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 3.1](#openssl-31)
14 - [OpenSSL 3.0](#openssl-30)
15 - [OpenSSL 1.1.1](#openssl-111)
16 - [OpenSSL 1.1.0](#openssl-110)
17 - [OpenSSL 1.0.2](#openssl-102)
18 - [OpenSSL 1.0.1](#openssl-101)
19 - [OpenSSL 1.0.0](#openssl-100)
20 - [OpenSSL 0.9.x](#openssl-09x)
25 ### Changes between 3.0 and 3.1 [xx XXX xxxx]
27 * Major refactor of the libssl record layer
31 * Added a new BIO_s_dgram_mem() to read/write datagrams to memory
35 * Add a mac salt length option for the pkcs12 command.
39 * Add more SRTP protection profiles from RFC8723 and RFC8269.
43 * Extended Kernel TLS (KTLS) to support TLS 1.3 receive offload.
45 *Daiki Ueno, John Baldwin and Dmitry Podgorny*
47 * Add support for TCP Fast Open (RFC7413) to macOS, Linux, and FreeBSD where
48 supported and enabled.
52 * Add ciphersuites based on DHE_PSK (RFC 4279) and ECDHE_PSK (RFC 5489)
53 to the list of ciphersuites providing Perfect Forward Secrecy as
54 required by SECLEVEL >= 3.
56 *Dmitry Belyavskiy, Nicola Tuveri*
58 * Add new SSL APIs to aid in efficiently implementing TLS/SSL fingerprinting.
59 The SSL_CTRL_GET_IANA_GROUPS control code, exposed as the
60 SSL_get0_iana_groups() function-like macro, retrieves the list of
61 supported groups sent by the peer.
62 The function SSL_client_hello_get_extension_order() populates
63 a caller-supplied array with the list of extension types present in the
64 ClientHello, in order of appearance.
68 * Fixed PEM_write_bio_PKCS8PrivateKey() and PEM_write_bio_PKCS8PrivateKey_nid()
69 to make it possible to use empty passphrase strings.
73 * RNDR and RNDRRS support in provider functions to provide
74 random number generation for Arm CPUs (aarch64).
78 * s_client and s_server apps now explicitly say when the TLS version
79 does not include the renegotiation mechanism. This avoids confusion
80 between that scenario versus when the TLS version includes secure
81 renegotiation but the peer lacks support for it.
85 * AES-GCM enabled with AVX512 vAES and vPCLMULQDQ.
87 *Tomasz Kantecki, Andrey Matyukov*
89 * The default SSL/TLS security level has been changed from 1 to 2. RSA,
90 DSA and DH keys of 1024 bits and above and less than 2048 bits and ECC keys
91 of 160 bits and above and less than 224 bits were previously accepted by
92 default but are now no longer allowed. By default TLS compression was
93 already disabled in previous OpenSSL versions. At security level 2 it cannot
98 * The SSL_CTX_set_cipher_list family functions now accept ciphers using their
103 * The PVK key derivation function has been moved from b2i_PVK_bio_ex() into
104 the legacy crypto provider as an EVP_KDF. Applications requiring this KDF
105 will need to load the legacy crypto provider.
109 * The various OBJ_* functions have been made thread safe.
113 * CCM8 cipher suites in TLS have been downgraded to security level zero
114 because they use a short authentication tag which lowers their strength.
118 * Subject or issuer names in X.509 objects are now displayed as UTF-8 strings
123 * Parallel dual-prime 1536/2048-bit modular exponentiation for
124 AVX512_IFMA capable processors.
126 *Sergey Kirillov, Andrey Matyukov (Intel Corp)*
128 * The functions `OPENSSL_LH_stats`, `OPENSSL_LH_node_stats`,
129 `OPENSSL_LH_node_usage_stats`, `OPENSSL_LH_stats_bio`,
130 `OPENSSL_LH_node_stats_bio` and `OPENSSL_LH_node_usage_stats_bio` are now
131 marked deprecated from OpenSSL 3.1 onwards and can be disabled by defining
132 `OPENSSL_NO_DEPRECATED_3_1`.
134 The macro `DEFINE_LHASH_OF` is now deprecated in favour of the macro
135 `DEFINE_LHASH_OF_EX`, which omits the corresponding type-specific function
136 definitions for these functions regardless of whether
137 `OPENSSL_NO_DEPRECATED_3_1` is defined.
139 Users of `DEFINE_LHASH_OF` may start receiving deprecation warnings for these
140 functions regardless of whether they are using them. It is recommended that
141 users transition to the new macro, `DEFINE_LHASH_OF_EX`.
145 * When generating safe-prime DH parameters set the recommended private key
146 length equivalent to minimum key lengths as in RFC 7919.
150 * Fix and extend certificate handling and the apps `x509`, `verify` etc.
151 such as adding a trace facility for debugging certificate chain building.
155 * Various fixes and extensions to the CMP+CRMF implementation and the `cmp` app
156 in particular supporting requests for central key generation, generalized
157 polling, and various types of genm/genp exchanges defined in CMP Updates.
161 * Fixes and extensions to the HTTP client and to the HTTP server in `apps/`
162 like correcting the TLS and proxy support and adding tracing for debugging.
166 * Extended the CMS API for handling `CMS_SignedData` and `CMS_EnvelopedData`.
170 * Fixed and extended `util/check-format.pl` for checking adherence to the
171 coding style <https://www.openssl.org/policies/technical/coding-style.html>.
172 The checks are meanwhile more complete and yield fewer false positives.
179 For OpenSSL 3.0 a [Migration guide][] has been added, so the CHANGES entries
180 listed here are only a brief description.
181 The migration guide contains more detailed information related to new features,
182 breaking changes, and mappings for the large list of deprecated functions.
184 [Migration guide]: https://github.com/openssl/openssl/tree/master/doc/man7/migration_guide.pod
186 ### Changes between 3.0.3 and 3.0.4 [21 June 2022]
188 * In addition to the c_rehash shell command injection identified in
189 CVE-2022-1292, further bugs where the c_rehash script does not
190 properly sanitise shell metacharacters to prevent command injection have been
193 When the CVE-2022-1292 was fixed it was not discovered that there
194 are other places in the script where the file names of certificates
195 being hashed were possibly passed to a command executed through the shell.
197 This script is distributed by some operating systems in a manner where
198 it is automatically executed. On such operating systems, an attacker
199 could execute arbitrary commands with the privileges of the script.
201 Use of the c_rehash script is considered obsolete and should be replaced
202 by the OpenSSL rehash command line tool.
205 *Daniel Fiala, Tomáš Mráz*
207 * Case insensitive string comparison no longer uses locales. It has instead
208 been directly implemented.
212 ### Changes between 3.0.2 and 3.0.3 [3 May 2022]
214 * Case insensitive string comparison is reimplemented via new locale-agnostic
215 comparison functions OPENSSL_str[n]casecmp always using the POSIX locale for
216 comparison. The previous implementation had problems when the Turkish locale
221 * Fixed a bug in the c_rehash script which was not properly sanitising shell
222 metacharacters to prevent command injection. This script is distributed by
223 some operating systems in a manner where it is automatically executed. On
224 such operating systems, an attacker could execute arbitrary commands with the
225 privileges of the script.
227 Use of the c_rehash script is considered obsolete and should be replaced
228 by the OpenSSL rehash command line tool.
233 * Fixed a bug in the function `OCSP_basic_verify` that verifies the signer
234 certificate on an OCSP response. The bug caused the function in the case
235 where the (non-default) flag OCSP_NOCHECKS is used to return a postivie
236 response (meaning a successful verification) even in the case where the
237 response signing certificate fails to verify.
239 It is anticipated that most users of `OCSP_basic_verify` will not use the
240 OCSP_NOCHECKS flag. In this case the `OCSP_basic_verify` function will return
241 a negative value (indicating a fatal error) in the case of a certificate
242 verification failure. The normal expected return value in this case would be
245 This issue also impacts the command line OpenSSL "ocsp" application. When
246 verifying an ocsp response with the "-no_cert_checks" option the command line
247 application will report that the verification is successful even though it
248 has in fact failed. In this case the incorrect successful response will also
249 be accompanied by error messages showing the failure and contradicting the
250 apparently successful result.
255 * Fixed a bug where the RC4-MD5 ciphersuite incorrectly used the
256 AAD data as the MAC key. This made the MAC key trivially predictable.
258 An attacker could exploit this issue by performing a man-in-the-middle attack
259 to modify data being sent from one endpoint to an OpenSSL 3.0 recipient such
260 that the modified data would still pass the MAC integrity check.
262 Note that data sent from an OpenSSL 3.0 endpoint to a non-OpenSSL 3.0
263 endpoint will always be rejected by the recipient and the connection will
264 fail at that point. Many application protocols require data to be sent from
265 the client to the server first. Therefore, in such a case, only an OpenSSL
266 3.0 server would be impacted when talking to a non-OpenSSL 3.0 client.
268 If both endpoints are OpenSSL 3.0 then the attacker could modify data being
269 sent in both directions. In this case both clients and servers could be
270 affected, regardless of the application protocol.
272 Note that in the absence of an attacker this bug means that an OpenSSL 3.0
273 endpoint communicating with a non-OpenSSL 3.0 endpoint will fail to complete
274 the handshake when using this ciphersuite.
276 The confidentiality of data is not impacted by this issue, i.e. an attacker
277 cannot decrypt data that has been encrypted using this ciphersuite - they can
280 In order for this attack to work both endpoints must legitimately negotiate
281 the RC4-MD5 ciphersuite. This ciphersuite is not compiled by default in
282 OpenSSL 3.0, and is not available within the default provider or the default
283 ciphersuite list. This ciphersuite will never be used if TLSv1.3 has been
284 negotiated. In order for an OpenSSL 3.0 endpoint to use this ciphersuite the
285 following must have occurred:
287 1) OpenSSL must have been compiled with the (non-default) compile time option
288 enable-weak-ssl-ciphers
290 2) OpenSSL must have had the legacy provider explicitly loaded (either
291 through application code or via configuration)
293 3) The ciphersuite must have been explicitly added to the ciphersuite list
295 4) The libssl security level must have been set to 0 (default is 1)
297 5) A version of SSL/TLS below TLSv1.3 must have been negotiated
299 6) Both endpoints must negotiate the RC4-MD5 ciphersuite in preference to any
300 others that both endpoints have in common
305 * Fix a bug in the OPENSSL_LH_flush() function that breaks reuse of the memory
306 occuppied by the removed hash table entries.
308 This function is used when decoding certificates or keys. If a long lived
309 process periodically decodes certificates or keys its memory usage will
310 expand without bounds and the process might be terminated by the operating
311 system causing a denial of service. Also traversing the empty hash table
312 entries will take increasingly more time.
314 Typically such long lived processes might be TLS clients or TLS servers
315 configured to accept client certificate authentication.
318 *Hugo Landau, Aliaksei Levin*
320 * The functions `OPENSSL_LH_stats` and `OPENSSL_LH_stats_bio` now only report
321 the `num_items`, `num_nodes` and `num_alloc_nodes` statistics. All other
322 statistics are no longer supported. For compatibility, these statistics are
323 still listed in the output but are now always reported as zero.
327 ### Changes between 3.0.1 and 3.0.2 [15 Mar 2022]
329 * Fixed a bug in the BN_mod_sqrt() function that can cause it to loop forever
330 for non-prime moduli.
332 Internally this function is used when parsing certificates that contain
333 elliptic curve public keys in compressed form or explicit elliptic curve
334 parameters with a base point encoded in compressed form.
336 It is possible to trigger the infinite loop by crafting a certificate that
337 has invalid explicit curve parameters.
339 Since certificate parsing happens prior to verification of the certificate
340 signature, any process that parses an externally supplied certificate may thus
341 be subject to a denial of service attack. The infinite loop can also be
342 reached when parsing crafted private keys as they can contain explicit
343 elliptic curve parameters.
345 Thus vulnerable situations include:
347 - TLS clients consuming server certificates
348 - TLS servers consuming client certificates
349 - Hosting providers taking certificates or private keys from customers
350 - Certificate authorities parsing certification requests from subscribers
351 - Anything else which parses ASN.1 elliptic curve parameters
353 Also any other applications that use the BN_mod_sqrt() where the attacker
354 can control the parameter values are vulnerable to this DoS issue.
359 * Add ciphersuites based on DHE_PSK (RFC 4279) and ECDHE_PSK (RFC 5489)
360 to the list of ciphersuites providing Perfect Forward Secrecy as
361 required by SECLEVEL >= 3.
363 *Dmitry Belyavskiy, Nicola Tuveri*
365 * Made the AES constant time code for no-asm configurations
366 optional due to the resulting 95% performance degradation.
367 The AES constant time code can be enabled, for no assembly
368 builds, with: ./config no-asm -DOPENSSL_AES_CONST_TIME
372 * Fixed PEM_write_bio_PKCS8PrivateKey() to make it possible to use empty
377 * The negative return value handling of the certificate verification callback
378 was reverted. The replacement is to set the verification retry state with
379 the SSL_set_retry_verify() function.
383 ### Changes between 3.0.0 and 3.0.1 [14 Dec 2021]
385 * Fixed invalid handling of X509_verify_cert() internal errors in libssl
386 Internally libssl in OpenSSL calls X509_verify_cert() on the client side to
387 verify a certificate supplied by a server. That function may return a
388 negative return value to indicate an internal error (for example out of
389 memory). Such a negative return value is mishandled by OpenSSL and will cause
390 an IO function (such as SSL_connect() or SSL_do_handshake()) to not indicate
391 success and a subsequent call to SSL_get_error() to return the value
392 SSL_ERROR_WANT_RETRY_VERIFY. This return value is only supposed to be
393 returned by OpenSSL if the application has previously called
394 SSL_CTX_set_cert_verify_callback(). Since most applications do not do this
395 the SSL_ERROR_WANT_RETRY_VERIFY return value from SSL_get_error() will be
396 totally unexpected and applications may not behave correctly as a result. The
397 exact behaviour will depend on the application but it could result in
398 crashes, infinite loops or other similar incorrect responses.
400 This issue is made more serious in combination with a separate bug in OpenSSL
401 3.0 that will cause X509_verify_cert() to indicate an internal error when
402 processing a certificate chain. This will occur where a certificate does not
403 include the Subject Alternative Name extension but where a Certificate
404 Authority has enforced name constraints. This issue can occur even with valid
410 * Corrected a few file name and file reference bugs in the build,
411 installation and setup scripts, which lead to installation verification
412 failures. Slightly enhanced the installation verification script.
416 * Fixed EVP_PKEY_eq() to make it possible to use it with strictly private
421 * Fixed PVK encoder to properly query for the passphrase.
425 * Multiple fixes in the OSSL_HTTP API functions.
429 * Allow sign extension in OSSL_PARAM_allocate_from_text() for the
430 OSSL_PARAM_INTEGER data type and return error on negative numbers
431 used with the OSSL_PARAM_UNSIGNED_INTEGER data type. Make
432 OSSL_PARAM_BLD_push_BN{,_pad}() return an error on negative numbers.
436 * Allow copying uninitialized digest contexts with EVP_MD_CTX_copy_ex.
440 * Fixed detection of ARMv7 and ARM64 CPU features on FreeBSD.
444 * Multiple threading fixes.
448 * Added NULL digest implementation to keep compatibility with 1.1.1 version.
452 * Allow fetching an operation from the provider that owns an unexportable key
453 as a fallback if that is still allowed by the property query.
457 ### Changes between 1.1.1 and 3.0.0 [7 Sep 2021]
459 * TLS_MAX_VERSION, DTLS_MAX_VERSION and DTLS_MIN_VERSION constants are now
464 * The `OPENSSL_s390xcap` environment variable can be used to set bits in the
465 S390X capability vector to zero. This simplifies testing of different code
466 paths on S390X architecture.
470 * Encrypting more than 2^64 TLS records with AES-GCM is disallowed
471 as per FIPS 140-2 IG A.5 "Key/IV Pair Uniqueness Requirements from
472 SP 800-38D". The communication will fail at this point.
476 * The EC_GROUP_clear_free() function is deprecated as there is nothing
477 confidential in EC_GROUP data.
481 * The byte order mark (BOM) character is ignored if encountered at the
482 beginning of a PEM-formatted file.
486 * Added CMS support for the Russian GOST algorithms.
490 * Due to move of the implementation of cryptographic operations
491 to the providers, validation of various operation parameters can
492 be postponed until the actual operation is executed where previously
493 it happened immediately when an operation parameter was set.
495 For example when setting an unsupported curve with
496 EVP_PKEY_CTX_set_ec_paramgen_curve_nid() this function call will not
497 fail but later keygen operations with the EVP_PKEY_CTX will fail.
499 *OpenSSL team members and many third party contributors*
501 * The EVP_get_cipherbyname() function will return NULL for algorithms such as
502 "AES-128-SIV", "AES-128-CBC-CTS" and "CAMELLIA-128-CBC-CTS" which were
503 previously only accessible via low level interfaces. Use EVP_CIPHER_fetch()
504 instead to retrieve these algorithms from a provider.
508 * On build targets where the multilib postfix is set in the build
509 configuration the libdir directory was changing based on whether
510 the lib directory with the multilib postfix exists on the system
511 or not. This unpredictable behavior was removed and eventual
512 multilib postfix is now always added to the default libdir. Use
513 `--libdir=lib` to override the libdir if adding the postfix is
518 * The triple DES key wrap functionality now conforms to RFC 3217 but is
519 no longer interoperable with OpenSSL 1.1.1.
523 * The ERR_GET_FUNC() function was removed. With the loss of meaningful
524 function codes, this function can only cause problems for calling
529 * Add a configurable flag to output date formats as ISO 8601. Does not
530 change the default date format.
534 * Version of MSVC earlier than 1300 could get link warnings, which could
535 be suppressed if the undocumented -DI_CAN_LIVE_WITH_LNK4049 was set.
536 Support for this flag has been removed.
540 * Rework and make DEBUG macros consistent. Remove unused -DCONF_DEBUG,
541 -DBN_CTX_DEBUG, and REF_PRINT. Add a new tracing category and use it for
542 printing reference counts. Rename -DDEBUG_UNUSED to -DUNUSED_RESULT_DEBUG
543 Fix BN_DEBUG_RAND so it compiles and, when set, force DEBUG_RAND to be set
544 also. Rename engine_debug_ref to be ENGINE_REF_PRINT also for consistency.
548 * The signatures of the functions to get and set options on SSL and
549 SSL_CTX objects changed from "unsigned long" to "uint64_t" type.
550 Some source code changes may be required.
554 * The public definitions of conf_method_st and conf_st have been
555 deprecated. They will be made opaque in a future release.
557 *Rich Salz and Tomáš Mráz*
559 * Client-initiated renegotiation is disabled by default. To allow it, use
560 the -client_renegotiation option, the SSL_OP_ALLOW_CLIENT_RENEGOTIATION
561 flag, or the "ClientRenegotiation" config parameter as appropriate.
565 * Add "abspath" and "includedir" pragma's to config files, to prevent,
566 or modify relative pathname inclusion.
570 * OpenSSL includes a cryptographic module that is intended to be FIPS 140-2
571 validated. Please consult the README-FIPS and
572 README-PROVIDERS files, as well as the migration guide.
574 *OpenSSL team members and many third party contributors*
576 * For the key types DH and DHX the allowed settable parameters are now different.
580 * The openssl commands that read keys, certificates, and CRLs now
581 automatically detect the PEM or DER format of the input files.
583 *David von Oheimb, Richard Levitte, and Tomáš Mráz*
585 * Added enhanced PKCS#12 APIs which accept a library context.
589 * The default manual page suffix ($MANSUFFIX) has been changed to "ossl"
593 * Added support for Kernel TLS (KTLS).
595 *Boris Pismenny, John Baldwin and Andrew Gallatin*
597 * Support for RFC 5746 secure renegotiation is now required by default for
598 SSL or TLS connections to succeed.
602 * The signature of the `copy` functional parameter of the
603 EVP_PKEY_meth_set_copy() function has changed so its `src` argument is
604 now `const EVP_PKEY_CTX *` instead of `EVP_PKEY_CTX *`. Similarly
605 the signature of the `pub_decode` functional parameter of the
606 EVP_PKEY_asn1_set_public() function has changed so its `pub` argument is
607 now `const X509_PUBKEY *` instead of `X509_PUBKEY *`.
611 * The error return values from some control calls (ctrl) have changed.
615 * A public key check is now performed during EVP_PKEY_derive_set_peer().
619 * Many functions in the EVP_ namespace that are getters of values from
620 implementations or contexts were renamed to include get or get0 in their
621 names. Old names are provided as macro aliases for compatibility and
626 * The EVP_PKEY_CTRL_PKCS7_ENCRYPT, EVP_PKEY_CTRL_PKCS7_DECRYPT,
627 EVP_PKEY_CTRL_PKCS7_SIGN, EVP_PKEY_CTRL_CMS_ENCRYPT,
628 EVP_PKEY_CTRL_CMS_DECRYPT, and EVP_PKEY_CTRL_CMS_SIGN control operations
633 * The EVP_PKEY_public_check() and EVP_PKEY_param_check() functions now work for
636 * The output from the command line applications may have minor
641 * The output from numerous "printing" may have minor changes.
645 * Windows thread synchronization uses read/write primitives (SRWLock) when
646 supported by the OS, otherwise CriticalSection continues to be used.
650 * Add filter BIO BIO_f_readbuffer() that allows BIO_tell() and BIO_seek() to
651 work on read only BIO source/sinks that do not support these functions.
652 This allows piping or redirection of a file BIO using stdin to be buffered
653 into memory. This is used internally in OSSL_DECODER_from_bio().
657 * OSSL_STORE_INFO_get_type() may now return an additional value. In 1.1.1
658 this function would return one of the values OSSL_STORE_INFO_NAME,
659 OSSL_STORE_INFO_PKEY, OSSL_STORE_INFO_PARAMS, OSSL_STORE_INFO_CERT or
660 OSSL_STORE_INFO_CRL. Decoded public keys would previously have been reported
661 as type OSSL_STORE_INFO_PKEY in 1.1.1. In 3.0 decoded public keys are now
662 reported as having the new type OSSL_STORE_INFO_PUBKEY. Applications
663 using this function should be amended to handle the changed return value.
667 * Improved adherence to Enhanced Security Services (ESS, RFC 2634 and RFC 5035)
668 for the TSP and CMS Advanced Electronic Signatures (CAdES) implementations.
669 As required by RFC 5035 check both ESSCertID and ESSCertIDv2 if both present.
670 Correct the semantics of checking the validation chain in case ESSCertID{,v2}
671 contains more than one certificate identifier: This means that all
672 certificates referenced there MUST be part of the validation chain.
676 * The implementation of older EVP ciphers related to CAST, IDEA, SEED, RC2, RC4,
677 RC5, DESX and DES have been moved to the legacy provider.
681 * The implementation of the EVP digests MD2, MD4, MDC2, WHIRLPOOL and
682 RIPEMD-160 have been moved to the legacy provider.
686 * The deprecated function EVP_PKEY_get0() now returns NULL being called for a
691 * The deprecated functions EVP_PKEY_get0_RSA(),
692 EVP_PKEY_get0_DSA(), EVP_PKEY_get0_EC_KEY(), EVP_PKEY_get0_DH(),
693 EVP_PKEY_get0_hmac(), EVP_PKEY_get0_poly1305() and EVP_PKEY_get0_siphash() as
694 well as the similarly named "get1" functions behave differently in
699 * A number of functions handling low-level keys or engines were deprecated
700 including EVP_PKEY_set1_engine(), EVP_PKEY_get0_engine(), EVP_PKEY_assign(),
701 EVP_PKEY_get0(), EVP_PKEY_get0_hmac(), EVP_PKEY_get0_poly1305() and
702 EVP_PKEY_get0_siphash().
706 * PKCS#5 PBKDF1 key derivation has been moved from PKCS5_PBE_keyivgen() into
707 the legacy crypto provider as an EVP_KDF. Applications requiring this KDF
708 will need to load the legacy crypto provider. This includes these PBE
709 algorithms which use this KDF:
710 - NID_pbeWithMD2AndDES_CBC
711 - NID_pbeWithMD5AndDES_CBC
712 - NID_pbeWithSHA1AndRC2_CBC
713 - NID_pbeWithMD2AndRC2_CBC
714 - NID_pbeWithMD5AndRC2_CBC
715 - NID_pbeWithSHA1AndDES_CBC
719 * Deprecated obsolete BIO_set_callback(), BIO_get_callback(), and
720 BIO_debug_callback() functions.
724 * Deprecated obsolete EVP_PKEY_CTX_get0_dh_kdf_ukm() and
725 EVP_PKEY_CTX_get0_ecdh_kdf_ukm() functions.
729 * The RAND_METHOD APIs have been deprecated.
733 * The SRP APIs have been deprecated.
737 * Add a compile time option to prevent the caching of provider fetched
738 algorithms. This is enabled by including the no-cached-fetch option
739 at configuration time.
743 * pkcs12 now uses defaults of PBKDF2, AES and SHA-256, with a MAC iteration
744 count of PKCS12_DEFAULT_ITER.
746 *Tomáš Mráz and Sahana Prasad*
748 * The openssl speed command does not use low-level API calls anymore.
752 * Parallel dual-prime 1024-bit modular exponentiation for AVX512_IFMA
755 *Ilya Albrekht, Sergey Kirillov, Andrey Matyukov (Intel Corp)*
757 * Combining the Configure options no-ec and no-dh no longer disables TLSv1.3.
761 * Implemented support for fully "pluggable" TLSv1.3 groups. This means that
762 providers may supply their own group implementations (using either the "key
763 exchange" or the "key encapsulation" methods) which will automatically be
764 detected and used by libssl.
766 *Matt Caswell, Nicola Tuveri*
768 * The undocumented function X509_certificate_type() has been deprecated;
772 * Deprecated the obsolete BN_pseudo_rand() and BN_pseudo_rand_range().
776 * Removed RSA padding mode for SSLv23 (which was only used for
777 SSLv2). This includes the functions RSA_padding_check_SSLv23() and
778 RSA_padding_add_SSLv23() and the `-ssl` option in the deprecated
783 * Deprecated the obsolete X9.31 RSA key generation related functions.
785 * While a callback function set via `SSL_CTX_set_cert_verify_callback()`
786 is not allowed to return a value > 1, this is no more taken as failure.
788 *Viktor Dukhovni and David von Oheimb*
790 * Deprecated the obsolete X9.31 RSA key generation related functions
791 BN_X931_generate_Xpq(), BN_X931_derive_prime_ex(), and
792 BN_X931_generate_prime_ex().
796 * The default key generation method for the regular 2-prime RSA keys was
797 changed to the FIPS 186-4 B.3.6 method.
801 * Deprecated the BN_is_prime_ex() and BN_is_prime_fasttest_ex() functions.
805 * Deprecated EVP_MD_CTX_set_update_fn() and EVP_MD_CTX_update_fn().
809 * Deprecated the type OCSP_REQ_CTX and the functions OCSP_REQ_CTX_*() and
810 replaced with OSSL_HTTP_REQ_CTX and the functions OSSL_HTTP_REQ_CTX_*().
812 *Rich Salz, Richard Levitte, and David von Oheimb*
814 * Deprecated `X509_http_nbio()` and `X509_CRL_http_nbio()`.
818 * Deprecated `OCSP_parse_url()`.
822 * Validation of SM2 keys has been separated from the validation of regular EC
827 * Behavior of the `pkey` app is changed, when using the `-check` or `-pubcheck`
828 switches: a validation failure triggers an early exit, returning a failure
829 exit status to the parent process.
833 * Changed behavior of SSL_CTX_set_ciphersuites() and SSL_set_ciphersuites()
834 to ignore unknown ciphers.
838 * The `-cipher-commands` and `-digest-commands` options
839 of the command line utility `list` have been deprecated.
840 Instead use the `-cipher-algorithms` and `-digest-algorithms` options.
844 * Added convenience functions for generating asymmetric key pairs:
845 The 'quick' one-shot (yet somewhat limited) function L<EVP_PKEY_Q_keygen(3)>
846 and macros for the most common cases: <EVP_RSA_gen(3)> and L<EVP_EC_gen(3)>.
850 * All of the low level EC_KEY functions have been deprecated.
852 *Shane Lontis, Paul Dale, Richard Levitte, and Tomáš Mráz*
854 * Deprecated all the libcrypto and libssl error string loading
859 * The functions SSL_CTX_set_tmp_dh_callback and SSL_set_tmp_dh_callback, as
860 well as the macros SSL_CTX_set_tmp_dh() and SSL_set_tmp_dh() have been
865 * The `-crypt` option to the `passwd` command line tool has been removed.
869 * The -C option to the `x509`, `dhparam`, `dsaparam`, and `ecparam` commands
874 * Add support for AES Key Wrap inverse ciphers to the EVP layer.
878 * Deprecated EVP_PKEY_set1_tls_encodedpoint() and
879 EVP_PKEY_get1_tls_encodedpoint().
883 * The security callback, which can be customised by application code, supports
884 the security operation SSL_SECOP_TMP_DH. One location of the "other" parameter
885 was incorrectly passing a DH object. It now passed an EVP_PKEY in all cases.
889 * Add PKCS7_get_octet_string() and PKCS7_type_is_other() to the public
890 interface. Their functionality remains unchanged.
894 * Added new option for 'openssl list', '-providers', which will display the
895 list of loaded providers, their names, version and status. It optionally
896 displays their gettable parameters.
900 * Removed EVP_PKEY_set_alias_type().
904 * Deprecated `EVP_PKEY_CTX_set_rsa_keygen_pubexp()` and introduced
905 `EVP_PKEY_CTX_set1_rsa_keygen_pubexp()`, which is now preferred.
909 * Changed all "STACK" functions to be macros instead of inline functions. Macro
910 parameters are still checked for type safety at compile time via helper
915 * Remove the RAND_DRBG API
917 *Paul Dale and Matthias St. Pierre*
919 * Allow `SSL_set1_host()` and `SSL_add1_host()` to take IP literal addresses
920 as well as actual hostnames.
924 * The 'MinProtocol' and 'MaxProtocol' configuration commands now silently
925 ignore TLS protocol version bounds when configuring DTLS-based contexts, and
926 conversely, silently ignore DTLS protocol version bounds when configuring
927 TLS-based contexts. The commands can be repeated to set bounds of both
928 types. The same applies with the corresponding "min_protocol" and
929 "max_protocol" command-line switches, in case some application uses both TLS
932 SSL_CTX instances that are created for a fixed protocol version (e.g.
933 `TLSv1_server_method()`) also silently ignore version bounds. Previously
934 attempts to apply bounds to these protocol versions would result in an
935 error. Now only the "version-flexible" SSL_CTX instances are subject to
936 limits in configuration files in command-line options.
940 * Deprecated the `ENGINE` API. Engines should be replaced with providers
945 * Reworked the recorded ERR codes to make better space for system errors.
946 To distinguish them, the macro `ERR_SYSTEM_ERROR()` indicates if the
947 given code is a system error (true) or an OpenSSL error (false).
951 * Reworked the test perl framework to better allow parallel testing.
953 *Nicola Tuveri and David von Oheimb*
955 * Added ciphertext stealing algorithms AES-128-CBC-CTS, AES-192-CBC-CTS and
956 AES-256-CBC-CTS to the providers. CS1, CS2 and CS3 variants are supported.
960 * 'Configure' has been changed to figure out the configuration target if
961 none is given on the command line. Consequently, the 'config' script is
962 now only a mere wrapper. All documentation is changed to only mention
965 *Rich Salz and Richard Levitte*
967 * Added a library context `OSSL_LIB_CTX` that applications as well as
968 other libraries can use to form a separate context within which
969 libcrypto operations are performed.
973 * Added various `_ex` functions to the OpenSSL API that support using
974 a non-default `OSSL_LIB_CTX`.
978 * Handshake now fails if Extended Master Secret extension is dropped
983 * Dropped interactive mode from the `openssl` program.
987 * Deprecated `EVP_PKEY_cmp()` and `EVP_PKEY_cmp_parameters()`.
989 *David von Oheimb and Shane Lontis*
991 * Deprecated `EC_METHOD_get_field_type()`.
995 * Deprecated EC_GFp_simple_method(), EC_GFp_mont_method(),
996 EC_GF2m_simple_method(), EC_GFp_nist_method(), EC_GFp_nistp224_method()
997 EC_GFp_nistp256_method(), and EC_GFp_nistp521_method().
1001 * Deprecated EC_GROUP_new(), EC_GROUP_method_of(), and EC_POINT_method_of().
1005 * Add CAdES-BES signature verification support, mostly derived
1006 from ESSCertIDv2 TS (RFC 5816) contribution by Marek Klein.
1008 *Filipe Raimundo da Silva*
1010 * Add CAdES-BES signature scheme and attributes support (RFC 5126) to CMS API.
1014 * Added the AuthEnvelopedData content type structure (RFC 5083) with AES-GCM
1015 parameter (RFC 5084) for the Cryptographic Message Syntax (CMS).
1019 * Deprecated EC_POINT_make_affine() and EC_POINTs_make_affine().
1023 * Deprecated EC_GROUP_precompute_mult(), EC_GROUP_have_precompute_mult(), and
1024 EC_KEY_precompute_mult().
1028 * Deprecated EC_POINTs_mul().
1032 * Removed FIPS_mode() and FIPS_mode_set().
1036 * The SSL option SSL_OP_IGNORE_UNEXPECTED_EOF is introduced.
1040 * Deprecated EC_POINT_set_Jprojective_coordinates_GFp() and
1041 EC_POINT_get_Jprojective_coordinates_GFp().
1045 * Added OSSL_PARAM_BLD to the public interface. This allows OSSL_PARAM
1046 arrays to be more easily constructed via a series of utility functions.
1047 Create a parameter builder using OSSL_PARAM_BLD_new(), add parameters using
1048 the various push functions and finally convert to a passable OSSL_PARAM
1049 array using OSSL_PARAM_BLD_to_param().
1053 * The security strength of SHA1 and MD5 based signatures in TLS has been
1058 * Added EVP_PKEY_set_type_by_keymgmt(), to initialise an EVP_PKEY to
1059 contain a provider side internal key.
1063 * ASN1_verify(), ASN1_digest() and ASN1_sign() have been deprecated.
1067 * Project text documents not yet having a proper file name extension
1068 (`HACKING`, `LICENSE`, `NOTES*`, `README*`, `VERSION`) have been renamed to
1069 `*.md` as far as reasonable, else `*.txt`, for better use with file managers.
1073 * The main project documents (README, NEWS, CHANGES, INSTALL, SUPPORT)
1074 have been converted to Markdown with the goal to produce documents
1075 which not only look pretty when viewed online in the browser, but
1076 remain well readable inside a plain text editor.
1078 To achieve this goal, a 'minimalistic' Markdown style has been applied
1079 which avoids formatting elements that interfere too much with the
1080 reading flow in the text file. For example, it
1082 * avoids [ATX headings][] and uses [setext headings][] instead
1083 (which works for `<h1>` and `<h2>` headings only).
1084 * avoids [inline links][] and uses [reference links][] instead.
1085 * avoids [fenced code blocks][] and uses [indented code blocks][] instead.
1087 [ATX headings]: https://github.github.com/gfm/#atx-headings
1088 [setext headings]: https://github.github.com/gfm/#setext-headings
1089 [inline links]: https://github.github.com/gfm/#inline-link
1090 [reference links]: https://github.github.com/gfm/#reference-link
1091 [fenced code blocks]: https://github.github.com/gfm/#fenced-code-blocks
1092 [indented code blocks]: https://github.github.com/gfm/#indented-code-blocks
1094 *Matthias St. Pierre*
1096 * The test suite is changed to preserve results of each test recipe.
1097 A new directory test-runs/ with subdirectories named like the
1098 test recipes are created in the build tree for this purpose.
1102 * Added an implementation of CMP and CRMF (RFC 4210, RFC 4211 RFC 6712).
1103 This adds `crypto/cmp/`, `crpyto/crmf/`, `apps/cmp.c`, and `test/cmp_*`.
1104 See L<openssl-cmp(1)> and L<OSSL_CMP_exec_IR_ses(3)> as starting points.
1106 *David von Oheimb, Martin Peylo*
1108 * Generalized the HTTP client code from `crypto/ocsp/` into `crpyto/http/`.
1109 It supports arbitrary request and response content types, GET redirection,
1110 TLS, connections via HTTP(S) proxies, connections and exchange via
1111 user-defined BIOs (allowing implicit connections), persistent connections,
1112 and timeout checks. See L<OSSL_HTTP_transfer(3)> etc. for details.
1113 The legacy OCSP-focused (and only partly documented) API
1114 is retained for backward compatibility, while most of it is deprecated.
1118 * Added `util/check-format.pl`, a tool for checking adherence to the
1119 OpenSSL coding style <https://www.openssl.org/policies/codingstyle.html>.
1120 The checks performed are incomplete and yield some false positives.
1121 Still the tool should be useful for detecting most typical glitches.
1125 * `BIO_do_connect()` and `BIO_do_handshake()` have been extended:
1126 If domain name resolution yields multiple IP addresses all of them are tried
1127 after `connect()` failures.
1131 * All of the low level RSA functions have been deprecated.
1135 * X509 certificates signed using SHA1 are no longer allowed at security
1140 * The command line utilities dhparam, dsa, gendsa and dsaparam have been
1141 modified to use PKEY APIs. These commands are now in maintenance mode
1142 and no new features will be added to them.
1146 * The command line utility rsautl has been deprecated.
1150 * The command line utilities genrsa and rsa have been modified to use PKEY
1151 APIs. They now write PKCS#8 keys by default. These commands are now in
1152 maintenance mode and no new features will be added to them.
1156 * All of the low level DH functions have been deprecated.
1158 *Paul Dale and Matt Caswell*
1160 * All of the low level DSA functions have been deprecated.
1164 * Reworked the treatment of EC EVP_PKEYs with the SM2 curve to
1165 automatically become EVP_PKEY_SM2 rather than EVP_PKEY_EC.
1169 * Deprecated low level ECDH and ECDSA functions.
1173 * Deprecated EVP_PKEY_decrypt_old() and EVP_PKEY_encrypt_old().
1177 * Enhanced the documentation of EVP_PKEY_get_size(), EVP_PKEY_get_bits()
1178 and EVP_PKEY_get_security_bits(). Especially EVP_PKEY_get_size() needed
1179 a new formulation to include all the things it can be used for,
1180 as well as words of caution.
1184 * The SSL_CTX_set_tlsext_ticket_key_cb(3) function has been deprecated.
1188 * All of the low level HMAC functions have been deprecated.
1190 *Paul Dale and David von Oheimb*
1192 * Over two thousand fixes were made to the documentation, including:
1193 - Common options (such as -rand/-writerand, TLS version control, etc)
1194 were refactored and point to newly-enhanced descriptions in openssl.pod.
1195 - Added style conformance for all options (with help from Richard Levitte),
1196 documented all reported missing options, added a CI build to check
1197 that all options are documented and that no unimplemented options
1199 - Documented some internals, such as all use of environment variables.
1200 - Addressed all internal broken L<> references.
1204 * All of the low level CMAC functions have been deprecated.
1208 * The low-level MD2, MD4, MD5, MDC2, RIPEMD160 and Whirlpool digest
1209 functions have been deprecated.
1211 *Paul Dale and David von Oheimb*
1213 * Corrected the documentation of the return values from the `EVP_DigestSign*`
1214 set of functions. The documentation mentioned negative values for some
1215 errors, but this was never the case, so the mention of negative values
1218 Code that followed the documentation and thereby check with something
1219 like `EVP_DigestSignInit(...) <= 0` will continue to work undisturbed.
1223 * All of the low level cipher functions have been deprecated.
1225 *Matt Caswell and Paul Dale*
1227 * Removed include/openssl/opensslconf.h.in and replaced it with
1228 include/openssl/configuration.h.in, which differs in not including
1229 <openssl/macros.h>. A short header include/openssl/opensslconf.h
1230 was added to include both.
1232 This allows internal hacks where one might need to modify the set
1233 of configured macros, for example this if deprecated symbols are
1234 still supposed to be available internally:
1236 #include <openssl/configuration.h>
1238 #undef OPENSSL_NO_DEPRECATED
1239 #define OPENSSL_SUPPRESS_DEPRECATED
1241 #include <openssl/macros.h>
1243 This should not be used by applications that use the exported
1244 symbols, as that will lead to linking errors.
1248 * Fixed an overflow bug in the x64_64 Montgomery squaring procedure
1249 used in exponentiation with 512-bit moduli. No EC algorithms are
1250 affected. Analysis suggests that attacks against 2-prime RSA1024,
1251 3-prime RSA1536, and DSA1024 as a result of this defect would be very
1252 difficult to perform and are not believed likely. Attacks against DH512
1253 are considered just feasible. However, for an attack the target would
1254 have to re-use the DH512 private key, which is not recommended anyway.
1255 Also applications directly using the low-level API BN_mod_exp may be
1256 affected if they use BN_FLG_CONSTTIME.
1261 * Most memory-debug features have been deprecated, and the functionality
1262 replaced with no-ops.
1266 * Added documentation for the STACK API.
1270 * Introduced a new method type and API, OSSL_ENCODER, to represent
1271 generic encoders. These do the same sort of job that PEM writers
1272 and d2i functions do, but with support for methods supplied by
1273 providers, and the possibility for providers to support other
1278 * Introduced a new method type and API, OSSL_DECODER, to represent
1279 generic decoders. These do the same sort of job that PEM readers
1280 and i2d functions do, but with support for methods supplied by
1281 providers, and the possibility for providers to support other
1286 * Added a .pragma directive to the syntax of configuration files, to
1287 allow varying behavior in a supported and predictable manner.
1288 Currently added pragma:
1292 This allows dollar signs to be a keyword character unless it's
1293 followed by a opening brace or parenthesis. This is useful for
1294 platforms where dollar signs are commonly used in names, such as
1295 volume names and system directory names on VMS.
1299 * Added functionality to create an EVP_PKEY from user data.
1303 * Change the interpretation of the '--api' configuration option to
1304 mean that this is a desired API compatibility level with no
1305 further meaning. The previous interpretation, that this would
1306 also mean to remove all deprecated symbols up to and including
1307 the given version, no requires that 'no-deprecated' is also used
1308 in the configuration.
1310 When building applications, the desired API compatibility level
1311 can be set with the OPENSSL_API_COMPAT macro like before. For
1312 API compatibility version below 3.0, the old style numerical
1313 value is valid as before, such as -DOPENSSL_API_COMPAT=0x10100000L.
1314 For version 3.0 and on, the value is expected to be the decimal
1315 value calculated from the major and minor version like this:
1317 MAJOR * 10000 + MINOR * 100
1321 -DOPENSSL_API_COMPAT=30000 For 3.0
1322 -DOPENSSL_API_COMPAT=30200 For 3.2
1324 To hide declarations that are deprecated up to and including the
1325 given API compatibility level, -DOPENSSL_NO_DEPRECATED must be
1326 given when building the application as well.
1330 * Added the X509_LOOKUP_METHOD called X509_LOOKUP_store, to allow
1331 access to certificate and CRL stores via URIs and OSSL_STORE
1334 This adds the following functions:
1336 - X509_LOOKUP_store()
1337 - X509_STORE_load_file()
1338 - X509_STORE_load_path()
1339 - X509_STORE_load_store()
1340 - SSL_add_store_cert_subjects_to_stack()
1341 - SSL_CTX_set_default_verify_store()
1342 - SSL_CTX_load_verify_file()
1343 - SSL_CTX_load_verify_dir()
1344 - SSL_CTX_load_verify_store()
1348 * Added a new method to gather entropy on VMS, based on SYS$GET_ENTROPY.
1349 The presence of this system service is determined at run-time.
1353 * Added functionality to create an EVP_PKEY context based on data
1354 for methods from providers. This takes an algorithm name and a
1355 property query string and simply stores them, with the intent
1356 that any operation that uses this context will use those strings
1357 to fetch the needed methods implicitly, thereby making the port
1358 of application written for pre-3.0 OpenSSL easier.
1362 * The undocumented function NCONF_WIN32() has been deprecated; for
1363 conversion details see the HISTORY section of doc/man5/config.pod
1367 * Introduced the new functions EVP_DigestSignInit_ex() and
1368 EVP_DigestVerifyInit_ex(). The macros EVP_DigestSignUpdate() and
1369 EVP_DigestVerifyUpdate() have been converted to functions. See the man
1370 pages for further details.
1374 * Over two thousand fixes were made to the documentation, including:
1375 adding missing command flags, better style conformance, documentation
1378 *Rich Salz, Richard Levitte*
1380 * s390x assembly pack: add hardware-support for P-256, P-384, P-521,
1381 X25519, X448, Ed25519 and Ed448.
1385 * Print all values for a PKCS#12 attribute with 'openssl pkcs12', not just
1390 * Deprecated the public definition of `ERR_STATE` as well as the function
1391 `ERR_get_state()`. This is done in preparation of making `ERR_STATE` an
1396 * Added ERR functionality to give callers access to the stored function
1397 names that have replaced the older function code based functions.
1399 New functions are ERR_peek_error_func(), ERR_peek_last_error_func(),
1400 ERR_peek_error_data(), ERR_peek_last_error_data(), ERR_get_error_all(),
1401 ERR_peek_error_all() and ERR_peek_last_error_all().
1403 Deprecate ERR functions ERR_get_error_line(), ERR_get_error_line_data(),
1404 ERR_peek_error_line_data(), ERR_peek_last_error_line_data() and
1405 ERR_func_error_string().
1409 * Extended testing to be verbose for failing tests only. The make variables
1410 VERBOSE_FAILURE or VF can be used to enable this:
1412 $ make VF=1 test # Unix
1413 $ mms /macro=(VF=1) test ! OpenVMS
1414 $ nmake VF=1 test # Windows
1418 * Added the `-copy_extensions` option to the `x509` command for use with
1419 `-req` and `-x509toreq`. When given with the `copy` or `copyall` argument,
1420 all extensions in the request are copied to the certificate or vice versa.
1422 *David von Oheimb*, *Kirill Stefanenkov <kirill_stefanenkov@rambler.ru>*
1424 * Added the `-copy_extensions` option to the `req` command for use with
1425 `-x509`. When given with the `copy` or `copyall` argument,
1426 all extensions in the certification request are copied to the certificate.
1430 * The `x509`, `req`, and `ca` commands now make sure that X.509v3 certificates
1431 they generate are by default RFC 5280 compliant in the following sense:
1432 There is a subjectKeyIdentifier extension with a hash value of the public key
1433 and for not self-signed certs there is an authorityKeyIdentifier extension
1434 with a keyIdentifier field or issuer information identifying the signing key.
1435 This is done unless some configuration overrides the new default behavior,
1436 such as `subjectKeyIdentifier = none` and `authorityKeyIdentifier = none`.
1440 * Added several checks to `X509_verify_cert()` according to requirements in
1441 RFC 5280 in case `X509_V_FLAG_X509_STRICT` is set
1442 (which may be done by using the CLI option `-x509_strict`):
1443 * The basicConstraints of CA certificates must be marked critical.
1444 * CA certificates must explicitly include the keyUsage extension.
1445 * If a pathlenConstraint is given the key usage keyCertSign must be allowed.
1446 * The issuer name of any certificate must not be empty.
1447 * The subject name of CA certs, certs with keyUsage crlSign,
1448 and certs without subjectAlternativeName must not be empty.
1449 * If a subjectAlternativeName extension is given it must not be empty.
1450 * The signatureAlgorithm field and the cert signature must be consistent.
1451 * Any given authorityKeyIdentifier and any given subjectKeyIdentifier
1452 must not be marked critical.
1453 * The authorityKeyIdentifier must be given for X.509v3 certs
1454 unless they are self-signed.
1455 * The subjectKeyIdentifier must be given for all X.509v3 CA certs.
1459 * Certificate verification using `X509_verify_cert()` meanwhile rejects EC keys
1460 with explicit curve parameters (specifiedCurve) as required by RFC 5480.
1464 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
1465 used even when parsing explicit parameters, when loading a encoded key
1466 or calling `EC_GROUP_new_from_ecpkparameters()`/
1467 `EC_GROUP_new_from_ecparameters()`.
1468 This prevents bypass of security hardening and performance gains,
1469 especially for curves with specialized EC_METHODs.
1470 By default, if a key encoded with explicit parameters is loaded and later
1471 encoded, the output is still encoded with explicit parameters, even if
1472 internally a "named" EC_GROUP is used for computation.
1476 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
1477 this change, EC_GROUP_set_generator would accept order and/or cofactor as
1478 NULL. After this change, only the cofactor parameter can be NULL. It also
1479 does some minimal sanity checks on the passed order.
1484 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
1485 An attack is simple, if the first CMS_recipientInfo is valid but the
1486 second CMS_recipientInfo is chosen ciphertext. If the second
1487 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
1488 encryption key will be replaced by garbage, and the message cannot be
1489 decoded, but if the RSA decryption fails, the correct encryption key is
1490 used and the recipient will not notice the attack.
1491 As a work around for this potential attack the length of the decrypted
1492 key must be equal to the cipher default key length, in case the
1493 certifiate is not given and all recipientInfo are tried out.
1494 The old behaviour can be re-enabled in the CMS code by setting the
1495 CMS_DEBUG_DECRYPT flag.
1499 * Early start up entropy quality from the DEVRANDOM seed source has been
1500 improved for older Linux systems. The RAND subsystem will wait for
1501 /dev/random to be producing output before seeding from /dev/urandom.
1502 The seeded state is stored for future library initialisations using
1503 a system global shared memory segment. The shared memory identifier
1504 can be configured by defining OPENSSL_RAND_SEED_DEVRANDOM_SHM_ID to
1505 the desired value. The default identifier is 114.
1509 * Revised BN_generate_prime_ex to not avoid factors 2..17863 in p-1
1510 when primes for RSA keys are computed.
1511 Since we previously always generated primes == 2 (mod 3) for RSA keys,
1512 the 2-prime and 3-prime RSA modules were easy to distinguish, since
1513 `N = p*q = 1 (mod 3)`, but `N = p*q*r = 2 (mod 3)`. Therefore fingerprinting
1514 2-prime vs. 3-prime RSA keys was possible by computing N mod 3.
1515 This avoids possible fingerprinting of newly generated RSA modules.
1519 * Correct the extended master secret constant on EBCDIC systems. Without this
1520 fix TLS connections between an EBCDIC system and a non-EBCDIC system that
1521 negotiate EMS will fail. Unfortunately this also means that TLS connections
1522 between EBCDIC systems with this fix, and EBCDIC systems without this
1523 fix will fail if they negotiate EMS.
1527 * Changed the library initialisation so that the config file is now loaded
1528 by default. This was already the case for libssl. It now occurs for both
1529 libcrypto and libssl. Use the OPENSSL_INIT_NO_LOAD_CONFIG option to
1530 `OPENSSL_init_crypto()` to suppress automatic loading of a config file.
1534 * Introduced new error raising macros, `ERR_raise()` and `ERR_raise_data()`,
1535 where the former acts as a replacement for `ERR_put_error()`, and the
1536 latter replaces the combination `ERR_put_error()` + `ERR_add_error_data()`.
1537 `ERR_raise_data()` adds more flexibility by taking a format string and
1538 an arbitrary number of arguments following it, to be processed with
1543 * Introduced a new function, `OSSL_PROVIDER_available()`, which can be used
1544 to check if a named provider is loaded and available. When called, it
1545 will also activate all fallback providers if such are still present.
1549 * Enforce a minimum DH modulus size of 512 bits.
1553 * Changed DH parameters to generate the order q subgroup instead of 2q.
1554 Previously generated DH parameters are still accepted by DH_check
1555 but DH_generate_key works around that by clearing bit 0 of the
1556 private key for those. This avoids leaking bit 0 of the private key.
1560 * Significantly reduce secure memory usage by the randomness pools.
1564 * `{CRYPTO,OPENSSL}_mem_debug_{push,pop}` are now no-ops and have been
1569 * A new type, EVP_KEYEXCH, has been introduced to represent key exchange
1570 algorithms. An implementation of a key exchange algorithm can be obtained
1571 by using the function EVP_KEYEXCH_fetch(). An EVP_KEYEXCH algorithm can be
1572 used in a call to EVP_PKEY_derive_init_ex() which works in a similar way to
1573 the older EVP_PKEY_derive_init() function. See the man pages for the new
1574 functions for further details.
1578 * The EVP_PKEY_CTX_set_dh_pad() macro has now been converted to a function.
1582 * Removed the function names from error messages and deprecated the
1587 * Removed NextStep support and the macro OPENSSL_UNISTD
1591 * Removed DES_check_key. Also removed OPENSSL_IMPLEMENT_GLOBAL,
1592 OPENSSL_GLOBAL_REF, OPENSSL_DECLARE_GLOBAL.
1593 Also removed "export var as function" capability; we do not export
1594 variables, only functions.
1598 * RC5_32_set_key has been changed to return an int type, with 0 indicating
1599 an error and 1 indicating success. In previous versions of OpenSSL this
1600 was a void type. If a key was set longer than the maximum possible this
1605 * Support SM2 signing and verification schemes with X509 certificate.
1609 * Use SHA256 as the default digest for TS query in the `ts` app.
1613 * Change PBKDF2 to conform to SP800-132 instead of the older PKCS5 RFC2898.
1617 * Default cipher lists/suites are now available via a function, the
1618 #defines are deprecated.
1622 * Add target VC-WIN32-UWP, VC-WIN64A-UWP, VC-WIN32-ARM-UWP and
1623 VC-WIN64-ARM-UWP in Windows OneCore target for making building libraries
1624 for Windows Store apps easier. Also, the "no-uplink" option has been added.
1628 * Join the directories crypto/x509 and crypto/x509v3
1632 * Added command 'openssl kdf' that uses the EVP_KDF API.
1636 * Added command 'openssl mac' that uses the EVP_MAC API.
1640 * Added OPENSSL_info() to get diverse built-in OpenSSL data, such
1641 as default directories. Also added the command 'openssl info'
1642 for scripting purposes.
1646 * The functions AES_ige_encrypt() and AES_bi_ige_encrypt() have been
1651 * Add prediction resistance to the DRBG reseeding process.
1655 * Limit the number of blocks in a data unit for AES-XTS to 2^20 as
1656 mandated by IEEE Std 1619-2018.
1660 * Added newline escaping functionality to a filename when using openssl dgst.
1661 This output format is to replicate the output format found in the `*sum`
1662 checksum programs. This aims to preserve backward compatibility.
1664 *Matt Eaton, Richard Levitte, and Paul Dale*
1666 * Removed the heartbeat message in DTLS feature, as it has very
1667 little usage and doesn't seem to fulfill a valuable purpose.
1668 The configuration option is now deprecated.
1672 * Changed the output of 'openssl {digestname} < file' to display the
1673 digest name in its output.
1677 * Added a new generic trace API which provides support for enabling
1678 instrumentation through trace output.
1680 *Richard Levitte & Matthias St. Pierre*
1682 * Added build tests for C++. These are generated files that only do one
1683 thing, to include one public OpenSSL head file each. This tests that
1684 the public header files can be usefully included in a C++ application.
1686 This test isn't enabled by default. It can be enabled with the option
1687 'enable-buildtest-c++'.
1691 * Added KB KDF (EVP_KDF_KB) to EVP_KDF.
1695 * Added SSH KDF (EVP_KDF_SSHKDF) and KRB5 KDF (EVP_KDF_KRB5KDF) to EVP_KDF.
1699 * Added Single Step KDF (EVP_KDF_SS), X963 KDF, and X942 KDF to EVP_KDF.
1703 * Added KMAC to EVP_MAC.
1707 * Added property based algorithm implementation selection framework to
1712 * Added SCA hardening for modular field inversion in EC_GROUP through
1713 a new dedicated field_inv() pointer in EC_METHOD.
1714 This also addresses a leakage affecting conversions from projective
1715 to affine coordinates.
1717 *Billy Bob Brumley, Nicola Tuveri*
1719 * Added EVP_KDF, an EVP layer KDF API, to simplify adding KDF and PRF
1720 implementations. This includes an EVP_PKEY to EVP_KDF bridge for
1721 those algorithms that were already supported through the EVP_PKEY API
1722 (scrypt, TLS1 PRF and HKDF). The low-level KDF functions for PBKDF2
1723 and scrypt are now wrappers that call EVP_KDF.
1727 * Build devcrypto engine as a dynamic engine.
1729 *Eneas U de Queiroz*
1731 * Add keyed BLAKE2 to EVP_MAC.
1735 * Fix a bug in the computation of the endpoint-pair shared secret used
1736 by DTLS over SCTP. This breaks interoperability with older versions
1737 of OpenSSL like OpenSSL 1.1.0 and OpenSSL 1.0.2. There is a runtime
1738 switch SSL_MODE_DTLS_SCTP_LABEL_LENGTH_BUG (off by default) enabling
1739 interoperability with such broken implementations. However, enabling
1740 this switch breaks interoperability with correct implementations.
1742 * Fix a use after free bug in d2i_X509_PUBKEY when overwriting a
1743 re-used X509_PUBKEY object if the second PUBKEY is malformed.
1747 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
1751 * Changed the license to the Apache License v2.0.
1755 * Switch to a new version scheme using three numbers MAJOR.MINOR.PATCH.
1757 - Major releases (indicated by incrementing the MAJOR release number)
1758 may introduce incompatible API/ABI changes.
1759 - Minor releases (indicated by incrementing the MINOR release number)
1760 may introduce new features but retain API/ABI compatibility.
1761 - Patch releases (indicated by incrementing the PATCH number)
1762 are intended for bug fixes and other improvements of existing
1763 features only (like improving performance or adding documentation)
1764 and retain API/ABI compatibility.
1768 * Add support for RFC5297 SIV mode (siv128), including AES-SIV.
1772 * Remove the 'dist' target and add a tarball building script. The
1773 'dist' target has fallen out of use, and it shouldn't be
1774 necessary to configure just to create a source distribution.
1778 * Recreate the OS390-Unix config target. It no longer relies on a
1779 special script like it did for OpenSSL pre-1.1.0.
1783 * Instead of having the source directories listed in Configure, add
1784 a 'build.info' keyword SUBDIRS to indicate what sub-directories to
1789 * Add GMAC to EVP_MAC.
1793 * Ported the HMAC, CMAC and SipHash EVP_PKEY_METHODs to EVP_MAC.
1797 * Added EVP_MAC, an EVP layer MAC API, to simplify adding MAC
1798 implementations. This includes a generic EVP_PKEY to EVP_MAC bridge,
1799 to facilitate the continued use of MACs through raw private keys in
1800 functionality such as `EVP_DigestSign*` and `EVP_DigestVerify*`.
1804 * Deprecate ECDH_KDF_X9_62().
1808 * Added EVP_PKEY_ECDH_KDF_X9_63 and ecdh_KDF_X9_63() as replacements for
1809 the EVP_PKEY_ECDH_KDF_X9_62 KDF type and ECDH_KDF_X9_62(). The old names
1810 are retained for backwards compatibility.
1814 * AES-XTS mode now enforces that its two keys are different to mitigate
1815 the attacked described in "Efficient Instantiations of Tweakable
1816 Blockciphers and Refinements to Modes OCB and PMAC" by Phillip Rogaway.
1817 Details of this attack can be obtained from:
1818 <http://web.cs.ucdavis.edu/%7Erogaway/papers/offsets.pdf>
1822 * Rename the object files, i.e. give them other names than in previous
1823 versions. Their names now include the name of the final product, as
1824 well as its type mnemonic (bin, lib, shlib).
1828 * Added new option for 'openssl list', '-objects', which will display the
1829 list of built in objects, i.e. OIDs with names.
1833 * Added the options `-crl_lastupdate` and `-crl_nextupdate` to `openssl ca`,
1834 allowing the `lastUpdate` and `nextUpdate` fields in the generated CRL to
1839 * Added support for Linux Kernel TLS data-path. The Linux Kernel data-path
1840 improves application performance by removing data copies and providing
1841 applications with zero-copy system calls such as sendfile and splice.
1845 * The SSL option SSL_OP_CLEANSE_PLAINTEXT is introduced.
1849 * `PKCS12_parse` now maintains the order of the parsed certificates
1850 when outputting them via `*ca` (rather than reversing it).
1854 * Deprecated pthread fork support methods.
1858 * Added support for FFDHE key exchange in TLS 1.3.
1862 * Added a new concept for OpenSSL plugability: providers. This
1863 functionality is designed to replace the ENGINE API and ENGINE
1864 implementations, and to be much more dynamic, allowing provider
1865 authors to introduce new algorithms among other things, as long as
1866 there's an API that supports the algorithm type.
1868 With this concept comes a new core API for interaction between
1869 libcrypto and provider implementations. Public libcrypto functions
1870 that want to use providers do so through this core API.
1872 The main documentation for this core API is found in
1873 doc/man7/provider.pod, doc/man7/provider-base.pod, and they in turn
1874 refer to other manuals describing the API specific for supported
1875 algorithm types (also called operations).
1882 ### Changes between 1.1.1m and 1.1.1n [xx XXX xxxx]
1884 ### Changes between 1.1.1l and 1.1.1m [14 Dec 2021]
1886 * Avoid loading of a dynamic engine twice.
1890 * Prioritise DANE TLSA issuer certs over peer certs
1894 * Fixed random API for MacOS prior to 10.12
1896 These MacOS versions don't support the CommonCrypto APIs
1900 ### Changes between 1.1.1k and 1.1.1l [24 Aug 2021]
1902 * Fixed an SM2 Decryption Buffer Overflow.
1904 In order to decrypt SM2 encrypted data an application is expected to
1905 call the API function EVP_PKEY_decrypt(). Typically an application will
1906 call this function twice. The first time, on entry, the "out" parameter
1907 can be NULL and, on exit, the "outlen" parameter is populated with the
1908 buffer size required to hold the decrypted plaintext. The application
1909 can then allocate a sufficiently sized buffer and call EVP_PKEY_decrypt()
1910 again, but this time passing a non-NULL value for the "out" parameter.
1912 A bug in the implementation of the SM2 decryption code means that the
1913 calculation of the buffer size required to hold the plaintext returned
1914 by the first call to EVP_PKEY_decrypt() can be smaller than the actual
1915 size required by the second call. This can lead to a buffer overflow
1916 when EVP_PKEY_decrypt() is called by the application a second time with
1917 a buffer that is too small.
1919 A malicious attacker who is able present SM2 content for decryption to
1920 an application could cause attacker chosen data to overflow the buffer
1921 by up to a maximum of 62 bytes altering the contents of other data held
1922 after the buffer, possibly changing application behaviour or causing
1923 the application to crash. The location of the buffer is application
1924 dependent but is typically heap allocated.
1929 * Fixed various read buffer overruns processing ASN.1 strings
1931 ASN.1 strings are represented internally within OpenSSL as an ASN1_STRING
1932 structure which contains a buffer holding the string data and a field
1933 holding the buffer length. This contrasts with normal C strings which
1934 are repesented as a buffer for the string data which is terminated
1935 with a NUL (0) byte.
1937 Although not a strict requirement, ASN.1 strings that are parsed using
1938 OpenSSL's own "d2i" functions (and other similar parsing functions) as
1939 well as any string whose value has been set with the ASN1_STRING_set()
1940 function will additionally NUL terminate the byte array in the
1941 ASN1_STRING structure.
1943 However, it is possible for applications to directly construct valid
1944 ASN1_STRING structures which do not NUL terminate the byte array by
1945 directly setting the "data" and "length" fields in the ASN1_STRING
1946 array. This can also happen by using the ASN1_STRING_set0() function.
1948 Numerous OpenSSL functions that print ASN.1 data have been found to
1949 assume that the ASN1_STRING byte array will be NUL terminated, even
1950 though this is not guaranteed for strings that have been directly
1951 constructed. Where an application requests an ASN.1 structure to be
1952 printed, and where that ASN.1 structure contains ASN1_STRINGs that have
1953 been directly constructed by the application without NUL terminating
1954 the "data" field, then a read buffer overrun can occur.
1956 The same thing can also occur during name constraints processing
1957 of certificates (for example if a certificate has been directly
1958 constructed by the application instead of loading it via the OpenSSL
1959 parsing functions, and the certificate contains non NUL terminated
1960 ASN1_STRING structures). It can also occur in the X509_get1_email(),
1961 X509_REQ_get1_email() and X509_get1_ocsp() functions.
1963 If a malicious actor can cause an application to directly construct an
1964 ASN1_STRING and then process it through one of the affected OpenSSL
1965 functions then this issue could be hit. This might result in a crash
1966 (causing a Denial of Service attack). It could also result in the
1967 disclosure of private memory contents (such as private keys, or
1968 sensitive plaintext).
1973 ### Changes between 1.1.1j and 1.1.1k [25 Mar 2021]
1975 * Fixed a problem with verifying a certificate chain when using the
1976 X509_V_FLAG_X509_STRICT flag. This flag enables additional security checks of
1977 the certificates present in a certificate chain. It is not set by default.
1979 Starting from OpenSSL version 1.1.1h a check to disallow certificates in
1980 the chain that have explicitly encoded elliptic curve parameters was added
1981 as an additional strict check.
1983 An error in the implementation of this check meant that the result of a
1984 previous check to confirm that certificates in the chain are valid CA
1985 certificates was overwritten. This effectively bypasses the check
1986 that non-CA certificates must not be able to issue other certificates.
1988 If a "purpose" has been configured then there is a subsequent opportunity
1989 for checks that the certificate is a valid CA. All of the named "purpose"
1990 values implemented in libcrypto perform this check. Therefore, where
1991 a purpose is set the certificate chain will still be rejected even when the
1992 strict flag has been used. A purpose is set by default in libssl client and
1993 server certificate verification routines, but it can be overridden or
1994 removed by an application.
1996 In order to be affected, an application must explicitly set the
1997 X509_V_FLAG_X509_STRICT verification flag and either not set a purpose
1998 for the certificate verification or, in the case of TLS client or server
1999 applications, override the default purpose.
2004 * Fixed an issue where an OpenSSL TLS server may crash if sent a maliciously
2005 crafted renegotiation ClientHello message from a client. If a TLSv1.2
2006 renegotiation ClientHello omits the signature_algorithms extension (where it
2007 was present in the initial ClientHello), but includes a
2008 signature_algorithms_cert extension then a NULL pointer dereference will
2009 result, leading to a crash and a denial of service attack.
2011 A server is only vulnerable if it has TLSv1.2 and renegotiation enabled
2012 (which is the default configuration). OpenSSL TLS clients are not impacted by
2016 *Peter Kästle and Samuel Sapalski*
2018 ### Changes between 1.1.1i and 1.1.1j [16 Feb 2021]
2020 * Fixed the X509_issuer_and_serial_hash() function. It attempts to
2021 create a unique hash value based on the issuer and serial number data
2022 contained within an X509 certificate. However it was failing to correctly
2023 handle any errors that may occur while parsing the issuer field (which might
2024 occur if the issuer field is maliciously constructed). This may subsequently
2025 result in a NULL pointer deref and a crash leading to a potential denial of
2031 * Fixed the RSA_padding_check_SSLv23() function and the RSA_SSLV23_PADDING
2032 padding mode to correctly check for rollback attacks. This is considered a
2033 bug in OpenSSL 1.1.1 because it does not support SSLv2. In 1.0.2 this is
2038 Fixed the EVP_CipherUpdate, EVP_EncryptUpdate and EVP_DecryptUpdate
2039 functions. Previously they could overflow the output length argument in some
2040 cases where the input length is close to the maximum permissable length for
2041 an integer on the platform. In such cases the return value from the function
2042 call would be 1 (indicating success), but the output length value would be
2043 negative. This could cause applications to behave incorrectly or crash.
2048 * Fixed SRP_Calc_client_key so that it runs in constant time. The previous
2049 implementation called BN_mod_exp without setting BN_FLG_CONSTTIME. This
2050 could be exploited in a side channel attack to recover the password. Since
2051 the attack is local host only this is outside of the current OpenSSL
2052 threat model and therefore no CVE is assigned.
2054 Thanks to Mohammed Sabt and Daniel De Almeida Braga for reporting this
2059 ### Changes between 1.1.1h and 1.1.1i [8 Dec 2020]
2061 * Fixed NULL pointer deref in the GENERAL_NAME_cmp function
2062 This function could crash if both GENERAL_NAMEs contain an EDIPARTYNAME.
2063 If an attacker can control both items being compared then this could lead
2064 to a possible denial of service attack. OpenSSL itself uses the
2065 GENERAL_NAME_cmp function for two purposes:
2066 1) Comparing CRL distribution point names between an available CRL and a
2067 CRL distribution point embedded in an X509 certificate
2068 2) When verifying that a timestamp response token signer matches the
2069 timestamp authority name (exposed via the API functions
2070 TS_RESP_verify_response and TS_RESP_verify_token)
2075 ### Changes between 1.1.1g and 1.1.1h [22 Sep 2020]
2077 * Certificates with explicit curve parameters are now disallowed in
2078 verification chains if the X509_V_FLAG_X509_STRICT flag is used.
2082 * The 'MinProtocol' and 'MaxProtocol' configuration commands now silently
2083 ignore TLS protocol version bounds when configuring DTLS-based contexts, and
2084 conversely, silently ignore DTLS protocol version bounds when configuring
2085 TLS-based contexts. The commands can be repeated to set bounds of both
2086 types. The same applies with the corresponding "min_protocol" and
2087 "max_protocol" command-line switches, in case some application uses both TLS
2090 SSL_CTX instances that are created for a fixed protocol version (e.g.
2091 TLSv1_server_method()) also silently ignore version bounds. Previously
2092 attempts to apply bounds to these protocol versions would result in an
2093 error. Now only the "version-flexible" SSL_CTX instances are subject to
2094 limits in configuration files in command-line options.
2098 * Handshake now fails if Extended Master Secret extension is dropped
2103 * The Oracle Developer Studio compiler will start reporting deprecated APIs
2105 ### Changes between 1.1.1f and 1.1.1g [21 Apr 2020]
2107 * Fixed segmentation fault in SSL_check_chain()
2108 Server or client applications that call the SSL_check_chain() function
2109 during or after a TLS 1.3 handshake may crash due to a NULL pointer
2110 dereference as a result of incorrect handling of the
2111 "signature_algorithms_cert" TLS extension. The crash occurs if an invalid
2112 or unrecognised signature algorithm is received from the peer. This could
2113 be exploited by a malicious peer in a Denial of Service attack.
2118 * Added AES consttime code for no-asm configurations
2119 an optional constant time support for AES was added
2120 when building openssl for no-asm.
2121 Enable with: ./config no-asm -DOPENSSL_AES_CONST_TIME
2122 Disable with: ./config no-asm -DOPENSSL_NO_AES_CONST_TIME
2123 At this time this feature is by default disabled.
2124 It will be enabled by default in 3.0.
2128 ### Changes between 1.1.1e and 1.1.1f [31 Mar 2020]
2130 * Revert the change of EOF detection while reading in libssl to avoid
2131 regressions in applications depending on the current way of reporting
2132 the EOF. As the existing method is not fully accurate the change to
2133 reporting the EOF via SSL_ERROR_SSL is kept on the current development
2134 branch and will be present in the 3.0 release.
2138 * Revised BN_generate_prime_ex to not avoid factors 3..17863 in p-1
2139 when primes for RSA keys are computed.
2140 Since we previously always generated primes == 2 (mod 3) for RSA keys,
2141 the 2-prime and 3-prime RSA modules were easy to distinguish, since
2142 N = p*q = 1 (mod 3), but N = p*q*r = 2 (mod 3). Therefore fingerprinting
2143 2-prime vs. 3-prime RSA keys was possible by computing N mod 3.
2144 This avoids possible fingerprinting of newly generated RSA modules.
2148 ### Changes between 1.1.1d and 1.1.1e [17 Mar 2020]
2150 * Properly detect EOF while reading in libssl. Previously if we hit an EOF
2151 while reading in libssl then we would report an error back to the
2152 application (SSL_ERROR_SYSCALL) but errno would be 0. We now add
2153 an error to the stack (which means we instead return SSL_ERROR_SSL) and
2154 therefore give a hint as to what went wrong.
2158 * Check that ed25519 and ed448 are allowed by the security level. Previously
2159 signature algorithms not using an MD were not being checked that they were
2160 allowed by the security level.
2164 * Fixed SSL_get_servername() behaviour. The behaviour of SSL_get_servername()
2165 was not quite right. The behaviour was not consistent between resumption
2166 and normal handshakes, and also not quite consistent with historical
2167 behaviour. The behaviour in various scenarios has been clarified and
2168 it has been updated to make it match historical behaviour as closely as
2173 * *[VMS only]* The header files that the VMS compilers include automatically,
2174 `__DECC_INCLUDE_PROLOGUE.H` and `__DECC_INCLUDE_EPILOGUE.H`, use pragmas
2175 that the C++ compiler doesn't understand. This is a shortcoming in the
2176 compiler, but can be worked around with `__cplusplus` guards.
2178 C++ applications that use OpenSSL libraries must be compiled using the
2179 qualifier `/NAMES=(AS_IS,SHORTENED)` to be able to use all the OpenSSL
2180 functions. Otherwise, only functions with symbols of less than 31
2181 characters can be used, as the linker will not be able to successfully
2182 resolve symbols with longer names.
2186 * Added a new method to gather entropy on VMS, based on SYS$GET_ENTROPY.
2187 The presence of this system service is determined at run-time.
2191 * Print all values for a PKCS#12 attribute with 'openssl pkcs12', not just
2196 ### Changes between 1.1.1c and 1.1.1d [10 Sep 2019]
2198 * Fixed a fork protection issue. OpenSSL 1.1.1 introduced a rewritten random
2199 number generator (RNG). This was intended to include protection in the
2200 event of a fork() system call in order to ensure that the parent and child
2201 processes did not share the same RNG state. However this protection was not
2202 being used in the default case.
2204 A partial mitigation for this issue is that the output from a high
2205 precision timer is mixed into the RNG state so the likelihood of a parent
2206 and child process sharing state is significantly reduced.
2208 If an application already calls OPENSSL_init_crypto() explicitly using
2209 OPENSSL_INIT_ATFORK then this problem does not occur at all.
2212 *Matthias St. Pierre*
2214 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
2215 used even when parsing explicit parameters, when loading a encoded key
2216 or calling `EC_GROUP_new_from_ecpkparameters()`/
2217 `EC_GROUP_new_from_ecparameters()`.
2218 This prevents bypass of security hardening and performance gains,
2219 especially for curves with specialized EC_METHODs.
2220 By default, if a key encoded with explicit parameters is loaded and later
2221 encoded, the output is still encoded with explicit parameters, even if
2222 internally a "named" EC_GROUP is used for computation.
2226 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
2227 this change, EC_GROUP_set_generator would accept order and/or cofactor as
2228 NULL. After this change, only the cofactor parameter can be NULL. It also
2229 does some minimal sanity checks on the passed order.
2234 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
2235 An attack is simple, if the first CMS_recipientInfo is valid but the
2236 second CMS_recipientInfo is chosen ciphertext. If the second
2237 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
2238 encryption key will be replaced by garbage, and the message cannot be
2239 decoded, but if the RSA decryption fails, the correct encryption key is
2240 used and the recipient will not notice the attack.
2241 As a work around for this potential attack the length of the decrypted
2242 key must be equal to the cipher default key length, in case the
2243 certifiate is not given and all recipientInfo are tried out.
2244 The old behaviour can be re-enabled in the CMS code by setting the
2245 CMS_DEBUG_DECRYPT flag.
2250 * Early start up entropy quality from the DEVRANDOM seed source has been
2251 improved for older Linux systems. The RAND subsystem will wait for
2252 /dev/random to be producing output before seeding from /dev/urandom.
2253 The seeded state is stored for future library initialisations using
2254 a system global shared memory segment. The shared memory identifier
2255 can be configured by defining OPENSSL_RAND_SEED_DEVRANDOM_SHM_ID to
2256 the desired value. The default identifier is 114.
2260 * Correct the extended master secret constant on EBCDIC systems. Without this
2261 fix TLS connections between an EBCDIC system and a non-EBCDIC system that
2262 negotiate EMS will fail. Unfortunately this also means that TLS connections
2263 between EBCDIC systems with this fix, and EBCDIC systems without this
2264 fix will fail if they negotiate EMS.
2268 * Use Windows installation paths in the mingw builds
2270 Mingw isn't a POSIX environment per se, which means that Windows
2271 paths should be used for installation.
2276 * Changed DH_check to accept parameters with order q and 2q subgroups.
2277 With order 2q subgroups the bit 0 of the private key is not secret
2278 but DH_generate_key works around that by clearing bit 0 of the
2279 private key for those. This avoids leaking bit 0 of the private key.
2283 * Significantly reduce secure memory usage by the randomness pools.
2287 * Revert the DEVRANDOM_WAIT feature for Linux systems
2289 The DEVRANDOM_WAIT feature added a select() call to wait for the
2290 /dev/random device to become readable before reading from the
2291 /dev/urandom device.
2293 It turned out that this change had negative side effects on
2294 performance which were not acceptable. After some discussion it
2295 was decided to revert this feature and leave it up to the OS
2296 resp. the platform maintainer to ensure a proper initialization
2297 during early boot time.
2299 *Matthias St. Pierre*
2301 ### Changes between 1.1.1b and 1.1.1c [28 May 2019]
2303 * Add build tests for C++. These are generated files that only do one
2304 thing, to include one public OpenSSL head file each. This tests that
2305 the public header files can be usefully included in a C++ application.
2307 This test isn't enabled by default. It can be enabled with the option
2308 'enable-buildtest-c++'.
2312 * Enable SHA3 pre-hashing for ECDSA and DSA.
2316 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
2317 This changes the size when using the `genpkey` command when no size is given.
2318 It fixes an omission in earlier changes that changed all RSA, DSA and DH
2319 generation commands to use 2048 bits by default.
2323 * Reorganize the manual pages to consistently have RETURN VALUES,
2324 EXAMPLES, SEE ALSO and HISTORY come in that order, and adjust
2325 util/fix-doc-nits accordingly.
2327 *Paul Yang, Joshua Lock*
2329 * Add the missing accessor EVP_PKEY_get0_engine()
2333 * Have commands like `s_client` and `s_server` output the signature scheme
2334 along with other cipher suite parameters when debugging.
2338 * Make OPENSSL_config() error agnostic again.
2342 * Do the error handling in RSA decryption constant time.
2346 * Prevent over long nonces in ChaCha20-Poly1305.
2348 ChaCha20-Poly1305 is an AEAD cipher, and requires a unique nonce input
2349 for every encryption operation. RFC 7539 specifies that the nonce value
2350 (IV) should be 96 bits (12 bytes). OpenSSL allows a variable nonce length
2351 and front pads the nonce with 0 bytes if it is less than 12
2352 bytes. However it also incorrectly allows a nonce to be set of up to 16
2353 bytes. In this case only the last 12 bytes are significant and any
2354 additional leading bytes are ignored.
2356 It is a requirement of using this cipher that nonce values are
2357 unique. Messages encrypted using a reused nonce value are susceptible to
2358 serious confidentiality and integrity attacks. If an application changes
2359 the default nonce length to be longer than 12 bytes and then makes a
2360 change to the leading bytes of the nonce expecting the new value to be a
2361 new unique nonce then such an application could inadvertently encrypt
2362 messages with a reused nonce.
2364 Additionally the ignored bytes in a long nonce are not covered by the
2365 integrity guarantee of this cipher. Any application that relies on the
2366 integrity of these ignored leading bytes of a long nonce may be further
2367 affected. Any OpenSSL internal use of this cipher, including in SSL/TLS,
2368 is safe because no such use sets such a long nonce value. However user
2369 applications that use this cipher directly and set a non-default nonce
2370 length to be longer than 12 bytes may be vulnerable.
2372 This issue was reported to OpenSSL on 16th of March 2019 by Joran Dirk
2378 * Add DEVRANDOM_WAIT feature for Linux systems
2380 On older Linux systems where the getrandom() system call is not available,
2381 OpenSSL normally uses the /dev/urandom device for seeding its CSPRNG.
2382 Contrary to getrandom(), the /dev/urandom device will not block during
2383 early boot when the kernel CSPRNG has not been seeded yet.
2385 To mitigate this known weakness, use select() to wait for /dev/random to
2386 become readable before reading from /dev/urandom.
2388 * Ensure that SM2 only uses SM3 as digest algorithm
2392 ### Changes between 1.1.1a and 1.1.1b [26 Feb 2019]
2394 * Change the info callback signals for the start and end of a post-handshake
2395 message exchange in TLSv1.3. In 1.1.1/1.1.1a we used SSL_CB_HANDSHAKE_START
2396 and SSL_CB_HANDSHAKE_DONE. Experience has shown that many applications get
2397 confused by this and assume that a TLSv1.2 renegotiation has started. This
2398 can break KeyUpdate handling. Instead we no longer signal the start and end
2399 of a post handshake message exchange (although the messages themselves are
2400 still signalled). This could break some applications that were expecting
2401 the old signals. However without this KeyUpdate is not usable for many
2406 ### Changes between 1.1.1 and 1.1.1a [20 Nov 2018]
2408 * Timing vulnerability in DSA signature generation
2410 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
2411 timing side channel attack. An attacker could use variations in the signing
2412 algorithm to recover the private key.
2414 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
2419 * Timing vulnerability in ECDSA signature generation
2421 The OpenSSL ECDSA signature algorithm has been shown to be vulnerable to a
2422 timing side channel attack. An attacker could use variations in the signing
2423 algorithm to recover the private key.
2425 This issue was reported to OpenSSL on 25th October 2018 by Samuel Weiser.
2430 * Fixed the issue that RAND_add()/RAND_seed() silently discards random input
2431 if its length exceeds 4096 bytes. The limit has been raised to a buffer size
2432 of two gigabytes and the error handling improved.
2434 This issue was reported to OpenSSL by Dr. Falko Strenzke. It has been
2435 categorized as a normal bug, not a security issue, because the DRBG reseeds
2436 automatically and is fully functional even without additional randomness
2437 provided by the application.
2439 ### Changes between 1.1.0i and 1.1.1 [11 Sep 2018]
2441 * Add a new ClientHello callback. Provides a callback interface that gives
2442 the application the ability to adjust the nascent SSL object at the
2443 earliest stage of ClientHello processing, immediately after extensions have
2444 been collected but before they have been processed. In particular, this
2445 callback can adjust the supported TLS versions in response to the contents
2450 * Add SM2 base algorithm support.
2454 * s390x assembly pack: add (improved) hardware-support for the following
2455 cryptographic primitives: sha3, shake, aes-gcm, aes-ccm, aes-ctr, aes-ofb,
2456 aes-cfb/cfb8, aes-ecb.
2460 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
2461 parameter is no longer accepted, as it leads to a corrupt table. NULL
2462 pem_str is reserved for alias entries only.
2466 * Use the new ec_scalar_mul_ladder scaffold to implement a specialized ladder
2467 step for prime curves. The new implementation is based on formulae from
2468 differential addition-and-doubling in homogeneous projective coordinates
2469 from Izu-Takagi "A fast parallel elliptic curve multiplication resistant
2470 against side channel attacks" and Brier-Joye "Weierstrass Elliptic Curves
2471 and Side-Channel Attacks" Eq. (8) for y-coordinate recovery, modified
2472 to work in projective coordinates.
2474 *Billy Bob Brumley, Nicola Tuveri*
2476 * Change generating and checking of primes so that the error rate of not
2477 being prime depends on the intended use based on the size of the input.
2478 For larger primes this will result in more rounds of Miller-Rabin.
2479 The maximal error rate for primes with more than 1080 bits is lowered
2482 *Kurt Roeckx, Annie Yousar*
2484 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
2488 * The 'tsget' script is renamed to 'tsget.pl', to avoid confusion when
2489 moving between systems, and to avoid confusion when a Windows build is
2490 done with mingw vs with MSVC. For POSIX installs, there's still a
2491 symlink or copy named 'tsget' to avoid that confusion as well.
2495 * Revert blinding in ECDSA sign and instead make problematic addition
2496 length-invariant. Switch even to fixed-length Montgomery multiplication.
2500 * Use the new ec_scalar_mul_ladder scaffold to implement a specialized ladder
2501 step for binary curves. The new implementation is based on formulae from
2502 differential addition-and-doubling in mixed Lopez-Dahab projective
2503 coordinates, modified to independently blind the operands.
2505 *Billy Bob Brumley, Sohaib ul Hassan, Nicola Tuveri*
2507 * Add a scaffold to optionally enhance the Montgomery ladder implementation
2508 for `ec_scalar_mul_ladder` (formerly `ec_mul_consttime`) allowing
2509 EC_METHODs to implement their own specialized "ladder step", to take
2510 advantage of more favorable coordinate systems or more efficient
2511 differential addition-and-doubling algorithms.
2513 *Billy Bob Brumley, Sohaib ul Hassan, Nicola Tuveri*
2515 * Modified the random device based seed sources to keep the relevant
2516 file descriptors open rather than reopening them on each access.
2517 This allows such sources to operate in a chroot() jail without
2518 the associated device nodes being available. This behaviour can be
2519 controlled using RAND_keep_random_devices_open().
2523 * Numerous side-channel attack mitigations have been applied. This may have
2524 performance impacts for some algorithms for the benefit of improved
2525 security. Specific changes are noted in this change log by their respective
2530 * AIX shared library support overhaul. Switch to AIX "natural" way of
2531 handling shared libraries, which means collecting shared objects of
2532 different versions and bitnesses in one common archive. This allows to
2533 mitigate conflict between 1.0 and 1.1 side-by-side installations. It
2534 doesn't affect the way 3rd party applications are linked, only how
2535 multi-version installation is managed.
2539 * Make ec_group_do_inverse_ord() more robust and available to other
2540 EC cryptosystems, so that irrespective of BN_FLG_CONSTTIME, SCA
2541 mitigations are applied to the fallback BN_mod_inverse().
2542 When using this function rather than BN_mod_inverse() directly, new
2543 EC cryptosystem implementations are then safer-by-default.
2547 * Add coordinate blinding for EC_POINT and implement projective
2548 coordinate blinding for generic prime curves as a countermeasure to
2549 chosen point SCA attacks.
2551 *Sohaib ul Hassan, Nicola Tuveri, Billy Bob Brumley*
2553 * Add blinding to ECDSA and DSA signatures to protect against side channel
2554 attacks discovered by Keegan Ryan (NCC Group).
2558 * Enforce checking in the `pkeyutl` command to ensure that the input
2559 length does not exceed the maximum supported digest length when performing
2560 a sign, verify or verifyrecover operation.
2564 * SSL_MODE_AUTO_RETRY is enabled by default. Applications that use blocking
2565 I/O in combination with something like select() or poll() will hang. This
2566 can be turned off again using SSL_CTX_clear_mode().
2567 Many applications do not properly handle non-application data records, and
2568 TLS 1.3 sends more of such records. Setting SSL_MODE_AUTO_RETRY works
2569 around the problems in those applications, but can also break some.
2570 It's recommended to read the manpages about SSL_read(), SSL_write(),
2571 SSL_get_error(), SSL_shutdown(), SSL_CTX_set_mode() and
2572 SSL_CTX_set_read_ahead() again.
2576 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
2577 now allow empty (zero character) pass phrases.
2581 * Apply blinding to binary field modular inversion and remove patent
2582 pending (OPENSSL_SUN_GF2M_DIV) BN_GF2m_mod_div implementation.
2586 * Deprecate ec2_mult.c and unify scalar multiplication code paths for
2587 binary and prime elliptic curves.
2591 * Remove ECDSA nonce padding: EC_POINT_mul is now responsible for
2592 constant time fixed point multiplication.
2596 * Revise elliptic curve scalar multiplication with timing attack
2597 defenses: ec_wNAF_mul redirects to a constant time implementation
2598 when computing fixed point and variable point multiplication (which
2599 in OpenSSL are mostly used with secret scalars in keygen, sign,
2600 ECDH derive operations).
2601 *Billy Bob Brumley, Nicola Tuveri, Cesar Pereida García,
2604 * Updated CONTRIBUTING
2608 * Updated DRBG / RAND to request nonce and additional low entropy
2609 randomness from the system.
2611 *Matthias St. Pierre*
2613 * Updated 'openssl rehash' to use OpenSSL consistent default.
2617 * Moved the load of the ssl_conf module to libcrypto, which helps
2618 loading engines that libssl uses before libssl is initialised.
2622 * Added EVP_PKEY_sign() and EVP_PKEY_verify() for EdDSA
2626 * Fixed X509_NAME_ENTRY_set to get multi-valued RDNs right in all cases.
2628 *Ingo Schwarze, Rich Salz*
2630 * Added output of accepting IP address and port for 'openssl s_server'
2634 * Added a new API for TLSv1.3 ciphersuites:
2635 SSL_CTX_set_ciphersuites()
2636 SSL_set_ciphersuites()
2640 * Memory allocation failures consistently add an error to the error
2645 * Don't use OPENSSL_ENGINES and OPENSSL_CONF environment values
2646 in libcrypto when run as setuid/setgid.
2650 * Load any config file by default when libssl is used.
2654 * Added new public header file <openssl/rand_drbg.h> and documentation
2655 for the RAND_DRBG API. See manual page RAND_DRBG(7) for an overview.
2657 *Matthias St. Pierre*
2659 * QNX support removed (cannot find contributors to get their approval
2660 for the license change).
2664 * TLSv1.3 replay protection for early data has been implemented. See the
2665 SSL_read_early_data() man page for further details.
2669 * Separated TLSv1.3 ciphersuite configuration out from TLSv1.2 ciphersuite
2670 configuration. TLSv1.3 ciphersuites are not compatible with TLSv1.2 and
2671 below. Similarly TLSv1.2 ciphersuites are not compatible with TLSv1.3.
2672 In order to avoid issues where legacy TLSv1.2 ciphersuite configuration
2673 would otherwise inadvertently disable all TLSv1.3 ciphersuites the
2674 configuration has been separated out. See the ciphers man page or the
2675 SSL_CTX_set_ciphersuites() man page for more information.
2679 * On POSIX (BSD, Linux, ...) systems the ocsp(1) command running
2680 in responder mode now supports the new "-multi" option, which
2681 spawns the specified number of child processes to handle OCSP
2682 requests. The "-timeout" option now also limits the OCSP
2683 responder's patience to wait to receive the full client request
2684 on a newly accepted connection. Child processes are respawned
2685 as needed, and the CA index file is automatically reloaded
2686 when changed. This makes it possible to run the "ocsp" responder
2687 as a long-running service, making the OpenSSL CA somewhat more
2688 feature-complete. In this mode, most diagnostic messages logged
2689 after entering the event loop are logged via syslog(3) rather than
2694 * Added support for X448 and Ed448. Heavily based on original work by
2699 * Extend OSSL_STORE with capabilities to search and to narrow the set of
2700 objects loaded. This adds the functions OSSL_STORE_expect() and
2701 OSSL_STORE_find() as well as needed tools to construct searches and
2702 get the search data out of them.
2706 * Support for TLSv1.3 added. Note that users upgrading from an earlier
2707 version of OpenSSL should review their configuration settings to ensure
2708 that they are still appropriate for TLSv1.3. For further information see:
2709 <https://wiki.openssl.org/index.php/TLS1.3>
2713 * Grand redesign of the OpenSSL random generator
2715 The default RAND method now utilizes an AES-CTR DRBG according to
2716 NIST standard SP 800-90Ar1. The new random generator is essentially
2717 a port of the default random generator from the OpenSSL FIPS 2.0
2718 object module. It is a hybrid deterministic random bit generator
2719 using an AES-CTR bit stream and which seeds and reseeds itself
2720 automatically using trusted system entropy sources.
2722 Some of its new features are:
2723 - Support for multiple DRBG instances with seed chaining.
2724 - The default RAND method makes use of a DRBG.
2725 - There is a public and private DRBG instance.
2726 - The DRBG instances are fork-safe.
2727 - Keep all global DRBG instances on the secure heap if it is enabled.
2728 - The public and private DRBG instance are per thread for lock free
2731 *Paul Dale, Benjamin Kaduk, Kurt Roeckx, Rich Salz, Matthias St. Pierre*
2733 * Changed Configure so it only says what it does and doesn't dump
2734 so much data. Instead, ./configdata.pm should be used as a script
2735 to display all sorts of configuration data.
2739 * Added processing of "make variables" to Configure.
2743 * Added SHA512/224 and SHA512/256 algorithm support.
2747 * The last traces of Netware support, first removed in 1.1.0, have
2752 * Get rid of Makefile.shared, and in the process, make the processing
2753 of certain files (rc.obj, or the .def/.map/.opt files produced from
2754 the ordinal files) more visible and hopefully easier to trace and
2755 debug (or make silent).
2759 * Make it possible to have environment variable assignments as
2760 arguments to config / Configure.
2764 * Add multi-prime RSA (RFC 8017) support.
2768 * Add SM3 implemented according to GB/T 32905-2016
2769 *Jack Lloyd <jack.lloyd@ribose.com>,*
2770 *Ronald Tse <ronald.tse@ribose.com>,*
2771 *Erick Borsboom <erick.borsboom@ribose.com>*
2773 * Add 'Maximum Fragment Length' TLS extension negotiation and support
2774 as documented in RFC6066.
2775 Based on a patch from Tomasz Moń
2777 *Filipe Raimundo da Silva*
2779 * Add SM4 implemented according to GB/T 32907-2016.
2780 *Jack Lloyd <jack.lloyd@ribose.com>,*
2781 *Ronald Tse <ronald.tse@ribose.com>,*
2782 *Erick Borsboom <erick.borsboom@ribose.com>*
2784 * Reimplement -newreq-nodes and ERR_error_string_n; the
2785 original author does not agree with the license change.
2789 * Add ARIA AEAD TLS support.
2793 * Some macro definitions to support VS6 have been removed. Visual
2794 Studio 6 has not worked since 1.1.0
2798 * Add ERR_clear_last_mark(), to allow callers to clear the last mark
2799 without clearing the errors.
2803 * Add "atfork" functions. If building on a system that without
2804 pthreads, see doc/man3/OPENSSL_fork_prepare.pod for application
2805 requirements. The RAND facility now uses/requires this.
2813 * The UI API becomes a permanent and integral part of libcrypto, i.e.
2814 not possible to disable entirely. However, it's still possible to
2815 disable the console reading UI method, UI_OpenSSL() (use UI_null()
2818 To disable, configure with 'no-ui-console'. 'no-ui' is still
2819 possible to use as an alias. Check at compile time with the
2820 macro OPENSSL_NO_UI_CONSOLE. The macro OPENSSL_NO_UI is still
2821 possible to check and is an alias for OPENSSL_NO_UI_CONSOLE.
2825 * Add a STORE module, which implements a uniform and URI based reader of
2826 stores that can contain keys, certificates, CRLs and numerous other
2827 objects. The main API is loosely based on a few stdio functions,
2828 and includes OSSL_STORE_open, OSSL_STORE_load, OSSL_STORE_eof,
2829 OSSL_STORE_error and OSSL_STORE_close.
2830 The implementation uses backends called "loaders" to implement arbitrary
2831 URI schemes. There is one built in "loader" for the 'file' scheme.
2835 * Add devcrypto engine. This has been implemented against cryptodev-linux,
2836 then adjusted to work on FreeBSD 8.4 as well.
2837 Enable by configuring with 'enable-devcryptoeng'. This is done by default
2838 on BSD implementations, as cryptodev.h is assumed to exist on all of them.
2842 * Module names can prefixed with OSSL_ or OPENSSL_. This affects
2843 util/mkerr.pl, which is adapted to allow those prefixes, leading to
2844 error code calls like this:
2846 OSSL_FOOerr(OSSL_FOO_F_SOMETHING, OSSL_FOO_R_WHATEVER);
2848 With this change, we claim the namespaces OSSL and OPENSSL in a manner
2849 that can be encoded in C. For the foreseeable future, this will only
2852 *Richard Levitte and Tim Hudson*
2854 * Removed BSD cryptodev engine.
2858 * Add a build target 'build_all_generated', to build all generated files
2859 and only that. This can be used to prepare everything that requires
2860 things like perl for a system that lacks perl and then move everything
2861 to that system and do the rest of the build there.
2865 * In the UI interface, make it possible to duplicate the user data. This
2866 can be used by engines that need to retain the data for a longer time
2867 than just the call where this user data is passed.
2871 * Ignore the '-named_curve auto' value for compatibility of applications
2874 *Tomáš Mráz <tmraz@fedoraproject.org>*
2876 * Fragmented SSL/TLS alerts are no longer accepted. An alert message is 2
2877 bytes long. In theory it is permissible in SSLv3 - TLSv1.2 to fragment such
2878 alerts across multiple records (some of which could be empty). In practice
2879 it make no sense to send an empty alert record, or to fragment one. TLSv1.3
2880 prohibits this altogether and other libraries (BoringSSL, NSS) do not
2881 support this at all. Supporting it adds significant complexity to the
2882 record layer, and its removal is unlikely to cause interoperability
2887 * Add the ASN.1 types INT32, UINT32, INT64, UINT64 and variants prefixed
2888 with Z. These are meant to replace LONG and ZLONG and to be size safe.
2889 The use of LONG and ZLONG is discouraged and scheduled for deprecation
2894 * Add the 'z' and 'j' modifiers to BIO_printf() et al formatting string,
2895 'z' is to be used for [s]size_t, and 'j' - with [u]int64_t.
2897 *Richard Levitte, Andy Polyakov*
2899 * Add EC_KEY_get0_engine(), which does for EC_KEY what RSA_get0_engine()
2904 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
2905 platform rather than 'mingw'.
2909 * The functions X509_STORE_add_cert and X509_STORE_add_crl return
2910 success if they are asked to add an object which already exists
2911 in the store. This change cascades to other functions which load
2912 certificates and CRLs.
2916 * x86_64 assembly pack: annotate code with DWARF CFI directives to
2917 facilitate stack unwinding even from assembly subroutines.
2921 * Remove VAX C specific definitions of OPENSSL_EXPORT, OPENSSL_EXTERN.
2922 Also remove OPENSSL_GLOBAL entirely, as it became a no-op.
2926 * Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
2927 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
2928 which is the minimum version we support.
2932 * Certificate time validation (X509_cmp_time) enforces stricter
2933 compliance with RFC 5280. Fractional seconds and timezone offsets
2934 are no longer allowed.
2938 * Add support for ARIA
2942 * s_client will now send the Server Name Indication (SNI) extension by
2943 default unless the new "-noservername" option is used. The server name is
2944 based on the host provided to the "-connect" option unless overridden by
2945 using "-servername".
2949 * Add support for SipHash
2953 * OpenSSL now fails if it receives an unrecognised record type in TLS1.0
2954 or TLS1.1. Previously this only happened in SSLv3 and TLS1.2. This is to
2955 prevent issues where no progress is being made and the peer continually
2956 sends unrecognised record types, using up resources processing them.
2960 * 'openssl passwd' can now produce SHA256 and SHA512 based output,
2961 using the algorithm defined in
2962 <https://www.akkadia.org/drepper/SHA-crypt.txt>
2966 * Heartbeat support has been removed; the ABI is changed for now.
2968 *Richard Levitte, Rich Salz*
2970 * Support for SSL_OP_NO_ENCRYPT_THEN_MAC in SSL_CONF_cmd.
2974 * The RSA "null" method, which was partially supported to avoid patent
2975 issues, has been replaced to always returns NULL.
2982 ### Changes between 1.1.0k and 1.1.0l [10 Sep 2019]
2984 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
2985 used even when parsing explicit parameters, when loading a encoded key
2986 or calling `EC_GROUP_new_from_ecpkparameters()`/
2987 `EC_GROUP_new_from_ecparameters()`.
2988 This prevents bypass of security hardening and performance gains,
2989 especially for curves with specialized EC_METHODs.
2990 By default, if a key encoded with explicit parameters is loaded and later
2991 encoded, the output is still encoded with explicit parameters, even if
2992 internally a "named" EC_GROUP is used for computation.
2996 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
2997 this change, EC_GROUP_set_generator would accept order and/or cofactor as
2998 NULL. After this change, only the cofactor parameter can be NULL. It also
2999 does some minimal sanity checks on the passed order.
3004 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
3005 An attack is simple, if the first CMS_recipientInfo is valid but the
3006 second CMS_recipientInfo is chosen ciphertext. If the second
3007 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
3008 encryption key will be replaced by garbage, and the message cannot be
3009 decoded, but if the RSA decryption fails, the correct encryption key is
3010 used and the recipient will not notice the attack.
3011 As a work around for this potential attack the length of the decrypted
3012 key must be equal to the cipher default key length, in case the
3013 certifiate is not given and all recipientInfo are tried out.
3014 The old behaviour can be re-enabled in the CMS code by setting the
3015 CMS_DEBUG_DECRYPT flag.
3020 * Use Windows installation paths in the mingw builds
3022 Mingw isn't a POSIX environment per se, which means that Windows
3023 paths should be used for installation.
3028 ### Changes between 1.1.0j and 1.1.0k [28 May 2019]
3030 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
3031 This changes the size when using the `genpkey` command when no size is given.
3032 It fixes an omission in earlier changes that changed all RSA, DSA and DH
3033 generation commands to use 2048 bits by default.
3037 * Prevent over long nonces in ChaCha20-Poly1305.
3039 ChaCha20-Poly1305 is an AEAD cipher, and requires a unique nonce input
3040 for every encryption operation. RFC 7539 specifies that the nonce value
3041 (IV) should be 96 bits (12 bytes). OpenSSL allows a variable nonce length
3042 and front pads the nonce with 0 bytes if it is less than 12
3043 bytes. However it also incorrectly allows a nonce to be set of up to 16
3044 bytes. In this case only the last 12 bytes are significant and any
3045 additional leading bytes are ignored.
3047 It is a requirement of using this cipher that nonce values are
3048 unique. Messages encrypted using a reused nonce value are susceptible to
3049 serious confidentiality and integrity attacks. If an application changes
3050 the default nonce length to be longer than 12 bytes and then makes a
3051 change to the leading bytes of the nonce expecting the new value to be a
3052 new unique nonce then such an application could inadvertently encrypt
3053 messages with a reused nonce.
3055 Additionally the ignored bytes in a long nonce are not covered by the
3056 integrity guarantee of this cipher. Any application that relies on the
3057 integrity of these ignored leading bytes of a long nonce may be further
3058 affected. Any OpenSSL internal use of this cipher, including in SSL/TLS,
3059 is safe because no such use sets such a long nonce value. However user
3060 applications that use this cipher directly and set a non-default nonce
3061 length to be longer than 12 bytes may be vulnerable.
3063 This issue was reported to OpenSSL on 16th of March 2019 by Joran Dirk
3069 * Added SCA hardening for modular field inversion in EC_GROUP through
3070 a new dedicated field_inv() pointer in EC_METHOD.
3071 This also addresses a leakage affecting conversions from projective
3072 to affine coordinates.
3074 *Billy Bob Brumley, Nicola Tuveri*
3076 * Fix a use after free bug in d2i_X509_PUBKEY when overwriting a
3077 re-used X509_PUBKEY object if the second PUBKEY is malformed.
3081 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
3085 * Remove the 'dist' target and add a tarball building script. The
3086 'dist' target has fallen out of use, and it shouldn't be
3087 necessary to configure just to create a source distribution.
3091 ### Changes between 1.1.0i and 1.1.0j [20 Nov 2018]
3093 * Timing vulnerability in DSA signature generation
3095 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
3096 timing side channel attack. An attacker could use variations in the signing
3097 algorithm to recover the private key.
3099 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
3104 * Timing vulnerability in ECDSA signature generation
3106 The OpenSSL ECDSA signature algorithm has been shown to be vulnerable to a
3107 timing side channel attack. An attacker could use variations in the signing
3108 algorithm to recover the private key.
3110 This issue was reported to OpenSSL on 25th October 2018 by Samuel Weiser.
3115 * Add coordinate blinding for EC_POINT and implement projective
3116 coordinate blinding for generic prime curves as a countermeasure to
3117 chosen point SCA attacks.
3119 *Sohaib ul Hassan, Nicola Tuveri, Billy Bob Brumley*
3121 ### Changes between 1.1.0h and 1.1.0i [14 Aug 2018]
3123 * Client DoS due to large DH parameter
3125 During key agreement in a TLS handshake using a DH(E) based ciphersuite a
3126 malicious server can send a very large prime value to the client. This will
3127 cause the client to spend an unreasonably long period of time generating a
3128 key for this prime resulting in a hang until the client has finished. This
3129 could be exploited in a Denial Of Service attack.
3131 This issue was reported to OpenSSL on 5th June 2018 by Guido Vranken
3136 * Cache timing vulnerability in RSA Key Generation
3138 The OpenSSL RSA Key generation algorithm has been shown to be vulnerable to
3139 a cache timing side channel attack. An attacker with sufficient access to
3140 mount cache timing attacks during the RSA key generation process could
3141 recover the private key.
3143 This issue was reported to OpenSSL on 4th April 2018 by Alejandro Cabrera
3144 Aldaya, Billy Brumley, Cesar Pereida Garcia and Luis Manuel Alvarez Tapia.
3149 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
3150 parameter is no longer accepted, as it leads to a corrupt table. NULL
3151 pem_str is reserved for alias entries only.
3155 * Revert blinding in ECDSA sign and instead make problematic addition
3156 length-invariant. Switch even to fixed-length Montgomery multiplication.
3160 * Change generating and checking of primes so that the error rate of not
3161 being prime depends on the intended use based on the size of the input.
3162 For larger primes this will result in more rounds of Miller-Rabin.
3163 The maximal error rate for primes with more than 1080 bits is lowered
3166 *Kurt Roeckx, Annie Yousar*
3168 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
3172 * Add blinding to ECDSA and DSA signatures to protect against side channel
3173 attacks discovered by Keegan Ryan (NCC Group).
3177 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
3178 now allow empty (zero character) pass phrases.
3182 * Certificate time validation (X509_cmp_time) enforces stricter
3183 compliance with RFC 5280. Fractional seconds and timezone offsets
3184 are no longer allowed.
3188 * Fixed a text canonicalisation bug in CMS
3190 Where a CMS detached signature is used with text content the text goes
3191 through a canonicalisation process first prior to signing or verifying a
3192 signature. This process strips trailing space at the end of lines, converts
3193 line terminators to CRLF and removes additional trailing line terminators
3194 at the end of a file. A bug in the canonicalisation process meant that
3195 some characters, such as form-feed, were incorrectly treated as whitespace
3196 and removed. This is contrary to the specification (RFC5485). This fix
3197 could mean that detached text data signed with an earlier version of
3198 OpenSSL 1.1.0 may fail to verify using the fixed version, or text data
3199 signed with a fixed OpenSSL may fail to verify with an earlier version of
3200 OpenSSL 1.1.0. A workaround is to only verify the canonicalised text data
3201 and use the "-binary" flag (for the "cms" command line application) or set
3202 the SMIME_BINARY/PKCS7_BINARY/CMS_BINARY flags (if using CMS_verify()).
3206 ### Changes between 1.1.0g and 1.1.0h [27 Mar 2018]
3208 * Constructed ASN.1 types with a recursive definition could exceed the stack
3210 Constructed ASN.1 types with a recursive definition (such as can be found
3211 in PKCS7) could eventually exceed the stack given malicious input with
3212 excessive recursion. This could result in a Denial Of Service attack. There
3213 are no such structures used within SSL/TLS that come from untrusted sources
3214 so this is considered safe.
3216 This issue was reported to OpenSSL on 4th January 2018 by the OSS-fuzz
3222 * Incorrect CRYPTO_memcmp on HP-UX PA-RISC
3224 Because of an implementation bug the PA-RISC CRYPTO_memcmp function is
3225 effectively reduced to only comparing the least significant bit of each
3226 byte. This allows an attacker to forge messages that would be considered as
3227 authenticated in an amount of tries lower than that guaranteed by the
3228 security claims of the scheme. The module can only be compiled by the
3229 HP-UX assembler, so that only HP-UX PA-RISC targets are affected.
3231 This issue was reported to OpenSSL on 2nd March 2018 by Peter Waltenberg
3237 * Add a build target 'build_all_generated', to build all generated files
3238 and only that. This can be used to prepare everything that requires
3239 things like perl for a system that lacks perl and then move everything
3240 to that system and do the rest of the build there.
3244 * Backport SSL_OP_NO_RENGOTIATION
3246 OpenSSL 1.0.2 and below had the ability to disable renegotiation using the
3247 (undocumented) SSL3_FLAGS_NO_RENEGOTIATE_CIPHERS flag. Due to the opacity
3248 changes this is no longer possible in 1.1.0. Therefore the new
3249 SSL_OP_NO_RENEGOTIATION option from 1.1.1-dev has been backported to
3250 1.1.0 to provide equivalent functionality.
3252 Note that if an application built against 1.1.0h headers (or above) is run
3253 using an older version of 1.1.0 (prior to 1.1.0h) then the option will be
3254 accepted but nothing will happen, i.e. renegotiation will not be prevented.
3258 * Removed the OS390-Unix config target. It relied on a script that doesn't
3263 * rsaz_1024_mul_avx2 overflow bug on x86_64
3265 There is an overflow bug in the AVX2 Montgomery multiplication procedure
3266 used in exponentiation with 1024-bit moduli. No EC algorithms are affected.
3267 Analysis suggests that attacks against RSA and DSA as a result of this
3268 defect would be very difficult to perform and are not believed likely.
3269 Attacks against DH1024 are considered just feasible, because most of the
3270 work necessary to deduce information about a private key may be performed
3271 offline. The amount of resources required for such an attack would be
3272 significant. However, for an attack on TLS to be meaningful, the server
3273 would have to share the DH1024 private key among multiple clients, which is
3274 no longer an option since CVE-2016-0701.
3276 This only affects processors that support the AVX2 but not ADX extensions
3277 like Intel Haswell (4th generation).
3279 This issue was reported to OpenSSL by David Benjamin (Google). The issue
3280 was originally found via the OSS-Fuzz project.
3285 ### Changes between 1.1.0f and 1.1.0g [2 Nov 2017]
3287 * bn_sqrx8x_internal carry bug on x86_64
3289 There is a carry propagating bug in the x86_64 Montgomery squaring
3290 procedure. No EC algorithms are affected. Analysis suggests that attacks
3291 against RSA and DSA as a result of this defect would be very difficult to
3292 perform and are not believed likely. Attacks against DH are considered just
3293 feasible (although very difficult) because most of the work necessary to
3294 deduce information about a private key may be performed offline. The amount
3295 of resources required for such an attack would be very significant and
3296 likely only accessible to a limited number of attackers. An attacker would
3297 additionally need online access to an unpatched system using the target
3298 private key in a scenario with persistent DH parameters and a private
3299 key that is shared between multiple clients.
3301 This only affects processors that support the BMI1, BMI2 and ADX extensions
3302 like Intel Broadwell (5th generation) and later or AMD Ryzen.
3304 This issue was reported to OpenSSL by the OSS-Fuzz project.
3309 * Malformed X.509 IPAddressFamily could cause OOB read
3311 If an X.509 certificate has a malformed IPAddressFamily extension,
3312 OpenSSL could do a one-byte buffer overread. The most likely result
3313 would be an erroneous display of the certificate in text format.
3315 This issue was reported to OpenSSL by the OSS-Fuzz project.
3320 ### Changes between 1.1.0e and 1.1.0f [25 May 2017]
3322 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
3323 platform rather than 'mingw'.
3327 * Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
3328 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
3329 which is the minimum version we support.
3333 ### Changes between 1.1.0d and 1.1.0e [16 Feb 2017]
3335 * Encrypt-Then-Mac renegotiation crash
3337 During a renegotiation handshake if the Encrypt-Then-Mac extension is
3338 negotiated where it was not in the original handshake (or vice-versa) then
3339 this can cause OpenSSL to crash (dependant on ciphersuite). Both clients
3340 and servers are affected.
3342 This issue was reported to OpenSSL by Joe Orton (Red Hat).
3347 ### Changes between 1.1.0c and 1.1.0d [26 Jan 2017]
3349 * Truncated packet could crash via OOB read
3351 If one side of an SSL/TLS path is running on a 32-bit host and a specific
3352 cipher is being used, then a truncated packet can cause that host to
3353 perform an out-of-bounds read, usually resulting in a crash.
3355 This issue was reported to OpenSSL by Robert Święcki of Google.
3360 * Bad (EC)DHE parameters cause a client crash
3362 If a malicious server supplies bad parameters for a DHE or ECDHE key
3363 exchange then this can result in the client attempting to dereference a
3364 NULL pointer leading to a client crash. This could be exploited in a Denial
3367 This issue was reported to OpenSSL by Guido Vranken.
3372 * BN_mod_exp may produce incorrect results on x86_64
3374 There is a carry propagating bug in the x86_64 Montgomery squaring
3375 procedure. No EC algorithms are affected. Analysis suggests that attacks
3376 against RSA and DSA as a result of this defect would be very difficult to
3377 perform and are not believed likely. Attacks against DH are considered just
3378 feasible (although very difficult) because most of the work necessary to
3379 deduce information about a private key may be performed offline. The amount
3380 of resources required for such an attack would be very significant and
3381 likely only accessible to a limited number of attackers. An attacker would
3382 additionally need online access to an unpatched system using the target
3383 private key in a scenario with persistent DH parameters and a private
3384 key that is shared between multiple clients. For example this can occur by
3385 default in OpenSSL DHE based SSL/TLS ciphersuites. Note: This issue is very
3386 similar to CVE-2015-3193 but must be treated as a separate problem.
3388 This issue was reported to OpenSSL by the OSS-Fuzz project.
3393 ### Changes between 1.1.0b and 1.1.0c [10 Nov 2016]
3395 * ChaCha20/Poly1305 heap-buffer-overflow
3397 TLS connections using `*-CHACHA20-POLY1305` ciphersuites are susceptible to
3398 a DoS attack by corrupting larger payloads. This can result in an OpenSSL
3399 crash. This issue is not considered to be exploitable beyond a DoS.
3401 This issue was reported to OpenSSL by Robert Święcki (Google Security Team)
3406 * CMS Null dereference
3408 Applications parsing invalid CMS structures can crash with a NULL pointer
3409 dereference. This is caused by a bug in the handling of the ASN.1 CHOICE
3410 type in OpenSSL 1.1.0 which can result in a NULL value being passed to the
3411 structure callback if an attempt is made to free certain invalid encodings.
3412 Only CHOICE structures using a callback which do not handle NULL value are
3415 This issue was reported to OpenSSL by Tyler Nighswander of ForAllSecure.
3420 * Montgomery multiplication may produce incorrect results
3422 There is a carry propagating bug in the Broadwell-specific Montgomery
3423 multiplication procedure that handles input lengths divisible by, but
3424 longer than 256 bits. Analysis suggests that attacks against RSA, DSA
3425 and DH private keys are impossible. This is because the subroutine in
3426 question is not used in operations with the private key itself and an input
3427 of the attacker's direct choice. Otherwise the bug can manifest itself as
3428 transient authentication and key negotiation failures or reproducible
3429 erroneous outcome of public-key operations with specially crafted input.
3430 Among EC algorithms only Brainpool P-512 curves are affected and one
3431 presumably can attack ECDH key negotiation. Impact was not analyzed in
3432 detail, because pre-requisites for attack are considered unlikely. Namely
3433 multiple clients have to choose the curve in question and the server has to
3434 share the private key among them, neither of which is default behaviour.
3435 Even then only clients that chose the curve will be affected.
3437 This issue was publicly reported as transient failures and was not
3438 initially recognized as a security issue. Thanks to Richard Morgan for
3439 providing reproducible case.
3444 * Removed automatic addition of RPATH in shared libraries and executables,
3445 as this was a remainder from OpenSSL 1.0.x and isn't needed any more.
3449 ### Changes between 1.1.0a and 1.1.0b [26 Sep 2016]
3451 * Fix Use After Free for large message sizes
3453 The patch applied to address CVE-2016-6307 resulted in an issue where if a
3454 message larger than approx 16k is received then the underlying buffer to
3455 store the incoming message is reallocated and moved. Unfortunately a
3456 dangling pointer to the old location is left which results in an attempt to
3457 write to the previously freed location. This is likely to result in a
3458 crash, however it could potentially lead to execution of arbitrary code.
3460 This issue only affects OpenSSL 1.1.0a.
3462 This issue was reported to OpenSSL by Robert Święcki.
3467 ### Changes between 1.1.0 and 1.1.0a [22 Sep 2016]
3469 * OCSP Status Request extension unbounded memory growth
3471 A malicious client can send an excessively large OCSP Status Request
3472 extension. If that client continually requests renegotiation, sending a
3473 large OCSP Status Request extension each time, then there will be unbounded
3474 memory growth on the server. This will eventually lead to a Denial Of
3475 Service attack through memory exhaustion. Servers with a default
3476 configuration are vulnerable even if they do not support OCSP. Builds using
3477 the "no-ocsp" build time option are not affected.
3479 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
3484 * SSL_peek() hang on empty record
3486 OpenSSL 1.1.0 SSL/TLS will hang during a call to SSL_peek() if the peer
3487 sends an empty record. This could be exploited by a malicious peer in a
3488 Denial Of Service attack.
3490 This issue was reported to OpenSSL by Alex Gaynor.
3495 * Excessive allocation of memory in tls_get_message_header() and
3496 dtls1_preprocess_fragment()
3498 A (D)TLS message includes 3 bytes for its length in the header for the
3499 message. This would allow for messages up to 16Mb in length. Messages of
3500 this length are excessive and OpenSSL includes a check to ensure that a
3501 peer is sending reasonably sized messages in order to avoid too much memory
3502 being consumed to service a connection. A flaw in the logic of version
3503 1.1.0 means that memory for the message is allocated too early, prior to
3504 the excessive message length check. Due to way memory is allocated in
3505 OpenSSL this could mean an attacker could force up to 21Mb to be allocated
3506 to service a connection. This could lead to a Denial of Service through
3507 memory exhaustion. However, the excessive message length check still takes
3508 place, and this would cause the connection to immediately fail. Assuming
3509 that the application calls SSL_free() on the failed connection in a timely
3510 manner then the 21Mb of allocated memory will then be immediately freed
3511 again. Therefore the excessive memory allocation will be transitory in
3512 nature. This then means that there is only a security impact if:
3514 1) The application does not call SSL_free() in a timely manner in the event
3515 that the connection fails
3517 2) The application is working in a constrained environment where there is
3518 very little free memory
3520 3) The attacker initiates multiple connection attempts such that there are
3521 multiple connections in a state where memory has been allocated for the
3522 connection; SSL_free() has not yet been called; and there is insufficient
3523 memory to service the multiple requests.
3525 Except in the instance of (1) above any Denial Of Service is likely to be
3526 transitory because as soon as the connection fails the memory is
3527 subsequently freed again in the SSL_free() call. However there is an
3528 increased risk during this period of application crashes due to the lack of
3529 memory - which would then mean a more serious Denial of Service.
3531 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
3532 (CVE-2016-6307 and CVE-2016-6308)
3536 * solaris-x86-cc, i.e. 32-bit configuration with vendor compiler,
3537 had to be removed. Primary reason is that vendor assembler can't
3538 assemble our modules with -KPIC flag. As result it, assembly
3539 support, was not even available as option. But its lack means
3540 lack of side-channel resistant code, which is incompatible with
3541 security by todays standards. Fortunately gcc is readily available
3542 prepackaged option, which we firmly point at...
3546 ### Changes between 1.0.2h and 1.1.0 [25 Aug 2016]
3548 * Windows command-line tool supports UTF-8 opt-in option for arguments
3549 and console input. Setting OPENSSL_WIN32_UTF8 environment variable
3550 (to any value) allows Windows user to access PKCS#12 file generated
3551 with Windows CryptoAPI and protected with non-ASCII password, as well
3552 as files generated under UTF-8 locale on Linux also protected with
3557 * To mitigate the SWEET32 attack ([CVE-2016-2183]), 3DES cipher suites
3558 have been disabled by default and removed from DEFAULT, just like RC4.
3559 See the RC4 item below to re-enable both.
3563 * The method for finding the storage location for the Windows RAND seed file
3564 has changed. First we check %RANDFILE%. If that is not set then we check
3565 the directories %HOME%, %USERPROFILE% and %SYSTEMROOT% in that order. If
3566 all else fails we fall back to C:\.
3570 * The EVP_EncryptUpdate() function has had its return type changed from void
3571 to int. A return of 0 indicates and error while a return of 1 indicates
3576 * The flags RSA_FLAG_NO_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME and
3577 DH_FLAG_NO_EXP_CONSTTIME which previously provided the ability to switch
3578 off the constant time implementation for RSA, DSA and DH have been made
3579 no-ops and deprecated.
3583 * Windows RAND implementation was simplified to only get entropy by
3584 calling CryptGenRandom(). Various other RAND-related tickets
3587 *Joseph Wylie Yandle, Rich Salz*
3589 * The stack and lhash API's were renamed to start with `OPENSSL_SK_`
3590 and `OPENSSL_LH_`, respectively. The old names are available
3591 with API compatibility. They new names are now completely documented.
3595 * Unify TYPE_up_ref(obj) methods signature.
3596 SSL_CTX_up_ref(), SSL_up_ref(), X509_up_ref(), EVP_PKEY_up_ref(),
3597 X509_CRL_up_ref(), X509_OBJECT_up_ref_count() methods are now returning an
3598 int (instead of void) like all others TYPE_up_ref() methods.
3599 So now these methods also check the return value of CRYPTO_atomic_add(),
3600 and the validity of object reference counter.
3602 *fdasilvayy@gmail.com*
3604 * With Windows Visual Studio builds, the .pdb files are installed
3605 alongside the installed libraries and executables. For a static
3606 library installation, ossl_static.pdb is the associate compiler
3607 generated .pdb file to be used when linking programs.
3611 * Remove openssl.spec. Packaging files belong with the packagers.
3615 * Automatic Darwin/OSX configuration has had a refresh, it will now
3616 recognise x86_64 architectures automatically. You can still decide
3617 to build for a different bitness with the environment variable
3618 KERNEL_BITS (can be 32 or 64), for example:
3620 KERNEL_BITS=32 ./config
3624 * Change default algorithms in pkcs8 utility to use PKCS#5 v2.0,
3625 256 bit AES and HMAC with SHA256.
3629 * Remove support for MIPS o32 ABI on IRIX (and IRIX only).
3633 * Triple-DES ciphers have been moved from HIGH to MEDIUM.
3637 * To enable users to have their own config files and build file templates,
3638 Configure looks in the directory indicated by the environment variable
3639 OPENSSL_LOCAL_CONFIG_DIR as well as the in-source Configurations/
3640 directory. On VMS, OPENSSL_LOCAL_CONFIG_DIR is expected to be a logical
3641 name and is used as is.
3645 * The following datatypes were made opaque: X509_OBJECT, X509_STORE_CTX,
3646 X509_STORE, X509_LOOKUP, and X509_LOOKUP_METHOD. The unused type
3647 X509_CERT_FILE_CTX was removed.
3651 * "shared" builds are now the default. To create only static libraries use
3652 the "no-shared" Configure option.
3656 * Remove the no-aes, no-hmac, no-rsa, no-sha and no-md5 Configure options.
3657 All of these option have not worked for some while and are fundamental
3662 * Make various cleanup routines no-ops and mark them as deprecated. Most
3663 global cleanup functions are no longer required because they are handled
3664 via auto-deinit (see OPENSSL_init_crypto and OPENSSL_init_ssl man pages).
3665 Explicitly de-initing can cause problems (e.g. where a library that uses
3666 OpenSSL de-inits, but an application is still using it). The affected
3667 functions are CONF_modules_free(), ENGINE_cleanup(), OBJ_cleanup(),
3668 EVP_cleanup(), BIO_sock_cleanup(), CRYPTO_cleanup_all_ex_data(),
3669 RAND_cleanup(), SSL_COMP_free_compression_methods(), ERR_free_strings() and
3670 COMP_zlib_cleanup().
3674 * --strict-warnings no longer enables runtime debugging options
3675 such as REF_DEBUG. Instead, debug options are automatically
3676 enabled with '--debug' builds.
3678 *Andy Polyakov, Emilia Käsper*
3680 * Made DH and DH_METHOD opaque. The structures for managing DH objects
3681 have been moved out of the public header files. New functions for managing
3682 these have been added.
3686 * Made RSA and RSA_METHOD opaque. The structures for managing RSA
3687 objects have been moved out of the public header files. New
3688 functions for managing these have been added.
3692 * Made DSA and DSA_METHOD opaque. The structures for managing DSA objects
3693 have been moved out of the public header files. New functions for managing
3694 these have been added.
3698 * Made BIO and BIO_METHOD opaque. The structures for managing BIOs have been
3699 moved out of the public header files. New functions for managing these
3704 * Removed no-rijndael as a config option. Rijndael is an old name for AES.
3708 * Removed the mk1mf build scripts.
3712 * Headers are now wrapped, if necessary, with OPENSSL_NO_xxx, so
3713 it is always safe to #include a header now.
3717 * Removed the aged BC-32 config and all its supporting scripts
3721 * Removed support for Ultrix, Netware, and OS/2.
3725 * Add support for HKDF.
3727 *Alessandro Ghedini*
3729 * Add support for blake2b and blake2s
3733 * Added support for "pipelining". Ciphers that have the
3734 EVP_CIPH_FLAG_PIPELINE flag set have a capability to process multiple
3735 encryptions/decryptions simultaneously. There are currently no built-in
3736 ciphers with this property but the expectation is that engines will be able
3737 to offer it to significantly improve throughput. Support has been extended
3738 into libssl so that multiple records for a single connection can be
3739 processed in one go (for >=TLS 1.1).
3743 * Added the AFALG engine. This is an async capable engine which is able to
3744 offload work to the Linux kernel. In this initial version it only supports
3745 AES128-CBC. The kernel must be version 4.1.0 or greater.
3749 * OpenSSL now uses a new threading API. It is no longer necessary to
3750 set locking callbacks to use OpenSSL in a multi-threaded environment. There
3751 are two supported threading models: pthreads and windows threads. It is
3752 also possible to configure OpenSSL at compile time for "no-threads". The
3753 old threading API should no longer be used. The functions have been
3754 replaced with "no-op" compatibility macros.
3756 *Alessandro Ghedini, Matt Caswell*
3758 * Modify behavior of ALPN to invoke callback after SNI/servername
3759 callback, such that updates to the SSL_CTX affect ALPN.
3763 * Add SSL_CIPHER queries for authentication and key-exchange.
3767 * Changes to the DEFAULT cipherlist:
3768 - Prefer (EC)DHE handshakes over plain RSA.
3769 - Prefer AEAD ciphers over legacy ciphers.
3770 - Prefer ECDSA over RSA when both certificates are available.
3771 - Prefer TLSv1.2 ciphers/PRF.
3772 - Remove DSS, SEED, IDEA, CAMELLIA, and AES-CCM from the
3777 * Change the ECC default curve list to be this, in order: x25519,
3778 secp256r1, secp521r1, secp384r1.
3782 * RC4 based libssl ciphersuites are now classed as "weak" ciphers and are
3783 disabled by default. They can be re-enabled using the
3784 enable-weak-ssl-ciphers option to Configure.
3788 * If the server has ALPN configured, but supports no protocols that the
3789 client advertises, send a fatal "no_application_protocol" alert.
3790 This behaviour is SHALL in RFC 7301, though it isn't universally
3791 implemented by other servers.
3795 * Add X25519 support.
3796 Add ASN.1 and EVP_PKEY methods for X25519. This includes support
3797 for public and private key encoding using the format documented in
3798 draft-ietf-curdle-pkix-02. The corresponding EVP_PKEY method supports
3799 key generation and key derivation.
3801 TLS support complies with draft-ietf-tls-rfc4492bis-08 and uses
3806 * Deprecate SRP_VBASE_get_by_user.
3807 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
3808 In order to fix an unavoidable memory leak ([CVE-2016-0798]),
3809 SRP_VBASE_get_by_user was changed to ignore the "fake user" SRP
3810 seed, even if the seed is configured.
3812 Users should use SRP_VBASE_get1_by_user instead. Note that in
3813 SRP_VBASE_get1_by_user, caller must free the returned value. Note
3814 also that even though configuring the SRP seed attempts to hide
3815 invalid usernames by continuing the handshake with fake
3816 credentials, this behaviour is not constant time and no strong
3817 guarantees are made that the handshake is indistinguishable from
3818 that of a valid user.
3822 * Configuration change; it's now possible to build dynamic engines
3823 without having to build shared libraries and vice versa. This
3824 only applies to the engines in `engines/`, those in `crypto/engine/`
3825 will always be built into libcrypto (i.e. "static").
3827 Building dynamic engines is enabled by default; to disable, use
3828 the configuration option "disable-dynamic-engine".
3830 The only requirements for building dynamic engines are the
3831 presence of the DSO module and building with position independent
3832 code, so they will also automatically be disabled if configuring
3833 with "disable-dso" or "disable-pic".
3835 The macros OPENSSL_NO_STATIC_ENGINE and OPENSSL_NO_DYNAMIC_ENGINE
3836 are also taken away from openssl/opensslconf.h, as they are
3841 * Configuration change; if there is a known flag to compile
3842 position independent code, it will always be applied on the
3843 libcrypto and libssl object files, and never on the application
3844 object files. This means other libraries that use routines from
3845 libcrypto / libssl can be made into shared libraries regardless
3846 of how OpenSSL was configured.
3848 If this isn't desirable, the configuration options "disable-pic"
3849 or "no-pic" can be used to disable the use of PIC. This will
3850 also disable building shared libraries and dynamic engines.
3854 * Removed JPAKE code. It was experimental and has no wide use.
3858 * The INSTALL_PREFIX Makefile variable has been renamed to
3859 DESTDIR. That makes for less confusion on what this variable
3860 is for. Also, the configuration option --install_prefix is
3865 * Heartbeat for TLS has been removed and is disabled by default
3866 for DTLS; configure with enable-heartbeats. Code that uses the
3867 old #define's might need to be updated.
3869 *Emilia Käsper, Rich Salz*
3871 * Rename REF_CHECK to REF_DEBUG.
3875 * New "unified" build system
3877 The "unified" build system is aimed to be a common system for all
3878 platforms we support. With it comes new support for VMS.
3880 This system builds supports building in a different directory tree
3881 than the source tree. It produces one Makefile (for unix family
3882 or lookalikes), or one descrip.mms (for VMS).
3884 The source of information to make the Makefile / descrip.mms is
3885 small files called 'build.info', holding the necessary
3886 information for each directory with source to compile, and a
3887 template in Configurations, like unix-Makefile.tmpl or
3890 With this change, the library names were also renamed on Windows
3891 and on VMS. They now have names that are closer to the standard
3892 on Unix, and include the major version number, and in certain
3893 cases, the architecture they are built for. See "Notes on shared
3894 libraries" in INSTALL.
3896 We rely heavily on the perl module Text::Template.
3900 * Added support for auto-initialisation and de-initialisation of the library.
3901 OpenSSL no longer requires explicit init or deinit routines to be called,
3902 except in certain circumstances. See the OPENSSL_init_crypto() and
3903 OPENSSL_init_ssl() man pages for further information.
3907 * The arguments to the DTLSv1_listen function have changed. Specifically the
3908 "peer" argument is now expected to be a BIO_ADDR object.
3910 * Rewrite of BIO networking library. The BIO library lacked consistent
3911 support of IPv6, and adding it required some more extensive
3912 modifications. This introduces the BIO_ADDR and BIO_ADDRINFO types,
3913 which hold all types of addresses and chains of address information.
3914 It also introduces a new API, with functions like BIO_socket,
3915 BIO_connect, BIO_listen, BIO_lookup and a rewrite of BIO_accept.
3916 The source/sink BIOs BIO_s_connect, BIO_s_accept and BIO_s_datagram
3917 have been adapted accordingly.
3921 * RSA_padding_check_PKCS1_type_1 now accepts inputs with and without
3926 * CRIME protection: disable compression by default, even if OpenSSL is
3927 compiled with zlib enabled. Applications can still enable compression
3928 by calling SSL_CTX_clear_options(ctx, SSL_OP_NO_COMPRESSION), or by
3929 using the SSL_CONF library to configure compression.
3933 * The signature of the session callback configured with
3934 SSL_CTX_sess_set_get_cb was changed. The read-only input buffer
3935 was explicitly marked as `const unsigned char*` instead of
3940 * Always DPURIFY. Remove the use of uninitialized memory in the
3941 RNG, and other conditional uses of DPURIFY. This makes -DPURIFY a no-op.
3945 * Removed many obsolete configuration items, including
3946 DES_PTR, DES_RISC1, DES_RISC2, DES_INT
3947 MD2_CHAR, MD2_INT, MD2_LONG
3949 IDEA_SHORT, IDEA_LONG
3950 RC2_SHORT, RC2_LONG, RC4_LONG, RC4_CHUNK, RC4_INDEX
3952 *Rich Salz, with advice from Andy Polyakov*
3954 * Many BN internals have been moved to an internal header file.
3956 *Rich Salz with help from Andy Polyakov*
3958 * Configuration and writing out the results from it has changed.
3959 Files such as Makefile include/openssl/opensslconf.h and are now
3960 produced through general templates, such as Makefile.in and
3961 crypto/opensslconf.h.in and some help from the perl module
3964 Also, the center of configuration information is no longer
3965 Makefile. Instead, Configure produces a perl module in
3966 configdata.pm which holds most of the config data (in the hash
3967 table %config), the target data that comes from the target
3968 configuration in one of the `Configurations/*.conf` files (in
3973 * To clarify their intended purposes, the Configure options
3974 --prefix and --openssldir change their semantics, and become more
3975 straightforward and less interdependent.
3977 --prefix shall be used exclusively to give the location INSTALLTOP
3978 where programs, scripts, libraries, include files and manuals are
3979 going to be installed. The default is now /usr/local.
3981 --openssldir shall be used exclusively to give the default
3982 location OPENSSLDIR where certificates, private keys, CRLs are
3983 managed. This is also where the default openssl.cnf gets
3985 If the directory given with this option is a relative path, the
3986 values of both the --prefix value and the --openssldir value will
3987 be combined to become OPENSSLDIR.
3988 The default for --openssldir is INSTALLTOP/ssl.
3990 Anyone who uses --openssldir to specify where OpenSSL is to be
3991 installed MUST change to use --prefix instead.
3995 * The GOST engine was out of date and therefore it has been removed. An up
3996 to date GOST engine is now being maintained in an external repository.
3997 See: <https://wiki.openssl.org/index.php/Binaries>. Libssl still retains
3998 support for GOST ciphersuites (these are only activated if a GOST engine
4003 * EGD is no longer supported by default; use enable-egd when
4006 *Ben Kaduk and Rich Salz*
4008 * The distribution now has Makefile.in files, which are used to
4009 create Makefile's when Configure is run. *Configure must be run
4010 before trying to build now.*
4014 * The return value for SSL_CIPHER_description() for error conditions
4019 * Support for RFC6698/RFC7671 DANE TLSA peer authentication.
4021 Obtaining and performing DNSSEC validation of TLSA records is
4022 the application's responsibility. The application provides
4023 the TLSA records of its choice to OpenSSL, and these are then
4024 used to authenticate the peer.
4026 The TLSA records need not even come from DNS. They can, for
4027 example, be used to implement local end-entity certificate or
4028 trust-anchor "pinning", where the "pin" data takes the form
4029 of TLSA records, which can augment or replace verification
4030 based on the usual WebPKI public certification authorities.
4034 * Revert default OPENSSL_NO_DEPRECATED setting. Instead OpenSSL
4035 continues to support deprecated interfaces in default builds.
4036 However, applications are strongly advised to compile their
4037 source files with -DOPENSSL_API_COMPAT=0x10100000L, which hides
4038 the declarations of all interfaces deprecated in 0.9.8, 1.0.0
4039 or the 1.1.0 releases.
4041 In environments in which all applications have been ported to
4042 not use any deprecated interfaces OpenSSL's Configure script
4043 should be used with the --api=1.1.0 option to entirely remove
4044 support for the deprecated features from the library and
4045 unconditionally disable them in the installed headers.
4046 Essentially the same effect can be achieved with the "no-deprecated"
4047 argument to Configure, except that this will always restrict
4048 the build to just the latest API, rather than a fixed API
4051 As applications are ported to future revisions of the API,
4052 they should update their compile-time OPENSSL_API_COMPAT define
4053 accordingly, but in most cases should be able to continue to
4054 compile with later releases.
4056 The OPENSSL_API_COMPAT versions for 1.0.0, and 0.9.8 are
4057 0x10000000L and 0x00908000L, respectively. However those
4058 versions did not support the OPENSSL_API_COMPAT feature, and
4059 so applications are not typically tested for explicit support
4060 of just the undeprecated features of either release.
4064 * Add support for setting the minimum and maximum supported protocol.
4065 It can bet set via the SSL_set_min_proto_version() and
4066 SSL_set_max_proto_version(), or via the SSL_CONF's MinProtocol and
4067 MaxProtocol. It's recommended to use the new APIs to disable
4068 protocols instead of disabling individual protocols using
4069 SSL_set_options() or SSL_CONF's Protocol. This change also
4070 removes support for disabling TLS 1.2 in the OpenSSL TLS
4071 client at compile time by defining OPENSSL_NO_TLS1_2_CLIENT.
4075 * Support for ChaCha20 and Poly1305 added to libcrypto and libssl.
4079 * New EC_KEY_METHOD, this replaces the older ECDSA_METHOD and ECDH_METHOD
4080 and integrates ECDSA and ECDH functionality into EC. Implementations can
4081 now redirect key generation and no longer need to convert to or from
4084 Note: the ecdsa.h and ecdh.h headers are now no longer needed and just
4085 include the ec.h header file instead.
4089 * Remove support for all 40 and 56 bit ciphers. This includes all the export
4090 ciphers who are no longer supported and drops support the ephemeral RSA key
4091 exchange. The LOW ciphers currently doesn't have any ciphers in it.
4095 * Made EVP_MD_CTX, EVP_MD, EVP_CIPHER_CTX, EVP_CIPHER and HMAC_CTX
4096 opaque. For HMAC_CTX, the following constructors and destructors
4099 HMAC_CTX *HMAC_CTX_new(void);
4100 void HMAC_CTX_free(HMAC_CTX *ctx);
4102 For EVP_MD and EVP_CIPHER, complete APIs to create, fill and
4103 destroy such methods has been added. See EVP_MD_meth_new(3) and
4104 EVP_CIPHER_meth_new(3) for documentation.
4107 1) `EVP_MD_CTX_cleanup()`, `EVP_CIPHER_CTX_cleanup()` and
4108 `HMAC_CTX_cleanup()` were removed. `HMAC_CTX_reset()` and
4109 `EVP_MD_CTX_reset()` should be called instead to reinitialise
4110 an already created structure.
4111 2) For consistency with the majority of our object creators and
4112 destructors, `EVP_MD_CTX_(create|destroy)` were renamed to
4113 `EVP_MD_CTX_(new|free)`. The old names are retained as macros
4114 for deprecated builds.
4118 * Added ASYNC support. Libcrypto now includes the async sub-library to enable
4119 cryptographic operations to be performed asynchronously as long as an
4120 asynchronous capable engine is used. See the ASYNC_start_job() man page for
4121 further details. Libssl has also had this capability integrated with the
4122 introduction of the new mode SSL_MODE_ASYNC and associated error
4123 SSL_ERROR_WANT_ASYNC. See the SSL_CTX_set_mode() and SSL_get_error() man
4124 pages. This work was developed in partnership with Intel Corp.
4128 * SSL_{CTX_}set_ecdh_auto() has been removed and ECDH is support is
4129 always enabled now. If you want to disable the support you should
4130 exclude it using the list of supported ciphers. This also means that the
4131 "-no_ecdhe" option has been removed from s_server.
4135 * SSL_{CTX}_set_tmp_ecdh() which can set 1 EC curve now internally calls
4136 SSL_{CTX_}set1_curves() which can set a list.
4140 * Remove support for SSL_{CTX_}set_tmp_ecdh_callback(). You should set the
4141 curve you want to support using SSL_{CTX_}set1_curves().
4145 * State machine rewrite. The state machine code has been significantly
4146 refactored in order to remove much duplication of code and solve issues
4147 with the old code (see [ssl/statem/README.md](ssl/statem/README.md) for
4148 further details). This change does have some associated API changes.
4149 Notably the SSL_state() function has been removed and replaced by
4150 SSL_get_state which now returns an "OSSL_HANDSHAKE_STATE" instead of an int.
4151 SSL_set_state() has been removed altogether. The previous handshake states
4152 defined in ssl.h and ssl3.h have also been removed.
4156 * All instances of the string "ssleay" in the public API were replaced
4157 with OpenSSL (case-matching; e.g., OPENSSL_VERSION for #define's)
4158 Some error codes related to internal RSA_eay API's were renamed.
4162 * The demo files in crypto/threads were moved to demo/threads.
4166 * Removed obsolete engines: 4758cca, aep, atalla, cswift, nuron, gmp,
4169 *Matt Caswell, Rich Salz*
4171 * New ASN.1 embed macro.
4173 New ASN.1 macro ASN1_EMBED. This is the same as ASN1_SIMPLE except the
4174 structure is not allocated: it is part of the parent. That is instead of
4182 This reduces memory fragmentation and make it impossible to accidentally
4183 set a mandatory field to NULL.
4185 This currently only works for some fields specifically a SEQUENCE, CHOICE,
4186 or ASN1_STRING type which is part of a parent SEQUENCE. Since it is
4187 equivalent to ASN1_SIMPLE it cannot be tagged, OPTIONAL, SET OF or
4192 * Remove EVP_CHECK_DES_KEY, a compile-time option that never compiled.
4196 * Removed DES and RC4 ciphersuites from DEFAULT. Also removed RC2 although
4197 in 1.0.2 EXPORT was already removed and the only RC2 ciphersuite is also
4198 an EXPORT one. COMPLEMENTOFDEFAULT has been updated accordingly to add
4199 DES and RC4 ciphersuites.
4203 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
4204 This changes the decoding behaviour for some invalid messages,
4205 though the change is mostly in the more lenient direction, and
4206 legacy behaviour is preserved as much as possible.
4210 * Fix no-stdio build.
4211 *David Woodhouse <David.Woodhouse@intel.com> and also*
4212 *Ivan Nestlerode <ivan.nestlerode@sonos.com>*
4214 * New testing framework
4215 The testing framework has been largely rewritten and is now using
4216 perl and the perl modules Test::Harness and an extended variant of
4217 Test::More called OpenSSL::Test to do its work. All test scripts in
4218 test/ have been rewritten into test recipes, and all direct calls to
4219 executables in test/Makefile have become individual recipes using the
4220 simplified testing OpenSSL::Test::Simple.
4222 For documentation on our testing modules, do:
4224 perldoc test/testlib/OpenSSL/Test/Simple.pm
4225 perldoc test/testlib/OpenSSL/Test.pm
4229 * Revamped memory debug; only -DCRYPTO_MDEBUG and -DCRYPTO_MDEBUG_ABORT
4230 are used; the latter aborts on memory leaks (usually checked on exit).
4231 Some undocumented "set malloc, etc., hooks" functions were removed
4232 and others were changed. All are now documented.
4236 * In DSA_generate_parameters_ex, if the provided seed is too short,
4239 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
4241 * Rewrite PSK to support ECDHE_PSK, DHE_PSK and RSA_PSK. Add ciphersuites
4242 from RFC4279, RFC4785, RFC5487, RFC5489.
4244 Thanks to Christian J. Dietrich and Giuseppe D'Angelo for the
4245 original RSA_PSK patch.
4249 * Dropped support for the SSL3_FLAGS_DELAY_CLIENT_FINISHED flag. This SSLeay
4250 era flag was never set throughout the codebase (only read). Also removed
4251 SSL3_FLAGS_POP_BUFFER which was only used if
4252 SSL3_FLAGS_DELAY_CLIENT_FINISHED was also set.
4256 * Changed the default name options in the "ca", "crl", "req" and "x509"
4257 to be "oneline" instead of "compat".
4261 * Remove SSL_OP_TLS_BLOCK_PADDING_BUG. This is SSLeay legacy, we're
4262 not aware of clients that still exhibit this bug, and the workaround
4263 hasn't been working properly for a while.
4267 * The return type of BIO_number_read() and BIO_number_written() as well as
4268 the corresponding num_read and num_write members in the BIO structure has
4269 changed from unsigned long to uint64_t. On platforms where an unsigned
4270 long is 32 bits (e.g. Windows) these counters could overflow if >4Gb is
4275 * Given the pervasive nature of TLS extensions it is inadvisable to run
4276 OpenSSL without support for them. It also means that maintaining
4277 the OPENSSL_NO_TLSEXT option within the code is very invasive (and probably
4278 not well tested). Therefore the OPENSSL_NO_TLSEXT option has been removed.
4282 * Removed support for the two export grade static DH ciphersuites
4283 EXP-DH-RSA-DES-CBC-SHA and EXP-DH-DSS-DES-CBC-SHA. These two ciphersuites
4284 were newly added (along with a number of other static DH ciphersuites) to
4285 1.0.2. However the two export ones have *never* worked since they were
4286 introduced. It seems strange in any case to be adding new export
4287 ciphersuites, and given "logjam" it also does not seem correct to fix them.
4291 * Version negotiation has been rewritten. In particular SSLv23_method(),
4292 SSLv23_client_method() and SSLv23_server_method() have been deprecated,
4293 and turned into macros which simply call the new preferred function names
4294 TLS_method(), TLS_client_method() and TLS_server_method(). All new code
4295 should use the new names instead. Also as part of this change the ssl23.h
4296 header file has been removed.
4300 * Support for Kerberos ciphersuites in TLS (RFC2712) has been removed. This
4301 code and the associated standard is no longer considered fit-for-purpose.
4305 * RT2547 was closed. When generating a private key, try to make the
4306 output file readable only by the owner. This behavior change might
4307 be noticeable when interacting with other software.
4309 * Documented all exdata functions. Added CRYPTO_free_ex_index.
4314 * Added HTTP GET support to the ocsp command.
4318 * Changed default digest for the dgst and enc commands from MD5 to
4323 * RAND_pseudo_bytes has been deprecated. Users should use RAND_bytes instead.
4327 * Added support for TLS extended master secret from
4328 draft-ietf-tls-session-hash-03.txt. Thanks for Alfredo Pironti for an
4329 initial patch which was a great help during development.
4333 * All libssl internal structures have been removed from the public header
4334 files, and the OPENSSL_NO_SSL_INTERN option has been removed (since it is
4335 now redundant). Users should not attempt to access internal structures
4336 directly. Instead they should use the provided API functions.
4340 * config has been changed so that by default OPENSSL_NO_DEPRECATED is used.
4341 Access to deprecated functions can be re-enabled by running config with
4342 "enable-deprecated". In addition applications wishing to use deprecated
4343 functions must define OPENSSL_USE_DEPRECATED. Note that this new behaviour
4344 will, by default, disable some transitive includes that previously existed
4345 in the header files (e.g. ec.h will no longer, by default, include bn.h)
4349 * Added support for OCB mode. OpenSSL has been granted a patent license
4350 compatible with the OpenSSL license for use of OCB. Details are available
4351 at <https://www.openssl.org/source/OCB-patent-grant-OpenSSL.pdf>. Support
4352 for OCB can be removed by calling config with no-ocb.
4356 * SSLv2 support has been removed. It still supports receiving a SSLv2
4357 compatible client hello.
4361 * Increased the minimal RSA keysize from 256 to 512 bits [Rich Salz],
4362 done while fixing the error code for the key-too-small case.
4364 *Annie Yousar <a.yousar@informatik.hu-berlin.de>*
4366 * CA.sh has been removed; use CA.pl instead.
4370 * Removed old DES API.
4374 * Remove various unsupported platforms:
4380 Sinix/ReliantUNIX RM400
4385 16-bit platforms such as WIN16
4389 * Clean up OPENSSL_NO_xxx #define's
4390 - Use setbuf() and remove OPENSSL_NO_SETVBUF_IONBF
4391 - Rename OPENSSL_SYSNAME_xxx to OPENSSL_SYS_xxx
4392 - OPENSSL_NO_EC{DH,DSA} merged into OPENSSL_NO_EC
4393 - OPENSSL_NO_RIPEMD160, OPENSSL_NO_RIPEMD merged into OPENSSL_NO_RMD160
4394 - OPENSSL_NO_FP_API merged into OPENSSL_NO_STDIO
4395 - Remove OPENSSL_NO_BIO OPENSSL_NO_BUFFER OPENSSL_NO_CHAIN_VERIFY
4396 OPENSSL_NO_EVP OPENSSL_NO_FIPS_ERR OPENSSL_NO_HASH_COMP
4397 OPENSSL_NO_LHASH OPENSSL_NO_OBJECT OPENSSL_NO_SPEED OPENSSL_NO_STACK
4398 OPENSSL_NO_X509 OPENSSL_NO_X509_VERIFY
4399 - Remove MS_STATIC; it's a relic from platforms <32 bits.
4403 * Cleaned up dead code
4404 Remove all but one '#ifdef undef' which is to be looked at.
4408 * Clean up calling of xxx_free routines.
4409 Just like free(), fix most of the xxx_free routines to accept
4410 NULL. Remove the non-null checks from callers. Save much code.
4414 * Add secure heap for storage of private keys (when possible).
4415 Add BIO_s_secmem(), CBIGNUM, etc.
4416 Contributed by Akamai Technologies under our Corporate CLA.
4420 * Experimental support for a new, fast, unbiased prime candidate generator,
4421 bn_probable_prime_dh_coprime(). Not currently used by any prime generator.
4423 *Felix Laurie von Massenbach <felix@erbridge.co.uk>*
4425 * New output format NSS in the sess_id command line tool. This allows
4426 exporting the session id and the master key in NSS keylog format.
4428 *Martin Kaiser <martin@kaiser.cx>*
4430 * Harmonize version and its documentation. -f flag is used to display
4433 *mancha <mancha1@zoho.com>*
4435 * Fix eckey_priv_encode so it immediately returns an error upon a failure
4436 in i2d_ECPrivateKey. Thanks to Ted Unangst for feedback on this issue.
4438 *mancha <mancha1@zoho.com>*
4440 * Fix some double frees. These are not thought to be exploitable.
4442 *mancha <mancha1@zoho.com>*
4444 * A missing bounds check in the handling of the TLS heartbeat extension
4445 can be used to reveal up to 64k of memory to a connected client or
4448 Thanks for Neel Mehta of Google Security for discovering this bug and to
4449 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
4450 preparing the fix ([CVE-2014-0160])
4452 *Adam Langley, Bodo Moeller*
4454 * Fix for the attack described in the paper "Recovering OpenSSL
4455 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
4456 by Yuval Yarom and Naomi Benger. Details can be obtained from:
4457 <http://eprint.iacr.org/2014/140>
4459 Thanks to Yuval Yarom and Naomi Benger for discovering this
4460 flaw and to Yuval Yarom for supplying a fix ([CVE-2014-0076])
4462 *Yuval Yarom and Naomi Benger*
4464 * Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
4465 this fixes a limitation in previous versions of OpenSSL.
4469 * Experimental encrypt-then-mac support.
4471 Experimental support for encrypt then mac from
4472 draft-gutmann-tls-encrypt-then-mac-02.txt
4474 To enable it set the appropriate extension number (0x42 for the test
4475 server) using e.g. -DTLSEXT_TYPE_encrypt_then_mac=0x42
4477 For non-compliant peers (i.e. just about everything) this should have no
4480 WARNING: EXPERIMENTAL, SUBJECT TO CHANGE.
4484 * Add EVP support for key wrapping algorithms, to avoid problems with
4485 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
4486 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
4487 algorithms and include tests cases.
4491 * Extend CMS code to support RSA-PSS signatures and RSA-OAEP for
4496 * Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
4497 MGF1 digest and OAEP label.
4501 * Make openssl verify return errors.
4503 *Chris Palmer <palmer@google.com> and Ben Laurie*
4505 * New function ASN1_TIME_diff to calculate the difference between two
4506 ASN1_TIME structures or one structure and the current time.
4510 * Update fips_test_suite to support multiple command line options. New
4511 test to induce all self test errors in sequence and check expected
4516 * Add FIPS_{rsa,dsa,ecdsa}_{sign,verify} functions which digest and
4517 sign or verify all in one operation.
4521 * Add fips_algvs: a multicall fips utility incorporating all the algorithm
4522 test programs and fips_test_suite. Includes functionality to parse
4523 the minimal script output of fipsalgest.pl directly.
4527 * Add authorisation parameter to FIPS_module_mode_set().
4531 * Add FIPS selftest for ECDH algorithm using P-224 and B-233 curves.
4535 * Use separate DRBG fields for internal and external flags. New function
4536 FIPS_drbg_health_check() to perform on demand health checking. Add
4537 generation tests to fips_test_suite with reduced health check interval to
4538 demonstrate periodic health checking. Add "nodh" option to
4539 fips_test_suite to skip very slow DH test.
4543 * New function FIPS_get_cipherbynid() to lookup FIPS supported ciphers
4548 * More extensive health check for DRBG checking many more failure modes.
4549 New function FIPS_selftest_drbg_all() to handle every possible DRBG
4550 combination: call this in fips_test_suite.
4554 * Add support for canonical generation of DSA parameter 'g'. See
4557 * Add support for HMAC DRBG from SP800-90. Update DRBG algorithm test and
4558 POST to handle HMAC cases.
4562 * Add functions FIPS_module_version() and FIPS_module_version_text()
4563 to return numerical and string versions of the FIPS module number.
4567 * Rename FIPS_mode_set and FIPS_mode to FIPS_module_mode_set and
4568 FIPS_module_mode. FIPS_mode and FIPS_mode_set will be implemented
4569 outside the validated module in the FIPS capable OpenSSL.
4573 * Minor change to DRBG entropy callback semantics. In some cases
4574 there is no multiple of the block length between min_len and
4575 max_len. Allow the callback to return more than max_len bytes
4576 of entropy but discard any extra: it is the callback's responsibility
4577 to ensure that the extra data discarded does not impact the
4578 requested amount of entropy.
4582 * Add PRNG security strength checks to RSA, DSA and ECDSA using
4583 information in FIPS186-3, SP800-57 and SP800-131A.
4587 * CCM support via EVP. Interface is very similar to GCM case except we
4588 must supply all data in one chunk (i.e. no update, final) and the
4589 message length must be supplied if AAD is used. Add algorithm test
4594 * Initial version of POST overhaul. Add POST callback to allow the status
4595 of POST to be monitored and/or failures induced. Modify fips_test_suite
4596 to use callback. Always run all selftests even if one fails.
4600 * XTS support including algorithm test driver in the fips_gcmtest program.
4601 Note: this does increase the maximum key length from 32 to 64 bytes but
4602 there should be no binary compatibility issues as existing applications
4603 will never use XTS mode.
4607 * Extensive reorganisation of FIPS PRNG behaviour. Remove all dependencies
4608 to OpenSSL RAND code and replace with a tiny FIPS RAND API which also
4609 performs algorithm blocking for unapproved PRNG types. Also do not
4610 set PRNG type in FIPS_mode_set(): leave this to the application.
4611 Add default OpenSSL DRBG handling: sets up FIPS PRNG and seeds with
4612 the standard OpenSSL PRNG: set additional data to a date time vector.
4616 * Rename old X9.31 PRNG functions of the form `FIPS_rand*` to `FIPS_x931*`.
4617 This shouldn't present any incompatibility problems because applications
4618 shouldn't be using these directly and any that are will need to rethink
4619 anyway as the X9.31 PRNG is now deprecated by FIPS 140-2
4623 * Extensive self tests and health checking required by SP800-90 DRBG.
4624 Remove strength parameter from FIPS_drbg_instantiate and always
4625 instantiate at maximum supported strength.
4629 * Add ECDH code to fips module and fips_ecdhvs for primitives only testing.
4633 * New algorithm test program fips_dhvs to handle DH primitives only testing.
4637 * New function DH_compute_key_padded() to compute a DH key and pad with
4638 leading zeroes if needed: this complies with SP800-56A et al.
4642 * Initial implementation of SP800-90 DRBGs for Hash and CTR. Not used by
4643 anything, incomplete, subject to change and largely untested at present.
4647 * Modify fipscanisteronly build option to only build the necessary object
4648 files by filtering FIPS_EX_OBJ through a perl script in crypto/Makefile.
4652 * Add experimental option FIPSSYMS to give all symbols in
4653 fipscanister.o and FIPS or fips prefix. This will avoid
4654 conflicts with future versions of OpenSSL. Add perl script
4655 util/fipsas.pl to preprocess assembly language source files
4656 and rename any affected symbols.
4660 * Add selftest checks and algorithm block of non-fips algorithms in
4661 FIPS mode. Remove DES2 from selftests.
4665 * Add ECDSA code to fips module. Add tiny fips_ecdsa_check to just
4666 return internal method without any ENGINE dependencies. Add new
4667 tiny fips sign and verify functions.
4671 * New build option no-ec2m to disable characteristic 2 code.
4675 * New build option "fipscanisteronly". This only builds fipscanister.o
4676 and (currently) associated fips utilities. Uses the file Makefile.fips
4677 instead of Makefile.org as the prototype.
4681 * Add some FIPS mode restrictions to GCM. Add internal IV generator.
4682 Update fips_gcmtest to use IV generator.
4686 * Initial, experimental EVP support for AES-GCM. AAD can be input by
4687 setting output buffer to NULL. The `*Final` function must be
4688 called although it will not retrieve any additional data. The tag
4689 can be set or retrieved with a ctrl. The IV length is by default 12
4690 bytes (96 bits) but can be set to an alternative value. If the IV
4691 length exceeds the maximum IV length (currently 16 bytes) it cannot be
4696 * New flag in ciphers: EVP_CIPH_FLAG_CUSTOM_CIPHER. This means the
4697 underlying do_cipher function handles all cipher semantics itself
4698 including padding and finalisation. This is useful if (for example)
4699 an ENGINE cipher handles block padding itself. The behaviour of
4700 do_cipher is subtly changed if this flag is set: the return value
4701 is the number of characters written to the output buffer (zero is
4702 no longer an error code) or a negative error code. Also if the
4703 input buffer is NULL and length 0 finalisation should be performed.
4707 * If a candidate issuer certificate is already part of the constructed
4708 path ignore it: new debug notification X509_V_ERR_PATH_LOOP for this case.
4712 * Improve forward-security support: add functions
4714 void SSL_CTX_set_not_resumable_session_callback(
4715 SSL_CTX *ctx, int (*cb)(SSL *ssl, int is_forward_secure))
4716 void SSL_set_not_resumable_session_callback(
4717 SSL *ssl, int (*cb)(SSL *ssl, int is_forward_secure))
4719 for use by SSL/TLS servers; the callback function will be called whenever a
4720 new session is created, and gets to decide whether the session may be
4721 cached to make it resumable (return 0) or not (return 1). (As by the
4722 SSL/TLS protocol specifications, the session_id sent by the server will be
4723 empty to indicate that the session is not resumable; also, the server will
4724 not generate RFC 4507 (RFC 5077) session tickets.)
4726 A simple reasonable callback implementation is to return is_forward_secure.
4727 This parameter will be set to 1 or 0 depending on the ciphersuite selected
4728 by the SSL/TLS server library, indicating whether it can provide forward
4731 *Emilia Käsper <emilia.kasper@esat.kuleuven.be> (Google)*
4733 * New -verify_name option in command line utilities to set verification
4738 * Initial CMAC implementation. WARNING: EXPERIMENTAL, API MAY CHANGE.
4739 Add CMAC pkey methods.
4743 * Experimental renegotiation in s_server -www mode. If the client
4744 browses /reneg connection is renegotiated. If /renegcert it is
4745 renegotiated requesting a certificate.
4749 * Add an "external" session cache for debugging purposes to s_server. This
4750 should help trace issues which normally are only apparent in deployed
4751 multi-process servers.
4755 * Extensive audit of libcrypto with DEBUG_UNUSED. Fix many cases where
4756 return value is ignored. NB. The functions RAND_add(), RAND_seed(),
4757 BIO_set_cipher() and some obscure PEM functions were changed so they
4758 can now return an error. The RAND changes required a change to the
4759 RAND_METHOD structure.
4763 * New macro `__owur` for "OpenSSL Warn Unused Result". This makes use of
4764 a gcc attribute to warn if the result of a function is ignored. This
4765 is enable if DEBUG_UNUSED is set. Add to several functions in evp.h
4766 whose return value is often ignored.
4770 * New -noct, -requestct, -requirect and -ctlogfile options for s_client.
4771 These allow SCTs (signed certificate timestamps) to be requested and
4772 validated when establishing a connection.
4774 *Rob Percival <robpercival@google.com>*
4779 ### Changes between 1.0.2s and 1.0.2t [10 Sep 2019]
4781 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
4782 used even when parsing explicit parameters, when loading a encoded key
4783 or calling `EC_GROUP_new_from_ecpkparameters()`/
4784 `EC_GROUP_new_from_ecparameters()`.
4785 This prevents bypass of security hardening and performance gains,
4786 especially for curves with specialized EC_METHODs.
4787 By default, if a key encoded with explicit parameters is loaded and later
4788 encoded, the output is still encoded with explicit parameters, even if
4789 internally a "named" EC_GROUP is used for computation.
4793 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
4794 this change, EC_GROUP_set_generator would accept order and/or cofactor as
4795 NULL. After this change, only the cofactor parameter can be NULL. It also
4796 does some minimal sanity checks on the passed order.
4801 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
4802 An attack is simple, if the first CMS_recipientInfo is valid but the
4803 second CMS_recipientInfo is chosen ciphertext. If the second
4804 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
4805 encryption key will be replaced by garbage, and the message cannot be
4806 decoded, but if the RSA decryption fails, the correct encryption key is
4807 used and the recipient will not notice the attack.
4808 As a work around for this potential attack the length of the decrypted
4809 key must be equal to the cipher default key length, in case the
4810 certifiate is not given and all recipientInfo are tried out.
4811 The old behaviour can be re-enabled in the CMS code by setting the
4812 CMS_DEBUG_DECRYPT flag.
4817 * Document issue with installation paths in diverse Windows builds
4819 '/usr/local/ssl' is an unsafe prefix for location to install OpenSSL
4820 binaries and run-time config file.
4825 ### Changes between 1.0.2r and 1.0.2s [28 May 2019]
4827 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
4828 This changes the size when using the `genpkey` command when no size is given.
4829 It fixes an omission in earlier changes that changed all RSA, DSA and DH
4830 generation commands to use 2048 bits by default.
4834 * Add FIPS support for Android Arm 64-bit
4836 Support for Android Arm 64-bit was added to the OpenSSL FIPS Object
4837 Module in Version 2.0.10. For some reason, the corresponding target
4838 'android64-aarch64' was missing OpenSSL 1.0.2, whence it could not be
4839 built with FIPS support on Android Arm 64-bit. This omission has been
4842 *Matthias St. Pierre*
4844 ### Changes between 1.0.2q and 1.0.2r [26 Feb 2019]
4846 * 0-byte record padding oracle
4848 If an application encounters a fatal protocol error and then calls
4849 SSL_shutdown() twice (once to send a close_notify, and once to receive one)
4850 then OpenSSL can respond differently to the calling application if a 0 byte
4851 record is received with invalid padding compared to if a 0 byte record is
4852 received with an invalid MAC. If the application then behaves differently
4853 based on that in a way that is detectable to the remote peer, then this
4854 amounts to a padding oracle that could be used to decrypt data.
4856 In order for this to be exploitable "non-stitched" ciphersuites must be in
4857 use. Stitched ciphersuites are optimised implementations of certain
4858 commonly used ciphersuites. Also the application must call SSL_shutdown()
4859 twice even if a protocol error has occurred (applications should not do
4860 this but some do anyway).
4862 This issue was discovered by Juraj Somorovsky, Robert Merget and Nimrod
4863 Aviram, with additional investigation by Steven Collison and Andrew
4864 Hourselt. It was reported to OpenSSL on 10th December 2018.
4869 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
4873 ### Changes between 1.0.2p and 1.0.2q [20 Nov 2018]
4875 * Microarchitecture timing vulnerability in ECC scalar multiplication
4877 OpenSSL ECC scalar multiplication, used in e.g. ECDSA and ECDH, has been
4878 shown to be vulnerable to a microarchitecture timing side channel attack.
4879 An attacker with sufficient access to mount local timing attacks during
4880 ECDSA signature generation could recover the private key.
4882 This issue was reported to OpenSSL on 26th October 2018 by Alejandro
4883 Cabrera Aldaya, Billy Brumley, Sohaib ul Hassan, Cesar Pereida Garcia and
4889 * Timing vulnerability in DSA signature generation
4891 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
4892 timing side channel attack. An attacker could use variations in the signing
4893 algorithm to recover the private key.
4895 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
4900 * Resolve a compatibility issue in EC_GROUP handling with the FIPS Object
4901 Module, accidentally introduced while backporting security fixes from the
4902 development branch and hindering the use of ECC in FIPS mode.
4906 ### Changes between 1.0.2o and 1.0.2p [14 Aug 2018]
4908 * Client DoS due to large DH parameter
4910 During key agreement in a TLS handshake using a DH(E) based ciphersuite a
4911 malicious server can send a very large prime value to the client. This will
4912 cause the client to spend an unreasonably long period of time generating a
4913 key for this prime resulting in a hang until the client has finished. This
4914 could be exploited in a Denial Of Service attack.
4916 This issue was reported to OpenSSL on 5th June 2018 by Guido Vranken
4921 * Cache timing vulnerability in RSA Key Generation
4923 The OpenSSL RSA Key generation algorithm has been shown to be vulnerable to
4924 a cache timing side channel attack. An attacker with sufficient access to
4925 mount cache timing attacks during the RSA key generation process could
4926 recover the private key.
4928 This issue was reported to OpenSSL on 4th April 2018 by Alejandro Cabrera
4929 Aldaya, Billy Brumley, Cesar Pereida Garcia and Luis Manuel Alvarez Tapia.
4934 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
4935 parameter is no longer accepted, as it leads to a corrupt table. NULL
4936 pem_str is reserved for alias entries only.
4940 * Revert blinding in ECDSA sign and instead make problematic addition
4941 length-invariant. Switch even to fixed-length Montgomery multiplication.
4945 * Change generating and checking of primes so that the error rate of not
4946 being prime depends on the intended use based on the size of the input.
4947 For larger primes this will result in more rounds of Miller-Rabin.
4948 The maximal error rate for primes with more than 1080 bits is lowered
4951 *Kurt Roeckx, Annie Yousar*
4953 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
4957 * Add blinding to ECDSA and DSA signatures to protect against side channel
4958 attacks discovered by Keegan Ryan (NCC Group).
4962 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
4963 now allow empty (zero character) pass phrases.
4967 * Certificate time validation (X509_cmp_time) enforces stricter
4968 compliance with RFC 5280. Fractional seconds and timezone offsets
4969 are no longer allowed.
4973 ### Changes between 1.0.2n and 1.0.2o [27 Mar 2018]
4975 * Constructed ASN.1 types with a recursive definition could exceed the stack
4977 Constructed ASN.1 types with a recursive definition (such as can be found
4978 in PKCS7) could eventually exceed the stack given malicious input with
4979 excessive recursion. This could result in a Denial Of Service attack. There
4980 are no such structures used within SSL/TLS that come from untrusted sources
4981 so this is considered safe.
4983 This issue was reported to OpenSSL on 4th January 2018 by the OSS-fuzz
4989 ### Changes between 1.0.2m and 1.0.2n [7 Dec 2017]
4991 * Read/write after SSL object in error state
4993 OpenSSL 1.0.2 (starting from version 1.0.2b) introduced an "error state"
4994 mechanism. The intent was that if a fatal error occurred during a handshake
4995 then OpenSSL would move into the error state and would immediately fail if
4996 you attempted to continue the handshake. This works as designed for the
4997 explicit handshake functions (SSL_do_handshake(), SSL_accept() and
4998 SSL_connect()), however due to a bug it does not work correctly if
4999 SSL_read() or SSL_write() is called directly. In that scenario, if the
5000 handshake fails then a fatal error will be returned in the initial function
5001 call. If SSL_read()/SSL_write() is subsequently called by the application
5002 for the same SSL object then it will succeed and the data is passed without
5003 being decrypted/encrypted directly from the SSL/TLS record layer.
5005 In order to exploit this issue an application bug would have to be present
5006 that resulted in a call to SSL_read()/SSL_write() being issued after having
5007 already received a fatal error.
5009 This issue was reported to OpenSSL by David Benjamin (Google).
5014 * rsaz_1024_mul_avx2 overflow bug on x86_64
5016 There is an overflow bug in the AVX2 Montgomery multiplication procedure
5017 used in exponentiation with 1024-bit moduli. No EC algorithms are affected.
5018 Analysis suggests that attacks against RSA and DSA as a result of this
5019 defect would be very difficult to perform and are not believed likely.
5020 Attacks against DH1024 are considered just feasible, because most of the
5021 work necessary to deduce information about a private key may be performed
5022 offline. The amount of resources required for such an attack would be
5023 significant. However, for an attack on TLS to be meaningful, the server
5024 would have to share the DH1024 private key among multiple clients, which is
5025 no longer an option since CVE-2016-0701.
5027 This only affects processors that support the AVX2 but not ADX extensions
5028 like Intel Haswell (4th generation).
5030 This issue was reported to OpenSSL by David Benjamin (Google). The issue
5031 was originally found via the OSS-Fuzz project.
5036 ### Changes between 1.0.2l and 1.0.2m [2 Nov 2017]
5038 * bn_sqrx8x_internal carry bug on x86_64
5040 There is a carry propagating bug in the x86_64 Montgomery squaring
5041 procedure. No EC algorithms are affected. Analysis suggests that attacks
5042 against RSA and DSA as a result of this defect would be very difficult to
5043 perform and are not believed likely. Attacks against DH are considered just
5044 feasible (although very difficult) because most of the work necessary to
5045 deduce information about a private key may be performed offline. The amount
5046 of resources required for such an attack would be very significant and
5047 likely only accessible to a limited number of attackers. An attacker would
5048 additionally need online access to an unpatched system using the target
5049 private key in a scenario with persistent DH parameters and a private
5050 key that is shared between multiple clients.
5052 This only affects processors that support the BMI1, BMI2 and ADX extensions
5053 like Intel Broadwell (5th generation) and later or AMD Ryzen.
5055 This issue was reported to OpenSSL by the OSS-Fuzz project.
5060 * Malformed X.509 IPAddressFamily could cause OOB read
5062 If an X.509 certificate has a malformed IPAddressFamily extension,
5063 OpenSSL could do a one-byte buffer overread. The most likely result
5064 would be an erroneous display of the certificate in text format.
5066 This issue was reported to OpenSSL by the OSS-Fuzz project.
5070 ### Changes between 1.0.2k and 1.0.2l [25 May 2017]
5072 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
5073 platform rather than 'mingw'.
5077 ### Changes between 1.0.2j and 1.0.2k [26 Jan 2017]
5079 * Truncated packet could crash via OOB read
5081 If one side of an SSL/TLS path is running on a 32-bit host and a specific
5082 cipher is being used, then a truncated packet can cause that host to
5083 perform an out-of-bounds read, usually resulting in a crash.
5085 This issue was reported to OpenSSL by Robert Święcki of Google.
5090 * BN_mod_exp may produce incorrect results on x86_64
5092 There is a carry propagating bug in the x86_64 Montgomery squaring
5093 procedure. No EC algorithms are affected. Analysis suggests that attacks
5094 against RSA and DSA as a result of this defect would be very difficult to
5095 perform and are not believed likely. Attacks against DH are considered just
5096 feasible (although very difficult) because most of the work necessary to
5097 deduce information about a private key may be performed offline. The amount
5098 of resources required for such an attack would be very significant and
5099 likely only accessible to a limited number of attackers. An attacker would
5100 additionally need online access to an unpatched system using the target
5101 private key in a scenario with persistent DH parameters and a private
5102 key that is shared between multiple clients. For example this can occur by
5103 default in OpenSSL DHE based SSL/TLS ciphersuites. Note: This issue is very
5104 similar to CVE-2015-3193 but must be treated as a separate problem.
5106 This issue was reported to OpenSSL by the OSS-Fuzz project.
5111 * Montgomery multiplication may produce incorrect results
5113 There is a carry propagating bug in the Broadwell-specific Montgomery
5114 multiplication procedure that handles input lengths divisible by, but
5115 longer than 256 bits. Analysis suggests that attacks against RSA, DSA
5116 and DH private keys are impossible. This is because the subroutine in
5117 question is not used in operations with the private key itself and an input
5118 of the attacker's direct choice. Otherwise the bug can manifest itself as
5119 transient authentication and key negotiation failures or reproducible
5120 erroneous outcome of public-key operations with specially crafted input.
5121 Among EC algorithms only Brainpool P-512 curves are affected and one
5122 presumably can attack ECDH key negotiation. Impact was not analyzed in
5123 detail, because pre-requisites for attack are considered unlikely. Namely
5124 multiple clients have to choose the curve in question and the server has to
5125 share the private key among them, neither of which is default behaviour.
5126 Even then only clients that chose the curve will be affected.
5128 This issue was publicly reported as transient failures and was not
5129 initially recognized as a security issue. Thanks to Richard Morgan for
5130 providing reproducible case.
5135 * OpenSSL now fails if it receives an unrecognised record type in TLS1.0
5136 or TLS1.1. Previously this only happened in SSLv3 and TLS1.2. This is to
5137 prevent issues where no progress is being made and the peer continually
5138 sends unrecognised record types, using up resources processing them.
5142 ### Changes between 1.0.2i and 1.0.2j [26 Sep 2016]
5144 * Missing CRL sanity check
5146 A bug fix which included a CRL sanity check was added to OpenSSL 1.1.0
5147 but was omitted from OpenSSL 1.0.2i. As a result any attempt to use
5148 CRLs in OpenSSL 1.0.2i will crash with a null pointer exception.
5150 This issue only affects the OpenSSL 1.0.2i
5155 ### Changes between 1.0.2h and 1.0.2i [22 Sep 2016]
5157 * OCSP Status Request extension unbounded memory growth
5159 A malicious client can send an excessively large OCSP Status Request
5160 extension. If that client continually requests renegotiation, sending a
5161 large OCSP Status Request extension each time, then there will be unbounded
5162 memory growth on the server. This will eventually lead to a Denial Of
5163 Service attack through memory exhaustion. Servers with a default
5164 configuration are vulnerable even if they do not support OCSP. Builds using
5165 the "no-ocsp" build time option are not affected.
5167 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5172 * In order to mitigate the SWEET32 attack, the DES ciphers were moved from
5175 This issue was reported to OpenSSL Karthikeyan Bhargavan and Gaetan
5181 * OOB write in MDC2_Update()
5183 An overflow can occur in MDC2_Update() either if called directly or
5184 through the EVP_DigestUpdate() function using MDC2. If an attacker
5185 is able to supply very large amounts of input data after a previous
5186 call to EVP_EncryptUpdate() with a partial block then a length check
5187 can overflow resulting in a heap corruption.
5189 The amount of data needed is comparable to SIZE_MAX which is impractical
5192 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5197 * Malformed SHA512 ticket DoS
5199 If a server uses SHA512 for TLS session ticket HMAC it is vulnerable to a
5200 DoS attack where a malformed ticket will result in an OOB read which will
5203 The use of SHA512 in TLS session tickets is comparatively rare as it requires
5204 a custom server callback and ticket lookup mechanism.
5206 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5211 * OOB write in BN_bn2dec()
5213 The function BN_bn2dec() does not check the return value of BN_div_word().
5214 This can cause an OOB write if an application uses this function with an
5215 overly large BIGNUM. This could be a problem if an overly large certificate
5216 or CRL is printed out from an untrusted source. TLS is not affected because
5217 record limits will reject an oversized certificate before it is parsed.
5219 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5224 * OOB read in TS_OBJ_print_bio()
5226 The function TS_OBJ_print_bio() misuses OBJ_obj2txt(): the return value is
5227 the total length the OID text representation would use and not the amount
5228 of data written. This will result in OOB reads when large OIDs are
5231 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5236 * Pointer arithmetic undefined behaviour
5238 Avoid some undefined pointer arithmetic
5240 A common idiom in the codebase is to check limits in the following manner:
5243 Where "p" points to some malloc'd data of SIZE bytes and
5246 "len" here could be from some externally supplied data (e.g. from a TLS
5249 The rules of C pointer arithmetic are such that "p + len" is only well
5250 defined where len <= SIZE. Therefore the above idiom is actually
5251 undefined behaviour.
5253 For example this could cause problems if some malloc implementation
5254 provides an address for "p" such that "p + len" actually overflows for
5255 values of len that are too big and therefore p + len < limit.
5257 This issue was reported to OpenSSL by Guido Vranken
5262 * Constant time flag not preserved in DSA signing
5264 Operations in the DSA signing algorithm should run in constant time in
5265 order to avoid side channel attacks. A flaw in the OpenSSL DSA
5266 implementation means that a non-constant time codepath is followed for
5267 certain operations. This has been demonstrated through a cache-timing
5268 attack to be sufficient for an attacker to recover the private DSA key.
5270 This issue was reported by César Pereida (Aalto University), Billy Brumley
5271 (Tampere University of Technology), and Yuval Yarom (The University of
5272 Adelaide and NICTA).
5277 * DTLS buffered message DoS
5279 In a DTLS connection where handshake messages are delivered out-of-order
5280 those messages that OpenSSL is not yet ready to process will be buffered
5281 for later use. Under certain circumstances, a flaw in the logic means that
5282 those messages do not get removed from the buffer even though the handshake
5283 has been completed. An attacker could force up to approx. 15 messages to
5284 remain in the buffer when they are no longer required. These messages will
5285 be cleared when the DTLS connection is closed. The default maximum size for
5286 a message is 100k. Therefore the attacker could force an additional 1500k
5287 to be consumed per connection. By opening many simulataneous connections an
5288 attacker could cause a DoS attack through memory exhaustion.
5290 This issue was reported to OpenSSL by Quan Luo.
5295 * DTLS replay protection DoS
5297 A flaw in the DTLS replay attack protection mechanism means that records
5298 that arrive for future epochs update the replay protection "window" before
5299 the MAC for the record has been validated. This could be exploited by an
5300 attacker by sending a record for the next epoch (which does not have to
5301 decrypt or have a valid MAC), with a very large sequence number. This means
5302 that all subsequent legitimate packets are dropped causing a denial of
5303 service for a specific DTLS connection.
5305 This issue was reported to OpenSSL by the OCAP audit team.
5310 * Certificate message OOB reads
5312 In OpenSSL 1.0.2 and earlier some missing message length checks can result
5313 in OOB reads of up to 2 bytes beyond an allocated buffer. There is a
5314 theoretical DoS risk but this has not been observed in practice on common
5317 The messages affected are client certificate, client certificate request
5318 and server certificate. As a result the attack can only be performed
5319 against a client or a server which enables client authentication.
5321 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5326 ### Changes between 1.0.2g and 1.0.2h [3 May 2016]
5328 * Prevent padding oracle in AES-NI CBC MAC check
5330 A MITM attacker can use a padding oracle attack to decrypt traffic
5331 when the connection uses an AES CBC cipher and the server support
5334 This issue was introduced as part of the fix for Lucky 13 padding
5335 attack ([CVE-2013-0169]). The padding check was rewritten to be in
5336 constant time by making sure that always the same bytes are read and
5337 compared against either the MAC or padding bytes. But it no longer
5338 checked that there was enough data to have both the MAC and padding
5341 This issue was reported by Juraj Somorovsky using TLS-Attacker.
5345 * Fix EVP_EncodeUpdate overflow
5347 An overflow can occur in the EVP_EncodeUpdate() function which is used for
5348 Base64 encoding of binary data. If an attacker is able to supply very large
5349 amounts of input data then a length check can overflow resulting in a heap
5352 Internally to OpenSSL the EVP_EncodeUpdate() function is primarily used by
5353 the `PEM_write_bio*` family of functions. These are mainly used within the
5354 OpenSSL command line applications, so any application which processes data
5355 from an untrusted source and outputs it as a PEM file should be considered
5356 vulnerable to this issue. User applications that call these APIs directly
5357 with large amounts of untrusted data may also be vulnerable.
5359 This issue was reported by Guido Vranken.
5364 * Fix EVP_EncryptUpdate overflow
5366 An overflow can occur in the EVP_EncryptUpdate() function. If an attacker
5367 is able to supply very large amounts of input data after a previous call to
5368 EVP_EncryptUpdate() with a partial block then a length check can overflow
5369 resulting in a heap corruption. Following an analysis of all OpenSSL
5370 internal usage of the EVP_EncryptUpdate() function all usage is one of two
5371 forms. The first form is where the EVP_EncryptUpdate() call is known to be
5372 the first called function after an EVP_EncryptInit(), and therefore that
5373 specific call must be safe. The second form is where the length passed to
5374 EVP_EncryptUpdate() can be seen from the code to be some small value and
5375 therefore there is no possibility of an overflow. Since all instances are
5376 one of these two forms, it is believed that there can be no overflows in
5377 internal code due to this problem. It should be noted that
5378 EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths.
5379 Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances
5380 of these calls have also been analysed too and it is believed there are no
5381 instances in internal usage where an overflow could occur.
5383 This issue was reported by Guido Vranken.
5388 * Prevent ASN.1 BIO excessive memory allocation
5390 When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio()
5391 a short invalid encoding can cause allocation of large amounts of memory
5392 potentially consuming excessive resources or exhausting memory.
5394 Any application parsing untrusted data through d2i BIO functions is
5395 affected. The memory based functions such as d2i_X509() are *not* affected.
5396 Since the memory based functions are used by the TLS library, TLS
5397 applications are not affected.
5399 This issue was reported by Brian Carpenter.
5406 ASN1 Strings that are over 1024 bytes can cause an overread in applications
5407 using the X509_NAME_oneline() function on EBCDIC systems. This could result
5408 in arbitrary stack data being returned in the buffer.
5410 This issue was reported by Guido Vranken.
5415 * Modify behavior of ALPN to invoke callback after SNI/servername
5416 callback, such that updates to the SSL_CTX affect ALPN.
5420 * Remove LOW from the DEFAULT cipher list. This removes singles DES from the
5425 * Only remove the SSLv2 methods with the no-ssl2-method option. When the
5426 methods are enabled and ssl2 is disabled the methods return NULL.
5430 ### Changes between 1.0.2f and 1.0.2g [1 Mar 2016]
5432 * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL.
5433 Builds that are not configured with "enable-weak-ssl-ciphers" will not
5434 provide any "EXPORT" or "LOW" strength ciphers.
5438 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2
5439 is by default disabled at build-time. Builds that are not configured with
5440 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used,
5441 users who want to negotiate SSLv2 via the version-flexible SSLv23_method()
5442 will need to explicitly call either of:
5444 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2);
5446 SSL_clear_options(ssl, SSL_OP_NO_SSLv2);
5448 as appropriate. Even if either of those is used, or the application
5449 explicitly uses the version-specific SSLv2_method() or its client and
5450 server variants, SSLv2 ciphers vulnerable to exhaustive search key
5451 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT
5452 ciphers, and SSLv2 56-bit DES are no longer available.
5457 * Fix a double-free in DSA code
5459 A double free bug was discovered when OpenSSL parses malformed DSA private
5460 keys and could lead to a DoS attack or memory corruption for applications
5461 that receive DSA private keys from untrusted sources. This scenario is
5464 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using
5470 * Disable SRP fake user seed to address a server memory leak.
5472 Add a new method SRP_VBASE_get1_by_user that handles the seed properly.
5474 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
5475 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user
5476 was changed to ignore the "fake user" SRP seed, even if the seed
5479 Users should use SRP_VBASE_get1_by_user instead. Note that in
5480 SRP_VBASE_get1_by_user, caller must free the returned value. Note
5481 also that even though configuring the SRP seed attempts to hide
5482 invalid usernames by continuing the handshake with fake
5483 credentials, this behaviour is not constant time and no strong
5484 guarantees are made that the handshake is indistinguishable from
5485 that of a valid user.
5490 * Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption
5492 In the BN_hex2bn function the number of hex digits is calculated using an
5493 int value `i`. Later `bn_expand` is called with a value of `i * 4`. For
5494 large values of `i` this can result in `bn_expand` not allocating any
5495 memory because `i * 4` is negative. This can leave the internal BIGNUM data
5496 field as NULL leading to a subsequent NULL ptr deref. For very large values
5497 of `i`, the calculation `i * 4` could be a positive value smaller than `i`.
5498 In this case memory is allocated to the internal BIGNUM data field, but it
5499 is insufficiently sized leading to heap corruption. A similar issue exists
5500 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn
5501 is ever called by user applications with very large untrusted hex/dec data.
5502 This is anticipated to be a rare occurrence.
5504 All OpenSSL internal usage of these functions use data that is not expected
5505 to be untrusted, e.g. config file data or application command line
5506 arguments. If user developed applications generate config file data based
5507 on untrusted data then it is possible that this could also lead to security
5508 consequences. This is also anticipated to be rare.
5510 This issue was reported to OpenSSL by Guido Vranken.
5515 * Fix memory issues in `BIO_*printf` functions
5517 The internal `fmtstr` function used in processing a "%s" format string in
5518 the `BIO_*printf` functions could overflow while calculating the length of a
5519 string and cause an OOB read when printing very long strings.
5521 Additionally the internal `doapr_outch` function can attempt to write to an
5522 OOB memory location (at an offset from the NULL pointer) in the event of a
5523 memory allocation failure. In 1.0.2 and below this could be caused where
5524 the size of a buffer to be allocated is greater than INT_MAX. E.g. this
5525 could be in processing a very long "%s" format string. Memory leaks can
5528 The first issue may mask the second issue dependent on compiler behaviour.
5529 These problems could enable attacks where large amounts of untrusted data
5530 is passed to the `BIO_*printf` functions. If applications use these functions
5531 in this way then they could be vulnerable. OpenSSL itself uses these
5532 functions when printing out human-readable dumps of ASN.1 data. Therefore
5533 applications that print this data could be vulnerable if the data is from
5534 untrusted sources. OpenSSL command line applications could also be
5535 vulnerable where they print out ASN.1 data, or if untrusted data is passed
5536 as command line arguments.
5538 Libssl is not considered directly vulnerable. Additionally certificates etc
5539 received via remote connections via libssl are also unlikely to be able to
5540 trigger these issues because of message size limits enforced within libssl.
5542 This issue was reported to OpenSSL Guido Vranken.
5547 * Side channel attack on modular exponentiation
5549 A side-channel attack was found which makes use of cache-bank conflicts on
5550 the Intel Sandy-Bridge microarchitecture which could lead to the recovery
5551 of RSA keys. The ability to exploit this issue is limited as it relies on
5552 an attacker who has control of code in a thread running on the same
5553 hyper-threaded core as the victim thread which is performing decryptions.
5555 This issue was reported to OpenSSL by Yuval Yarom, The University of
5556 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and
5557 Nadia Heninger, University of Pennsylvania with more information at
5558 <http://cachebleed.info>.
5563 * Change the `req` command to generate a 2048-bit RSA/DSA key by default,
5564 if no keysize is specified with default_bits. This fixes an
5565 omission in an earlier change that changed all RSA/DSA key generation
5566 commands to use 2048 bits by default.
5570 ### Changes between 1.0.2e and 1.0.2f [28 Jan 2016]
5572 * DH small subgroups
5574 Historically OpenSSL only ever generated DH parameters based on "safe"
5575 primes. More recently (in version 1.0.2) support was provided for
5576 generating X9.42 style parameter files such as those required for RFC 5114
5577 support. The primes used in such files may not be "safe". Where an
5578 application is using DH configured with parameters based on primes that are
5579 not "safe" then an attacker could use this fact to find a peer's private
5580 DH exponent. This attack requires that the attacker complete multiple
5581 handshakes in which the peer uses the same private DH exponent. For example
5582 this could be used to discover a TLS server's private DH exponent if it's
5583 reusing the private DH exponent or it's using a static DH ciphersuite.
5585 OpenSSL provides the option SSL_OP_SINGLE_DH_USE for ephemeral DH (DHE) in
5586 TLS. It is not on by default. If the option is not set then the server
5587 reuses the same private DH exponent for the life of the server process and
5588 would be vulnerable to this attack. It is believed that many popular
5589 applications do set this option and would therefore not be at risk.
5591 The fix for this issue adds an additional check where a "q" parameter is
5592 available (as is the case in X9.42 based parameters). This detects the
5593 only known attack, and is the only possible defense for static DH
5594 ciphersuites. This could have some performance impact.
5596 Additionally the SSL_OP_SINGLE_DH_USE option has been switched on by
5597 default and cannot be disabled. This could have some performance impact.
5599 This issue was reported to OpenSSL by Antonio Sanso (Adobe).
5604 * SSLv2 doesn't block disabled ciphers
5606 A malicious client can negotiate SSLv2 ciphers that have been disabled on
5607 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
5608 been disabled, provided that the SSLv2 protocol was not also disabled via
5611 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
5612 and Sebastian Schinzel.
5617 ### Changes between 1.0.2d and 1.0.2e [3 Dec 2015]
5619 * BN_mod_exp may produce incorrect results on x86_64
5621 There is a carry propagating bug in the x86_64 Montgomery squaring
5622 procedure. No EC algorithms are affected. Analysis suggests that attacks
5623 against RSA and DSA as a result of this defect would be very difficult to
5624 perform and are not believed likely. Attacks against DH are considered just
5625 feasible (although very difficult) because most of the work necessary to
5626 deduce information about a private key may be performed offline. The amount
5627 of resources required for such an attack would be very significant and
5628 likely only accessible to a limited number of attackers. An attacker would
5629 additionally need online access to an unpatched system using the target
5630 private key in a scenario with persistent DH parameters and a private
5631 key that is shared between multiple clients. For example this can occur by
5632 default in OpenSSL DHE based SSL/TLS ciphersuites.
5634 This issue was reported to OpenSSL by Hanno Böck.
5639 * Certificate verify crash with missing PSS parameter
5641 The signature verification routines will crash with a NULL pointer
5642 dereference if presented with an ASN.1 signature using the RSA PSS
5643 algorithm and absent mask generation function parameter. Since these
5644 routines are used to verify certificate signature algorithms this can be
5645 used to crash any certificate verification operation and exploited in a
5646 DoS attack. Any application which performs certificate verification is
5647 vulnerable including OpenSSL clients and servers which enable client
5650 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
5655 * X509_ATTRIBUTE memory leak
5657 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
5658 memory. This structure is used by the PKCS#7 and CMS routines so any
5659 application which reads PKCS#7 or CMS data from untrusted sources is
5660 affected. SSL/TLS is not affected.
5662 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
5668 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
5669 This changes the decoding behaviour for some invalid messages,
5670 though the change is mostly in the more lenient direction, and
5671 legacy behaviour is preserved as much as possible.
5675 * In DSA_generate_parameters_ex, if the provided seed is too short,
5678 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
5680 ### Changes between 1.0.2c and 1.0.2d [9 Jul 2015]
5682 * Alternate chains certificate forgery
5684 During certificate verification, OpenSSL will attempt to find an
5685 alternative certificate chain if the first attempt to build such a chain
5686 fails. An error in the implementation of this logic can mean that an
5687 attacker could cause certain checks on untrusted certificates to be
5688 bypassed, such as the CA flag, enabling them to use a valid leaf
5689 certificate to act as a CA and "issue" an invalid certificate.
5691 This issue was reported to OpenSSL by Adam Langley/David Benjamin
5696 ### Changes between 1.0.2b and 1.0.2c [12 Jun 2015]
5698 * Fix HMAC ABI incompatibility. The previous version introduced an ABI
5699 incompatibility in the handling of HMAC. The previous ABI has now been
5704 ### Changes between 1.0.2a and 1.0.2b [11 Jun 2015]
5706 * Malformed ECParameters causes infinite loop
5708 When processing an ECParameters structure OpenSSL enters an infinite loop
5709 if the curve specified is over a specially malformed binary polynomial
5712 This can be used to perform denial of service against any
5713 system which processes public keys, certificate requests or
5714 certificates. This includes TLS clients and TLS servers with
5715 client authentication enabled.
5717 This issue was reported to OpenSSL by Joseph Barr-Pixton.
5722 * Exploitable out-of-bounds read in X509_cmp_time
5724 X509_cmp_time does not properly check the length of the ASN1_TIME
5725 string and can read a few bytes out of bounds. In addition,
5726 X509_cmp_time accepts an arbitrary number of fractional seconds in the
5729 An attacker can use this to craft malformed certificates and CRLs of
5730 various sizes and potentially cause a segmentation fault, resulting in
5731 a DoS on applications that verify certificates or CRLs. TLS clients
5732 that verify CRLs are affected. TLS clients and servers with client
5733 authentication enabled may be affected if they use custom verification
5736 This issue was reported to OpenSSL by Robert Swiecki (Google), and
5737 independently by Hanno Böck.
5742 * PKCS7 crash with missing EnvelopedContent
5744 The PKCS#7 parsing code does not handle missing inner EncryptedContent
5745 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
5746 with missing content and trigger a NULL pointer dereference on parsing.
5748 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
5749 structures from untrusted sources are affected. OpenSSL clients and
5750 servers are not affected.
5752 This issue was reported to OpenSSL by Michal Zalewski (Google).
5757 * CMS verify infinite loop with unknown hash function
5759 When verifying a signedData message the CMS code can enter an infinite loop
5760 if presented with an unknown hash function OID. This can be used to perform
5761 denial of service against any system which verifies signedData messages using
5763 This issue was reported to OpenSSL by Johannes Bauer.
5768 * Race condition handling NewSessionTicket
5770 If a NewSessionTicket is received by a multi-threaded client when attempting to
5771 reuse a previous ticket then a race condition can occur potentially leading to
5772 a double free of the ticket data.
5777 * Only support 256-bit or stronger elliptic curves with the
5778 'ecdh_auto' setting (server) or by default (client). Of supported
5779 curves, prefer P-256 (both).
5783 ### Changes between 1.0.2 and 1.0.2a [19 Mar 2015]
5785 * ClientHello sigalgs DoS fix
5787 If a client connects to an OpenSSL 1.0.2 server and renegotiates with an
5788 invalid signature algorithms extension a NULL pointer dereference will
5789 occur. This can be exploited in a DoS attack against the server.
5791 This issue was was reported to OpenSSL by David Ramos of Stanford
5795 *Stephen Henson and Matt Caswell*
5797 * Multiblock corrupted pointer fix
5799 OpenSSL 1.0.2 introduced the "multiblock" performance improvement. This
5800 feature only applies on 64 bit x86 architecture platforms that support AES
5801 NI instructions. A defect in the implementation of "multiblock" can cause
5802 OpenSSL's internal write buffer to become incorrectly set to NULL when
5803 using non-blocking IO. Typically, when the user application is using a
5804 socket BIO for writing, this will only result in a failed connection.
5805 However if some other BIO is used then it is likely that a segmentation
5806 fault will be triggered, thus enabling a potential DoS attack.
5808 This issue was reported to OpenSSL by Daniel Danner and Rainer Mueller.
5813 * Segmentation fault in DTLSv1_listen fix
5815 The DTLSv1_listen function is intended to be stateless and processes the
5816 initial ClientHello from many peers. It is common for user code to loop
5817 over the call to DTLSv1_listen until a valid ClientHello is received with
5818 an associated cookie. A defect in the implementation of DTLSv1_listen means
5819 that state is preserved in the SSL object from one invocation to the next
5820 that can lead to a segmentation fault. Errors processing the initial
5821 ClientHello can trigger this scenario. An example of such an error could be
5822 that a DTLS1.0 only client is attempting to connect to a DTLS1.2 only
5825 This issue was reported to OpenSSL by Per Allansson.
5830 * Segmentation fault in ASN1_TYPE_cmp fix
5832 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
5833 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
5834 certificate signature algorithm consistency this can be used to crash any
5835 certificate verification operation and exploited in a DoS attack. Any
5836 application which performs certificate verification is vulnerable including
5837 OpenSSL clients and servers which enable client authentication.
5842 * Segmentation fault for invalid PSS parameters fix
5844 The signature verification routines will crash with a NULL pointer
5845 dereference if presented with an ASN.1 signature using the RSA PSS
5846 algorithm and invalid parameters. Since these routines are used to verify
5847 certificate signature algorithms this can be used to crash any
5848 certificate verification operation and exploited in a DoS attack. Any
5849 application which performs certificate verification is vulnerable including
5850 OpenSSL clients and servers which enable client authentication.
5852 This issue was was reported to OpenSSL by Brian Carpenter.
5857 * ASN.1 structure reuse memory corruption fix
5859 Reusing a structure in ASN.1 parsing may allow an attacker to cause
5860 memory corruption via an invalid write. Such reuse is and has been
5861 strongly discouraged and is believed to be rare.
5863 Applications that parse structures containing CHOICE or ANY DEFINED BY
5864 components may be affected. Certificate parsing (d2i_X509 and related
5865 functions) are however not affected. OpenSSL clients and servers are
5871 * PKCS7 NULL pointer dereferences fix
5873 The PKCS#7 parsing code does not handle missing outer ContentInfo
5874 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
5875 missing content and trigger a NULL pointer dereference on parsing.
5877 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
5878 otherwise parse PKCS#7 structures from untrusted sources are
5879 affected. OpenSSL clients and servers are not affected.
5881 This issue was reported to OpenSSL by Michal Zalewski (Google).
5886 * DoS via reachable assert in SSLv2 servers fix
5888 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
5889 servers that both support SSLv2 and enable export cipher suites by sending
5890 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
5892 This issue was discovered by Sean Burford (Google) and Emilia Käsper
5893 (OpenSSL development team).
5898 * Empty CKE with client auth and DHE fix
5900 If client auth is used then a server can seg fault in the event of a DHE
5901 ciphersuite being selected and a zero length ClientKeyExchange message
5902 being sent by the client. This could be exploited in a DoS attack.
5907 * Handshake with unseeded PRNG fix
5909 Under certain conditions an OpenSSL 1.0.2 client can complete a handshake
5910 with an unseeded PRNG. The conditions are:
5911 - The client is on a platform where the PRNG has not been seeded
5912 automatically, and the user has not seeded manually
5913 - A protocol specific client method version has been used (i.e. not
5914 SSL_client_methodv23)
5915 - A ciphersuite is used that does not require additional random data from
5916 the PRNG beyond the initial ClientHello client random (e.g. PSK-RC4-SHA).
5918 If the handshake succeeds then the client random that has been used will
5919 have been generated from a PRNG with insufficient entropy and therefore the
5920 output may be predictable.
5922 For example using the following command with an unseeded openssl will
5923 succeed on an unpatched platform:
5925 openssl s_client -psk 1a2b3c4d -tls1_2 -cipher PSK-RC4-SHA
5930 * Use After Free following d2i_ECPrivatekey error fix
5932 A malformed EC private key file consumed via the d2i_ECPrivateKey function
5933 could cause a use after free condition. This, in turn, could cause a double
5934 free in several private key parsing functions (such as d2i_PrivateKey
5935 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
5936 for applications that receive EC private keys from untrusted
5937 sources. This scenario is considered rare.
5939 This issue was discovered by the BoringSSL project and fixed in their
5945 * X509_to_X509_REQ NULL pointer deref fix
5947 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
5948 the certificate key is invalid. This function is rarely used in practice.
5950 This issue was discovered by Brian Carpenter.
5955 * Removed the export ciphers from the DEFAULT ciphers
5959 ### Changes between 1.0.1l and 1.0.2 [22 Jan 2015]
5961 * Facilitate "universal" ARM builds targeting range of ARM ISAs, e.g.
5962 ARMv5 through ARMv8, as opposite to "locking" it to single one.
5963 So far those who have to target multiple platforms would compromise
5964 and argue that binary targeting say ARMv5 would still execute on
5965 ARMv8. "Universal" build resolves this compromise by providing
5966 near-optimal performance even on newer platforms.
5970 * Accelerated NIST P-256 elliptic curve implementation for x86_64
5971 (other platforms pending).
5973 *Shay Gueron & Vlad Krasnov (Intel Corp), Andy Polyakov*
5975 * Add support for the SignedCertificateTimestampList certificate and
5976 OCSP response extensions from RFC6962.
5980 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
5981 for corner cases. (Certain input points at infinity could lead to
5982 bogus results, with non-infinity inputs mapped to infinity too.)
5986 * Initial support for PowerISA 2.0.7, first implemented in POWER8.
5987 This covers AES, SHA256/512 and GHASH. "Initial" means that most
5988 common cases are optimized and there still is room for further
5989 improvements. Vector Permutation AES for Altivec is also added.
5993 * Add support for little-endian ppc64 Linux target.
5995 *Marcelo Cerri (IBM)*
5997 * Initial support for AMRv8 ISA crypto extensions. This covers AES,
5998 SHA1, SHA256 and GHASH. "Initial" means that most common cases
5999 are optimized and there still is room for further improvements.
6000 Both 32- and 64-bit modes are supported.
6002 *Andy Polyakov, Ard Biesheuvel (Linaro)*
6004 * Improved ARMv7 NEON support.
6008 * Support for SPARC Architecture 2011 crypto extensions, first
6009 implemented in SPARC T4. This covers AES, DES, Camellia, SHA1,
6010 SHA256/512, MD5, GHASH and modular exponentiation.
6012 *Andy Polyakov, David Miller*
6014 * Accelerated modular exponentiation for Intel processors, a.k.a.
6017 *Shay Gueron & Vlad Krasnov (Intel Corp)*
6019 * Support for new and upcoming Intel processors, including AVX2,
6020 BMI and SHA ISA extensions. This includes additional "stitched"
6021 implementations, AESNI-SHA256 and GCM, and multi-buffer support
6024 This work was sponsored by Intel Corp.
6028 * Support for DTLS 1.2. This adds two sets of DTLS methods: DTLS_*_method()
6029 supports both DTLS 1.2 and 1.0 and should use whatever version the peer
6030 supports and DTLSv1_2_*_method() which supports DTLS 1.2 only.
6034 * Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
6035 this fixes a limitation in previous versions of OpenSSL.
6039 * Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
6040 MGF1 digest and OAEP label.
6044 * Add EVP support for key wrapping algorithms, to avoid problems with
6045 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
6046 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
6047 algorithms and include tests cases.
6051 * Add functions to allocate and set the fields of an ECDSA_METHOD
6054 *Douglas E. Engert, Steve Henson*
6056 * New functions OPENSSL_gmtime_diff and ASN1_TIME_diff to find the
6057 difference in days and seconds between two tm or ASN1_TIME structures.
6061 * Add -rev test option to s_server to just reverse order of characters
6062 received by client and send back to server. Also prints an abbreviated
6063 summary of the connection parameters.
6067 * New option -brief for s_client and s_server to print out a brief summary
6068 of connection parameters.
6072 * Add callbacks for arbitrary TLS extensions.
6074 *Trevor Perrin <trevp@trevp.net> and Ben Laurie*
6076 * New option -crl_download in several openssl utilities to download CRLs
6077 from CRLDP extension in certificates.
6081 * New options -CRL and -CRLform for s_client and s_server for CRLs.
6085 * New function X509_CRL_diff to generate a delta CRL from the difference
6086 of two full CRLs. Add support to "crl" utility.
6090 * New functions to set lookup_crls function and to retrieve
6091 X509_STORE from X509_STORE_CTX.
6095 * Print out deprecated issuer and subject unique ID fields in
6100 * Extend OCSP I/O functions so they can be used for simple general purpose
6101 HTTP as well as OCSP. New wrapper function which can be used to download
6102 CRLs using the OCSP API.
6106 * Delegate command line handling in s_client/s_server to SSL_CONF APIs.
6110 * `SSL_CONF*` functions. These provide a common framework for application
6111 configuration using configuration files or command lines.
6115 * SSL/TLS tracing code. This parses out SSL/TLS records using the
6116 message callback and prints the results. Needs compile time option
6117 "enable-ssl-trace". New options to s_client and s_server to enable
6122 * New ctrl and macro to retrieve supported points extensions.
6123 Print out extension in s_server and s_client.
6127 * New functions to retrieve certificate signature and signature
6132 * Add functions to retrieve and manipulate the raw cipherlist sent by a
6137 * New Suite B modes for TLS code. These use and enforce the requirements
6138 of RFC6460: restrict ciphersuites, only permit Suite B algorithms and
6139 only use Suite B curves. The Suite B modes can be set by using the
6140 strings "SUITEB128", "SUITEB192" or "SUITEB128ONLY" for the cipherstring.
6144 * New chain verification flags for Suite B levels of security. Check
6145 algorithms are acceptable when flags are set in X509_verify_cert.
6149 * Make tls1_check_chain return a set of flags indicating checks passed
6150 by a certificate chain. Add additional tests to handle client
6151 certificates: checks for matching certificate type and issuer name
6156 * If an attempt is made to use a signature algorithm not in the peer
6157 preference list abort the handshake. If client has no suitable
6158 signature algorithms in response to a certificate request do not
6159 use the certificate.
6163 * If server EC tmp key is not in client preference list abort handshake.
6167 * Add support for certificate stores in CERT structure. This makes it
6168 possible to have different stores per SSL structure or one store in
6169 the parent SSL_CTX. Include distinct stores for certificate chain
6170 verification and chain building. New ctrl SSL_CTRL_BUILD_CERT_CHAIN
6171 to build and store a certificate chain in CERT structure: returning
6172 an error if the chain cannot be built: this will allow applications
6173 to test if a chain is correctly configured.
6175 Note: if the CERT based stores are not set then the parent SSL_CTX
6176 store is used to retain compatibility with existing behaviour.
6180 * New function ssl_set_client_disabled to set a ciphersuite disabled
6181 mask based on the current session, check mask when sending client
6182 hello and checking the requested ciphersuite.
6186 * New ctrls to retrieve and set certificate types in a certificate
6187 request message. Print out received values in s_client. If certificate
6188 types is not set with custom values set sensible values based on
6189 supported signature algorithms.
6193 * Support for distinct client and server supported signature algorithms.
6197 * Add certificate callback. If set this is called whenever a certificate
6198 is required by client or server. An application can decide which
6199 certificate chain to present based on arbitrary criteria: for example
6200 supported signature algorithms. Add very simple example to s_server.
6201 This fixes many of the problems and restrictions of the existing client
6202 certificate callback: for example you can now clear an existing
6203 certificate and specify the whole chain.
6207 * Add new "valid_flags" field to CERT_PKEY structure which determines what
6208 the certificate can be used for (if anything). Set valid_flags field
6209 in new tls1_check_chain function. Simplify ssl_set_cert_masks which used
6210 to have similar checks in it.
6212 Add new "cert_flags" field to CERT structure and include a "strict mode".
6213 This enforces some TLS certificate requirements (such as only permitting
6214 certificate signature algorithms contained in the supported algorithms
6215 extension) which some implementations ignore: this option should be used
6216 with caution as it could cause interoperability issues.
6220 * Update and tidy signature algorithm extension processing. Work out
6221 shared signature algorithms based on preferences and peer algorithms
6222 and print them out in s_client and s_server. Abort handshake if no
6223 shared signature algorithms.
6227 * Add new functions to allow customised supported signature algorithms
6228 for SSL and SSL_CTX structures. Add options to s_client and s_server
6233 * New function SSL_certs_clear() to delete all references to certificates
6234 from an SSL structure. Before this once a certificate had been added
6235 it couldn't be removed.
6239 * Integrate hostname, email address and IP address checking with certificate
6240 verification. New verify options supporting checking in openssl utility.
6244 * Fixes and wildcard matching support to hostname and email checking
6245 functions. Add manual page.
6247 *Florian Weimer (Red Hat Product Security Team)*
6249 * New functions to check a hostname email or IP address against a
6250 certificate. Add options x509 utility to print results of checks against
6255 * Fix OCSP checking.
6257 *Rob Stradling <rob.stradling@comodo.com> and Ben Laurie*
6259 * Initial experimental support for explicitly trusted non-root CAs.
6260 OpenSSL still tries to build a complete chain to a root but if an
6261 intermediate CA has a trust setting included that is used. The first
6262 setting is used: whether to trust (e.g., -addtrust option to the x509
6267 * Add -trusted_first option which attempts to find certificates in the
6268 trusted store even if an untrusted chain is also supplied.
6272 * MIPS assembly pack updates: support for MIPS32r2 and SmartMIPS ASE,
6273 platform support for Linux and Android.
6277 * Support for linux-x32, ILP32 environment in x86_64 framework.
6281 * Experimental multi-implementation support for FIPS capable OpenSSL.
6282 When in FIPS mode the approved implementations are used as normal,
6283 when not in FIPS mode the internal unapproved versions are used instead.
6284 This means that the FIPS capable OpenSSL isn't forced to use the
6285 (often lower performance) FIPS implementations outside FIPS mode.
6289 * Transparently support X9.42 DH parameters when calling
6290 PEM_read_bio_DHparameters. This means existing applications can handle
6291 the new parameter format automatically.
6295 * Initial experimental support for X9.42 DH parameter format: mainly
6296 to support use of 'q' parameter for RFC5114 parameters.
6300 * Add DH parameters from RFC5114 including test data to dhtest.
6304 * Support for automatic EC temporary key parameter selection. If enabled
6305 the most preferred EC parameters are automatically used instead of
6306 hardcoded fixed parameters. Now a server just has to call:
6307 SSL_CTX_set_ecdh_auto(ctx, 1) and the server will automatically
6308 support ECDH and use the most appropriate parameters.
6312 * Enhance and tidy EC curve and point format TLS extension code. Use
6313 static structures instead of allocation if default values are used.
6314 New ctrls to set curves we wish to support and to retrieve shared curves.
6315 Print out shared curves in s_server. New options to s_server and s_client
6316 to set list of supported curves.
6320 * New ctrls to retrieve supported signature algorithms and
6321 supported curve values as an array of NIDs. Extend openssl utility
6322 to print out received values.
6326 * Add new APIs EC_curve_nist2nid and EC_curve_nid2nist which convert
6327 between NIDs and the more common NIST names such as "P-256". Enhance
6328 ecparam utility and ECC method to recognise the NIST names for curves.
6332 * Enhance SSL/TLS certificate chain handling to support different
6333 chains for each certificate instead of one chain in the parent SSL_CTX.
6337 * Support for fixed DH ciphersuite client authentication: where both
6338 server and client use DH certificates with common parameters.
6342 * Support for fixed DH ciphersuites: those requiring DH server
6347 * New function i2d_re_X509_tbs for re-encoding the TBS portion of
6349 Note: Related 1.0.2-beta specific macros X509_get_cert_info,
6350 X509_CINF_set_modified, X509_CINF_get_issuer, X509_CINF_get_extensions and
6351 X509_CINF_get_signature were reverted post internal team review.
6356 ### Changes between 1.0.1t and 1.0.1u [22 Sep 2016]
6358 * OCSP Status Request extension unbounded memory growth
6360 A malicious client can send an excessively large OCSP Status Request
6361 extension. If that client continually requests renegotiation, sending a
6362 large OCSP Status Request extension each time, then there will be unbounded
6363 memory growth on the server. This will eventually lead to a Denial Of
6364 Service attack through memory exhaustion. Servers with a default
6365 configuration are vulnerable even if they do not support OCSP. Builds using
6366 the "no-ocsp" build time option are not affected.
6368 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6373 * In order to mitigate the SWEET32 attack, the DES ciphers were moved from
6376 This issue was reported to OpenSSL Karthikeyan Bhargavan and Gaetan
6382 * OOB write in MDC2_Update()
6384 An overflow can occur in MDC2_Update() either if called directly or
6385 through the EVP_DigestUpdate() function using MDC2. If an attacker
6386 is able to supply very large amounts of input data after a previous
6387 call to EVP_EncryptUpdate() with a partial block then a length check
6388 can overflow resulting in a heap corruption.
6390 The amount of data needed is comparable to SIZE_MAX which is impractical
6393 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6398 * Malformed SHA512 ticket DoS
6400 If a server uses SHA512 for TLS session ticket HMAC it is vulnerable to a
6401 DoS attack where a malformed ticket will result in an OOB read which will
6404 The use of SHA512 in TLS session tickets is comparatively rare as it requires
6405 a custom server callback and ticket lookup mechanism.
6407 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6412 * OOB write in BN_bn2dec()
6414 The function BN_bn2dec() does not check the return value of BN_div_word().
6415 This can cause an OOB write if an application uses this function with an
6416 overly large BIGNUM. This could be a problem if an overly large certificate
6417 or CRL is printed out from an untrusted source. TLS is not affected because
6418 record limits will reject an oversized certificate before it is parsed.
6420 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6425 * OOB read in TS_OBJ_print_bio()
6427 The function TS_OBJ_print_bio() misuses OBJ_obj2txt(): the return value is
6428 the total length the OID text representation would use and not the amount
6429 of data written. This will result in OOB reads when large OIDs are
6432 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6437 * Pointer arithmetic undefined behaviour
6439 Avoid some undefined pointer arithmetic
6441 A common idiom in the codebase is to check limits in the following manner:
6444 Where "p" points to some malloc'd data of SIZE bytes and
6447 "len" here could be from some externally supplied data (e.g. from a TLS
6450 The rules of C pointer arithmetic are such that "p + len" is only well
6451 defined where len <= SIZE. Therefore the above idiom is actually
6452 undefined behaviour.
6454 For example this could cause problems if some malloc implementation
6455 provides an address for "p" such that "p + len" actually overflows for
6456 values of len that are too big and therefore p + len < limit.
6458 This issue was reported to OpenSSL by Guido Vranken
6463 * Constant time flag not preserved in DSA signing
6465 Operations in the DSA signing algorithm should run in constant time in
6466 order to avoid side channel attacks. A flaw in the OpenSSL DSA
6467 implementation means that a non-constant time codepath is followed for
6468 certain operations. This has been demonstrated through a cache-timing
6469 attack to be sufficient for an attacker to recover the private DSA key.
6471 This issue was reported by César Pereida (Aalto University), Billy Brumley
6472 (Tampere University of Technology), and Yuval Yarom (The University of
6473 Adelaide and NICTA).
6478 * DTLS buffered message DoS
6480 In a DTLS connection where handshake messages are delivered out-of-order
6481 those messages that OpenSSL is not yet ready to process will be buffered
6482 for later use. Under certain circumstances, a flaw in the logic means that
6483 those messages do not get removed from the buffer even though the handshake
6484 has been completed. An attacker could force up to approx. 15 messages to
6485 remain in the buffer when they are no longer required. These messages will
6486 be cleared when the DTLS connection is closed. The default maximum size for
6487 a message is 100k. Therefore the attacker could force an additional 1500k
6488 to be consumed per connection. By opening many simulataneous connections an
6489 attacker could cause a DoS attack through memory exhaustion.
6491 This issue was reported to OpenSSL by Quan Luo.
6496 * DTLS replay protection DoS
6498 A flaw in the DTLS replay attack protection mechanism means that records
6499 that arrive for future epochs update the replay protection "window" before
6500 the MAC for the record has been validated. This could be exploited by an
6501 attacker by sending a record for the next epoch (which does not have to
6502 decrypt or have a valid MAC), with a very large sequence number. This means
6503 that all subsequent legitimate packets are dropped causing a denial of
6504 service for a specific DTLS connection.
6506 This issue was reported to OpenSSL by the OCAP audit team.
6511 * Certificate message OOB reads
6513 In OpenSSL 1.0.2 and earlier some missing message length checks can result
6514 in OOB reads of up to 2 bytes beyond an allocated buffer. There is a
6515 theoretical DoS risk but this has not been observed in practice on common
6518 The messages affected are client certificate, client certificate request
6519 and server certificate. As a result the attack can only be performed
6520 against a client or a server which enables client authentication.
6522 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6527 ### Changes between 1.0.1s and 1.0.1t [3 May 2016]
6529 * Prevent padding oracle in AES-NI CBC MAC check
6531 A MITM attacker can use a padding oracle attack to decrypt traffic
6532 when the connection uses an AES CBC cipher and the server support
6535 This issue was introduced as part of the fix for Lucky 13 padding
6536 attack ([CVE-2013-0169]). The padding check was rewritten to be in
6537 constant time by making sure that always the same bytes are read and
6538 compared against either the MAC or padding bytes. But it no longer
6539 checked that there was enough data to have both the MAC and padding
6542 This issue was reported by Juraj Somorovsky using TLS-Attacker.
6547 * Fix EVP_EncodeUpdate overflow
6549 An overflow can occur in the EVP_EncodeUpdate() function which is used for
6550 Base64 encoding of binary data. If an attacker is able to supply very large
6551 amounts of input data then a length check can overflow resulting in a heap
6554 Internally to OpenSSL the EVP_EncodeUpdate() function is primarly used by
6555 the `PEM_write_bio*` family of functions. These are mainly used within the
6556 OpenSSL command line applications, so any application which processes data
6557 from an untrusted source and outputs it as a PEM file should be considered
6558 vulnerable to this issue. User applications that call these APIs directly
6559 with large amounts of untrusted data may also be vulnerable.
6561 This issue was reported by Guido Vranken.
6566 * Fix EVP_EncryptUpdate overflow
6568 An overflow can occur in the EVP_EncryptUpdate() function. If an attacker
6569 is able to supply very large amounts of input data after a previous call to
6570 EVP_EncryptUpdate() with a partial block then a length check can overflow
6571 resulting in a heap corruption. Following an analysis of all OpenSSL
6572 internal usage of the EVP_EncryptUpdate() function all usage is one of two
6573 forms. The first form is where the EVP_EncryptUpdate() call is known to be
6574 the first called function after an EVP_EncryptInit(), and therefore that
6575 specific call must be safe. The second form is where the length passed to
6576 EVP_EncryptUpdate() can be seen from the code to be some small value and
6577 therefore there is no possibility of an overflow. Since all instances are
6578 one of these two forms, it is believed that there can be no overflows in
6579 internal code due to this problem. It should be noted that
6580 EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths.
6581 Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances
6582 of these calls have also been analysed too and it is believed there are no
6583 instances in internal usage where an overflow could occur.
6585 This issue was reported by Guido Vranken.
6590 * Prevent ASN.1 BIO excessive memory allocation
6592 When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio()
6593 a short invalid encoding can casuse allocation of large amounts of memory
6594 potentially consuming excessive resources or exhausting memory.
6596 Any application parsing untrusted data through d2i BIO functions is
6597 affected. The memory based functions such as d2i_X509() are *not* affected.
6598 Since the memory based functions are used by the TLS library, TLS
6599 applications are not affected.
6601 This issue was reported by Brian Carpenter.
6608 ASN1 Strings that are over 1024 bytes can cause an overread in applications
6609 using the X509_NAME_oneline() function on EBCDIC systems. This could result
6610 in arbitrary stack data being returned in the buffer.
6612 This issue was reported by Guido Vranken.
6617 * Modify behavior of ALPN to invoke callback after SNI/servername
6618 callback, such that updates to the SSL_CTX affect ALPN.
6622 * Remove LOW from the DEFAULT cipher list. This removes singles DES from the
6627 * Only remove the SSLv2 methods with the no-ssl2-method option. When the
6628 methods are enabled and ssl2 is disabled the methods return NULL.
6632 ### Changes between 1.0.1r and 1.0.1s [1 Mar 2016]
6634 * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL.
6635 Builds that are not configured with "enable-weak-ssl-ciphers" will not
6636 provide any "EXPORT" or "LOW" strength ciphers.
6640 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2
6641 is by default disabled at build-time. Builds that are not configured with
6642 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used,
6643 users who want to negotiate SSLv2 via the version-flexible SSLv23_method()
6644 will need to explicitly call either of:
6646 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2);
6648 SSL_clear_options(ssl, SSL_OP_NO_SSLv2);
6650 as appropriate. Even if either of those is used, or the application
6651 explicitly uses the version-specific SSLv2_method() or its client and
6652 server variants, SSLv2 ciphers vulnerable to exhaustive search key
6653 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT
6654 ciphers, and SSLv2 56-bit DES are no longer available.
6659 * Fix a double-free in DSA code
6661 A double free bug was discovered when OpenSSL parses malformed DSA private
6662 keys and could lead to a DoS attack or memory corruption for applications
6663 that receive DSA private keys from untrusted sources. This scenario is
6666 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using
6672 * Disable SRP fake user seed to address a server memory leak.
6674 Add a new method SRP_VBASE_get1_by_user that handles the seed properly.
6676 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
6677 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user
6678 was changed to ignore the "fake user" SRP seed, even if the seed
6681 Users should use SRP_VBASE_get1_by_user instead. Note that in
6682 SRP_VBASE_get1_by_user, caller must free the returned value. Note
6683 also that even though configuring the SRP seed attempts to hide
6684 invalid usernames by continuing the handshake with fake
6685 credentials, this behaviour is not constant time and no strong
6686 guarantees are made that the handshake is indistinguishable from
6687 that of a valid user.
6692 * Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption
6694 In the BN_hex2bn function the number of hex digits is calculated using an
6695 int value `i`. Later `bn_expand` is called with a value of `i * 4`. For
6696 large values of `i` this can result in `bn_expand` not allocating any
6697 memory because `i * 4` is negative. This can leave the internal BIGNUM data
6698 field as NULL leading to a subsequent NULL ptr deref. For very large values
6699 of `i`, the calculation `i * 4` could be a positive value smaller than `i`.
6700 In this case memory is allocated to the internal BIGNUM data field, but it
6701 is insufficiently sized leading to heap corruption. A similar issue exists
6702 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn
6703 is ever called by user applications with very large untrusted hex/dec data.
6704 This is anticipated to be a rare occurrence.
6706 All OpenSSL internal usage of these functions use data that is not expected
6707 to be untrusted, e.g. config file data or application command line
6708 arguments. If user developed applications generate config file data based
6709 on untrusted data then it is possible that this could also lead to security
6710 consequences. This is also anticipated to be rare.
6712 This issue was reported to OpenSSL by Guido Vranken.
6717 * Fix memory issues in `BIO_*printf` functions
6719 The internal `fmtstr` function used in processing a "%s" format string in
6720 the `BIO_*printf` functions could overflow while calculating the length of a
6721 string and cause an OOB read when printing very long strings.
6723 Additionally the internal `doapr_outch` function can attempt to write to an
6724 OOB memory location (at an offset from the NULL pointer) in the event of a
6725 memory allocation failure. In 1.0.2 and below this could be caused where
6726 the size of a buffer to be allocated is greater than INT_MAX. E.g. this
6727 could be in processing a very long "%s" format string. Memory leaks can
6730 The first issue may mask the second issue dependent on compiler behaviour.
6731 These problems could enable attacks where large amounts of untrusted data
6732 is passed to the `BIO_*printf` functions. If applications use these functions
6733 in this way then they could be vulnerable. OpenSSL itself uses these
6734 functions when printing out human-readable dumps of ASN.1 data. Therefore
6735 applications that print this data could be vulnerable if the data is from
6736 untrusted sources. OpenSSL command line applications could also be
6737 vulnerable where they print out ASN.1 data, or if untrusted data is passed
6738 as command line arguments.
6740 Libssl is not considered directly vulnerable. Additionally certificates etc
6741 received via remote connections via libssl are also unlikely to be able to
6742 trigger these issues because of message size limits enforced within libssl.
6744 This issue was reported to OpenSSL Guido Vranken.
6749 * Side channel attack on modular exponentiation
6751 A side-channel attack was found which makes use of cache-bank conflicts on
6752 the Intel Sandy-Bridge microarchitecture which could lead to the recovery
6753 of RSA keys. The ability to exploit this issue is limited as it relies on
6754 an attacker who has control of code in a thread running on the same
6755 hyper-threaded core as the victim thread which is performing decryptions.
6757 This issue was reported to OpenSSL by Yuval Yarom, The University of
6758 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and
6759 Nadia Heninger, University of Pennsylvania with more information at
6760 <http://cachebleed.info>.
6765 * Change the req command to generate a 2048-bit RSA/DSA key by default,
6766 if no keysize is specified with default_bits. This fixes an
6767 omission in an earlier change that changed all RSA/DSA key generation
6768 commands to use 2048 bits by default.
6772 ### Changes between 1.0.1q and 1.0.1r [28 Jan 2016]
6774 * Protection for DH small subgroup attacks
6776 As a precautionary measure the SSL_OP_SINGLE_DH_USE option has been
6777 switched on by default and cannot be disabled. This could have some
6782 * SSLv2 doesn't block disabled ciphers
6784 A malicious client can negotiate SSLv2 ciphers that have been disabled on
6785 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
6786 been disabled, provided that the SSLv2 protocol was not also disabled via
6789 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
6790 and Sebastian Schinzel.
6795 * Reject DH handshakes with parameters shorter than 1024 bits.
6799 ### Changes between 1.0.1p and 1.0.1q [3 Dec 2015]
6801 * Certificate verify crash with missing PSS parameter
6803 The signature verification routines will crash with a NULL pointer
6804 dereference if presented with an ASN.1 signature using the RSA PSS
6805 algorithm and absent mask generation function parameter. Since these
6806 routines are used to verify certificate signature algorithms this can be
6807 used to crash any certificate verification operation and exploited in a
6808 DoS attack. Any application which performs certificate verification is
6809 vulnerable including OpenSSL clients and servers which enable client
6812 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
6817 * X509_ATTRIBUTE memory leak
6819 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
6820 memory. This structure is used by the PKCS#7 and CMS routines so any
6821 application which reads PKCS#7 or CMS data from untrusted sources is
6822 affected. SSL/TLS is not affected.
6824 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
6830 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
6831 This changes the decoding behaviour for some invalid messages,
6832 though the change is mostly in the more lenient direction, and
6833 legacy behaviour is preserved as much as possible.
6837 * In DSA_generate_parameters_ex, if the provided seed is too short,
6838 use a random seed, as already documented.
6840 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
6842 ### Changes between 1.0.1o and 1.0.1p [9 Jul 2015]
6844 * Alternate chains certificate forgery
6846 During certificate verfification, OpenSSL will attempt to find an
6847 alternative certificate chain if the first attempt to build such a chain
6848 fails. An error in the implementation of this logic can mean that an
6849 attacker could cause certain checks on untrusted certificates to be
6850 bypassed, such as the CA flag, enabling them to use a valid leaf
6851 certificate to act as a CA and "issue" an invalid certificate.
6853 This issue was reported to OpenSSL by Adam Langley/David Benjamin
6859 * Race condition handling PSK identify hint
6861 If PSK identity hints are received by a multi-threaded client then
6862 the values are wrongly updated in the parent SSL_CTX structure. This can
6863 result in a race condition potentially leading to a double free of the
6869 ### Changes between 1.0.1n and 1.0.1o [12 Jun 2015]
6871 * Fix HMAC ABI incompatibility. The previous version introduced an ABI
6872 incompatibility in the handling of HMAC. The previous ABI has now been
6875 ### Changes between 1.0.1m and 1.0.1n [11 Jun 2015]
6877 * Malformed ECParameters causes infinite loop
6879 When processing an ECParameters structure OpenSSL enters an infinite loop
6880 if the curve specified is over a specially malformed binary polynomial
6883 This can be used to perform denial of service against any
6884 system which processes public keys, certificate requests or
6885 certificates. This includes TLS clients and TLS servers with
6886 client authentication enabled.
6888 This issue was reported to OpenSSL by Joseph Barr-Pixton.
6893 * Exploitable out-of-bounds read in X509_cmp_time
6895 X509_cmp_time does not properly check the length of the ASN1_TIME
6896 string and can read a few bytes out of bounds. In addition,
6897 X509_cmp_time accepts an arbitrary number of fractional seconds in the
6900 An attacker can use this to craft malformed certificates and CRLs of
6901 various sizes and potentially cause a segmentation fault, resulting in
6902 a DoS on applications that verify certificates or CRLs. TLS clients
6903 that verify CRLs are affected. TLS clients and servers with client
6904 authentication enabled may be affected if they use custom verification
6907 This issue was reported to OpenSSL by Robert Swiecki (Google), and
6908 independently by Hanno Böck.
6913 * PKCS7 crash with missing EnvelopedContent
6915 The PKCS#7 parsing code does not handle missing inner EncryptedContent
6916 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
6917 with missing content and trigger a NULL pointer dereference on parsing.
6919 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
6920 structures from untrusted sources are affected. OpenSSL clients and
6921 servers are not affected.
6923 This issue was reported to OpenSSL by Michal Zalewski (Google).
6928 * CMS verify infinite loop with unknown hash function
6930 When verifying a signedData message the CMS code can enter an infinite loop
6931 if presented with an unknown hash function OID. This can be used to perform
6932 denial of service against any system which verifies signedData messages using
6934 This issue was reported to OpenSSL by Johannes Bauer.
6939 * Race condition handling NewSessionTicket
6941 If a NewSessionTicket is received by a multi-threaded client when attempting to
6942 reuse a previous ticket then a race condition can occur potentially leading to
6943 a double free of the ticket data.
6948 * Reject DH handshakes with parameters shorter than 768 bits.
6950 *Kurt Roeckx and Emilia Kasper*
6952 * dhparam: generate 2048-bit parameters by default.
6954 *Kurt Roeckx and Emilia Kasper*
6956 ### Changes between 1.0.1l and 1.0.1m [19 Mar 2015]
6958 * Segmentation fault in ASN1_TYPE_cmp fix
6960 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
6961 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
6962 certificate signature algorithm consistency this can be used to crash any
6963 certificate verification operation and exploited in a DoS attack. Any
6964 application which performs certificate verification is vulnerable including
6965 OpenSSL clients and servers which enable client authentication.
6970 * ASN.1 structure reuse memory corruption fix
6972 Reusing a structure in ASN.1 parsing may allow an attacker to cause
6973 memory corruption via an invalid write. Such reuse is and has been
6974 strongly discouraged and is believed to be rare.
6976 Applications that parse structures containing CHOICE or ANY DEFINED BY
6977 components may be affected. Certificate parsing (d2i_X509 and related
6978 functions) are however not affected. OpenSSL clients and servers are
6984 * PKCS7 NULL pointer dereferences fix
6986 The PKCS#7 parsing code does not handle missing outer ContentInfo
6987 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
6988 missing content and trigger a NULL pointer dereference on parsing.
6990 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
6991 otherwise parse PKCS#7 structures from untrusted sources are
6992 affected. OpenSSL clients and servers are not affected.
6994 This issue was reported to OpenSSL by Michal Zalewski (Google).
6999 * DoS via reachable assert in SSLv2 servers fix
7001 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
7002 servers that both support SSLv2 and enable export cipher suites by sending
7003 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
7005 This issue was discovered by Sean Burford (Google) and Emilia Käsper
7006 (OpenSSL development team).
7011 * Use After Free following d2i_ECPrivatekey error fix
7013 A malformed EC private key file consumed via the d2i_ECPrivateKey function
7014 could cause a use after free condition. This, in turn, could cause a double
7015 free in several private key parsing functions (such as d2i_PrivateKey
7016 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
7017 for applications that receive EC private keys from untrusted
7018 sources. This scenario is considered rare.
7020 This issue was discovered by the BoringSSL project and fixed in their
7026 * X509_to_X509_REQ NULL pointer deref fix
7028 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
7029 the certificate key is invalid. This function is rarely used in practice.
7031 This issue was discovered by Brian Carpenter.
7036 * Removed the export ciphers from the DEFAULT ciphers
7040 ### Changes between 1.0.1k and 1.0.1l [15 Jan 2015]
7042 * Build fixes for the Windows and OpenVMS platforms
7044 *Matt Caswell and Richard Levitte*
7046 ### Changes between 1.0.1j and 1.0.1k [8 Jan 2015]
7048 * Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
7049 message can cause a segmentation fault in OpenSSL due to a NULL pointer
7050 dereference. This could lead to a Denial Of Service attack. Thanks to
7051 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
7056 * Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
7057 dtls1_buffer_record function under certain conditions. In particular this
7058 could occur if an attacker sent repeated DTLS records with the same
7059 sequence number but for the next epoch. The memory leak could be exploited
7060 by an attacker in a Denial of Service attack through memory exhaustion.
7061 Thanks to Chris Mueller for reporting this issue.
7066 * Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
7067 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
7068 method would be set to NULL which could later result in a NULL pointer
7069 dereference. Thanks to Frank Schmirler for reporting this issue.
7074 * Abort handshake if server key exchange message is omitted for ephemeral
7077 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
7078 reporting this issue.
7083 * Remove non-export ephemeral RSA code on client and server. This code
7084 violated the TLS standard by allowing the use of temporary RSA keys in
7085 non-export ciphersuites and could be used by a server to effectively
7086 downgrade the RSA key length used to a value smaller than the server
7087 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
7088 INRIA or reporting this issue.
7093 * Fixed issue where DH client certificates are accepted without verification.
7094 An OpenSSL server will accept a DH certificate for client authentication
7095 without the certificate verify message. This effectively allows a client to
7096 authenticate without the use of a private key. This only affects servers
7097 which trust a client certificate authority which issues certificates
7098 containing DH keys: these are extremely rare and hardly ever encountered.
7099 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
7105 * Ensure that the session ID context of an SSL is updated when its
7106 SSL_CTX is updated via SSL_set_SSL_CTX.
7108 The session ID context is typically set from the parent SSL_CTX,
7109 and can vary with the CTX.
7113 * Fix various certificate fingerprint issues.
7115 By using non-DER or invalid encodings outside the signed portion of a
7116 certificate the fingerprint can be changed without breaking the signature.
7117 Although no details of the signed portion of the certificate can be changed
7118 this can cause problems with some applications: e.g. those using the
7119 certificate fingerprint for blacklists.
7121 1. Reject signatures with non zero unused bits.
7123 If the BIT STRING containing the signature has non zero unused bits reject
7124 the signature. All current signature algorithms require zero unused bits.
7126 2. Check certificate algorithm consistency.
7128 Check the AlgorithmIdentifier inside TBS matches the one in the
7129 certificate signature. NB: this will result in signature failure
7130 errors for some broken certificates.
7132 Thanks to Konrad Kraszewski from Google for reporting this issue.
7134 3. Check DSA/ECDSA signatures use DER.
7136 Re-encode DSA/ECDSA signatures and compare with the original received
7137 signature. Return an error if there is a mismatch.
7139 This will reject various cases including garbage after signature
7140 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
7141 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
7142 (negative or with leading zeroes).
7144 Further analysis was conducted and fixes were developed by Stephen Henson
7145 of the OpenSSL core team.
7151 * Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
7152 results on some platforms, including x86_64. This bug occurs at random
7153 with a very low probability, and is not known to be exploitable in any
7154 way, though its exact impact is difficult to determine. Thanks to Pieter
7155 Wuille (Blockstream) who reported this issue and also suggested an initial
7156 fix. Further analysis was conducted by the OpenSSL development team and
7157 Adam Langley of Google. The final fix was developed by Andy Polyakov of
7158 the OpenSSL core team.
7163 * Do not resume sessions on the server if the negotiated protocol
7164 version does not match the session's version. Resuming with a different
7165 version, while not strictly forbidden by the RFC, is of questionable
7166 sanity and breaks all known clients.
7168 *David Benjamin, Emilia Käsper*
7170 * Tighten handling of the ChangeCipherSpec (CCS) message: reject
7171 early CCS messages during renegotiation. (Note that because
7172 renegotiation is encrypted, this early CCS was not exploitable.)
7176 * Tighten client-side session ticket handling during renegotiation:
7177 ensure that the client only accepts a session ticket if the server sends
7178 the extension anew in the ServerHello. Previously, a TLS client would
7179 reuse the old extension state and thus accept a session ticket if one was
7180 announced in the initial ServerHello.
7182 Similarly, ensure that the client requires a session ticket if one
7183 was advertised in the ServerHello. Previously, a TLS client would
7184 ignore a missing NewSessionTicket message.
7188 ### Changes between 1.0.1i and 1.0.1j [15 Oct 2014]
7192 A flaw in the DTLS SRTP extension parsing code allows an attacker, who
7193 sends a carefully crafted handshake message, to cause OpenSSL to fail
7194 to free up to 64k of memory causing a memory leak. This could be
7195 exploited in a Denial Of Service attack. This issue affects OpenSSL
7196 1.0.1 server implementations for both SSL/TLS and DTLS regardless of
7197 whether SRTP is used or configured. Implementations of OpenSSL that
7198 have been compiled with OPENSSL_NO_SRTP defined are not affected.
7200 The fix was developed by the OpenSSL team.
7205 * Session Ticket Memory Leak.
7207 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
7208 integrity of that ticket is first verified. In the event of a session
7209 ticket integrity check failing, OpenSSL will fail to free memory
7210 causing a memory leak. By sending a large number of invalid session
7211 tickets an attacker could exploit this issue in a Denial Of Service
7217 * Build option no-ssl3 is incomplete.
7219 When OpenSSL is configured with "no-ssl3" as a build option, servers
7220 could accept and complete a SSL 3.0 handshake, and clients could be
7221 configured to send them.
7224 *Akamai and the OpenSSL team*
7226 * Add support for TLS_FALLBACK_SCSV.
7227 Client applications doing fallback retries should call
7228 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
7231 *Adam Langley, Bodo Moeller*
7233 * Add additional DigestInfo checks.
7235 Re-encode DigestInto in DER and check against the original when
7236 verifying RSA signature: this will reject any improperly encoded
7237 DigestInfo structures.
7239 Note: this is a precautionary measure and no attacks are currently known.
7243 ### Changes between 1.0.1h and 1.0.1i [6 Aug 2014]
7245 * Fix SRP buffer overrun vulnerability. Invalid parameters passed to the
7246 SRP code can be overrun an internal buffer. Add sanity check that
7247 g, A, B < N to SRP code.
7249 Thanks to Sean Devlin and Watson Ladd of Cryptography Services, NCC
7250 Group for discovering this issue.
7255 * A flaw in the OpenSSL SSL/TLS server code causes the server to negotiate
7256 TLS 1.0 instead of higher protocol versions when the ClientHello message
7257 is badly fragmented. This allows a man-in-the-middle attacker to force a
7258 downgrade to TLS 1.0 even if both the server and the client support a
7259 higher protocol version, by modifying the client's TLS records.
7261 Thanks to David Benjamin and Adam Langley (Google) for discovering and
7262 researching this issue.
7267 * OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
7268 to a denial of service attack. A malicious server can crash the client
7269 with a null pointer dereference (read) by specifying an anonymous (EC)DH
7270 ciphersuite and sending carefully crafted handshake messages.
7272 Thanks to Felix Gröbert (Google) for discovering and researching this
7278 * By sending carefully crafted DTLS packets an attacker could cause openssl
7279 to leak memory. This can be exploited through a Denial of Service attack.
7280 Thanks to Adam Langley for discovering and researching this issue.
7285 * An attacker can force openssl to consume large amounts of memory whilst
7286 processing DTLS handshake messages. This can be exploited through a
7287 Denial of Service attack.
7288 Thanks to Adam Langley for discovering and researching this issue.
7293 * An attacker can force an error condition which causes openssl to crash
7294 whilst processing DTLS packets due to memory being freed twice. This
7295 can be exploited through a Denial of Service attack.
7296 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
7302 * If a multithreaded client connects to a malicious server using a resumed
7303 session and the server sends an ec point format extension it could write
7304 up to 255 bytes to freed memory.
7306 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
7312 * A malicious server can crash an OpenSSL client with a null pointer
7313 dereference (read) by specifying an SRP ciphersuite even though it was not
7314 properly negotiated with the client. This can be exploited through a
7315 Denial of Service attack.
7317 Thanks to Joonas Kuorilehto and Riku Hietamäki (Codenomicon) for
7318 discovering and researching this issue.
7323 * A flaw in OBJ_obj2txt may cause pretty printing functions such as
7324 X509_name_oneline, X509_name_print_ex et al. to leak some information
7325 from the stack. Applications may be affected if they echo pretty printing
7326 output to the attacker.
7328 Thanks to Ivan Fratric (Google) for discovering this issue.
7331 *Emilia Käsper, and Steve Henson*
7333 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
7334 for corner cases. (Certain input points at infinity could lead to
7335 bogus results, with non-infinity inputs mapped to infinity too.)
7339 ### Changes between 1.0.1g and 1.0.1h [5 Jun 2014]
7341 * Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
7342 handshake can force the use of weak keying material in OpenSSL
7343 SSL/TLS clients and servers.
7345 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
7346 researching this issue. ([CVE-2014-0224])
7348 *KIKUCHI Masashi, Steve Henson*
7350 * Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
7351 OpenSSL DTLS client the code can be made to recurse eventually crashing
7354 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
7357 *Imre Rad, Steve Henson*
7359 * Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
7360 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
7361 client or server. This is potentially exploitable to run arbitrary
7362 code on a vulnerable client or server.
7364 Thanks to Jüri Aedla for reporting this issue. ([CVE-2014-0195])
7366 *Jüri Aedla, Steve Henson*
7368 * Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
7369 are subject to a denial of service attack.
7371 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
7372 this issue. ([CVE-2014-3470])
7374 *Felix Gröbert, Ivan Fratric, Steve Henson*
7376 * Harmonize version and its documentation. -f flag is used to display
7379 *mancha <mancha1@zoho.com>*
7381 * Fix eckey_priv_encode so it immediately returns an error upon a failure
7382 in i2d_ECPrivateKey.
7384 *mancha <mancha1@zoho.com>*
7386 * Fix some double frees. These are not thought to be exploitable.
7388 *mancha <mancha1@zoho.com>*
7390 ### Changes between 1.0.1f and 1.0.1g [7 Apr 2014]
7392 * A missing bounds check in the handling of the TLS heartbeat extension
7393 can be used to reveal up to 64k of memory to a connected client or
7396 Thanks for Neel Mehta of Google Security for discovering this bug and to
7397 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
7398 preparing the fix ([CVE-2014-0160])
7400 *Adam Langley, Bodo Moeller*
7402 * Fix for the attack described in the paper "Recovering OpenSSL
7403 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
7404 by Yuval Yarom and Naomi Benger. Details can be obtained from:
7405 <http://eprint.iacr.org/2014/140>
7407 Thanks to Yuval Yarom and Naomi Benger for discovering this
7408 flaw and to Yuval Yarom for supplying a fix ([CVE-2014-0076])
7410 *Yuval Yarom and Naomi Benger*
7412 * TLS pad extension: draft-agl-tls-padding-03
7414 Workaround for the "TLS hang bug" (see FAQ and PR#2771): if the
7415 TLS client Hello record length value would otherwise be > 255 and
7416 less that 512 pad with a dummy extension containing zeroes so it
7417 is at least 512 bytes long.
7419 *Adam Langley, Steve Henson*
7421 ### Changes between 1.0.1e and 1.0.1f [6 Jan 2014]
7423 * Fix for TLS record tampering bug. A carefully crafted invalid
7424 handshake could crash OpenSSL with a NULL pointer exception.
7425 Thanks to Anton Johansson for reporting this issues.
7428 * Keep original DTLS digest and encryption contexts in retransmission
7429 structures so we can use the previous session parameters if they need
7430 to be resent. ([CVE-2013-6450])
7434 * Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
7435 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
7436 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
7437 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
7438 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
7439 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
7441 *Rob Stradling, Adam Langley*
7443 ### Changes between 1.0.1d and 1.0.1e [11 Feb 2013]
7445 * Correct fix for CVE-2013-0169. The original didn't work on AES-NI
7446 supporting platforms or when small records were transferred.
7448 *Andy Polyakov, Steve Henson*
7450 ### Changes between 1.0.1c and 1.0.1d [5 Feb 2013]
7452 * Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
7454 This addresses the flaw in CBC record processing discovered by
7455 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
7456 at: <http://www.isg.rhul.ac.uk/tls/>
7458 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
7459 Security Group at Royal Holloway, University of London
7460 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
7461 Emilia Käsper for the initial patch.
7464 *Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson*
7466 * Fix flaw in AESNI handling of TLS 1.2 and 1.1 records for CBC mode
7467 ciphersuites which can be exploited in a denial of service attack.
7468 Thanks go to and to Adam Langley <agl@chromium.org> for discovering
7469 and detecting this bug and to Wolfgang Ettlinger
7470 <wolfgang.ettlinger@gmail.com> for independently discovering this issue.
7475 * Return an error when checking OCSP signatures when key is NULL.
7476 This fixes a DoS attack. ([CVE-2013-0166])
7480 * Make openssl verify return errors.
7482 *Chris Palmer <palmer@google.com> and Ben Laurie*
7484 * Call OCSP Stapling callback after ciphersuite has been chosen, so
7485 the right response is stapled. Also change SSL_get_certificate()
7486 so it returns the certificate actually sent.
7487 See <http://rt.openssl.org/Ticket/Display.html?id=2836>.
7489 *Rob Stradling <rob.stradling@comodo.com>*
7491 * Fix possible deadlock when decoding public keys.
7495 * Don't use TLS 1.0 record version number in initial client hello
7500 ### Changes between 1.0.1b and 1.0.1c [10 May 2012]
7502 * Sanity check record length before skipping explicit IV in TLS
7503 1.2, 1.1 and DTLS to fix DoS attack.
7505 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
7506 fuzzing as a service testing platform.
7511 * Initialise tkeylen properly when encrypting CMS messages.
7512 Thanks to Solar Designer of Openwall for reporting this issue.
7516 * In FIPS mode don't try to use composite ciphers as they are not
7521 ### Changes between 1.0.1a and 1.0.1b [26 Apr 2012]
7523 * OpenSSL 1.0.0 sets SSL_OP_ALL to 0x80000FFFL and OpenSSL 1.0.1 and
7524 1.0.1a set SSL_OP_NO_TLSv1_1 to 0x00000400L which would unfortunately
7525 mean any application compiled against OpenSSL 1.0.0 headers setting
7526 SSL_OP_ALL would also set SSL_OP_NO_TLSv1_1, unintentionally disabling
7527 TLS 1.1 also. Fix this by changing the value of SSL_OP_NO_TLSv1_1 to
7528 0x10000000L Any application which was previously compiled against
7529 OpenSSL 1.0.1 or 1.0.1a headers and which cares about SSL_OP_NO_TLSv1_1
7530 will need to be recompiled as a result. Letting be results in
7531 inability to disable specifically TLS 1.1 and in client context,
7532 in unlike event, limit maximum offered version to TLS 1.0 [see below].
7536 * In order to ensure interoperability SSL_OP_NO_protocolX does not
7537 disable just protocol X, but all protocols above X *if* there are
7538 protocols *below* X still enabled. In more practical terms it means
7539 that if application wants to disable TLS1.0 in favor of TLS1.1 and
7540 above, it's not sufficient to pass `SSL_OP_NO_TLSv1`, one has to pass
7541 `SSL_OP_NO_TLSv1|SSL_OP_NO_SSLv3|SSL_OP_NO_SSLv2`. This applies to
7546 ### Changes between 1.0.1 and 1.0.1a [19 Apr 2012]
7548 * Check for potentially exploitable overflows in asn1_d2i_read_bio
7549 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
7550 in CRYPTO_realloc_clean.
7552 Thanks to Tavis Ormandy, Google Security Team, for discovering this
7553 issue and to Adam Langley <agl@chromium.org> for fixing it.
7556 *Adam Langley (Google), Tavis Ormandy, Google Security Team*
7558 * Don't allow TLS 1.2 SHA-256 ciphersuites in TLS 1.0, 1.1 connections.
7562 * Workarounds for some broken servers that "hang" if a client hello
7563 record length exceeds 255 bytes.
7565 1. Do not use record version number > TLS 1.0 in initial client
7566 hello: some (but not all) hanging servers will now work.
7567 2. If we set OPENSSL_MAX_TLS1_2_CIPHER_LENGTH this will truncate
7568 the number of ciphers sent in the client hello. This should be
7569 set to an even number, such as 50, for example by passing:
7570 -DOPENSSL_MAX_TLS1_2_CIPHER_LENGTH=50 to config or Configure.
7571 Most broken servers should now work.
7572 3. If all else fails setting OPENSSL_NO_TLS1_2_CLIENT will disable
7573 TLS 1.2 client support entirely.
7577 * Fix SEGV in Vector Permutation AES module observed in OpenSSH.
7581 ### Changes between 1.0.0h and 1.0.1 [14 Mar 2012]
7583 * Add compatibility with old MDC2 signatures which use an ASN1 OCTET
7584 STRING form instead of a DigestInfo.
7588 * The format used for MDC2 RSA signatures is inconsistent between EVP
7589 and the RSA_sign/RSA_verify functions. This was made more apparent when
7590 OpenSSL used RSA_sign/RSA_verify for some RSA signatures in particular
7591 those which went through EVP_PKEY_METHOD in 1.0.0 and later. Detect
7592 the correct format in RSA_verify so both forms transparently work.
7596 * Some servers which support TLS 1.0 can choke if we initially indicate
7597 support for TLS 1.2 and later renegotiate using TLS 1.0 in the RSA
7598 encrypted premaster secret. As a workaround use the maximum permitted
7599 client version in client hello, this should keep such servers happy
7600 and still work with previous versions of OpenSSL.
7604 * Add support for TLS/DTLS heartbeats.
7606 *Robin Seggelmann <seggelmann@fh-muenster.de>*
7608 * Add support for SCTP.
7610 *Robin Seggelmann <seggelmann@fh-muenster.de>*
7612 * Improved PRNG seeding for VOS.
7614 *Paul Green <Paul.Green@stratus.com>*
7616 * Extensive assembler packs updates, most notably:
7618 - x86[_64]: AES-NI, PCLMULQDQ, RDRAND support;
7619 - x86[_64]: SSSE3 support (SHA1, vector-permutation AES);
7620 - x86_64: bit-sliced AES implementation;
7621 - ARM: NEON support, contemporary platforms optimizations;
7622 - s390x: z196 support;
7623 - `*`: GHASH and GF(2^m) multiplication implementations;
7627 * Make TLS-SRP code conformant with RFC 5054 API cleanup
7628 (removal of unnecessary code)
7630 *Peter Sylvester <peter.sylvester@edelweb.fr>*
7632 * Add TLS key material exporter from RFC 5705.
7636 * Add DTLS-SRTP negotiation from RFC 5764.
7640 * Add Next Protocol Negotiation,
7641 <http://tools.ietf.org/html/draft-agl-tls-nextprotoneg-00>. Can be
7642 disabled with a no-npn flag to config or Configure. Code donated
7645 *Adam Langley <agl@google.com> and Ben Laurie*
7647 * Add optional 64-bit optimized implementations of elliptic curves NIST-P224,
7648 NIST-P256, NIST-P521, with constant-time single point multiplication on
7649 typical inputs. Compiler support for the nonstandard type `__uint128_t` is
7650 required to use this (present in gcc 4.4 and later, for 64-bit builds).
7651 Code made available under Apache License version 2.0.
7653 Specify "enable-ec_nistp_64_gcc_128" on the Configure (or config) command
7654 line to include this in your build of OpenSSL, and run "make depend" (or
7655 "make update"). This enables the following EC_METHODs:
7657 EC_GFp_nistp224_method()
7658 EC_GFp_nistp256_method()
7659 EC_GFp_nistp521_method()
7661 EC_GROUP_new_by_curve_name() will automatically use these (while
7662 EC_GROUP_new_curve_GFp() currently prefers the more flexible
7665 *Emilia Käsper, Adam Langley, Bodo Moeller (Google)*
7667 * Use type ossl_ssize_t instead of ssize_t which isn't available on
7668 all platforms. Move ssize_t definition from e_os.h to the public
7669 header file e_os2.h as it now appears in public header file cms.h
7673 * New -sigopt option to the ca, req and x509 utilities. Additional
7674 signature parameters can be passed using this option and in
7679 * Add RSA PSS signing function. This will generate and set the
7680 appropriate AlgorithmIdentifiers for PSS based on those in the
7681 corresponding EVP_MD_CTX structure. No application support yet.
7685 * Support for companion algorithm specific ASN1 signing routines.
7686 New function ASN1_item_sign_ctx() signs a pre-initialised
7687 EVP_MD_CTX structure and sets AlgorithmIdentifiers based on
7688 the appropriate parameters.
7692 * Add new algorithm specific ASN1 verification initialisation function
7693 to EVP_PKEY_ASN1_METHOD: this is not in EVP_PKEY_METHOD since the ASN1
7694 handling will be the same no matter what EVP_PKEY_METHOD is used.
7695 Add a PSS handler to support verification of PSS signatures: checked
7696 against a number of sample certificates.
7700 * Add signature printing for PSS. Add PSS OIDs.
7702 *Steve Henson, Martin Kaiser <lists@kaiser.cx>*
7704 * Add algorithm specific signature printing. An individual ASN1 method
7705 can now print out signatures instead of the standard hex dump.
7707 More complex signatures (e.g. PSS) can print out more meaningful
7708 information. Include DSA version that prints out the signature
7713 * Password based recipient info support for CMS library: implementing
7718 * Split password based encryption into PBES2 and PBKDF2 functions. This
7719 neatly separates the code into cipher and PBE sections and is required
7720 for some algorithms that split PBES2 into separate pieces (such as
7721 password based CMS).
7725 * Session-handling fixes:
7726 - Fix handling of connections that are resuming with a session ID,
7727 but also support Session Tickets.
7728 - Fix a bug that suppressed issuing of a new ticket if the client
7729 presented a ticket with an expired session.
7730 - Try to set the ticket lifetime hint to something reasonable.
7731 - Make tickets shorter by excluding irrelevant information.
7732 - On the client side, don't ignore renewed tickets.
7734 *Adam Langley, Bodo Moeller (Google)*
7736 * Fix PSK session representation.
7740 * Add RC4-MD5 and AESNI-SHA1 "stitched" implementations.
7742 This work was sponsored by Intel.
7746 * Add GCM support to TLS library. Some custom code is needed to split
7747 the IV between the fixed (from PRF) and explicit (from TLS record)
7748 portions. This adds all GCM ciphersuites supported by RFC5288 and
7749 RFC5289. Generalise some `AES*` cipherstrings to include GCM and
7750 add a special AESGCM string for GCM only.
7754 * Expand range of ctrls for AES GCM. Permit setting invocation
7755 field on decrypt and retrieval of invocation field only on encrypt.
7759 * Add HMAC ECC ciphersuites from RFC5289. Include SHA384 PRF support.
7760 As required by RFC5289 these ciphersuites cannot be used if for
7761 versions of TLS earlier than 1.2.
7765 * For FIPS capable OpenSSL interpret a NULL default public key method
7766 as unset and return the appropriate default but do *not* set the default.
7767 This means we can return the appropriate method in applications that
7768 switch between FIPS and non-FIPS modes.
7772 * Redirect HMAC and CMAC operations to FIPS module in FIPS mode. If an
7773 ENGINE is used then we cannot handle that in the FIPS module so we
7774 keep original code iff non-FIPS operations are allowed.
7778 * Add -attime option to openssl utilities.
7780 *Peter Eckersley <pde@eff.org>, Ben Laurie and Steve Henson*
7782 * Redirect DSA and DH operations to FIPS module in FIPS mode.
7786 * Redirect ECDSA and ECDH operations to FIPS module in FIPS mode. Also use
7787 FIPS EC methods unconditionally for now.
7791 * New build option no-ec2m to disable characteristic 2 code.
7795 * Backport libcrypto audit of return value checking from 1.1.0-dev; not
7796 all cases can be covered as some introduce binary incompatibilities.
7800 * Redirect RSA operations to FIPS module including keygen,
7801 encrypt, decrypt, sign and verify. Block use of non FIPS RSA methods.
7805 * Add similar low-level API blocking to ciphers.
7809 * low-level digest APIs are not approved in FIPS mode: any attempt
7810 to use these will cause a fatal error. Applications that *really* want
7811 to use them can use the `private_*` version instead.
7815 * Redirect cipher operations to FIPS module for FIPS builds.
7819 * Redirect digest operations to FIPS module for FIPS builds.
7823 * Update build system to add "fips" flag which will link in fipscanister.o
7824 for static and shared library builds embedding a signature if needed.
7828 * Output TLS supported curves in preference order instead of numerical
7829 order. This is currently hardcoded for the highest order curves first.
7830 This should be configurable so applications can judge speed vs strength.
7834 * Add TLS v1.2 server support for client authentication.
7838 * Add support for FIPS mode in ssl library: disable SSLv3, non-FIPS ciphers
7843 * Functions FIPS_mode_set() and FIPS_mode() which call the underlying
7844 FIPS modules versions.
7848 * Add TLS v1.2 client side support for client authentication. Keep cache
7849 of handshake records longer as we don't know the hash algorithm to use
7850 until after the certificate request message is received.
7854 * Initial TLS v1.2 client support. Add a default signature algorithms
7855 extension including all the algorithms we support. Parse new signature
7856 format in client key exchange. Relax some ECC signing restrictions for
7857 TLS v1.2 as indicated in RFC5246.
7861 * Add server support for TLS v1.2 signature algorithms extension. Switch
7862 to new signature format when needed using client digest preference.
7863 All server ciphersuites should now work correctly in TLS v1.2. No client
7864 support yet and no support for client certificates.
7868 * Initial TLS v1.2 support. Add new SHA256 digest to ssl code, switch
7869 to SHA256 for PRF when using TLS v1.2 and later. Add new SHA256 based
7870 ciphersuites. At present only RSA key exchange ciphersuites work with
7871 TLS v1.2. Add new option for TLS v1.2 replacing the old and obsolete
7872 SSL_OP_PKCS1_CHECK flags with SSL_OP_NO_TLSv1_2. New TLSv1.2 methods
7873 and version checking.
7877 * New option OPENSSL_NO_SSL_INTERN. If an application can be compiled
7878 with this defined it will not be affected by any changes to ssl internal
7879 structures. Add several utility functions to allow openssl application
7880 to work with OPENSSL_NO_SSL_INTERN defined.
7884 * A long standing patch to add support for SRP from EdelWeb (Peter
7885 Sylvester and Christophe Renou) was integrated.
7886 *Christophe Renou <christophe.renou@edelweb.fr>, Peter Sylvester
7887 <peter.sylvester@edelweb.fr>, Tom Wu <tjw@cs.stanford.edu>, and
7890 * Add functions to copy EVP_PKEY_METHOD and retrieve flags and id.
7894 * Permit abbreviated handshakes when renegotiating using the function
7895 SSL_renegotiate_abbreviated().
7897 *Robin Seggelmann <seggelmann@fh-muenster.de>*
7899 * Add call to ENGINE_register_all_complete() to
7900 ENGINE_load_builtin_engines(), so some implementations get used
7901 automatically instead of needing explicit application support.
7905 * Add support for TLS key exporter as described in RFC5705.
7907 *Robin Seggelmann <seggelmann@fh-muenster.de>, Steve Henson*
7909 * Initial TLSv1.1 support. Since TLSv1.1 is very similar to TLS v1.0 only
7910 a few changes are required:
7912 Add SSL_OP_NO_TLSv1_1 flag.
7913 Add TLSv1_1 methods.
7914 Update version checking logic to handle version 1.1.
7915 Add explicit IV handling (ported from DTLS code).
7916 Add command line options to s_client/s_server.
7923 ### Changes between 1.0.0s and 1.0.0t [3 Dec 2015]
7925 * X509_ATTRIBUTE memory leak
7927 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
7928 memory. This structure is used by the PKCS#7 and CMS routines so any
7929 application which reads PKCS#7 or CMS data from untrusted sources is
7930 affected. SSL/TLS is not affected.
7932 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
7938 * Race condition handling PSK identify hint
7940 If PSK identity hints are received by a multi-threaded client then
7941 the values are wrongly updated in the parent SSL_CTX structure. This can
7942 result in a race condition potentially leading to a double free of the
7948 ### Changes between 1.0.0r and 1.0.0s [11 Jun 2015]
7950 * Malformed ECParameters causes infinite loop
7952 When processing an ECParameters structure OpenSSL enters an infinite loop
7953 if the curve specified is over a specially malformed binary polynomial
7956 This can be used to perform denial of service against any
7957 system which processes public keys, certificate requests or
7958 certificates. This includes TLS clients and TLS servers with
7959 client authentication enabled.
7961 This issue was reported to OpenSSL by Joseph Barr-Pixton.
7966 * Exploitable out-of-bounds read in X509_cmp_time
7968 X509_cmp_time does not properly check the length of the ASN1_TIME
7969 string and can read a few bytes out of bounds. In addition,
7970 X509_cmp_time accepts an arbitrary number of fractional seconds in the
7973 An attacker can use this to craft malformed certificates and CRLs of
7974 various sizes and potentially cause a segmentation fault, resulting in
7975 a DoS on applications that verify certificates or CRLs. TLS clients
7976 that verify CRLs are affected. TLS clients and servers with client
7977 authentication enabled may be affected if they use custom verification
7980 This issue was reported to OpenSSL by Robert Swiecki (Google), and
7981 independently by Hanno Böck.
7986 * PKCS7 crash with missing EnvelopedContent
7988 The PKCS#7 parsing code does not handle missing inner EncryptedContent
7989 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
7990 with missing content and trigger a NULL pointer dereference on parsing.
7992 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
7993 structures from untrusted sources are affected. OpenSSL clients and
7994 servers are not affected.
7996 This issue was reported to OpenSSL by Michal Zalewski (Google).
8001 * CMS verify infinite loop with unknown hash function
8003 When verifying a signedData message the CMS code can enter an infinite loop
8004 if presented with an unknown hash function OID. This can be used to perform
8005 denial of service against any system which verifies signedData messages using
8007 This issue was reported to OpenSSL by Johannes Bauer.
8012 * Race condition handling NewSessionTicket
8014 If a NewSessionTicket is received by a multi-threaded client when attempting to
8015 reuse a previous ticket then a race condition can occur potentially leading to
8016 a double free of the ticket data.
8021 ### Changes between 1.0.0q and 1.0.0r [19 Mar 2015]
8023 * Segmentation fault in ASN1_TYPE_cmp fix
8025 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
8026 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
8027 certificate signature algorithm consistency this can be used to crash any
8028 certificate verification operation and exploited in a DoS attack. Any
8029 application which performs certificate verification is vulnerable including
8030 OpenSSL clients and servers which enable client authentication.
8035 * ASN.1 structure reuse memory corruption fix
8037 Reusing a structure in ASN.1 parsing may allow an attacker to cause
8038 memory corruption via an invalid write. Such reuse is and has been
8039 strongly discouraged and is believed to be rare.
8041 Applications that parse structures containing CHOICE or ANY DEFINED BY
8042 components may be affected. Certificate parsing (d2i_X509 and related
8043 functions) are however not affected. OpenSSL clients and servers are
8049 * PKCS7 NULL pointer dereferences fix
8051 The PKCS#7 parsing code does not handle missing outer ContentInfo
8052 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
8053 missing content and trigger a NULL pointer dereference on parsing.
8055 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
8056 otherwise parse PKCS#7 structures from untrusted sources are
8057 affected. OpenSSL clients and servers are not affected.
8059 This issue was reported to OpenSSL by Michal Zalewski (Google).
8064 * DoS via reachable assert in SSLv2 servers fix
8066 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
8067 servers that both support SSLv2 and enable export cipher suites by sending
8068 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
8070 This issue was discovered by Sean Burford (Google) and Emilia Käsper
8071 (OpenSSL development team).
8076 * Use After Free following d2i_ECPrivatekey error fix
8078 A malformed EC private key file consumed via the d2i_ECPrivateKey function
8079 could cause a use after free condition. This, in turn, could cause a double
8080 free in several private key parsing functions (such as d2i_PrivateKey
8081 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
8082 for applications that receive EC private keys from untrusted
8083 sources. This scenario is considered rare.
8085 This issue was discovered by the BoringSSL project and fixed in their
8091 * X509_to_X509_REQ NULL pointer deref fix
8093 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
8094 the certificate key is invalid. This function is rarely used in practice.
8096 This issue was discovered by Brian Carpenter.
8101 * Removed the export ciphers from the DEFAULT ciphers
8105 ### Changes between 1.0.0p and 1.0.0q [15 Jan 2015]
8107 * Build fixes for the Windows and OpenVMS platforms
8109 *Matt Caswell and Richard Levitte*
8111 ### Changes between 1.0.0o and 1.0.0p [8 Jan 2015]
8113 * Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
8114 message can cause a segmentation fault in OpenSSL due to a NULL pointer
8115 dereference. This could lead to a Denial Of Service attack. Thanks to
8116 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
8121 * Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
8122 dtls1_buffer_record function under certain conditions. In particular this
8123 could occur if an attacker sent repeated DTLS records with the same
8124 sequence number but for the next epoch. The memory leak could be exploited
8125 by an attacker in a Denial of Service attack through memory exhaustion.
8126 Thanks to Chris Mueller for reporting this issue.
8131 * Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
8132 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
8133 method would be set to NULL which could later result in a NULL pointer
8134 dereference. Thanks to Frank Schmirler for reporting this issue.
8139 * Abort handshake if server key exchange message is omitted for ephemeral
8142 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
8143 reporting this issue.
8148 * Remove non-export ephemeral RSA code on client and server. This code
8149 violated the TLS standard by allowing the use of temporary RSA keys in
8150 non-export ciphersuites and could be used by a server to effectively
8151 downgrade the RSA key length used to a value smaller than the server
8152 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
8153 INRIA or reporting this issue.
8158 * Fixed issue where DH client certificates are accepted without verification.
8159 An OpenSSL server will accept a DH certificate for client authentication
8160 without the certificate verify message. This effectively allows a client to
8161 authenticate without the use of a private key. This only affects servers
8162 which trust a client certificate authority which issues certificates
8163 containing DH keys: these are extremely rare and hardly ever encountered.
8164 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
8170 * Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
8171 results on some platforms, including x86_64. This bug occurs at random
8172 with a very low probability, and is not known to be exploitable in any
8173 way, though its exact impact is difficult to determine. Thanks to Pieter
8174 Wuille (Blockstream) who reported this issue and also suggested an initial
8175 fix. Further analysis was conducted by the OpenSSL development team and
8176 Adam Langley of Google. The final fix was developed by Andy Polyakov of
8177 the OpenSSL core team.
8182 * Fix various certificate fingerprint issues.
8184 By using non-DER or invalid encodings outside the signed portion of a
8185 certificate the fingerprint can be changed without breaking the signature.
8186 Although no details of the signed portion of the certificate can be changed
8187 this can cause problems with some applications: e.g. those using the
8188 certificate fingerprint for blacklists.
8190 1. Reject signatures with non zero unused bits.
8192 If the BIT STRING containing the signature has non zero unused bits reject
8193 the signature. All current signature algorithms require zero unused bits.
8195 2. Check certificate algorithm consistency.
8197 Check the AlgorithmIdentifier inside TBS matches the one in the
8198 certificate signature. NB: this will result in signature failure
8199 errors for some broken certificates.
8201 Thanks to Konrad Kraszewski from Google for reporting this issue.
8203 3. Check DSA/ECDSA signatures use DER.
8205 Reencode DSA/ECDSA signatures and compare with the original received
8206 signature. Return an error if there is a mismatch.
8208 This will reject various cases including garbage after signature
8209 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
8210 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
8211 (negative or with leading zeroes).
8213 Further analysis was conducted and fixes were developed by Stephen Henson
8214 of the OpenSSL core team.
8220 ### Changes between 1.0.0n and 1.0.0o [15 Oct 2014]
8222 * Session Ticket Memory Leak.
8224 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
8225 integrity of that ticket is first verified. In the event of a session
8226 ticket integrity check failing, OpenSSL will fail to free memory
8227 causing a memory leak. By sending a large number of invalid session
8228 tickets an attacker could exploit this issue in a Denial Of Service
8234 * Build option no-ssl3 is incomplete.
8236 When OpenSSL is configured with "no-ssl3" as a build option, servers
8237 could accept and complete a SSL 3.0 handshake, and clients could be
8238 configured to send them.
8241 *Akamai and the OpenSSL team*
8243 * Add support for TLS_FALLBACK_SCSV.
8244 Client applications doing fallback retries should call
8245 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
8248 *Adam Langley, Bodo Moeller*
8250 * Add additional DigestInfo checks.
8252 Reencode DigestInto in DER and check against the original when
8253 verifying RSA signature: this will reject any improperly encoded
8254 DigestInfo structures.
8256 Note: this is a precautionary measure and no attacks are currently known.
8260 ### Changes between 1.0.0m and 1.0.0n [6 Aug 2014]
8262 * OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
8263 to a denial of service attack. A malicious server can crash the client
8264 with a null pointer dereference (read) by specifying an anonymous (EC)DH
8265 ciphersuite and sending carefully crafted handshake messages.
8267 Thanks to Felix Gröbert (Google) for discovering and researching this
8273 * By sending carefully crafted DTLS packets an attacker could cause openssl
8274 to leak memory. This can be exploited through a Denial of Service attack.
8275 Thanks to Adam Langley for discovering and researching this issue.
8280 * An attacker can force openssl to consume large amounts of memory whilst
8281 processing DTLS handshake messages. This can be exploited through a
8282 Denial of Service attack.
8283 Thanks to Adam Langley for discovering and researching this issue.
8288 * An attacker can force an error condition which causes openssl to crash
8289 whilst processing DTLS packets due to memory being freed twice. This
8290 can be exploited through a Denial of Service attack.
8291 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
8297 * If a multithreaded client connects to a malicious server using a resumed
8298 session and the server sends an ec point format extension it could write
8299 up to 255 bytes to freed memory.
8301 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
8307 * A flaw in OBJ_obj2txt may cause pretty printing functions such as
8308 X509_name_oneline, X509_name_print_ex et al. to leak some information
8309 from the stack. Applications may be affected if they echo pretty printing
8310 output to the attacker.
8312 Thanks to Ivan Fratric (Google) for discovering this issue.
8315 *Emilia Käsper, and Steve Henson*
8317 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
8318 for corner cases. (Certain input points at infinity could lead to
8319 bogus results, with non-infinity inputs mapped to infinity too.)
8323 ### Changes between 1.0.0l and 1.0.0m [5 Jun 2014]
8325 * Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
8326 handshake can force the use of weak keying material in OpenSSL
8327 SSL/TLS clients and servers.
8329 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
8330 researching this issue. ([CVE-2014-0224])
8332 *KIKUCHI Masashi, Steve Henson*
8334 * Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
8335 OpenSSL DTLS client the code can be made to recurse eventually crashing
8338 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
8341 *Imre Rad, Steve Henson*
8343 * Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
8344 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
8345 client or server. This is potentially exploitable to run arbitrary
8346 code on a vulnerable client or server.
8348 Thanks to Jüri Aedla for reporting this issue. ([CVE-2014-0195])
8350 *Jüri Aedla, Steve Henson*
8352 * Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
8353 are subject to a denial of service attack.
8355 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
8356 this issue. ([CVE-2014-3470])
8358 *Felix Gröbert, Ivan Fratric, Steve Henson*
8360 * Harmonize version and its documentation. -f flag is used to display
8363 *mancha <mancha1@zoho.com>*
8365 * Fix eckey_priv_encode so it immediately returns an error upon a failure
8366 in i2d_ECPrivateKey.
8368 *mancha <mancha1@zoho.com>*
8370 * Fix some double frees. These are not thought to be exploitable.
8372 *mancha <mancha1@zoho.com>*
8374 * Fix for the attack described in the paper "Recovering OpenSSL
8375 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
8376 by Yuval Yarom and Naomi Benger. Details can be obtained from:
8377 <http://eprint.iacr.org/2014/140>
8379 Thanks to Yuval Yarom and Naomi Benger for discovering this
8380 flaw and to Yuval Yarom for supplying a fix ([CVE-2014-0076])
8382 *Yuval Yarom and Naomi Benger*
8384 ### Changes between 1.0.0k and 1.0.0l [6 Jan 2014]
8386 * Keep original DTLS digest and encryption contexts in retransmission
8387 structures so we can use the previous session parameters if they need
8388 to be resent. ([CVE-2013-6450])
8392 * Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
8393 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
8394 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
8395 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
8396 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
8397 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
8399 *Rob Stradling, Adam Langley*
8401 ### Changes between 1.0.0j and 1.0.0k [5 Feb 2013]
8403 * Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
8405 This addresses the flaw in CBC record processing discovered by
8406 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
8407 at: <http://www.isg.rhul.ac.uk/tls/>
8409 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
8410 Security Group at Royal Holloway, University of London
8411 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
8412 Emilia Käsper for the initial patch.
8415 *Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson*
8417 * Return an error when checking OCSP signatures when key is NULL.
8418 This fixes a DoS attack. ([CVE-2013-0166])
8422 * Call OCSP Stapling callback after ciphersuite has been chosen, so
8423 the right response is stapled. Also change SSL_get_certificate()
8424 so it returns the certificate actually sent.
8425 See <http://rt.openssl.org/Ticket/Display.html?id=2836>.
8426 (This is a backport)
8428 *Rob Stradling <rob.stradling@comodo.com>*
8430 * Fix possible deadlock when decoding public keys.
8434 ### Changes between 1.0.0i and 1.0.0j [10 May 2012]
8436 [NB: OpenSSL 1.0.0i and later 1.0.0 patch levels were released after
8439 * Sanity check record length before skipping explicit IV in DTLS
8442 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
8443 fuzzing as a service testing platform.
8448 * Initialise tkeylen properly when encrypting CMS messages.
8449 Thanks to Solar Designer of Openwall for reporting this issue.
8453 ### Changes between 1.0.0h and 1.0.0i [19 Apr 2012]
8455 * Check for potentially exploitable overflows in asn1_d2i_read_bio
8456 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
8457 in CRYPTO_realloc_clean.
8459 Thanks to Tavis Ormandy, Google Security Team, for discovering this
8460 issue and to Adam Langley <agl@chromium.org> for fixing it.
8463 *Adam Langley (Google), Tavis Ormandy, Google Security Team*
8465 ### Changes between 1.0.0g and 1.0.0h [12 Mar 2012]
8467 * Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness
8468 in CMS and PKCS7 code. When RSA decryption fails use a random key for
8469 content decryption and always return the same error. Note: this attack
8470 needs on average 2^20 messages so it only affects automated senders. The
8471 old behaviour can be re-enabled in the CMS code by setting the
8472 CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where
8473 an MMA defence is not necessary.
8474 Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering
8475 this issue. ([CVE-2012-0884])
8479 * Fix CVE-2011-4619: make sure we really are receiving a
8480 client hello before rejecting multiple SGC restarts. Thanks to
8481 Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug.
8485 ### Changes between 1.0.0f and 1.0.0g [18 Jan 2012]
8487 * Fix for DTLS DoS issue introduced by fix for CVE-2011-4109.
8488 Thanks to Antonio Martin, Enterprise Secure Access Research and
8489 Development, Cisco Systems, Inc. for discovering this bug and
8490 preparing a fix. ([CVE-2012-0050])
8494 ### Changes between 1.0.0e and 1.0.0f [4 Jan 2012]
8496 * Nadhem Alfardan and Kenny Paterson have discovered an extension
8497 of the Vaudenay padding oracle attack on CBC mode encryption
8498 which enables an efficient plaintext recovery attack against
8499 the OpenSSL implementation of DTLS. Their attack exploits timing
8500 differences arising during decryption processing. A research
8501 paper describing this attack can be found at:
8502 <http://www.isg.rhul.ac.uk/~kp/dtls.pdf>
8503 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
8504 Security Group at Royal Holloway, University of London
8505 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
8506 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
8507 for preparing the fix. ([CVE-2011-4108])
8509 *Robin Seggelmann, Michael Tuexen*
8511 * Clear bytes used for block padding of SSL 3.0 records.
8514 *Adam Langley (Google)*
8516 * Only allow one SGC handshake restart for SSL/TLS. Thanks to George
8517 Kadianakis <desnacked@gmail.com> for discovering this issue and
8518 Adam Langley for preparing the fix. ([CVE-2011-4619])
8520 *Adam Langley (Google)*
8522 * Check parameters are not NULL in GOST ENGINE. ([CVE-2012-0027])
8524 *Andrey Kulikov <amdeich@gmail.com>*
8526 * Prevent malformed RFC3779 data triggering an assertion failure.
8527 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
8528 and Rob Austein <sra@hactrn.net> for fixing it. ([CVE-2011-4577])
8530 *Rob Austein <sra@hactrn.net>*
8532 * Improved PRNG seeding for VOS.
8534 *Paul Green <Paul.Green@stratus.com>*
8536 * Fix ssl_ciph.c set-up race.
8538 *Adam Langley (Google)*
8540 * Fix spurious failures in ecdsatest.c.
8542 *Emilia Käsper (Google)*
8544 * Fix the BIO_f_buffer() implementation (which was mixing different
8545 interpretations of the `..._len` fields).
8547 *Adam Langley (Google)*
8549 * Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
8550 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
8551 threads won't reuse the same blinding coefficients.
8553 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
8554 lock to call BN_BLINDING_invert_ex, and avoids one use of
8555 BN_BLINDING_update for each BN_BLINDING structure (previously,
8556 the last update always remained unused).
8558 *Emilia Käsper (Google)*
8560 * In ssl3_clear, preserve s3->init_extra along with s3->rbuf.
8562 *Bob Buckholz (Google)*
8564 ### Changes between 1.0.0d and 1.0.0e [6 Sep 2011]
8566 * Fix bug where CRLs with nextUpdate in the past are sometimes accepted
8567 by initialising X509_STORE_CTX properly. ([CVE-2011-3207])
8569 *Kaspar Brand <ossl@velox.ch>*
8571 * Fix SSL memory handling for (EC)DH ciphersuites, in particular
8572 for multi-threaded use of ECDH. ([CVE-2011-3210])
8574 *Adam Langley (Google)*
8576 * Fix x509_name_ex_d2i memory leak on bad inputs.
8580 * Remove hard coded ecdsaWithSHA1 signature tests in ssl code and check
8581 signature public key algorithm by using OID xref utilities instead.
8582 Before this you could only use some ECC ciphersuites with SHA1 only.
8586 * Add protection against ECDSA timing attacks as mentioned in the paper
8587 by Billy Bob Brumley and Nicola Tuveri, see:
8588 <http://eprint.iacr.org/2011/232.pdf>
8590 *Billy Bob Brumley and Nicola Tuveri*
8592 ### Changes between 1.0.0c and 1.0.0d [8 Feb 2011]
8594 * Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
8596 *Neel Mehta, Adam Langley, Bodo Moeller (Google)*
8598 * Fix bug in string printing code: if *any* escaping is enabled we must
8599 escape the escape character (backslash) or the resulting string is
8604 ### Changes between 1.0.0b and 1.0.0c [2 Dec 2010]
8606 * Disable code workaround for ancient and obsolete Netscape browsers
8607 and servers: an attacker can use it in a ciphersuite downgrade attack.
8608 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
8612 * Fixed J-PAKE implementation error, originally discovered by
8613 Sebastien Martini, further info and confirmation from Stefan
8614 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
8618 ### Changes between 1.0.0a and 1.0.0b [16 Nov 2010]
8620 * Fix extension code to avoid race conditions which can result in a buffer
8621 overrun vulnerability: resumed sessions must not be modified as they can
8622 be shared by multiple threads. CVE-2010-3864
8626 * Fix WIN32 build system to correctly link an ENGINE directory into
8631 ### Changes between 1.0.0 and 1.0.0a [01 Jun 2010]
8633 * Check return value of int_rsa_verify in pkey_rsa_verifyrecover
8636 *Steve Henson, Peter-Michael Hager <hager@dortmund.net>*
8638 ### Changes between 0.9.8n and 1.0.0 [29 Mar 2010]
8640 * Add "missing" function EVP_CIPHER_CTX_copy(). This copies a cipher
8641 context. The operation can be customised via the ctrl mechanism in
8642 case ENGINEs want to include additional functionality.
8646 * Tolerate yet another broken PKCS#8 key format: private key value negative.
8650 * Add new -subject_hash_old and -issuer_hash_old options to x509 utility to
8651 output hashes compatible with older versions of OpenSSL.
8653 *Willy Weisz <weisz@vcpc.univie.ac.at>*
8655 * Fix compression algorithm handling: if resuming a session use the
8656 compression algorithm of the resumed session instead of determining
8657 it from client hello again. Don't allow server to change algorithm.
8661 * Add load_crls() function to commands tidying load_certs() too. Add option
8662 to verify utility to allow additional CRLs to be included.
8666 * Update OCSP request code to permit adding custom headers to the request:
8667 some responders need this.
8671 * The function EVP_PKEY_sign() returns <=0 on error: check return code
8674 *Julia Lawall <julia@diku.dk>*
8676 * Update verify callback code in `apps/s_cb.c` and `apps/verify.c`, it
8677 needlessly dereferenced structures, used obsolete functions and
8678 didn't handle all updated verify codes correctly.
8682 * Disable MD2 in the default configuration.
8686 * In BIO_pop() and BIO_push() use the ctrl argument (which was NULL) to
8687 indicate the initial BIO being pushed or popped. This makes it possible
8688 to determine whether the BIO is the one explicitly called or as a result
8689 of the ctrl being passed down the chain. Fix BIO_pop() and SSL BIOs so
8690 it handles reference counts correctly and doesn't zero out the I/O bio
8691 when it is not being explicitly popped. WARNING: applications which
8692 included workarounds for the old buggy behaviour will need to be modified
8693 or they could free up already freed BIOs.
8697 * Extend the uni2asc/asc2uni => OPENSSL_uni2asc/OPENSSL_asc2uni
8698 renaming to all platforms (within the 0.9.8 branch, this was
8699 done conditionally on Netware platforms to avoid a name clash).
8701 *Guenter <lists@gknw.net>*
8703 * Add ECDHE and PSK support to DTLS.
8705 *Michael Tuexen <tuexen@fh-muenster.de>*
8707 * Add CHECKED_STACK_OF macro to safestack.h, otherwise safestack can't
8712 * Add "missing" function EVP_MD_flags() (without this the only way to
8713 retrieve a digest flags is by accessing the structure directly. Update
8714 `EVP_MD_do_all*()` and `EVP_CIPHER_do_all*()` to include the name a digest
8715 or cipher is registered as in the "from" argument. Print out all
8716 registered digests in the dgst usage message instead of manually
8717 attempting to work them out.
8721 * If no SSLv2 ciphers are used don't use an SSLv2 compatible client hello:
8722 this allows the use of compression and extensions. Change default cipher
8723 string to remove SSLv2 ciphersuites. This effectively avoids ancient SSLv2
8724 by default unless an application cipher string requests it.
8728 * Alter match criteria in PKCS12_parse(). It used to try to use local
8729 key ids to find matching certificates and keys but some PKCS#12 files
8730 don't follow the (somewhat unwritten) rules and this strategy fails.
8731 Now just gather all certificates together and the first private key
8732 then look for the first certificate that matches the key.
8736 * Support use of registered digest and cipher names for dgst and cipher
8737 commands instead of having to add each one as a special case. So now
8744 openssl dgst -sha256 foo
8746 and this works for ENGINE based algorithms too.
8750 * Update Gost ENGINE to support parameter files.
8752 *Victor B. Wagner <vitus@cryptocom.ru>*
8754 * Support GeneralizedTime in ca utility.
8756 *Oliver Martin <oliver@volatilevoid.net>, Steve Henson*
8758 * Enhance the hash format used for certificate directory links. The new
8759 form uses the canonical encoding (meaning equivalent names will work
8760 even if they aren't identical) and uses SHA1 instead of MD5. This form
8761 is incompatible with the older format and as a result c_rehash should
8762 be used to rebuild symbolic links.
8766 * Make PKCS#8 the default write format for private keys, replacing the
8767 traditional format. This form is standardised, more secure and doesn't
8768 include an implicit MD5 dependency.
8772 * Add a $gcc_devteam_warn option to Configure. The idea is that any code
8773 committed to OpenSSL should pass this lot as a minimum.
8777 * Add session ticket override functionality for use by EAP-FAST.
8779 *Jouni Malinen <j@w1.fi>*
8781 * Modify HMAC functions to return a value. Since these can be implemented
8782 in an ENGINE errors can occur.
8786 * Type-checked OBJ_bsearch_ex.
8790 * Type-checked OBJ_bsearch. Also some constification necessitated
8791 by type-checking. Still to come: TXT_DB, bsearch(?),
8792 OBJ_bsearch_ex, qsort, CRYPTO_EX_DATA, ASN1_VALUE, ASN1_STRING,
8797 * New function OPENSSL_gmtime_adj() to add a specific number of days and
8798 seconds to a tm structure directly, instead of going through OS
8799 specific date routines. This avoids any issues with OS routines such
8800 as the year 2038 bug. New `*_adj()` functions for ASN1 time structures
8801 and X509_time_adj_ex() to cover the extended range. The existing
8802 X509_time_adj() is still usable and will no longer have any date issues.
8806 * Delta CRL support. New use deltas option which will attempt to locate
8807 and search any appropriate delta CRLs available.
8809 This work was sponsored by Google.
8813 * Support for CRLs partitioned by reason code. Reorganise CRL processing
8814 code and add additional score elements. Validate alternate CRL paths
8815 as part of the CRL checking and indicate a new error "CRL path validation
8816 error" in this case. Applications wanting additional details can use
8817 the verify callback and check the new "parent" field. If this is not
8818 NULL CRL path validation is taking place. Existing applications won't
8819 see this because it requires extended CRL support which is off by
8822 This work was sponsored by Google.
8826 * Support for freshest CRL extension.
8828 This work was sponsored by Google.
8832 * Initial indirect CRL support. Currently only supported in the CRLs
8833 passed directly and not via lookup. Process certificate issuer
8834 CRL entry extension and lookup CRL entries by bother issuer name
8835 and serial number. Check and process CRL issuer entry in IDP extension.
8837 This work was sponsored by Google.
8841 * Add support for distinct certificate and CRL paths. The CRL issuer
8842 certificate is validated separately in this case. Only enabled if
8843 an extended CRL support flag is set: this flag will enable additional
8844 CRL functionality in future.
8846 This work was sponsored by Google.
8850 * Add support for policy mappings extension.
8852 This work was sponsored by Google.
8856 * Fixes to pathlength constraint, self issued certificate handling,
8857 policy processing to align with RFC3280 and PKITS tests.
8859 This work was sponsored by Google.
8863 * Support for name constraints certificate extension. DN, email, DNS
8864 and URI types are currently supported.
8866 This work was sponsored by Google.
8870 * To cater for systems that provide a pointer-based thread ID rather
8871 than numeric, deprecate the current numeric thread ID mechanism and
8872 replace it with a structure and associated callback type. This
8873 mechanism allows a numeric "hash" to be extracted from a thread ID in
8874 either case, and on platforms where pointers are larger than 'long',
8875 mixing is done to help ensure the numeric 'hash' is usable even if it
8876 can't be guaranteed unique. The default mechanism is to use "&errno"
8877 as a pointer-based thread ID to distinguish between threads.
8879 Applications that want to provide their own thread IDs should now use
8880 CRYPTO_THREADID_set_callback() to register a callback that will call
8881 either CRYPTO_THREADID_set_numeric() or CRYPTO_THREADID_set_pointer().
8883 Note that ERR_remove_state() is now deprecated, because it is tied
8884 to the assumption that thread IDs are numeric. ERR_remove_state(0)
8885 to free the current thread's error state should be replaced by
8886 ERR_remove_thread_state(NULL).
8888 (This new approach replaces the functions CRYPTO_set_idptr_callback(),
8889 CRYPTO_get_idptr_callback(), and CRYPTO_thread_idptr() that existed in
8890 OpenSSL 0.9.9-dev between June 2006 and August 2008. Also, if an
8891 application was previously providing a numeric thread callback that
8892 was inappropriate for distinguishing threads, then uniqueness might
8893 have been obtained with &errno that happened immediately in the
8894 intermediate development versions of OpenSSL; this is no longer the
8895 case, the numeric thread callback will now override the automatic use
8898 *Geoff Thorpe, with help from Bodo Moeller*
8900 * Initial support for different CRL issuing certificates. This covers a
8901 simple case where the self issued certificates in the chain exist and
8902 the real CRL issuer is higher in the existing chain.
8904 This work was sponsored by Google.
8908 * Removed effectively defunct crypto/store from the build.
8912 * Revamp of STACK to provide stronger type-checking. Still to come:
8913 TXT_DB, bsearch(?), OBJ_bsearch, qsort, CRYPTO_EX_DATA, ASN1_VALUE,
8914 ASN1_STRING, CONF_VALUE.
8918 * Add a new SSL_MODE_RELEASE_BUFFERS mode flag to release unused buffer
8919 RAM on SSL connections. This option can save about 34k per idle SSL.
8923 * Revamp of LHASH to provide stronger type-checking. Still to come:
8924 STACK, TXT_DB, bsearch, qsort.
8928 * Initial support for Cryptographic Message Syntax (aka CMS) based
8929 on RFC3850, RFC3851 and RFC3852. New cms directory and cms utility,
8930 support for data, signedData, compressedData, digestedData and
8931 encryptedData, envelopedData types included. Scripts to check against
8932 RFC4134 examples draft and interop and consistency checks of many
8933 content types and variants.
8937 * Add options to enc utility to support use of zlib compression BIO.
8941 * Extend mk1mf to support importing of options and assembly language
8942 files from Configure script, currently only included in VC-WIN32.
8943 The assembly language rules can now optionally generate the source
8944 files from the associated perl scripts.
8948 * Implement remaining functionality needed to support GOST ciphersuites.
8949 Interop testing has been performed using CryptoPro implementations.
8951 *Victor B. Wagner <vitus@cryptocom.ru>*
8953 * s390x assembler pack.
8957 * ARMv4 assembler pack. ARMv4 refers to v4 and later ISA, not CPU
8962 * Implement Opaque PRF Input TLS extension as specified in
8963 draft-rescorla-tls-opaque-prf-input-00.txt. Since this is not an
8964 official specification yet and no extension type assignment by
8965 IANA exists, this extension (for now) will have to be explicitly
8966 enabled when building OpenSSL by providing the extension number
8967 to use. For example, specify an option
8969 -DTLSEXT_TYPE_opaque_prf_input=0x9527
8971 to the "config" or "Configure" script to enable the extension,
8972 assuming extension number 0x9527 (which is a completely arbitrary
8973 and unofficial assignment based on the MD5 hash of the Internet
8974 Draft). Note that by doing so, you potentially lose
8975 interoperability with other TLS implementations since these might
8976 be using the same extension number for other purposes.
8978 SSL_set_tlsext_opaque_prf_input(ssl, src, len) is used to set the
8979 opaque PRF input value to use in the handshake. This will create
8980 an internal copy of the length-'len' string at 'src', and will
8981 return non-zero for success.
8983 To get more control and flexibility, provide a callback function
8986 SSL_CTX_set_tlsext_opaque_prf_input_callback(ctx, cb)
8987 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg(ctx, arg)
8991 int (*cb)(SSL *, void *peerinput, size_t len, void *arg);
8994 Callback function 'cb' will be called in handshakes, and is
8995 expected to use SSL_set_tlsext_opaque_prf_input() as appropriate.
8996 Argument 'arg' is for application purposes (the value as given to
8997 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg() will directly
8998 be provided to the callback function). The callback function
8999 has to return non-zero to report success: usually 1 to use opaque
9000 PRF input just if possible, or 2 to enforce use of the opaque PRF
9001 input. In the latter case, the library will abort the handshake
9002 if opaque PRF input is not successfully negotiated.
9004 Arguments 'peerinput' and 'len' given to the callback function
9005 will always be NULL and 0 in the case of a client. A server will
9006 see the client's opaque PRF input through these variables if
9007 available (NULL and 0 otherwise). Note that if the server
9008 provides an opaque PRF input, the length must be the same as the
9009 length of the client's opaque PRF input.
9011 Note that the callback function will only be called when creating
9012 a new session (session resumption can resume whatever was
9013 previously negotiated), and will not be called in SSL 2.0
9014 handshakes; thus, SSL_CTX_set_options(ctx, SSL_OP_NO_SSLv2) or
9015 SSL_set_options(ssl, SSL_OP_NO_SSLv2) is especially recommended
9016 for applications that need to enforce opaque PRF input.
9020 * Update ssl code to support digests other than SHA1+MD5 for handshake
9023 *Victor B. Wagner <vitus@cryptocom.ru>*
9025 * Add RFC4507 support to OpenSSL. This includes the corrections in
9026 RFC4507bis. The encrypted ticket format is an encrypted encoded
9027 SSL_SESSION structure, that way new session features are automatically
9030 If a client application caches session in an SSL_SESSION structure
9031 support is transparent because tickets are now stored in the encoded
9034 The SSL_CTX structure automatically generates keys for ticket
9035 protection in servers so again support should be possible
9036 with no application modification.
9038 If a client or server wishes to disable RFC4507 support then the option
9039 SSL_OP_NO_TICKET can be set.
9041 Add a TLS extension debugging callback to allow the contents of any client
9042 or server extensions to be examined.
9044 This work was sponsored by Google.
9048 * Final changes to avoid use of pointer pointer casts in OpenSSL.
9049 OpenSSL should now compile cleanly on gcc 4.2
9051 *Peter Hartley <pdh@utter.chaos.org.uk>, Steve Henson*
9053 * Update SSL library to use new EVP_PKEY MAC API. Include generic MAC
9054 support including streaming MAC support: this is required for GOST
9055 ciphersuite support.
9057 *Victor B. Wagner <vitus@cryptocom.ru>, Steve Henson*
9059 * Add option -stream to use PKCS#7 streaming in smime utility. New
9060 function i2d_PKCS7_bio_stream() and PEM_write_PKCS7_bio_stream()
9061 to output in BER and PEM format.
9065 * Experimental support for use of HMAC via EVP_PKEY interface. This
9066 allows HMAC to be handled via the `EVP_DigestSign*()` interface. The
9067 EVP_PKEY "key" in this case is the HMAC key, potentially allowing
9068 ENGINE support for HMAC keys which are unextractable. New -mac and
9069 -macopt options to dgst utility.
9073 * New option -sigopt to dgst utility. Update dgst to use
9074 `EVP_Digest{Sign,Verify}*`. These two changes make it possible to use
9075 alternative signing parameters such as X9.31 or PSS in the dgst
9080 * Change ssl_cipher_apply_rule(), the internal function that does
9081 the work each time a ciphersuite string requests enabling
9082 ("foo+bar"), moving ("+foo+bar"), disabling ("-foo+bar", or
9083 removing ("!foo+bar") a class of ciphersuites: Now it maintains
9084 the order of disabled ciphersuites such that those ciphersuites
9085 that most recently went from enabled to disabled not only stay
9086 in order with respect to each other, but also have higher priority
9087 than other disabled ciphersuites the next time ciphersuites are
9090 This means that you can now say, e.g., "PSK:-PSK:HIGH" to enable
9091 the same ciphersuites as with "HIGH" alone, but in a specific
9092 order where the PSK ciphersuites come first (since they are the
9093 most recently disabled ciphersuites when "HIGH" is parsed).
9095 Also, change ssl_create_cipher_list() (using this new
9096 functionality) such that between otherwise identical
9097 ciphersuites, ephemeral ECDH is preferred over ephemeral DH in
9102 * Change ssl_create_cipher_list() so that it automatically
9103 arranges the ciphersuites in reasonable order before starting
9104 to process the rule string. Thus, the definition for "DEFAULT"
9105 (SSL_DEFAULT_CIPHER_LIST) now is just "ALL:!aNULL:!eNULL", but
9106 remains equivalent to `"AES:ALL:!aNULL:!eNULL:+aECDH:+kRSA:+RC4:@STRENGTH"`.
9107 This makes it much easier to arrive at a reasonable default order
9108 in applications for which anonymous ciphers are OK (meaning
9109 that you can't actually use DEFAULT).
9111 *Bodo Moeller; suggested by Victor Duchovni*
9113 * Split the SSL/TLS algorithm mask (as used for ciphersuite string
9114 processing) into multiple integers instead of setting
9115 "SSL_MKEY_MASK" bits, "SSL_AUTH_MASK" bits, "SSL_ENC_MASK",
9116 "SSL_MAC_MASK", and "SSL_SSL_MASK" bits all in a single integer.
9117 (These masks as well as the individual bit definitions are hidden
9118 away into the non-exported interface ssl/ssl_locl.h, so this
9119 change to the definition of the SSL_CIPHER structure shouldn't
9120 affect applications.) This give us more bits for each of these
9121 categories, so there is no longer a need to coagulate AES128 and
9122 AES256 into a single algorithm bit, and to coagulate Camellia128
9123 and Camellia256 into a single algorithm bit, which has led to all
9126 Thus, among other things, the kludge introduced in 0.9.7m and
9127 0.9.8e for masking out AES256 independently of AES128 or masking
9128 out Camellia256 independently of AES256 is not needed here in 0.9.9.
9130 With the change, we also introduce new ciphersuite aliases that
9131 so far were missing: "AES128", "AES256", "CAMELLIA128", and
9136 * Add support for dsa-with-SHA224 and dsa-with-SHA256.
9137 Use the leftmost N bytes of the signature input if the input is
9138 larger than the prime q (with N being the size in bytes of q).
9142 * Very *very* experimental PKCS#7 streaming encoder support. Nothing uses
9143 it yet and it is largely untested.
9147 * Add support for the ecdsa-with-SHA224/256/384/512 signature types.
9151 * Initial incomplete changes to avoid need for function casts in OpenSSL
9152 some compilers (gcc 4.2 and later) reject their use. Safestack is
9153 reimplemented. Update ASN1 to avoid use of legacy functions.
9157 * Win32/64 targets are linked with Winsock2.
9161 * Add an X509_CRL_METHOD structure to allow CRL processing to be redirected
9162 to external functions. This can be used to increase CRL handling
9163 efficiency especially when CRLs are very large by (for example) storing
9164 the CRL revoked certificates in a database.
9168 * Overhaul of by_dir code. Add support for dynamic loading of CRLs so
9169 new CRLs added to a directory can be used. New command line option
9170 -verify_return_error to s_client and s_server. This causes real errors
9171 to be returned by the verify callback instead of carrying on no matter
9172 what. This reflects the way a "real world" verify callback would behave.
9176 * GOST engine, supporting several GOST algorithms and public key formats.
9177 Kindly donated by Cryptocom.
9181 * Partial support for Issuing Distribution Point CRL extension. CRLs
9182 partitioned by DP are handled but no indirect CRL or reason partitioning
9183 (yet). Complete overhaul of CRL handling: now the most suitable CRL is
9184 selected via a scoring technique which handles IDP and AKID in CRLs.
9188 * New X509_STORE_CTX callbacks lookup_crls() and lookup_certs() which
9189 will ultimately be used for all verify operations: this will remove the
9190 X509_STORE dependency on certificate verification and allow alternative
9191 lookup methods. X509_STORE based implementations of these two callbacks.
9195 * Allow multiple CRLs to exist in an X509_STORE with matching issuer names.
9196 Modify get_crl() to find a valid (unexpired) CRL if possible.
9200 * New function X509_CRL_match() to check if two CRLs are identical. Normally
9201 this would be called X509_CRL_cmp() but that name is already used by
9202 a function that just compares CRL issuer names. Cache several CRL
9203 extensions in X509_CRL structure and cache CRLDP in X509.
9207 * Store a "canonical" representation of X509_NAME structure (ASN1 Name)
9208 this maps equivalent X509_NAME structures into a consistent structure.
9209 Name comparison can then be performed rapidly using memcmp().
9213 * Non-blocking OCSP request processing. Add -timeout option to ocsp
9218 * Allow digests to supply their own micalg string for S/MIME type using
9219 the ctrl EVP_MD_CTRL_MICALG.
9223 * During PKCS7 signing pass the PKCS7 SignerInfo structure to the
9224 EVP_PKEY_METHOD before and after signing via the EVP_PKEY_CTRL_PKCS7_SIGN
9225 ctrl. It can then customise the structure before and/or after signing
9230 * New function OBJ_add_sigid() to allow application defined signature OIDs
9231 to be added to OpenSSLs internal tables. New function OBJ_sigid_free()
9232 to free up any added signature OIDs.
9236 * New functions EVP_CIPHER_do_all(), EVP_CIPHER_do_all_sorted(),
9237 EVP_MD_do_all() and EVP_MD_do_all_sorted() to enumerate internal
9238 digest and cipher tables. New options added to openssl utility:
9239 list-message-digest-algorithms and list-cipher-algorithms.
9243 * Change the array representation of binary polynomials: the list
9244 of degrees of non-zero coefficients is now terminated with -1.
9245 Previously it was terminated with 0, which was also part of the
9246 value; thus, the array representation was not applicable to
9247 polynomials where t^0 has coefficient zero. This change makes
9248 the array representation useful in a more general context.
9252 * Various modifications and fixes to SSL/TLS cipher string
9253 handling. For ECC, the code now distinguishes between fixed ECDH
9254 with RSA certificates on the one hand and with ECDSA certificates
9255 on the other hand, since these are separate ciphersuites. The
9256 unused code for Fortezza ciphersuites has been removed.
9258 For consistency with EDH, ephemeral ECDH is now called "EECDH"
9259 (not "ECDHE"). For consistency with the code for DH
9260 certificates, use of ECDH certificates is now considered ECDH
9261 authentication, not RSA or ECDSA authentication (the latter is
9262 merely the CA's signing algorithm and not actively used in the
9265 The temporary ciphersuite alias "ECCdraft" is no longer
9266 available, and ECC ciphersuites are no longer excluded from "ALL"
9267 and "DEFAULT". The following aliases now exist for RFC 4492
9268 ciphersuites, most of these by analogy with the DH case:
9270 kECDHr - ECDH cert, signed with RSA
9271 kECDHe - ECDH cert, signed with ECDSA
9272 kECDH - ECDH cert (signed with either RSA or ECDSA)
9273 kEECDH - ephemeral ECDH
9274 ECDH - ECDH cert or ephemeral ECDH
9280 AECDH - anonymous ECDH
9281 EECDH - non-anonymous ephemeral ECDH (equivalent to "kEECDH:-AECDH")
9285 * Add additional S/MIME capabilities for AES and GOST ciphers if supported.
9286 Use correct micalg parameters depending on digest(s) in signed message.
9290 * Add engine support for EVP_PKEY_ASN1_METHOD. Add functions to process
9291 an ENGINE asn1 method. Support ENGINE lookups in the ASN1 code.
9295 * Initial engine support for EVP_PKEY_METHOD. New functions to permit
9296 an engine to register a method. Add ENGINE lookups for methods and
9297 functional reference processing.
9301 * New functions `EVP_Digest{Sign,Verify)*`. These are enhanced versions of
9302 `EVP_{Sign,Verify}*` which allow an application to customise the signature
9307 * New -resign option to smime utility. This adds one or more signers
9308 to an existing PKCS#7 signedData structure. Also -md option to use an
9309 alternative message digest algorithm for signing.
9313 * Tidy up PKCS#7 routines and add new functions to make it easier to
9314 create PKCS7 structures containing multiple signers. Update smime
9315 application to support multiple signers.
9319 * New -macalg option to pkcs12 utility to allow setting of an alternative
9324 * Initial support for PKCS#5 v2.0 PRFs other than default SHA1 HMAC.
9325 Reorganize PBE internals to lookup from a static table using NIDs,
9326 add support for HMAC PBE OID translation. Add a EVP_CIPHER ctrl:
9327 EVP_CTRL_PBE_PRF_NID this allows a cipher to specify an alternative
9328 PRF which will be automatically used with PBES2.
9332 * Replace the algorithm specific calls to generate keys in "req" with the
9337 * Update PKCS#7 enveloped data routines to use new API. This is now
9338 supported by any public key method supporting the encrypt operation. A
9339 ctrl is added to allow the public key algorithm to examine or modify
9340 the PKCS#7 RecipientInfo structure if it needs to: for RSA this is
9345 * Add a ctrl to asn1 method to allow a public key algorithm to express
9346 a default digest type to use. In most cases this will be SHA1 but some
9347 algorithms (such as GOST) need to specify an alternative digest. The
9348 return value indicates how strong the preference is 1 means optional and
9349 2 is mandatory (that is it is the only supported type). Modify
9350 ASN1_item_sign() to accept a NULL digest argument to indicate it should
9351 use the default md. Update openssl utilities to use the default digest
9352 type for signing if it is not explicitly indicated.
9356 * Use OID cross reference table in ASN1_sign() and ASN1_verify(). New
9357 EVP_MD flag EVP_MD_FLAG_PKEY_METHOD_SIGNATURE. This uses the relevant
9358 signing method from the key type. This effectively removes the link
9359 between digests and public key types.
9363 * Add an OID cross reference table and utility functions. Its purpose is to
9364 translate between signature OIDs such as SHA1WithrsaEncryption and SHA1,
9365 rsaEncryption. This will allow some of the algorithm specific hackery
9366 needed to use the correct OID to be removed.
9370 * Remove algorithm specific dependencies when setting PKCS7_SIGNER_INFO
9371 structures for PKCS7_sign(). They are now set up by the relevant public
9376 * Add provisional EC pkey method with support for ECDSA and ECDH.
9380 * Add support for key derivation (agreement) in the API, DH method and
9385 * Add DSA pkey method and DH pkey methods, extend DH ASN1 method to support
9386 public and private key formats. As a side effect these add additional
9387 command line functionality not previously available: DSA signatures can be
9388 generated and verified using pkeyutl and DH key support and generation in
9395 *Oliver Tappe <zooey@hirschkaefer.de>*
9397 * New make target "install_html_docs" installs HTML renditions of the
9400 *Oliver Tappe <zooey@hirschkaefer.de>*
9402 * New utility "genpkey" this is analogous to "genrsa" etc except it can
9403 generate keys for any algorithm. Extend and update EVP_PKEY_METHOD to
9404 support key and parameter generation and add initial key generation
9405 functionality for RSA.
9409 * Add functions for main EVP_PKEY_method operations. The undocumented
9410 functions `EVP_PKEY_{encrypt,decrypt}` have been renamed to
9411 `EVP_PKEY_{encrypt,decrypt}_old`.
9415 * Initial definitions for EVP_PKEY_METHOD. This will be a high level public
9416 key API, doesn't do much yet.
9420 * New function EVP_PKEY_asn1_get0_info() to retrieve information about
9421 public key algorithms. New option to openssl utility:
9422 "list-public-key-algorithms" to print out info.
9426 * Implement the Supported Elliptic Curves Extension for
9427 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
9431 * Don't free up OIDs in OBJ_cleanup() if they are in use by EVP_MD or
9432 EVP_CIPHER structures to avoid later problems in EVP_cleanup().
9436 * New utilities pkey and pkeyparam. These are similar to algorithm specific
9437 utilities such as rsa, dsa, dsaparam etc except they process any key
9442 * Transfer public key printing routines to EVP_PKEY_ASN1_METHOD. New
9443 functions EVP_PKEY_print_public(), EVP_PKEY_print_private(),
9444 EVP_PKEY_print_param() to print public key data from an EVP_PKEY
9449 * Initial support for pluggable public key ASN1.
9450 De-spaghettify the public key ASN1 handling. Move public and private
9451 key ASN1 handling to a new EVP_PKEY_ASN1_METHOD structure. Relocate
9452 algorithm specific handling to a single module within the relevant
9453 algorithm directory. Add functions to allow (near) opaque processing
9454 of public and private key structures.
9458 * Implement the Supported Point Formats Extension for
9459 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
9463 * Add initial support for RFC 4279 PSK TLS ciphersuites. Add members
9464 for the psk identity [hint] and the psk callback functions to the
9465 SSL_SESSION, SSL and SSL_CTX structure.
9468 PSK-RC4-SHA, PSK-3DES-EDE-CBC-SHA, PSK-AES128-CBC-SHA,
9472 SSL_CTX_use_psk_identity_hint
9473 SSL_get_psk_identity_hint
9474 SSL_get_psk_identity
9475 SSL_use_psk_identity_hint
9477 *Mika Kousa and Pasi Eronen of Nokia Corporation*
9479 * Add RFC 3161 compliant time stamp request creation, response generation
9480 and response verification functionality.
9482 *Zoltán Glózik <zglozik@opentsa.org>, The OpenTSA Project*
9484 * Add initial support for TLS extensions, specifically for the server_name
9485 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
9486 have new members for a host name. The SSL data structure has an
9487 additional member `SSL_CTX *initial_ctx` so that new sessions can be
9488 stored in that context to allow for session resumption, even after the
9489 SSL has been switched to a new SSL_CTX in reaction to a client's
9490 server_name extension.
9492 New functions (subject to change):
9494 SSL_get_servername()
9495 SSL_get_servername_type()
9498 New CTRL codes and macros (subject to change):
9500 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
9501 - SSL_CTX_set_tlsext_servername_callback()
9502 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
9503 - SSL_CTX_set_tlsext_servername_arg()
9504 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
9506 openssl s_client has a new '-servername ...' option.
9508 openssl s_server has new options '-servername_host ...', '-cert2 ...',
9509 '-key2 ...', '-servername_fatal' (subject to change). This allows
9510 testing the HostName extension for a specific single host name ('-cert'
9511 and '-key' remain fallbacks for handshakes without HostName
9512 negotiation). If the unrecognized_name alert has to be sent, this by
9513 default is a warning; it becomes fatal with the '-servername_fatal'
9516 *Peter Sylvester, Remy Allais, Christophe Renou*
9518 * Whirlpool hash implementation is added.
9522 * BIGNUM code on 64-bit SPARCv9 targets is switched from bn(64,64) to
9523 bn(64,32). Because of instruction set limitations it doesn't have
9524 any negative impact on performance. This was done mostly in order
9525 to make it possible to share assembler modules, such as bn_mul_mont
9526 implementations, between 32- and 64-bit builds without hassle.
9530 * Move code previously exiled into file crypto/ec/ec2_smpt.c
9531 to ec2_smpl.c, and no longer require the OPENSSL_EC_BIN_PT_COMP
9536 * New candidate for BIGNUM assembler implementation, bn_mul_mont,
9537 dedicated Montgomery multiplication procedure, is introduced.
9538 BN_MONT_CTX is modified to allow bn_mul_mont to reach for higher
9539 "64-bit" performance on certain 32-bit targets.
9543 * New option SSL_OP_NO_COMP to disable use of compression selectively
9544 in SSL structures. New SSL ctrl to set maximum send fragment size.
9545 Save memory by setting the I/O buffer sizes dynamically instead of
9546 using the maximum available value.
9550 * New option -V for 'openssl ciphers'. This prints the ciphersuite code
9551 in addition to the text details.
9555 * Very, very preliminary EXPERIMENTAL support for printing of general
9556 ASN1 structures. This currently produces rather ugly output and doesn't
9557 handle several customised structures at all.
9561 * Integrated support for PVK file format and some related formats such
9562 as MS PUBLICKEYBLOB and PRIVATEKEYBLOB. Command line switches to support
9563 these in the 'rsa' and 'dsa' utilities.
9567 * Support for PKCS#1 RSAPublicKey format on rsa utility command line.
9571 * Remove the ancient ASN1_METHOD code. This was only ever used in one
9572 place for the (very old) "NETSCAPE" format certificates which are now
9573 handled using new ASN1 code equivalents.
9577 * Let the TLSv1_method() etc. functions return a 'const' SSL_METHOD
9578 pointer and make the SSL_METHOD parameter in SSL_CTX_new,
9579 SSL_CTX_set_ssl_version and SSL_set_ssl_method 'const'.
9583 * Modify CRL distribution points extension code to print out previously
9584 unsupported fields. Enhance extension setting code to allow setting of
9589 * Add print and set support for Issuing Distribution Point CRL extension.
9593 * Change 'Configure' script to enable Camellia by default.
9600 ### Changes between 0.9.8m and 0.9.8n [24 Mar 2010]
9602 * When rejecting SSL/TLS records due to an incorrect version number, never
9603 update s->server with a new major version number. As of
9604 - OpenSSL 0.9.8m if 'short' is a 16-bit type,
9605 - OpenSSL 0.9.8f if 'short' is longer than 16 bits,
9606 the previous behavior could result in a read attempt at NULL when
9607 receiving specific incorrect SSL/TLS records once record payload
9608 protection is active. ([CVE-2010-0740])
9610 *Bodo Moeller, Adam Langley <agl@chromium.org>*
9612 * Fix for CVE-2010-0433 where some kerberos enabled versions of OpenSSL
9613 could be crashed if the relevant tables were not present (e.g. chrooted).
9615 *Tomas Hoger <thoger@redhat.com>*
9617 ### Changes between 0.9.8l and 0.9.8m [25 Feb 2010]
9619 * Always check bn_wexpand() return values for failure. ([CVE-2009-3245])
9621 *Martin Olsson, Neel Mehta*
9623 * Fix X509_STORE locking: Every 'objs' access requires a lock (to
9624 accommodate for stack sorting, always a write lock!).
9628 * On some versions of WIN32 Heap32Next is very slow. This can cause
9629 excessive delays in the RAND_poll(): over a minute. As a workaround
9630 include a time check in the inner Heap32Next loop too.
9634 * The code that handled flushing of data in SSL/TLS originally used the
9635 BIO_CTRL_INFO ctrl to see if any data was pending first. This caused
9636 the problem outlined in PR#1949. The fix suggested there however can
9637 trigger problems with buggy BIO_CTRL_WPENDING (e.g. some versions
9638 of Apache). So instead simplify the code to flush unconditionally.
9639 This should be fine since flushing with no data to flush is a no op.
9643 * Handle TLS versions 2.0 and later properly and correctly use the
9644 highest version of TLS/SSL supported. Although TLS >= 2.0 is some way
9645 off ancient servers have a habit of sticking around for a while...
9649 * Modify compression code so it frees up structures without using the
9650 ex_data callbacks. This works around a problem where some applications
9651 call CRYPTO_cleanup_all_ex_data() before application exit (e.g. when
9652 restarting) then use compression (e.g. SSL with compression) later.
9653 This results in significant per-connection memory leaks and
9654 has caused some security issues including CVE-2008-1678 and
9659 * Constify crypto/cast (i.e., <openssl/cast.h>): a CAST_KEY doesn't
9660 change when encrypting or decrypting.
9664 * Add option SSL_OP_LEGACY_SERVER_CONNECT which will allow clients to
9665 connect and renegotiate with servers which do not support RI.
9666 Until RI is more widely deployed this option is enabled by default.
9670 * Add "missing" ssl ctrls to clear options and mode.
9674 * If client attempts to renegotiate and doesn't support RI respond with
9675 a no_renegotiation alert as required by RFC5746. Some renegotiating
9676 TLS clients will continue a connection gracefully when they receive
9677 the alert. Unfortunately OpenSSL mishandled this alert and would hang
9678 waiting for a server hello which it will never receive. Now we treat a
9679 received no_renegotiation alert as a fatal error. This is because
9680 applications requesting a renegotiation might well expect it to succeed
9681 and would have no code in place to handle the server denying it so the
9682 only safe thing to do is to terminate the connection.
9686 * Add ctrl macro SSL_get_secure_renegotiation_support() which returns 1 if
9687 peer supports secure renegotiation and 0 otherwise. Print out peer
9688 renegotiation support in s_client/s_server.
9692 * Replace the highly broken and deprecated SPKAC certification method with
9693 the updated NID creation version. This should correctly handle UTF8.
9697 * Implement RFC5746. Re-enable renegotiation but require the extension
9698 as needed. Unfortunately, SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION
9699 turns out to be a bad idea. It has been replaced by
9700 SSL_OP_ALLOW_UNSAFE_LEGACY_RENEGOTIATION which can be set with
9701 SSL_CTX_set_options(). This is really not recommended unless you
9702 know what you are doing.
9704 *Eric Rescorla <ekr@networkresonance.com>, Ben Laurie, Steve Henson*
9706 * Fixes to stateless session resumption handling. Use initial_ctx when
9707 issuing and attempting to decrypt tickets in case it has changed during
9708 servername handling. Use a non-zero length session ID when attempting
9709 stateless session resumption: this makes it possible to determine if
9710 a resumption has occurred immediately after receiving server hello
9711 (several places in OpenSSL subtly assume this) instead of later in
9716 * The functions ENGINE_ctrl(), OPENSSL_isservice(),
9717 CMS_get1_RecipientRequest() and RAND_bytes() can return <=0 on error
9718 fixes for a few places where the return code is not checked
9721 *Julia Lawall <julia@diku.dk>*
9723 * Add --strict-warnings option to Configure script to include devteam
9724 warnings in other configurations.
9728 * Add support for --libdir option and LIBDIR variable in makefiles. This
9729 makes it possible to install openssl libraries in locations which
9730 have names other than "lib", for example "/usr/lib64" which some
9733 *Steve Henson, based on patch from Jeremy Utley*
9735 * Don't allow the use of leading 0x80 in OIDs. This is a violation of
9736 X690 8.9.12 and can produce some misleading textual output of OIDs.
9738 *Steve Henson, reported by Dan Kaminsky*
9740 * Delete MD2 from algorithm tables. This follows the recommendation in
9741 several standards that it is not used in new applications due to
9742 several cryptographic weaknesses. For binary compatibility reasons
9743 the MD2 API is still compiled in by default.
9747 * Add compression id to {d2i,i2d}_SSL_SESSION so it is correctly saved
9752 * Rename uni2asc and asc2uni functions to OPENSSL_uni2asc and
9753 OPENSSL_asc2uni conditionally on Netware platforms to avoid a name
9756 *Guenter <lists@gknw.net>*
9758 * Fix the server certificate chain building code to use X509_verify_cert(),
9759 it used to have an ad-hoc builder which was unable to cope with anything
9760 other than a simple chain.
9762 *David Woodhouse <dwmw2@infradead.org>, Steve Henson*
9764 * Don't check self signed certificate signatures in X509_verify_cert()
9765 by default (a flag can override this): it just wastes time without
9766 adding any security. As a useful side effect self signed root CAs
9767 with non-FIPS digests are now usable in FIPS mode.
9771 * In dtls1_process_out_of_seq_message() the check if the current message
9772 is already buffered was missing. For every new message was memory
9773 allocated, allowing an attacker to perform an denial of service attack
9774 with sending out of seq handshake messages until there is no memory
9775 left. Additionally every future message was buffered, even if the
9776 sequence number made no sense and would be part of another handshake.
9777 So only messages with sequence numbers less than 10 in advance will be
9778 buffered. ([CVE-2009-1378])
9780 *Robin Seggelmann, discovered by Daniel Mentz*
9782 * Records are buffered if they arrive with a future epoch to be
9783 processed after finishing the corresponding handshake. There is
9784 currently no limitation to this buffer allowing an attacker to perform
9785 a DOS attack with sending records with future epochs until there is no
9786 memory left. This patch adds the pqueue_size() function to determine
9787 the size of a buffer and limits the record buffer to 100 entries.
9790 *Robin Seggelmann, discovered by Daniel Mentz*
9792 * Keep a copy of frag->msg_header.frag_len so it can be used after the
9793 parent structure is freed. ([CVE-2009-1379])
9797 * Handle non-blocking I/O properly in SSL_shutdown() call.
9799 *Darryl Miles <darryl-mailinglists@netbauds.net>*
9801 * Add `2.5.4.*` OIDs
9803 *Ilya O. <vrghost@gmail.com>*
9805 ### Changes between 0.9.8k and 0.9.8l [5 Nov 2009]
9807 * Disable renegotiation completely - this fixes a severe security
9808 problem ([CVE-2009-3555]) at the cost of breaking all
9809 renegotiation. Renegotiation can be re-enabled by setting
9810 SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION in s3->flags at
9811 run-time. This is really not recommended unless you know what
9816 ### Changes between 0.9.8j and 0.9.8k [25 Mar 2009]
9818 * Don't set val to NULL when freeing up structures, it is freed up by
9819 underlying code. If `sizeof(void *) > sizeof(long)` this can result in
9820 zeroing past the valid field. ([CVE-2009-0789])
9822 *Paolo Ganci <Paolo.Ganci@AdNovum.CH>*
9824 * Fix bug where return value of CMS_SignerInfo_verify_content() was not
9825 checked correctly. This would allow some invalid signed attributes to
9826 appear to verify correctly. ([CVE-2009-0591])
9828 *Ivan Nestlerode <inestlerode@us.ibm.com>*
9830 * Reject UniversalString and BMPString types with invalid lengths. This
9831 prevents a crash in ASN1_STRING_print_ex() which assumes the strings have
9832 a legal length. ([CVE-2009-0590])
9836 * Set S/MIME signing as the default purpose rather than setting it
9837 unconditionally. This allows applications to override it at the store
9842 * Permit restricted recursion of ASN1 strings. This is needed in practice
9843 to handle some structures.
9847 * Improve efficiency of mem_gets: don't search whole buffer each time
9850 *Jeremy Shapiro <jnshapir@us.ibm.com>*
9852 * New -hex option for openssl rand.
9856 * Print out UTF8String and NumericString when parsing ASN1.
9860 * Support NumericString type for name components.
9864 * Allow CC in the environment to override the automatically chosen
9865 compiler. Note that nothing is done to ensure flags work with the
9870 ### Changes between 0.9.8i and 0.9.8j [07 Jan 2009]
9872 * Properly check EVP_VerifyFinal() and similar return values
9875 *Ben Laurie, Bodo Moeller, Google Security Team*
9877 * Enable TLS extensions by default.
9881 * Allow the CHIL engine to be loaded, whether the application is
9882 multithreaded or not. (This does not release the developer from the
9883 obligation to set up the dynamic locking callbacks.)
9885 *Sander Temme <sander@temme.net>*
9887 * Use correct exit code if there is an error in dgst command.
9889 *Steve Henson; problem pointed out by Roland Dirlewanger*
9891 * Tweak Configure so that you need to say "experimental-jpake" to enable
9892 JPAKE, and need to use -DOPENSSL_EXPERIMENTAL_JPAKE in applications.
9896 * Add experimental JPAKE support, including demo authentication in
9897 s_client and s_server.
9901 * Set the comparison function in v3_addr_canonize().
9903 *Rob Austein <sra@hactrn.net>*
9905 * Add support for XMPP STARTTLS in s_client.
9907 *Philip Paeps <philip@freebsd.org>*
9909 * Change the server-side SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG behavior
9910 to ensure that even with this option, only ciphersuites in the
9911 server's preference list will be accepted. (Note that the option
9912 applies only when resuming a session, so the earlier behavior was
9913 just about the algorithm choice for symmetric cryptography.)
9917 ### Changes between 0.9.8h and 0.9.8i [15 Sep 2008]
9919 * Fix NULL pointer dereference if a DTLS server received
9920 ChangeCipherSpec as first record ([CVE-2009-1386]).
9924 * Fix a state transition in s3_srvr.c and d1_srvr.c
9925 (was using SSL3_ST_CW_CLNT_HELLO_B, should be `..._ST_SW_SRVR_...`).
9929 * The fix in 0.9.8c that supposedly got rid of unsafe
9930 double-checked locking was incomplete for RSA blinding,
9931 addressing just one layer of what turns out to have been
9932 doubly unsafe triple-checked locking.
9934 So now fix this for real by retiring the MONT_HELPER macro
9935 in crypto/rsa/rsa_eay.c.
9937 *Bodo Moeller; problem pointed out by Marius Schilder*
9939 * Various precautionary measures:
9941 - Avoid size_t integer overflow in HASH_UPDATE (md32_common.h).
9943 - Avoid a buffer overflow in d2i_SSL_SESSION() (ssl_asn1.c).
9944 (NB: This would require knowledge of the secret session ticket key
9945 to exploit, in which case you'd be SOL either way.)
9947 - Change bn_nist.c so that it will properly handle input BIGNUMs
9948 outside the expected range.
9950 - Enforce the 'num' check in BN_div() (bn_div.c) for non-BN_DEBUG
9953 *Neel Mehta, Bodo Moeller*
9955 * Allow engines to be "soft loaded" - i.e. optionally don't die if
9956 the load fails. Useful for distros.
9958 *Ben Laurie and the FreeBSD team*
9960 * Add support for Local Machine Keyset attribute in PKCS#12 files.
9964 * Fix BN_GF2m_mod_arr() top-bit cleanup code.
9968 * Expand ENGINE to support engine supplied SSL client certificate functions.
9970 This work was sponsored by Logica.
9974 * Add CryptoAPI ENGINE to support use of RSA and DSA keys held in Windows
9975 keystores. Support for SSL/TLS client authentication too.
9976 Not compiled unless enable-capieng specified to Configure.
9978 This work was sponsored by Logica.
9982 * Fix bug in X509_ATTRIBUTE creation: don't set attribute using
9983 ASN1_TYPE_set1 if MBSTRING flag set. This bug would crash certain
9984 attribute creation routines such as certificate requests and PKCS#12
9989 ### Changes between 0.9.8g and 0.9.8h [28 May 2008]
9991 * Fix flaw if 'Server Key exchange message' is omitted from a TLS
9992 handshake which could lead to a client crash as found using the
9993 Codenomicon TLS test suite ([CVE-2008-1672])
9995 *Steve Henson, Mark Cox*
9997 * Fix double free in TLS server name extensions which could lead to
9998 a remote crash found by Codenomicon TLS test suite ([CVE-2008-0891])
10002 * Clear error queue in SSL_CTX_use_certificate_chain_file()
10004 Clear the error queue to ensure that error entries left from
10005 older function calls do not interfere with the correct operation.
10007 *Lutz Jaenicke, Erik de Castro Lopo*
10009 * Remove root CA certificates of commercial CAs:
10011 The OpenSSL project does not recommend any specific CA and does not
10012 have any policy with respect to including or excluding any CA.
10013 Therefore it does not make any sense to ship an arbitrary selection
10014 of root CA certificates with the OpenSSL software.
10018 * RSA OAEP patches to fix two separate invalid memory reads.
10019 The first one involves inputs when 'lzero' is greater than
10020 'SHA_DIGEST_LENGTH' (it would read about SHA_DIGEST_LENGTH bytes
10021 before the beginning of from). The second one involves inputs where
10022 the 'db' section contains nothing but zeroes (there is a one-byte
10023 invalid read after the end of 'db').
10025 *Ivan Nestlerode <inestlerode@us.ibm.com>*
10027 * Partial backport from 0.9.9-dev:
10029 Introduce bn_mul_mont (dedicated Montgomery multiplication
10030 procedure) as a candidate for BIGNUM assembler implementation.
10031 While 0.9.9-dev uses assembler for various architectures, only
10032 x86_64 is available by default here in the 0.9.8 branch, and
10033 32-bit x86 is available through a compile-time setting.
10035 To try the 32-bit x86 assembler implementation, use Configure
10036 option "enable-montasm" (which exists only for this backport).
10038 As "enable-montasm" for 32-bit x86 disclaims code stability
10039 anyway, in this constellation we activate additional code
10040 backported from 0.9.9-dev for further performance improvements,
10041 namely BN_from_montgomery_word. (To enable this otherwise,
10042 e.g. x86_64, try `-DMONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD`.)
10044 *Andy Polyakov (backport partially by Bodo Moeller)*
10046 * Add TLS session ticket callback. This allows an application to set
10047 TLS ticket cipher and HMAC keys rather than relying on hardcoded fixed
10048 values. This is useful for key rollover for example where several key
10049 sets may exist with different names.
10053 * Reverse ENGINE-internal logic for caching default ENGINE handles.
10054 This was broken until now in 0.9.8 releases, such that the only way
10055 a registered ENGINE could be used (assuming it initialises
10056 successfully on the host) was to explicitly set it as the default
10057 for the relevant algorithms. This is in contradiction with 0.9.7
10058 behaviour and the documentation. With this fix, when an ENGINE is
10059 registered into a given algorithm's table of implementations, the
10060 'uptodate' flag is reset so that auto-discovery will be used next
10061 time a new context for that algorithm attempts to select an
10064 *Ian Lister (tweaked by Geoff Thorpe)*
10066 * Backport of CMS code to OpenSSL 0.9.8. This differs from the 0.9.9
10067 implementation in the following ways:
10069 Lack of EVP_PKEY_ASN1_METHOD means algorithm parameters have to be
10072 Lack of BER streaming support means one pass streaming processing is
10073 only supported if data is detached: setting the streaming flag is
10074 ignored for embedded content.
10076 CMS support is disabled by default and must be explicitly enabled
10077 with the enable-cms configuration option.
10081 * Update the GMP engine glue to do direct copies between BIGNUM and
10082 mpz_t when openssl and GMP use the same limb size. Otherwise the
10083 existing "conversion via a text string export" trick is still used.
10085 *Paul Sheer <paulsheer@gmail.com>*
10087 * Zlib compression BIO. This is a filter BIO which compressed and
10088 uncompresses any data passed through it.
10092 * Add AES_wrap_key() and AES_unwrap_key() functions to implement
10093 RFC3394 compatible AES key wrapping.
10097 * Add utility functions to handle ASN1 structures. ASN1_STRING_set0():
10098 sets string data without copying. X509_ALGOR_set0() and
10099 X509_ALGOR_get0(): set and retrieve X509_ALGOR (AlgorithmIdentifier)
10100 data. Attribute function X509at_get0_data_by_OBJ(): retrieves data
10101 from an X509_ATTRIBUTE structure optionally checking it occurs only
10102 once. ASN1_TYPE_set1(): set and ASN1_TYPE structure copying supplied
10107 * Fix BN flag handling in RSA_eay_mod_exp() and BN_MONT_CTX_set()
10108 to get the expected BN_FLG_CONSTTIME behavior.
10110 *Bodo Moeller (Google)*
10114 - fixed wrong usage of ioctlsocket() when build for LIBC BSD sockets
10115 - fixed do_tests.pl to run the test suite with CLIB builds too (CLIB_OPT)
10116 - added some more tests to do_tests.pl
10117 - fixed RunningProcess usage so that it works with newer LIBC NDKs too
10118 - removed usage of BN_LLONG for CLIB builds to avoid runtime dependency
10119 - added new Configure targets netware-clib-bsdsock, netware-clib-gcc,
10120 netware-clib-bsdsock-gcc, netware-libc-bsdsock-gcc
10121 - various changes to netware.pl to enable gcc-cross builds on Win32
10123 - changed crypto/bio/b_sock.c to work with macro functions (CLIB BSD)
10124 - various changes to fix missing prototype warnings
10125 - fixed x86nasm.pl to create correct asm files for NASM COFF output
10126 - added AES, WHIRLPOOL and CPUID assembler code to build files
10127 - added missing AES assembler make rules to mk1mf.pl
10128 - fixed order of includes in `apps/ocsp.c` so that `e_os.h` settings apply
10130 *Guenter Knauf <eflash@gmx.net>*
10132 * Implement certificate status request TLS extension defined in RFC3546.
10133 A client can set the appropriate parameters and receive the encoded
10134 OCSP response via a callback. A server can query the supplied parameters
10135 and set the encoded OCSP response in the callback. Add simplified examples
10136 to s_client and s_server.
10140 ### Changes between 0.9.8f and 0.9.8g [19 Oct 2007]
10142 * Fix various bugs:
10143 + Binary incompatibility of ssl_ctx_st structure
10144 + DTLS interoperation with non-compliant servers
10145 + Don't call get_session_cb() without proposed session
10146 + Fix ia64 assembler code
10148 *Andy Polyakov, Steve Henson*
10150 ### Changes between 0.9.8e and 0.9.8f [11 Oct 2007]
10152 * DTLS Handshake overhaul. There were longstanding issues with
10153 OpenSSL DTLS implementation, which were making it impossible for
10154 RFC 4347 compliant client to communicate with OpenSSL server.
10155 Unfortunately just fixing these incompatibilities would "cut off"
10156 pre-0.9.8f clients. To allow for hassle free upgrade post-0.9.8e
10157 server keeps tolerating non RFC compliant syntax. The opposite is
10158 not true, 0.9.8f client can not communicate with earlier server.
10159 This update even addresses CVE-2007-4995.
10163 * Changes to avoid need for function casts in OpenSSL: some compilers
10164 (gcc 4.2 and later) reject their use.
10165 *Kurt Roeckx <kurt@roeckx.be>, Peter Hartley <pdh@utter.chaos.org.uk>,
10168 * Add RFC4507 support to OpenSSL. This includes the corrections in
10169 RFC4507bis. The encrypted ticket format is an encrypted encoded
10170 SSL_SESSION structure, that way new session features are automatically
10173 If a client application caches session in an SSL_SESSION structure
10174 support is transparent because tickets are now stored in the encoded
10177 The SSL_CTX structure automatically generates keys for ticket
10178 protection in servers so again support should be possible
10179 with no application modification.
10181 If a client or server wishes to disable RFC4507 support then the option
10182 SSL_OP_NO_TICKET can be set.
10184 Add a TLS extension debugging callback to allow the contents of any client
10185 or server extensions to be examined.
10187 This work was sponsored by Google.
10191 * Add initial support for TLS extensions, specifically for the server_name
10192 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
10193 have new members for a host name. The SSL data structure has an
10194 additional member `SSL_CTX *initial_ctx` so that new sessions can be
10195 stored in that context to allow for session resumption, even after the
10196 SSL has been switched to a new SSL_CTX in reaction to a client's
10197 server_name extension.
10199 New functions (subject to change):
10201 SSL_get_servername()
10202 SSL_get_servername_type()
10205 New CTRL codes and macros (subject to change):
10207 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
10208 - SSL_CTX_set_tlsext_servername_callback()
10209 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
10210 - SSL_CTX_set_tlsext_servername_arg()
10211 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
10213 openssl s_client has a new '-servername ...' option.
10215 openssl s_server has new options '-servername_host ...', '-cert2 ...',
10216 '-key2 ...', '-servername_fatal' (subject to change). This allows
10217 testing the HostName extension for a specific single host name ('-cert'
10218 and '-key' remain fallbacks for handshakes without HostName
10219 negotiation). If the unrecognized_name alert has to be sent, this by
10220 default is a warning; it becomes fatal with the '-servername_fatal'
10223 *Peter Sylvester, Remy Allais, Christophe Renou, Steve Henson*
10225 * Add AES and SSE2 assembly language support to VC++ build.
10229 * Mitigate attack on final subtraction in Montgomery reduction.
10233 * Fix crypto/ec/ec_mult.c to work properly with scalars of value 0
10234 (which previously caused an internal error).
10238 * Squeeze another 10% out of IGE mode when in != out.
10242 * AES IGE mode speedup.
10244 *Dean Gaudet (Google)*
10246 * Add the Korean symmetric 128-bit cipher SEED (see
10247 <http://www.kisa.or.kr/kisa/seed/jsp/seed_eng.jsp>) and
10248 add SEED ciphersuites from RFC 4162:
10250 TLS_RSA_WITH_SEED_CBC_SHA = "SEED-SHA"
10251 TLS_DHE_DSS_WITH_SEED_CBC_SHA = "DHE-DSS-SEED-SHA"
10252 TLS_DHE_RSA_WITH_SEED_CBC_SHA = "DHE-RSA-SEED-SHA"
10253 TLS_DH_anon_WITH_SEED_CBC_SHA = "ADH-SEED-SHA"
10255 To minimize changes between patchlevels in the OpenSSL 0.9.8
10256 series, SEED remains excluded from compilation unless OpenSSL
10257 is configured with 'enable-seed'.
10259 *KISA, Bodo Moeller*
10261 * Mitigate branch prediction attacks, which can be practical if a
10262 single processor is shared, allowing a spy process to extract
10263 information. For detailed background information, see
10264 <http://eprint.iacr.org/2007/039> (O. Aciicmez, S. Gueron,
10265 J.-P. Seifert, "New Branch Prediction Vulnerabilities in OpenSSL
10266 and Necessary Software Countermeasures"). The core of the change
10267 are new versions BN_div_no_branch() and
10268 BN_mod_inverse_no_branch() of BN_div() and BN_mod_inverse(),
10269 respectively, which are slower, but avoid the security-relevant
10270 conditional branches. These are automatically called by BN_div()
10271 and BN_mod_inverse() if the flag BN_FLG_CONSTTIME is set for one
10272 of the input BIGNUMs. Also, BN_is_bit_set() has been changed to
10273 remove a conditional branch.
10275 BN_FLG_CONSTTIME is the new name for the previous
10276 BN_FLG_EXP_CONSTTIME flag, since it now affects more than just
10277 modular exponentiation. (Since OpenSSL 0.9.7h, setting this flag
10278 in the exponent causes BN_mod_exp_mont() to use the alternative
10279 implementation in BN_mod_exp_mont_consttime().) The old name
10280 remains as a deprecated alias.
10282 Similarly, RSA_FLAG_NO_EXP_CONSTTIME is replaced by a more general
10283 RSA_FLAG_NO_CONSTTIME flag since the RSA implementation now uses
10284 constant-time implementations for more than just exponentiation.
10285 Here too the old name is kept as a deprecated alias.
10287 BN_BLINDING_new() will now use BN_dup() for the modulus so that
10288 the BN_BLINDING structure gets an independent copy of the
10289 modulus. This means that the previous `BIGNUM *m` argument to
10290 BN_BLINDING_new() and to BN_BLINDING_create_param() now
10291 essentially becomes `const BIGNUM *m`, although we can't actually
10292 change this in the header file before 0.9.9. It allows
10293 RSA_setup_blinding() to use BN_with_flags() on the modulus to
10294 enable BN_FLG_CONSTTIME.
10296 *Matthew D Wood (Intel Corp)*
10298 * In the SSL/TLS server implementation, be strict about session ID
10299 context matching (which matters if an application uses a single
10300 external cache for different purposes). Previously,
10301 out-of-context reuse was forbidden only if SSL_VERIFY_PEER was
10302 set. This did ensure strict client verification, but meant that,
10303 with applications using a single external cache for quite
10304 different requirements, clients could circumvent ciphersuite
10305 restrictions for a given session ID context by starting a session
10306 in a different context.
10310 * Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
10311 a ciphersuite string such as "DEFAULT:RSA" cannot enable
10312 authentication-only ciphersuites.
10316 * Update the SSL_get_shared_ciphers() fix CVE-2006-3738 which was
10317 not complete and could lead to a possible single byte overflow
10318 ([CVE-2007-5135]) [Ben Laurie]
10320 ### Changes between 0.9.8d and 0.9.8e [23 Feb 2007]
10322 * Since AES128 and AES256 (and similarly Camellia128 and
10323 Camellia256) share a single mask bit in the logic of
10324 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
10325 kludge to work properly if AES128 is available and AES256 isn't
10326 (or if Camellia128 is available and Camellia256 isn't).
10330 * Fix the BIT STRING encoding generated by crypto/ec/ec_asn1.c
10331 (within i2d_ECPrivateKey, i2d_ECPKParameters, i2d_ECParameters):
10332 When a point or a seed is encoded in a BIT STRING, we need to
10333 prevent the removal of trailing zero bits to get the proper DER
10334 encoding. (By default, crypto/asn1/a_bitstr.c assumes the case
10335 of a NamedBitList, for which trailing 0 bits need to be removed.)
10339 * Have SSL/TLS server implementation tolerate "mismatched" record
10340 protocol version while receiving ClientHello even if the
10341 ClientHello is fragmented. (The server can't insist on the
10342 particular protocol version it has chosen before the ServerHello
10343 message has informed the client about his choice.)
10347 * Add RFC 3779 support.
10349 *Rob Austein for ARIN, Ben Laurie*
10351 * Load error codes if they are not already present instead of using a
10352 static variable. This allows them to be cleanly unloaded and reloaded.
10353 Improve header file function name parsing.
10357 * extend SMTP and IMAP protocol emulation in s_client to use EHLO
10358 or CAPABILITY handshake as required by RFCs.
10360 *Goetz Babin-Ebell*
10362 ### Changes between 0.9.8c and 0.9.8d [28 Sep 2006]
10364 * Introduce limits to prevent malicious keys being able to
10365 cause a denial of service. ([CVE-2006-2940])
10367 *Steve Henson, Bodo Moeller*
10369 * Fix ASN.1 parsing of certain invalid structures that can result
10370 in a denial of service. ([CVE-2006-2937]) [Steve Henson]
10372 * Fix buffer overflow in SSL_get_shared_ciphers() function.
10373 ([CVE-2006-3738]) [Tavis Ormandy and Will Drewry, Google Security Team]
10375 * Fix SSL client code which could crash if connecting to a
10376 malicious SSLv2 server. ([CVE-2006-4343])
10378 *Tavis Ormandy and Will Drewry, Google Security Team*
10380 * Since 0.9.8b, ciphersuite strings naming explicit ciphersuites
10381 match only those. Before that, "AES256-SHA" would be interpreted
10382 as a pattern and match "AES128-SHA" too (since AES128-SHA got
10383 the same strength classification in 0.9.7h) as we currently only
10384 have a single AES bit in the ciphersuite description bitmap.
10385 That change, however, also applied to ciphersuite strings such as
10386 "RC4-MD5" that intentionally matched multiple ciphersuites --
10387 namely, SSL 2.0 ciphersuites in addition to the more common ones
10388 from SSL 3.0/TLS 1.0.
10390 So we change the selection algorithm again: Naming an explicit
10391 ciphersuite selects this one ciphersuite, and any other similar
10392 ciphersuite (same bitmap) from *other* protocol versions.
10393 Thus, "RC4-MD5" again will properly select both the SSL 2.0
10394 ciphersuite and the SSL 3.0/TLS 1.0 ciphersuite.
10396 Since SSL 2.0 does not have any ciphersuites for which the
10397 128/256 bit distinction would be relevant, this works for now.
10398 The proper fix will be to use different bits for AES128 and
10399 AES256, which would have avoided the problems from the beginning;
10400 however, bits are scarce, so we can only do this in a new release
10401 (not just a patchlevel) when we can change the SSL_CIPHER
10402 definition to split the single 'unsigned long mask' bitmap into
10403 multiple values to extend the available space.
10407 ### Changes between 0.9.8b and 0.9.8c [05 Sep 2006]
10409 * Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
10410 ([CVE-2006-4339]) [Ben Laurie and Google Security Team]
10412 * Add AES IGE and biIGE modes.
10416 * Change the Unix randomness entropy gathering to use poll() when
10417 possible instead of select(), since the latter has some
10418 undesirable limitations.
10420 *Darryl Miles via Richard Levitte and Bodo Moeller*
10422 * Disable "ECCdraft" ciphersuites more thoroughly. Now special
10423 treatment in ssl/ssl_ciph.s makes sure that these ciphersuites
10424 cannot be implicitly activated as part of, e.g., the "AES" alias.
10425 However, please upgrade to OpenSSL 0.9.9[-dev] for
10426 non-experimental use of the ECC ciphersuites to get TLS extension
10427 support, which is required for curve and point format negotiation
10428 to avoid potential handshake problems.
10432 * Disable rogue ciphersuites:
10434 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
10435 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
10436 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
10438 The latter two were purportedly from
10439 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
10442 Also deactivate the remaining ciphersuites from
10443 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
10444 unofficial, and the ID has long expired.
10448 * Fix RSA blinding Heisenbug (problems sometimes occurred on
10449 dual-core machines) and other potential thread-safety issues.
10453 * Add the symmetric cipher Camellia (128-bit, 192-bit, 256-bit key
10454 versions), which is now available for royalty-free use
10455 (see <http://info.isl.ntt.co.jp/crypt/eng/info/chiteki.html>).
10456 Also, add Camellia TLS ciphersuites from RFC 4132.
10458 To minimize changes between patchlevels in the OpenSSL 0.9.8
10459 series, Camellia remains excluded from compilation unless OpenSSL
10460 is configured with 'enable-camellia'.
10464 * Disable the padding bug check when compression is in use. The padding
10465 bug check assumes the first packet is of even length, this is not
10466 necessarily true if compression is enabled and can result in false
10467 positives causing handshake failure. The actual bug test is ancient
10468 code so it is hoped that implementations will either have fixed it by
10469 now or any which still have the bug do not support compression.
10473 ### Changes between 0.9.8a and 0.9.8b [04 May 2006]
10475 * When applying a cipher rule check to see if string match is an explicit
10476 cipher suite and only match that one cipher suite if it is.
10480 * Link in manifests for VC++ if needed.
10482 *Austin Ziegler <halostatue@gmail.com>*
10484 * Update support for ECC-based TLS ciphersuites according to
10485 draft-ietf-tls-ecc-12.txt with proposed changes (but without
10486 TLS extensions, which are supported starting with the 0.9.9
10487 branch, not in the OpenSSL 0.9.8 branch).
10491 * New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support
10492 opaque EVP_CIPHER_CTX handling.
10496 * Fixes and enhancements to zlib compression code. We now only use
10497 "zlib1.dll" and use the default `__cdecl` calling convention on Win32
10498 to conform with the standards mentioned here:
10499 <http://www.zlib.net/DLL_FAQ.txt>
10500 Static zlib linking now works on Windows and the new --with-zlib-include
10501 --with-zlib-lib options to Configure can be used to supply the location
10502 of the headers and library. Gracefully handle case where zlib library
10507 * Several fixes and enhancements to the OID generation code. The old code
10508 sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't
10509 handle numbers larger than ULONG_MAX, truncated printing and had a
10510 non standard OBJ_obj2txt() behaviour.
10514 * Add support for building of engines under engine/ as shared libraries
10515 under VC++ build system.
10519 * Corrected the numerous bugs in the Win32 path splitter in DSO.
10520 Hopefully, we will not see any false combination of paths any more.
10524 ### Changes between 0.9.8 and 0.9.8a [11 Oct 2005]
10526 * Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
10527 (part of SSL_OP_ALL). This option used to disable the
10528 countermeasure against man-in-the-middle protocol-version
10529 rollback in the SSL 2.0 server implementation, which is a bad
10530 idea. ([CVE-2005-2969])
10532 *Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
10533 for Information Security, National Institute of Advanced Industrial
10534 Science and Technology [AIST], Japan)*
10536 * Add two function to clear and return the verify parameter flags.
10540 * Keep cipherlists sorted in the source instead of sorting them at
10541 runtime, thus removing the need for a lock.
10545 * Avoid some small subgroup attacks in Diffie-Hellman.
10547 *Nick Mathewson and Ben Laurie*
10549 * Add functions for well-known primes.
10553 * Extended Windows CE support.
10555 *Satoshi Nakamura and Andy Polyakov*
10557 * Initialize SSL_METHOD structures at compile time instead of during
10558 runtime, thus removing the need for a lock.
10562 * Make PKCS7_decrypt() work even if no certificate is supplied by
10563 attempting to decrypt each encrypted key in turn. Add support to
10568 ### Changes between 0.9.7h and 0.9.8 [05 Jul 2005]
10570 [NB: OpenSSL 0.9.7i and later 0.9.7 patch levels were released after
10573 * Add libcrypto.pc and libssl.pc for those who feel they need them.
10577 * Change CA.sh and CA.pl so they don't bundle the CSR and the private
10578 key into the same file any more.
10582 * Add initial support for Win64, both IA64 and AMD64/x64 flavors.
10586 * Add -utf8 command line and config file option to 'ca'.
10588 *Stefan <stf@udoma.org*
10590 * Removed the macro des_crypt(), as it seems to conflict with some
10591 libraries. Use DES_crypt().
10595 * Correct naming of the 'chil' and '4758cca' ENGINEs. This
10596 involves renaming the source and generated shared-libs for
10597 both. The engines will accept the corrected or legacy ids
10598 ('ncipher' and '4758_cca' respectively) when binding. NB,
10599 this only applies when building 'shared'.
10601 *Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe*
10603 * Add attribute functions to EVP_PKEY structure. Modify
10604 PKCS12_create() to recognize a CSP name attribute and
10605 use it. Make -CSP option work again in pkcs12 utility.
10609 * Add new functionality to the bn blinding code:
10610 - automatic re-creation of the BN_BLINDING parameters after
10611 a fixed number of uses (currently 32)
10612 - add new function for parameter creation
10613 - introduce flags to control the update behaviour of the
10614 BN_BLINDING parameters
10615 - hide BN_BLINDING structure
10616 Add a second BN_BLINDING slot to the RSA structure to improve
10617 performance when a single RSA object is shared among several
10622 * Add support for DTLS.
10624 *Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie*
10626 * Add support for DER encoded private keys (SSL_FILETYPE_ASN1)
10627 to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file()
10631 * Remove buggy and incomplete DH cert support from
10632 ssl/ssl_rsa.c and ssl/s3_both.c
10636 * Use SHA-1 instead of MD5 as the default digest algorithm for
10637 the `apps/openssl` commands.
10641 * Compile clean with "-Wall -Wmissing-prototypes
10642 -Wstrict-prototypes -Wmissing-declarations -Werror". Currently
10643 DEBUG_SAFESTACK must also be set.
10647 * Change ./Configure so that certain algorithms can be disabled by default.
10648 The new counterpiece to "no-xxx" is "enable-xxx".
10650 The patented RC5 and MDC2 algorithms will now be disabled unless
10651 "enable-rc5" and "enable-mdc2", respectively, are specified.
10653 (IDEA remains enabled despite being patented. This is because IDEA
10654 is frequently required for interoperability, and there is no license
10655 fee for non-commercial use. As before, "no-idea" can be used to
10656 avoid this algorithm.)
10660 * Add processing of proxy certificates (see RFC 3820). This work was
10661 sponsored by KTH (The Royal Institute of Technology in Stockholm) and
10662 EGEE (Enabling Grids for E-science in Europe).
10666 * RC4 performance overhaul on modern architectures/implementations, such
10667 as Intel P4, IA-64 and AMD64.
10671 * New utility extract-section.pl. This can be used specify an alternative
10672 section number in a pod file instead of having to treat each file as
10673 a separate case in Makefile. This can be done by adding two lines to the
10676 =for comment openssl_section:XXX
10678 The blank line is mandatory.
10682 * New arguments -certform, -keyform and -pass for s_client and s_server
10683 to allow alternative format key and certificate files and passphrase
10688 * New structure X509_VERIFY_PARAM which combines current verify parameters,
10689 update associated structures and add various utility functions.
10691 Add new policy related verify parameters, include policy checking in
10692 standard verify code. Enhance 'smime' application with extra parameters
10693 to support policy checking and print out.
10697 * Add a new engine to support VIA PadLock ACE extensions in the VIA C3
10698 Nehemiah processors. These extensions support AES encryption in hardware
10699 as well as RNG (though RNG support is currently disabled).
10701 *Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov*
10703 * Deprecate `BN_[get|set]_params()` functions (they were ignored internally).
10707 * New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented.
10709 *Andy Polyakov and a number of other people*
10711 * Improved PowerPC platform support. Most notably BIGNUM assembler
10712 implementation contributed by IBM.
10714 *Suresh Chari, Peter Waltenberg, Andy Polyakov*
10716 * The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public
10717 exponent rather than 'unsigned long'. There is a corresponding change to
10718 the new 'rsa_keygen' element of the RSA_METHOD structure.
10720 *Jelte Jansen, Geoff Thorpe*
10722 * Functionality for creating the initial serial number file is now
10723 moved from CA.pl to the 'ca' utility with a new option -create_serial.
10725 (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial
10726 number file to 1, which is bound to cause problems. To avoid
10727 the problems while respecting compatibility between different 0.9.7
10728 patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in
10729 CA.pl for serial number initialization. With the new release 0.9.8,
10730 we can fix the problem directly in the 'ca' utility.)
10734 * Reduced header interdependencies by declaring more opaque objects in
10735 ossl_typ.h. As a consequence, including some headers (eg. engine.h) will
10736 give fewer recursive includes, which could break lazy source code - so
10737 this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always,
10738 developers should define this symbol when building and using openssl to
10739 ensure they track the recommended behaviour, interfaces, [etc], but
10740 backwards-compatible behaviour prevails when this isn't defined.
10744 * New function X509_POLICY_NODE_print() which prints out policy nodes.
10748 * Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality.
10749 This will generate a random key of the appropriate length based on the
10750 cipher context. The EVP_CIPHER can provide its own random key generation
10751 routine to support keys of a specific form. This is used in the des and
10752 3des routines to generate a key of the correct parity. Update S/MIME
10753 code to use new functions and hence generate correct parity DES keys.
10754 Add EVP_CHECK_DES_KEY #define to return an error if the key is not
10755 valid (weak or incorrect parity).
10759 * Add a local set of CRLs that can be used by X509_verify_cert() as well
10760 as looking them up. This is useful when the verified structure may contain
10761 CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs
10762 present unless the new PKCS7_NO_CRL flag is asserted.
10766 * Extend ASN1 oid configuration module. It now additionally accepts the
10769 shortName = some long name, 1.2.3.4
10773 * Reimplemented the BN_CTX implementation. There is now no more static
10774 limitation on the number of variables it can handle nor the depth of the
10775 "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack
10776 information can now expand as required, and rather than having a single
10777 static array of bignums, BN_CTX now uses a linked-list of such arrays
10778 allowing it to expand on demand whilst maintaining the usefulness of
10779 BN_CTX's "bundling".
10783 * Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD
10784 to allow all RSA operations to function using a single BN_CTX.
10788 * Preliminary support for certificate policy evaluation and checking. This
10789 is initially intended to pass the tests outlined in "Conformance Testing
10790 of Relying Party Client Certificate Path Processing Logic" v1.07.
10794 * bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and
10795 remained unused and not that useful. A variety of other little bignum
10796 tweaks and fixes have also been made continuing on from the audit (see
10801 * Constify all or almost all d2i, c2i, s2i and r2i functions, along with
10802 associated ASN1, EVP and SSL functions and old ASN1 macros.
10806 * BN_zero() only needs to set 'top' and 'neg' to zero for correct results,
10807 and this should never fail. So the return value from the use of
10808 BN_set_word() (which can fail due to needless expansion) is now deprecated;
10809 if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro.
10813 * BN_CTX_get() should return zero-valued bignums, providing the same
10814 initialised value as BN_new().
10816 *Geoff Thorpe, suggested by Ulf Möller*
10818 * Support for inhibitAnyPolicy certificate extension.
10822 * An audit of the BIGNUM code is underway, for which debugging code is
10823 enabled when BN_DEBUG is defined. This makes stricter enforcements on what
10824 is considered valid when processing BIGNUMs, and causes execution to
10825 assert() when a problem is discovered. If BN_DEBUG_RAND is defined,
10826 further steps are taken to deliberately pollute unused data in BIGNUM
10827 structures to try and expose faulty code further on. For now, openssl will
10828 (in its default mode of operation) continue to tolerate the inconsistent
10829 forms that it has tolerated in the past, but authors and packagers should
10830 consider trying openssl and their own applications when compiled with
10831 these debugging symbols defined. It will help highlight potential bugs in
10832 their own code, and will improve the test coverage for OpenSSL itself. At
10833 some point, these tighter rules will become openssl's default to improve
10834 maintainability, though the assert()s and other overheads will remain only
10835 in debugging configurations. See bn.h for more details.
10837 *Geoff Thorpe, Nils Larsch, Ulf Möller*
10839 * BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure
10840 that can only be obtained through BN_CTX_new() (which implicitly
10841 initialises it). The presence of this function only made it possible
10842 to overwrite an existing structure (and cause memory leaks).
10846 * Because of the callback-based approach for implementing LHASH as a
10847 template type, lh_insert() adds opaque objects to hash-tables and
10848 lh_doall() or lh_doall_arg() are typically used with a destructor callback
10849 to clean up those corresponding objects before destroying the hash table
10850 (and losing the object pointers). So some over-zealous constifications in
10851 LHASH have been relaxed so that lh_insert() does not take (nor store) the
10852 objects as "const" and the `lh_doall[_arg]` callback wrappers are not
10853 prototyped to have "const" restrictions on the object pointers they are
10854 given (and so aren't required to cast them away any more).
10858 * The tmdiff.h API was so ugly and minimal that our own timing utility
10859 (speed) prefers to use its own implementation. The two implementations
10860 haven't been consolidated as yet (volunteers?) but the tmdiff API has had
10861 its object type properly exposed (MS_TM) instead of casting to/from
10862 `char *`. This may still change yet if someone realises MS_TM and
10864 aren't necessarily the greatest nomenclatures - but this is what was used
10865 internally to the implementation so I've used that for now.
10869 * Ensure that deprecated functions do not get compiled when
10870 OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of
10871 the self-tests were still using deprecated key-generation functions so
10872 these have been updated also.
10876 * Reorganise PKCS#7 code to separate the digest location functionality
10877 into PKCS7_find_digest(), digest addition into PKCS7_bio_add_digest().
10878 New function PKCS7_set_digest() to set the digest type for PKCS#7
10879 digestedData type. Add additional code to correctly generate the
10880 digestedData type and add support for this type in PKCS7 initialization
10885 * New function PKCS7_set0_type_other() this initializes a PKCS7
10886 structure of type "other".
10890 * Fix prime generation loop in crypto/bn/bn_prime.pl by making
10891 sure the loop does correctly stop and breaking ("division by zero")
10892 modulus operations are not performed. The (pre-generated) prime
10893 table crypto/bn/bn_prime.h was already correct, but it could not be
10894 re-generated on some platforms because of the "division by zero"
10895 situation in the script.
10897 *Ralf S. Engelschall*
10899 * Update support for ECC-based TLS ciphersuites according to
10900 draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with
10901 SHA-1 now is only used for "small" curves (where the
10902 representation of a field element takes up to 24 bytes); for
10903 larger curves, the field element resulting from ECDH is directly
10904 used as premaster secret.
10906 *Douglas Stebila (Sun Microsystems Laboratories)*
10908 * Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2
10909 curve secp160r1 to the tests.
10911 *Douglas Stebila (Sun Microsystems Laboratories)*
10913 * Add the possibility to load symbols globally with DSO.
10915 *Götz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte*
10917 * Add the functions ERR_set_mark() and ERR_pop_to_mark() for better
10918 control of the error stack.
10922 * Add support for STORE in ENGINE.
10926 * Add the STORE type. The intention is to provide a common interface
10927 to certificate and key stores, be they simple file-based stores, or
10928 HSM-type store, or LDAP stores, or...
10929 NOTE: The code is currently UNTESTED and isn't really used anywhere.
10933 * Add a generic structure called OPENSSL_ITEM. This can be used to
10934 pass a list of arguments to any function as well as provide a way
10935 for a function to pass data back to the caller.
10939 * Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup()
10940 works like BUF_strdup() but can be used to duplicate a portion of
10941 a string. The copy gets NUL-terminated. BUF_memdup() duplicates
10946 * Add the function sk_find_ex() which works like sk_find(), but will
10947 return an index to an element even if an exact match couldn't be
10948 found. The index is guaranteed to point at the element where the
10949 searched-for key would be inserted to preserve sorting order.
10953 * Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but
10954 takes an extra flags argument for optional functionality. Currently,
10955 the following flags are defined:
10957 OBJ_BSEARCH_VALUE_ON_NOMATCH
10958 This one gets OBJ_bsearch_ex() to return a pointer to the first
10959 element where the comparing function returns a negative or zero
10962 OBJ_BSEARCH_FIRST_VALUE_ON_MATCH
10963 This one gets OBJ_bsearch_ex() to return a pointer to the first
10964 element where the comparing function returns zero. This is useful
10965 if there are more than one element where the comparing function
10970 * Make it possible to create self-signed certificates with 'openssl ca'
10971 in such a way that the self-signed certificate becomes part of the
10972 CA database and uses the same mechanisms for serial number generation
10973 as all other certificate signing. The new flag '-selfsign' enables
10974 this functionality. Adapt CA.sh and CA.pl.in.
10978 * Add functionality to check the public key of a certificate request
10979 against a given private. This is useful to check that a certificate
10980 request can be signed by that key (self-signing).
10984 * Make it possible to have multiple active certificates with the same
10985 subject in the CA index file. This is done only if the keyword
10986 'unique_subject' is set to 'no' in the main CA section (default
10987 if 'CA_default') of the configuration file. The value is saved
10988 with the database itself in a separate index attribute file,
10989 named like the index file with '.attr' appended to the name.
10993 * Generate multi-valued AVAs using '+' notation in config files for
10998 * Support for nameConstraints certificate extension.
11002 * Support for policyConstraints certificate extension.
11006 * Support for policyMappings certificate extension.
11010 * Make sure the default DSA_METHOD implementation only uses its
11011 dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL,
11012 and change its own handlers to be NULL so as to remove unnecessary
11013 indirection. This lets alternative implementations fallback to the
11014 default implementation more easily.
11018 * Support for directoryName in GeneralName related extensions
11023 * Make it possible to link applications using Makefile.shared.
11024 Make that possible even when linking against static libraries!
11028 * Support for single pass processing for S/MIME signing. This now
11029 means that S/MIME signing can be done from a pipe, in addition
11030 cleartext signing (multipart/signed type) is effectively streaming
11031 and the signed data does not need to be all held in memory.
11033 This is done with a new flag PKCS7_STREAM. When this flag is set
11034 PKCS7_sign() only initializes the PKCS7 structure and the actual signing
11035 is done after the data is output (and digests calculated) in
11036 SMIME_write_PKCS7().
11040 * Add full support for -rpath/-R, both in shared libraries and
11041 applications, at least on the platforms where it's known how
11046 * In crypto/ec/ec_mult.c, implement fast point multiplication with
11047 precomputation, based on wNAF splitting: EC_GROUP_precompute_mult()
11048 will now compute a table of multiples of the generator that
11049 makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul()
11050 faster (notably in the case of a single point multiplication,
11051 scalar * generator).
11053 *Nils Larsch, Bodo Moeller*
11055 * IPv6 support for certificate extensions. The various extensions
11056 which use the IP:a.b.c.d can now take IPv6 addresses using the
11057 formats of RFC1884 2.2 . IPv6 addresses are now also displayed
11062 * Added an ENGINE that implements RSA by performing private key
11063 exponentiations with the GMP library. The conversions to and from
11064 GMP's mpz_t format aren't optimised nor are any montgomery forms
11065 cached, and on x86 it appears OpenSSL's own performance has caught up.
11066 However there are likely to be other architectures where GMP could
11067 provide a boost. This ENGINE is not built in by default, but it can be
11068 specified at Configure time and should be accompanied by the necessary
11069 linker additions, eg;
11070 ./config -DOPENSSL_USE_GMP -lgmp
11074 * "openssl engine" will not display ENGINE/DSO load failure errors when
11075 testing availability of engines with "-t" - the old behaviour is
11076 produced by increasing the feature's verbosity with "-tt".
11080 * ECDSA routines: under certain error conditions uninitialized BN objects
11081 could be freed. Solution: make sure initialization is performed early
11082 enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de>
11087 * Key-generation can now be implemented in RSA_METHOD, DSA_METHOD
11088 and DH_METHOD (eg. by ENGINE implementations) to override the normal
11089 software implementations. For DSA and DH, parameter generation can
11090 also be overridden by providing the appropriate method callbacks.
11094 * Change the "progress" mechanism used in key-generation and
11095 primality testing to functions that take a new BN_GENCB pointer in
11096 place of callback/argument pairs. The new API functions have `_ex`
11097 postfixes and the older functions are reimplemented as wrappers for
11098 the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide
11099 declarations of the old functions to help (graceful) attempts to
11100 migrate to the new functions. Also, the new key-generation API
11101 functions operate on a caller-supplied key-structure and return
11102 success/failure rather than returning a key or NULL - this is to
11103 help make "keygen" another member function of RSA_METHOD etc.
11105 Example for using the new callback interface:
11107 int (*my_callback)(int a, int b, BN_GENCB *cb) = ...;
11108 void *my_arg = ...;
11111 BN_GENCB_set(&my_cb, my_callback, my_arg);
11113 return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb);
11114 /* For the meaning of a, b in calls to my_callback(), see the
11115 * documentation of the function that calls the callback.
11116 * cb will point to my_cb; my_arg can be retrieved as cb->arg.
11117 * my_callback should return 1 if it wants BN_is_prime_ex()
11118 * to continue, or 0 to stop.
11123 * Change the ZLIB compression method to be stateful, and make it
11124 available to TLS with the number defined in
11125 draft-ietf-tls-compression-04.txt.
11129 * Add the ASN.1 structures and functions for CertificatePair, which
11130 is defined as follows (according to X.509_4thEditionDraftV6.pdf):
11132 CertificatePair ::= SEQUENCE {
11133 forward [0] Certificate OPTIONAL,
11134 reverse [1] Certificate OPTIONAL,
11135 -- at least one of the pair shall be present -- }
11137 Also implement the PEM functions to read and write certificate
11138 pairs, and defined the PEM tag as "CERTIFICATE PAIR".
11140 This needed to be defined, mostly for the sake of the LDAP
11141 attribute crossCertificatePair, but may prove useful elsewhere as
11146 * Make it possible to inhibit symlinking of shared libraries in
11147 Makefile.shared, for Cygwin's sake.
11151 * Extend the BIGNUM API by creating a function
11152 void BN_set_negative(BIGNUM *a, int neg);
11153 and a macro that behave like
11154 int BN_is_negative(const BIGNUM *a);
11156 to avoid the need to access 'a->neg' directly in applications.
11160 * Implement fast modular reduction for pseudo-Mersenne primes
11161 used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c).
11162 EC_GROUP_new_curve_GFp() will now automatically use this
11165 *Nils Larsch <nla@trustcenter.de>*
11167 * Add new lock type (CRYPTO_LOCK_BN).
11171 * Change the ENGINE framework to automatically load engines
11172 dynamically from specific directories unless they could be
11173 found to already be built in or loaded. Move all the
11174 current engines except for the cryptodev one to a new
11175 directory engines/.
11176 The engines in engines/ are built as shared libraries if
11177 the "shared" options was given to ./Configure or ./config.
11178 Otherwise, they are inserted in libcrypto.a.
11179 /usr/local/ssl/engines is the default directory for dynamic
11180 engines, but that can be overridden at configure time through
11181 the usual use of --prefix and/or --openssldir, and at run
11182 time with the environment variable OPENSSL_ENGINES.
11184 *Geoff Thorpe and Richard Levitte*
11186 * Add Makefile.shared, a helper makefile to build shared
11187 libraries. Adapt Makefile.org.
11191 * Add version info to Win32 DLLs.
11193 *Peter 'Luna' Runestig" <peter@runestig.com>*
11195 * Add new 'medium level' PKCS#12 API. Certificates and keys
11196 can be added using this API to created arbitrary PKCS#12
11197 files while avoiding the low-level API.
11199 New options to PKCS12_create(), key or cert can be NULL and
11200 will then be omitted from the output file. The encryption
11201 algorithm NIDs can be set to -1 for no encryption, the mac
11202 iteration count can be set to 0 to omit the mac.
11204 Enhance pkcs12 utility by making the -nokeys and -nocerts
11205 options work when creating a PKCS#12 file. New option -nomac
11206 to omit the mac, NONE can be set for an encryption algorithm.
11207 New code is modified to use the enhanced PKCS12_create()
11208 instead of the low-level API.
11212 * Extend ASN1 encoder to support indefinite length constructed
11213 encoding. This can output sequences tags and octet strings in
11214 this form. Modify pk7_asn1.c to support indefinite length
11215 encoding. This is experimental and needs additional code to
11216 be useful, such as an ASN1 bio and some enhanced streaming
11219 Extend template encode functionality so that tagging is passed
11220 down to the template encoder.
11224 * Let 'openssl req' fail if an argument to '-newkey' is not
11225 recognized instead of using RSA as a default.
11229 * Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt.
11230 As these are not official, they are not included in "ALL";
11231 the "ECCdraft" ciphersuite group alias can be used to select them.
11233 *Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)*
11235 * Add ECDH engine support.
11237 *Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)*
11239 * Add ECDH in new directory crypto/ecdh/.
11241 *Douglas Stebila (Sun Microsystems Laboratories)*
11243 * Let BN_rand_range() abort with an error after 100 iterations
11244 without success (which indicates a broken PRNG).
11248 * Change BN_mod_sqrt() so that it verifies that the input value
11249 is really the square of the return value. (Previously,
11250 BN_mod_sqrt would show GIGO behaviour.)
11254 * Add named elliptic curves over binary fields from X9.62, SECG,
11255 and WAP/WTLS; add OIDs that were still missing.
11257 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
11259 * Extend the EC library for elliptic curves over binary fields
11260 (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/).
11263 EC_GF2m_simple_method
11267 EC_GROUP_new_curve_GF2m
11268 EC_GROUP_set_curve_GF2m
11269 EC_GROUP_get_curve_GF2m
11270 EC_POINT_set_affine_coordinates_GF2m
11271 EC_POINT_get_affine_coordinates_GF2m
11272 EC_POINT_set_compressed_coordinates_GF2m
11274 Point compression for binary fields is disabled by default for
11275 patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to
11278 As binary polynomials are represented as BIGNUMs, various members
11279 of the EC_GROUP and EC_POINT data structures can be shared
11280 between the implementations for prime fields and binary fields;
11281 the above `..._GF2m functions` (except for EX_GROUP_new_curve_GF2m)
11282 are essentially identical to their `..._GFp` counterparts.
11283 (For simplicity, the `..._GFp` prefix has been dropped from
11284 various internal method names.)
11286 An internal 'field_div' method (similar to 'field_mul' and
11287 'field_sqr') has been added; this is used only for binary fields.
11289 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
11291 * Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult()
11292 through methods ('mul', 'precompute_mult').
11294 The generic implementations (now internally called 'ec_wNAF_mul'
11295 and 'ec_wNAF_precomputed_mult') remain the default if these
11296 methods are undefined.
11298 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
11300 * New function EC_GROUP_get_degree, which is defined through
11301 EC_METHOD. For curves over prime fields, this returns the bit
11302 length of the modulus.
11304 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
11306 * New functions EC_GROUP_dup, EC_POINT_dup.
11307 (These simply call ..._new and ..._copy).
11309 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
11311 * Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c.
11312 Polynomials are represented as BIGNUMs (where the sign bit is not
11313 used) in the following functions [macros]:
11316 BN_GF2m_sub [= BN_GF2m_add]
11317 BN_GF2m_mod [wrapper for BN_GF2m_mod_arr]
11318 BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr]
11319 BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr]
11321 BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr]
11322 BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr]
11323 BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr]
11324 BN_GF2m_cmp [= BN_ucmp]
11326 (Note that only the 'mod' functions are actually for fields GF(2^m).
11327 BN_GF2m_add() is misnomer, but this is for the sake of consistency.)
11329 For some functions, an the irreducible polynomial defining a
11330 field can be given as an 'unsigned int[]' with strictly
11331 decreasing elements giving the indices of those bits that are set;
11332 i.e., p[] represents the polynomial
11333 f(t) = t^p[0] + t^p[1] + ... + t^p[k]
11335 p[0] > p[1] > ... > p[k] = 0.
11336 This applies to the following functions:
11339 BN_GF2m_mod_mul_arr
11340 BN_GF2m_mod_sqr_arr
11341 BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv]
11342 BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div]
11343 BN_GF2m_mod_exp_arr
11344 BN_GF2m_mod_sqrt_arr
11345 BN_GF2m_mod_solve_quad_arr
11349 Conversion can be performed by the following functions:
11354 bntest.c has additional tests for binary polynomial arithmetic.
11356 Two implementations for BN_GF2m_mod_div() are available.
11357 The default algorithm simply uses BN_GF2m_mod_inv() and
11358 BN_GF2m_mod_mul(). The alternative algorithm is compiled in only
11359 if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the
11360 copyright notice in crypto/bn/bn_gf2m.c before enabling it).
11362 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
11364 * Add new error code 'ERR_R_DISABLED' that can be used when some
11365 functionality is disabled at compile-time.
11367 *Douglas Stebila <douglas.stebila@sun.com>*
11369 * Change default behaviour of 'openssl asn1parse' so that more
11370 information is visible when viewing, e.g., a certificate:
11372 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump'
11373 mode the content of non-printable OCTET STRINGs is output in a
11374 style similar to INTEGERs, but with '[HEX DUMP]' prepended to
11375 avoid the appearance of a printable string.
11377 *Nils Larsch <nla@trustcenter.de>*
11379 * Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access
11381 EC_GROUP_set_asn1_flag()
11382 EC_GROUP_get_asn1_flag()
11383 EC_GROUP_set_point_conversion_form()
11384 EC_GROUP_get_point_conversion_form()
11385 These control ASN1 encoding details:
11386 - Curves (i.e., groups) are encoded explicitly unless asn1_flag
11387 has been set to OPENSSL_EC_NAMED_CURVE.
11388 - Points are encoded in uncompressed form by default; options for
11389 asn1_for are as for point2oct, namely
11390 POINT_CONVERSION_COMPRESSED
11391 POINT_CONVERSION_UNCOMPRESSED
11392 POINT_CONVERSION_HYBRID
11394 Also add 'seed' and 'seed_len' members to EC_GROUP with access
11396 EC_GROUP_set_seed()
11397 EC_GROUP_get0_seed()
11398 EC_GROUP_get_seed_len()
11399 This is used only for ASN1 purposes (so far).
11401 *Nils Larsch <nla@trustcenter.de>*
11403 * Add 'field_type' member to EC_METHOD, which holds the NID
11404 of the appropriate field type OID. The new function
11405 EC_METHOD_get_field_type() returns this value.
11407 *Nils Larsch <nla@trustcenter.de>*
11410 EC_POINT_point2bn()
11411 EC_POINT_bn2point()
11412 EC_POINT_point2hex()
11413 EC_POINT_hex2point()
11414 providing useful interfaces to EC_POINT_point2oct() and
11415 EC_POINT_oct2point().
11417 *Nils Larsch <nla@trustcenter.de>*
11419 * Change internals of the EC library so that the functions
11420 EC_GROUP_set_generator()
11421 EC_GROUP_get_generator()
11422 EC_GROUP_get_order()
11423 EC_GROUP_get_cofactor()
11424 are implemented directly in crypto/ec/ec_lib.c and not dispatched
11425 to methods, which would lead to unnecessary code duplication when
11426 adding different types of curves.
11428 *Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller*
11430 * Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM
11431 arithmetic, and such that modified wNAFs are generated
11432 (which avoid length expansion in many cases).
11436 * Add a function EC_GROUP_check_discriminant() (defined via
11437 EC_METHOD) that verifies that the curve discriminant is non-zero.
11439 Add a function EC_GROUP_check() that makes some sanity tests
11440 on a EC_GROUP, its generator and order. This includes
11441 EC_GROUP_check_discriminant().
11443 *Nils Larsch <nla@trustcenter.de>*
11445 * Add ECDSA in new directory crypto/ecdsa/.
11447 Add applications 'openssl ecparam' and 'openssl ecdsa'
11448 (these are based on 'openssl dsaparam' and 'openssl dsa').
11450 ECDSA support is also included in various other files across the
11451 library. Most notably,
11452 - 'openssl req' now has a '-newkey ecdsa:file' option;
11453 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA;
11454 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and
11455 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make
11456 them suitable for ECDSA where domain parameters must be
11457 extracted before the specific public key;
11458 - ECDSA engine support has been added.
11460 *Nils Larsch <nla@trustcenter.de>*
11462 * Include some named elliptic curves, and add OIDs from X9.62,
11463 SECG, and WAP/WTLS. Each curve can be obtained from the new
11465 EC_GROUP_new_by_curve_name(),
11466 and the list of available named curves can be obtained with
11467 EC_get_builtin_curves().
11468 Also add a 'curve_name' member to EC_GROUP objects, which can be
11470 EC_GROUP_set_curve_name()
11471 EC_GROUP_get_curve_name()
11473 *Nils Larsch <larsch@trustcenter.de, Bodo Moeller*
11475 * Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
11476 was actually never needed) and in BN_mul(). The removal in BN_mul()
11477 required a small change in bn_mul_part_recursive() and the addition
11478 of the functions bn_cmp_part_words(), bn_sub_part_words() and
11479 bn_add_part_words(), which do the same thing as bn_cmp_words(),
11480 bn_sub_words() and bn_add_words() except they take arrays with
11485 ### Changes between 0.9.7l and 0.9.7m [23 Feb 2007]
11487 * Cleanse PEM buffers before freeing them since they may contain
11490 *Benjamin Bennett <ben@psc.edu>*
11492 * Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
11493 a ciphersuite string such as "DEFAULT:RSA" cannot enable
11494 authentication-only ciphersuites.
11498 * Since AES128 and AES256 share a single mask bit in the logic of
11499 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
11500 kludge to work properly if AES128 is available and AES256 isn't.
11504 * Expand security boundary to match 1.1.1 module.
11508 * Remove redundant features: hash file source, editing of test vectors
11509 modify fipsld to use external fips_premain.c signature.
11513 * New perl script mkfipsscr.pl to create shell scripts or batch files to
11514 run algorithm test programs.
11518 * Make algorithm test programs more tolerant of whitespace.
11522 * Have SSL/TLS server implementation tolerate "mismatched" record
11523 protocol version while receiving ClientHello even if the
11524 ClientHello is fragmented. (The server can't insist on the
11525 particular protocol version it has chosen before the ServerHello
11526 message has informed the client about his choice.)
11530 * Load error codes if they are not already present instead of using a
11531 static variable. This allows them to be cleanly unloaded and reloaded.
11535 ### Changes between 0.9.7k and 0.9.7l [28 Sep 2006]
11537 * Introduce limits to prevent malicious keys being able to
11538 cause a denial of service. ([CVE-2006-2940])
11540 *Steve Henson, Bodo Moeller*
11542 * Fix ASN.1 parsing of certain invalid structures that can result
11543 in a denial of service. ([CVE-2006-2937]) [Steve Henson]
11545 * Fix buffer overflow in SSL_get_shared_ciphers() function.
11546 ([CVE-2006-3738]) [Tavis Ormandy and Will Drewry, Google Security Team]
11548 * Fix SSL client code which could crash if connecting to a
11549 malicious SSLv2 server. ([CVE-2006-4343])
11551 *Tavis Ormandy and Will Drewry, Google Security Team*
11553 * Change ciphersuite string processing so that an explicit
11554 ciphersuite selects this one ciphersuite (so that "AES256-SHA"
11555 will no longer include "AES128-SHA"), and any other similar
11556 ciphersuite (same bitmap) from *other* protocol versions (so that
11557 "RC4-MD5" will still include both the SSL 2.0 ciphersuite and the
11558 SSL 3.0/TLS 1.0 ciphersuite). This is a backport combining
11559 changes from 0.9.8b and 0.9.8d.
11563 ### Changes between 0.9.7j and 0.9.7k [05 Sep 2006]
11565 * Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
11566 ([CVE-2006-4339]) [Ben Laurie and Google Security Team]
11568 * Change the Unix randomness entropy gathering to use poll() when
11569 possible instead of select(), since the latter has some
11570 undesirable limitations.
11572 *Darryl Miles via Richard Levitte and Bodo Moeller*
11574 * Disable rogue ciphersuites:
11576 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
11577 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
11578 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
11580 The latter two were purportedly from
11581 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
11584 Also deactivate the remaining ciphersuites from
11585 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
11586 unofficial, and the ID has long expired.
11590 * Fix RSA blinding Heisenbug (problems sometimes occurred on
11591 dual-core machines) and other potential thread-safety issues.
11595 ### Changes between 0.9.7i and 0.9.7j [04 May 2006]
11597 * Adapt fipsld and the build system to link against the validated FIPS
11598 module in FIPS mode.
11602 * Fixes for VC++ 2005 build under Windows.
11606 * Add new Windows build target VC-32-GMAKE for VC++. This uses GNU make
11607 from a Windows bash shell such as MSYS. It is autodetected from the
11608 "config" script when run from a VC++ environment. Modify standard VC++
11609 build to use fipscanister.o from the GNU make build.
11613 ### Changes between 0.9.7h and 0.9.7i [14 Oct 2005]
11615 * Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS.
11616 The value now differs depending on if you build for FIPS or not.
11617 BEWARE! A program linked with a shared FIPSed libcrypto can't be
11618 safely run with a non-FIPSed libcrypto, as it may crash because of
11619 the difference induced by this change.
11623 ### Changes between 0.9.7g and 0.9.7h [11 Oct 2005]
11625 * Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
11626 (part of SSL_OP_ALL). This option used to disable the
11627 countermeasure against man-in-the-middle protocol-version
11628 rollback in the SSL 2.0 server implementation, which is a bad
11629 idea. ([CVE-2005-2969])
11631 *Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
11632 for Information Security, National Institute of Advanced Industrial
11633 Science and Technology [AIST, Japan)]*
11635 * Minimal support for X9.31 signatures and PSS padding modes. This is
11636 mainly for FIPS compliance and not fully integrated at this stage.
11640 * For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform
11641 the exponentiation using a fixed-length exponent. (Otherwise,
11642 the information leaked through timing could expose the secret key
11643 after many signatures; cf. Bleichenbacher's attack on DSA with
11648 * Make a new fixed-window mod_exp implementation the default for
11649 RSA, DSA, and DH private-key operations so that the sequence of
11650 squares and multiplies and the memory access pattern are
11651 independent of the particular secret key. This will mitigate
11652 cache-timing and potential related attacks.
11654 BN_mod_exp_mont_consttime() is the new exponentiation implementation,
11655 and this is automatically used by BN_mod_exp_mont() if the new flag
11656 BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH
11657 will use this BN flag for private exponents unless the flag
11658 RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or
11659 DH_FLAG_NO_EXP_CONSTTIME, respectively, is set.
11661 *Matthew D Wood (Intel Corp), with some changes by Bodo Moeller*
11663 * Change the client implementation for SSLv23_method() and
11664 SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0
11665 Client Hello message format if the SSL_OP_NO_SSLv2 option is set.
11666 (Previously, the SSL 2.0 backwards compatible Client Hello
11667 message format would be used even with SSL_OP_NO_SSLv2.)
11671 * Add support for smime-type MIME parameter in S/MIME messages which some
11676 * New function BN_MONT_CTX_set_locked() to set montgomery parameters in
11677 a threadsafe manner. Modify rsa code to use new function and add calls
11678 to dsa and dh code (which had race conditions before).
11682 * Include the fixed error library code in the C error file definitions
11683 instead of fixing them up at runtime. This keeps the error code
11684 structures constant.
11688 ### Changes between 0.9.7f and 0.9.7g [11 Apr 2005]
11690 [NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after
11693 * Fixes for newer kerberos headers. NB: the casts are needed because
11694 the 'length' field is signed on one version and unsigned on another
11695 with no (?) obvious way to tell the difference, without these VC++
11696 complains. Also the "definition" of FAR (blank) is no longer included
11697 nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up
11698 some needed definitions.
11702 * Undo Cygwin change.
11706 * Added support for proxy certificates according to RFC 3820.
11707 Because they may be a security thread to unaware applications,
11708 they must be explicitly allowed in run-time. See
11709 docs/HOWTO/proxy_certificates.txt for further information.
11713 ### Changes between 0.9.7e and 0.9.7f [22 Mar 2005]
11715 * Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating
11716 server and client random values. Previously
11717 (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in
11718 less random data when sizeof(time_t) > 4 (some 64 bit platforms).
11720 This change has negligible security impact because:
11722 1. Server and client random values still have 24 bytes of pseudo random
11725 2. Server and client random values are sent in the clear in the initial
11728 3. The master secret is derived using the premaster secret (48 bytes in
11729 size for static RSA ciphersuites) as well as client server and random
11732 The OpenSSL team would like to thank the UK NISCC for bringing this issue
11735 *Stephen Henson, reported by UK NISCC*
11737 * Use Windows randomness collection on Cygwin.
11741 * Fix hang in EGD/PRNGD query when communication socket is closed
11742 prematurely by EGD/PRNGD.
11744 *Darren Tucker <dtucker@zip.com.au> via Lutz Jänicke, resolves #1014*
11746 * Prompt for pass phrases when appropriate for PKCS12 input format.
11750 * Back-port of selected performance improvements from development
11751 branch, as well as improved support for PowerPC platforms.
11755 * Add lots of checks for memory allocation failure, error codes to indicate
11756 failure and freeing up memory if a failure occurs.
11758 *Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson*
11760 * Add new -passin argument to dgst.
11764 * Perform some character comparisons of different types in X509_NAME_cmp:
11765 this is needed for some certificates that re-encode DNs into UTF8Strings
11766 (in violation of RFC3280) and can't or won't issue name rollover
11771 * Make an explicit check during certificate validation to see that
11772 the CA setting in each certificate on the chain is correct. As a
11773 side effect always do the following basic checks on extensions,
11774 not just when there's an associated purpose to the check:
11776 - if there is an unhandled critical extension (unless the user
11777 has chosen to ignore this fault)
11778 - if the path length has been exceeded (if one is set at all)
11779 - that certain extensions fit the associated purpose (if one has
11784 ### Changes between 0.9.7d and 0.9.7e [25 Oct 2004]
11786 * Avoid a race condition when CRLs are checked in a multi threaded
11787 environment. This would happen due to the reordering of the revoked
11788 entries during signature checking and serial number lookup. Now the
11789 encoding is cached and the serial number sort performed under a lock.
11790 Add new STACK function sk_is_sorted().
11794 * Add Delta CRL to the extension code.
11798 * Various fixes to s3_pkt.c so alerts are sent properly.
11800 *David Holmes <d.holmes@f5.com>*
11802 * Reduce the chances of duplicate issuer name and serial numbers (in
11803 violation of RFC3280) using the OpenSSL certificate creation utilities.
11804 This is done by creating a random 64 bit value for the initial serial
11805 number when a serial number file is created or when a self signed
11806 certificate is created using 'openssl req -x509'. The initial serial
11807 number file is created using 'openssl x509 -next_serial' in CA.pl
11808 rather than being initialized to 1.
11812 ### Changes between 0.9.7c and 0.9.7d [17 Mar 2004]
11814 * Fix null-pointer assignment in do_change_cipher_spec() revealed
11815 by using the Codenomicon TLS Test Tool ([CVE-2004-0079])
11817 *Joe Orton, Steve Henson*
11819 * Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites
11822 *Joe Orton, Steve Henson*
11824 * Make it possible to have multiple active certificates with the same
11825 subject in the CA index file. This is done only if the keyword
11826 'unique_subject' is set to 'no' in the main CA section (default
11827 if 'CA_default') of the configuration file. The value is saved
11828 with the database itself in a separate index attribute file,
11829 named like the index file with '.attr' appended to the name.
11833 * X509 verify fixes. Disable broken certificate workarounds when
11834 X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if
11835 keyUsage extension present. Don't accept CRLs with unhandled critical
11836 extensions: since verify currently doesn't process CRL extensions this
11837 rejects a CRL with *any* critical extensions. Add new verify error codes
11842 * When creating an OCSP nonce use an OCTET STRING inside the extnValue.
11843 A clarification of RFC2560 will require the use of OCTET STRINGs and
11844 some implementations cannot handle the current raw format. Since OpenSSL
11845 copies and compares OCSP nonces as opaque blobs without any attempt at
11846 parsing them this should not create any compatibility issues.
11850 * New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when
11851 calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without
11852 this HMAC (and other) operations are several times slower than OpenSSL
11857 * Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex().
11859 *Peter Sylvester <Peter.Sylvester@EdelWeb.fr>*
11861 * Use the correct content when signing type "other".
11865 ### Changes between 0.9.7b and 0.9.7c [30 Sep 2003]
11867 * Fix various bugs revealed by running the NISCC test suite:
11869 Stop out of bounds reads in the ASN1 code when presented with
11870 invalid tags (CVE-2003-0543 and CVE-2003-0544).
11872 Free up ASN1_TYPE correctly if ANY type is invalid ([CVE-2003-0545]).
11874 If verify callback ignores invalid public key errors don't try to check
11875 certificate signature with the NULL public key.
11879 * New -ignore_err option in ocsp application to stop the server
11880 exiting on the first error in a request.
11884 * In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
11885 if the server requested one: as stated in TLS 1.0 and SSL 3.0
11890 * In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
11891 extra data after the compression methods not only for TLS 1.0
11892 but also for SSL 3.0 (as required by the specification).
11894 *Bodo Moeller; problem pointed out by Matthias Loepfe*
11896 * Change X509_certificate_type() to mark the key as exported/exportable
11897 when it's 512 *bits* long, not 512 bytes.
11901 * Change AES_cbc_encrypt() so it outputs exact multiple of
11902 blocks during encryption.
11906 * Various fixes to base64 BIO and non blocking I/O. On write
11907 flushes were not handled properly if the BIO retried. On read
11908 data was not being buffered properly and had various logic bugs.
11909 This also affects blocking I/O when the data being decoded is a
11914 * Various S/MIME bugfixes and compatibility changes:
11915 output correct application/pkcs7 MIME type if
11916 PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures.
11917 Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening
11918 of files as .eml work). Correctly handle very long lines in MIME
11923 ### Changes between 0.9.7a and 0.9.7b [10 Apr 2003]
11925 * Countermeasure against the Klima-Pokorny-Rosa extension of
11926 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
11927 a protocol version number mismatch like a decryption error
11928 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
11932 * Turn on RSA blinding by default in the default implementation
11933 to avoid a timing attack. Applications that don't want it can call
11934 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
11935 They would be ill-advised to do so in most cases.
11937 *Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller*
11939 * Change RSA blinding code so that it works when the PRNG is not
11940 seeded (in this case, the secret RSA exponent is abused as
11941 an unpredictable seed -- if it is not unpredictable, there
11942 is no point in blinding anyway). Make RSA blinding thread-safe
11943 by remembering the creator's thread ID in rsa->blinding and
11944 having all other threads use local one-time blinding factors
11945 (this requires more computation than sharing rsa->blinding, but
11946 avoids excessive locking; and if an RSA object is not shared
11947 between threads, blinding will still be very fast).
11951 * Fixed a typo bug that would cause ENGINE_set_default() to set an
11952 ENGINE as defaults for all supported algorithms irrespective of
11953 the 'flags' parameter. 'flags' is now honoured, so applications
11954 should make sure they are passing it correctly.
11958 * Target "mingw" now allows native Windows code to be generated in
11959 the Cygwin environment as well as with the MinGW compiler.
11963 ### Changes between 0.9.7 and 0.9.7a [19 Feb 2003]
11965 * In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
11966 via timing by performing a MAC computation even if incorrect
11967 block cipher padding has been found. This is a countermeasure
11968 against active attacks where the attacker has to distinguish
11969 between bad padding and a MAC verification error. ([CVE-2003-0078])
11971 *Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
11972 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
11973 Martin Vuagnoux (EPFL, Ilion)*
11975 * Make the no-err option work as intended. The intention with no-err
11976 is not to have the whole error stack handling routines removed from
11977 libcrypto, it's only intended to remove all the function name and
11978 reason texts, thereby removing some of the footprint that may not
11979 be interesting if those errors aren't displayed anyway.
11981 NOTE: it's still possible for any application or module to have its
11982 own set of error texts inserted. The routines are there, just not
11983 used by default when no-err is given.
11987 * Add support for FreeBSD on IA64.
11989 *dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454*
11991 * Adjust DES_cbc_cksum() so it returns the same value as the MIT
11992 Kerberos function mit_des_cbc_cksum(). Before this change,
11993 the value returned by DES_cbc_cksum() was like the one from
11994 mit_des_cbc_cksum(), except the bytes were swapped.
11996 *Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte*
11998 * Allow an application to disable the automatic SSL chain building.
11999 Before this a rather primitive chain build was always performed in
12000 ssl3_output_cert_chain(): an application had no way to send the
12001 correct chain if the automatic operation produced an incorrect result.
12003 Now the chain builder is disabled if either:
12005 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert().
12007 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set.
12009 The reasoning behind this is that an application would not want the
12010 auto chain building to take place if extra chain certificates are
12011 present and it might also want a means of sending no additional
12012 certificates (for example the chain has two certificates and the
12017 * Add the possibility to build without the ENGINE framework.
12019 *Steven Reddie <smr@essemer.com.au> via Richard Levitte*
12021 * Under Win32 gmtime() can return NULL: check return value in
12022 OPENSSL_gmtime(). Add error code for case where gmtime() fails.
12026 * DSA routines: under certain error conditions uninitialized BN objects
12027 could be freed. Solution: make sure initialization is performed early
12028 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>,
12029 Nils Larsch <nla@trustcenter.de> via PR#459)
12033 * Another fix for SSLv2 session ID handling: the session ID was incorrectly
12034 checked on reconnect on the client side, therefore session resumption
12035 could still fail with a "ssl session id is different" error. This
12036 behaviour is masked when SSL_OP_ALL is used due to
12037 SSL_OP_MICROSOFT_SESS_ID_BUG being set.
12038 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
12039 followup to PR #377.
12043 * IA-32 assembler support enhancements: unified ELF targets, support
12044 for SCO/Caldera platforms, fix for Cygwin shared build.
12048 * Add support for FreeBSD on sparc64. As a consequence, support for
12049 FreeBSD on non-x86 processors is separate from x86 processors on
12050 the config script, much like the NetBSD support.
12052 *Richard Levitte & Kris Kennaway <kris@obsecurity.org>*
12054 ### Changes between 0.9.6h and 0.9.7 [31 Dec 2002]
12056 [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after
12059 * Fix session ID handling in SSLv2 client code: the SERVER FINISHED
12060 code (06) was taken as the first octet of the session ID and the last
12061 octet was ignored consequently. As a result SSLv2 client side session
12062 caching could not have worked due to the session ID mismatch between
12064 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
12069 * Change the declaration of needed Kerberos libraries to use EX_LIBS
12070 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is
12075 * The hw_ncipher.c engine requires dynamic locks. Unfortunately, it
12076 seems that in spite of existing for more than a year, many application
12077 author have done nothing to provide the necessary callbacks, which
12078 means that this particular engine will not work properly anywhere.
12079 This is a very unfortunate situation which forces us, in the name
12080 of usability, to give the hw_ncipher.c a static lock, which is part
12082 NOTE: This is for the 0.9.7 series ONLY. This hack will never
12083 appear in 0.9.8 or later. We EXPECT application authors to have
12084 dealt properly with this when 0.9.8 is released (unless we actually
12085 make such changes in the libcrypto locking code that changes will
12086 have to be made anyway).
12090 * In asn1_d2i_read_bio() repeatedly call BIO_read() until all content
12091 octets have been read, EOF or an error occurs. Without this change
12092 some truncated ASN1 structures will not produce an error.
12096 * Disable Heimdal support, since it hasn't been fully implemented.
12097 Still give the possibility to force the use of Heimdal, but with
12098 warnings and a request that patches get sent to openssl-dev.
12102 * Add the VC-CE target, introduce the WINCE sysname, and add
12103 INSTALL.WCE and appropriate conditionals to make it build.
12105 *Steven Reddie <smr@essemer.com.au> via Richard Levitte*
12107 * Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and
12108 cygssl-x.y.z.dll, where x, y and z are the major, minor and
12109 edit numbers of the version.
12111 *Corinna Vinschen <vinschen@redhat.com> and Richard Levitte*
12113 * Introduce safe string copy and catenation functions
12114 (BUF_strlcpy() and BUF_strlcat()).
12116 *Ben Laurie (CHATS) and Richard Levitte*
12118 * Avoid using fixed-size buffers for one-line DNs.
12120 *Ben Laurie (CHATS)*
12122 * Add BUF_MEM_grow_clean() to avoid information leakage when
12123 resizing buffers containing secrets, and use where appropriate.
12125 *Ben Laurie (CHATS)*
12127 * Avoid using fixed size buffers for configuration file location.
12129 *Ben Laurie (CHATS)*
12131 * Avoid filename truncation for various CA files.
12133 *Ben Laurie (CHATS)*
12135 * Use sizeof in preference to magic numbers.
12137 *Ben Laurie (CHATS)*
12139 * Avoid filename truncation in cert requests.
12141 *Ben Laurie (CHATS)*
12143 * Add assertions to check for (supposedly impossible) buffer
12146 *Ben Laurie (CHATS)*
12148 * Don't cache truncated DNS entries in the local cache (this could
12149 potentially lead to a spoofing attack).
12151 *Ben Laurie (CHATS)*
12153 * Fix various buffers to be large enough for hex/decimal
12154 representations in a platform independent manner.
12156 *Ben Laurie (CHATS)*
12158 * Add CRYPTO_realloc_clean() to avoid information leakage when
12159 resizing buffers containing secrets, and use where appropriate.
12161 *Ben Laurie (CHATS)*
12163 * Add BIO_indent() to avoid much slightly worrying code to do
12166 *Ben Laurie (CHATS)*
12168 * Convert sprintf()/BIO_puts() to BIO_printf().
12170 *Ben Laurie (CHATS)*
12172 * buffer_gets() could terminate with the buffer only half
12175 *Ben Laurie (CHATS)*
12177 * Add assertions to prevent user-supplied crypto functions from
12178 overflowing internal buffers by having large block sizes, etc.
12180 *Ben Laurie (CHATS)*
12182 * New OPENSSL_assert() macro (similar to assert(), but enabled
12185 *Ben Laurie (CHATS)*
12187 * Eliminate unused copy of key in RC4.
12189 *Ben Laurie (CHATS)*
12191 * Eliminate unused and incorrectly sized buffers for IV in pem.h.
12193 *Ben Laurie (CHATS)*
12195 * Fix off-by-one error in EGD path.
12197 *Ben Laurie (CHATS)*
12199 * If RANDFILE path is too long, ignore instead of truncating.
12201 *Ben Laurie (CHATS)*
12203 * Eliminate unused and incorrectly sized X.509 structure
12206 *Ben Laurie (CHATS)*
12208 * Eliminate unused and dangerous function knumber().
12210 *Ben Laurie (CHATS)*
12212 * Eliminate unused and dangerous structure, KSSL_ERR.
12214 *Ben Laurie (CHATS)*
12216 * Protect against overlong session ID context length in an encoded
12217 session object. Since these are local, this does not appear to be
12220 *Ben Laurie (CHATS)*
12222 * Change from security patch (see 0.9.6e below) that did not affect
12223 the 0.9.6 release series:
12225 Remote buffer overflow in SSL3 protocol - an attacker could
12226 supply an oversized master key in Kerberos-enabled versions.
12229 *Ben Laurie (CHATS)*
12231 * Change the SSL kerb5 codes to match RFC 2712.
12235 * Make -nameopt work fully for req and add -reqopt switch.
12237 *Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson*
12239 * The "block size" for block ciphers in CFB and OFB mode should be 1.
12241 *Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>*
12243 * Make sure tests can be performed even if the corresponding algorithms
12244 have been removed entirely. This was also the last step to make
12245 OpenSSL compilable with DJGPP under all reasonable conditions.
12247 *Richard Levitte, Doug Kaufman <dkaufman@rahul.net>*
12249 * Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT
12250 to allow version independent disabling of normally unselected ciphers,
12251 which may be activated as a side-effect of selecting a single cipher.
12253 (E.g., cipher list string "RSA" enables ciphersuites that are left
12254 out of "ALL" because they do not provide symmetric encryption.
12255 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.)
12257 *Lutz Jaenicke, Bodo Moeller*
12259 * Add appropriate support for separate platform-dependent build
12260 directories. The recommended way to make a platform-dependent
12261 build directory is the following (tested on Linux), maybe with
12264 # Place yourself outside of the OpenSSL source tree. In
12265 # this example, the environment variable OPENSSL_SOURCE
12266 # is assumed to contain the absolute OpenSSL source directory.
12267 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`"
12268 cd objtree/"`uname -s`-`uname -r`-`uname -m`"
12269 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do
12270 mkdir -p `dirname $F`
12271 ln -s $OPENSSL_SOURCE/$F $F
12274 To be absolutely sure not to disturb the source tree, a "make clean"
12275 is a good thing. If it isn't successful, don't worry about it,
12276 it probably means the source directory is very clean.
12280 * Make sure any ENGINE control commands make local copies of string
12281 pointers passed to them whenever necessary. Otherwise it is possible
12282 the caller may have overwritten (or deallocated) the original string
12283 data when a later ENGINE operation tries to use the stored values.
12285 *Götz Babin-Ebell <babinebell@trustcenter.de>*
12287 * Improve diagnostics in file reading and command-line digests.
12289 *Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>*
12291 * Add AES modes CFB and OFB to the object database. Correct an
12292 error in AES-CFB decryption.
12296 * Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this
12297 allows existing EVP_CIPHER_CTX structures to be reused after
12298 calling `EVP_*Final()`. This behaviour is used by encryption
12299 BIOs and some applications. This has the side effect that
12300 applications must explicitly clean up cipher contexts with
12301 EVP_CIPHER_CTX_cleanup() or they will leak memory.
12305 * Check the values of dna and dnb in bn_mul_recursive before calling
12306 bn_mul_comba (a non zero value means the a or b arrays do not contain
12307 n2 elements) and fallback to bn_mul_normal if either is not zero.
12311 * Fix escaping of non-ASCII characters when using the -subj option
12312 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>)
12316 * Make object definitions compliant to LDAP (RFC2256): SN is the short
12317 form for "surname", serialNumber has no short form.
12318 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798;
12319 therefore remove "mail" short name for "internet 7".
12320 The OID for unique identifiers in X509 certificates is
12321 x500UniqueIdentifier, not uniqueIdentifier.
12322 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>)
12326 * Add an "init" command to the ENGINE config module and auto initialize
12327 ENGINEs. Without any "init" command the ENGINE will be initialized
12328 after all ctrl commands have been executed on it. If init=1 the
12329 ENGINE is initialized at that point (ctrls before that point are run
12330 on the uninitialized ENGINE and after on the initialized one). If
12331 init=0 then the ENGINE will not be initialized at all.
12335 * Fix the 'app_verify_callback' interface so that the user-defined
12336 argument is actually passed to the callback: In the
12337 SSL_CTX_set_cert_verify_callback() prototype, the callback
12338 declaration has been changed from
12341 int (*cb)(X509_STORE_CTX *,void *);
12342 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
12343 i=s->ctx->app_verify_callback(&ctx)
12344 has been changed into
12345 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
12347 To update applications using SSL_CTX_set_cert_verify_callback(),
12348 a dummy argument can be added to their callback functions.
12350 *D. K. Smetters <smetters@parc.xerox.com>*
12352 * Added the '4758cca' ENGINE to support IBM 4758 cards.
12354 *Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe*
12356 * Add and OPENSSL_LOAD_CONF define which will cause
12357 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
12358 This allows older applications to transparently support certain
12359 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
12360 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
12361 load the config file and OPENSSL_add_all_algorithms_conf() which will
12362 always load it have also been added.
12366 * Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
12367 Adjust NIDs and EVP layer.
12369 *Stephen Sprunk <stephen@sprunk.org> and Richard Levitte*
12371 * Config modules support in openssl utility.
12373 Most commands now load modules from the config file,
12374 though in a few (such as version) this isn't done
12375 because it couldn't be used for anything.
12377 In the case of ca and req the config file used is
12378 the same as the utility itself: that is the -config
12379 command line option can be used to specify an
12384 * Move default behaviour from OPENSSL_config(). If appname is NULL
12385 use "openssl_conf" if filename is NULL use default openssl config file.
12389 * Add an argument to OPENSSL_config() to allow the use of an alternative
12390 config section name. Add a new flag to tolerate a missing config file
12391 and move code to CONF_modules_load_file().
12395 * Support for crypto accelerator cards from Accelerated Encryption
12396 Processing, www.aep.ie. (Use engine 'aep')
12397 The support was copied from 0.9.6c [engine] and adapted/corrected
12398 to work with the new engine framework.
12400 *AEP Inc. and Richard Levitte*
12402 * Support for SureWare crypto accelerator cards from Baltimore
12403 Technologies. (Use engine 'sureware')
12404 The support was copied from 0.9.6c [engine] and adapted
12405 to work with the new engine framework.
12409 * Have the CHIL engine fork-safe (as defined by nCipher) and actually
12410 make the newer ENGINE framework commands for the CHIL engine work.
12412 *Toomas Kiisk <vix@cyber.ee> and Richard Levitte*
12414 * Make it possible to produce shared libraries on ReliantUNIX.
12416 *Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte*
12418 * Add the configuration target debug-linux-ppro.
12419 Make 'openssl rsa' use the general key loading routines
12420 implemented in `apps.c`, and make those routines able to
12421 handle the key format FORMAT_NETSCAPE and the variant
12424 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
12426 * Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
12428 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
12430 * Add -keyform to rsautl, and document -engine.
12432 *Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>*
12434 * Change BIO_new_file (crypto/bio/bss_file.c) to use new
12435 BIO_R_NO_SUCH_FILE error code rather than the generic
12436 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
12440 * Add new functions
12441 ERR_peek_last_error
12442 ERR_peek_last_error_line
12443 ERR_peek_last_error_line_data.
12444 These are similar to
12446 ERR_peek_error_line
12447 ERR_peek_error_line_data,
12448 but report on the latest error recorded rather than the first one
12449 still in the error queue.
12451 *Ben Laurie, Bodo Moeller*
12453 * default_algorithms option in ENGINE config module. This allows things
12455 default_algorithms = ALL
12456 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
12460 * Preliminary ENGINE config module.
12464 * New experimental application configuration code.
12468 * Change the AES code to follow the same name structure as all other
12469 symmetric ciphers, and behave the same way. Move everything to
12470 the directory crypto/aes, thereby obsoleting crypto/rijndael.
12472 *Stephen Sprunk <stephen@sprunk.org> and Richard Levitte*
12474 * SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
12476 *Ben Laurie and Theo de Raadt*
12478 * Add option to output public keys in req command.
12480 *Massimiliano Pala madwolf@openca.org*
12482 * Use wNAFs in EC_POINTs_mul() for improved efficiency
12483 (up to about 10% better than before for P-192 and P-224).
12487 * New functions/macros
12489 SSL_CTX_set_msg_callback(ctx, cb)
12490 SSL_CTX_set_msg_callback_arg(ctx, arg)
12491 SSL_set_msg_callback(ssl, cb)
12492 SSL_set_msg_callback_arg(ssl, arg)
12494 to request calling a callback function
12496 void cb(int write_p, int version, int content_type,
12497 const void *buf, size_t len, SSL *ssl, void *arg)
12499 whenever a protocol message has been completely received
12500 (write_p == 0) or sent (write_p == 1). Here 'version' is the
12501 protocol version according to which the SSL library interprets
12502 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
12503 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
12504 the content type as defined in the SSL 3.0/TLS 1.0 protocol
12505 specification (change_cipher_spec(20), alert(21), handshake(22)).
12506 'buf' and 'len' point to the actual message, 'ssl' to the
12507 SSL object, and 'arg' is the application-defined value set by
12508 SSL[_CTX]_set_msg_callback_arg().
12510 'openssl s_client' and 'openssl s_server' have new '-msg' options
12511 to enable a callback that displays all protocol messages.
12515 * Change the shared library support so shared libraries are built as
12516 soon as the corresponding static library is finished, and thereby get
12517 openssl and the test programs linked against the shared library.
12518 This still only happens when the keyword "shard" has been given to
12519 the configuration scripts.
12521 NOTE: shared library support is still an experimental thing, and
12522 backward binary compatibility is still not guaranteed.
12524 *"Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte*
12526 * Add support for Subject Information Access extension.
12528 *Peter Sylvester <Peter.Sylvester@EdelWeb.fr>*
12530 * Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
12531 additional bytes when new memory had to be allocated, not just
12532 when reusing an existing buffer.
12536 * New command line and configuration option 'utf8' for the req command.
12537 This allows field values to be specified as UTF8 strings.
12541 * Add -multi and -mr options to "openssl speed" - giving multiple parallel
12542 runs for the former and machine-readable output for the latter.
12546 * Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
12547 of the e-mail address in the DN (i.e., it will go into a certificate
12548 extension only). The new configuration file option 'email_in_dn = no'
12549 has the same effect.
12551 *Massimiliano Pala madwolf@openca.org*
12553 * Change all functions with names starting with `des_` to be starting
12554 with `DES_` instead. Add wrappers that are compatible with libdes,
12555 but are named `_ossl_old_des_*`. Finally, add macros that map the
12556 `des_*` symbols to the corresponding `_ossl_old_des_*` if libdes
12557 compatibility is desired. If OpenSSL 0.9.6c compatibility is
12558 desired, the `des_*` symbols will be mapped to `DES_*`, with one
12561 Since we provide two compatibility mappings, the user needs to
12562 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
12563 compatibility is desired. The default (i.e., when that macro
12564 isn't defined) is OpenSSL 0.9.6c compatibility.
12566 There are also macros that enable and disable the support of old
12567 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
12568 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
12569 are defined, the default will apply: to support the old des routines.
12571 In either case, one must include openssl/des.h to get the correct
12572 definitions. Do not try to just include openssl/des_old.h, that
12575 NOTE: This is a major break of an old API into a new one. Software
12576 authors are encouraged to switch to the `DES_` style functions. Some
12577 time in the future, des_old.h and the libdes compatibility functions
12578 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
12579 default), and then completely removed.
12583 * Test for certificates which contain unsupported critical extensions.
12584 If such a certificate is found during a verify operation it is
12585 rejected by default: this behaviour can be overridden by either
12586 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
12587 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
12588 X509_supported_extension() has also been added which returns 1 if a
12589 particular extension is supported.
12593 * Modify the behaviour of EVP cipher functions in similar way to digests
12594 to retain compatibility with existing code.
12598 * Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
12599 compatibility with existing code. In particular the 'ctx' parameter does
12600 not have to be to be initialized before the call to EVP_DigestInit() and
12601 it is tidied up after a call to EVP_DigestFinal(). New function
12602 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
12603 EVP_MD_CTX_copy() changed to not require the destination to be
12604 initialized valid and new function EVP_MD_CTX_copy_ex() added which
12605 requires the destination to be valid.
12607 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
12608 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
12612 * Change ssl3_get_message (ssl/s3_both.c) and the functions using it
12613 so that complete 'Handshake' protocol structures are kept in memory
12614 instead of overwriting 'msg_type' and 'length' with 'body' data.
12618 * Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
12620 *Massimo Santin via Richard Levitte*
12622 * Major restructuring to the underlying ENGINE code. This includes
12623 reduction of linker bloat, separation of pure "ENGINE" manipulation
12624 (initialisation, etc) from functionality dealing with implementations
12625 of specific crypto interfaces. This change also introduces integrated
12626 support for symmetric ciphers and digest implementations - so ENGINEs
12627 can now accelerate these by providing EVP_CIPHER and EVP_MD
12628 implementations of their own. This is detailed in
12629 [crypto/engine/README.md](crypto/engine/README.md)
12630 as it couldn't be adequately described here. However, there are a few
12631 API changes worth noting - some RSA, DSA, DH, and RAND functions that
12632 were changed in the original introduction of ENGINE code have now
12633 reverted back - the hooking from this code to ENGINE is now a good
12634 deal more passive and at run-time, operations deal directly with
12635 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
12636 dereferencing through an ENGINE pointer any more. Also, the ENGINE
12637 functions dealing with `BN_MOD_EXP[_CRT]` handlers have been removed -
12638 they were not being used by the framework as there is no concept of a
12639 BIGNUM_METHOD and they could not be generalised to the new
12640 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
12641 ENGINE_cpy() has been removed as it cannot be consistently defined in
12646 * Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
12650 * Change mkdef.pl to sort symbols that get the same entry number,
12651 and make sure the automatically generated functions `ERR_load_*`
12652 become part of libeay.num as well.
12656 * New function SSL_renegotiate_pending(). This returns true once
12657 renegotiation has been requested (either SSL_renegotiate() call
12658 or HelloRequest/ClientHello received from the peer) and becomes
12659 false once a handshake has been completed.
12660 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
12661 sends a HelloRequest, but does not ensure that a handshake takes
12662 place. SSL_renegotiate_pending() is useful for checking if the
12663 client has followed the request.)
12667 * New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
12668 By default, clients may request session resumption even during
12669 renegotiation (if session ID contexts permit); with this option,
12670 session resumption is possible only in the first handshake.
12672 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes
12673 more bits available for options that should not be part of
12674 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION).
12678 * Add some demos for certificate and certificate request creation.
12682 * Make maximum certificate chain size accepted from the peer application
12683 settable (`SSL*_get/set_max_cert_list()`), as proposed by
12684 "Douglas E. Engert" <deengert@anl.gov>.
12688 * Add support for shared libraries for Unixware-7
12689 (Boyd Lynn Gerber <gerberb@zenez.com>).
12693 * Add a "destroy" handler to ENGINEs that allows structural cleanup to
12694 be done prior to destruction. Use this to unload error strings from
12695 ENGINEs that load their own error strings. NB: This adds two new API
12696 functions to "get" and "set" this destroy handler in an ENGINE.
12700 * Alter all existing ENGINE implementations (except "openssl" and
12701 "openbsd") to dynamically instantiate their own error strings. This
12702 makes them more flexible to be built both as statically-linked ENGINEs
12703 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
12704 Also, add stub code to each that makes building them as self-contained
12705 shared-libraries easier (see [README-Engine.md](README-Engine.md)).
12709 * Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
12710 implementations into applications that are completely implemented in
12711 self-contained shared-libraries. The "dynamic" ENGINE exposes control
12712 commands that can be used to configure what shared-library to load and
12713 to control aspects of the way it is handled. Also, made an update to
12714 the [README-Engine.md](README-Engine.md) file
12715 that brings its information up-to-date and
12716 provides some information and instructions on the "dynamic" ENGINE
12717 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
12721 * Make it possible to unload ranges of ERR strings with a new
12722 "ERR_unload_strings" function.
12726 * Add a copy() function to EVP_MD.
12730 * Make EVP_MD routines take a context pointer instead of just the
12731 md_data void pointer.
12735 * Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
12736 that the digest can only process a single chunk of data
12737 (typically because it is provided by a piece of
12738 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
12739 is only going to provide a single chunk of data, and hence the
12740 framework needn't accumulate the data for oneshot drivers.
12744 * As with "ERR", make it possible to replace the underlying "ex_data"
12745 functions. This change also alters the storage and management of global
12746 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
12747 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
12748 index counters. The API functions that use this state have been changed
12749 to take a "class_index" rather than pointers to the class's local STACK
12750 and counter, and there is now an API function to dynamically create new
12751 classes. This centralisation allows us to (a) plug a lot of the
12752 thread-safety problems that existed, and (b) makes it possible to clean
12753 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
12754 such data would previously have always leaked in application code and
12755 workarounds were in place to make the memory debugging turn a blind eye
12756 to it. Application code that doesn't use this new function will still
12757 leak as before, but their memory debugging output will announce it now
12758 rather than letting it slide.
12760 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
12761 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
12762 has a return value to indicate success or failure.
12766 * Make it possible to replace the underlying "ERR" functions such that the
12767 global state (2 LHASH tables and 2 locks) is only used by the "default"
12768 implementation. This change also adds two functions to "get" and "set"
12769 the implementation prior to it being automatically set the first time
12770 any other ERR function takes place. Ie. an application can call "get",
12771 pass the return value to a module it has just loaded, and that module
12772 can call its own "set" function using that value. This means the
12773 module's "ERR" operations will use (and modify) the error state in the
12774 application and not in its own statically linked copy of OpenSSL code.
12778 * Give DH, DSA, and RSA types their own `*_up_ref()` function to increment
12779 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
12780 the operation, and provides a more encapsulated way for external code
12781 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
12782 to use these functions rather than manually incrementing the counts.
12784 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
12788 * Add EVP test program.
12792 * Add symmetric cipher support to ENGINE. Expect the API to change!
12796 * New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
12797 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
12798 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
12799 These allow a CRL to be built without having to access X509_CRL fields
12800 directly. Modify 'ca' application to use new functions.
12804 * Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
12805 bug workarounds. Rollback attack detection is a security feature.
12806 The problem will only arise on OpenSSL servers when TLSv1 is not
12807 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
12808 Software authors not wanting to support TLSv1 will have special reasons
12809 for their choice and can explicitly enable this option.
12811 *Bodo Moeller, Lutz Jaenicke*
12813 * Rationalise EVP so it can be extended: don't include a union of
12814 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX
12815 (similar to those existing for EVP_CIPHER_CTX).
12820 EVP_MD_CTX_init(&md); /* new function call */
12821 EVP_DigestInit(&md, EVP_sha1());
12822 EVP_DigestUpdate(&md, in, len);
12823 EVP_DigestFinal(&md, out, NULL);
12824 EVP_MD_CTX_cleanup(&md); /* new function call */
12828 * Make DES key schedule conform to the usual scheme, as well as
12829 correcting its structure. This means that calls to DES functions
12830 now have to pass a pointer to a des_key_schedule instead of a
12831 plain des_key_schedule (which was actually always a pointer
12834 des_key_schedule ks;
12836 des_set_key_checked(..., &ks);
12837 des_ncbc_encrypt(..., &ks, ...);
12839 (Note that a later change renames 'des_...' into 'DES_...'.)
12843 * Initial reduction of linker bloat: the use of some functions, such as
12844 PEM causes large amounts of unused functions to be linked in due to
12845 poor organisation. For example pem_all.c contains every PEM function
12846 which has a knock on effect of linking in large amounts of (unused)
12847 ASN1 code. Grouping together similar functions and splitting unrelated
12848 functions prevents this.
12852 * Cleanup of EVP macros.
12856 * Change historical references to `{NID,SN,LN}_des_ede` and ede3 to add the
12857 correct `_ecb suffix`.
12861 * Add initial OCSP responder support to ocsp application. The
12862 revocation information is handled using the text based index
12863 use by the ca application. The responder can either handle
12864 requests generated internally, supplied in files (for example
12865 via a CGI script) or using an internal minimal server.
12869 * Add configuration choices to get zlib compression for TLS.
12873 * Changes to Kerberos SSL for RFC 2712 compliance:
12874 1. Implemented real KerberosWrapper, instead of just using
12875 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
12876 2. Implemented optional authenticator field of KerberosWrapper.
12878 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
12879 and authenticator structs; see crypto/krb5/.
12881 Generalized Kerberos calls to support multiple Kerberos libraries.
12882 *Vern Staats <staatsvr@asc.hpc.mil>, Jeffrey Altman <jaltman@columbia.edu>
12883 via Richard Levitte*
12885 * Cause 'openssl speed' to use fully hard-coded DSA keys as it
12886 already does with RSA. testdsa.h now has 'priv_key/pub_key'
12887 values for each of the key sizes rather than having just
12888 parameters (and 'speed' generating keys each time).
12892 * Speed up EVP routines.
12895 pe 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
12896 s-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
12897 s-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
12898 s-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
12900 s-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
12901 s-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
12902 s-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
12905 s-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
12907 s-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
12911 * Added the OS2-EMX target.
12913 *"Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte*
12915 * Rewrite commands to use `NCONF` routines instead of the old `CONF`.
12916 New functions to support `NCONF` routines in extension code.
12917 New function `CONF_set_nconf()`
12918 to allow functions which take an `NCONF` to also handle the old `LHASH`
12919 structure: this means that the old `CONF` compatible routines can be
12920 retained (in particular w.rt. extensions) without having to duplicate the
12921 code. New function `X509V3_add_ext_nconf_sk()` to add extensions to a stack.
12925 * Enhance the general user interface with mechanisms for inner control
12926 and with possibilities to have yes/no kind of prompts.
12930 * Change all calls to low-level digest routines in the library and
12931 applications to use EVP. Add missing calls to HMAC_cleanup() and
12932 don't assume HMAC_CTX can be copied using memcpy().
12934 *Verdon Walker <VWalker@novell.com>, Steve Henson*
12936 * Add the possibility to control engines through control names but with
12937 arbitrary arguments instead of just a string.
12938 Change the key loaders to take a UI_METHOD instead of a callback
12939 function pointer. NOTE: this breaks binary compatibility with earlier
12940 versions of OpenSSL [engine].
12941 Adapt the nCipher code for these new conditions and add a card insertion
12946 * Enhance the general user interface with mechanisms to better support
12947 dialog box interfaces, application-defined prompts, the possibility
12948 to use defaults (for example default passwords from somewhere else)
12949 and interrupts/cancellations.
12953 * Tidy up PKCS#12 attribute handling. Add support for the CSP name
12954 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
12958 * Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
12959 tidy up some unnecessarily weird code in 'sk_new()').
12961 *Geoff, reported by Diego Tartara <dtartara@novamens.com>*
12963 * Change the key loading routines for ENGINEs to use the same kind
12964 callback (pem_password_cb) as all other routines that need this
12969 * Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
12970 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
12971 than this minimum value is recommended.
12975 * New random seeder for OpenVMS, using the system process statistics
12976 that are easily reachable.
12980 * Windows apparently can't transparently handle global
12981 variables defined in DLLs. Initialisations such as:
12983 const ASN1_ITEM *it = &ASN1_INTEGER_it;
12985 won't compile. This is used by the any applications that need to
12986 declare their own ASN1 modules. This was fixed by adding the option
12987 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
12988 needed for static libraries under Win32.
12992 * New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
12993 setting of purpose and trust fields. New X509_STORE trust and
12994 purpose functions and tidy up setting in other SSL functions.
12998 * Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
12999 structure. These are inherited by X509_STORE_CTX when it is
13000 initialised. This allows various defaults to be set in the
13001 X509_STORE structure (such as flags for CRL checking and custom
13002 purpose or trust settings) for functions which only use X509_STORE_CTX
13003 internally such as S/MIME.
13005 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
13006 trust settings if they are not set in X509_STORE. This allows X509_STORE
13007 purposes and trust (in S/MIME for example) to override any set by default.
13009 Add command line options for CRL checking to smime, s_client and s_server
13014 * Initial CRL based revocation checking. If the CRL checking flag(s)
13015 are set then the CRL is looked up in the X509_STORE structure and
13016 its validity and signature checked, then if the certificate is found
13017 in the CRL the verify fails with a revoked error.
13019 Various new CRL related callbacks added to X509_STORE_CTX structure.
13021 Command line options added to 'verify' application to support this.
13023 This needs some additional work, such as being able to handle multiple
13024 CRLs with different times, extension based lookup (rather than just
13025 by subject name) and ultimately more complete V2 CRL extension
13030 * Add a general user interface API (crypto/ui/). This is designed
13031 to replace things like des_read_password and friends (backward
13032 compatibility functions using this new API are provided).
13033 The purpose is to remove prompting functions from the DES code
13034 section as well as provide for prompting through dialog boxes in
13035 a window system and the like.
13039 * Add "ex_data" support to ENGINE so implementations can add state at a
13040 per-structure level rather than having to store it globally.
13044 * Make it possible for ENGINE structures to be copied when retrieved by
13045 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
13046 This causes the "original" ENGINE structure to act like a template,
13047 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
13048 operational state can be localised to each ENGINE structure, despite the
13049 fact they all share the same "methods". New ENGINE structures returned in
13050 this case have no functional references and the return value is the single
13051 structural reference. This matches the single structural reference returned
13052 by ENGINE_by_id() normally, when it is incremented on the pre-existing
13057 * Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
13058 needs to match any other type at all we need to manually clear the
13063 * Changes to the "openssl engine" utility to include;
13064 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
13065 about an ENGINE's available control commands.
13066 - executing control commands from command line arguments using the
13067 '-pre' and '-post' switches. '-post' is only used if '-t' is
13068 specified and the ENGINE is successfully initialised. The syntax for
13069 the individual commands are colon-separated, for example;
13070 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
13074 * New dynamic control command support for ENGINEs. ENGINEs can now
13075 declare their own commands (numbers), names (strings), descriptions,
13076 and input types for run-time discovery by calling applications. A
13077 subset of these commands are implicitly classed as "executable"
13078 depending on their input type, and only these can be invoked through
13079 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
13080 can be based on user input, config files, etc). The distinction is
13081 that "executable" commands cannot return anything other than a boolean
13082 result and can only support numeric or string input, whereas some
13083 discoverable commands may only be for direct use through
13084 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
13085 pointers, or other custom uses. The "executable" commands are to
13086 support parameterisations of ENGINE behaviour that can be
13087 unambiguously defined by ENGINEs and used consistently across any
13088 OpenSSL-based application. Commands have been added to all the
13089 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
13090 control over shared-library paths without source code alterations.
13094 * Changed all ENGINE implementations to dynamically allocate their
13095 ENGINEs rather than declaring them statically. Apart from this being
13096 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
13097 this also allows the implementations to compile without using the
13098 internal engine_int.h header.
13102 * Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
13103 'const' value. Any code that should be able to modify a RAND_METHOD
13104 should already have non-const pointers to it (ie. they should only
13105 modify their own ones).
13109 * Made a variety of little tweaks to the ENGINE code.
13110 - "atalla" and "ubsec" string definitions were moved from header files
13111 to C code. "nuron" string definitions were placed in variables
13112 rather than hard-coded - allowing parameterisation of these values
13113 later on via ctrl() commands.
13114 - Removed unused "#if 0"'d code.
13115 - Fixed engine list iteration code so it uses ENGINE_free() to release
13116 structural references.
13117 - Constified the RAND_METHOD element of ENGINE structures.
13118 - Constified various get/set functions as appropriate and added
13119 missing functions (including a catch-all ENGINE_cpy that duplicates
13120 all ENGINE values onto a new ENGINE except reference counts/state).
13121 - Removed NULL parameter checks in get/set functions. Setting a method
13122 or function to NULL is a way of cancelling out a previously set
13123 value. Passing a NULL ENGINE parameter is just plain stupid anyway
13124 and doesn't justify the extra error symbols and code.
13125 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
13126 flags from engine_int.h to engine.h.
13127 - Changed prototypes for ENGINE handler functions (init(), finish(),
13128 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
13132 * Implement binary inversion algorithm for BN_mod_inverse in addition
13133 to the algorithm using long division. The binary algorithm can be
13134 used only if the modulus is odd. On 32-bit systems, it is faster
13135 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
13136 roughly 5-15% for 256-bit moduli), so we use it only for moduli
13137 up to 450 bits. In 64-bit environments, the binary algorithm
13138 appears to be advantageous for much longer moduli; here we use it
13139 for moduli up to 2048 bits.
13143 * Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
13144 could not support the combine flag in choice fields.
13148 * Add a 'copy_extensions' option to the 'ca' utility. This copies
13149 extensions from a certificate request to the certificate.
13153 * Allow multiple 'certopt' and 'nameopt' options to be separated
13154 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
13155 file: this allows the display of the certificate about to be
13156 signed to be customised, to allow certain fields to be included
13157 or excluded and extension details. The old system didn't display
13158 multicharacter strings properly, omitted fields not in the policy
13159 and couldn't display additional details such as extensions.
13163 * Function EC_POINTs_mul for multiple scalar multiplication
13164 of an arbitrary number of elliptic curve points
13165 \sum scalars[i]*points[i],
13166 optionally including the generator defined for the EC_GROUP:
13167 scalar*generator + \sum scalars[i]*points[i].
13169 EC_POINT_mul is a simple wrapper function for the typical case
13170 that the point list has just one item (besides the optional
13175 * First EC_METHODs for curves over GF(p):
13177 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
13178 operations and provides various method functions that can also
13179 operate with faster implementations of modular arithmetic.
13181 EC_GFp_mont_method() reuses most functions that are part of
13182 EC_GFp_simple_method, but uses Montgomery arithmetic.
13184 *Bodo Moeller; point addition and point doubling
13185 implementation directly derived from source code provided by
13186 Lenka Fibikova <fibikova@exp-math.uni-essen.de>*
13188 * Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
13189 crypto/ec/ec_lib.c):
13191 Curves are EC_GROUP objects (with an optional group generator)
13192 based on EC_METHODs that are built into the library.
13194 Points are EC_POINT objects based on EC_GROUP objects.
13196 Most of the framework would be able to handle curves over arbitrary
13197 finite fields, but as there are no obvious types for fields other
13198 than GF(p), some functions are limited to that for now.
13202 * Add the -HTTP option to s_server. It is similar to -WWW, but requires
13203 that the file contains a complete HTTP response.
13207 * Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
13208 change the def and num file printf format specifier from "%-40sXXX"
13209 to "%-39s XXX". The latter will always guarantee a space after the
13210 field while the former will cause them to run together if the field
13211 is 40 of more characters long.
13215 * Constify the cipher and digest 'method' functions and structures
13216 and modify related functions to take constant EVP_MD and EVP_CIPHER
13221 * Hide BN_CTX structure details in bn_lcl.h instead of publishing them
13222 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
13226 * Modify `EVP_Digest*()` routines so they now return values. Although the
13227 internal software routines can never fail additional hardware versions
13232 * Clean up crypto/err/err.h and change some error codes to avoid conflicts:
13234 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
13235 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
13238 ERR_R_NESTED_ASN1_ERROR
13240 ERR_R_MISSING_ASN1_EOS
13241 were 4 .. 9, conflicting with
13242 ERR_LIB_RSA (= ERR_R_RSA_LIB)
13244 ERR_LIB_PEM (= ERR_R_PEM_LIB).
13245 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
13247 Add new error code 'ERR_R_INTERNAL_ERROR'.
13251 * Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
13256 * New option '-subj arg' for 'openssl req' and 'openssl ca'. This
13257 sets the subject name for a new request or supersedes the
13258 subject name in a given request. Formats that can be parsed are
13259 'CN=Some Name, OU=myOU, C=IT'
13261 'CN=Some Name/OU=myOU/C=IT'.
13263 Add options '-batch' and '-verbose' to 'openssl req'.
13265 *Massimiliano Pala <madwolf@hackmasters.net>*
13267 * Introduce the possibility to access global variables through
13268 functions on platform were that's the best way to handle exporting
13269 global variables in shared libraries. To enable this functionality,
13270 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
13271 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
13272 is normally done by Configure or something similar).
13274 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
13275 in the source file (foo.c) like this:
13277 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
13278 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
13280 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
13281 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
13283 OPENSSL_DECLARE_GLOBAL(int,foo);
13284 #define foo OPENSSL_GLOBAL_REF(foo)
13285 OPENSSL_DECLARE_GLOBAL(double,bar);
13286 #define bar OPENSSL_GLOBAL_REF(bar)
13288 The #defines are very important, and therefore so is including the
13289 header file everywhere where the defined globals are used.
13291 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
13292 of ASN.1 items, but that structure is a bit different.
13294 The largest change is in util/mkdef.pl which has been enhanced with
13295 better and easier to understand logic to choose which symbols should
13296 go into the Windows .def files as well as a number of fixes and code
13297 cleanup (among others, algorithm keywords are now sorted
13298 lexicographically to avoid constant rewrites).
13302 * In BN_div() keep a copy of the sign of 'num' before writing the
13303 result to 'rm' because if rm==num the value will be overwritten
13304 and produce the wrong result if 'num' is negative: this caused
13305 problems with BN_mod() and BN_nnmod().
13309 * Function OCSP_request_verify(). This checks the signature on an
13310 OCSP request and verifies the signer certificate. The signer
13311 certificate is just checked for a generic purpose and OCSP request
13316 * Add OCSP_check_validity() function to check the validity of OCSP
13317 responses. OCSP responses are prepared in real time and may only
13318 be a few seconds old. Simply checking that the current time lies
13319 between thisUpdate and nextUpdate max reject otherwise valid responses
13320 caused by either OCSP responder or client clock inaccuracy. Instead
13321 we allow thisUpdate and nextUpdate to fall within a certain period of
13322 the current time. The age of the response can also optionally be
13323 checked. Two new options -validity_period and -status_age added to
13328 * If signature or public key algorithm is unrecognized print out its
13329 OID rather that just UNKNOWN.
13333 * Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
13334 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
13335 ID to be generated from the issuer certificate alone which can then be
13336 passed to OCSP_id_issuer_cmp().
13340 * New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
13341 ASN1 modules to export functions returning ASN1_ITEM pointers
13342 instead of the ASN1_ITEM structures themselves. This adds several
13343 new macros which allow the underlying ASN1 function/structure to
13344 be accessed transparently. As a result code should not use ASN1_ITEM
13345 references directly (such as &X509_it) but instead use the relevant
13346 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
13347 use of the new ASN1 code on platforms where exporting structures
13348 is problematical (for example in shared libraries) but exporting
13349 functions returning pointers to structures is not.
13353 * Add support for overriding the generation of SSL/TLS session IDs.
13354 These callbacks can be registered either in an SSL_CTX or per SSL.
13355 The purpose of this is to allow applications to control, if they wish,
13356 the arbitrary values chosen for use as session IDs, particularly as it
13357 can be useful for session caching in multiple-server environments. A
13358 command-line switch for testing this (and any client code that wishes
13359 to use such a feature) has been added to "s_server".
13361 *Geoff Thorpe, Lutz Jaenicke*
13363 * Modify mkdef.pl to recognise and parse preprocessor conditionals
13364 of the form `#if defined(...) || defined(...) || ...` and
13365 `#if !defined(...) && !defined(...) && ...`. This also avoids
13366 the growing number of special cases it was previously handling.
13370 * Make all configuration macros available for application by making
13371 sure they are available in opensslconf.h, by giving them names starting
13372 with `OPENSSL_` to avoid conflicts with other packages and by making
13373 sure e_os2.h will cover all platform-specific cases together with
13375 Additionally, it is now possible to define configuration/platform-
13376 specific names (called "system identities"). In the C code, these
13377 are prefixed with `OPENSSL_SYSNAME_`. e_os2.h will create another
13378 macro with the name beginning with `OPENSSL_SYS_`, which is determined
13379 from `OPENSSL_SYSNAME_*` or compiler-specific macros depending on
13384 * New option -set_serial to 'req' and 'x509' this allows the serial
13385 number to use to be specified on the command line. Previously self
13386 signed certificates were hard coded with serial number 0 and the
13387 CA options of 'x509' had to use a serial number in a file which was
13392 * New options to 'ca' utility to support V2 CRL entry extensions.
13393 Currently CRL reason, invalidity date and hold instruction are
13394 supported. Add new CRL extensions to V3 code and some new objects.
13398 * New function EVP_CIPHER_CTX_set_padding() this is used to
13399 disable standard block padding (aka PKCS#5 padding) in the EVP
13400 API, which was previously mandatory. This means that the data is
13401 not padded in any way and so the total length much be a multiple
13402 of the block size, otherwise an error occurs.
13406 * Initial (incomplete) OCSP SSL support.
13410 * New function OCSP_parse_url(). This splits up a URL into its host,
13411 port and path components: primarily to parse OCSP URLs. New -url
13412 option to ocsp utility.
13416 * New nonce behavior. The return value of OCSP_check_nonce() now
13417 reflects the various checks performed. Applications can decide
13418 whether to tolerate certain situations such as an absent nonce
13419 in a response when one was present in a request: the ocsp application
13420 just prints out a warning. New function OCSP_add1_basic_nonce()
13421 this is to allow responders to include a nonce in a response even if
13422 the request is nonce-less.
13426 * Disable stdin buffering in `load_cert()` (`apps/apps.c`) so that no certs are
13427 skipped when using openssl x509 multiple times on a single input file,
13428 e.g. `(openssl x509 -out cert1; openssl x509 -out cert2) <certs`.
13432 * Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
13433 set string type: to handle setting ASN1_TIME structures. Fix ca
13434 utility to correctly initialize revocation date of CRLs.
13438 * New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
13439 the clients preferred ciphersuites and rather use its own preferences.
13440 Should help to work around M$ SGC (Server Gated Cryptography) bug in
13441 Internet Explorer by ensuring unchanged hash method during stepup.
13442 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
13446 * Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
13447 to aes and add a new 'exist' option to print out symbols that don't
13452 * Additional options to ocsp utility to allow flags to be set and
13453 additional certificates supplied.
13457 * Add the option -VAfile to 'openssl ocsp', so the user can give the
13458 OCSP client a number of certificate to only verify the response
13463 * Update Rijndael code to version 3.0 and change EVP AES ciphers to
13464 handle the new API. Currently only ECB, CBC modes supported. Add new
13467 Add TLS AES ciphersuites as described in RFC3268, "Advanced
13468 Encryption Standard (AES) Ciphersuites for Transport Layer
13469 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were
13470 not enabled by default and were not part of the "ALL" ciphersuite
13471 alias because they were not yet official; they could be
13472 explicitly requested by specifying the "AESdraft" ciphersuite
13473 group alias. In the final release of OpenSSL 0.9.7, the group
13474 alias is called "AES" and is part of "ALL".)
13476 *Ben Laurie, Steve Henson, Bodo Moeller*
13478 * New function OCSP_copy_nonce() to copy nonce value (if present) from
13479 request to response.
13483 * Functions for OCSP responders. OCSP_request_onereq_count(),
13484 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
13485 extract information from a certificate request. OCSP_response_create()
13486 creates a response and optionally adds a basic response structure.
13487 OCSP_basic_add1_status() adds a complete single response to a basic
13488 response and returns the OCSP_SINGLERESP structure just added (to allow
13489 extensions to be included for example). OCSP_basic_add1_cert() adds a
13490 certificate to a basic response and OCSP_basic_sign() signs a basic
13491 response with various flags. New helper functions ASN1_TIME_check()
13492 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
13493 (converts ASN1_TIME to GeneralizedTime).
13497 * Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
13498 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
13499 structure from a certificate. X509_pubkey_digest() digests the public_key
13500 contents: this is used in various key identifiers.
13504 * Make sk_sort() tolerate a NULL argument.
13506 *Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>*
13508 * New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
13509 passed by the function are trusted implicitly. If any of them signed the
13510 response then it is assumed to be valid and is not verified.
13514 * In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
13515 to data. This was previously part of the PKCS7 ASN1 code. This
13516 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
13517 *Steve Henson, reported by Kenneth R. Robinette
13518 <support@securenetterm.com>*
13520 * Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
13521 routines: without these tracing memory leaks is very painful.
13522 Fix leaks in PKCS12 and PKCS7 routines.
13526 * Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
13527 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
13528 effectively meant GeneralizedTime would never be used. Now it
13529 is initialised to -1 but X509_time_adj() now has to check the value
13530 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
13531 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
13532 *Steve Henson, reported by Kenneth R. Robinette
13533 <support@securenetterm.com>*
13535 * Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
13536 result in a zero length in the ASN1_INTEGER structure which was
13537 not consistent with the structure when d2i_ASN1_INTEGER() was used
13538 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
13539 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
13540 where it did not print out a minus for negative ASN1_INTEGER.
13544 * Add summary printout to ocsp utility. The various functions which
13545 convert status values to strings have been renamed to:
13546 OCSP_response_status_str(), OCSP_cert_status_str() and
13547 OCSP_crl_reason_str() and are no longer static. New options
13548 to verify nonce values and to disable verification. OCSP response
13549 printout format cleaned up.
13553 * Add additional OCSP certificate checks. These are those specified
13554 in RFC2560. This consists of two separate checks: the CA of the
13555 certificate being checked must either be the OCSP signer certificate
13556 or the issuer of the OCSP signer certificate. In the latter case the
13557 OCSP signer certificate must contain the OCSP signing extended key
13558 usage. This check is performed by attempting to match the OCSP
13559 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
13560 in the OCSP_CERTID structures of the response.
13564 * Initial OCSP certificate verification added to OCSP_basic_verify()
13565 and related routines. This uses the standard OpenSSL certificate
13566 verify routines to perform initial checks (just CA validity) and
13567 to obtain the certificate chain. Then additional checks will be
13568 performed on the chain. Currently the root CA is checked to see
13569 if it is explicitly trusted for OCSP signing. This is used to set
13570 a root CA as a global signing root: that is any certificate that
13571 chains to that CA is an acceptable OCSP signing certificate.
13575 * New '-extfile ...' option to 'openssl ca' for reading X.509v3
13576 extensions from a separate configuration file.
13577 As when reading extensions from the main configuration file,
13578 the '-extensions ...' option may be used for specifying the
13581 *Massimiliano Pala <madwolf@comune.modena.it>*
13583 * New OCSP utility. Allows OCSP requests to be generated or
13584 read. The request can be sent to a responder and the output
13585 parsed, outputted or printed in text form. Not complete yet:
13586 still needs to check the OCSP response validity.
13590 * New subcommands for 'openssl ca':
13591 `openssl ca -status <serial>` prints the status of the cert with
13592 the given serial number (according to the index file).
13593 `openssl ca -updatedb` updates the expiry status of certificates
13596 *Massimiliano Pala <madwolf@comune.modena.it>*
13598 * New '-newreq-nodes' command option to CA.pl. This is like
13599 '-newreq', but calls 'openssl req' with the '-nodes' option
13600 so that the resulting key is not encrypted.
13602 *Damien Miller <djm@mindrot.org>*
13604 * New configuration for the GNU Hurd.
13606 *Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte*
13608 * Initial code to implement OCSP basic response verify. This
13609 is currently incomplete. Currently just finds the signer's
13610 certificate and verifies the signature on the response.
13614 * New SSLeay_version code SSLEAY_DIR to determine the compiled-in
13615 value of OPENSSLDIR. This is available via the new '-d' option
13616 to 'openssl version', and is also included in 'openssl version -a'.
13620 * Allowing defining memory allocation callbacks that will be given
13621 file name and line number information in additional arguments
13622 (a `const char*` and an int). The basic functionality remains, as
13623 well as the original possibility to just replace malloc(),
13624 realloc() and free() by functions that do not know about these
13625 additional arguments. To register and find out the current
13626 settings for extended allocation functions, the following
13627 functions are provided:
13629 CRYPTO_set_mem_ex_functions
13630 CRYPTO_set_locked_mem_ex_functions
13631 CRYPTO_get_mem_ex_functions
13632 CRYPTO_get_locked_mem_ex_functions
13634 These work the same way as CRYPTO_set_mem_functions and friends.
13635 `CRYPTO_get_[locked_]mem_functions` now writes 0 where such an
13636 extended allocation function is enabled.
13637 Similarly, `CRYPTO_get_[locked_]mem_ex_functions` writes 0 where
13638 a conventional allocation function is enabled.
13640 *Richard Levitte, Bodo Moeller*
13642 * Finish off removing the remaining LHASH function pointer casts.
13643 There should no longer be any prototype-casting required when using
13644 the LHASH abstraction, and any casts that remain are "bugs". See
13645 the callback types and macros at the head of lhash.h for details
13646 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
13650 * Add automatic query of EGD sockets in RAND_poll() for the unix variant.
13651 If /dev/[u]random devices are not available or do not return enough
13652 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
13654 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
13655 /etc/entropy will be queried once each in this sequence, querying stops
13656 when enough entropy was collected without querying more sockets.
13660 * Change the Unix RAND_poll() variant to be able to poll several
13661 random devices, as specified by DEVRANDOM, until a sufficient amount
13662 of data has been collected. We spend at most 10 ms on each file
13663 (select timeout) and read in non-blocking mode. DEVRANDOM now
13664 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
13665 (previously it was just the string "/dev/urandom"), so on typical
13666 platforms the 10 ms delay will never occur.
13667 Also separate out the Unix variant to its own file, rand_unix.c.
13668 For VMS, there's a currently-empty rand_vms.c.
13672 * Move OCSP client related routines to ocsp_cl.c. These
13673 provide utility functions which an application needing
13674 to issue a request to an OCSP responder and analyse the
13675 response will typically need: as opposed to those which an
13676 OCSP responder itself would need which will be added later.
13678 OCSP_request_sign() signs an OCSP request with an API similar
13679 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
13680 response. OCSP_response_get1_basic() extracts basic response
13681 from response. OCSP_resp_find_status(): finds and extracts status
13682 information from an OCSP_CERTID structure (which will be created
13683 when the request structure is built). These are built from lower
13684 level functions which work on OCSP_SINGLERESP structures but
13685 won't normally be used unless the application wishes to examine
13686 extensions in the OCSP response for example.
13688 Replace nonce routines with a pair of functions.
13689 OCSP_request_add1_nonce() adds a nonce value and optionally
13690 generates a random value. OCSP_check_nonce() checks the
13691 validity of the nonce in an OCSP response.
13695 * Change function OCSP_request_add() to OCSP_request_add0_id().
13696 This doesn't copy the supplied OCSP_CERTID and avoids the
13697 need to free up the newly created id. Change return type
13698 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
13699 This can then be used to add extensions to the request.
13700 Deleted OCSP_request_new(), since most of its functionality
13701 is now in OCSP_REQUEST_new() (and the case insensitive name
13702 clash) apart from the ability to set the request name which
13703 will be added elsewhere.
13707 * Update OCSP API. Remove obsolete extensions argument from
13708 various functions. Extensions are now handled using the new
13709 OCSP extension code. New simple OCSP HTTP function which
13710 can be used to send requests and parse the response.
13714 * Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
13715 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
13716 uses the special reorder version of SET OF to sort the attributes
13717 and reorder them to match the encoded order. This resolves a long
13718 standing problem: a verify on a PKCS7 structure just after signing
13719 it used to fail because the attribute order did not match the
13720 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
13721 it uses the received order. This is necessary to tolerate some broken
13722 software that does not order SET OF. This is handled by encoding
13723 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
13724 to produce the required SET OF.
13728 * Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
13729 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
13730 files to get correct declarations of the ASN.1 item variables.
13734 * Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
13735 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
13736 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
13737 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
13738 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
13739 ASN1_ITEM and no wrapper functions.
13743 * New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
13744 replace the old function pointer based I/O routines. Change most of
13745 the `*_d2i_bio()` and `*_d2i_fp()` functions to use these.
13749 * Enhance mkdef.pl to be more accepting about spacing in C preprocessor
13750 lines, recognize more "algorithms" that can be deselected, and make
13751 it complain about algorithm deselection that isn't recognised.
13755 * New ASN1 functions to handle dup, sign, verify, digest, pack and
13756 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
13757 to use new functions. Add NO_ASN1_OLD which can be set to remove
13758 some old style ASN1 functions: this can be used to determine if old
13759 code will still work when these eventually go away.
13763 * New extension functions for OCSP structures, these follow the
13764 same conventions as certificates and CRLs.
13768 * New function X509V3_add1_i2d(). This automatically encodes and
13769 adds an extension. Its behaviour can be customised with various
13770 flags to append, replace or delete. Various wrappers added for
13771 certificates and CRLs.
13775 * Fix to avoid calling the underlying ASN1 print routine when
13776 an extension cannot be parsed. Correct a typo in the
13777 OCSP_SERVICELOC extension. Tidy up print OCSP format.
13781 * Make mkdef.pl parse some of the ASN1 macros and add appropriate
13782 entries for variables.
13786 * Add functionality to `apps/openssl.c` for detecting locking
13787 problems: As the program is single-threaded, all we have
13788 to do is register a locking callback using an array for
13789 storing which locks are currently held by the program.
13793 * Use a lock around the call to CRYPTO_get_ex_new_index() in
13794 SSL_get_ex_data_X509_STORE_idx(), which is used in
13795 ssl_verify_cert_chain() and thus can be called at any time
13796 during TLS/SSL handshakes so that thread-safety is essential.
13797 Unfortunately, the ex_data design is not at all suited
13798 for multi-threaded use, so it probably should be abolished.
13802 * Added Broadcom "ubsec" ENGINE to OpenSSL.
13804 *Broadcom, tweaked and integrated by Geoff Thorpe*
13806 * Move common extension printing code to new function
13807 X509V3_print_extensions(). Reorganise OCSP print routines and
13808 implement some needed OCSP ASN1 functions. Add OCSP extensions.
13812 * New function X509_signature_print() to remove duplication in some
13817 * Add a special meaning when SET OF and SEQUENCE OF flags are both
13818 set (this was treated exactly the same as SET OF previously). This
13819 is used to reorder the STACK representing the structure to match the
13820 encoding. This will be used to get round a problem where a PKCS7
13821 structure which was signed could not be verified because the STACK
13822 order did not reflect the encoded order.
13826 * Reimplement the OCSP ASN1 module using the new code.
13830 * Update the X509V3 code to permit the use of an ASN1_ITEM structure
13831 for its ASN1 operations. The old style function pointers still exist
13832 for now but they will eventually go away.
13836 * Merge in replacement ASN1 code from the ASN1 branch. This almost
13837 completely replaces the old ASN1 functionality with a table driven
13838 encoder and decoder which interprets an ASN1_ITEM structure describing
13839 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
13840 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
13841 has also been converted to the new form.
13845 * Change BN_mod_exp_recp so that negative moduli are tolerated
13846 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
13847 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
13848 for negative moduli.
13852 * Fix BN_uadd and BN_usub: Always return non-negative results instead
13853 of not touching the result's sign bit.
13857 * BN_div bugfix: If the result is 0, the sign (res->neg) must not be
13862 * Changed the LHASH code to use prototypes for callbacks, and created
13863 macros to declare and implement thin (optionally static) functions
13864 that provide type-safety and avoid function pointer casting for the
13865 type-specific callbacks.
13869 * Added Kerberos Cipher Suites to be used with TLS, as written in
13871 *Veers Staats <staatsvr@asc.hpc.mil>,
13872 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte*
13874 * Reformat the FAQ so the different questions and answers can be divided
13875 in sections depending on the subject.
13879 * Have the zlib compression code load ZLIB.DLL dynamically under
13884 * New function BN_mod_sqrt for computing square roots modulo a prime
13885 (using the probabilistic Tonelli-Shanks algorithm unless
13886 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
13887 be handled deterministically).
13889 *Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller*
13891 * Make BN_mod_inverse faster by explicitly handling small quotients
13892 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
13893 512 bits], about 30% for larger ones [1024 or 2048 bits].)
13897 * New function BN_kronecker.
13901 * Fix BN_gcd so that it works on negative inputs; the result is
13902 positive unless both parameters are zero.
13903 Previously something reasonably close to an infinite loop was
13904 possible because numbers could be growing instead of shrinking
13905 in the implementation of Euclid's algorithm.
13909 * Fix BN_is_word() and BN_is_one() macros to take into account the
13910 sign of the number in question.
13912 Fix BN_is_word(a,w) to work correctly for w == 0.
13914 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
13915 because its test if the absolute value of 'a' equals 'w'.
13916 Note that BN_abs_is_word does *not* handle w == 0 reliably;
13917 it exists mostly for use in the implementations of BN_is_zero(),
13918 BN_is_one(), and BN_is_word().
13922 * New function BN_swap.
13926 * Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
13927 the exponentiation functions are more likely to produce reasonable
13928 results on negative inputs.
13932 * Change BN_mod_mul so that the result is always non-negative.
13933 Previously, it could be negative if one of the factors was negative;
13934 I don't think anyone really wanted that behaviour.
13938 * Move `BN_mod_...` functions into new file `crypto/bn/bn_mod.c`
13939 (except for exponentiation, which stays in `crypto/bn/bn_exp.c`,
13940 and `BN_mod_mul_reciprocal`, which stays in `crypto/bn/bn_recp.c`)
13941 and add new functions:
13950 BN_mod_lshift1_quick
13952 BN_mod_lshift_quick
13954 These functions always generate non-negative results.
13956 `BN_nnmod` otherwise is `like BN_mod` (if `BN_mod` computes a remainder `r`
13957 such that `|m| < r < 0`, `BN_nnmod` will output `rem + |m|` instead).
13959 `BN_mod_XXX_quick(r, a, [b,] m)` generates the same result as
13960 `BN_mod_XXX(r, a, [b,] m, ctx)`, but requires that `a` [and `b`]
13961 be reduced modulo `m`.
13963 *Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller*
13966 The following entry accidentally appeared in the CHANGES file
13967 distributed with OpenSSL 0.9.7. The modifications described in
13968 it do *not* apply to OpenSSL 0.9.7.
13970 * Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
13971 was actually never needed) and in BN_mul(). The removal in BN_mul()
13972 required a small change in bn_mul_part_recursive() and the addition
13973 of the functions bn_cmp_part_words(), bn_sub_part_words() and
13974 bn_add_part_words(), which do the same thing as bn_cmp_words(),
13975 bn_sub_words() and bn_add_words() except they take arrays with
13981 * In 'openssl passwd', verify passwords read from the terminal
13982 unless the '-salt' option is used (which usually means that
13983 verification would just waste user's time since the resulting
13984 hash is going to be compared with some given password hash)
13985 or the new '-noverify' option is used.
13987 This is an incompatible change, but it does not affect
13988 non-interactive use of 'openssl passwd' (passwords on the command
13989 line, '-stdin' option, '-in ...' option) and thus should not
13990 cause any problems.
13994 * Remove all references to RSAref, since there's no more need for it.
13998 * Make DSO load along a path given through an environment variable
13999 (SHLIB_PATH) with shl_load().
14003 * Constify the ENGINE code as a result of BIGNUM constification.
14004 Also constify the RSA code and most things related to it. In a
14005 few places, most notable in the depth of the ASN.1 code, ugly
14006 casts back to non-const were required (to be solved at a later
14011 * Make it so the openssl application has all engines loaded by default.
14015 * Constify the BIGNUM routines a little more.
14019 * Add the following functions:
14021 ENGINE_load_cswift()
14023 ENGINE_load_atalla()
14024 ENGINE_load_nuron()
14025 ENGINE_load_builtin_engines()
14027 That way, an application can itself choose if external engines that
14028 are built-in in OpenSSL shall ever be used or not. The benefit is
14029 that applications won't have to be linked with libdl or other dso
14030 libraries unless it's really needed.
14032 Changed 'openssl engine' to load all engines on demand.
14033 Changed the engine header files to avoid the duplication of some
14034 declarations (they differed!).
14038 * 'openssl engine' can now list capabilities.
14042 * Better error reporting in 'openssl engine'.
14046 * Never call load_dh_param(NULL) in s_server.
14050 * Add engine application. It can currently list engines by name and
14051 identity, and test if they are actually available.
14055 * Improve RPM specification file by forcing symbolic linking and making
14056 sure the installed documentation is also owned by root.root.
14058 *Damien Miller <djm@mindrot.org>*
14060 * Give the OpenSSL applications more possibilities to make use of
14061 keys (public as well as private) handled by engines.
14065 * Add OCSP code that comes from CertCo.
14069 * Add VMS support for the Rijndael code.
14073 * Added untested support for Nuron crypto accelerator.
14077 * Add support for external cryptographic devices. This code was
14078 previously distributed separately as the "engine" branch.
14080 *Geoff Thorpe, Richard Levitte*
14082 * Rework the filename-translation in the DSO code. It is now possible to
14083 have far greater control over how a "name" is turned into a filename
14084 depending on the operating environment and any oddities about the
14085 different shared library filenames on each system.
14089 * Support threads on FreeBSD-elf in Configure.
14093 * Fix for SHA1 assembly problem with MASM: it produces
14094 warnings about corrupt line number information when assembling
14095 with debugging information. This is caused by the overlapping
14098 *Bernd Matthes <mainbug@celocom.de>, Steve Henson*
14101 NCONF_get_number() has no error checking at all. As a replacement,
14102 NCONF_get_number_e() is defined (`_e` for "error checking") and is
14103 promoted strongly. The old NCONF_get_number is kept around for
14104 binary backward compatibility.
14105 Make it possible for methods to load from something other than a BIO,
14106 by providing a function pointer that is given a name instead of a BIO.
14107 For example, this could be used to load configuration data from an
14112 * Fix for non blocking accept BIOs. Added new I/O special reason
14113 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
14114 with non blocking I/O was not possible because no retry code was
14115 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
14120 * Added the beginnings of Rijndael support.
14124 * Fix for bug in DirectoryString mask setting. Add support for
14125 X509_NAME_print_ex() in 'req' and X509_print_ex() function
14126 to allow certificate printing to more controllable, additional
14127 'certopt' option to 'x509' to allow new printing options to be
14132 * Clean old EAY MD5 hack from e_os.h.
14136 ### Changes between 0.9.6l and 0.9.6m [17 Mar 2004]
14138 * Fix null-pointer assignment in do_change_cipher_spec() revealed
14139 by using the Codenomicon TLS Test Tool ([CVE-2004-0079])
14141 *Joe Orton, Steve Henson*
14143 ### Changes between 0.9.6k and 0.9.6l [04 Nov 2003]
14145 * Fix additional bug revealed by the NISCC test suite:
14147 Stop bug triggering large recursion when presented with
14148 certain ASN.1 tags ([CVE-2003-0851])
14152 ### Changes between 0.9.6j and 0.9.6k [30 Sep 2003]
14154 * Fix various bugs revealed by running the NISCC test suite:
14156 Stop out of bounds reads in the ASN1 code when presented with
14157 invalid tags (CVE-2003-0543 and CVE-2003-0544).
14159 If verify callback ignores invalid public key errors don't try to check
14160 certificate signature with the NULL public key.
14164 * In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
14165 if the server requested one: as stated in TLS 1.0 and SSL 3.0
14170 * In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
14171 extra data after the compression methods not only for TLS 1.0
14172 but also for SSL 3.0 (as required by the specification).
14174 *Bodo Moeller; problem pointed out by Matthias Loepfe*
14176 * Change X509_certificate_type() to mark the key as exported/exportable
14177 when it's 512 *bits* long, not 512 bytes.
14181 ### Changes between 0.9.6i and 0.9.6j [10 Apr 2003]
14183 * Countermeasure against the Klima-Pokorny-Rosa extension of
14184 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
14185 a protocol version number mismatch like a decryption error
14186 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
14190 * Turn on RSA blinding by default in the default implementation
14191 to avoid a timing attack. Applications that don't want it can call
14192 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
14193 They would be ill-advised to do so in most cases.
14195 *Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller*
14197 * Change RSA blinding code so that it works when the PRNG is not
14198 seeded (in this case, the secret RSA exponent is abused as
14199 an unpredictable seed -- if it is not unpredictable, there
14200 is no point in blinding anyway). Make RSA blinding thread-safe
14201 by remembering the creator's thread ID in rsa->blinding and
14202 having all other threads use local one-time blinding factors
14203 (this requires more computation than sharing rsa->blinding, but
14204 avoids excessive locking; and if an RSA object is not shared
14205 between threads, blinding will still be very fast).
14209 ### Changes between 0.9.6h and 0.9.6i [19 Feb 2003]
14211 * In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
14212 via timing by performing a MAC computation even if incorrect
14213 block cipher padding has been found. This is a countermeasure
14214 against active attacks where the attacker has to distinguish
14215 between bad padding and a MAC verification error. ([CVE-2003-0078])
14217 *Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
14218 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
14219 Martin Vuagnoux (EPFL, Ilion)*
14221 ### Changes between 0.9.6g and 0.9.6h [5 Dec 2002]
14223 * New function OPENSSL_cleanse(), which is used to cleanse a section of
14224 memory from its contents. This is done with a counter that will
14225 place alternating values in each byte. This can be used to solve
14226 two issues: 1) the removal of calls to memset() by highly optimizing
14227 compilers, and 2) cleansing with other values than 0, since those can
14228 be read through on certain media, for example a swap space on disk.
14232 * Bugfix: client side session caching did not work with external caching,
14233 because the session->cipher setting was not restored when reloading
14234 from the external cache. This problem was masked, when
14235 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set.
14236 (Found by Steve Haslam <steve@araqnid.ddts.net>.)
14240 * Fix client_certificate (ssl/s2_clnt.c): The permissible total
14241 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33.
14243 *Zeev Lieber <zeev-l@yahoo.com>*
14245 * Undo an undocumented change introduced in 0.9.6e which caused
14246 repeated calls to OpenSSL_add_all_ciphers() and
14247 OpenSSL_add_all_digests() to be ignored, even after calling
14252 * Change the default configuration reader to deal with last line not
14253 being properly terminated.
14257 * Change X509_NAME_cmp() so it applies the special rules on handling
14258 DN values that are of type PrintableString, as well as RDNs of type
14259 emailAddress where the value has the type ia5String.
14261 *stefank@valicert.com via Richard Levitte*
14263 * Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half
14264 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently
14265 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be
14266 the bitwise-OR of the two for use by the majority of applications
14267 wanting this behaviour, and update the docs. The documented
14268 behaviour and actual behaviour were inconsistent and had been
14269 changing anyway, so this is more a bug-fix than a behavioural
14272 *Geoff Thorpe, diagnosed by Nadav Har'El*
14274 * Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c
14275 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes).
14279 * Fix initialization code race conditions in
14280 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(),
14281 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(),
14282 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(),
14283 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(),
14284 ssl2_get_cipher_by_char(),
14285 ssl3_get_cipher_by_char().
14287 *Patrick McCormick <patrick@tellme.com>, Bodo Moeller*
14289 * Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after
14290 the cached sessions are flushed, as the remove_cb() might use ex_data
14291 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com>
14292 (see [openssl.org #212]).
14294 *Geoff Thorpe, Lutz Jaenicke*
14296 * Fix typo in OBJ_txt2obj which incorrectly passed the content
14297 length, instead of the encoding length to d2i_ASN1_OBJECT.
14301 ### Changes between 0.9.6f and 0.9.6g [9 Aug 2002]
14303 * [In 0.9.6g-engine release:]
14304 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use `_stdcall`).
14306 *Lynn Gazis <lgazis@rainbow.com>*
14308 ### Changes between 0.9.6e and 0.9.6f [8 Aug 2002]
14310 * Fix ASN1 checks. Check for overflow by comparing with LONG_MAX
14311 and get fix the header length calculation.
14312 *Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>,
14313 Alon Kantor <alonk@checkpoint.com> (and others), Steve Henson*
14315 * Use proper error handling instead of 'assertions' in buffer
14316 overflow checks added in 0.9.6e. This prevents DoS (the
14317 assertions could call abort()).
14319 *Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller*
14321 ### Changes between 0.9.6d and 0.9.6e [30 Jul 2002]
14323 * Add various sanity checks to asn1_get_length() to reject
14324 the ASN1 length bytes if they exceed sizeof(long), will appear
14325 negative or the content length exceeds the length of the
14328 *Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>*
14330 * Fix cipher selection routines: ciphers without encryption had no flags
14331 for the cipher strength set and where therefore not handled correctly
14332 by the selection routines (PR #130).
14336 * Fix EVP_dsa_sha macro.
14341 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS
14342 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure
14343 that was added in OpenSSL 0.9.6d.
14345 As the countermeasure turned out to be incompatible with some
14346 broken SSL implementations, the new option is part of SSL_OP_ALL.
14347 SSL_OP_ALL is usually employed when compatibility with weird SSL
14348 implementations is desired (e.g. '-bugs' option to 's_client' and
14349 's_server'), so the new option is automatically set in many
14354 * Changes in security patch:
14356 Changes marked "(CHATS)" were sponsored by the Defense Advanced
14357 Research Projects Agency (DARPA) and Air Force Research Laboratory,
14358 Air Force Materiel Command, USAF, under agreement number
14361 * Add various sanity checks to asn1_get_length() to reject
14362 the ASN1 length bytes if they exceed sizeof(long), will appear
14363 negative or the content length exceeds the length of the
14364 supplied buffer. ([CVE-2002-0659])
14366 *Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>*
14368 * Assertions for various potential buffer overflows, not known to
14369 happen in practice.
14371 *Ben Laurie (CHATS)*
14373 * Various temporary buffers to hold ASCII versions of integers were
14374 too small for 64 bit platforms. ([CVE-2002-0655])
14375 *Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>*
14377 * Remote buffer overflow in SSL3 protocol - an attacker could
14378 supply an oversized session ID to a client. ([CVE-2002-0656])
14380 *Ben Laurie (CHATS)*
14382 * Remote buffer overflow in SSL2 protocol - an attacker could
14383 supply an oversized client master key. ([CVE-2002-0656])
14385 *Ben Laurie (CHATS)*
14387 ### Changes between 0.9.6c and 0.9.6d [9 May 2002]
14389 * Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
14390 encoded as NULL) with id-dsa-with-sha1.
14392 *Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller*
14394 * Check various `X509_...()` return values in `apps/req.c`.
14396 *Nils Larsch <nla@trustcenter.de>*
14398 * Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
14399 an end-of-file condition would erroneously be flagged, when the CRLF
14400 was just at the end of a processed block. The bug was discovered when
14401 processing data through a buffering memory BIO handing the data to a
14402 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
14403 <ptsekov@syntrex.com> and Nedelcho Stanev.
14407 * Implement a countermeasure against a vulnerability recently found
14408 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
14409 before application data chunks to avoid the use of known IVs
14410 with data potentially chosen by the attacker.
14414 * Fix length checks in ssl3_get_client_hello().
14418 * TLS/SSL library bugfix: use s->s3->in_read_app_data differently
14419 to prevent ssl3_read_internal() from incorrectly assuming that
14420 ssl3_read_bytes() found application data while handshake
14421 processing was enabled when in fact s->s3->in_read_app_data was
14422 merely automatically cleared during the initial handshake.
14424 *Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>*
14426 * Fix object definitions for Private and Enterprise: they were not
14427 recognized in their shortname (=lowercase) representation. Extend
14428 obj_dat.pl to issue an error when using undefined keywords instead
14429 of silently ignoring the problem (Svenning Sorensen
14430 <sss@sss.dnsalias.net>).
14434 * Fix DH_generate_parameters() so that it works for 'non-standard'
14435 generators, i.e. generators other than 2 and 5. (Previously, the
14436 code did not properly initialise the 'add' and 'rem' values to
14437 BN_generate_prime().)
14439 In the new general case, we do not insist that 'generator' is
14440 actually a primitive root: This requirement is rather pointless;
14441 a generator of the order-q subgroup is just as good, if not
14446 * Map new X509 verification errors to alerts. Discovered and submitted by
14447 Tom Wu <tom@arcot.com>.
14451 * Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
14452 returning non-zero before the data has been completely received
14453 when using non-blocking I/O.
14455 *Bodo Moeller; problem pointed out by John Hughes*
14457 * Some of the ciphers missed the strength entry (SSL_LOW etc).
14459 *Ben Laurie, Lutz Jaenicke*
14461 * Fix bug in SSL_clear(): bad sessions were not removed (found by
14462 Yoram Zahavi <YoramZ@gilian.com>).
14466 * Add information about CygWin 1.3 and on, and preserve proper
14467 configuration for the versions before that.
14469 *Corinna Vinschen <vinschen@redhat.com> and Richard Levitte*
14471 * Make removal from session cache (SSL_CTX_remove_session()) more robust:
14472 check whether we deal with a copy of a session and do not delete from
14473 the cache in this case. Problem reported by "Izhar Shoshani Levi"
14474 <izhar@checkpoint.com>.
14478 * Do not store session data into the internal session cache, if it
14479 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
14480 flag is set). Proposed by Aslam <aslam@funk.com>.
14484 * Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
14489 * [In 0.9.6d-engine release:]
14490 Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
14492 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
14494 * Add the configuration target linux-s390x.
14496 *Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte*
14498 * The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
14499 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
14500 variable as an indication that a ClientHello message has been
14501 received. As the flag value will be lost between multiple
14502 invocations of ssl3_accept when using non-blocking I/O, the
14503 function may not be aware that a handshake has actually taken
14504 place, thus preventing a new session from being added to the
14507 To avoid this problem, we now set s->new_session to 2 instead of
14508 using a local variable.
14510 *Lutz Jaenicke, Bodo Moeller*
14512 * Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
14513 if the SSL_R_LENGTH_MISMATCH error is detected.
14515 *Geoff Thorpe, Bodo Moeller*
14517 * New 'shared_ldflag' column in Configure platform table.
14521 * Fix EVP_CIPHER_mode macro.
14523 *"Dan S. Camper" <dan@bti.net>*
14525 * Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
14526 type, we must throw them away by setting rr->length to 0.
14528 *D P Chang <dpc@qualys.com>*
14530 ### Changes between 0.9.6b and 0.9.6c [21 dec 2001]
14532 * Fix BN_rand_range bug pointed out by Dominikus Scherkl
14533 <Dominikus.Scherkl@biodata.com>. (The previous implementation
14534 worked incorrectly for those cases where range = `10..._2` and
14535 `3*range` is two bits longer than range.)
14539 * Only add signing time to PKCS7 structures if it is not already
14544 * Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
14545 OBJ_ld_ce should be OBJ_id_ce.
14546 Also some ip-pda OIDs in crypto/objects/objects.txt were
14547 incorrect (cf. RFC 3039).
14549 *Matt Cooper, Frederic Giudicelli, Bodo Moeller*
14551 * Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
14552 returns early because it has nothing to do.
14554 *Andy Schneider <andy.schneider@bjss.co.uk>*
14556 * [In 0.9.6c-engine release:]
14557 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
14559 *Andy Schneider <andy.schneider@bjss.co.uk>*
14561 * [In 0.9.6c-engine release:]
14562 Add support for Cryptographic Appliance's keyserver technology.
14563 (Use engine 'keyclient')
14565 *Cryptographic Appliances and Geoff Thorpe*
14567 * Add a configuration entry for OS/390 Unix. The C compiler 'c89'
14568 is called via tools/c89.sh because arguments have to be
14569 rearranged (all '-L' options must appear before the first object
14572 *Richard Shapiro <rshapiro@abinitio.com>*
14574 * [In 0.9.6c-engine release:]
14575 Add support for Broadcom crypto accelerator cards, backported
14578 *Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox*
14580 * [In 0.9.6c-engine release:]
14581 Add support for SureWare crypto accelerator cards from
14582 Baltimore Technologies. (Use engine 'sureware')
14584 *Baltimore Technologies and Mark Cox*
14586 * [In 0.9.6c-engine release:]
14587 Add support for crypto accelerator cards from Accelerated
14588 Encryption Processing, www.aep.ie. (Use engine 'aep')
14590 *AEP Inc. and Mark Cox*
14592 * Add a configuration entry for gcc on UnixWare.
14594 *Gary Benson <gbenson@redhat.com>*
14596 * Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
14597 messages are stored in a single piece (fixed-length part and
14598 variable-length part combined) and fix various bugs found on the way.
14602 * Disable caching in BIO_gethostbyname(), directly use gethostbyname()
14603 instead. BIO_gethostbyname() does not know what timeouts are
14604 appropriate, so entries would stay in cache even when they have
14606 *Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>*
14608 * Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
14609 faced with a pathologically small ClientHello fragment that does
14610 not contain client_version: Instead of aborting with an error,
14611 simply choose the highest available protocol version (i.e.,
14612 TLS 1.0 unless it is disabled). In practice, ClientHello
14613 messages are never sent like this, but this change gives us
14614 strictly correct behaviour at least for TLS.
14618 * Fix SSL handshake functions and SSL_clear() such that SSL_clear()
14619 never resets s->method to s->ctx->method when called from within
14620 one of the SSL handshake functions.
14622 *Bodo Moeller; problem pointed out by Niko Baric*
14624 * In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
14625 (sent using the client's version number) if client_version is
14626 smaller than the protocol version in use. Also change
14627 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
14628 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
14629 the client will at least see that alert.
14633 * Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
14638 * Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
14639 client receives HelloRequest while in a handshake.
14641 *Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>*
14643 * Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
14644 should end in 'break', not 'goto end' which circumvents various
14645 cleanups done in state SSL_ST_OK. But session related stuff
14646 must be disabled for SSL_ST_OK in the case that we just sent a
14649 Also avoid some overhead by not calling ssl_init_wbio_buffer()
14650 before just sending a HelloRequest.
14652 *Bodo Moeller, Eric Rescorla <ekr@rtfm.com>*
14654 * Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
14655 reveal whether illegal block cipher padding was found or a MAC
14656 verification error occurred. (Neither SSLerr() codes nor alerts
14657 are directly visible to potential attackers, but the information
14658 may leak via logfiles.)
14660 Similar changes are not required for the SSL 2.0 implementation
14661 because the number of padding bytes is sent in clear for SSL 2.0,
14662 and the extra bytes are just ignored. However ssl/s2_pkt.c
14663 failed to verify that the purported number of padding bytes is in
14668 * Add OpenUNIX-8 support including shared libraries
14669 (Boyd Lynn Gerber <gerberb@zenez.com>).
14673 * Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
14674 'wristwatch attack' using huge encoding parameters (cf.
14675 James H. Manger's CRYPTO 2001 paper). Note that the
14676 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
14677 encoding parameters and hence was not vulnerable.
14681 * BN_sqr() bug fix.
14683 *Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>*
14685 * Rabin-Miller test analyses assume uniformly distributed witnesses,
14686 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
14687 followed by modular reduction.
14689 *Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>*
14691 * Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
14692 equivalent based on BN_pseudo_rand() instead of BN_rand().
14696 * s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
14697 This function was broken, as the check for a new client hello message
14698 to handle SGC did not allow these large messages.
14699 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
14703 * Add alert descriptions for TLSv1 to `SSL_alert_desc_string[_long]()`.
14707 * Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
14708 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
14712 * Rework the configuration and shared library support for Tru64 Unix.
14713 The configuration part makes use of modern compiler features and
14714 still retains old compiler behavior for those that run older versions
14715 of the OS. The shared library support part includes a variant that
14716 uses the RPATH feature, and is available through the special
14717 configuration target "alpha-cc-rpath", which will never be selected
14720 *Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte*
14722 * In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
14723 with the same message size as in ssl3_get_certificate_request().
14724 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
14725 messages might inadvertently be reject as too long.
14727 *Petr Lampa <lampa@fee.vutbr.cz>*
14729 * Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
14733 * Modified SSL library such that the verify_callback that has been set
14734 specifically for an SSL object with SSL_set_verify() is actually being
14735 used. Before the change, a verify_callback set with this function was
14736 ignored and the verify_callback() set in the SSL_CTX at the time of
14737 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
14738 to allow the necessary settings.
14742 * Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
14743 explicitly to NULL, as at least on Solaris 8 this seems not always to be
14744 done automatically (in contradiction to the requirements of the C
14745 standard). This made problems when used from OpenSSH.
14749 * In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
14750 dh->length and always used
14752 BN_rand_range(priv_key, dh->p).
14754 BN_rand_range() is not necessary for Diffie-Hellman, and this
14755 specific range makes Diffie-Hellman unnecessarily inefficient if
14756 dh->length (recommended exponent length) is much smaller than the
14757 length of dh->p. We could use BN_rand_range() if the order of
14758 the subgroup was stored in the DH structure, but we only have
14763 BN_rand(priv_key, l, ...)
14765 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
14772 RSA_eay_public_encrypt
14773 RSA_eay_private_decrypt
14774 RSA_eay_private_encrypt (signing)
14775 RSA_eay_public_decrypt (signature verification)
14777 (default implementations for RSA_public_encrypt,
14778 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
14779 always reject numbers >= n.
14783 * In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
14784 to synchronize access to 'locking_thread'. This is necessary on
14785 systems where access to 'locking_thread' (an 'unsigned long'
14786 variable) is not atomic.
14790 * In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
14791 *before* setting the 'crypto_lock_rand' flag. The previous code had
14792 a race condition if 0 is a valid thread ID.
14794 *Travis Vitek <vitek@roguewave.com>*
14796 * Add support for shared libraries under Irix.
14798 *Albert Chin-A-Young <china@thewrittenword.com>*
14800 * Add configuration option to build on Linux on both big-endian and
14801 little-endian MIPS.
14803 *Ralf Baechle <ralf@uni-koblenz.de>*
14805 * Add the possibility to create shared libraries on HP-UX.
14809 ### Changes between 0.9.6a and 0.9.6b [9 Jul 2001]
14811 * Change ssleay_rand_bytes (crypto/rand/md_rand.c)
14812 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
14813 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
14814 PRNG state recovery was possible based on the output of
14815 one PRNG request appropriately sized to gain knowledge on
14816 'md' followed by enough consecutive 1-byte PRNG requests
14817 to traverse all of 'state'.
14819 1. When updating 'md_local' (the current thread's copy of 'md')
14820 during PRNG output generation, hash all of the previous
14821 'md_local' value, not just the half used for PRNG output.
14823 2. Make the number of bytes from 'state' included into the hash
14824 independent from the number of PRNG bytes requested.
14826 The first measure alone would be sufficient to avoid
14827 Markku-Juhani's attack. (Actually it had never occurred
14828 to me that the half of 'md_local' used for chaining was the
14829 half from which PRNG output bytes were taken -- I had always
14830 assumed that the secret half would be used.) The second
14831 measure makes sure that additional data from 'state' is never
14832 mixed into 'md_local' in small portions; this heuristically
14833 further strengthens the PRNG.
14837 * Fix crypto/bn/asm/mips3.s.
14841 * When only the key is given to "enc", the IV is undefined. Print out
14842 an error message in this case.
14846 * Handle special case when X509_NAME is empty in X509 printing routines.
14850 * In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
14851 positive and less than q.
14855 * Don't change `*pointer` in CRYPTO_add_lock() is add_lock_callback is
14856 used: it isn't thread safe and the add_lock_callback should handle
14859 *Paul Rose <Paul.Rose@bridge.com>*
14861 * Verify that incoming data obeys the block size in
14862 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
14868 *Ulf Möller, Bodo Möller*
14870 * The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
14871 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
14872 when fixing the server behaviour for backwards-compatible 'client
14873 hello' messages. (Note that the attack is impractical against
14874 SSL 3.0 and TLS 1.0 anyway because length and version checking
14875 means that the probability of guessing a valid ciphertext is
14876 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
14879 Before 0.9.5, the countermeasure (hide the error by generating a
14880 random 'decryption result') did not work properly because
14881 ERR_clear_error() was missing, meaning that SSL_get_error() would
14882 detect the supposedly ignored error.
14884 Both problems are now fixed.
14888 * In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
14889 (previously it was 1024).
14893 * Fix for compatibility mode trust settings: ignore trust settings
14894 unless some valid trust or reject settings are present.
14898 * Fix for blowfish EVP: its a variable length cipher.
14902 * Fix various bugs related to DSA S/MIME verification. Handle missing
14903 parameters in DSA public key structures and return an error in the
14904 DSA routines if parameters are absent.
14908 * In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
14909 in the current directory if neither $RANDFILE nor $HOME was set.
14910 RAND_file_name() in 0.9.6a returned NULL in this case. This has
14911 caused some confusion to Windows users who haven't defined $HOME.
14912 Thus RAND_file_name() is changed again: e_os.h can define a
14913 DEFAULT_HOME, which will be used if $HOME is not set.
14914 For Windows, we use "C:"; on other platforms, we still require
14915 environment variables.
14917 * Move 'if (!initialized) RAND_poll()' into regions protected by
14918 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
14919 having multiple threads call RAND_poll() concurrently.
14923 * In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
14924 combination of a flag and a thread ID variable.
14925 Otherwise while one thread is in ssleay_rand_bytes (which sets the
14926 flag), *other* threads can enter ssleay_add_bytes without obeying
14927 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
14928 that they do not hold after the first thread unsets add_do_not_lock).
14932 * Change bctest again: '-x' expressions are not available in all
14933 versions of 'test'.
14937 ### Changes between 0.9.6 and 0.9.6a [5 Apr 2001]
14939 * Fix a couple of memory leaks in PKCS7_dataDecode()
14941 *Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>*
14943 * Change Configure and Makefiles to provide EXE_EXT, which will contain
14944 the default extension for executables, if any. Also, make the perl
14945 scripts that use symlink() to test if it really exists and use "cp"
14946 if it doesn't. All this made OpenSSL compilable and installable in
14951 * Fix for asn1_GetSequence() for indefinite length constructed data.
14952 If SEQUENCE is length is indefinite just set c->slen to the total
14953 amount of data available.
14955 *Steve Henson, reported by shige@FreeBSD.org*
14957 *This change does not apply to 0.9.7.*
14959 * Change bctest to avoid here-documents inside command substitution
14960 (workaround for FreeBSD /bin/sh bug).
14961 For compatibility with Ultrix, avoid shell functions (introduced
14962 in the bctest version that searches along $PATH).
14966 * Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
14967 with des_encrypt() defined on some operating systems, like Solaris
14972 * Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
14973 On the Importance of Eliminating Errors in Cryptographic
14974 Computations, J. Cryptology 14 (2001) 2, 101-119,
14975 <http://theory.stanford.edu/~dabo/papers/faults.ps.gz>).
14979 * MIPS assembler BIGNUM division bug fix.
14983 * Disabled incorrect Alpha assembler code.
14987 * Fix PKCS#7 decode routines so they correctly update the length
14988 after reading an EOC for the EXPLICIT tag.
14992 *This change does not apply to 0.9.7.*
14994 * Fix bug in PKCS#12 key generation routines. This was triggered
14995 if a 3DES key was generated with a 0 initial byte. Include
14996 PKCS12_BROKEN_KEYGEN compilation option to retain the old
14997 (but broken) behaviour.
15001 * Enhance bctest to search for a working bc along $PATH and print
15004 *Tim Rice <tim@multitalents.net> via Richard Levitte*
15006 * Fix memory leaks in err.c: free err_data string if necessary;
15007 don't write to the wrong index in ERR_set_error_data.
15011 * Implement ssl23_peek (analogous to ssl23_read), which previously
15016 * Replace rdtsc with `_emit` statements for VC++ version 5.
15018 *Jeremy Cooper <jeremy@baymoo.org>*
15020 * Make it possible to reuse SSLv2 sessions.
15024 * In copy_email() check for >= 0 as a return value for
15025 X509_NAME_get_index_by_NID() since 0 is a valid index.
15027 *Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>*
15029 * Avoid coredump with unsupported or invalid public keys by checking if
15030 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
15031 PKCS7_verify() fails with non detached data.
15035 * Don't use getenv in library functions when run as setuid/setgid.
15036 New function OPENSSL_issetugid().
15040 * Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
15041 due to incorrect handling of multi-threading:
15043 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
15045 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
15047 3. Count how many times MemCheck_off() has been called so that
15048 nested use can be treated correctly. This also avoids
15049 inband-signalling in the previous code (which relied on the
15050 assumption that thread ID 0 is impossible).
15054 * Add "-rand" option also to s_client and s_server.
15058 * Fix CPU detection on Irix 6.x.
15059 *Kurt Hockenbury <khockenb@stevens-tech.edu> and
15060 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>*
15062 * Fix X509_NAME bug which produced incorrect encoding if X509_NAME
15067 *This change does not apply to 0.9.7.*
15069 * Use the cached encoding of an X509_NAME structure rather than
15070 copying it. This is apparently the reason for the libsafe "errors"
15071 but the code is actually correct.
15075 * Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
15076 Bleichenbacher's DSA attack.
15077 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
15078 to be set and top=0 forces the highest bit to be set; top=-1 is new
15079 and leaves the highest bit random.
15081 *Ulf Moeller, Bodo Moeller*
15083 * In the `NCONF_...`-based implementations for `CONF_...` queries
15084 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
15085 a temporary CONF structure with the data component set to NULL
15086 (which gives segmentation faults in lh_retrieve).
15087 Instead, use NULL for the CONF pointer in CONF_get_string and
15088 CONF_get_number (which may use environment variables) and directly
15089 return NULL from CONF_get_section.
15093 * Fix potential buffer overrun for EBCDIC.
15097 * Tolerate nonRepudiation as being valid for S/MIME signing and certSign
15098 keyUsage if basicConstraints absent for a CA.
15102 * Make SMIME_write_PKCS7() write mail header values with a format that
15103 is more generally accepted (no spaces before the semicolon), since
15104 some programs can't parse those values properly otherwise. Also make
15105 sure BIO's that break lines after each write do not create invalid
15110 * Make the CRL encoding routines work with empty SEQUENCE OF. The
15111 macros previously used would not encode an empty SEQUENCE OF
15112 and break the signature.
15116 *This change does not apply to 0.9.7.*
15118 * Zero the premaster secret after deriving the master secret in
15123 * Add some EVP_add_digest_alias registrations (as found in
15124 OpenSSL_add_all_digests()) to SSL_library_init()
15125 aka OpenSSL_add_ssl_algorithms(). This provides improved
15126 compatibility with peers using X.509 certificates
15127 with unconventional AlgorithmIdentifier OIDs.
15131 * Fix for Irix with NO_ASM.
15133 *"Bruce W. Forsberg" <bruce.forsberg@baesystems.com>*
15135 * ./config script fixes.
15137 *Ulf Moeller, Richard Levitte*
15139 * Fix 'openssl passwd -1'.
15143 * Change PKCS12_key_gen_asc() so it can cope with non null
15144 terminated strings whose length is passed in the passlen
15145 parameter, for example from PEM callbacks. This was done
15146 by adding an extra length parameter to asc2uni().
15148 *Steve Henson, reported by <oddissey@samsung.co.kr>*
15150 * Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
15151 call failed, free the DSA structure.
15155 * Fix to uni2asc() to cope with zero length Unicode strings.
15156 These are present in some PKCS#12 files.
15160 * Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
15161 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
15162 when writing a 32767 byte record.
15164 *Bodo Moeller; problem reported by Eric Day <eday@concentric.net>*
15166 * In `RSA_eay_public_{en,ed}crypt` and RSA_eay_mod_exp (rsa_eay.c),
15167 obtain lock CRYPTO_LOCK_RSA before setting `rsa->_method_mod_{n,p,q}`.
15169 (RSA objects have a reference count access to which is protected
15170 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
15171 so they are meant to be shared between threads.)
15172 *Bodo Moeller, Geoff Thorpe; original patch submitted by
15173 "Reddie, Steven" <Steven.Reddie@ca.com>*
15175 * Fix a deadlock in CRYPTO_mem_leaks().
15179 * Use better test patterns in bntest.
15183 * rand_win.c fix for Borland C.
15187 * BN_rshift bugfix for n == 0.
15191 * Add a 'bctest' script that checks for some known 'bc' bugs
15192 so that 'make test' does not abort just because 'bc' is broken.
15196 * Store verify_result within SSL_SESSION also for client side to
15197 avoid potential security hole. (Re-used sessions on the client side
15198 always resulted in verify_result==X509_V_OK, not using the original
15199 result of the server certificate verification.)
15203 * Fix ssl3_pending: If the record in s->s3->rrec is not of type
15204 SSL3_RT_APPLICATION_DATA, return 0.
15205 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
15210 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
15211 releases, have been re-implemented by renaming the previous
15212 implementations of ssl2_read and ssl3_read to ssl2_read_internal
15213 and ssl3_read_internal, respectively, and adding 'peek' parameters
15214 to them. The new ssl[23]_{read,peek} functions are calls to
15215 ssl[23]_read_internal with the 'peek' flag set appropriately.
15216 A 'peek' parameter has also been added to ssl3_read_bytes, which
15217 does the actual work for ssl3_read_internal.
15221 * Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
15222 the method-specific "init()" handler. Also clean up ex_data after
15223 calling the method-specific "finish()" handler. Previously, this was
15224 happening the other way round.
15228 * Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
15229 The previous value, 12, was not always sufficient for BN_mod_exp().
15233 * Make sure that shared libraries get the internal name engine with
15234 the full version number and not just 0. This should mark the
15235 shared libraries as not backward compatible. Of course, this should
15236 be changed again when we can guarantee backward binary compatibility.
15240 * Fix typo in get_cert_by_subject() in by_dir.c
15242 *Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>*
15244 * Rework the system to generate shared libraries:
15246 - Make note of the expected extension for the shared libraries and
15247 if there is a need for symbolic links from for example libcrypto.so.0
15248 to libcrypto.so.0.9.7. There is extended info in Configure for
15251 - Make as few rebuilds of the shared libraries as possible.
15253 - Still avoid linking the OpenSSL programs with the shared libraries.
15255 - When installing, install the shared libraries separately from the
15260 * Fix SSL_CTX_set_read_ahead macro to actually use its argument.
15262 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
15263 and not in SSL_clear because the latter is also used by the
15264 accept/connect functions; previously, the settings made by
15265 SSL_set_read_ahead would be lost during the handshake.
15267 *Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>*
15269 * Correct util/mkdef.pl to be selective about disabled algorithms.
15270 Previously, it would create entries for disabled algorithms no
15275 * Added several new manual pages for SSL_* function.
15279 ### Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
15281 * In ssl23_get_client_hello, generate an error message when faced
15282 with an initial SSL 3.0/TLS record that is too small to contain the
15283 first two bytes of the ClientHello message, i.e. client_version.
15284 (Note that this is a pathologic case that probably has never happened
15285 in real life.) The previous approach was to use the version number
15286 from the record header as a substitute; but our protocol choice
15287 should not depend on that one because it is not authenticated
15288 by the Finished messages.
15292 * More robust randomness gathering functions for Windows.
15294 *Jeffrey Altman <jaltman@columbia.edu>*
15296 * For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
15297 not set then we don't setup the error code for issuer check errors
15298 to avoid possibly overwriting other errors which the callback does
15299 handle. If an application does set the flag then we assume it knows
15300 what it is doing and can handle the new informational codes
15305 * Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
15306 a general "ANY" type, as such it should be able to decode anything
15307 including tagged types. However it didn't check the class so it would
15308 wrongly interpret tagged types in the same way as their universal
15309 counterpart and unknown types were just rejected. Changed so that the
15310 tagged and unknown types are handled in the same way as a SEQUENCE:
15311 that is the encoding is stored intact. There is also a new type
15312 "V_ASN1_OTHER" which is used when the class is not universal, in this
15313 case we have no idea what the actual type is so we just lump them all
15318 * On VMS, stdout may very well lead to a file that is written to
15319 in a record-oriented fashion. That means that every write() will
15320 write a separate record, which will be read separately by the
15321 programs trying to read from it. This can be very confusing.
15323 The solution is to put a BIO filter in the way that will buffer
15324 text until a linefeed is reached, and then write everything a
15325 line at a time, so every record written will be an actual line,
15326 not chunks of lines and not (usually doesn't happen, but I've
15327 seen it once) several lines in one record. BIO_f_linebuffer() is
15330 Currently, it's a VMS-only method, because that's where it has
15331 been tested well enough.
15335 * Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
15336 it can return incorrect results.
15337 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
15338 but it was in 0.9.6-beta[12].)
15342 * Disable the check for content being present when verifying detached
15343 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
15344 include zero length content when signing messages.
15348 * New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
15349 BIO_ctrl (for BIO pairs).
15353 * Add DSO method for VMS.
15357 * Bug fix: Montgomery multiplication could produce results with the
15362 * Add RPM specification openssl.spec and modify it to build three
15363 packages. The default package contains applications, application
15364 documentation and run-time libraries. The devel package contains
15365 include files, static libraries and function documentation. The
15366 doc package contains the contents of the doc directory. The original
15367 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
15371 * Add a large number of documentation files for many SSL routines.
15373 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>*
15375 * Add a configuration entry for Sony News 4.
15377 *NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>*
15379 * Don't set the two most significant bits to one when generating a
15380 random number < q in the DSA library.
15384 * New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
15385 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
15386 the underlying transport is blocking) if a handshake took place.
15387 (The default behaviour is needed by applications such as s_client
15388 and s_server that use select() to determine when to use SSL_read;
15389 but for applications that know in advance when to expect data, it
15390 just makes things more complicated.)
15394 * Add RAND_egd_bytes(), which gives control over the number of bytes read
15399 * Add a few more EBCDIC conditionals that make `req` and `x509`
15400 work better on such systems.
15402 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
15404 * Add two demo programs for PKCS12_parse() and PKCS12_create().
15405 Update PKCS12_parse() so it copies the friendlyName and the
15406 keyid to the certificates aux info.
15410 * Fix bug in PKCS7_verify() which caused an infinite loop
15411 if there was more than one signature.
15413 *Sven Uszpelkat <su@celocom.de>*
15415 * Major change in util/mkdef.pl to include extra information
15416 about each symbol, as well as presenting variables as well
15417 as functions. This change means that there's n more need
15418 to rebuild the .num files when some algorithms are excluded.
15422 * Allow the verify time to be set by an application,
15423 rather than always using the current time.
15427 * Phase 2 verify code reorganisation. The certificate
15428 verify code now looks up an issuer certificate by a
15429 number of criteria: subject name, authority key id
15430 and key usage. It also verifies self signed certificates
15431 by the same criteria. The main comparison function is
15432 X509_check_issued() which performs these checks.
15434 Lot of changes were necessary in order to support this
15435 without completely rewriting the lookup code.
15437 Authority and subject key identifier are now cached.
15439 The LHASH 'certs' is X509_STORE has now been replaced
15440 by a STACK_OF(X509_OBJECT). This is mainly because an
15441 LHASH can't store or retrieve multiple objects with
15442 the same hash value.
15444 As a result various functions (which were all internal
15445 use only) have changed to handle the new X509_STORE
15446 structure. This will break anything that messed round
15447 with X509_STORE internally.
15449 The functions X509_STORE_add_cert() now checks for an
15450 exact match, rather than just subject name.
15452 The X509_STORE API doesn't directly support the retrieval
15453 of multiple certificates matching a given criteria, however
15454 this can be worked round by performing a lookup first
15455 (which will fill the cache with candidate certificates)
15456 and then examining the cache for matches. This is probably
15457 the best we can do without throwing out X509_LOOKUP
15458 entirely (maybe later...).
15460 The X509_VERIFY_CTX structure has been enhanced considerably.
15462 All certificate lookup operations now go via a get_issuer()
15463 callback. Although this currently uses an X509_STORE it
15464 can be replaced by custom lookups. This is a simple way
15465 to bypass the X509_STORE hackery necessary to make this
15466 work and makes it possible to use more efficient techniques
15467 in future. A very simple version which uses a simple
15468 STACK for its trusted certificate store is also provided
15469 using X509_STORE_CTX_trusted_stack().
15471 The verify_cb() and verify() callbacks now have equivalents
15472 in the X509_STORE_CTX structure.
15474 X509_STORE_CTX also has a 'flags' field which can be used
15475 to customise the verify behaviour.
15479 * Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
15480 excludes S/MIME capabilities.
15484 * When a certificate request is read in keep a copy of the
15485 original encoding of the signed data and use it when outputting
15486 again. Signatures then use the original encoding rather than
15487 a decoded, encoded version which may cause problems if the
15488 request is improperly encoded.
15492 * For consistency with other BIO_puts implementations, call
15493 buffer_write(b, ...) directly in buffer_puts instead of calling
15496 In BIO_puts, increment b->num_write as in BIO_write.
15498 *Peter.Sylvester@EdelWeb.fr*
15500 * Fix BN_mul_word for the case where the word is 0. (We have to use
15501 BN_zero, we may not return a BIGNUM with an array consisting of
15502 words set to zero.)
15506 * Avoid calling abort() from within the library when problems are
15507 detected, except if preprocessor symbols have been defined
15508 (such as REF_CHECK, BN_DEBUG etc.).
15512 * New openssl application 'rsautl'. This utility can be
15513 used for low-level RSA operations. DER public key
15514 BIO/fp routines also added.
15518 * New Configure entry and patches for compiling on QNX 4.
15520 *Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>*
15522 * A demo state-machine implementation was sponsored by
15523 Nuron (<http://www.nuron.com/>) and is now available in
15524 demos/state_machine.
15528 * New options added to the 'dgst' utility for signature
15529 generation and verification.
15533 * Unrecognized PKCS#7 content types are now handled via a
15534 catch all ASN1_TYPE structure. This allows unsupported
15535 types to be stored as a "blob" and an application can
15536 encode and decode it manually.
15540 * Fix various signed/unsigned issues to make a_strex.c
15541 compile under VC++.
15543 *Oscar Jacobsson <oscar.jacobsson@celocom.com>*
15545 * ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
15546 length if passed a buffer. ASN1_INTEGER_to_BN failed
15547 if passed a NULL BN and its argument was negative.
15549 *Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>*
15551 * Modification to PKCS#7 encoding routines to output definite
15552 length encoding. Since currently the whole structures are in
15553 memory there's not real point in using indefinite length
15554 constructed encoding. However if OpenSSL is compiled with
15555 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
15559 * Added BIO_vprintf() and BIO_vsnprintf().
15563 * Added more prefixes to parse for in the strings written
15564 through a logging bio, to cover all the levels that are available
15565 through syslog. The prefixes are now:
15567 PANIC, EMERG, EMR => LOG_EMERG
15568 ALERT, ALR => LOG_ALERT
15569 CRIT, CRI => LOG_CRIT
15570 ERROR, ERR => LOG_ERR
15571 WARNING, WARN, WAR => LOG_WARNING
15572 NOTICE, NOTE, NOT => LOG_NOTICE
15573 INFO, INF => LOG_INFO
15574 DEBUG, DBG => LOG_DEBUG
15576 and as before, if none of those prefixes are present at the
15577 beginning of the string, LOG_ERR is chosen.
15579 On Win32, the `LOG_*` levels are mapped according to this:
15581 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
15582 LOG_WARNING => EVENTLOG_WARNING_TYPE
15583 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
15587 * Made it possible to reconfigure with just the configuration
15588 argument "reconf" or "reconfigure". The command line arguments
15589 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
15590 and are retrieved from there when reconfiguring.
15596 *Assar Westerlund <assar@sics.se>, Richard Levitte*
15598 * Add the arguments -CAfile and -CApath to the pkcs12 utility.
15602 * The obj_dat.pl script was messing up the sorting of object
15603 names. The reason was that it compared the quoted version
15604 of strings as a result "OCSP" > "OCSP Signing" because
15605 " > SPACE. Changed script to store unquoted versions of
15606 names and add quotes on output. It was also omitting some
15607 names from the lookup table if they were given a default
15608 value (that is if SN is missing it is given the same
15609 value as LN and vice versa), these are now added on the
15610 grounds that if an object has a name we should be able to
15611 look it up. Finally added warning output when duplicate
15612 short or long names are found.
15616 * Changes needed for Tandem NSK.
15618 *Scott Uroff <scott@xypro.com>*
15620 * Fix SSL 2.0 rollback checking: Due to an off-by-one error in
15621 RSA_padding_check_SSLv23(), special padding was never detected
15622 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
15623 version rollback attacks was not effective.
15625 In s23_clnt.c, don't use special rollback-attack detection padding
15626 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
15627 client; similarly, in s23_srvr.c, don't do the rollback check if
15628 SSL 2.0 is the only protocol enabled in the server.
15632 * Make it possible to get hexdumps of unprintable data with 'openssl
15633 asn1parse'. By implication, the functions ASN1_parse_dump() and
15634 BIO_dump_indent() are added.
15638 * New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
15639 these print out strings and name structures based on various
15640 flags including RFC2253 support and proper handling of
15641 multibyte characters. Added options to the 'x509' utility
15642 to allow the various flags to be set.
15646 * Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
15647 Also change the functions X509_cmp_current_time() and
15648 X509_gmtime_adj() work with an ASN1_TIME structure,
15649 this will enable certificates using GeneralizedTime in validity
15650 dates to be checked.
15654 * Make the NEG_PUBKEY_BUG code (which tolerates invalid
15655 negative public key encodings) on by default,
15656 NO_NEG_PUBKEY_BUG can be set to disable it.
15660 * New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
15661 content octets. An i2c_ASN1_OBJECT is unnecessary because
15662 the encoding can be trivially obtained from the structure.
15666 * crypto/err.c locking bugfix: Use write locks (`CRYPTO_w_[un]lock`),
15667 not read locks (`CRYPTO_r_[un]lock`).
15671 * A first attempt at creating official support for shared
15672 libraries through configuration. I've kept it so the
15673 default is static libraries only, and the OpenSSL programs
15674 are always statically linked for now, but there are
15675 preparations for dynamic linking in place.
15676 This has been tested on Linux and Tru64.
15680 * Randomness polling function for Win9x, as described in:
15681 Peter Gutmann, Software Generation of Practically Strong
15686 * Fix so PRNG is seeded in req if using an already existing
15691 * New options to smime application. -inform and -outform
15692 allow alternative formats for the S/MIME message including
15693 PEM and DER. The -content option allows the content to be
15694 specified separately. This should allow things like Netscape
15695 form signing output easier to verify.
15699 * Fix the ASN1 encoding of tags using the 'long form'.
15703 * New ASN1 functions, `i2c_*` and `c2i_*` for INTEGER and BIT
15704 STRING types. These convert content octets to and from the
15705 underlying type. The actual tag and length octets are
15706 already assumed to have been read in and checked. These
15707 are needed because all other string types have virtually
15708 identical handling apart from the tag. By having versions
15709 of the ASN1 functions that just operate on content octets
15710 IMPLICIT tagging can be handled properly. It also allows
15711 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
15712 and ASN1_INTEGER are identical apart from the tag.
15716 * Change the handling of OID objects as follows:
15718 - New object identifiers are inserted in objects.txt, following
15719 the syntax given in [crypto/objects/README.md](crypto/objects/README.md).
15720 - objects.pl is used to process obj_mac.num and create a new
15722 - obj_dat.pl is used to create a new obj_dat.h, using the data in
15725 This is currently kind of a hack, and the perl code in objects.pl
15726 isn't very elegant, but it works as I intended. The simplest way
15727 to check that it worked correctly is to look in obj_dat.h and
15728 check the array nid_objs and make sure the objects haven't moved
15729 around (this is important!). Additions are OK, as well as
15730 consistent name changes.
15734 * Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
15738 * Addition of the command line parameter '-rand file' to 'openssl req'.
15739 The given file adds to whatever has already been seeded into the
15740 random pool through the RANDFILE configuration file option or
15741 environment variable, or the default random state file.
15745 * mkstack.pl now sorts each macro group into lexical order.
15746 Previously the output order depended on the order the files
15747 appeared in the directory, resulting in needless rewriting
15752 * Patches to make OpenSSL compile under Win32 again. Mostly
15753 work arounds for the VC++ problem that it treats func() as
15754 func(void). Also stripped out the parts of mkdef.pl that
15755 added extra typesafe functions: these no longer exist.
15759 * Reorganisation of the stack code. The macros are now all
15760 collected in safestack.h . Each macro is defined in terms of
15761 a "stack macro" of the form `SKM_<name>(type, a, b)`. The
15762 DEBUG_SAFESTACK is now handled in terms of function casts,
15763 this has the advantage of retaining type safety without the
15764 use of additional functions. If DEBUG_SAFESTACK is not defined
15765 then the non typesafe macros are used instead. Also modified the
15766 mkstack.pl script to handle the new form. Needs testing to see
15767 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
15768 the default if no major problems. Similar behaviour for ASN1_SET_OF
15769 and PKCS12_STACK_OF.
15773 * When some versions of IIS use the 'NET' form of private key the
15774 key derivation algorithm is different. Normally MD5(password) is
15775 used as a 128 bit RC4 key. In the modified case
15776 MD5(MD5(password) + "SGCKEYSALT") is used instead. Added some
15777 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
15778 as the old Netscape_RSA functions except they have an additional
15779 'sgckey' parameter which uses the modified algorithm. Also added
15780 an -sgckey command line option to the rsa utility. Thanks to
15781 Adrian Peck <bertie@ncipher.com> for posting details of the modified
15782 algorithm to openssl-dev.
15786 * The evp_local.h macros were using 'c.##kname' which resulted in
15787 invalid expansion on some systems (SCO 5.0.5 for example).
15788 Corrected to 'c.kname'.
15790 *Phillip Porch <root@theporch.com>*
15792 * New X509_get1_email() and X509_REQ_get1_email() functions that return
15793 a STACK of email addresses from a certificate or request, these look
15794 in the subject name and the subject alternative name extensions and
15795 omit any duplicate addresses.
15799 * Re-implement BN_mod_exp2_mont using independent (and larger) windows.
15800 This makes DSA verification about 2 % faster.
15804 * Increase maximum window size in `BN_mod_exp_...` to 6 bits instead of 5
15805 (meaning that now 2^5 values will be precomputed, which is only 4 KB
15806 plus overhead for 1024 bit moduli).
15807 This makes exponentiations about 0.5 % faster for 1024 bit
15808 exponents (as measured by "openssl speed rsa2048").
15812 * Rename memory handling macros to avoid conflicts with other
15814 Malloc => OPENSSL_malloc
15815 Malloc_locked => OPENSSL_malloc_locked
15816 Realloc => OPENSSL_realloc
15817 Free => OPENSSL_free
15821 * New function BN_mod_exp_mont_word for small bases (roughly 15%
15822 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
15826 * CygWin32 support.
15828 *John Jarvie <jjarvie@newsguy.com>*
15830 * The type-safe stack code has been rejigged. It is now only compiled
15831 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
15832 by default all type-specific stack functions are "#define"d back to
15833 standard stack functions. This results in more streamlined output
15834 but retains the type-safety checking possibilities of the original
15839 * The STACK code has been cleaned up, and certain type declarations
15840 that didn't make a lot of sense have been brought in line. This has
15841 also involved a cleanup of sorts in safestack.h to more correctly
15842 map type-safe stack functions onto their plain stack counterparts.
15843 This work has also resulted in a variety of "const"ifications of
15844 lots of the code, especially `_cmp` operations which should normally
15845 be prototyped with "const" parameters anyway.
15849 * When generating bytes for the first time in md_rand.c, 'stir the pool'
15850 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
15851 (The PRNG state consists of two parts, the large pool 'state' and 'md',
15852 where all of 'md' is used each time the PRNG is used, but 'state'
15853 is used only indexed by a cyclic counter. As entropy may not be
15854 well distributed from the beginning, 'md' is important as a
15855 chaining variable. However, the output function chains only half
15856 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
15857 all of 'md', and seeding with STATE_SIZE dummy bytes will result
15858 in all of 'state' being rewritten, with the new values depending
15859 on virtually all of 'md'. This overcomes the 80 bit limitation.)
15863 * In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
15864 the handshake is continued after ssl_verify_cert_chain();
15865 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
15866 can lead to 'unexplainable' connection aborts later.
15868 *Bodo Moeller; problem tracked down by Lutz Jaenicke*
15870 * Major EVP API cipher revision.
15871 Add hooks for extra EVP features. This allows various cipher
15872 parameters to be set in the EVP interface. Support added for variable
15873 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
15874 setting of RC2 and RC5 parameters.
15876 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
15879 Remove lots of duplicated code from the EVP library. For example *every*
15880 cipher init() function handles the 'iv' in the same way according to the
15881 cipher mode. They also all do nothing if the 'key' parameter is NULL and
15882 for CFB and OFB modes they zero ctx->num.
15884 New functionality allows removal of S/MIME code RC2 hack.
15886 Most of the routines have the same form and so can be declared in terms
15889 By shifting this to the top level EVP_CipherInit() it can be removed from
15890 all individual ciphers. If the cipher wants to handle IVs or keys
15891 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
15894 Change lots of functions like EVP_EncryptUpdate() to now return a
15895 value: although software versions of the algorithms cannot fail
15896 any installed hardware versions can.
15900 * Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
15901 this option is set, tolerate broken clients that send the negotiated
15902 protocol version number instead of the requested protocol version
15907 * Call dh_tmp_cb (set by `..._TMP_DH_CB`) with correct 'is_export' flag;
15908 i.e. non-zero for export ciphersuites, zero otherwise.
15909 Previous versions had this flag inverted, inconsistent with
15910 rsa_tmp_cb (..._TMP_RSA_CB).
15912 *Bodo Moeller; problem reported by Amit Chopra*
15914 * Add missing DSA library text string. Work around for some IIS
15915 key files with invalid SEQUENCE encoding.
15919 * Add a document (doc/standards.txt) that list all kinds of standards
15920 and so on that are implemented in OpenSSL.
15924 * Enhance c_rehash script. Old version would mishandle certificates
15925 with the same subject name hash and wouldn't handle CRLs at all.
15926 Added -fingerprint option to crl utility, to support new c_rehash
15931 * Eliminate non-ANSI declarations in crypto.h and stack.h.
15935 * Fix for SSL server purpose checking. Server checking was
15936 rejecting certificates which had extended key usage present
15937 but no ssl client purpose.
15939 *Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>*
15941 * Make PKCS#12 code work with no password. The PKCS#12 spec
15942 is a little unclear about how a blank password is handled.
15943 Since the password in encoded as a BMPString with terminating
15944 double NULL a zero length password would end up as just the
15945 double NULL. However no password at all is different and is
15946 handled differently in the PKCS#12 key generation code. NS
15947 treats a blank password as zero length. MSIE treats it as no
15948 password on export: but it will try both on import. We now do
15949 the same: PKCS12_parse() tries zero length and no password if
15950 the password is set to "" or NULL (NULL is now a valid password:
15951 it wasn't before) as does the pkcs12 application.
15955 * Bugfixes in `apps/x509.c`: Avoid a memory leak; and don't use
15956 perror when PEM_read_bio_X509_REQ fails, the error message must
15957 be obtained from the error queue.
15961 * Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
15962 it in ERR_remove_state if appropriate, and change ERR_get_state
15963 accordingly to avoid race conditions (this is necessary because
15964 thread_hash is no longer constant once set).
15968 * Bugfix for linux-elf makefile.one.
15972 * RSA_get_default_method() will now cause a default
15973 RSA_METHOD to be chosen if one doesn't exist already.
15974 Previously this was only set during a call to RSA_new()
15975 or RSA_new_method(NULL) meaning it was possible for
15976 RSA_get_default_method() to return NULL.
15980 * Added native name translation to the existing DSO code
15981 that will convert (if the flag to do so is set) filenames
15982 that are sufficiently small and have no path information
15983 into a canonical native form. Eg. "blah" converted to
15984 "libblah.so" or "blah.dll" etc.
15988 * New function ERR_error_string_n(e, buf, len) which is like
15989 ERR_error_string(e, buf), but writes at most 'len' bytes
15990 including the 0 terminator. For ERR_error_string_n, 'buf'
15993 *Damien Miller <djm@mindrot.org>, Bodo Moeller*
15995 * CONF library reworked to become more general. A new CONF
15996 configuration file reader "class" is implemented as well as a
15997 new functions (`NCONF_*`, for "New CONF") to handle it. The now
15998 old `CONF_*` functions are still there, but are reimplemented to
15999 work in terms of the new functions. Also, a set of functions
16000 to handle the internal storage of the configuration data is
16001 provided to make it easier to write new configuration file
16002 reader "classes" (I can definitely see something reading a
16003 configuration file in XML format, for example), called `_CONF_*`,
16004 or "the configuration storage API"...
16006 The new configuration file reading functions are:
16008 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
16009 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
16011 NCONF_default, NCONF_WIN32
16013 NCONF_dump_fp, NCONF_dump_bio
16015 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
16016 NCONF_new creates a new CONF object. This works in the same way
16017 as other interfaces in OpenSSL, like the BIO interface.
16018 `NCONF_dump_*` dump the internal storage of the configuration file,
16019 which is useful for debugging. All other functions take the same
16020 arguments as the old `CONF_*` functions with the exception of the
16021 first that must be a `CONF *` instead of a `LHASH *`.
16023 To make it easier to use the new classes with the old `CONF_*` functions,
16024 the function CONF_set_default_method is provided.
16028 * Add '-tls1' option to 'openssl ciphers', which was already
16029 mentioned in the documentation but had not been implemented.
16030 (This option is not yet really useful because even the additional
16031 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
16035 * Initial DSO code added into libcrypto for letting OpenSSL (and
16036 OpenSSL-based applications) load shared libraries and bind to
16037 them in a portable way.
16039 *Geoff Thorpe, with contributions from Richard Levitte*
16041 ### Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
16043 * Make sure _lrotl and _lrotr are only used with MSVC.
16045 * Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
16046 (the default implementation of RAND_status).
16048 * Rename openssl x509 option '-crlext', which was added in 0.9.5,
16049 to '-clrext' (= clear extensions), as intended and documented.
16050 *Bodo Moeller; inconsistency pointed out by Michael Attili
16051 <attili@amaxo.com>*
16053 * Fix for HMAC. It wasn't zeroing the rest of the block if the key length
16054 was larger than the MD block size.
16056 *Steve Henson, pointed out by Yost William <YostW@tce.com>*
16058 * Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
16059 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
16060 using the passed key: if the passed key was a private key the result
16061 of X509_print(), for example, would be to print out all the private key
16066 * des_quad_cksum() byte order bug fix.
16067 *Ulf Möller, using the problem description in krb4-0.9.7, where
16068 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>*
16070 * Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
16073 *Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>*
16075 * For easily testing in shell scripts whether some command
16076 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
16077 returns with exit code 0 iff no command of the given name is available.
16078 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
16079 the output goes to stdout and nothing is printed to stderr.
16080 Additional arguments are always ignored.
16082 Since for each cipher there is a command of the same name,
16083 the 'no-cipher' compilation switches can be tested this way.
16085 ('openssl no-XXX' is not able to detect pseudo-commands such
16086 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
16090 * Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
16094 * For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
16095 is set; it will be thrown away anyway because each handshake creates
16097 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
16098 to parameters -- in previous versions (since OpenSSL 0.9.3) the
16099 'default key' from SSL_CTX_set_tmp_dh would always be lost, meaning
16100 you effectively got SSL_OP_SINGLE_DH_USE when using this macro.
16104 * New s_client option -ign_eof: EOF at stdin is ignored, and
16105 'Q' and 'R' lose their special meanings (quit/renegotiate).
16106 This is part of what -quiet does; unlike -quiet, -ign_eof
16107 does not suppress any output.
16111 * Add compatibility options to the purpose and trust code. The
16112 purpose X509_PURPOSE_ANY is "any purpose" which automatically
16113 accepts a certificate or CA, this was the previous behaviour,
16114 with all the associated security issues.
16116 X509_TRUST_COMPAT is the old trust behaviour: only and
16117 automatically trust self signed roots in certificate store. A
16118 new trust setting X509_TRUST_DEFAULT is used to specify that
16119 a purpose has no associated trust setting and it should instead
16120 use the value in the default purpose.
16124 * Fix the PKCS#8 DSA private key code so it decodes keys again
16125 and fix a memory leak.
16129 * In util/mkerr.pl (which implements 'make errors'), preserve
16130 reason strings from the previous version of the .c file, as
16131 the default to have only downcase letters (and digits) in
16132 automatically generated reasons codes is not always appropriate.
16136 * In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
16137 using strerror. Previously, ERR_reason_error_string() returned
16138 library names as reason strings for SYSerr; but SYSerr is a special
16139 case where small numbers are errno values, not library numbers.
16143 * Add '-dsaparam' option to 'openssl dhparam' application. This
16144 converts DSA parameters into DH parameters. (When creating parameters,
16145 DSA_generate_parameters is used.)
16149 * Include 'length' (recommended exponent length) in C code generated
16150 by 'openssl dhparam -C'.
16154 * The second argument to set_label in perlasm was already being used
16155 so couldn't be used as a "file scope" flag. Moved to third argument
16160 * In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
16161 instead of RAND_bytes for encryption IVs and salts.
16165 * Include RAND_status() into RAND_METHOD instead of implementing
16166 it only for md_rand.c Otherwise replacing the PRNG by calling
16167 RAND_set_rand_method would be impossible.
16171 * Don't let DSA_generate_key() enter an infinite loop if the random
16172 number generation fails.
16176 * New 'rand' application for creating pseudo-random output.
16180 * Added configuration support for Linux/IA64
16182 *Rolf Haberrecker <rolf@suse.de>*
16184 * Assembler module support for Mingw32.
16188 * Shared library support for HPUX (in shlib/).
16190 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous*
16192 * Shared library support for Solaris gcc.
16194 *Lutz Behnke <behnke@trustcenter.de>*
16196 ### Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
16198 * PKCS7_encrypt() was adding text MIME headers twice because they
16199 were added manually and by SMIME_crlf_copy().
16203 * In bntest.c don't call BN_rand with zero bits argument.
16205 *Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>*
16207 * BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
16208 case was implemented. This caused BN_div_recp() to fail occasionally.
16212 * Add an optional second argument to the set_label() in the perl
16213 assembly language builder. If this argument exists and is set
16214 to 1 it signals that the assembler should use a symbol whose
16215 scope is the entire file, not just the current function. This
16216 is needed with MASM which uses the format label:: for this scope.
16218 *Steve Henson, pointed out by Peter Runestig <peter@runestig.com>*
16220 * Change the ASN1 types so they are typedefs by default. Before
16221 almost all types were #define'd to ASN1_STRING which was causing
16222 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
16227 * Change names of new functions to the new get1/get0 naming
16228 convention: After 'get1', the caller owns a reference count
16229 and has to call `..._free`; 'get0' returns a pointer to some
16230 data structure without incrementing reference counters.
16231 (Some of the existing 'get' functions increment a reference
16232 counter, some don't.)
16233 Similarly, 'set1' and 'add1' functions increase reference
16234 counters or duplicate objects.
16238 * Allow for the possibility of temp RSA key generation failure:
16239 the code used to assume it always worked and crashed on failure.
16243 * Fix potential buffer overrun problem in BIO_printf().
16244 *Ulf Möller, using public domain code by Patrick Powell; problem
16245 pointed out by David Sacerdote <das33@cornell.edu>*
16247 * Support EGD <http://www.lothar.com/tech/crypto/>. New functions
16248 RAND_egd() and RAND_status(). In the command line application,
16249 the EGD socket can be specified like a seed file using RANDFILE
16254 * Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
16255 Some CAs (e.g. Verisign) distribute certificates in this form.
16259 * Remove the SSL_ALLOW_ADH compile option and set the default cipher
16260 list to exclude them. This means that no special compilation option
16261 is needed to use anonymous DH: it just needs to be included in the
16266 * Change the EVP_MD_CTX_type macro so its meaning consistent with
16267 EVP_MD_type. The old functionality is available in a new macro called
16268 EVP_MD_md(). Change code that uses it and update docs.
16272 * `..._ctrl` functions now have corresponding `..._callback_ctrl` functions
16273 where the `void *` argument is replaced by a function pointer argument.
16274 Previously `void *` was abused to point to functions, which works on
16275 many platforms, but is not correct. As these functions are usually
16276 called by macros defined in OpenSSL header files, most source code
16277 should work without changes.
16281 * `<openssl/opensslconf.h>` (which is created by Configure) now contains
16282 sections with information on -D... compiler switches used for
16283 compiling the library so that applications can see them. To enable
16284 one of these sections, a pre-processor symbol `OPENSSL_..._DEFINES`
16285 must be defined. E.g.,
16286 #define OPENSSL_ALGORITHM_DEFINES
16287 #include <openssl/opensslconf.h>
16288 defines all pertinent `NO_<algo>` symbols, such as NO_IDEA, NO_RSA, etc.
16290 *Richard Levitte, Ulf and Bodo Möller*
16292 * Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
16297 * Change the 'other' type in certificate aux info to a STACK_OF
16298 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
16299 the required ASN1 format: arbitrary types determined by an OID.
16303 * Add some PEM_write_X509_REQ_NEW() functions and a command line
16304 argument to 'req'. This is not because the function is newer or
16305 better than others it just uses the work 'NEW' in the certificate
16306 request header lines. Some software needs this.
16310 * Reorganise password command line arguments: now passwords can be
16311 obtained from various sources. Delete the PEM_cb function and make
16312 it the default behaviour: i.e. if the callback is NULL and the
16313 usrdata argument is not NULL interpret it as a null terminated pass
16314 phrase. If usrdata and the callback are NULL then the pass phrase
16315 is prompted for as usual.
16319 * Add support for the Compaq Atalla crypto accelerator. If it is installed,
16320 the support is automatically enabled. The resulting binaries will
16321 autodetect the card and use it if present.
16323 *Ben Laurie and Compaq Inc.*
16325 * Work around for Netscape hang bug. This sends certificate request
16326 and server done in one record. Since this is perfectly legal in the
16327 SSL/TLS protocol it isn't a "bug" option and is on by default. See
16328 the bugs/SSLv3 entry for more info.
16332 * HP-UX tune-up: new unified configs, HP C compiler bug workaround.
16336 * Add -rand argument to smime and pkcs12 applications and read/write
16341 * New 'passwd' tool for crypt(3) and apr1 password hashes.
16345 * Add command line password options to the remaining applications.
16349 * Bug fix for BN_div_recp() for numerators with an even number of
16354 * More tests in bntest.c, and changed test_bn output.
16358 * ./config recognizes MacOS X now.
16362 * Bug fix for BN_div() when the first words of num and divisor are
16363 equal (it gave wrong results if `(rem=(n1-q*d0)&BN_MASK2) < d0)`.
16367 * Add support for various broken PKCS#8 formats, and command line
16368 options to produce them.
16372 * New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
16373 get temporary BIGNUMs from a BN_CTX.
16377 * Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
16382 * Change the `SSLeay_add_all_*()` functions to `OpenSSL_add_all_*()` and
16383 include a #define from the old name to the new. The original intent
16384 was that statically linked binaries could for example just call
16385 SSLeay_add_all_ciphers() to just add ciphers to the table and not
16386 link with digests. This never worked because SSLeay_add_all_digests()
16387 and SSLeay_add_all_ciphers() were in the same source file so calling
16388 one would link with the other. They are now in separate source files.
16392 * Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
16396 * Use a less unusual form of the Miller-Rabin primality test (it used
16397 a binary algorithm for exponentiation integrated into the Miller-Rabin
16398 loop, our standard modexp algorithms are faster).
16402 * Support for the EBCDIC character set completed.
16404 *Martin Kraemer <Martin.Kraemer@Mch.SNI.De>*
16406 * Source code cleanups: use const where appropriate, eliminate casts,
16407 use `void *` instead of `char *` in lhash.
16411 * Bugfix: ssl3_send_server_key_exchange was not restartable
16412 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
16413 this the server could overwrite ephemeral keys that the client
16418 * Turn DSA_is_prime into a macro that calls BN_is_prime,
16419 using 50 iterations of the Rabin-Miller test.
16421 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
16422 iterations of the Rabin-Miller test as required by the appendix
16423 to FIPS PUB 186[-1]) instead of DSA_is_prime.
16424 As BN_is_prime_fasttest includes trial division, DSA parameter
16425 generation becomes much faster.
16427 This implies a change for the callback functions in DSA_is_prime
16428 and DSA_generate_parameters: The callback function is called once
16429 for each positive witness in the Rabin-Miller test, not just
16430 occasionally in the inner loop; and the parameters to the
16431 callback function now provide an iteration count for the outer
16432 loop rather than for the current invocation of the inner loop.
16433 DSA_generate_parameters additionally can call the callback
16434 function with an 'iteration count' of -1, meaning that a
16435 candidate has passed the trial division test (when q is generated
16436 from an application-provided seed, trial division is skipped).
16440 * New function BN_is_prime_fasttest that optionally does trial
16441 division before starting the Rabin-Miller test and has
16442 an additional BN_CTX * argument (whereas BN_is_prime always
16443 has to allocate at least one BN_CTX).
16444 'callback(1, -1, cb_arg)' is called when a number has passed the
16445 trial division stage.
16449 * Fix for bug in CRL encoding. The validity dates weren't being handled
16454 * New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
16458 * New function BN_pseudo_rand().
16462 * Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
16463 bignum version of BN_from_montgomery() with the working code from
16464 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
16469 * Avoid a race condition in s2_clnt.c (function get_server_hello) that
16470 made it impossible to use the same SSL_SESSION data structure in
16471 SSL2 clients in multiple threads.
16475 * The return value of RAND_load_file() no longer counts bytes obtained
16476 by stat(). RAND_load_file(..., -1) is new and uses the complete file
16477 to seed the PRNG (previously an explicit byte count was required).
16479 *Ulf Möller, Bodo Möller*
16481 * Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
16482 used `char *` instead of `void *` and had casts all over the place.
16486 * Make BN_generate_prime() return NULL on error if ret!=NULL.
16490 * Retain source code compatibility for BN_prime_checks macro:
16491 BN_is_prime(..., BN_prime_checks, ...) now uses
16492 BN_prime_checks_for_size to determine the appropriate number of
16493 Rabin-Miller iterations.
16497 * Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
16498 DH_CHECK_P_NOT_SAFE_PRIME.
16499 (Check if this is true? OpenPGP calls them "strong".)
16503 * Merge the functionality of "dh" and "gendh" programs into a new program
16504 "dhparam". The old programs are retained for now but will handle DH keys
16505 (instead of parameters) in future.
16509 * Make the ciphers, s_server and s_client programs check the return values
16510 when a new cipher list is set.
16514 * Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
16515 ciphers. Before when the 56bit ciphers were enabled the sorting was
16518 The syntax for the cipher sorting has been extended to support sorting by
16519 cipher-strength (using the strength_bits hard coded in the tables).
16520 The new command is `@STRENGTH` (see also `doc/apps/ciphers.pod`).
16522 Fix a bug in the cipher-command parser: when supplying a cipher command
16523 string with an "undefined" symbol (neither command nor alphanumeric
16524 *A-Za-z0-9*, ssl_set_cipher_list used to hang in an endless loop. Now
16525 an error is flagged.
16527 Due to the strength-sorting extension, the code of the
16528 ssl_create_cipher_list() function was completely rearranged. I hope that
16529 the readability was also increased :-)
16531 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>*
16533 * Minor change to 'x509' utility. The -CAcreateserial option now uses 1
16534 for the first serial number and places 2 in the serial number file. This
16535 avoids problems when the root CA is created with serial number zero and
16536 the first user certificate has the same issuer name and serial number
16541 * Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
16542 the new code. Add documentation for this stuff.
16546 * Changes to X509_ATTRIBUTE utilities. These have been renamed from
16547 `X509_*()` to `X509at_*()` on the grounds that they don't handle X509
16548 structures and behave in an analogous way to the X509v3 functions:
16549 they shouldn't be called directly but wrapper functions should be used
16552 So we also now have some wrapper functions that call the X509at functions
16553 when passed certificate requests. (TO DO: similar things can be done with
16554 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
16555 things. Some of these need some d2i or i2d and print functionality
16556 because they handle more complex structures.)
16560 * Add missing #ifndefs that caused missing symbols when building libssl
16561 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
16562 NO_RSA in `ssl/s2*.c`.
16564 *Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller*
16566 * Precautions against using the PRNG uninitialized: RAND_bytes() now
16567 has a return value which indicates the quality of the random data
16568 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
16569 error queue. New function RAND_pseudo_bytes() generates output that is
16570 guaranteed to be unique but not unpredictable. RAND_add is like
16571 RAND_seed, but takes an extra argument for an entropy estimate
16572 (RAND_seed always assumes full entropy).
16576 * Do more iterations of Rabin-Miller probable prime test (specifically,
16577 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
16578 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
16579 in crypto/bn/bn_prime.c for the complete table). This guarantees a
16580 false-positive rate of at most 2^-80 for random input.
16584 * Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
16588 * New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
16589 in the 0.9.5 release), this returns the chain
16590 from an X509_CTX structure with a dup of the stack and all
16591 the X509 reference counts upped: so the stack will exist
16592 after X509_CTX_cleanup() has been called. Modify pkcs12.c
16595 Also make SSL_SESSION_print() print out the verify return
16600 * Add manpage for the pkcs12 command. Also change the default
16601 behaviour so MAC iteration counts are used unless the new
16602 -nomaciter option is used. This improves file security and
16603 only older versions of MSIE (4.0 for example) need it.
16607 * Honor the no-xxx Configure options when creating .DEF files.
16611 * Add PKCS#10 attributes to field table: challengePassword,
16612 unstructuredName and unstructuredAddress. These are taken from
16613 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
16614 international characters are used.
16616 More changes to X509_ATTRIBUTE code: allow the setting of types
16617 based on strings. Remove the 'loc' parameter when adding
16618 attributes because these will be a SET OF encoding which is sorted
16623 * Initial changes to the 'req' utility to allow request generation
16624 automation. This will allow an application to just generate a template
16625 file containing all the field values and have req construct the
16628 Initial support for X509_ATTRIBUTE handling. Stacks of these are
16629 used all over the place including certificate requests and PKCS#7
16630 structures. They are currently handled manually where necessary with
16631 some primitive wrappers for PKCS#7. The new functions behave in a
16632 manner analogous to the X509 extension functions: they allow
16633 attributes to be looked up by NID and added.
16635 Later something similar to the X509V3 code would be desirable to
16636 automatically handle the encoding, decoding and printing of the
16637 more complex types. The string types like challengePassword can
16638 be handled by the string table functions.
16640 Also modified the multi byte string table handling. Now there is
16641 a 'global mask' which masks out certain types. The table itself
16642 can use the flag STABLE_NO_MASK to ignore the mask setting: this
16643 is useful when for example there is only one permissible type
16644 (as in countryName) and using the mask might result in no valid
16649 * Clean up 'Finished' handling, and add functions SSL_get_finished and
16650 SSL_get_peer_finished to allow applications to obtain the latest
16651 Finished messages sent to the peer or expected from the peer,
16652 respectively. (SSL_get_peer_finished is usually the Finished message
16653 actually received from the peer, otherwise the protocol will be aborted.)
16655 As the Finished message are message digests of the complete handshake
16656 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
16657 be used for external authentication procedures when the authentication
16658 provided by SSL/TLS is not desired or is not enough.
16662 * Enhanced support for Alpha Linux is added. Now ./config checks if
16663 the host supports BWX extension and if Compaq C is present on the
16664 $PATH. Just exploiting of the BWX extension results in 20-30%
16665 performance kick for some algorithms, e.g. DES and RC4 to mention
16666 a couple. Compaq C in turn generates ~20% faster code for MD5 and
16671 * Add support for MS "fast SGC". This is arguably a violation of the
16672 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
16673 weak crypto and after checking the certificate is SGC a second one
16674 with strong crypto. MS SGC stops the first handshake after receiving
16675 the server certificate message and sends a second client hello. Since
16676 a server will typically do all the time consuming operations before
16677 expecting any further messages from the client (server key exchange
16678 is the most expensive) there is little difference between the two.
16680 To get OpenSSL to support MS SGC we have to permit a second client
16681 hello message after we have sent server done. In addition we have to
16682 reset the MAC if we do get this second client hello.
16686 * Add a function 'd2i_AutoPrivateKey()' this will automatically decide
16687 if a DER encoded private key is RSA or DSA traditional format. Changed
16688 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
16689 format DER encoded private key. Newer code should use PKCS#8 format which
16690 has the key type encoded in the ASN1 structure. Added DER private key
16691 support to pkcs8 application.
16695 * SSL 3/TLS 1 servers now don't request certificates when an anonymous
16696 ciphersuites has been selected (as required by the SSL 3/TLS 1
16697 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
16698 is set, we interpret this as a request to violate the specification
16699 (the worst that can happen is a handshake failure, and 'correct'
16700 behaviour would result in a handshake failure anyway).
16704 * In SSL_CTX_add_session, take into account that there might be multiple
16705 SSL_SESSION structures with the same session ID (e.g. when two threads
16706 concurrently obtain them from an external cache).
16707 The internal cache can handle only one SSL_SESSION with a given ID,
16708 so if there's a conflict, we now throw out the old one to achieve
16713 * Add OIDs for idea and blowfish in CBC mode. This will allow both
16714 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
16715 some routines that use cipher OIDs: some ciphers do not have OIDs
16716 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
16721 * Simplify the trust setting structure and code. Now we just have
16722 two sequences of OIDs for trusted and rejected settings. These will
16723 typically have values the same as the extended key usage extension
16724 and any application specific purposes.
16726 The trust checking code now has a default behaviour: it will just
16727 check for an object with the same NID as the passed id. Functions can
16728 be provided to override either the default behaviour or the behaviour
16729 for a given id. SSL client, server and email already have functions
16730 in place for compatibility: they check the NID and also return "trusted"
16731 if the certificate is self signed.
16735 * Add d2i,i2d bio/fp functions for PrivateKey: these convert the
16736 traditional format into an EVP_PKEY structure.
16740 * Add a password callback function PEM_cb() which either prompts for
16741 a password if usr_data is NULL or otherwise assumes it is a null
16742 terminated password. Allow passwords to be passed on command line
16743 environment or config files in a few more utilities.
16747 * Add a bunch of DER and PEM functions to handle PKCS#8 format private
16748 keys. Add some short names for PKCS#8 PBE algorithms and allow them
16749 to be specified on the command line for the pkcs8 and pkcs12 utilities.
16750 Update documentation.
16754 * Support for ASN1 "NULL" type. This could be handled before by using
16755 ASN1_TYPE but there wasn't any function that would try to read a NULL
16756 and produce an error if it couldn't. For compatibility we also have
16757 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
16758 don't allocate anything because they don't need to.
16762 * Initial support for MacOS is now provided. Examine INSTALL.MacOS
16765 *Andy Polyakov, Roy Woods <roy@centicsystems.ca>*
16767 * Rebuild of the memory allocation routines used by OpenSSL code and
16768 possibly others as well. The purpose is to make an interface that
16769 provide hooks so anyone can build a separate set of allocation and
16770 deallocation routines to be used by OpenSSL, for example memory
16771 pool implementations, or something else, which was previously hard
16772 since Malloc(), Realloc() and Free() were defined as macros having
16773 the values malloc, realloc and free, respectively (except for Win32
16774 compilations). The same is provided for memory debugging code.
16775 OpenSSL already comes with functionality to find memory leaks, but
16776 this gives people a chance to debug other memory problems.
16778 With these changes, a new set of functions and macros have appeared:
16780 CRYPTO_set_mem_debug_functions() [F]
16781 CRYPTO_get_mem_debug_functions() [F]
16782 CRYPTO_dbg_set_options() [F]
16783 CRYPTO_dbg_get_options() [F]
16784 CRYPTO_malloc_debug_init() [M]
16786 The memory debug functions are NULL by default, unless the library
16787 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
16788 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
16789 gives the standard debugging functions that come with OpenSSL) or
16790 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
16791 provided by the library user) must be used. When the standard
16792 debugging functions are used, CRYPTO_dbg_set_options can be used to
16793 request additional information:
16794 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
16795 the CRYPTO_MDEBUG_xxx macro when compiling the library.
16797 Also, things like CRYPTO_set_mem_functions will always give the
16798 expected result (the new set of functions is used for allocation
16799 and deallocation) at all times, regardless of platform and compiler
16802 To finish it up, some functions that were never use in any other
16803 way than through macros have a new API and new semantic:
16805 CRYPTO_dbg_malloc()
16806 CRYPTO_dbg_realloc()
16809 All macros of value have retained their old syntax.
16811 *Richard Levitte and Bodo Moeller*
16813 * Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
16814 ordering of SMIMECapabilities wasn't in "strength order" and there
16815 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
16820 * Some ASN1 types with illegal zero length encoding (INTEGER,
16821 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
16823 *Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson*
16825 * Merge in my S/MIME library for OpenSSL. This provides a simple
16826 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
16827 functionality to handle multipart/signed properly) and a utility
16828 called 'smime' to call all this stuff. This is based on code I
16829 originally wrote for Celo who have kindly allowed it to be
16830 included in OpenSSL.
16834 * Add variants des_set_key_checked and des_set_key_unchecked of
16835 des_set_key (aka des_key_sched). Global variable des_check_key
16836 decides which of these is called by des_set_key; this way
16837 des_check_key behaves as it always did, but applications and
16838 the library itself, which was buggy for des_check_key == 1,
16839 have a cleaner way to pick the version they need.
16843 * New function PKCS12_newpass() which changes the password of a
16848 * Modify X509_TRUST and X509_PURPOSE so it also uses a static and
16849 dynamic mix. In both cases the ids can be used as an index into the
16850 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
16851 functions so they accept a list of the field values and the
16852 application doesn't need to directly manipulate the X509_TRUST
16857 * Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
16862 * Modify the way the V3 extension code looks up extensions. This now
16863 works in a similar way to the object code: we have some "standard"
16864 extensions in a static table which is searched with OBJ_bsearch()
16865 and the application can add dynamic ones if needed. The file
16866 crypto/x509v3/ext_dat.h now has the info: this file needs to be
16867 updated whenever a new extension is added to the core code and kept
16868 in ext_nid order. There is a simple program 'tabtest.c' which checks
16869 this. New extensions are not added too often so this file can readily
16870 be maintained manually.
16872 There are two big advantages in doing things this way. The extensions
16873 can be looked up immediately and no longer need to be "added" using
16874 X509V3_add_standard_extensions(): this function now does nothing.
16875 Side note: I get *lots* of email saying the extension code doesn't
16876 work because people forget to call this function.
16877 Also no dynamic allocation is done unless new extensions are added:
16878 so if we don't add custom extensions there is no need to call
16879 X509V3_EXT_cleanup().
16883 * Modify enc utility's salting as follows: make salting the default. Add a
16884 magic header, so unsalted files fail gracefully instead of just decrypting
16885 to garbage. This is because not salting is a big security hole, so people
16886 should be discouraged from doing it.
16890 * Fixes and enhancements to the 'x509' utility. It allowed a message
16891 digest to be passed on the command line but it only used this
16892 parameter when signing a certificate. Modified so all relevant
16893 operations are affected by the digest parameter including the
16894 -fingerprint and -x509toreq options. Also -x509toreq choked if a
16895 DSA key was used because it didn't fix the digest.
16899 * Initial certificate chain verify code. Currently tests the untrusted
16900 certificates for consistency with the verify purpose (which is set
16901 when the X509_STORE_CTX structure is set up) and checks the pathlength.
16903 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
16904 this is because it will reject chains with invalid extensions whereas
16905 every previous version of OpenSSL and SSLeay made no checks at all.
16907 Trust code: checks the root CA for the relevant trust settings. Trust
16908 settings have an initial value consistent with the verify purpose: e.g.
16909 if the verify purpose is for SSL client use it expects the CA to be
16910 trusted for SSL client use. However the default value can be changed to
16911 permit custom trust settings: one example of this would be to only trust
16912 certificates from a specific "secure" set of CAs.
16914 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
16915 which should be used for version portability: especially since the
16916 verify structure is likely to change more often now.
16918 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
16919 to set them. If not set then assume SSL clients will verify SSL servers
16922 Two new options to the verify program: -untrusted allows a set of
16923 untrusted certificates to be passed in and -purpose which sets the
16924 intended purpose of the certificate. If a purpose is set then the
16925 new chain verify code is used to check extension consistency.
16929 * Support for the authority information access extension.
16933 * Modify RSA and DSA PEM read routines to transparently handle
16934 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
16935 public keys in a format compatible with certificate
16936 SubjectPublicKeyInfo structures. Unfortunately there were already
16937 functions called *_PublicKey_* which used various odd formats so
16938 these are retained for compatibility: however the DSA variants were
16939 never in a public release so they have been deleted. Changed dsa/rsa
16940 utilities to handle the new format: note no releases ever handled public
16941 keys so we should be OK.
16943 The primary motivation for this change is to avoid the same fiasco
16944 that dogs private keys: there are several incompatible private key
16945 formats some of which are standard and some OpenSSL specific and
16946 require various evil hacks to allow partial transparent handling and
16947 even then it doesn't work with DER formats. Given the option anything
16948 other than PKCS#8 should be dumped: but the other formats have to
16949 stay in the name of compatibility.
16951 With public keys and the benefit of hindsight one standard format
16952 is used which works with EVP_PKEY, RSA or DSA structures: though
16953 it clearly returns an error if you try to read the wrong kind of key.
16955 Added a -pubkey option to the 'x509' utility to output the public key.
16956 Also rename the `EVP_PKEY_get_*()` to `EVP_PKEY_rget_*()`
16957 (renamed to `EVP_PKEY_get1_*()` in the OpenSSL 0.9.5 release) and add
16958 `EVP_PKEY_rset_*()` functions (renamed to `EVP_PKEY_set1_*()`)
16959 that do the same as the `EVP_PKEY_assign_*()` except they up the
16960 reference count of the added key (they don't "swallow" the
16965 * Fixes to crypto/x509/by_file.c the code to read in certificates and
16966 CRLs would fail if the file contained no certificates or no CRLs:
16967 added a new function to read in both types and return the number
16968 read: this means that if none are read it will be an error. The
16969 DER versions of the certificate and CRL reader would always fail
16970 because it isn't possible to mix certificates and CRLs in DER format
16971 without choking one or the other routine. Changed this to just read
16972 a certificate: this is the best we can do. Also modified the code
16973 in `apps/verify.c` to take notice of return codes: it was previously
16974 attempting to read in certificates from NULL pointers and ignoring
16975 any errors: this is one reason why the cert and CRL reader seemed
16976 to work. It doesn't check return codes from the default certificate
16977 routines: these may well fail if the certificates aren't installed.
16981 * Code to support otherName option in GeneralName.
16985 * First update to verify code. Change the verify utility
16986 so it warns if it is passed a self signed certificate:
16987 for consistency with the normal behaviour. X509_verify
16988 has been modified to it will now verify a self signed
16989 certificate if *exactly* the same certificate appears
16990 in the store: it was previously impossible to trust a
16991 single self signed certificate. This means that:
16992 openssl verify ss.pem
16993 now gives a warning about a self signed certificate but
16994 openssl verify -CAfile ss.pem ss.pem
16999 * For servers, store verify_result in SSL_SESSION data structure
17000 (and add it to external session representation).
17001 This is needed when client certificate verifications fails,
17002 but an application-provided verification callback (set by
17003 SSL_CTX_set_cert_verify_callback) allows accepting the session
17004 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
17005 but returns 1): When the session is reused, we have to set
17006 ssl->verify_result to the appropriate error code to avoid
17009 *Bodo Moeller, problem pointed out by Lutz Jaenicke*
17011 * Fix a bug in the new PKCS#7 code: it didn't consider the
17012 case in PKCS7_dataInit() where the signed PKCS7 structure
17013 didn't contain any existing data because it was being created.
17015 *Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson*
17017 * Add a salt to the key derivation routines in enc.c. This
17018 forms the first 8 bytes of the encrypted file. Also add a
17019 -S option to allow a salt to be input on the command line.
17023 * New function X509_cmp(). Oddly enough there wasn't a function
17024 to compare two certificates. We do this by working out the SHA1
17025 hash and comparing that. X509_cmp() will be needed by the trust
17030 * SSL_get1_session() is like SSL_get_session(), but increments
17031 the reference count in the SSL_SESSION returned.
17033 *Geoff Thorpe <geoff@eu.c2.net>*
17035 * Fix for 'req': it was adding a null to request attributes.
17036 Also change the X509_LOOKUP and X509_INFO code to handle
17037 certificate auxiliary information.
17041 * Add support for 40 and 64 bit RC2 and RC4 algorithms: document
17046 * Add the possibility to add extra information to the memory leak
17047 detecting output, to form tracebacks, showing from where each
17048 allocation was originated: CRYPTO_push_info("constant string") adds
17049 the string plus current file name and line number to a per-thread
17050 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
17051 is like calling CYRPTO_pop_info() until the stack is empty.
17052 Also updated memory leak detection code to be multi-thread-safe.
17056 * Add options -text and -noout to pkcs7 utility and delete the
17057 encryption options which never did anything. Update docs.
17061 * Add options to some of the utilities to allow the pass phrase
17062 to be included on either the command line (not recommended on
17063 OSes like Unix) or read from the environment. Update the
17064 manpages and fix a few bugs.
17068 * Add a few manpages for some of the openssl commands.
17072 * Fix the -revoke option in ca. It was freeing up memory twice,
17073 leaking and not finding already revoked certificates.
17077 * Extensive changes to support certificate auxiliary information.
17078 This involves the use of X509_CERT_AUX structure and X509_AUX
17079 functions. An X509_AUX function such as PEM_read_X509_AUX()
17080 can still read in a certificate file in the usual way but it
17081 will also read in any additional "auxiliary information". By
17082 doing things this way a fair degree of compatibility can be
17083 retained: existing certificates can have this information added
17084 using the new 'x509' options.
17086 Current auxiliary information includes an "alias" and some trust
17087 settings. The trust settings will ultimately be used in enhanced
17088 certificate chain verification routines: currently a certificate
17089 can only be trusted if it is self signed and then it is trusted
17094 * Fix assembler for Alpha (tested only on DEC OSF not Linux or `*BSD`).
17095 The problem was that one of the replacement routines had not been working
17096 since SSLeay releases. For now the offending routine has been replaced
17097 with non-optimised assembler. Even so, this now gives around 95%
17098 performance improvement for 1024 bit RSA signs.
17102 * Hack to fix PKCS#7 decryption when used with some unorthodox RC2
17103 handling. Most clients have the effective key size in bits equal to
17104 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
17105 A few however don't do this and instead use the size of the decrypted key
17106 to determine the RC2 key length and the AlgorithmIdentifier to determine
17107 the effective key length. In this case the effective key length can still
17108 be 40 bits but the key length can be 168 bits for example. This is fixed
17109 by manually forcing an RC2 key into the EVP_PKEY structure because the
17110 EVP code can't currently handle unusual RC2 key sizes: it always assumes
17111 the key length and effective key length are equal.
17115 * Add a bunch of functions that should simplify the creation of
17116 X509_NAME structures. Now you should be able to do:
17117 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
17118 and have it automatically work out the correct field type and fill in
17119 the structures. The more adventurous can try:
17120 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
17121 and it will (hopefully) work out the correct multibyte encoding.
17125 * Change the 'req' utility to use the new field handling and multibyte
17126 copy routines. Before the DN field creation was handled in an ad hoc
17127 way in req, ca, and x509 which was rather broken and didn't support
17128 BMPStrings or UTF8Strings. Since some software doesn't implement
17129 BMPStrings or UTF8Strings yet, they can be enabled using the config file
17130 using the dirstring_type option. See the new comment in the default
17131 openssl.cnf for more info.
17135 * Make crypto/rand/md_rand.c more robust:
17136 - Assure unique random numbers after fork().
17137 - Make sure that concurrent threads access the global counter and
17138 md serializably so that we never lose entropy in them
17139 or use exactly the same state in multiple threads.
17140 Access to the large state is not always serializable because
17141 the additional locking could be a performance killer, and
17142 md should be large enough anyway.
17146 * New file `apps/app_rand.c` with commonly needed functionality
17147 for handling the random seed file.
17149 Use the random seed file in some applications that previously did not:
17151 dsaparam -genkey (which also ignored its '-rand' option),
17154 x509 (when signing).
17155 Except on systems with /dev/urandom, it is crucial to have a random
17156 seed file at least for key creation, DSA signing, and for DH exchanges;
17157 for RSA signatures we could do without one.
17159 gendh and gendsa (unlike genrsa) used to read only the first byte
17160 of each file listed in the '-rand' option. The function as previously
17161 found in genrsa is now in app_rand.c and is used by all programs
17162 that support '-rand'.
17166 * In RAND_write_file, use mode 0600 for creating files;
17167 don't just chmod when it may be too late.
17171 * Report an error from X509_STORE_load_locations
17172 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
17176 * New function ASN1_mbstring_copy() this copies a string in either
17177 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
17178 into an ASN1_STRING type. A mask of permissible types is passed
17179 and it chooses the "minimal" type to use or an error if not type
17184 * Add function equivalents to the various macros in asn1.h. The old
17185 macros are retained with an `M_` prefix. Code inside the library can
17186 use the `M_` macros. External code (including the openssl utility)
17187 should *NOT* in order to be "shared library friendly".
17191 * Add various functions that can check a certificate's extensions
17192 to see if it usable for various purposes such as SSL client,
17193 server or S/MIME and CAs of these types. This is currently
17194 VERY EXPERIMENTAL but will ultimately be used for certificate chain
17195 verification. Also added a -purpose flag to x509 utility to
17196 print out all the purposes.
17200 * Add a CRYPTO_EX_DATA to X509 certificate structure and associated
17205 * New `X509V3_{X509,CRL,REVOKED}_get_d2i()` functions. These will search
17206 for, obtain and decode and extension and obtain its critical flag.
17207 This allows all the necessary extension code to be handled in a
17208 single function call.
17212 * RC4 tune-up featuring 30-40% performance improvement on most RISC
17213 platforms. See crypto/rc4/rc4_enc.c for further details.
17217 * New -noout option to asn1parse. This causes no output to be produced
17218 its main use is when combined with -strparse and -out to extract data
17219 from a file (which may not be in ASN.1 format).
17223 * Fix for pkcs12 program. It was hashing an invalid certificate pointer
17224 when producing the local key id.
17226 *Richard Levitte <levitte@stacken.kth.se>*
17228 * New option -dhparam in s_server. This allows a DH parameter file to be
17229 stated explicitly. If it is not stated then it tries the first server
17230 certificate file. The previous behaviour hard coded the filename
17235 * Add -pubin and -pubout options to the rsa and dsa commands. These allow
17236 a public key to be input or output. For example:
17237 openssl rsa -in key.pem -pubout -out pubkey.pem
17238 Also added necessary DSA public key functions to handle this.
17242 * Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
17243 in the message. This was handled by allowing
17244 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
17246 *Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>*
17248 * Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
17249 to the end of the strings whereas this didn't. This would cause problems
17250 if strings read with d2i_ASN1_bytes() were later modified.
17252 *Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>*
17254 * Fix for base64 decode bug. When a base64 bio reads only one line of
17255 data and it contains EOF it will end up returning an error. This is
17256 caused by input 46 bytes long. The cause is due to the way base64
17257 BIOs find the start of base64 encoded data. They do this by trying a
17258 trial decode on each line until they find one that works. When they
17259 do a flag is set and it starts again knowing it can pass all the
17260 data directly through the decoder. Unfortunately it doesn't reset
17261 the context it uses. This means that if EOF is reached an attempt
17262 is made to pass two EOFs through the context and this causes the
17263 resulting error. This can also cause other problems as well. As is
17264 usual with these problems it takes *ages* to find and the fix is
17265 trivial: move one line.
17267 *Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer)*
17269 * Ugly workaround to get s_client and s_server working under Windows. The
17270 old code wouldn't work because it needed to select() on sockets and the
17271 tty (for keypresses and to see if data could be written). Win32 only
17272 supports select() on sockets so we select() with a 1s timeout on the
17273 sockets and then see if any characters are waiting to be read, if none
17274 are present then we retry, we also assume we can always write data to
17275 the tty. This isn't nice because the code then blocks until we've
17276 received a complete line of data and it is effectively polling the
17277 keyboard at 1s intervals: however it's quite a bit better than not
17278 working at all :-) A dedicated Windows application might handle this
17279 with an event loop for example.
17283 * Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
17284 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
17285 will be called when RSA_sign() and RSA_verify() are used. This is useful
17286 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
17287 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
17288 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
17289 This necessitated the support of an extra signature type NID_md5_sha1
17290 for SSL signatures and modifications to the SSL library to use it instead
17291 of calling RSA_public_decrypt() and RSA_private_encrypt().
17295 * Add new -verify -CAfile and -CApath options to the crl program, these
17296 will lookup a CRL issuers certificate and verify the signature in a
17297 similar way to the verify program. Tidy up the crl program so it
17298 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
17299 less strict. It will now permit CRL extensions even if it is not
17300 a V2 CRL: this will allow it to tolerate some broken CRLs.
17304 * Initialize all non-automatic variables each time one of the openssl
17305 sub-programs is started (this is necessary as they may be started
17306 multiple times from the "OpenSSL>" prompt).
17308 *Lennart Bang, Bodo Moeller*
17310 * Preliminary compilation option RSA_NULL which disables RSA crypto without
17311 removing all other RSA functionality (this is what NO_RSA does). This
17312 is so (for example) those in the US can disable those operations covered
17313 by the RSA patent while allowing storage and parsing of RSA keys and RSA
17318 * Non-copying interface to BIO pairs.
17319 (still largely untested)
17323 * New function ASN1_tag2str() to convert an ASN1 tag to a descriptive
17324 ASCII string. This was handled independently in various places before.
17328 * New functions UTF8_getc() and UTF8_putc() that parse and generate
17329 UTF8 strings a character at a time.
17333 * Use client_version from client hello to select the protocol
17334 (s23_srvr.c) and for RSA client key exchange verification
17335 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
17339 * Add various utility functions to handle SPKACs, these were previously
17340 handled by poking round in the structure internals. Added new function
17341 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
17342 print, verify and generate SPKACs. Based on an original idea from
17343 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
17347 * RIPEMD160 is operational on all platforms and is back in 'make test'.
17351 * Allow the config file extension section to be overwritten on the
17352 command line. Based on an original idea from Massimiliano Pala
17353 <madwolf@comune.modena.it>. The new option is called -extensions
17354 and can be applied to ca, req and x509. Also -reqexts to override
17355 the request extensions in req and -crlexts to override the crl extensions
17360 * Add new feature to the SPKAC handling in ca. Now you can include
17361 the same field multiple times by preceding it by "XXXX." for example:
17364 this is the same syntax as used in the req config file.
17368 * Allow certificate extensions to be added to certificate requests. These
17369 are specified in a 'req_extensions' option of the req section of the
17370 config file. They can be printed out with the -text option to req but
17371 are otherwise ignored at present.
17375 * Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
17376 data read consists of only the final block it would not decrypted because
17377 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
17378 A misplaced 'break' also meant the decrypted final block might not be
17379 copied until the next read.
17383 * Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
17384 a few extra parameters to the DH structure: these will be useful if
17385 for example we want the value of 'q' or implement X9.42 DH.
17389 * Initial support for DSA_METHOD. This is based on the RSA_METHOD and
17390 provides hooks that allow the default DSA functions or functions on a
17391 "per key" basis to be replaced. This allows hardware acceleration and
17392 hardware key storage to be handled without major modification to the
17393 library. Also added low-level modexp hooks and CRYPTO_EX structure and
17394 associated functions.
17398 * Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
17399 as "read only": it can't be written to and the buffer it points to will
17400 not be freed. Reading from a read only BIO is much more efficient than
17401 a normal memory BIO. This was added because there are several times when
17402 an area of memory needs to be read from a BIO. The previous method was
17403 to create a memory BIO and write the data to it, this results in two
17404 copies of the data and an O(n^2) reading algorithm. There is a new
17405 function BIO_new_mem_buf() which creates a read only memory BIO from
17406 an area of memory. Also modified the PKCS#7 routines to use read only
17411 * Bugfix: ssl23_get_client_hello did not work properly when called in
17412 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
17413 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
17414 but a retry condition occurred while trying to read the rest.
17418 * The PKCS7_ENC_CONTENT_new() function was setting the content type as
17419 NID_pkcs7_encrypted by default: this was wrong since this should almost
17420 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
17421 the encrypted data type: this is a more sensible place to put it and it
17422 allows the PKCS#12 code to be tidied up that duplicated this
17427 * Changed obj_dat.pl script so it takes its input and output files on
17428 the command line. This should avoid shell escape redirection problems
17433 * Initial support for certificate extension requests, these are included
17434 in things like Xenroll certificate requests. Included functions to allow
17435 extensions to be obtained and added.
17439 * -crlf option to s_client and s_server for sending newlines as
17440 CRLF (as required by many protocols).
17444 ### Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
17446 * Install libRSAglue.a when OpenSSL is built with RSAref.
17448 *Ralf S. Engelschall*
17450 * A few more `#ifndef NO_FP_API / #endif` pairs for consistency.
17452 *Andrija Antonijevic <TheAntony2@bigfoot.com>*
17454 * Fix -startdate and -enddate (which was missing) arguments to 'ca'
17459 * New function DSA_dup_DH, which duplicates DSA parameters/keys as
17460 DH parameters/keys (q is lost during that conversion, but the resulting
17461 DH parameters contain its length).
17463 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
17464 much faster than DH_generate_parameters (which creates parameters
17465 where `p = 2*q + 1`), and also the smaller q makes DH computations
17466 much more efficient (160-bit exponentiation instead of 1024-bit
17467 exponentiation); so this provides a convenient way to support DHE
17468 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
17469 utter importance to use
17470 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
17472 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
17473 when such DH parameters are used, because otherwise small subgroup
17474 attacks may become possible!
17478 * Avoid memory leak in i2d_DHparams.
17482 * Allow the -k option to be used more than once in the enc program:
17483 this allows the same encrypted message to be read by multiple recipients.
17487 * New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
17488 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
17489 it will always use the numerical form of the OID, even if it has a short
17494 * Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
17495 method only got called if p,q,dmp1,dmq1,iqmp components were present,
17496 otherwise bn_mod_exp was called. In the case of hardware keys for example
17497 no private key components need be present and it might store extra data
17498 in the RSA structure, which cannot be accessed from bn_mod_exp.
17499 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
17500 private key operations.
17504 * Added support for SPARC Linux.
17508 * pem_password_cb function type incompatibly changed from
17509 typedef int pem_password_cb(char *buf, int size, int rwflag);
17511 ....(char *buf, int size, int rwflag, void *userdata);
17512 so that applications can pass data to their callbacks:
17513 The `PEM[_ASN1]_{read,write}...` functions and macros now take an
17514 additional void * argument, which is just handed through whenever
17515 the password callback is called.
17517 *Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller*
17519 New function SSL_CTX_set_default_passwd_cb_userdata.
17521 Compatibility note: As many C implementations push function arguments
17522 onto the stack in reverse order, the new library version is likely to
17523 interoperate with programs that have been compiled with the old
17524 pem_password_cb definition (PEM_whatever takes some data that
17525 happens to be on the stack as its last argument, and the callback
17526 just ignores this garbage); but there is no guarantee whatsoever that
17529 * The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
17530 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
17531 problems not only on Windows, but also on some Unix platforms.
17532 To avoid problematic command lines, these definitions are now in an
17533 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
17534 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
17538 * MIPS III/IV assembler module is reimplemented.
17542 * More DES library cleanups: remove references to srand/rand and
17543 delete an unused file.
17547 * Add support for the free Netwide assembler (NASM) under Win32,
17548 since not many people have MASM (ml) and it can be hard to obtain.
17549 This is currently experimental but it seems to work OK and pass all
17550 the tests. Check out INSTALL.W32 for info.
17554 * Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
17555 without temporary keys kept an extra copy of the server key,
17556 and connections with temporary keys did not free everything in case
17561 * New function RSA_check_key and new openssl rsa option -check
17562 for verifying the consistency of RSA keys.
17564 *Ulf Moeller, Bodo Moeller*
17566 * Various changes to make Win32 compile work:
17567 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
17568 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
17569 comparison" warnings.
17570 3. Add `sk_<TYPE>_sort` to DEF file generator and do make update.
17574 * Add a debugging option to PKCS#5 v2 key generation function: when
17575 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
17576 derived keys are printed to stderr.
17580 * Copy the flags in ASN1_STRING_dup().
17582 *Roman E. Pavlov <pre@mo.msk.ru>*
17584 * The x509 application mishandled signing requests containing DSA
17585 keys when the signing key was also DSA and the parameters didn't match.
17587 It was supposed to omit the parameters when they matched the signing key:
17588 the verifying software was then supposed to automatically use the CA's
17589 parameters if they were absent from the end user certificate.
17591 Omitting parameters is no longer recommended. The test was also
17592 the wrong way round! This was probably due to unusual behaviour in
17593 EVP_cmp_parameters() which returns 1 if the parameters match.
17594 This meant that parameters were omitted when they *didn't* match and
17595 the certificate was useless. Certificates signed with 'ca' didn't have
17598 *Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>*
17600 * Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
17601 The interface is as follows:
17602 Applications can use
17603 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
17604 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
17605 "off" is now the default.
17606 The library internally uses
17607 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
17608 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
17609 to disable memory-checking temporarily.
17611 Some inconsistent states that previously were possible (and were
17612 even the default) are now avoided.
17614 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
17615 with each memory chunk allocated; this is occasionally more helpful
17616 than just having a counter.
17618 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
17620 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
17625 * Introduce "mode" for SSL structures (with defaults in SSL_CTX),
17626 which largely parallels "options", but is for changing API behaviour,
17627 whereas "options" are about protocol behaviour.
17628 Initial "mode" flags are:
17630 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
17631 a single record has been written.
17632 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
17633 retries use the same buffer location.
17634 (But all of the contents must be
17639 * Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
17642 * Fix problems with no-hmac etc.
17644 *Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>*
17646 * New functions RSA_get_default_method(), RSA_set_method() and
17647 RSA_get_method(). These allows replacement of RSA_METHODs without having
17648 to mess around with the internals of an RSA structure.
17652 * Fix memory leaks in DSA_do_sign and DSA_is_prime.
17653 Also really enable memory leak checks in openssl.c and in some
17656 *Chad C. Mulligan, Bodo Moeller*
17658 * Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
17659 up the length of negative integers. This has now been simplified to just
17660 store the length when it is first determined and use it later, rather
17661 than trying to keep track of where data is copied and updating it to
17663 *Steve Henson, reported by Brien Wheeler <bwheeler@authentica-security.com>*
17665 * Add a new function PKCS7_signatureVerify. This allows the verification
17666 of a PKCS#7 signature but with the signing certificate passed to the
17667 function itself. This contrasts with PKCS7_dataVerify which assumes the
17668 certificate is present in the PKCS#7 structure. This isn't always the
17669 case: certificates can be omitted from a PKCS#7 structure and be
17670 distributed by "out of band" means (such as a certificate database).
17674 * Complete the `PEM_*` macros with DECLARE_PEM versions to replace the
17675 function prototypes in pem.h, also change util/mkdef.pl to add the
17676 necessary function names.
17680 * mk1mf.pl (used by Windows builds) did not properly read the
17681 options set by Configure in the top level Makefile, and Configure
17682 was not even able to write more than one option correctly.
17683 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
17687 * New functions CONF_load_bio() and CONF_load_fp() to allow a config
17688 file to be loaded from a BIO or FILE pointer. The BIO version will
17689 for example allow memory BIOs to contain config info.
17693 * New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
17694 Whoever hopes to achieve shared-library compatibility across versions
17695 must use this, not the compile-time macro.
17696 (Exercise 0.9.4: Which is the minimum library version required by
17698 Note: All this applies only to multi-threaded programs, others don't
17703 * Add missing case to s3_clnt.c state machine -- one of the new SSL tests
17704 through a BIO pair triggered the default case, i.e.
17705 SSLerr(...,SSL_R_UNKNOWN_STATE).
17709 * New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
17710 can use the SSL library even if none of the specific BIOs is
17715 * Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
17716 for the encoded length.
17718 *Jeon KyoungHo <khjeon@sds.samsung.co.kr>*
17720 * Add initial documentation of the X509V3 functions.
17724 * Add a new pair of functions PEM_write_PKCS8PrivateKey() and
17725 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
17726 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
17727 secure PKCS#8 private key format with a high iteration count.
17731 * Fix determination of Perl interpreter: A perl or perl5
17732 *directory* in $PATH was also accepted as the interpreter.
17734 *Ralf S. Engelschall*
17736 * Fix demos/sign/sign.c: well there wasn't anything strictly speaking
17737 wrong with it but it was very old and did things like calling
17738 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
17739 unusual formatting.
17743 * Fix demos/selfsign.c: it used obsolete and deleted functions, changed
17744 to use the new extension code.
17748 * Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
17749 with macros. This should make it easier to change their form, add extra
17750 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
17755 * Add to configuration table a new entry that can specify an alternative
17756 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
17757 according to Mark Crispin <MRC@Panda.COM>.
17761 * DES CBC did not update the IV. Weird.
17765 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
17766 Changing the behaviour of the former might break existing programs --
17767 where IV updating is needed, des_ncbc_encrypt can be used.
17770 * When bntest is run from "make test" it drives bc to check its
17771 calculations, as well as internally checking them. If an internal check
17772 fails, it needs to cause bc to give a non-zero result or make test carries
17773 on without noticing the failure. Fixed.
17777 * DES library cleanups.
17781 * Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
17782 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
17783 ciphers. NOTE: although the key derivation function has been verified
17784 against some published test vectors it has not been extensively tested
17785 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
17790 * Instead of "mkdir -p", which is not fully portable, use new
17791 Perl script "util/mkdir-p.pl".
17795 * Rewrite the way password based encryption (PBE) is handled. It used to
17796 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
17797 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
17798 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
17799 the 'parameter' field of the AlgorithmIdentifier is passed to the
17800 underlying key generation function so it must do its own ASN1 parsing.
17801 This has also changed the EVP_PBE_CipherInit() function which now has a
17802 'parameter' argument instead of literal salt and iteration count values
17803 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
17807 * Support for PKCS#5 v1.5 compatible password based encryption algorithms
17808 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
17809 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
17810 KEY" because this clashed with PKCS#8 unencrypted string. Since this
17811 value was just used as a "magic string" and not used directly its
17812 value doesn't matter.
17816 * Introduce some semblance of const correctness to BN. Shame C doesn't
17821 * "linux-sparc64" configuration (ultrapenguin).
17823 *Ray Miller <ray.miller@oucs.ox.ac.uk>*
17824 "linux-sparc" configuration.
17826 *Christian Forster <fo@hawo.stw.uni-erlangen.de>*
17828 * config now generates no-xxx options for missing ciphers.
17832 * Support the EBCDIC character set (work in progress).
17833 File ebcdic.c not yet included because it has a different license.
17835 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
17837 * Support BS2000/OSD-POSIX.
17839 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
17841 * Make callbacks for key generation use `void *` instead of `char *`.
17845 * Make S/MIME samples compile (not yet tested).
17849 * Additional typesafe stacks.
17853 * New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
17857 ### Changes between 0.9.3 and 0.9.3a [29 May 1999]
17859 * New configuration variant "sco5-gcc".
17861 * Updated some demos.
17863 *Sean O Riordain, Wade Scholine*
17865 * Add missing BIO_free at exit of pkcs12 application.
17869 * Fix memory leak in conf.c.
17873 * Updates for Win32 to assembler version of MD5.
17877 * Set #! path to perl in `apps/der_chop` to where we found it
17878 instead of using a fixed path.
17882 * SHA library changes for irix64-mips4-cc.
17886 * Improvements for VMS support.
17890 ### Changes between 0.9.2b and 0.9.3 [24 May 1999]
17892 * Bignum library bug fix. IRIX 6 passes "make test" now!
17893 This also avoids the problems with SC4.2 and unpatched SC5.
17895 *Andy Polyakov <appro@fy.chalmers.se>*
17897 * New functions sk_num, sk_value and sk_set to replace the previous macros.
17898 These are required because of the typesafe stack would otherwise break
17899 existing code. If old code used a structure member which used to be STACK
17900 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
17901 sk_num or sk_value it would produce an error because the num, data members
17902 are not present in STACK_OF. Now it just produces a warning. sk_set
17903 replaces the old method of assigning a value to sk_value
17904 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
17905 that does this will no longer work (and should use sk_set instead) but
17906 this could be regarded as a "questionable" behaviour anyway.
17910 * Fix most of the other PKCS#7 bugs. The "experimental" code can now
17911 correctly handle encrypted S/MIME data.
17915 * Change type of various DES function arguments from des_cblock
17916 (which means, in function argument declarations, pointer to char)
17917 to des_cblock * (meaning pointer to array with 8 char elements),
17918 which allows the compiler to do more typechecking; it was like
17919 that back in SSLeay, but with lots of ugly casts.
17921 Introduce new type const_des_cblock.
17925 * Reorganise the PKCS#7 library and get rid of some of the more obvious
17926 problems: find RecipientInfo structure that matches recipient certificate
17927 and initialise the ASN1 structures properly based on passed cipher.
17931 * Belatedly make the BN tests actually check the results.
17935 * Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
17936 to and from BNs: it was completely broken. New compilation option
17937 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
17938 key elements as negative integers.
17942 * Reorganize and speed up MD5.
17944 *Andy Polyakov <appro@fy.chalmers.se>*
17948 *Richard Levitte <richard@levitte.org>*
17950 * New option -out to asn1parse to allow the parsed structure to be
17951 output to a file. This is most useful when combined with the -strparse
17952 option to examine the output of things like OCTET STRINGS.
17956 * Make SSL library a little more fool-proof by not requiring any longer
17957 that `SSL_set_{accept,connect}_state` be called before
17958 `SSL_{accept,connect}` may be used (`SSL_set_..._state` is omitted
17959 in many applications because usually everything *appeared* to work as
17960 intended anyway -- now it really works as intended).
17964 * Move openssl.cnf out of lib/.
17968 * Fix various things to let OpenSSL even pass "egcc -pipe -O2 -Wall
17969 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
17970 -Wmissing-declarations -Wnested-externs -Winline" with EGCS 1.1.2+
17972 *Ralf S. Engelschall*
17974 * Various fixes to the EVP and PKCS#7 code. It may now be able to
17975 handle PKCS#7 enveloped data properly.
17977 *Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve*
17979 * Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
17980 copying pointers. The cert_st handling is changed by this in
17981 various ways (and thus what used to be known as ctx->default_cert
17982 is now called ctx->cert, since we don't resort to `s->ctx->[default_]cert`
17983 any longer when s->cert does not give us what we need).
17984 ssl_cert_instantiate becomes obsolete by this change.
17985 As soon as we've got the new code right (possibly it already is?),
17986 we have solved a couple of bugs of the earlier code where s->cert
17987 was used as if it could not have been shared with other SSL structures.
17989 Note that using the SSL API in certain dirty ways now will result
17990 in different behaviour than observed with earlier library versions:
17991 Changing settings for an `SSL_CTX *ctx` after having done s = SSL_new(ctx)
17992 does not influence s as it used to.
17994 In order to clean up things more thoroughly, inside SSL_SESSION
17995 we don't use CERT any longer, but a new structure SESS_CERT
17996 that holds per-session data (if available); currently, this is
17997 the peer's certificate chain and, for clients, the server's certificate
17998 and temporary key. CERT holds only those values that can have
17999 meaningful defaults in an SSL_CTX.
18003 * New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
18004 from the internal representation. Various PKCS#7 fixes: remove some
18005 evil casts and set the enc_dig_alg field properly based on the signing
18010 * Allow PKCS#12 password to be set from the command line or the
18011 environment. Let 'ca' get its config file name from the environment
18012 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
18017 * Allow certificate policies extension to use an IA5STRING for the
18018 organization field. This is contrary to the PKIX definition but
18019 VeriSign uses it and IE5 only recognises this form. Document 'x509'
18024 * Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
18025 without disallowing inline assembler and the like for non-pedantic builds.
18029 * Support Borland C++ builder.
18031 *Janez Jere <jj@void.si>, modified by Ulf Möller*
18037 * SHA-1 cleanups and performance enhancements.
18039 *Andy Polyakov <appro@fy.chalmers.se>*
18041 * Sparc v8plus assembler for the bignum library.
18043 *Andy Polyakov <appro@fy.chalmers.se>*
18045 * Accept any -xxx and +xxx compiler options in Configure.
18049 * Update HPUX configuration.
18053 * Add missing `sk_<type>_unshift()` function to safestack.h
18055 *Ralf S. Engelschall*
18057 * New function SSL_CTX_use_certificate_chain_file that sets the
18058 "extra_cert"s in addition to the certificate. (This makes sense
18059 only for "PEM" format files, as chains as a whole are not
18064 * Support verify_depth from the SSL API.
18065 x509_vfy.c had what can be considered an off-by-one-error:
18066 Its depth (which was not part of the external interface)
18067 was actually counting the number of certificates in a chain;
18068 now it really counts the depth.
18072 * Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
18073 instead of X509err, which often resulted in confusing error
18074 messages since the error codes are not globally unique
18075 (e.g. an alleged error in ssl3_accept when a certificate
18076 didn't match the private key).
18078 * New function SSL_CTX_set_session_id_context that allows to set a default
18079 value (so that you don't need SSL_set_session_id_context for each
18080 connection using the SSL_CTX).
18084 * OAEP decoding bug fix.
18088 * Support INSTALL_PREFIX for package builders, as proposed by
18093 * New Configure options "threads" and "no-threads". For systems
18094 where the proper compiler options are known (currently Solaris
18095 and Linux), "threads" is the default.
18099 * New script util/mklink.pl as a faster substitute for util/mklink.sh.
18103 * Install various scripts to $(OPENSSLDIR)/misc, not to
18104 $(INSTALLTOP)/bin -- they shouldn't clutter directories
18105 such as /usr/local/bin.
18109 * "make linux-shared" to build shared libraries.
18111 *Niels Poppe <niels@netbox.org>*
18113 * New Configure option `no-<cipher>` (rsa, idea, rc5, ...).
18117 * Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
18118 extension adding in x509 utility.
18122 * Remove NOPROTO sections and error code comments.
18126 * Partial rewrite of the DEF file generator to now parse the ANSI
18131 * New Configure options --prefix=DIR and --openssldir=DIR.
18135 * Complete rewrite of the error code script(s). It is all now handled
18136 by one script at the top level which handles error code gathering,
18137 header rewriting and C source file generation. It should be much better
18138 than the old method: it now uses a modified version of Ulf's parser to
18139 read the ANSI prototypes in all header files (thus the old K&R definitions
18140 aren't needed for error creation any more) and do a better job of
18141 translating function codes into names. The old 'ASN1 error code embedded
18142 in a comment' is no longer necessary and it doesn't use .err files which
18143 have now been deleted. Also the error code call doesn't have to appear all
18144 on one line (which resulted in some large lines...).
18148 * Change #include filenames from `<foo.h>` to `<openssl/foo.h>`.
18152 * Change behaviour of ssl2_read when facing length-0 packets: Don't return
18153 0 (which usually indicates a closed connection), but continue reading.
18157 * Fix some race conditions.
18161 * Add support for CRL distribution points extension. Add Certificate
18162 Policies and CRL distribution points documentation.
18166 * Move the autogenerated header file parts to crypto/opensslconf.h.
18170 * Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
18171 8 of keying material. Merlin has also confirmed interop with this fix
18172 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
18174 *Merlin Hughes <merlin@baltimore.ie>*
18176 * Fix lots of warnings.
18178 *Richard Levitte <levitte@stacken.kth.se>*
18180 * In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
18181 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
18183 *Richard Levitte <levitte@stacken.kth.se>*
18185 * Fix problems with sizeof(long) == 8.
18187 *Andy Polyakov <appro@fy.chalmers.se>*
18189 * Change functions to ANSI C.
18193 * Fix typos in error codes.
18195 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller*
18197 * Remove defunct assembler files from Configure.
18201 * SPARC v8 assembler BIGNUM implementation.
18203 *Andy Polyakov <appro@fy.chalmers.se>*
18205 * Support for Certificate Policies extension: both print and set.
18206 Various additions to support the r2i method this uses.
18210 * A lot of constification, and fix a bug in X509_NAME_oneline() that could
18211 return a const string when you are expecting an allocated buffer.
18215 * Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
18216 types DirectoryString and DisplayText.
18220 * Add code to allow r2i extensions to access the configuration database,
18221 add an LHASH database driver and add several ctx helper functions.
18225 * Fix an evil bug in bn_expand2() which caused various BN functions to
18226 fail when they extended the size of a BIGNUM.
18230 * Various utility functions to handle SXNet extension. Modify mkdef.pl to
18231 support typesafe stack.
18235 * Fix typo in SSL_[gs]et_options().
18237 *Nils Frostberg <nils@medcom.se>*
18239 * Delete various functions and files that belonged to the (now obsolete)
18240 old X509V3 handling code.
18244 * New Configure option "rsaref".
18248 * Don't auto-generate pem.h.
18252 * Introduce type-safe ASN.1 SETs.
18256 * Convert various additional casted stacks to type-safe STACK_OF() variants.
18258 *Ben Laurie, Ralf S. Engelschall, Steve Henson*
18260 * Introduce type-safe STACKs. This will almost certainly break lots of code
18261 that links with OpenSSL (well at least cause lots of warnings), but fear
18262 not: the conversion is trivial, and it eliminates loads of evil casts. A
18263 few STACKed things have been converted already. Feel free to convert more.
18264 In the fullness of time, I'll do away with the STACK type altogether.
18268 * Add `openssl ca -revoke <certfile>` facility which revokes a certificate
18269 specified in `<certfile>` by updating the entry in the index.txt file.
18270 This way one no longer has to edit the index.txt file manually for
18271 revoking a certificate. The -revoke option does the gory details now.
18273 *Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall*
18275 * Fix `openssl crl -noout -text` combination where `-noout` killed the
18276 `-text` option at all and this way the `-noout -text` combination was
18277 inconsistent in `openssl crl` with the friends in `openssl x509|rsa|dsa`.
18279 *Ralf S. Engelschall*
18281 * Make sure a corresponding plain text error message exists for the
18282 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
18283 verify callback function determined that a certificate was revoked.
18285 *Ralf S. Engelschall*
18287 * Bugfix: In test/testenc, don't test `openssl <cipher>` for
18288 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
18289 all available ciphers including rc5, which was forgotten until now.
18290 In order to let the testing shell script know which algorithms
18291 are available, a new (up to now undocumented) command
18292 `openssl list-cipher-commands` is used.
18296 * Bugfix: s_client occasionally would sleep in select() when
18297 it should have checked SSL_pending() first.
18301 * New functions DSA_do_sign and DSA_do_verify to provide access to
18302 the raw DSA values prior to ASN.1 encoding.
18306 * Tweaks to Configure
18308 *Niels Poppe <niels@netbox.org>*
18310 * Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
18315 * New variables $(RANLIB) and $(PERL) in the Makefiles.
18319 * New config option to avoid instructions that are illegal on the 80386.
18320 The default code is faster, but requires at least a 486.
18324 * Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
18325 SSL2_SERVER_VERSION (not used at all) macros, which are now the
18326 same as SSL2_VERSION anyway.
18330 * New "-showcerts" option for s_client.
18334 * Still more PKCS#12 integration. Add pkcs12 application to openssl
18335 application. Various cleanups and fixes.
18339 * More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
18340 modify error routines to work internally. Add error codes and PBE init
18341 to library startup routines.
18345 * Further PKCS#12 integration. Added password based encryption, PKCS#8 and
18346 packing functions to asn1 and evp. Changed function names and error
18347 codes along the way.
18351 * PKCS12 integration: and so it begins... First of several patches to
18352 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
18353 objects to objects.h
18357 * Add a new 'indent' option to some X509V3 extension code. Initial ASN1
18358 and display support for Thawte strong extranet extension.
18362 * Add LinuxPPC support.
18364 *Jeff Dubrule <igor@pobox.org>*
18366 * Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
18367 bn_div_words in alpha.s.
18369 *Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie*
18371 * Make sure the RSA OAEP test is skipped under -DRSAref because
18372 OAEP isn't supported when OpenSSL is built with RSAref.
18374 *Ulf Moeller <ulf@fitug.de>*
18376 * Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
18377 so they no longer are missing under -DNOPROTO.
18379 *Soren S. Jorvang <soren@t.dk>*
18381 ### Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
18383 * Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
18384 doesn't work when the session is reused. Coming soon!
18388 * Fix a security hole, that allows sessions to be reused in the wrong
18389 context thus bypassing client cert protection! All software that uses
18390 client certs and session caches in multiple contexts NEEDS PATCHING to
18391 allow session reuse! A fuller solution is in the works.
18393 *Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)*
18395 * Some more source tree cleanups (removed obsolete files
18396 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
18397 permission on "config" script to be executable) and a fix for the INSTALL
18400 *Ulf Moeller <ulf@fitug.de>*
18402 * Remove some legacy and erroneous uses of malloc, free instead of
18405 *Lennart Bang <lob@netstream.se>, with minor changes by Steve*
18407 * Make rsa_oaep_test return non-zero on error.
18409 *Ulf Moeller <ulf@fitug.de>*
18411 * Add support for native Solaris shared libraries. Configure
18412 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
18413 if someone would make that last step automatic.
18415 *Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>*
18417 * ctx_size was not built with the right compiler during "make links". Fixed.
18421 * Change the meaning of 'ALL' in the cipher list. It now means "everything
18422 except NULL ciphers". This means the default cipher list will no longer
18423 enable NULL ciphers. They need to be specifically enabled e.g. with
18424 the string "DEFAULT:eNULL".
18428 * Fix to RSA private encryption routines: if p < q then it would
18429 occasionally produce an invalid result. This will only happen with
18430 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
18434 * Be less restrictive and allow also `perl util/perlpath.pl
18435 /path/to/bin/perl` in addition to `perl util/perlpath.pl /path/to/bin`,
18436 because this way one can also use an interpreter named `perl5` (which is
18437 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
18438 installed as `perl`).
18440 *Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
18442 * Let util/clean-depend.pl work also with older Perl 5.00x versions.
18444 *Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
18446 * Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
18447 advapi32.lib to Win32 build and change the pem test comparison
18448 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
18449 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
18450 and crypto/des/ede_cbcm_enc.c.
18454 * DES quad checksum was broken on big-endian architectures. Fixed.
18458 * Comment out two functions in bio.h that aren't implemented. Fix up the
18459 Win32 test batch file so it (might) work again. The Win32 test batch file
18460 is horrible: I feel ill....
18464 * Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
18465 in e_os.h. Audit of header files to check ANSI and non ANSI
18466 sections: 10 functions were absent from non ANSI section and not exported
18467 from Windows DLLs. Fixed up libeay.num for new functions.
18471 * Make `openssl version` output lines consistent.
18473 *Ralf S. Engelschall*
18475 * Fix Win32 symbol export lists for BIO functions: Added
18476 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
18477 to ms/libeay{16,32}.def.
18479 *Ralf S. Engelschall*
18481 * Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
18482 fine under Unix and passes some trivial tests I've now added. But the
18483 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
18484 added to make sure no one expects that this stuff really works in the
18485 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
18486 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
18489 *Ralf S. Engelschall*
18491 * Fix the generation of two part addresses in perl.
18493 *Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie*
18495 * Add config entry for Linux on MIPS.
18497 *John Tobey <jtobey@channel1.com>*
18499 * Make links whenever Configure is run, unless we are on Windoze.
18503 * Permit extensions to be added to CRLs using crl_section in openssl.cnf.
18504 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
18509 * Add a useful kludge to allow package maintainers to specify compiler and
18510 other platforms details on the command line without having to patch the
18511 Configure script every time: One now can use
18512 `perl Configure <id>:<details>`,
18513 i.e. platform ids are allowed to have details appended
18514 to them (separated by colons). This is treated as there would be a static
18515 pre-configured entry in Configure's %table under key `<id>` with value
18516 `<details>` and `perl Configure <id>` is called. So, when you want to
18517 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
18518 assembler stuff you can use `perl Configure "FreeBSD-elf:pgcc:-O6:::"`
18519 now, which overrides the FreeBSD-elf entry on-the-fly.
18521 *Ralf S. Engelschall*
18523 * Disable new TLS1 ciphersuites by default: they aren't official yet.
18527 * Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
18528 on the `perl Configure ...` command line. This way one can compile
18529 OpenSSL libraries with Position Independent Code (PIC) which is needed
18530 for linking it into DSOs.
18532 *Ralf S. Engelschall*
18534 * Remarkably, export ciphers were totally broken and no-one had noticed!
18539 * Cleaned up the LICENSE document: The official contact for any license
18540 questions now is the OpenSSL core team under openssl-core@openssl.org.
18541 And add a paragraph about the dual-license situation to make sure people
18542 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
18543 to the OpenSSL toolkit.
18545 *Ralf S. Engelschall*
18547 * General source tree makefile cleanups: Made `making xxx in yyy...`
18548 display consistent in the source tree and replaced `/bin/rm` by `rm`.
18549 Additionally cleaned up the `make links` target: Remove unnecessary
18550 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
18551 to speed processing and no longer clutter the display with confusing
18552 stuff. Instead only the actually done links are displayed.
18554 *Ralf S. Engelschall*
18556 * Permit null encryption ciphersuites, used for authentication only. It used
18557 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
18558 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
18563 * Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
18564 signed attributes when verifying signatures (this would break them),
18565 the detached data encoding was wrong and public keys obtained using
18566 X509_get_pubkey() weren't freed.
18570 * Add text documentation for the BUFFER functions. Also added a work around
18571 to a Win95 console bug. This was triggered by the password read stuff: the
18572 last character typed gets carried over to the next fread(). If you were
18573 generating a new cert request using 'req' for example then the last
18574 character of the passphrase would be CR which would then enter the first
18579 * Added the new 'Includes OpenSSL Cryptography Software' button as
18580 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
18581 button and can be used by applications based on OpenSSL to show the
18582 relationship to the OpenSSL project.
18584 *Ralf S. Engelschall*
18586 * Remove confusing variables in function signatures in files
18587 ssl/ssl_lib.c and ssl/ssl.h.
18589 *Lennart Bong <lob@kulthea.stacken.kth.se>*
18591 * Don't install bss_file.c under PREFIX/include/
18593 *Lennart Bong <lob@kulthea.stacken.kth.se>*
18595 * Get the Win32 compile working again. Modify mkdef.pl so it can handle
18596 functions that return function pointers and has support for NT specific
18597 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
18598 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
18599 unsigned to signed types: this was killing the Win32 compile.
18603 * Add new certificate file to stack functions,
18604 SSL_add_dir_cert_subjects_to_stack() and
18605 SSL_add_file_cert_subjects_to_stack(). These largely supplant
18606 SSL_load_client_CA_file(), and can be used to add multiple certs easily
18607 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
18608 This means that Apache-SSL and similar packages don't have to mess around
18609 to add as many CAs as they want to the preferred list.
18613 * Experiment with doxygen documentation. Currently only partially applied to
18615 See <http://www.stack.nl/~dimitri/doxygen/index.html>, and run doxygen with
18616 openssl.doxy as the configuration file.
18620 * Get rid of remaining C++-style comments which strict C compilers hate.
18622 *Ralf S. Engelschall, pointed out by Carlos Amengual*
18624 * Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
18625 compiled in by default: it has problems with large keys.
18629 * Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
18630 DH private keys and/or callback functions which directly correspond to
18631 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
18632 is needed for applications which have to configure certificates on a
18633 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
18635 For the RSA certificate situation is makes no difference, but
18636 for the DSA certificate situation this fixes the "no shared cipher"
18637 problem where the OpenSSL cipher selection procedure failed because the
18638 temporary keys were not overtaken from the context and the API provided
18639 no way to reconfigure them.
18640 The new functions now let applications reconfigure the stuff and they
18641 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
18642 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
18643 non-public-API function ssl_cert_instantiate() is used as a helper
18644 function and also to reduce code redundancy inside ssl_rsa.c.
18646 *Ralf S. Engelschall*
18648 * Move s_server -dcert and -dkey options out of the undocumented feature
18649 area because they are useful for the DSA situation and should be
18650 recognized by the users.
18652 *Ralf S. Engelschall*
18654 * Fix the cipher decision scheme for export ciphers: the export bits are
18655 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
18656 SSL_EXP_MASK. So, the original variable has to be used instead of the
18657 already masked variable.
18659 *Richard Levitte <levitte@stacken.kth.se>*
18661 * Fix `port` variable from `int` to `unsigned int` in crypto/bio/b_sock.c
18663 *Richard Levitte <levitte@stacken.kth.se>*
18665 * Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
18666 from `int` to `unsigned int` because it is a length and initialized by
18667 EVP_DigestFinal() which expects an `unsigned int *`.
18669 *Richard Levitte <levitte@stacken.kth.se>*
18671 * Don't hard-code path to Perl interpreter on shebang line of Configure
18672 script. Instead use the usual Shell->Perl transition trick.
18674 *Ralf S. Engelschall*
18676 * Make `openssl x509 -noout -modulus`' functional also for DSA certificates
18677 (in addition to RSA certificates) to match the behaviour of `openssl dsa
18678 -noout -modulus` as it's already the case for `openssl rsa -noout
18679 -modulus`. For RSA the -modulus is the real "modulus" while for DSA
18680 currently the public key is printed (a decision which was already done by
18681 `openssl dsa -modulus` in the past) which serves a similar purpose.
18682 Additionally the NO_RSA no longer completely removes the whole -modulus
18683 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
18686 *Ralf S. Engelschall*
18688 * Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
18689 BIO. See the source (crypto/evp/bio_ok.c) for more info.
18691 *Arne Ansper <arne@ats.cyber.ee>*
18693 * Dump the old yucky req code that tried (and failed) to allow raw OIDs
18694 to be added. Now both 'req' and 'ca' can use new objects defined in the
18699 * Add cool BIO that does syslog (or event log on NT).
18701 *Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie*
18703 * Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
18704 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
18705 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
18706 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
18710 * Add preliminary config info for new extension code.
18714 * Make RSA_NO_PADDING really use no padding.
18716 *Ulf Moeller <ulf@fitug.de>*
18718 * Generate errors when private/public key check is done.
18722 * Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
18723 for some CRL extensions and new objects added.
18727 * Really fix the ASN1 IMPLICIT bug this time... Partial support for private
18728 key usage extension and fuller support for authority key id.
18732 * Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
18733 padding method for RSA, which is recommended for new applications in PKCS
18734 #1 v2.0 (RFC 2437, October 1998).
18735 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
18736 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
18737 against Bleichbacher's attack on RSA.
18738 *Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
18741 * Updates to the new SSL compression code
18743 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
18745 * Fix so that the version number in the master secret, when passed
18746 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
18747 (because the server will not accept higher), that the version number
18748 is 0x03,0x01, not 0x03,0x00
18750 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
18752 * Run extensive memory leak checks on SSL commands. Fixed *lots* of memory
18753 leaks in `ssl/` relating to new `X509_get_pubkey()` behaviour. Also fixes
18754 in `apps/` and an unrelated leak in `crypto/dsa/dsa_vrf.c`.
18758 * Support for RAW extensions where an arbitrary extension can be
18759 created by including its DER encoding. See `apps/openssl.cnf` for
18764 * Make sure latest Perl versions don't interpret some generated C array
18765 code as Perl array code in the crypto/err/err_genc.pl script.
18767 *Lars Weber <3weber@informatik.uni-hamburg.de>*
18769 * Modify ms/do_ms.bat to not generate assembly language makefiles since
18770 not many people have the assembler. Various Win32 compilation fixes and
18771 update to the INSTALL.W32 file with (hopefully) more accurate Win32
18772 build instructions.
18776 * Modify configure script 'Configure' to automatically create crypto/date.h
18777 file under Win32 and also build pem.h from pem.org. New script
18778 util/mkfiles.pl to create the MINFO file on environments that can't do a
18779 'make files': perl util/mkfiles.pl >MINFO should work.
18783 * Major rework of DES function declarations, in the pursuit of correctness
18784 and purity. As a result, many evil casts evaporated, and some weirdness,
18785 too. You may find this causes warnings in your code. Zapping your evil
18786 casts will probably fix them. Mostly.
18790 * Fix for a typo in asn1.h. Bug fix to object creation script
18791 obj_dat.pl. It considered a zero in an object definition to mean
18792 "end of object": none of the objects in objects.h have any zeros
18793 so it wasn't spotted.
18795 *Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>*
18797 * Add support for Triple DES Cipher Block Chaining with Output Feedback
18798 Masking (CBCM). In the absence of test vectors, the best I have been able
18799 to do is check that the decrypt undoes the encrypt, so far. Send me test
18800 vectors if you have them.
18804 * Correct calculation of key length for export ciphers (too much space was
18805 allocated for null ciphers). This has not been tested!
18809 * Modifications to the mkdef.pl for Win32 DEF file creation. The usage
18810 message is now correct (it understands "crypto" and "ssl" on its
18811 command line). There is also now an "update" option. This will update
18812 the util/ssleay.num and util/libeay.num files with any new functions.
18814 perl util/mkdef.pl crypto ssl update
18815 it will update them.
18819 * Overhauled the Perl interface:
18820 - ported BN stuff to OpenSSL's different BN library
18821 - made the perl/ source tree CVS-aware
18822 - renamed the package from SSLeay to OpenSSL (the files still contain
18823 their history because I've copied them in the repository)
18824 - removed obsolete files (the test scripts will be replaced
18825 by better Test::Harness variants in the future)
18827 *Ralf S. Engelschall*
18829 * First cut for a very conservative source tree cleanup:
18830 1. merge various obsolete readme texts into doc/ssleay.txt
18831 where we collect the old documents and readme texts.
18832 2. remove the first part of files where I'm already sure that we no
18833 longer need them because of three reasons: either they are just temporary
18834 files which were left by Eric or they are preserved original files where
18835 I've verified that the diff is also available in the CVS via "cvs diff
18836 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
18837 the crypto/md/ stuff).
18839 *Ralf S. Engelschall*
18841 * More extension code. Incomplete support for subject and issuer alt
18842 name, issuer and authority key id. Change the i2v function parameters
18843 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
18844 what that's for :-) Fix to ASN1 macro which messed up
18845 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
18849 * Preliminary support for ENUMERATED type. This is largely copied from the
18854 * Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
18856 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
18858 * Make sure `make rehash` target really finds the `openssl` program.
18860 *Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
18862 * Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
18863 like to hear about it if this slows down other processors.
18867 * Add CygWin32 platform information to Configure script.
18869 *Alan Batie <batie@aahz.jf.intel.com>*
18871 * Fixed ms/32all.bat script: `no_asm` -> `no-asm`
18873 *Rainer W. Gerling <gerling@mpg-gv.mpg.de>*
18875 * New program nseq to manipulate netscape certificate sequences
18879 * Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
18884 * Fixes to BN code. Previously the default was to define BN_RECURSION
18885 but the BN code had some problems that would cause failures when
18886 doing certificate verification and some other functions.
18888 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
18890 * Add ASN1 and PEM code to support netscape certificate sequences.
18894 * Add ASN1 and PEM code to support netscape certificate sequences.
18898 * Add several PKIX and private extended key usage OIDs.
18902 * Modify the 'ca' program to handle the new extension code. Modify
18903 openssl.cnf for new extension format, add comments.
18907 * More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
18908 and add a sample to openssl.cnf so req -x509 now adds appropriate
18913 * Continued X509 V3 changes. Add to other makefiles, integrate with the
18914 error code, add initial support to X509_print() and x509 application.
18918 * Takes a deep breath and start adding X509 V3 extension support code. Add
18919 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
18920 stuff is currently isolated and isn't even compiled yet.
18924 * Continuing patches for GeneralizedTime. Fix up certificate and CRL
18925 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
18926 Removed the versions check from X509 routines when loading extensions:
18927 this allows certain broken certificates that don't set the version
18928 properly to be processed.
18932 * Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
18933 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
18934 can still be regenerated with "make depend".
18938 * Spelling mistake in C version of CAST-128.
18940 *Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>*
18942 * Changes to the error generation code. The perl script err-code.pl
18943 now reads in the old error codes and retains the old numbers, only
18944 adding new ones if necessary. It also only changes the .err files if new
18945 codes are added. The makefiles have been modified to only insert errors
18946 when needed (to avoid needlessly modifying header files). This is done
18947 by only inserting errors if the .err file is newer than the auto generated
18948 C file. To rebuild all the error codes from scratch (the old behaviour)
18949 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
18950 or delete all the .err files.
18954 * CAST-128 was incorrectly implemented for short keys. The C version has
18955 been fixed, but is untested. The assembler versions are also fixed, but
18956 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
18957 to regenerate it if needed.
18958 *Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
18959 Hagino <itojun@kame.net>*
18961 * File was opened incorrectly in randfile.c.
18963 *Ulf Möller <ulf@fitug.de>*
18965 * Beginning of support for GeneralizedTime. d2i, i2d, check and print
18966 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
18967 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
18968 al: it's just almost always a UTCTime. Note this patch adds new error
18969 codes so do a "make errors" if there are problems.
18973 * Correct Linux 1 recognition in config.
18975 *Ulf Möller <ulf@fitug.de>*
18977 * Remove pointless MD5 hash when using DSA keys in ca.
18979 *Anonymous <nobody@replay.com>*
18981 * Generate an error if given an empty string as a cert directory. Also
18982 generate an error if handed NULL (previously returned 0 to indicate an
18983 error, but didn't set one).
18985 *Ben Laurie, reported by Anonymous <nobody@replay.com>*
18987 * Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
18991 * Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
18992 parameters. This was causing a warning which killed off the Win32 compile.
18996 * Remove C++ style comments from crypto/bn/bn_local.h.
18998 *Neil Costigan <neil.costigan@celocom.com>*
19000 * The function OBJ_txt2nid was broken. It was supposed to return a nid
19001 based on a text string, looking up short and long names and finally
19002 "dot" format. The "dot" format stuff didn't work. Added new function
19003 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
19004 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
19005 OID is not part of the table.
19009 * Add prototypes to X509 lookup/verify methods, fixing a bug in
19010 X509_LOOKUP_by_alias().
19014 * Sort openssl functions by name.
19018 * Get the `gendsa` command working and add it to the `list` command. Remove
19019 encryption from sample DSA keys (in case anyone is interested the password
19024 * Make *all* `*_free` functions accept a NULL pointer.
19026 *Frans Heymans <fheymans@isaserver.be>*
19028 * If a DH key is generated in s3_srvr.c, don't blow it by trying to use
19031 *Anonymous <nobody@replay.com>*
19033 * s_server should send the CAfile as acceptable CAs, not its own cert.
19035 *Bodo Moeller <3moeller@informatik.uni-hamburg.de>*
19037 * Don't blow it for numeric `-newkey` arguments to `apps/req`.
19039 *Bodo Moeller <3moeller@informatik.uni-hamburg.de>*
19041 * Temp key "for export" tests were wrong in s3_srvr.c.
19043 *Anonymous <nobody@replay.com>*
19045 * Add prototype for temp key callback functions
19046 SSL_CTX_set_tmp_{rsa,dh}_callback().
19050 * Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
19051 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
19055 * X509_name_add_entry() freed the wrong thing after an error.
19057 *Arne Ansper <arne@ats.cyber.ee>*
19059 * rsa_eay.c would attempt to free a NULL context.
19061 *Arne Ansper <arne@ats.cyber.ee>*
19063 * BIO_s_socket() had a broken should_retry() on Windoze.
19065 *Arne Ansper <arne@ats.cyber.ee>*
19067 * BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
19069 *Arne Ansper <arne@ats.cyber.ee>*
19071 * Make sure the already existing X509_STORE->depth variable is initialized
19072 in X509_STORE_new(), but document the fact that this variable is still
19073 unused in the certificate verification process.
19075 *Ralf S. Engelschall*
19077 * Fix the various library and `apps/` files to free up pkeys obtained from
19078 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
19082 * Fix reference counting in X509_PUBKEY_get(). This makes
19083 demos/maurice/example2.c work, amongst others, probably.
19085 *Steve Henson and Ben Laurie*
19087 * First cut of a cleanup for `apps/`. First the `ssleay` program is now named
19088 `openssl` and second, the shortcut symlinks for the `openssl <command>`
19089 are no longer created. This way we have a single and consistent command
19090 line interface `openssl <command>`, similar to `cvs <command>`.
19092 *Ralf S. Engelschall, Paul Sutton and Ben Laurie*
19094 * ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
19095 BIT STRING wrapper always have zero unused bits.
19099 * Add CA.pl, perl version of CA.sh, add extended key usage OID.
19103 * Make the top-level INSTALL documentation easier to understand.
19107 * Makefiles updated to exit if an error occurs in a sub-directory
19108 make (including if user presses ^C) [Paul Sutton]
19110 * Make Montgomery context stuff explicit in RSA data structure.
19114 * Fix build order of pem and err to allow for generated pem.h.
19118 * Fix renumbering bug in X509_NAME_delete_entry().
19122 * Enhanced the err-ins.pl script so it makes the error library number
19123 global and can add a library name. This is needed for external ASN1 and
19124 other error libraries.
19128 * Fixed sk_insert which never worked properly.
19132 * Fix ASN1 macros so they can handle indefinite length constructed
19133 EXPLICIT tags. Some non standard certificates use these: they can now
19138 * Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
19139 into a single doc/ssleay.txt bundle. This way the information is still
19140 preserved but no longer messes up this directory. Now it's new room for
19141 the new set of documentation files.
19143 *Ralf S. Engelschall*
19145 * SETs were incorrectly DER encoded. This was a major pain, because they
19146 shared code with SEQUENCEs, which aren't coded the same. This means that
19147 almost everything to do with SETs or SEQUENCEs has either changed name or
19148 number of arguments.
19150 *Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>*
19152 * Fix test data to work with the above.
19156 * Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
19157 was already fixed by Eric for 0.9.1 it seems.
19159 *Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>*
19161 * Autodetect FreeBSD3.
19165 * Fix various bugs in Configure. This affects the following platforms:
19169 unixware-2.0-pentium
19174 * Eliminate generated files from CVS. Reorder tests to regenerate files
19175 before they are needed.
19179 * Generate Makefile.ssl from Makefile.org (to keep CVS happy).
19183 ### Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
19185 * Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
19186 changed SSLeay to OpenSSL in version strings.
19188 *Ralf S. Engelschall*
19190 * Some fixups to the top-level documents.
19194 * Fixed the nasty bug where rsaref.h was not found under compile-time
19195 because the symlink to include/ was missing.
19197 *Ralf S. Engelschall*
19199 * Incorporated the popular no-RSA/DSA-only patches
19200 which allow to compile a RSA-free SSLeay.
19202 *Andrew Cooke / Interrader Ldt., Ralf S. Engelschall*
19204 * Fixed nasty rehash problem under `make -f Makefile.ssl links`
19205 when "ssleay" is still not found.
19207 *Ralf S. Engelschall*
19209 * Added more platforms to Configure: Cray T3E, HPUX 11,
19211 *Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>*
19213 * Updated the README file.
19215 *Ralf S. Engelschall*
19217 * Added various .cvsignore files in the CVS repository subdirs
19218 to make a "cvs update" really silent.
19220 *Ralf S. Engelschall*
19222 * Recompiled the error-definition header files and added
19223 missing symbols to the Win32 linker tables.
19225 *Ralf S. Engelschall*
19227 * Cleaned up the top-level documents;
19228 o new files: CHANGES and LICENSE
19229 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
19230 o merged COPYRIGHT into LICENSE
19231 o removed obsolete TODO file
19232 o renamed MICROSOFT to INSTALL.W32
19234 *Ralf S. Engelschall*
19236 * Removed dummy files from the 0.9.1b source tree:
19237 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
19238 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
19239 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
19240 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
19241 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
19243 *Ralf S. Engelschall*
19245 * Added various platform portability fixes.
19249 * The Genesis of the OpenSSL rpject:
19250 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
19251 Young and Tim J. Hudson created while they were working for C2Net until
19254 *The OpenSSL Project*
19256 ### Changes between 0.9.0b and 0.9.1b [not released]
19258 * Updated a few CA certificates under certs/
19262 * Changed some BIGNUM api stuff.
19266 * Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
19267 DGUX x86, Linux Alpha, etc.
19271 * New COMP library [crypto/comp/] for SSL Record Layer Compression:
19272 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
19277 * Add -strparse option to asn1pars program which parses nested
19280 *Dr Stephen Henson <shenson@bigfoot.com>*
19282 * Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
19286 * DSA fix for "ca" program.
19290 * Added "-genkey" option to "dsaparam" program.
19294 * Added RIPE MD160 (rmd160) message digest.
19298 * Added -a (all) option to "ssleay version" command.
19302 * Added PLATFORM define which is the id given to Configure.
19306 * Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
19310 * Extended the ASN.1 parser routines.
19314 * Extended BIO routines to support REUSEADDR, seek, tell, etc.
19318 * Added a BN_CTX to the BN library.
19322 * Fixed the weak key values in DES library
19326 * Changed API in EVP library for cipher aliases.
19330 * Added support for RC2/64bit cipher.
19334 * Converted the lhash library to the crypto/mem.c functions.
19338 * Added more recognized ASN.1 object ids.
19342 * Added more RSA padding checks for SSL/TLS.
19346 * Added BIO proxy/filter functionality.
19350 * Added extra_certs to SSL_CTX which can be used
19351 send extra CA certificates to the client in the CA cert chain sending
19352 process. It can be configured with SSL_CTX_add_extra_chain_cert().
19356 * Now Fortezza is denied in the authentication phase because
19357 this is key exchange mechanism is not supported by SSLeay at all.
19361 * Additional PKCS1 checks.
19365 * Support the string "TLSv1" for all TLS v1 ciphers.
19369 * Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
19370 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
19374 * Fixed a few memory leaks.
19378 * Fixed various code and comment typos.
19382 * A minor bug in ssl/s3_clnt.c where there would always be 4 0
19383 bytes sent in the client random.
19385 *Edward Bishop <ebishop@spyglass.com>*
19389 [CVE-2020-1971]: https://www.openssl.org/news/vulnerabilities.html#CVE-2020-1971
19390 [CVE-2020-1967]: https://www.openssl.org/news/vulnerabilities.html#CVE-2020-1967
19391 [CVE-2019-1563]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1563
19392 [CVE-2019-1559]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1559
19393 [CVE-2019-1552]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1552
19394 [CVE-2019-1551]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1551
19395 [CVE-2019-1549]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1549
19396 [CVE-2019-1547]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1547
19397 [CVE-2019-1543]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1543
19398 [CVE-2018-5407]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-5407
19399 [CVE-2018-0739]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0739
19400 [CVE-2018-0737]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0737
19401 [CVE-2018-0735]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0735
19402 [CVE-2018-0734]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0734
19403 [CVE-2018-0733]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0733
19404 [CVE-2018-0732]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0732
19405 [CVE-2017-3738]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3738
19406 [CVE-2017-3737]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3737
19407 [CVE-2017-3736]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3736
19408 [CVE-2017-3735]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3735
19409 [CVE-2017-3733]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3733
19410 [CVE-2017-3732]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3732
19411 [CVE-2017-3731]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3731
19412 [CVE-2017-3730]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3730
19413 [CVE-2016-7055]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7055
19414 [CVE-2016-7054]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7054
19415 [CVE-2016-7053]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7053
19416 [CVE-2016-7052]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7052
19417 [CVE-2016-6309]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6309
19418 [CVE-2016-6308]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6308
19419 [CVE-2016-6307]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6307
19420 [CVE-2016-6306]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6306
19421 [CVE-2016-6305]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6305
19422 [CVE-2016-6304]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6304
19423 [CVE-2016-6303]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6303
19424 [CVE-2016-6302]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6302
19425 [CVE-2016-2183]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2183
19426 [CVE-2016-2182]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2182
19427 [CVE-2016-2181]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2181
19428 [CVE-2016-2180]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2180
19429 [CVE-2016-2179]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2179
19430 [CVE-2016-2178]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2178
19431 [CVE-2016-2177]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2177
19432 [CVE-2016-2176]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2176
19433 [CVE-2016-2109]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2109
19434 [CVE-2016-2107]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2107
19435 [CVE-2016-2106]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2106
19436 [CVE-2016-2105]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2105
19437 [CVE-2016-0800]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0800
19438 [CVE-2016-0799]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0799
19439 [CVE-2016-0798]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0798
19440 [CVE-2016-0797]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0797
19441 [CVE-2016-0705]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0705
19442 [CVE-2016-0702]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0702
19443 [CVE-2016-0701]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0701
19444 [CVE-2015-3197]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3197
19445 [CVE-2015-3196]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3196
19446 [CVE-2015-3195]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3195
19447 [CVE-2015-3194]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3194
19448 [CVE-2015-3193]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3193
19449 [CVE-2015-1793]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1793
19450 [CVE-2015-1792]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1792
19451 [CVE-2015-1791]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1791
19452 [CVE-2015-1790]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1790
19453 [CVE-2015-1789]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1789
19454 [CVE-2015-1788]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1788
19455 [CVE-2015-1787]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1787
19456 [CVE-2015-0293]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0293
19457 [CVE-2015-0291]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0291
19458 [CVE-2015-0290]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0290
19459 [CVE-2015-0289]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0289
19460 [CVE-2015-0288]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0288
19461 [CVE-2015-0287]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0287
19462 [CVE-2015-0286]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0286
19463 [CVE-2015-0285]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0285
19464 [CVE-2015-0209]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0209
19465 [CVE-2015-0208]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0208
19466 [CVE-2015-0207]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0207
19467 [CVE-2015-0206]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0206
19468 [CVE-2015-0205]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0205
19469 [CVE-2015-0204]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0204
19470 [CVE-2014-8275]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-8275
19471 [CVE-2014-5139]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-5139
19472 [CVE-2014-3572]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3572
19473 [CVE-2014-3571]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3571
19474 [CVE-2014-3570]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3570
19475 [CVE-2014-3569]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3569
19476 [CVE-2014-3568]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3568
19477 [CVE-2014-3567]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3567
19478 [CVE-2014-3566]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3566
19479 [CVE-2014-3513]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3513
19480 [CVE-2014-3512]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3512
19481 [CVE-2014-3511]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3511
19482 [CVE-2014-3510]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3510
19483 [CVE-2014-3509]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3509
19484 [CVE-2014-3508]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3508
19485 [CVE-2014-3507]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3507
19486 [CVE-2014-3506]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3506
19487 [CVE-2014-3505]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3505
19488 [CVE-2014-3470]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3470
19489 [CVE-2014-0224]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0224
19490 [CVE-2014-0221]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0221
19491 [CVE-2014-0195]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0195
19492 [CVE-2014-0160]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0160
19493 [CVE-2014-0076]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0076
19494 [CVE-2013-6450]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-6450
19495 [CVE-2013-4353]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-4353
19496 [CVE-2013-0169]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-0169
19497 [CVE-2013-0166]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-0166
19498 [CVE-2012-2686]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2686
19499 [CVE-2012-2333]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2333
19500 [CVE-2012-2110]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2110
19501 [CVE-2012-0884]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0884
19502 [CVE-2012-0050]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0050
19503 [CVE-2012-0027]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0027
19504 [CVE-2011-4619]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4619
19505 [CVE-2011-4577]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4577
19506 [CVE-2011-4576]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4576
19507 [CVE-2011-4109]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4109
19508 [CVE-2011-4108]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4108
19509 [CVE-2011-3210]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-3210
19510 [CVE-2011-3207]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-3207
19511 [CVE-2011-0014]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-0014
19512 [CVE-2010-4252]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-4252
19513 [CVE-2010-4180]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-4180
19514 [CVE-2010-3864]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-3864
19515 [CVE-2010-1633]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-1633
19516 [CVE-2010-0740]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-0740
19517 [CVE-2010-0433]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-0433
19518 [CVE-2009-4355]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-4355
19519 [CVE-2009-3555]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-3555
19520 [CVE-2009-3245]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-3245
19521 [CVE-2009-1386]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1386
19522 [CVE-2009-1379]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1379
19523 [CVE-2009-1378]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1378
19524 [CVE-2009-1377]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1377
19525 [CVE-2009-0789]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0789
19526 [CVE-2009-0591]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0591
19527 [CVE-2009-0590]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0590
19528 [CVE-2008-5077]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-5077
19529 [CVE-2008-1678]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-1678
19530 [CVE-2008-1672]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-1672
19531 [CVE-2008-0891]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-0891
19532 [CVE-2007-5135]: https://www.openssl.org/news/vulnerabilities.html#CVE-2007-5135
19533 [CVE-2007-4995]: https://www.openssl.org/news/vulnerabilities.html#CVE-2007-4995
19534 [CVE-2006-4343]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-4343
19535 [CVE-2006-4339]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-4339
19536 [CVE-2006-3738]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-3738
19537 [CVE-2006-2940]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-2940
19538 [CVE-2006-2937]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-2937
19539 [CVE-2005-2969]: https://www.openssl.org/news/vulnerabilities.html#CVE-2005-2969
19540 [CVE-2004-0112]: https://www.openssl.org/news/vulnerabilities.html#CVE-2004-0112
19541 [CVE-2004-0079]: https://www.openssl.org/news/vulnerabilities.html#CVE-2004-0079
19542 [CVE-2003-0851]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0851
19543 [CVE-2003-0545]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0545
19544 [CVE-2003-0544]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0544
19545 [CVE-2003-0543]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0543
19546 [CVE-2003-0078]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0078
19547 [CVE-2002-0659]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0659
19548 [CVE-2002-0657]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0657
19549 [CVE-2002-0656]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0656
19550 [CVE-2002-0655]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0655