277 users who want to negotiate SSLv2 via the version-flexible SSLv23_method()
6278 will need to explicitly call either of:
6280 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2);
6282 SSL_clear_options(ssl, SSL_OP_NO_SSLv2);
6284 as appropriate. Even if either of those is used, or the application
6285 explicitly uses the version-specific SSLv2_method() or its client and
6286 server variants, SSLv2 ciphers vulnerable to exhaustive search key
6287 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT
6288 ciphers, and SSLv2 56-bit DES are no longer available.
6293 * Fix a double-free in DSA code
6295 A double free bug was discovered when OpenSSL parses malformed DSA private
6296 keys and could lead to a DoS attack or memory corruption for applications
6297 that receive DSA private keys from untrusted sources. This scenario is
6300 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using
6306 * Disable SRP fake user seed to address a server memory leak.
6308 Add a new method SRP_VBASE_get1_by_user that handles the seed properly.
6310 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
6311 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user
6312 was changed to ignore the "fake user" SRP seed, even if the seed
6315 Users should use SRP_VBASE_get1_by_user instead. Note that in
6316 SRP_VBASE_get1_by_user, caller must free the returned value. Note
6317 also that even though configuring the SRP seed attempts to hide
6318 invalid usernames by continuing the handshake with fake
6319 credentials, this behaviour is not constant time and no strong
6320 guarantees are made that the handshake is indistinguishable from
6321 that of a valid user.
6326 * Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption
6328 In the BN_hex2bn function the number of hex digits is calculated using an
6329 int value `i`. Later `bn_expand` is called with a value of `i * 4`. For
6330 large values of `i` this can result in `bn_expand` not allocating any
6331 memory because `i * 4` is negative. This can leave the internal BIGNUM data
6332 field as NULL leading to a subsequent NULL ptr deref. For very large values
6333 of `i`, the calculation `i * 4` could be a positive value smaller than `i`.
6334 In this case memory is allocated to the internal BIGNUM data field, but it
6335 is insufficiently sized leading to heap corruption. A similar issue exists
6336 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn
6337 is ever called by user applications with very large untrusted hex/dec data.
6338 This is anticipated to be a rare occurrence.
6340 All OpenSSL internal usage of these functions use data that is not expected
6341 to be untrusted, e.g. config file data or application command line
6342 arguments. If user developed applications generate config file data based
6343 on untrusted data then it is possible that this could also lead to security
6344 consequences. This is also anticipated to be rare.
6346 This issue was reported to OpenSSL by Guido Vranken.
6351 * Fix memory issues in `BIO_*printf` functions
6353 The internal `fmtstr` function used in processing a "%s" format string in
6354 the `BIO_*printf` functions could overflow while calculating the length of a
6355 string and cause an OOB read when printing very long strings.
6357 Additionally the internal `doapr_outch` function can attempt to write to an
6358 OOB memory location (at an offset from the NULL pointer) in the event of a
6359 memory allocation failure. In 1.0.2 and below this could be caused where
6360 the size of a buffer to be allocated is greater than INT_MAX. E.g. this
6361 could be in processing a very long "%s" format string. Memory leaks can
6364 The first issue may mask the second issue dependent on compiler behaviour.
6365 These problems could enable attacks where large amounts of untrusted data
6366 is passed to the `BIO_*printf` functions. If applications use these functions
6367 in this way then they could be vulnerable. OpenSSL itself uses these
6368 functions when printing out human-readable dumps of ASN.1 data. Therefore
6369 applications that print this data could be vulnerable if the data is from
6370 untrusted sources. OpenSSL command line applications could also be
6371 vulnerable where they print out ASN.1 data, or if untrusted data is passed
6372 as command line arguments.
6374 Libssl is not considered directly vulnerable. Additionally certificates etc
6375 received via remote connections via libssl are also unlikely to be able to
6376 trigger these issues because of message size limits enforced within libssl.
6378 This issue was reported to OpenSSL Guido Vranken.
6383 * Side channel attack on modular exponentiation
6385 A side-channel attack was found which makes use of cache-bank conflicts on
6386 the Intel Sandy-Bridge microarchitecture which could lead to the recovery
6387 of RSA keys. The ability to exploit this issue is limited as it relies on
6388 an attacker who has control of code in a thread running on the same
6389 hyper-threaded core as the victim thread which is performing decryptions.
6391 This issue was reported to OpenSSL by Yuval Yarom, The University of
6392 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and
6393 Nadia Heninger, University of Pennsylvania with more information at
6394 <http://cachebleed.info>.
6399 * Change the `req` command to generate a 2048-bit RSA/DSA key by default,
6400 if no keysize is specified with default_bits. This fixes an
6401 omission in an earlier change that changed all RSA/DSA key generation
6402 commands to use 2048 bits by default.
6406 ### Changes between 1.0.2e and 1.0.2f [28 Jan 2016]
6408 * DH small subgroups
6410 Historically OpenSSL only ever generated DH parameters based on "safe"
6411 primes. More recently (in version 1.0.2) support was provided for
6412 generating X9.42 style parameter files such as those required for RFC 5114
6413 support. The primes used in such files may not be "safe". Where an
6414 application is using DH configured with parameters based on primes that are
6415 not "safe" then an attacker could use this fact to find a peer's private
6416 DH exponent. This attack requires that the attacker complete multiple
6417 handshakes in which the peer uses the same private DH exponent. For example
6418 this could be used to discover a TLS server's private DH exponent if it's
6419 reusing the private DH exponent or it's using a static DH ciphersuite.
6421 OpenSSL provides the option SSL_OP_SINGLE_DH_USE for ephemeral DH (DHE) in
6422 TLS. It is not on by default. If the option is not set then the server
6423 reuses the same private DH exponent for the life of the server process and
6424 would be vulnerable to this attack. It is believed that many popular
6425 applications do set this option and would therefore not be at risk.
6427 The fix for this issue adds an additional check where a "q" parameter is
6428 available (as is the case in X9.42 based parameters). This detects the
6429 only known attack, and is the only possible defense for static DH
6430 ciphersuites. This could have some performance impact.
6432 Additionally the SSL_OP_SINGLE_DH_USE option has been switched on by
6433 default and cannot be disabled. This could have some performance impact.
6435 This issue was reported to OpenSSL by Antonio Sanso (Adobe).
6440 * SSLv2 doesn't block disabled ciphers
6442 A malicious client can negotiate SSLv2 ciphers that have been disabled on
6443 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
6444 been disabled, provided that the SSLv2 protocol was not also disabled via
6447 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
6448 and Sebastian Schinzel.
6453 ### Changes between 1.0.2d and 1.0.2e [3 Dec 2015]
6455 * BN_mod_exp may produce incorrect results on x86_64
6457 There is a carry propagating bug in the x86_64 Montgomery squaring
6458 procedure. No EC algorithms are affected. Analysis suggests that attacks
6459 against RSA and DSA as a result of this defect would be very difficult to
6460 perform and are not believed likely. Attacks against DH are considered just
6461 feasible (although very difficult) because most of the work necessary to
6462 deduce information about a private key may be performed offline. The amount
6463 of resources required for such an attack would be very significant and
6464 likely only accessible to a limited number of attackers. An attacker would
6465 additionally need online access to an unpatched system using the target
6466 private key in a scenario with persistent DH parameters and a private
6467 key that is shared between multiple clients. For example this can occur by
6468 default in OpenSSL DHE based SSL/TLS ciphersuites.
6470 This issue was reported to OpenSSL by Hanno Böck.
6475 * Certificate verify crash with missing PSS parameter
6477 The signature verification routines will crash with a NULL pointer
6478 dereference if presented with an ASN.1 signature using the RSA PSS
6479 algorithm and absent mask generation function parameter. Since these
6480 routines are used to verify certificate signature algorithms this can be
6481 used to crash any certificate verification operation and exploited in a
6482 DoS attack. Any application which performs certificate verification is
6483 vulnerable including OpenSSL clients and servers which enable client
6486 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
6491 * X509_ATTRIBUTE memory leak
6493 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
6494 memory. This structure is used by the PKCS#7 and CMS routines so any
6495 application which reads PKCS#7 or CMS data from untrusted sources is
6496 affected. SSL/TLS is not affected.
6498 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
6504 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
6505 This changes the decoding behaviour for some invalid messages,
6506 though the change is mostly in the more lenient direction, and
6507 legacy behaviour is preserved as much as possible.
6511 * In DSA_generate_parameters_ex, if the provided seed is too short,
6514 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
6516 ### Changes between 1.0.2c and 1.0.2d [9 Jul 2015]
6518 * Alternate chains certificate forgery
6520 During certificate verification, OpenSSL will attempt to find an
6521 alternative certificate chain if the first attempt to build such a chain
6522 fails. An error in the implementation of this logic can mean that an
6523 attacker could cause certain checks on untrusted certificates to be
6524 bypassed, such as the CA flag, enabling them to use a valid leaf
6525 certificate to act as a CA and "issue" an invalid certificate.
6527 This issue was reported to OpenSSL by Adam Langley/David Benjamin
6532 ### Changes between 1.0.2b and 1.0.2c [12 Jun 2015]
6534 * Fix HMAC ABI incompatibility. The previous version introduced an ABI
6535 incompatibility in the handling of HMAC. The previous ABI has now been
6540 ### Changes between 1.0.2a and 1.0.2b [11 Jun 2015]
6542 * Malformed ECParameters causes infinite loop
6544 When processing an ECParameters structure OpenSSL enters an infinite loop
6545 if the curve specified is over a specially malformed binary polynomial
6548 This can be used to perform denial of service against any
6549 system which processes public keys, certificate requests or
6550 certificates. This includes TLS clients and TLS servers with
6551 client authentication enabled.
6553 This issue was reported to OpenSSL by Joseph Barr-Pixton.
6558 * Exploitable out-of-bounds read in X509_cmp_time
6560 X509_cmp_time does not properly check the length of the ASN1_TIME
6561 string and can read a few bytes out of bounds. In addition,
6562 X509_cmp_time accepts an arbitrary number of fractional seconds in the
6565 An attacker can use this to craft malformed certificates and CRLs of
6566 various sizes and potentially cause a segmentation fault, resulting in
6567 a DoS on applications that verify certificates or CRLs. TLS clients
6568 that verify CRLs are affected. TLS clients and servers with client
6569 authentication enabled may be affected if they use custom verification
6572 This issue was reported to OpenSSL by Robert Swiecki (Google), and
6573 independently by Hanno Böck.
6578 * PKCS7 crash with missing EnvelopedContent
6580 The PKCS#7 parsing code does not handle missing inner EncryptedContent
6581 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
6582 with missing content and trigger a NULL pointer dereference on parsing.
6584 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
6585 structures from untrusted sources are affected. OpenSSL clients and
6586 servers are not affected.
6588 This issue was reported to OpenSSL by Michal Zalewski (Google).
6593 * CMS verify infinite loop with unknown hash function
6595 When verifying a signedData message the CMS code can enter an infinite loop
6596 if presented with an unknown hash function OID. This can be used to perform
6597 denial of service against any system which verifies signedData messages using
6599 This issue was reported to OpenSSL by Johannes Bauer.
6604 * Race condition handling NewSessionTicket
6606 If a NewSessionTicket is received by a multi-threaded client when attempting to
6607 reuse a previous ticket then a race condition can occur potentially leading to
6608 a double free of the ticket data.
6613 * Only support 256-bit or stronger elliptic curves with the
6614 'ecdh_auto' setting (server) or by default (client). Of supported
6615 curves, prefer P-256 (both).
6619 ### Changes between 1.0.2 and 1.0.2a [19 Mar 2015]
6621 * ClientHello sigalgs DoS fix
6623 If a client connects to an OpenSSL 1.0.2 server and renegotiates with an
6624 invalid signature algorithms extension a NULL pointer dereference will
6625 occur. This can be exploited in a DoS attack against the server.
6627 This issue was was reported to OpenSSL by David Ramos of Stanford
6631 *Stephen Henson and Matt Caswell*
6633 * Multiblock corrupted pointer fix
6635 OpenSSL 1.0.2 introduced the "multiblock" performance improvement. This
6636 feature only applies on 64 bit x86 architecture platforms that support AES
6637 NI instructions. A defect in the implementation of "multiblock" can cause
6638 OpenSSL's internal write buffer to become incorrectly set to NULL when
6639 using non-blocking IO. Typically, when the user application is using a
6640 socket BIO for writing, this will only result in a failed connection.
6641 However if some other BIO is used then it is likely that a segmentation
6642 fault will be triggered, thus enabling a potential DoS attack.
6644 This issue was reported to OpenSSL by Daniel Danner and Rainer Mueller.
6649 * Segmentation fault in DTLSv1_listen fix
6651 The DTLSv1_listen function is intended to be stateless and processes the
6652 initial ClientHello from many peers. It is common for user code to loop
6653 over the call to DTLSv1_listen until a valid ClientHello is received with
6654 an associated cookie. A defect in the implementation of DTLSv1_listen means
6655 that state is preserved in the SSL object from one invocation to the next
6656 that can lead to a segmentation fault. Errors processing the initial
6657 ClientHello can trigger this scenario. An example of such an error could be
6658 that a DTLS1.0 only client is attempting to connect to a DTLS1.2 only
6661 This issue was reported to OpenSSL by Per Allansson.
6666 * Segmentation fault in ASN1_TYPE_cmp fix
6668 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
6669 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
6670 certificate signature algorithm consistency this can be used to crash any
6671 certificate verification operation and exploited in a DoS attack. Any
6672 application which performs certificate verification is vulnerable including
6673 OpenSSL clients and servers which enable client authentication.
6678 * Segmentation fault for invalid PSS parameters fix
6680 The signature verification routines will crash with a NULL pointer
6681 dereference if presented with an ASN.1 signature using the RSA PSS
6682 algorithm and invalid parameters. Since these routines are used to verify
6683 certificate signature algorithms this can be used to crash any
6684 certificate verification operation and exploited in a DoS attack. Any
6685 application which performs certificate verification is vulnerable including
6686 OpenSSL clients and servers which enable client authentication.
6688 This issue was was reported to OpenSSL by Brian Carpenter.
6693 * ASN.1 structure reuse memory corruption fix
6695 Reusing a structure in ASN.1 parsing may allow an attacker to cause
6696 memory corruption via an invalid write. Such reuse is and has been
6697 strongly discouraged and is believed to be rare.
6699 Applications that parse structures containing CHOICE or ANY DEFINED BY
6700 components may be affected. Certificate parsing (d2i_X509 and related
6701 functions) are however not affected. OpenSSL clients and servers are
6707 * PKCS7 NULL pointer dereferences fix
6709 The PKCS#7 parsing code does not handle missing outer ContentInfo
6710 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
6711 missing content and trigger a NULL pointer dereference on parsing.
6713 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
6714 otherwise parse PKCS#7 structures from untrusted sources are
6715 affected. OpenSSL clients and servers are not affected.
6717 This issue was reported to OpenSSL by Michal Zalewski (Google).
6722 * DoS via reachable assert in SSLv2 servers fix
6724 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
6725 servers that both support SSLv2 and enable export cipher suites by sending
6726 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
6728 This issue was discovered by Sean Burford (Google) and Emilia Käsper
6729 (OpenSSL development team).
6734 * Empty CKE with client auth and DHE fix
6736 If client auth is used then a server can seg fault in the event of a DHE
6737 ciphersuite being selected and a zero length ClientKeyExchange message
6738 being sent by the client. This could be exploited in a DoS attack.
6743 * Handshake with unseeded PRNG fix
6745 Under certain conditions an OpenSSL 1.0.2 client can complete a handshake
6746 with an unseeded PRNG. The conditions are:
6747 - The client is on a platform where the PRNG has not been seeded
6748 automatically, and the user has not seeded manually
6749 - A protocol specific client method version has been used (i.e. not
6750 SSL_client_methodv23)
6751 - A ciphersuite is used that does not require additional random data from
6752 the PRNG beyond the initial ClientHello client random (e.g. PSK-RC4-SHA).
6754 If the handshake succeeds then the client random that has been used will
6755 have been generated from a PRNG with insufficient entropy and therefore the
6756 output may be predictable.
6758 For example using the following command with an unseeded openssl will
6759 succeed on an unpatched platform:
6761 openssl s_client -psk 1a2b3c4d -tls1_2 -cipher PSK-RC4-SHA
6766 * Use After Free following d2i_ECPrivatekey error fix
6768 A malformed EC private key file consumed via the d2i_ECPrivateKey function
6769 could cause a use after free condition. This, in turn, could cause a double
6770 free in several private key parsing functions (such as d2i_PrivateKey
6771 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
6772 for applications that receive EC private keys from untrusted
6773 sources. This scenario is considered rare.
6775 This issue was discovered by the BoringSSL project and fixed in their
6781 * X509_to_X509_REQ NULL pointer deref fix
6783 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
6784 the certificate key is invalid. This function is rarely used in practice.
6786 This issue was discovered by Brian Carpenter.
6791 * Removed the export ciphers from the DEFAULT ciphers
6795 ### Changes between 1.0.1l and 1.0.2 [22 Jan 2015]
6797 * Facilitate "universal" ARM builds targeting range of ARM ISAs, e.g.
6798 ARMv5 through ARMv8, as opposite to "locking" it to single one.
6799 So far those who have to target multiple platforms would compromise
6800 and argue that binary targeting say ARMv5 would still execute on
6801 ARMv8. "Universal" build resolves this compromise by providing
6802 near-optimal performance even on newer platforms.
6806 * Accelerated NIST P-256 elliptic curve implementation for x86_64
6807 (other platforms pending).
6809 *Shay Gueron & Vlad Krasnov (Intel Corp), Andy Polyakov*
6811 * Add support for the SignedCertificateTimestampList certificate and
6812 OCSP response extensions from RFC6962.
6816 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
6817 for corner cases. (Certain input points at infinity could lead to
6818 bogus results, with non-infinity inputs mapped to infinity too.)
6822 * Initial support for PowerISA 2.0.7, first implemented in POWER8.
6823 This covers AES, SHA256/512 and GHASH. "Initial" means that most
6824 common cases are optimized and there still is room for further
6825 improvements. Vector Permutation AES for Altivec is also added.
6829 * Add support for little-endian ppc64 Linux target.
6831 *Marcelo Cerri (IBM)*
6833 * Initial support for AMRv8 ISA crypto extensions. This covers AES,
6834 SHA1, SHA256 and GHASH. "Initial" means that most common cases
6835 are optimized and there still is room for further improvements.
6836 Both 32- and 64-bit modes are supported.
6838 *Andy Polyakov, Ard Biesheuvel (Linaro)*
6840 * Improved ARMv7 NEON support.
6844 * Support for SPARC Architecture 2011 crypto extensions, first
6845 implemented in SPARC T4. This covers AES, DES, Camellia, SHA1,
6846 SHA256/512, MD5, GHASH and modular exponentiation.
6848 *Andy Polyakov, David Miller*
6850 * Accelerated modular exponentiation for Intel processors, a.k.a.
6853 *Shay Gueron & Vlad Krasnov (Intel Corp)*
6855 * Support for new and upcoming Intel processors, including AVX2,
6856 BMI and SHA ISA extensions. This includes additional "stitched"
6857 implementations, AESNI-SHA256 and GCM, and multi-buffer support
6860 This work was sponsored by Intel Corp.
6864 * Support for DTLS 1.2. This adds two sets of DTLS methods: DTLS_*_method()
6865 supports both DTLS 1.2 and 1.0 and should use whatever version the peer
6866 supports and DTLSv1_2_*_method() which supports DTLS 1.2 only.
6870 * Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
6871 this fixes a limitation in previous versions of OpenSSL.
6875 * Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
6876 MGF1 digest and OAEP label.
6880 * Add EVP support for key wrapping algorithms, to avoid problems with
6881 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
6882 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
6883 algorithms and include tests cases.
6887 * Add functions to allocate and set the fields of an ECDSA_METHOD
6890 *Douglas E. Engert, Steve Henson*
6892 * New functions OPENSSL_gmtime_diff and ASN1_TIME_diff to find the
6893 difference in days and seconds between two tm or ASN1_TIME structures.
6897 * Add -rev test option to s_server to just reverse order of characters
6898 received by client and send back to server. Also prints an abbreviated
6899 summary of the connection parameters.
6903 * New option -brief for s_client and s_server to print out a brief summary
6904 of connection parameters.
6908 * Add callbacks for arbitrary TLS extensions.
6910 *Trevor Perrin <trevp@trevp.net> and Ben Laurie*
6912 * New option -crl_download in several openssl utilities to download CRLs
6913 from CRLDP extension in certificates.
6917 * New options -CRL and -CRLform for s_client and s_server for CRLs.
6921 * New function X509_CRL_diff to generate a delta CRL from the difference
6922 of two full CRLs. Add support to "crl" utility.
6926 * New functions to set lookup_crls function and to retrieve
6927 X509_STORE from X509_STORE_CTX.
6931 * Print out deprecated issuer and subject unique ID fields in
6936 * Extend OCSP I/O functions so they can be used for simple general purpose
6937 HTTP as well as OCSP. New wrapper function which can be used to download
6938 CRLs using the OCSP API.
6942 * Delegate command line handling in s_client/s_server to SSL_CONF APIs.
6946 * `SSL_CONF*` functions. These provide a common framework for application
6947 configuration using configuration files or command lines.
6951 * SSL/TLS tracing code. This parses out SSL/TLS records using the
6952 message callback and prints the results. Needs compile time option
6953 "enable-ssl-trace". New options to s_client and s_server to enable
6958 * New ctrl and macro to retrieve supported points extensions.
6959 Print out extension in s_server and s_client.
6963 * New functions to retrieve certificate signature and signature
6968 * Add functions to retrieve and manipulate the raw cipherlist sent by a
6973 * New Suite B modes for TLS code. These use and enforce the requirements
6974 of RFC6460: restrict ciphersuites, only permit Suite B algorithms and
6975 only use Suite B curves. The Suite B modes can be set by using the
6976 strings "SUITEB128", "SUITEB192" or "SUITEB128ONLY" for the cipherstring.
6980 * New chain verification flags for Suite B levels of security. Check
6981 algorithms are acceptable when flags are set in X509_verify_cert.
6985 * Make tls1_check_chain return a set of flags indicating checks passed
6986 by a certificate chain. Add additional tests to handle client
6987 certificates: checks for matching certificate type and issuer name
6992 * If an attempt is made to use a signature algorithm not in the peer
6993 preference list abort the handshake. If client has no suitable
6994 signature algorithms in response to a certificate request do not
6995 use the certificate.
6999 * If server EC tmp key is not in client preference list abort handshake.
7003 * Add support for certificate stores in CERT structure. This makes it
7004 possible to have different stores per SSL structure or one store in
7005 the parent SSL_CTX. Include distinct stores for certificate chain
7006 verification and chain building. New ctrl SSL_CTRL_BUILD_CERT_CHAIN
7007 to build and store a certificate chain in CERT structure: returning
7008 an error if the chain cannot be built: this will allow applications
7009 to test if a chain is correctly configured.
7011 Note: if the CERT based stores are not set then the parent SSL_CTX
7012 store is used to retain compatibility with existing behaviour.
7016 * New function ssl_set_client_disabled to set a ciphersuite disabled
7017 mask based on the current session, check mask when sending client
7018 hello and checking the requested ciphersuite.
7022 * New ctrls to retrieve and set certificate types in a certificate
7023 request message. Print out received values in s_client. If certificate
7024 types is not set with custom values set sensible values based on
7025 supported signature algorithms.
7029 * Support for distinct client and server supported signature algorithms.
7033 * Add certificate callback. If set this is called whenever a certificate
7034 is required by client or server. An application can decide which
7035 certificate chain to present based on arbitrary criteria: for example
7036 supported signature algorithms. Add very simple example to s_server.
7037 This fixes many of the problems and restrictions of the existing client
7038 certificate callback: for example you can now clear an existing
7039 certificate and specify the whole chain.
7043 * Add new "valid_flags" field to CERT_PKEY structure which determines what
7044 the certificate can be used for (if anything). Set valid_flags field
7045 in new tls1_check_chain function. Simplify ssl_set_cert_masks which used
7046 to have similar checks in it.
7048 Add new "cert_flags" field to CERT structure and include a "strict mode".
7049 This enforces some TLS certificate requirements (such as only permitting
7050 certificate signature algorithms contained in the supported algorithms
7051 extension) which some implementations ignore: this option should be used
7052 with caution as it could cause interoperability issues.
7056 * Update and tidy signature algorithm extension processing. Work out
7057 shared signature algorithms based on preferences and peer algorithms
7058 and print them out in s_client and s_server. Abort handshake if no
7059 shared signature algorithms.
7063 * Add new functions to allow customised supported signature algorithms
7064 for SSL and SSL_CTX structures. Add options to s_client and s_server
7069 * New function SSL_certs_clear() to delete all references to certificates
7070 from an SSL structure. Before this once a certificate had been added
7071 it couldn't be removed.
7075 * Integrate hostname, email address and IP address checking with certificate
7076 verification. New verify options supporting checking in openssl utility.
7080 * Fixes and wildcard matching support to hostname and email checking
7081 functions. Add manual page.
7083 *Florian Weimer (Red Hat Product Security Team)*
7085 * New functions to check a hostname email or IP address against a
7086 certificate. Add options x509 utility to print results of checks against
7091 * Fix OCSP checking.
7093 *Rob Stradling <rob.stradling@comodo.com> and Ben Laurie*
7095 * Initial experimental support for explicitly trusted non-root CAs.
7096 OpenSSL still tries to build a complete chain to a root but if an
7097 intermediate CA has a trust setting included that is used. The first
7098 setting is used: whether to trust (e.g., -addtrust option to the x509
7103 * Add -trusted_first option which attempts to find certificates in the
7104 trusted store even if an untrusted chain is also supplied.
7108 * MIPS assembly pack updates: support for MIPS32r2 and SmartMIPS ASE,
7109 platform support for Linux and Android.
7113 * Support for linux-x32, ILP32 environment in x86_64 framework.
7117 * Experimental multi-implementation support for FIPS capable OpenSSL.
7118 When in FIPS mode the approved implementations are used as normal,
7119 when not in FIPS mode the internal unapproved versions are used instead.
7120 This means that the FIPS capable OpenSSL isn't forced to use the
7121 (often lower performance) FIPS implementations outside FIPS mode.
7125 * Transparently support X9.42 DH parameters when calling
7126 PEM_read_bio_DHparameters. This means existing applications can handle
7127 the new parameter format automatically.
7131 * Initial experimental support for X9.42 DH parameter format: mainly
7132 to support use of 'q' parameter for RFC5114 parameters.
7136 * Add DH parameters from RFC5114 including test data to dhtest.
7140 * Support for automatic EC temporary key parameter selection. If enabled
7141 the most preferred EC parameters are automatically used instead of
7142 hardcoded fixed parameters. Now a server just has to call:
7143 SSL_CTX_set_ecdh_auto(ctx, 1) and the server will automatically
7144 support ECDH and use the most appropriate parameters.
7148 * Enhance and tidy EC curve and point format TLS extension code. Use
7149 static structures instead of allocation if default values are used.
7150 New ctrls to set curves we wish to support and to retrieve shared curves.
7151 Print out shared curves in s_server. New options to s_server and s_client
7152 to set list of supported curves.
7156 * New ctrls to retrieve supported signature algorithms and
7157 supported curve values as an array of NIDs. Extend openssl utility
7158 to print out received values.
7162 * Add new APIs EC_curve_nist2nid and EC_curve_nid2nist which convert
7163 between NIDs and the more common NIST names such as "P-256". Enhance
7164 ecparam utility and ECC method to recognise the NIST names for curves.
7168 * Enhance SSL/TLS certificate chain handling to support different
7169 chains for each certificate instead of one chain in the parent SSL_CTX.
7173 * Support for fixed DH ciphersuite client authentication: where both
7174 server and client use DH certificates with common parameters.
7178 * Support for fixed DH ciphersuites: those requiring DH server
7183 * New function i2d_re_X509_tbs for re-encoding the TBS portion of
7185 Note: Related 1.0.2-beta specific macros X509_get_cert_info,
7186 X509_CINF_set_modified, X509_CINF_get_issuer, X509_CINF_get_extensions and
7187 X509_CINF_get_signature were reverted post internal team review.
7192 ### Changes between 1.0.1t and 1.0.1u [22 Sep 2016]
7194 * OCSP Status Request extension unbounded memory growth
7196 A malicious client can send an excessively large OCSP Status Request
7197 extension. If that client continually requests renegotiation, sending a
7198 large OCSP Status Request extension each time, then there will be unbounded
7199 memory growth on the server. This will eventually lead to a Denial Of
7200 Service attack through memory exhaustion. Servers with a default
7201 configuration are vulnerable even if they do not support OCSP. Builds using
7202 the "no-ocsp" build time option are not affected.
7204 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
7209 * In order to mitigate the SWEET32 attack, the DES ciphers were moved from
7212 This issue was reported to OpenSSL Karthikeyan Bhargavan and Gaetan
7218 * OOB write in MDC2_Update()
7220 An overflow can occur in MDC2_Update() either if called directly or
7221 through the EVP_DigestUpdate() function using MDC2. If an attacker
7222 is able to supply very large amounts of input data after a previous
7223 call to EVP_EncryptUpdate() with a partial block then a length check
7224 can overflow resulting in a heap corruption.
7226 The amount of data needed is comparable to SIZE_MAX which is impractical
7229 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
7234 * Malformed SHA512 ticket DoS
7236 If a server uses SHA512 for TLS session ticket HMAC it is vulnerable to a
7237 DoS attack where a malformed ticket will result in an OOB read which will
7240 The use of SHA512 in TLS session tickets is comparatively rare as it requires
7241 a custom server callback and ticket lookup mechanism.
7243 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
7248 * OOB write in BN_bn2dec()
7250 The function BN_bn2dec() does not check the return value of BN_div_word().
7251 This can cause an OOB write if an application uses this function with an
7252 overly large BIGNUM. This could be a problem if an overly large certificate
7253 or CRL is printed out from an untrusted source. TLS is not affected because
7254 record limits will reject an oversized certificate before it is parsed.
7256 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
7261 * OOB read in TS_OBJ_print_bio()
7263 The function TS_OBJ_print_bio() misuses OBJ_obj2txt(): the return value is
7264 the total length the OID text representation would use and not the amount
7265 of data written. This will result in OOB reads when large OIDs are
7268 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
7273 * Pointer arithmetic undefined behaviour
7275 Avoid some undefined pointer arithmetic
7277 A common idiom in the codebase is to check limits in the following manner:
7280 Where "p" points to some malloc'd data of SIZE bytes and
7283 "len" here could be from some externally supplied data (e.g. from a TLS
7286 The rules of C pointer arithmetic are such that "p + len" is only well
7287 defined where len <= SIZE. Therefore, the above idiom is actually
7288 undefined behaviour.
7290 For example this could cause problems if some malloc implementation
7291 provides an address for "p" such that "p + len" actually overflows for
7292 values of len that are too big and therefore p + len < limit.
7294 This issue was reported to OpenSSL by Guido Vranken
7299 * Constant time flag not preserved in DSA signing
7301 Operations in the DSA signing algorithm should run in constant time in
7302 order to avoid side channel attacks. A flaw in the OpenSSL DSA
7303 implementation means that a non-constant time codepath is followed for
7304 certain operations. This has been demonstrated through a cache-timing
7305 attack to be sufficient for an attacker to recover the private DSA key.
7307 This issue was reported by César Pereida (Aalto University), Billy Brumley
7308 (Tampere University of Technology), and Yuval Yarom (The University of
7309 Adelaide and NICTA).
7314 * DTLS buffered message DoS
7316 In a DTLS connection where handshake messages are delivered out-of-order
7317 those messages that OpenSSL is not yet ready to process will be buffered
7318 for later use. Under certain circumstances, a flaw in the logic means that
7319 those messages do not get removed from the buffer even though the handshake
7320 has been completed. An attacker could force up to approx. 15 messages to
7321 remain in the buffer when they are no longer required. These messages will
7322 be cleared when the DTLS connection is closed. The default maximum size for
7323 a message is 100k. Therefore, the attacker could force an additional 1500k
7324 to be consumed per connection. By opening many simultaneous connections an
7325 attacker could cause a DoS attack through memory exhaustion.
7327 This issue was reported to OpenSSL by Quan Luo.
7332 * DTLS replay protection DoS
7334 A flaw in the DTLS replay attack protection mechanism means that records
7335 that arrive for future epochs update the replay protection "window" before
7336 the MAC for the record has been validated. This could be exploited by an
7337 attacker by sending a record for the next epoch (which does not have to
7338 decrypt or have a valid MAC), with a very large sequence number. This means
7339 that all subsequent legitimate packets are dropped causing a denial of
7340 service for a specific DTLS connection.
7342 This issue was reported to OpenSSL by the OCAP audit team.
7347 * Certificate message OOB reads
7349 In OpenSSL 1.0.2 and earlier some missing message length checks can result
7350 in OOB reads of up to 2 bytes beyond an allocated buffer. There is a
7351 theoretical DoS risk but this has not been observed in practice on common
7354 The messages affected are client certificate, client certificate request
7355 and server certificate. As a result the attack can only be performed
7356 against a client or a server which enables client authentication.
7358 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
7363 ### Changes between 1.0.1s and 1.0.1t [3 May 2016]
7365 * Prevent padding oracle in AES-NI CBC MAC check
7367 A MITM attacker can use a padding oracle attack to decrypt traffic
7368 when the connection uses an AES CBC cipher and the server support
7371 This issue was introduced as part of the fix for Lucky 13 padding
7372 attack ([CVE-2013-0169]). The padding check was rewritten to be in
7373 constant time by making sure that always the same bytes are read and
7374 compared against either the MAC or padding bytes. But it no longer
7375 checked that there was enough data to have both the MAC and padding
7378 This issue was reported by Juraj Somorovsky using TLS-Attacker.
7383 * Fix EVP_EncodeUpdate overflow
7385 An overflow can occur in the EVP_EncodeUpdate() function which is used for
7386 Base64 encoding of binary data. If an attacker is able to supply very large
7387 amounts of input data then a length check can overflow resulting in a heap
7390 Internally to OpenSSL the EVP_EncodeUpdate() function is primarily used by
7391 the `PEM_write_bio*` family of functions. These are mainly used within the
7392 OpenSSL command line applications, so any application which processes data
7393 from an untrusted source and outputs it as a PEM file should be considered
7394 vulnerable to this issue. User applications that call these APIs directly
7395 with large amounts of untrusted data may also be vulnerable.
7397 This issue was reported by Guido Vranken.
7402 * Fix EVP_EncryptUpdate overflow
7404 An overflow can occur in the EVP_EncryptUpdate() function. If an attacker
7405 is able to supply very large amounts of input data after a previous call to
7406 EVP_EncryptUpdate() with a partial block then a length check can overflow
7407 resulting in a heap corruption. Following an analysis of all OpenSSL
7408 internal usage of the EVP_EncryptUpdate() function all usage is one of two
7409 forms. The first form is where the EVP_EncryptUpdate() call is known to be
7410 the first called function after an EVP_EncryptInit(), and therefore that
7411 specific call must be safe. The second form is where the length passed to
7412 EVP_EncryptUpdate() can be seen from the code to be some small value and
7413 therefore there is no possibility of an overflow. Since all instances are
7414 one of these two forms, it is believed that there can be no overflows in
7415 internal code due to this problem. It should be noted that
7416 EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths.
7417 Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances
7418 of these calls have also been analysed too and it is believed there are no
7419 instances in internal usage where an overflow could occur.
7421 This issue was reported by Guido Vranken.
7426 * Prevent ASN.1 BIO excessive memory allocation
7428 When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio()
7429 a short invalid encoding can casuse allocation of large amounts of memory
7430 potentially consuming excessive resources or exhausting memory.
7432 Any application parsing untrusted data through d2i BIO functions is
7433 affected. The memory based functions such as d2i_X509() are *not* affected.
7434 Since the memory based functions are used by the TLS library, TLS
7435 applications are not affected.
7437 This issue was reported by Brian Carpenter.
7444 ASN1 Strings that are over 1024 bytes can cause an overread in applications
7445 using the X509_NAME_oneline() function on EBCDIC systems. This could result
7446 in arbitrary stack data being returned in the buffer.
7448 This issue was reported by Guido Vranken.
7453 * Modify behavior of ALPN to invoke callback after SNI/servername
7454 callback, such that updates to the SSL_CTX affect ALPN.
7458 * Remove LOW from the DEFAULT cipher list. This removes singles DES from the
7463 * Only remove the SSLv2 methods with the no-ssl2-method option. When the
7464 methods are enabled and ssl2 is disabled the methods return NULL.
7468 ### Changes between 1.0.1r and 1.0.1s [1 Mar 2016]
7470 * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL.
7471 Builds that are not configured with "enable-weak-ssl-ciphers" will not
7472 provide any "EXPORT" or "LOW" strength ciphers.
7476 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2
7477 is by default disabled at build-time. Builds that are not configured with
7478 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used,
7479 users who want to negotiate SSLv2 via the version-flexible SSLv23_method()
7480 will need to explicitly call either of:
7482 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2);
7484 SSL_clear_options(ssl, SSL_OP_NO_SSLv2);
7486 as appropriate. Even if either of those is used, or the application
7487 explicitly uses the version-specific SSLv2_method() or its client and
7488 server variants, SSLv2 ciphers vulnerable to exhaustive search key
7489 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT
7490 ciphers, and SSLv2 56-bit DES are no longer available.
7495 * Fix a double-free in DSA code
7497 A double free bug was discovered when OpenSSL parses malformed DSA private
7498 keys and could lead to a DoS attack or memory corruption for applications
7499 that receive DSA private keys from untrusted sources. This scenario is
7502 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using
7508 * Disable SRP fake user seed to address a server memory leak.
7510 Add a new method SRP_VBASE_get1_by_user that handles the seed properly.
7512 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
7513 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user
7514 was changed to ignore the "fake user" SRP seed, even if the seed
7517 Users should use SRP_VBASE_get1_by_user instead. Note that in
7518 SRP_VBASE_get1_by_user, caller must free the returned value. Note
7519 also that even though configuring the SRP seed attempts to hide
7520 invalid usernames by continuing the handshake with fake
7521 credentials, this behaviour is not constant time and no strong
7522 guarantees are made that the handshake is indistinguishable from
7523 that of a valid user.
7528 * Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption
7530 In the BN_hex2bn function the number of hex digits is calculated using an
7531 int value `i`. Later `bn_expand` is called with a value of `i * 4`. For
7532 large values of `i` this can result in `bn_expand` not allocating any
7533 memory because `i * 4` is negative. This can leave the internal BIGNUM data
7534 field as NULL leading to a subsequent NULL ptr deref. For very large values
7535 of `i`, the calculation `i * 4` could be a positive value smaller than `i`.
7536 In this case memory is allocated to the internal BIGNUM data field, but it
7537 is insufficiently sized leading to heap corruption. A similar issue exists
7538 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn
7539 is ever called by user applications with very large untrusted hex/dec data.
7540 This is anticipated to be a rare occurrence.
7542 All OpenSSL internal usage of these functions use data that is not expected
7543 to be untrusted, e.g. config file data or application command line
7544 arguments. If user developed applications generate config file data based
7545 on untrusted data then it is possible that this could also lead to security
7546 consequences. This is also anticipated to be rare.
7548 This issue was reported to OpenSSL by Guido Vranken.
7553 * Fix memory issues in `BIO_*printf` functions
7555 The internal `fmtstr` function used in processing a "%s" format string in
7556 the `BIO_*printf` functions could overflow while calculating the length of a
7557 string and cause an OOB read when printing very long strings.
7559 Additionally the internal `doapr_outch` function can attempt to write to an
7560 OOB memory location (at an offset from the NULL pointer) in the event of a
7561 memory allocation failure. In 1.0.2 and below this could be caused where
7562 the size of a buffer to be allocated is greater than INT_MAX. E.g. this
7563 could be in processing a very long "%s" format string. Memory leaks can
7566 The first issue may mask the second issue dependent on compiler behaviour.
7567 These problems could enable attacks where large amounts of untrusted data
7568 is passed to the `BIO_*printf` functions. If applications use these functions
7569 in this way then they could be vulnerable. OpenSSL itself uses these
7570 functions when printing out human-readable dumps of ASN.1 data. Therefore
7571 applications that print this data could be vulnerable if the data is from
7572 untrusted sources. OpenSSL command line applications could also be
7573 vulnerable where they print out ASN.1 data, or if untrusted data is passed
7574 as command line arguments.
7576 Libssl is not considered directly vulnerable. Additionally certificates etc
7577 received via remote connections via libssl are also unlikely to be able to
7578 trigger these issues because of message size limits enforced within libssl.
7580 This issue was reported to OpenSSL Guido Vranken.
7585 * Side channel attack on modular exponentiation
7587 A side-channel attack was found which makes use of cache-bank conflicts on
7588 the Intel Sandy-Bridge microarchitecture which could lead to the recovery
7589 of RSA keys. The ability to exploit this issue is limited as it relies on
7590 an attacker who has control of code in a thread running on the same
7591 hyper-threaded core as the victim thread which is performing decryptions.
7593 This issue was reported to OpenSSL by Yuval Yarom, The University of
7594 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and
7595 Nadia Heninger, University of Pennsylvania with more information at
7596 <http://cachebleed.info>.
7601 * Change the req command to generate a 2048-bit RSA/DSA key by default,
7602 if no keysize is specified with default_bits. This fixes an
7603 omission in an earlier change that changed all RSA/DSA key generation
7604 commands to use 2048 bits by default.
7608 ### Changes between 1.0.1q and 1.0.1r [28 Jan 2016]
7610 * Protection for DH small subgroup attacks
7612 As a precautionary measure the SSL_OP_SINGLE_DH_USE option has been
7613 switched on by default and cannot be disabled. This could have some
7618 * SSLv2 doesn't block disabled ciphers
7620 A malicious client can negotiate SSLv2 ciphers that have been disabled on
7621 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
7622 been disabled, provided that the SSLv2 protocol was not also disabled via
7625 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
7626 and Sebastian Schinzel.
7631 * Reject DH handshakes with parameters shorter than 1024 bits.
7635 ### Changes between 1.0.1p and 1.0.1q [3 Dec 2015]
7637 * Certificate verify crash with missing PSS parameter
7639 The signature verification routines will crash with a NULL pointer
7640 dereference if presented with an ASN.1 signature using the RSA PSS
7641 algorithm and absent mask generation function parameter. Since these
7642 routines are used to verify certificate signature algorithms this can be
7643 used to crash any certificate verification operation and exploited in a
7644 DoS attack. Any application which performs certificate verification is
7645 vulnerable including OpenSSL clients and servers which enable client
7648 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
7653 * X509_ATTRIBUTE memory leak
7655 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
7656 memory. This structure is used by the PKCS#7 and CMS routines so any
7657 application which reads PKCS#7 or CMS data from untrusted sources is
7658 affected. SSL/TLS is not affected.
7660 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
7666 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
7667 This changes the decoding behaviour for some invalid messages,
7668 though the change is mostly in the more lenient direction, and
7669 legacy behaviour is preserved as much as possible.
7673 * In DSA_generate_parameters_ex, if the provided seed is too short,
7674 use a random seed, as already documented.
7676 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
7678 ### Changes between 1.0.1o and 1.0.1p [9 Jul 2015]
7680 * Alternate chains certificate forgery
7682 During certificate verification, OpenSSL will attempt to find an
7683 alternative certificate chain if the first attempt to build such a chain
7684 fails. An error in the implementation of this logic can mean that an
7685 attacker could cause certain checks on untrusted certificates to be
7686 bypassed, such as the CA flag, enabling them to use a valid leaf
7687 certificate to act as a CA and "issue" an invalid certificate.
7689 This issue was reported to OpenSSL by Adam Langley/David Benjamin
7695 * Race condition handling PSK identify hint
7697 If PSK identity hints are received by a multi-threaded client then
7698 the values are wrongly updated in the parent SSL_CTX structure. This can
7699 result in a race condition potentially leading to a double free of the
7705 ### Changes between 1.0.1n and 1.0.1o [12 Jun 2015]
7707 * Fix HMAC ABI incompatibility. The previous version introduced an ABI
7708 incompatibility in the handling of HMAC. The previous ABI has now been
7711 ### Changes between 1.0.1m and 1.0.1n [11 Jun 2015]
7713 * Malformed ECParameters causes infinite loop
7715 When processing an ECParameters structure OpenSSL enters an infinite loop
7716 if the curve specified is over a specially malformed binary polynomial
7719 This can be used to perform denial of service against any
7720 system which processes public keys, certificate requests or
7721 certificates. This includes TLS clients and TLS servers with
7722 client authentication enabled.
7724 This issue was reported to OpenSSL by Joseph Barr-Pixton.
7729 * Exploitable out-of-bounds read in X509_cmp_time
7731 X509_cmp_time does not properly check the length of the ASN1_TIME
7732 string and can read a few bytes out of bounds. In addition,
7733 X509_cmp_time accepts an arbitrary number of fractional seconds in the
7736 An attacker can use this to craft malformed certificates and CRLs of
7737 various sizes and potentially cause a segmentation fault, resulting in
7738 a DoS on applications that verify certificates or CRLs. TLS clients
7739 that verify CRLs are affected. TLS clients and servers with client
7740 authentication enabled may be affected if they use custom verification
7743 This issue was reported to OpenSSL by Robert Swiecki (Google), and
7744 independently by Hanno Böck.
7749 * PKCS7 crash with missing EnvelopedContent
7751 The PKCS#7 parsing code does not handle missing inner EncryptedContent
7752 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
7753 with missing content and trigger a NULL pointer dereference on parsing.
7755 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
7756 structures from untrusted sources are affected. OpenSSL clients and
7757 servers are not affected.
7759 This issue was reported to OpenSSL by Michal Zalewski (Google).
7764 * CMS verify infinite loop with unknown hash function
7766 When verifying a signedData message the CMS code can enter an infinite loop
7767 if presented with an unknown hash function OID. This can be used to perform
7768 denial of service against any system which verifies signedData messages using
7770 This issue was reported to OpenSSL by Johannes Bauer.
7775 * Race condition handling NewSessionTicket
7777 If a NewSessionTicket is received by a multi-threaded client when attempting to
7778 reuse a previous ticket then a race condition can occur potentially leading to
7779 a double free of the ticket data.
7784 * Reject DH handshakes with parameters shorter than 768 bits.
7786 *Kurt Roeckx and Emilia Kasper*
7788 * dhparam: generate 2048-bit parameters by default.
7790 *Kurt Roeckx and Emilia Kasper*
7792 ### Changes between 1.0.1l and 1.0.1m [19 Mar 2015]
7794 * Segmentation fault in ASN1_TYPE_cmp fix
7796 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
7797 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
7798 certificate signature algorithm consistency this can be used to crash any
7799 certificate verification operation and exploited in a DoS attack. Any
7800 application which performs certificate verification is vulnerable including
7801 OpenSSL clients and servers which enable client authentication.
7806 * ASN.1 structure reuse memory corruption fix
7808 Reusing a structure in ASN.1 parsing may allow an attacker to cause
7809 memory corruption via an invalid write. Such reuse is and has been
7810 strongly discouraged and is believed to be rare.
7812 Applications that parse structures containing CHOICE or ANY DEFINED BY
7813 components may be affected. Certificate parsing (d2i_X509 and related
7814 functions) are however not affected. OpenSSL clients and servers are
7820 * PKCS7 NULL pointer dereferences fix
7822 The PKCS#7 parsing code does not handle missing outer ContentInfo
7823 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
7824 missing content and trigger a NULL pointer dereference on parsing.
7826 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
7827 otherwise parse PKCS#7 structures from untrusted sources are
7828 affected. OpenSSL clients and servers are not affected.
7830 This issue was reported to OpenSSL by Michal Zalewski (Google).
7835 * DoS via reachable assert in SSLv2 servers fix
7837 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
7838 servers that both support SSLv2 and enable export cipher suites by sending
7839 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
7841 This issue was discovered by Sean Burford (Google) and Emilia Käsper
7842 (OpenSSL development team).
7847 * Use After Free following d2i_ECPrivatekey error fix
7849 A malformed EC private key file consumed via the d2i_ECPrivateKey function
7850 could cause a use after free condition. This, in turn, could cause a double
7851 free in several private key parsing functions (such as d2i_PrivateKey
7852 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
7853 for applications that receive EC private keys from untrusted
7854 sources. This scenario is considered rare.
7856 This issue was discovered by the BoringSSL project and fixed in their
7862 * X509_to_X509_REQ NULL pointer deref fix
7864 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
7865 the certificate key is invalid. This function is rarely used in practice.
7867 This issue was discovered by Brian Carpenter.
7872 * Removed the export ciphers from the DEFAULT ciphers
7876 ### Changes between 1.0.1k and 1.0.1l [15 Jan 2015]
7878 * Build fixes for the Windows and OpenVMS platforms
7880 *Matt Caswell and Richard Levitte*
7882 ### Changes between 1.0.1j and 1.0.1k [8 Jan 2015]
7884 * Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
7885 message can cause a segmentation fault in OpenSSL due to a NULL pointer
7886 dereference. This could lead to a Denial Of Service attack. Thanks to
7887 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
7892 * Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
7893 dtls1_buffer_record function under certain conditions. In particular this
7894 could occur if an attacker sent repeated DTLS records with the same
7895 sequence number but for the next epoch. The memory leak could be exploited
7896 by an attacker in a Denial of Service attack through memory exhaustion.
7897 Thanks to Chris Mueller for reporting this issue.
7902 * Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
7903 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
7904 method would be set to NULL which could later result in a NULL pointer
7905 dereference. Thanks to Frank Schmirler for reporting this issue.
7910 * Abort handshake if server key exchange message is omitted for ephemeral
7913 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
7914 reporting this issue.
7919 * Remove non-export ephemeral RSA code on client and server. This code
7920 violated the TLS standard by allowing the use of temporary RSA keys in
7921 non-export ciphersuites and could be used by a server to effectively
7922 downgrade the RSA key length used to a value smaller than the server
7923 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
7924 INRIA or reporting this issue.
7929 * Fixed issue where DH client certificates are accepted without verification.
7930 An OpenSSL server will accept a DH certificate for client authentication
7931 without the certificate verify message. This effectively allows a client to
7932 authenticate without the use of a private key. This only affects servers
7933 which trust a client certificate authority which issues certificates
7934 containing DH keys: these are extremely rare and hardly ever encountered.
7935 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
7941 * Ensure that the session ID context of an SSL is updated when its
7942 SSL_CTX is updated via SSL_set_SSL_CTX.
7944 The session ID context is typically set from the parent SSL_CTX,
7945 and can vary with the CTX.
7949 * Fix various certificate fingerprint issues.
7951 By using non-DER or invalid encodings outside the signed portion of a
7952 certificate the fingerprint can be changed without breaking the signature.
7953 Although no details of the signed portion of the certificate can be changed
7954 this can cause problems with some applications: e.g. those using the
7955 certificate fingerprint for blacklists.
7957 1. Reject signatures with non zero unused bits.
7959 If the BIT STRING containing the signature has non zero unused bits reject
7960 the signature. All current signature algorithms require zero unused bits.
7962 2. Check certificate algorithm consistency.
7964 Check the AlgorithmIdentifier inside TBS matches the one in the
7965 certificate signature. NB: this will result in signature failure
7966 errors for some broken certificates.
7968 Thanks to Konrad Kraszewski from Google for reporting this issue.
7970 3. Check DSA/ECDSA signatures use DER.
7972 Re-encode DSA/ECDSA signatures and compare with the original received
7973 signature. Return an error if there is a mismatch.
7975 This will reject various cases including garbage after signature
7976 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
7977 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
7978 (negative or with leading zeroes).
7980 Further analysis was conducted and fixes were developed by Stephen Henson
7981 of the OpenSSL core team.
7987 * Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
7988 results on some platforms, including x86_64. This bug occurs at random
7989 with a very low probability, and is not known to be exploitable in any
7990 way, though its exact impact is difficult to determine. Thanks to Pieter
7991 Wuille (Blockstream) who reported this issue and also suggested an initial
7992 fix. Further analysis was conducted by the OpenSSL development team and
7993 Adam Langley of Google. The final fix was developed by Andy Polyakov of
7994 the OpenSSL core team.
7999 * Do not resume sessions on the server if the negotiated protocol
8000 version does not match the session's version. Resuming with a different
8001 version, while not strictly forbidden by the RFC, is of questionable
8002 sanity and breaks all known clients.
8004 *David Benjamin, Emilia Käsper*
8006 * Tighten handling of the ChangeCipherSpec (CCS) message: reject
8007 early CCS messages during renegotiation. (Note that because
8008 renegotiation is encrypted, this early CCS was not exploitable.)
8012 * Tighten client-side session ticket handling during renegotiation:
8013 ensure that the client only accepts a session ticket if the server sends
8014 the extension anew in the ServerHello. Previously, a TLS client would
8015 reuse the old extension state and thus accept a session ticket if one was
8016 announced in the initial ServerHello.
8018 Similarly, ensure that the client requires a session ticket if one
8019 was advertised in the ServerHello. Previously, a TLS client would
8020 ignore a missing NewSessionTicket message.
8024 ### Changes between 1.0.1i and 1.0.1j [15 Oct 2014]
8028 A flaw in the DTLS SRTP extension parsing code allows an attacker, who
8029 sends a carefully crafted handshake message, to cause OpenSSL to fail
8030 to free up to 64k of memory causing a memory leak. This could be
8031 exploited in a Denial Of Service attack. This issue affects OpenSSL
8032 1.0.1 server implementations for both SSL/TLS and DTLS regardless of
8033 whether SRTP is used or configured. Implementations of OpenSSL that
8034 have been compiled with OPENSSL_NO_SRTP defined are not affected.
8036 The fix was developed by the OpenSSL team.
8041 * Session Ticket Memory Leak.
8043 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
8044 integrity of that ticket is first verified. In the event of a session
8045 ticket integrity check failing, OpenSSL will fail to free memory
8046 causing a memory leak. By sending a large number of invalid session
8047 tickets an attacker could exploit this issue in a Denial Of Service
8053 * Build option no-ssl3 is incomplete.
8055 When OpenSSL is configured with "no-ssl3" as a build option, servers
8056 could accept and complete an SSL 3.0 handshake, and clients could be
8057 configured to send them.
8060 *Akamai and the OpenSSL team*
8062 * Add support for TLS_FALLBACK_SCSV.
8063 Client applications doing fallback retries should call
8064 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
8067 *Adam Langley, Bodo Moeller*
8069 * Add additional DigestInfo checks.
8071 Re-encode DigestInto in DER and check against the original when
8072 verifying RSA signature: this will reject any improperly encoded
8073 DigestInfo structures.
8075 Note: this is a precautionary measure and no attacks are currently known.
8079 ### Changes between 1.0.1h and 1.0.1i [6 Aug 2014]
8081 * Fix SRP buffer overrun vulnerability. Invalid parameters passed to the
8082 SRP code can be overrun an internal buffer. Add sanity check that
8083 g, A, B < N to SRP code.
8085 Thanks to Sean Devlin and Watson Ladd of Cryptography Services, NCC
8086 Group for discovering this issue.
8091 * A flaw in the OpenSSL SSL/TLS server code causes the server to negotiate
8092 TLS 1.0 instead of higher protocol versions when the ClientHello message
8093 is badly fragmented. This allows a man-in-the-middle attacker to force a
8094 downgrade to TLS 1.0 even if both the server and the client support a
8095 higher protocol version, by modifying the client's TLS records.
8097 Thanks to David Benjamin and Adam Langley (Google) for discovering and
8098 researching this issue.
8103 * OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
8104 to a denial of service attack. A malicious server can crash the client
8105 with a null pointer dereference (read) by specifying an anonymous (EC)DH
8106 ciphersuite and sending carefully crafted handshake messages.
8108 Thanks to Felix Gröbert (Google) for discovering and researching this
8114 * By sending carefully crafted DTLS packets an attacker could cause openssl
8115 to leak memory. This can be exploited through a Denial of Service attack.
8116 Thanks to Adam Langley for discovering and researching this issue.
8121 * An attacker can force openssl to consume large amounts of memory whilst
8122 processing DTLS handshake messages. This can be exploited through a
8123 Denial of Service attack.
8124 Thanks to Adam Langley for discovering and researching this issue.
8129 * An attacker can force an error condition which causes openssl to crash
8130 whilst processing DTLS packets due to memory being freed twice. This
8131 can be exploited through a Denial of Service attack.
8132 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
8138 * If a multithreaded client connects to a malicious server using a resumed
8139 session and the server sends an ec point format extension it could write
8140 up to 255 bytes to freed memory.
8142 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
8148 * A malicious server can crash an OpenSSL client with a null pointer
8149 dereference (read) by specifying an SRP ciphersuite even though it was not
8150 properly negotiated with the client. This can be exploited through a
8151 Denial of Service attack.
8153 Thanks to Joonas Kuorilehto and Riku Hietamäki (Codenomicon) for
8154 discovering and researching this issue.
8159 * A flaw in OBJ_obj2txt may cause pretty printing functions such as
8160 X509_name_oneline, X509_name_print_ex et al. to leak some information
8161 from the stack. Applications may be affected if they echo pretty printing
8162 output to the attacker.
8164 Thanks to Ivan Fratric (Google) for discovering this issue.
8167 *Emilia Käsper, and Steve Henson*
8169 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
8170 for corner cases. (Certain input points at infinity could lead to
8171 bogus results, with non-infinity inputs mapped to infinity too.)
8175 ### Changes between 1.0.1g and 1.0.1h [5 Jun 2014]
8177 * Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
8178 handshake can force the use of weak keying material in OpenSSL
8179 SSL/TLS clients and servers.
8181 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
8182 researching this issue. ([CVE-2014-0224])
8184 *KIKUCHI Masashi, Steve Henson*
8186 * Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
8187 OpenSSL DTLS client the code can be made to recurse eventually crashing
8190 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
8193 *Imre Rad, Steve Henson*
8195 * Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
8196 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
8197 client or server. This is potentially exploitable to run arbitrary
8198 code on a vulnerable client or server.
8200 Thanks to Jüri Aedla for reporting this issue. ([CVE-2014-0195])
8202 *Jüri Aedla, Steve Henson*
8204 * Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
8205 are subject to a denial of service attack.
8207 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
8208 this issue. ([CVE-2014-3470])
8210 *Felix Gröbert, Ivan Fratric, Steve Henson*
8212 * Harmonize version and its documentation. -f flag is used to display
8215 *mancha <mancha1@zoho.com>*
8217 * Fix eckey_priv_encode so it immediately returns an error upon a failure
8218 in i2d_ECPrivateKey.
8220 *mancha <mancha1@zoho.com>*
8222 * Fix some double frees. These are not thought to be exploitable.
8224 *mancha <mancha1@zoho.com>*
8226 ### Changes between 1.0.1f and 1.0.1g [7 Apr 2014]
8228 * A missing bounds check in the handling of the TLS heartbeat extension
8229 can be used to reveal up to 64k of memory to a connected client or
8232 Thanks for Neel Mehta of Google Security for discovering this bug and to
8233 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
8234 preparing the fix ([CVE-2014-0160])
8236 *Adam Langley, Bodo Moeller*
8238 * Fix for the attack described in the paper "Recovering OpenSSL
8239 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
8240 by Yuval Yarom and Naomi Benger. Details can be obtained from:
8241 <http://eprint.iacr.org/2014/140>
8243 Thanks to Yuval Yarom and Naomi Benger for discovering this
8244 flaw and to Yuval Yarom for supplying a fix ([CVE-2014-0076])
8246 *Yuval Yarom and Naomi Benger*
8248 * TLS pad extension: draft-agl-tls-padding-03
8250 Workaround for the "TLS hang bug" (see FAQ and PR#2771): if the
8251 TLS client Hello record length value would otherwise be > 255 and
8252 less that 512 pad with a dummy extension containing zeroes so it
8253 is at least 512 bytes long.
8255 *Adam Langley, Steve Henson*
8257 ### Changes between 1.0.1e and 1.0.1f [6 Jan 2014]
8259 * Fix for TLS record tampering bug. A carefully crafted invalid
8260 handshake could crash OpenSSL with a NULL pointer exception.
8261 Thanks to Anton Johansson for reporting this issues.
8264 * Keep original DTLS digest and encryption contexts in retransmission
8265 structures so we can use the previous session parameters if they need
8266 to be resent. ([CVE-2013-6450])
8270 * Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
8271 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
8272 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
8273 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
8274 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
8275 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
8277 *Rob Stradling, Adam Langley*
8279 ### Changes between 1.0.1d and 1.0.1e [11 Feb 2013]
8281 * Correct fix for CVE-2013-0169. The original didn't work on AES-NI
8282 supporting platforms or when small records were transferred.
8284 *Andy Polyakov, Steve Henson*
8286 ### Changes between 1.0.1c and 1.0.1d [5 Feb 2013]
8288 * Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
8290 This addresses the flaw in CBC record processing discovered by
8291 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
8292 at: <http://www.isg.rhul.ac.uk/tls/>
8294 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
8295 Security Group at Royal Holloway, University of London
8296 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
8297 Emilia Käsper for the initial patch.
8300 *Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson*
8302 * Fix flaw in AESNI handling of TLS 1.2 and 1.1 records for CBC mode
8303 ciphersuites which can be exploited in a denial of service attack.
8304 Thanks go to and to Adam Langley <agl@chromium.org> for discovering
8305 and detecting this bug and to Wolfgang Ettlinger
8306 <wolfgang.ettlinger@gmail.com> for independently discovering this issue.
8311 * Return an error when checking OCSP signatures when key is NULL.
8312 This fixes a DoS attack. ([CVE-2013-0166])
8316 * Make openssl verify return errors.
8318 *Chris Palmer <palmer@google.com> and Ben Laurie*
8320 * Call OCSP Stapling callback after ciphersuite has been chosen, so
8321 the right response is stapled. Also change SSL_get_certificate()
8322 so it returns the certificate actually sent.
8323 See <http://rt.openssl.org/Ticket/Display.html?id=2836>.
8325 *Rob Stradling <rob.stradling@comodo.com>*
8327 * Fix possible deadlock when decoding public keys.
8331 * Don't use TLS 1.0 record version number in initial client hello
8336 ### Changes between 1.0.1b and 1.0.1c [10 May 2012]
8338 * Sanity check record length before skipping explicit IV in TLS
8339 1.2, 1.1 and DTLS to fix DoS attack.
8341 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
8342 fuzzing as a service testing platform.
8347 * Initialise tkeylen properly when encrypting CMS messages.
8348 Thanks to Solar Designer of Openwall for reporting this issue.
8352 * In FIPS mode don't try to use composite ciphers as they are not
8357 ### Changes between 1.0.1a and 1.0.1b [26 Apr 2012]
8359 * OpenSSL 1.0.0 sets SSL_OP_ALL to 0x80000FFFL and OpenSSL 1.0.1 and
8360 1.0.1a set SSL_OP_NO_TLSv1_1 to 0x00000400L which would unfortunately
8361 mean any application compiled against OpenSSL 1.0.0 headers setting
8362 SSL_OP_ALL would also set SSL_OP_NO_TLSv1_1, unintentionally disabling
8363 TLS 1.1 also. Fix this by changing the value of SSL_OP_NO_TLSv1_1 to
8364 0x10000000L Any application which was previously compiled against
8365 OpenSSL 1.0.1 or 1.0.1a headers and which cares about SSL_OP_NO_TLSv1_1
8366 will need to be recompiled as a result. Letting be results in
8367 inability to disable specifically TLS 1.1 and in client context,
8368 in unlike event, limit maximum offered version to TLS 1.0 [see below].
8372 * In order to ensure interoperability SSL_OP_NO_protocolX does not
8373 disable just protocol X, but all protocols above X *if* there are
8374 protocols *below* X still enabled. In more practical terms it means
8375 that if application wants to disable TLS1.0 in favor of TLS1.1 and
8376 above, it's not sufficient to pass `SSL_OP_NO_TLSv1`, one has to pass
8377 `SSL_OP_NO_TLSv1|SSL_OP_NO_SSLv3|SSL_OP_NO_SSLv2`. This applies to
8382 ### Changes between 1.0.1 and 1.0.1a [19 Apr 2012]
8384 * Check for potentially exploitable overflows in asn1_d2i_read_bio
8385 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
8386 in CRYPTO_realloc_clean.
8388 Thanks to Tavis Ormandy, Google Security Team, for discovering this
8389 issue and to Adam Langley <agl@chromium.org> for fixing it.
8392 *Adam Langley (Google), Tavis Ormandy, Google Security Team*
8394 * Don't allow TLS 1.2 SHA-256 ciphersuites in TLS 1.0, 1.1 connections.
8398 * Workarounds for some broken servers that "hang" if a client hello
8399 record length exceeds 255 bytes.
8401 1. Do not use record version number > TLS 1.0 in initial client
8402 hello: some (but not all) hanging servers will now work.
8403 2. If we set OPENSSL_MAX_TLS1_2_CIPHER_LENGTH this will truncate
8404 the number of ciphers sent in the client hello. This should be
8405 set to an even number, such as 50, for example by passing:
8406 -DOPENSSL_MAX_TLS1_2_CIPHER_LENGTH=50 to config or Configure.
8407 Most broken servers should now work.
8408 3. If all else fails setting OPENSSL_NO_TLS1_2_CLIENT will disable
8409 TLS 1.2 client support entirely.
8413 * Fix SEGV in Vector Permutation AES module observed in OpenSSH.
8417 ### Changes between 1.0.0h and 1.0.1 [14 Mar 2012]
8419 * Add compatibility with old MDC2 signatures which use an ASN1 OCTET
8420 STRING form instead of a DigestInfo.
8424 * The format used for MDC2 RSA signatures is inconsistent between EVP
8425 and the RSA_sign/RSA_verify functions. This was made more apparent when
8426 OpenSSL used RSA_sign/RSA_verify for some RSA signatures in particular
8427 those which went through EVP_PKEY_METHOD in 1.0.0 and later. Detect
8428 the correct format in RSA_verify so both forms transparently work.
8432 * Some servers which support TLS 1.0 can choke if we initially indicate
8433 support for TLS 1.2 and later renegotiate using TLS 1.0 in the RSA
8434 encrypted premaster secret. As a workaround use the maximum permitted
8435 client version in client hello, this should keep such servers happy
8436 and still work with previous versions of OpenSSL.
8440 * Add support for TLS/DTLS heartbeats.
8442 *Robin Seggelmann <seggelmann@fh-muenster.de>*
8444 * Add support for SCTP.
8446 *Robin Seggelmann <seggelmann@fh-muenster.de>*
8448 * Improved PRNG seeding for VOS.
8450 *Paul Green <Paul.Green@stratus.com>*
8452 * Extensive assembler packs updates, most notably:
8454 - x86[_64]: AES-NI, PCLMULQDQ, RDRAND support;
8455 - x86[_64]: SSSE3 support (SHA1, vector-permutation AES);
8456 - x86_64: bit-sliced AES implementation;
8457 - ARM: NEON support, contemporary platforms optimizations;
8458 - s390x: z196 support;
8459 - `*`: GHASH and GF(2^m) multiplication implementations;
8463 * Make TLS-SRP code conformant with RFC 5054 API cleanup
8464 (removal of unnecessary code)
8466 *Peter Sylvester <peter.sylvester@edelweb.fr>*
8468 * Add TLS key material exporter from RFC 5705.
8472 * Add DTLS-SRTP negotiation from RFC 5764.
8476 * Add Next Protocol Negotiation,
8477 <http://tools.ietf.org/html/draft-agl-tls-nextprotoneg-00>. Can be
8478 disabled with a no-npn flag to config or Configure. Code donated
8481 *Adam Langley <agl@google.com> and Ben Laurie*
8483 * Add optional 64-bit optimized implementations of elliptic curves NIST-P224,
8484 NIST-P256, NIST-P521, with constant-time single point multiplication on
8485 typical inputs. Compiler support for the nonstandard type `__uint128_t` is
8486 required to use this (present in gcc 4.4 and later, for 64-bit builds).
8487 Code made available under Apache License version 2.0.
8489 Specify "enable-ec_nistp_64_gcc_128" on the Configure (or config) command
8490 line to include this in your build of OpenSSL, and run "make depend" (or
8491 "make update"). This enables the following EC_METHODs:
8493 EC_GFp_nistp224_method()
8494 EC_GFp_nistp256_method()
8495 EC_GFp_nistp521_method()
8497 EC_GROUP_new_by_curve_name() will automatically use these (while
8498 EC_GROUP_new_curve_GFp() currently prefers the more flexible
8501 *Emilia Käsper, Adam Langley, Bodo Moeller (Google)*
8503 * Use type ossl_ssize_t instead of ssize_t which isn't available on
8504 all platforms. Move ssize_t definition from e_os.h to the public
8505 header file e_os2.h as it now appears in public header file cms.h
8509 * New -sigopt option to the ca, req and x509 utilities. Additional
8510 signature parameters can be passed using this option and in
8515 * Add RSA PSS signing function. This will generate and set the
8516 appropriate AlgorithmIdentifiers for PSS based on those in the
8517 corresponding EVP_MD_CTX structure. No application support yet.
8521 * Support for companion algorithm specific ASN1 signing routines.
8522 New function ASN1_item_sign_ctx() signs a pre-initialised
8523 EVP_MD_CTX structure and sets AlgorithmIdentifiers based on
8524 the appropriate parameters.
8528 * Add new algorithm specific ASN1 verification initialisation function
8529 to EVP_PKEY_ASN1_METHOD: this is not in EVP_PKEY_METHOD since the ASN1
8530 handling will be the same no matter what EVP_PKEY_METHOD is used.
8531 Add a PSS handler to support verification of PSS signatures: checked
8532 against a number of sample certificates.
8536 * Add signature printing for PSS. Add PSS OIDs.
8538 *Steve Henson, Martin Kaiser <lists@kaiser.cx>*
8540 * Add algorithm specific signature printing. An individual ASN1 method
8541 can now print out signatures instead of the standard hex dump.
8543 More complex signatures (e.g. PSS) can print out more meaningful
8544 information. Include DSA version that prints out the signature
8549 * Password based recipient info support for CMS library: implementing
8554 * Split password based encryption into PBES2 and PBKDF2 functions. This
8555 neatly separates the code into cipher and PBE sections and is required
8556 for some algorithms that split PBES2 into separate pieces (such as
8557 password based CMS).
8561 * Session-handling fixes:
8562 - Fix handling of connections that are resuming with a session ID,
8563 but also support Session Tickets.
8564 - Fix a bug that suppressed issuing of a new ticket if the client
8565 presented a ticket with an expired session.
8566 - Try to set the ticket lifetime hint to something reasonable.
8567 - Make tickets shorter by excluding irrelevant information.
8568 - On the client side, don't ignore renewed tickets.
8570 *Adam Langley, Bodo Moeller (Google)*
8572 * Fix PSK session representation.
8576 * Add RC4-MD5 and AESNI-SHA1 "stitched" implementations.
8578 This work was sponsored by Intel.
8582 * Add GCM support to TLS library. Some custom code is needed to split
8583 the IV between the fixed (from PRF) and explicit (from TLS record)
8584 portions. This adds all GCM ciphersuites supported by RFC5288 and
8585 RFC5289. Generalise some `AES*` cipherstrings to include GCM and
8586 add a special AESGCM string for GCM only.
8590 * Expand range of ctrls for AES GCM. Permit setting invocation
8591 field on decrypt and retrieval of invocation field only on encrypt.
8595 * Add HMAC ECC ciphersuites from RFC5289. Include SHA384 PRF support.
8596 As required by RFC5289 these ciphersuites cannot be used if for
8597 versions of TLS earlier than 1.2.
8601 * For FIPS capable OpenSSL interpret a NULL default public key method
8602 as unset and return the appropriate default but do *not* set the default.
8603 This means we can return the appropriate method in applications that
8604 switch between FIPS and non-FIPS modes.
8608 * Redirect HMAC and CMAC operations to FIPS module in FIPS mode. If an
8609 ENGINE is used then we cannot handle that in the FIPS module so we
8610 keep original code iff non-FIPS operations are allowed.
8614 * Add -attime option to openssl utilities.
8616 *Peter Eckersley <pde@eff.org>, Ben Laurie and Steve Henson*
8618 * Redirect DSA and DH operations to FIPS module in FIPS mode.
8622 * Redirect ECDSA and ECDH operations to FIPS module in FIPS mode. Also use
8623 FIPS EC methods unconditionally for now.
8627 * New build option no-ec2m to disable characteristic 2 code.
8631 * Backport libcrypto audit of return value checking from 1.1.0-dev; not
8632 all cases can be covered as some introduce binary incompatibilities.
8636 * Redirect RSA operations to FIPS module including keygen,
8637 encrypt, decrypt, sign and verify. Block use of non FIPS RSA methods.
8641 * Add similar low-level API blocking to ciphers.
8645 * low-level digest APIs are not approved in FIPS mode: any attempt
8646 to use these will cause a fatal error. Applications that *really* want
8647 to use them can use the `private_*` version instead.
8651 * Redirect cipher operations to FIPS module for FIPS builds.
8655 * Redirect digest operations to FIPS module for FIPS builds.
8659 * Update build system to add "fips" flag which will link in fipscanister.o
8660 for static and shared library builds embedding a signature if needed.
8664 * Output TLS supported curves in preference order instead of numerical
8665 order. This is currently hardcoded for the highest order curves first.
8666 This should be configurable so applications can judge speed vs strength.
8670 * Add TLS v1.2 server support for client authentication.
8674 * Add support for FIPS mode in ssl library: disable SSLv3, non-FIPS ciphers
8679 * Functions FIPS_mode_set() and FIPS_mode() which call the underlying
8680 FIPS modules versions.
8684 * Add TLS v1.2 client side support for client authentication. Keep cache
8685 of handshake records longer as we don't know the hash algorithm to use
8686 until after the certificate request message is received.
8690 * Initial TLS v1.2 client support. Add a default signature algorithms
8691 extension including all the algorithms we support. Parse new signature
8692 format in client key exchange. Relax some ECC signing restrictions for
8693 TLS v1.2 as indicated in RFC5246.
8697 * Add server support for TLS v1.2 signature algorithms extension. Switch
8698 to new signature format when needed using client digest preference.
8699 All server ciphersuites should now work correctly in TLS v1.2. No client
8700 support yet and no support for client certificates.
8704 * Initial TLS v1.2 support. Add new SHA256 digest to ssl code, switch
8705 to SHA256 for PRF when using TLS v1.2 and later. Add new SHA256 based
8706 ciphersuites. At present only RSA key exchange ciphersuites work with
8707 TLS v1.2. Add new option for TLS v1.2 replacing the old and obsolete
8708 SSL_OP_PKCS1_CHECK flags with SSL_OP_NO_TLSv1_2. New TLSv1.2 methods
8709 and version checking.
8713 * New option OPENSSL_NO_SSL_INTERN. If an application can be compiled
8714 with this defined it will not be affected by any changes to ssl internal
8715 structures. Add several utility functions to allow openssl application
8716 to work with OPENSSL_NO_SSL_INTERN defined.
8720 * A long standing patch to add support for SRP from EdelWeb (Peter
8721 Sylvester and Christophe Renou) was integrated.
8722 *Christophe Renou <christophe.renou@edelweb.fr>, Peter Sylvester
8723 <peter.sylvester@edelweb.fr>, Tom Wu <tjw@cs.stanford.edu>, and
8726 * Add functions to copy EVP_PKEY_METHOD and retrieve flags and id.
8730 * Permit abbreviated handshakes when renegotiating using the function
8731 SSL_renegotiate_abbreviated().
8733 *Robin Seggelmann <seggelmann@fh-muenster.de>*
8735 * Add call to ENGINE_register_all_complete() to
8736 ENGINE_load_builtin_engines(), so some implementations get used
8737 automatically instead of needing explicit application support.
8741 * Add support for TLS key exporter as described in RFC5705.
8743 *Robin Seggelmann <seggelmann@fh-muenster.de>, Steve Henson*
8745 * Initial TLSv1.1 support. Since TLSv1.1 is very similar to TLS v1.0 only
8746 a few changes are required:
8748 Add SSL_OP_NO_TLSv1_1 flag.
8749 Add TLSv1_1 methods.
8750 Update version checking logic to handle version 1.1.
8751 Add explicit IV handling (ported from DTLS code).
8752 Add command line options to s_client/s_server.
8759 ### Changes between 1.0.0s and 1.0.0t [3 Dec 2015]
8761 * X509_ATTRIBUTE memory leak
8763 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
8764 memory. This structure is used by the PKCS#7 and CMS routines so any
8765 application which reads PKCS#7 or CMS data from untrusted sources is
8766 affected. SSL/TLS is not affected.
8768 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
8774 * Race condition handling PSK identify hint
8776 If PSK identity hints are received by a multi-threaded client then
8777 the values are wrongly updated in the parent SSL_CTX structure. This can
8778 result in a race condition potentially leading to a double free of the
8784 ### Changes between 1.0.0r and 1.0.0s [11 Jun 2015]
8786 * Malformed ECParameters causes infinite loop
8788 When processing an ECParameters structure OpenSSL enters an infinite loop
8789 if the curve specified is over a specially malformed binary polynomial
8792 This can be used to perform denial of service against any
8793 system which processes public keys, certificate requests or
8794 certificates. This includes TLS clients and TLS servers with
8795 client authentication enabled.
8797 This issue was reported to OpenSSL by Joseph Barr-Pixton.
8802 * Exploitable out-of-bounds read in X509_cmp_time
8804 X509_cmp_time does not properly check the length of the ASN1_TIME
8805 string and can read a few bytes out of bounds. In addition,
8806 X509_cmp_time accepts an arbitrary number of fractional seconds in the
8809 An attacker can use this to craft malformed certificates and CRLs of
8810 various sizes and potentially cause a segmentation fault, resulting in
8811 a DoS on applications that verify certificates or CRLs. TLS clients
8812 that verify CRLs are affected. TLS clients and servers with client
8813 authentication enabled may be affected if they use custom verification
8816 This issue was reported to OpenSSL by Robert Swiecki (Google), and
8817 independently by Hanno Böck.
8822 * PKCS7 crash with missing EnvelopedContent
8824 The PKCS#7 parsing code does not handle missing inner EncryptedContent
8825 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
8826 with missing content and trigger a NULL pointer dereference on parsing.
8828 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
8829 structures from untrusted sources are affected. OpenSSL clients and
8830 servers are not affected.
8832 This issue was reported to OpenSSL by Michal Zalewski (Google).
8837 * CMS verify infinite loop with unknown hash function
8839 When verifying a signedData message the CMS code can enter an infinite loop
8840 if presented with an unknown hash function OID. This can be used to perform
8841 denial of service against any system which verifies signedData messages using
8843 This issue was reported to OpenSSL by Johannes Bauer.
8848 * Race condition handling NewSessionTicket
8850 If a NewSessionTicket is received by a multi-threaded client when attempting to
8851 reuse a previous ticket then a race condition can occur potentially leading to
8852 a double free of the ticket data.
8857 ### Changes between 1.0.0q and 1.0.0r [19 Mar 2015]
8859 * Segmentation fault in ASN1_TYPE_cmp fix
8861 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
8862 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
8863 certificate signature algorithm consistency this can be used to crash any
8864 certificate verification operation and exploited in a DoS attack. Any
8865 application which performs certificate verification is vulnerable including
8866 OpenSSL clients and servers which enable client authentication.
8871 * ASN.1 structure reuse memory corruption fix
8873 Reusing a structure in ASN.1 parsing may allow an attacker to cause
8874 memory corruption via an invalid write. Such reuse is and has been
8875 strongly discouraged and is believed to be rare.
8877 Applications that parse structures containing CHOICE or ANY DEFINED BY
8878 components may be affected. Certificate parsing (d2i_X509 and related
8879 functions) are however not affected. OpenSSL clients and servers are
8885 * PKCS7 NULL pointer dereferences fix
8887 The PKCS#7 parsing code does not handle missing outer ContentInfo
8888 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
8889 missing content and trigger a NULL pointer dereference on parsing.
8891 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
8892 otherwise parse PKCS#7 structures from untrusted sources are
8893 affected. OpenSSL clients and servers are not affected.
8895 This issue was reported to OpenSSL by Michal Zalewski (Google).
8900 * DoS via reachable assert in SSLv2 servers fix
8902 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
8903 servers that both support SSLv2 and enable export cipher suites by sending
8904 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
8906 This issue was discovered by Sean Burford (Google) and Emilia Käsper
8907 (OpenSSL development team).
8912 * Use After Free following d2i_ECPrivatekey error fix
8914 A malformed EC private key file consumed via the d2i_ECPrivateKey function
8915 could cause a use after free condition. This, in turn, could cause a double
8916 free in several private key parsing functions (such as d2i_PrivateKey
8917 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
8918 for applications that receive EC private keys from untrusted
8919 sources. This scenario is considered rare.
8921 This issue was discovered by the BoringSSL project and fixed in their
8927 * X509_to_X509_REQ NULL pointer deref fix
8929 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
8930 the certificate key is invalid. This function is rarely used in practice.
8932 This issue was discovered by Brian Carpenter.
8937 * Removed the export ciphers from the DEFAULT ciphers
8941 ### Changes between 1.0.0p and 1.0.0q [15 Jan 2015]
8943 * Build fixes for the Windows and OpenVMS platforms
8945 *Matt Caswell and Richard Levitte*
8947 ### Changes between 1.0.0o and 1.0.0p [8 Jan 2015]
8949 * Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
8950 message can cause a segmentation fault in OpenSSL due to a NULL pointer
8951 dereference. This could lead to a Denial Of Service attack. Thanks to
8952 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
8957 * Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
8958 dtls1_buffer_record function under certain conditions. In particular this
8959 could occur if an attacker sent repeated DTLS records with the same
8960 sequence number but for the next epoch. The memory leak could be exploited
8961 by an attacker in a Denial of Service attack through memory exhaustion.
8962 Thanks to Chris Mueller for reporting this issue.
8967 * Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
8968 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
8969 method would be set to NULL which could later result in a NULL pointer
8970 dereference. Thanks to Frank Schmirler for reporting this issue.
8975 * Abort handshake if server key exchange message is omitted for ephemeral
8978 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
8979 reporting this issue.
8984 * Remove non-export ephemeral RSA code on client and server. This code
8985 violated the TLS standard by allowing the use of temporary RSA keys in
8986 non-export ciphersuites and could be used by a server to effectively
8987 downgrade the RSA key length used to a value smaller than the server
8988 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
8989 INRIA or reporting this issue.
8994 * Fixed issue where DH client certificates are accepted without verification.
8995 An OpenSSL server will accept a DH certificate for client authentication
8996 without the certificate verify message. This effectively allows a client to
8997 authenticate without the use of a private key. This only affects servers
8998 which trust a client certificate authority which issues certificates
8999 containing DH keys: these are extremely rare and hardly ever encountered.
9000 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
9006 * Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
9007 results on some platforms, including x86_64. This bug occurs at random
9008 with a very low probability, and is not known to be exploitable in any
9009 way, though its exact impact is difficult to determine. Thanks to Pieter
9010 Wuille (Blockstream) who reported this issue and also suggested an initial
9011 fix. Further analysis was conducted by the OpenSSL development team and
9012 Adam Langley of Google. The final fix was developed by Andy Polyakov of
9013 the OpenSSL core team.
9018 * Fix various certificate fingerprint issues.
9020 By using non-DER or invalid encodings outside the signed portion of a
9021 certificate the fingerprint can be changed without breaking the signature.
9022 Although no details of the signed portion of the certificate can be changed
9023 this can cause problems with some applications: e.g. those using the
9024 certificate fingerprint for blacklists.
9026 1. Reject signatures with non zero unused bits.
9028 If the BIT STRING containing the signature has non zero unused bits reject
9029 the signature. All current signature algorithms require zero unused bits.
9031 2. Check certificate algorithm consistency.
9033 Check the AlgorithmIdentifier inside TBS matches the one in the
9034 certificate signature. NB: this will result in signature failure
9035 errors for some broken certificates.
9037 Thanks to Konrad Kraszewski from Google for reporting this issue.
9039 3. Check DSA/ECDSA signatures use DER.
9041 Re-encode DSA/ECDSA signatures and compare with the original received
9042 signature. Return an error if there is a mismatch.
9044 This will reject various cases including garbage after signature
9045 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
9046 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
9047 (negative or with leading zeroes).
9049 Further analysis was conducted and fixes were developed by Stephen Henson
9050 of the OpenSSL core team.
9056 ### Changes between 1.0.0n and 1.0.0o [15 Oct 2014]
9058 * Session Ticket Memory Leak.
9060 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
9061 integrity of that ticket is first verified. In the event of a session
9062 ticket integrity check failing, OpenSSL will fail to free memory
9063 causing a memory leak. By sending a large number of invalid session
9064 tickets an attacker could exploit this issue in a Denial Of Service
9070 * Build option no-ssl3 is incomplete.
9072 When OpenSSL is configured with "no-ssl3" as a build option, servers
9073 could accept and complete an SSL 3.0 handshake, and clients could be
9074 configured to send them.
9077 *Akamai and the OpenSSL team*
9079 * Add support for TLS_FALLBACK_SCSV.
9080 Client applications doing fallback retries should call
9081 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
9084 *Adam Langley, Bodo Moeller*
9086 * Add additional DigestInfo checks.
9088 Re-encode DigestInto in DER and check against the original when
9089 verifying RSA signature: this will reject any improperly encoded
9090 DigestInfo structures.
9092 Note: this is a precautionary measure and no attacks are currently known.
9096 ### Changes between 1.0.0m and 1.0.0n [6 Aug 2014]
9098 * OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
9099 to a denial of service attack. A malicious server can crash the client
9100 with a null pointer dereference (read) by specifying an anonymous (EC)DH
9101 ciphersuite and sending carefully crafted handshake messages.
9103 Thanks to Felix Gröbert (Google) for discovering and researching this
9109 * By sending carefully crafted DTLS packets an attacker could cause openssl
9110 to leak memory. This can be exploited through a Denial of Service attack.
9111 Thanks to Adam Langley for discovering and researching this issue.
9116 * An attacker can force openssl to consume large amounts of memory whilst
9117 processing DTLS handshake messages. This can be exploited through a
9118 Denial of Service attack.
9119 Thanks to Adam Langley for discovering and researching this issue.
9124 * An attacker can force an error condition which causes openssl to crash
9125 whilst processing DTLS packets due to memory being freed twice. This
9126 can be exploited through a Denial of Service attack.
9127 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
9133 * If a multithreaded client connects to a malicious server using a resumed
9134 session and the server sends an ec point format extension it could write
9135 up to 255 bytes to freed memory.
9137 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
9143 * A flaw in OBJ_obj2txt may cause pretty printing functions such as
9144 X509_name_oneline, X509_name_print_ex et al. to leak some information
9145 from the stack. Applications may be affected if they echo pretty printing
9146 output to the attacker.
9148 Thanks to Ivan Fratric (Google) for discovering this issue.
9151 *Emilia Käsper, and Steve Henson*
9153 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
9154 for corner cases. (Certain input points at infinity could lead to
9155 bogus results, with non-infinity inputs mapped to infinity too.)
9159 ### Changes between 1.0.0l and 1.0.0m [5 Jun 2014]
9161 * Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
9162 handshake can force the use of weak keying material in OpenSSL
9163 SSL/TLS clients and servers.
9165 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
9166 researching this issue. ([CVE-2014-0224])
9168 *KIKUCHI Masashi, Steve Henson*
9170 * Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
9171 OpenSSL DTLS client the code can be made to recurse eventually crashing
9174 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
9177 *Imre Rad, Steve Henson*
9179 * Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
9180 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
9181 client or server. This is potentially exploitable to run arbitrary
9182 code on a vulnerable client or server.
9184 Thanks to Jüri Aedla for reporting this issue. ([CVE-2014-0195])
9186 *Jüri Aedla, Steve Henson*
9188 * Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
9189 are subject to a denial of service attack.
9191 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
9192 this issue. ([CVE-2014-3470])
9194 *Felix Gröbert, Ivan Fratric, Steve Henson*
9196 * Harmonize version and its documentation. -f flag is used to display
9199 *mancha <mancha1@zoho.com>*
9201 * Fix eckey_priv_encode so it immediately returns an error upon a failure
9202 in i2d_ECPrivateKey.
9204 *mancha <mancha1@zoho.com>*
9206 * Fix some double frees. These are not thought to be exploitable.
9208 *mancha <mancha1@zoho.com>*
9210 * Fix for the attack described in the paper "Recovering OpenSSL
9211 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
9212 by Yuval Yarom and Naomi Benger. Details can be obtained from:
9213 <http://eprint.iacr.org/2014/140>
9215 Thanks to Yuval Yarom and Naomi Benger for discovering this
9216 flaw and to Yuval Yarom for supplying a fix ([CVE-2014-0076])
9218 *Yuval Yarom and Naomi Benger*
9220 ### Changes between 1.0.0k and 1.0.0l [6 Jan 2014]
9222 * Keep original DTLS digest and encryption contexts in retransmission
9223 structures so we can use the previous session parameters if they need
9224 to be resent. ([CVE-2013-6450])
9228 * Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
9229 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
9230 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
9231 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
9232 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
9233 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
9235 *Rob Stradling, Adam Langley*
9237 ### Changes between 1.0.0j and 1.0.0k [5 Feb 2013]
9239 * Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
9241 This addresses the flaw in CBC record processing discovered by
9242 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
9243 at: <http://www.isg.rhul.ac.uk/tls/>
9245 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
9246 Security Group at Royal Holloway, University of London
9247 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
9248 Emilia Käsper for the initial patch.
9251 *Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson*
9253 * Return an error when checking OCSP signatures when key is NULL.
9254 This fixes a DoS attack. ([CVE-2013-0166])
9258 * Call OCSP Stapling callback after ciphersuite has been chosen, so
9259 the right response is stapled. Also change SSL_get_certificate()
9260 so it returns the certificate actually sent.
9261 See <http://rt.openssl.org/Ticket/Display.html?id=2836>.
9262 (This is a backport)
9264 *Rob Stradling <rob.stradling@comodo.com>*
9266 * Fix possible deadlock when decoding public keys.
9270 ### Changes between 1.0.0i and 1.0.0j [10 May 2012]
9272 [NB: OpenSSL 1.0.0i and later 1.0.0 patch levels were released after
9275 * Sanity check record length before skipping explicit IV in DTLS
9278 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
9279 fuzzing as a service testing platform.
9284 * Initialise tkeylen properly when encrypting CMS messages.
9285 Thanks to Solar Designer of Openwall for reporting this issue.
9289 ### Changes between 1.0.0h and 1.0.0i [19 Apr 2012]
9291 * Check for potentially exploitable overflows in asn1_d2i_read_bio
9292 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
9293 in CRYPTO_realloc_clean.
9295 Thanks to Tavis Ormandy, Google Security Team, for discovering this
9296 issue and to Adam Langley <agl@chromium.org> for fixing it.
9299 *Adam Langley (Google), Tavis Ormandy, Google Security Team*
9301 ### Changes between 1.0.0g and 1.0.0h [12 Mar 2012]
9303 * Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness
9304 in CMS and PKCS7 code. When RSA decryption fails use a random key for
9305 content decryption and always return the same error. Note: this attack
9306 needs on average 2^20 messages so it only affects automated senders. The
9307 old behaviour can be re-enabled in the CMS code by setting the
9308 CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where
9309 an MMA defence is not necessary.
9310 Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering
9311 this issue. ([CVE-2012-0884])
9315 * Fix CVE-2011-4619: make sure we really are receiving a
9316 client hello before rejecting multiple SGC restarts. Thanks to
9317 Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug.
9321 ### Changes between 1.0.0f and 1.0.0g [18 Jan 2012]
9323 * Fix for DTLS DoS issue introduced by fix for CVE-2011-4109.
9324 Thanks to Antonio Martin, Enterprise Secure Access Research and
9325 Development, Cisco Systems, Inc. for discovering this bug and
9326 preparing a fix. ([CVE-2012-0050])
9330 ### Changes between 1.0.0e and 1.0.0f [4 Jan 2012]
9332 * Nadhem Alfardan and Kenny Paterson have discovered an extension
9333 of the Vaudenay padding oracle attack on CBC mode encryption
9334 which enables an efficient plaintext recovery attack against
9335 the OpenSSL implementation of DTLS. Their attack exploits timing
9336 differences arising during decryption processing. A research
9337 paper describing this attack can be found at:
9338 <http://www.isg.rhul.ac.uk/~kp/dtls.pdf>
9339 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
9340 Security Group at Royal Holloway, University of London
9341 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
9342 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
9343 for preparing the fix. ([CVE-2011-4108])
9345 *Robin Seggelmann, Michael Tuexen*
9347 * Clear bytes used for block padding of SSL 3.0 records.
9350 *Adam Langley (Google)*
9352 * Only allow one SGC handshake restart for SSL/TLS. Thanks to George
9353 Kadianakis <desnacked@gmail.com> for discovering this issue and
9354 Adam Langley for preparing the fix. ([CVE-2011-4619])
9356 *Adam Langley (Google)*
9358 * Check parameters are not NULL in GOST ENGINE. ([CVE-2012-0027])
9360 *Andrey Kulikov <amdeich@gmail.com>*
9362 * Prevent malformed RFC3779 data triggering an assertion failure.
9363 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
9364 and Rob Austein <sra@hactrn.net> for fixing it. ([CVE-2011-4577])
9366 *Rob Austein <sra@hactrn.net>*
9368 * Improved PRNG seeding for VOS.
9370 *Paul Green <Paul.Green@stratus.com>*
9372 * Fix ssl_ciph.c set-up race.
9374 *Adam Langley (Google)*
9376 * Fix spurious failures in ecdsatest.c.
9378 *Emilia Käsper (Google)*
9380 * Fix the BIO_f_buffer() implementation (which was mixing different
9381 interpretations of the `..._len` fields).
9383 *Adam Langley (Google)*
9385 * Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
9386 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
9387 threads won't reuse the same blinding coefficients.
9389 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
9390 lock to call BN_BLINDING_invert_ex, and avoids one use of
9391 BN_BLINDING_update for each BN_BLINDING structure (previously,
9392 the last update always remained unused).
9394 *Emilia Käsper (Google)*
9396 * In ssl3_clear, preserve s3->init_extra along with s3->rbuf.
9398 *Bob Buckholz (Google)*
9400 ### Changes between 1.0.0d and 1.0.0e [6 Sep 2011]
9402 * Fix bug where CRLs with nextUpdate in the past are sometimes accepted
9403 by initialising X509_STORE_CTX properly. ([CVE-2011-3207])
9405 *Kaspar Brand <ossl@velox.ch>*
9407 * Fix SSL memory handling for (EC)DH ciphersuites, in particular
9408 for multi-threaded use of ECDH. ([CVE-2011-3210])
9410 *Adam Langley (Google)*
9412 * Fix x509_name_ex_d2i memory leak on bad inputs.
9416 * Remove hard coded ecdsaWithSHA1 signature tests in ssl code and check
9417 signature public key algorithm by using OID xref utilities instead.
9418 Before this you could only use some ECC ciphersuites with SHA1 only.
9422 * Add protection against ECDSA timing attacks as mentioned in the paper
9423 by Billy Bob Brumley and Nicola Tuveri, see:
9424 <http://eprint.iacr.org/2011/232.pdf>
9426 *Billy Bob Brumley and Nicola Tuveri*
9428 ### Changes between 1.0.0c and 1.0.0d [8 Feb 2011]
9430 * Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
9432 *Neel Mehta, Adam Langley, Bodo Moeller (Google)*
9434 * Fix bug in string printing code: if *any* escaping is enabled we must
9435 escape the escape character (backslash) or the resulting string is
9440 ### Changes between 1.0.0b and 1.0.0c [2 Dec 2010]
9442 * Disable code workaround for ancient and obsolete Netscape browsers
9443 and servers: an attacker can use it in a ciphersuite downgrade attack.
9444 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
9448 * Fixed J-PAKE implementation error, originally discovered by
9449 Sebastien Martini, further info and confirmation from Stefan
9450 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
9454 ### Changes between 1.0.0a and 1.0.0b [16 Nov 2010]
9456 * Fix extension code to avoid race conditions which can result in a buffer
9457 overrun vulnerability: resumed sessions must not be modified as they can
9458 be shared by multiple threads. CVE-2010-3864
9462 * Fix WIN32 build system to correctly link an ENGINE directory into
9467 ### Changes between 1.0.0 and 1.0.0a [01 Jun 2010]
9469 * Check return value of int_rsa_verify in pkey_rsa_verifyrecover
9472 *Steve Henson, Peter-Michael Hager <hager@dortmund.net>*
9474 ### Changes between 0.9.8n and 1.0.0 [29 Mar 2010]
9476 * Add "missing" function EVP_CIPHER_CTX_copy(). This copies a cipher
9477 context. The operation can be customised via the ctrl mechanism in
9478 case ENGINEs want to include additional functionality.
9482 * Tolerate yet another broken PKCS#8 key format: private key value negative.
9486 * Add new -subject_hash_old and -issuer_hash_old options to x509 utility to
9487 output hashes compatible with older versions of OpenSSL.
9489 *Willy Weisz <weisz@vcpc.univie.ac.at>*
9491 * Fix compression algorithm handling: if resuming a session use the
9492 compression algorithm of the resumed session instead of determining
9493 it from client hello again. Don't allow server to change algorithm.
9497 * Add load_crls() function to commands tidying load_certs() too. Add option
9498 to verify utility to allow additional CRLs to be included.
9502 * Update OCSP request code to permit adding custom headers to the request:
9503 some responders need this.
9507 * The function EVP_PKEY_sign() returns <=0 on error: check return code
9510 *Julia Lawall <julia@diku.dk>*
9512 * Update verify callback code in `apps/s_cb.c` and `apps/verify.c`, it
9513 needlessly dereferenced structures, used obsolete functions and
9514 didn't handle all updated verify codes correctly.
9518 * Disable MD2 in the default configuration.
9522 * In BIO_pop() and BIO_push() use the ctrl argument (which was NULL) to
9523 indicate the initial BIO being pushed or popped. This makes it possible
9524 to determine whether the BIO is the one explicitly called or as a result
9525 of the ctrl being passed down the chain. Fix BIO_pop() and SSL BIOs so
9526 it handles reference counts correctly and doesn't zero out the I/O bio
9527 when it is not being explicitly popped. WARNING: applications which
9528 included workarounds for the old buggy behaviour will need to be modified
9529 or they could free up already freed BIOs.
9533 * Extend the uni2asc/asc2uni => OPENSSL_uni2asc/OPENSSL_asc2uni
9534 renaming to all platforms (within the 0.9.8 branch, this was
9535 done conditionally on Netware platforms to avoid a name clash).
9537 *Guenter <lists@gknw.net>*
9539 * Add ECDHE and PSK support to DTLS.
9541 *Michael Tuexen <tuexen@fh-muenster.de>*
9543 * Add CHECKED_STACK_OF macro to safestack.h, otherwise safestack can't
9548 * Add "missing" function EVP_MD_flags() (without this the only way to
9549 retrieve a digest flags is by accessing the structure directly. Update
9550 `EVP_MD_do_all*()` and `EVP_CIPHER_do_all*()` to include the name a digest
9551 or cipher is registered as in the "from" argument. Print out all
9552 registered digests in the dgst usage message instead of manually
9553 attempting to work them out.
9557 * If no SSLv2 ciphers are used don't use an SSLv2 compatible client hello:
9558 this allows the use of compression and extensions. Change default cipher
9559 string to remove SSLv2 ciphersuites. This effectively avoids ancient SSLv2
9560 by default unless an application cipher string requests it.
9564 * Alter match criteria in PKCS12_parse(). It used to try to use local
9565 key ids to find matching certificates and keys but some PKCS#12 files
9566 don't follow the (somewhat unwritten) rules and this strategy fails.
9567 Now just gather all certificates together and the first private key
9568 then look for the first certificate that matches the key.
9572 * Support use of registered digest and cipher names for dgst and cipher
9573 commands instead of having to add each one as a special case. So now
9580 openssl dgst -sha256 foo
9582 and this works for ENGINE based algorithms too.
9586 * Update Gost ENGINE to support parameter files.
9588 *Victor B. Wagner <vitus@cryptocom.ru>*
9590 * Support GeneralizedTime in ca utility.
9592 *Oliver Martin <oliver@volatilevoid.net>, Steve Henson*
9594 * Enhance the hash format used for certificate directory links. The new
9595 form uses the canonical encoding (meaning equivalent names will work
9596 even if they aren't identical) and uses SHA1 instead of MD5. This form
9597 is incompatible with the older format and as a result c_rehash should
9598 be used to rebuild symbolic links.
9602 * Make PKCS#8 the default write format for private keys, replacing the
9603 traditional format. This form is standardised, more secure and doesn't
9604 include an implicit MD5 dependency.
9608 * Add a $gcc_devteam_warn option to Configure. The idea is that any code
9609 committed to OpenSSL should pass this lot as a minimum.
9613 * Add session ticket override functionality for use by EAP-FAST.
9615 *Jouni Malinen <j@w1.fi>*
9617 * Modify HMAC functions to return a value. Since these can be implemented
9618 in an ENGINE errors can occur.
9622 * Type-checked OBJ_bsearch_ex.
9626 * Type-checked OBJ_bsearch. Also some constification necessitated
9627 by type-checking. Still to come: TXT_DB, bsearch(?),
9628 OBJ_bsearch_ex, qsort, CRYPTO_EX_DATA, ASN1_VALUE, ASN1_STRING,
9633 * New function OPENSSL_gmtime_adj() to add a specific number of days and
9634 seconds to a tm structure directly, instead of going through OS
9635 specific date routines. This avoids any issues with OS routines such
9636 as the year 2038 bug. New `*_adj()` functions for ASN1 time structures
9637 and X509_time_adj_ex() to cover the extended range. The existing
9638 X509_time_adj() is still usable and will no longer have any date issues.
9642 * Delta CRL support. New use deltas option which will attempt to locate
9643 and search any appropriate delta CRLs available.
9645 This work was sponsored by Google.
9649 * Support for CRLs partitioned by reason code. Reorganise CRL processing
9650 code and add additional score elements. Validate alternate CRL paths
9651 as part of the CRL checking and indicate a new error "CRL path validation
9652 error" in this case. Applications wanting additional details can use
9653 the verify callback and check the new "parent" field. If this is not
9654 NULL CRL path validation is taking place. Existing applications won't
9655 see this because it requires extended CRL support which is off by
9658 This work was sponsored by Google.
9662 * Support for freshest CRL extension.
9664 This work was sponsored by Google.
9668 * Initial indirect CRL support. Currently only supported in the CRLs
9669 passed directly and not via lookup. Process certificate issuer
9670 CRL entry extension and lookup CRL entries by bother issuer name
9671 and serial number. Check and process CRL issuer entry in IDP extension.
9673 This work was sponsored by Google.
9677 * Add support for distinct certificate and CRL paths. The CRL issuer
9678 certificate is validated separately in this case. Only enabled if
9679 an extended CRL support flag is set: this flag will enable additional
9680 CRL functionality in future.
9682 This work was sponsored by Google.
9686 * Add support for policy mappings extension.
9688 This work was sponsored by Google.
9692 * Fixes to pathlength constraint, self issued certificate handling,
9693 policy processing to align with RFC3280 and PKITS tests.
9695 This work was sponsored by Google.
9699 * Support for name constraints certificate extension. DN, email, DNS
9700 and URI types are currently supported.
9702 This work was sponsored by Google.
9706 * To cater for systems that provide a pointer-based thread ID rather
9707 than numeric, deprecate the current numeric thread ID mechanism and
9708 replace it with a structure and associated callback type. This
9709 mechanism allows a numeric "hash" to be extracted from a thread ID in
9710 either case, and on platforms where pointers are larger than 'long',
9711 mixing is done to help ensure the numeric 'hash' is usable even if it
9712 can't be guaranteed unique. The default mechanism is to use "&errno"
9713 as a pointer-based thread ID to distinguish between threads.
9715 Applications that want to provide their own thread IDs should now use
9716 CRYPTO_THREADID_set_callback() to register a callback that will call
9717 either CRYPTO_THREADID_set_numeric() or CRYPTO_THREADID_set_pointer().
9719 Note that ERR_remove_state() is now deprecated, because it is tied
9720 to the assumption that thread IDs are numeric. ERR_remove_state(0)
9721 to free the current thread's error state should be replaced by
9722 ERR_remove_thread_state(NULL).
9724 (This new approach replaces the functions CRYPTO_set_idptr_callback(),
9725 CRYPTO_get_idptr_callback(), and CRYPTO_thread_idptr() that existed in
9726 OpenSSL 0.9.9-dev between June 2006 and August 2008. Also, if an
9727 application was previously providing a numeric thread callback that
9728 was inappropriate for distinguishing threads, then uniqueness might
9729 have been obtained with &errno that happened immediately in the
9730 intermediate development versions of OpenSSL; this is no longer the
9731 case, the numeric thread callback will now override the automatic use
9734 *Geoff Thorpe, with help from Bodo Moeller*
9736 * Initial support for different CRL issuing certificates. This covers a
9737 simple case where the self issued certificates in the chain exist and
9738 the real CRL issuer is higher in the existing chain.
9740 This work was sponsored by Google.
9744 * Removed effectively defunct crypto/store from the build.
9748 * Revamp of STACK to provide stronger type-checking. Still to come:
9749 TXT_DB, bsearch(?), OBJ_bsearch, qsort, CRYPTO_EX_DATA, ASN1_VALUE,
9750 ASN1_STRING, CONF_VALUE.
9754 * Add a new SSL_MODE_RELEASE_BUFFERS mode flag to release unused buffer
9755 RAM on SSL connections. This option can save about 34k per idle SSL.
9759 * Revamp of LHASH to provide stronger type-checking. Still to come:
9760 STACK, TXT_DB, bsearch, qsort.
9764 * Initial support for Cryptographic Message Syntax (aka CMS) based
9765 on RFC3850, RFC3851 and RFC3852. New cms directory and cms utility,
9766 support for data, signedData, compressedData, digestedData and
9767 encryptedData, envelopedData types included. Scripts to check against
9768 RFC4134 examples draft and interop and consistency checks of many
9769 content types and variants.
9773 * Add options to enc utility to support use of zlib compression BIO.
9777 * Extend mk1mf to support importing of options and assembly language
9778 files from Configure script, currently only included in VC-WIN32.
9779 The assembly language rules can now optionally generate the source
9780 files from the associated perl scripts.
9784 * Implement remaining functionality needed to support GOST ciphersuites.
9785 Interop testing has been performed using CryptoPro implementations.
9787 *Victor B. Wagner <vitus@cryptocom.ru>*
9789 * s390x assembler pack.
9793 * ARMv4 assembler pack. ARMv4 refers to v4 and later ISA, not CPU
9798 * Implement Opaque PRF Input TLS extension as specified in
9799 draft-rescorla-tls-opaque-prf-input-00.txt. Since this is not an
9800 official specification yet and no extension type assignment by
9801 IANA exists, this extension (for now) will have to be explicitly
9802 enabled when building OpenSSL by providing the extension number
9803 to use. For example, specify an option
9805 -DTLSEXT_TYPE_opaque_prf_input=0x9527
9807 to the "config" or "Configure" script to enable the extension,
9808 assuming extension number 0x9527 (which is a completely arbitrary
9809 and unofficial assignment based on the MD5 hash of the Internet
9810 Draft). Note that by doing so, you potentially lose
9811 interoperability with other TLS implementations since these might
9812 be using the same extension number for other purposes.
9814 SSL_set_tlsext_opaque_prf_input(ssl, src, len) is used to set the
9815 opaque PRF input value to use in the handshake. This will create
9816 an internal copy of the length-'len' string at 'src', and will
9817 return non-zero for success.
9819 To get more control and flexibility, provide a callback function
9822 SSL_CTX_set_tlsext_opaque_prf_input_callback(ctx, cb)
9823 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg(ctx, arg)
9827 int (*cb)(SSL *, void *peerinput, size_t len, void *arg);
9830 Callback function 'cb' will be called in handshakes, and is
9831 expected to use SSL_set_tlsext_opaque_prf_input() as appropriate.
9832 Argument 'arg' is for application purposes (the value as given to
9833 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg() will directly
9834 be provided to the callback function). The callback function
9835 has to return non-zero to report success: usually 1 to use opaque
9836 PRF input just if possible, or 2 to enforce use of the opaque PRF
9837 input. In the latter case, the library will abort the handshake
9838 if opaque PRF input is not successfully negotiated.
9840 Arguments 'peerinput' and 'len' given to the callback function
9841 will always be NULL and 0 in the case of a client. A server will
9842 see the client's opaque PRF input through these variables if
9843 available (NULL and 0 otherwise). Note that if the server
9844 provides an opaque PRF input, the length must be the same as the
9845 length of the client's opaque PRF input.
9847 Note that the callback function will only be called when creating
9848 a new session (session resumption can resume whatever was
9849 previously negotiated), and will not be called in SSL 2.0
9850 handshakes; thus, SSL_CTX_set_options(ctx, SSL_OP_NO_SSLv2) or
9851 SSL_set_options(ssl, SSL_OP_NO_SSLv2) is especially recommended
9852 for applications that need to enforce opaque PRF input.
9856 * Update ssl code to support digests other than SHA1+MD5 for handshake
9859 *Victor B. Wagner <vitus@cryptocom.ru>*
9861 * Add RFC4507 support to OpenSSL. This includes the corrections in
9862 RFC4507bis. The encrypted ticket format is an encrypted encoded
9863 SSL_SESSION structure, that way new session features are automatically
9866 If a client application caches session in an SSL_SESSION structure
9867 support is transparent because tickets are now stored in the encoded
9870 The SSL_CTX structure automatically generates keys for ticket
9871 protection in servers so again support should be possible
9872 with no application modification.
9874 If a client or server wishes to disable RFC4507 support then the option
9875 SSL_OP_NO_TICKET can be set.
9877 Add a TLS extension debugging callback to allow the contents of any client
9878 or server extensions to be examined.
9880 This work was sponsored by Google.
9884 * Final changes to avoid use of pointer pointer casts in OpenSSL.
9885 OpenSSL should now compile cleanly on gcc 4.2
9887 *Peter Hartley <pdh@utter.chaos.org.uk>, Steve Henson*
9889 * Update SSL library to use new EVP_PKEY MAC API. Include generic MAC
9890 support including streaming MAC support: this is required for GOST
9891 ciphersuite support.
9893 *Victor B. Wagner <vitus@cryptocom.ru>, Steve Henson*
9895 * Add option -stream to use PKCS#7 streaming in smime utility. New
9896 function i2d_PKCS7_bio_stream() and PEM_write_PKCS7_bio_stream()
9897 to output in BER and PEM format.
9901 * Experimental support for use of HMAC via EVP_PKEY interface. This
9902 allows HMAC to be handled via the `EVP_DigestSign*()` interface. The
9903 EVP_PKEY "key" in this case is the HMAC key, potentially allowing
9904 ENGINE support for HMAC keys which are unextractable. New -mac and
9905 -macopt options to dgst utility.
9909 * New option -sigopt to dgst utility. Update dgst to use
9910 `EVP_Digest{Sign,Verify}*`. These two changes make it possible to use
9911 alternative signing parameters such as X9.31 or PSS in the dgst
9916 * Change ssl_cipher_apply_rule(), the internal function that does
9917 the work each time a ciphersuite string requests enabling
9918 ("foo+bar"), moving ("+foo+bar"), disabling ("-foo+bar", or
9919 removing ("!foo+bar") a class of ciphersuites: Now it maintains
9920 the order of disabled ciphersuites such that those ciphersuites
9921 that most recently went from enabled to disabled not only stay
9922 in order with respect to each other, but also have higher priority
9923 than other disabled ciphersuites the next time ciphersuites are
9926 This means that you can now say, e.g., "PSK:-PSK:HIGH" to enable
9927 the same ciphersuites as with "HIGH" alone, but in a specific
9928 order where the PSK ciphersuites come first (since they are the
9929 most recently disabled ciphersuites when "HIGH" is parsed).
9931 Also, change ssl_create_cipher_list() (using this new
9932 functionality) such that between otherwise identical
9933 ciphersuites, ephemeral ECDH is preferred over ephemeral DH in
9938 * Change ssl_create_cipher_list() so that it automatically
9939 arranges the ciphersuites in reasonable order before starting
9940 to process the rule string. Thus, the definition for "DEFAULT"
9941 (SSL_DEFAULT_CIPHER_LIST) now is just "ALL:!aNULL:!eNULL", but
9942 remains equivalent to `"AES:ALL:!aNULL:!eNULL:+aECDH:+kRSA:+RC4:@STRENGTH"`.
9943 This makes it much easier to arrive at a reasonable default order
9944 in applications for which anonymous ciphers are OK (meaning
9945 that you can't actually use DEFAULT).
9947 *Bodo Moeller; suggested by Victor Duchovni*
9949 * Split the SSL/TLS algorithm mask (as used for ciphersuite string
9950 processing) into multiple integers instead of setting
9951 "SSL_MKEY_MASK" bits, "SSL_AUTH_MASK" bits, "SSL_ENC_MASK",
9952 "SSL_MAC_MASK", and "SSL_SSL_MASK" bits all in a single integer.
9953 (These masks as well as the individual bit definitions are hidden
9954 away into the non-exported interface ssl/ssl_locl.h, so this
9955 change to the definition of the SSL_CIPHER structure shouldn't
9956 affect applications.) This give us more bits for each of these
9957 categories, so there is no longer a need to coagulate AES128 and
9958 AES256 into a single algorithm bit, and to coagulate Camellia128
9959 and Camellia256 into a single algorithm bit, which has led to all
9962 Thus, among other things, the kludge introduced in 0.9.7m and
9963 0.9.8e for masking out AES256 independently of AES128 or masking
9964 out Camellia256 independently of AES256 is not needed here in 0.9.9.
9966 With the change, we also introduce new ciphersuite aliases that
9967 so far were missing: "AES128", "AES256", "CAMELLIA128", and
9972 * Add support for dsa-with-SHA224 and dsa-with-SHA256.
9973 Use the leftmost N bytes of the signature input if the input is
9974 larger than the prime q (with N being the size in bytes of q).
9978 * Very *very* experimental PKCS#7 streaming encoder support. Nothing uses
9979 it yet and it is largely untested.
9983 * Add support for the ecdsa-with-SHA224/256/384/512 signature types.
9987 * Initial incomplete changes to avoid need for function casts in OpenSSL
9988 some compilers (gcc 4.2 and later) reject their use. Safestack is
9989 reimplemented. Update ASN1 to avoid use of legacy functions.
9993 * Win32/64 targets are linked with Winsock2.
9997 * Add an X509_CRL_METHOD structure to allow CRL processing to be redirected
9998 to external functions. This can be used to increase CRL handling
9999 efficiency especially when CRLs are very large by (for example) storing
10000 the CRL revoked certificates in a database.
10004 * Overhaul of by_dir code. Add support for dynamic loading of CRLs so
10005 new CRLs added to a directory can be used. New command line option
10006 -verify_return_error to s_client and s_server. This causes real errors
10007 to be returned by the verify callback instead of carrying on no matter
10008 what. This reflects the way a "real world" verify callback would behave.
10012 * GOST engine, supporting several GOST algorithms and public key formats.
10013 Kindly donated by Cryptocom.
10017 * Partial support for Issuing Distribution Point CRL extension. CRLs
10018 partitioned by DP are handled but no indirect CRL or reason partitioning
10019 (yet). Complete overhaul of CRL handling: now the most suitable CRL is
10020 selected via a scoring technique which handles IDP and AKID in CRLs.
10024 * New X509_STORE_CTX callbacks lookup_crls() and lookup_certs() which
10025 will ultimately be used for all verify operations: this will remove the
10026 X509_STORE dependency on certificate verification and allow alternative
10027 lookup methods. X509_STORE based implementations of these two callbacks.
10031 * Allow multiple CRLs to exist in an X509_STORE with matching issuer names.
10032 Modify get_crl() to find a valid (unexpired) CRL if possible.
10036 * New function X509_CRL_match() to check if two CRLs are identical. Normally
10037 this would be called X509_CRL_cmp() but that name is already used by
10038 a function that just compares CRL issuer names. Cache several CRL
10039 extensions in X509_CRL structure and cache CRLDP in X509.
10043 * Store a "canonical" representation of X509_NAME structure (ASN1 Name)
10044 this maps equivalent X509_NAME structures into a consistent structure.
10045 Name comparison can then be performed rapidly using memcmp().
10049 * Non-blocking OCSP request processing. Add -timeout option to ocsp
10054 * Allow digests to supply their own micalg string for S/MIME type using
10055 the ctrl EVP_MD_CTRL_MICALG.
10059 * During PKCS7 signing pass the PKCS7 SignerInfo structure to the
10060 EVP_PKEY_METHOD before and after signing via the EVP_PKEY_CTRL_PKCS7_SIGN
10061 ctrl. It can then customise the structure before and/or after signing
10066 * New function OBJ_add_sigid() to allow application defined signature OIDs
10067 to be added to OpenSSLs internal tables. New function OBJ_sigid_free()
10068 to free up any added signature OIDs.
10072 * New functions EVP_CIPHER_do_all(), EVP_CIPHER_do_all_sorted(),
10073 EVP_MD_do_all() and EVP_MD_do_all_sorted() to enumerate internal
10074 digest and cipher tables. New options added to openssl utility:
10075 list-message-digest-algorithms and list-cipher-algorithms.
10079 * Change the array representation of binary polynomials: the list
10080 of degrees of non-zero coefficients is now terminated with -1.
10081 Previously it was terminated with 0, which was also part of the
10082 value; thus, the array representation was not applicable to
10083 polynomials where t^0 has coefficient zero. This change makes
10084 the array representation useful in a more general context.
10088 * Various modifications and fixes to SSL/TLS cipher string
10089 handling. For ECC, the code now distinguishes between fixed ECDH
10090 with RSA certificates on the one hand and with ECDSA certificates
10091 on the other hand, since these are separate ciphersuites. The
10092 unused code for Fortezza ciphersuites has been removed.
10094 For consistency with EDH, ephemeral ECDH is now called "EECDH"
10095 (not "ECDHE"). For consistency with the code for DH
10096 certificates, use of ECDH certificates is now considered ECDH
10097 authentication, not RSA or ECDSA authentication (the latter is
10098 merely the CA's signing algorithm and not actively used in the
10101 The temporary ciphersuite alias "ECCdraft" is no longer
10102 available, and ECC ciphersuites are no longer excluded from "ALL"
10103 and "DEFAULT". The following aliases now exist for RFC 4492
10104 ciphersuites, most of these by analogy with the DH case:
10106 kECDHr - ECDH cert, signed with RSA
10107 kECDHe - ECDH cert, signed with ECDSA
10108 kECDH - ECDH cert (signed with either RSA or ECDSA)
10109 kEECDH - ephemeral ECDH
10110 ECDH - ECDH cert or ephemeral ECDH
10113 aECDSA - ECDSA cert
10116 AECDH - anonymous ECDH
10117 EECDH - non-anonymous ephemeral ECDH (equivalent to "kEECDH:-AECDH")
10121 * Add additional S/MIME capabilities for AES and GOST ciphers if supported.
10122 Use correct micalg parameters depending on digest(s) in signed message.
10126 * Add engine support for EVP_PKEY_ASN1_METHOD. Add functions to process
10127 an ENGINE asn1 method. Support ENGINE lookups in the ASN1 code.
10131 * Initial engine support for EVP_PKEY_METHOD. New functions to permit
10132 an engine to register a method. Add ENGINE lookups for methods and
10133 functional reference processing.
10137 * New functions `EVP_Digest{Sign,Verify)*`. These are enhanced versions of
10138 `EVP_{Sign,Verify}*` which allow an application to customise the signature
10143 * New -resign option to smime utility. This adds one or more signers
10144 to an existing PKCS#7 signedData structure. Also -md option to use an
10145 alternative message digest algorithm for signing.
10149 * Tidy up PKCS#7 routines and add new functions to make it easier to
10150 create PKCS7 structures containing multiple signers. Update smime
10151 application to support multiple signers.
10155 * New -macalg option to pkcs12 utility to allow setting of an alternative
10160 * Initial support for PKCS#5 v2.0 PRFs other than default SHA1 HMAC.
10161 Reorganize PBE internals to lookup from a static table using NIDs,
10162 add support for HMAC PBE OID translation. Add a EVP_CIPHER ctrl:
10163 EVP_CTRL_PBE_PRF_NID this allows a cipher to specify an alternative
10164 PRF which will be automatically used with PBES2.
10168 * Replace the algorithm specific calls to generate keys in "req" with the
10173 * Update PKCS#7 enveloped data routines to use new API. This is now
10174 supported by any public key method supporting the encrypt operation. A
10175 ctrl is added to allow the public key algorithm to examine or modify
10176 the PKCS#7 RecipientInfo structure if it needs to: for RSA this is
10181 * Add a ctrl to asn1 method to allow a public key algorithm to express
10182 a default digest type to use. In most cases this will be SHA1 but some
10183 algorithms (such as GOST) need to specify an alternative digest. The
10184 return value indicates how strong the preference is 1 means optional and
10185 2 is mandatory (that is it is the only supported type). Modify
10186 ASN1_item_sign() to accept a NULL digest argument to indicate it should
10187 use the default md. Update openssl utilities to use the default digest
10188 type for signing if it is not explicitly indicated.
10192 * Use OID cross reference table in ASN1_sign() and ASN1_verify(). New
10193 EVP_MD flag EVP_MD_FLAG_PKEY_METHOD_SIGNATURE. This uses the relevant
10194 signing method from the key type. This effectively removes the link
10195 between digests and public key types.
10199 * Add an OID cross reference table and utility functions. Its purpose is to
10200 translate between signature OIDs such as SHA1WithrsaEncryption and SHA1,
10201 rsaEncryption. This will allow some of the algorithm specific hackery
10202 needed to use the correct OID to be removed.
10206 * Remove algorithm specific dependencies when setting PKCS7_SIGNER_INFO
10207 structures for PKCS7_sign(). They are now set up by the relevant public
10212 * Add provisional EC pkey method with support for ECDSA and ECDH.
10216 * Add support for key derivation (agreement) in the API, DH method and
10221 * Add DSA pkey method and DH pkey methods, extend DH ASN1 method to support
10222 public and private key formats. As a side effect these add additional
10223 command line functionality not previously available: DSA signatures can be
10224 generated and verified using pkeyutl and DH key support and generation in
10231 *Oliver Tappe <zooey@hirschkaefer.de>*
10233 * New make target "install_html_docs" installs HTML renditions of the
10236 *Oliver Tappe <zooey@hirschkaefer.de>*
10238 * New utility "genpkey" this is analogous to "genrsa" etc except it can
10239 generate keys for any algorithm. Extend and update EVP_PKEY_METHOD to
10240 support key and parameter generation and add initial key generation
10241 functionality for RSA.
10245 * Add functions for main EVP_PKEY_method operations. The undocumented
10246 functions `EVP_PKEY_{encrypt,decrypt}` have been renamed to
10247 `EVP_PKEY_{encrypt,decrypt}_old`.
10251 * Initial definitions for EVP_PKEY_METHOD. This will be a high level public
10252 key API, doesn't do much yet.
10256 * New function EVP_PKEY_asn1_get0_info() to retrieve information about
10257 public key algorithms. New option to openssl utility:
10258 "list-public-key-algorithms" to print out info.
10262 * Implement the Supported Elliptic Curves Extension for
10263 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
10267 * Don't free up OIDs in OBJ_cleanup() if they are in use by EVP_MD or
10268 EVP_CIPHER structures to avoid later problems in EVP_cleanup().
10272 * New utilities pkey and pkeyparam. These are similar to algorithm specific
10273 utilities such as rsa, dsa, dsaparam etc except they process any key
10278 * Transfer public key printing routines to EVP_PKEY_ASN1_METHOD. New
10279 functions EVP_PKEY_print_public(), EVP_PKEY_print_private(),
10280 EVP_PKEY_print_param() to print public key data from an EVP_PKEY
10285 * Initial support for pluggable public key ASN1.
10286 De-spaghettify the public key ASN1 handling. Move public and private
10287 key ASN1 handling to a new EVP_PKEY_ASN1_METHOD structure. Relocate
10288 algorithm specific handling to a single module within the relevant
10289 algorithm directory. Add functions to allow (near) opaque processing
10290 of public and private key structures.
10294 * Implement the Supported Point Formats Extension for
10295 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
10299 * Add initial support for RFC 4279 PSK TLS ciphersuites. Add members
10300 for the psk identity [hint] and the psk callback functions to the
10301 SSL_SESSION, SSL and SSL_CTX structure.
10304 PSK-RC4-SHA, PSK-3DES-EDE-CBC-SHA, PSK-AES128-CBC-SHA,
10308 SSL_CTX_use_psk_identity_hint
10309 SSL_get_psk_identity_hint
10310 SSL_get_psk_identity
10311 SSL_use_psk_identity_hint
10313 *Mika Kousa and Pasi Eronen of Nokia Corporation*
10315 * Add RFC 3161 compliant time stamp request creation, response generation
10316 and response verification functionality.
10318 *Zoltán Glózik <zglozik@opentsa.org>, The OpenTSA Project*
10320 * Add initial support for TLS extensions, specifically for the server_name
10321 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
10322 have new members for a hostname. The SSL data structure has an
10323 additional member `SSL_CTX *initial_ctx` so that new sessions can be
10324 stored in that context to allow for session resumption, even after the
10325 SSL has been switched to a new SSL_CTX in reaction to a client's
10326 server_name extension.
10328 New functions (subject to change):
10330 SSL_get_servername()
10331 SSL_get_servername_type()
10334 New CTRL codes and macros (subject to change):
10336 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
10337 - SSL_CTX_set_tlsext_servername_callback()
10338 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
10339 - SSL_CTX_set_tlsext_servername_arg()
10340 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
10342 openssl s_client has a new '-servername ...' option.
10344 openssl s_server has new options '-servername_host ...', '-cert2 ...',
10345 '-key2 ...', '-servername_fatal' (subject to change). This allows
10346 testing the HostName extension for a specific single hostname ('-cert'
10347 and '-key' remain fallbacks for handshakes without HostName
10348 negotiation). If the unrecognized_name alert has to be sent, this by
10349 default is a warning; it becomes fatal with the '-servername_fatal'
10352 *Peter Sylvester, Remy Allais, Christophe Renou*
10354 * Whirlpool hash implementation is added.
10358 * BIGNUM code on 64-bit SPARCv9 targets is switched from bn(64,64) to
10359 bn(64,32). Because of instruction set limitations it doesn't have
10360 any negative impact on performance. This was done mostly in order
10361 to make it possible to share assembler modules, such as bn_mul_mont
10362 implementations, between 32- and 64-bit builds without hassle.
10366 * Move code previously exiled into file crypto/ec/ec2_smpt.c
10367 to ec2_smpl.c, and no longer require the OPENSSL_EC_BIN_PT_COMP
10372 * New candidate for BIGNUM assembler implementation, bn_mul_mont,
10373 dedicated Montgomery multiplication procedure, is introduced.
10374 BN_MONT_CTX is modified to allow bn_mul_mont to reach for higher
10375 "64-bit" performance on certain 32-bit targets.
10379 * New option SSL_OP_NO_COMP to disable use of compression selectively
10380 in SSL structures. New SSL ctrl to set maximum send fragment size.
10381 Save memory by setting the I/O buffer sizes dynamically instead of
10382 using the maximum available value.
10386 * New option -V for 'openssl ciphers'. This prints the ciphersuite code
10387 in addition to the text details.
10391 * Very, very preliminary EXPERIMENTAL support for printing of general
10392 ASN1 structures. This currently produces rather ugly output and doesn't
10393 handle several customised structures at all.
10397 * Integrated support for PVK file format and some related formats such
10398 as MS PUBLICKEYBLOB and PRIVATEKEYBLOB. Command line switches to support
10399 these in the 'rsa' and 'dsa' utilities.
10403 * Support for PKCS#1 RSAPublicKey format on rsa utility command line.
10407 * Remove the ancient ASN1_METHOD code. This was only ever used in one
10408 place for the (very old) "NETSCAPE" format certificates which are now
10409 handled using new ASN1 code equivalents.
10413 * Let the TLSv1_method() etc. functions return a 'const' SSL_METHOD
10414 pointer and make the SSL_METHOD parameter in SSL_CTX_new,
10415 SSL_CTX_set_ssl_version and SSL_set_ssl_method 'const'.
10419 * Modify CRL distribution points extension code to print out previously
10420 unsupported fields. Enhance extension setting code to allow setting of
10425 * Add print and set support for Issuing Distribution Point CRL extension.
10429 * Change 'Configure' script to enable Camellia by default.
10436 ### Changes between 0.9.8m and 0.9.8n [24 Mar 2010]
10438 * When rejecting SSL/TLS records due to an incorrect version number, never
10439 update s->server with a new major version number. As of
10440 - OpenSSL 0.9.8m if 'short' is a 16-bit type,
10441 - OpenSSL 0.9.8f if 'short' is longer than 16 bits,
10442 the previous behavior could result in a read attempt at NULL when
10443 receiving specific incorrect SSL/TLS records once record payload
10444 protection is active. ([CVE-2010-0740])
10446 *Bodo Moeller, Adam Langley <agl@chromium.org>*
10448 * Fix for CVE-2010-0433 where some kerberos enabled versions of OpenSSL
10449 could be crashed if the relevant tables were not present (e.g. chrooted).
10451 *Tomas Hoger <thoger@redhat.com>*
10453 ### Changes between 0.9.8l and 0.9.8m [25 Feb 2010]
10455 * Always check bn_wexpand() return values for failure. ([CVE-2009-3245])
10457 *Martin Olsson, Neel Mehta*
10459 * Fix X509_STORE locking: Every 'objs' access requires a lock (to
10460 accommodate for stack sorting, always a write lock!).
10464 * On some versions of WIN32 Heap32Next is very slow. This can cause
10465 excessive delays in the RAND_poll(): over a minute. As a workaround
10466 include a time check in the inner Heap32Next loop too.
10470 * The code that handled flushing of data in SSL/TLS originally used the
10471 BIO_CTRL_INFO ctrl to see if any data was pending first. This caused
10472 the problem outlined in PR#1949. The fix suggested there however can
10473 trigger problems with buggy BIO_CTRL_WPENDING (e.g. some versions
10474 of Apache). So instead simplify the code to flush unconditionally.
10475 This should be fine since flushing with no data to flush is a no op.
10479 * Handle TLS versions 2.0 and later properly and correctly use the
10480 highest version of TLS/SSL supported. Although TLS >= 2.0 is some way
10481 off ancient servers have a habit of sticking around for a while...
10485 * Modify compression code so it frees up structures without using the
10486 ex_data callbacks. This works around a problem where some applications
10487 call CRYPTO_cleanup_all_ex_data() before application exit (e.g. when
10488 restarting) then use compression (e.g. SSL with compression) later.
10489 This results in significant per-connection memory leaks and
10490 has caused some security issues including CVE-2008-1678 and
10495 * Constify crypto/cast (i.e., <openssl/cast.h>): a CAST_KEY doesn't
10496 change when encrypting or decrypting.
10500 * Add option SSL_OP_LEGACY_SERVER_CONNECT which will allow clients to
10501 connect and renegotiate with servers which do not support RI.
10502 Until RI is more widely deployed this option is enabled by default.
10506 * Add "missing" ssl ctrls to clear options and mode.
10510 * If client attempts to renegotiate and doesn't support RI respond with
10511 a no_renegotiation alert as required by RFC5746. Some renegotiating
10512 TLS clients will continue a connection gracefully when they receive
10513 the alert. Unfortunately OpenSSL mishandled this alert and would hang
10514 waiting for a server hello which it will never receive. Now we treat a
10515 received no_renegotiation alert as a fatal error. This is because
10516 applications requesting a renegotiation might well expect it to succeed
10517 and would have no code in place to handle the server denying it so the
10518 only safe thing to do is to terminate the connection.
10522 * Add ctrl macro SSL_get_secure_renegotiation_support() which returns 1 if
10523 peer supports secure renegotiation and 0 otherwise. Print out peer
10524 renegotiation support in s_client/s_server.
10528 * Replace the highly broken and deprecated SPKAC certification method with
10529 the updated NID creation version. This should correctly handle UTF8.
10533 * Implement RFC5746. Re-enable renegotiation but require the extension
10534 as needed. Unfortunately, SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION
10535 turns out to be a bad idea. It has been replaced by
10536 SSL_OP_ALLOW_UNSAFE_LEGACY_RENEGOTIATION which can be set with
10537 SSL_CTX_set_options(). This is really not recommended unless you
10538 know what you are doing.
10540 *Eric Rescorla <ekr@networkresonance.com>, Ben Laurie, Steve Henson*
10542 * Fixes to stateless session resumption handling. Use initial_ctx when
10543 issuing and attempting to decrypt tickets in case it has changed during
10544 servername handling. Use a non-zero length session ID when attempting
10545 stateless session resumption: this makes it possible to determine if
10546 a resumption has occurred immediately after receiving server hello
10547 (several places in OpenSSL subtly assume this) instead of later in
10552 * The functions ENGINE_ctrl(), OPENSSL_isservice(),
10553 CMS_get1_RecipientRequest() and RAND_bytes() can return <=0 on error
10554 fixes for a few places where the return code is not checked
10557 *Julia Lawall <julia@diku.dk>*
10559 * Add --strict-warnings option to Configure script to include devteam
10560 warnings in other configurations.
10564 * Add support for --libdir option and LIBDIR variable in makefiles. This
10565 makes it possible to install openssl libraries in locations which
10566 have names other than "lib", for example "/usr/lib64" which some
10569 *Steve Henson, based on patch from Jeremy Utley*
10571 * Don't allow the use of leading 0x80 in OIDs. This is a violation of
10572 X690 8.9.12 and can produce some misleading textual output of OIDs.
10574 *Steve Henson, reported by Dan Kaminsky*
10576 * Delete MD2 from algorithm tables. This follows the recommendation in
10577 several standards that it is not used in new applications due to
10578 several cryptographic weaknesses. For binary compatibility reasons
10579 the MD2 API is still compiled in by default.
10583 * Add compression id to {d2i,i2d}_SSL_SESSION so it is correctly saved
10588 * Rename uni2asc and asc2uni functions to OPENSSL_uni2asc and
10589 OPENSSL_asc2uni conditionally on Netware platforms to avoid a name
10592 *Guenter <lists@gknw.net>*
10594 * Fix the server certificate chain building code to use X509_verify_cert(),
10595 it used to have an ad-hoc builder which was unable to cope with anything
10596 other than a simple chain.
10598 *David Woodhouse <dwmw2@infradead.org>, Steve Henson*
10600 * Don't check self signed certificate signatures in X509_verify_cert()
10601 by default (a flag can override this): it just wastes time without
10602 adding any security. As a useful side effect self signed root CAs
10603 with non-FIPS digests are now usable in FIPS mode.
10607 * In dtls1_process_out_of_seq_message() the check if the current message
10608 is already buffered was missing. For every new message was memory
10609 allocated, allowing an attacker to perform an denial of service attack
10610 with sending out of seq handshake messages until there is no memory
10611 left. Additionally every future message was buffered, even if the
10612 sequence number made no sense and would be part of another handshake.
10613 So only messages with sequence numbers less than 10 in advance will be
10614 buffered. ([CVE-2009-1378])
10616 *Robin Seggelmann, discovered by Daniel Mentz*
10618 * Records are buffered if they arrive with a future epoch to be
10619 processed after finishing the corresponding handshake. There is
10620 currently no limitation to this buffer allowing an attacker to perform
10621 a DOS attack with sending records with future epochs until there is no
10622 memory left. This patch adds the pqueue_size() function to determine
10623 the size of a buffer and limits the record buffer to 100 entries.
10626 *Robin Seggelmann, discovered by Daniel Mentz*
10628 * Keep a copy of frag->msg_header.frag_len so it can be used after the
10629 parent structure is freed. ([CVE-2009-1379])
10633 * Handle non-blocking I/O properly in SSL_shutdown() call.
10635 *Darryl Miles <darryl-mailinglists@netbauds.net>*
10637 * Add `2.5.4.*` OIDs
10639 *Ilya O. <vrghost@gmail.com>*
10641 ### Changes between 0.9.8k and 0.9.8l [5 Nov 2009]
10643 * Disable renegotiation completely - this fixes a severe security
10644 problem ([CVE-2009-3555]) at the cost of breaking all
10645 renegotiation. Renegotiation can be re-enabled by setting
10646 SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION in s3->flags at
10647 run-time. This is really not recommended unless you know what
10652 ### Changes between 0.9.8j and 0.9.8k [25 Mar 2009]
10654 * Don't set val to NULL when freeing up structures, it is freed up by
10655 underlying code. If `sizeof(void *) > sizeof(long)` this can result in
10656 zeroing past the valid field. ([CVE-2009-0789])
10658 *Paolo Ganci <Paolo.Ganci@AdNovum.CH>*
10660 * Fix bug where return value of CMS_SignerInfo_verify_content() was not
10661 checked correctly. This would allow some invalid signed attributes to
10662 appear to verify correctly. ([CVE-2009-0591])
10664 *Ivan Nestlerode <inestlerode@us.ibm.com>*
10666 * Reject UniversalString and BMPString types with invalid lengths. This
10667 prevents a crash in ASN1_STRING_print_ex() which assumes the strings have
10668 a legal length. ([CVE-2009-0590])
10672 * Set S/MIME signing as the default purpose rather than setting it
10673 unconditionally. This allows applications to override it at the store
10678 * Permit restricted recursion of ASN1 strings. This is needed in practice
10679 to handle some structures.
10683 * Improve efficiency of mem_gets: don't search whole buffer each time
10686 *Jeremy Shapiro <jnshapir@us.ibm.com>*
10688 * New -hex option for openssl rand.
10692 * Print out UTF8String and NumericString when parsing ASN1.
10696 * Support NumericString type for name components.
10700 * Allow CC in the environment to override the automatically chosen
10701 compiler. Note that nothing is done to ensure flags work with the
10706 ### Changes between 0.9.8i and 0.9.8j [07 Jan 2009]
10708 * Properly check EVP_VerifyFinal() and similar return values
10711 *Ben Laurie, Bodo Moeller, Google Security Team*
10713 * Enable TLS extensions by default.
10717 * Allow the CHIL engine to be loaded, whether the application is
10718 multithreaded or not. (This does not release the developer from the
10719 obligation to set up the dynamic locking callbacks.)
10721 *Sander Temme <sander@temme.net>*
10723 * Use correct exit code if there is an error in dgst command.
10725 *Steve Henson; problem pointed out by Roland Dirlewanger*
10727 * Tweak Configure so that you need to say "experimental-jpake" to enable
10728 JPAKE, and need to use -DOPENSSL_EXPERIMENTAL_JPAKE in applications.
10732 * Add experimental JPAKE support, including demo authentication in
10733 s_client and s_server.
10737 * Set the comparison function in v3_addr_canonize().
10739 *Rob Austein <sra@hactrn.net>*
10741 * Add support for XMPP STARTTLS in s_client.
10743 *Philip Paeps <philip@freebsd.org>*
10745 * Change the server-side SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG behavior
10746 to ensure that even with this option, only ciphersuites in the
10747 server's preference list will be accepted. (Note that the option
10748 applies only when resuming a session, so the earlier behavior was
10749 just about the algorithm choice for symmetric cryptography.)
10753 ### Changes between 0.9.8h and 0.9.8i [15 Sep 2008]
10755 * Fix NULL pointer dereference if a DTLS server received
10756 ChangeCipherSpec as first record ([CVE-2009-1386]).
10760 * Fix a state transition in s3_srvr.c and d1_srvr.c
10761 (was using SSL3_ST_CW_CLNT_HELLO_B, should be `..._ST_SW_SRVR_...`).
10763 *Nagendra Modadugu*
10765 * The fix in 0.9.8c that supposedly got rid of unsafe
10766 double-checked locking was incomplete for RSA blinding,
10767 addressing just one layer of what turns out to have been
10768 doubly unsafe triple-checked locking.
10770 So now fix this for real by retiring the MONT_HELPER macro
10771 in crypto/rsa/rsa_eay.c.
10773 *Bodo Moeller; problem pointed out by Marius Schilder*
10775 * Various precautionary measures:
10777 - Avoid size_t integer overflow in HASH_UPDATE (md32_common.h).
10779 - Avoid a buffer overflow in d2i_SSL_SESSION() (ssl_asn1.c).
10780 (NB: This would require knowledge of the secret session ticket key
10781 to exploit, in which case you'd be SOL either way.)
10783 - Change bn_nist.c so that it will properly handle input BIGNUMs
10784 outside the expected range.
10786 - Enforce the 'num' check in BN_div() (bn_div.c) for non-BN_DEBUG
10789 *Neel Mehta, Bodo Moeller*
10791 * Allow engines to be "soft loaded" - i.e. optionally don't die if
10792 the load fails. Useful for distros.
10794 *Ben Laurie and the FreeBSD team*
10796 * Add support for Local Machine Keyset attribute in PKCS#12 files.
10800 * Fix BN_GF2m_mod_arr() top-bit cleanup code.
10804 * Expand ENGINE to support engine supplied SSL client certificate functions.
10806 This work was sponsored by Logica.
10810 * Add CryptoAPI ENGINE to support use of RSA and DSA keys held in Windows
10811 keystores. Support for SSL/TLS client authentication too.
10812 Not compiled unless enable-capieng specified to Configure.
10814 This work was sponsored by Logica.
10818 * Fix bug in X509_ATTRIBUTE creation: don't set attribute using
10819 ASN1_TYPE_set1 if MBSTRING flag set. This bug would crash certain
10820 attribute creation routines such as certificate requests and PKCS#12
10825 ### Changes between 0.9.8g and 0.9.8h [28 May 2008]
10827 * Fix flaw if 'Server Key exchange message' is omitted from a TLS
10828 handshake which could lead to a client crash as found using the
10829 Codenomicon TLS test suite ([CVE-2008-1672])
10831 *Steve Henson, Mark Cox*
10833 * Fix double free in TLS server name extensions which could lead to
10834 a remote crash found by Codenomicon TLS test suite ([CVE-2008-0891])
10838 * Clear error queue in SSL_CTX_use_certificate_chain_file()
10840 Clear the error queue to ensure that error entries left from
10841 older function calls do not interfere with the correct operation.
10843 *Lutz Jaenicke, Erik de Castro Lopo*
10845 * Remove root CA certificates of commercial CAs:
10847 The OpenSSL project does not recommend any specific CA and does not
10848 have any policy with respect to including or excluding any CA.
10849 Therefore, it does not make any sense to ship an arbitrary selection
10850 of root CA certificates with the OpenSSL software.
10854 * RSA OAEP patches to fix two separate invalid memory reads.
10855 The first one involves inputs when 'lzero' is greater than
10856 'SHA_DIGEST_LENGTH' (it would read about SHA_DIGEST_LENGTH bytes
10857 before the beginning of from). The second one involves inputs where
10858 the 'db' section contains nothing but zeroes (there is a one-byte
10859 invalid read after the end of 'db').
10861 *Ivan Nestlerode <inestlerode@us.ibm.com>*
10863 * Partial backport from 0.9.9-dev:
10865 Introduce bn_mul_mont (dedicated Montgomery multiplication
10866 procedure) as a candidate for BIGNUM assembler implementation.
10867 While 0.9.9-dev uses assembler for various architectures, only
10868 x86_64 is available by default here in the 0.9.8 branch, and
10869 32-bit x86 is available through a compile-time setting.
10871 To try the 32-bit x86 assembler implementation, use Configure
10872 option "enable-montasm" (which exists only for this backport).
10874 As "enable-montasm" for 32-bit x86 disclaims code stability
10875 anyway, in this constellation we activate additional code
10876 backported from 0.9.9-dev for further performance improvements,
10877 namely BN_from_montgomery_word. (To enable this otherwise,
10878 e.g. x86_64, try `-DMONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD`.)
10880 *Andy Polyakov (backport partially by Bodo Moeller)*
10882 * Add TLS session ticket callback. This allows an application to set
10883 TLS ticket cipher and HMAC keys rather than relying on hardcoded fixed
10884 values. This is useful for key rollover for example where several key
10885 sets may exist with different names.
10889 * Reverse ENGINE-internal logic for caching default ENGINE handles.
10890 This was broken until now in 0.9.8 releases, such that the only way
10891 a registered ENGINE could be used (assuming it initialises
10892 successfully on the host) was to explicitly set it as the default
10893 for the relevant algorithms. This is in contradiction with 0.9.7
10894 behaviour and the documentation. With this fix, when an ENGINE is
10895 registered into a given algorithm's table of implementations, the
10896 'uptodate' flag is reset so that auto-discovery will be used next
10897 time a new context for that algorithm attempts to select an
10900 *Ian Lister (tweaked by Geoff Thorpe)*
10902 * Backport of CMS code to OpenSSL 0.9.8. This differs from the 0.9.9
10903 implementation in the following ways:
10905 Lack of EVP_PKEY_ASN1_METHOD means algorithm parameters have to be
10908 Lack of BER streaming support means one pass streaming processing is
10909 only supported if data is detached: setting the streaming flag is
10910 ignored for embedded content.
10912 CMS support is disabled by default and must be explicitly enabled
10913 with the enable-cms configuration option.
10917 * Update the GMP engine glue to do direct copies between BIGNUM and
10918 mpz_t when openssl and GMP use the same limb size. Otherwise the
10919 existing "conversion via a text string export" trick is still used.
10921 *Paul Sheer <paulsheer@gmail.com>*
10923 * Zlib compression BIO. This is a filter BIO which compressed and
10924 uncompresses any data passed through it.
10928 * Add AES_wrap_key() and AES_unwrap_key() functions to implement
10929 RFC3394 compatible AES key wrapping.
10933 * Add utility functions to handle ASN1 structures. ASN1_STRING_set0():
10934 sets string data without copying. X509_ALGOR_set0() and
10935 X509_ALGOR_get0(): set and retrieve X509_ALGOR (AlgorithmIdentifier)
10936 data. Attribute function X509at_get0_data_by_OBJ(): retrieves data
10937 from an X509_ATTRIBUTE structure optionally checking it occurs only
10938 once. ASN1_TYPE_set1(): set and ASN1_TYPE structure copying supplied
10943 * Fix BN flag handling in RSA_eay_mod_exp() and BN_MONT_CTX_set()
10944 to get the expected BN_FLG_CONSTTIME behavior.
10946 *Bodo Moeller (Google)*
10950 - fixed wrong usage of ioctlsocket() when build for LIBC BSD sockets
10951 - fixed do_tests.pl to run the test suite with CLIB builds too (CLIB_OPT)
10952 - added some more tests to do_tests.pl
10953 - fixed RunningProcess usage so that it works with newer LIBC NDKs too
10954 - removed usage of BN_LLONG for CLIB builds to avoid runtime dependency
10955 - added new Configure targets netware-clib-bsdsock, netware-clib-gcc,
10956 netware-clib-bsdsock-gcc, netware-libc-bsdsock-gcc
10957 - various changes to netware.pl to enable gcc-cross builds on Win32
10959 - changed crypto/bio/b_sock.c to work with macro functions (CLIB BSD)
10960 - various changes to fix missing prototype warnings
10961 - fixed x86nasm.pl to create correct asm files for NASM COFF output
10962 - added AES, WHIRLPOOL and CPUID assembler code to build files
10963 - added missing AES assembler make rules to mk1mf.pl
10964 - fixed order of includes in `apps/ocsp.c` so that `e_os.h` settings apply
10966 *Guenter Knauf <eflash@gmx.net>*
10968 * Implement certificate status request TLS extension defined in RFC3546.
10969 A client can set the appropriate parameters and receive the encoded
10970 OCSP response via a callback. A server can query the supplied parameters
10971 and set the encoded OCSP response in the callback. Add simplified examples
10972 to s_client and s_server.
10976 ### Changes between 0.9.8f and 0.9.8g [19 Oct 2007]
10978 * Fix various bugs:
10979 + Binary incompatibility of ssl_ctx_st structure
10980 + DTLS interoperation with non-compliant servers
10981 + Don't call get_session_cb() without proposed session
10982 + Fix ia64 assembler code
10984 *Andy Polyakov, Steve Henson*
10986 ### Changes between 0.9.8e and 0.9.8f [11 Oct 2007]
10988 * DTLS Handshake overhaul. There were longstanding issues with
10989 OpenSSL DTLS implementation, which were making it impossible for
10990 RFC 4347 compliant client to communicate with OpenSSL server.
10991 Unfortunately just fixing these incompatibilities would "cut off"
10992 pre-0.9.8f clients. To allow for hassle free upgrade post-0.9.8e
10993 server keeps tolerating non RFC compliant syntax. The opposite is
10994 not true, 0.9.8f client can not communicate with earlier server.
10995 This update even addresses CVE-2007-4995.
10999 * Changes to avoid need for function casts in OpenSSL: some compilers
11000 (gcc 4.2 and later) reject their use.
11001 *Kurt Roeckx <kurt@roeckx.be>, Peter Hartley <pdh@utter.chaos.org.uk>,
11004 * Add RFC4507 support to OpenSSL. This includes the corrections in
11005 RFC4507bis. The encrypted ticket format is an encrypted encoded
11006 SSL_SESSION structure, that way new session features are automatically
11009 If a client application caches session in an SSL_SESSION structure
11010 support is transparent because tickets are now stored in the encoded
11013 The SSL_CTX structure automatically generates keys for ticket
11014 protection in servers so again support should be possible
11015 with no application modification.
11017 If a client or server wishes to disable RFC4507 support then the option
11018 SSL_OP_NO_TICKET can be set.
11020 Add a TLS extension debugging callback to allow the contents of any client
11021 or server extensions to be examined.
11023 This work was sponsored by Google.
11027 * Add initial support for TLS extensions, specifically for the server_name
11028 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
11029 have new members for a hostname. The SSL data structure has an
11030 additional member `SSL_CTX *initial_ctx` so that new sessions can be
11031 stored in that context to allow for session resumption, even after the
11032 SSL has been switched to a new SSL_CTX in reaction to a client's
11033 server_name extension.
11035 New functions (subject to change):
11037 SSL_get_servername()
11038 SSL_get_servername_type()
11041 New CTRL codes and macros (subject to change):
11043 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
11044 - SSL_CTX_set_tlsext_servername_callback()
11045 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
11046 - SSL_CTX_set_tlsext_servername_arg()
11047 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
11049 openssl s_client has a new '-servername ...' option.
11051 openssl s_server has new options '-servername_host ...', '-cert2 ...',
11052 '-key2 ...', '-servername_fatal' (subject to change). This allows
11053 testing the HostName extension for a specific single hostname ('-cert'
11054 and '-key' remain fallbacks for handshakes without HostName
11055 negotiation). If the unrecognized_name alert has to be sent, this by
11056 default is a warning; it becomes fatal with the '-servername_fatal'
11059 *Peter Sylvester, Remy Allais, Christophe Renou, Steve Henson*
11061 * Add AES and SSE2 assembly language support to VC++ build.
11065 * Mitigate attack on final subtraction in Montgomery reduction.
11069 * Fix crypto/ec/ec_mult.c to work properly with scalars of value 0
11070 (which previously caused an internal error).
11074 * Squeeze another 10% out of IGE mode when in != out.
11078 * AES IGE mode speedup.
11080 *Dean Gaudet (Google)*
11082 * Add the Korean symmetric 128-bit cipher SEED (see
11083 <http://www.kisa.or.kr/kisa/seed/jsp/seed_eng.jsp>) and
11084 add SEED ciphersuites from RFC 4162:
11086 TLS_RSA_WITH_SEED_CBC_SHA = "SEED-SHA"
11087 TLS_DHE_DSS_WITH_SEED_CBC_SHA = "DHE-DSS-SEED-SHA"
11088 TLS_DHE_RSA_WITH_SEED_CBC_SHA = "DHE-RSA-SEED-SHA"
11089 TLS_DH_anon_WITH_SEED_CBC_SHA = "ADH-SEED-SHA"
11091 To minimize changes between patchlevels in the OpenSSL 0.9.8
11092 series, SEED remains excluded from compilation unless OpenSSL
11093 is configured with 'enable-seed'.
11095 *KISA, Bodo Moeller*
11097 * Mitigate branch prediction attacks, which can be practical if a
11098 single processor is shared, allowing a spy process to extract
11099 information. For detailed background information, see
11100 <http://eprint.iacr.org/2007/039> (O. Aciicmez, S. Gueron,
11101 J.-P. Seifert, "New Branch Prediction Vulnerabilities in OpenSSL
11102 and Necessary Software Countermeasures"). The core of the change
11103 are new versions BN_div_no_branch() and
11104 BN_mod_inverse_no_branch() of BN_div() and BN_mod_inverse(),
11105 respectively, which are slower, but avoid the security-relevant
11106 conditional branches. These are automatically called by BN_div()
11107 and BN_mod_inverse() if the flag BN_FLG_CONSTTIME is set for one
11108 of the input BIGNUMs. Also, BN_is_bit_set() has been changed to
11109 remove a conditional branch.
11111 BN_FLG_CONSTTIME is the new name for the previous
11112 BN_FLG_EXP_CONSTTIME flag, since it now affects more than just
11113 modular exponentiation. (Since OpenSSL 0.9.7h, setting this flag
11114 in the exponent causes BN_mod_exp_mont() to use the alternative
11115 implementation in BN_mod_exp_mont_consttime().) The old name
11116 remains as a deprecated alias.
11118 Similarly, RSA_FLAG_NO_EXP_CONSTTIME is replaced by a more general
11119 RSA_FLAG_NO_CONSTTIME flag since the RSA implementation now uses
11120 constant-time implementations for more than just exponentiation.
11121 Here too the old name is kept as a deprecated alias.
11123 BN_BLINDING_new() will now use BN_dup() for the modulus so that
11124 the BN_BLINDING structure gets an independent copy of the
11125 modulus. This means that the previous `BIGNUM *m` argument to
11126 BN_BLINDING_new() and to BN_BLINDING_create_param() now
11127 essentially becomes `const BIGNUM *m`, although we can't actually
11128 change this in the header file before 0.9.9. It allows
11129 RSA_setup_blinding() to use BN_with_flags() on the modulus to
11130 enable BN_FLG_CONSTTIME.
11132 *Matthew D Wood (Intel Corp)*
11134 * In the SSL/TLS server implementation, be strict about session ID
11135 context matching (which matters if an application uses a single
11136 external cache for different purposes). Previously,
11137 out-of-context reuse was forbidden only if SSL_VERIFY_PEER was
11138 set. This did ensure strict client verification, but meant that,
11139 with applications using a single external cache for quite
11140 different requirements, clients could circumvent ciphersuite
11141 restrictions for a given session ID context by starting a session
11142 in a different context.
11146 * Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
11147 a ciphersuite string such as "DEFAULT:RSA" cannot enable
11148 authentication-only ciphersuites.
11152 * Update the SSL_get_shared_ciphers() fix CVE-2006-3738 which was
11153 not complete and could lead to a possible single byte overflow
11154 ([CVE-2007-5135]) [Ben Laurie]
11156 ### Changes between 0.9.8d and 0.9.8e [23 Feb 2007]
11158 * Since AES128 and AES256 (and similarly Camellia128 and
11159 Camellia256) share a single mask bit in the logic of
11160 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
11161 kludge to work properly if AES128 is available and AES256 isn't
11162 (or if Camellia128 is available and Camellia256 isn't).
11166 * Fix the BIT STRING encoding generated by crypto/ec/ec_asn1.c
11167 (within i2d_ECPrivateKey, i2d_ECPKParameters, i2d_ECParameters):
11168 When a point or a seed is encoded in a BIT STRING, we need to
11169 prevent the removal of trailing zero bits to get the proper DER
11170 encoding. (By default, crypto/asn1/a_bitstr.c assumes the case
11171 of a NamedBitList, for which trailing 0 bits need to be removed.)
11175 * Have SSL/TLS server implementation tolerate "mismatched" record
11176 protocol version while receiving ClientHello even if the
11177 ClientHello is fragmented. (The server can't insist on the
11178 particular protocol version it has chosen before the ServerHello
11179 message has informed the client about his choice.)
11183 * Add RFC 3779 support.
11185 *Rob Austein for ARIN, Ben Laurie*
11187 * Load error codes if they are not already present instead of using a
11188 static variable. This allows them to be cleanly unloaded and reloaded.
11189 Improve header file function name parsing.
11193 * extend SMTP and IMAP protocol emulation in s_client to use EHLO
11194 or CAPABILITY handshake as required by RFCs.
11196 *Goetz Babin-Ebell*
11198 ### Changes between 0.9.8c and 0.9.8d [28 Sep 2006]
11200 * Introduce limits to prevent malicious keys being able to
11201 cause a denial of service. ([CVE-2006-2940])
11203 *Steve Henson, Bodo Moeller*
11205 * Fix ASN.1 parsing of certain invalid structures that can result
11206 in a denial of service. ([CVE-2006-2937]) [Steve Henson]
11208 * Fix buffer overflow in SSL_get_shared_ciphers() function.
11209 ([CVE-2006-3738]) [Tavis Ormandy and Will Drewry, Google Security Team]
11211 * Fix SSL client code which could crash if connecting to a
11212 malicious SSLv2 server. ([CVE-2006-4343])
11214 *Tavis Ormandy and Will Drewry, Google Security Team*
11216 * Since 0.9.8b, ciphersuite strings naming explicit ciphersuites
11217 match only those. Before that, "AES256-SHA" would be interpreted
11218 as a pattern and match "AES128-SHA" too (since AES128-SHA got
11219 the same strength classification in 0.9.7h) as we currently only
11220 have a single AES bit in the ciphersuite description bitmap.
11221 That change, however, also applied to ciphersuite strings such as
11222 "RC4-MD5" that intentionally matched multiple ciphersuites --
11223 namely, SSL 2.0 ciphersuites in addition to the more common ones
11224 from SSL 3.0/TLS 1.0.
11226 So we change the selection algorithm again: Naming an explicit
11227 ciphersuite selects this one ciphersuite, and any other similar
11228 ciphersuite (same bitmap) from *other* protocol versions.
11229 Thus, "RC4-MD5" again will properly select both the SSL 2.0
11230 ciphersuite and the SSL 3.0/TLS 1.0 ciphersuite.
11232 Since SSL 2.0 does not have any ciphersuites for which the
11233 128/256 bit distinction would be relevant, this works for now.
11234 The proper fix will be to use different bits for AES128 and
11235 AES256, which would have avoided the problems from the beginning;
11236 however, bits are scarce, so we can only do this in a new release
11237 (not just a patchlevel) when we can change the SSL_CIPHER
11238 definition to split the single 'unsigned long mask' bitmap into
11239 multiple values to extend the available space.
11243 ### Changes between 0.9.8b and 0.9.8c [05 Sep 2006]
11245 * Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
11246 ([CVE-2006-4339]) [Ben Laurie and Google Security Team]
11248 * Add AES IGE and biIGE modes.
11252 * Change the Unix randomness entropy gathering to use poll() when
11253 possible instead of select(), since the latter has some
11254 undesirable limitations.
11256 *Darryl Miles via Richard Levitte and Bodo Moeller*
11258 * Disable "ECCdraft" ciphersuites more thoroughly. Now special
11259 treatment in ssl/ssl_ciph.s makes sure that these ciphersuites
11260 cannot be implicitly activated as part of, e.g., the "AES" alias.
11261 However, please upgrade to OpenSSL 0.9.9[-dev] for
11262 non-experimental use of the ECC ciphersuites to get TLS extension
11263 support, which is required for curve and point format negotiation
11264 to avoid potential handshake problems.
11268 * Disable rogue ciphersuites:
11270 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
11271 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
11272 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
11274 The latter two were purportedly from
11275 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
11278 Also deactivate the remaining ciphersuites from
11279 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
11280 unofficial, and the ID has long expired.
11284 * Fix RSA blinding Heisenbug (problems sometimes occurred on
11285 dual-core machines) and other potential thread-safety issues.
11289 * Add the symmetric cipher Camellia (128-bit, 192-bit, 256-bit key
11290 versions), which is now available for royalty-free use
11291 (see <http://info.isl.ntt.co.jp/crypt/eng/info/chiteki.html>).
11292 Also, add Camellia TLS ciphersuites from RFC 4132.
11294 To minimize changes between patchlevels in the OpenSSL 0.9.8
11295 series, Camellia remains excluded from compilation unless OpenSSL
11296 is configured with 'enable-camellia'.
11300 * Disable the padding bug check when compression is in use. The padding
11301 bug check assumes the first packet is of even length, this is not
11302 necessarily true if compression is enabled and can result in false
11303 positives causing handshake failure. The actual bug test is ancient
11304 code so it is hoped that implementations will either have fixed it by
11305 now or any which still have the bug do not support compression.
11309 ### Changes between 0.9.8a and 0.9.8b [04 May 2006]
11311 * When applying a cipher rule check to see if string match is an explicit
11312 cipher suite and only match that one cipher suite if it is.
11316 * Link in manifests for VC++ if needed.
11318 *Austin Ziegler <halostatue@gmail.com>*
11320 * Update support for ECC-based TLS ciphersuites according to
11321 draft-ietf-tls-ecc-12.txt with proposed changes (but without
11322 TLS extensions, which are supported starting with the 0.9.9
11323 branch, not in the OpenSSL 0.9.8 branch).
11327 * New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support
11328 opaque EVP_CIPHER_CTX handling.
11332 * Fixes and enhancements to zlib compression code. We now only use
11333 "zlib1.dll" and use the default `__cdecl` calling convention on Win32
11334 to conform with the standards mentioned here:
11335 <http://www.zlib.net/DLL_FAQ.txt>
11336 Static zlib linking now works on Windows and the new --with-zlib-include
11337 --with-zlib-lib options to Configure can be used to supply the location
11338 of the headers and library. Gracefully handle case where zlib library
11343 * Several fixes and enhancements to the OID generation code. The old code
11344 sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't
11345 handle numbers larger than ULONG_MAX, truncated printing and had a
11346 non standard OBJ_obj2txt() behaviour.
11350 * Add support for building of engines under engine/ as shared libraries
11351 under VC++ build system.
11355 * Corrected the numerous bugs in the Win32 path splitter in DSO.
11356 Hopefully, we will not see any false combination of paths any more.
11360 ### Changes between 0.9.8 and 0.9.8a [11 Oct 2005]
11362 * Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
11363 (part of SSL_OP_ALL). This option used to disable the
11364 countermeasure against man-in-the-middle protocol-version
11365 rollback in the SSL 2.0 server implementation, which is a bad
11366 idea. ([CVE-2005-2969])
11368 *Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
11369 for Information Security, National Institute of Advanced Industrial
11370 Science and Technology [AIST], Japan)*
11372 * Add two function to clear and return the verify parameter flags.
11376 * Keep cipherlists sorted in the source instead of sorting them at
11377 runtime, thus removing the need for a lock.
11381 * Avoid some small subgroup attacks in Diffie-Hellman.
11383 *Nick Mathewson and Ben Laurie*
11385 * Add functions for well-known primes.
11389 * Extended Windows CE support.
11391 *Satoshi Nakamura and Andy Polyakov*
11393 * Initialize SSL_METHOD structures at compile time instead of during
11394 runtime, thus removing the need for a lock.
11398 * Make PKCS7_decrypt() work even if no certificate is supplied by
11399 attempting to decrypt each encrypted key in turn. Add support to
11404 ### Changes between 0.9.7h and 0.9.8 [05 Jul 2005]
11406 [NB: OpenSSL 0.9.7i and later 0.9.7 patch levels were released after
11409 * Add libcrypto.pc and libssl.pc for those who feel they need them.
11413 * Change CA.sh and CA.pl so they don't bundle the CSR and the private
11414 key into the same file any more.
11418 * Add initial support for Win64, both IA64 and AMD64/x64 flavors.
11422 * Add -utf8 command line and config file option to 'ca'.
11424 *Stefan <stf@udoma.org*
11426 * Removed the macro des_crypt(), as it seems to conflict with some
11427 libraries. Use DES_crypt().
11431 * Correct naming of the 'chil' and '4758cca' ENGINEs. This
11432 involves renaming the source and generated shared-libs for
11433 both. The engines will accept the corrected or legacy ids
11434 ('ncipher' and '4758_cca' respectively) when binding. NB,
11435 this only applies when building 'shared'.
11437 *Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe*
11439 * Add attribute functions to EVP_PKEY structure. Modify
11440 PKCS12_create() to recognize a CSP name attribute and
11441 use it. Make -CSP option work again in pkcs12 utility.
11445 * Add new functionality to the bn blinding code:
11446 - automatic re-creation of the BN_BLINDING parameters after
11447 a fixed number of uses (currently 32)
11448 - add new function for parameter creation
11449 - introduce flags to control the update behaviour of the
11450 BN_BLINDING parameters
11451 - hide BN_BLINDING structure
11452 Add a second BN_BLINDING slot to the RSA structure to improve
11453 performance when a single RSA object is shared among several
11458 * Add support for DTLS.
11460 *Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie*
11462 * Add support for DER encoded private keys (SSL_FILETYPE_ASN1)
11463 to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file()
11467 * Remove buggy and incomplete DH cert support from
11468 ssl/ssl_rsa.c and ssl/s3_both.c
11472 * Use SHA-1 instead of MD5 as the default digest algorithm for
11473 the `apps/openssl` commands.
11477 * Compile clean with "-Wall -Wmissing-prototypes
11478 -Wstrict-prototypes -Wmissing-declarations -Werror". Currently
11479 DEBUG_SAFESTACK must also be set.
11483 * Change ./Configure so that certain algorithms can be disabled by default.
11484 The new counterpiece to "no-xxx" is "enable-xxx".
11486 The patented RC5 and MDC2 algorithms will now be disabled unless
11487 "enable-rc5" and "enable-mdc2", respectively, are specified.
11489 (IDEA remains enabled despite being patented. This is because IDEA
11490 is frequently required for interoperability, and there is no license
11491 fee for non-commercial use. As before, "no-idea" can be used to
11492 avoid this algorithm.)
11496 * Add processing of proxy certificates (see RFC 3820). This work was
11497 sponsored by KTH (The Royal Institute of Technology in Stockholm) and
11498 EGEE (Enabling Grids for E-science in Europe).
11502 * RC4 performance overhaul on modern architectures/implementations, such
11503 as Intel P4, IA-64 and AMD64.
11507 * New utility extract-section.pl. This can be used specify an alternative
11508 section number in a pod file instead of having to treat each file as
11509 a separate case in Makefile. This can be done by adding two lines to the
11512 =for comment openssl_section:XXX
11514 The blank line is mandatory.
11518 * New arguments -certform, -keyform and -pass for s_client and s_server
11519 to allow alternative format key and certificate files and passphrase
11524 * New structure X509_VERIFY_PARAM which combines current verify parameters,
11525 update associated structures and add various utility functions.
11527 Add new policy related verify parameters, include policy checking in
11528 standard verify code. Enhance 'smime' application with extra parameters
11529 to support policy checking and print out.
11533 * Add a new engine to support VIA PadLock ACE extensions in the VIA C3
11534 Nehemiah processors. These extensions support AES encryption in hardware
11535 as well as RNG (though RNG support is currently disabled).
11537 *Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov*
11539 * Deprecate `BN_[get|set]_params()` functions (they were ignored internally).
11543 * New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented.
11545 *Andy Polyakov and a number of other people*
11547 * Improved PowerPC platform support. Most notably BIGNUM assembler
11548 implementation contributed by IBM.
11550 *Suresh Chari, Peter Waltenberg, Andy Polyakov*
11552 * The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public
11553 exponent rather than 'unsigned long'. There is a corresponding change to
11554 the new 'rsa_keygen' element of the RSA_METHOD structure.
11556 *Jelte Jansen, Geoff Thorpe*
11558 * Functionality for creating the initial serial number file is now
11559 moved from CA.pl to the 'ca' utility with a new option -create_serial.
11561 (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial
11562 number file to 1, which is bound to cause problems. To avoid
11563 the problems while respecting compatibility between different 0.9.7
11564 patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in
11565 CA.pl for serial number initialization. With the new release 0.9.8,
11566 we can fix the problem directly in the 'ca' utility.)
11570 * Reduced header interdependencies by declaring more opaque objects in
11571 ossl_typ.h. As a consequence, including some headers (eg. engine.h) will
11572 give fewer recursive includes, which could break lazy source code - so
11573 this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always,
11574 developers should define this symbol when building and using openssl to
11575 ensure they track the recommended behaviour, interfaces, [etc], but
11576 backwards-compatible behaviour prevails when this isn't defined.
11580 * New function X509_POLICY_NODE_print() which prints out policy nodes.
11584 * Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality.
11585 This will generate a random key of the appropriate length based on the
11586 cipher context. The EVP_CIPHER can provide its own random key generation
11587 routine to support keys of a specific form. This is used in the des and
11588 3des routines to generate a key of the correct parity. Update S/MIME
11589 code to use new functions and hence generate correct parity DES keys.
11590 Add EVP_CHECK_DES_KEY #define to return an error if the key is not
11591 valid (weak or incorrect parity).
11595 * Add a local set of CRLs that can be used by X509_verify_cert() as well
11596 as looking them up. This is useful when the verified structure may contain
11597 CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs
11598 present unless the new PKCS7_NO_CRL flag is asserted.
11602 * Extend ASN1 oid configuration module. It now additionally accepts the
11605 shortName = some long name, 1.2.3.4
11609 * Reimplemented the BN_CTX implementation. There is now no more static
11610 limitation on the number of variables it can handle nor the depth of the
11611 "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack
11612 information can now expand as required, and rather than having a single
11613 static array of bignums, BN_CTX now uses a linked-list of such arrays
11614 allowing it to expand on demand whilst maintaining the usefulness of
11615 BN_CTX's "bundling".
11619 * Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD
11620 to allow all RSA operations to function using a single BN_CTX.
11624 * Preliminary support for certificate policy evaluation and checking. This
11625 is initially intended to pass the tests outlined in "Conformance Testing
11626 of Relying Party Client Certificate Path Processing Logic" v1.07.
11630 * bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and
11631 remained unused and not that useful. A variety of other little bignum
11632 tweaks and fixes have also been made continuing on from the audit (see
11637 * Constify all or almost all d2i, c2i, s2i and r2i functions, along with
11638 associated ASN1, EVP and SSL functions and old ASN1 macros.
11642 * BN_zero() only needs to set 'top' and 'neg' to zero for correct results,
11643 and this should never fail. So the return value from the use of
11644 BN_set_word() (which can fail due to needless expansion) is now deprecated;
11645 if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro.
11649 * BN_CTX_get() should return zero-valued bignums, providing the same
11650 initialised value as BN_new().
11652 *Geoff Thorpe, suggested by Ulf Möller*
11654 * Support for inhibitAnyPolicy certificate extension.
11658 * An audit of the BIGNUM code is underway, for which debugging code is
11659 enabled when BN_DEBUG is defined. This makes stricter enforcements on what
11660 is considered valid when processing BIGNUMs, and causes execution to
11661 assert() when a problem is discovered. If BN_DEBUG_RAND is defined,
11662 further steps are taken to deliberately pollute unused data in BIGNUM
11663 structures to try and expose faulty code further on. For now, openssl will
11664 (in its default mode of operation) continue to tolerate the inconsistent
11665 forms that it has tolerated in the past, but authors and packagers should
11666 consider trying openssl and their own applications when compiled with
11667 these debugging symbols defined. It will help highlight potential bugs in
11668 their own code, and will improve the test coverage for OpenSSL itself. At
11669 some point, these tighter rules will become openssl's default to improve
11670 maintainability, though the assert()s and other overheads will remain only
11671 in debugging configurations. See bn.h for more details.
11673 *Geoff Thorpe, Nils Larsch, Ulf Möller*
11675 * BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure
11676 that can only be obtained through BN_CTX_new() (which implicitly
11677 initialises it). The presence of this function only made it possible
11678 to overwrite an existing structure (and cause memory leaks).
11682 * Because of the callback-based approach for implementing LHASH as a
11683 template type, lh_insert() adds opaque objects to hash-tables and
11684 lh_doall() or lh_doall_arg() are typically used with a destructor callback
11685 to clean up those corresponding objects before destroying the hash table
11686 (and losing the object pointers). So some over-zealous constifications in
11687 LHASH have been relaxed so that lh_insert() does not take (nor store) the
11688 objects as "const" and the `lh_doall[_arg]` callback wrappers are not
11689 prototyped to have "const" restrictions on the object pointers they are
11690 given (and so aren't required to cast them away any more).
11694 * The tmdiff.h API was so ugly and minimal that our own timing utility
11695 (speed) prefers to use its own implementation. The two implementations
11696 haven't been consolidated as yet (volunteers?) but the tmdiff API has had
11697 its object type properly exposed (MS_TM) instead of casting to/from
11698 `char *`. This may still change yet if someone realises MS_TM and
11700 aren't necessarily the greatest nomenclatures - but this is what was used
11701 internally to the implementation so I've used that for now.
11705 * Ensure that deprecated functions do not get compiled when
11706 OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of
11707 the self-tests were still using deprecated key-generation functions so
11708 these have been updated also.
11712 * Reorganise PKCS#7 code to separate the digest location functionality
11713 into PKCS7_find_digest(), digest addition into PKCS7_bio_add_digest().
11714 New function PKCS7_set_digest() to set the digest type for PKCS#7
11715 digestedData type. Add additional code to correctly generate the
11716 digestedData type and add support for this type in PKCS7 initialization
11721 * New function PKCS7_set0_type_other() this initializes a PKCS7
11722 structure of type "other".
11726 * Fix prime generation loop in crypto/bn/bn_prime.pl by making
11727 sure the loop does correctly stop and breaking ("division by zero")
11728 modulus operations are not performed. The (pre-generated) prime
11729 table crypto/bn/bn_prime.h was already correct, but it could not be
11730 re-generated on some platforms because of the "division by zero"
11731 situation in the script.
11733 *Ralf S. Engelschall*
11735 * Update support for ECC-based TLS ciphersuites according to
11736 draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with
11737 SHA-1 now is only used for "small" curves (where the
11738 representation of a field element takes up to 24 bytes); for
11739 larger curves, the field element resulting from ECDH is directly
11740 used as premaster secret.
11742 *Douglas Stebila (Sun Microsystems Laboratories)*
11744 * Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2
11745 curve secp160r1 to the tests.
11747 *Douglas Stebila (Sun Microsystems Laboratories)*
11749 * Add the possibility to load symbols globally with DSO.
11751 *Götz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte*
11753 * Add the functions ERR_set_mark() and ERR_pop_to_mark() for better
11754 control of the error stack.
11758 * Add support for STORE in ENGINE.
11762 * Add the STORE type. The intention is to provide a common interface
11763 to certificate and key stores, be they simple file-based stores, or
11764 HSM-type store, or LDAP stores, or...
11765 NOTE: The code is currently UNTESTED and isn't really used anywhere.
11769 * Add a generic structure called OPENSSL_ITEM. This can be used to
11770 pass a list of arguments to any function as well as provide a way
11771 for a function to pass data back to the caller.
11775 * Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup()
11776 works like BUF_strdup() but can be used to duplicate a portion of
11777 a string. The copy gets NUL-terminated. BUF_memdup() duplicates
11782 * Add the function sk_find_ex() which works like sk_find(), but will
11783 return an index to an element even if an exact match couldn't be
11784 found. The index is guaranteed to point at the element where the
11785 searched-for key would be inserted to preserve sorting order.
11789 * Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but
11790 takes an extra flags argument for optional functionality. Currently,
11791 the following flags are defined:
11793 OBJ_BSEARCH_VALUE_ON_NOMATCH
11794 This one gets OBJ_bsearch_ex() to return a pointer to the first
11795 element where the comparing function returns a negative or zero
11798 OBJ_BSEARCH_FIRST_VALUE_ON_MATCH
11799 This one gets OBJ_bsearch_ex() to return a pointer to the first
11800 element where the comparing function returns zero. This is useful
11801 if there are more than one element where the comparing function
11806 * Make it possible to create self-signed certificates with 'openssl ca'
11807 in such a way that the self-signed certificate becomes part of the
11808 CA database and uses the same mechanisms for serial number generation
11809 as all other certificate signing. The new flag '-selfsign' enables
11810 this functionality. Adapt CA.sh and CA.pl.in.
11814 * Add functionality to check the public key of a certificate request
11815 against a given private. This is useful to check that a certificate
11816 request can be signed by that key (self-signing).
11820 * Make it possible to have multiple active certificates with the same
11821 subject in the CA index file. This is done only if the keyword
11822 'unique_subject' is set to 'no' in the main CA section (default
11823 if 'CA_default') of the configuration file. The value is saved
11824 with the database itself in a separate index attribute file,
11825 named like the index file with '.attr' appended to the name.
11829 * Generate multi-valued AVAs using '+' notation in config files for
11834 * Support for nameConstraints certificate extension.
11838 * Support for policyConstraints certificate extension.
11842 * Support for policyMappings certificate extension.
11846 * Make sure the default DSA_METHOD implementation only uses its
11847 dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL,
11848 and change its own handlers to be NULL so as to remove unnecessary
11849 indirection. This lets alternative implementations fallback to the
11850 default implementation more easily.
11854 * Support for directoryName in GeneralName related extensions
11859 * Make it possible to link applications using Makefile.shared.
11860 Make that possible even when linking against static libraries!
11864 * Support for single pass processing for S/MIME signing. This now
11865 means that S/MIME signing can be done from a pipe, in addition
11866 cleartext signing (multipart/signed type) is effectively streaming
11867 and the signed data does not need to be all held in memory.
11869 This is done with a new flag PKCS7_STREAM. When this flag is set
11870 PKCS7_sign() only initializes the PKCS7 structure and the actual signing
11871 is done after the data is output (and digests calculated) in
11872 SMIME_write_PKCS7().
11876 * Add full support for -rpath/-R, both in shared libraries and
11877 applications, at least on the platforms where it's known how
11882 * In crypto/ec/ec_mult.c, implement fast point multiplication with
11883 precomputation, based on wNAF splitting: EC_GROUP_precompute_mult()
11884 will now compute a table of multiples of the generator that
11885 makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul()
11886 faster (notably in the case of a single point multiplication,
11887 scalar * generator).
11889 *Nils Larsch, Bodo Moeller*
11891 * IPv6 support for certificate extensions. The various extensions
11892 which use the IP:a.b.c.d can now take IPv6 addresses using the
11893 formats of RFC1884 2.2 . IPv6 addresses are now also displayed
11898 * Added an ENGINE that implements RSA by performing private key
11899 exponentiations with the GMP library. The conversions to and from
11900 GMP's mpz_t format aren't optimised nor are any montgomery forms
11901 cached, and on x86 it appears OpenSSL's own performance has caught up.
11902 However there are likely to be other architectures where GMP could
11903 provide a boost. This ENGINE is not built in by default, but it can be
11904 specified at Configure time and should be accompanied by the necessary
11905 linker additions, eg;
11906 ./config -DOPENSSL_USE_GMP -lgmp
11910 * "openssl engine" will not display ENGINE/DSO load failure errors when
11911 testing availability of engines with "-t" - the old behaviour is
11912 produced by increasing the feature's verbosity with "-tt".
11916 * ECDSA routines: under certain error conditions uninitialized BN objects
11917 could be freed. Solution: make sure initialization is performed early
11918 enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de>
11923 * Key-generation can now be implemented in RSA_METHOD, DSA_METHOD
11924 and DH_METHOD (eg. by ENGINE implementations) to override the normal
11925 software implementations. For DSA and DH, parameter generation can
11926 also be overridden by providing the appropriate method callbacks.
11930 * Change the "progress" mechanism used in key-generation and
11931 primality testing to functions that take a new BN_GENCB pointer in
11932 place of callback/argument pairs. The new API functions have `_ex`
11933 postfixes and the older functions are reimplemented as wrappers for
11934 the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide
11935 declarations of the old functions to help (graceful) attempts to
11936 migrate to the new functions. Also, the new key-generation API
11937 functions operate on a caller-supplied key-structure and return
11938 success/failure rather than returning a key or NULL - this is to
11939 help make "keygen" another member function of RSA_METHOD etc.
11941 Example for using the new callback interface:
11943 int (*my_callback)(int a, int b, BN_GENCB *cb) = ...;
11944 void *my_arg = ...;
11947 BN_GENCB_set(&my_cb, my_callback, my_arg);
11949 return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb);
11950 /* For the meaning of a, b in calls to my_callback(), see the
11951 * documentation of the function that calls the callback.
11952 * cb will point to my_cb; my_arg can be retrieved as cb->arg.
11953 * my_callback should return 1 if it wants BN_is_prime_ex()
11954 * to continue, or 0 to stop.
11959 * Change the ZLIB compression method to be stateful, and make it
11960 available to TLS with the number defined in
11961 draft-ietf-tls-compression-04.txt.
11965 * Add the ASN.1 structures and functions for CertificatePair, which
11966 is defined as follows (according to X.509_4thEditionDraftV6.pdf):
11968 CertificatePair ::= SEQUENCE {
11969 forward [0] Certificate OPTIONAL,
11970 reverse [1] Certificate OPTIONAL,
11971 -- at least one of the pair shall be present -- }
11973 Also implement the PEM functions to read and write certificate
11974 pairs, and defined the PEM tag as "CERTIFICATE PAIR".
11976 This needed to be defined, mostly for the sake of the LDAP
11977 attribute crossCertificatePair, but may prove useful elsewhere as
11982 * Make it possible to inhibit symlinking of shared libraries in
11983 Makefile.shared, for Cygwin's sake.
11987 * Extend the BIGNUM API by creating a function
11988 void BN_set_negative(BIGNUM *a, int neg);
11989 and a macro that behave like
11990 int BN_is_negative(const BIGNUM *a);
11992 to avoid the need to access 'a->neg' directly in applications.
11996 * Implement fast modular reduction for pseudo-Mersenne primes
11997 used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c).
11998 EC_GROUP_new_curve_GFp() will now automatically use this
12001 *Nils Larsch <nla@trustcenter.de>*
12003 * Add new lock type (CRYPTO_LOCK_BN).
12007 * Change the ENGINE framework to automatically load engines
12008 dynamically from specific directories unless they could be
12009 found to already be built in or loaded. Move all the
12010 current engines except for the cryptodev one to a new
12011 directory engines/.
12012 The engines in engines/ are built as shared libraries if
12013 the "shared" options was given to ./Configure or ./config.
12014 Otherwise, they are inserted in libcrypto.a.
12015 /usr/local/ssl/engines is the default directory for dynamic
12016 engines, but that can be overridden at configure time through
12017 the usual use of --prefix and/or --openssldir, and at run
12018 time with the environment variable OPENSSL_ENGINES.
12020 *Geoff Thorpe and Richard Levitte*
12022 * Add Makefile.shared, a helper makefile to build shared
12023 libraries. Adapt Makefile.org.
12027 * Add version info to Win32 DLLs.
12029 *Peter 'Luna' Runestig" <peter@runestig.com>*
12031 * Add new 'medium level' PKCS#12 API. Certificates and keys
12032 can be added using this API to created arbitrary PKCS#12
12033 files while avoiding the low-level API.
12035 New options to PKCS12_create(), key or cert can be NULL and
12036 will then be omitted from the output file. The encryption
12037 algorithm NIDs can be set to -1 for no encryption, the mac
12038 iteration count can be set to 0 to omit the mac.
12040 Enhance pkcs12 utility by making the -nokeys and -nocerts
12041 options work when creating a PKCS#12 file. New option -nomac
12042 to omit the mac, NONE can be set for an encryption algorithm.
12043 New code is modified to use the enhanced PKCS12_create()
12044 instead of the low-level API.
12048 * Extend ASN1 encoder to support indefinite length constructed
12049 encoding. This can output sequences tags and octet strings in
12050 this form. Modify pk7_asn1.c to support indefinite length
12051 encoding. This is experimental and needs additional code to
12052 be useful, such as an ASN1 bio and some enhanced streaming
12055 Extend template encode functionality so that tagging is passed
12056 down to the template encoder.
12060 * Let 'openssl req' fail if an argument to '-newkey' is not
12061 recognized instead of using RSA as a default.
12065 * Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt.
12066 As these are not official, they are not included in "ALL";
12067 the "ECCdraft" ciphersuite group alias can be used to select them.
12069 *Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)*
12071 * Add ECDH engine support.
12073 *Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)*
12075 * Add ECDH in new directory crypto/ecdh/.
12077 *Douglas Stebila (Sun Microsystems Laboratories)*
12079 * Let BN_rand_range() abort with an error after 100 iterations
12080 without success (which indicates a broken PRNG).
12084 * Change BN_mod_sqrt() so that it verifies that the input value
12085 is really the square of the return value. (Previously,
12086 BN_mod_sqrt would show GIGO behaviour.)
12090 * Add named elliptic curves over binary fields from X9.62, SECG,
12091 and WAP/WTLS; add OIDs that were still missing.
12093 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
12095 * Extend the EC library for elliptic curves over binary fields
12096 (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/).
12099 EC_GF2m_simple_method
12103 EC_GROUP_new_curve_GF2m
12104 EC_GROUP_set_curve_GF2m
12105 EC_GROUP_get_curve_GF2m
12106 EC_POINT_set_affine_coordinates_GF2m
12107 EC_POINT_get_affine_coordinates_GF2m
12108 EC_POINT_set_compressed_coordinates_GF2m
12110 Point compression for binary fields is disabled by default for
12111 patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to
12114 As binary polynomials are represented as BIGNUMs, various members
12115 of the EC_GROUP and EC_POINT data structures can be shared
12116 between the implementations for prime fields and binary fields;
12117 the above `..._GF2m functions` (except for EX_GROUP_new_curve_GF2m)
12118 are essentially identical to their `..._GFp` counterparts.
12119 (For simplicity, the `..._GFp` prefix has been dropped from
12120 various internal method names.)
12122 An internal 'field_div' method (similar to 'field_mul' and
12123 'field_sqr') has been added; this is used only for binary fields.
12125 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
12127 * Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult()
12128 through methods ('mul', 'precompute_mult').
12130 The generic implementations (now internally called 'ec_wNAF_mul'
12131 and 'ec_wNAF_precomputed_mult') remain the default if these
12132 methods are undefined.
12134 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
12136 * New function EC_GROUP_get_degree, which is defined through
12137 EC_METHOD. For curves over prime fields, this returns the bit
12138 length of the modulus.
12140 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
12142 * New functions EC_GROUP_dup, EC_POINT_dup.
12143 (These simply call ..._new and ..._copy).
12145 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
12147 * Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c.
12148 Polynomials are represented as BIGNUMs (where the sign bit is not
12149 used) in the following functions [macros]:
12152 BN_GF2m_sub [= BN_GF2m_add]
12153 BN_GF2m_mod [wrapper for BN_GF2m_mod_arr]
12154 BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr]
12155 BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr]
12157 BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr]
12158 BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr]
12159 BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr]
12160 BN_GF2m_cmp [= BN_ucmp]
12162 (Note that only the 'mod' functions are actually for fields GF(2^m).
12163 BN_GF2m_add() is misnomer, but this is for the sake of consistency.)
12165 For some functions, an the irreducible polynomial defining a
12166 field can be given as an 'unsigned int[]' with strictly
12167 decreasing elements giving the indices of those bits that are set;
12168 i.e., p[] represents the polynomial
12169 f(t) = t^p[0] + t^p[1] + ... + t^p[k]
12171 p[0] > p[1] > ... > p[k] = 0.
12172 This applies to the following functions:
12175 BN_GF2m_mod_mul_arr
12176 BN_GF2m_mod_sqr_arr
12177 BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv]
12178 BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div]
12179 BN_GF2m_mod_exp_arr
12180 BN_GF2m_mod_sqrt_arr
12181 BN_GF2m_mod_solve_quad_arr
12185 Conversion can be performed by the following functions:
12190 bntest.c has additional tests for binary polynomial arithmetic.
12192 Two implementations for BN_GF2m_mod_div() are available.
12193 The default algorithm simply uses BN_GF2m_mod_inv() and
12194 BN_GF2m_mod_mul(). The alternative algorithm is compiled in only
12195 if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the
12196 copyright notice in crypto/bn/bn_gf2m.c before enabling it).
12198 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
12200 * Add new error code 'ERR_R_DISABLED' that can be used when some
12201 functionality is disabled at compile-time.
12203 *Douglas Stebila <douglas.stebila@sun.com>*
12205 * Change default behaviour of 'openssl asn1parse' so that more
12206 information is visible when viewing, e.g., a certificate:
12208 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump'
12209 mode the content of non-printable OCTET STRINGs is output in a
12210 style similar to INTEGERs, but with '[HEX DUMP]' prepended to
12211 avoid the appearance of a printable string.
12213 *Nils Larsch <nla@trustcenter.de>*
12215 * Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access
12217 EC_GROUP_set_asn1_flag()
12218 EC_GROUP_get_asn1_flag()
12219 EC_GROUP_set_point_conversion_form()
12220 EC_GROUP_get_point_conversion_form()
12221 These control ASN1 encoding details:
12222 - Curves (i.e., groups) are encoded explicitly unless asn1_flag
12223 has been set to OPENSSL_EC_NAMED_CURVE.
12224 - Points are encoded in uncompressed form by default; options for
12225 asn1_for are as for point2oct, namely
12226 POINT_CONVERSION_COMPRESSED
12227 POINT_CONVERSION_UNCOMPRESSED
12228 POINT_CONVERSION_HYBRID
12230 Also add 'seed' and 'seed_len' members to EC_GROUP with access
12232 EC_GROUP_set_seed()
12233 EC_GROUP_get0_seed()
12234 EC_GROUP_get_seed_len()
12235 This is used only for ASN1 purposes (so far).
12237 *Nils Larsch <nla@trustcenter.de>*
12239 * Add 'field_type' member to EC_METHOD, which holds the NID
12240 of the appropriate field type OID. The new function
12241 EC_METHOD_get_field_type() returns this value.
12243 *Nils Larsch <nla@trustcenter.de>*
12246 EC_POINT_point2bn()
12247 EC_POINT_bn2point()
12248 EC_POINT_point2hex()
12249 EC_POINT_hex2point()
12250 providing useful interfaces to EC_POINT_point2oct() and
12251 EC_POINT_oct2point().
12253 *Nils Larsch <nla@trustcenter.de>*
12255 * Change internals of the EC library so that the functions
12256 EC_GROUP_set_generator()
12257 EC_GROUP_get_generator()
12258 EC_GROUP_get_order()
12259 EC_GROUP_get_cofactor()
12260 are implemented directly in crypto/ec/ec_lib.c and not dispatched
12261 to methods, which would lead to unnecessary code duplication when
12262 adding different types of curves.
12264 *Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller*
12266 * Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM
12267 arithmetic, and such that modified wNAFs are generated
12268 (which avoid length expansion in many cases).
12272 * Add a function EC_GROUP_check_discriminant() (defined via
12273 EC_METHOD) that verifies that the curve discriminant is non-zero.
12275 Add a function EC_GROUP_check() that makes some sanity tests
12276 on a EC_GROUP, its generator and order. This includes
12277 EC_GROUP_check_discriminant().
12279 *Nils Larsch <nla@trustcenter.de>*
12281 * Add ECDSA in new directory crypto/ecdsa/.
12283 Add applications 'openssl ecparam' and 'openssl ecdsa'
12284 (these are based on 'openssl dsaparam' and 'openssl dsa').
12286 ECDSA support is also included in various other files across the
12287 library. Most notably,
12288 - 'openssl req' now has a '-newkey ecdsa:file' option;
12289 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA;
12290 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and
12291 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make
12292 them suitable for ECDSA where domain parameters must be
12293 extracted before the specific public key;
12294 - ECDSA engine support has been added.
12296 *Nils Larsch <nla@trustcenter.de>*
12298 * Include some named elliptic curves, and add OIDs from X9.62,
12299 SECG, and WAP/WTLS. Each curve can be obtained from the new
12301 EC_GROUP_new_by_curve_name(),
12302 and the list of available named curves can be obtained with
12303 EC_get_builtin_curves().
12304 Also add a 'curve_name' member to EC_GROUP objects, which can be
12306 EC_GROUP_set_curve_name()
12307 EC_GROUP_get_curve_name()
12309 *Nils Larsch <larsch@trustcenter.de, Bodo Moeller*
12311 * Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
12312 was actually never needed) and in BN_mul(). The removal in BN_mul()
12313 required a small change in bn_mul_part_recursive() and the addition
12314 of the functions bn_cmp_part_words(), bn_sub_part_words() and
12315 bn_add_part_words(), which do the same thing as bn_cmp_words(),
12316 bn_sub_words() and bn_add_words() except they take arrays with
12321 ### Changes between 0.9.7l and 0.9.7m [23 Feb 2007]
12323 * Cleanse PEM buffers before freeing them since they may contain
12326 *Benjamin Bennett <ben@psc.edu>*
12328 * Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
12329 a ciphersuite string such as "DEFAULT:RSA" cannot enable
12330 authentication-only ciphersuites.
12334 * Since AES128 and AES256 share a single mask bit in the logic of
12335 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
12336 kludge to work properly if AES128 is available and AES256 isn't.
12340 * Expand security boundary to match 1.1.1 module.
12344 * Remove redundant features: hash file source, editing of test vectors
12345 modify fipsld to use external fips_premain.c signature.
12349 * New perl script mkfipsscr.pl to create shell scripts or batch files to
12350 run algorithm test programs.
12354 * Make algorithm test programs more tolerant of whitespace.
12358 * Have SSL/TLS server implementation tolerate "mismatched" record
12359 protocol version while receiving ClientHello even if the
12360 ClientHello is fragmented. (The server can't insist on the
12361 particular protocol version it has chosen before the ServerHello
12362 message has informed the client about his choice.)
12366 * Load error codes if they are not already present instead of using a
12367 static variable. This allows them to be cleanly unloaded and reloaded.
12371 ### Changes between 0.9.7k and 0.9.7l [28 Sep 2006]
12373 * Introduce limits to prevent malicious keys being able to
12374 cause a denial of service. ([CVE-2006-2940])
12376 *Steve Henson, Bodo Moeller*
12378 * Fix ASN.1 parsing of certain invalid structures that can result
12379 in a denial of service. ([CVE-2006-2937]) [Steve Henson]
12381 * Fix buffer overflow in SSL_get_shared_ciphers() function.
12382 ([CVE-2006-3738]) [Tavis Ormandy and Will Drewry, Google Security Team]
12384 * Fix SSL client code which could crash if connecting to a
12385 malicious SSLv2 server. ([CVE-2006-4343])
12387 *Tavis Ormandy and Will Drewry, Google Security Team*
12389 * Change ciphersuite string processing so that an explicit
12390 ciphersuite selects this one ciphersuite (so that "AES256-SHA"
12391 will no longer include "AES128-SHA"), and any other similar
12392 ciphersuite (same bitmap) from *other* protocol versions (so that
12393 "RC4-MD5" will still include both the SSL 2.0 ciphersuite and the
12394 SSL 3.0/TLS 1.0 ciphersuite). This is a backport combining
12395 changes from 0.9.8b and 0.9.8d.
12399 ### Changes between 0.9.7j and 0.9.7k [05 Sep 2006]
12401 * Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
12402 ([CVE-2006-4339]) [Ben Laurie and Google Security Team]
12404 * Change the Unix randomness entropy gathering to use poll() when
12405 possible instead of select(), since the latter has some
12406 undesirable limitations.
12408 *Darryl Miles via Richard Levitte and Bodo Moeller*
12410 * Disable rogue ciphersuites:
12412 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
12413 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
12414 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
12416 The latter two were purportedly from
12417 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
12420 Also deactivate the remaining ciphersuites from
12421 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
12422 unofficial, and the ID has long expired.
12426 * Fix RSA blinding Heisenbug (problems sometimes occurred on
12427 dual-core machines) and other potential thread-safety issues.
12431 ### Changes between 0.9.7i and 0.9.7j [04 May 2006]
12433 * Adapt fipsld and the build system to link against the validated FIPS
12434 module in FIPS mode.
12438 * Fixes for VC++ 2005 build under Windows.
12442 * Add new Windows build target VC-32-GMAKE for VC++. This uses GNU make
12443 from a Windows bash shell such as MSYS. It is autodetected from the
12444 "config" script when run from a VC++ environment. Modify standard VC++
12445 build to use fipscanister.o from the GNU make build.
12449 ### Changes between 0.9.7h and 0.9.7i [14 Oct 2005]
12451 * Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS.
12452 The value now differs depending on if you build for FIPS or not.
12453 BEWARE! A program linked with a shared FIPSed libcrypto can't be
12454 safely run with a non-FIPSed libcrypto, as it may crash because of
12455 the difference induced by this change.
12459 ### Changes between 0.9.7g and 0.9.7h [11 Oct 2005]
12461 * Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
12462 (part of SSL_OP_ALL). This option used to disable the
12463 countermeasure against man-in-the-middle protocol-version
12464 rollback in the SSL 2.0 server implementation, which is a bad
12465 idea. ([CVE-2005-2969])
12467 *Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
12468 for Information Security, National Institute of Advanced Industrial
12469 Science and Technology [AIST, Japan)]*
12471 * Minimal support for X9.31 signatures and PSS padding modes. This is
12472 mainly for FIPS compliance and not fully integrated at this stage.
12476 * For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform
12477 the exponentiation using a fixed-length exponent. (Otherwise,
12478 the information leaked through timing could expose the secret key
12479 after many signatures; cf. Bleichenbacher's attack on DSA with
12484 * Make a new fixed-window mod_exp implementation the default for
12485 RSA, DSA, and DH private-key operations so that the sequence of
12486 squares and multiplies and the memory access pattern are
12487 independent of the particular secret key. This will mitigate
12488 cache-timing and potential related attacks.
12490 BN_mod_exp_mont_consttime() is the new exponentiation implementation,
12491 and this is automatically used by BN_mod_exp_mont() if the new flag
12492 BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH
12493 will use this BN flag for private exponents unless the flag
12494 RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or
12495 DH_FLAG_NO_EXP_CONSTTIME, respectively, is set.
12497 *Matthew D Wood (Intel Corp), with some changes by Bodo Moeller*
12499 * Change the client implementation for SSLv23_method() and
12500 SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0
12501 Client Hello message format if the SSL_OP_NO_SSLv2 option is set.
12502 (Previously, the SSL 2.0 backwards compatible Client Hello
12503 message format would be used even with SSL_OP_NO_SSLv2.)
12507 * Add support for smime-type MIME parameter in S/MIME messages which some
12512 * New function BN_MONT_CTX_set_locked() to set montgomery parameters in
12513 a threadsafe manner. Modify rsa code to use new function and add calls
12514 to dsa and dh code (which had race conditions before).
12518 * Include the fixed error library code in the C error file definitions
12519 instead of fixing them up at runtime. This keeps the error code
12520 structures constant.
12524 ### Changes between 0.9.7f and 0.9.7g [11 Apr 2005]
12526 [NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after
12529 * Fixes for newer kerberos headers. NB: the casts are needed because
12530 the 'length' field is signed on one version and unsigned on another
12531 with no (?) obvious way to tell the difference, without these VC++
12532 complains. Also the "definition" of FAR (blank) is no longer included
12533 nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up
12534 some needed definitions.
12538 * Undo Cygwin change.
12542 * Added support for proxy certificates according to RFC 3820.
12543 Because they may be a security thread to unaware applications,
12544 they must be explicitly allowed in run-time. See
12545 docs/HOWTO/proxy_certificates.txt for further information.
12549 ### Changes between 0.9.7e and 0.9.7f [22 Mar 2005]
12551 * Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating
12552 server and client random values. Previously
12553 (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in
12554 less random data when sizeof(time_t) > 4 (some 64 bit platforms).
12556 This change has negligible security impact because:
12558 1. Server and client random values still have 24 bytes of pseudo random
12561 2. Server and client random values are sent in the clear in the initial
12564 3. The master secret is derived using the premaster secret (48 bytes in
12565 size for static RSA ciphersuites) as well as client server and random
12568 The OpenSSL team would like to thank the UK NISCC for bringing this issue
12571 *Stephen Henson, reported by UK NISCC*
12573 * Use Windows randomness collection on Cygwin.
12577 * Fix hang in EGD/PRNGD query when communication socket is closed
12578 prematurely by EGD/PRNGD.
12580 *Darren Tucker <dtucker@zip.com.au> via Lutz Jänicke, resolves #1014*
12582 * Prompt for pass phrases when appropriate for PKCS12 input format.
12586 * Back-port of selected performance improvements from development
12587 branch, as well as improved support for PowerPC platforms.
12591 * Add lots of checks for memory allocation failure, error codes to indicate
12592 failure and freeing up memory if a failure occurs.
12594 *Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson*
12596 * Add new -passin argument to dgst.
12600 * Perform some character comparisons of different types in X509_NAME_cmp:
12601 this is needed for some certificates that re-encode DNs into UTF8Strings
12602 (in violation of RFC3280) and can't or won't issue name rollover
12607 * Make an explicit check during certificate validation to see that
12608 the CA setting in each certificate on the chain is correct. As a
12609 side effect always do the following basic checks on extensions,
12610 not just when there's an associated purpose to the check:
12612 - if there is an unhandled critical extension (unless the user
12613 has chosen to ignore this fault)
12614 - if the path length has been exceeded (if one is set at all)
12615 - that certain extensions fit the associated purpose (if one has
12620 ### Changes between 0.9.7d and 0.9.7e [25 Oct 2004]
12622 * Avoid a race condition when CRLs are checked in a multi threaded
12623 environment. This would happen due to the reordering of the revoked
12624 entries during signature checking and serial number lookup. Now the
12625 encoding is cached and the serial number sort performed under a lock.
12626 Add new STACK function sk_is_sorted().
12630 * Add Delta CRL to the extension code.
12634 * Various fixes to s3_pkt.c so alerts are sent properly.
12636 *David Holmes <d.holmes@f5.com>*
12638 * Reduce the chances of duplicate issuer name and serial numbers (in
12639 violation of RFC3280) using the OpenSSL certificate creation utilities.
12640 This is done by creating a random 64 bit value for the initial serial
12641 number when a serial number file is created or when a self signed
12642 certificate is created using 'openssl req -x509'. The initial serial
12643 number file is created using 'openssl x509 -next_serial' in CA.pl
12644 rather than being initialized to 1.
12648 ### Changes between 0.9.7c and 0.9.7d [17 Mar 2004]
12650 * Fix null-pointer assignment in do_change_cipher_spec() revealed
12651 by using the Codenomicon TLS Test Tool ([CVE-2004-0079])
12653 *Joe Orton, Steve Henson*
12655 * Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites
12658 *Joe Orton, Steve Henson*
12660 * Make it possible to have multiple active certificates with the same
12661 subject in the CA index file. This is done only if the keyword
12662 'unique_subject' is set to 'no' in the main CA section (default
12663 if 'CA_default') of the configuration file. The value is saved
12664 with the database itself in a separate index attribute file,
12665 named like the index file with '.attr' appended to the name.
12669 * X509 verify fixes. Disable broken certificate workarounds when
12670 X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if
12671 keyUsage extension present. Don't accept CRLs with unhandled critical
12672 extensions: since verify currently doesn't process CRL extensions this
12673 rejects a CRL with *any* critical extensions. Add new verify error codes
12678 * When creating an OCSP nonce use an OCTET STRING inside the extnValue.
12679 A clarification of RFC2560 will require the use of OCTET STRINGs and
12680 some implementations cannot handle the current raw format. Since OpenSSL
12681 copies and compares OCSP nonces as opaque blobs without any attempt at
12682 parsing them this should not create any compatibility issues.
12686 * New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when
12687 calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without
12688 this HMAC (and other) operations are several times slower than OpenSSL
12693 * Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex().
12695 *Peter Sylvester <Peter.Sylvester@EdelWeb.fr>*
12697 * Use the correct content when signing type "other".
12701 ### Changes between 0.9.7b and 0.9.7c [30 Sep 2003]
12703 * Fix various bugs revealed by running the NISCC test suite:
12705 Stop out of bounds reads in the ASN1 code when presented with
12706 invalid tags (CVE-2003-0543 and CVE-2003-0544).
12708 Free up ASN1_TYPE correctly if ANY type is invalid ([CVE-2003-0545]).
12710 If verify callback ignores invalid public key errors don't try to check
12711 certificate signature with the NULL public key.
12715 * New -ignore_err option in ocsp application to stop the server
12716 exiting on the first error in a request.
12720 * In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
12721 if the server requested one: as stated in TLS 1.0 and SSL 3.0
12726 * In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
12727 extra data after the compression methods not only for TLS 1.0
12728 but also for SSL 3.0 (as required by the specification).
12730 *Bodo Moeller; problem pointed out by Matthias Loepfe*
12732 * Change X509_certificate_type() to mark the key as exported/exportable
12733 when it's 512 *bits* long, not 512 bytes.
12737 * Change AES_cbc_encrypt() so it outputs exact multiple of
12738 blocks during encryption.
12742 * Various fixes to base64 BIO and non blocking I/O. On write
12743 flushes were not handled properly if the BIO retried. On read
12744 data was not being buffered properly and had various logic bugs.
12745 This also affects blocking I/O when the data being decoded is a
12750 * Various S/MIME bugfixes and compatibility changes:
12751 output correct application/pkcs7 MIME type if
12752 PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures.
12753 Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening
12754 of files as .eml work). Correctly handle very long lines in MIME
12759 ### Changes between 0.9.7a and 0.9.7b [10 Apr 2003]
12761 * Countermeasure against the Klima-Pokorny-Rosa extension of
12762 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
12763 a protocol version number mismatch like a decryption error
12764 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
12768 * Turn on RSA blinding by default in the default implementation
12769 to avoid a timing attack. Applications that don't want it can call
12770 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
12771 They would be ill-advised to do so in most cases.
12773 *Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller*
12775 * Change RSA blinding code so that it works when the PRNG is not
12776 seeded (in this case, the secret RSA exponent is abused as
12777 an unpredictable seed -- if it is not unpredictable, there
12778 is no point in blinding anyway). Make RSA blinding thread-safe
12779 by remembering the creator's thread ID in rsa->blinding and
12780 having all other threads use local one-time blinding factors
12781 (this requires more computation than sharing rsa->blinding, but
12782 avoids excessive locking; and if an RSA object is not shared
12783 between threads, blinding will still be very fast).
12787 * Fixed a typo bug that would cause ENGINE_set_default() to set an
12788 ENGINE as defaults for all supported algorithms irrespective of
12789 the 'flags' parameter. 'flags' is now honoured, so applications
12790 should make sure they are passing it correctly.
12794 * Target "mingw" now allows native Windows code to be generated in
12795 the Cygwin environment as well as with the MinGW compiler.
12799 ### Changes between 0.9.7 and 0.9.7a [19 Feb 2003]
12801 * In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
12802 via timing by performing a MAC computation even if incorrect
12803 block cipher padding has been found. This is a countermeasure
12804 against active attacks where the attacker has to distinguish
12805 between bad padding and a MAC verification error. ([CVE-2003-0078])
12807 *Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
12808 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
12809 Martin Vuagnoux (EPFL, Ilion)*
12811 * Make the no-err option work as intended. The intention with no-err
12812 is not to have the whole error stack handling routines removed from
12813 libcrypto, it's only intended to remove all the function name and
12814 reason texts, thereby removing some of the footprint that may not
12815 be interesting if those errors aren't displayed anyway.
12817 NOTE: it's still possible for any application or module to have its
12818 own set of error texts inserted. The routines are there, just not
12819 used by default when no-err is given.
12823 * Add support for FreeBSD on IA64.
12825 *dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454*
12827 * Adjust DES_cbc_cksum() so it returns the same value as the MIT
12828 Kerberos function mit_des_cbc_cksum(). Before this change,
12829 the value returned by DES_cbc_cksum() was like the one from
12830 mit_des_cbc_cksum(), except the bytes were swapped.
12832 *Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte*
12834 * Allow an application to disable the automatic SSL chain building.
12835 Before this a rather primitive chain build was always performed in
12836 ssl3_output_cert_chain(): an application had no way to send the
12837 correct chain if the automatic operation produced an incorrect result.
12839 Now the chain builder is disabled if either:
12841 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert().
12843 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set.
12845 The reasoning behind this is that an application would not want the
12846 auto chain building to take place if extra chain certificates are
12847 present and it might also want a means of sending no additional
12848 certificates (for example the chain has two certificates and the
12853 * Add the possibility to build without the ENGINE framework.
12855 *Steven Reddie <smr@essemer.com.au> via Richard Levitte*
12857 * Under Win32 gmtime() can return NULL: check return value in
12858 OPENSSL_gmtime(). Add error code for case where gmtime() fails.
12862 * DSA routines: under certain error conditions uninitialized BN objects
12863 could be freed. Solution: make sure initialization is performed early
12864 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>,
12865 Nils Larsch <nla@trustcenter.de> via PR#459)
12869 * Another fix for SSLv2 session ID handling: the session ID was incorrectly
12870 checked on reconnect on the client side, therefore session resumption
12871 could still fail with a "ssl session id is different" error. This
12872 behaviour is masked when SSL_OP_ALL is used due to
12873 SSL_OP_MICROSOFT_SESS_ID_BUG being set.
12874 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
12875 followup to PR #377.
12879 * IA-32 assembler support enhancements: unified ELF targets, support
12880 for SCO/Caldera platforms, fix for Cygwin shared build.
12884 * Add support for FreeBSD on sparc64. As a consequence, support for
12885 FreeBSD on non-x86 processors is separate from x86 processors on
12886 the config script, much like the NetBSD support.
12888 *Richard Levitte & Kris Kennaway <kris@obsecurity.org>*
12890 ### Changes between 0.9.6h and 0.9.7 [31 Dec 2002]
12892 [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after
12895 * Fix session ID handling in SSLv2 client code: the SERVER FINISHED
12896 code (06) was taken as the first octet of the session ID and the last
12897 octet was ignored consequently. As a result SSLv2 client side session
12898 caching could not have worked due to the session ID mismatch between
12900 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
12905 * Change the declaration of needed Kerberos libraries to use EX_LIBS
12906 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is
12911 * The hw_ncipher.c engine requires dynamic locks. Unfortunately, it
12912 seems that in spite of existing for more than a year, many application
12913 author have done nothing to provide the necessary callbacks, which
12914 means that this particular engine will not work properly anywhere.
12915 This is a very unfortunate situation which forces us, in the name
12916 of usability, to give the hw_ncipher.c a static lock, which is part
12918 NOTE: This is for the 0.9.7 series ONLY. This hack will never
12919 appear in 0.9.8 or later. We EXPECT application authors to have
12920 dealt properly with this when 0.9.8 is released (unless we actually
12921 make such changes in the libcrypto locking code that changes will
12922 have to be made anyway).
12926 * In asn1_d2i_read_bio() repeatedly call BIO_read() until all content
12927 octets have been read, EOF or an error occurs. Without this change
12928 some truncated ASN1 structures will not produce an error.
12932 * Disable Heimdal support, since it hasn't been fully implemented.
12933 Still give the possibility to force the use of Heimdal, but with
12934 warnings and a request that patches get sent to openssl-dev.
12938 * Add the VC-CE target, introduce the WINCE sysname, and add
12939 INSTALL.WCE and appropriate conditionals to make it build.
12941 *Steven Reddie <smr@essemer.com.au> via Richard Levitte*
12943 * Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and
12944 cygssl-x.y.z.dll, where x, y and z are the major, minor and
12945 edit numbers of the version.
12947 *Corinna Vinschen <vinschen@redhat.com> and Richard Levitte*
12949 * Introduce safe string copy and catenation functions
12950 (BUF_strlcpy() and BUF_strlcat()).
12952 *Ben Laurie (CHATS) and Richard Levitte*
12954 * Avoid using fixed-size buffers for one-line DNs.
12956 *Ben Laurie (CHATS)*
12958 * Add BUF_MEM_grow_clean() to avoid information leakage when
12959 resizing buffers containing secrets, and use where appropriate.
12961 *Ben Laurie (CHATS)*
12963 * Avoid using fixed size buffers for configuration file location.
12965 *Ben Laurie (CHATS)*
12967 * Avoid filename truncation for various CA files.
12969 *Ben Laurie (CHATS)*
12971 * Use sizeof in preference to magic numbers.
12973 *Ben Laurie (CHATS)*
12975 * Avoid filename truncation in cert requests.
12977 *Ben Laurie (CHATS)*
12979 * Add assertions to check for (supposedly impossible) buffer
12982 *Ben Laurie (CHATS)*
12984 * Don't cache truncated DNS entries in the local cache (this could
12985 potentially lead to a spoofing attack).
12987 *Ben Laurie (CHATS)*
12989 * Fix various buffers to be large enough for hex/decimal
12990 representations in a platform independent manner.
12992 *Ben Laurie (CHATS)*
12994 * Add CRYPTO_realloc_clean() to avoid information leakage when
12995 resizing buffers containing secrets, and use where appropriate.
12997 *Ben Laurie (CHATS)*
12999 * Add BIO_indent() to avoid much slightly worrying code to do
13002 *Ben Laurie (CHATS)*
13004 * Convert sprintf()/BIO_puts() to BIO_printf().
13006 *Ben Laurie (CHATS)*
13008 * buffer_gets() could terminate with the buffer only half
13011 *Ben Laurie (CHATS)*
13013 * Add assertions to prevent user-supplied crypto functions from
13014 overflowing internal buffers by having large block sizes, etc.
13016 *Ben Laurie (CHATS)*
13018 * New OPENSSL_assert() macro (similar to assert(), but enabled
13021 *Ben Laurie (CHATS)*
13023 * Eliminate unused copy of key in RC4.
13025 *Ben Laurie (CHATS)*
13027 * Eliminate unused and incorrectly sized buffers for IV in pem.h.
13029 *Ben Laurie (CHATS)*
13031 * Fix off-by-one error in EGD path.
13033 *Ben Laurie (CHATS)*
13035 * If RANDFILE path is too long, ignore instead of truncating.
13037 *Ben Laurie (CHATS)*
13039 * Eliminate unused and incorrectly sized X.509 structure
13042 *Ben Laurie (CHATS)*
13044 * Eliminate unused and dangerous function knumber().
13046 *Ben Laurie (CHATS)*
13048 * Eliminate unused and dangerous structure, KSSL_ERR.
13050 *Ben Laurie (CHATS)*
13052 * Protect against overlong session ID context length in an encoded
13053 session object. Since these are local, this does not appear to be
13056 *Ben Laurie (CHATS)*
13058 * Change from security patch (see 0.9.6e below) that did not affect
13059 the 0.9.6 release series:
13061 Remote buffer overflow in SSL3 protocol - an attacker could
13062 supply an oversized master key in Kerberos-enabled versions.
13065 *Ben Laurie (CHATS)*
13067 * Change the SSL kerb5 codes to match RFC 2712.
13071 * Make -nameopt work fully for req and add -reqopt switch.
13073 *Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson*
13075 * The "block size" for block ciphers in CFB and OFB mode should be 1.
13077 *Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>*
13079 * Make sure tests can be performed even if the corresponding algorithms
13080 have been removed entirely. This was also the last step to make
13081 OpenSSL compilable with DJGPP under all reasonable conditions.
13083 *Richard Levitte, Doug Kaufman <dkaufman@rahul.net>*
13085 * Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT
13086 to allow version independent disabling of normally unselected ciphers,
13087 which may be activated as a side-effect of selecting a single cipher.
13089 (E.g., cipher list string "RSA" enables ciphersuites that are left
13090 out of "ALL" because they do not provide symmetric encryption.
13091 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.)
13093 *Lutz Jaenicke, Bodo Moeller*
13095 * Add appropriate support for separate platform-dependent build
13096 directories. The recommended way to make a platform-dependent
13097 build directory is the following (tested on Linux), maybe with
13100 # Place yourself outside of the OpenSSL source tree. In
13101 # this example, the environment variable OPENSSL_SOURCE
13102 # is assumed to contain the absolute OpenSSL source directory.
13103 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`"
13104 cd objtree/"`uname -s`-`uname -r`-`uname -m`"
13105 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do
13106 mkdir -p `dirname $F`
13107 ln -s $OPENSSL_SOURCE/$F $F
13110 To be absolutely sure not to disturb the source tree, a "make clean"
13111 is a good thing. If it isn't successful, don't worry about it,
13112 it probably means the source directory is very clean.
13116 * Make sure any ENGINE control commands make local copies of string
13117 pointers passed to them whenever necessary. Otherwise it is possible
13118 the caller may have overwritten (or deallocated) the original string
13119 data when a later ENGINE operation tries to use the stored values.
13121 *Götz Babin-Ebell <babinebell@trustcenter.de>*
13123 * Improve diagnostics in file reading and command-line digests.
13125 *Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>*
13127 * Add AES modes CFB and OFB to the object database. Correct an
13128 error in AES-CFB decryption.
13132 * Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this
13133 allows existing EVP_CIPHER_CTX structures to be reused after
13134 calling `EVP_*Final()`. This behaviour is used by encryption
13135 BIOs and some applications. This has the side effect that
13136 applications must explicitly clean up cipher contexts with
13137 EVP_CIPHER_CTX_cleanup() or they will leak memory.
13141 * Check the values of dna and dnb in bn_mul_recursive before calling
13142 bn_mul_comba (a non zero value means the a or b arrays do not contain
13143 n2 elements) and fallback to bn_mul_normal if either is not zero.
13147 * Fix escaping of non-ASCII characters when using the -subj option
13148 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>)
13152 * Make object definitions compliant to LDAP (RFC2256): SN is the short
13153 form for "surname", serialNumber has no short form.
13154 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798;
13155 therefore remove "mail" short name for "internet 7".
13156 The OID for unique identifiers in X509 certificates is
13157 x500UniqueIdentifier, not uniqueIdentifier.
13158 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>)
13162 * Add an "init" command to the ENGINE config module and auto initialize
13163 ENGINEs. Without any "init" command the ENGINE will be initialized
13164 after all ctrl commands have been executed on it. If init=1 the
13165 ENGINE is initialized at that point (ctrls before that point are run
13166 on the uninitialized ENGINE and after on the initialized one). If
13167 init=0 then the ENGINE will not be initialized at all.
13171 * Fix the 'app_verify_callback' interface so that the user-defined
13172 argument is actually passed to the callback: In the
13173 SSL_CTX_set_cert_verify_callback() prototype, the callback
13174 declaration has been changed from
13177 int (*cb)(X509_STORE_CTX *,void *);
13178 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
13179 i=s->ctx->app_verify_callback(&ctx)
13180 has been changed into
13181 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
13183 To update applications using SSL_CTX_set_cert_verify_callback(),
13184 a dummy argument can be added to their callback functions.
13186 *D. K. Smetters <smetters@parc.xerox.com>*
13188 * Added the '4758cca' ENGINE to support IBM 4758 cards.
13190 *Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe*
13192 * Add and OPENSSL_LOAD_CONF define which will cause
13193 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
13194 This allows older applications to transparently support certain
13195 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
13196 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
13197 load the config file and OPENSSL_add_all_algorithms_conf() which will
13198 always load it have also been added.
13202 * Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
13203 Adjust NIDs and EVP layer.
13205 *Stephen Sprunk <stephen@sprunk.org> and Richard Levitte*
13207 * Config modules support in openssl utility.
13209 Most commands now load modules from the config file,
13210 though in a few (such as version) this isn't done
13211 because it couldn't be used for anything.
13213 In the case of ca and req the config file used is
13214 the same as the utility itself: that is the -config
13215 command line option can be used to specify an
13220 * Move default behaviour from OPENSSL_config(). If appname is NULL
13221 use "openssl_conf" if filename is NULL use default openssl config file.
13225 * Add an argument to OPENSSL_config() to allow the use of an alternative
13226 config section name. Add a new flag to tolerate a missing config file
13227 and move code to CONF_modules_load_file().
13231 * Support for crypto accelerator cards from Accelerated Encryption
13232 Processing, www.aep.ie. (Use engine 'aep')
13233 The support was copied from 0.9.6c [engine] and adapted/corrected
13234 to work with the new engine framework.
13236 *AEP Inc. and Richard Levitte*
13238 * Support for SureWare crypto accelerator cards from Baltimore
13239 Technologies. (Use engine 'sureware')
13240 The support was copied from 0.9.6c [engine] and adapted
13241 to work with the new engine framework.
13245 * Have the CHIL engine fork-safe (as defined by nCipher) and actually
13246 make the newer ENGINE framework commands for the CHIL engine work.
13248 *Toomas Kiisk <vix@cyber.ee> and Richard Levitte*
13250 * Make it possible to produce shared libraries on ReliantUNIX.
13252 *Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte*
13254 * Add the configuration target debug-linux-ppro.
13255 Make 'openssl rsa' use the general key loading routines
13256 implemented in `apps.c`, and make those routines able to
13257 handle the key format FORMAT_NETSCAPE and the variant
13260 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
13262 * Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
13264 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
13266 * Add -keyform to rsautl, and document -engine.
13268 *Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>*
13270 * Change BIO_new_file (crypto/bio/bss_file.c) to use new
13271 BIO_R_NO_SUCH_FILE error code rather than the generic
13272 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
13276 * Add new functions
13277 ERR_peek_last_error
13278 ERR_peek_last_error_line
13279 ERR_peek_last_error_line_data.
13280 These are similar to
13282 ERR_peek_error_line
13283 ERR_peek_error_line_data,
13284 but report on the latest error recorded rather than the first one
13285 still in the error queue.
13287 *Ben Laurie, Bodo Moeller*
13289 * default_algorithms option in ENGINE config module. This allows things
13291 default_algorithms = ALL
13292 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
13296 * Preliminary ENGINE config module.
13300 * New experimental application configuration code.
13304 * Change the AES code to follow the same name structure as all other
13305 symmetric ciphers, and behave the same way. Move everything to
13306 the directory crypto/aes, thereby obsoleting crypto/rijndael.
13308 *Stephen Sprunk <stephen@sprunk.org> and Richard Levitte*
13310 * SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
13312 *Ben Laurie and Theo de Raadt*
13314 * Add option to output public keys in req command.
13316 *Massimiliano Pala madwolf@openca.org*
13318 * Use wNAFs in EC_POINTs_mul() for improved efficiency
13319 (up to about 10% better than before for P-192 and P-224).
13323 * New functions/macros
13325 SSL_CTX_set_msg_callback(ctx, cb)
13326 SSL_CTX_set_msg_callback_arg(ctx, arg)
13327 SSL_set_msg_callback(ssl, cb)
13328 SSL_set_msg_callback_arg(ssl, arg)
13330 to request calling a callback function
13332 void cb(int write_p, int version, int content_type,
13333 const void *buf, size_t len, SSL *ssl, void *arg)
13335 whenever a protocol message has been completely received
13336 (write_p == 0) or sent (write_p == 1). Here 'version' is the
13337 protocol version according to which the SSL library interprets
13338 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
13339 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
13340 the content type as defined in the SSL 3.0/TLS 1.0 protocol
13341 specification (change_cipher_spec(20), alert(21), handshake(22)).
13342 'buf' and 'len' point to the actual message, 'ssl' to the
13343 SSL object, and 'arg' is the application-defined value set by
13344 SSL[_CTX]_set_msg_callback_arg().
13346 'openssl s_client' and 'openssl s_server' have new '-msg' options
13347 to enable a callback that displays all protocol messages.
13351 * Change the shared library support so shared libraries are built as
13352 soon as the corresponding static library is finished, and thereby get
13353 openssl and the test programs linked against the shared library.
13354 This still only happens when the keyword "shard" has been given to
13355 the configuration scripts.
13357 NOTE: shared library support is still an experimental thing, and
13358 backward binary compatibility is still not guaranteed.
13360 *"Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte*
13362 * Add support for Subject Information Access extension.
13364 *Peter Sylvester <Peter.Sylvester@EdelWeb.fr>*
13366 * Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
13367 additional bytes when new memory had to be allocated, not just
13368 when reusing an existing buffer.
13372 * New command line and configuration option 'utf8' for the req command.
13373 This allows field values to be specified as UTF8 strings.
13377 * Add -multi and -mr options to "openssl speed" - giving multiple parallel
13378 runs for the former and machine-readable output for the latter.
13382 * Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
13383 of the e-mail address in the DN (i.e., it will go into a certificate
13384 extension only). The new configuration file option 'email_in_dn = no'
13385 has the same effect.
13387 *Massimiliano Pala madwolf@openca.org*
13389 * Change all functions with names starting with `des_` to be starting
13390 with `DES_` instead. Add wrappers that are compatible with libdes,
13391 but are named `_ossl_old_des_*`. Finally, add macros that map the
13392 `des_*` symbols to the corresponding `_ossl_old_des_*` if libdes
13393 compatibility is desired. If OpenSSL 0.9.6c compatibility is
13394 desired, the `des_*` symbols will be mapped to `DES_*`, with one
13397 Since we provide two compatibility mappings, the user needs to
13398 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
13399 compatibility is desired. The default (i.e., when that macro
13400 isn't defined) is OpenSSL 0.9.6c compatibility.
13402 There are also macros that enable and disable the support of old
13403 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
13404 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
13405 are defined, the default will apply: to support the old des routines.
13407 In either case, one must include openssl/des.h to get the correct
13408 definitions. Do not try to just include openssl/des_old.h, that
13411 NOTE: This is a major break of an old API into a new one. Software
13412 authors are encouraged to switch to the `DES_` style functions. Some
13413 time in the future, des_old.h and the libdes compatibility functions
13414 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
13415 default), and then completely removed.
13419 * Test for certificates which contain unsupported critical extensions.
13420 If such a certificate is found during a verify operation it is
13421 rejected by default: this behaviour can be overridden by either
13422 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
13423 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
13424 X509_supported_extension() has also been added which returns 1 if a
13425 particular extension is supported.
13429 * Modify the behaviour of EVP cipher functions in similar way to digests
13430 to retain compatibility with existing code.
13434 * Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
13435 compatibility with existing code. In particular the 'ctx' parameter does
13436 not have to be to be initialized before the call to EVP_DigestInit() and
13437 it is tidied up after a call to EVP_DigestFinal(). New function
13438 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
13439 EVP_MD_CTX_copy() changed to not require the destination to be
13440 initialized valid and new function EVP_MD_CTX_copy_ex() added which
13441 requires the destination to be valid.
13443 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
13444 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
13448 * Change ssl3_get_message (ssl/s3_both.c) and the functions using it
13449 so that complete 'Handshake' protocol structures are kept in memory
13450 instead of overwriting 'msg_type' and 'length' with 'body' data.
13454 * Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
13456 *Massimo Santin via Richard Levitte*
13458 * Major restructuring to the underlying ENGINE code. This includes
13459 reduction of linker bloat, separation of pure "ENGINE" manipulation
13460 (initialisation, etc) from functionality dealing with implementations
13461 of specific crypto interfaces. This change also introduces integrated
13462 support for symmetric ciphers and digest implementations - so ENGINEs
13463 can now accelerate these by providing EVP_CIPHER and EVP_MD
13464 implementations of their own. This is detailed in
13465 [crypto/engine/README.md](crypto/engine/README.md)
13466 as it couldn't be adequately described here. However, there are a few
13467 API changes worth noting - some RSA, DSA, DH, and RAND functions that
13468 were changed in the original introduction of ENGINE code have now
13469 reverted back - the hooking from this code to ENGINE is now a good
13470 deal more passive and at run-time, operations deal directly with
13471 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
13472 dereferencing through an ENGINE pointer any more. Also, the ENGINE
13473 functions dealing with `BN_MOD_EXP[_CRT]` handlers have been removed -
13474 they were not being used by the framework as there is no concept of a
13475 BIGNUM_METHOD and they could not be generalised to the new
13476 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
13477 ENGINE_cpy() has been removed as it cannot be consistently defined in
13482 * Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
13486 * Change mkdef.pl to sort symbols that get the same entry number,
13487 and make sure the automatically generated functions `ERR_load_*`
13488 become part of libeay.num as well.
13492 * New function SSL_renegotiate_pending(). This returns true once
13493 renegotiation has been requested (either SSL_renegotiate() call
13494 or HelloRequest/ClientHello received from the peer) and becomes
13495 false once a handshake has been completed.
13496 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
13497 sends a HelloRequest, but does not ensure that a handshake takes
13498 place. SSL_renegotiate_pending() is useful for checking if the
13499 client has followed the request.)
13503 * New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
13504 By default, clients may request session resumption even during
13505 renegotiation (if session ID contexts permit); with this option,
13506 session resumption is possible only in the first handshake.
13508 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes
13509 more bits available for options that should not be part of
13510 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION).
13514 * Add some demos for certificate and certificate request creation.
13518 * Make maximum certificate chain size accepted from the peer application
13519 settable (`SSL*_get/set_max_cert_list()`), as proposed by
13520 "Douglas E. Engert" <deengert@anl.gov>.
13524 * Add support for shared libraries for Unixware-7
13525 (Boyd Lynn Gerber <gerberb@zenez.com>).
13529 * Add a "destroy" handler to ENGINEs that allows structural cleanup to
13530 be done prior to destruction. Use this to unload error strings from
13531 ENGINEs that load their own error strings. NB: This adds two new API
13532 functions to "get" and "set" this destroy handler in an ENGINE.
13536 * Alter all existing ENGINE implementations (except "openssl" and
13537 "openbsd") to dynamically instantiate their own error strings. This
13538 makes them more flexible to be built both as statically-linked ENGINEs
13539 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
13540 Also, add stub code to each that makes building them as self-contained
13541 shared-libraries easier (see [README-Engine.md](README-Engine.md)).
13545 * Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
13546 implementations into applications that are completely implemented in
13547 self-contained shared-libraries. The "dynamic" ENGINE exposes control
13548 commands that can be used to configure what shared-library to load and
13549 to control aspects of the way it is handled. Also, made an update to
13550 the [README-Engine.md](README-Engine.md) file
13551 that brings its information up-to-date and
13552 provides some information and instructions on the "dynamic" ENGINE
13553 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
13557 * Make it possible to unload ranges of ERR strings with a new
13558 "ERR_unload_strings" function.
13562 * Add a copy() function to EVP_MD.
13566 * Make EVP_MD routines take a context pointer instead of just the
13567 md_data void pointer.
13571 * Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
13572 that the digest can only process a single chunk of data
13573 (typically because it is provided by a piece of
13574 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
13575 is only going to provide a single chunk of data, and hence the
13576 framework needn't accumulate the data for oneshot drivers.
13580 * As with "ERR", make it possible to replace the underlying "ex_data"
13581 functions. This change also alters the storage and management of global
13582 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
13583 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
13584 index counters. The API functions that use this state have been changed
13585 to take a "class_index" rather than pointers to the class's local STACK
13586 and counter, and there is now an API function to dynamically create new
13587 classes. This centralisation allows us to (a) plug a lot of the
13588 thread-safety problems that existed, and (b) makes it possible to clean
13589 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
13590 such data would previously have always leaked in application code and
13591 workarounds were in place to make the memory debugging turn a blind eye
13592 to it. Application code that doesn't use this new function will still
13593 leak as before, but their memory debugging output will announce it now
13594 rather than letting it slide.
13596 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
13597 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
13598 has a return value to indicate success or failure.
13602 * Make it possible to replace the underlying "ERR" functions such that the
13603 global state (2 LHASH tables and 2 locks) is only used by the "default"
13604 implementation. This change also adds two functions to "get" and "set"
13605 the implementation prior to it being automatically set the first time
13606 any other ERR function takes place. Ie. an application can call "get",
13607 pass the return value to a module it has just loaded, and that module
13608 can call its own "set" function using that value. This means the
13609 module's "ERR" operations will use (and modify) the error state in the
13610 application and not in its own statically linked copy of OpenSSL code.
13614 * Give DH, DSA, and RSA types their own `*_up_ref()` function to increment
13615 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
13616 the operation, and provides a more encapsulated way for external code
13617 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
13618 to use these functions rather than manually incrementing the counts.
13620 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
13624 * Add EVP test program.
13628 * Add symmetric cipher support to ENGINE. Expect the API to change!
13632 * New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
13633 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
13634 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
13635 These allow a CRL to be built without having to access X509_CRL fields
13636 directly. Modify 'ca' application to use new functions.
13640 * Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
13641 bug workarounds. Rollback attack detection is a security feature.
13642 The problem will only arise on OpenSSL servers when TLSv1 is not
13643 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
13644 Software authors not wanting to support TLSv1 will have special reasons
13645 for their choice and can explicitly enable this option.
13647 *Bodo Moeller, Lutz Jaenicke*
13649 * Rationalise EVP so it can be extended: don't include a union of
13650 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX
13651 (similar to those existing for EVP_CIPHER_CTX).
13656 EVP_MD_CTX_init(&md); /* new function call */
13657 EVP_DigestInit(&md, EVP_sha1());
13658 EVP_DigestUpdate(&md, in, len);
13659 EVP_DigestFinal(&md, out, NULL);
13660 EVP_MD_CTX_cleanup(&md); /* new function call */
13664 * Make DES key schedule conform to the usual scheme, as well as
13665 correcting its structure. This means that calls to DES functions
13666 now have to pass a pointer to a des_key_schedule instead of a
13667 plain des_key_schedule (which was actually always a pointer
13670 des_key_schedule ks;
13672 des_set_key_checked(..., &ks);
13673 des_ncbc_encrypt(..., &ks, ...);
13675 (Note that a later change renames 'des_...' into 'DES_...'.)
13679 * Initial reduction of linker bloat: the use of some functions, such as
13680 PEM causes large amounts of unused functions to be linked in due to
13681 poor organisation. For example pem_all.c contains every PEM function
13682 which has a knock on effect of linking in large amounts of (unused)
13683 ASN1 code. Grouping together similar functions and splitting unrelated
13684 functions prevents this.
13688 * Cleanup of EVP macros.
13692 * Change historical references to `{NID,SN,LN}_des_ede` and ede3 to add the
13693 correct `_ecb suffix`.
13697 * Add initial OCSP responder support to ocsp application. The
13698 revocation information is handled using the text based index
13699 use by the ca application. The responder can either handle
13700 requests generated internally, supplied in files (for example
13701 via a CGI script) or using an internal minimal server.
13705 * Add configuration choices to get zlib compression for TLS.
13709 * Changes to Kerberos SSL for RFC 2712 compliance:
13710 1. Implemented real KerberosWrapper, instead of just using
13711 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
13712 2. Implemented optional authenticator field of KerberosWrapper.
13714 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
13715 and authenticator structs; see crypto/krb5/.
13717 Generalized Kerberos calls to support multiple Kerberos libraries.
13718 *Vern Staats <staatsvr@asc.hpc.mil>, Jeffrey Altman <jaltman@columbia.edu>
13719 via Richard Levitte*
13721 * Cause 'openssl speed' to use fully hard-coded DSA keys as it
13722 already does with RSA. testdsa.h now has 'priv_key/pub_key'
13723 values for each of the key sizes rather than having just
13724 parameters (and 'speed' generating keys each time).
13728 * Speed up EVP routines.
13731 pe 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
13732 s-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
13733 s-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
13734 s-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
13736 s-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
13737 s-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
13738 s-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
13741 s-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
13743 s-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
13747 * Added the OS2-EMX target.
13749 *"Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte*
13751 * Rewrite commands to use `NCONF` routines instead of the old `CONF`.
13752 New functions to support `NCONF` routines in extension code.
13753 New function `CONF_set_nconf()`
13754 to allow functions which take an `NCONF` to also handle the old `LHASH`
13755 structure: this means that the old `CONF` compatible routines can be
13756 retained (in particular w.rt. extensions) without having to duplicate the
13757 code. New function `X509V3_add_ext_nconf_sk()` to add extensions to a stack.
13761 * Enhance the general user interface with mechanisms for inner control
13762 and with possibilities to have yes/no kind of prompts.
13766 * Change all calls to low-level digest routines in the library and
13767 applications to use EVP. Add missing calls to HMAC_cleanup() and
13768 don't assume HMAC_CTX can be copied using memcpy().
13770 *Verdon Walker <VWalker@novell.com>, Steve Henson*
13772 * Add the possibility to control engines through control names but with
13773 arbitrary arguments instead of just a string.
13774 Change the key loaders to take a UI_METHOD instead of a callback
13775 function pointer. NOTE: this breaks binary compatibility with earlier
13776 versions of OpenSSL [engine].
13777 Adapt the nCipher code for these new conditions and add a card insertion
13782 * Enhance the general user interface with mechanisms to better support
13783 dialog box interfaces, application-defined prompts, the possibility
13784 to use defaults (for example default passwords from somewhere else)
13785 and interrupts/cancellations.
13789 * Tidy up PKCS#12 attribute handling. Add support for the CSP name
13790 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
13794 * Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
13795 tidy up some unnecessarily weird code in 'sk_new()').
13797 *Geoff, reported by Diego Tartara <dtartara@novamens.com>*
13799 * Change the key loading routines for ENGINEs to use the same kind
13800 callback (pem_password_cb) as all other routines that need this
13805 * Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
13806 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
13807 than this minimum value is recommended.
13811 * New random seeder for OpenVMS, using the system process statistics
13812 that are easily reachable.
13816 * Windows apparently can't transparently handle global
13817 variables defined in DLLs. Initialisations such as:
13819 const ASN1_ITEM *it = &ASN1_INTEGER_it;
13821 won't compile. This is used by the any applications that need to
13822 declare their own ASN1 modules. This was fixed by adding the option
13823 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
13824 needed for static libraries under Win32.
13828 * New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
13829 setting of purpose and trust fields. New X509_STORE trust and
13830 purpose functions and tidy up setting in other SSL functions.
13834 * Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
13835 structure. These are inherited by X509_STORE_CTX when it is
13836 initialised. This allows various defaults to be set in the
13837 X509_STORE structure (such as flags for CRL checking and custom
13838 purpose or trust settings) for functions which only use X509_STORE_CTX
13839 internally such as S/MIME.
13841 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
13842 trust settings if they are not set in X509_STORE. This allows X509_STORE
13843 purposes and trust (in S/MIME for example) to override any set by default.
13845 Add command line options for CRL checking to smime, s_client and s_server
13850 * Initial CRL based revocation checking. If the CRL checking flag(s)
13851 are set then the CRL is looked up in the X509_STORE structure and
13852 its validity and signature checked, then if the certificate is found
13853 in the CRL the verify fails with a revoked error.
13855 Various new CRL related callbacks added to X509_STORE_CTX structure.
13857 Command line options added to 'verify' application to support this.
13859 This needs some additional work, such as being able to handle multiple
13860 CRLs with different times, extension based lookup (rather than just
13861 by subject name) and ultimately more complete V2 CRL extension
13866 * Add a general user interface API (crypto/ui/). This is designed
13867 to replace things like des_read_password and friends (backward
13868 compatibility functions using this new API are provided).
13869 The purpose is to remove prompting functions from the DES code
13870 section as well as provide for prompting through dialog boxes in
13871 a window system and the like.
13875 * Add "ex_data" support to ENGINE so implementations can add state at a
13876 per-structure level rather than having to store it globally.
13880 * Make it possible for ENGINE structures to be copied when retrieved by
13881 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
13882 This causes the "original" ENGINE structure to act like a template,
13883 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
13884 operational state can be localised to each ENGINE structure, despite the
13885 fact they all share the same "methods". New ENGINE structures returned in
13886 this case have no functional references and the return value is the single
13887 structural reference. This matches the single structural reference returned
13888 by ENGINE_by_id() normally, when it is incremented on the pre-existing
13893 * Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
13894 needs to match any other type at all we need to manually clear the
13899 * Changes to the "openssl engine" utility to include;
13900 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
13901 about an ENGINE's available control commands.
13902 - executing control commands from command line arguments using the
13903 '-pre' and '-post' switches. '-post' is only used if '-t' is
13904 specified and the ENGINE is successfully initialised. The syntax for
13905 the individual commands are colon-separated, for example;
13906 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
13910 * New dynamic control command support for ENGINEs. ENGINEs can now
13911 declare their own commands (numbers), names (strings), descriptions,
13912 and input types for run-time discovery by calling applications. A
13913 subset of these commands are implicitly classed as "executable"
13914 depending on their input type, and only these can be invoked through
13915 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
13916 can be based on user input, config files, etc). The distinction is
13917 that "executable" commands cannot return anything other than a boolean
13918 result and can only support numeric or string input, whereas some
13919 discoverable commands may only be for direct use through
13920 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
13921 pointers, or other custom uses. The "executable" commands are to
13922 support parameterisations of ENGINE behaviour that can be
13923 unambiguously defined by ENGINEs and used consistently across any
13924 OpenSSL-based application. Commands have been added to all the
13925 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
13926 control over shared-library paths without source code alterations.
13930 * Changed all ENGINE implementations to dynamically allocate their
13931 ENGINEs rather than declaring them statically. Apart from this being
13932 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
13933 this also allows the implementations to compile without using the
13934 internal engine_int.h header.
13938 * Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
13939 'const' value. Any code that should be able to modify a RAND_METHOD
13940 should already have non-const pointers to it (ie. they should only
13941 modify their own ones).
13945 * Made a variety of little tweaks to the ENGINE code.
13946 - "atalla" and "ubsec" string definitions were moved from header files
13947 to C code. "nuron" string definitions were placed in variables
13948 rather than hard-coded - allowing parameterisation of these values
13949 later on via ctrl() commands.
13950 - Removed unused "#if 0"'d code.
13951 - Fixed engine list iteration code so it uses ENGINE_free() to release
13952 structural references.
13953 - Constified the RAND_METHOD element of ENGINE structures.
13954 - Constified various get/set functions as appropriate and added
13955 missing functions (including a catch-all ENGINE_cpy that duplicates
13956 all ENGINE values onto a new ENGINE except reference counts/state).
13957 - Removed NULL parameter checks in get/set functions. Setting a method
13958 or function to NULL is a way of cancelling out a previously set
13959 value. Passing a NULL ENGINE parameter is just plain stupid anyway
13960 and doesn't justify the extra error symbols and code.
13961 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
13962 flags from engine_int.h to engine.h.
13963 - Changed prototypes for ENGINE handler functions (init(), finish(),
13964 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
13968 * Implement binary inversion algorithm for BN_mod_inverse in addition
13969 to the algorithm using long division. The binary algorithm can be
13970 used only if the modulus is odd. On 32-bit systems, it is faster
13971 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
13972 roughly 5-15% for 256-bit moduli), so we use it only for moduli
13973 up to 450 bits. In 64-bit environments, the binary algorithm
13974 appears to be advantageous for much longer moduli; here we use it
13975 for moduli up to 2048 bits.
13979 * Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
13980 could not support the combine flag in choice fields.
13984 * Add a 'copy_extensions' option to the 'ca' utility. This copies
13985 extensions from a certificate request to the certificate.
13989 * Allow multiple 'certopt' and 'nameopt' options to be separated
13990 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
13991 file: this allows the display of the certificate about to be
13992 signed to be customised, to allow certain fields to be included
13993 or excluded and extension details. The old system didn't display
13994 multicharacter strings properly, omitted fields not in the policy
13995 and couldn't display additional details such as extensions.
13999 * Function EC_POINTs_mul for multiple scalar multiplication
14000 of an arbitrary number of elliptic curve points
14001 \sum scalars[i]*points[i],
14002 optionally including the generator defined for the EC_GROUP:
14003 scalar*generator + \sum scalars[i]*points[i].
14005 EC_POINT_mul is a simple wrapper function for the typical case
14006 that the point list has just one item (besides the optional
14011 * First EC_METHODs for curves over GF(p):
14013 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
14014 operations and provides various method functions that can also
14015 operate with faster implementations of modular arithmetic.
14017 EC_GFp_mont_method() reuses most functions that are part of
14018 EC_GFp_simple_method, but uses Montgomery arithmetic.
14020 *Bodo Moeller; point addition and point doubling
14021 implementation directly derived from source code provided by
14022 Lenka Fibikova <fibikova@exp-math.uni-essen.de>*
14024 * Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
14025 crypto/ec/ec_lib.c):
14027 Curves are EC_GROUP objects (with an optional group generator)
14028 based on EC_METHODs that are built into the library.
14030 Points are EC_POINT objects based on EC_GROUP objects.
14032 Most of the framework would be able to handle curves over arbitrary
14033 finite fields, but as there are no obvious types for fields other
14034 than GF(p), some functions are limited to that for now.
14038 * Add the -HTTP option to s_server. It is similar to -WWW, but requires
14039 that the file contains a complete HTTP response.
14043 * Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
14044 change the def and num file printf format specifier from "%-40sXXX"
14045 to "%-39s XXX". The latter will always guarantee a space after the
14046 field while the former will cause them to run together if the field
14047 is 40 of more characters long.
14051 * Constify the cipher and digest 'method' functions and structures
14052 and modify related functions to take constant EVP_MD and EVP_CIPHER
14057 * Hide BN_CTX structure details in bn_lcl.h instead of publishing them
14058 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
14062 * Modify `EVP_Digest*()` routines so they now return values. Although the
14063 internal software routines can never fail additional hardware versions
14068 * Clean up crypto/err/err.h and change some error codes to avoid conflicts:
14070 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
14071 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
14074 ERR_R_NESTED_ASN1_ERROR
14076 ERR_R_MISSING_ASN1_EOS
14077 were 4 .. 9, conflicting with
14078 ERR_LIB_RSA (= ERR_R_RSA_LIB)
14080 ERR_LIB_PEM (= ERR_R_PEM_LIB).
14081 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
14083 Add new error code 'ERR_R_INTERNAL_ERROR'.
14087 * Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
14092 * New option '-subj arg' for 'openssl req' and 'openssl ca'. This
14093 sets the subject name for a new request or supersedes the
14094 subject name in a given request. Formats that can be parsed are
14095 'CN=Some Name, OU=myOU, C=IT'
14097 'CN=Some Name/OU=myOU/C=IT'.
14099 Add options '-batch' and '-verbose' to 'openssl req'.
14101 *Massimiliano Pala <madwolf@hackmasters.net>*
14103 * Introduce the possibility to access global variables through
14104 functions on platform were that's the best way to handle exporting
14105 global variables in shared libraries. To enable this functionality,
14106 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
14107 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
14108 is normally done by Configure or something similar).
14110 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
14111 in the source file (foo.c) like this:
14113 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
14114 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
14116 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
14117 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
14119 OPENSSL_DECLARE_GLOBAL(int,foo);
14120 #define foo OPENSSL_GLOBAL_REF(foo)
14121 OPENSSL_DECLARE_GLOBAL(double,bar);
14122 #define bar OPENSSL_GLOBAL_REF(bar)
14124 The #defines are very important, and therefore so is including the
14125 header file everywhere where the defined globals are used.
14127 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
14128 of ASN.1 items, but that structure is a bit different.
14130 The largest change is in util/mkdef.pl which has been enhanced with
14131 better and easier to understand logic to choose which symbols should
14132 go into the Windows .def files as well as a number of fixes and code
14133 cleanup (among others, algorithm keywords are now sorted
14134 lexicographically to avoid constant rewrites).
14138 * In BN_div() keep a copy of the sign of 'num' before writing the
14139 result to 'rm' because if rm==num the value will be overwritten
14140 and produce the wrong result if 'num' is negative: this caused
14141 problems with BN_mod() and BN_nnmod().
14145 * Function OCSP_request_verify(). This checks the signature on an
14146 OCSP request and verifies the signer certificate. The signer
14147 certificate is just checked for a generic purpose and OCSP request
14152 * Add OCSP_check_validity() function to check the validity of OCSP
14153 responses. OCSP responses are prepared in real time and may only
14154 be a few seconds old. Simply checking that the current time lies
14155 between thisUpdate and nextUpdate max reject otherwise valid responses
14156 caused by either OCSP responder or client clock inaccuracy. Instead
14157 we allow thisUpdate and nextUpdate to fall within a certain period of
14158 the current time. The age of the response can also optionally be
14159 checked. Two new options -validity_period and -status_age added to
14164 * If signature or public key algorithm is unrecognized print out its
14165 OID rather that just UNKNOWN.
14169 * Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
14170 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
14171 ID to be generated from the issuer certificate alone which can then be
14172 passed to OCSP_id_issuer_cmp().
14176 * New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
14177 ASN1 modules to export functions returning ASN1_ITEM pointers
14178 instead of the ASN1_ITEM structures themselves. This adds several
14179 new macros which allow the underlying ASN1 function/structure to
14180 be accessed transparently. As a result code should not use ASN1_ITEM
14181 references directly (such as &X509_it) but instead use the relevant
14182 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
14183 use of the new ASN1 code on platforms where exporting structures
14184 is problematical (for example in shared libraries) but exporting
14185 functions returning pointers to structures is not.
14189 * Add support for overriding the generation of SSL/TLS session IDs.
14190 These callbacks can be registered either in an SSL_CTX or per SSL.
14191 The purpose of this is to allow applications to control, if they wish,
14192 the arbitrary values chosen for use as session IDs, particularly as it
14193 can be useful for session caching in multiple-server environments. A
14194 command-line switch for testing this (and any client code that wishes
14195 to use such a feature) has been added to "s_server".
14197 *Geoff Thorpe, Lutz Jaenicke*
14199 * Modify mkdef.pl to recognise and parse preprocessor conditionals
14200 of the form `#if defined(...) || defined(...) || ...` and
14201 `#if !defined(...) && !defined(...) && ...`. This also avoids
14202 the growing number of special cases it was previously handling.
14206 * Make all configuration macros available for application by making
14207 sure they are available in opensslconf.h, by giving them names starting
14208 with `OPENSSL_` to avoid conflicts with other packages and by making
14209 sure e_os2.h will cover all platform-specific cases together with
14211 Additionally, it is now possible to define configuration/platform-
14212 specific names (called "system identities"). In the C code, these
14213 are prefixed with `OPENSSL_SYSNAME_`. e_os2.h will create another
14214 macro with the name beginning with `OPENSSL_SYS_`, which is determined
14215 from `OPENSSL_SYSNAME_*` or compiler-specific macros depending on
14220 * New option -set_serial to 'req' and 'x509' this allows the serial
14221 number to use to be specified on the command line. Previously self
14222 signed certificates were hard coded with serial number 0 and the
14223 CA options of 'x509' had to use a serial number in a file which was
14228 * New options to 'ca' utility to support V2 CRL entry extensions.
14229 Currently CRL reason, invalidity date and hold instruction are
14230 supported. Add new CRL extensions to V3 code and some new objects.
14234 * New function EVP_CIPHER_CTX_set_padding() this is used to
14235 disable standard block padding (aka PKCS#5 padding) in the EVP
14236 API, which was previously mandatory. This means that the data is
14237 not padded in any way and so the total length much be a multiple
14238 of the block size, otherwise an error occurs.
14242 * Initial (incomplete) OCSP SSL support.
14246 * New function OCSP_parse_url(). This splits up a URL into its host,
14247 port and path components: primarily to parse OCSP URLs. New -url
14248 option to ocsp utility.
14252 * New nonce behavior. The return value of OCSP_check_nonce() now
14253 reflects the various checks performed. Applications can decide
14254 whether to tolerate certain situations such as an absent nonce
14255 in a response when one was present in a request: the ocsp application
14256 just prints out a warning. New function OCSP_add1_basic_nonce()
14257 this is to allow responders to include a nonce in a response even if
14258 the request is nonce-less.
14262 * Disable stdin buffering in `load_cert()` (`apps/apps.c`) so that no certs are
14263 skipped when using openssl x509 multiple times on a single input file,
14264 e.g. `(openssl x509 -out cert1; openssl x509 -out cert2) <certs`.
14268 * Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
14269 set string type: to handle setting ASN1_TIME structures. Fix ca
14270 utility to correctly initialize revocation date of CRLs.
14274 * New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
14275 the clients preferred ciphersuites and rather use its own preferences.
14276 Should help to work around M$ SGC (Server Gated Cryptography) bug in
14277 Internet Explorer by ensuring unchanged hash method during stepup.
14278 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
14282 * Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
14283 to aes and add a new 'exist' option to print out symbols that don't
14288 * Additional options to ocsp utility to allow flags to be set and
14289 additional certificates supplied.
14293 * Add the option -VAfile to 'openssl ocsp', so the user can give the
14294 OCSP client a number of certificate to only verify the response
14299 * Update Rijndael code to version 3.0 and change EVP AES ciphers to
14300 handle the new API. Currently only ECB, CBC modes supported. Add new
14303 Add TLS AES ciphersuites as described in RFC3268, "Advanced
14304 Encryption Standard (AES) Ciphersuites for Transport Layer
14305 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were
14306 not enabled by default and were not part of the "ALL" ciphersuite
14307 alias because they were not yet official; they could be
14308 explicitly requested by specifying the "AESdraft" ciphersuite
14309 group alias. In the final release of OpenSSL 0.9.7, the group
14310 alias is called "AES" and is part of "ALL".)
14312 *Ben Laurie, Steve Henson, Bodo Moeller*
14314 * New function OCSP_copy_nonce() to copy nonce value (if present) from
14315 request to response.
14319 * Functions for OCSP responders. OCSP_request_onereq_count(),
14320 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
14321 extract information from a certificate request. OCSP_response_create()
14322 creates a response and optionally adds a basic response structure.
14323 OCSP_basic_add1_status() adds a complete single response to a basic
14324 response and returns the OCSP_SINGLERESP structure just added (to allow
14325 extensions to be included for example). OCSP_basic_add1_cert() adds a
14326 certificate to a basic response and OCSP_basic_sign() signs a basic
14327 response with various flags. New helper functions ASN1_TIME_check()
14328 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
14329 (converts ASN1_TIME to GeneralizedTime).
14333 * Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
14334 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
14335 structure from a certificate. X509_pubkey_digest() digests the public_key
14336 contents: this is used in various key identifiers.
14340 * Make sk_sort() tolerate a NULL argument.
14342 *Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>*
14344 * New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
14345 passed by the function are trusted implicitly. If any of them signed the
14346 response then it is assumed to be valid and is not verified.
14350 * In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
14351 to data. This was previously part of the PKCS7 ASN1 code. This
14352 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
14353 *Steve Henson, reported by Kenneth R. Robinette
14354 <support@securenetterm.com>*
14356 * Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
14357 routines: without these tracing memory leaks is very painful.
14358 Fix leaks in PKCS12 and PKCS7 routines.
14362 * Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
14363 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
14364 effectively meant GeneralizedTime would never be used. Now it
14365 is initialised to -1 but X509_time_adj() now has to check the value
14366 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
14367 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
14368 *Steve Henson, reported by Kenneth R. Robinette
14369 <support@securenetterm.com>*
14371 * Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
14372 result in a zero length in the ASN1_INTEGER structure which was
14373 not consistent with the structure when d2i_ASN1_INTEGER() was used
14374 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
14375 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
14376 where it did not print out a minus for negative ASN1_INTEGER.
14380 * Add summary printout to ocsp utility. The various functions which
14381 convert status values to strings have been renamed to:
14382 OCSP_response_status_str(), OCSP_cert_status_str() and
14383 OCSP_crl_reason_str() and are no longer static. New options
14384 to verify nonce values and to disable verification. OCSP response
14385 printout format cleaned up.
14389 * Add additional OCSP certificate checks. These are those specified
14390 in RFC2560. This consists of two separate checks: the CA of the
14391 certificate being checked must either be the OCSP signer certificate
14392 or the issuer of the OCSP signer certificate. In the latter case the
14393 OCSP signer certificate must contain the OCSP signing extended key
14394 usage. This check is performed by attempting to match the OCSP
14395 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
14396 in the OCSP_CERTID structures of the response.
14400 * Initial OCSP certificate verification added to OCSP_basic_verify()
14401 and related routines. This uses the standard OpenSSL certificate
14402 verify routines to perform initial checks (just CA validity) and
14403 to obtain the certificate chain. Then additional checks will be
14404 performed on the chain. Currently the root CA is checked to see
14405 if it is explicitly trusted for OCSP signing. This is used to set
14406 a root CA as a global signing root: that is any certificate that
14407 chains to that CA is an acceptable OCSP signing certificate.
14411 * New '-extfile ...' option to 'openssl ca' for reading X.509v3
14412 extensions from a separate configuration file.
14413 As when reading extensions from the main configuration file,
14414 the '-extensions ...' option may be used for specifying the
14417 *Massimiliano Pala <madwolf@comune.modena.it>*
14419 * New OCSP utility. Allows OCSP requests to be generated or
14420 read. The request can be sent to a responder and the output
14421 parsed, outputted or printed in text form. Not complete yet:
14422 still needs to check the OCSP response validity.
14426 * New subcommands for 'openssl ca':
14427 `openssl ca -status <serial>` prints the status of the cert with
14428 the given serial number (according to the index file).
14429 `openssl ca -updatedb` updates the expiry status of certificates
14432 *Massimiliano Pala <madwolf@comune.modena.it>*
14434 * New '-newreq-nodes' command option to CA.pl. This is like
14435 '-newreq', but calls 'openssl req' with the '-nodes' option
14436 so that the resulting key is not encrypted.
14438 *Damien Miller <djm@mindrot.org>*
14440 * New configuration for the GNU Hurd.
14442 *Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte*
14444 * Initial code to implement OCSP basic response verify. This
14445 is currently incomplete. Currently just finds the signer's
14446 certificate and verifies the signature on the response.
14450 * New SSLeay_version code SSLEAY_DIR to determine the compiled-in
14451 value of OPENSSLDIR. This is available via the new '-d' option
14452 to 'openssl version', and is also included in 'openssl version -a'.
14456 * Allowing defining memory allocation callbacks that will be given
14457 file name and line number information in additional arguments
14458 (a `const char*` and an int). The basic functionality remains, as
14459 well as the original possibility to just replace malloc(),
14460 realloc() and free() by functions that do not know about these
14461 additional arguments. To register and find out the current
14462 settings for extended allocation functions, the following
14463 functions are provided:
14465 CRYPTO_set_mem_ex_functions
14466 CRYPTO_set_locked_mem_ex_functions
14467 CRYPTO_get_mem_ex_functions
14468 CRYPTO_get_locked_mem_ex_functions
14470 These work the same way as CRYPTO_set_mem_functions and friends.
14471 `CRYPTO_get_[locked_]mem_functions` now writes 0 where such an
14472 extended allocation function is enabled.
14473 Similarly, `CRYPTO_get_[locked_]mem_ex_functions` writes 0 where
14474 a conventional allocation function is enabled.
14476 *Richard Levitte, Bodo Moeller*
14478 * Finish off removing the remaining LHASH function pointer casts.
14479 There should no longer be any prototype-casting required when using
14480 the LHASH abstraction, and any casts that remain are "bugs". See
14481 the callback types and macros at the head of lhash.h for details
14482 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
14486 * Add automatic query of EGD sockets in RAND_poll() for the unix variant.
14487 If /dev/[u]random devices are not available or do not return enough
14488 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
14490 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
14491 /etc/entropy will be queried once each in this sequence, querying stops
14492 when enough entropy was collected without querying more sockets.
14496 * Change the Unix RAND_poll() variant to be able to poll several
14497 random devices, as specified by DEVRANDOM, until a sufficient amount
14498 of data has been collected. We spend at most 10 ms on each file
14499 (select timeout) and read in non-blocking mode. DEVRANDOM now
14500 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
14501 (previously it was just the string "/dev/urandom"), so on typical
14502 platforms the 10 ms delay will never occur.
14503 Also separate out the Unix variant to its own file, rand_unix.c.
14504 For VMS, there's a currently-empty rand_vms.c.
14508 * Move OCSP client related routines to ocsp_cl.c. These
14509 provide utility functions which an application needing
14510 to issue a request to an OCSP responder and analyse the
14511 response will typically need: as opposed to those which an
14512 OCSP responder itself would need which will be added later.
14514 OCSP_request_sign() signs an OCSP request with an API similar
14515 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
14516 response. OCSP_response_get1_basic() extracts basic response
14517 from response. OCSP_resp_find_status(): finds and extracts status
14518 information from an OCSP_CERTID structure (which will be created
14519 when the request structure is built). These are built from lower
14520 level functions which work on OCSP_SINGLERESP structures but
14521 won't normally be used unless the application wishes to examine
14522 extensions in the OCSP response for example.
14524 Replace nonce routines with a pair of functions.
14525 OCSP_request_add1_nonce() adds a nonce value and optionally
14526 generates a random value. OCSP_check_nonce() checks the
14527 validity of the nonce in an OCSP response.
14531 * Change function OCSP_request_add() to OCSP_request_add0_id().
14532 This doesn't copy the supplied OCSP_CERTID and avoids the
14533 need to free up the newly created id. Change return type
14534 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
14535 This can then be used to add extensions to the request.
14536 Deleted OCSP_request_new(), since most of its functionality
14537 is now in OCSP_REQUEST_new() (and the case insensitive name
14538 clash) apart from the ability to set the request name which
14539 will be added elsewhere.
14543 * Update OCSP API. Remove obsolete extensions argument from
14544 various functions. Extensions are now handled using the new
14545 OCSP extension code. New simple OCSP HTTP function which
14546 can be used to send requests and parse the response.
14550 * Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
14551 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
14552 uses the special reorder version of SET OF to sort the attributes
14553 and reorder them to match the encoded order. This resolves a long
14554 standing problem: a verify on a PKCS7 structure just after signing
14555 it used to fail because the attribute order did not match the
14556 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
14557 it uses the received order. This is necessary to tolerate some broken
14558 software that does not order SET OF. This is handled by encoding
14559 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
14560 to produce the required SET OF.
14564 * Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
14565 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
14566 files to get correct declarations of the ASN.1 item variables.
14570 * Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
14571 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
14572 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
14573 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
14574 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
14575 ASN1_ITEM and no wrapper functions.
14579 * New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
14580 replace the old function pointer based I/O routines. Change most of
14581 the `*_d2i_bio()` and `*_d2i_fp()` functions to use these.
14585 * Enhance mkdef.pl to be more accepting about spacing in C preprocessor
14586 lines, recognize more "algorithms" that can be deselected, and make
14587 it complain about algorithm deselection that isn't recognised.
14591 * New ASN1 functions to handle dup, sign, verify, digest, pack and
14592 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
14593 to use new functions. Add NO_ASN1_OLD which can be set to remove
14594 some old style ASN1 functions: this can be used to determine if old
14595 code will still work when these eventually go away.
14599 * New extension functions for OCSP structures, these follow the
14600 same conventions as certificates and CRLs.
14604 * New function X509V3_add1_i2d(). This automatically encodes and
14605 adds an extension. Its behaviour can be customised with various
14606 flags to append, replace or delete. Various wrappers added for
14607 certificates and CRLs.
14611 * Fix to avoid calling the underlying ASN1 print routine when
14612 an extension cannot be parsed. Correct a typo in the
14613 OCSP_SERVICELOC extension. Tidy up print OCSP format.
14617 * Make mkdef.pl parse some of the ASN1 macros and add appropriate
14618 entries for variables.
14622 * Add functionality to `apps/openssl.c` for detecting locking
14623 problems: As the program is single-threaded, all we have
14624 to do is register a locking callback using an array for
14625 storing which locks are currently held by the program.
14629 * Use a lock around the call to CRYPTO_get_ex_new_index() in
14630 SSL_get_ex_data_X509_STORE_idx(), which is used in
14631 ssl_verify_cert_chain() and thus can be called at any time
14632 during TLS/SSL handshakes so that thread-safety is essential.
14633 Unfortunately, the ex_data design is not at all suited
14634 for multi-threaded use, so it probably should be abolished.
14638 * Added Broadcom "ubsec" ENGINE to OpenSSL.
14640 *Broadcom, tweaked and integrated by Geoff Thorpe*
14642 * Move common extension printing code to new function
14643 X509V3_print_extensions(). Reorganise OCSP print routines and
14644 implement some needed OCSP ASN1 functions. Add OCSP extensions.
14648 * New function X509_signature_print() to remove duplication in some
14653 * Add a special meaning when SET OF and SEQUENCE OF flags are both
14654 set (this was treated exactly the same as SET OF previously). This
14655 is used to reorder the STACK representing the structure to match the
14656 encoding. This will be used to get round a problem where a PKCS7
14657 structure which was signed could not be verified because the STACK
14658 order did not reflect the encoded order.
14662 * Reimplement the OCSP ASN1 module using the new code.
14666 * Update the X509V3 code to permit the use of an ASN1_ITEM structure
14667 for its ASN1 operations. The old style function pointers still exist
14668 for now but they will eventually go away.
14672 * Merge in replacement ASN1 code from the ASN1 branch. This almost
14673 completely replaces the old ASN1 functionality with a table driven
14674 encoder and decoder which interprets an ASN1_ITEM structure describing
14675 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
14676 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
14677 has also been converted to the new form.
14681 * Change BN_mod_exp_recp so that negative moduli are tolerated
14682 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
14683 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
14684 for negative moduli.
14688 * Fix BN_uadd and BN_usub: Always return non-negative results instead
14689 of not touching the result's sign bit.
14693 * BN_div bugfix: If the result is 0, the sign (res->neg) must not be
14698 * Changed the LHASH code to use prototypes for callbacks, and created
14699 macros to declare and implement thin (optionally static) functions
14700 that provide type-safety and avoid function pointer casting for the
14701 type-specific callbacks.
14705 * Added Kerberos Cipher Suites to be used with TLS, as written in
14707 *Veers Staats <staatsvr@asc.hpc.mil>,
14708 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte*
14710 * Reformat the FAQ so the different questions and answers can be divided
14711 in sections depending on the subject.
14715 * Have the zlib compression code load ZLIB.DLL dynamically under
14720 * New function BN_mod_sqrt for computing square roots modulo a prime
14721 (using the probabilistic Tonelli-Shanks algorithm unless
14722 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
14723 be handled deterministically).
14725 *Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller*
14727 * Make BN_mod_inverse faster by explicitly handling small quotients
14728 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
14729 512 bits], about 30% for larger ones [1024 or 2048 bits].)
14733 * New function BN_kronecker.
14737 * Fix BN_gcd so that it works on negative inputs; the result is
14738 positive unless both parameters are zero.
14739 Previously something reasonably close to an infinite loop was
14740 possible because numbers could be growing instead of shrinking
14741 in the implementation of Euclid's algorithm.
14745 * Fix BN_is_word() and BN_is_one() macros to take into account the
14746 sign of the number in question.
14748 Fix BN_is_word(a,w) to work correctly for w == 0.
14750 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
14751 because its test if the absolute value of 'a' equals 'w'.
14752 Note that BN_abs_is_word does *not* handle w == 0 reliably;
14753 it exists mostly for use in the implementations of BN_is_zero(),
14754 BN_is_one(), and BN_is_word().
14758 * New function BN_swap.
14762 * Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
14763 the exponentiation functions are more likely to produce reasonable
14764 results on negative inputs.
14768 * Change BN_mod_mul so that the result is always non-negative.
14769 Previously, it could be negative if one of the factors was negative;
14770 I don't think anyone really wanted that behaviour.
14774 * Move `BN_mod_...` functions into new file `crypto/bn/bn_mod.c`
14775 (except for exponentiation, which stays in `crypto/bn/bn_exp.c`,
14776 and `BN_mod_mul_reciprocal`, which stays in `crypto/bn/bn_recp.c`)
14777 and add new functions:
14786 BN_mod_lshift1_quick
14788 BN_mod_lshift_quick
14790 These functions always generate non-negative results.
14792 `BN_nnmod` otherwise is `like BN_mod` (if `BN_mod` computes a remainder `r`
14793 such that `|m| < r < 0`, `BN_nnmod` will output `rem + |m|` instead).
14795 `BN_mod_XXX_quick(r, a, [b,] m)` generates the same result as
14796 `BN_mod_XXX(r, a, [b,] m, ctx)`, but requires that `a` [and `b`]
14797 be reduced modulo `m`.
14799 *Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller*
14802 The following entry accidentally appeared in the CHANGES file
14803 distributed with OpenSSL 0.9.7. The modifications described in
14804 it do *not* apply to OpenSSL 0.9.7.
14806 * Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
14807 was actually never needed) and in BN_mul(). The removal in BN_mul()
14808 required a small change in bn_mul_part_recursive() and the addition
14809 of the functions bn_cmp_part_words(), bn_sub_part_words() and
14810 bn_add_part_words(), which do the same thing as bn_cmp_words(),
14811 bn_sub_words() and bn_add_words() except they take arrays with
14817 * In 'openssl passwd', verify passwords read from the terminal
14818 unless the '-salt' option is used (which usually means that
14819 verification would just waste user's time since the resulting
14820 hash is going to be compared with some given password hash)
14821 or the new '-noverify' option is used.
14823 This is an incompatible change, but it does not affect
14824 non-interactive use of 'openssl passwd' (passwords on the command
14825 line, '-stdin' option, '-in ...' option) and thus should not
14826 cause any problems.
14830 * Remove all references to RSAref, since there's no more need for it.
14834 * Make DSO load along a path given through an environment variable
14835 (SHLIB_PATH) with shl_load().
14839 * Constify the ENGINE code as a result of BIGNUM constification.
14840 Also constify the RSA code and most things related to it. In a
14841 few places, most notable in the depth of the ASN.1 code, ugly
14842 casts back to non-const were required (to be solved at a later
14847 * Make it so the openssl application has all engines loaded by default.
14851 * Constify the BIGNUM routines a little more.
14855 * Add the following functions:
14857 ENGINE_load_cswift()
14859 ENGINE_load_atalla()
14860 ENGINE_load_nuron()
14861 ENGINE_load_builtin_engines()
14863 That way, an application can itself choose if external engines that
14864 are built-in in OpenSSL shall ever be used or not. The benefit is
14865 that applications won't have to be linked with libdl or other dso
14866 libraries unless it's really needed.
14868 Changed 'openssl engine' to load all engines on demand.
14869 Changed the engine header files to avoid the duplication of some
14870 declarations (they differed!).
14874 * 'openssl engine' can now list capabilities.
14878 * Better error reporting in 'openssl engine'.
14882 * Never call load_dh_param(NULL) in s_server.
14886 * Add engine application. It can currently list engines by name and
14887 identity, and test if they are actually available.
14891 * Improve RPM specification file by forcing symbolic linking and making
14892 sure the installed documentation is also owned by root.root.
14894 *Damien Miller <djm@mindrot.org>*
14896 * Give the OpenSSL applications more possibilities to make use of
14897 keys (public as well as private) handled by engines.
14901 * Add OCSP code that comes from CertCo.
14905 * Add VMS support for the Rijndael code.
14909 * Added untested support for Nuron crypto accelerator.
14913 * Add support for external cryptographic devices. This code was
14914 previously distributed separately as the "engine" branch.
14916 *Geoff Thorpe, Richard Levitte*
14918 * Rework the filename-translation in the DSO code. It is now possible to
14919 have far greater control over how a "name" is turned into a filename
14920 depending on the operating environment and any oddities about the
14921 different shared library filenames on each system.
14925 * Support threads on FreeBSD-elf in Configure.
14929 * Fix for SHA1 assembly problem with MASM: it produces
14930 warnings about corrupt line number information when assembling
14931 with debugging information. This is caused by the overlapping
14934 *Bernd Matthes <mainbug@celocom.de>, Steve Henson*
14937 NCONF_get_number() has no error checking at all. As a replacement,
14938 NCONF_get_number_e() is defined (`_e` for "error checking") and is
14939 promoted strongly. The old NCONF_get_number is kept around for
14940 binary backward compatibility.
14941 Make it possible for methods to load from something other than a BIO,
14942 by providing a function pointer that is given a name instead of a BIO.
14943 For example, this could be used to load configuration data from an
14948 * Fix for non blocking accept BIOs. Added new I/O special reason
14949 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
14950 with non blocking I/O was not possible because no retry code was
14951 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
14956 * Added the beginnings of Rijndael support.
14960 * Fix for bug in DirectoryString mask setting. Add support for
14961 X509_NAME_print_ex() in 'req' and X509_print_ex() function
14962 to allow certificate printing to more controllable, additional
14963 'certopt' option to 'x509' to allow new printing options to be
14968 * Clean old EAY MD5 hack from e_os.h.
14972 ### Changes between 0.9.6l and 0.9.6m [17 Mar 2004]
14974 * Fix null-pointer assignment in do_change_cipher_spec() revealed
14975 by using the Codenomicon TLS Test Tool ([CVE-2004-0079])
14977 *Joe Orton, Steve Henson*
14979 ### Changes between 0.9.6k and 0.9.6l [04 Nov 2003]
14981 * Fix additional bug revealed by the NISCC test suite:
14983 Stop bug triggering large recursion when presented with
14984 certain ASN.1 tags ([CVE-2003-0851])
14988 ### Changes between 0.9.6j and 0.9.6k [30 Sep 2003]
14990 * Fix various bugs revealed by running the NISCC test suite:
14992 Stop out of bounds reads in the ASN1 code when presented with
14993 invalid tags (CVE-2003-0543 and CVE-2003-0544).
14995 If verify callback ignores invalid public key errors don't try to check
14996 certificate signature with the NULL public key.
15000 * In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
15001 if the server requested one: as stated in TLS 1.0 and SSL 3.0
15006 * In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
15007 extra data after the compression methods not only for TLS 1.0
15008 but also for SSL 3.0 (as required by the specification).
15010 *Bodo Moeller; problem pointed out by Matthias Loepfe*
15012 * Change X509_certificate_type() to mark the key as exported/exportable
15013 when it's 512 *bits* long, not 512 bytes.
15017 ### Changes between 0.9.6i and 0.9.6j [10 Apr 2003]
15019 * Countermeasure against the Klima-Pokorny-Rosa extension of
15020 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
15021 a protocol version number mismatch like a decryption error
15022 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
15026 * Turn on RSA blinding by default in the default implementation
15027 to avoid a timing attack. Applications that don't want it can call
15028 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
15029 They would be ill-advised to do so in most cases.
15031 *Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller*
15033 * Change RSA blinding code so that it works when the PRNG is not
15034 seeded (in this case, the secret RSA exponent is abused as
15035 an unpredictable seed -- if it is not unpredictable, there
15036 is no point in blinding anyway). Make RSA blinding thread-safe
15037 by remembering the creator's thread ID in rsa->blinding and
15038 having all other threads use local one-time blinding factors
15039 (this requires more computation than sharing rsa->blinding, but
15040 avoids excessive locking; and if an RSA object is not shared
15041 between threads, blinding will still be very fast).
15045 ### Changes between 0.9.6h and 0.9.6i [19 Feb 2003]
15047 * In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
15048 via timing by performing a MAC computation even if incorrect
15049 block cipher padding has been found. This is a countermeasure
15050 against active attacks where the attacker has to distinguish
15051 between bad padding and a MAC verification error. ([CVE-2003-0078])
15053 *Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
15054 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
15055 Martin Vuagnoux (EPFL, Ilion)*
15057 ### Changes between 0.9.6g and 0.9.6h [5 Dec 2002]
15059 * New function OPENSSL_cleanse(), which is used to cleanse a section of
15060 memory from its contents. This is done with a counter that will
15061 place alternating values in each byte. This can be used to solve
15062 two issues: 1) the removal of calls to memset() by highly optimizing
15063 compilers, and 2) cleansing with other values than 0, since those can
15064 be read through on certain media, for example a swap space on disk.
15068 * Bugfix: client side session caching did not work with external caching,
15069 because the session->cipher setting was not restored when reloading
15070 from the external cache. This problem was masked, when
15071 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set.
15072 (Found by Steve Haslam <steve@araqnid.ddts.net>.)
15076 * Fix client_certificate (ssl/s2_clnt.c): The permissible total
15077 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33.
15079 *Zeev Lieber <zeev-l@yahoo.com>*
15081 * Undo an undocumented change introduced in 0.9.6e which caused
15082 repeated calls to OpenSSL_add_all_ciphers() and
15083 OpenSSL_add_all_digests() to be ignored, even after calling
15088 * Change the default configuration reader to deal with last line not
15089 being properly terminated.
15093 * Change X509_NAME_cmp() so it applies the special rules on handling
15094 DN values that are of type PrintableString, as well as RDNs of type
15095 emailAddress where the value has the type ia5String.
15097 *stefank@valicert.com via Richard Levitte*
15099 * Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half
15100 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently
15101 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be
15102 the bitwise-OR of the two for use by the majority of applications
15103 wanting this behaviour, and update the docs. The documented
15104 behaviour and actual behaviour were inconsistent and had been
15105 changing anyway, so this is more a bug-fix than a behavioural
15108 *Geoff Thorpe, diagnosed by Nadav Har'El*
15110 * Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c
15111 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes).
15115 * Fix initialization code race conditions in
15116 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(),
15117 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(),
15118 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(),
15119 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(),
15120 ssl2_get_cipher_by_char(),
15121 ssl3_get_cipher_by_char().
15123 *Patrick McCormick <patrick@tellme.com>, Bodo Moeller*
15125 * Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after
15126 the cached sessions are flushed, as the remove_cb() might use ex_data
15127 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com>
15128 (see [openssl.org #212]).
15130 *Geoff Thorpe, Lutz Jaenicke*
15132 * Fix typo in OBJ_txt2obj which incorrectly passed the content
15133 length, instead of the encoding length to d2i_ASN1_OBJECT.
15137 ### Changes between 0.9.6f and 0.9.6g [9 Aug 2002]
15139 * [In 0.9.6g-engine release:]
15140 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use `_stdcall`).
15142 *Lynn Gazis <lgazis@rainbow.com>*
15144 ### Changes between 0.9.6e and 0.9.6f [8 Aug 2002]
15146 * Fix ASN1 checks. Check for overflow by comparing with LONG_MAX
15147 and get fix the header length calculation.
15148 *Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>,
15149 Alon Kantor <alonk@checkpoint.com> (and others), Steve Henson*
15151 * Use proper error handling instead of 'assertions' in buffer
15152 overflow checks added in 0.9.6e. This prevents DoS (the
15153 assertions could call abort()).
15155 *Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller*
15157 ### Changes between 0.9.6d and 0.9.6e [30 Jul 2002]
15159 * Add various sanity checks to asn1_get_length() to reject
15160 the ASN1 length bytes if they exceed sizeof(long), will appear
15161 negative or the content length exceeds the length of the
15164 *Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>*
15166 * Fix cipher selection routines: ciphers without encryption had no flags
15167 for the cipher strength set and where therefore not handled correctly
15168 by the selection routines (PR #130).
15172 * Fix EVP_dsa_sha macro.
15177 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS
15178 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure
15179 that was added in OpenSSL 0.9.6d.
15181 As the countermeasure turned out to be incompatible with some
15182 broken SSL implementations, the new option is part of SSL_OP_ALL.
15183 SSL_OP_ALL is usually employed when compatibility with weird SSL
15184 implementations is desired (e.g. '-bugs' option to 's_client' and
15185 's_server'), so the new option is automatically set in many
15190 * Changes in security patch:
15192 Changes marked "(CHATS)" were sponsored by the Defense Advanced
15193 Research Projects Agency (DARPA) and Air Force Research Laboratory,
15194 Air Force Materiel Command, USAF, under agreement number
15197 * Add various sanity checks to asn1_get_length() to reject
15198 the ASN1 length bytes if they exceed sizeof(long), will appear
15199 negative or the content length exceeds the length of the
15200 supplied buffer. ([CVE-2002-0659])
15202 *Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>*
15204 * Assertions for various potential buffer overflows, not known to
15205 happen in practice.
15207 *Ben Laurie (CHATS)*
15209 * Various temporary buffers to hold ASCII versions of integers were
15210 too small for 64 bit platforms. ([CVE-2002-0655])
15211 *Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>*
15213 * Remote buffer overflow in SSL3 protocol - an attacker could
15214 supply an oversized session ID to a client. ([CVE-2002-0656])
15216 *Ben Laurie (CHATS)*
15218 * Remote buffer overflow in SSL2 protocol - an attacker could
15219 supply an oversized client master key. ([CVE-2002-0656])
15221 *Ben Laurie (CHATS)*
15223 ### Changes between 0.9.6c and 0.9.6d [9 May 2002]
15225 * Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
15226 encoded as NULL) with id-dsa-with-sha1.
15228 *Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller*
15230 * Check various `X509_...()` return values in `apps/req.c`.
15232 *Nils Larsch <nla@trustcenter.de>*
15234 * Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
15235 an end-of-file condition would erroneously be flagged, when the CRLF
15236 was just at the end of a processed block. The bug was discovered when
15237 processing data through a buffering memory BIO handing the data to a
15238 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
15239 <ptsekov@syntrex.com> and Nedelcho Stanev.
15243 * Implement a countermeasure against a vulnerability recently found
15244 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
15245 before application data chunks to avoid the use of known IVs
15246 with data potentially chosen by the attacker.
15250 * Fix length checks in ssl3_get_client_hello().
15254 * TLS/SSL library bugfix: use s->s3->in_read_app_data differently
15255 to prevent ssl3_read_internal() from incorrectly assuming that
15256 ssl3_read_bytes() found application data while handshake
15257 processing was enabled when in fact s->s3->in_read_app_data was
15258 merely automatically cleared during the initial handshake.
15260 *Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>*
15262 * Fix object definitions for Private and Enterprise: they were not
15263 recognized in their shortname (=lowercase) representation. Extend
15264 obj_dat.pl to issue an error when using undefined keywords instead
15265 of silently ignoring the problem (Svenning Sorensen
15266 <sss@sss.dnsalias.net>).
15270 * Fix DH_generate_parameters() so that it works for 'non-standard'
15271 generators, i.e. generators other than 2 and 5. (Previously, the
15272 code did not properly initialise the 'add' and 'rem' values to
15273 BN_generate_prime().)
15275 In the new general case, we do not insist that 'generator' is
15276 actually a primitive root: This requirement is rather pointless;
15277 a generator of the order-q subgroup is just as good, if not
15282 * Map new X509 verification errors to alerts. Discovered and submitted by
15283 Tom Wu <tom@arcot.com>.
15287 * Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
15288 returning non-zero before the data has been completely received
15289 when using non-blocking I/O.
15291 *Bodo Moeller; problem pointed out by John Hughes*
15293 * Some of the ciphers missed the strength entry (SSL_LOW etc).
15295 *Ben Laurie, Lutz Jaenicke*
15297 * Fix bug in SSL_clear(): bad sessions were not removed (found by
15298 Yoram Zahavi <YoramZ@gilian.com>).
15302 * Add information about CygWin 1.3 and on, and preserve proper
15303 configuration for the versions before that.
15305 *Corinna Vinschen <vinschen@redhat.com> and Richard Levitte*
15307 * Make removal from session cache (SSL_CTX_remove_session()) more robust:
15308 check whether we deal with a copy of a session and do not delete from
15309 the cache in this case. Problem reported by "Izhar Shoshani Levi"
15310 <izhar@checkpoint.com>.
15314 * Do not store session data into the internal session cache, if it
15315 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
15316 flag is set). Proposed by Aslam <aslam@funk.com>.
15320 * Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
15325 * [In 0.9.6d-engine release:]
15326 Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
15328 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
15330 * Add the configuration target linux-s390x.
15332 *Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte*
15334 * The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
15335 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
15336 variable as an indication that a ClientHello message has been
15337 received. As the flag value will be lost between multiple
15338 invocations of ssl3_accept when using non-blocking I/O, the
15339 function may not be aware that a handshake has actually taken
15340 place, thus preventing a new session from being added to the
15343 To avoid this problem, we now set s->new_session to 2 instead of
15344 using a local variable.
15346 *Lutz Jaenicke, Bodo Moeller*
15348 * Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
15349 if the SSL_R_LENGTH_MISMATCH error is detected.
15351 *Geoff Thorpe, Bodo Moeller*
15353 * New 'shared_ldflag' column in Configure platform table.
15357 * Fix EVP_CIPHER_mode macro.
15359 *"Dan S. Camper" <dan@bti.net>*
15361 * Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
15362 type, we must throw them away by setting rr->length to 0.
15364 *D P Chang <dpc@qualys.com>*
15366 ### Changes between 0.9.6b and 0.9.6c [21 dec 2001]
15368 * Fix BN_rand_range bug pointed out by Dominikus Scherkl
15369 <Dominikus.Scherkl@biodata.com>. (The previous implementation
15370 worked incorrectly for those cases where range = `10..._2` and
15371 `3*range` is two bits longer than range.)
15375 * Only add signing time to PKCS7 structures if it is not already
15380 * Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
15381 OBJ_ld_ce should be OBJ_id_ce.
15382 Also some ip-pda OIDs in crypto/objects/objects.txt were
15383 incorrect (cf. RFC 3039).
15385 *Matt Cooper, Frederic Giudicelli, Bodo Moeller*
15387 * Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
15388 returns early because it has nothing to do.
15390 *Andy Schneider <andy.schneider@bjss.co.uk>*
15392 * [In 0.9.6c-engine release:]
15393 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
15395 *Andy Schneider <andy.schneider@bjss.co.uk>*
15397 * [In 0.9.6c-engine release:]
15398 Add support for Cryptographic Appliance's keyserver technology.
15399 (Use engine 'keyclient')
15401 *Cryptographic Appliances and Geoff Thorpe*
15403 * Add a configuration entry for OS/390 Unix. The C compiler 'c89'
15404 is called via tools/c89.sh because arguments have to be
15405 rearranged (all '-L' options must appear before the first object
15408 *Richard Shapiro <rshapiro@abinitio.com>*
15410 * [In 0.9.6c-engine release:]
15411 Add support for Broadcom crypto accelerator cards, backported
15414 *Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox*
15416 * [In 0.9.6c-engine release:]
15417 Add support for SureWare crypto accelerator cards from
15418 Baltimore Technologies. (Use engine 'sureware')
15420 *Baltimore Technologies and Mark Cox*
15422 * [In 0.9.6c-engine release:]
15423 Add support for crypto accelerator cards from Accelerated
15424 Encryption Processing, www.aep.ie. (Use engine 'aep')
15426 *AEP Inc. and Mark Cox*
15428 * Add a configuration entry for gcc on UnixWare.
15430 *Gary Benson <gbenson@redhat.com>*
15432 * Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
15433 messages are stored in a single piece (fixed-length part and
15434 variable-length part combined) and fix various bugs found on the way.
15438 * Disable caching in BIO_gethostbyname(), directly use gethostbyname()
15439 instead. BIO_gethostbyname() does not know what timeouts are
15440 appropriate, so entries would stay in cache even when they have
15442 *Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>*
15444 * Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
15445 faced with a pathologically small ClientHello fragment that does
15446 not contain client_version: Instead of aborting with an error,
15447 simply choose the highest available protocol version (i.e.,
15448 TLS 1.0 unless it is disabled). In practice, ClientHello
15449 messages are never sent like this, but this change gives us
15450 strictly correct behaviour at least for TLS.
15454 * Fix SSL handshake functions and SSL_clear() such that SSL_clear()
15455 never resets s->method to s->ctx->method when called from within
15456 one of the SSL handshake functions.
15458 *Bodo Moeller; problem pointed out by Niko Baric*
15460 * In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
15461 (sent using the client's version number) if client_version is
15462 smaller than the protocol version in use. Also change
15463 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
15464 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
15465 the client will at least see that alert.
15469 * Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
15474 * Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
15475 client receives HelloRequest while in a handshake.
15477 *Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>*
15479 * Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
15480 should end in 'break', not 'goto end' which circumvents various
15481 cleanups done in state SSL_ST_OK. But session related stuff
15482 must be disabled for SSL_ST_OK in the case that we just sent a
15485 Also avoid some overhead by not calling ssl_init_wbio_buffer()
15486 before just sending a HelloRequest.
15488 *Bodo Moeller, Eric Rescorla <ekr@rtfm.com>*
15490 * Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
15491 reveal whether illegal block cipher padding was found or a MAC
15492 verification error occurred. (Neither SSLerr() codes nor alerts
15493 are directly visible to potential attackers, but the information
15494 may leak via logfiles.)
15496 Similar changes are not required for the SSL 2.0 implementation
15497 because the number of padding bytes is sent in clear for SSL 2.0,
15498 and the extra bytes are just ignored. However ssl/s2_pkt.c
15499 failed to verify that the purported number of padding bytes is in
15504 * Add OpenUNIX-8 support including shared libraries
15505 (Boyd Lynn Gerber <gerberb@zenez.com>).
15509 * Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
15510 'wristwatch attack' using huge encoding parameters (cf.
15511 James H. Manger's CRYPTO 2001 paper). Note that the
15512 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
15513 encoding parameters and hence was not vulnerable.
15517 * BN_sqr() bug fix.
15519 *Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>*
15521 * Rabin-Miller test analyses assume uniformly distributed witnesses,
15522 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
15523 followed by modular reduction.
15525 *Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>*
15527 * Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
15528 equivalent based on BN_pseudo_rand() instead of BN_rand().
15532 * s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
15533 This function was broken, as the check for a new client hello message
15534 to handle SGC did not allow these large messages.
15535 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
15539 * Add alert descriptions for TLSv1 to `SSL_alert_desc_string[_long]()`.
15543 * Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
15544 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
15548 * Rework the configuration and shared library support for Tru64 Unix.
15549 The configuration part makes use of modern compiler features and
15550 still retains old compiler behavior for those that run older versions
15551 of the OS. The shared library support part includes a variant that
15552 uses the RPATH feature, and is available through the special
15553 configuration target "alpha-cc-rpath", which will never be selected
15556 *Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte*
15558 * In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
15559 with the same message size as in ssl3_get_certificate_request().
15560 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
15561 messages might inadvertently be reject as too long.
15563 *Petr Lampa <lampa@fee.vutbr.cz>*
15565 * Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
15569 * Modified SSL library such that the verify_callback that has been set
15570 specifically for an SSL object with SSL_set_verify() is actually being
15571 used. Before the change, a verify_callback set with this function was
15572 ignored and the verify_callback() set in the SSL_CTX at the time of
15573 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
15574 to allow the necessary settings.
15578 * Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
15579 explicitly to NULL, as at least on Solaris 8 this seems not always to be
15580 done automatically (in contradiction to the requirements of the C
15581 standard). This made problems when used from OpenSSH.
15585 * In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
15586 dh->length and always used
15588 BN_rand_range(priv_key, dh->p).
15590 BN_rand_range() is not necessary for Diffie-Hellman, and this
15591 specific range makes Diffie-Hellman unnecessarily inefficient if
15592 dh->length (recommended exponent length) is much smaller than the
15593 length of dh->p. We could use BN_rand_range() if the order of
15594 the subgroup was stored in the DH structure, but we only have
15599 BN_rand(priv_key, l, ...)
15601 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
15608 RSA_eay_public_encrypt
15609 RSA_eay_private_decrypt
15610 RSA_eay_private_encrypt (signing)
15611 RSA_eay_public_decrypt (signature verification)
15613 (default implementations for RSA_public_encrypt,
15614 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
15615 always reject numbers >= n.
15619 * In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
15620 to synchronize access to 'locking_thread'. This is necessary on
15621 systems where access to 'locking_thread' (an 'unsigned long'
15622 variable) is not atomic.
15626 * In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
15627 *before* setting the 'crypto_lock_rand' flag. The previous code had
15628 a race condition if 0 is a valid thread ID.
15630 *Travis Vitek <vitek@roguewave.com>*
15632 * Add support for shared libraries under Irix.
15634 *Albert Chin-A-Young <china@thewrittenword.com>*
15636 * Add configuration option to build on Linux on both big-endian and
15637 little-endian MIPS.
15639 *Ralf Baechle <ralf@uni-koblenz.de>*
15641 * Add the possibility to create shared libraries on HP-UX.
15645 ### Changes between 0.9.6a and 0.9.6b [9 Jul 2001]
15647 * Change ssleay_rand_bytes (crypto/rand/md_rand.c)
15648 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
15649 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
15650 PRNG state recovery was possible based on the output of
15651 one PRNG request appropriately sized to gain knowledge on
15652 'md' followed by enough consecutive 1-byte PRNG requests
15653 to traverse all of 'state'.
15655 1. When updating 'md_local' (the current thread's copy of 'md')
15656 during PRNG output generation, hash all of the previous
15657 'md_local' value, not just the half used for PRNG output.
15659 2. Make the number of bytes from 'state' included into the hash
15660 independent from the number of PRNG bytes requested.
15662 The first measure alone would be sufficient to avoid
15663 Markku-Juhani's attack. (Actually it had never occurred
15664 to me that the half of 'md_local' used for chaining was the
15665 half from which PRNG output bytes were taken -- I had always
15666 assumed that the secret half would be used.) The second
15667 measure makes sure that additional data from 'state' is never
15668 mixed into 'md_local' in small portions; this heuristically
15669 further strengthens the PRNG.
15673 * Fix crypto/bn/asm/mips3.s.
15677 * When only the key is given to "enc", the IV is undefined. Print out
15678 an error message in this case.
15682 * Handle special case when X509_NAME is empty in X509 printing routines.
15686 * In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
15687 positive and less than q.
15691 * Don't change `*pointer` in CRYPTO_add_lock() is add_lock_callback is
15692 used: it isn't thread safe and the add_lock_callback should handle
15695 *Paul Rose <Paul.Rose@bridge.com>*
15697 * Verify that incoming data obeys the block size in
15698 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
15704 *Ulf Möller, Bodo Möller*
15706 * The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
15707 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
15708 when fixing the server behaviour for backwards-compatible 'client
15709 hello' messages. (Note that the attack is impractical against
15710 SSL 3.0 and TLS 1.0 anyway because length and version checking
15711 means that the probability of guessing a valid ciphertext is
15712 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
15715 Before 0.9.5, the countermeasure (hide the error by generating a
15716 random 'decryption result') did not work properly because
15717 ERR_clear_error() was missing, meaning that SSL_get_error() would
15718 detect the supposedly ignored error.
15720 Both problems are now fixed.
15724 * In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
15725 (previously it was 1024).
15729 * Fix for compatibility mode trust settings: ignore trust settings
15730 unless some valid trust or reject settings are present.
15734 * Fix for blowfish EVP: its a variable length cipher.
15738 * Fix various bugs related to DSA S/MIME verification. Handle missing
15739 parameters in DSA public key structures and return an error in the
15740 DSA routines if parameters are absent.
15744 * In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
15745 in the current directory if neither $RANDFILE nor $HOME was set.
15746 RAND_file_name() in 0.9.6a returned NULL in this case. This has
15747 caused some confusion to Windows users who haven't defined $HOME.
15748 Thus RAND_file_name() is changed again: e_os.h can define a
15749 DEFAULT_HOME, which will be used if $HOME is not set.
15750 For Windows, we use "C:"; on other platforms, we still require
15751 environment variables.
15753 * Move 'if (!initialized) RAND_poll()' into regions protected by
15754 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
15755 having multiple threads call RAND_poll() concurrently.
15759 * In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
15760 combination of a flag and a thread ID variable.
15761 Otherwise while one thread is in ssleay_rand_bytes (which sets the
15762 flag), *other* threads can enter ssleay_add_bytes without obeying
15763 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
15764 that they do not hold after the first thread unsets add_do_not_lock).
15768 * Change bctest again: '-x' expressions are not available in all
15769 versions of 'test'.
15773 ### Changes between 0.9.6 and 0.9.6a [5 Apr 2001]
15775 * Fix a couple of memory leaks in PKCS7_dataDecode()
15777 *Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>*
15779 * Change Configure and Makefiles to provide EXE_EXT, which will contain
15780 the default extension for executables, if any. Also, make the perl
15781 scripts that use symlink() to test if it really exists and use "cp"
15782 if it doesn't. All this made OpenSSL compilable and installable in
15787 * Fix for asn1_GetSequence() for indefinite length constructed data.
15788 If SEQUENCE is length is indefinite just set c->slen to the total
15789 amount of data available.
15791 *Steve Henson, reported by shige@FreeBSD.org*
15793 *This change does not apply to 0.9.7.*
15795 * Change bctest to avoid here-documents inside command substitution
15796 (workaround for FreeBSD /bin/sh bug).
15797 For compatibility with Ultrix, avoid shell functions (introduced
15798 in the bctest version that searches along $PATH).
15802 * Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
15803 with des_encrypt() defined on some operating systems, like Solaris
15808 * Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
15809 On the Importance of Eliminating Errors in Cryptographic
15810 Computations, J. Cryptology 14 (2001) 2, 101-119,
15811 <http://theory.stanford.edu/~dabo/papers/faults.ps.gz>).
15815 * MIPS assembler BIGNUM division bug fix.
15819 * Disabled incorrect Alpha assembler code.
15823 * Fix PKCS#7 decode routines so they correctly update the length
15824 after reading an EOC for the EXPLICIT tag.
15828 *This change does not apply to 0.9.7.*
15830 * Fix bug in PKCS#12 key generation routines. This was triggered
15831 if a 3DES key was generated with a 0 initial byte. Include
15832 PKCS12_BROKEN_KEYGEN compilation option to retain the old
15833 (but broken) behaviour.
15837 * Enhance bctest to search for a working bc along $PATH and print
15840 *Tim Rice <tim@multitalents.net> via Richard Levitte*
15842 * Fix memory leaks in err.c: free err_data string if necessary;
15843 don't write to the wrong index in ERR_set_error_data.
15847 * Implement ssl23_peek (analogous to ssl23_read), which previously
15852 * Replace rdtsc with `_emit` statements for VC++ version 5.
15854 *Jeremy Cooper <jeremy@baymoo.org>*
15856 * Make it possible to reuse SSLv2 sessions.
15860 * In copy_email() check for >= 0 as a return value for
15861 X509_NAME_get_index_by_NID() since 0 is a valid index.
15863 *Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>*
15865 * Avoid coredump with unsupported or invalid public keys by checking if
15866 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
15867 PKCS7_verify() fails with non detached data.
15871 * Don't use getenv in library functions when run as setuid/setgid.
15872 New function OPENSSL_issetugid().
15876 * Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
15877 due to incorrect handling of multi-threading:
15879 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
15881 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
15883 3. Count how many times MemCheck_off() has been called so that
15884 nested use can be treated correctly. This also avoids
15885 inband-signalling in the previous code (which relied on the
15886 assumption that thread ID 0 is impossible).
15890 * Add "-rand" option also to s_client and s_server.
15894 * Fix CPU detection on Irix 6.x.
15895 *Kurt Hockenbury <khockenb@stevens-tech.edu> and
15896 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>*
15898 * Fix X509_NAME bug which produced incorrect encoding if X509_NAME
15903 *This change does not apply to 0.9.7.*
15905 * Use the cached encoding of an X509_NAME structure rather than
15906 copying it. This is apparently the reason for the libsafe "errors"
15907 but the code is actually correct.
15911 * Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
15912 Bleichenbacher's DSA attack.
15913 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
15914 to be set and top=0 forces the highest bit to be set; top=-1 is new
15915 and leaves the highest bit random.
15917 *Ulf Moeller, Bodo Moeller*
15919 * In the `NCONF_...`-based implementations for `CONF_...` queries
15920 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
15921 a temporary CONF structure with the data component set to NULL
15922 (which gives segmentation faults in lh_retrieve).
15923 Instead, use NULL for the CONF pointer in CONF_get_string and
15924 CONF_get_number (which may use environment variables) and directly
15925 return NULL from CONF_get_section.
15929 * Fix potential buffer overrun for EBCDIC.
15933 * Tolerate nonRepudiation as being valid for S/MIME signing and certSign
15934 keyUsage if basicConstraints absent for a CA.
15938 * Make SMIME_write_PKCS7() write mail header values with a format that
15939 is more generally accepted (no spaces before the semicolon), since
15940 some programs can't parse those values properly otherwise. Also make
15941 sure BIO's that break lines after each write do not create invalid
15946 * Make the CRL encoding routines work with empty SEQUENCE OF. The
15947 macros previously used would not encode an empty SEQUENCE OF
15948 and break the signature.
15952 *This change does not apply to 0.9.7.*
15954 * Zero the premaster secret after deriving the master secret in
15959 * Add some EVP_add_digest_alias registrations (as found in
15960 OpenSSL_add_all_digests()) to SSL_library_init()
15961 aka OpenSSL_add_ssl_algorithms(). This provides improved
15962 compatibility with peers using X.509 certificates
15963 with unconventional AlgorithmIdentifier OIDs.
15967 * Fix for Irix with NO_ASM.
15969 *"Bruce W. Forsberg" <bruce.forsberg@baesystems.com>*
15971 * ./config script fixes.
15973 *Ulf Moeller, Richard Levitte*
15975 * Fix 'openssl passwd -1'.
15979 * Change PKCS12_key_gen_asc() so it can cope with non null
15980 terminated strings whose length is passed in the passlen
15981 parameter, for example from PEM callbacks. This was done
15982 by adding an extra length parameter to asc2uni().
15984 *Steve Henson, reported by <oddissey@samsung.co.kr>*
15986 * Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
15987 call failed, free the DSA structure.
15991 * Fix to uni2asc() to cope with zero length Unicode strings.
15992 These are present in some PKCS#12 files.
15996 * Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
15997 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
15998 when writing a 32767 byte record.
16000 *Bodo Moeller; problem reported by Eric Day <eday@concentric.net>*
16002 * In `RSA_eay_public_{en,ed}crypt` and RSA_eay_mod_exp (rsa_eay.c),
16003 obtain lock CRYPTO_LOCK_RSA before setting `rsa->_method_mod_{n,p,q}`.
16005 (RSA objects have a reference count access to which is protected
16006 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
16007 so they are meant to be shared between threads.)
16008 *Bodo Moeller, Geoff Thorpe; original patch submitted by
16009 "Reddie, Steven" <Steven.Reddie@ca.com>*
16011 * Fix a deadlock in CRYPTO_mem_leaks().
16015 * Use better test patterns in bntest.
16019 * rand_win.c fix for Borland C.
16023 * BN_rshift bugfix for n == 0.
16027 * Add a 'bctest' script that checks for some known 'bc' bugs
16028 so that 'make test' does not abort just because 'bc' is broken.
16032 * Store verify_result within SSL_SESSION also for client side to
16033 avoid potential security hole. (Re-used sessions on the client side
16034 always resulted in verify_result==X509_V_OK, not using the original
16035 result of the server certificate verification.)
16039 * Fix ssl3_pending: If the record in s->s3->rrec is not of type
16040 SSL3_RT_APPLICATION_DATA, return 0.
16041 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
16046 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
16047 releases, have been re-implemented by renaming the previous
16048 implementations of ssl2_read and ssl3_read to ssl2_read_internal
16049 and ssl3_read_internal, respectively, and adding 'peek' parameters
16050 to them. The new ssl[23]_{read,peek} functions are calls to
16051 ssl[23]_read_internal with the 'peek' flag set appropriately.
16052 A 'peek' parameter has also been added to ssl3_read_bytes, which
16053 does the actual work for ssl3_read_internal.
16057 * Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
16058 the method-specific "init()" handler. Also clean up ex_data after
16059 calling the method-specific "finish()" handler. Previously, this was
16060 happening the other way round.
16064 * Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
16065 The previous value, 12, was not always sufficient for BN_mod_exp().
16069 * Make sure that shared libraries get the internal name engine with
16070 the full version number and not just 0. This should mark the
16071 shared libraries as not backward compatible. Of course, this should
16072 be changed again when we can guarantee backward binary compatibility.
16076 * Fix typo in get_cert_by_subject() in by_dir.c
16078 *Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>*
16080 * Rework the system to generate shared libraries:
16082 - Make note of the expected extension for the shared libraries and
16083 if there is a need for symbolic links from for example libcrypto.so.0
16084 to libcrypto.so.0.9.7. There is extended info in Configure for
16087 - Make as few rebuilds of the shared libraries as possible.
16089 - Still avoid linking the OpenSSL programs with the shared libraries.
16091 - When installing, install the shared libraries separately from the
16096 * Fix SSL_CTX_set_read_ahead macro to actually use its argument.
16098 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
16099 and not in SSL_clear because the latter is also used by the
16100 accept/connect functions; previously, the settings made by
16101 SSL_set_read_ahead would be lost during the handshake.
16103 *Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>*
16105 * Correct util/mkdef.pl to be selective about disabled algorithms.
16106 Previously, it would create entries for disabled algorithms no
16111 * Added several new manual pages for SSL_* function.
16115 ### Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
16117 * In ssl23_get_client_hello, generate an error message when faced
16118 with an initial SSL 3.0/TLS record that is too small to contain the
16119 first two bytes of the ClientHello message, i.e. client_version.
16120 (Note that this is a pathologic case that probably has never happened
16121 in real life.) The previous approach was to use the version number
16122 from the record header as a substitute; but our protocol choice
16123 should not depend on that one because it is not authenticated
16124 by the Finished messages.
16128 * More robust randomness gathering functions for Windows.
16130 *Jeffrey Altman <jaltman@columbia.edu>*
16132 * For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
16133 not set then we don't setup the error code for issuer check errors
16134 to avoid possibly overwriting other errors which the callback does
16135 handle. If an application does set the flag then we assume it knows
16136 what it is doing and can handle the new informational codes
16141 * Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
16142 a general "ANY" type, as such it should be able to decode anything
16143 including tagged types. However it didn't check the class so it would
16144 wrongly interpret tagged types in the same way as their universal
16145 counterpart and unknown types were just rejected. Changed so that the
16146 tagged and unknown types are handled in the same way as a SEQUENCE:
16147 that is the encoding is stored intact. There is also a new type
16148 "V_ASN1_OTHER" which is used when the class is not universal, in this
16149 case we have no idea what the actual type is so we just lump them all
16154 * On VMS, stdout may very well lead to a file that is written to
16155 in a record-oriented fashion. That means that every write() will
16156 write a separate record, which will be read separately by the
16157 programs trying to read from it. This can be very confusing.
16159 The solution is to put a BIO filter in the way that will buffer
16160 text until a linefeed is reached, and then write everything a
16161 line at a time, so every record written will be an actual line,
16162 not chunks of lines and not (usually doesn't happen, but I've
16163 seen it once) several lines in one record. BIO_f_linebuffer() is
16166 Currently, it's a VMS-only method, because that's where it has
16167 been tested well enough.
16171 * Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
16172 it can return incorrect results.
16173 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
16174 but it was in 0.9.6-beta[12].)
16178 * Disable the check for content being present when verifying detached
16179 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
16180 include zero length content when signing messages.
16184 * New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
16185 BIO_ctrl (for BIO pairs).
16189 * Add DSO method for VMS.
16193 * Bug fix: Montgomery multiplication could produce results with the
16198 * Add RPM specification openssl.spec and modify it to build three
16199 packages. The default package contains applications, application
16200 documentation and run-time libraries. The devel package contains
16201 include files, static libraries and function documentation. The
16202 doc package contains the contents of the doc directory. The original
16203 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
16207 * Add a large number of documentation files for many SSL routines.
16209 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>*
16211 * Add a configuration entry for Sony News 4.
16213 *NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>*
16215 * Don't set the two most significant bits to one when generating a
16216 random number < q in the DSA library.
16220 * New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
16221 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
16222 the underlying transport is blocking) if a handshake took place.
16223 (The default behaviour is needed by applications such as s_client
16224 and s_server that use select() to determine when to use SSL_read;
16225 but for applications that know in advance when to expect data, it
16226 just makes things more complicated.)
16230 * Add RAND_egd_bytes(), which gives control over the number of bytes read
16235 * Add a few more EBCDIC conditionals that make `req` and `x509`
16236 work better on such systems.
16238 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
16240 * Add two demo programs for PKCS12_parse() and PKCS12_create().
16241 Update PKCS12_parse() so it copies the friendlyName and the
16242 keyid to the certificates aux info.
16246 * Fix bug in PKCS7_verify() which caused an infinite loop
16247 if there was more than one signature.
16249 *Sven Uszpelkat <su@celocom.de>*
16251 * Major change in util/mkdef.pl to include extra information
16252 about each symbol, as well as presenting variables as well
16253 as functions. This change means that there's n more need
16254 to rebuild the .num files when some algorithms are excluded.
16258 * Allow the verify time to be set by an application,
16259 rather than always using the current time.
16263 * Phase 2 verify code reorganisation. The certificate
16264 verify code now looks up an issuer certificate by a
16265 number of criteria: subject name, authority key id
16266 and key usage. It also verifies self signed certificates
16267 by the same criteria. The main comparison function is
16268 X509_check_issued() which performs these checks.
16270 Lot of changes were necessary in order to support this
16271 without completely rewriting the lookup code.
16273 Authority and subject key identifier are now cached.
16275 The LHASH 'certs' is X509_STORE has now been replaced
16276 by a STACK_OF(X509_OBJECT). This is mainly because an
16277 LHASH can't store or retrieve multiple objects with
16278 the same hash value.
16280 As a result various functions (which were all internal
16281 use only) have changed to handle the new X509_STORE
16282 structure. This will break anything that messed round
16283 with X509_STORE internally.
16285 The functions X509_STORE_add_cert() now checks for an
16286 exact match, rather than just subject name.
16288 The X509_STORE API doesn't directly support the retrieval
16289 of multiple certificates matching a given criteria, however
16290 this can be worked round by performing a lookup first
16291 (which will fill the cache with candidate certificates)
16292 and then examining the cache for matches. This is probably
16293 the best we can do without throwing out X509_LOOKUP
16294 entirely (maybe later...).
16296 The X509_VERIFY_CTX structure has been enhanced considerably.
16298 All certificate lookup operations now go via a get_issuer()
16299 callback. Although this currently uses an X509_STORE it
16300 can be replaced by custom lookups. This is a simple way
16301 to bypass the X509_STORE hackery necessary to make this
16302 work and makes it possible to use more efficient techniques
16303 in future. A very simple version which uses a simple
16304 STACK for its trusted certificate store is also provided
16305 using X509_STORE_CTX_trusted_stack().
16307 The verify_cb() and verify() callbacks now have equivalents
16308 in the X509_STORE_CTX structure.
16310 X509_STORE_CTX also has a 'flags' field which can be used
16311 to customise the verify behaviour.
16315 * Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
16316 excludes S/MIME capabilities.
16320 * When a certificate request is read in keep a copy of the
16321 original encoding of the signed data and use it when outputting
16322 again. Signatures then use the original encoding rather than
16323 a decoded, encoded version which may cause problems if the
16324 request is improperly encoded.
16328 * For consistency with other BIO_puts implementations, call
16329 buffer_write(b, ...) directly in buffer_puts instead of calling
16332 In BIO_puts, increment b->num_write as in BIO_write.
16334 *Peter.Sylvester@EdelWeb.fr*
16336 * Fix BN_mul_word for the case where the word is 0. (We have to use
16337 BN_zero, we may not return a BIGNUM with an array consisting of
16338 words set to zero.)
16342 * Avoid calling abort() from within the library when problems are
16343 detected, except if preprocessor symbols have been defined
16344 (such as REF_CHECK, BN_DEBUG etc.).
16348 * New openssl application 'rsautl'. This utility can be
16349 used for low-level RSA operations. DER public key
16350 BIO/fp routines also added.
16354 * New Configure entry and patches for compiling on QNX 4.
16356 *Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>*
16358 * A demo state-machine implementation was sponsored by
16359 Nuron (<http://www.nuron.com/>) and is now available in
16360 demos/state_machine.
16364 * New options added to the 'dgst' utility for signature
16365 generation and verification.
16369 * Unrecognized PKCS#7 content types are now handled via a
16370 catch all ASN1_TYPE structure. This allows unsupported
16371 types to be stored as a "blob" and an application can
16372 encode and decode it manually.
16376 * Fix various signed/unsigned issues to make a_strex.c
16377 compile under VC++.
16379 *Oscar Jacobsson <oscar.jacobsson@celocom.com>*
16381 * ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
16382 length if passed a buffer. ASN1_INTEGER_to_BN failed
16383 if passed a NULL BN and its argument was negative.
16385 *Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>*
16387 * Modification to PKCS#7 encoding routines to output definite
16388 length encoding. Since currently the whole structures are in
16389 memory there's not real point in using indefinite length
16390 constructed encoding. However if OpenSSL is compiled with
16391 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
16395 * Added BIO_vprintf() and BIO_vsnprintf().
16399 * Added more prefixes to parse for in the strings written
16400 through a logging bio, to cover all the levels that are available
16401 through syslog. The prefixes are now:
16403 PANIC, EMERG, EMR => LOG_EMERG
16404 ALERT, ALR => LOG_ALERT
16405 CRIT, CRI => LOG_CRIT
16406 ERROR, ERR => LOG_ERR
16407 WARNING, WARN, WAR => LOG_WARNING
16408 NOTICE, NOTE, NOT => LOG_NOTICE
16409 INFO, INF => LOG_INFO
16410 DEBUG, DBG => LOG_DEBUG
16412 and as before, if none of those prefixes are present at the
16413 beginning of the string, LOG_ERR is chosen.
16415 On Win32, the `LOG_*` levels are mapped according to this:
16417 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
16418 LOG_WARNING => EVENTLOG_WARNING_TYPE
16419 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
16423 * Made it possible to reconfigure with just the configuration
16424 argument "reconf" or "reconfigure". The command line arguments
16425 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
16426 and are retrieved from there when reconfiguring.
16432 *Assar Westerlund <assar@sics.se>, Richard Levitte*
16434 * Add the arguments -CAfile and -CApath to the pkcs12 utility.
16438 * The obj_dat.pl script was messing up the sorting of object
16439 names. The reason was that it compared the quoted version
16440 of strings as a result "OCSP" > "OCSP Signing" because
16441 " > SPACE. Changed script to store unquoted versions of
16442 names and add quotes on output. It was also omitting some
16443 names from the lookup table if they were given a default
16444 value (that is if SN is missing it is given the same
16445 value as LN and vice versa), these are now added on the
16446 grounds that if an object has a name we should be able to
16447 look it up. Finally added warning output when duplicate
16448 short or long names are found.
16452 * Changes needed for Tandem NSK.
16454 *Scott Uroff <scott@xypro.com>*
16456 * Fix SSL 2.0 rollback checking: Due to an off-by-one error in
16457 RSA_padding_check_SSLv23(), special padding was never detected
16458 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
16459 version rollback attacks was not effective.
16461 In s23_clnt.c, don't use special rollback-attack detection padding
16462 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
16463 client; similarly, in s23_srvr.c, don't do the rollback check if
16464 SSL 2.0 is the only protocol enabled in the server.
16468 * Make it possible to get hexdumps of unprintable data with 'openssl
16469 asn1parse'. By implication, the functions ASN1_parse_dump() and
16470 BIO_dump_indent() are added.
16474 * New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
16475 these print out strings and name structures based on various
16476 flags including RFC2253 support and proper handling of
16477 multibyte characters. Added options to the 'x509' utility
16478 to allow the various flags to be set.
16482 * Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
16483 Also change the functions X509_cmp_current_time() and
16484 X509_gmtime_adj() work with an ASN1_TIME structure,
16485 this will enable certificates using GeneralizedTime in validity
16486 dates to be checked.
16490 * Make the NEG_PUBKEY_BUG code (which tolerates invalid
16491 negative public key encodings) on by default,
16492 NO_NEG_PUBKEY_BUG can be set to disable it.
16496 * New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
16497 content octets. An i2c_ASN1_OBJECT is unnecessary because
16498 the encoding can be trivially obtained from the structure.
16502 * crypto/err.c locking bugfix: Use write locks (`CRYPTO_w_[un]lock`),
16503 not read locks (`CRYPTO_r_[un]lock`).
16507 * A first attempt at creating official support for shared
16508 libraries through configuration. I've kept it so the
16509 default is static libraries only, and the OpenSSL programs
16510 are always statically linked for now, but there are
16511 preparations for dynamic linking in place.
16512 This has been tested on Linux and Tru64.
16516 * Randomness polling function for Win9x, as described in:
16517 Peter Gutmann, Software Generation of Practically Strong
16522 * Fix so PRNG is seeded in req if using an already existing
16527 * New options to smime application. -inform and -outform
16528 allow alternative formats for the S/MIME message including
16529 PEM and DER. The -content option allows the content to be
16530 specified separately. This should allow things like Netscape
16531 form signing output easier to verify.
16535 * Fix the ASN1 encoding of tags using the 'long form'.
16539 * New ASN1 functions, `i2c_*` and `c2i_*` for INTEGER and BIT
16540 STRING types. These convert content octets to and from the
16541 underlying type. The actual tag and length octets are
16542 already assumed to have been read in and checked. These
16543 are needed because all other string types have virtually
16544 identical handling apart from the tag. By having versions
16545 of the ASN1 functions that just operate on content octets
16546 IMPLICIT tagging can be handled properly. It also allows
16547 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
16548 and ASN1_INTEGER are identical apart from the tag.
16552 * Change the handling of OID objects as follows:
16554 - New object identifiers are inserted in objects.txt, following
16555 the syntax given in [crypto/objects/README.md](crypto/objects/README.md).
16556 - objects.pl is used to process obj_mac.num and create a new
16558 - obj_dat.pl is used to create a new obj_dat.h, using the data in
16561 This is currently kind of a hack, and the perl code in objects.pl
16562 isn't very elegant, but it works as I intended. The simplest way
16563 to check that it worked correctly is to look in obj_dat.h and
16564 check the array nid_objs and make sure the objects haven't moved
16565 around (this is important!). Additions are OK, as well as
16566 consistent name changes.
16570 * Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
16574 * Addition of the command line parameter '-rand file' to 'openssl req'.
16575 The given file adds to whatever has already been seeded into the
16576 random pool through the RANDFILE configuration file option or
16577 environment variable, or the default random state file.
16581 * mkstack.pl now sorts each macro group into lexical order.
16582 Previously the output order depended on the order the files
16583 appeared in the directory, resulting in needless rewriting
16588 * Patches to make OpenSSL compile under Win32 again. Mostly
16589 work arounds for the VC++ problem that it treats func() as
16590 func(void). Also stripped out the parts of mkdef.pl that
16591 added extra typesafe functions: these no longer exist.
16595 * Reorganisation of the stack code. The macros are now all
16596 collected in safestack.h . Each macro is defined in terms of
16597 a "stack macro" of the form `SKM_<name>(type, a, b)`. The
16598 DEBUG_SAFESTACK is now handled in terms of function casts,
16599 this has the advantage of retaining type safety without the
16600 use of additional functions. If DEBUG_SAFESTACK is not defined
16601 then the non typesafe macros are used instead. Also modified the
16602 mkstack.pl script to handle the new form. Needs testing to see
16603 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
16604 the default if no major problems. Similar behaviour for ASN1_SET_OF
16605 and PKCS12_STACK_OF.
16609 * When some versions of IIS use the 'NET' form of private key the
16610 key derivation algorithm is different. Normally MD5(password) is
16611 used as a 128 bit RC4 key. In the modified case
16612 MD5(MD5(password) + "SGCKEYSALT") is used instead. Added some
16613 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
16614 as the old Netscape_RSA functions except they have an additional
16615 'sgckey' parameter which uses the modified algorithm. Also added
16616 an -sgckey command line option to the rsa utility. Thanks to
16617 Adrian Peck <bertie@ncipher.com> for posting details of the modified
16618 algorithm to openssl-dev.
16622 * The evp_local.h macros were using 'c.##kname' which resulted in
16623 invalid expansion on some systems (SCO 5.0.5 for example).
16624 Corrected to 'c.kname'.
16626 *Phillip Porch <root@theporch.com>*
16628 * New X509_get1_email() and X509_REQ_get1_email() functions that return
16629 a STACK of email addresses from a certificate or request, these look
16630 in the subject name and the subject alternative name extensions and
16631 omit any duplicate addresses.
16635 * Re-implement BN_mod_exp2_mont using independent (and larger) windows.
16636 This makes DSA verification about 2 % faster.
16640 * Increase maximum window size in `BN_mod_exp_...` to 6 bits instead of 5
16641 (meaning that now 2^5 values will be precomputed, which is only 4 KB
16642 plus overhead for 1024 bit moduli).
16643 This makes exponentiations about 0.5 % faster for 1024 bit
16644 exponents (as measured by "openssl speed rsa2048").
16648 * Rename memory handling macros to avoid conflicts with other
16650 Malloc => OPENSSL_malloc
16651 Malloc_locked => OPENSSL_malloc_locked
16652 Realloc => OPENSSL_realloc
16653 Free => OPENSSL_free
16657 * New function BN_mod_exp_mont_word for small bases (roughly 15%
16658 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
16662 * CygWin32 support.
16664 *John Jarvie <jjarvie@newsguy.com>*
16666 * The type-safe stack code has been rejigged. It is now only compiled
16667 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
16668 by default all type-specific stack functions are "#define"d back to
16669 standard stack functions. This results in more streamlined output
16670 but retains the type-safety checking possibilities of the original
16675 * The STACK code has been cleaned up, and certain type declarations
16676 that didn't make a lot of sense have been brought in line. This has
16677 also involved a cleanup of sorts in safestack.h to more correctly
16678 map type-safe stack functions onto their plain stack counterparts.
16679 This work has also resulted in a variety of "const"ifications of
16680 lots of the code, especially `_cmp` operations which should normally
16681 be prototyped with "const" parameters anyway.
16685 * When generating bytes for the first time in md_rand.c, 'stir the pool'
16686 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
16687 (The PRNG state consists of two parts, the large pool 'state' and 'md',
16688 where all of 'md' is used each time the PRNG is used, but 'state'
16689 is used only indexed by a cyclic counter. As entropy may not be
16690 well distributed from the beginning, 'md' is important as a
16691 chaining variable. However, the output function chains only half
16692 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
16693 all of 'md', and seeding with STATE_SIZE dummy bytes will result
16694 in all of 'state' being rewritten, with the new values depending
16695 on virtually all of 'md'. This overcomes the 80 bit limitation.)
16699 * In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
16700 the handshake is continued after ssl_verify_cert_chain();
16701 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
16702 can lead to 'unexplainable' connection aborts later.
16704 *Bodo Moeller; problem tracked down by Lutz Jaenicke*
16706 * Major EVP API cipher revision.
16707 Add hooks for extra EVP features. This allows various cipher
16708 parameters to be set in the EVP interface. Support added for variable
16709 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
16710 setting of RC2 and RC5 parameters.
16712 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
16715 Remove lots of duplicated code from the EVP library. For example *every*
16716 cipher init() function handles the 'iv' in the same way according to the
16717 cipher mode. They also all do nothing if the 'key' parameter is NULL and
16718 for CFB and OFB modes they zero ctx->num.
16720 New functionality allows removal of S/MIME code RC2 hack.
16722 Most of the routines have the same form and so can be declared in terms
16725 By shifting this to the top level EVP_CipherInit() it can be removed from
16726 all individual ciphers. If the cipher wants to handle IVs or keys
16727 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
16730 Change lots of functions like EVP_EncryptUpdate() to now return a
16731 value: although software versions of the algorithms cannot fail
16732 any installed hardware versions can.
16736 * Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
16737 this option is set, tolerate broken clients that send the negotiated
16738 protocol version number instead of the requested protocol version
16743 * Call dh_tmp_cb (set by `..._TMP_DH_CB`) with correct 'is_export' flag;
16744 i.e. non-zero for export ciphersuites, zero otherwise.
16745 Previous versions had this flag inverted, inconsistent with
16746 rsa_tmp_cb (..._TMP_RSA_CB).
16748 *Bodo Moeller; problem reported by Amit Chopra*
16750 * Add missing DSA library text string. Work around for some IIS
16751 key files with invalid SEQUENCE encoding.
16755 * Add a document (doc/standards.txt) that list all kinds of standards
16756 and so on that are implemented in OpenSSL.
16760 * Enhance c_rehash script. Old version would mishandle certificates
16761 with the same subject name hash and wouldn't handle CRLs at all.
16762 Added -fingerprint option to crl utility, to support new c_rehash
16767 * Eliminate non-ANSI declarations in crypto.h and stack.h.
16771 * Fix for SSL server purpose checking. Server checking was
16772 rejecting certificates which had extended key usage present
16773 but no ssl client purpose.
16775 *Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>*
16777 * Make PKCS#12 code work with no password. The PKCS#12 spec
16778 is a little unclear about how a blank password is handled.
16779 Since the password in encoded as a BMPString with terminating
16780 double NULL a zero length password would end up as just the
16781 double NULL. However no password at all is different and is
16782 handled differently in the PKCS#12 key generation code. NS
16783 treats a blank password as zero length. MSIE treats it as no
16784 password on export: but it will try both on import. We now do
16785 the same: PKCS12_parse() tries zero length and no password if
16786 the password is set to "" or NULL (NULL is now a valid password:
16787 it wasn't before) as does the pkcs12 application.
16791 * Bugfixes in `apps/x509.c`: Avoid a memory leak; and don't use
16792 perror when PEM_read_bio_X509_REQ fails, the error message must
16793 be obtained from the error queue.
16797 * Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
16798 it in ERR_remove_state if appropriate, and change ERR_get_state
16799 accordingly to avoid race conditions (this is necessary because
16800 thread_hash is no longer constant once set).
16804 * Bugfix for linux-elf makefile.one.
16808 * RSA_get_default_method() will now cause a default
16809 RSA_METHOD to be chosen if one doesn't exist already.
16810 Previously this was only set during a call to RSA_new()
16811 or RSA_new_method(NULL) meaning it was possible for
16812 RSA_get_default_method() to return NULL.
16816 * Added native name translation to the existing DSO code
16817 that will convert (if the flag to do so is set) filenames
16818 that are sufficiently small and have no path information
16819 into a canonical native form. Eg. "blah" converted to
16820 "libblah.so" or "blah.dll" etc.
16824 * New function ERR_error_string_n(e, buf, len) which is like
16825 ERR_error_string(e, buf), but writes at most 'len' bytes
16826 including the 0 terminator. For ERR_error_string_n, 'buf'
16829 *Damien Miller <djm@mindrot.org>, Bodo Moeller*
16831 * CONF library reworked to become more general. A new CONF
16832 configuration file reader "class" is implemented as well as a
16833 new functions (`NCONF_*`, for "New CONF") to handle it. The now
16834 old `CONF_*` functions are still there, but are reimplemented to
16835 work in terms of the new functions. Also, a set of functions
16836 to handle the internal storage of the configuration data is
16837 provided to make it easier to write new configuration file
16838 reader "classes" (I can definitely see something reading a
16839 configuration file in XML format, for example), called `_CONF_*`,
16840 or "the configuration storage API"...
16842 The new configuration file reading functions are:
16844 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
16845 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
16847 NCONF_default, NCONF_WIN32
16849 NCONF_dump_fp, NCONF_dump_bio
16851 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
16852 NCONF_new creates a new CONF object. This works in the same way
16853 as other interfaces in OpenSSL, like the BIO interface.
16854 `NCONF_dump_*` dump the internal storage of the configuration file,
16855 which is useful for debugging. All other functions take the same
16856 arguments as the old `CONF_*` functions with the exception of the
16857 first that must be a `CONF *` instead of a `LHASH *`.
16859 To make it easier to use the new classes with the old `CONF_*` functions,
16860 the function CONF_set_default_method is provided.
16864 * Add '-tls1' option to 'openssl ciphers', which was already
16865 mentioned in the documentation but had not been implemented.
16866 (This option is not yet really useful because even the additional
16867 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
16871 * Initial DSO code added into libcrypto for letting OpenSSL (and
16872 OpenSSL-based applications) load shared libraries and bind to
16873 them in a portable way.
16875 *Geoff Thorpe, with contributions from Richard Levitte*
16877 ### Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
16879 * Make sure _lrotl and _lrotr are only used with MSVC.
16881 * Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
16882 (the default implementation of RAND_status).
16884 * Rename openssl x509 option '-crlext', which was added in 0.9.5,
16885 to '-clrext' (= clear extensions), as intended and documented.
16886 *Bodo Moeller; inconsistency pointed out by Michael Attili
16887 <attili@amaxo.com>*
16889 * Fix for HMAC. It wasn't zeroing the rest of the block if the key length
16890 was larger than the MD block size.
16892 *Steve Henson, pointed out by Yost William <YostW@tce.com>*
16894 * Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
16895 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
16896 using the passed key: if the passed key was a private key the result
16897 of X509_print(), for example, would be to print out all the private key
16902 * des_quad_cksum() byte order bug fix.
16903 *Ulf Möller, using the problem description in krb4-0.9.7, where
16904 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>*
16906 * Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
16909 *Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>*
16911 * For easily testing in shell scripts whether some command
16912 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
16913 returns with exit code 0 iff no command of the given name is available.
16914 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
16915 the output goes to stdout and nothing is printed to stderr.
16916 Additional arguments are always ignored.
16918 Since for each cipher there is a command of the same name,
16919 the 'no-cipher' compilation switches can be tested this way.
16921 ('openssl no-XXX' is not able to detect pseudo-commands such
16922 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
16926 * Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
16930 * For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
16931 is set; it will be thrown away anyway because each handshake creates
16933 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
16934 to parameters -- in previous versions (since OpenSSL 0.9.3) the
16935 'default key' from SSL_CTX_set_tmp_dh would always be lost, meaning
16936 you effectively got SSL_OP_SINGLE_DH_USE when using this macro.
16940 * New s_client option -ign_eof: EOF at stdin is ignored, and
16941 'Q' and 'R' lose their special meanings (quit/renegotiate).
16942 This is part of what -quiet does; unlike -quiet, -ign_eof
16943 does not suppress any output.
16947 * Add compatibility options to the purpose and trust code. The
16948 purpose X509_PURPOSE_ANY is "any purpose" which automatically
16949 accepts a certificate or CA, this was the previous behaviour,
16950 with all the associated security issues.
16952 X509_TRUST_COMPAT is the old trust behaviour: only and
16953 automatically trust self signed roots in certificate store. A
16954 new trust setting X509_TRUST_DEFAULT is used to specify that
16955 a purpose has no associated trust setting and it should instead
16956 use the value in the default purpose.
16960 * Fix the PKCS#8 DSA private key code so it decodes keys again
16961 and fix a memory leak.
16965 * In util/mkerr.pl (which implements 'make errors'), preserve
16966 reason strings from the previous version of the .c file, as
16967 the default to have only downcase letters (and digits) in
16968 automatically generated reasons codes is not always appropriate.
16972 * In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
16973 using strerror. Previously, ERR_reason_error_string() returned
16974 library names as reason strings for SYSerr; but SYSerr is a special
16975 case where small numbers are errno values, not library numbers.
16979 * Add '-dsaparam' option to 'openssl dhparam' application. This
16980 converts DSA parameters into DH parameters. (When creating parameters,
16981 DSA_generate_parameters is used.)
16985 * Include 'length' (recommended exponent length) in C code generated
16986 by 'openssl dhparam -C'.
16990 * The second argument to set_label in perlasm was already being used
16991 so couldn't be used as a "file scope" flag. Moved to third argument
16996 * In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
16997 instead of RAND_bytes for encryption IVs and salts.
17001 * Include RAND_status() into RAND_METHOD instead of implementing
17002 it only for md_rand.c Otherwise replacing the PRNG by calling
17003 RAND_set_rand_method would be impossible.
17007 * Don't let DSA_generate_key() enter an infinite loop if the random
17008 number generation fails.
17012 * New 'rand' application for creating pseudo-random output.
17016 * Added configuration support for Linux/IA64
17018 *Rolf Haberrecker <rolf@suse.de>*
17020 * Assembler module support for Mingw32.
17024 * Shared library support for HPUX (in shlib/).
17026 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous*
17028 * Shared library support for Solaris gcc.
17030 *Lutz Behnke <behnke@trustcenter.de>*
17032 ### Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
17034 * PKCS7_encrypt() was adding text MIME headers twice because they
17035 were added manually and by SMIME_crlf_copy().
17039 * In bntest.c don't call BN_rand with zero bits argument.
17041 *Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>*
17043 * BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
17044 case was implemented. This caused BN_div_recp() to fail occasionally.
17048 * Add an optional second argument to the set_label() in the perl
17049 assembly language builder. If this argument exists and is set
17050 to 1 it signals that the assembler should use a symbol whose
17051 scope is the entire file, not just the current function. This
17052 is needed with MASM which uses the format label:: for this scope.
17054 *Steve Henson, pointed out by Peter Runestig <peter@runestig.com>*
17056 * Change the ASN1 types so they are typedefs by default. Before
17057 almost all types were #define'd to ASN1_STRING which was causing
17058 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
17063 * Change names of new functions to the new get1/get0 naming
17064 convention: After 'get1', the caller owns a reference count
17065 and has to call `..._free`; 'get0' returns a pointer to some
17066 data structure without incrementing reference counters.
17067 (Some of the existing 'get' functions increment a reference
17068 counter, some don't.)
17069 Similarly, 'set1' and 'add1' functions increase reference
17070 counters or duplicate objects.
17074 * Allow for the possibility of temp RSA key generation failure:
17075 the code used to assume it always worked and crashed on failure.
17079 * Fix potential buffer overrun problem in BIO_printf().
17080 *Ulf Möller, using public domain code by Patrick Powell; problem
17081 pointed out by David Sacerdote <das33@cornell.edu>*
17083 * Support EGD <http://www.lothar.com/tech/crypto/>. New functions
17084 RAND_egd() and RAND_status(). In the command line application,
17085 the EGD socket can be specified like a seed file using RANDFILE
17090 * Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
17091 Some CAs (e.g. Verisign) distribute certificates in this form.
17095 * Remove the SSL_ALLOW_ADH compile option and set the default cipher
17096 list to exclude them. This means that no special compilation option
17097 is needed to use anonymous DH: it just needs to be included in the
17102 * Change the EVP_MD_CTX_type macro so its meaning consistent with
17103 EVP_MD_type. The old functionality is available in a new macro called
17104 EVP_MD_md(). Change code that uses it and update docs.
17108 * `..._ctrl` functions now have corresponding `..._callback_ctrl` functions
17109 where the `void *` argument is replaced by a function pointer argument.
17110 Previously `void *` was abused to point to functions, which works on
17111 many platforms, but is not correct. As these functions are usually
17112 called by macros defined in OpenSSL header files, most source code
17113 should work without changes.
17117 * `<openssl/opensslconf.h>` (which is created by Configure) now contains
17118 sections with information on -D... compiler switches used for
17119 compiling the library so that applications can see them. To enable
17120 one of these sections, a pre-processor symbol `OPENSSL_..._DEFINES`
17121 must be defined. E.g.,
17122 #define OPENSSL_ALGORITHM_DEFINES
17123 #include <openssl/opensslconf.h>
17124 defines all pertinent `NO_<algo>` symbols, such as NO_IDEA, NO_RSA, etc.
17126 *Richard Levitte, Ulf and Bodo Möller*
17128 * Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
17133 * Change the 'other' type in certificate aux info to a STACK_OF
17134 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
17135 the required ASN1 format: arbitrary types determined by an OID.
17139 * Add some PEM_write_X509_REQ_NEW() functions and a command line
17140 argument to 'req'. This is not because the function is newer or
17141 better than others it just uses the work 'NEW' in the certificate
17142 request header lines. Some software needs this.
17146 * Reorganise password command line arguments: now passwords can be
17147 obtained from various sources. Delete the PEM_cb function and make
17148 it the default behaviour: i.e. if the callback is NULL and the
17149 usrdata argument is not NULL interpret it as a null terminated pass
17150 phrase. If usrdata and the callback are NULL then the pass phrase
17151 is prompted for as usual.
17155 * Add support for the Compaq Atalla crypto accelerator. If it is installed,
17156 the support is automatically enabled. The resulting binaries will
17157 autodetect the card and use it if present.
17159 *Ben Laurie and Compaq Inc.*
17161 * Work around for Netscape hang bug. This sends certificate request
17162 and server done in one record. Since this is perfectly legal in the
17163 SSL/TLS protocol it isn't a "bug" option and is on by default. See
17164 the bugs/SSLv3 entry for more info.
17168 * HP-UX tune-up: new unified configs, HP C compiler bug workaround.
17172 * Add -rand argument to smime and pkcs12 applications and read/write
17177 * New 'passwd' tool for crypt(3) and apr1 password hashes.
17181 * Add command line password options to the remaining applications.
17185 * Bug fix for BN_div_recp() for numerators with an even number of
17190 * More tests in bntest.c, and changed test_bn output.
17194 * ./config recognizes MacOS X now.
17198 * Bug fix for BN_div() when the first words of num and divisor are
17199 equal (it gave wrong results if `(rem=(n1-q*d0)&BN_MASK2) < d0)`.
17203 * Add support for various broken PKCS#8 formats, and command line
17204 options to produce them.
17208 * New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
17209 get temporary BIGNUMs from a BN_CTX.
17213 * Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
17218 * Change the `SSLeay_add_all_*()` functions to `OpenSSL_add_all_*()` and
17219 include a #define from the old name to the new. The original intent
17220 was that statically linked binaries could for example just call
17221 SSLeay_add_all_ciphers() to just add ciphers to the table and not
17222 link with digests. This never worked because SSLeay_add_all_digests()
17223 and SSLeay_add_all_ciphers() were in the same source file so calling
17224 one would link with the other. They are now in separate source files.
17228 * Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
17232 * Use a less unusual form of the Miller-Rabin primality test (it used
17233 a binary algorithm for exponentiation integrated into the Miller-Rabin
17234 loop, our standard modexp algorithms are faster).
17238 * Support for the EBCDIC character set completed.
17240 *Martin Kraemer <Martin.Kraemer@Mch.SNI.De>*
17242 * Source code cleanups: use const where appropriate, eliminate casts,
17243 use `void *` instead of `char *` in lhash.
17247 * Bugfix: ssl3_send_server_key_exchange was not restartable
17248 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
17249 this the server could overwrite ephemeral keys that the client
17254 * Turn DSA_is_prime into a macro that calls BN_is_prime,
17255 using 50 iterations of the Rabin-Miller test.
17257 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
17258 iterations of the Rabin-Miller test as required by the appendix
17259 to FIPS PUB 186[-1]) instead of DSA_is_prime.
17260 As BN_is_prime_fasttest includes trial division, DSA parameter
17261 generation becomes much faster.
17263 This implies a change for the callback functions in DSA_is_prime
17264 and DSA_generate_parameters: The callback function is called once
17265 for each positive witness in the Rabin-Miller test, not just
17266 occasionally in the inner loop; and the parameters to the
17267 callback function now provide an iteration count for the outer
17268 loop rather than for the current invocation of the inner loop.
17269 DSA_generate_parameters additionally can call the callback
17270 function with an 'iteration count' of -1, meaning that a
17271 candidate has passed the trial division test (when q is generated
17272 from an application-provided seed, trial division is skipped).
17276 * New function BN_is_prime_fasttest that optionally does trial
17277 division before starting the Rabin-Miller test and has
17278 an additional BN_CTX * argument (whereas BN_is_prime always
17279 has to allocate at least one BN_CTX).
17280 'callback(1, -1, cb_arg)' is called when a number has passed the
17281 trial division stage.
17285 * Fix for bug in CRL encoding. The validity dates weren't being handled
17290 * New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
17294 * New function BN_pseudo_rand().
17298 * Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
17299 bignum version of BN_from_montgomery() with the working code from
17300 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
17305 * Avoid a race condition in s2_clnt.c (function get_server_hello) that
17306 made it impossible to use the same SSL_SESSION data structure in
17307 SSL2 clients in multiple threads.
17311 * The return value of RAND_load_file() no longer counts bytes obtained
17312 by stat(). RAND_load_file(..., -1) is new and uses the complete file
17313 to seed the PRNG (previously an explicit byte count was required).
17315 *Ulf Möller, Bodo Möller*
17317 * Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
17318 used `char *` instead of `void *` and had casts all over the place.
17322 * Make BN_generate_prime() return NULL on error if ret!=NULL.
17326 * Retain source code compatibility for BN_prime_checks macro:
17327 BN_is_prime(..., BN_prime_checks, ...) now uses
17328 BN_prime_checks_for_size to determine the appropriate number of
17329 Rabin-Miller iterations.
17333 * Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
17334 DH_CHECK_P_NOT_SAFE_PRIME.
17335 (Check if this is true? OpenPGP calls them "strong".)
17339 * Merge the functionality of "dh" and "gendh" programs into a new program
17340 "dhparam". The old programs are retained for now but will handle DH keys
17341 (instead of parameters) in future.
17345 * Make the ciphers, s_server and s_client programs check the return values
17346 when a new cipher list is set.
17350 * Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
17351 ciphers. Before when the 56bit ciphers were enabled the sorting was
17354 The syntax for the cipher sorting has been extended to support sorting by
17355 cipher-strength (using the strength_bits hard coded in the tables).
17356 The new command is `@STRENGTH` (see also `doc/apps/ciphers.pod`).
17358 Fix a bug in the cipher-command parser: when supplying a cipher command
17359 string with an "undefined" symbol (neither command nor alphanumeric
17360 *A-Za-z0-9*, ssl_set_cipher_list used to hang in an endless loop. Now
17361 an error is flagged.
17363 Due to the strength-sorting extension, the code of the
17364 ssl_create_cipher_list() function was completely rearranged. I hope that
17365 the readability was also increased :-)
17367 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>*
17369 * Minor change to 'x509' utility. The -CAcreateserial option now uses 1
17370 for the first serial number and places 2 in the serial number file. This
17371 avoids problems when the root CA is created with serial number zero and
17372 the first user certificate has the same issuer name and serial number
17377 * Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
17378 the new code. Add documentation for this stuff.
17382 * Changes to X509_ATTRIBUTE utilities. These have been renamed from
17383 `X509_*()` to `X509at_*()` on the grounds that they don't handle X509
17384 structures and behave in an analogous way to the X509v3 functions:
17385 they shouldn't be called directly but wrapper functions should be used
17388 So we also now have some wrapper functions that call the X509at functions
17389 when passed certificate requests. (TO DO: similar things can be done with
17390 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
17391 things. Some of these need some d2i or i2d and print functionality
17392 because they handle more complex structures.)
17396 * Add missing #ifndefs that caused missing symbols when building libssl
17397 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
17398 NO_RSA in `ssl/s2*.c`.
17400 *Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller*
17402 * Precautions against using the PRNG uninitialized: RAND_bytes() now
17403 has a return value which indicates the quality of the random data
17404 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
17405 error queue. New function RAND_pseudo_bytes() generates output that is
17406 guaranteed to be unique but not unpredictable. RAND_add is like
17407 RAND_seed, but takes an extra argument for an entropy estimate
17408 (RAND_seed always assumes full entropy).
17412 * Do more iterations of Rabin-Miller probable prime test (specifically,
17413 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
17414 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
17415 in crypto/bn/bn_prime.c for the complete table). This guarantees a
17416 false-positive rate of at most 2^-80 for random input.
17420 * Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
17424 * New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
17425 in the 0.9.5 release), this returns the chain
17426 from an X509_CTX structure with a dup of the stack and all
17427 the X509 reference counts upped: so the stack will exist
17428 after X509_CTX_cleanup() has been called. Modify pkcs12.c
17431 Also make SSL_SESSION_print() print out the verify return
17436 * Add manpage for the pkcs12 command. Also change the default
17437 behaviour so MAC iteration counts are used unless the new
17438 -nomaciter option is used. This improves file security and
17439 only older versions of MSIE (4.0 for example) need it.
17443 * Honor the no-xxx Configure options when creating .DEF files.
17447 * Add PKCS#10 attributes to field table: challengePassword,
17448 unstructuredName and unstructuredAddress. These are taken from
17449 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
17450 international characters are used.
17452 More changes to X509_ATTRIBUTE code: allow the setting of types
17453 based on strings. Remove the 'loc' parameter when adding
17454 attributes because these will be a SET OF encoding which is sorted
17459 * Initial changes to the 'req' utility to allow request generation
17460 automation. This will allow an application to just generate a template
17461 file containing all the field values and have req construct the
17464 Initial support for X509_ATTRIBUTE handling. Stacks of these are
17465 used all over the place including certificate requests and PKCS#7
17466 structures. They are currently handled manually where necessary with
17467 some primitive wrappers for PKCS#7. The new functions behave in a
17468 manner analogous to the X509 extension functions: they allow
17469 attributes to be looked up by NID and added.
17471 Later something similar to the X509V3 code would be desirable to
17472 automatically handle the encoding, decoding and printing of the
17473 more complex types. The string types like challengePassword can
17474 be handled by the string table functions.
17476 Also modified the multi byte string table handling. Now there is
17477 a 'global mask' which masks out certain types. The table itself
17478 can use the flag STABLE_NO_MASK to ignore the mask setting: this
17479 is useful when for example there is only one permissible type
17480 (as in countryName) and using the mask might result in no valid
17485 * Clean up 'Finished' handling, and add functions SSL_get_finished and
17486 SSL_get_peer_finished to allow applications to obtain the latest
17487 Finished messages sent to the peer or expected from the peer,
17488 respectively. (SSL_get_peer_finished is usually the Finished message
17489 actually received from the peer, otherwise the protocol will be aborted.)
17491 As the Finished message are message digests of the complete handshake
17492 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
17493 be used for external authentication procedures when the authentication
17494 provided by SSL/TLS is not desired or is not enough.
17498 * Enhanced support for Alpha Linux is added. Now ./config checks if
17499 the host supports BWX extension and if Compaq C is present on the
17500 $PATH. Just exploiting of the BWX extension results in 20-30%
17501 performance kick for some algorithms, e.g. DES and RC4 to mention
17502 a couple. Compaq C in turn generates ~20% faster code for MD5 and
17507 * Add support for MS "fast SGC". This is arguably a violation of the
17508 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
17509 weak crypto and after checking the certificate is SGC a second one
17510 with strong crypto. MS SGC stops the first handshake after receiving
17511 the server certificate message and sends a second client hello. Since
17512 a server will typically do all the time consuming operations before
17513 expecting any further messages from the client (server key exchange
17514 is the most expensive) there is little difference between the two.
17516 To get OpenSSL to support MS SGC we have to permit a second client
17517 hello message after we have sent server done. In addition we have to
17518 reset the MAC if we do get this second client hello.
17522 * Add a function 'd2i_AutoPrivateKey()' this will automatically decide
17523 if a DER encoded private key is RSA or DSA traditional format. Changed
17524 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
17525 format DER encoded private key. Newer code should use PKCS#8 format which
17526 has the key type encoded in the ASN1 structure. Added DER private key
17527 support to pkcs8 application.
17531 * SSL 3/TLS 1 servers now don't request certificates when an anonymous
17532 ciphersuites has been selected (as required by the SSL 3/TLS 1
17533 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
17534 is set, we interpret this as a request to violate the specification
17535 (the worst that can happen is a handshake failure, and 'correct'
17536 behaviour would result in a handshake failure anyway).
17540 * In SSL_CTX_add_session, take into account that there might be multiple
17541 SSL_SESSION structures with the same session ID (e.g. when two threads
17542 concurrently obtain them from an external cache).
17543 The internal cache can handle only one SSL_SESSION with a given ID,
17544 so if there's a conflict, we now throw out the old one to achieve
17549 * Add OIDs for idea and blowfish in CBC mode. This will allow both
17550 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
17551 some routines that use cipher OIDs: some ciphers do not have OIDs
17552 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
17557 * Simplify the trust setting structure and code. Now we just have
17558 two sequences of OIDs for trusted and rejected settings. These will
17559 typically have values the same as the extended key usage extension
17560 and any application specific purposes.
17562 The trust checking code now has a default behaviour: it will just
17563 check for an object with the same NID as the passed id. Functions can
17564 be provided to override either the default behaviour or the behaviour
17565 for a given id. SSL client, server and email already have functions
17566 in place for compatibility: they check the NID and also return "trusted"
17567 if the certificate is self signed.
17571 * Add d2i,i2d bio/fp functions for PrivateKey: these convert the
17572 traditional format into an EVP_PKEY structure.
17576 * Add a password callback function PEM_cb() which either prompts for
17577 a password if usr_data is NULL or otherwise assumes it is a null
17578 terminated password. Allow passwords to be passed on command line
17579 environment or config files in a few more utilities.
17583 * Add a bunch of DER and PEM functions to handle PKCS#8 format private
17584 keys. Add some short names for PKCS#8 PBE algorithms and allow them
17585 to be specified on the command line for the pkcs8 and pkcs12 utilities.
17586 Update documentation.
17590 * Support for ASN1 "NULL" type. This could be handled before by using
17591 ASN1_TYPE but there wasn't any function that would try to read a NULL
17592 and produce an error if it couldn't. For compatibility we also have
17593 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
17594 don't allocate anything because they don't need to.
17598 * Initial support for MacOS is now provided. Examine INSTALL.MacOS
17601 *Andy Polyakov, Roy Woods <roy@centicsystems.ca>*
17603 * Rebuild of the memory allocation routines used by OpenSSL code and
17604 possibly others as well. The purpose is to make an interface that
17605 provide hooks so anyone can build a separate set of allocation and
17606 deallocation routines to be used by OpenSSL, for example memory
17607 pool implementations, or something else, which was previously hard
17608 since Malloc(), Realloc() and Free() were defined as macros having
17609 the values malloc, realloc and free, respectively (except for Win32
17610 compilations). The same is provided for memory debugging code.
17611 OpenSSL already comes with functionality to find memory leaks, but
17612 this gives people a chance to debug other memory problems.
17614 With these changes, a new set of functions and macros have appeared:
17616 CRYPTO_set_mem_debug_functions() [F]
17617 CRYPTO_get_mem_debug_functions() [F]
17618 CRYPTO_dbg_set_options() [F]
17619 CRYPTO_dbg_get_options() [F]
17620 CRYPTO_malloc_debug_init() [M]
17622 The memory debug functions are NULL by default, unless the library
17623 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
17624 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
17625 gives the standard debugging functions that come with OpenSSL) or
17626 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
17627 provided by the library user) must be used. When the standard
17628 debugging functions are used, CRYPTO_dbg_set_options can be used to
17629 request additional information:
17630 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
17631 the CRYPTO_MDEBUG_xxx macro when compiling the library.
17633 Also, things like CRYPTO_set_mem_functions will always give the
17634 expected result (the new set of functions is used for allocation
17635 and deallocation) at all times, regardless of platform and compiler
17638 To finish it up, some functions that were never use in any other
17639 way than through macros have a new API and new semantic:
17641 CRYPTO_dbg_malloc()
17642 CRYPTO_dbg_realloc()
17645 All macros of value have retained their old syntax.
17647 *Richard Levitte and Bodo Moeller*
17649 * Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
17650 ordering of SMIMECapabilities wasn't in "strength order" and there
17651 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
17656 * Some ASN1 types with illegal zero length encoding (INTEGER,
17657 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
17659 *Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson*
17661 * Merge in my S/MIME library for OpenSSL. This provides a simple
17662 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
17663 functionality to handle multipart/signed properly) and a utility
17664 called 'smime' to call all this stuff. This is based on code I
17665 originally wrote for Celo who have kindly allowed it to be
17666 included in OpenSSL.
17670 * Add variants des_set_key_checked and des_set_key_unchecked of
17671 des_set_key (aka des_key_sched). Global variable des_check_key
17672 decides which of these is called by des_set_key; this way
17673 des_check_key behaves as it always did, but applications and
17674 the library itself, which was buggy for des_check_key == 1,
17675 have a cleaner way to pick the version they need.
17679 * New function PKCS12_newpass() which changes the password of a
17684 * Modify X509_TRUST and X509_PURPOSE so it also uses a static and
17685 dynamic mix. In both cases the ids can be used as an index into the
17686 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
17687 functions so they accept a list of the field values and the
17688 application doesn't need to directly manipulate the X509_TRUST
17693 * Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
17698 * Modify the way the V3 extension code looks up extensions. This now
17699 works in a similar way to the object code: we have some "standard"
17700 extensions in a static table which is searched with OBJ_bsearch()
17701 and the application can add dynamic ones if needed. The file
17702 crypto/x509v3/ext_dat.h now has the info: this file needs to be
17703 updated whenever a new extension is added to the core code and kept
17704 in ext_nid order. There is a simple program 'tabtest.c' which checks
17705 this. New extensions are not added too often so this file can readily
17706 be maintained manually.
17708 There are two big advantages in doing things this way. The extensions
17709 can be looked up immediately and no longer need to be "added" using
17710 X509V3_add_standard_extensions(): this function now does nothing.
17711 Side note: I get *lots* of email saying the extension code doesn't
17712 work because people forget to call this function.
17713 Also no dynamic allocation is done unless new extensions are added:
17714 so if we don't add custom extensions there is no need to call
17715 X509V3_EXT_cleanup().
17719 * Modify enc utility's salting as follows: make salting the default. Add a
17720 magic header, so unsalted files fail gracefully instead of just decrypting
17721 to garbage. This is because not salting is a big security hole, so people
17722 should be discouraged from doing it.
17726 * Fixes and enhancements to the 'x509' utility. It allowed a message
17727 digest to be passed on the command line but it only used this
17728 parameter when signing a certificate. Modified so all relevant
17729 operations are affected by the digest parameter including the
17730 -fingerprint and -x509toreq options. Also -x509toreq choked if a
17731 DSA key was used because it didn't fix the digest.
17735 * Initial certificate chain verify code. Currently tests the untrusted
17736 certificates for consistency with the verify purpose (which is set
17737 when the X509_STORE_CTX structure is set up) and checks the pathlength.
17739 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
17740 this is because it will reject chains with invalid extensions whereas
17741 every previous version of OpenSSL and SSLeay made no checks at all.
17743 Trust code: checks the root CA for the relevant trust settings. Trust
17744 settings have an initial value consistent with the verify purpose: e.g.
17745 if the verify purpose is for SSL client use it expects the CA to be
17746 trusted for SSL client use. However the default value can be changed to
17747 permit custom trust settings: one example of this would be to only trust
17748 certificates from a specific "secure" set of CAs.
17750 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
17751 which should be used for version portability: especially since the
17752 verify structure is likely to change more often now.
17754 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
17755 to set them. If not set then assume SSL clients will verify SSL servers
17758 Two new options to the verify program: -untrusted allows a set of
17759 untrusted certificates to be passed in and -purpose which sets the
17760 intended purpose of the certificate. If a purpose is set then the
17761 new chain verify code is used to check extension consistency.
17765 * Support for the authority information access extension.
17769 * Modify RSA and DSA PEM read routines to transparently handle
17770 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
17771 public keys in a format compatible with certificate
17772 SubjectPublicKeyInfo structures. Unfortunately there were already
17773 functions called *_PublicKey_* which used various odd formats so
17774 these are retained for compatibility: however the DSA variants were
17775 never in a public release so they have been deleted. Changed dsa/rsa
17776 utilities to handle the new format: note no releases ever handled public
17777 keys so we should be OK.
17779 The primary motivation for this change is to avoid the same fiasco
17780 that dogs private keys: there are several incompatible private key
17781 formats some of which are standard and some OpenSSL specific and
17782 require various evil hacks to allow partial transparent handling and
17783 even then it doesn't work with DER formats. Given the option anything
17784 other than PKCS#8 should be dumped: but the other formats have to
17785 stay in the name of compatibility.
17787 With public keys and the benefit of hindsight one standard format
17788 is used which works with EVP_PKEY, RSA or DSA structures: though
17789 it clearly returns an error if you try to read the wrong kind of key.
17791 Added a -pubkey option to the 'x509' utility to output the public key.
17792 Also rename the `EVP_PKEY_get_*()` to `EVP_PKEY_rget_*()`
17793 (renamed to `EVP_PKEY_get1_*()` in the OpenSSL 0.9.5 release) and add
17794 `EVP_PKEY_rset_*()` functions (renamed to `EVP_PKEY_set1_*()`)
17795 that do the same as the `EVP_PKEY_assign_*()` except they up the
17796 reference count of the added key (they don't "swallow" the
17801 * Fixes to crypto/x509/by_file.c the code to read in certificates and
17802 CRLs would fail if the file contained no certificates or no CRLs:
17803 added a new function to read in both types and return the number
17804 read: this means that if none are read it will be an error. The
17805 DER versions of the certificate and CRL reader would always fail
17806 because it isn't possible to mix certificates and CRLs in DER format
17807 without choking one or the other routine. Changed this to just read
17808 a certificate: this is the best we can do. Also modified the code
17809 in `apps/verify.c` to take notice of return codes: it was previously
17810 attempting to read in certificates from NULL pointers and ignoring
17811 any errors: this is one reason why the cert and CRL reader seemed
17812 to work. It doesn't check return codes from the default certificate
17813 routines: these may well fail if the certificates aren't installed.
17817 * Code to support otherName option in GeneralName.
17821 * First update to verify code. Change the verify utility
17822 so it warns if it is passed a self signed certificate:
17823 for consistency with the normal behaviour. X509_verify
17824 has been modified to it will now verify a self signed
17825 certificate if *exactly* the same certificate appears
17826 in the store: it was previously impossible to trust a
17827 single self signed certificate. This means that:
17828 openssl verify ss.pem
17829 now gives a warning about a self signed certificate but
17830 openssl verify -CAfile ss.pem ss.pem
17835 * For servers, store verify_result in SSL_SESSION data structure
17836 (and add it to external session representation).
17837 This is needed when client certificate verifications fails,
17838 but an application-provided verification callback (set by
17839 SSL_CTX_set_cert_verify_callback) allows accepting the session
17840 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
17841 but returns 1): When the session is reused, we have to set
17842 ssl->verify_result to the appropriate error code to avoid
17845 *Bodo Moeller, problem pointed out by Lutz Jaenicke*
17847 * Fix a bug in the new PKCS#7 code: it didn't consider the
17848 case in PKCS7_dataInit() where the signed PKCS7 structure
17849 didn't contain any existing data because it was being created.
17851 *Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson*
17853 * Add a salt to the key derivation routines in enc.c. This
17854 forms the first 8 bytes of the encrypted file. Also add a
17855 -S option to allow a salt to be input on the command line.
17859 * New function X509_cmp(). Oddly enough there wasn't a function
17860 to compare two certificates. We do this by working out the SHA1
17861 hash and comparing that. X509_cmp() will be needed by the trust
17866 * SSL_get1_session() is like SSL_get_session(), but increments
17867 the reference count in the SSL_SESSION returned.
17869 *Geoff Thorpe <geoff@eu.c2.net>*
17871 * Fix for 'req': it was adding a null to request attributes.
17872 Also change the X509_LOOKUP and X509_INFO code to handle
17873 certificate auxiliary information.
17877 * Add support for 40 and 64 bit RC2 and RC4 algorithms: document
17882 * Add the possibility to add extra information to the memory leak
17883 detecting output, to form tracebacks, showing from where each
17884 allocation was originated: CRYPTO_push_info("constant string") adds
17885 the string plus current file name and line number to a per-thread
17886 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
17887 is like calling CYRPTO_pop_info() until the stack is empty.
17888 Also updated memory leak detection code to be multi-thread-safe.
17892 * Add options -text and -noout to pkcs7 utility and delete the
17893 encryption options which never did anything. Update docs.
17897 * Add options to some of the utilities to allow the pass phrase
17898 to be included on either the command line (not recommended on
17899 OSes like Unix) or read from the environment. Update the
17900 manpages and fix a few bugs.
17904 * Add a few manpages for some of the openssl commands.
17908 * Fix the -revoke option in ca. It was freeing up memory twice,
17909 leaking and not finding already revoked certificates.
17913 * Extensive changes to support certificate auxiliary information.
17914 This involves the use of X509_CERT_AUX structure and X509_AUX
17915 functions. An X509_AUX function such as PEM_read_X509_AUX()
17916 can still read in a certificate file in the usual way but it
17917 will also read in any additional "auxiliary information". By
17918 doing things this way a fair degree of compatibility can be
17919 retained: existing certificates can have this information added
17920 using the new 'x509' options.
17922 Current auxiliary information includes an "alias" and some trust
17923 settings. The trust settings will ultimately be used in enhanced
17924 certificate chain verification routines: currently a certificate
17925 can only be trusted if it is self signed and then it is trusted
17930 * Fix assembler for Alpha (tested only on DEC OSF not Linux or `*BSD`).
17931 The problem was that one of the replacement routines had not been working
17932 since SSLeay releases. For now the offending routine has been replaced
17933 with non-optimised assembler. Even so, this now gives around 95%
17934 performance improvement for 1024 bit RSA signs.
17938 * Hack to fix PKCS#7 decryption when used with some unorthodox RC2
17939 handling. Most clients have the effective key size in bits equal to
17940 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
17941 A few however don't do this and instead use the size of the decrypted key
17942 to determine the RC2 key length and the AlgorithmIdentifier to determine
17943 the effective key length. In this case the effective key length can still
17944 be 40 bits but the key length can be 168 bits for example. This is fixed
17945 by manually forcing an RC2 key into the EVP_PKEY structure because the
17946 EVP code can't currently handle unusual RC2 key sizes: it always assumes
17947 the key length and effective key length are equal.
17951 * Add a bunch of functions that should simplify the creation of
17952 X509_NAME structures. Now you should be able to do:
17953 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
17954 and have it automatically work out the correct field type and fill in
17955 the structures. The more adventurous can try:
17956 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
17957 and it will (hopefully) work out the correct multibyte encoding.
17961 * Change the 'req' utility to use the new field handling and multibyte
17962 copy routines. Before the DN field creation was handled in an ad hoc
17963 way in req, ca, and x509 which was rather broken and didn't support
17964 BMPStrings or UTF8Strings. Since some software doesn't implement
17965 BMPStrings or UTF8Strings yet, they can be enabled using the config file
17966 using the dirstring_type option. See the new comment in the default
17967 openssl.cnf for more info.
17971 * Make crypto/rand/md_rand.c more robust:
17972 - Assure unique random numbers after fork().
17973 - Make sure that concurrent threads access the global counter and
17974 md serializably so that we never lose entropy in them
17975 or use exactly the same state in multiple threads.
17976 Access to the large state is not always serializable because
17977 the additional locking could be a performance killer, and
17978 md should be large enough anyway.
17982 * New file `apps/app_rand.c` with commonly needed functionality
17983 for handling the random seed file.
17985 Use the random seed file in some applications that previously did not:
17987 dsaparam -genkey (which also ignored its '-rand' option),
17990 x509 (when signing).
17991 Except on systems with /dev/urandom, it is crucial to have a random
17992 seed file at least for key creation, DSA signing, and for DH exchanges;
17993 for RSA signatures we could do without one.
17995 gendh and gendsa (unlike genrsa) used to read only the first byte
17996 of each file listed in the '-rand' option. The function as previously
17997 found in genrsa is now in app_rand.c and is used by all programs
17998 that support '-rand'.
18002 * In RAND_write_file, use mode 0600 for creating files;
18003 don't just chmod when it may be too late.
18007 * Report an error from X509_STORE_load_locations
18008 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
18012 * New function ASN1_mbstring_copy() this copies a string in either
18013 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
18014 into an ASN1_STRING type. A mask of permissible types is passed
18015 and it chooses the "minimal" type to use or an error if not type
18020 * Add function equivalents to the various macros in asn1.h. The old
18021 macros are retained with an `M_` prefix. Code inside the library can
18022 use the `M_` macros. External code (including the openssl utility)
18023 should *NOT* in order to be "shared library friendly".
18027 * Add various functions that can check a certificate's extensions
18028 to see if it usable for various purposes such as SSL client,
18029 server or S/MIME and CAs of these types. This is currently
18030 VERY EXPERIMENTAL but will ultimately be used for certificate chain
18031 verification. Also added a -purpose flag to x509 utility to
18032 print out all the purposes.
18036 * Add a CRYPTO_EX_DATA to X509 certificate structure and associated
18041 * New `X509V3_{X509,CRL,REVOKED}_get_d2i()` functions. These will search
18042 for, obtain and decode and extension and obtain its critical flag.
18043 This allows all the necessary extension code to be handled in a
18044 single function call.
18048 * RC4 tune-up featuring 30-40% performance improvement on most RISC
18049 platforms. See crypto/rc4/rc4_enc.c for further details.
18053 * New -noout option to asn1parse. This causes no output to be produced
18054 its main use is when combined with -strparse and -out to extract data
18055 from a file (which may not be in ASN.1 format).
18059 * Fix for pkcs12 program. It was hashing an invalid certificate pointer
18060 when producing the local key id.
18062 *Richard Levitte <levitte@stacken.kth.se>*
18064 * New option -dhparam in s_server. This allows a DH parameter file to be
18065 stated explicitly. If it is not stated then it tries the first server
18066 certificate file. The previous behaviour hard coded the filename
18071 * Add -pubin and -pubout options to the rsa and dsa commands. These allow
18072 a public key to be input or output. For example:
18073 openssl rsa -in key.pem -pubout -out pubkey.pem
18074 Also added necessary DSA public key functions to handle this.
18078 * Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
18079 in the message. This was handled by allowing
18080 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
18082 *Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>*
18084 * Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
18085 to the end of the strings whereas this didn't. This would cause problems
18086 if strings read with d2i_ASN1_bytes() were later modified.
18088 *Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>*
18090 * Fix for base64 decode bug. When a base64 bio reads only one line of
18091 data and it contains EOF it will end up returning an error. This is
18092 caused by input 46 bytes long. The cause is due to the way base64
18093 BIOs find the start of base64 encoded data. They do this by trying a
18094 trial decode on each line until they find one that works. When they
18095 do a flag is set and it starts again knowing it can pass all the
18096 data directly through the decoder. Unfortunately it doesn't reset
18097 the context it uses. This means that if EOF is reached an attempt
18098 is made to pass two EOFs through the context and this causes the
18099 resulting error. This can also cause other problems as well. As is
18100 usual with these problems it takes *ages* to find and the fix is
18101 trivial: move one line.
18103 *Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer)*
18105 * Ugly workaround to get s_client and s_server working under Windows. The
18106 old code wouldn't work because it needed to select() on sockets and the
18107 tty (for keypresses and to see if data could be written). Win32 only
18108 supports select() on sockets so we select() with a 1s timeout on the
18109 sockets and then see if any characters are waiting to be read, if none
18110 are present then we retry, we also assume we can always write data to
18111 the tty. This isn't nice because the code then blocks until we've
18112 received a complete line of data and it is effectively polling the
18113 keyboard at 1s intervals: however it's quite a bit better than not
18114 working at all :-) A dedicated Windows application might handle this
18115 with an event loop for example.
18119 * Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
18120 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
18121 will be called when RSA_sign() and RSA_verify() are used. This is useful
18122 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
18123 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
18124 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
18125 This necessitated the support of an extra signature type NID_md5_sha1
18126 for SSL signatures and modifications to the SSL library to use it instead
18127 of calling RSA_public_decrypt() and RSA_private_encrypt().
18131 * Add new -verify -CAfile and -CApath options to the crl program, these
18132 will lookup a CRL issuers certificate and verify the signature in a
18133 similar way to the verify program. Tidy up the crl program so it
18134 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
18135 less strict. It will now permit CRL extensions even if it is not
18136 a V2 CRL: this will allow it to tolerate some broken CRLs.
18140 * Initialize all non-automatic variables each time one of the openssl
18141 sub-programs is started (this is necessary as they may be started
18142 multiple times from the "OpenSSL>" prompt).
18144 *Lennart Bang, Bodo Moeller*
18146 * Preliminary compilation option RSA_NULL which disables RSA crypto without
18147 removing all other RSA functionality (this is what NO_RSA does). This
18148 is so (for example) those in the US can disable those operations covered
18149 by the RSA patent while allowing storage and parsing of RSA keys and RSA
18154 * Non-copying interface to BIO pairs.
18155 (still largely untested)
18159 * New function ASN1_tag2str() to convert an ASN1 tag to a descriptive
18160 ASCII string. This was handled independently in various places before.
18164 * New functions UTF8_getc() and UTF8_putc() that parse and generate
18165 UTF8 strings a character at a time.
18169 * Use client_version from client hello to select the protocol
18170 (s23_srvr.c) and for RSA client key exchange verification
18171 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
18175 * Add various utility functions to handle SPKACs, these were previously
18176 handled by poking round in the structure internals. Added new function
18177 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
18178 print, verify and generate SPKACs. Based on an original idea from
18179 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
18183 * RIPEMD160 is operational on all platforms and is back in 'make test'.
18187 * Allow the config file extension section to be overwritten on the
18188 command line. Based on an original idea from Massimiliano Pala
18189 <madwolf@comune.modena.it>. The new option is called -extensions
18190 and can be applied to ca, req and x509. Also -reqexts to override
18191 the request extensions in req and -crlexts to override the crl extensions
18196 * Add new feature to the SPKAC handling in ca. Now you can include
18197 the same field multiple times by preceding it by "XXXX." for example:
18200 this is the same syntax as used in the req config file.
18204 * Allow certificate extensions to be added to certificate requests. These
18205 are specified in a 'req_extensions' option of the req section of the
18206 config file. They can be printed out with the -text option to req but
18207 are otherwise ignored at present.
18211 * Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
18212 data read consists of only the final block it would not decrypted because
18213 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
18214 A misplaced 'break' also meant the decrypted final block might not be
18215 copied until the next read.
18219 * Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
18220 a few extra parameters to the DH structure: these will be useful if
18221 for example we want the value of 'q' or implement X9.42 DH.
18225 * Initial support for DSA_METHOD. This is based on the RSA_METHOD and
18226 provides hooks that allow the default DSA functions or functions on a
18227 "per key" basis to be replaced. This allows hardware acceleration and
18228 hardware key storage to be handled without major modification to the
18229 library. Also added low-level modexp hooks and CRYPTO_EX structure and
18230 associated functions.
18234 * Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
18235 as "read only": it can't be written to and the buffer it points to will
18236 not be freed. Reading from a read only BIO is much more efficient than
18237 a normal memory BIO. This was added because there are several times when
18238 an area of memory needs to be read from a BIO. The previous method was
18239 to create a memory BIO and write the data to it, this results in two
18240 copies of the data and an O(n^2) reading algorithm. There is a new
18241 function BIO_new_mem_buf() which creates a read only memory BIO from
18242 an area of memory. Also modified the PKCS#7 routines to use read only
18247 * Bugfix: ssl23_get_client_hello did not work properly when called in
18248 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
18249 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
18250 but a retry condition occurred while trying to read the rest.
18254 * The PKCS7_ENC_CONTENT_new() function was setting the content type as
18255 NID_pkcs7_encrypted by default: this was wrong since this should almost
18256 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
18257 the encrypted data type: this is a more sensible place to put it and it
18258 allows the PKCS#12 code to be tidied up that duplicated this
18263 * Changed obj_dat.pl script so it takes its input and output files on
18264 the command line. This should avoid shell escape redirection problems
18269 * Initial support for certificate extension requests, these are included
18270 in things like Xenroll certificate requests. Included functions to allow
18271 extensions to be obtained and added.
18275 * -crlf option to s_client and s_server for sending newlines as
18276 CRLF (as required by many protocols).
18280 ### Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
18282 * Install libRSAglue.a when OpenSSL is built with RSAref.
18284 *Ralf S. Engelschall*
18286 * A few more `#ifndef NO_FP_API / #endif` pairs for consistency.
18288 *Andrija Antonijevic <TheAntony2@bigfoot.com>*
18290 * Fix -startdate and -enddate (which was missing) arguments to 'ca'
18295 * New function DSA_dup_DH, which duplicates DSA parameters/keys as
18296 DH parameters/keys (q is lost during that conversion, but the resulting
18297 DH parameters contain its length).
18299 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
18300 much faster than DH_generate_parameters (which creates parameters
18301 where `p = 2*q + 1`), and also the smaller q makes DH computations
18302 much more efficient (160-bit exponentiation instead of 1024-bit
18303 exponentiation); so this provides a convenient way to support DHE
18304 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
18305 utter importance to use
18306 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
18308 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
18309 when such DH parameters are used, because otherwise small subgroup
18310 attacks may become possible!
18314 * Avoid memory leak in i2d_DHparams.
18318 * Allow the -k option to be used more than once in the enc program:
18319 this allows the same encrypted message to be read by multiple recipients.
18323 * New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
18324 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
18325 it will always use the numerical form of the OID, even if it has a short
18330 * Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
18331 method only got called if p,q,dmp1,dmq1,iqmp components were present,
18332 otherwise bn_mod_exp was called. In the case of hardware keys for example
18333 no private key components need be present and it might store extra data
18334 in the RSA structure, which cannot be accessed from bn_mod_exp.
18335 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
18336 private key operations.
18340 * Added support for SPARC Linux.
18344 * pem_password_cb function type incompatibly changed from
18345 typedef int pem_password_cb(char *buf, int size, int rwflag);
18347 ....(char *buf, int size, int rwflag, void *userdata);
18348 so that applications can pass data to their callbacks:
18349 The `PEM[_ASN1]_{read,write}...` functions and macros now take an
18350 additional void * argument, which is just handed through whenever
18351 the password callback is called.
18353 *Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller*
18355 New function SSL_CTX_set_default_passwd_cb_userdata.
18357 Compatibility note: As many C implementations push function arguments
18358 onto the stack in reverse order, the new library version is likely to
18359 interoperate with programs that have been compiled with the old
18360 pem_password_cb definition (PEM_whatever takes some data that
18361 happens to be on the stack as its last argument, and the callback
18362 just ignores this garbage); but there is no guarantee whatsoever that
18365 * The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
18366 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
18367 problems not only on Windows, but also on some Unix platforms.
18368 To avoid problematic command lines, these definitions are now in an
18369 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
18370 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
18374 * MIPS III/IV assembler module is reimplemented.
18378 * More DES library cleanups: remove references to srand/rand and
18379 delete an unused file.
18383 * Add support for the free Netwide assembler (NASM) under Win32,
18384 since not many people have MASM (ml) and it can be hard to obtain.
18385 This is currently experimental but it seems to work OK and pass all
18386 the tests. Check out INSTALL.W32 for info.
18390 * Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
18391 without temporary keys kept an extra copy of the server key,
18392 and connections with temporary keys did not free everything in case
18397 * New function RSA_check_key and new openssl rsa option -check
18398 for verifying the consistency of RSA keys.
18400 *Ulf Moeller, Bodo Moeller*
18402 * Various changes to make Win32 compile work:
18403 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
18404 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
18405 comparison" warnings.
18406 3. Add `sk_<TYPE>_sort` to DEF file generator and do make update.
18410 * Add a debugging option to PKCS#5 v2 key generation function: when
18411 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
18412 derived keys are printed to stderr.
18416 * Copy the flags in ASN1_STRING_dup().
18418 *Roman E. Pavlov <pre@mo.msk.ru>*
18420 * The x509 application mishandled signing requests containing DSA
18421 keys when the signing key was also DSA and the parameters didn't match.
18423 It was supposed to omit the parameters when they matched the signing key:
18424 the verifying software was then supposed to automatically use the CA's
18425 parameters if they were absent from the end user certificate.
18427 Omitting parameters is no longer recommended. The test was also
18428 the wrong way round! This was probably due to unusual behaviour in
18429 EVP_cmp_parameters() which returns 1 if the parameters match.
18430 This meant that parameters were omitted when they *didn't* match and
18431 the certificate was useless. Certificates signed with 'ca' didn't have
18434 *Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>*
18436 * Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
18437 The interface is as follows:
18438 Applications can use
18439 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
18440 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
18441 "off" is now the default.
18442 The library internally uses
18443 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
18444 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
18445 to disable memory-checking temporarily.
18447 Some inconsistent states that previously were possible (and were
18448 even the default) are now avoided.
18450 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
18451 with each memory chunk allocated; this is occasionally more helpful
18452 than just having a counter.
18454 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
18456 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
18461 * Introduce "mode" for SSL structures (with defaults in SSL_CTX),
18462 which largely parallels "options", but is for changing API behaviour,
18463 whereas "options" are about protocol behaviour.
18464 Initial "mode" flags are:
18466 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
18467 a single record has been written.
18468 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
18469 retries use the same buffer location.
18470 (But all of the contents must be
18475 * Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
18478 * Fix problems with no-hmac etc.
18480 *Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>*
18482 * New functions RSA_get_default_method(), RSA_set_method() and
18483 RSA_get_method(). These allows replacement of RSA_METHODs without having
18484 to mess around with the internals of an RSA structure.
18488 * Fix memory leaks in DSA_do_sign and DSA_is_prime.
18489 Also really enable memory leak checks in openssl.c and in some
18492 *Chad C. Mulligan, Bodo Moeller*
18494 * Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
18495 up the length of negative integers. This has now been simplified to just
18496 store the length when it is first determined and use it later, rather
18497 than trying to keep track of where data is copied and updating it to
18499 *Steve Henson, reported by Brien Wheeler <bwheeler@authentica-security.com>*
18501 * Add a new function PKCS7_signatureVerify. This allows the verification
18502 of a PKCS#7 signature but with the signing certificate passed to the
18503 function itself. This contrasts with PKCS7_dataVerify which assumes the
18504 certificate is present in the PKCS#7 structure. This isn't always the
18505 case: certificates can be omitted from a PKCS#7 structure and be
18506 distributed by "out of band" means (such as a certificate database).
18510 * Complete the `PEM_*` macros with DECLARE_PEM versions to replace the
18511 function prototypes in pem.h, also change util/mkdef.pl to add the
18512 necessary function names.
18516 * mk1mf.pl (used by Windows builds) did not properly read the
18517 options set by Configure in the top level Makefile, and Configure
18518 was not even able to write more than one option correctly.
18519 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
18523 * New functions CONF_load_bio() and CONF_load_fp() to allow a config
18524 file to be loaded from a BIO or FILE pointer. The BIO version will
18525 for example allow memory BIOs to contain config info.
18529 * New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
18530 Whoever hopes to achieve shared-library compatibility across versions
18531 must use this, not the compile-time macro.
18532 (Exercise 0.9.4: Which is the minimum library version required by
18534 Note: All this applies only to multi-threaded programs, others don't
18539 * Add missing case to s3_clnt.c state machine -- one of the new SSL tests
18540 through a BIO pair triggered the default case, i.e.
18541 SSLerr(...,SSL_R_UNKNOWN_STATE).
18545 * New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
18546 can use the SSL library even if none of the specific BIOs is
18551 * Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
18552 for the encoded length.
18554 *Jeon KyoungHo <khjeon@sds.samsung.co.kr>*
18556 * Add initial documentation of the X509V3 functions.
18560 * Add a new pair of functions PEM_write_PKCS8PrivateKey() and
18561 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
18562 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
18563 secure PKCS#8 private key format with a high iteration count.
18567 * Fix determination of Perl interpreter: A perl or perl5
18568 *directory* in $PATH was also accepted as the interpreter.
18570 *Ralf S. Engelschall*
18572 * Fix demos/sign/sign.c: well there wasn't anything strictly speaking
18573 wrong with it but it was very old and did things like calling
18574 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
18575 unusual formatting.
18579 * Fix demos/selfsign.c: it used obsolete and deleted functions, changed
18580 to use the new extension code.
18584 * Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
18585 with macros. This should make it easier to change their form, add extra
18586 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
18591 * Add to configuration table a new entry that can specify an alternative
18592 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
18593 according to Mark Crispin <MRC@Panda.COM>.
18597 * DES CBC did not update the IV. Weird.
18601 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
18602 Changing the behaviour of the former might break existing programs --
18603 where IV updating is needed, des_ncbc_encrypt can be used.
18606 * When bntest is run from "make test" it drives bc to check its
18607 calculations, as well as internally checking them. If an internal check
18608 fails, it needs to cause bc to give a non-zero result or make test carries
18609 on without noticing the failure. Fixed.
18613 * DES library cleanups.
18617 * Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
18618 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
18619 ciphers. NOTE: although the key derivation function has been verified
18620 against some published test vectors it has not been extensively tested
18621 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
18626 * Instead of "mkdir -p", which is not fully portable, use new
18627 Perl script "util/mkdir-p.pl".
18631 * Rewrite the way password based encryption (PBE) is handled. It used to
18632 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
18633 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
18634 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
18635 the 'parameter' field of the AlgorithmIdentifier is passed to the
18636 underlying key generation function so it must do its own ASN1 parsing.
18637 This has also changed the EVP_PBE_CipherInit() function which now has a
18638 'parameter' argument instead of literal salt and iteration count values
18639 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
18643 * Support for PKCS#5 v1.5 compatible password based encryption algorithms
18644 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
18645 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
18646 KEY" because this clashed with PKCS#8 unencrypted string. Since this
18647 value was just used as a "magic string" and not used directly its
18648 value doesn't matter.
18652 * Introduce some semblance of const correctness to BN. Shame C doesn't
18657 * "linux-sparc64" configuration (ultrapenguin).
18659 *Ray Miller <ray.miller@oucs.ox.ac.uk>*
18660 "linux-sparc" configuration.
18662 *Christian Forster <fo@hawo.stw.uni-erlangen.de>*
18664 * config now generates no-xxx options for missing ciphers.
18668 * Support the EBCDIC character set (work in progress).
18669 File ebcdic.c not yet included because it has a different license.
18671 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
18673 * Support BS2000/OSD-POSIX.
18675 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
18677 * Make callbacks for key generation use `void *` instead of `char *`.
18681 * Make S/MIME samples compile (not yet tested).
18685 * Additional typesafe stacks.
18689 * New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
18693 ### Changes between 0.9.3 and 0.9.3a [29 May 1999]
18695 * New configuration variant "sco5-gcc".
18697 * Updated some demos.
18699 *Sean O Riordain, Wade Scholine*
18701 * Add missing BIO_free at exit of pkcs12 application.
18705 * Fix memory leak in conf.c.
18709 * Updates for Win32 to assembler version of MD5.
18713 * Set #! path to perl in `apps/der_chop` to where we found it
18714 instead of using a fixed path.
18718 * SHA library changes for irix64-mips4-cc.
18722 * Improvements for VMS support.
18726 ### Changes between 0.9.2b and 0.9.3 [24 May 1999]
18728 * Bignum library bug fix. IRIX 6 passes "make test" now!
18729 This also avoids the problems with SC4.2 and unpatched SC5.
18731 *Andy Polyakov <appro@fy.chalmers.se>*
18733 * New functions sk_num, sk_value and sk_set to replace the previous macros.
18734 These are required because of the typesafe stack would otherwise break
18735 existing code. If old code used a structure member which used to be STACK
18736 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
18737 sk_num or sk_value it would produce an error because the num, data members
18738 are not present in STACK_OF. Now it just produces a warning. sk_set
18739 replaces the old method of assigning a value to sk_value
18740 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
18741 that does this will no longer work (and should use sk_set instead) but
18742 this could be regarded as a "questionable" behaviour anyway.
18746 * Fix most of the other PKCS#7 bugs. The "experimental" code can now
18747 correctly handle encrypted S/MIME data.
18751 * Change type of various DES function arguments from des_cblock
18752 (which means, in function argument declarations, pointer to char)
18753 to des_cblock * (meaning pointer to array with 8 char elements),
18754 which allows the compiler to do more typechecking; it was like
18755 that back in SSLeay, but with lots of ugly casts.
18757 Introduce new type const_des_cblock.
18761 * Reorganise the PKCS#7 library and get rid of some of the more obvious
18762 problems: find RecipientInfo structure that matches recipient certificate
18763 and initialise the ASN1 structures properly based on passed cipher.
18767 * Belatedly make the BN tests actually check the results.
18771 * Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
18772 to and from BNs: it was completely broken. New compilation option
18773 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
18774 key elements as negative integers.
18778 * Reorganize and speed up MD5.
18780 *Andy Polyakov <appro@fy.chalmers.se>*
18784 *Richard Levitte <richard@levitte.org>*
18786 * New option -out to asn1parse to allow the parsed structure to be
18787 output to a file. This is most useful when combined with the -strparse
18788 option to examine the output of things like OCTET STRINGS.
18792 * Make SSL library a little more fool-proof by not requiring any longer
18793 that `SSL_set_{accept,connect}_state` be called before
18794 `SSL_{accept,connect}` may be used (`SSL_set_..._state` is omitted
18795 in many applications because usually everything *appeared* to work as
18796 intended anyway -- now it really works as intended).
18800 * Move openssl.cnf out of lib/.
18804 * Fix various things to let OpenSSL even pass "egcc -pipe -O2 -Wall
18805 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
18806 -Wmissing-declarations -Wnested-externs -Winline" with EGCS 1.1.2+
18808 *Ralf S. Engelschall*
18810 * Various fixes to the EVP and PKCS#7 code. It may now be able to
18811 handle PKCS#7 enveloped data properly.
18813 *Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve*
18815 * Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
18816 copying pointers. The cert_st handling is changed by this in
18817 various ways (and thus what used to be known as ctx->default_cert
18818 is now called ctx->cert, since we don't resort to `s->ctx->[default_]cert`
18819 any longer when s->cert does not give us what we need).
18820 ssl_cert_instantiate becomes obsolete by this change.
18821 As soon as we've got the new code right (possibly it already is?),
18822 we have solved a couple of bugs of the earlier code where s->cert
18823 was used as if it could not have been shared with other SSL structures.
18825 Note that using the SSL API in certain dirty ways now will result
18826 in different behaviour than observed with earlier library versions:
18827 Changing settings for an `SSL_CTX *ctx` after having done s = SSL_new(ctx)
18828 does not influence s as it used to.
18830 In order to clean up things more thoroughly, inside SSL_SESSION
18831 we don't use CERT any longer, but a new structure SESS_CERT
18832 that holds per-session data (if available); currently, this is
18833 the peer's certificate chain and, for clients, the server's certificate
18834 and temporary key. CERT holds only those values that can have
18835 meaningful defaults in an SSL_CTX.
18839 * New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
18840 from the internal representation. Various PKCS#7 fixes: remove some
18841 evil casts and set the enc_dig_alg field properly based on the signing
18846 * Allow PKCS#12 password to be set from the command line or the
18847 environment. Let 'ca' get its config file name from the environment
18848 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
18853 * Allow certificate policies extension to use an IA5STRING for the
18854 organization field. This is contrary to the PKIX definition but
18855 VeriSign uses it and IE5 only recognises this form. Document 'x509'
18860 * Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
18861 without disallowing inline assembler and the like for non-pedantic builds.
18865 * Support Borland C++ builder.
18867 *Janez Jere <jj@void.si>, modified by Ulf Möller*
18873 * SHA-1 cleanups and performance enhancements.
18875 *Andy Polyakov <appro@fy.chalmers.se>*
18877 * Sparc v8plus assembler for the bignum library.
18879 *Andy Polyakov <appro@fy.chalmers.se>*
18881 * Accept any -xxx and +xxx compiler options in Configure.
18885 * Update HPUX configuration.
18889 * Add missing `sk_<type>_unshift()` function to safestack.h
18891 *Ralf S. Engelschall*
18893 * New function SSL_CTX_use_certificate_chain_file that sets the
18894 "extra_cert"s in addition to the certificate. (This makes sense
18895 only for "PEM" format files, as chains as a whole are not
18900 * Support verify_depth from the SSL API.
18901 x509_vfy.c had what can be considered an off-by-one-error:
18902 Its depth (which was not part of the external interface)
18903 was actually counting the number of certificates in a chain;
18904 now it really counts the depth.
18908 * Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
18909 instead of X509err, which often resulted in confusing error
18910 messages since the error codes are not globally unique
18911 (e.g. an alleged error in ssl3_accept when a certificate
18912 didn't match the private key).
18914 * New function SSL_CTX_set_session_id_context that allows to set a default
18915 value (so that you don't need SSL_set_session_id_context for each
18916 connection using the SSL_CTX).
18920 * OAEP decoding bug fix.
18924 * Support INSTALL_PREFIX for package builders, as proposed by
18929 * New Configure options "threads" and "no-threads". For systems
18930 where the proper compiler options are known (currently Solaris
18931 and Linux), "threads" is the default.
18935 * New script util/mklink.pl as a faster substitute for util/mklink.sh.
18939 * Install various scripts to $(OPENSSLDIR)/misc, not to
18940 $(INSTALLTOP)/bin -- they shouldn't clutter directories
18941 such as /usr/local/bin.
18945 * "make linux-shared" to build shared libraries.
18947 *Niels Poppe <niels@netbox.org>*
18949 * New Configure option `no-<cipher>` (rsa, idea, rc5, ...).
18953 * Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
18954 extension adding in x509 utility.
18958 * Remove NOPROTO sections and error code comments.
18962 * Partial rewrite of the DEF file generator to now parse the ANSI
18967 * New Configure options --prefix=DIR and --openssldir=DIR.
18971 * Complete rewrite of the error code script(s). It is all now handled
18972 by one script at the top level which handles error code gathering,
18973 header rewriting and C source file generation. It should be much better
18974 than the old method: it now uses a modified version of Ulf's parser to
18975 read the ANSI prototypes in all header files (thus the old K&R definitions
18976 aren't needed for error creation any more) and do a better job of
18977 translating function codes into names. The old 'ASN1 error code embedded
18978 in a comment' is no longer necessary and it doesn't use .err files which
18979 have now been deleted. Also the error code call doesn't have to appear all
18980 on one line (which resulted in some large lines...).
18984 * Change #include filenames from `<foo.h>` to `<openssl/foo.h>`.
18988 * Change behaviour of ssl2_read when facing length-0 packets: Don't return
18989 0 (which usually indicates a closed connection), but continue reading.
18993 * Fix some race conditions.
18997 * Add support for CRL distribution points extension. Add Certificate
18998 Policies and CRL distribution points documentation.
19002 * Move the autogenerated header file parts to crypto/opensslconf.h.
19006 * Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
19007 8 of keying material. Merlin has also confirmed interop with this fix
19008 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
19010 *Merlin Hughes <merlin@baltimore.ie>*
19012 * Fix lots of warnings.
19014 *Richard Levitte <levitte@stacken.kth.se>*
19016 * In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
19017 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
19019 *Richard Levitte <levitte@stacken.kth.se>*
19021 * Fix problems with sizeof(long) == 8.
19023 *Andy Polyakov <appro@fy.chalmers.se>*
19025 * Change functions to ANSI C.
19029 * Fix typos in error codes.
19031 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller*
19033 * Remove defunct assembler files from Configure.
19037 * SPARC v8 assembler BIGNUM implementation.
19039 *Andy Polyakov <appro@fy.chalmers.se>*
19041 * Support for Certificate Policies extension: both print and set.
19042 Various additions to support the r2i method this uses.
19046 * A lot of constification, and fix a bug in X509_NAME_oneline() that could
19047 return a const string when you are expecting an allocated buffer.
19051 * Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
19052 types DirectoryString and DisplayText.
19056 * Add code to allow r2i extensions to access the configuration database,
19057 add an LHASH database driver and add several ctx helper functions.
19061 * Fix an evil bug in bn_expand2() which caused various BN functions to
19062 fail when they extended the size of a BIGNUM.
19066 * Various utility functions to handle SXNet extension. Modify mkdef.pl to
19067 support typesafe stack.
19071 * Fix typo in SSL_[gs]et_options().
19073 *Nils Frostberg <nils@medcom.se>*
19075 * Delete various functions and files that belonged to the (now obsolete)
19076 old X509V3 handling code.
19080 * New Configure option "rsaref".
19084 * Don't auto-generate pem.h.
19088 * Introduce type-safe ASN.1 SETs.
19092 * Convert various additional casted stacks to type-safe STACK_OF() variants.
19094 *Ben Laurie, Ralf S. Engelschall, Steve Henson*
19096 * Introduce type-safe STACKs. This will almost certainly break lots of code
19097 that links with OpenSSL (well at least cause lots of warnings), but fear
19098 not: the conversion is trivial, and it eliminates loads of evil casts. A
19099 few STACKed things have been converted already. Feel free to convert more.
19100 In the fullness of time, I'll do away with the STACK type altogether.
19104 * Add `openssl ca -revoke <certfile>` facility which revokes a certificate
19105 specified in `<certfile>` by updating the entry in the index.txt file.
19106 This way one no longer has to edit the index.txt file manually for
19107 revoking a certificate. The -revoke option does the gory details now.
19109 *Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall*
19111 * Fix `openssl crl -noout -text` combination where `-noout` killed the
19112 `-text` option at all and this way the `-noout -text` combination was
19113 inconsistent in `openssl crl` with the friends in `openssl x509|rsa|dsa`.
19115 *Ralf S. Engelschall*
19117 * Make sure a corresponding plain text error message exists for the
19118 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
19119 verify callback function determined that a certificate was revoked.
19121 *Ralf S. Engelschall*
19123 * Bugfix: In test/testenc, don't test `openssl <cipher>` for
19124 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
19125 all available ciphers including rc5, which was forgotten until now.
19126 In order to let the testing shell script know which algorithms
19127 are available, a new (up to now undocumented) command
19128 `openssl list-cipher-commands` is used.
19132 * Bugfix: s_client occasionally would sleep in select() when
19133 it should have checked SSL_pending() first.
19137 * New functions DSA_do_sign and DSA_do_verify to provide access to
19138 the raw DSA values prior to ASN.1 encoding.
19142 * Tweaks to Configure
19144 *Niels Poppe <niels@netbox.org>*
19146 * Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
19151 * New variables $(RANLIB) and $(PERL) in the Makefiles.
19155 * New config option to avoid instructions that are illegal on the 80386.
19156 The default code is faster, but requires at least a 486.
19160 * Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
19161 SSL2_SERVER_VERSION (not used at all) macros, which are now the
19162 same as SSL2_VERSION anyway.
19166 * New "-showcerts" option for s_client.
19170 * Still more PKCS#12 integration. Add pkcs12 application to openssl
19171 application. Various cleanups and fixes.
19175 * More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
19176 modify error routines to work internally. Add error codes and PBE init
19177 to library startup routines.
19181 * Further PKCS#12 integration. Added password based encryption, PKCS#8 and
19182 packing functions to asn1 and evp. Changed function names and error
19183 codes along the way.
19187 * PKCS12 integration: and so it begins... First of several patches to
19188 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
19189 objects to objects.h
19193 * Add a new 'indent' option to some X509V3 extension code. Initial ASN1
19194 and display support for Thawte strong extranet extension.
19198 * Add LinuxPPC support.
19200 *Jeff Dubrule <igor@pobox.org>*
19202 * Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
19203 bn_div_words in alpha.s.
19205 *Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie*
19207 * Make sure the RSA OAEP test is skipped under -DRSAref because
19208 OAEP isn't supported when OpenSSL is built with RSAref.
19210 *Ulf Moeller <ulf@fitug.de>*
19212 * Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
19213 so they no longer are missing under -DNOPROTO.
19215 *Soren S. Jorvang <soren@t.dk>*
19217 ### Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
19219 * Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
19220 doesn't work when the session is reused. Coming soon!
19224 * Fix a security hole, that allows sessions to be reused in the wrong
19225 context thus bypassing client cert protection! All software that uses
19226 client certs and session caches in multiple contexts NEEDS PATCHING to
19227 allow session reuse! A fuller solution is in the works.
19229 *Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)*
19231 * Some more source tree cleanups (removed obsolete files
19232 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
19233 permission on "config" script to be executable) and a fix for the INSTALL
19236 *Ulf Moeller <ulf@fitug.de>*
19238 * Remove some legacy and erroneous uses of malloc, free instead of
19241 *Lennart Bang <lob@netstream.se>, with minor changes by Steve*
19243 * Make rsa_oaep_test return non-zero on error.
19245 *Ulf Moeller <ulf@fitug.de>*
19247 * Add support for native Solaris shared libraries. Configure
19248 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
19249 if someone would make that last step automatic.
19251 *Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>*
19253 * ctx_size was not built with the right compiler during "make links". Fixed.
19257 * Change the meaning of 'ALL' in the cipher list. It now means "everything
19258 except NULL ciphers". This means the default cipher list will no longer
19259 enable NULL ciphers. They need to be specifically enabled e.g. with
19260 the string "DEFAULT:eNULL".
19264 * Fix to RSA private encryption routines: if p < q then it would
19265 occasionally produce an invalid result. This will only happen with
19266 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
19270 * Be less restrictive and allow also `perl util/perlpath.pl
19271 /path/to/bin/perl` in addition to `perl util/perlpath.pl /path/to/bin`,
19272 because this way one can also use an interpreter named `perl5` (which is
19273 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
19274 installed as `perl`).
19276 *Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
19278 * Let util/clean-depend.pl work also with older Perl 5.00x versions.
19280 *Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
19282 * Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
19283 advapi32.lib to Win32 build and change the pem test comparison
19284 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
19285 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
19286 and crypto/des/ede_cbcm_enc.c.
19290 * DES quad checksum was broken on big-endian architectures. Fixed.
19294 * Comment out two functions in bio.h that aren't implemented. Fix up the
19295 Win32 test batch file so it (might) work again. The Win32 test batch file
19296 is horrible: I feel ill....
19300 * Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
19301 in e_os.h. Audit of header files to check ANSI and non ANSI
19302 sections: 10 functions were absent from non ANSI section and not exported
19303 from Windows DLLs. Fixed up libeay.num for new functions.
19307 * Make `openssl version` output lines consistent.
19309 *Ralf S. Engelschall*
19311 * Fix Win32 symbol export lists for BIO functions: Added
19312 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
19313 to ms/libeay{16,32}.def.
19315 *Ralf S. Engelschall*
19317 * Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
19318 fine under Unix and passes some trivial tests I've now added. But the
19319 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
19320 added to make sure no one expects that this stuff really works in the
19321 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
19322 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
19325 *Ralf S. Engelschall*
19327 * Fix the generation of two part addresses in perl.
19329 *Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie*
19331 * Add config entry for Linux on MIPS.
19333 *John Tobey <jtobey@channel1.com>*
19335 * Make links whenever Configure is run, unless we are on Windoze.
19339 * Permit extensions to be added to CRLs using crl_section in openssl.cnf.
19340 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
19345 * Add a useful kludge to allow package maintainers to specify compiler and
19346 other platforms details on the command line without having to patch the
19347 Configure script every time: One now can use
19348 `perl Configure <id>:<details>`,
19349 i.e. platform ids are allowed to have details appended
19350 to them (separated by colons). This is treated as there would be a static
19351 pre-configured entry in Configure's %table under key `<id>` with value
19352 `<details>` and `perl Configure <id>` is called. So, when you want to
19353 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
19354 assembler stuff you can use `perl Configure "FreeBSD-elf:pgcc:-O6:::"`
19355 now, which overrides the FreeBSD-elf entry on-the-fly.
19357 *Ralf S. Engelschall*
19359 * Disable new TLS1 ciphersuites by default: they aren't official yet.
19363 * Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
19364 on the `perl Configure ...` command line. This way one can compile
19365 OpenSSL libraries with Position Independent Code (PIC) which is needed
19366 for linking it into DSOs.
19368 *Ralf S. Engelschall*
19370 * Remarkably, export ciphers were totally broken and no-one had noticed!
19375 * Cleaned up the LICENSE document: The official contact for any license
19376 questions now is the OpenSSL core team under openssl-core@openssl.org.
19377 And add a paragraph about the dual-license situation to make sure people
19378 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
19379 to the OpenSSL toolkit.
19381 *Ralf S. Engelschall*
19383 * General source tree makefile cleanups: Made `making xxx in yyy...`
19384 display consistent in the source tree and replaced `/bin/rm` by `rm`.
19385 Additionally cleaned up the `make links` target: Remove unnecessary
19386 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
19387 to speed processing and no longer clutter the display with confusing
19388 stuff. Instead only the actually done links are displayed.
19390 *Ralf S. Engelschall*
19392 * Permit null encryption ciphersuites, used for authentication only. It used
19393 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
19394 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
19399 * Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
19400 signed attributes when verifying signatures (this would break them),
19401 the detached data encoding was wrong and public keys obtained using
19402 X509_get_pubkey() weren't freed.
19406 * Add text documentation for the BUFFER functions. Also added a work around
19407 to a Win95 console bug. This was triggered by the password read stuff: the
19408 last character typed gets carried over to the next fread(). If you were
19409 generating a new cert request using 'req' for example then the last
19410 character of the passphrase would be CR which would then enter the first
19415 * Added the new 'Includes OpenSSL Cryptography Software' button as
19416 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
19417 button and can be used by applications based on OpenSSL to show the
19418 relationship to the OpenSSL project.
19420 *Ralf S. Engelschall*
19422 * Remove confusing variables in function signatures in files
19423 ssl/ssl_lib.c and ssl/ssl.h.
19425 *Lennart Bong <lob@kulthea.stacken.kth.se>*
19427 * Don't install bss_file.c under PREFIX/include/
19429 *Lennart Bong <lob@kulthea.stacken.kth.se>*
19431 * Get the Win32 compile working again. Modify mkdef.pl so it can handle
19432 functions that return function pointers and has support for NT specific
19433 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
19434 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
19435 unsigned to signed types: this was killing the Win32 compile.
19439 * Add new certificate file to stack functions,
19440 SSL_add_dir_cert_subjects_to_stack() and
19441 SSL_add_file_cert_subjects_to_stack(). These largely supplant
19442 SSL_load_client_CA_file(), and can be used to add multiple certs easily
19443 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
19444 This means that Apache-SSL and similar packages don't have to mess around
19445 to add as many CAs as they want to the preferred list.
19449 * Experiment with doxygen documentation. Currently only partially applied to
19451 See <http://www.stack.nl/~dimitri/doxygen/index.html>, and run doxygen with
19452 openssl.doxy as the configuration file.
19456 * Get rid of remaining C++-style comments which strict C compilers hate.
19458 *Ralf S. Engelschall, pointed out by Carlos Amengual*
19460 * Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
19461 compiled in by default: it has problems with large keys.
19465 * Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
19466 DH private keys and/or callback functions which directly correspond to
19467 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
19468 is needed for applications which have to configure certificates on a
19469 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
19471 For the RSA certificate situation is makes no difference, but
19472 for the DSA certificate situation this fixes the "no shared cipher"
19473 problem where the OpenSSL cipher selection procedure failed because the
19474 temporary keys were not overtaken from the context and the API provided
19475 no way to reconfigure them.
19476 The new functions now let applications reconfigure the stuff and they
19477 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
19478 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
19479 non-public-API function ssl_cert_instantiate() is used as a helper
19480 function and also to reduce code redundancy inside ssl_rsa.c.
19482 *Ralf S. Engelschall*
19484 * Move s_server -dcert and -dkey options out of the undocumented feature
19485 area because they are useful for the DSA situation and should be
19486 recognized by the users.
19488 *Ralf S. Engelschall*
19490 * Fix the cipher decision scheme for export ciphers: the export bits are
19491 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
19492 SSL_EXP_MASK. So, the original variable has to be used instead of the
19493 already masked variable.
19495 *Richard Levitte <levitte@stacken.kth.se>*
19497 * Fix `port` variable from `int` to `unsigned int` in crypto/bio/b_sock.c
19499 *Richard Levitte <levitte@stacken.kth.se>*
19501 * Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
19502 from `int` to `unsigned int` because it is a length and initialized by
19503 EVP_DigestFinal() which expects an `unsigned int *`.
19505 *Richard Levitte <levitte@stacken.kth.se>*
19507 * Don't hard-code path to Perl interpreter on shebang line of Configure
19508 script. Instead use the usual Shell->Perl transition trick.
19510 *Ralf S. Engelschall*
19512 * Make `openssl x509 -noout -modulus`' functional also for DSA certificates
19513 (in addition to RSA certificates) to match the behaviour of `openssl dsa
19514 -noout -modulus` as it's already the case for `openssl rsa -noout
19515 -modulus`. For RSA the -modulus is the real "modulus" while for DSA
19516 currently the public key is printed (a decision which was already done by
19517 `openssl dsa -modulus` in the past) which serves a similar purpose.
19518 Additionally the NO_RSA no longer completely removes the whole -modulus
19519 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
19522 *Ralf S. Engelschall*
19524 * Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
19525 BIO. See the source (crypto/evp/bio_ok.c) for more info.
19527 *Arne Ansper <arne@ats.cyber.ee>*
19529 * Dump the old yucky req code that tried (and failed) to allow raw OIDs
19530 to be added. Now both 'req' and 'ca' can use new objects defined in the
19535 * Add cool BIO that does syslog (or event log on NT).
19537 *Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie*
19539 * Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
19540 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
19541 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
19542 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
19546 * Add preliminary config info for new extension code.
19550 * Make RSA_NO_PADDING really use no padding.
19552 *Ulf Moeller <ulf@fitug.de>*
19554 * Generate errors when private/public key check is done.
19558 * Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
19559 for some CRL extensions and new objects added.
19563 * Really fix the ASN1 IMPLICIT bug this time... Partial support for private
19564 key usage extension and fuller support for authority key id.
19568 * Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
19569 padding method for RSA, which is recommended for new applications in PKCS
19570 #1 v2.0 (RFC 2437, October 1998).
19571 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
19572 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
19573 against Bleichbacher's attack on RSA.
19574 *Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
19577 * Updates to the new SSL compression code
19579 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
19581 * Fix so that the version number in the master secret, when passed
19582 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
19583 (because the server will not accept higher), that the version number
19584 is 0x03,0x01, not 0x03,0x00
19586 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
19588 * Run extensive memory leak checks on SSL commands. Fixed *lots* of memory
19589 leaks in `ssl/` relating to new `X509_get_pubkey()` behaviour. Also fixes
19590 in `apps/` and an unrelated leak in `crypto/dsa/dsa_vrf.c`.
19594 * Support for RAW extensions where an arbitrary extension can be
19595 created by including its DER encoding. See `apps/openssl.cnf` for
19600 * Make sure latest Perl versions don't interpret some generated C array
19601 code as Perl array code in the crypto/err/err_genc.pl script.
19603 *Lars Weber <3weber@informatik.uni-hamburg.de>*
19605 * Modify ms/do_ms.bat to not generate assembly language makefiles since
19606 not many people have the assembler. Various Win32 compilation fixes and
19607 update to the INSTALL.W32 file with (hopefully) more accurate Win32
19608 build instructions.
19612 * Modify configure script 'Configure' to automatically create crypto/date.h
19613 file under Win32 and also build pem.h from pem.org. New script
19614 util/mkfiles.pl to create the MINFO file on environments that can't do a
19615 'make files': perl util/mkfiles.pl >MINFO should work.
19619 * Major rework of DES function declarations, in the pursuit of correctness
19620 and purity. As a result, many evil casts evaporated, and some weirdness,
19621 too. You may find this causes warnings in your code. Zapping your evil
19622 casts will probably fix them. Mostly.
19626 * Fix for a typo in asn1.h. Bug fix to object creation script
19627 obj_dat.pl. It considered a zero in an object definition to mean
19628 "end of object": none of the objects in objects.h have any zeros
19629 so it wasn't spotted.
19631 *Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>*
19633 * Add support for Triple DES Cipher Block Chaining with Output Feedback
19634 Masking (CBCM). In the absence of test vectors, the best I have been able
19635 to do is check that the decrypt undoes the encrypt, so far. Send me test
19636 vectors if you have them.
19640 * Correct calculation of key length for export ciphers (too much space was
19641 allocated for null ciphers). This has not been tested!
19645 * Modifications to the mkdef.pl for Win32 DEF file creation. The usage
19646 message is now correct (it understands "crypto" and "ssl" on its
19647 command line). There is also now an "update" option. This will update
19648 the util/ssleay.num and util/libeay.num files with any new functions.
19650 perl util/mkdef.pl crypto ssl update
19651 it will update them.
19655 * Overhauled the Perl interface:
19656 - ported BN stuff to OpenSSL's different BN library
19657 - made the perl/ source tree CVS-aware
19658 - renamed the package from SSLeay to OpenSSL (the files still contain
19659 their history because I've copied them in the repository)
19660 - removed obsolete files (the test scripts will be replaced
19661 by better Test::Harness variants in the future)
19663 *Ralf S. Engelschall*
19665 * First cut for a very conservative source tree cleanup:
19666 1. merge various obsolete readme texts into doc/ssleay.txt
19667 where we collect the old documents and readme texts.
19668 2. remove the first part of files where I'm already sure that we no
19669 longer need them because of three reasons: either they are just temporary
19670 files which were left by Eric or they are preserved original files where
19671 I've verified that the diff is also available in the CVS via "cvs diff
19672 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
19673 the crypto/md/ stuff).
19675 *Ralf S. Engelschall*
19677 * More extension code. Incomplete support for subject and issuer alt
19678 name, issuer and authority key id. Change the i2v function parameters
19679 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
19680 what that's for :-) Fix to ASN1 macro which messed up
19681 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
19685 * Preliminary support for ENUMERATED type. This is largely copied from the
19690 * Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
19692 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
19694 * Make sure `make rehash` target really finds the `openssl` program.
19696 *Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
19698 * Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
19699 like to hear about it if this slows down other processors.
19703 * Add CygWin32 platform information to Configure script.
19705 *Alan Batie <batie@aahz.jf.intel.com>*
19707 * Fixed ms/32all.bat script: `no_asm` -> `no-asm`
19709 *Rainer W. Gerling <gerling@mpg-gv.mpg.de>*
19711 * New program nseq to manipulate netscape certificate sequences
19715 * Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
19720 * Fixes to BN code. Previously the default was to define BN_RECURSION
19721 but the BN code had some problems that would cause failures when
19722 doing certificate verification and some other functions.
19724 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
19726 * Add ASN1 and PEM code to support netscape certificate sequences.
19730 * Add ASN1 and PEM code to support netscape certificate sequences.
19734 * Add several PKIX and private extended key usage OIDs.
19738 * Modify the 'ca' program to handle the new extension code. Modify
19739 openssl.cnf for new extension format, add comments.
19743 * More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
19744 and add a sample to openssl.cnf so req -x509 now adds appropriate
19749 * Continued X509 V3 changes. Add to other makefiles, integrate with the
19750 error code, add initial support to X509_print() and x509 application.
19754 * Takes a deep breath and start adding X509 V3 extension support code. Add
19755 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
19756 stuff is currently isolated and isn't even compiled yet.
19760 * Continuing patches for GeneralizedTime. Fix up certificate and CRL
19761 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
19762 Removed the versions check from X509 routines when loading extensions:
19763 this allows certain broken certificates that don't set the version
19764 properly to be processed.
19768 * Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
19769 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
19770 can still be regenerated with "make depend".
19774 * Spelling mistake in C version of CAST-128.
19776 *Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>*
19778 * Changes to the error generation code. The perl script err-code.pl
19779 now reads in the old error codes and retains the old numbers, only
19780 adding new ones if necessary. It also only changes the .err files if new
19781 codes are added. The makefiles have been modified to only insert errors
19782 when needed (to avoid needlessly modifying header files). This is done
19783 by only inserting errors if the .err file is newer than the auto generated
19784 C file. To rebuild all the error codes from scratch (the old behaviour)
19785 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
19786 or delete all the .err files.
19790 * CAST-128 was incorrectly implemented for short keys. The C version has
19791 been fixed, but is untested. The assembler versions are also fixed, but
19792 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
19793 to regenerate it if needed.
19794 *Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
19795 Hagino <itojun@kame.net>*
19797 * File was opened incorrectly in randfile.c.
19799 *Ulf Möller <ulf@fitug.de>*
19801 * Beginning of support for GeneralizedTime. d2i, i2d, check and print
19802 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
19803 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
19804 al: it's just almost always a UTCTime. Note this patch adds new error
19805 codes so do a "make errors" if there are problems.
19809 * Correct Linux 1 recognition in config.
19811 *Ulf Möller <ulf@fitug.de>*
19813 * Remove pointless MD5 hash when using DSA keys in ca.
19815 *Anonymous <nobody@replay.com>*
19817 * Generate an error if given an empty string as a cert directory. Also
19818 generate an error if handed NULL (previously returned 0 to indicate an
19819 error, but didn't set one).
19821 *Ben Laurie, reported by Anonymous <nobody@replay.com>*
19823 * Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
19827 * Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
19828 parameters. This was causing a warning which killed off the Win32 compile.
19832 * Remove C++ style comments from crypto/bn/bn_local.h.
19834 *Neil Costigan <neil.costigan@celocom.com>*
19836 * The function OBJ_txt2nid was broken. It was supposed to return a nid
19837 based on a text string, looking up short and long names and finally
19838 "dot" format. The "dot" format stuff didn't work. Added new function
19839 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
19840 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
19841 OID is not part of the table.
19845 * Add prototypes to X509 lookup/verify methods, fixing a bug in
19846 X509_LOOKUP_by_alias().
19850 * Sort openssl functions by name.
19854 * Get the `gendsa` command working and add it to the `list` command. Remove
19855 encryption from sample DSA keys (in case anyone is interested the password
19860 * Make *all* `*_free` functions accept a NULL pointer.
19862 *Frans Heymans <fheymans@isaserver.be>*
19864 * If a DH key is generated in s3_srvr.c, don't blow it by trying to use
19867 *Anonymous <nobody@replay.com>*
19869 * s_server should send the CAfile as acceptable CAs, not its own cert.
19871 *Bodo Moeller <3moeller@informatik.uni-hamburg.de>*
19873 * Don't blow it for numeric `-newkey` arguments to `apps/req`.
19875 *Bodo Moeller <3moeller@informatik.uni-hamburg.de>*
19877 * Temp key "for export" tests were wrong in s3_srvr.c.
19879 *Anonymous <nobody@replay.com>*
19881 * Add prototype for temp key callback functions
19882 SSL_CTX_set_tmp_{rsa,dh}_callback().
19886 * Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
19887 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
19891 * X509_name_add_entry() freed the wrong thing after an error.
19893 *Arne Ansper <arne@ats.cyber.ee>*
19895 * rsa_eay.c would attempt to free a NULL context.
19897 *Arne Ansper <arne@ats.cyber.ee>*
19899 * BIO_s_socket() had a broken should_retry() on Windoze.
19901 *Arne Ansper <arne@ats.cyber.ee>*
19903 * BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
19905 *Arne Ansper <arne@ats.cyber.ee>*
19907 * Make sure the already existing X509_STORE->depth variable is initialized
19908 in X509_STORE_new(), but document the fact that this variable is still
19909 unused in the certificate verification process.
19911 *Ralf S. Engelschall*
19913 * Fix the various library and `apps/` files to free up pkeys obtained from
19914 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
19918 * Fix reference counting in X509_PUBKEY_get(). This makes
19919 demos/maurice/example2.c work, amongst others, probably.
19921 *Steve Henson and Ben Laurie*
19923 * First cut of a cleanup for `apps/`. First the `ssleay` program is now named
19924 `openssl` and second, the shortcut symlinks for the `openssl <command>`
19925 are no longer created. This way we have a single and consistent command
19926 line interface `openssl <command>`, similar to `cvs <command>`.
19928 *Ralf S. Engelschall, Paul Sutton and Ben Laurie*
19930 * ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
19931 BIT STRING wrapper always have zero unused bits.
19935 * Add CA.pl, perl version of CA.sh, add extended key usage OID.
19939 * Make the top-level INSTALL documentation easier to understand.
19943 * Makefiles updated to exit if an error occurs in a sub-directory
19944 make (including if user presses ^C) [Paul Sutton]
19946 * Make Montgomery context stuff explicit in RSA data structure.
19950 * Fix build order of pem and err to allow for generated pem.h.
19954 * Fix renumbering bug in X509_NAME_delete_entry().
19958 * Enhanced the err-ins.pl script so it makes the error library number
19959 global and can add a library name. This is needed for external ASN1 and
19960 other error libraries.
19964 * Fixed sk_insert which never worked properly.
19968 * Fix ASN1 macros so they can handle indefinite length constructed
19969 EXPLICIT tags. Some non standard certificates use these: they can now
19974 * Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
19975 into a single doc/ssleay.txt bundle. This way the information is still
19976 preserved but no longer messes up this directory. Now it's new room for
19977 the new set of documentation files.
19979 *Ralf S. Engelschall*
19981 * SETs were incorrectly DER encoded. This was a major pain, because they
19982 shared code with SEQUENCEs, which aren't coded the same. This means that
19983 almost everything to do with SETs or SEQUENCEs has either changed name or
19984 number of arguments.
19986 *Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>*
19988 * Fix test data to work with the above.
19992 * Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
19993 was already fixed by Eric for 0.9.1 it seems.
19995 *Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>*
19997 * Autodetect FreeBSD3.
20001 * Fix various bugs in Configure. This affects the following platforms:
20005 unixware-2.0-pentium
20010 * Eliminate generated files from CVS. Reorder tests to regenerate files
20011 before they are needed.
20015 * Generate Makefile.ssl from Makefile.org (to keep CVS happy).
20019 ### Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
20021 * Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
20022 changed SSLeay to OpenSSL in version strings.
20024 *Ralf S. Engelschall*
20026 * Some fixups to the top-level documents.
20030 * Fixed the nasty bug where rsaref.h was not found under compile-time
20031 because the symlink to include/ was missing.
20033 *Ralf S. Engelschall*
20035 * Incorporated the popular no-RSA/DSA-only patches
20036 which allow to compile an RSA-free SSLeay.
20038 *Andrew Cooke / Interrader Ldt., Ralf S. Engelschall*
20040 * Fixed nasty rehash problem under `make -f Makefile.ssl links`
20041 when "ssleay" is still not found.
20043 *Ralf S. Engelschall*
20045 * Added more platforms to Configure: Cray T3E, HPUX 11,
20047 *Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>*
20049 * Updated the README file.
20051 *Ralf S. Engelschall*
20053 * Added various .cvsignore files in the CVS repository subdirs
20054 to make a "cvs update" really silent.
20056 *Ralf S. Engelschall*
20058 * Recompiled the error-definition header files and added
20059 missing symbols to the Win32 linker tables.
20061 *Ralf S. Engelschall*
20063 * Cleaned up the top-level documents;
20064 o new files: CHANGES and LICENSE
20065 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
20066 o merged COPYRIGHT into LICENSE
20067 o removed obsolete TODO file
20068 o renamed MICROSOFT to INSTALL.W32
20070 *Ralf S. Engelschall*
20072 * Removed dummy files from the 0.9.1b source tree:
20073 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
20074 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
20075 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
20076 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
20077 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
20079 *Ralf S. Engelschall*
20081 * Added various platform portability fixes.
20085 * The Genesis of the OpenSSL rpject:
20086 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
20087 Young and Tim J. Hudson created while they were working for C2Net until
20090 *The OpenSSL Project*
20092 ### Changes between 0.9.0b and 0.9.1b [not released]
20094 * Updated a few CA certificates under certs/
20098 * Changed some BIGNUM api stuff.
20102 * Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
20103 DGUX x86, Linux Alpha, etc.
20107 * New COMP library [crypto/comp/] for SSL Record Layer Compression:
20108 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
20113 * Add -strparse option to asn1pars program which parses nested
20116 *Dr Stephen Henson <shenson@bigfoot.com>*
20118 * Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
20122 * DSA fix for "ca" program.
20126 * Added "-genkey" option to "dsaparam" program.
20130 * Added RIPE MD160 (rmd160) message digest.
20134 * Added -a (all) option to "ssleay version" command.
20138 * Added PLATFORM define which is the id given to Configure.
20142 * Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
20146 * Extended the ASN.1 parser routines.
20150 * Extended BIO routines to support REUSEADDR, seek, tell, etc.
20154 * Added a BN_CTX to the BN library.
20158 * Fixed the weak key values in DES library
20162 * Changed API in EVP library for cipher aliases.
20166 * Added support for RC2/64bit cipher.
20170 * Converted the lhash library to the crypto/mem.c functions.
20174 * Added more recognized ASN.1 object ids.
20178 * Added more RSA padding checks for SSL/TLS.
20182 * Added BIO proxy/filter functionality.
20186 * Added extra_certs to SSL_CTX which can be used
20187 send extra CA certificates to the client in the CA cert chain sending
20188 process. It can be configured with SSL_CTX_add_extra_chain_cert().
20192 * Now Fortezza is denied in the authentication phase because
20193 this is key exchange mechanism is not supported by SSLeay at all.
20197 * Additional PKCS1 checks.
20201 * Support the string "TLSv1" for all TLS v1 ciphers.
20205 * Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
20206 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
20210 * Fixed a few memory leaks.
20214 * Fixed various code and comment typos.
20218 * A minor bug in ssl/s3_clnt.c where there would always be 4 0
20219 bytes sent in the client random.
20221 *Edward Bishop <ebishop@spyglass.com>*
20225 [CVE-2023-3817]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-3817
20226 [CVE-2023-3446]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-3446
20227 [CVE-2023-2975]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-2975
20228 [RFC 2578 (STD 58), section 3.5]: https://datatracker.ietf.org/doc/html/rfc2578#section-3.5
20229 [CVE-2023-2650]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-2650
20230 [CVE-2023-1255]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-1255
20231 [CVE-2023-0466]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-0466
20232 [CVE-2023-0465]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-0465
20233 [CVE-2023-0464]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-0464
20234 [CVE-2023-0401]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-0401
20235 [CVE-2023-0286]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-0286
20236 [CVE-2023-0217]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-0217
20237 [CVE-2023-0216]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-0216
20238 [CVE-2023-0215]: https://www.openssl.org/news/vulnerabilities.html#CVE-2023-0215
20239 [CVE-2022-4450]: https://www.openssl.org/news/vulnerabilities.html#CVE-2022-4450
20240 [CVE-2022-4304]: https://www.openssl.org/news/vulnerabilities.html#CVE-2022-4304
20241 [CVE-2022-4203]: https://www.openssl.org/news/vulnerabilities.html#CVE-2022-4203
20242 [CVE-2022-3996]: https://www.openssl.org/news/vulnerabilities.html#CVE-2022-3996
20243 [CVE-2022-2274]: https://www.openssl.org/news/vulnerabilities.html#CVE-2022-2274
20244 [CVE-2022-2097]: https://www.openssl.org/news/vulnerabilities.html#CVE-2022-2097
20245 [CVE-2020-1971]: https://www.openssl.org/news/vulnerabilities.html#CVE-2020-1971
20246 [CVE-2020-1967]: https://www.openssl.org/news/vulnerabilities.html#CVE-2020-1967
20247 [CVE-2019-1563]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1563
20248 [CVE-2019-1559]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1559
20249 [CVE-2019-1552]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1552
20250 [CVE-2019-1551]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1551
20251 [CVE-2019-1549]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1549
20252 [CVE-2019-1547]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1547
20253 [CVE-2019-1543]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1543
20254 [CVE-2018-5407]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-5407
20255 [CVE-2018-0739]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0739
20256 [CVE-2018-0737]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0737
20257 [CVE-2018-0735]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0735
20258 [CVE-2018-0734]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0734
20259 [CVE-2018-0733]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0733
20260 [CVE-2018-0732]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0732
20261 [CVE-2017-3738]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3738
20262 [CVE-2017-3737]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3737
20263 [CVE-2017-3736]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3736
20264 [CVE-2017-3735]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3735
20265 [CVE-2017-3733]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3733
20266 [CVE-2017-3732]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3732
20267 [CVE-2017-3731]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3731
20268 [CVE-2017-3730]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3730
20269 [CVE-2016-7055]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7055
20270 [CVE-2016-7054]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7054
20271 [CVE-2016-7053]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7053
20272 [CVE-2016-7052]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7052
20273 [CVE-2016-6309]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6309
20274 [CVE-2016-6308]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6308
20275 [CVE-2016-6307]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6307
20276 [CVE-2016-6306]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6306
20277 [CVE-2016-6305]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6305
20278 [CVE-2016-6304]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6304
20279 [CVE-2016-6303]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6303
20280 [CVE-2016-6302]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6302
20281 [CVE-2016-2183]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2183
20282 [CVE-2016-2182]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2182
20283 [CVE-2016-2181]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2181
20284 [CVE-2016-2180]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2180
20285 [CVE-2016-2179]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2179
20286 [CVE-2016-2178]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2178
20287 [CVE-2016-2177]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2177
20288 [CVE-2016-2176]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2176
20289 [CVE-2016-2109]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2109
20290 [CVE-2016-2107]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2107
20291 [CVE-2016-2106]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2106
20292 [CVE-2016-2105]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2105
20293 [CVE-2016-0800]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0800
20294 [CVE-2016-0799]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0799
20295 [CVE-2016-0798]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0798
20296 [CVE-2016-0797]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0797
20297 [CVE-2016-0705]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0705
20298 [CVE-2016-0702]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0702
20299 [CVE-2016-0701]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0701
20300 [CVE-2015-3197]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3197
20301 [CVE-2015-3196]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3196
20302 [CVE-2015-3195]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3195
20303 [CVE-2015-3194]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3194
20304 [CVE-2015-3193]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3193
20305 [CVE-2015-1793]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1793
20306 [CVE-2015-1792]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1792
20307 [CVE-2015-1791]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1791
20308 [CVE-2015-1790]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1790
20309 [CVE-2015-1789]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1789
20310 [CVE-2015-1788]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1788
20311 [CVE-2015-1787]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1787
20312 [CVE-2015-0293]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0293
20313 [CVE-2015-0291]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0291
20314 [CVE-2015-0290]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0290
20315 [CVE-2015-0289]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0289
20316 [CVE-2015-0288]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0288
20317 [CVE-2015-0287]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0287
20318 [CVE-2015-0286]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0286
20319 [CVE-2015-0285]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0285
20320 [CVE-2015-0209]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0209
20321 [CVE-2015-0208]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0208
20322 [CVE-2015-0207]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0207
20323 [CVE-2015-0206]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0206
20324 [CVE-2015-0205]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0205
20325 [CVE-2015-0204]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0204
20326 [CVE-2014-8275]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-8275
20327 [CVE-2014-5139]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-5139
20328 [CVE-2014-3572]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3572
20329 [CVE-2014-3571]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3571
20330 [CVE-2014-3570]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3570
20331 [CVE-2014-3569]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3569
20332 [CVE-2014-3568]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3568
20333 [CVE-2014-3567]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3567
20334 [CVE-2014-3566]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3566
20335 [CVE-2014-3513]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3513
20336 [CVE-2014-3512]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3512
20337 [CVE-2014-3511]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3511
20338 [CVE-2014-3510]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3510
20339 [CVE-2014-3509]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3509
20340 [CVE-2014-3508]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3508
20341 [CVE-2014-3507]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3507
20342 [CVE-2014-3506]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3506
20343 [CVE-2014-3505]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3505
20344 [CVE-2014-3470]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3470
20345 [CVE-2014-0224]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0224
20346 [CVE-2014-0221]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0221
20347 [CVE-2014-0195]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0195
20348 [CVE-2014-0160]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0160
20349 [CVE-2014-0076]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0076
20350 [CVE-2013-6450]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-6450
20351 [CVE-2013-4353]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-4353
20352 [CVE-2013-0169]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-0169
20353 [CVE-2013-0166]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-0166
20354 [CVE-2012-2686]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2686
20355 [CVE-2012-2333]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2333
20356 [CVE-2012-2110]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2110
20357 [CVE-2012-0884]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0884
20358 [CVE-2012-0050]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0050
20359 [CVE-2012-0027]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0027
20360 [CVE-2011-4619]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4619
20361 [CVE-2011-4577]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4577
20362 [CVE-2011-4576]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4576
20363 [CVE-2011-4109]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4109
20364 [CVE-2011-4108]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4108
20365 [CVE-2011-3210]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-3210
20366 [CVE-2011-3207]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-3207
20367 [CVE-2011-0014]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-0014
20368 [CVE-2010-4252]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-4252
20369 [CVE-2010-4180]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-4180
20370 [CVE-2010-3864]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-3864
20371 [CVE-2010-1633]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-1633
20372 [CVE-2010-0740]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-0740
20373 [CVE-2010-0433]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-0433
20374 [CVE-2009-4355]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-4355
20375 [CVE-2009-3555]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-3555
20376 [CVE-2009-3245]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-3245
20377 [CVE-2009-1386]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1386
20378 [CVE-2009-1379]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1379
20379 [CVE-2009-1378]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1378
20380 [CVE-2009-1377]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1377
20381 [CVE-2009-0789]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0789
20382 [CVE-2009-0591]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0591
20383 [CVE-2009-0590]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0590
20384 [CVE-2008-5077]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-5077
20385 [CVE-2008-1678]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-1678
20386 [CVE-2008-1672]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-1672
20387 [CVE-2008-0891]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-0891
20388 [CVE-2007-5135]: https://www.openssl.org/news/vulnerabilities.html#CVE-2007-5135
20389 [CVE-2007-4995]: https://www.openssl.org/news/vulnerabilities.html#CVE-2007-4995
20390 [CVE-2006-4343]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-4343
20391 [CVE-2006-4339]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-4339
20392 [CVE-2006-3738]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-3738
20393 [CVE-2006-2940]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-2940
20394 [CVE-2006-2937]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-2937
20395 [CVE-2005-2969]: https://www.openssl.org/news/vulnerabilities.html#CVE-2005-2969
20396 [CVE-2004-0112]: https://www.openssl.org/news/vulnerabilities.html#CVE-2004-0112
20397 [CVE-2004-0079]: https://www.openssl.org/news/vulnerabilities.html#CVE-2004-0079
20398 [CVE-2003-0851]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0851
20399 [CVE-2003-0545]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0545
20400 [CVE-2003-0544]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0544
20401 [CVE-2003-0543]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0543
20402 [CVE-2003-0078]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0078
20403 [CVE-2002-0659]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0659
20404 [CVE-2002-0657]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0657
20405 [CVE-2002-0656]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0656
20406 [CVE-2002-0655]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0655