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 * Add new SSL APIs to aid in efficiently implementing TLS/SSL fingerprinting. The
28 SSL_CTRL_GET_IANA_GROUPS control code, exposed as the SSL_get0_iana_groups()
29 function-like macro, retrieves the list of supported groups sent by the peer,
30 and the function SSL_client_hello_get_extension_order() populates a caller-supplied
31 array with the list of extension types present in the ClientHello, in order of
36 * Fixed PEM_write_bio_PKCS8PrivateKey() and PEM_write_bio_PKCS8PrivateKey_nid()
37 to make it possible to use empty passphrase strings.
41 * RNDR and RNDRRS support in provider functions to provide
42 random number generation for Arm CPUs (aarch64).
46 * s_client and s_server apps now explicitly say when the TLS version
47 does not include the renegotiation mechanism. This avoids confusion
48 between that scenario versus when the TLS version includes secure
49 renegotiation but the peer lacks support for it.
53 * The default SSL/TLS security level has been changed from 1 to 2. RSA,
54 DSA and DH keys of 1024 bits and above and less than 2048 bits and ECC keys
55 of 160 bits and above and less than 224 bits were previously accepted by
56 default but are now no longer allowed. By default TLS compression was
57 already disabled in previous OpenSSL versions. At security level 2 it cannot
62 * The SSL_CTX_set_cipher_list family functions now accept ciphers using their
67 * The PVK key derivation function has been moved from b2i_PVK_bio_ex() into
68 the legacy crypto provider as an EVP_KDF. Applications requiring this KDF
69 will need to load the legacy crypto provider.
73 * The various OBJ_* functions have been made thread safe.
77 * CCM8 cipher suites in TLS have been downgraded to security level zero
78 because they use a short authentication tag which lowers their strength.
82 * Subject or issuer names in X.509 objects are now displayed as UTF-8 strings
87 * Parallel dual-prime 1536/2048-bit modular exponentiation for
88 AVX512_IFMA capable processors.
90 *Sergey Kirillov, Andrey Matyukov (Intel Corp)*
95 For OpenSSL 3.0 a [Migration guide][] has been added, so the CHANGES entries
96 listed here are only a brief description.
97 The migration guide contains more detailed information related to new features,
98 breaking changes, and mappings for the large list of deprecated functions.
100 [Migration guide]: https://github.com/openssl/openssl/tree/master/doc/man7/migration_guide.pod
102 ### Changes between 3.0.1 and 3.0.2 [xx XXX xxxx]
104 * Made the AES constant time code for no-asm configurations
105 optional due to the resulting 95% performance degradation.
106 The AES constant time code can be enabled, for no assembly
107 builds, with: ./config no-asm -DOPENSSL_AES_CONST_TIME
111 ### Changes between 3.0.0 and 3.0.1 [14 dec 2021]
113 * Fixed invalid handling of X509_verify_cert() internal errors in libssl
114 Internally libssl in OpenSSL calls X509_verify_cert() on the client side to
115 verify a certificate supplied by a server. That function may return a
116 negative return value to indicate an internal error (for example out of
117 memory). Such a negative return value is mishandled by OpenSSL and will cause
118 an IO function (such as SSL_connect() or SSL_do_handshake()) to not indicate
119 success and a subsequent call to SSL_get_error() to return the value
120 SSL_ERROR_WANT_RETRY_VERIFY. This return value is only supposed to be
121 returned by OpenSSL if the application has previously called
122 SSL_CTX_set_cert_verify_callback(). Since most applications do not do this
123 the SSL_ERROR_WANT_RETRY_VERIFY return value from SSL_get_error() will be
124 totally unexpected and applications may not behave correctly as a result. The
125 exact behaviour will depend on the application but it could result in
126 crashes, infinite loops or other similar incorrect responses.
128 This issue is made more serious in combination with a separate bug in OpenSSL
129 3.0 that will cause X509_verify_cert() to indicate an internal error when
130 processing a certificate chain. This will occur where a certificate does not
131 include the Subject Alternative Name extension but where a Certificate
132 Authority has enforced name constraints. This issue can occur even with valid
138 * Corrected a few file name and file reference bugs in the build,
139 installation and setup scripts, which lead to installation verification
140 failures. Slightly enhanced the installation verification script.
144 * Fixed EVP_PKEY_eq() to make it possible to use it with strictly private
149 * Fixed PVK encoder to properly query for the passphrase.
153 * Multiple fixes in the OSSL_HTTP API functions.
157 * Allow sign extension in OSSL_PARAM_allocate_from_text() for the
158 OSSL_PARAM_INTEGER data type and return error on negative numbers
159 used with the OSSL_PARAM_UNSIGNED_INTEGER data type. Make
160 OSSL_PARAM_BLD_push_BN{,_pad}() return an error on negative numbers.
164 * Allow copying uninitialized digest contexts with EVP_MD_CTX_copy_ex.
168 * Fixed detection of ARMv7 and ARM64 CPU features on FreeBSD.
172 * Multiple threading fixes.
176 * Added NULL digest implementation to keep compatibility with 1.1.1 version.
180 * Allow fetching an operation from the provider that owns an unexportable key
181 as a fallback if that is still allowed by the property query.
185 ### Changes between 1.1.1 and 3.0.0 [7 sep 2021]
187 * TLS_MAX_VERSION, DTLS_MAX_VERSION and DTLS_MIN_VERSION constants are now
192 * The `OPENSSL_s390xcap` environment variable can be used to set bits in the
193 S390X capability vector to zero. This simplifies testing of different code
194 paths on S390X architecture.
198 * Encrypting more than 2^64 TLS records with AES-GCM is disallowed
199 as per FIPS 140-2 IG A.5 "Key/IV Pair Uniqueness Requirements from
200 SP 800-38D". The communication will fail at this point.
204 * The EC_GROUP_clear_free() function is deprecated as there is nothing
205 confidential in EC_GROUP data.
209 * The byte order mark (BOM) character is ignored if encountered at the
210 beginning of a PEM-formatted file.
214 * Added CMS support for the Russian GOST algorithms.
218 * Due to move of the implementation of cryptographic operations
219 to the providers, validation of various operation parameters can
220 be postponed until the actual operation is executed where previously
221 it happened immediately when an operation parameter was set.
223 For example when setting an unsupported curve with
224 EVP_PKEY_CTX_set_ec_paramgen_curve_nid() this function call will not
225 fail but later keygen operations with the EVP_PKEY_CTX will fail.
227 *OpenSSL team members and many third party contributors*
229 * The EVP_get_cipherbyname() function will return NULL for algorithms such as
230 "AES-128-SIV", "AES-128-CBC-CTS" and "CAMELLIA-128-CBC-CTS" which were
231 previously only accessible via low level interfaces. Use EVP_CIPHER_fetch()
232 instead to retrieve these algorithms from a provider.
236 * On build targets where the multilib postfix is set in the build
237 configuration the libdir directory was changing based on whether
238 the lib directory with the multilib postfix exists on the system
239 or not. This unpredictable behavior was removed and eventual
240 multilib postfix is now always added to the default libdir. Use
241 `--libdir=lib` to override the libdir if adding the postfix is
246 * The triple DES key wrap functionality now conforms to RFC 3217 but is
247 no longer interoperable with OpenSSL 1.1.1.
251 * The ERR_GET_FUNC() function was removed. With the loss of meaningful
252 function codes, this function can only cause problems for calling
257 * Add a configurable flag to output date formats as ISO 8601. Does not
258 change the default date format.
262 * Version of MSVC earlier than 1300 could get link warnings, which could
263 be suppressed if the undocumented -DI_CAN_LIVE_WITH_LNK4049 was set.
264 Support for this flag has been removed.
268 * Rework and make DEBUG macros consistent. Remove unused -DCONF_DEBUG,
269 -DBN_CTX_DEBUG, and REF_PRINT. Add a new tracing category and use it for
270 printing reference counts. Rename -DDEBUG_UNUSED to -DUNUSED_RESULT_DEBUG
271 Fix BN_DEBUG_RAND so it compiles and, when set, force DEBUG_RAND to be set
272 also. Rename engine_debug_ref to be ENGINE_REF_PRINT also for consistency.
276 * The signatures of the functions to get and set options on SSL and
277 SSL_CTX objects changed from "unsigned long" to "uint64_t" type.
278 Some source code changes may be required.
282 * The public definitions of conf_method_st and conf_st have been
283 deprecated. They will be made opaque in a future release.
285 *Rich Salz and Tomáš Mráz*
287 * Client-initiated renegotiation is disabled by default. To allow it, use
288 the -client_renegotiation option, the SSL_OP_ALLOW_CLIENT_RENEGOTIATION
289 flag, or the "ClientRenegotiation" config parameter as appropriate.
293 * Add "abspath" and "includedir" pragma's to config files, to prevent,
294 or modify relative pathname inclusion.
298 * OpenSSL includes a cryptographic module that is intended to be FIPS 140-2
299 validated. Please consult the README-FIPS and
300 README-PROVIDERS files, as well as the migration guide.
302 *OpenSSL team members and many third party contributors*
304 * For the key types DH and DHX the allowed settable parameters are now different.
308 * The openssl commands that read keys, certificates, and CRLs now
309 automatically detect the PEM or DER format of the input files.
311 *David von Oheimb, Richard Levitte, and Tomáš Mráz*
313 * Added enhanced PKCS#12 APIs which accept a library context.
317 * The default manual page suffix ($MANSUFFIX) has been changed to "ossl"
321 * Added support for Kernel TLS (KTLS).
323 *Boris Pismenny, John Baldwin and Andrew Gallatin*
325 * Support for RFC 5746 secure renegotiation is now required by default for
326 SSL or TLS connections to succeed.
330 * The signature of the `copy` functional parameter of the
331 EVP_PKEY_meth_set_copy() function has changed so its `src` argument is
332 now `const EVP_PKEY_CTX *` instead of `EVP_PKEY_CTX *`. Similarly
333 the signature of the `pub_decode` functional parameter of the
334 EVP_PKEY_asn1_set_public() function has changed so its `pub` argument is
335 now `const X509_PUBKEY *` instead of `X509_PUBKEY *`.
339 * The error return values from some control calls (ctrl) have changed.
343 * A public key check is now performed during EVP_PKEY_derive_set_peer().
347 * Many functions in the EVP_ namespace that are getters of values from
348 implementations or contexts were renamed to include get or get0 in their
349 names. Old names are provided as macro aliases for compatibility and
354 * The EVP_PKEY_CTRL_PKCS7_ENCRYPT, EVP_PKEY_CTRL_PKCS7_DECRYPT,
355 EVP_PKEY_CTRL_PKCS7_SIGN, EVP_PKEY_CTRL_CMS_ENCRYPT,
356 EVP_PKEY_CTRL_CMS_DECRYPT, and EVP_PKEY_CTRL_CMS_SIGN control operations
361 * The EVP_PKEY_public_check() and EVP_PKEY_param_check() functions now work for
364 * The output from the command line applications may have minor
369 * The output from numerous "printing" may have minor changes.
373 * Windows thread synchronization uses read/write primitives (SRWLock) when
374 supported by the OS, otherwise CriticalSection continues to be used.
378 * Add filter BIO BIO_f_readbuffer() that allows BIO_tell() and BIO_seek() to
379 work on read only BIO source/sinks that do not support these functions.
380 This allows piping or redirection of a file BIO using stdin to be buffered
381 into memory. This is used internally in OSSL_DECODER_from_bio().
385 * OSSL_STORE_INFO_get_type() may now return an additional value. In 1.1.1
386 this function would return one of the values OSSL_STORE_INFO_NAME,
387 OSSL_STORE_INFO_PKEY, OSSL_STORE_INFO_PARAMS, OSSL_STORE_INFO_CERT or
388 OSSL_STORE_INFO_CRL. Decoded public keys would previously have been reported
389 as type OSSL_STORE_INFO_PKEY in 1.1.1. In 3.0 decoded public keys are now
390 reported as having the new type OSSL_STORE_INFO_PUBKEY. Applications
391 using this function should be amended to handle the changed return value.
395 * Improved adherence to Enhanced Security Services (ESS, RFC 2634 and RFC 5035)
396 for the TSP and CMS Advanced Electronic Signatures (CAdES) implementations.
397 As required by RFC 5035 check both ESSCertID and ESSCertIDv2 if both present.
398 Correct the semantics of checking the validation chain in case ESSCertID{,v2}
399 contains more than one certificate identifier: This means that all
400 certificates referenced there MUST be part of the validation chain.
404 * The implementation of older EVP ciphers related to CAST, IDEA, SEED, RC2, RC4,
405 RC5, DESX and DES have been moved to the legacy provider.
409 * The implementation of the EVP digests MD2, MD4, MDC2, WHIRLPOOL and
410 RIPEMD-160 have been moved to the legacy provider.
414 * The deprecated function EVP_PKEY_get0() now returns NULL being called for a
419 * The deprecated functions EVP_PKEY_get0_RSA(),
420 EVP_PKEY_get0_DSA(), EVP_PKEY_get0_EC_KEY(), EVP_PKEY_get0_DH(),
421 EVP_PKEY_get0_hmac(), EVP_PKEY_get0_poly1305() and EVP_PKEY_get0_siphash() as
422 well as the similarly named "get1" functions behave differently in
427 * A number of functions handling low-level keys or engines were deprecated
428 including EVP_PKEY_set1_engine(), EVP_PKEY_get0_engine(), EVP_PKEY_assign(),
429 EVP_PKEY_get0(), EVP_PKEY_get0_hmac(), EVP_PKEY_get0_poly1305() and
430 EVP_PKEY_get0_siphash().
434 * PKCS#5 PBKDF1 key derivation has been moved from PKCS5_PBE_keyivgen() into
435 the legacy crypto provider as an EVP_KDF. Applications requiring this KDF
436 will need to load the legacy crypto provider. This includes these PBE
437 algorithms which use this KDF:
438 - NID_pbeWithMD2AndDES_CBC
439 - NID_pbeWithMD5AndDES_CBC
440 - NID_pbeWithSHA1AndRC2_CBC
441 - NID_pbeWithMD2AndRC2_CBC
442 - NID_pbeWithMD5AndRC2_CBC
443 - NID_pbeWithSHA1AndDES_CBC
447 * Deprecated obsolete BIO_set_callback(), BIO_get_callback(), and
448 BIO_debug_callback() functions.
452 * Deprecated obsolete EVP_PKEY_CTX_get0_dh_kdf_ukm() and
453 EVP_PKEY_CTX_get0_ecdh_kdf_ukm() functions.
457 * The RAND_METHOD APIs have been deprecated.
461 * The SRP APIs have been deprecated.
465 * Add a compile time option to prevent the caching of provider fetched
466 algorithms. This is enabled by including the no-cached-fetch option
467 at configuration time.
471 * pkcs12 now uses defaults of PBKDF2, AES and SHA-256, with a MAC iteration
472 count of PKCS12_DEFAULT_ITER.
474 *Tomáš Mráz and Sahana Prasad*
476 * The openssl speed command does not use low-level API calls anymore.
480 * Parallel dual-prime 1024-bit modular exponentiation for AVX512_IFMA
483 *Ilya Albrekht, Sergey Kirillov, Andrey Matyukov (Intel Corp)*
485 * Combining the Configure options no-ec and no-dh no longer disables TLSv1.3.
489 * Implemented support for fully "pluggable" TLSv1.3 groups. This means that
490 providers may supply their own group implementations (using either the "key
491 exchange" or the "key encapsulation" methods) which will automatically be
492 detected and used by libssl.
494 *Matt Caswell, Nicola Tuveri*
496 * The undocumented function X509_certificate_type() has been deprecated;
500 * Deprecated the obsolete BN_pseudo_rand() and BN_pseudo_rand_range().
504 * Removed RSA padding mode for SSLv23 (which was only used for
505 SSLv2). This includes the functions RSA_padding_check_SSLv23() and
506 RSA_padding_add_SSLv23() and the `-ssl` option in the deprecated
511 * Deprecated the obsolete X9.31 RSA key generation related functions.
513 * While a callback function set via `SSL_CTX_set_cert_verify_callback()`
514 is not allowed to return a value > 1, this is no more taken as failure.
516 *Viktor Dukhovni and David von Oheimb*
518 * Deprecated the obsolete X9.31 RSA key generation related functions
519 BN_X931_generate_Xpq(), BN_X931_derive_prime_ex(), and
520 BN_X931_generate_prime_ex().
524 * The default key generation method for the regular 2-prime RSA keys was
525 changed to the FIPS 186-4 B.3.6 method.
529 * Deprecated the BN_is_prime_ex() and BN_is_prime_fasttest_ex() functions.
533 * Deprecated EVP_MD_CTX_set_update_fn() and EVP_MD_CTX_update_fn().
537 * Deprecated the type OCSP_REQ_CTX and the functions OCSP_REQ_CTX_*() and
538 replaced with OSSL_HTTP_REQ_CTX and the functions OSSL_HTTP_REQ_CTX_*().
540 *Rich Salz, Richard Levitte, and David von Oheimb*
542 * Deprecated `X509_http_nbio()` and `X509_CRL_http_nbio()`.
546 * Deprecated `OCSP_parse_url()`.
550 * Validation of SM2 keys has been separated from the validation of regular EC
555 * Behavior of the `pkey` app is changed, when using the `-check` or `-pubcheck`
556 switches: a validation failure triggers an early exit, returning a failure
557 exit status to the parent process.
561 * Changed behavior of SSL_CTX_set_ciphersuites() and SSL_set_ciphersuites()
562 to ignore unknown ciphers.
566 * The `-cipher-commands` and `-digest-commands` options
567 of the command line utility `list` have been deprecated.
568 Instead use the `-cipher-algorithms` and `-digest-algorithms` options.
572 * Added convenience functions for generating asymmetric key pairs:
573 The 'quick' one-shot (yet somewhat limited) function L<EVP_PKEY_Q_keygen(3)>
574 and macros for the most common cases: <EVP_RSA_gen(3)> and L<EVP_EC_gen(3)>.
578 * All of the low level EC_KEY functions have been deprecated.
580 *Shane Lontis, Paul Dale, Richard Levitte, and Tomáš Mráz*
582 * Deprecated all the libcrypto and libssl error string loading
587 * The functions SSL_CTX_set_tmp_dh_callback and SSL_set_tmp_dh_callback, as
588 well as the macros SSL_CTX_set_tmp_dh() and SSL_set_tmp_dh() have been
593 * The `-crypt` option to the `passwd` command line tool has been removed.
597 * The -C option to the `x509`, `dhparam`, `dsaparam`, and `ecparam` commands
602 * Add support for AES Key Wrap inverse ciphers to the EVP layer.
606 * Deprecated EVP_PKEY_set1_tls_encodedpoint() and
607 EVP_PKEY_get1_tls_encodedpoint().
611 * The security callback, which can be customised by application code, supports
612 the security operation SSL_SECOP_TMP_DH. One location of the "other" parameter
613 was incorrectly passing a DH object. It now passed an EVP_PKEY in all cases.
617 * Add PKCS7_get_octet_string() and PKCS7_type_is_other() to the public
618 interface. Their functionality remains unchanged.
622 * Added new option for 'openssl list', '-providers', which will display the
623 list of loaded providers, their names, version and status. It optionally
624 displays their gettable parameters.
628 * Removed EVP_PKEY_set_alias_type().
632 * Deprecated `EVP_PKEY_CTX_set_rsa_keygen_pubexp()` and introduced
633 `EVP_PKEY_CTX_set1_rsa_keygen_pubexp()`, which is now preferred.
637 * Changed all "STACK" functions to be macros instead of inline functions. Macro
638 parameters are still checked for type safety at compile time via helper
643 * Remove the RAND_DRBG API
645 *Paul Dale and Matthias St. Pierre*
647 * Allow `SSL_set1_host()` and `SSL_add1_host()` to take IP literal addresses
648 as well as actual hostnames.
652 * The 'MinProtocol' and 'MaxProtocol' configuration commands now silently
653 ignore TLS protocol version bounds when configuring DTLS-based contexts, and
654 conversely, silently ignore DTLS protocol version bounds when configuring
655 TLS-based contexts. The commands can be repeated to set bounds of both
656 types. The same applies with the corresponding "min_protocol" and
657 "max_protocol" command-line switches, in case some application uses both TLS
660 SSL_CTX instances that are created for a fixed protocol version (e.g.
661 `TLSv1_server_method()`) also silently ignore version bounds. Previously
662 attempts to apply bounds to these protocol versions would result in an
663 error. Now only the "version-flexible" SSL_CTX instances are subject to
664 limits in configuration files in command-line options.
668 * Deprecated the `ENGINE` API. Engines should be replaced with providers
673 * Reworked the recorded ERR codes to make better space for system errors.
674 To distinguish them, the macro `ERR_SYSTEM_ERROR()` indicates if the
675 given code is a system error (true) or an OpenSSL error (false).
679 * Reworked the test perl framework to better allow parallel testing.
681 *Nicola Tuveri and David von Oheimb*
683 * Added ciphertext stealing algorithms AES-128-CBC-CTS, AES-192-CBC-CTS and
684 AES-256-CBC-CTS to the providers. CS1, CS2 and CS3 variants are supported.
688 * 'Configure' has been changed to figure out the configuration target if
689 none is given on the command line. Consequently, the 'config' script is
690 now only a mere wrapper. All documentation is changed to only mention
693 *Rich Salz and Richard Levitte*
695 * Added a library context `OSSL_LIB_CTX` that applications as well as
696 other libraries can use to form a separate context within which
697 libcrypto operations are performed.
701 * Added various `_ex` functions to the OpenSSL API that support using
702 a non-default `OSSL_LIB_CTX`.
706 * Handshake now fails if Extended Master Secret extension is dropped
711 * Dropped interactive mode from the `openssl` program.
715 * Deprecated `EVP_PKEY_cmp()` and `EVP_PKEY_cmp_parameters()`.
717 *David von Oheimb and Shane Lontis*
719 * Deprecated `EC_METHOD_get_field_type()`.
723 * Deprecated EC_GFp_simple_method(), EC_GFp_mont_method(),
724 EC_GF2m_simple_method(), EC_GFp_nist_method(), EC_GFp_nistp224_method()
725 EC_GFp_nistp256_method(), and EC_GFp_nistp521_method().
729 * Deprecated EC_GROUP_new(), EC_GROUP_method_of(), and EC_POINT_method_of().
733 * Add CAdES-BES signature verification support, mostly derived
734 from ESSCertIDv2 TS (RFC 5816) contribution by Marek Klein.
736 *Filipe Raimundo da Silva*
738 * Add CAdES-BES signature scheme and attributes support (RFC 5126) to CMS API.
742 * Added the AuthEnvelopedData content type structure (RFC 5083) with AES-GCM
743 parameter (RFC 5084) for the Cryptographic Message Syntax (CMS).
747 * Deprecated EC_POINT_make_affine() and EC_POINTs_make_affine().
751 * Deprecated EC_GROUP_precompute_mult(), EC_GROUP_have_precompute_mult(), and
752 EC_KEY_precompute_mult().
756 * Deprecated EC_POINTs_mul().
760 * Removed FIPS_mode() and FIPS_mode_set().
764 * The SSL option SSL_OP_IGNORE_UNEXPECTED_EOF is introduced.
768 * Deprecated EC_POINT_set_Jprojective_coordinates_GFp() and
769 EC_POINT_get_Jprojective_coordinates_GFp().
773 * Added OSSL_PARAM_BLD to the public interface. This allows OSSL_PARAM
774 arrays to be more easily constructed via a series of utility functions.
775 Create a parameter builder using OSSL_PARAM_BLD_new(), add parameters using
776 the various push functions and finally convert to a passable OSSL_PARAM
777 array using OSSL_PARAM_BLD_to_param().
781 * The security strength of SHA1 and MD5 based signatures in TLS has been
786 * Added EVP_PKEY_set_type_by_keymgmt(), to initialise an EVP_PKEY to
787 contain a provider side internal key.
791 * ASN1_verify(), ASN1_digest() and ASN1_sign() have been deprecated.
795 * Project text documents not yet having a proper file name extension
796 (`HACKING`, `LICENSE`, `NOTES*`, `README*`, `VERSION`) have been renamed to
797 `*.md` as far as reasonable, else `*.txt`, for better use with file managers.
801 * The main project documents (README, NEWS, CHANGES, INSTALL, SUPPORT)
802 have been converted to Markdown with the goal to produce documents
803 which not only look pretty when viewed online in the browser, but
804 remain well readable inside a plain text editor.
806 To achieve this goal, a 'minimalistic' Markdown style has been applied
807 which avoids formatting elements that interfere too much with the
808 reading flow in the text file. For example, it
810 * avoids [ATX headings][] and uses [setext headings][] instead
811 (which works for `<h1>` and `<h2>` headings only).
812 * avoids [inline links][] and uses [reference links][] instead.
813 * avoids [fenced code blocks][] and uses [indented code blocks][] instead.
815 [ATX headings]: https://github.github.com/gfm/#atx-headings
816 [setext headings]: https://github.github.com/gfm/#setext-headings
817 [inline links]: https://github.github.com/gfm/#inline-link
818 [reference links]: https://github.github.com/gfm/#reference-link
819 [fenced code blocks]: https://github.github.com/gfm/#fenced-code-blocks
820 [indented code blocks]: https://github.github.com/gfm/#indented-code-blocks
822 *Matthias St. Pierre*
824 * The test suite is changed to preserve results of each test recipe.
825 A new directory test-runs/ with subdirectories named like the
826 test recipes are created in the build tree for this purpose.
830 * Added an implementation of CMP and CRMF (RFC 4210, RFC 4211 RFC 6712).
831 This adds `crypto/cmp/`, `crpyto/crmf/`, `apps/cmp.c`, and `test/cmp_*`.
832 See L<openssl-cmp(1)> and L<OSSL_CMP_exec_IR_ses(3)> as starting points.
834 *David von Oheimb, Martin Peylo*
836 * Generalized the HTTP client code from `crypto/ocsp/` into `crpyto/http/`.
837 It supports arbitrary request and response content types, GET redirection,
838 TLS, connections via HTTP(S) proxies, connections and exchange via
839 user-defined BIOs (allowing implicit connections), persistent connections,
840 and timeout checks. See L<OSSL_HTTP_transfer(3)> etc. for details.
841 The legacy OCSP-focused (and only partly documented) API
842 is retained for backward compatibility, while most of it is deprecated.
846 * Added `util/check-format.pl`, a tool for checking adherence to the
847 OpenSSL coding style <https://www.openssl.org/policies/codingstyle.html>.
848 The checks performed are incomplete and yield some false positives.
849 Still the tool should be useful for detecting most typical glitches.
853 * `BIO_do_connect()` and `BIO_do_handshake()` have been extended:
854 If domain name resolution yields multiple IP addresses all of them are tried
855 after `connect()` failures.
859 * All of the low level RSA functions have been deprecated.
863 * X509 certificates signed using SHA1 are no longer allowed at security
868 * The command line utilities dhparam, dsa, gendsa and dsaparam have been
869 modified to use PKEY APIs. These commands are now in maintenance mode
870 and no new features will be added to them.
874 * The command line utility rsautl has been deprecated.
878 * The command line utilities genrsa and rsa have been modified to use PKEY
879 APIs. They now write PKCS#8 keys by default. These commands are now in
880 maintenance mode and no new features will be added to them.
884 * All of the low level DH functions have been deprecated.
886 *Paul Dale and Matt Caswell*
888 * All of the low level DSA functions have been deprecated.
892 * Reworked the treatment of EC EVP_PKEYs with the SM2 curve to
893 automatically become EVP_PKEY_SM2 rather than EVP_PKEY_EC.
897 * Deprecated low level ECDH and ECDSA functions.
901 * Deprecated EVP_PKEY_decrypt_old() and EVP_PKEY_encrypt_old().
905 * Enhanced the documentation of EVP_PKEY_get_size(), EVP_PKEY_get_bits()
906 and EVP_PKEY_get_security_bits(). Especially EVP_PKEY_get_size() needed
907 a new formulation to include all the things it can be used for,
908 as well as words of caution.
912 * The SSL_CTX_set_tlsext_ticket_key_cb(3) function has been deprecated.
916 * All of the low level HMAC functions have been deprecated.
918 *Paul Dale and David von Oheimb*
920 * Over two thousand fixes were made to the documentation, including:
921 - Common options (such as -rand/-writerand, TLS version control, etc)
922 were refactored and point to newly-enhanced descriptions in openssl.pod.
923 - Added style conformance for all options (with help from Richard Levitte),
924 documented all reported missing options, added a CI build to check
925 that all options are documented and that no unimplemented options
927 - Documented some internals, such as all use of environment variables.
928 - Addressed all internal broken L<> references.
932 * All of the low level CMAC functions have been deprecated.
936 * The low-level MD2, MD4, MD5, MDC2, RIPEMD160 and Whirlpool digest
937 functions have been deprecated.
939 *Paul Dale and David von Oheimb*
941 * Corrected the documentation of the return values from the `EVP_DigestSign*`
942 set of functions. The documentation mentioned negative values for some
943 errors, but this was never the case, so the mention of negative values
946 Code that followed the documentation and thereby check with something
947 like `EVP_DigestSignInit(...) <= 0` will continue to work undisturbed.
951 * All of the low level cipher functions have been deprecated.
953 *Matt Caswell and Paul Dale*
955 * Removed include/openssl/opensslconf.h.in and replaced it with
956 include/openssl/configuration.h.in, which differs in not including
957 <openssl/macros.h>. A short header include/openssl/opensslconf.h
958 was added to include both.
960 This allows internal hacks where one might need to modify the set
961 of configured macros, for example this if deprecated symbols are
962 still supposed to be available internally:
964 #include <openssl/configuration.h>
966 #undef OPENSSL_NO_DEPRECATED
967 #define OPENSSL_SUPPRESS_DEPRECATED
969 #include <openssl/macros.h>
971 This should not be used by applications that use the exported
972 symbols, as that will lead to linking errors.
976 * Fixed an overflow bug in the x64_64 Montgomery squaring procedure
977 used in exponentiation with 512-bit moduli. No EC algorithms are
978 affected. Analysis suggests that attacks against 2-prime RSA1024,
979 3-prime RSA1536, and DSA1024 as a result of this defect would be very
980 difficult to perform and are not believed likely. Attacks against DH512
981 are considered just feasible. However, for an attack the target would
982 have to re-use the DH512 private key, which is not recommended anyway.
983 Also applications directly using the low-level API BN_mod_exp may be
984 affected if they use BN_FLG_CONSTTIME.
989 * Most memory-debug features have been deprecated, and the functionality
990 replaced with no-ops.
994 * Added documentation for the STACK API.
998 * Introduced a new method type and API, OSSL_ENCODER, to represent
999 generic encoders. These do the same sort of job that PEM writers
1000 and d2i functions do, but with support for methods supplied by
1001 providers, and the possibility for providers to support other
1006 * Introduced a new method type and API, OSSL_DECODER, to represent
1007 generic decoders. These do the same sort of job that PEM readers
1008 and i2d functions do, but with support for methods supplied by
1009 providers, and the possibility for providers to support other
1014 * Added a .pragma directive to the syntax of configuration files, to
1015 allow varying behavior in a supported and predictable manner.
1016 Currently added pragma:
1020 This allows dollar signs to be a keyword character unless it's
1021 followed by a opening brace or parenthesis. This is useful for
1022 platforms where dollar signs are commonly used in names, such as
1023 volume names and system directory names on VMS.
1027 * Added functionality to create an EVP_PKEY from user data.
1031 * Change the interpretation of the '--api' configuration option to
1032 mean that this is a desired API compatibility level with no
1033 further meaning. The previous interpretation, that this would
1034 also mean to remove all deprecated symbols up to and including
1035 the given version, no requires that 'no-deprecated' is also used
1036 in the configuration.
1038 When building applications, the desired API compatibility level
1039 can be set with the OPENSSL_API_COMPAT macro like before. For
1040 API compatibility version below 3.0, the old style numerical
1041 value is valid as before, such as -DOPENSSL_API_COMPAT=0x10100000L.
1042 For version 3.0 and on, the value is expected to be the decimal
1043 value calculated from the major and minor version like this:
1045 MAJOR * 10000 + MINOR * 100
1049 -DOPENSSL_API_COMPAT=30000 For 3.0
1050 -DOPENSSL_API_COMPAT=30200 For 3.2
1052 To hide declarations that are deprecated up to and including the
1053 given API compatibility level, -DOPENSSL_NO_DEPRECATED must be
1054 given when building the application as well.
1058 * Added the X509_LOOKUP_METHOD called X509_LOOKUP_store, to allow
1059 access to certificate and CRL stores via URIs and OSSL_STORE
1062 This adds the following functions:
1064 - X509_LOOKUP_store()
1065 - X509_STORE_load_file()
1066 - X509_STORE_load_path()
1067 - X509_STORE_load_store()
1068 - SSL_add_store_cert_subjects_to_stack()
1069 - SSL_CTX_set_default_verify_store()
1070 - SSL_CTX_load_verify_file()
1071 - SSL_CTX_load_verify_dir()
1072 - SSL_CTX_load_verify_store()
1076 * Added a new method to gather entropy on VMS, based on SYS$GET_ENTROPY.
1077 The presence of this system service is determined at run-time.
1081 * Added functionality to create an EVP_PKEY context based on data
1082 for methods from providers. This takes an algorithm name and a
1083 property query string and simply stores them, with the intent
1084 that any operation that uses this context will use those strings
1085 to fetch the needed methods implicitly, thereby making the port
1086 of application written for pre-3.0 OpenSSL easier.
1090 * The undocumented function NCONF_WIN32() has been deprecated; for
1091 conversion details see the HISTORY section of doc/man5/config.pod
1095 * Introduced the new functions EVP_DigestSignInit_ex() and
1096 EVP_DigestVerifyInit_ex(). The macros EVP_DigestSignUpdate() and
1097 EVP_DigestVerifyUpdate() have been converted to functions. See the man
1098 pages for further details.
1102 * Over two thousand fixes were made to the documentation, including:
1103 adding missing command flags, better style conformance, documentation
1106 *Rich Salz, Richard Levitte*
1108 * s390x assembly pack: add hardware-support for P-256, P-384, P-521,
1109 X25519, X448, Ed25519 and Ed448.
1113 * Print all values for a PKCS#12 attribute with 'openssl pkcs12', not just
1118 * Deprecated the public definition of `ERR_STATE` as well as the function
1119 `ERR_get_state()`. This is done in preparation of making `ERR_STATE` an
1124 * Added ERR functionality to give callers access to the stored function
1125 names that have replaced the older function code based functions.
1127 New functions are ERR_peek_error_func(), ERR_peek_last_error_func(),
1128 ERR_peek_error_data(), ERR_peek_last_error_data(), ERR_get_error_all(),
1129 ERR_peek_error_all() and ERR_peek_last_error_all().
1131 Deprecate ERR functions ERR_get_error_line(), ERR_get_error_line_data(),
1132 ERR_peek_error_line_data(), ERR_peek_last_error_line_data() and
1133 ERR_func_error_string().
1137 * Extended testing to be verbose for failing tests only. The make variables
1138 VERBOSE_FAILURE or VF can be used to enable this:
1140 $ make VF=1 test # Unix
1141 $ mms /macro=(VF=1) test ! OpenVMS
1142 $ nmake VF=1 test # Windows
1146 * Added the `-copy_extensions` option to the `x509` command for use with
1147 `-req` and `-x509toreq`. When given with the `copy` or `copyall` argument,
1148 all extensions in the request are copied to the certificate or vice versa.
1150 *David von Oheimb*, *Kirill Stefanenkov <kirill_stefanenkov@rambler.ru>*
1152 * Added the `-copy_extensions` option to the `req` command for use with
1153 `-x509`. When given with the `copy` or `copyall` argument,
1154 all extensions in the certification request are copied to the certificate.
1158 * The `x509`, `req`, and `ca` commands now make sure that X.509v3 certificates
1159 they generate are by default RFC 5280 compliant in the following sense:
1160 There is a subjectKeyIdentifier extension with a hash value of the public key
1161 and for not self-signed certs there is an authorityKeyIdentifier extension
1162 with a keyIdentifier field or issuer information identifying the signing key.
1163 This is done unless some configuration overrides the new default behavior,
1164 such as `subjectKeyIdentifier = none` and `authorityKeyIdentifier = none`.
1168 * Added several checks to `X509_verify_cert()` according to requirements in
1169 RFC 5280 in case `X509_V_FLAG_X509_STRICT` is set
1170 (which may be done by using the CLI option `-x509_strict`):
1171 * The basicConstraints of CA certificates must be marked critical.
1172 * CA certificates must explicitly include the keyUsage extension.
1173 * If a pathlenConstraint is given the key usage keyCertSign must be allowed.
1174 * The issuer name of any certificate must not be empty.
1175 * The subject name of CA certs, certs with keyUsage crlSign,
1176 and certs without subjectAlternativeName must not be empty.
1177 * If a subjectAlternativeName extension is given it must not be empty.
1178 * The signatureAlgorithm field and the cert signature must be consistent.
1179 * Any given authorityKeyIdentifier and any given subjectKeyIdentifier
1180 must not be marked critical.
1181 * The authorityKeyIdentifier must be given for X.509v3 certs
1182 unless they are self-signed.
1183 * The subjectKeyIdentifier must be given for all X.509v3 CA certs.
1187 * Certificate verification using `X509_verify_cert()` meanwhile rejects EC keys
1188 with explicit curve parameters (specifiedCurve) as required by RFC 5480.
1192 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
1193 used even when parsing explicit parameters, when loading a encoded key
1194 or calling `EC_GROUP_new_from_ecpkparameters()`/
1195 `EC_GROUP_new_from_ecparameters()`.
1196 This prevents bypass of security hardening and performance gains,
1197 especially for curves with specialized EC_METHODs.
1198 By default, if a key encoded with explicit parameters is loaded and later
1199 encoded, the output is still encoded with explicit parameters, even if
1200 internally a "named" EC_GROUP is used for computation.
1204 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
1205 this change, EC_GROUP_set_generator would accept order and/or cofactor as
1206 NULL. After this change, only the cofactor parameter can be NULL. It also
1207 does some minimal sanity checks on the passed order.
1212 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
1213 An attack is simple, if the first CMS_recipientInfo is valid but the
1214 second CMS_recipientInfo is chosen ciphertext. If the second
1215 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
1216 encryption key will be replaced by garbage, and the message cannot be
1217 decoded, but if the RSA decryption fails, the correct encryption key is
1218 used and the recipient will not notice the attack.
1219 As a work around for this potential attack the length of the decrypted
1220 key must be equal to the cipher default key length, in case the
1221 certifiate is not given and all recipientInfo are tried out.
1222 The old behaviour can be re-enabled in the CMS code by setting the
1223 CMS_DEBUG_DECRYPT flag.
1227 * Early start up entropy quality from the DEVRANDOM seed source has been
1228 improved for older Linux systems. The RAND subsystem will wait for
1229 /dev/random to be producing output before seeding from /dev/urandom.
1230 The seeded state is stored for future library initialisations using
1231 a system global shared memory segment. The shared memory identifier
1232 can be configured by defining OPENSSL_RAND_SEED_DEVRANDOM_SHM_ID to
1233 the desired value. The default identifier is 114.
1237 * Revised BN_generate_prime_ex to not avoid factors 2..17863 in p-1
1238 when primes for RSA keys are computed.
1239 Since we previously always generated primes == 2 (mod 3) for RSA keys,
1240 the 2-prime and 3-prime RSA modules were easy to distinguish, since
1241 `N = p*q = 1 (mod 3)`, but `N = p*q*r = 2 (mod 3)`. Therefore fingerprinting
1242 2-prime vs. 3-prime RSA keys was possible by computing N mod 3.
1243 This avoids possible fingerprinting of newly generated RSA modules.
1247 * Correct the extended master secret constant on EBCDIC systems. Without this
1248 fix TLS connections between an EBCDIC system and a non-EBCDIC system that
1249 negotiate EMS will fail. Unfortunately this also means that TLS connections
1250 between EBCDIC systems with this fix, and EBCDIC systems without this
1251 fix will fail if they negotiate EMS.
1255 * Changed the library initialisation so that the config file is now loaded
1256 by default. This was already the case for libssl. It now occurs for both
1257 libcrypto and libssl. Use the OPENSSL_INIT_NO_LOAD_CONFIG option to
1258 `OPENSSL_init_crypto()` to suppress automatic loading of a config file.
1262 * Introduced new error raising macros, `ERR_raise()` and `ERR_raise_data()`,
1263 where the former acts as a replacement for `ERR_put_error()`, and the
1264 latter replaces the combination `ERR_put_error()` + `ERR_add_error_data()`.
1265 `ERR_raise_data()` adds more flexibility by taking a format string and
1266 an arbitrary number of arguments following it, to be processed with
1271 * Introduced a new function, `OSSL_PROVIDER_available()`, which can be used
1272 to check if a named provider is loaded and available. When called, it
1273 will also activate all fallback providers if such are still present.
1277 * Enforce a minimum DH modulus size of 512 bits.
1281 * Changed DH parameters to generate the order q subgroup instead of 2q.
1282 Previously generated DH parameters are still accepted by DH_check
1283 but DH_generate_key works around that by clearing bit 0 of the
1284 private key for those. This avoids leaking bit 0 of the private key.
1288 * Significantly reduce secure memory usage by the randomness pools.
1292 * `{CRYPTO,OPENSSL}_mem_debug_{push,pop}` are now no-ops and have been
1297 * A new type, EVP_KEYEXCH, has been introduced to represent key exchange
1298 algorithms. An implementation of a key exchange algorithm can be obtained
1299 by using the function EVP_KEYEXCH_fetch(). An EVP_KEYEXCH algorithm can be
1300 used in a call to EVP_PKEY_derive_init_ex() which works in a similar way to
1301 the older EVP_PKEY_derive_init() function. See the man pages for the new
1302 functions for further details.
1306 * The EVP_PKEY_CTX_set_dh_pad() macro has now been converted to a function.
1310 * Removed the function names from error messages and deprecated the
1315 * Removed NextStep support and the macro OPENSSL_UNISTD
1319 * Removed DES_check_key. Also removed OPENSSL_IMPLEMENT_GLOBAL,
1320 OPENSSL_GLOBAL_REF, OPENSSL_DECLARE_GLOBAL.
1321 Also removed "export var as function" capability; we do not export
1322 variables, only functions.
1326 * RC5_32_set_key has been changed to return an int type, with 0 indicating
1327 an error and 1 indicating success. In previous versions of OpenSSL this
1328 was a void type. If a key was set longer than the maximum possible this
1333 * Support SM2 signing and verification schemes with X509 certificate.
1337 * Use SHA256 as the default digest for TS query in the `ts` app.
1341 * Change PBKDF2 to conform to SP800-132 instead of the older PKCS5 RFC2898.
1345 * Default cipher lists/suites are now available via a function, the
1346 #defines are deprecated.
1350 * Add target VC-WIN32-UWP, VC-WIN64A-UWP, VC-WIN32-ARM-UWP and
1351 VC-WIN64-ARM-UWP in Windows OneCore target for making building libraries
1352 for Windows Store apps easier. Also, the "no-uplink" option has been added.
1356 * Join the directories crypto/x509 and crypto/x509v3
1360 * Added command 'openssl kdf' that uses the EVP_KDF API.
1364 * Added command 'openssl mac' that uses the EVP_MAC API.
1368 * Added OPENSSL_info() to get diverse built-in OpenSSL data, such
1369 as default directories. Also added the command 'openssl info'
1370 for scripting purposes.
1374 * The functions AES_ige_encrypt() and AES_bi_ige_encrypt() have been
1379 * Add prediction resistance to the DRBG reseeding process.
1383 * Limit the number of blocks in a data unit for AES-XTS to 2^20 as
1384 mandated by IEEE Std 1619-2018.
1388 * Added newline escaping functionality to a filename when using openssl dgst.
1389 This output format is to replicate the output format found in the `*sum`
1390 checksum programs. This aims to preserve backward compatibility.
1392 *Matt Eaton, Richard Levitte, and Paul Dale*
1394 * Removed the heartbeat message in DTLS feature, as it has very
1395 little usage and doesn't seem to fulfill a valuable purpose.
1396 The configuration option is now deprecated.
1400 * Changed the output of 'openssl {digestname} < file' to display the
1401 digest name in its output.
1405 * Added a new generic trace API which provides support for enabling
1406 instrumentation through trace output.
1408 *Richard Levitte & Matthias St. Pierre*
1410 * Added build tests for C++. These are generated files that only do one
1411 thing, to include one public OpenSSL head file each. This tests that
1412 the public header files can be usefully included in a C++ application.
1414 This test isn't enabled by default. It can be enabled with the option
1415 'enable-buildtest-c++'.
1419 * Added KB KDF (EVP_KDF_KB) to EVP_KDF.
1423 * Added SSH KDF (EVP_KDF_SSHKDF) and KRB5 KDF (EVP_KDF_KRB5KDF) to EVP_KDF.
1427 * Added Single Step KDF (EVP_KDF_SS), X963 KDF, and X942 KDF to EVP_KDF.
1431 * Added KMAC to EVP_MAC.
1435 * Added property based algorithm implementation selection framework to
1440 * Added SCA hardening for modular field inversion in EC_GROUP through
1441 a new dedicated field_inv() pointer in EC_METHOD.
1442 This also addresses a leakage affecting conversions from projective
1443 to affine coordinates.
1445 *Billy Bob Brumley, Nicola Tuveri*
1447 * Added EVP_KDF, an EVP layer KDF API, to simplify adding KDF and PRF
1448 implementations. This includes an EVP_PKEY to EVP_KDF bridge for
1449 those algorithms that were already supported through the EVP_PKEY API
1450 (scrypt, TLS1 PRF and HKDF). The low-level KDF functions for PBKDF2
1451 and scrypt are now wrappers that call EVP_KDF.
1455 * Build devcrypto engine as a dynamic engine.
1457 *Eneas U de Queiroz*
1459 * Add keyed BLAKE2 to EVP_MAC.
1463 * Fix a bug in the computation of the endpoint-pair shared secret used
1464 by DTLS over SCTP. This breaks interoperability with older versions
1465 of OpenSSL like OpenSSL 1.1.0 and OpenSSL 1.0.2. There is a runtime
1466 switch SSL_MODE_DTLS_SCTP_LABEL_LENGTH_BUG (off by default) enabling
1467 interoperability with such broken implementations. However, enabling
1468 this switch breaks interoperability with correct implementations.
1470 * Fix a use after free bug in d2i_X509_PUBKEY when overwriting a
1471 re-used X509_PUBKEY object if the second PUBKEY is malformed.
1475 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
1479 * Changed the license to the Apache License v2.0.
1483 * Switch to a new version scheme using three numbers MAJOR.MINOR.PATCH.
1485 - Major releases (indicated by incrementing the MAJOR release number)
1486 may introduce incompatible API/ABI changes.
1487 - Minor releases (indicated by incrementing the MINOR release number)
1488 may introduce new features but retain API/ABI compatibility.
1489 - Patch releases (indicated by incrementing the PATCH number)
1490 are intended for bug fixes and other improvements of existing
1491 features only (like improving performance or adding documentation)
1492 and retain API/ABI compatibility.
1496 * Add support for RFC5297 SIV mode (siv128), including AES-SIV.
1500 * Remove the 'dist' target and add a tarball building script. The
1501 'dist' target has fallen out of use, and it shouldn't be
1502 necessary to configure just to create a source distribution.
1506 * Recreate the OS390-Unix config target. It no longer relies on a
1507 special script like it did for OpenSSL pre-1.1.0.
1511 * Instead of having the source directories listed in Configure, add
1512 a 'build.info' keyword SUBDIRS to indicate what sub-directories to
1517 * Add GMAC to EVP_MAC.
1521 * Ported the HMAC, CMAC and SipHash EVP_PKEY_METHODs to EVP_MAC.
1525 * Added EVP_MAC, an EVP layer MAC API, to simplify adding MAC
1526 implementations. This includes a generic EVP_PKEY to EVP_MAC bridge,
1527 to facilitate the continued use of MACs through raw private keys in
1528 functionality such as `EVP_DigestSign*` and `EVP_DigestVerify*`.
1532 * Deprecate ECDH_KDF_X9_62().
1536 * Added EVP_PKEY_ECDH_KDF_X9_63 and ecdh_KDF_X9_63() as replacements for
1537 the EVP_PKEY_ECDH_KDF_X9_62 KDF type and ECDH_KDF_X9_62(). The old names
1538 are retained for backwards compatibility.
1542 * AES-XTS mode now enforces that its two keys are different to mitigate
1543 the attacked described in "Efficient Instantiations of Tweakable
1544 Blockciphers and Refinements to Modes OCB and PMAC" by Phillip Rogaway.
1545 Details of this attack can be obtained from:
1546 <http://web.cs.ucdavis.edu/%7Erogaway/papers/offsets.pdf>
1550 * Rename the object files, i.e. give them other names than in previous
1551 versions. Their names now include the name of the final product, as
1552 well as its type mnemonic (bin, lib, shlib).
1556 * Added new option for 'openssl list', '-objects', which will display the
1557 list of built in objects, i.e. OIDs with names.
1561 * Added the options `-crl_lastupdate` and `-crl_nextupdate` to `openssl ca`,
1562 allowing the `lastUpdate` and `nextUpdate` fields in the generated CRL to
1567 * Added support for Linux Kernel TLS data-path. The Linux Kernel data-path
1568 improves application performance by removing data copies and providing
1569 applications with zero-copy system calls such as sendfile and splice.
1573 * The SSL option SSL_OP_CLEANSE_PLAINTEXT is introduced.
1577 * `PKCS12_parse` now maintains the order of the parsed certificates
1578 when outputting them via `*ca` (rather than reversing it).
1582 * Deprecated pthread fork support methods.
1586 * Added support for FFDHE key exchange in TLS 1.3.
1590 * Added a new concept for OpenSSL plugability: providers. This
1591 functionality is designed to replace the ENGINE API and ENGINE
1592 implementations, and to be much more dynamic, allowing provider
1593 authors to introduce new algorithms among other things, as long as
1594 there's an API that supports the algorithm type.
1596 With this concept comes a new core API for interaction between
1597 libcrypto and provider implementations. Public libcrypto functions
1598 that want to use providers do so through this core API.
1600 The main documentation for this core API is found in
1601 doc/man7/provider.pod, doc/man7/provider-base.pod, and they in turn
1602 refer to other manuals describing the API specific for supported
1603 algorithm types (also called operations).
1610 ### Changes between 1.1.1l and 1.1.1m [xx XXX xxxx]
1612 * Avoid loading of a dynamic engine twice.
1616 * Prioritise DANE TLSA issuer certs over peer certs
1620 * Fixed random API for MacOS prior to 10.12
1622 These MacOS versions don't support the CommonCrypto APIs
1626 ### Changes between 1.1.1k and 1.1.1l [24 Aug 2021]
1628 * Fixed an SM2 Decryption Buffer Overflow.
1630 In order to decrypt SM2 encrypted data an application is expected to
1631 call the API function EVP_PKEY_decrypt(). Typically an application will
1632 call this function twice. The first time, on entry, the "out" parameter
1633 can be NULL and, on exit, the "outlen" parameter is populated with the
1634 buffer size required to hold the decrypted plaintext. The application
1635 can then allocate a sufficiently sized buffer and call EVP_PKEY_decrypt()
1636 again, but this time passing a non-NULL value for the "out" parameter.
1638 A bug in the implementation of the SM2 decryption code means that the
1639 calculation of the buffer size required to hold the plaintext returned
1640 by the first call to EVP_PKEY_decrypt() can be smaller than the actual
1641 size required by the second call. This can lead to a buffer overflow
1642 when EVP_PKEY_decrypt() is called by the application a second time with
1643 a buffer that is too small.
1645 A malicious attacker who is able present SM2 content for decryption to
1646 an application could cause attacker chosen data to overflow the buffer
1647 by up to a maximum of 62 bytes altering the contents of other data held
1648 after the buffer, possibly changing application behaviour or causing
1649 the application to crash. The location of the buffer is application
1650 dependent but is typically heap allocated.
1655 * Fixed various read buffer overruns processing ASN.1 strings
1657 ASN.1 strings are represented internally within OpenSSL as an ASN1_STRING
1658 structure which contains a buffer holding the string data and a field
1659 holding the buffer length. This contrasts with normal C strings which
1660 are repesented as a buffer for the string data which is terminated
1661 with a NUL (0) byte.
1663 Although not a strict requirement, ASN.1 strings that are parsed using
1664 OpenSSL's own "d2i" functions (and other similar parsing functions) as
1665 well as any string whose value has been set with the ASN1_STRING_set()
1666 function will additionally NUL terminate the byte array in the
1667 ASN1_STRING structure.
1669 However, it is possible for applications to directly construct valid
1670 ASN1_STRING structures which do not NUL terminate the byte array by
1671 directly setting the "data" and "length" fields in the ASN1_STRING
1672 array. This can also happen by using the ASN1_STRING_set0() function.
1674 Numerous OpenSSL functions that print ASN.1 data have been found to
1675 assume that the ASN1_STRING byte array will be NUL terminated, even
1676 though this is not guaranteed for strings that have been directly
1677 constructed. Where an application requests an ASN.1 structure to be
1678 printed, and where that ASN.1 structure contains ASN1_STRINGs that have
1679 been directly constructed by the application without NUL terminating
1680 the "data" field, then a read buffer overrun can occur.
1682 The same thing can also occur during name constraints processing
1683 of certificates (for example if a certificate has been directly
1684 constructed by the application instead of loading it via the OpenSSL
1685 parsing functions, and the certificate contains non NUL terminated
1686 ASN1_STRING structures). It can also occur in the X509_get1_email(),
1687 X509_REQ_get1_email() and X509_get1_ocsp() functions.
1689 If a malicious actor can cause an application to directly construct an
1690 ASN1_STRING and then process it through one of the affected OpenSSL
1691 functions then this issue could be hit. This might result in a crash
1692 (causing a Denial of Service attack). It could also result in the
1693 disclosure of private memory contents (such as private keys, or
1694 sensitive plaintext).
1699 ### Changes between 1.1.1j and 1.1.1k [25 Mar 2021]
1701 * Fixed a problem with verifying a certificate chain when using the
1702 X509_V_FLAG_X509_STRICT flag. This flag enables additional security checks of
1703 the certificates present in a certificate chain. It is not set by default.
1705 Starting from OpenSSL version 1.1.1h a check to disallow certificates in
1706 the chain that have explicitly encoded elliptic curve parameters was added
1707 as an additional strict check.
1709 An error in the implementation of this check meant that the result of a
1710 previous check to confirm that certificates in the chain are valid CA
1711 certificates was overwritten. This effectively bypasses the check
1712 that non-CA certificates must not be able to issue other certificates.
1714 If a "purpose" has been configured then there is a subsequent opportunity
1715 for checks that the certificate is a valid CA. All of the named "purpose"
1716 values implemented in libcrypto perform this check. Therefore, where
1717 a purpose is set the certificate chain will still be rejected even when the
1718 strict flag has been used. A purpose is set by default in libssl client and
1719 server certificate verification routines, but it can be overridden or
1720 removed by an application.
1722 In order to be affected, an application must explicitly set the
1723 X509_V_FLAG_X509_STRICT verification flag and either not set a purpose
1724 for the certificate verification or, in the case of TLS client or server
1725 applications, override the default purpose.
1730 * Fixed an issue where an OpenSSL TLS server may crash if sent a maliciously
1731 crafted renegotiation ClientHello message from a client. If a TLSv1.2
1732 renegotiation ClientHello omits the signature_algorithms extension (where it
1733 was present in the initial ClientHello), but includes a
1734 signature_algorithms_cert extension then a NULL pointer dereference will
1735 result, leading to a crash and a denial of service attack.
1737 A server is only vulnerable if it has TLSv1.2 and renegotiation enabled
1738 (which is the default configuration). OpenSSL TLS clients are not impacted by
1742 *Peter Kästle and Samuel Sapalski*
1744 ### Changes between 1.1.1i and 1.1.1j [16 Feb 2021]
1746 * Fixed the X509_issuer_and_serial_hash() function. It attempts to
1747 create a unique hash value based on the issuer and serial number data
1748 contained within an X509 certificate. However it was failing to correctly
1749 handle any errors that may occur while parsing the issuer field (which might
1750 occur if the issuer field is maliciously constructed). This may subsequently
1751 result in a NULL pointer deref and a crash leading to a potential denial of
1757 * Fixed the RSA_padding_check_SSLv23() function and the RSA_SSLV23_PADDING
1758 padding mode to correctly check for rollback attacks. This is considered a
1759 bug in OpenSSL 1.1.1 because it does not support SSLv2. In 1.0.2 this is
1764 Fixed the EVP_CipherUpdate, EVP_EncryptUpdate and EVP_DecryptUpdate
1765 functions. Previously they could overflow the output length argument in some
1766 cases where the input length is close to the maximum permissable length for
1767 an integer on the platform. In such cases the return value from the function
1768 call would be 1 (indicating success), but the output length value would be
1769 negative. This could cause applications to behave incorrectly or crash.
1774 * Fixed SRP_Calc_client_key so that it runs in constant time. The previous
1775 implementation called BN_mod_exp without setting BN_FLG_CONSTTIME. This
1776 could be exploited in a side channel attack to recover the password. Since
1777 the attack is local host only this is outside of the current OpenSSL
1778 threat model and therefore no CVE is assigned.
1780 Thanks to Mohammed Sabt and Daniel De Almeida Braga for reporting this
1785 ### Changes between 1.1.1h and 1.1.1i [8 Dec 2020]
1787 * Fixed NULL pointer deref in the GENERAL_NAME_cmp function
1788 This function could crash if both GENERAL_NAMEs contain an EDIPARTYNAME.
1789 If an attacker can control both items being compared then this could lead
1790 to a possible denial of service attack. OpenSSL itself uses the
1791 GENERAL_NAME_cmp function for two purposes:
1792 1) Comparing CRL distribution point names between an available CRL and a
1793 CRL distribution point embedded in an X509 certificate
1794 2) When verifying that a timestamp response token signer matches the
1795 timestamp authority name (exposed via the API functions
1796 TS_RESP_verify_response and TS_RESP_verify_token)
1801 ### Changes between 1.1.1g and 1.1.1h [22 Sep 2020]
1803 * Certificates with explicit curve parameters are now disallowed in
1804 verification chains if the X509_V_FLAG_X509_STRICT flag is used.
1808 * The 'MinProtocol' and 'MaxProtocol' configuration commands now silently
1809 ignore TLS protocol version bounds when configuring DTLS-based contexts, and
1810 conversely, silently ignore DTLS protocol version bounds when configuring
1811 TLS-based contexts. The commands can be repeated to set bounds of both
1812 types. The same applies with the corresponding "min_protocol" and
1813 "max_protocol" command-line switches, in case some application uses both TLS
1816 SSL_CTX instances that are created for a fixed protocol version (e.g.
1817 TLSv1_server_method()) also silently ignore version bounds. Previously
1818 attempts to apply bounds to these protocol versions would result in an
1819 error. Now only the "version-flexible" SSL_CTX instances are subject to
1820 limits in configuration files in command-line options.
1824 * Handshake now fails if Extended Master Secret extension is dropped
1829 * The Oracle Developer Studio compiler will start reporting deprecated APIs
1831 ### Changes between 1.1.1f and 1.1.1g [21 Apr 2020]
1833 * Fixed segmentation fault in SSL_check_chain()
1834 Server or client applications that call the SSL_check_chain() function
1835 during or after a TLS 1.3 handshake may crash due to a NULL pointer
1836 dereference as a result of incorrect handling of the
1837 "signature_algorithms_cert" TLS extension. The crash occurs if an invalid
1838 or unrecognised signature algorithm is received from the peer. This could
1839 be exploited by a malicious peer in a Denial of Service attack.
1844 * Added AES consttime code for no-asm configurations
1845 an optional constant time support for AES was added
1846 when building openssl for no-asm.
1847 Enable with: ./config no-asm -DOPENSSL_AES_CONST_TIME
1848 Disable with: ./config no-asm -DOPENSSL_NO_AES_CONST_TIME
1849 At this time this feature is by default disabled.
1850 It will be enabled by default in 3.0.
1854 ### Changes between 1.1.1e and 1.1.1f [31 Mar 2020]
1856 * Revert the change of EOF detection while reading in libssl to avoid
1857 regressions in applications depending on the current way of reporting
1858 the EOF. As the existing method is not fully accurate the change to
1859 reporting the EOF via SSL_ERROR_SSL is kept on the current development
1860 branch and will be present in the 3.0 release.
1864 * Revised BN_generate_prime_ex to not avoid factors 3..17863 in p-1
1865 when primes for RSA keys are computed.
1866 Since we previously always generated primes == 2 (mod 3) for RSA keys,
1867 the 2-prime and 3-prime RSA modules were easy to distinguish, since
1868 N = p*q = 1 (mod 3), but N = p*q*r = 2 (mod 3). Therefore fingerprinting
1869 2-prime vs. 3-prime RSA keys was possible by computing N mod 3.
1870 This avoids possible fingerprinting of newly generated RSA modules.
1874 ### Changes between 1.1.1d and 1.1.1e [17 Mar 2020]
1876 * Properly detect EOF while reading in libssl. Previously if we hit an EOF
1877 while reading in libssl then we would report an error back to the
1878 application (SSL_ERROR_SYSCALL) but errno would be 0. We now add
1879 an error to the stack (which means we instead return SSL_ERROR_SSL) and
1880 therefore give a hint as to what went wrong.
1884 * Check that ed25519 and ed448 are allowed by the security level. Previously
1885 signature algorithms not using an MD were not being checked that they were
1886 allowed by the security level.
1890 * Fixed SSL_get_servername() behaviour. The behaviour of SSL_get_servername()
1891 was not quite right. The behaviour was not consistent between resumption
1892 and normal handshakes, and also not quite consistent with historical
1893 behaviour. The behaviour in various scenarios has been clarified and
1894 it has been updated to make it match historical behaviour as closely as
1899 * *[VMS only]* The header files that the VMS compilers include automatically,
1900 `__DECC_INCLUDE_PROLOGUE.H` and `__DECC_INCLUDE_EPILOGUE.H`, use pragmas
1901 that the C++ compiler doesn't understand. This is a shortcoming in the
1902 compiler, but can be worked around with `__cplusplus` guards.
1904 C++ applications that use OpenSSL libraries must be compiled using the
1905 qualifier `/NAMES=(AS_IS,SHORTENED)` to be able to use all the OpenSSL
1906 functions. Otherwise, only functions with symbols of less than 31
1907 characters can be used, as the linker will not be able to successfully
1908 resolve symbols with longer names.
1912 * Added a new method to gather entropy on VMS, based on SYS$GET_ENTROPY.
1913 The presence of this system service is determined at run-time.
1917 * Print all values for a PKCS#12 attribute with 'openssl pkcs12', not just
1922 ### Changes between 1.1.1c and 1.1.1d [10 Sep 2019]
1924 * Fixed a fork protection issue. OpenSSL 1.1.1 introduced a rewritten random
1925 number generator (RNG). This was intended to include protection in the
1926 event of a fork() system call in order to ensure that the parent and child
1927 processes did not share the same RNG state. However this protection was not
1928 being used in the default case.
1930 A partial mitigation for this issue is that the output from a high
1931 precision timer is mixed into the RNG state so the likelihood of a parent
1932 and child process sharing state is significantly reduced.
1934 If an application already calls OPENSSL_init_crypto() explicitly using
1935 OPENSSL_INIT_ATFORK then this problem does not occur at all.
1938 *Matthias St. Pierre*
1940 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
1941 used even when parsing explicit parameters, when loading a encoded key
1942 or calling `EC_GROUP_new_from_ecpkparameters()`/
1943 `EC_GROUP_new_from_ecparameters()`.
1944 This prevents bypass of security hardening and performance gains,
1945 especially for curves with specialized EC_METHODs.
1946 By default, if a key encoded with explicit parameters is loaded and later
1947 encoded, the output is still encoded with explicit parameters, even if
1948 internally a "named" EC_GROUP is used for computation.
1952 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
1953 this change, EC_GROUP_set_generator would accept order and/or cofactor as
1954 NULL. After this change, only the cofactor parameter can be NULL. It also
1955 does some minimal sanity checks on the passed order.
1960 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
1961 An attack is simple, if the first CMS_recipientInfo is valid but the
1962 second CMS_recipientInfo is chosen ciphertext. If the second
1963 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
1964 encryption key will be replaced by garbage, and the message cannot be
1965 decoded, but if the RSA decryption fails, the correct encryption key is
1966 used and the recipient will not notice the attack.
1967 As a work around for this potential attack the length of the decrypted
1968 key must be equal to the cipher default key length, in case the
1969 certifiate is not given and all recipientInfo are tried out.
1970 The old behaviour can be re-enabled in the CMS code by setting the
1971 CMS_DEBUG_DECRYPT flag.
1976 * Early start up entropy quality from the DEVRANDOM seed source has been
1977 improved for older Linux systems. The RAND subsystem will wait for
1978 /dev/random to be producing output before seeding from /dev/urandom.
1979 The seeded state is stored for future library initialisations using
1980 a system global shared memory segment. The shared memory identifier
1981 can be configured by defining OPENSSL_RAND_SEED_DEVRANDOM_SHM_ID to
1982 the desired value. The default identifier is 114.
1986 * Correct the extended master secret constant on EBCDIC systems. Without this
1987 fix TLS connections between an EBCDIC system and a non-EBCDIC system that
1988 negotiate EMS will fail. Unfortunately this also means that TLS connections
1989 between EBCDIC systems with this fix, and EBCDIC systems without this
1990 fix will fail if they negotiate EMS.
1994 * Use Windows installation paths in the mingw builds
1996 Mingw isn't a POSIX environment per se, which means that Windows
1997 paths should be used for installation.
2002 * Changed DH_check to accept parameters with order q and 2q subgroups.
2003 With order 2q subgroups the bit 0 of the private key is not secret
2004 but DH_generate_key works around that by clearing bit 0 of the
2005 private key for those. This avoids leaking bit 0 of the private key.
2009 * Significantly reduce secure memory usage by the randomness pools.
2013 * Revert the DEVRANDOM_WAIT feature for Linux systems
2015 The DEVRANDOM_WAIT feature added a select() call to wait for the
2016 /dev/random device to become readable before reading from the
2017 /dev/urandom device.
2019 It turned out that this change had negative side effects on
2020 performance which were not acceptable. After some discussion it
2021 was decided to revert this feature and leave it up to the OS
2022 resp. the platform maintainer to ensure a proper initialization
2023 during early boot time.
2025 *Matthias St. Pierre*
2027 ### Changes between 1.1.1b and 1.1.1c [28 May 2019]
2029 * Add build tests for C++. These are generated files that only do one
2030 thing, to include one public OpenSSL head file each. This tests that
2031 the public header files can be usefully included in a C++ application.
2033 This test isn't enabled by default. It can be enabled with the option
2034 'enable-buildtest-c++'.
2038 * Enable SHA3 pre-hashing for ECDSA and DSA.
2042 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
2043 This changes the size when using the `genpkey` command when no size is given.
2044 It fixes an omission in earlier changes that changed all RSA, DSA and DH
2045 generation commands to use 2048 bits by default.
2049 * Reorganize the manual pages to consistently have RETURN VALUES,
2050 EXAMPLES, SEE ALSO and HISTORY come in that order, and adjust
2051 util/fix-doc-nits accordingly.
2053 *Paul Yang, Joshua Lock*
2055 * Add the missing accessor EVP_PKEY_get0_engine()
2059 * Have commands like `s_client` and `s_server` output the signature scheme
2060 along with other cipher suite parameters when debugging.
2064 * Make OPENSSL_config() error agnostic again.
2068 * Do the error handling in RSA decryption constant time.
2072 * Prevent over long nonces in ChaCha20-Poly1305.
2074 ChaCha20-Poly1305 is an AEAD cipher, and requires a unique nonce input
2075 for every encryption operation. RFC 7539 specifies that the nonce value
2076 (IV) should be 96 bits (12 bytes). OpenSSL allows a variable nonce length
2077 and front pads the nonce with 0 bytes if it is less than 12
2078 bytes. However it also incorrectly allows a nonce to be set of up to 16
2079 bytes. In this case only the last 12 bytes are significant and any
2080 additional leading bytes are ignored.
2082 It is a requirement of using this cipher that nonce values are
2083 unique. Messages encrypted using a reused nonce value are susceptible to
2084 serious confidentiality and integrity attacks. If an application changes
2085 the default nonce length to be longer than 12 bytes and then makes a
2086 change to the leading bytes of the nonce expecting the new value to be a
2087 new unique nonce then such an application could inadvertently encrypt
2088 messages with a reused nonce.
2090 Additionally the ignored bytes in a long nonce are not covered by the
2091 integrity guarantee of this cipher. Any application that relies on the
2092 integrity of these ignored leading bytes of a long nonce may be further
2093 affected. Any OpenSSL internal use of this cipher, including in SSL/TLS,
2094 is safe because no such use sets such a long nonce value. However user
2095 applications that use this cipher directly and set a non-default nonce
2096 length to be longer than 12 bytes may be vulnerable.
2098 This issue was reported to OpenSSL on 16th of March 2019 by Joran Dirk
2104 * Add DEVRANDOM_WAIT feature for Linux systems
2106 On older Linux systems where the getrandom() system call is not available,
2107 OpenSSL normally uses the /dev/urandom device for seeding its CSPRNG.
2108 Contrary to getrandom(), the /dev/urandom device will not block during
2109 early boot when the kernel CSPRNG has not been seeded yet.
2111 To mitigate this known weakness, use select() to wait for /dev/random to
2112 become readable before reading from /dev/urandom.
2114 * Ensure that SM2 only uses SM3 as digest algorithm
2118 ### Changes between 1.1.1a and 1.1.1b [26 Feb 2019]
2120 * Change the info callback signals for the start and end of a post-handshake
2121 message exchange in TLSv1.3. In 1.1.1/1.1.1a we used SSL_CB_HANDSHAKE_START
2122 and SSL_CB_HANDSHAKE_DONE. Experience has shown that many applications get
2123 confused by this and assume that a TLSv1.2 renegotiation has started. This
2124 can break KeyUpdate handling. Instead we no longer signal the start and end
2125 of a post handshake message exchange (although the messages themselves are
2126 still signalled). This could break some applications that were expecting
2127 the old signals. However without this KeyUpdate is not usable for many
2132 ### Changes between 1.1.1 and 1.1.1a [20 Nov 2018]
2134 * Timing vulnerability in DSA signature generation
2136 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
2137 timing side channel attack. An attacker could use variations in the signing
2138 algorithm to recover the private key.
2140 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
2145 * Timing vulnerability in ECDSA signature generation
2147 The OpenSSL ECDSA signature algorithm has been shown to be vulnerable to a
2148 timing side channel attack. An attacker could use variations in the signing
2149 algorithm to recover the private key.
2151 This issue was reported to OpenSSL on 25th October 2018 by Samuel Weiser.
2156 * Fixed the issue that RAND_add()/RAND_seed() silently discards random input
2157 if its length exceeds 4096 bytes. The limit has been raised to a buffer size
2158 of two gigabytes and the error handling improved.
2160 This issue was reported to OpenSSL by Dr. Falko Strenzke. It has been
2161 categorized as a normal bug, not a security issue, because the DRBG reseeds
2162 automatically and is fully functional even without additional randomness
2163 provided by the application.
2165 ### Changes between 1.1.0i and 1.1.1 [11 Sep 2018]
2167 * Add a new ClientHello callback. Provides a callback interface that gives
2168 the application the ability to adjust the nascent SSL object at the
2169 earliest stage of ClientHello processing, immediately after extensions have
2170 been collected but before they have been processed. In particular, this
2171 callback can adjust the supported TLS versions in response to the contents
2176 * Add SM2 base algorithm support.
2180 * s390x assembly pack: add (improved) hardware-support for the following
2181 cryptographic primitives: sha3, shake, aes-gcm, aes-ccm, aes-ctr, aes-ofb,
2182 aes-cfb/cfb8, aes-ecb.
2186 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
2187 parameter is no longer accepted, as it leads to a corrupt table. NULL
2188 pem_str is reserved for alias entries only.
2192 * Use the new ec_scalar_mul_ladder scaffold to implement a specialized ladder
2193 step for prime curves. The new implementation is based on formulae from
2194 differential addition-and-doubling in homogeneous projective coordinates
2195 from Izu-Takagi "A fast parallel elliptic curve multiplication resistant
2196 against side channel attacks" and Brier-Joye "Weierstrass Elliptic Curves
2197 and Side-Channel Attacks" Eq. (8) for y-coordinate recovery, modified
2198 to work in projective coordinates.
2200 *Billy Bob Brumley, Nicola Tuveri*
2202 * Change generating and checking of primes so that the error rate of not
2203 being prime depends on the intended use based on the size of the input.
2204 For larger primes this will result in more rounds of Miller-Rabin.
2205 The maximal error rate for primes with more than 1080 bits is lowered
2208 *Kurt Roeckx, Annie Yousar*
2210 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
2214 * The 'tsget' script is renamed to 'tsget.pl', to avoid confusion when
2215 moving between systems, and to avoid confusion when a Windows build is
2216 done with mingw vs with MSVC. For POSIX installs, there's still a
2217 symlink or copy named 'tsget' to avoid that confusion as well.
2221 * Revert blinding in ECDSA sign and instead make problematic addition
2222 length-invariant. Switch even to fixed-length Montgomery multiplication.
2226 * Use the new ec_scalar_mul_ladder scaffold to implement a specialized ladder
2227 step for binary curves. The new implementation is based on formulae from
2228 differential addition-and-doubling in mixed Lopez-Dahab projective
2229 coordinates, modified to independently blind the operands.
2231 *Billy Bob Brumley, Sohaib ul Hassan, Nicola Tuveri*
2233 * Add a scaffold to optionally enhance the Montgomery ladder implementation
2234 for `ec_scalar_mul_ladder` (formerly `ec_mul_consttime`) allowing
2235 EC_METHODs to implement their own specialized "ladder step", to take
2236 advantage of more favorable coordinate systems or more efficient
2237 differential addition-and-doubling algorithms.
2239 *Billy Bob Brumley, Sohaib ul Hassan, Nicola Tuveri*
2241 * Modified the random device based seed sources to keep the relevant
2242 file descriptors open rather than reopening them on each access.
2243 This allows such sources to operate in a chroot() jail without
2244 the associated device nodes being available. This behaviour can be
2245 controlled using RAND_keep_random_devices_open().
2249 * Numerous side-channel attack mitigations have been applied. This may have
2250 performance impacts for some algorithms for the benefit of improved
2251 security. Specific changes are noted in this change log by their respective
2256 * AIX shared library support overhaul. Switch to AIX "natural" way of
2257 handling shared libraries, which means collecting shared objects of
2258 different versions and bitnesses in one common archive. This allows to
2259 mitigate conflict between 1.0 and 1.1 side-by-side installations. It
2260 doesn't affect the way 3rd party applications are linked, only how
2261 multi-version installation is managed.
2265 * Make ec_group_do_inverse_ord() more robust and available to other
2266 EC cryptosystems, so that irrespective of BN_FLG_CONSTTIME, SCA
2267 mitigations are applied to the fallback BN_mod_inverse().
2268 When using this function rather than BN_mod_inverse() directly, new
2269 EC cryptosystem implementations are then safer-by-default.
2273 * Add coordinate blinding for EC_POINT and implement projective
2274 coordinate blinding for generic prime curves as a countermeasure to
2275 chosen point SCA attacks.
2277 *Sohaib ul Hassan, Nicola Tuveri, Billy Bob Brumley*
2279 * Add blinding to ECDSA and DSA signatures to protect against side channel
2280 attacks discovered by Keegan Ryan (NCC Group).
2284 * Enforce checking in the `pkeyutl` command to ensure that the input
2285 length does not exceed the maximum supported digest length when performing
2286 a sign, verify or verifyrecover operation.
2290 * SSL_MODE_AUTO_RETRY is enabled by default. Applications that use blocking
2291 I/O in combination with something like select() or poll() will hang. This
2292 can be turned off again using SSL_CTX_clear_mode().
2293 Many applications do not properly handle non-application data records, and
2294 TLS 1.3 sends more of such records. Setting SSL_MODE_AUTO_RETRY works
2295 around the problems in those applications, but can also break some.
2296 It's recommended to read the manpages about SSL_read(), SSL_write(),
2297 SSL_get_error(), SSL_shutdown(), SSL_CTX_set_mode() and
2298 SSL_CTX_set_read_ahead() again.
2302 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
2303 now allow empty (zero character) pass phrases.
2307 * Apply blinding to binary field modular inversion and remove patent
2308 pending (OPENSSL_SUN_GF2M_DIV) BN_GF2m_mod_div implementation.
2312 * Deprecate ec2_mult.c and unify scalar multiplication code paths for
2313 binary and prime elliptic curves.
2317 * Remove ECDSA nonce padding: EC_POINT_mul is now responsible for
2318 constant time fixed point multiplication.
2322 * Revise elliptic curve scalar multiplication with timing attack
2323 defenses: ec_wNAF_mul redirects to a constant time implementation
2324 when computing fixed point and variable point multiplication (which
2325 in OpenSSL are mostly used with secret scalars in keygen, sign,
2326 ECDH derive operations).
2327 *Billy Bob Brumley, Nicola Tuveri, Cesar Pereida García,
2330 * Updated CONTRIBUTING
2334 * Updated DRBG / RAND to request nonce and additional low entropy
2335 randomness from the system.
2337 *Matthias St. Pierre*
2339 * Updated 'openssl rehash' to use OpenSSL consistent default.
2343 * Moved the load of the ssl_conf module to libcrypto, which helps
2344 loading engines that libssl uses before libssl is initialised.
2348 * Added EVP_PKEY_sign() and EVP_PKEY_verify() for EdDSA
2352 * Fixed X509_NAME_ENTRY_set to get multi-valued RDNs right in all cases.
2354 *Ingo Schwarze, Rich Salz*
2356 * Added output of accepting IP address and port for 'openssl s_server'
2360 * Added a new API for TLSv1.3 ciphersuites:
2361 SSL_CTX_set_ciphersuites()
2362 SSL_set_ciphersuites()
2366 * Memory allocation failures consistently add an error to the error
2371 * Don't use OPENSSL_ENGINES and OPENSSL_CONF environment values
2372 in libcrypto when run as setuid/setgid.
2376 * Load any config file by default when libssl is used.
2380 * Added new public header file <openssl/rand_drbg.h> and documentation
2381 for the RAND_DRBG API. See manual page RAND_DRBG(7) for an overview.
2383 *Matthias St. Pierre*
2385 * QNX support removed (cannot find contributors to get their approval
2386 for the license change).
2390 * TLSv1.3 replay protection for early data has been implemented. See the
2391 SSL_read_early_data() man page for further details.
2395 * Separated TLSv1.3 ciphersuite configuration out from TLSv1.2 ciphersuite
2396 configuration. TLSv1.3 ciphersuites are not compatible with TLSv1.2 and
2397 below. Similarly TLSv1.2 ciphersuites are not compatible with TLSv1.3.
2398 In order to avoid issues where legacy TLSv1.2 ciphersuite configuration
2399 would otherwise inadvertently disable all TLSv1.3 ciphersuites the
2400 configuration has been separated out. See the ciphers man page or the
2401 SSL_CTX_set_ciphersuites() man page for more information.
2405 * On POSIX (BSD, Linux, ...) systems the ocsp(1) command running
2406 in responder mode now supports the new "-multi" option, which
2407 spawns the specified number of child processes to handle OCSP
2408 requests. The "-timeout" option now also limits the OCSP
2409 responder's patience to wait to receive the full client request
2410 on a newly accepted connection. Child processes are respawned
2411 as needed, and the CA index file is automatically reloaded
2412 when changed. This makes it possible to run the "ocsp" responder
2413 as a long-running service, making the OpenSSL CA somewhat more
2414 feature-complete. In this mode, most diagnostic messages logged
2415 after entering the event loop are logged via syslog(3) rather than
2420 * Added support for X448 and Ed448. Heavily based on original work by
2425 * Extend OSSL_STORE with capabilities to search and to narrow the set of
2426 objects loaded. This adds the functions OSSL_STORE_expect() and
2427 OSSL_STORE_find() as well as needed tools to construct searches and
2428 get the search data out of them.
2432 * Support for TLSv1.3 added. Note that users upgrading from an earlier
2433 version of OpenSSL should review their configuration settings to ensure
2434 that they are still appropriate for TLSv1.3. For further information see:
2435 <https://wiki.openssl.org/index.php/TLS1.3>
2439 * Grand redesign of the OpenSSL random generator
2441 The default RAND method now utilizes an AES-CTR DRBG according to
2442 NIST standard SP 800-90Ar1. The new random generator is essentially
2443 a port of the default random generator from the OpenSSL FIPS 2.0
2444 object module. It is a hybrid deterministic random bit generator
2445 using an AES-CTR bit stream and which seeds and reseeds itself
2446 automatically using trusted system entropy sources.
2448 Some of its new features are:
2449 - Support for multiple DRBG instances with seed chaining.
2450 - The default RAND method makes use of a DRBG.
2451 - There is a public and private DRBG instance.
2452 - The DRBG instances are fork-safe.
2453 - Keep all global DRBG instances on the secure heap if it is enabled.
2454 - The public and private DRBG instance are per thread for lock free
2457 *Paul Dale, Benjamin Kaduk, Kurt Roeckx, Rich Salz, Matthias St. Pierre*
2459 * Changed Configure so it only says what it does and doesn't dump
2460 so much data. Instead, ./configdata.pm should be used as a script
2461 to display all sorts of configuration data.
2465 * Added processing of "make variables" to Configure.
2469 * Added SHA512/224 and SHA512/256 algorithm support.
2473 * The last traces of Netware support, first removed in 1.1.0, have
2478 * Get rid of Makefile.shared, and in the process, make the processing
2479 of certain files (rc.obj, or the .def/.map/.opt files produced from
2480 the ordinal files) more visible and hopefully easier to trace and
2481 debug (or make silent).
2485 * Make it possible to have environment variable assignments as
2486 arguments to config / Configure.
2490 * Add multi-prime RSA (RFC 8017) support.
2494 * Add SM3 implemented according to GB/T 32905-2016
2495 *Jack Lloyd <jack.lloyd@ribose.com>,*
2496 *Ronald Tse <ronald.tse@ribose.com>,*
2497 *Erick Borsboom <erick.borsboom@ribose.com>*
2499 * Add 'Maximum Fragment Length' TLS extension negotiation and support
2500 as documented in RFC6066.
2501 Based on a patch from Tomasz Moń
2503 *Filipe Raimundo da Silva*
2505 * Add SM4 implemented according to GB/T 32907-2016.
2506 *Jack Lloyd <jack.lloyd@ribose.com>,*
2507 *Ronald Tse <ronald.tse@ribose.com>,*
2508 *Erick Borsboom <erick.borsboom@ribose.com>*
2510 * Reimplement -newreq-nodes and ERR_error_string_n; the
2511 original author does not agree with the license change.
2515 * Add ARIA AEAD TLS support.
2519 * Some macro definitions to support VS6 have been removed. Visual
2520 Studio 6 has not worked since 1.1.0
2524 * Add ERR_clear_last_mark(), to allow callers to clear the last mark
2525 without clearing the errors.
2529 * Add "atfork" functions. If building on a system that without
2530 pthreads, see doc/man3/OPENSSL_fork_prepare.pod for application
2531 requirements. The RAND facility now uses/requires this.
2539 * The UI API becomes a permanent and integral part of libcrypto, i.e.
2540 not possible to disable entirely. However, it's still possible to
2541 disable the console reading UI method, UI_OpenSSL() (use UI_null()
2544 To disable, configure with 'no-ui-console'. 'no-ui' is still
2545 possible to use as an alias. Check at compile time with the
2546 macro OPENSSL_NO_UI_CONSOLE. The macro OPENSSL_NO_UI is still
2547 possible to check and is an alias for OPENSSL_NO_UI_CONSOLE.
2551 * Add a STORE module, which implements a uniform and URI based reader of
2552 stores that can contain keys, certificates, CRLs and numerous other
2553 objects. The main API is loosely based on a few stdio functions,
2554 and includes OSSL_STORE_open, OSSL_STORE_load, OSSL_STORE_eof,
2555 OSSL_STORE_error and OSSL_STORE_close.
2556 The implementation uses backends called "loaders" to implement arbitrary
2557 URI schemes. There is one built in "loader" for the 'file' scheme.
2561 * Add devcrypto engine. This has been implemented against cryptodev-linux,
2562 then adjusted to work on FreeBSD 8.4 as well.
2563 Enable by configuring with 'enable-devcryptoeng'. This is done by default
2564 on BSD implementations, as cryptodev.h is assumed to exist on all of them.
2568 * Module names can prefixed with OSSL_ or OPENSSL_. This affects
2569 util/mkerr.pl, which is adapted to allow those prefixes, leading to
2570 error code calls like this:
2572 OSSL_FOOerr(OSSL_FOO_F_SOMETHING, OSSL_FOO_R_WHATEVER);
2574 With this change, we claim the namespaces OSSL and OPENSSL in a manner
2575 that can be encoded in C. For the foreseeable future, this will only
2578 *Richard Levitte and Tim Hudson*
2580 * Removed BSD cryptodev engine.
2584 * Add a build target 'build_all_generated', to build all generated files
2585 and only that. This can be used to prepare everything that requires
2586 things like perl for a system that lacks perl and then move everything
2587 to that system and do the rest of the build there.
2591 * In the UI interface, make it possible to duplicate the user data. This
2592 can be used by engines that need to retain the data for a longer time
2593 than just the call where this user data is passed.
2597 * Ignore the '-named_curve auto' value for compatibility of applications
2600 *Tomáš Mráz <tmraz@fedoraproject.org>*
2602 * Fragmented SSL/TLS alerts are no longer accepted. An alert message is 2
2603 bytes long. In theory it is permissible in SSLv3 - TLSv1.2 to fragment such
2604 alerts across multiple records (some of which could be empty). In practice
2605 it make no sense to send an empty alert record, or to fragment one. TLSv1.3
2606 prohibits this altogether and other libraries (BoringSSL, NSS) do not
2607 support this at all. Supporting it adds significant complexity to the
2608 record layer, and its removal is unlikely to cause interoperability
2613 * Add the ASN.1 types INT32, UINT32, INT64, UINT64 and variants prefixed
2614 with Z. These are meant to replace LONG and ZLONG and to be size safe.
2615 The use of LONG and ZLONG is discouraged and scheduled for deprecation
2620 * Add the 'z' and 'j' modifiers to BIO_printf() et al formatting string,
2621 'z' is to be used for [s]size_t, and 'j' - with [u]int64_t.
2623 *Richard Levitte, Andy Polyakov*
2625 * Add EC_KEY_get0_engine(), which does for EC_KEY what RSA_get0_engine()
2630 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
2631 platform rather than 'mingw'.
2635 * The functions X509_STORE_add_cert and X509_STORE_add_crl return
2636 success if they are asked to add an object which already exists
2637 in the store. This change cascades to other functions which load
2638 certificates and CRLs.
2642 * x86_64 assembly pack: annotate code with DWARF CFI directives to
2643 facilitate stack unwinding even from assembly subroutines.
2647 * Remove VAX C specific definitions of OPENSSL_EXPORT, OPENSSL_EXTERN.
2648 Also remove OPENSSL_GLOBAL entirely, as it became a no-op.
2652 * Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
2653 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
2654 which is the minimum version we support.
2658 * Certificate time validation (X509_cmp_time) enforces stricter
2659 compliance with RFC 5280. Fractional seconds and timezone offsets
2660 are no longer allowed.
2664 * Add support for ARIA
2668 * s_client will now send the Server Name Indication (SNI) extension by
2669 default unless the new "-noservername" option is used. The server name is
2670 based on the host provided to the "-connect" option unless overridden by
2671 using "-servername".
2675 * Add support for SipHash
2679 * OpenSSL now fails if it receives an unrecognised record type in TLS1.0
2680 or TLS1.1. Previously this only happened in SSLv3 and TLS1.2. This is to
2681 prevent issues where no progress is being made and the peer continually
2682 sends unrecognised record types, using up resources processing them.
2686 * 'openssl passwd' can now produce SHA256 and SHA512 based output,
2687 using the algorithm defined in
2688 <https://www.akkadia.org/drepper/SHA-crypt.txt>
2692 * Heartbeat support has been removed; the ABI is changed for now.
2694 *Richard Levitte, Rich Salz*
2696 * Support for SSL_OP_NO_ENCRYPT_THEN_MAC in SSL_CONF_cmd.
2700 * The RSA "null" method, which was partially supported to avoid patent
2701 issues, has been replaced to always returns NULL.
2708 ### Changes between 1.1.0k and 1.1.0l [10 Sep 2019]
2710 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
2711 used even when parsing explicit parameters, when loading a encoded key
2712 or calling `EC_GROUP_new_from_ecpkparameters()`/
2713 `EC_GROUP_new_from_ecparameters()`.
2714 This prevents bypass of security hardening and performance gains,
2715 especially for curves with specialized EC_METHODs.
2716 By default, if a key encoded with explicit parameters is loaded and later
2717 encoded, the output is still encoded with explicit parameters, even if
2718 internally a "named" EC_GROUP is used for computation.
2722 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
2723 this change, EC_GROUP_set_generator would accept order and/or cofactor as
2724 NULL. After this change, only the cofactor parameter can be NULL. It also
2725 does some minimal sanity checks on the passed order.
2730 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
2731 An attack is simple, if the first CMS_recipientInfo is valid but the
2732 second CMS_recipientInfo is chosen ciphertext. If the second
2733 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
2734 encryption key will be replaced by garbage, and the message cannot be
2735 decoded, but if the RSA decryption fails, the correct encryption key is
2736 used and the recipient will not notice the attack.
2737 As a work around for this potential attack the length of the decrypted
2738 key must be equal to the cipher default key length, in case the
2739 certifiate is not given and all recipientInfo are tried out.
2740 The old behaviour can be re-enabled in the CMS code by setting the
2741 CMS_DEBUG_DECRYPT flag.
2746 * Use Windows installation paths in the mingw builds
2748 Mingw isn't a POSIX environment per se, which means that Windows
2749 paths should be used for installation.
2754 ### Changes between 1.1.0j and 1.1.0k [28 May 2019]
2756 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
2757 This changes the size when using the `genpkey` command when no size is given.
2758 It fixes an omission in earlier changes that changed all RSA, DSA and DH
2759 generation commands to use 2048 bits by default.
2763 * Prevent over long nonces in ChaCha20-Poly1305.
2765 ChaCha20-Poly1305 is an AEAD cipher, and requires a unique nonce input
2766 for every encryption operation. RFC 7539 specifies that the nonce value
2767 (IV) should be 96 bits (12 bytes). OpenSSL allows a variable nonce length
2768 and front pads the nonce with 0 bytes if it is less than 12
2769 bytes. However it also incorrectly allows a nonce to be set of up to 16
2770 bytes. In this case only the last 12 bytes are significant and any
2771 additional leading bytes are ignored.
2773 It is a requirement of using this cipher that nonce values are
2774 unique. Messages encrypted using a reused nonce value are susceptible to
2775 serious confidentiality and integrity attacks. If an application changes
2776 the default nonce length to be longer than 12 bytes and then makes a
2777 change to the leading bytes of the nonce expecting the new value to be a
2778 new unique nonce then such an application could inadvertently encrypt
2779 messages with a reused nonce.
2781 Additionally the ignored bytes in a long nonce are not covered by the
2782 integrity guarantee of this cipher. Any application that relies on the
2783 integrity of these ignored leading bytes of a long nonce may be further
2784 affected. Any OpenSSL internal use of this cipher, including in SSL/TLS,
2785 is safe because no such use sets such a long nonce value. However user
2786 applications that use this cipher directly and set a non-default nonce
2787 length to be longer than 12 bytes may be vulnerable.
2789 This issue was reported to OpenSSL on 16th of March 2019 by Joran Dirk
2795 * Added SCA hardening for modular field inversion in EC_GROUP through
2796 a new dedicated field_inv() pointer in EC_METHOD.
2797 This also addresses a leakage affecting conversions from projective
2798 to affine coordinates.
2800 *Billy Bob Brumley, Nicola Tuveri*
2802 * Fix a use after free bug in d2i_X509_PUBKEY when overwriting a
2803 re-used X509_PUBKEY object if the second PUBKEY is malformed.
2807 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
2811 * Remove the 'dist' target and add a tarball building script. The
2812 'dist' target has fallen out of use, and it shouldn't be
2813 necessary to configure just to create a source distribution.
2817 ### Changes between 1.1.0i and 1.1.0j [20 Nov 2018]
2819 * Timing vulnerability in DSA signature generation
2821 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
2822 timing side channel attack. An attacker could use variations in the signing
2823 algorithm to recover the private key.
2825 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
2830 * Timing vulnerability in ECDSA signature generation
2832 The OpenSSL ECDSA signature algorithm has been shown to be vulnerable to a
2833 timing side channel attack. An attacker could use variations in the signing
2834 algorithm to recover the private key.
2836 This issue was reported to OpenSSL on 25th October 2018 by Samuel Weiser.
2841 * Add coordinate blinding for EC_POINT and implement projective
2842 coordinate blinding for generic prime curves as a countermeasure to
2843 chosen point SCA attacks.
2845 *Sohaib ul Hassan, Nicola Tuveri, Billy Bob Brumley*
2847 ### Changes between 1.1.0h and 1.1.0i [14 Aug 2018]
2849 * Client DoS due to large DH parameter
2851 During key agreement in a TLS handshake using a DH(E) based ciphersuite a
2852 malicious server can send a very large prime value to the client. This will
2853 cause the client to spend an unreasonably long period of time generating a
2854 key for this prime resulting in a hang until the client has finished. This
2855 could be exploited in a Denial Of Service attack.
2857 This issue was reported to OpenSSL on 5th June 2018 by Guido Vranken
2862 * Cache timing vulnerability in RSA Key Generation
2864 The OpenSSL RSA Key generation algorithm has been shown to be vulnerable to
2865 a cache timing side channel attack. An attacker with sufficient access to
2866 mount cache timing attacks during the RSA key generation process could
2867 recover the private key.
2869 This issue was reported to OpenSSL on 4th April 2018 by Alejandro Cabrera
2870 Aldaya, Billy Brumley, Cesar Pereida Garcia and Luis Manuel Alvarez Tapia.
2875 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
2876 parameter is no longer accepted, as it leads to a corrupt table. NULL
2877 pem_str is reserved for alias entries only.
2881 * Revert blinding in ECDSA sign and instead make problematic addition
2882 length-invariant. Switch even to fixed-length Montgomery multiplication.
2886 * Change generating and checking of primes so that the error rate of not
2887 being prime depends on the intended use based on the size of the input.
2888 For larger primes this will result in more rounds of Miller-Rabin.
2889 The maximal error rate for primes with more than 1080 bits is lowered
2892 *Kurt Roeckx, Annie Yousar*
2894 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
2898 * Add blinding to ECDSA and DSA signatures to protect against side channel
2899 attacks discovered by Keegan Ryan (NCC Group).
2903 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
2904 now allow empty (zero character) pass phrases.
2908 * Certificate time validation (X509_cmp_time) enforces stricter
2909 compliance with RFC 5280. Fractional seconds and timezone offsets
2910 are no longer allowed.
2914 * Fixed a text canonicalisation bug in CMS
2916 Where a CMS detached signature is used with text content the text goes
2917 through a canonicalisation process first prior to signing or verifying a
2918 signature. This process strips trailing space at the end of lines, converts
2919 line terminators to CRLF and removes additional trailing line terminators
2920 at the end of a file. A bug in the canonicalisation process meant that
2921 some characters, such as form-feed, were incorrectly treated as whitespace
2922 and removed. This is contrary to the specification (RFC5485). This fix
2923 could mean that detached text data signed with an earlier version of
2924 OpenSSL 1.1.0 may fail to verify using the fixed version, or text data
2925 signed with a fixed OpenSSL may fail to verify with an earlier version of
2926 OpenSSL 1.1.0. A workaround is to only verify the canonicalised text data
2927 and use the "-binary" flag (for the "cms" command line application) or set
2928 the SMIME_BINARY/PKCS7_BINARY/CMS_BINARY flags (if using CMS_verify()).
2932 ### Changes between 1.1.0g and 1.1.0h [27 Mar 2018]
2934 * Constructed ASN.1 types with a recursive definition could exceed the stack
2936 Constructed ASN.1 types with a recursive definition (such as can be found
2937 in PKCS7) could eventually exceed the stack given malicious input with
2938 excessive recursion. This could result in a Denial Of Service attack. There
2939 are no such structures used within SSL/TLS that come from untrusted sources
2940 so this is considered safe.
2942 This issue was reported to OpenSSL on 4th January 2018 by the OSS-fuzz
2948 * Incorrect CRYPTO_memcmp on HP-UX PA-RISC
2950 Because of an implementation bug the PA-RISC CRYPTO_memcmp function is
2951 effectively reduced to only comparing the least significant bit of each
2952 byte. This allows an attacker to forge messages that would be considered as
2953 authenticated in an amount of tries lower than that guaranteed by the
2954 security claims of the scheme. The module can only be compiled by the
2955 HP-UX assembler, so that only HP-UX PA-RISC targets are affected.
2957 This issue was reported to OpenSSL on 2nd March 2018 by Peter Waltenberg
2963 * Add a build target 'build_all_generated', to build all generated files
2964 and only that. This can be used to prepare everything that requires
2965 things like perl for a system that lacks perl and then move everything
2966 to that system and do the rest of the build there.
2970 * Backport SSL_OP_NO_RENGOTIATION
2972 OpenSSL 1.0.2 and below had the ability to disable renegotiation using the
2973 (undocumented) SSL3_FLAGS_NO_RENEGOTIATE_CIPHERS flag. Due to the opacity
2974 changes this is no longer possible in 1.1.0. Therefore the new
2975 SSL_OP_NO_RENEGOTIATION option from 1.1.1-dev has been backported to
2976 1.1.0 to provide equivalent functionality.
2978 Note that if an application built against 1.1.0h headers (or above) is run
2979 using an older version of 1.1.0 (prior to 1.1.0h) then the option will be
2980 accepted but nothing will happen, i.e. renegotiation will not be prevented.
2984 * Removed the OS390-Unix config target. It relied on a script that doesn't
2989 * rsaz_1024_mul_avx2 overflow bug on x86_64
2991 There is an overflow bug in the AVX2 Montgomery multiplication procedure
2992 used in exponentiation with 1024-bit moduli. No EC algorithms are affected.
2993 Analysis suggests that attacks against RSA and DSA as a result of this
2994 defect would be very difficult to perform and are not believed likely.
2995 Attacks against DH1024 are considered just feasible, because most of the
2996 work necessary to deduce information about a private key may be performed
2997 offline. The amount of resources required for such an attack would be
2998 significant. However, for an attack on TLS to be meaningful, the server
2999 would have to share the DH1024 private key among multiple clients, which is
3000 no longer an option since CVE-2016-0701.
3002 This only affects processors that support the AVX2 but not ADX extensions
3003 like Intel Haswell (4th generation).
3005 This issue was reported to OpenSSL by David Benjamin (Google). The issue
3006 was originally found via the OSS-Fuzz project.
3011 ### Changes between 1.1.0f and 1.1.0g [2 Nov 2017]
3013 * bn_sqrx8x_internal carry bug on x86_64
3015 There is a carry propagating bug in the x86_64 Montgomery squaring
3016 procedure. No EC algorithms are affected. Analysis suggests that attacks
3017 against RSA and DSA as a result of this defect would be very difficult to
3018 perform and are not believed likely. Attacks against DH are considered just
3019 feasible (although very difficult) because most of the work necessary to
3020 deduce information about a private key may be performed offline. The amount
3021 of resources required for such an attack would be very significant and
3022 likely only accessible to a limited number of attackers. An attacker would
3023 additionally need online access to an unpatched system using the target
3024 private key in a scenario with persistent DH parameters and a private
3025 key that is shared between multiple clients.
3027 This only affects processors that support the BMI1, BMI2 and ADX extensions
3028 like Intel Broadwell (5th generation) and later or AMD Ryzen.
3030 This issue was reported to OpenSSL by the OSS-Fuzz project.
3035 * Malformed X.509 IPAddressFamily could cause OOB read
3037 If an X.509 certificate has a malformed IPAddressFamily extension,
3038 OpenSSL could do a one-byte buffer overread. The most likely result
3039 would be an erroneous display of the certificate in text format.
3041 This issue was reported to OpenSSL by the OSS-Fuzz project.
3046 ### Changes between 1.1.0e and 1.1.0f [25 May 2017]
3048 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
3049 platform rather than 'mingw'.
3053 * Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
3054 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
3055 which is the minimum version we support.
3059 ### Changes between 1.1.0d and 1.1.0e [16 Feb 2017]
3061 * Encrypt-Then-Mac renegotiation crash
3063 During a renegotiation handshake if the Encrypt-Then-Mac extension is
3064 negotiated where it was not in the original handshake (or vice-versa) then
3065 this can cause OpenSSL to crash (dependant on ciphersuite). Both clients
3066 and servers are affected.
3068 This issue was reported to OpenSSL by Joe Orton (Red Hat).
3073 ### Changes between 1.1.0c and 1.1.0d [26 Jan 2017]
3075 * Truncated packet could crash via OOB read
3077 If one side of an SSL/TLS path is running on a 32-bit host and a specific
3078 cipher is being used, then a truncated packet can cause that host to
3079 perform an out-of-bounds read, usually resulting in a crash.
3081 This issue was reported to OpenSSL by Robert Święcki of Google.
3086 * Bad (EC)DHE parameters cause a client crash
3088 If a malicious server supplies bad parameters for a DHE or ECDHE key
3089 exchange then this can result in the client attempting to dereference a
3090 NULL pointer leading to a client crash. This could be exploited in a Denial
3093 This issue was reported to OpenSSL by Guido Vranken.
3098 * BN_mod_exp may produce incorrect results on x86_64
3100 There is a carry propagating bug in the x86_64 Montgomery squaring
3101 procedure. No EC algorithms are affected. Analysis suggests that attacks
3102 against RSA and DSA as a result of this defect would be very difficult to
3103 perform and are not believed likely. Attacks against DH are considered just
3104 feasible (although very difficult) because most of the work necessary to
3105 deduce information about a private key may be performed offline. The amount
3106 of resources required for such an attack would be very significant and
3107 likely only accessible to a limited number of attackers. An attacker would
3108 additionally need online access to an unpatched system using the target
3109 private key in a scenario with persistent DH parameters and a private
3110 key that is shared between multiple clients. For example this can occur by
3111 default in OpenSSL DHE based SSL/TLS ciphersuites. Note: This issue is very
3112 similar to CVE-2015-3193 but must be treated as a separate problem.
3114 This issue was reported to OpenSSL by the OSS-Fuzz project.
3119 ### Changes between 1.1.0b and 1.1.0c [10 Nov 2016]
3121 * ChaCha20/Poly1305 heap-buffer-overflow
3123 TLS connections using `*-CHACHA20-POLY1305` ciphersuites are susceptible to
3124 a DoS attack by corrupting larger payloads. This can result in an OpenSSL
3125 crash. This issue is not considered to be exploitable beyond a DoS.
3127 This issue was reported to OpenSSL by Robert Święcki (Google Security Team)
3132 * CMS Null dereference
3134 Applications parsing invalid CMS structures can crash with a NULL pointer
3135 dereference. This is caused by a bug in the handling of the ASN.1 CHOICE
3136 type in OpenSSL 1.1.0 which can result in a NULL value being passed to the
3137 structure callback if an attempt is made to free certain invalid encodings.
3138 Only CHOICE structures using a callback which do not handle NULL value are
3141 This issue was reported to OpenSSL by Tyler Nighswander of ForAllSecure.
3146 * Montgomery multiplication may produce incorrect results
3148 There is a carry propagating bug in the Broadwell-specific Montgomery
3149 multiplication procedure that handles input lengths divisible by, but
3150 longer than 256 bits. Analysis suggests that attacks against RSA, DSA
3151 and DH private keys are impossible. This is because the subroutine in
3152 question is not used in operations with the private key itself and an input
3153 of the attacker's direct choice. Otherwise the bug can manifest itself as
3154 transient authentication and key negotiation failures or reproducible
3155 erroneous outcome of public-key operations with specially crafted input.
3156 Among EC algorithms only Brainpool P-512 curves are affected and one
3157 presumably can attack ECDH key negotiation. Impact was not analyzed in
3158 detail, because pre-requisites for attack are considered unlikely. Namely
3159 multiple clients have to choose the curve in question and the server has to
3160 share the private key among them, neither of which is default behaviour.
3161 Even then only clients that chose the curve will be affected.
3163 This issue was publicly reported as transient failures and was not
3164 initially recognized as a security issue. Thanks to Richard Morgan for
3165 providing reproducible case.
3170 * Removed automatic addition of RPATH in shared libraries and executables,
3171 as this was a remainder from OpenSSL 1.0.x and isn't needed any more.
3175 ### Changes between 1.1.0a and 1.1.0b [26 Sep 2016]
3177 * Fix Use After Free for large message sizes
3179 The patch applied to address CVE-2016-6307 resulted in an issue where if a
3180 message larger than approx 16k is received then the underlying buffer to
3181 store the incoming message is reallocated and moved. Unfortunately a
3182 dangling pointer to the old location is left which results in an attempt to
3183 write to the previously freed location. This is likely to result in a
3184 crash, however it could potentially lead to execution of arbitrary code.
3186 This issue only affects OpenSSL 1.1.0a.
3188 This issue was reported to OpenSSL by Robert Święcki.
3193 ### Changes between 1.1.0 and 1.1.0a [22 Sep 2016]
3195 * OCSP Status Request extension unbounded memory growth
3197 A malicious client can send an excessively large OCSP Status Request
3198 extension. If that client continually requests renegotiation, sending a
3199 large OCSP Status Request extension each time, then there will be unbounded
3200 memory growth on the server. This will eventually lead to a Denial Of
3201 Service attack through memory exhaustion. Servers with a default
3202 configuration are vulnerable even if they do not support OCSP. Builds using
3203 the "no-ocsp" build time option are not affected.
3205 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
3210 * SSL_peek() hang on empty record
3212 OpenSSL 1.1.0 SSL/TLS will hang during a call to SSL_peek() if the peer
3213 sends an empty record. This could be exploited by a malicious peer in a
3214 Denial Of Service attack.
3216 This issue was reported to OpenSSL by Alex Gaynor.
3221 * Excessive allocation of memory in tls_get_message_header() and
3222 dtls1_preprocess_fragment()
3224 A (D)TLS message includes 3 bytes for its length in the header for the
3225 message. This would allow for messages up to 16Mb in length. Messages of
3226 this length are excessive and OpenSSL includes a check to ensure that a
3227 peer is sending reasonably sized messages in order to avoid too much memory
3228 being consumed to service a connection. A flaw in the logic of version
3229 1.1.0 means that memory for the message is allocated too early, prior to
3230 the excessive message length check. Due to way memory is allocated in
3231 OpenSSL this could mean an attacker could force up to 21Mb to be allocated
3232 to service a connection. This could lead to a Denial of Service through
3233 memory exhaustion. However, the excessive message length check still takes
3234 place, and this would cause the connection to immediately fail. Assuming
3235 that the application calls SSL_free() on the failed connection in a timely
3236 manner then the 21Mb of allocated memory will then be immediately freed
3237 again. Therefore the excessive memory allocation will be transitory in
3238 nature. This then means that there is only a security impact if:
3240 1) The application does not call SSL_free() in a timely manner in the event
3241 that the connection fails
3243 2) The application is working in a constrained environment where there is
3244 very little free memory
3246 3) The attacker initiates multiple connection attempts such that there are
3247 multiple connections in a state where memory has been allocated for the
3248 connection; SSL_free() has not yet been called; and there is insufficient
3249 memory to service the multiple requests.
3251 Except in the instance of (1) above any Denial Of Service is likely to be
3252 transitory because as soon as the connection fails the memory is
3253 subsequently freed again in the SSL_free() call. However there is an
3254 increased risk during this period of application crashes due to the lack of
3255 memory - which would then mean a more serious Denial of Service.
3257 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
3258 (CVE-2016-6307 and CVE-2016-6308)
3262 * solaris-x86-cc, i.e. 32-bit configuration with vendor compiler,
3263 had to be removed. Primary reason is that vendor assembler can't
3264 assemble our modules with -KPIC flag. As result it, assembly
3265 support, was not even available as option. But its lack means
3266 lack of side-channel resistant code, which is incompatible with
3267 security by todays standards. Fortunately gcc is readily available
3268 prepackaged option, which we firmly point at...
3272 ### Changes between 1.0.2h and 1.1.0 [25 Aug 2016]
3274 * Windows command-line tool supports UTF-8 opt-in option for arguments
3275 and console input. Setting OPENSSL_WIN32_UTF8 environment variable
3276 (to any value) allows Windows user to access PKCS#12 file generated
3277 with Windows CryptoAPI and protected with non-ASCII password, as well
3278 as files generated under UTF-8 locale on Linux also protected with
3283 * To mitigate the SWEET32 attack ([CVE-2016-2183]), 3DES cipher suites
3284 have been disabled by default and removed from DEFAULT, just like RC4.
3285 See the RC4 item below to re-enable both.
3289 * The method for finding the storage location for the Windows RAND seed file
3290 has changed. First we check %RANDFILE%. If that is not set then we check
3291 the directories %HOME%, %USERPROFILE% and %SYSTEMROOT% in that order. If
3292 all else fails we fall back to C:\.
3296 * The EVP_EncryptUpdate() function has had its return type changed from void
3297 to int. A return of 0 indicates and error while a return of 1 indicates
3302 * The flags RSA_FLAG_NO_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME and
3303 DH_FLAG_NO_EXP_CONSTTIME which previously provided the ability to switch
3304 off the constant time implementation for RSA, DSA and DH have been made
3305 no-ops and deprecated.
3309 * Windows RAND implementation was simplified to only get entropy by
3310 calling CryptGenRandom(). Various other RAND-related tickets
3313 *Joseph Wylie Yandle, Rich Salz*
3315 * The stack and lhash API's were renamed to start with `OPENSSL_SK_`
3316 and `OPENSSL_LH_`, respectively. The old names are available
3317 with API compatibility. They new names are now completely documented.
3321 * Unify TYPE_up_ref(obj) methods signature.
3322 SSL_CTX_up_ref(), SSL_up_ref(), X509_up_ref(), EVP_PKEY_up_ref(),
3323 X509_CRL_up_ref(), X509_OBJECT_up_ref_count() methods are now returning an
3324 int (instead of void) like all others TYPE_up_ref() methods.
3325 So now these methods also check the return value of CRYPTO_atomic_add(),
3326 and the validity of object reference counter.
3328 *fdasilvayy@gmail.com*
3330 * With Windows Visual Studio builds, the .pdb files are installed
3331 alongside the installed libraries and executables. For a static
3332 library installation, ossl_static.pdb is the associate compiler
3333 generated .pdb file to be used when linking programs.
3337 * Remove openssl.spec. Packaging files belong with the packagers.
3341 * Automatic Darwin/OSX configuration has had a refresh, it will now
3342 recognise x86_64 architectures automatically. You can still decide
3343 to build for a different bitness with the environment variable
3344 KERNEL_BITS (can be 32 or 64), for example:
3346 KERNEL_BITS=32 ./config
3350 * Change default algorithms in pkcs8 utility to use PKCS#5 v2.0,
3351 256 bit AES and HMAC with SHA256.
3355 * Remove support for MIPS o32 ABI on IRIX (and IRIX only).
3359 * Triple-DES ciphers have been moved from HIGH to MEDIUM.
3363 * To enable users to have their own config files and build file templates,
3364 Configure looks in the directory indicated by the environment variable
3365 OPENSSL_LOCAL_CONFIG_DIR as well as the in-source Configurations/
3366 directory. On VMS, OPENSSL_LOCAL_CONFIG_DIR is expected to be a logical
3367 name and is used as is.
3371 * The following datatypes were made opaque: X509_OBJECT, X509_STORE_CTX,
3372 X509_STORE, X509_LOOKUP, and X509_LOOKUP_METHOD. The unused type
3373 X509_CERT_FILE_CTX was removed.
3377 * "shared" builds are now the default. To create only static libraries use
3378 the "no-shared" Configure option.
3382 * Remove the no-aes, no-hmac, no-rsa, no-sha and no-md5 Configure options.
3383 All of these option have not worked for some while and are fundamental
3388 * Make various cleanup routines no-ops and mark them as deprecated. Most
3389 global cleanup functions are no longer required because they are handled
3390 via auto-deinit (see OPENSSL_init_crypto and OPENSSL_init_ssl man pages).
3391 Explicitly de-initing can cause problems (e.g. where a library that uses
3392 OpenSSL de-inits, but an application is still using it). The affected
3393 functions are CONF_modules_free(), ENGINE_cleanup(), OBJ_cleanup(),
3394 EVP_cleanup(), BIO_sock_cleanup(), CRYPTO_cleanup_all_ex_data(),
3395 RAND_cleanup(), SSL_COMP_free_compression_methods(), ERR_free_strings() and
3396 COMP_zlib_cleanup().
3400 * --strict-warnings no longer enables runtime debugging options
3401 such as REF_DEBUG. Instead, debug options are automatically
3402 enabled with '--debug' builds.
3404 *Andy Polyakov, Emilia Käsper*
3406 * Made DH and DH_METHOD opaque. The structures for managing DH objects
3407 have been moved out of the public header files. New functions for managing
3408 these have been added.
3412 * Made RSA and RSA_METHOD opaque. The structures for managing RSA
3413 objects have been moved out of the public header files. New
3414 functions for managing these have been added.
3418 * Made DSA and DSA_METHOD opaque. The structures for managing DSA objects
3419 have been moved out of the public header files. New functions for managing
3420 these have been added.
3424 * Made BIO and BIO_METHOD opaque. The structures for managing BIOs have been
3425 moved out of the public header files. New functions for managing these
3430 * Removed no-rijndael as a config option. Rijndael is an old name for AES.
3434 * Removed the mk1mf build scripts.
3438 * Headers are now wrapped, if necessary, with OPENSSL_NO_xxx, so
3439 it is always safe to #include a header now.
3443 * Removed the aged BC-32 config and all its supporting scripts
3447 * Removed support for Ultrix, Netware, and OS/2.
3451 * Add support for HKDF.
3453 *Alessandro Ghedini*
3455 * Add support for blake2b and blake2s
3459 * Added support for "pipelining". Ciphers that have the
3460 EVP_CIPH_FLAG_PIPELINE flag set have a capability to process multiple
3461 encryptions/decryptions simultaneously. There are currently no built-in
3462 ciphers with this property but the expectation is that engines will be able
3463 to offer it to significantly improve throughput. Support has been extended
3464 into libssl so that multiple records for a single connection can be
3465 processed in one go (for >=TLS 1.1).
3469 * Added the AFALG engine. This is an async capable engine which is able to
3470 offload work to the Linux kernel. In this initial version it only supports
3471 AES128-CBC. The kernel must be version 4.1.0 or greater.
3475 * OpenSSL now uses a new threading API. It is no longer necessary to
3476 set locking callbacks to use OpenSSL in a multi-threaded environment. There
3477 are two supported threading models: pthreads and windows threads. It is
3478 also possible to configure OpenSSL at compile time for "no-threads". The
3479 old threading API should no longer be used. The functions have been
3480 replaced with "no-op" compatibility macros.
3482 *Alessandro Ghedini, Matt Caswell*
3484 * Modify behavior of ALPN to invoke callback after SNI/servername
3485 callback, such that updates to the SSL_CTX affect ALPN.
3489 * Add SSL_CIPHER queries for authentication and key-exchange.
3493 * Changes to the DEFAULT cipherlist:
3494 - Prefer (EC)DHE handshakes over plain RSA.
3495 - Prefer AEAD ciphers over legacy ciphers.
3496 - Prefer ECDSA over RSA when both certificates are available.
3497 - Prefer TLSv1.2 ciphers/PRF.
3498 - Remove DSS, SEED, IDEA, CAMELLIA, and AES-CCM from the
3503 * Change the ECC default curve list to be this, in order: x25519,
3504 secp256r1, secp521r1, secp384r1.
3508 * RC4 based libssl ciphersuites are now classed as "weak" ciphers and are
3509 disabled by default. They can be re-enabled using the
3510 enable-weak-ssl-ciphers option to Configure.
3514 * If the server has ALPN configured, but supports no protocols that the
3515 client advertises, send a fatal "no_application_protocol" alert.
3516 This behaviour is SHALL in RFC 7301, though it isn't universally
3517 implemented by other servers.
3521 * Add X25519 support.
3522 Add ASN.1 and EVP_PKEY methods for X25519. This includes support
3523 for public and private key encoding using the format documented in
3524 draft-ietf-curdle-pkix-02. The corresponding EVP_PKEY method supports
3525 key generation and key derivation.
3527 TLS support complies with draft-ietf-tls-rfc4492bis-08 and uses
3532 * Deprecate SRP_VBASE_get_by_user.
3533 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
3534 In order to fix an unavoidable memory leak ([CVE-2016-0798]),
3535 SRP_VBASE_get_by_user was changed to ignore the "fake user" SRP
3536 seed, even if the seed is configured.
3538 Users should use SRP_VBASE_get1_by_user instead. Note that in
3539 SRP_VBASE_get1_by_user, caller must free the returned value. Note
3540 also that even though configuring the SRP seed attempts to hide
3541 invalid usernames by continuing the handshake with fake
3542 credentials, this behaviour is not constant time and no strong
3543 guarantees are made that the handshake is indistinguishable from
3544 that of a valid user.
3548 * Configuration change; it's now possible to build dynamic engines
3549 without having to build shared libraries and vice versa. This
3550 only applies to the engines in `engines/`, those in `crypto/engine/`
3551 will always be built into libcrypto (i.e. "static").
3553 Building dynamic engines is enabled by default; to disable, use
3554 the configuration option "disable-dynamic-engine".
3556 The only requirements for building dynamic engines are the
3557 presence of the DSO module and building with position independent
3558 code, so they will also automatically be disabled if configuring
3559 with "disable-dso" or "disable-pic".
3561 The macros OPENSSL_NO_STATIC_ENGINE and OPENSSL_NO_DYNAMIC_ENGINE
3562 are also taken away from openssl/opensslconf.h, as they are
3567 * Configuration change; if there is a known flag to compile
3568 position independent code, it will always be applied on the
3569 libcrypto and libssl object files, and never on the application
3570 object files. This means other libraries that use routines from
3571 libcrypto / libssl can be made into shared libraries regardless
3572 of how OpenSSL was configured.
3574 If this isn't desirable, the configuration options "disable-pic"
3575 or "no-pic" can be used to disable the use of PIC. This will
3576 also disable building shared libraries and dynamic engines.
3580 * Removed JPAKE code. It was experimental and has no wide use.
3584 * The INSTALL_PREFIX Makefile variable has been renamed to
3585 DESTDIR. That makes for less confusion on what this variable
3586 is for. Also, the configuration option --install_prefix is
3591 * Heartbeat for TLS has been removed and is disabled by default
3592 for DTLS; configure with enable-heartbeats. Code that uses the
3593 old #define's might need to be updated.
3595 *Emilia Käsper, Rich Salz*
3597 * Rename REF_CHECK to REF_DEBUG.
3601 * New "unified" build system
3603 The "unified" build system is aimed to be a common system for all
3604 platforms we support. With it comes new support for VMS.
3606 This system builds supports building in a different directory tree
3607 than the source tree. It produces one Makefile (for unix family
3608 or lookalikes), or one descrip.mms (for VMS).
3610 The source of information to make the Makefile / descrip.mms is
3611 small files called 'build.info', holding the necessary
3612 information for each directory with source to compile, and a
3613 template in Configurations, like unix-Makefile.tmpl or
3616 With this change, the library names were also renamed on Windows
3617 and on VMS. They now have names that are closer to the standard
3618 on Unix, and include the major version number, and in certain
3619 cases, the architecture they are built for. See "Notes on shared
3620 libraries" in INSTALL.
3622 We rely heavily on the perl module Text::Template.
3626 * Added support for auto-initialisation and de-initialisation of the library.
3627 OpenSSL no longer requires explicit init or deinit routines to be called,
3628 except in certain circumstances. See the OPENSSL_init_crypto() and
3629 OPENSSL_init_ssl() man pages for further information.
3633 * The arguments to the DTLSv1_listen function have changed. Specifically the
3634 "peer" argument is now expected to be a BIO_ADDR object.
3636 * Rewrite of BIO networking library. The BIO library lacked consistent
3637 support of IPv6, and adding it required some more extensive
3638 modifications. This introduces the BIO_ADDR and BIO_ADDRINFO types,
3639 which hold all types of addresses and chains of address information.
3640 It also introduces a new API, with functions like BIO_socket,
3641 BIO_connect, BIO_listen, BIO_lookup and a rewrite of BIO_accept.
3642 The source/sink BIOs BIO_s_connect, BIO_s_accept and BIO_s_datagram
3643 have been adapted accordingly.
3647 * RSA_padding_check_PKCS1_type_1 now accepts inputs with and without
3652 * CRIME protection: disable compression by default, even if OpenSSL is
3653 compiled with zlib enabled. Applications can still enable compression
3654 by calling SSL_CTX_clear_options(ctx, SSL_OP_NO_COMPRESSION), or by
3655 using the SSL_CONF library to configure compression.
3659 * The signature of the session callback configured with
3660 SSL_CTX_sess_set_get_cb was changed. The read-only input buffer
3661 was explicitly marked as `const unsigned char*` instead of
3666 * Always DPURIFY. Remove the use of uninitialized memory in the
3667 RNG, and other conditional uses of DPURIFY. This makes -DPURIFY a no-op.
3671 * Removed many obsolete configuration items, including
3672 DES_PTR, DES_RISC1, DES_RISC2, DES_INT
3673 MD2_CHAR, MD2_INT, MD2_LONG
3675 IDEA_SHORT, IDEA_LONG
3676 RC2_SHORT, RC2_LONG, RC4_LONG, RC4_CHUNK, RC4_INDEX
3678 *Rich Salz, with advice from Andy Polyakov*
3680 * Many BN internals have been moved to an internal header file.
3682 *Rich Salz with help from Andy Polyakov*
3684 * Configuration and writing out the results from it has changed.
3685 Files such as Makefile include/openssl/opensslconf.h and are now
3686 produced through general templates, such as Makefile.in and
3687 crypto/opensslconf.h.in and some help from the perl module
3690 Also, the center of configuration information is no longer
3691 Makefile. Instead, Configure produces a perl module in
3692 configdata.pm which holds most of the config data (in the hash
3693 table %config), the target data that comes from the target
3694 configuration in one of the `Configurations/*.conf` files (in
3699 * To clarify their intended purposes, the Configure options
3700 --prefix and --openssldir change their semantics, and become more
3701 straightforward and less interdependent.
3703 --prefix shall be used exclusively to give the location INSTALLTOP
3704 where programs, scripts, libraries, include files and manuals are
3705 going to be installed. The default is now /usr/local.
3707 --openssldir shall be used exclusively to give the default
3708 location OPENSSLDIR where certificates, private keys, CRLs are
3709 managed. This is also where the default openssl.cnf gets
3711 If the directory given with this option is a relative path, the
3712 values of both the --prefix value and the --openssldir value will
3713 be combined to become OPENSSLDIR.
3714 The default for --openssldir is INSTALLTOP/ssl.
3716 Anyone who uses --openssldir to specify where OpenSSL is to be
3717 installed MUST change to use --prefix instead.
3721 * The GOST engine was out of date and therefore it has been removed. An up
3722 to date GOST engine is now being maintained in an external repository.
3723 See: <https://wiki.openssl.org/index.php/Binaries>. Libssl still retains
3724 support for GOST ciphersuites (these are only activated if a GOST engine
3729 * EGD is no longer supported by default; use enable-egd when
3732 *Ben Kaduk and Rich Salz*
3734 * The distribution now has Makefile.in files, which are used to
3735 create Makefile's when Configure is run. *Configure must be run
3736 before trying to build now.*
3740 * The return value for SSL_CIPHER_description() for error conditions
3745 * Support for RFC6698/RFC7671 DANE TLSA peer authentication.
3747 Obtaining and performing DNSSEC validation of TLSA records is
3748 the application's responsibility. The application provides
3749 the TLSA records of its choice to OpenSSL, and these are then
3750 used to authenticate the peer.
3752 The TLSA records need not even come from DNS. They can, for
3753 example, be used to implement local end-entity certificate or
3754 trust-anchor "pinning", where the "pin" data takes the form
3755 of TLSA records, which can augment or replace verification
3756 based on the usual WebPKI public certification authorities.
3760 * Revert default OPENSSL_NO_DEPRECATED setting. Instead OpenSSL
3761 continues to support deprecated interfaces in default builds.
3762 However, applications are strongly advised to compile their
3763 source files with -DOPENSSL_API_COMPAT=0x10100000L, which hides
3764 the declarations of all interfaces deprecated in 0.9.8, 1.0.0
3765 or the 1.1.0 releases.
3767 In environments in which all applications have been ported to
3768 not use any deprecated interfaces OpenSSL's Configure script
3769 should be used with the --api=1.1.0 option to entirely remove
3770 support for the deprecated features from the library and
3771 unconditionally disable them in the installed headers.
3772 Essentially the same effect can be achieved with the "no-deprecated"
3773 argument to Configure, except that this will always restrict
3774 the build to just the latest API, rather than a fixed API
3777 As applications are ported to future revisions of the API,
3778 they should update their compile-time OPENSSL_API_COMPAT define
3779 accordingly, but in most cases should be able to continue to
3780 compile with later releases.
3782 The OPENSSL_API_COMPAT versions for 1.0.0, and 0.9.8 are
3783 0x10000000L and 0x00908000L, respectively. However those
3784 versions did not support the OPENSSL_API_COMPAT feature, and
3785 so applications are not typically tested for explicit support
3786 of just the undeprecated features of either release.
3790 * Add support for setting the minimum and maximum supported protocol.
3791 It can bet set via the SSL_set_min_proto_version() and
3792 SSL_set_max_proto_version(), or via the SSL_CONF's MinProtocol and
3793 MaxProtocol. It's recommended to use the new APIs to disable
3794 protocols instead of disabling individual protocols using
3795 SSL_set_options() or SSL_CONF's Protocol. This change also
3796 removes support for disabling TLS 1.2 in the OpenSSL TLS
3797 client at compile time by defining OPENSSL_NO_TLS1_2_CLIENT.
3801 * Support for ChaCha20 and Poly1305 added to libcrypto and libssl.
3805 * New EC_KEY_METHOD, this replaces the older ECDSA_METHOD and ECDH_METHOD
3806 and integrates ECDSA and ECDH functionality into EC. Implementations can
3807 now redirect key generation and no longer need to convert to or from
3810 Note: the ecdsa.h and ecdh.h headers are now no longer needed and just
3811 include the ec.h header file instead.
3815 * Remove support for all 40 and 56 bit ciphers. This includes all the export
3816 ciphers who are no longer supported and drops support the ephemeral RSA key
3817 exchange. The LOW ciphers currently doesn't have any ciphers in it.
3821 * Made EVP_MD_CTX, EVP_MD, EVP_CIPHER_CTX, EVP_CIPHER and HMAC_CTX
3822 opaque. For HMAC_CTX, the following constructors and destructors
3825 HMAC_CTX *HMAC_CTX_new(void);
3826 void HMAC_CTX_free(HMAC_CTX *ctx);
3828 For EVP_MD and EVP_CIPHER, complete APIs to create, fill and
3829 destroy such methods has been added. See EVP_MD_meth_new(3) and
3830 EVP_CIPHER_meth_new(3) for documentation.
3833 1) `EVP_MD_CTX_cleanup()`, `EVP_CIPHER_CTX_cleanup()` and
3834 `HMAC_CTX_cleanup()` were removed. `HMAC_CTX_reset()` and
3835 `EVP_MD_CTX_reset()` should be called instead to reinitialise
3836 an already created structure.
3837 2) For consistency with the majority of our object creators and
3838 destructors, `EVP_MD_CTX_(create|destroy)` were renamed to
3839 `EVP_MD_CTX_(new|free)`. The old names are retained as macros
3840 for deprecated builds.
3844 * Added ASYNC support. Libcrypto now includes the async sub-library to enable
3845 cryptographic operations to be performed asynchronously as long as an
3846 asynchronous capable engine is used. See the ASYNC_start_job() man page for
3847 further details. Libssl has also had this capability integrated with the
3848 introduction of the new mode SSL_MODE_ASYNC and associated error
3849 SSL_ERROR_WANT_ASYNC. See the SSL_CTX_set_mode() and SSL_get_error() man
3850 pages. This work was developed in partnership with Intel Corp.
3854 * SSL_{CTX_}set_ecdh_auto() has been removed and ECDH is support is
3855 always enabled now. If you want to disable the support you should
3856 exclude it using the list of supported ciphers. This also means that the
3857 "-no_ecdhe" option has been removed from s_server.
3861 * SSL_{CTX}_set_tmp_ecdh() which can set 1 EC curve now internally calls
3862 SSL_{CTX_}set1_curves() which can set a list.
3866 * Remove support for SSL_{CTX_}set_tmp_ecdh_callback(). You should set the
3867 curve you want to support using SSL_{CTX_}set1_curves().
3871 * State machine rewrite. The state machine code has been significantly
3872 refactored in order to remove much duplication of code and solve issues
3873 with the old code (see [ssl/statem/README.md](ssl/statem/README.md) for
3874 further details). This change does have some associated API changes.
3875 Notably the SSL_state() function has been removed and replaced by
3876 SSL_get_state which now returns an "OSSL_HANDSHAKE_STATE" instead of an int.
3877 SSL_set_state() has been removed altogether. The previous handshake states
3878 defined in ssl.h and ssl3.h have also been removed.
3882 * All instances of the string "ssleay" in the public API were replaced
3883 with OpenSSL (case-matching; e.g., OPENSSL_VERSION for #define's)
3884 Some error codes related to internal RSA_eay API's were renamed.
3888 * The demo files in crypto/threads were moved to demo/threads.
3892 * Removed obsolete engines: 4758cca, aep, atalla, cswift, nuron, gmp,
3895 *Matt Caswell, Rich Salz*
3897 * New ASN.1 embed macro.
3899 New ASN.1 macro ASN1_EMBED. This is the same as ASN1_SIMPLE except the
3900 structure is not allocated: it is part of the parent. That is instead of
3908 This reduces memory fragmentation and make it impossible to accidentally
3909 set a mandatory field to NULL.
3911 This currently only works for some fields specifically a SEQUENCE, CHOICE,
3912 or ASN1_STRING type which is part of a parent SEQUENCE. Since it is
3913 equivalent to ASN1_SIMPLE it cannot be tagged, OPTIONAL, SET OF or
3918 * Remove EVP_CHECK_DES_KEY, a compile-time option that never compiled.
3922 * Removed DES and RC4 ciphersuites from DEFAULT. Also removed RC2 although
3923 in 1.0.2 EXPORT was already removed and the only RC2 ciphersuite is also
3924 an EXPORT one. COMPLEMENTOFDEFAULT has been updated accordingly to add
3925 DES and RC4 ciphersuites.
3929 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
3930 This changes the decoding behaviour for some invalid messages,
3931 though the change is mostly in the more lenient direction, and
3932 legacy behaviour is preserved as much as possible.
3936 * Fix no-stdio build.
3937 *David Woodhouse <David.Woodhouse@intel.com> and also*
3938 *Ivan Nestlerode <ivan.nestlerode@sonos.com>*
3940 * New testing framework
3941 The testing framework has been largely rewritten and is now using
3942 perl and the perl modules Test::Harness and an extended variant of
3943 Test::More called OpenSSL::Test to do its work. All test scripts in
3944 test/ have been rewritten into test recipes, and all direct calls to
3945 executables in test/Makefile have become individual recipes using the
3946 simplified testing OpenSSL::Test::Simple.
3948 For documentation on our testing modules, do:
3950 perldoc test/testlib/OpenSSL/Test/Simple.pm
3951 perldoc test/testlib/OpenSSL/Test.pm
3955 * Revamped memory debug; only -DCRYPTO_MDEBUG and -DCRYPTO_MDEBUG_ABORT
3956 are used; the latter aborts on memory leaks (usually checked on exit).
3957 Some undocumented "set malloc, etc., hooks" functions were removed
3958 and others were changed. All are now documented.
3962 * In DSA_generate_parameters_ex, if the provided seed is too short,
3965 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
3967 * Rewrite PSK to support ECDHE_PSK, DHE_PSK and RSA_PSK. Add ciphersuites
3968 from RFC4279, RFC4785, RFC5487, RFC5489.
3970 Thanks to Christian J. Dietrich and Giuseppe D'Angelo for the
3971 original RSA_PSK patch.
3975 * Dropped support for the SSL3_FLAGS_DELAY_CLIENT_FINISHED flag. This SSLeay
3976 era flag was never set throughout the codebase (only read). Also removed
3977 SSL3_FLAGS_POP_BUFFER which was only used if
3978 SSL3_FLAGS_DELAY_CLIENT_FINISHED was also set.
3982 * Changed the default name options in the "ca", "crl", "req" and "x509"
3983 to be "oneline" instead of "compat".
3987 * Remove SSL_OP_TLS_BLOCK_PADDING_BUG. This is SSLeay legacy, we're
3988 not aware of clients that still exhibit this bug, and the workaround
3989 hasn't been working properly for a while.
3993 * The return type of BIO_number_read() and BIO_number_written() as well as
3994 the corresponding num_read and num_write members in the BIO structure has
3995 changed from unsigned long to uint64_t. On platforms where an unsigned
3996 long is 32 bits (e.g. Windows) these counters could overflow if >4Gb is
4001 * Given the pervasive nature of TLS extensions it is inadvisable to run
4002 OpenSSL without support for them. It also means that maintaining
4003 the OPENSSL_NO_TLSEXT option within the code is very invasive (and probably
4004 not well tested). Therefore the OPENSSL_NO_TLSEXT option has been removed.
4008 * Removed support for the two export grade static DH ciphersuites
4009 EXP-DH-RSA-DES-CBC-SHA and EXP-DH-DSS-DES-CBC-SHA. These two ciphersuites
4010 were newly added (along with a number of other static DH ciphersuites) to
4011 1.0.2. However the two export ones have *never* worked since they were
4012 introduced. It seems strange in any case to be adding new export
4013 ciphersuites, and given "logjam" it also does not seem correct to fix them.
4017 * Version negotiation has been rewritten. In particular SSLv23_method(),
4018 SSLv23_client_method() and SSLv23_server_method() have been deprecated,
4019 and turned into macros which simply call the new preferred function names
4020 TLS_method(), TLS_client_method() and TLS_server_method(). All new code
4021 should use the new names instead. Also as part of this change the ssl23.h
4022 header file has been removed.
4026 * Support for Kerberos ciphersuites in TLS (RFC2712) has been removed. This
4027 code and the associated standard is no longer considered fit-for-purpose.
4031 * RT2547 was closed. When generating a private key, try to make the
4032 output file readable only by the owner. This behavior change might
4033 be noticeable when interacting with other software.
4035 * Documented all exdata functions. Added CRYPTO_free_ex_index.
4040 * Added HTTP GET support to the ocsp command.
4044 * Changed default digest for the dgst and enc commands from MD5 to
4049 * RAND_pseudo_bytes has been deprecated. Users should use RAND_bytes instead.
4053 * Added support for TLS extended master secret from
4054 draft-ietf-tls-session-hash-03.txt. Thanks for Alfredo Pironti for an
4055 initial patch which was a great help during development.
4059 * All libssl internal structures have been removed from the public header
4060 files, and the OPENSSL_NO_SSL_INTERN option has been removed (since it is
4061 now redundant). Users should not attempt to access internal structures
4062 directly. Instead they should use the provided API functions.
4066 * config has been changed so that by default OPENSSL_NO_DEPRECATED is used.
4067 Access to deprecated functions can be re-enabled by running config with
4068 "enable-deprecated". In addition applications wishing to use deprecated
4069 functions must define OPENSSL_USE_DEPRECATED. Note that this new behaviour
4070 will, by default, disable some transitive includes that previously existed
4071 in the header files (e.g. ec.h will no longer, by default, include bn.h)
4075 * Added support for OCB mode. OpenSSL has been granted a patent license
4076 compatible with the OpenSSL license for use of OCB. Details are available
4077 at <https://www.openssl.org/source/OCB-patent-grant-OpenSSL.pdf>. Support
4078 for OCB can be removed by calling config with no-ocb.
4082 * SSLv2 support has been removed. It still supports receiving a SSLv2
4083 compatible client hello.
4087 * Increased the minimal RSA keysize from 256 to 512 bits [Rich Salz],
4088 done while fixing the error code for the key-too-small case.
4090 *Annie Yousar <a.yousar@informatik.hu-berlin.de>*
4092 * CA.sh has been removed; use CA.pl instead.
4096 * Removed old DES API.
4100 * Remove various unsupported platforms:
4106 Sinix/ReliantUNIX RM400
4111 16-bit platforms such as WIN16
4115 * Clean up OPENSSL_NO_xxx #define's
4116 - Use setbuf() and remove OPENSSL_NO_SETVBUF_IONBF
4117 - Rename OPENSSL_SYSNAME_xxx to OPENSSL_SYS_xxx
4118 - OPENSSL_NO_EC{DH,DSA} merged into OPENSSL_NO_EC
4119 - OPENSSL_NO_RIPEMD160, OPENSSL_NO_RIPEMD merged into OPENSSL_NO_RMD160
4120 - OPENSSL_NO_FP_API merged into OPENSSL_NO_STDIO
4121 - Remove OPENSSL_NO_BIO OPENSSL_NO_BUFFER OPENSSL_NO_CHAIN_VERIFY
4122 OPENSSL_NO_EVP OPENSSL_NO_FIPS_ERR OPENSSL_NO_HASH_COMP
4123 OPENSSL_NO_LHASH OPENSSL_NO_OBJECT OPENSSL_NO_SPEED OPENSSL_NO_STACK
4124 OPENSSL_NO_X509 OPENSSL_NO_X509_VERIFY
4125 - Remove MS_STATIC; it's a relic from platforms <32 bits.
4129 * Cleaned up dead code
4130 Remove all but one '#ifdef undef' which is to be looked at.
4134 * Clean up calling of xxx_free routines.
4135 Just like free(), fix most of the xxx_free routines to accept
4136 NULL. Remove the non-null checks from callers. Save much code.
4140 * Add secure heap for storage of private keys (when possible).
4141 Add BIO_s_secmem(), CBIGNUM, etc.
4142 Contributed by Akamai Technologies under our Corporate CLA.
4146 * Experimental support for a new, fast, unbiased prime candidate generator,
4147 bn_probable_prime_dh_coprime(). Not currently used by any prime generator.
4149 *Felix Laurie von Massenbach <felix@erbridge.co.uk>*
4151 * New output format NSS in the sess_id command line tool. This allows
4152 exporting the session id and the master key in NSS keylog format.
4154 *Martin Kaiser <martin@kaiser.cx>*
4156 * Harmonize version and its documentation. -f flag is used to display
4159 *mancha <mancha1@zoho.com>*
4161 * Fix eckey_priv_encode so it immediately returns an error upon a failure
4162 in i2d_ECPrivateKey. Thanks to Ted Unangst for feedback on this issue.
4164 *mancha <mancha1@zoho.com>*
4166 * Fix some double frees. These are not thought to be exploitable.
4168 *mancha <mancha1@zoho.com>*
4170 * A missing bounds check in the handling of the TLS heartbeat extension
4171 can be used to reveal up to 64k of memory to a connected client or
4174 Thanks for Neel Mehta of Google Security for discovering this bug and to
4175 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
4176 preparing the fix ([CVE-2014-0160])
4178 *Adam Langley, Bodo Moeller*
4180 * Fix for the attack described in the paper "Recovering OpenSSL
4181 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
4182 by Yuval Yarom and Naomi Benger. Details can be obtained from:
4183 <http://eprint.iacr.org/2014/140>
4185 Thanks to Yuval Yarom and Naomi Benger for discovering this
4186 flaw and to Yuval Yarom for supplying a fix ([CVE-2014-0076])
4188 *Yuval Yarom and Naomi Benger*
4190 * Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
4191 this fixes a limitation in previous versions of OpenSSL.
4195 * Experimental encrypt-then-mac support.
4197 Experimental support for encrypt then mac from
4198 draft-gutmann-tls-encrypt-then-mac-02.txt
4200 To enable it set the appropriate extension number (0x42 for the test
4201 server) using e.g. -DTLSEXT_TYPE_encrypt_then_mac=0x42
4203 For non-compliant peers (i.e. just about everything) this should have no
4206 WARNING: EXPERIMENTAL, SUBJECT TO CHANGE.
4210 * Add EVP support for key wrapping algorithms, to avoid problems with
4211 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
4212 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
4213 algorithms and include tests cases.
4217 * Extend CMS code to support RSA-PSS signatures and RSA-OAEP for
4222 * Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
4223 MGF1 digest and OAEP label.
4227 * Make openssl verify return errors.
4229 *Chris Palmer <palmer@google.com> and Ben Laurie*
4231 * New function ASN1_TIME_diff to calculate the difference between two
4232 ASN1_TIME structures or one structure and the current time.
4236 * Update fips_test_suite to support multiple command line options. New
4237 test to induce all self test errors in sequence and check expected
4242 * Add FIPS_{rsa,dsa,ecdsa}_{sign,verify} functions which digest and
4243 sign or verify all in one operation.
4247 * Add fips_algvs: a multicall fips utility incorporating all the algorithm
4248 test programs and fips_test_suite. Includes functionality to parse
4249 the minimal script output of fipsalgest.pl directly.
4253 * Add authorisation parameter to FIPS_module_mode_set().
4257 * Add FIPS selftest for ECDH algorithm using P-224 and B-233 curves.
4261 * Use separate DRBG fields for internal and external flags. New function
4262 FIPS_drbg_health_check() to perform on demand health checking. Add
4263 generation tests to fips_test_suite with reduced health check interval to
4264 demonstrate periodic health checking. Add "nodh" option to
4265 fips_test_suite to skip very slow DH test.
4269 * New function FIPS_get_cipherbynid() to lookup FIPS supported ciphers
4274 * More extensive health check for DRBG checking many more failure modes.
4275 New function FIPS_selftest_drbg_all() to handle every possible DRBG
4276 combination: call this in fips_test_suite.
4280 * Add support for canonical generation of DSA parameter 'g'. See
4283 * Add support for HMAC DRBG from SP800-90. Update DRBG algorithm test and
4284 POST to handle HMAC cases.
4288 * Add functions FIPS_module_version() and FIPS_module_version_text()
4289 to return numerical and string versions of the FIPS module number.
4293 * Rename FIPS_mode_set and FIPS_mode to FIPS_module_mode_set and
4294 FIPS_module_mode. FIPS_mode and FIPS_mode_set will be implemented
4295 outside the validated module in the FIPS capable OpenSSL.
4299 * Minor change to DRBG entropy callback semantics. In some cases
4300 there is no multiple of the block length between min_len and
4301 max_len. Allow the callback to return more than max_len bytes
4302 of entropy but discard any extra: it is the callback's responsibility
4303 to ensure that the extra data discarded does not impact the
4304 requested amount of entropy.
4308 * Add PRNG security strength checks to RSA, DSA and ECDSA using
4309 information in FIPS186-3, SP800-57 and SP800-131A.
4313 * CCM support via EVP. Interface is very similar to GCM case except we
4314 must supply all data in one chunk (i.e. no update, final) and the
4315 message length must be supplied if AAD is used. Add algorithm test
4320 * Initial version of POST overhaul. Add POST callback to allow the status
4321 of POST to be monitored and/or failures induced. Modify fips_test_suite
4322 to use callback. Always run all selftests even if one fails.
4326 * XTS support including algorithm test driver in the fips_gcmtest program.
4327 Note: this does increase the maximum key length from 32 to 64 bytes but
4328 there should be no binary compatibility issues as existing applications
4329 will never use XTS mode.
4333 * Extensive reorganisation of FIPS PRNG behaviour. Remove all dependencies
4334 to OpenSSL RAND code and replace with a tiny FIPS RAND API which also
4335 performs algorithm blocking for unapproved PRNG types. Also do not
4336 set PRNG type in FIPS_mode_set(): leave this to the application.
4337 Add default OpenSSL DRBG handling: sets up FIPS PRNG and seeds with
4338 the standard OpenSSL PRNG: set additional data to a date time vector.
4342 * Rename old X9.31 PRNG functions of the form `FIPS_rand*` to `FIPS_x931*`.
4343 This shouldn't present any incompatibility problems because applications
4344 shouldn't be using these directly and any that are will need to rethink
4345 anyway as the X9.31 PRNG is now deprecated by FIPS 140-2
4349 * Extensive self tests and health checking required by SP800-90 DRBG.
4350 Remove strength parameter from FIPS_drbg_instantiate and always
4351 instantiate at maximum supported strength.
4355 * Add ECDH code to fips module and fips_ecdhvs for primitives only testing.
4359 * New algorithm test program fips_dhvs to handle DH primitives only testing.
4363 * New function DH_compute_key_padded() to compute a DH key and pad with
4364 leading zeroes if needed: this complies with SP800-56A et al.
4368 * Initial implementation of SP800-90 DRBGs for Hash and CTR. Not used by
4369 anything, incomplete, subject to change and largely untested at present.
4373 * Modify fipscanisteronly build option to only build the necessary object
4374 files by filtering FIPS_EX_OBJ through a perl script in crypto/Makefile.
4378 * Add experimental option FIPSSYMS to give all symbols in
4379 fipscanister.o and FIPS or fips prefix. This will avoid
4380 conflicts with future versions of OpenSSL. Add perl script
4381 util/fipsas.pl to preprocess assembly language source files
4382 and rename any affected symbols.
4386 * Add selftest checks and algorithm block of non-fips algorithms in
4387 FIPS mode. Remove DES2 from selftests.
4391 * Add ECDSA code to fips module. Add tiny fips_ecdsa_check to just
4392 return internal method without any ENGINE dependencies. Add new
4393 tiny fips sign and verify functions.
4397 * New build option no-ec2m to disable characteristic 2 code.
4401 * New build option "fipscanisteronly". This only builds fipscanister.o
4402 and (currently) associated fips utilities. Uses the file Makefile.fips
4403 instead of Makefile.org as the prototype.
4407 * Add some FIPS mode restrictions to GCM. Add internal IV generator.
4408 Update fips_gcmtest to use IV generator.
4412 * Initial, experimental EVP support for AES-GCM. AAD can be input by
4413 setting output buffer to NULL. The `*Final` function must be
4414 called although it will not retrieve any additional data. The tag
4415 can be set or retrieved with a ctrl. The IV length is by default 12
4416 bytes (96 bits) but can be set to an alternative value. If the IV
4417 length exceeds the maximum IV length (currently 16 bytes) it cannot be
4422 * New flag in ciphers: EVP_CIPH_FLAG_CUSTOM_CIPHER. This means the
4423 underlying do_cipher function handles all cipher semantics itself
4424 including padding and finalisation. This is useful if (for example)
4425 an ENGINE cipher handles block padding itself. The behaviour of
4426 do_cipher is subtly changed if this flag is set: the return value
4427 is the number of characters written to the output buffer (zero is
4428 no longer an error code) or a negative error code. Also if the
4429 input buffer is NULL and length 0 finalisation should be performed.
4433 * If a candidate issuer certificate is already part of the constructed
4434 path ignore it: new debug notification X509_V_ERR_PATH_LOOP for this case.
4438 * Improve forward-security support: add functions
4440 void SSL_CTX_set_not_resumable_session_callback(
4441 SSL_CTX *ctx, int (*cb)(SSL *ssl, int is_forward_secure))
4442 void SSL_set_not_resumable_session_callback(
4443 SSL *ssl, int (*cb)(SSL *ssl, int is_forward_secure))
4445 for use by SSL/TLS servers; the callback function will be called whenever a
4446 new session is created, and gets to decide whether the session may be
4447 cached to make it resumable (return 0) or not (return 1). (As by the
4448 SSL/TLS protocol specifications, the session_id sent by the server will be
4449 empty to indicate that the session is not resumable; also, the server will
4450 not generate RFC 4507 (RFC 5077) session tickets.)
4452 A simple reasonable callback implementation is to return is_forward_secure.
4453 This parameter will be set to 1 or 0 depending on the ciphersuite selected
4454 by the SSL/TLS server library, indicating whether it can provide forward
4457 *Emilia Käsper <emilia.kasper@esat.kuleuven.be> (Google)*
4459 * New -verify_name option in command line utilities to set verification
4464 * Initial CMAC implementation. WARNING: EXPERIMENTAL, API MAY CHANGE.
4465 Add CMAC pkey methods.
4469 * Experimental renegotiation in s_server -www mode. If the client
4470 browses /reneg connection is renegotiated. If /renegcert it is
4471 renegotiated requesting a certificate.
4475 * Add an "external" session cache for debugging purposes to s_server. This
4476 should help trace issues which normally are only apparent in deployed
4477 multi-process servers.
4481 * Extensive audit of libcrypto with DEBUG_UNUSED. Fix many cases where
4482 return value is ignored. NB. The functions RAND_add(), RAND_seed(),
4483 BIO_set_cipher() and some obscure PEM functions were changed so they
4484 can now return an error. The RAND changes required a change to the
4485 RAND_METHOD structure.
4489 * New macro `__owur` for "OpenSSL Warn Unused Result". This makes use of
4490 a gcc attribute to warn if the result of a function is ignored. This
4491 is enable if DEBUG_UNUSED is set. Add to several functions in evp.h
4492 whose return value is often ignored.
4496 * New -noct, -requestct, -requirect and -ctlogfile options for s_client.
4497 These allow SCTs (signed certificate timestamps) to be requested and
4498 validated when establishing a connection.
4500 *Rob Percival <robpercival@google.com>*
4505 ### Changes between 1.0.2s and 1.0.2t [10 Sep 2019]
4507 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
4508 used even when parsing explicit parameters, when loading a encoded key
4509 or calling `EC_GROUP_new_from_ecpkparameters()`/
4510 `EC_GROUP_new_from_ecparameters()`.
4511 This prevents bypass of security hardening and performance gains,
4512 especially for curves with specialized EC_METHODs.
4513 By default, if a key encoded with explicit parameters is loaded and later
4514 encoded, the output is still encoded with explicit parameters, even if
4515 internally a "named" EC_GROUP is used for computation.
4519 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
4520 this change, EC_GROUP_set_generator would accept order and/or cofactor as
4521 NULL. After this change, only the cofactor parameter can be NULL. It also
4522 does some minimal sanity checks on the passed order.
4527 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
4528 An attack is simple, if the first CMS_recipientInfo is valid but the
4529 second CMS_recipientInfo is chosen ciphertext. If the second
4530 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
4531 encryption key will be replaced by garbage, and the message cannot be
4532 decoded, but if the RSA decryption fails, the correct encryption key is
4533 used and the recipient will not notice the attack.
4534 As a work around for this potential attack the length of the decrypted
4535 key must be equal to the cipher default key length, in case the
4536 certifiate is not given and all recipientInfo are tried out.
4537 The old behaviour can be re-enabled in the CMS code by setting the
4538 CMS_DEBUG_DECRYPT flag.
4543 * Document issue with installation paths in diverse Windows builds
4545 '/usr/local/ssl' is an unsafe prefix for location to install OpenSSL
4546 binaries and run-time config file.
4551 ### Changes between 1.0.2r and 1.0.2s [28 May 2019]
4553 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
4554 This changes the size when using the `genpkey` command when no size is given.
4555 It fixes an omission in earlier changes that changed all RSA, DSA and DH
4556 generation commands to use 2048 bits by default.
4560 * Add FIPS support for Android Arm 64-bit
4562 Support for Android Arm 64-bit was added to the OpenSSL FIPS Object
4563 Module in Version 2.0.10. For some reason, the corresponding target
4564 'android64-aarch64' was missing OpenSSL 1.0.2, whence it could not be
4565 built with FIPS support on Android Arm 64-bit. This omission has been
4568 *Matthias St. Pierre*
4570 ### Changes between 1.0.2q and 1.0.2r [26 Feb 2019]
4572 * 0-byte record padding oracle
4574 If an application encounters a fatal protocol error and then calls
4575 SSL_shutdown() twice (once to send a close_notify, and once to receive one)
4576 then OpenSSL can respond differently to the calling application if a 0 byte
4577 record is received with invalid padding compared to if a 0 byte record is
4578 received with an invalid MAC. If the application then behaves differently
4579 based on that in a way that is detectable to the remote peer, then this
4580 amounts to a padding oracle that could be used to decrypt data.
4582 In order for this to be exploitable "non-stitched" ciphersuites must be in
4583 use. Stitched ciphersuites are optimised implementations of certain
4584 commonly used ciphersuites. Also the application must call SSL_shutdown()
4585 twice even if a protocol error has occurred (applications should not do
4586 this but some do anyway).
4588 This issue was discovered by Juraj Somorovsky, Robert Merget and Nimrod
4589 Aviram, with additional investigation by Steven Collison and Andrew
4590 Hourselt. It was reported to OpenSSL on 10th December 2018.
4595 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
4599 ### Changes between 1.0.2p and 1.0.2q [20 Nov 2018]
4601 * Microarchitecture timing vulnerability in ECC scalar multiplication
4603 OpenSSL ECC scalar multiplication, used in e.g. ECDSA and ECDH, has been
4604 shown to be vulnerable to a microarchitecture timing side channel attack.
4605 An attacker with sufficient access to mount local timing attacks during
4606 ECDSA signature generation could recover the private key.
4608 This issue was reported to OpenSSL on 26th October 2018 by Alejandro
4609 Cabrera Aldaya, Billy Brumley, Sohaib ul Hassan, Cesar Pereida Garcia and
4615 * Timing vulnerability in DSA signature generation
4617 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
4618 timing side channel attack. An attacker could use variations in the signing
4619 algorithm to recover the private key.
4621 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
4626 * Resolve a compatibility issue in EC_GROUP handling with the FIPS Object
4627 Module, accidentally introduced while backporting security fixes from the
4628 development branch and hindering the use of ECC in FIPS mode.
4632 ### Changes between 1.0.2o and 1.0.2p [14 Aug 2018]
4634 * Client DoS due to large DH parameter
4636 During key agreement in a TLS handshake using a DH(E) based ciphersuite a
4637 malicious server can send a very large prime value to the client. This will
4638 cause the client to spend an unreasonably long period of time generating a
4639 key for this prime resulting in a hang until the client has finished. This
4640 could be exploited in a Denial Of Service attack.
4642 This issue was reported to OpenSSL on 5th June 2018 by Guido Vranken
4647 * Cache timing vulnerability in RSA Key Generation
4649 The OpenSSL RSA Key generation algorithm has been shown to be vulnerable to
4650 a cache timing side channel attack. An attacker with sufficient access to
4651 mount cache timing attacks during the RSA key generation process could
4652 recover the private key.
4654 This issue was reported to OpenSSL on 4th April 2018 by Alejandro Cabrera
4655 Aldaya, Billy Brumley, Cesar Pereida Garcia and Luis Manuel Alvarez Tapia.
4660 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
4661 parameter is no longer accepted, as it leads to a corrupt table. NULL
4662 pem_str is reserved for alias entries only.
4666 * Revert blinding in ECDSA sign and instead make problematic addition
4667 length-invariant. Switch even to fixed-length Montgomery multiplication.
4671 * Change generating and checking of primes so that the error rate of not
4672 being prime depends on the intended use based on the size of the input.
4673 For larger primes this will result in more rounds of Miller-Rabin.
4674 The maximal error rate for primes with more than 1080 bits is lowered
4677 *Kurt Roeckx, Annie Yousar*
4679 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
4683 * Add blinding to ECDSA and DSA signatures to protect against side channel
4684 attacks discovered by Keegan Ryan (NCC Group).
4688 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
4689 now allow empty (zero character) pass phrases.
4693 * Certificate time validation (X509_cmp_time) enforces stricter
4694 compliance with RFC 5280. Fractional seconds and timezone offsets
4695 are no longer allowed.
4699 ### Changes between 1.0.2n and 1.0.2o [27 Mar 2018]
4701 * Constructed ASN.1 types with a recursive definition could exceed the stack
4703 Constructed ASN.1 types with a recursive definition (such as can be found
4704 in PKCS7) could eventually exceed the stack given malicious input with
4705 excessive recursion. This could result in a Denial Of Service attack. There
4706 are no such structures used within SSL/TLS that come from untrusted sources
4707 so this is considered safe.
4709 This issue was reported to OpenSSL on 4th January 2018 by the OSS-fuzz
4715 ### Changes between 1.0.2m and 1.0.2n [7 Dec 2017]
4717 * Read/write after SSL object in error state
4719 OpenSSL 1.0.2 (starting from version 1.0.2b) introduced an "error state"
4720 mechanism. The intent was that if a fatal error occurred during a handshake
4721 then OpenSSL would move into the error state and would immediately fail if
4722 you attempted to continue the handshake. This works as designed for the
4723 explicit handshake functions (SSL_do_handshake(), SSL_accept() and
4724 SSL_connect()), however due to a bug it does not work correctly if
4725 SSL_read() or SSL_write() is called directly. In that scenario, if the
4726 handshake fails then a fatal error will be returned in the initial function
4727 call. If SSL_read()/SSL_write() is subsequently called by the application
4728 for the same SSL object then it will succeed and the data is passed without
4729 being decrypted/encrypted directly from the SSL/TLS record layer.
4731 In order to exploit this issue an application bug would have to be present
4732 that resulted in a call to SSL_read()/SSL_write() being issued after having
4733 already received a fatal error.
4735 This issue was reported to OpenSSL by David Benjamin (Google).
4740 * rsaz_1024_mul_avx2 overflow bug on x86_64
4742 There is an overflow bug in the AVX2 Montgomery multiplication procedure
4743 used in exponentiation with 1024-bit moduli. No EC algorithms are affected.
4744 Analysis suggests that attacks against RSA and DSA as a result of this
4745 defect would be very difficult to perform and are not believed likely.
4746 Attacks against DH1024 are considered just feasible, because most of the
4747 work necessary to deduce information about a private key may be performed
4748 offline. The amount of resources required for such an attack would be
4749 significant. However, for an attack on TLS to be meaningful, the server
4750 would have to share the DH1024 private key among multiple clients, which is
4751 no longer an option since CVE-2016-0701.
4753 This only affects processors that support the AVX2 but not ADX extensions
4754 like Intel Haswell (4th generation).
4756 This issue was reported to OpenSSL by David Benjamin (Google). The issue
4757 was originally found via the OSS-Fuzz project.
4762 ### Changes between 1.0.2l and 1.0.2m [2 Nov 2017]
4764 * bn_sqrx8x_internal carry bug on x86_64
4766 There is a carry propagating bug in the x86_64 Montgomery squaring
4767 procedure. No EC algorithms are affected. Analysis suggests that attacks
4768 against RSA and DSA as a result of this defect would be very difficult to
4769 perform and are not believed likely. Attacks against DH are considered just
4770 feasible (although very difficult) because most of the work necessary to
4771 deduce information about a private key may be performed offline. The amount
4772 of resources required for such an attack would be very significant and
4773 likely only accessible to a limited number of attackers. An attacker would
4774 additionally need online access to an unpatched system using the target
4775 private key in a scenario with persistent DH parameters and a private
4776 key that is shared between multiple clients.
4778 This only affects processors that support the BMI1, BMI2 and ADX extensions
4779 like Intel Broadwell (5th generation) and later or AMD Ryzen.
4781 This issue was reported to OpenSSL by the OSS-Fuzz project.
4786 * Malformed X.509 IPAddressFamily could cause OOB read
4788 If an X.509 certificate has a malformed IPAddressFamily extension,
4789 OpenSSL could do a one-byte buffer overread. The most likely result
4790 would be an erroneous display of the certificate in text format.
4792 This issue was reported to OpenSSL by the OSS-Fuzz project.
4796 ### Changes between 1.0.2k and 1.0.2l [25 May 2017]
4798 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
4799 platform rather than 'mingw'.
4803 ### Changes between 1.0.2j and 1.0.2k [26 Jan 2017]
4805 * Truncated packet could crash via OOB read
4807 If one side of an SSL/TLS path is running on a 32-bit host and a specific
4808 cipher is being used, then a truncated packet can cause that host to
4809 perform an out-of-bounds read, usually resulting in a crash.
4811 This issue was reported to OpenSSL by Robert Święcki of Google.
4816 * BN_mod_exp may produce incorrect results on x86_64
4818 There is a carry propagating bug in the x86_64 Montgomery squaring
4819 procedure. No EC algorithms are affected. Analysis suggests that attacks
4820 against RSA and DSA as a result of this defect would be very difficult to
4821 perform and are not believed likely. Attacks against DH are considered just
4822 feasible (although very difficult) because most of the work necessary to
4823 deduce information about a private key may be performed offline. The amount
4824 of resources required for such an attack would be very significant and
4825 likely only accessible to a limited number of attackers. An attacker would
4826 additionally need online access to an unpatched system using the target
4827 private key in a scenario with persistent DH parameters and a private
4828 key that is shared between multiple clients. For example this can occur by
4829 default in OpenSSL DHE based SSL/TLS ciphersuites. Note: This issue is very
4830 similar to CVE-2015-3193 but must be treated as a separate problem.
4832 This issue was reported to OpenSSL by the OSS-Fuzz project.
4837 * Montgomery multiplication may produce incorrect results
4839 There is a carry propagating bug in the Broadwell-specific Montgomery
4840 multiplication procedure that handles input lengths divisible by, but
4841 longer than 256 bits. Analysis suggests that attacks against RSA, DSA
4842 and DH private keys are impossible. This is because the subroutine in
4843 question is not used in operations with the private key itself and an input
4844 of the attacker's direct choice. Otherwise the bug can manifest itself as
4845 transient authentication and key negotiation failures or reproducible
4846 erroneous outcome of public-key operations with specially crafted input.
4847 Among EC algorithms only Brainpool P-512 curves are affected and one
4848 presumably can attack ECDH key negotiation. Impact was not analyzed in
4849 detail, because pre-requisites for attack are considered unlikely. Namely
4850 multiple clients have to choose the curve in question and the server has to
4851 share the private key among them, neither of which is default behaviour.
4852 Even then only clients that chose the curve will be affected.
4854 This issue was publicly reported as transient failures and was not
4855 initially recognized as a security issue. Thanks to Richard Morgan for
4856 providing reproducible case.
4861 * OpenSSL now fails if it receives an unrecognised record type in TLS1.0
4862 or TLS1.1. Previously this only happened in SSLv3 and TLS1.2. This is to
4863 prevent issues where no progress is being made and the peer continually
4864 sends unrecognised record types, using up resources processing them.
4868 ### Changes between 1.0.2i and 1.0.2j [26 Sep 2016]
4870 * Missing CRL sanity check
4872 A bug fix which included a CRL sanity check was added to OpenSSL 1.1.0
4873 but was omitted from OpenSSL 1.0.2i. As a result any attempt to use
4874 CRLs in OpenSSL 1.0.2i will crash with a null pointer exception.
4876 This issue only affects the OpenSSL 1.0.2i
4881 ### Changes between 1.0.2h and 1.0.2i [22 Sep 2016]
4883 * OCSP Status Request extension unbounded memory growth
4885 A malicious client can send an excessively large OCSP Status Request
4886 extension. If that client continually requests renegotiation, sending a
4887 large OCSP Status Request extension each time, then there will be unbounded
4888 memory growth on the server. This will eventually lead to a Denial Of
4889 Service attack through memory exhaustion. Servers with a default
4890 configuration are vulnerable even if they do not support OCSP. Builds using
4891 the "no-ocsp" build time option are not affected.
4893 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4898 * In order to mitigate the SWEET32 attack, the DES ciphers were moved from
4901 This issue was reported to OpenSSL Karthikeyan Bhargavan and Gaetan
4907 * OOB write in MDC2_Update()
4909 An overflow can occur in MDC2_Update() either if called directly or
4910 through the EVP_DigestUpdate() function using MDC2. If an attacker
4911 is able to supply very large amounts of input data after a previous
4912 call to EVP_EncryptUpdate() with a partial block then a length check
4913 can overflow resulting in a heap corruption.
4915 The amount of data needed is comparable to SIZE_MAX which is impractical
4918 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4923 * Malformed SHA512 ticket DoS
4925 If a server uses SHA512 for TLS session ticket HMAC it is vulnerable to a
4926 DoS attack where a malformed ticket will result in an OOB read which will
4929 The use of SHA512 in TLS session tickets is comparatively rare as it requires
4930 a custom server callback and ticket lookup mechanism.
4932 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4937 * OOB write in BN_bn2dec()
4939 The function BN_bn2dec() does not check the return value of BN_div_word().
4940 This can cause an OOB write if an application uses this function with an
4941 overly large BIGNUM. This could be a problem if an overly large certificate
4942 or CRL is printed out from an untrusted source. TLS is not affected because
4943 record limits will reject an oversized certificate before it is parsed.
4945 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4950 * OOB read in TS_OBJ_print_bio()
4952 The function TS_OBJ_print_bio() misuses OBJ_obj2txt(): the return value is
4953 the total length the OID text representation would use and not the amount
4954 of data written. This will result in OOB reads when large OIDs are
4957 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4962 * Pointer arithmetic undefined behaviour
4964 Avoid some undefined pointer arithmetic
4966 A common idiom in the codebase is to check limits in the following manner:
4969 Where "p" points to some malloc'd data of SIZE bytes and
4972 "len" here could be from some externally supplied data (e.g. from a TLS
4975 The rules of C pointer arithmetic are such that "p + len" is only well
4976 defined where len <= SIZE. Therefore the above idiom is actually
4977 undefined behaviour.
4979 For example this could cause problems if some malloc implementation
4980 provides an address for "p" such that "p + len" actually overflows for
4981 values of len that are too big and therefore p + len < limit.
4983 This issue was reported to OpenSSL by Guido Vranken
4988 * Constant time flag not preserved in DSA signing
4990 Operations in the DSA signing algorithm should run in constant time in
4991 order to avoid side channel attacks. A flaw in the OpenSSL DSA
4992 implementation means that a non-constant time codepath is followed for
4993 certain operations. This has been demonstrated through a cache-timing
4994 attack to be sufficient for an attacker to recover the private DSA key.
4996 This issue was reported by César Pereida (Aalto University), Billy Brumley
4997 (Tampere University of Technology), and Yuval Yarom (The University of
4998 Adelaide and NICTA).
5003 * DTLS buffered message DoS
5005 In a DTLS connection where handshake messages are delivered out-of-order
5006 those messages that OpenSSL is not yet ready to process will be buffered
5007 for later use. Under certain circumstances, a flaw in the logic means that
5008 those messages do not get removed from the buffer even though the handshake
5009 has been completed. An attacker could force up to approx. 15 messages to
5010 remain in the buffer when they are no longer required. These messages will
5011 be cleared when the DTLS connection is closed. The default maximum size for
5012 a message is 100k. Therefore the attacker could force an additional 1500k
5013 to be consumed per connection. By opening many simulataneous connections an
5014 attacker could cause a DoS attack through memory exhaustion.
5016 This issue was reported to OpenSSL by Quan Luo.
5021 * DTLS replay protection DoS
5023 A flaw in the DTLS replay attack protection mechanism means that records
5024 that arrive for future epochs update the replay protection "window" before
5025 the MAC for the record has been validated. This could be exploited by an
5026 attacker by sending a record for the next epoch (which does not have to
5027 decrypt or have a valid MAC), with a very large sequence number. This means
5028 that all subsequent legitimate packets are dropped causing a denial of
5029 service for a specific DTLS connection.
5031 This issue was reported to OpenSSL by the OCAP audit team.
5036 * Certificate message OOB reads
5038 In OpenSSL 1.0.2 and earlier some missing message length checks can result
5039 in OOB reads of up to 2 bytes beyond an allocated buffer. There is a
5040 theoretical DoS risk but this has not been observed in practice on common
5043 The messages affected are client certificate, client certificate request
5044 and server certificate. As a result the attack can only be performed
5045 against a client or a server which enables client authentication.
5047 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5052 ### Changes between 1.0.2g and 1.0.2h [3 May 2016]
5054 * Prevent padding oracle in AES-NI CBC MAC check
5056 A MITM attacker can use a padding oracle attack to decrypt traffic
5057 when the connection uses an AES CBC cipher and the server support
5060 This issue was introduced as part of the fix for Lucky 13 padding
5061 attack ([CVE-2013-0169]). The padding check was rewritten to be in
5062 constant time by making sure that always the same bytes are read and
5063 compared against either the MAC or padding bytes. But it no longer
5064 checked that there was enough data to have both the MAC and padding
5067 This issue was reported by Juraj Somorovsky using TLS-Attacker.
5071 * Fix EVP_EncodeUpdate overflow
5073 An overflow can occur in the EVP_EncodeUpdate() function which is used for
5074 Base64 encoding of binary data. If an attacker is able to supply very large
5075 amounts of input data then a length check can overflow resulting in a heap
5078 Internally to OpenSSL the EVP_EncodeUpdate() function is primarily used by
5079 the `PEM_write_bio*` family of functions. These are mainly used within the
5080 OpenSSL command line applications, so any application which processes data
5081 from an untrusted source and outputs it as a PEM file should be considered
5082 vulnerable to this issue. User applications that call these APIs directly
5083 with large amounts of untrusted data may also be vulnerable.
5085 This issue was reported by Guido Vranken.
5090 * Fix EVP_EncryptUpdate overflow
5092 An overflow can occur in the EVP_EncryptUpdate() function. If an attacker
5093 is able to supply very large amounts of input data after a previous call to
5094 EVP_EncryptUpdate() with a partial block then a length check can overflow
5095 resulting in a heap corruption. Following an analysis of all OpenSSL
5096 internal usage of the EVP_EncryptUpdate() function all usage is one of two
5097 forms. The first form is where the EVP_EncryptUpdate() call is known to be
5098 the first called function after an EVP_EncryptInit(), and therefore that
5099 specific call must be safe. The second form is where the length passed to
5100 EVP_EncryptUpdate() can be seen from the code to be some small value and
5101 therefore there is no possibility of an overflow. Since all instances are
5102 one of these two forms, it is believed that there can be no overflows in
5103 internal code due to this problem. It should be noted that
5104 EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths.
5105 Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances
5106 of these calls have also been analysed too and it is believed there are no
5107 instances in internal usage where an overflow could occur.
5109 This issue was reported by Guido Vranken.
5114 * Prevent ASN.1 BIO excessive memory allocation
5116 When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio()
5117 a short invalid encoding can cause allocation of large amounts of memory
5118 potentially consuming excessive resources or exhausting memory.
5120 Any application parsing untrusted data through d2i BIO functions is
5121 affected. The memory based functions such as d2i_X509() are *not* affected.
5122 Since the memory based functions are used by the TLS library, TLS
5123 applications are not affected.
5125 This issue was reported by Brian Carpenter.
5132 ASN1 Strings that are over 1024 bytes can cause an overread in applications
5133 using the X509_NAME_oneline() function on EBCDIC systems. This could result
5134 in arbitrary stack data being returned in the buffer.
5136 This issue was reported by Guido Vranken.
5141 * Modify behavior of ALPN to invoke callback after SNI/servername
5142 callback, such that updates to the SSL_CTX affect ALPN.
5146 * Remove LOW from the DEFAULT cipher list. This removes singles DES from the
5151 * Only remove the SSLv2 methods with the no-ssl2-method option. When the
5152 methods are enabled and ssl2 is disabled the methods return NULL.
5156 ### Changes between 1.0.2f and 1.0.2g [1 Mar 2016]
5158 * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL.
5159 Builds that are not configured with "enable-weak-ssl-ciphers" will not
5160 provide any "EXPORT" or "LOW" strength ciphers.
5164 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2
5165 is by default disabled at build-time. Builds that are not configured with
5166 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used,
5167 users who want to negotiate SSLv2 via the version-flexible SSLv23_method()
5168 will need to explicitly call either of:
5170 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2);
5172 SSL_clear_options(ssl, SSL_OP_NO_SSLv2);
5174 as appropriate. Even if either of those is used, or the application
5175 explicitly uses the version-specific SSLv2_method() or its client and
5176 server variants, SSLv2 ciphers vulnerable to exhaustive search key
5177 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT
5178 ciphers, and SSLv2 56-bit DES are no longer available.
5183 * Fix a double-free in DSA code
5185 A double free bug was discovered when OpenSSL parses malformed DSA private
5186 keys and could lead to a DoS attack or memory corruption for applications
5187 that receive DSA private keys from untrusted sources. This scenario is
5190 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using
5196 * Disable SRP fake user seed to address a server memory leak.
5198 Add a new method SRP_VBASE_get1_by_user that handles the seed properly.
5200 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
5201 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user
5202 was changed to ignore the "fake user" SRP seed, even if the seed
5205 Users should use SRP_VBASE_get1_by_user instead. Note that in
5206 SRP_VBASE_get1_by_user, caller must free the returned value. Note
5207 also that even though configuring the SRP seed attempts to hide
5208 invalid usernames by continuing the handshake with fake
5209 credentials, this behaviour is not constant time and no strong
5210 guarantees are made that the handshake is indistinguishable from
5211 that of a valid user.
5216 * Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption
5218 In the BN_hex2bn function the number of hex digits is calculated using an
5219 int value `i`. Later `bn_expand` is called with a value of `i * 4`. For
5220 large values of `i` this can result in `bn_expand` not allocating any
5221 memory because `i * 4` is negative. This can leave the internal BIGNUM data
5222 field as NULL leading to a subsequent NULL ptr deref. For very large values
5223 of `i`, the calculation `i * 4` could be a positive value smaller than `i`.
5224 In this case memory is allocated to the internal BIGNUM data field, but it
5225 is insufficiently sized leading to heap corruption. A similar issue exists
5226 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn
5227 is ever called by user applications with very large untrusted hex/dec data.
5228 This is anticipated to be a rare occurrence.
5230 All OpenSSL internal usage of these functions use data that is not expected
5231 to be untrusted, e.g. config file data or application command line
5232 arguments. If user developed applications generate config file data based
5233 on untrusted data then it is possible that this could also lead to security
5234 consequences. This is also anticipated to be rare.
5236 This issue was reported to OpenSSL by Guido Vranken.
5241 * Fix memory issues in `BIO_*printf` functions
5243 The internal `fmtstr` function used in processing a "%s" format string in
5244 the `BIO_*printf` functions could overflow while calculating the length of a
5245 string and cause an OOB read when printing very long strings.
5247 Additionally the internal `doapr_outch` function can attempt to write to an
5248 OOB memory location (at an offset from the NULL pointer) in the event of a
5249 memory allocation failure. In 1.0.2 and below this could be caused where
5250 the size of a buffer to be allocated is greater than INT_MAX. E.g. this
5251 could be in processing a very long "%s" format string. Memory leaks can
5254 The first issue may mask the second issue dependent on compiler behaviour.
5255 These problems could enable attacks where large amounts of untrusted data
5256 is passed to the `BIO_*printf` functions. If applications use these functions
5257 in this way then they could be vulnerable. OpenSSL itself uses these
5258 functions when printing out human-readable dumps of ASN.1 data. Therefore
5259 applications that print this data could be vulnerable if the data is from
5260 untrusted sources. OpenSSL command line applications could also be
5261 vulnerable where they print out ASN.1 data, or if untrusted data is passed
5262 as command line arguments.
5264 Libssl is not considered directly vulnerable. Additionally certificates etc
5265 received via remote connections via libssl are also unlikely to be able to
5266 trigger these issues because of message size limits enforced within libssl.
5268 This issue was reported to OpenSSL Guido Vranken.
5273 * Side channel attack on modular exponentiation
5275 A side-channel attack was found which makes use of cache-bank conflicts on
5276 the Intel Sandy-Bridge microarchitecture which could lead to the recovery
5277 of RSA keys. The ability to exploit this issue is limited as it relies on
5278 an attacker who has control of code in a thread running on the same
5279 hyper-threaded core as the victim thread which is performing decryptions.
5281 This issue was reported to OpenSSL by Yuval Yarom, The University of
5282 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and
5283 Nadia Heninger, University of Pennsylvania with more information at
5284 <http://cachebleed.info>.
5289 * Change the `req` command to generate a 2048-bit RSA/DSA key by default,
5290 if no keysize is specified with default_bits. This fixes an
5291 omission in an earlier change that changed all RSA/DSA key generation
5292 commands to use 2048 bits by default.
5296 ### Changes between 1.0.2e and 1.0.2f [28 Jan 2016]
5298 * DH small subgroups
5300 Historically OpenSSL only ever generated DH parameters based on "safe"
5301 primes. More recently (in version 1.0.2) support was provided for
5302 generating X9.42 style parameter files such as those required for RFC 5114
5303 support. The primes used in such files may not be "safe". Where an
5304 application is using DH configured with parameters based on primes that are
5305 not "safe" then an attacker could use this fact to find a peer's private
5306 DH exponent. This attack requires that the attacker complete multiple
5307 handshakes in which the peer uses the same private DH exponent. For example
5308 this could be used to discover a TLS server's private DH exponent if it's
5309 reusing the private DH exponent or it's using a static DH ciphersuite.
5311 OpenSSL provides the option SSL_OP_SINGLE_DH_USE for ephemeral DH (DHE) in
5312 TLS. It is not on by default. If the option is not set then the server
5313 reuses the same private DH exponent for the life of the server process and
5314 would be vulnerable to this attack. It is believed that many popular
5315 applications do set this option and would therefore not be at risk.
5317 The fix for this issue adds an additional check where a "q" parameter is
5318 available (as is the case in X9.42 based parameters). This detects the
5319 only known attack, and is the only possible defense for static DH
5320 ciphersuites. This could have some performance impact.
5322 Additionally the SSL_OP_SINGLE_DH_USE option has been switched on by
5323 default and cannot be disabled. This could have some performance impact.
5325 This issue was reported to OpenSSL by Antonio Sanso (Adobe).
5330 * SSLv2 doesn't block disabled ciphers
5332 A malicious client can negotiate SSLv2 ciphers that have been disabled on
5333 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
5334 been disabled, provided that the SSLv2 protocol was not also disabled via
5337 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
5338 and Sebastian Schinzel.
5343 ### Changes between 1.0.2d and 1.0.2e [3 Dec 2015]
5345 * BN_mod_exp may produce incorrect results on x86_64
5347 There is a carry propagating bug in the x86_64 Montgomery squaring
5348 procedure. No EC algorithms are affected. Analysis suggests that attacks
5349 against RSA and DSA as a result of this defect would be very difficult to
5350 perform and are not believed likely. Attacks against DH are considered just
5351 feasible (although very difficult) because most of the work necessary to
5352 deduce information about a private key may be performed offline. The amount
5353 of resources required for such an attack would be very significant and
5354 likely only accessible to a limited number of attackers. An attacker would
5355 additionally need online access to an unpatched system using the target
5356 private key in a scenario with persistent DH parameters and a private
5357 key that is shared between multiple clients. For example this can occur by
5358 default in OpenSSL DHE based SSL/TLS ciphersuites.
5360 This issue was reported to OpenSSL by Hanno Böck.
5365 * Certificate verify crash with missing PSS parameter
5367 The signature verification routines will crash with a NULL pointer
5368 dereference if presented with an ASN.1 signature using the RSA PSS
5369 algorithm and absent mask generation function parameter. Since these
5370 routines are used to verify certificate signature algorithms this can be
5371 used to crash any certificate verification operation and exploited in a
5372 DoS attack. Any application which performs certificate verification is
5373 vulnerable including OpenSSL clients and servers which enable client
5376 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
5381 * X509_ATTRIBUTE memory leak
5383 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
5384 memory. This structure is used by the PKCS#7 and CMS routines so any
5385 application which reads PKCS#7 or CMS data from untrusted sources is
5386 affected. SSL/TLS is not affected.
5388 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
5394 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
5395 This changes the decoding behaviour for some invalid messages,
5396 though the change is mostly in the more lenient direction, and
5397 legacy behaviour is preserved as much as possible.
5401 * In DSA_generate_parameters_ex, if the provided seed is too short,
5404 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
5406 ### Changes between 1.0.2c and 1.0.2d [9 Jul 2015]
5408 * Alternate chains certificate forgery
5410 During certificate verification, OpenSSL will attempt to find an
5411 alternative certificate chain if the first attempt to build such a chain
5412 fails. An error in the implementation of this logic can mean that an
5413 attacker could cause certain checks on untrusted certificates to be
5414 bypassed, such as the CA flag, enabling them to use a valid leaf
5415 certificate to act as a CA and "issue" an invalid certificate.
5417 This issue was reported to OpenSSL by Adam Langley/David Benjamin
5422 ### Changes between 1.0.2b and 1.0.2c [12 Jun 2015]
5424 * Fix HMAC ABI incompatibility. The previous version introduced an ABI
5425 incompatibility in the handling of HMAC. The previous ABI has now been
5430 ### Changes between 1.0.2a and 1.0.2b [11 Jun 2015]
5432 * Malformed ECParameters causes infinite loop
5434 When processing an ECParameters structure OpenSSL enters an infinite loop
5435 if the curve specified is over a specially malformed binary polynomial
5438 This can be used to perform denial of service against any
5439 system which processes public keys, certificate requests or
5440 certificates. This includes TLS clients and TLS servers with
5441 client authentication enabled.
5443 This issue was reported to OpenSSL by Joseph Barr-Pixton.
5448 * Exploitable out-of-bounds read in X509_cmp_time
5450 X509_cmp_time does not properly check the length of the ASN1_TIME
5451 string and can read a few bytes out of bounds. In addition,
5452 X509_cmp_time accepts an arbitrary number of fractional seconds in the
5455 An attacker can use this to craft malformed certificates and CRLs of
5456 various sizes and potentially cause a segmentation fault, resulting in
5457 a DoS on applications that verify certificates or CRLs. TLS clients
5458 that verify CRLs are affected. TLS clients and servers with client
5459 authentication enabled may be affected if they use custom verification
5462 This issue was reported to OpenSSL by Robert Swiecki (Google), and
5463 independently by Hanno Böck.
5468 * PKCS7 crash with missing EnvelopedContent
5470 The PKCS#7 parsing code does not handle missing inner EncryptedContent
5471 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
5472 with missing content and trigger a NULL pointer dereference on parsing.
5474 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
5475 structures from untrusted sources are affected. OpenSSL clients and
5476 servers are not affected.
5478 This issue was reported to OpenSSL by Michal Zalewski (Google).
5483 * CMS verify infinite loop with unknown hash function
5485 When verifying a signedData message the CMS code can enter an infinite loop
5486 if presented with an unknown hash function OID. This can be used to perform
5487 denial of service against any system which verifies signedData messages using
5489 This issue was reported to OpenSSL by Johannes Bauer.
5494 * Race condition handling NewSessionTicket
5496 If a NewSessionTicket is received by a multi-threaded client when attempting to
5497 reuse a previous ticket then a race condition can occur potentially leading to
5498 a double free of the ticket data.
5503 * Only support 256-bit or stronger elliptic curves with the
5504 'ecdh_auto' setting (server) or by default (client). Of supported
5505 curves, prefer P-256 (both).
5509 ### Changes between 1.0.2 and 1.0.2a [19 Mar 2015]
5511 * ClientHello sigalgs DoS fix
5513 If a client connects to an OpenSSL 1.0.2 server and renegotiates with an
5514 invalid signature algorithms extension a NULL pointer dereference will
5515 occur. This can be exploited in a DoS attack against the server.
5517 This issue was was reported to OpenSSL by David Ramos of Stanford
5521 *Stephen Henson and Matt Caswell*
5523 * Multiblock corrupted pointer fix
5525 OpenSSL 1.0.2 introduced the "multiblock" performance improvement. This
5526 feature only applies on 64 bit x86 architecture platforms that support AES
5527 NI instructions. A defect in the implementation of "multiblock" can cause
5528 OpenSSL's internal write buffer to become incorrectly set to NULL when
5529 using non-blocking IO. Typically, when the user application is using a
5530 socket BIO for writing, this will only result in a failed connection.
5531 However if some other BIO is used then it is likely that a segmentation
5532 fault will be triggered, thus enabling a potential DoS attack.
5534 This issue was reported to OpenSSL by Daniel Danner and Rainer Mueller.
5539 * Segmentation fault in DTLSv1_listen fix
5541 The DTLSv1_listen function is intended to be stateless and processes the
5542 initial ClientHello from many peers. It is common for user code to loop
5543 over the call to DTLSv1_listen until a valid ClientHello is received with
5544 an associated cookie. A defect in the implementation of DTLSv1_listen means
5545 that state is preserved in the SSL object from one invocation to the next
5546 that can lead to a segmentation fault. Errors processing the initial
5547 ClientHello can trigger this scenario. An example of such an error could be
5548 that a DTLS1.0 only client is attempting to connect to a DTLS1.2 only
5551 This issue was reported to OpenSSL by Per Allansson.
5556 * Segmentation fault in ASN1_TYPE_cmp fix
5558 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
5559 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
5560 certificate signature algorithm consistency this can be used to crash any
5561 certificate verification operation and exploited in a DoS attack. Any
5562 application which performs certificate verification is vulnerable including
5563 OpenSSL clients and servers which enable client authentication.
5568 * Segmentation fault for invalid PSS parameters fix
5570 The signature verification routines will crash with a NULL pointer
5571 dereference if presented with an ASN.1 signature using the RSA PSS
5572 algorithm and invalid parameters. Since these routines are used to verify
5573 certificate signature algorithms this can be used to crash any
5574 certificate verification operation and exploited in a DoS attack. Any
5575 application which performs certificate verification is vulnerable including
5576 OpenSSL clients and servers which enable client authentication.
5578 This issue was was reported to OpenSSL by Brian Carpenter.
5583 * ASN.1 structure reuse memory corruption fix
5585 Reusing a structure in ASN.1 parsing may allow an attacker to cause
5586 memory corruption via an invalid write. Such reuse is and has been
5587 strongly discouraged and is believed to be rare.
5589 Applications that parse structures containing CHOICE or ANY DEFINED BY
5590 components may be affected. Certificate parsing (d2i_X509 and related
5591 functions) are however not affected. OpenSSL clients and servers are
5597 * PKCS7 NULL pointer dereferences fix
5599 The PKCS#7 parsing code does not handle missing outer ContentInfo
5600 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
5601 missing content and trigger a NULL pointer dereference on parsing.
5603 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
5604 otherwise parse PKCS#7 structures from untrusted sources are
5605 affected. OpenSSL clients and servers are not affected.
5607 This issue was reported to OpenSSL by Michal Zalewski (Google).
5612 * DoS via reachable assert in SSLv2 servers fix
5614 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
5615 servers that both support SSLv2 and enable export cipher suites by sending
5616 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
5618 This issue was discovered by Sean Burford (Google) and Emilia Käsper
5619 (OpenSSL development team).
5624 * Empty CKE with client auth and DHE fix
5626 If client auth is used then a server can seg fault in the event of a DHE
5627 ciphersuite being selected and a zero length ClientKeyExchange message
5628 being sent by the client. This could be exploited in a DoS attack.
5633 * Handshake with unseeded PRNG fix
5635 Under certain conditions an OpenSSL 1.0.2 client can complete a handshake
5636 with an unseeded PRNG. The conditions are:
5637 - The client is on a platform where the PRNG has not been seeded
5638 automatically, and the user has not seeded manually
5639 - A protocol specific client method version has been used (i.e. not
5640 SSL_client_methodv23)
5641 - A ciphersuite is used that does not require additional random data from
5642 the PRNG beyond the initial ClientHello client random (e.g. PSK-RC4-SHA).
5644 If the handshake succeeds then the client random that has been used will
5645 have been generated from a PRNG with insufficient entropy and therefore the
5646 output may be predictable.
5648 For example using the following command with an unseeded openssl will
5649 succeed on an unpatched platform:
5651 openssl s_client -psk 1a2b3c4d -tls1_2 -cipher PSK-RC4-SHA
5656 * Use After Free following d2i_ECPrivatekey error fix
5658 A malformed EC private key file consumed via the d2i_ECPrivateKey function
5659 could cause a use after free condition. This, in turn, could cause a double
5660 free in several private key parsing functions (such as d2i_PrivateKey
5661 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
5662 for applications that receive EC private keys from untrusted
5663 sources. This scenario is considered rare.
5665 This issue was discovered by the BoringSSL project and fixed in their
5671 * X509_to_X509_REQ NULL pointer deref fix
5673 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
5674 the certificate key is invalid. This function is rarely used in practice.
5676 This issue was discovered by Brian Carpenter.
5681 * Removed the export ciphers from the DEFAULT ciphers
5685 ### Changes between 1.0.1l and 1.0.2 [22 Jan 2015]
5687 * Facilitate "universal" ARM builds targeting range of ARM ISAs, e.g.
5688 ARMv5 through ARMv8, as opposite to "locking" it to single one.
5689 So far those who have to target multiple platforms would compromise
5690 and argue that binary targeting say ARMv5 would still execute on
5691 ARMv8. "Universal" build resolves this compromise by providing
5692 near-optimal performance even on newer platforms.
5696 * Accelerated NIST P-256 elliptic curve implementation for x86_64
5697 (other platforms pending).
5699 *Shay Gueron & Vlad Krasnov (Intel Corp), Andy Polyakov*
5701 * Add support for the SignedCertificateTimestampList certificate and
5702 OCSP response extensions from RFC6962.
5706 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
5707 for corner cases. (Certain input points at infinity could lead to
5708 bogus results, with non-infinity inputs mapped to infinity too.)
5712 * Initial support for PowerISA 2.0.7, first implemented in POWER8.
5713 This covers AES, SHA256/512 and GHASH. "Initial" means that most
5714 common cases are optimized and there still is room for further
5715 improvements. Vector Permutation AES for Altivec is also added.
5719 * Add support for little-endian ppc64 Linux target.
5721 *Marcelo Cerri (IBM)*
5723 * Initial support for AMRv8 ISA crypto extensions. This covers AES,
5724 SHA1, SHA256 and GHASH. "Initial" means that most common cases
5725 are optimized and there still is room for further improvements.
5726 Both 32- and 64-bit modes are supported.
5728 *Andy Polyakov, Ard Biesheuvel (Linaro)*
5730 * Improved ARMv7 NEON support.
5734 * Support for SPARC Architecture 2011 crypto extensions, first
5735 implemented in SPARC T4. This covers AES, DES, Camellia, SHA1,
5736 SHA256/512, MD5, GHASH and modular exponentiation.
5738 *Andy Polyakov, David Miller*
5740 * Accelerated modular exponentiation for Intel processors, a.k.a.
5743 *Shay Gueron & Vlad Krasnov (Intel Corp)*
5745 * Support for new and upcoming Intel processors, including AVX2,
5746 BMI and SHA ISA extensions. This includes additional "stitched"
5747 implementations, AESNI-SHA256 and GCM, and multi-buffer support
5750 This work was sponsored by Intel Corp.
5754 * Support for DTLS 1.2. This adds two sets of DTLS methods: DTLS_*_method()
5755 supports both DTLS 1.2 and 1.0 and should use whatever version the peer
5756 supports and DTLSv1_2_*_method() which supports DTLS 1.2 only.
5760 * Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
5761 this fixes a limitation in previous versions of OpenSSL.
5765 * Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
5766 MGF1 digest and OAEP label.
5770 * Add EVP support for key wrapping algorithms, to avoid problems with
5771 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
5772 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
5773 algorithms and include tests cases.
5777 * Add functions to allocate and set the fields of an ECDSA_METHOD
5780 *Douglas E. Engert, Steve Henson*
5782 * New functions OPENSSL_gmtime_diff and ASN1_TIME_diff to find the
5783 difference in days and seconds between two tm or ASN1_TIME structures.
5787 * Add -rev test option to s_server to just reverse order of characters
5788 received by client and send back to server. Also prints an abbreviated
5789 summary of the connection parameters.
5793 * New option -brief for s_client and s_server to print out a brief summary
5794 of connection parameters.
5798 * Add callbacks for arbitrary TLS extensions.
5800 *Trevor Perrin <trevp@trevp.net> and Ben Laurie*
5802 * New option -crl_download in several openssl utilities to download CRLs
5803 from CRLDP extension in certificates.
5807 * New options -CRL and -CRLform for s_client and s_server for CRLs.
5811 * New function X509_CRL_diff to generate a delta CRL from the difference
5812 of two full CRLs. Add support to "crl" utility.
5816 * New functions to set lookup_crls function and to retrieve
5817 X509_STORE from X509_STORE_CTX.
5821 * Print out deprecated issuer and subject unique ID fields in
5826 * Extend OCSP I/O functions so they can be used for simple general purpose
5827 HTTP as well as OCSP. New wrapper function which can be used to download
5828 CRLs using the OCSP API.
5832 * Delegate command line handling in s_client/s_server to SSL_CONF APIs.
5836 * `SSL_CONF*` functions. These provide a common framework for application
5837 configuration using configuration files or command lines.
5841 * SSL/TLS tracing code. This parses out SSL/TLS records using the
5842 message callback and prints the results. Needs compile time option
5843 "enable-ssl-trace". New options to s_client and s_server to enable
5848 * New ctrl and macro to retrieve supported points extensions.
5849 Print out extension in s_server and s_client.
5853 * New functions to retrieve certificate signature and signature
5858 * Add functions to retrieve and manipulate the raw cipherlist sent by a
5863 * New Suite B modes for TLS code. These use and enforce the requirements
5864 of RFC6460: restrict ciphersuites, only permit Suite B algorithms and
5865 only use Suite B curves. The Suite B modes can be set by using the
5866 strings "SUITEB128", "SUITEB192" or "SUITEB128ONLY" for the cipherstring.
5870 * New chain verification flags for Suite B levels of security. Check
5871 algorithms are acceptable when flags are set in X509_verify_cert.
5875 * Make tls1_check_chain return a set of flags indicating checks passed
5876 by a certificate chain. Add additional tests to handle client
5877 certificates: checks for matching certificate type and issuer name
5882 * If an attempt is made to use a signature algorithm not in the peer
5883 preference list abort the handshake. If client has no suitable
5884 signature algorithms in response to a certificate request do not
5885 use the certificate.
5889 * If server EC tmp key is not in client preference list abort handshake.
5893 * Add support for certificate stores in CERT structure. This makes it
5894 possible to have different stores per SSL structure or one store in
5895 the parent SSL_CTX. Include distinct stores for certificate chain
5896 verification and chain building. New ctrl SSL_CTRL_BUILD_CERT_CHAIN
5897 to build and store a certificate chain in CERT structure: returning
5898 an error if the chain cannot be built: this will allow applications
5899 to test if a chain is correctly configured.
5901 Note: if the CERT based stores are not set then the parent SSL_CTX
5902 store is used to retain compatibility with existing behaviour.
5906 * New function ssl_set_client_disabled to set a ciphersuite disabled
5907 mask based on the current session, check mask when sending client
5908 hello and checking the requested ciphersuite.
5912 * New ctrls to retrieve and set certificate types in a certificate
5913 request message. Print out received values in s_client. If certificate
5914 types is not set with custom values set sensible values based on
5915 supported signature algorithms.
5919 * Support for distinct client and server supported signature algorithms.
5923 * Add certificate callback. If set this is called whenever a certificate
5924 is required by client or server. An application can decide which
5925 certificate chain to present based on arbitrary criteria: for example
5926 supported signature algorithms. Add very simple example to s_server.
5927 This fixes many of the problems and restrictions of the existing client
5928 certificate callback: for example you can now clear an existing
5929 certificate and specify the whole chain.
5933 * Add new "valid_flags" field to CERT_PKEY structure which determines what
5934 the certificate can be used for (if anything). Set valid_flags field
5935 in new tls1_check_chain function. Simplify ssl_set_cert_masks which used
5936 to have similar checks in it.
5938 Add new "cert_flags" field to CERT structure and include a "strict mode".
5939 This enforces some TLS certificate requirements (such as only permitting
5940 certificate signature algorithms contained in the supported algorithms
5941 extension) which some implementations ignore: this option should be used
5942 with caution as it could cause interoperability issues.
5946 * Update and tidy signature algorithm extension processing. Work out
5947 shared signature algorithms based on preferences and peer algorithms
5948 and print them out in s_client and s_server. Abort handshake if no
5949 shared signature algorithms.
5953 * Add new functions to allow customised supported signature algorithms
5954 for SSL and SSL_CTX structures. Add options to s_client and s_server
5959 * New function SSL_certs_clear() to delete all references to certificates
5960 from an SSL structure. Before this once a certificate had been added
5961 it couldn't be removed.
5965 * Integrate hostname, email address and IP address checking with certificate
5966 verification. New verify options supporting checking in openssl utility.
5970 * Fixes and wildcard matching support to hostname and email checking
5971 functions. Add manual page.
5973 *Florian Weimer (Red Hat Product Security Team)*
5975 * New functions to check a hostname email or IP address against a
5976 certificate. Add options x509 utility to print results of checks against
5981 * Fix OCSP checking.
5983 *Rob Stradling <rob.stradling@comodo.com> and Ben Laurie*
5985 * Initial experimental support for explicitly trusted non-root CAs.
5986 OpenSSL still tries to build a complete chain to a root but if an
5987 intermediate CA has a trust setting included that is used. The first
5988 setting is used: whether to trust (e.g., -addtrust option to the x509
5993 * Add -trusted_first option which attempts to find certificates in the
5994 trusted store even if an untrusted chain is also supplied.
5998 * MIPS assembly pack updates: support for MIPS32r2 and SmartMIPS ASE,
5999 platform support for Linux and Android.
6003 * Support for linux-x32, ILP32 environment in x86_64 framework.
6007 * Experimental multi-implementation support for FIPS capable OpenSSL.
6008 When in FIPS mode the approved implementations are used as normal,
6009 when not in FIPS mode the internal unapproved versions are used instead.
6010 This means that the FIPS capable OpenSSL isn't forced to use the
6011 (often lower performance) FIPS implementations outside FIPS mode.
6015 * Transparently support X9.42 DH parameters when calling
6016 PEM_read_bio_DHparameters. This means existing applications can handle
6017 the new parameter format automatically.
6021 * Initial experimental support for X9.42 DH parameter format: mainly
6022 to support use of 'q' parameter for RFC5114 parameters.
6026 * Add DH parameters from RFC5114 including test data to dhtest.
6030 * Support for automatic EC temporary key parameter selection. If enabled
6031 the most preferred EC parameters are automatically used instead of
6032 hardcoded fixed parameters. Now a server just has to call:
6033 SSL_CTX_set_ecdh_auto(ctx, 1) and the server will automatically
6034 support ECDH and use the most appropriate parameters.
6038 * Enhance and tidy EC curve and point format TLS extension code. Use
6039 static structures instead of allocation if default values are used.
6040 New ctrls to set curves we wish to support and to retrieve shared curves.
6041 Print out shared curves in s_server. New options to s_server and s_client
6042 to set list of supported curves.
6046 * New ctrls to retrieve supported signature algorithms and
6047 supported curve values as an array of NIDs. Extend openssl utility
6048 to print out received values.
6052 * Add new APIs EC_curve_nist2nid and EC_curve_nid2nist which convert
6053 between NIDs and the more common NIST names such as "P-256". Enhance
6054 ecparam utility and ECC method to recognise the NIST names for curves.
6058 * Enhance SSL/TLS certificate chain handling to support different
6059 chains for each certificate instead of one chain in the parent SSL_CTX.
6063 * Support for fixed DH ciphersuite client authentication: where both
6064 server and client use DH certificates with common parameters.
6068 * Support for fixed DH ciphersuites: those requiring DH server
6073 * New function i2d_re_X509_tbs for re-encoding the TBS portion of
6075 Note: Related 1.0.2-beta specific macros X509_get_cert_info,
6076 X509_CINF_set_modified, X509_CINF_get_issuer, X509_CINF_get_extensions and
6077 X509_CINF_get_signature were reverted post internal team review.
6082 ### Changes between 1.0.1t and 1.0.1u [22 Sep 2016]
6084 * OCSP Status Request extension unbounded memory growth
6086 A malicious client can send an excessively large OCSP Status Request
6087 extension. If that client continually requests renegotiation, sending a
6088 large OCSP Status Request extension each time, then there will be unbounded
6089 memory growth on the server. This will eventually lead to a Denial Of
6090 Service attack through memory exhaustion. Servers with a default
6091 configuration are vulnerable even if they do not support OCSP. Builds using
6092 the "no-ocsp" build time option are not affected.
6094 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6099 * In order to mitigate the SWEET32 attack, the DES ciphers were moved from
6102 This issue was reported to OpenSSL Karthikeyan Bhargavan and Gaetan
6108 * OOB write in MDC2_Update()
6110 An overflow can occur in MDC2_Update() either if called directly or
6111 through the EVP_DigestUpdate() function using MDC2. If an attacker
6112 is able to supply very large amounts of input data after a previous
6113 call to EVP_EncryptUpdate() with a partial block then a length check
6114 can overflow resulting in a heap corruption.
6116 The amount of data needed is comparable to SIZE_MAX which is impractical
6119 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6124 * Malformed SHA512 ticket DoS
6126 If a server uses SHA512 for TLS session ticket HMAC it is vulnerable to a
6127 DoS attack where a malformed ticket will result in an OOB read which will
6130 The use of SHA512 in TLS session tickets is comparatively rare as it requires
6131 a custom server callback and ticket lookup mechanism.
6133 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6138 * OOB write in BN_bn2dec()
6140 The function BN_bn2dec() does not check the return value of BN_div_word().
6141 This can cause an OOB write if an application uses this function with an
6142 overly large BIGNUM. This could be a problem if an overly large certificate
6143 or CRL is printed out from an untrusted source. TLS is not affected because
6144 record limits will reject an oversized certificate before it is parsed.
6146 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6151 * OOB read in TS_OBJ_print_bio()
6153 The function TS_OBJ_print_bio() misuses OBJ_obj2txt(): the return value is
6154 the total length the OID text representation would use and not the amount
6155 of data written. This will result in OOB reads when large OIDs are
6158 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6163 * Pointer arithmetic undefined behaviour
6165 Avoid some undefined pointer arithmetic
6167 A common idiom in the codebase is to check limits in the following manner:
6170 Where "p" points to some malloc'd data of SIZE bytes and
6173 "len" here could be from some externally supplied data (e.g. from a TLS
6176 The rules of C pointer arithmetic are such that "p + len" is only well
6177 defined where len <= SIZE. Therefore the above idiom is actually
6178 undefined behaviour.
6180 For example this could cause problems if some malloc implementation
6181 provides an address for "p" such that "p + len" actually overflows for
6182 values of len that are too big and therefore p + len < limit.
6184 This issue was reported to OpenSSL by Guido Vranken
6189 * Constant time flag not preserved in DSA signing
6191 Operations in the DSA signing algorithm should run in constant time in
6192 order to avoid side channel attacks. A flaw in the OpenSSL DSA
6193 implementation means that a non-constant time codepath is followed for
6194 certain operations. This has been demonstrated through a cache-timing
6195 attack to be sufficient for an attacker to recover the private DSA key.
6197 This issue was reported by César Pereida (Aalto University), Billy Brumley
6198 (Tampere University of Technology), and Yuval Yarom (The University of
6199 Adelaide and NICTA).
6204 * DTLS buffered message DoS
6206 In a DTLS connection where handshake messages are delivered out-of-order
6207 those messages that OpenSSL is not yet ready to process will be buffered
6208 for later use. Under certain circumstances, a flaw in the logic means that
6209 those messages do not get removed from the buffer even though the handshake
6210 has been completed. An attacker could force up to approx. 15 messages to
6211 remain in the buffer when they are no longer required. These messages will
6212 be cleared when the DTLS connection is closed. The default maximum size for
6213 a message is 100k. Therefore the attacker could force an additional 1500k
6214 to be consumed per connection. By opening many simulataneous connections an
6215 attacker could cause a DoS attack through memory exhaustion.
6217 This issue was reported to OpenSSL by Quan Luo.
6222 * DTLS replay protection DoS
6224 A flaw in the DTLS replay attack protection mechanism means that records
6225 that arrive for future epochs update the replay protection "window" before
6226 the MAC for the record has been validated. This could be exploited by an
6227 attacker by sending a record for the next epoch (which does not have to
6228 decrypt or have a valid MAC), with a very large sequence number. This means
6229 that all subsequent legitimate packets are dropped causing a denial of
6230 service for a specific DTLS connection.
6232 This issue was reported to OpenSSL by the OCAP audit team.
6237 * Certificate message OOB reads
6239 In OpenSSL 1.0.2 and earlier some missing message length checks can result
6240 in OOB reads of up to 2 bytes beyond an allocated buffer. There is a
6241 theoretical DoS risk but this has not been observed in practice on common
6244 The messages affected are client certificate, client certificate request
6245 and server certificate. As a result the attack can only be performed
6246 against a client or a server which enables client authentication.
6248 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6253 ### Changes between 1.0.1s and 1.0.1t [3 May 2016]
6255 * Prevent padding oracle in AES-NI CBC MAC check
6257 A MITM attacker can use a padding oracle attack to decrypt traffic
6258 when the connection uses an AES CBC cipher and the server support
6261 This issue was introduced as part of the fix for Lucky 13 padding
6262 attack ([CVE-2013-0169]). The padding check was rewritten to be in
6263 constant time by making sure that always the same bytes are read and
6264 compared against either the MAC or padding bytes. But it no longer
6265 checked that there was enough data to have both the MAC and padding
6268 This issue was reported by Juraj Somorovsky using TLS-Attacker.
6273 * Fix EVP_EncodeUpdate overflow
6275 An overflow can occur in the EVP_EncodeUpdate() function which is used for
6276 Base64 encoding of binary data. If an attacker is able to supply very large
6277 amounts of input data then a length check can overflow resulting in a heap
6280 Internally to OpenSSL the EVP_EncodeUpdate() function is primarly used by
6281 the `PEM_write_bio*` family of functions. These are mainly used within the
6282 OpenSSL command line applications, so any application which processes data
6283 from an untrusted source and outputs it as a PEM file should be considered
6284 vulnerable to this issue. User applications that call these APIs directly
6285 with large amounts of untrusted data may also be vulnerable.
6287 This issue was reported by Guido Vranken.
6292 * Fix EVP_EncryptUpdate overflow
6294 An overflow can occur in the EVP_EncryptUpdate() function. If an attacker
6295 is able to supply very large amounts of input data after a previous call to
6296 EVP_EncryptUpdate() with a partial block then a length check can overflow
6297 resulting in a heap corruption. Following an analysis of all OpenSSL
6298 internal usage of the EVP_EncryptUpdate() function all usage is one of two
6299 forms. The first form is where the EVP_EncryptUpdate() call is known to be
6300 the first called function after an EVP_EncryptInit(), and therefore that
6301 specific call must be safe. The second form is where the length passed to
6302 EVP_EncryptUpdate() can be seen from the code to be some small value and
6303 therefore there is no possibility of an overflow. Since all instances are
6304 one of these two forms, it is believed that there can be no overflows in
6305 internal code due to this problem. It should be noted that
6306 EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths.
6307 Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances
6308 of these calls have also been analysed too and it is believed there are no
6309 instances in internal usage where an overflow could occur.
6311 This issue was reported by Guido Vranken.
6316 * Prevent ASN.1 BIO excessive memory allocation
6318 When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio()
6319 a short invalid encoding can casuse allocation of large amounts of memory
6320 potentially consuming excessive resources or exhausting memory.
6322 Any application parsing untrusted data through d2i BIO functions is
6323 affected. The memory based functions such as d2i_X509() are *not* affected.
6324 Since the memory based functions are used by the TLS library, TLS
6325 applications are not affected.
6327 This issue was reported by Brian Carpenter.
6334 ASN1 Strings that are over 1024 bytes can cause an overread in applications
6335 using the X509_NAME_oneline() function on EBCDIC systems. This could result
6336 in arbitrary stack data being returned in the buffer.
6338 This issue was reported by Guido Vranken.
6343 * Modify behavior of ALPN to invoke callback after SNI/servername
6344 callback, such that updates to the SSL_CTX affect ALPN.
6348 * Remove LOW from the DEFAULT cipher list. This removes singles DES from the
6353 * Only remove the SSLv2 methods with the no-ssl2-method option. When the
6354 methods are enabled and ssl2 is disabled the methods return NULL.
6358 ### Changes between 1.0.1r and 1.0.1s [1 Mar 2016]
6360 * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL.
6361 Builds that are not configured with "enable-weak-ssl-ciphers" will not
6362 provide any "EXPORT" or "LOW" strength ciphers.
6366 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2
6367 is by default disabled at build-time. Builds that are not configured with
6368 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used,
6369 users who want to negotiate SSLv2 via the version-flexible SSLv23_method()
6370 will need to explicitly call either of:
6372 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2);
6374 SSL_clear_options(ssl, SSL_OP_NO_SSLv2);
6376 as appropriate. Even if either of those is used, or the application
6377 explicitly uses the version-specific SSLv2_method() or its client and
6378 server variants, SSLv2 ciphers vulnerable to exhaustive search key
6379 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT
6380 ciphers, and SSLv2 56-bit DES are no longer available.
6385 * Fix a double-free in DSA code
6387 A double free bug was discovered when OpenSSL parses malformed DSA private
6388 keys and could lead to a DoS attack or memory corruption for applications
6389 that receive DSA private keys from untrusted sources. This scenario is
6392 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using
6398 * Disable SRP fake user seed to address a server memory leak.
6400 Add a new method SRP_VBASE_get1_by_user that handles the seed properly.
6402 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
6403 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user
6404 was changed to ignore the "fake user" SRP seed, even if the seed
6407 Users should use SRP_VBASE_get1_by_user instead. Note that in
6408 SRP_VBASE_get1_by_user, caller must free the returned value. Note
6409 also that even though configuring the SRP seed attempts to hide
6410 invalid usernames by continuing the handshake with fake
6411 credentials, this behaviour is not constant time and no strong
6412 guarantees are made that the handshake is indistinguishable from
6413 that of a valid user.
6418 * Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption
6420 In the BN_hex2bn function the number of hex digits is calculated using an
6421 int value `i`. Later `bn_expand` is called with a value of `i * 4`. For
6422 large values of `i` this can result in `bn_expand` not allocating any
6423 memory because `i * 4` is negative. This can leave the internal BIGNUM data
6424 field as NULL leading to a subsequent NULL ptr deref. For very large values
6425 of `i`, the calculation `i * 4` could be a positive value smaller than `i`.
6426 In this case memory is allocated to the internal BIGNUM data field, but it
6427 is insufficiently sized leading to heap corruption. A similar issue exists
6428 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn
6429 is ever called by user applications with very large untrusted hex/dec data.
6430 This is anticipated to be a rare occurrence.
6432 All OpenSSL internal usage of these functions use data that is not expected
6433 to be untrusted, e.g. config file data or application command line
6434 arguments. If user developed applications generate config file data based
6435 on untrusted data then it is possible that this could also lead to security
6436 consequences. This is also anticipated to be rare.
6438 This issue was reported to OpenSSL by Guido Vranken.
6443 * Fix memory issues in `BIO_*printf` functions
6445 The internal `fmtstr` function used in processing a "%s" format string in
6446 the `BIO_*printf` functions could overflow while calculating the length of a
6447 string and cause an OOB read when printing very long strings.
6449 Additionally the internal `doapr_outch` function can attempt to write to an
6450 OOB memory location (at an offset from the NULL pointer) in the event of a
6451 memory allocation failure. In 1.0.2 and below this could be caused where
6452 the size of a buffer to be allocated is greater than INT_MAX. E.g. this
6453 could be in processing a very long "%s" format string. Memory leaks can
6456 The first issue may mask the second issue dependent on compiler behaviour.
6457 These problems could enable attacks where large amounts of untrusted data
6458 is passed to the `BIO_*printf` functions. If applications use these functions
6459 in this way then they could be vulnerable. OpenSSL itself uses these
6460 functions when printing out human-readable dumps of ASN.1 data. Therefore
6461 applications that print this data could be vulnerable if the data is from
6462 untrusted sources. OpenSSL command line applications could also be
6463 vulnerable where they print out ASN.1 data, or if untrusted data is passed
6464 as command line arguments.
6466 Libssl is not considered directly vulnerable. Additionally certificates etc
6467 received via remote connections via libssl are also unlikely to be able to
6468 trigger these issues because of message size limits enforced within libssl.
6470 This issue was reported to OpenSSL Guido Vranken.
6475 * Side channel attack on modular exponentiation
6477 A side-channel attack was found which makes use of cache-bank conflicts on
6478 the Intel Sandy-Bridge microarchitecture which could lead to the recovery
6479 of RSA keys. The ability to exploit this issue is limited as it relies on
6480 an attacker who has control of code in a thread running on the same
6481 hyper-threaded core as the victim thread which is performing decryptions.
6483 This issue was reported to OpenSSL by Yuval Yarom, The University of
6484 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and
6485 Nadia Heninger, University of Pennsylvania with more information at
6486 <http://cachebleed.info>.
6491 * Change the req command to generate a 2048-bit RSA/DSA key by default,
6492 if no keysize is specified with default_bits. This fixes an
6493 omission in an earlier change that changed all RSA/DSA key generation
6494 commands to use 2048 bits by default.
6498 ### Changes between 1.0.1q and 1.0.1r [28 Jan 2016]
6500 * Protection for DH small subgroup attacks
6502 As a precautionary measure the SSL_OP_SINGLE_DH_USE option has been
6503 switched on by default and cannot be disabled. This could have some
6508 * SSLv2 doesn't block disabled ciphers
6510 A malicious client can negotiate SSLv2 ciphers that have been disabled on
6511 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
6512 been disabled, provided that the SSLv2 protocol was not also disabled via
6515 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
6516 and Sebastian Schinzel.
6521 * Reject DH handshakes with parameters shorter than 1024 bits.
6525 ### Changes between 1.0.1p and 1.0.1q [3 Dec 2015]
6527 * Certificate verify crash with missing PSS parameter
6529 The signature verification routines will crash with a NULL pointer
6530 dereference if presented with an ASN.1 signature using the RSA PSS
6531 algorithm and absent mask generation function parameter. Since these
6532 routines are used to verify certificate signature algorithms this can be
6533 used to crash any certificate verification operation and exploited in a
6534 DoS attack. Any application which performs certificate verification is
6535 vulnerable including OpenSSL clients and servers which enable client
6538 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
6543 * X509_ATTRIBUTE memory leak
6545 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
6546 memory. This structure is used by the PKCS#7 and CMS routines so any
6547 application which reads PKCS#7 or CMS data from untrusted sources is
6548 affected. SSL/TLS is not affected.
6550 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
6556 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
6557 This changes the decoding behaviour for some invalid messages,
6558 though the change is mostly in the more lenient direction, and
6559 legacy behaviour is preserved as much as possible.
6563 * In DSA_generate_parameters_ex, if the provided seed is too short,
6564 use a random seed, as already documented.
6566 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
6568 ### Changes between 1.0.1o and 1.0.1p [9 Jul 2015]
6570 * Alternate chains certificate forgery
6572 During certificate verfification, OpenSSL will attempt to find an
6573 alternative certificate chain if the first attempt to build such a chain
6574 fails. An error in the implementation of this logic can mean that an
6575 attacker could cause certain checks on untrusted certificates to be
6576 bypassed, such as the CA flag, enabling them to use a valid leaf
6577 certificate to act as a CA and "issue" an invalid certificate.
6579 This issue was reported to OpenSSL by Adam Langley/David Benjamin
6585 * Race condition handling PSK identify hint
6587 If PSK identity hints are received by a multi-threaded client then
6588 the values are wrongly updated in the parent SSL_CTX structure. This can
6589 result in a race condition potentially leading to a double free of the
6595 ### Changes between 1.0.1n and 1.0.1o [12 Jun 2015]
6597 * Fix HMAC ABI incompatibility. The previous version introduced an ABI
6598 incompatibility in the handling of HMAC. The previous ABI has now been
6601 ### Changes between 1.0.1m and 1.0.1n [11 Jun 2015]
6603 * Malformed ECParameters causes infinite loop
6605 When processing an ECParameters structure OpenSSL enters an infinite loop
6606 if the curve specified is over a specially malformed binary polynomial
6609 This can be used to perform denial of service against any
6610 system which processes public keys, certificate requests or
6611 certificates. This includes TLS clients and TLS servers with
6612 client authentication enabled.
6614 This issue was reported to OpenSSL by Joseph Barr-Pixton.
6619 * Exploitable out-of-bounds read in X509_cmp_time
6621 X509_cmp_time does not properly check the length of the ASN1_TIME
6622 string and can read a few bytes out of bounds. In addition,
6623 X509_cmp_time accepts an arbitrary number of fractional seconds in the
6626 An attacker can use this to craft malformed certificates and CRLs of
6627 various sizes and potentially cause a segmentation fault, resulting in
6628 a DoS on applications that verify certificates or CRLs. TLS clients
6629 that verify CRLs are affected. TLS clients and servers with client
6630 authentication enabled may be affected if they use custom verification
6633 This issue was reported to OpenSSL by Robert Swiecki (Google), and
6634 independently by Hanno Böck.
6639 * PKCS7 crash with missing EnvelopedContent
6641 The PKCS#7 parsing code does not handle missing inner EncryptedContent
6642 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
6643 with missing content and trigger a NULL pointer dereference on parsing.
6645 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
6646 structures from untrusted sources are affected. OpenSSL clients and
6647 servers are not affected.
6649 This issue was reported to OpenSSL by Michal Zalewski (Google).
6654 * CMS verify infinite loop with unknown hash function
6656 When verifying a signedData message the CMS code can enter an infinite loop
6657 if presented with an unknown hash function OID. This can be used to perform
6658 denial of service against any system which verifies signedData messages using
6660 This issue was reported to OpenSSL by Johannes Bauer.
6665 * Race condition handling NewSessionTicket
6667 If a NewSessionTicket is received by a multi-threaded client when attempting to
6668 reuse a previous ticket then a race condition can occur potentially leading to
6669 a double free of the ticket data.
6674 * Reject DH handshakes with parameters shorter than 768 bits.
6676 *Kurt Roeckx and Emilia Kasper*
6678 * dhparam: generate 2048-bit parameters by default.
6680 *Kurt Roeckx and Emilia Kasper*
6682 ### Changes between 1.0.1l and 1.0.1m [19 Mar 2015]
6684 * Segmentation fault in ASN1_TYPE_cmp fix
6686 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
6687 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
6688 certificate signature algorithm consistency this can be used to crash any
6689 certificate verification operation and exploited in a DoS attack. Any
6690 application which performs certificate verification is vulnerable including
6691 OpenSSL clients and servers which enable client authentication.
6696 * ASN.1 structure reuse memory corruption fix
6698 Reusing a structure in ASN.1 parsing may allow an attacker to cause
6699 memory corruption via an invalid write. Such reuse is and has been
6700 strongly discouraged and is believed to be rare.
6702 Applications that parse structures containing CHOICE or ANY DEFINED BY
6703 components may be affected. Certificate parsing (d2i_X509 and related
6704 functions) are however not affected. OpenSSL clients and servers are
6710 * PKCS7 NULL pointer dereferences fix
6712 The PKCS#7 parsing code does not handle missing outer ContentInfo
6713 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
6714 missing content and trigger a NULL pointer dereference on parsing.
6716 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
6717 otherwise parse PKCS#7 structures from untrusted sources are
6718 affected. OpenSSL clients and servers are not affected.
6720 This issue was reported to OpenSSL by Michal Zalewski (Google).
6725 * DoS via reachable assert in SSLv2 servers fix
6727 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
6728 servers that both support SSLv2 and enable export cipher suites by sending
6729 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
6731 This issue was discovered by Sean Burford (Google) and Emilia Käsper
6732 (OpenSSL development team).
6737 * Use After Free following d2i_ECPrivatekey error fix
6739 A malformed EC private key file consumed via the d2i_ECPrivateKey function
6740 could cause a use after free condition. This, in turn, could cause a double
6741 free in several private key parsing functions (such as d2i_PrivateKey
6742 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
6743 for applications that receive EC private keys from untrusted
6744 sources. This scenario is considered rare.
6746 This issue was discovered by the BoringSSL project and fixed in their
6752 * X509_to_X509_REQ NULL pointer deref fix
6754 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
6755 the certificate key is invalid. This function is rarely used in practice.
6757 This issue was discovered by Brian Carpenter.
6762 * Removed the export ciphers from the DEFAULT ciphers
6766 ### Changes between 1.0.1k and 1.0.1l [15 Jan 2015]
6768 * Build fixes for the Windows and OpenVMS platforms
6770 *Matt Caswell and Richard Levitte*
6772 ### Changes between 1.0.1j and 1.0.1k [8 Jan 2015]
6774 * Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
6775 message can cause a segmentation fault in OpenSSL due to a NULL pointer
6776 dereference. This could lead to a Denial Of Service attack. Thanks to
6777 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
6782 * Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
6783 dtls1_buffer_record function under certain conditions. In particular this
6784 could occur if an attacker sent repeated DTLS records with the same
6785 sequence number but for the next epoch. The memory leak could be exploited
6786 by an attacker in a Denial of Service attack through memory exhaustion.
6787 Thanks to Chris Mueller for reporting this issue.
6792 * Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
6793 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
6794 method would be set to NULL which could later result in a NULL pointer
6795 dereference. Thanks to Frank Schmirler for reporting this issue.
6800 * Abort handshake if server key exchange message is omitted for ephemeral
6803 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
6804 reporting this issue.
6809 * Remove non-export ephemeral RSA code on client and server. This code
6810 violated the TLS standard by allowing the use of temporary RSA keys in
6811 non-export ciphersuites and could be used by a server to effectively
6812 downgrade the RSA key length used to a value smaller than the server
6813 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
6814 INRIA or reporting this issue.
6819 * Fixed issue where DH client certificates are accepted without verification.
6820 An OpenSSL server will accept a DH certificate for client authentication
6821 without the certificate verify message. This effectively allows a client to
6822 authenticate without the use of a private key. This only affects servers
6823 which trust a client certificate authority which issues certificates
6824 containing DH keys: these are extremely rare and hardly ever encountered.
6825 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
6831 * Ensure that the session ID context of an SSL is updated when its
6832 SSL_CTX is updated via SSL_set_SSL_CTX.
6834 The session ID context is typically set from the parent SSL_CTX,
6835 and can vary with the CTX.
6839 * Fix various certificate fingerprint issues.
6841 By using non-DER or invalid encodings outside the signed portion of a
6842 certificate the fingerprint can be changed without breaking the signature.
6843 Although no details of the signed portion of the certificate can be changed
6844 this can cause problems with some applications: e.g. those using the
6845 certificate fingerprint for blacklists.
6847 1. Reject signatures with non zero unused bits.
6849 If the BIT STRING containing the signature has non zero unused bits reject
6850 the signature. All current signature algorithms require zero unused bits.
6852 2. Check certificate algorithm consistency.
6854 Check the AlgorithmIdentifier inside TBS matches the one in the
6855 certificate signature. NB: this will result in signature failure
6856 errors for some broken certificates.
6858 Thanks to Konrad Kraszewski from Google for reporting this issue.
6860 3. Check DSA/ECDSA signatures use DER.
6862 Re-encode DSA/ECDSA signatures and compare with the original received
6863 signature. Return an error if there is a mismatch.
6865 This will reject various cases including garbage after signature
6866 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
6867 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
6868 (negative or with leading zeroes).
6870 Further analysis was conducted and fixes were developed by Stephen Henson
6871 of the OpenSSL core team.
6877 * Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
6878 results on some platforms, including x86_64. This bug occurs at random
6879 with a very low probability, and is not known to be exploitable in any
6880 way, though its exact impact is difficult to determine. Thanks to Pieter
6881 Wuille (Blockstream) who reported this issue and also suggested an initial
6882 fix. Further analysis was conducted by the OpenSSL development team and
6883 Adam Langley of Google. The final fix was developed by Andy Polyakov of
6884 the OpenSSL core team.
6889 * Do not resume sessions on the server if the negotiated protocol
6890 version does not match the session's version. Resuming with a different
6891 version, while not strictly forbidden by the RFC, is of questionable
6892 sanity and breaks all known clients.
6894 *David Benjamin, Emilia Käsper*
6896 * Tighten handling of the ChangeCipherSpec (CCS) message: reject
6897 early CCS messages during renegotiation. (Note that because
6898 renegotiation is encrypted, this early CCS was not exploitable.)
6902 * Tighten client-side session ticket handling during renegotiation:
6903 ensure that the client only accepts a session ticket if the server sends
6904 the extension anew in the ServerHello. Previously, a TLS client would
6905 reuse the old extension state and thus accept a session ticket if one was
6906 announced in the initial ServerHello.
6908 Similarly, ensure that the client requires a session ticket if one
6909 was advertised in the ServerHello. Previously, a TLS client would
6910 ignore a missing NewSessionTicket message.
6914 ### Changes between 1.0.1i and 1.0.1j [15 Oct 2014]
6918 A flaw in the DTLS SRTP extension parsing code allows an attacker, who
6919 sends a carefully crafted handshake message, to cause OpenSSL to fail
6920 to free up to 64k of memory causing a memory leak. This could be
6921 exploited in a Denial Of Service attack. This issue affects OpenSSL
6922 1.0.1 server implementations for both SSL/TLS and DTLS regardless of
6923 whether SRTP is used or configured. Implementations of OpenSSL that
6924 have been compiled with OPENSSL_NO_SRTP defined are not affected.
6926 The fix was developed by the OpenSSL team.
6931 * Session Ticket Memory Leak.
6933 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
6934 integrity of that ticket is first verified. In the event of a session
6935 ticket integrity check failing, OpenSSL will fail to free memory
6936 causing a memory leak. By sending a large number of invalid session
6937 tickets an attacker could exploit this issue in a Denial Of Service
6943 * Build option no-ssl3 is incomplete.
6945 When OpenSSL is configured with "no-ssl3" as a build option, servers
6946 could accept and complete a SSL 3.0 handshake, and clients could be
6947 configured to send them.
6950 *Akamai and the OpenSSL team*
6952 * Add support for TLS_FALLBACK_SCSV.
6953 Client applications doing fallback retries should call
6954 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
6957 *Adam Langley, Bodo Moeller*
6959 * Add additional DigestInfo checks.
6961 Re-encode DigestInto in DER and check against the original when
6962 verifying RSA signature: this will reject any improperly encoded
6963 DigestInfo structures.
6965 Note: this is a precautionary measure and no attacks are currently known.
6969 ### Changes between 1.0.1h and 1.0.1i [6 Aug 2014]
6971 * Fix SRP buffer overrun vulnerability. Invalid parameters passed to the
6972 SRP code can be overrun an internal buffer. Add sanity check that
6973 g, A, B < N to SRP code.
6975 Thanks to Sean Devlin and Watson Ladd of Cryptography Services, NCC
6976 Group for discovering this issue.
6981 * A flaw in the OpenSSL SSL/TLS server code causes the server to negotiate
6982 TLS 1.0 instead of higher protocol versions when the ClientHello message
6983 is badly fragmented. This allows a man-in-the-middle attacker to force a
6984 downgrade to TLS 1.0 even if both the server and the client support a
6985 higher protocol version, by modifying the client's TLS records.
6987 Thanks to David Benjamin and Adam Langley (Google) for discovering and
6988 researching this issue.
6993 * OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
6994 to a denial of service attack. A malicious server can crash the client
6995 with a null pointer dereference (read) by specifying an anonymous (EC)DH
6996 ciphersuite and sending carefully crafted handshake messages.
6998 Thanks to Felix Gröbert (Google) for discovering and researching this
7004 * By sending carefully crafted DTLS packets an attacker could cause openssl
7005 to leak memory. This can be exploited through a Denial of Service attack.
7006 Thanks to Adam Langley for discovering and researching this issue.
7011 * An attacker can force openssl to consume large amounts of memory whilst
7012 processing DTLS handshake messages. This can be exploited through a
7013 Denial of Service attack.
7014 Thanks to Adam Langley for discovering and researching this issue.
7019 * An attacker can force an error condition which causes openssl to crash
7020 whilst processing DTLS packets due to memory being freed twice. This
7021 can be exploited through a Denial of Service attack.
7022 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
7028 * If a multithreaded client connects to a malicious server using a resumed
7029 session and the server sends an ec point format extension it could write
7030 up to 255 bytes to freed memory.
7032 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
7038 * A malicious server can crash an OpenSSL client with a null pointer
7039 dereference (read) by specifying an SRP ciphersuite even though it was not
7040 properly negotiated with the client. This can be exploited through a
7041 Denial of Service attack.
7043 Thanks to Joonas Kuorilehto and Riku Hietamäki (Codenomicon) for
7044 discovering and researching this issue.
7049 * A flaw in OBJ_obj2txt may cause pretty printing functions such as
7050 X509_name_oneline, X509_name_print_ex et al. to leak some information
7051 from the stack. Applications may be affected if they echo pretty printing
7052 output to the attacker.
7054 Thanks to Ivan Fratric (Google) for discovering this issue.
7057 *Emilia Käsper, and Steve Henson*
7059 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
7060 for corner cases. (Certain input points at infinity could lead to
7061 bogus results, with non-infinity inputs mapped to infinity too.)
7065 ### Changes between 1.0.1g and 1.0.1h [5 Jun 2014]
7067 * Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
7068 handshake can force the use of weak keying material in OpenSSL
7069 SSL/TLS clients and servers.
7071 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
7072 researching this issue. ([CVE-2014-0224])
7074 *KIKUCHI Masashi, Steve Henson*
7076 * Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
7077 OpenSSL DTLS client the code can be made to recurse eventually crashing
7080 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
7083 *Imre Rad, Steve Henson*
7085 * Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
7086 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
7087 client or server. This is potentially exploitable to run arbitrary
7088 code on a vulnerable client or server.
7090 Thanks to Jüri Aedla for reporting this issue. ([CVE-2014-0195])
7092 *Jüri Aedla, Steve Henson*
7094 * Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
7095 are subject to a denial of service attack.
7097 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
7098 this issue. ([CVE-2014-3470])
7100 *Felix Gröbert, Ivan Fratric, Steve Henson*
7102 * Harmonize version and its documentation. -f flag is used to display
7105 *mancha <mancha1@zoho.com>*
7107 * Fix eckey_priv_encode so it immediately returns an error upon a failure
7108 in i2d_ECPrivateKey.
7110 *mancha <mancha1@zoho.com>*
7112 * Fix some double frees. These are not thought to be exploitable.
7114 *mancha <mancha1@zoho.com>*
7116 ### Changes between 1.0.1f and 1.0.1g [7 Apr 2014]
7118 * A missing bounds check in the handling of the TLS heartbeat extension
7119 can be used to reveal up to 64k of memory to a connected client or
7122 Thanks for Neel Mehta of Google Security for discovering this bug and to
7123 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
7124 preparing the fix ([CVE-2014-0160])
7126 *Adam Langley, Bodo Moeller*
7128 * Fix for the attack described in the paper "Recovering OpenSSL
7129 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
7130 by Yuval Yarom and Naomi Benger. Details can be obtained from:
7131 <http://eprint.iacr.org/2014/140>
7133 Thanks to Yuval Yarom and Naomi Benger for discovering this
7134 flaw and to Yuval Yarom for supplying a fix ([CVE-2014-0076])
7136 *Yuval Yarom and Naomi Benger*
7138 * TLS pad extension: draft-agl-tls-padding-03
7140 Workaround for the "TLS hang bug" (see FAQ and PR#2771): if the
7141 TLS client Hello record length value would otherwise be > 255 and
7142 less that 512 pad with a dummy extension containing zeroes so it
7143 is at least 512 bytes long.
7145 *Adam Langley, Steve Henson*
7147 ### Changes between 1.0.1e and 1.0.1f [6 Jan 2014]
7149 * Fix for TLS record tampering bug. A carefully crafted invalid
7150 handshake could crash OpenSSL with a NULL pointer exception.
7151 Thanks to Anton Johansson for reporting this issues.
7154 * Keep original DTLS digest and encryption contexts in retransmission
7155 structures so we can use the previous session parameters if they need
7156 to be resent. ([CVE-2013-6450])
7160 * Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
7161 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
7162 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
7163 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
7164 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
7165 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
7167 *Rob Stradling, Adam Langley*
7169 ### Changes between 1.0.1d and 1.0.1e [11 Feb 2013]
7171 * Correct fix for CVE-2013-0169. The original didn't work on AES-NI
7172 supporting platforms or when small records were transferred.
7174 *Andy Polyakov, Steve Henson*
7176 ### Changes between 1.0.1c and 1.0.1d [5 Feb 2013]
7178 * Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
7180 This addresses the flaw in CBC record processing discovered by
7181 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
7182 at: <http://www.isg.rhul.ac.uk/tls/>
7184 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
7185 Security Group at Royal Holloway, University of London
7186 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
7187 Emilia Käsper for the initial patch.
7190 *Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson*
7192 * Fix flaw in AESNI handling of TLS 1.2 and 1.1 records for CBC mode
7193 ciphersuites which can be exploited in a denial of service attack.
7194 Thanks go to and to Adam Langley <agl@chromium.org> for discovering
7195 and detecting this bug and to Wolfgang Ettlinger
7196 <wolfgang.ettlinger@gmail.com> for independently discovering this issue.
7201 * Return an error when checking OCSP signatures when key is NULL.
7202 This fixes a DoS attack. ([CVE-2013-0166])
7206 * Make openssl verify return errors.
7208 *Chris Palmer <palmer@google.com> and Ben Laurie*
7210 * Call OCSP Stapling callback after ciphersuite has been chosen, so
7211 the right response is stapled. Also change SSL_get_certificate()
7212 so it returns the certificate actually sent.
7213 See <http://rt.openssl.org/Ticket/Display.html?id=2836>.
7215 *Rob Stradling <rob.stradling@comodo.com>*
7217 * Fix possible deadlock when decoding public keys.
7221 * Don't use TLS 1.0 record version number in initial client hello
7226 ### Changes between 1.0.1b and 1.0.1c [10 May 2012]
7228 * Sanity check record length before skipping explicit IV in TLS
7229 1.2, 1.1 and DTLS to fix DoS attack.
7231 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
7232 fuzzing as a service testing platform.
7237 * Initialise tkeylen properly when encrypting CMS messages.
7238 Thanks to Solar Designer of Openwall for reporting this issue.
7242 * In FIPS mode don't try to use composite ciphers as they are not
7247 ### Changes between 1.0.1a and 1.0.1b [26 Apr 2012]
7249 * OpenSSL 1.0.0 sets SSL_OP_ALL to 0x80000FFFL and OpenSSL 1.0.1 and
7250 1.0.1a set SSL_OP_NO_TLSv1_1 to 0x00000400L which would unfortunately
7251 mean any application compiled against OpenSSL 1.0.0 headers setting
7252 SSL_OP_ALL would also set SSL_OP_NO_TLSv1_1, unintentionally disabling
7253 TLS 1.1 also. Fix this by changing the value of SSL_OP_NO_TLSv1_1 to
7254 0x10000000L Any application which was previously compiled against
7255 OpenSSL 1.0.1 or 1.0.1a headers and which cares about SSL_OP_NO_TLSv1_1
7256 will need to be recompiled as a result. Letting be results in
7257 inability to disable specifically TLS 1.1 and in client context,
7258 in unlike event, limit maximum offered version to TLS 1.0 [see below].
7262 * In order to ensure interoperability SSL_OP_NO_protocolX does not
7263 disable just protocol X, but all protocols above X *if* there are
7264 protocols *below* X still enabled. In more practical terms it means
7265 that if application wants to disable TLS1.0 in favor of TLS1.1 and
7266 above, it's not sufficient to pass `SSL_OP_NO_TLSv1`, one has to pass
7267 `SSL_OP_NO_TLSv1|SSL_OP_NO_SSLv3|SSL_OP_NO_SSLv2`. This applies to
7272 ### Changes between 1.0.1 and 1.0.1a [19 Apr 2012]
7274 * Check for potentially exploitable overflows in asn1_d2i_read_bio
7275 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
7276 in CRYPTO_realloc_clean.
7278 Thanks to Tavis Ormandy, Google Security Team, for discovering this
7279 issue and to Adam Langley <agl@chromium.org> for fixing it.
7282 *Adam Langley (Google), Tavis Ormandy, Google Security Team*
7284 * Don't allow TLS 1.2 SHA-256 ciphersuites in TLS 1.0, 1.1 connections.
7288 * Workarounds for some broken servers that "hang" if a client hello
7289 record length exceeds 255 bytes.
7291 1. Do not use record version number > TLS 1.0 in initial client
7292 hello: some (but not all) hanging servers will now work.
7293 2. If we set OPENSSL_MAX_TLS1_2_CIPHER_LENGTH this will truncate
7294 the number of ciphers sent in the client hello. This should be
7295 set to an even number, such as 50, for example by passing:
7296 -DOPENSSL_MAX_TLS1_2_CIPHER_LENGTH=50 to config or Configure.
7297 Most broken servers should now work.
7298 3. If all else fails setting OPENSSL_NO_TLS1_2_CLIENT will disable
7299 TLS 1.2 client support entirely.
7303 * Fix SEGV in Vector Permutation AES module observed in OpenSSH.
7307 ### Changes between 1.0.0h and 1.0.1 [14 Mar 2012]
7309 * Add compatibility with old MDC2 signatures which use an ASN1 OCTET
7310 STRING form instead of a DigestInfo.
7314 * The format used for MDC2 RSA signatures is inconsistent between EVP
7315 and the RSA_sign/RSA_verify functions. This was made more apparent when
7316 OpenSSL used RSA_sign/RSA_verify for some RSA signatures in particular
7317 those which went through EVP_PKEY_METHOD in 1.0.0 and later. Detect
7318 the correct format in RSA_verify so both forms transparently work.
7322 * Some servers which support TLS 1.0 can choke if we initially indicate
7323 support for TLS 1.2 and later renegotiate using TLS 1.0 in the RSA
7324 encrypted premaster secret. As a workaround use the maximum permitted
7325 client version in client hello, this should keep such servers happy
7326 and still work with previous versions of OpenSSL.
7330 * Add support for TLS/DTLS heartbeats.
7332 *Robin Seggelmann <seggelmann@fh-muenster.de>*
7334 * Add support for SCTP.
7336 *Robin Seggelmann <seggelmann@fh-muenster.de>*
7338 * Improved PRNG seeding for VOS.
7340 *Paul Green <Paul.Green@stratus.com>*
7342 * Extensive assembler packs updates, most notably:
7344 - x86[_64]: AES-NI, PCLMULQDQ, RDRAND support;
7345 - x86[_64]: SSSE3 support (SHA1, vector-permutation AES);
7346 - x86_64: bit-sliced AES implementation;
7347 - ARM: NEON support, contemporary platforms optimizations;
7348 - s390x: z196 support;
7349 - `*`: GHASH and GF(2^m) multiplication implementations;
7353 * Make TLS-SRP code conformant with RFC 5054 API cleanup
7354 (removal of unnecessary code)
7356 *Peter Sylvester <peter.sylvester@edelweb.fr>*
7358 * Add TLS key material exporter from RFC 5705.
7362 * Add DTLS-SRTP negotiation from RFC 5764.
7366 * Add Next Protocol Negotiation,
7367 <http://tools.ietf.org/html/draft-agl-tls-nextprotoneg-00>. Can be
7368 disabled with a no-npn flag to config or Configure. Code donated
7371 *Adam Langley <agl@google.com> and Ben Laurie*
7373 * Add optional 64-bit optimized implementations of elliptic curves NIST-P224,
7374 NIST-P256, NIST-P521, with constant-time single point multiplication on
7375 typical inputs. Compiler support for the nonstandard type `__uint128_t` is
7376 required to use this (present in gcc 4.4 and later, for 64-bit builds).
7377 Code made available under Apache License version 2.0.
7379 Specify "enable-ec_nistp_64_gcc_128" on the Configure (or config) command
7380 line to include this in your build of OpenSSL, and run "make depend" (or
7381 "make update"). This enables the following EC_METHODs:
7383 EC_GFp_nistp224_method()
7384 EC_GFp_nistp256_method()
7385 EC_GFp_nistp521_method()
7387 EC_GROUP_new_by_curve_name() will automatically use these (while
7388 EC_GROUP_new_curve_GFp() currently prefers the more flexible
7391 *Emilia Käsper, Adam Langley, Bodo Moeller (Google)*
7393 * Use type ossl_ssize_t instead of ssize_t which isn't available on
7394 all platforms. Move ssize_t definition from e_os.h to the public
7395 header file e_os2.h as it now appears in public header file cms.h
7399 * New -sigopt option to the ca, req and x509 utilities. Additional
7400 signature parameters can be passed using this option and in
7405 * Add RSA PSS signing function. This will generate and set the
7406 appropriate AlgorithmIdentifiers for PSS based on those in the
7407 corresponding EVP_MD_CTX structure. No application support yet.
7411 * Support for companion algorithm specific ASN1 signing routines.
7412 New function ASN1_item_sign_ctx() signs a pre-initialised
7413 EVP_MD_CTX structure and sets AlgorithmIdentifiers based on
7414 the appropriate parameters.
7418 * Add new algorithm specific ASN1 verification initialisation function
7419 to EVP_PKEY_ASN1_METHOD: this is not in EVP_PKEY_METHOD since the ASN1
7420 handling will be the same no matter what EVP_PKEY_METHOD is used.
7421 Add a PSS handler to support verification of PSS signatures: checked
7422 against a number of sample certificates.
7426 * Add signature printing for PSS. Add PSS OIDs.
7428 *Steve Henson, Martin Kaiser <lists@kaiser.cx>*
7430 * Add algorithm specific signature printing. An individual ASN1 method
7431 can now print out signatures instead of the standard hex dump.
7433 More complex signatures (e.g. PSS) can print out more meaningful
7434 information. Include DSA version that prints out the signature
7439 * Password based recipient info support for CMS library: implementing
7444 * Split password based encryption into PBES2 and PBKDF2 functions. This
7445 neatly separates the code into cipher and PBE sections and is required
7446 for some algorithms that split PBES2 into separate pieces (such as
7447 password based CMS).
7451 * Session-handling fixes:
7452 - Fix handling of connections that are resuming with a session ID,
7453 but also support Session Tickets.
7454 - Fix a bug that suppressed issuing of a new ticket if the client
7455 presented a ticket with an expired session.
7456 - Try to set the ticket lifetime hint to something reasonable.
7457 - Make tickets shorter by excluding irrelevant information.
7458 - On the client side, don't ignore renewed tickets.
7460 *Adam Langley, Bodo Moeller (Google)*
7462 * Fix PSK session representation.
7466 * Add RC4-MD5 and AESNI-SHA1 "stitched" implementations.
7468 This work was sponsored by Intel.
7472 * Add GCM support to TLS library. Some custom code is needed to split
7473 the IV between the fixed (from PRF) and explicit (from TLS record)
7474 portions. This adds all GCM ciphersuites supported by RFC5288 and
7475 RFC5289. Generalise some `AES*` cipherstrings to include GCM and
7476 add a special AESGCM string for GCM only.
7480 * Expand range of ctrls for AES GCM. Permit setting invocation
7481 field on decrypt and retrieval of invocation field only on encrypt.
7485 * Add HMAC ECC ciphersuites from RFC5289. Include SHA384 PRF support.
7486 As required by RFC5289 these ciphersuites cannot be used if for
7487 versions of TLS earlier than 1.2.
7491 * For FIPS capable OpenSSL interpret a NULL default public key method
7492 as unset and return the appropriate default but do *not* set the default.
7493 This means we can return the appropriate method in applications that
7494 switch between FIPS and non-FIPS modes.
7498 * Redirect HMAC and CMAC operations to FIPS module in FIPS mode. If an
7499 ENGINE is used then we cannot handle that in the FIPS module so we
7500 keep original code iff non-FIPS operations are allowed.
7504 * Add -attime option to openssl utilities.
7506 *Peter Eckersley <pde@eff.org>, Ben Laurie and Steve Henson*
7508 * Redirect DSA and DH operations to FIPS module in FIPS mode.
7512 * Redirect ECDSA and ECDH operations to FIPS module in FIPS mode. Also use
7513 FIPS EC methods unconditionally for now.
7517 * New build option no-ec2m to disable characteristic 2 code.
7521 * Backport libcrypto audit of return value checking from 1.1.0-dev; not
7522 all cases can be covered as some introduce binary incompatibilities.
7526 * Redirect RSA operations to FIPS module including keygen,
7527 encrypt, decrypt, sign and verify. Block use of non FIPS RSA methods.
7531 * Add similar low-level API blocking to ciphers.
7535 * low-level digest APIs are not approved in FIPS mode: any attempt
7536 to use these will cause a fatal error. Applications that *really* want
7537 to use them can use the `private_*` version instead.
7541 * Redirect cipher operations to FIPS module for FIPS builds.
7545 * Redirect digest operations to FIPS module for FIPS builds.
7549 * Update build system to add "fips" flag which will link in fipscanister.o
7550 for static and shared library builds embedding a signature if needed.
7554 * Output TLS supported curves in preference order instead of numerical
7555 order. This is currently hardcoded for the highest order curves first.
7556 This should be configurable so applications can judge speed vs strength.
7560 * Add TLS v1.2 server support for client authentication.
7564 * Add support for FIPS mode in ssl library: disable SSLv3, non-FIPS ciphers
7569 * Functions FIPS_mode_set() and FIPS_mode() which call the underlying
7570 FIPS modules versions.
7574 * Add TLS v1.2 client side support for client authentication. Keep cache
7575 of handshake records longer as we don't know the hash algorithm to use
7576 until after the certificate request message is received.
7580 * Initial TLS v1.2 client support. Add a default signature algorithms
7581 extension including all the algorithms we support. Parse new signature
7582 format in client key exchange. Relax some ECC signing restrictions for
7583 TLS v1.2 as indicated in RFC5246.
7587 * Add server support for TLS v1.2 signature algorithms extension. Switch
7588 to new signature format when needed using client digest preference.
7589 All server ciphersuites should now work correctly in TLS v1.2. No client
7590 support yet and no support for client certificates.
7594 * Initial TLS v1.2 support. Add new SHA256 digest to ssl code, switch
7595 to SHA256 for PRF when using TLS v1.2 and later. Add new SHA256 based
7596 ciphersuites. At present only RSA key exchange ciphersuites work with
7597 TLS v1.2. Add new option for TLS v1.2 replacing the old and obsolete
7598 SSL_OP_PKCS1_CHECK flags with SSL_OP_NO_TLSv1_2. New TLSv1.2 methods
7599 and version checking.
7603 * New option OPENSSL_NO_SSL_INTERN. If an application can be compiled
7604 with this defined it will not be affected by any changes to ssl internal
7605 structures. Add several utility functions to allow openssl application
7606 to work with OPENSSL_NO_SSL_INTERN defined.
7610 * A long standing patch to add support for SRP from EdelWeb (Peter
7611 Sylvester and Christophe Renou) was integrated.
7612 *Christophe Renou <christophe.renou@edelweb.fr>, Peter Sylvester
7613 <peter.sylvester@edelweb.fr>, Tom Wu <tjw@cs.stanford.edu>, and
7616 * Add functions to copy EVP_PKEY_METHOD and retrieve flags and id.
7620 * Permit abbreviated handshakes when renegotiating using the function
7621 SSL_renegotiate_abbreviated().
7623 *Robin Seggelmann <seggelmann@fh-muenster.de>*
7625 * Add call to ENGINE_register_all_complete() to
7626 ENGINE_load_builtin_engines(), so some implementations get used
7627 automatically instead of needing explicit application support.
7631 * Add support for TLS key exporter as described in RFC5705.
7633 *Robin Seggelmann <seggelmann@fh-muenster.de>, Steve Henson*
7635 * Initial TLSv1.1 support. Since TLSv1.1 is very similar to TLS v1.0 only
7636 a few changes are required:
7638 Add SSL_OP_NO_TLSv1_1 flag.
7639 Add TLSv1_1 methods.
7640 Update version checking logic to handle version 1.1.
7641 Add explicit IV handling (ported from DTLS code).
7642 Add command line options to s_client/s_server.
7649 ### Changes between 1.0.0s and 1.0.0t [3 Dec 2015]
7651 * X509_ATTRIBUTE memory leak
7653 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
7654 memory. This structure is used by the PKCS#7 and CMS routines so any
7655 application which reads PKCS#7 or CMS data from untrusted sources is
7656 affected. SSL/TLS is not affected.
7658 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
7664 * Race condition handling PSK identify hint
7666 If PSK identity hints are received by a multi-threaded client then
7667 the values are wrongly updated in the parent SSL_CTX structure. This can
7668 result in a race condition potentially leading to a double free of the
7674 ### Changes between 1.0.0r and 1.0.0s [11 Jun 2015]
7676 * Malformed ECParameters causes infinite loop
7678 When processing an ECParameters structure OpenSSL enters an infinite loop
7679 if the curve specified is over a specially malformed binary polynomial
7682 This can be used to perform denial of service against any
7683 system which processes public keys, certificate requests or
7684 certificates. This includes TLS clients and TLS servers with
7685 client authentication enabled.
7687 This issue was reported to OpenSSL by Joseph Barr-Pixton.
7692 * Exploitable out-of-bounds read in X509_cmp_time
7694 X509_cmp_time does not properly check the length of the ASN1_TIME
7695 string and can read a few bytes out of bounds. In addition,
7696 X509_cmp_time accepts an arbitrary number of fractional seconds in the
7699 An attacker can use this to craft malformed certificates and CRLs of
7700 various sizes and potentially cause a segmentation fault, resulting in
7701 a DoS on applications that verify certificates or CRLs. TLS clients
7702 that verify CRLs are affected. TLS clients and servers with client
7703 authentication enabled may be affected if they use custom verification
7706 This issue was reported to OpenSSL by Robert Swiecki (Google), and
7707 independently by Hanno Böck.
7712 * PKCS7 crash with missing EnvelopedContent
7714 The PKCS#7 parsing code does not handle missing inner EncryptedContent
7715 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
7716 with missing content and trigger a NULL pointer dereference on parsing.
7718 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
7719 structures from untrusted sources are affected. OpenSSL clients and
7720 servers are not affected.
7722 This issue was reported to OpenSSL by Michal Zalewski (Google).
7727 * CMS verify infinite loop with unknown hash function
7729 When verifying a signedData message the CMS code can enter an infinite loop
7730 if presented with an unknown hash function OID. This can be used to perform
7731 denial of service against any system which verifies signedData messages using
7733 This issue was reported to OpenSSL by Johannes Bauer.
7738 * Race condition handling NewSessionTicket
7740 If a NewSessionTicket is received by a multi-threaded client when attempting to
7741 reuse a previous ticket then a race condition can occur potentially leading to
7742 a double free of the ticket data.
7747 ### Changes between 1.0.0q and 1.0.0r [19 Mar 2015]
7749 * Segmentation fault in ASN1_TYPE_cmp fix
7751 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
7752 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
7753 certificate signature algorithm consistency this can be used to crash any
7754 certificate verification operation and exploited in a DoS attack. Any
7755 application which performs certificate verification is vulnerable including
7756 OpenSSL clients and servers which enable client authentication.
7761 * ASN.1 structure reuse memory corruption fix
7763 Reusing a structure in ASN.1 parsing may allow an attacker to cause
7764 memory corruption via an invalid write. Such reuse is and has been
7765 strongly discouraged and is believed to be rare.
7767 Applications that parse structures containing CHOICE or ANY DEFINED BY
7768 components may be affected. Certificate parsing (d2i_X509 and related
7769 functions) are however not affected. OpenSSL clients and servers are
7775 * PKCS7 NULL pointer dereferences fix
7777 The PKCS#7 parsing code does not handle missing outer ContentInfo
7778 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
7779 missing content and trigger a NULL pointer dereference on parsing.
7781 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
7782 otherwise parse PKCS#7 structures from untrusted sources are
7783 affected. OpenSSL clients and servers are not affected.
7785 This issue was reported to OpenSSL by Michal Zalewski (Google).
7790 * DoS via reachable assert in SSLv2 servers fix
7792 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
7793 servers that both support SSLv2 and enable export cipher suites by sending
7794 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
7796 This issue was discovered by Sean Burford (Google) and Emilia Käsper
7797 (OpenSSL development team).
7802 * Use After Free following d2i_ECPrivatekey error fix
7804 A malformed EC private key file consumed via the d2i_ECPrivateKey function
7805 could cause a use after free condition. This, in turn, could cause a double
7806 free in several private key parsing functions (such as d2i_PrivateKey
7807 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
7808 for applications that receive EC private keys from untrusted
7809 sources. This scenario is considered rare.
7811 This issue was discovered by the BoringSSL project and fixed in their
7817 * X509_to_X509_REQ NULL pointer deref fix
7819 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
7820 the certificate key is invalid. This function is rarely used in practice.
7822 This issue was discovered by Brian Carpenter.
7827 * Removed the export ciphers from the DEFAULT ciphers
7831 ### Changes between 1.0.0p and 1.0.0q [15 Jan 2015]
7833 * Build fixes for the Windows and OpenVMS platforms
7835 *Matt Caswell and Richard Levitte*
7837 ### Changes between 1.0.0o and 1.0.0p [8 Jan 2015]
7839 * Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
7840 message can cause a segmentation fault in OpenSSL due to a NULL pointer
7841 dereference. This could lead to a Denial Of Service attack. Thanks to
7842 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
7847 * Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
7848 dtls1_buffer_record function under certain conditions. In particular this
7849 could occur if an attacker sent repeated DTLS records with the same
7850 sequence number but for the next epoch. The memory leak could be exploited
7851 by an attacker in a Denial of Service attack through memory exhaustion.
7852 Thanks to Chris Mueller for reporting this issue.
7857 * Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
7858 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
7859 method would be set to NULL which could later result in a NULL pointer
7860 dereference. Thanks to Frank Schmirler for reporting this issue.
7865 * Abort handshake if server key exchange message is omitted for ephemeral
7868 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
7869 reporting this issue.
7874 * Remove non-export ephemeral RSA code on client and server. This code
7875 violated the TLS standard by allowing the use of temporary RSA keys in
7876 non-export ciphersuites and could be used by a server to effectively
7877 downgrade the RSA key length used to a value smaller than the server
7878 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
7879 INRIA or reporting this issue.
7884 * Fixed issue where DH client certificates are accepted without verification.
7885 An OpenSSL server will accept a DH certificate for client authentication
7886 without the certificate verify message. This effectively allows a client to
7887 authenticate without the use of a private key. This only affects servers
7888 which trust a client certificate authority which issues certificates
7889 containing DH keys: these are extremely rare and hardly ever encountered.
7890 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
7896 * Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
7897 results on some platforms, including x86_64. This bug occurs at random
7898 with a very low probability, and is not known to be exploitable in any
7899 way, though its exact impact is difficult to determine. Thanks to Pieter
7900 Wuille (Blockstream) who reported this issue and also suggested an initial
7901 fix. Further analysis was conducted by the OpenSSL development team and
7902 Adam Langley of Google. The final fix was developed by Andy Polyakov of
7903 the OpenSSL core team.
7908 * Fix various certificate fingerprint issues.
7910 By using non-DER or invalid encodings outside the signed portion of a
7911 certificate the fingerprint can be changed without breaking the signature.
7912 Although no details of the signed portion of the certificate can be changed
7913 this can cause problems with some applications: e.g. those using the
7914 certificate fingerprint for blacklists.
7916 1. Reject signatures with non zero unused bits.
7918 If the BIT STRING containing the signature has non zero unused bits reject
7919 the signature. All current signature algorithms require zero unused bits.
7921 2. Check certificate algorithm consistency.
7923 Check the AlgorithmIdentifier inside TBS matches the one in the
7924 certificate signature. NB: this will result in signature failure
7925 errors for some broken certificates.
7927 Thanks to Konrad Kraszewski from Google for reporting this issue.
7929 3. Check DSA/ECDSA signatures use DER.
7931 Reencode DSA/ECDSA signatures and compare with the original received
7932 signature. Return an error if there is a mismatch.
7934 This will reject various cases including garbage after signature
7935 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
7936 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
7937 (negative or with leading zeroes).
7939 Further analysis was conducted and fixes were developed by Stephen Henson
7940 of the OpenSSL core team.
7946 ### Changes between 1.0.0n and 1.0.0o [15 Oct 2014]
7948 * Session Ticket Memory Leak.
7950 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
7951 integrity of that ticket is first verified. In the event of a session
7952 ticket integrity check failing, OpenSSL will fail to free memory
7953 causing a memory leak. By sending a large number of invalid session
7954 tickets an attacker could exploit this issue in a Denial Of Service
7960 * Build option no-ssl3 is incomplete.
7962 When OpenSSL is configured with "no-ssl3" as a build option, servers
7963 could accept and complete a SSL 3.0 handshake, and clients could be
7964 configured to send them.
7967 *Akamai and the OpenSSL team*
7969 * Add support for TLS_FALLBACK_SCSV.
7970 Client applications doing fallback retries should call
7971 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
7974 *Adam Langley, Bodo Moeller*
7976 * Add additional DigestInfo checks.
7978 Reencode DigestInto in DER and check against the original when
7979 verifying RSA signature: this will reject any improperly encoded
7980 DigestInfo structures.
7982 Note: this is a precautionary measure and no attacks are currently known.
7986 ### Changes between 1.0.0m and 1.0.0n [6 Aug 2014]
7988 * OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
7989 to a denial of service attack. A malicious server can crash the client
7990 with a null pointer dereference (read) by specifying an anonymous (EC)DH
7991 ciphersuite and sending carefully crafted handshake messages.
7993 Thanks to Felix Gröbert (Google) for discovering and researching this
7999 * By sending carefully crafted DTLS packets an attacker could cause openssl
8000 to leak memory. This can be exploited through a Denial of Service attack.
8001 Thanks to Adam Langley for discovering and researching this issue.
8006 * An attacker can force openssl to consume large amounts of memory whilst
8007 processing DTLS handshake messages. This can be exploited through a
8008 Denial of Service attack.
8009 Thanks to Adam Langley for discovering and researching this issue.
8014 * An attacker can force an error condition which causes openssl to crash
8015 whilst processing DTLS packets due to memory being freed twice. This
8016 can be exploited through a Denial of Service attack.
8017 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
8023 * If a multithreaded client connects to a malicious server using a resumed
8024 session and the server sends an ec point format extension it could write
8025 up to 255 bytes to freed memory.
8027 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
8033 * A flaw in OBJ_obj2txt may cause pretty printing functions such as
8034 X509_name_oneline, X509_name_print_ex et al. to leak some information
8035 from the stack. Applications may be affected if they echo pretty printing
8036 output to the attacker.
8038 Thanks to Ivan Fratric (Google) for discovering this issue.
8041 *Emilia Käsper, and Steve Henson*
8043 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
8044 for corner cases. (Certain input points at infinity could lead to
8045 bogus results, with non-infinity inputs mapped to infinity too.)
8049 ### Changes between 1.0.0l and 1.0.0m [5 Jun 2014]
8051 * Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
8052 handshake can force the use of weak keying material in OpenSSL
8053 SSL/TLS clients and servers.
8055 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
8056 researching this issue. ([CVE-2014-0224])
8058 *KIKUCHI Masashi, Steve Henson*
8060 * Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
8061 OpenSSL DTLS client the code can be made to recurse eventually crashing
8064 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
8067 *Imre Rad, Steve Henson*
8069 * Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
8070 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
8071 client or server. This is potentially exploitable to run arbitrary
8072 code on a vulnerable client or server.
8074 Thanks to Jüri Aedla for reporting this issue. ([CVE-2014-0195])
8076 *Jüri Aedla, Steve Henson*
8078 * Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
8079 are subject to a denial of service attack.
8081 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
8082 this issue. ([CVE-2014-3470])
8084 *Felix Gröbert, Ivan Fratric, Steve Henson*
8086 * Harmonize version and its documentation. -f flag is used to display
8089 *mancha <mancha1@zoho.com>*
8091 * Fix eckey_priv_encode so it immediately returns an error upon a failure
8092 in i2d_ECPrivateKey.
8094 *mancha <mancha1@zoho.com>*
8096 * Fix some double frees. These are not thought to be exploitable.
8098 *mancha <mancha1@zoho.com>*
8100 * Fix for the attack described in the paper "Recovering OpenSSL
8101 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
8102 by Yuval Yarom and Naomi Benger. Details can be obtained from:
8103 <http://eprint.iacr.org/2014/140>
8105 Thanks to Yuval Yarom and Naomi Benger for discovering this
8106 flaw and to Yuval Yarom for supplying a fix ([CVE-2014-0076])
8108 *Yuval Yarom and Naomi Benger*
8110 ### Changes between 1.0.0k and 1.0.0l [6 Jan 2014]
8112 * Keep original DTLS digest and encryption contexts in retransmission
8113 structures so we can use the previous session parameters if they need
8114 to be resent. ([CVE-2013-6450])
8118 * Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
8119 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
8120 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
8121 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
8122 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
8123 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
8125 *Rob Stradling, Adam Langley*
8127 ### Changes between 1.0.0j and 1.0.0k [5 Feb 2013]
8129 * Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
8131 This addresses the flaw in CBC record processing discovered by
8132 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
8133 at: <http://www.isg.rhul.ac.uk/tls/>
8135 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
8136 Security Group at Royal Holloway, University of London
8137 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
8138 Emilia Käsper for the initial patch.
8141 *Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson*
8143 * Return an error when checking OCSP signatures when key is NULL.
8144 This fixes a DoS attack. ([CVE-2013-0166])
8148 * Call OCSP Stapling callback after ciphersuite has been chosen, so
8149 the right response is stapled. Also change SSL_get_certificate()
8150 so it returns the certificate actually sent.
8151 See <http://rt.openssl.org/Ticket/Display.html?id=2836>.
8152 (This is a backport)
8154 *Rob Stradling <rob.stradling@comodo.com>*
8156 * Fix possible deadlock when decoding public keys.
8160 ### Changes between 1.0.0i and 1.0.0j [10 May 2012]
8162 [NB: OpenSSL 1.0.0i and later 1.0.0 patch levels were released after
8165 * Sanity check record length before skipping explicit IV in DTLS
8168 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
8169 fuzzing as a service testing platform.
8174 * Initialise tkeylen properly when encrypting CMS messages.
8175 Thanks to Solar Designer of Openwall for reporting this issue.
8179 ### Changes between 1.0.0h and 1.0.0i [19 Apr 2012]
8181 * Check for potentially exploitable overflows in asn1_d2i_read_bio
8182 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
8183 in CRYPTO_realloc_clean.
8185 Thanks to Tavis Ormandy, Google Security Team, for discovering this
8186 issue and to Adam Langley <agl@chromium.org> for fixing it.
8189 *Adam Langley (Google), Tavis Ormandy, Google Security Team*
8191 ### Changes between 1.0.0g and 1.0.0h [12 Mar 2012]
8193 * Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness
8194 in CMS and PKCS7 code. When RSA decryption fails use a random key for
8195 content decryption and always return the same error. Note: this attack
8196 needs on average 2^20 messages so it only affects automated senders. The
8197 old behaviour can be re-enabled in the CMS code by setting the
8198 CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where
8199 an MMA defence is not necessary.
8200 Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering
8201 this issue. ([CVE-2012-0884])
8205 * Fix CVE-2011-4619: make sure we really are receiving a
8206 client hello before rejecting multiple SGC restarts. Thanks to
8207 Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug.
8211 ### Changes between 1.0.0f and 1.0.0g [18 Jan 2012]
8213 * Fix for DTLS DoS issue introduced by fix for CVE-2011-4109.
8214 Thanks to Antonio Martin, Enterprise Secure Access Research and
8215 Development, Cisco Systems, Inc. for discovering this bug and
8216 preparing a fix. ([CVE-2012-0050])
8220 ### Changes between 1.0.0e and 1.0.0f [4 Jan 2012]
8222 * Nadhem Alfardan and Kenny Paterson have discovered an extension
8223 of the Vaudenay padding oracle attack on CBC mode encryption
8224 which enables an efficient plaintext recovery attack against
8225 the OpenSSL implementation of DTLS. Their attack exploits timing
8226 differences arising during decryption processing. A research
8227 paper describing this attack can be found at:
8228 <http://www.isg.rhul.ac.uk/~kp/dtls.pdf>
8229 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
8230 Security Group at Royal Holloway, University of London
8231 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
8232 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
8233 for preparing the fix. ([CVE-2011-4108])
8235 *Robin Seggelmann, Michael Tuexen*
8237 * Clear bytes used for block padding of SSL 3.0 records.
8240 *Adam Langley (Google)*
8242 * Only allow one SGC handshake restart for SSL/TLS. Thanks to George
8243 Kadianakis <desnacked@gmail.com> for discovering this issue and
8244 Adam Langley for preparing the fix. ([CVE-2011-4619])
8246 *Adam Langley (Google)*
8248 * Check parameters are not NULL in GOST ENGINE. ([CVE-2012-0027])
8250 *Andrey Kulikov <amdeich@gmail.com>*
8252 * Prevent malformed RFC3779 data triggering an assertion failure.
8253 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
8254 and Rob Austein <sra@hactrn.net> for fixing it. ([CVE-2011-4577])
8256 *Rob Austein <sra@hactrn.net>*
8258 * Improved PRNG seeding for VOS.
8260 *Paul Green <Paul.Green@stratus.com>*
8262 * Fix ssl_ciph.c set-up race.
8264 *Adam Langley (Google)*
8266 * Fix spurious failures in ecdsatest.c.
8268 *Emilia Käsper (Google)*
8270 * Fix the BIO_f_buffer() implementation (which was mixing different
8271 interpretations of the `..._len` fields).
8273 *Adam Langley (Google)*
8275 * Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
8276 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
8277 threads won't reuse the same blinding coefficients.
8279 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
8280 lock to call BN_BLINDING_invert_ex, and avoids one use of
8281 BN_BLINDING_update for each BN_BLINDING structure (previously,
8282 the last update always remained unused).
8284 *Emilia Käsper (Google)*
8286 * In ssl3_clear, preserve s3->init_extra along with s3->rbuf.
8288 *Bob Buckholz (Google)*
8290 ### Changes between 1.0.0d and 1.0.0e [6 Sep 2011]
8292 * Fix bug where CRLs with nextUpdate in the past are sometimes accepted
8293 by initialising X509_STORE_CTX properly. ([CVE-2011-3207])
8295 *Kaspar Brand <ossl@velox.ch>*
8297 * Fix SSL memory handling for (EC)DH ciphersuites, in particular
8298 for multi-threaded use of ECDH. ([CVE-2011-3210])
8300 *Adam Langley (Google)*
8302 * Fix x509_name_ex_d2i memory leak on bad inputs.
8306 * Remove hard coded ecdsaWithSHA1 signature tests in ssl code and check
8307 signature public key algorithm by using OID xref utilities instead.
8308 Before this you could only use some ECC ciphersuites with SHA1 only.
8312 * Add protection against ECDSA timing attacks as mentioned in the paper
8313 by Billy Bob Brumley and Nicola Tuveri, see:
8314 <http://eprint.iacr.org/2011/232.pdf>
8316 *Billy Bob Brumley and Nicola Tuveri*
8318 ### Changes between 1.0.0c and 1.0.0d [8 Feb 2011]
8320 * Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
8322 *Neel Mehta, Adam Langley, Bodo Moeller (Google)*
8324 * Fix bug in string printing code: if *any* escaping is enabled we must
8325 escape the escape character (backslash) or the resulting string is
8330 ### Changes between 1.0.0b and 1.0.0c [2 Dec 2010]
8332 * Disable code workaround for ancient and obsolete Netscape browsers
8333 and servers: an attacker can use it in a ciphersuite downgrade attack.
8334 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
8338 * Fixed J-PAKE implementation error, originally discovered by
8339 Sebastien Martini, further info and confirmation from Stefan
8340 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
8344 ### Changes between 1.0.0a and 1.0.0b [16 Nov 2010]
8346 * Fix extension code to avoid race conditions which can result in a buffer
8347 overrun vulnerability: resumed sessions must not be modified as they can
8348 be shared by multiple threads. CVE-2010-3864
8352 * Fix WIN32 build system to correctly link an ENGINE directory into
8357 ### Changes between 1.0.0 and 1.0.0a [01 Jun 2010]
8359 * Check return value of int_rsa_verify in pkey_rsa_verifyrecover
8362 *Steve Henson, Peter-Michael Hager <hager@dortmund.net>*
8364 ### Changes between 0.9.8n and 1.0.0 [29 Mar 2010]
8366 * Add "missing" function EVP_CIPHER_CTX_copy(). This copies a cipher
8367 context. The operation can be customised via the ctrl mechanism in
8368 case ENGINEs want to include additional functionality.
8372 * Tolerate yet another broken PKCS#8 key format: private key value negative.
8376 * Add new -subject_hash_old and -issuer_hash_old options to x509 utility to
8377 output hashes compatible with older versions of OpenSSL.
8379 *Willy Weisz <weisz@vcpc.univie.ac.at>*
8381 * Fix compression algorithm handling: if resuming a session use the
8382 compression algorithm of the resumed session instead of determining
8383 it from client hello again. Don't allow server to change algorithm.
8387 * Add load_crls() function to commands tidying load_certs() too. Add option
8388 to verify utility to allow additional CRLs to be included.
8392 * Update OCSP request code to permit adding custom headers to the request:
8393 some responders need this.
8397 * The function EVP_PKEY_sign() returns <=0 on error: check return code
8400 *Julia Lawall <julia@diku.dk>*
8402 * Update verify callback code in `apps/s_cb.c` and `apps/verify.c`, it
8403 needlessly dereferenced structures, used obsolete functions and
8404 didn't handle all updated verify codes correctly.
8408 * Disable MD2 in the default configuration.
8412 * In BIO_pop() and BIO_push() use the ctrl argument (which was NULL) to
8413 indicate the initial BIO being pushed or popped. This makes it possible
8414 to determine whether the BIO is the one explicitly called or as a result
8415 of the ctrl being passed down the chain. Fix BIO_pop() and SSL BIOs so
8416 it handles reference counts correctly and doesn't zero out the I/O bio
8417 when it is not being explicitly popped. WARNING: applications which
8418 included workarounds for the old buggy behaviour will need to be modified
8419 or they could free up already freed BIOs.
8423 * Extend the uni2asc/asc2uni => OPENSSL_uni2asc/OPENSSL_asc2uni
8424 renaming to all platforms (within the 0.9.8 branch, this was
8425 done conditionally on Netware platforms to avoid a name clash).
8427 *Guenter <lists@gknw.net>*
8429 * Add ECDHE and PSK support to DTLS.
8431 *Michael Tuexen <tuexen@fh-muenster.de>*
8433 * Add CHECKED_STACK_OF macro to safestack.h, otherwise safestack can't
8438 * Add "missing" function EVP_MD_flags() (without this the only way to
8439 retrieve a digest flags is by accessing the structure directly. Update
8440 `EVP_MD_do_all*()` and `EVP_CIPHER_do_all*()` to include the name a digest
8441 or cipher is registered as in the "from" argument. Print out all
8442 registered digests in the dgst usage message instead of manually
8443 attempting to work them out.
8447 * If no SSLv2 ciphers are used don't use an SSLv2 compatible client hello:
8448 this allows the use of compression and extensions. Change default cipher
8449 string to remove SSLv2 ciphersuites. This effectively avoids ancient SSLv2
8450 by default unless an application cipher string requests it.
8454 * Alter match criteria in PKCS12_parse(). It used to try to use local
8455 key ids to find matching certificates and keys but some PKCS#12 files
8456 don't follow the (somewhat unwritten) rules and this strategy fails.
8457 Now just gather all certificates together and the first private key
8458 then look for the first certificate that matches the key.
8462 * Support use of registered digest and cipher names for dgst and cipher
8463 commands instead of having to add each one as a special case. So now
8470 openssl dgst -sha256 foo
8472 and this works for ENGINE based algorithms too.
8476 * Update Gost ENGINE to support parameter files.
8478 *Victor B. Wagner <vitus@cryptocom.ru>*
8480 * Support GeneralizedTime in ca utility.
8482 *Oliver Martin <oliver@volatilevoid.net>, Steve Henson*
8484 * Enhance the hash format used for certificate directory links. The new
8485 form uses the canonical encoding (meaning equivalent names will work
8486 even if they aren't identical) and uses SHA1 instead of MD5. This form
8487 is incompatible with the older format and as a result c_rehash should
8488 be used to rebuild symbolic links.
8492 * Make PKCS#8 the default write format for private keys, replacing the
8493 traditional format. This form is standardised, more secure and doesn't
8494 include an implicit MD5 dependency.
8498 * Add a $gcc_devteam_warn option to Configure. The idea is that any code
8499 committed to OpenSSL should pass this lot as a minimum.
8503 * Add session ticket override functionality for use by EAP-FAST.
8505 *Jouni Malinen <j@w1.fi>*
8507 * Modify HMAC functions to return a value. Since these can be implemented
8508 in an ENGINE errors can occur.
8512 * Type-checked OBJ_bsearch_ex.
8516 * Type-checked OBJ_bsearch. Also some constification necessitated
8517 by type-checking. Still to come: TXT_DB, bsearch(?),
8518 OBJ_bsearch_ex, qsort, CRYPTO_EX_DATA, ASN1_VALUE, ASN1_STRING,
8523 * New function OPENSSL_gmtime_adj() to add a specific number of days and
8524 seconds to a tm structure directly, instead of going through OS
8525 specific date routines. This avoids any issues with OS routines such
8526 as the year 2038 bug. New `*_adj()` functions for ASN1 time structures
8527 and X509_time_adj_ex() to cover the extended range. The existing
8528 X509_time_adj() is still usable and will no longer have any date issues.
8532 * Delta CRL support. New use deltas option which will attempt to locate
8533 and search any appropriate delta CRLs available.
8535 This work was sponsored by Google.
8539 * Support for CRLs partitioned by reason code. Reorganise CRL processing
8540 code and add additional score elements. Validate alternate CRL paths
8541 as part of the CRL checking and indicate a new error "CRL path validation
8542 error" in this case. Applications wanting additional details can use
8543 the verify callback and check the new "parent" field. If this is not
8544 NULL CRL path validation is taking place. Existing applications won't
8545 see this because it requires extended CRL support which is off by
8548 This work was sponsored by Google.
8552 * Support for freshest CRL extension.
8554 This work was sponsored by Google.
8558 * Initial indirect CRL support. Currently only supported in the CRLs
8559 passed directly and not via lookup. Process certificate issuer
8560 CRL entry extension and lookup CRL entries by bother issuer name
8561 and serial number. Check and process CRL issuer entry in IDP extension.
8563 This work was sponsored by Google.
8567 * Add support for distinct certificate and CRL paths. The CRL issuer
8568 certificate is validated separately in this case. Only enabled if
8569 an extended CRL support flag is set: this flag will enable additional
8570 CRL functionality in future.
8572 This work was sponsored by Google.
8576 * Add support for policy mappings extension.
8578 This work was sponsored by Google.
8582 * Fixes to pathlength constraint, self issued certificate handling,
8583 policy processing to align with RFC3280 and PKITS tests.
8585 This work was sponsored by Google.
8589 * Support for name constraints certificate extension. DN, email, DNS
8590 and URI types are currently supported.
8592 This work was sponsored by Google.
8596 * To cater for systems that provide a pointer-based thread ID rather
8597 than numeric, deprecate the current numeric thread ID mechanism and
8598 replace it with a structure and associated callback type. This
8599 mechanism allows a numeric "hash" to be extracted from a thread ID in
8600 either case, and on platforms where pointers are larger than 'long',
8601 mixing is done to help ensure the numeric 'hash' is usable even if it
8602 can't be guaranteed unique. The default mechanism is to use "&errno"
8603 as a pointer-based thread ID to distinguish between threads.
8605 Applications that want to provide their own thread IDs should now use
8606 CRYPTO_THREADID_set_callback() to register a callback that will call
8607 either CRYPTO_THREADID_set_numeric() or CRYPTO_THREADID_set_pointer().
8609 Note that ERR_remove_state() is now deprecated, because it is tied
8610 to the assumption that thread IDs are numeric. ERR_remove_state(0)
8611 to free the current thread's error state should be replaced by
8612 ERR_remove_thread_state(NULL).
8614 (This new approach replaces the functions CRYPTO_set_idptr_callback(),
8615 CRYPTO_get_idptr_callback(), and CRYPTO_thread_idptr() that existed in
8616 OpenSSL 0.9.9-dev between June 2006 and August 2008. Also, if an
8617 application was previously providing a numeric thread callback that
8618 was inappropriate for distinguishing threads, then uniqueness might
8619 have been obtained with &errno that happened immediately in the
8620 intermediate development versions of OpenSSL; this is no longer the
8621 case, the numeric thread callback will now override the automatic use
8624 *Geoff Thorpe, with help from Bodo Moeller*
8626 * Initial support for different CRL issuing certificates. This covers a
8627 simple case where the self issued certificates in the chain exist and
8628 the real CRL issuer is higher in the existing chain.
8630 This work was sponsored by Google.
8634 * Removed effectively defunct crypto/store from the build.
8638 * Revamp of STACK to provide stronger type-checking. Still to come:
8639 TXT_DB, bsearch(?), OBJ_bsearch, qsort, CRYPTO_EX_DATA, ASN1_VALUE,
8640 ASN1_STRING, CONF_VALUE.
8644 * Add a new SSL_MODE_RELEASE_BUFFERS mode flag to release unused buffer
8645 RAM on SSL connections. This option can save about 34k per idle SSL.
8649 * Revamp of LHASH to provide stronger type-checking. Still to come:
8650 STACK, TXT_DB, bsearch, qsort.
8654 * Initial support for Cryptographic Message Syntax (aka CMS) based
8655 on RFC3850, RFC3851 and RFC3852. New cms directory and cms utility,
8656 support for data, signedData, compressedData, digestedData and
8657 encryptedData, envelopedData types included. Scripts to check against
8658 RFC4134 examples draft and interop and consistency checks of many
8659 content types and variants.
8663 * Add options to enc utility to support use of zlib compression BIO.
8667 * Extend mk1mf to support importing of options and assembly language
8668 files from Configure script, currently only included in VC-WIN32.
8669 The assembly language rules can now optionally generate the source
8670 files from the associated perl scripts.
8674 * Implement remaining functionality needed to support GOST ciphersuites.
8675 Interop testing has been performed using CryptoPro implementations.
8677 *Victor B. Wagner <vitus@cryptocom.ru>*
8679 * s390x assembler pack.
8683 * ARMv4 assembler pack. ARMv4 refers to v4 and later ISA, not CPU
8688 * Implement Opaque PRF Input TLS extension as specified in
8689 draft-rescorla-tls-opaque-prf-input-00.txt. Since this is not an
8690 official specification yet and no extension type assignment by
8691 IANA exists, this extension (for now) will have to be explicitly
8692 enabled when building OpenSSL by providing the extension number
8693 to use. For example, specify an option
8695 -DTLSEXT_TYPE_opaque_prf_input=0x9527
8697 to the "config" or "Configure" script to enable the extension,
8698 assuming extension number 0x9527 (which is a completely arbitrary
8699 and unofficial assignment based on the MD5 hash of the Internet
8700 Draft). Note that by doing so, you potentially lose
8701 interoperability with other TLS implementations since these might
8702 be using the same extension number for other purposes.
8704 SSL_set_tlsext_opaque_prf_input(ssl, src, len) is used to set the
8705 opaque PRF input value to use in the handshake. This will create
8706 an internal copy of the length-'len' string at 'src', and will
8707 return non-zero for success.
8709 To get more control and flexibility, provide a callback function
8712 SSL_CTX_set_tlsext_opaque_prf_input_callback(ctx, cb)
8713 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg(ctx, arg)
8717 int (*cb)(SSL *, void *peerinput, size_t len, void *arg);
8720 Callback function 'cb' will be called in handshakes, and is
8721 expected to use SSL_set_tlsext_opaque_prf_input() as appropriate.
8722 Argument 'arg' is for application purposes (the value as given to
8723 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg() will directly
8724 be provided to the callback function). The callback function
8725 has to return non-zero to report success: usually 1 to use opaque
8726 PRF input just if possible, or 2 to enforce use of the opaque PRF
8727 input. In the latter case, the library will abort the handshake
8728 if opaque PRF input is not successfully negotiated.
8730 Arguments 'peerinput' and 'len' given to the callback function
8731 will always be NULL and 0 in the case of a client. A server will
8732 see the client's opaque PRF input through these variables if
8733 available (NULL and 0 otherwise). Note that if the server
8734 provides an opaque PRF input, the length must be the same as the
8735 length of the client's opaque PRF input.
8737 Note that the callback function will only be called when creating
8738 a new session (session resumption can resume whatever was
8739 previously negotiated), and will not be called in SSL 2.0
8740 handshakes; thus, SSL_CTX_set_options(ctx, SSL_OP_NO_SSLv2) or
8741 SSL_set_options(ssl, SSL_OP_NO_SSLv2) is especially recommended
8742 for applications that need to enforce opaque PRF input.
8746 * Update ssl code to support digests other than SHA1+MD5 for handshake
8749 *Victor B. Wagner <vitus@cryptocom.ru>*
8751 * Add RFC4507 support to OpenSSL. This includes the corrections in
8752 RFC4507bis. The encrypted ticket format is an encrypted encoded
8753 SSL_SESSION structure, that way new session features are automatically
8756 If a client application caches session in an SSL_SESSION structure
8757 support is transparent because tickets are now stored in the encoded
8760 The SSL_CTX structure automatically generates keys for ticket
8761 protection in servers so again support should be possible
8762 with no application modification.
8764 If a client or server wishes to disable RFC4507 support then the option
8765 SSL_OP_NO_TICKET can be set.
8767 Add a TLS extension debugging callback to allow the contents of any client
8768 or server extensions to be examined.
8770 This work was sponsored by Google.
8774 * Final changes to avoid use of pointer pointer casts in OpenSSL.
8775 OpenSSL should now compile cleanly on gcc 4.2
8777 *Peter Hartley <pdh@utter.chaos.org.uk>, Steve Henson*
8779 * Update SSL library to use new EVP_PKEY MAC API. Include generic MAC
8780 support including streaming MAC support: this is required for GOST
8781 ciphersuite support.
8783 *Victor B. Wagner <vitus@cryptocom.ru>, Steve Henson*
8785 * Add option -stream to use PKCS#7 streaming in smime utility. New
8786 function i2d_PKCS7_bio_stream() and PEM_write_PKCS7_bio_stream()
8787 to output in BER and PEM format.
8791 * Experimental support for use of HMAC via EVP_PKEY interface. This
8792 allows HMAC to be handled via the `EVP_DigestSign*()` interface. The
8793 EVP_PKEY "key" in this case is the HMAC key, potentially allowing
8794 ENGINE support for HMAC keys which are unextractable. New -mac and
8795 -macopt options to dgst utility.
8799 * New option -sigopt to dgst utility. Update dgst to use
8800 `EVP_Digest{Sign,Verify}*`. These two changes make it possible to use
8801 alternative signing parameters such as X9.31 or PSS in the dgst
8806 * Change ssl_cipher_apply_rule(), the internal function that does
8807 the work each time a ciphersuite string requests enabling
8808 ("foo+bar"), moving ("+foo+bar"), disabling ("-foo+bar", or
8809 removing ("!foo+bar") a class of ciphersuites: Now it maintains
8810 the order of disabled ciphersuites such that those ciphersuites
8811 that most recently went from enabled to disabled not only stay
8812 in order with respect to each other, but also have higher priority
8813 than other disabled ciphersuites the next time ciphersuites are
8816 This means that you can now say, e.g., "PSK:-PSK:HIGH" to enable
8817 the same ciphersuites as with "HIGH" alone, but in a specific
8818 order where the PSK ciphersuites come first (since they are the
8819 most recently disabled ciphersuites when "HIGH" is parsed).
8821 Also, change ssl_create_cipher_list() (using this new
8822 functionality) such that between otherwise identical
8823 ciphersuites, ephemeral ECDH is preferred over ephemeral DH in
8828 * Change ssl_create_cipher_list() so that it automatically
8829 arranges the ciphersuites in reasonable order before starting
8830 to process the rule string. Thus, the definition for "DEFAULT"
8831 (SSL_DEFAULT_CIPHER_LIST) now is just "ALL:!aNULL:!eNULL", but
8832 remains equivalent to `"AES:ALL:!aNULL:!eNULL:+aECDH:+kRSA:+RC4:@STRENGTH"`.
8833 This makes it much easier to arrive at a reasonable default order
8834 in applications for which anonymous ciphers are OK (meaning
8835 that you can't actually use DEFAULT).
8837 *Bodo Moeller; suggested by Victor Duchovni*
8839 * Split the SSL/TLS algorithm mask (as used for ciphersuite string
8840 processing) into multiple integers instead of setting
8841 "SSL_MKEY_MASK" bits, "SSL_AUTH_MASK" bits, "SSL_ENC_MASK",
8842 "SSL_MAC_MASK", and "SSL_SSL_MASK" bits all in a single integer.
8843 (These masks as well as the individual bit definitions are hidden
8844 away into the non-exported interface ssl/ssl_locl.h, so this
8845 change to the definition of the SSL_CIPHER structure shouldn't
8846 affect applications.) This give us more bits for each of these
8847 categories, so there is no longer a need to coagulate AES128 and
8848 AES256 into a single algorithm bit, and to coagulate Camellia128
8849 and Camellia256 into a single algorithm bit, which has led to all
8852 Thus, among other things, the kludge introduced in 0.9.7m and
8853 0.9.8e for masking out AES256 independently of AES128 or masking
8854 out Camellia256 independently of AES256 is not needed here in 0.9.9.
8856 With the change, we also introduce new ciphersuite aliases that
8857 so far were missing: "AES128", "AES256", "CAMELLIA128", and
8862 * Add support for dsa-with-SHA224 and dsa-with-SHA256.
8863 Use the leftmost N bytes of the signature input if the input is
8864 larger than the prime q (with N being the size in bytes of q).
8868 * Very *very* experimental PKCS#7 streaming encoder support. Nothing uses
8869 it yet and it is largely untested.
8873 * Add support for the ecdsa-with-SHA224/256/384/512 signature types.
8877 * Initial incomplete changes to avoid need for function casts in OpenSSL
8878 some compilers (gcc 4.2 and later) reject their use. Safestack is
8879 reimplemented. Update ASN1 to avoid use of legacy functions.
8883 * Win32/64 targets are linked with Winsock2.
8887 * Add an X509_CRL_METHOD structure to allow CRL processing to be redirected
8888 to external functions. This can be used to increase CRL handling
8889 efficiency especially when CRLs are very large by (for example) storing
8890 the CRL revoked certificates in a database.
8894 * Overhaul of by_dir code. Add support for dynamic loading of CRLs so
8895 new CRLs added to a directory can be used. New command line option
8896 -verify_return_error to s_client and s_server. This causes real errors
8897 to be returned by the verify callback instead of carrying on no matter
8898 what. This reflects the way a "real world" verify callback would behave.
8902 * GOST engine, supporting several GOST algorithms and public key formats.
8903 Kindly donated by Cryptocom.
8907 * Partial support for Issuing Distribution Point CRL extension. CRLs
8908 partitioned by DP are handled but no indirect CRL or reason partitioning
8909 (yet). Complete overhaul of CRL handling: now the most suitable CRL is
8910 selected via a scoring technique which handles IDP and AKID in CRLs.
8914 * New X509_STORE_CTX callbacks lookup_crls() and lookup_certs() which
8915 will ultimately be used for all verify operations: this will remove the
8916 X509_STORE dependency on certificate verification and allow alternative
8917 lookup methods. X509_STORE based implementations of these two callbacks.
8921 * Allow multiple CRLs to exist in an X509_STORE with matching issuer names.
8922 Modify get_crl() to find a valid (unexpired) CRL if possible.
8926 * New function X509_CRL_match() to check if two CRLs are identical. Normally
8927 this would be called X509_CRL_cmp() but that name is already used by
8928 a function that just compares CRL issuer names. Cache several CRL
8929 extensions in X509_CRL structure and cache CRLDP in X509.
8933 * Store a "canonical" representation of X509_NAME structure (ASN1 Name)
8934 this maps equivalent X509_NAME structures into a consistent structure.
8935 Name comparison can then be performed rapidly using memcmp().
8939 * Non-blocking OCSP request processing. Add -timeout option to ocsp
8944 * Allow digests to supply their own micalg string for S/MIME type using
8945 the ctrl EVP_MD_CTRL_MICALG.
8949 * During PKCS7 signing pass the PKCS7 SignerInfo structure to the
8950 EVP_PKEY_METHOD before and after signing via the EVP_PKEY_CTRL_PKCS7_SIGN
8951 ctrl. It can then customise the structure before and/or after signing
8956 * New function OBJ_add_sigid() to allow application defined signature OIDs
8957 to be added to OpenSSLs internal tables. New function OBJ_sigid_free()
8958 to free up any added signature OIDs.
8962 * New functions EVP_CIPHER_do_all(), EVP_CIPHER_do_all_sorted(),
8963 EVP_MD_do_all() and EVP_MD_do_all_sorted() to enumerate internal
8964 digest and cipher tables. New options added to openssl utility:
8965 list-message-digest-algorithms and list-cipher-algorithms.
8969 * Change the array representation of binary polynomials: the list
8970 of degrees of non-zero coefficients is now terminated with -1.
8971 Previously it was terminated with 0, which was also part of the
8972 value; thus, the array representation was not applicable to
8973 polynomials where t^0 has coefficient zero. This change makes
8974 the array representation useful in a more general context.
8978 * Various modifications and fixes to SSL/TLS cipher string
8979 handling. For ECC, the code now distinguishes between fixed ECDH
8980 with RSA certificates on the one hand and with ECDSA certificates
8981 on the other hand, since these are separate ciphersuites. The
8982 unused code for Fortezza ciphersuites has been removed.
8984 For consistency with EDH, ephemeral ECDH is now called "EECDH"
8985 (not "ECDHE"). For consistency with the code for DH
8986 certificates, use of ECDH certificates is now considered ECDH
8987 authentication, not RSA or ECDSA authentication (the latter is
8988 merely the CA's signing algorithm and not actively used in the
8991 The temporary ciphersuite alias "ECCdraft" is no longer
8992 available, and ECC ciphersuites are no longer excluded from "ALL"
8993 and "DEFAULT". The following aliases now exist for RFC 4492
8994 ciphersuites, most of these by analogy with the DH case:
8996 kECDHr - ECDH cert, signed with RSA
8997 kECDHe - ECDH cert, signed with ECDSA
8998 kECDH - ECDH cert (signed with either RSA or ECDSA)
8999 kEECDH - ephemeral ECDH
9000 ECDH - ECDH cert or ephemeral ECDH
9006 AECDH - anonymous ECDH
9007 EECDH - non-anonymous ephemeral ECDH (equivalent to "kEECDH:-AECDH")
9011 * Add additional S/MIME capabilities for AES and GOST ciphers if supported.
9012 Use correct micalg parameters depending on digest(s) in signed message.
9016 * Add engine support for EVP_PKEY_ASN1_METHOD. Add functions to process
9017 an ENGINE asn1 method. Support ENGINE lookups in the ASN1 code.
9021 * Initial engine support for EVP_PKEY_METHOD. New functions to permit
9022 an engine to register a method. Add ENGINE lookups for methods and
9023 functional reference processing.
9027 * New functions `EVP_Digest{Sign,Verify)*`. These are enhanced versions of
9028 `EVP_{Sign,Verify}*` which allow an application to customise the signature
9033 * New -resign option to smime utility. This adds one or more signers
9034 to an existing PKCS#7 signedData structure. Also -md option to use an
9035 alternative message digest algorithm for signing.
9039 * Tidy up PKCS#7 routines and add new functions to make it easier to
9040 create PKCS7 structures containing multiple signers. Update smime
9041 application to support multiple signers.
9045 * New -macalg option to pkcs12 utility to allow setting of an alternative
9050 * Initial support for PKCS#5 v2.0 PRFs other than default SHA1 HMAC.
9051 Reorganize PBE internals to lookup from a static table using NIDs,
9052 add support for HMAC PBE OID translation. Add a EVP_CIPHER ctrl:
9053 EVP_CTRL_PBE_PRF_NID this allows a cipher to specify an alternative
9054 PRF which will be automatically used with PBES2.
9058 * Replace the algorithm specific calls to generate keys in "req" with the
9063 * Update PKCS#7 enveloped data routines to use new API. This is now
9064 supported by any public key method supporting the encrypt operation. A
9065 ctrl is added to allow the public key algorithm to examine or modify
9066 the PKCS#7 RecipientInfo structure if it needs to: for RSA this is
9071 * Add a ctrl to asn1 method to allow a public key algorithm to express
9072 a default digest type to use. In most cases this will be SHA1 but some
9073 algorithms (such as GOST) need to specify an alternative digest. The
9074 return value indicates how strong the preference is 1 means optional and
9075 2 is mandatory (that is it is the only supported type). Modify
9076 ASN1_item_sign() to accept a NULL digest argument to indicate it should
9077 use the default md. Update openssl utilities to use the default digest
9078 type for signing if it is not explicitly indicated.
9082 * Use OID cross reference table in ASN1_sign() and ASN1_verify(). New
9083 EVP_MD flag EVP_MD_FLAG_PKEY_METHOD_SIGNATURE. This uses the relevant
9084 signing method from the key type. This effectively removes the link
9085 between digests and public key types.
9089 * Add an OID cross reference table and utility functions. Its purpose is to
9090 translate between signature OIDs such as SHA1WithrsaEncryption and SHA1,
9091 rsaEncryption. This will allow some of the algorithm specific hackery
9092 needed to use the correct OID to be removed.
9096 * Remove algorithm specific dependencies when setting PKCS7_SIGNER_INFO
9097 structures for PKCS7_sign(). They are now set up by the relevant public
9102 * Add provisional EC pkey method with support for ECDSA and ECDH.
9106 * Add support for key derivation (agreement) in the API, DH method and
9111 * Add DSA pkey method and DH pkey methods, extend DH ASN1 method to support
9112 public and private key formats. As a side effect these add additional
9113 command line functionality not previously available: DSA signatures can be
9114 generated and verified using pkeyutl and DH key support and generation in
9121 *Oliver Tappe <zooey@hirschkaefer.de>*
9123 * New make target "install_html_docs" installs HTML renditions of the
9126 *Oliver Tappe <zooey@hirschkaefer.de>*
9128 * New utility "genpkey" this is analogous to "genrsa" etc except it can
9129 generate keys for any algorithm. Extend and update EVP_PKEY_METHOD to
9130 support key and parameter generation and add initial key generation
9131 functionality for RSA.
9135 * Add functions for main EVP_PKEY_method operations. The undocumented
9136 functions `EVP_PKEY_{encrypt,decrypt}` have been renamed to
9137 `EVP_PKEY_{encrypt,decrypt}_old`.
9141 * Initial definitions for EVP_PKEY_METHOD. This will be a high level public
9142 key API, doesn't do much yet.
9146 * New function EVP_PKEY_asn1_get0_info() to retrieve information about
9147 public key algorithms. New option to openssl utility:
9148 "list-public-key-algorithms" to print out info.
9152 * Implement the Supported Elliptic Curves Extension for
9153 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
9157 * Don't free up OIDs in OBJ_cleanup() if they are in use by EVP_MD or
9158 EVP_CIPHER structures to avoid later problems in EVP_cleanup().
9162 * New utilities pkey and pkeyparam. These are similar to algorithm specific
9163 utilities such as rsa, dsa, dsaparam etc except they process any key
9168 * Transfer public key printing routines to EVP_PKEY_ASN1_METHOD. New
9169 functions EVP_PKEY_print_public(), EVP_PKEY_print_private(),
9170 EVP_PKEY_print_param() to print public key data from an EVP_PKEY
9175 * Initial support for pluggable public key ASN1.
9176 De-spaghettify the public key ASN1 handling. Move public and private
9177 key ASN1 handling to a new EVP_PKEY_ASN1_METHOD structure. Relocate
9178 algorithm specific handling to a single module within the relevant
9179 algorithm directory. Add functions to allow (near) opaque processing
9180 of public and private key structures.
9184 * Implement the Supported Point Formats Extension for
9185 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
9189 * Add initial support for RFC 4279 PSK TLS ciphersuites. Add members
9190 for the psk identity [hint] and the psk callback functions to the
9191 SSL_SESSION, SSL and SSL_CTX structure.
9194 PSK-RC4-SHA, PSK-3DES-EDE-CBC-SHA, PSK-AES128-CBC-SHA,
9198 SSL_CTX_use_psk_identity_hint
9199 SSL_get_psk_identity_hint
9200 SSL_get_psk_identity
9201 SSL_use_psk_identity_hint
9203 *Mika Kousa and Pasi Eronen of Nokia Corporation*
9205 * Add RFC 3161 compliant time stamp request creation, response generation
9206 and response verification functionality.
9208 *Zoltán Glózik <zglozik@opentsa.org>, The OpenTSA Project*
9210 * Add initial support for TLS extensions, specifically for the server_name
9211 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
9212 have new members for a host name. The SSL data structure has an
9213 additional member `SSL_CTX *initial_ctx` so that new sessions can be
9214 stored in that context to allow for session resumption, even after the
9215 SSL has been switched to a new SSL_CTX in reaction to a client's
9216 server_name extension.
9218 New functions (subject to change):
9220 SSL_get_servername()
9221 SSL_get_servername_type()
9224 New CTRL codes and macros (subject to change):
9226 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
9227 - SSL_CTX_set_tlsext_servername_callback()
9228 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
9229 - SSL_CTX_set_tlsext_servername_arg()
9230 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
9232 openssl s_client has a new '-servername ...' option.
9234 openssl s_server has new options '-servername_host ...', '-cert2 ...',
9235 '-key2 ...', '-servername_fatal' (subject to change). This allows
9236 testing the HostName extension for a specific single host name ('-cert'
9237 and '-key' remain fallbacks for handshakes without HostName
9238 negotiation). If the unrecognized_name alert has to be sent, this by
9239 default is a warning; it becomes fatal with the '-servername_fatal'
9242 *Peter Sylvester, Remy Allais, Christophe Renou*
9244 * Whirlpool hash implementation is added.
9248 * BIGNUM code on 64-bit SPARCv9 targets is switched from bn(64,64) to
9249 bn(64,32). Because of instruction set limitations it doesn't have
9250 any negative impact on performance. This was done mostly in order
9251 to make it possible to share assembler modules, such as bn_mul_mont
9252 implementations, between 32- and 64-bit builds without hassle.
9256 * Move code previously exiled into file crypto/ec/ec2_smpt.c
9257 to ec2_smpl.c, and no longer require the OPENSSL_EC_BIN_PT_COMP
9262 * New candidate for BIGNUM assembler implementation, bn_mul_mont,
9263 dedicated Montgomery multiplication procedure, is introduced.
9264 BN_MONT_CTX is modified to allow bn_mul_mont to reach for higher
9265 "64-bit" performance on certain 32-bit targets.
9269 * New option SSL_OP_NO_COMP to disable use of compression selectively
9270 in SSL structures. New SSL ctrl to set maximum send fragment size.
9271 Save memory by setting the I/O buffer sizes dynamically instead of
9272 using the maximum available value.
9276 * New option -V for 'openssl ciphers'. This prints the ciphersuite code
9277 in addition to the text details.
9281 * Very, very preliminary EXPERIMENTAL support for printing of general
9282 ASN1 structures. This currently produces rather ugly output and doesn't
9283 handle several customised structures at all.
9287 * Integrated support for PVK file format and some related formats such
9288 as MS PUBLICKEYBLOB and PRIVATEKEYBLOB. Command line switches to support
9289 these in the 'rsa' and 'dsa' utilities.
9293 * Support for PKCS#1 RSAPublicKey format on rsa utility command line.
9297 * Remove the ancient ASN1_METHOD code. This was only ever used in one
9298 place for the (very old) "NETSCAPE" format certificates which are now
9299 handled using new ASN1 code equivalents.
9303 * Let the TLSv1_method() etc. functions return a 'const' SSL_METHOD
9304 pointer and make the SSL_METHOD parameter in SSL_CTX_new,
9305 SSL_CTX_set_ssl_version and SSL_set_ssl_method 'const'.
9309 * Modify CRL distribution points extension code to print out previously
9310 unsupported fields. Enhance extension setting code to allow setting of
9315 * Add print and set support for Issuing Distribution Point CRL extension.
9319 * Change 'Configure' script to enable Camellia by default.
9326 ### Changes between 0.9.8m and 0.9.8n [24 Mar 2010]
9328 * When rejecting SSL/TLS records due to an incorrect version number, never
9329 update s->server with a new major version number. As of
9330 - OpenSSL 0.9.8m if 'short' is a 16-bit type,
9331 - OpenSSL 0.9.8f if 'short' is longer than 16 bits,
9332 the previous behavior could result in a read attempt at NULL when
9333 receiving specific incorrect SSL/TLS records once record payload
9334 protection is active. ([CVE-2010-0740])
9336 *Bodo Moeller, Adam Langley <agl@chromium.org>*
9338 * Fix for CVE-2010-0433 where some kerberos enabled versions of OpenSSL
9339 could be crashed if the relevant tables were not present (e.g. chrooted).
9341 *Tomas Hoger <thoger@redhat.com>*
9343 ### Changes between 0.9.8l and 0.9.8m [25 Feb 2010]
9345 * Always check bn_wexpand() return values for failure. ([CVE-2009-3245])
9347 *Martin Olsson, Neel Mehta*
9349 * Fix X509_STORE locking: Every 'objs' access requires a lock (to
9350 accommodate for stack sorting, always a write lock!).
9354 * On some versions of WIN32 Heap32Next is very slow. This can cause
9355 excessive delays in the RAND_poll(): over a minute. As a workaround
9356 include a time check in the inner Heap32Next loop too.
9360 * The code that handled flushing of data in SSL/TLS originally used the
9361 BIO_CTRL_INFO ctrl to see if any data was pending first. This caused
9362 the problem outlined in PR#1949. The fix suggested there however can
9363 trigger problems with buggy BIO_CTRL_WPENDING (e.g. some versions
9364 of Apache). So instead simplify the code to flush unconditionally.
9365 This should be fine since flushing with no data to flush is a no op.
9369 * Handle TLS versions 2.0 and later properly and correctly use the
9370 highest version of TLS/SSL supported. Although TLS >= 2.0 is some way
9371 off ancient servers have a habit of sticking around for a while...
9375 * Modify compression code so it frees up structures without using the
9376 ex_data callbacks. This works around a problem where some applications
9377 call CRYPTO_cleanup_all_ex_data() before application exit (e.g. when
9378 restarting) then use compression (e.g. SSL with compression) later.
9379 This results in significant per-connection memory leaks and
9380 has caused some security issues including CVE-2008-1678 and
9385 * Constify crypto/cast (i.e., <openssl/cast.h>): a CAST_KEY doesn't
9386 change when encrypting or decrypting.
9390 * Add option SSL_OP_LEGACY_SERVER_CONNECT which will allow clients to
9391 connect and renegotiate with servers which do not support RI.
9392 Until RI is more widely deployed this option is enabled by default.
9396 * Add "missing" ssl ctrls to clear options and mode.
9400 * If client attempts to renegotiate and doesn't support RI respond with
9401 a no_renegotiation alert as required by RFC5746. Some renegotiating
9402 TLS clients will continue a connection gracefully when they receive
9403 the alert. Unfortunately OpenSSL mishandled this alert and would hang
9404 waiting for a server hello which it will never receive. Now we treat a
9405 received no_renegotiation alert as a fatal error. This is because
9406 applications requesting a renegotiation might well expect it to succeed
9407 and would have no code in place to handle the server denying it so the
9408 only safe thing to do is to terminate the connection.
9412 * Add ctrl macro SSL_get_secure_renegotiation_support() which returns 1 if
9413 peer supports secure renegotiation and 0 otherwise. Print out peer
9414 renegotiation support in s_client/s_server.
9418 * Replace the highly broken and deprecated SPKAC certification method with
9419 the updated NID creation version. This should correctly handle UTF8.
9423 * Implement RFC5746. Re-enable renegotiation but require the extension
9424 as needed. Unfortunately, SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION
9425 turns out to be a bad idea. It has been replaced by
9426 SSL_OP_ALLOW_UNSAFE_LEGACY_RENEGOTIATION which can be set with
9427 SSL_CTX_set_options(). This is really not recommended unless you
9428 know what you are doing.
9430 *Eric Rescorla <ekr@networkresonance.com>, Ben Laurie, Steve Henson*
9432 * Fixes to stateless session resumption handling. Use initial_ctx when
9433 issuing and attempting to decrypt tickets in case it has changed during
9434 servername handling. Use a non-zero length session ID when attempting
9435 stateless session resumption: this makes it possible to determine if
9436 a resumption has occurred immediately after receiving server hello
9437 (several places in OpenSSL subtly assume this) instead of later in
9442 * The functions ENGINE_ctrl(), OPENSSL_isservice(),
9443 CMS_get1_RecipientRequest() and RAND_bytes() can return <=0 on error
9444 fixes for a few places where the return code is not checked
9447 *Julia Lawall <julia@diku.dk>*
9449 * Add --strict-warnings option to Configure script to include devteam
9450 warnings in other configurations.
9454 * Add support for --libdir option and LIBDIR variable in makefiles. This
9455 makes it possible to install openssl libraries in locations which
9456 have names other than "lib", for example "/usr/lib64" which some
9459 *Steve Henson, based on patch from Jeremy Utley*
9461 * Don't allow the use of leading 0x80 in OIDs. This is a violation of
9462 X690 8.9.12 and can produce some misleading textual output of OIDs.
9464 *Steve Henson, reported by Dan Kaminsky*
9466 * Delete MD2 from algorithm tables. This follows the recommendation in
9467 several standards that it is not used in new applications due to
9468 several cryptographic weaknesses. For binary compatibility reasons
9469 the MD2 API is still compiled in by default.
9473 * Add compression id to {d2i,i2d}_SSL_SESSION so it is correctly saved
9478 * Rename uni2asc and asc2uni functions to OPENSSL_uni2asc and
9479 OPENSSL_asc2uni conditionally on Netware platforms to avoid a name
9482 *Guenter <lists@gknw.net>*
9484 * Fix the server certificate chain building code to use X509_verify_cert(),
9485 it used to have an ad-hoc builder which was unable to cope with anything
9486 other than a simple chain.
9488 *David Woodhouse <dwmw2@infradead.org>, Steve Henson*
9490 * Don't check self signed certificate signatures in X509_verify_cert()
9491 by default (a flag can override this): it just wastes time without
9492 adding any security. As a useful side effect self signed root CAs
9493 with non-FIPS digests are now usable in FIPS mode.
9497 * In dtls1_process_out_of_seq_message() the check if the current message
9498 is already buffered was missing. For every new message was memory
9499 allocated, allowing an attacker to perform an denial of service attack
9500 with sending out of seq handshake messages until there is no memory
9501 left. Additionally every future message was buffered, even if the
9502 sequence number made no sense and would be part of another handshake.
9503 So only messages with sequence numbers less than 10 in advance will be
9504 buffered. ([CVE-2009-1378])
9506 *Robin Seggelmann, discovered by Daniel Mentz*
9508 * Records are buffered if they arrive with a future epoch to be
9509 processed after finishing the corresponding handshake. There is
9510 currently no limitation to this buffer allowing an attacker to perform
9511 a DOS attack with sending records with future epochs until there is no
9512 memory left. This patch adds the pqueue_size() function to determine
9513 the size of a buffer and limits the record buffer to 100 entries.
9516 *Robin Seggelmann, discovered by Daniel Mentz*
9518 * Keep a copy of frag->msg_header.frag_len so it can be used after the
9519 parent structure is freed. ([CVE-2009-1379])
9523 * Handle non-blocking I/O properly in SSL_shutdown() call.
9525 *Darryl Miles <darryl-mailinglists@netbauds.net>*
9527 * Add `2.5.4.*` OIDs
9529 *Ilya O. <vrghost@gmail.com>*
9531 ### Changes between 0.9.8k and 0.9.8l [5 Nov 2009]
9533 * Disable renegotiation completely - this fixes a severe security
9534 problem ([CVE-2009-3555]) at the cost of breaking all
9535 renegotiation. Renegotiation can be re-enabled by setting
9536 SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION in s3->flags at
9537 run-time. This is really not recommended unless you know what
9542 ### Changes between 0.9.8j and 0.9.8k [25 Mar 2009]
9544 * Don't set val to NULL when freeing up structures, it is freed up by
9545 underlying code. If `sizeof(void *) > sizeof(long)` this can result in
9546 zeroing past the valid field. ([CVE-2009-0789])
9548 *Paolo Ganci <Paolo.Ganci@AdNovum.CH>*
9550 * Fix bug where return value of CMS_SignerInfo_verify_content() was not
9551 checked correctly. This would allow some invalid signed attributes to
9552 appear to verify correctly. ([CVE-2009-0591])
9554 *Ivan Nestlerode <inestlerode@us.ibm.com>*
9556 * Reject UniversalString and BMPString types with invalid lengths. This
9557 prevents a crash in ASN1_STRING_print_ex() which assumes the strings have
9558 a legal length. ([CVE-2009-0590])
9562 * Set S/MIME signing as the default purpose rather than setting it
9563 unconditionally. This allows applications to override it at the store
9568 * Permit restricted recursion of ASN1 strings. This is needed in practice
9569 to handle some structures.
9573 * Improve efficiency of mem_gets: don't search whole buffer each time
9576 *Jeremy Shapiro <jnshapir@us.ibm.com>*
9578 * New -hex option for openssl rand.
9582 * Print out UTF8String and NumericString when parsing ASN1.
9586 * Support NumericString type for name components.
9590 * Allow CC in the environment to override the automatically chosen
9591 compiler. Note that nothing is done to ensure flags work with the
9596 ### Changes between 0.9.8i and 0.9.8j [07 Jan 2009]
9598 * Properly check EVP_VerifyFinal() and similar return values
9601 *Ben Laurie, Bodo Moeller, Google Security Team*
9603 * Enable TLS extensions by default.
9607 * Allow the CHIL engine to be loaded, whether the application is
9608 multithreaded or not. (This does not release the developer from the
9609 obligation to set up the dynamic locking callbacks.)
9611 *Sander Temme <sander@temme.net>*
9613 * Use correct exit code if there is an error in dgst command.
9615 *Steve Henson; problem pointed out by Roland Dirlewanger*
9617 * Tweak Configure so that you need to say "experimental-jpake" to enable
9618 JPAKE, and need to use -DOPENSSL_EXPERIMENTAL_JPAKE in applications.
9622 * Add experimental JPAKE support, including demo authentication in
9623 s_client and s_server.
9627 * Set the comparison function in v3_addr_canonize().
9629 *Rob Austein <sra@hactrn.net>*
9631 * Add support for XMPP STARTTLS in s_client.
9633 *Philip Paeps <philip@freebsd.org>*
9635 * Change the server-side SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG behavior
9636 to ensure that even with this option, only ciphersuites in the
9637 server's preference list will be accepted. (Note that the option
9638 applies only when resuming a session, so the earlier behavior was
9639 just about the algorithm choice for symmetric cryptography.)
9643 ### Changes between 0.9.8h and 0.9.8i [15 Sep 2008]
9645 * Fix NULL pointer dereference if a DTLS server received
9646 ChangeCipherSpec as first record ([CVE-2009-1386]).
9650 * Fix a state transition in s3_srvr.c and d1_srvr.c
9651 (was using SSL3_ST_CW_CLNT_HELLO_B, should be `..._ST_SW_SRVR_...`).
9655 * The fix in 0.9.8c that supposedly got rid of unsafe
9656 double-checked locking was incomplete for RSA blinding,
9657 addressing just one layer of what turns out to have been
9658 doubly unsafe triple-checked locking.
9660 So now fix this for real by retiring the MONT_HELPER macro
9661 in crypto/rsa/rsa_eay.c.
9663 *Bodo Moeller; problem pointed out by Marius Schilder*
9665 * Various precautionary measures:
9667 - Avoid size_t integer overflow in HASH_UPDATE (md32_common.h).
9669 - Avoid a buffer overflow in d2i_SSL_SESSION() (ssl_asn1.c).
9670 (NB: This would require knowledge of the secret session ticket key
9671 to exploit, in which case you'd be SOL either way.)
9673 - Change bn_nist.c so that it will properly handle input BIGNUMs
9674 outside the expected range.
9676 - Enforce the 'num' check in BN_div() (bn_div.c) for non-BN_DEBUG
9679 *Neel Mehta, Bodo Moeller*
9681 * Allow engines to be "soft loaded" - i.e. optionally don't die if
9682 the load fails. Useful for distros.
9684 *Ben Laurie and the FreeBSD team*
9686 * Add support for Local Machine Keyset attribute in PKCS#12 files.
9690 * Fix BN_GF2m_mod_arr() top-bit cleanup code.
9694 * Expand ENGINE to support engine supplied SSL client certificate functions.
9696 This work was sponsored by Logica.
9700 * Add CryptoAPI ENGINE to support use of RSA and DSA keys held in Windows
9701 keystores. Support for SSL/TLS client authentication too.
9702 Not compiled unless enable-capieng specified to Configure.
9704 This work was sponsored by Logica.
9708 * Fix bug in X509_ATTRIBUTE creation: don't set attribute using
9709 ASN1_TYPE_set1 if MBSTRING flag set. This bug would crash certain
9710 attribute creation routines such as certificate requests and PKCS#12
9715 ### Changes between 0.9.8g and 0.9.8h [28 May 2008]
9717 * Fix flaw if 'Server Key exchange message' is omitted from a TLS
9718 handshake which could lead to a client crash as found using the
9719 Codenomicon TLS test suite ([CVE-2008-1672])
9721 *Steve Henson, Mark Cox*
9723 * Fix double free in TLS server name extensions which could lead to
9724 a remote crash found by Codenomicon TLS test suite ([CVE-2008-0891])
9728 * Clear error queue in SSL_CTX_use_certificate_chain_file()
9730 Clear the error queue to ensure that error entries left from
9731 older function calls do not interfere with the correct operation.
9733 *Lutz Jaenicke, Erik de Castro Lopo*
9735 * Remove root CA certificates of commercial CAs:
9737 The OpenSSL project does not recommend any specific CA and does not
9738 have any policy with respect to including or excluding any CA.
9739 Therefore it does not make any sense to ship an arbitrary selection
9740 of root CA certificates with the OpenSSL software.
9744 * RSA OAEP patches to fix two separate invalid memory reads.
9745 The first one involves inputs when 'lzero' is greater than
9746 'SHA_DIGEST_LENGTH' (it would read about SHA_DIGEST_LENGTH bytes
9747 before the beginning of from). The second one involves inputs where
9748 the 'db' section contains nothing but zeroes (there is a one-byte
9749 invalid read after the end of 'db').
9751 *Ivan Nestlerode <inestlerode@us.ibm.com>*
9753 * Partial backport from 0.9.9-dev:
9755 Introduce bn_mul_mont (dedicated Montgomery multiplication
9756 procedure) as a candidate for BIGNUM assembler implementation.
9757 While 0.9.9-dev uses assembler for various architectures, only
9758 x86_64 is available by default here in the 0.9.8 branch, and
9759 32-bit x86 is available through a compile-time setting.
9761 To try the 32-bit x86 assembler implementation, use Configure
9762 option "enable-montasm" (which exists only for this backport).
9764 As "enable-montasm" for 32-bit x86 disclaims code stability
9765 anyway, in this constellation we activate additional code
9766 backported from 0.9.9-dev for further performance improvements,
9767 namely BN_from_montgomery_word. (To enable this otherwise,
9768 e.g. x86_64, try `-DMONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD`.)
9770 *Andy Polyakov (backport partially by Bodo Moeller)*
9772 * Add TLS session ticket callback. This allows an application to set
9773 TLS ticket cipher and HMAC keys rather than relying on hardcoded fixed
9774 values. This is useful for key rollover for example where several key
9775 sets may exist with different names.
9779 * Reverse ENGINE-internal logic for caching default ENGINE handles.
9780 This was broken until now in 0.9.8 releases, such that the only way
9781 a registered ENGINE could be used (assuming it initialises
9782 successfully on the host) was to explicitly set it as the default
9783 for the relevant algorithms. This is in contradiction with 0.9.7
9784 behaviour and the documentation. With this fix, when an ENGINE is
9785 registered into a given algorithm's table of implementations, the
9786 'uptodate' flag is reset so that auto-discovery will be used next
9787 time a new context for that algorithm attempts to select an
9790 *Ian Lister (tweaked by Geoff Thorpe)*
9792 * Backport of CMS code to OpenSSL 0.9.8. This differs from the 0.9.9
9793 implementation in the following ways:
9795 Lack of EVP_PKEY_ASN1_METHOD means algorithm parameters have to be
9798 Lack of BER streaming support means one pass streaming processing is
9799 only supported if data is detached: setting the streaming flag is
9800 ignored for embedded content.
9802 CMS support is disabled by default and must be explicitly enabled
9803 with the enable-cms configuration option.
9807 * Update the GMP engine glue to do direct copies between BIGNUM and
9808 mpz_t when openssl and GMP use the same limb size. Otherwise the
9809 existing "conversion via a text string export" trick is still used.
9811 *Paul Sheer <paulsheer@gmail.com>*
9813 * Zlib compression BIO. This is a filter BIO which compressed and
9814 uncompresses any data passed through it.
9818 * Add AES_wrap_key() and AES_unwrap_key() functions to implement
9819 RFC3394 compatible AES key wrapping.
9823 * Add utility functions to handle ASN1 structures. ASN1_STRING_set0():
9824 sets string data without copying. X509_ALGOR_set0() and
9825 X509_ALGOR_get0(): set and retrieve X509_ALGOR (AlgorithmIdentifier)
9826 data. Attribute function X509at_get0_data_by_OBJ(): retrieves data
9827 from an X509_ATTRIBUTE structure optionally checking it occurs only
9828 once. ASN1_TYPE_set1(): set and ASN1_TYPE structure copying supplied
9833 * Fix BN flag handling in RSA_eay_mod_exp() and BN_MONT_CTX_set()
9834 to get the expected BN_FLG_CONSTTIME behavior.
9836 *Bodo Moeller (Google)*
9840 - fixed wrong usage of ioctlsocket() when build for LIBC BSD sockets
9841 - fixed do_tests.pl to run the test suite with CLIB builds too (CLIB_OPT)
9842 - added some more tests to do_tests.pl
9843 - fixed RunningProcess usage so that it works with newer LIBC NDKs too
9844 - removed usage of BN_LLONG for CLIB builds to avoid runtime dependency
9845 - added new Configure targets netware-clib-bsdsock, netware-clib-gcc,
9846 netware-clib-bsdsock-gcc, netware-libc-bsdsock-gcc
9847 - various changes to netware.pl to enable gcc-cross builds on Win32
9849 - changed crypto/bio/b_sock.c to work with macro functions (CLIB BSD)
9850 - various changes to fix missing prototype warnings
9851 - fixed x86nasm.pl to create correct asm files for NASM COFF output
9852 - added AES, WHIRLPOOL and CPUID assembler code to build files
9853 - added missing AES assembler make rules to mk1mf.pl
9854 - fixed order of includes in `apps/ocsp.c` so that `e_os.h` settings apply
9856 *Guenter Knauf <eflash@gmx.net>*
9858 * Implement certificate status request TLS extension defined in RFC3546.
9859 A client can set the appropriate parameters and receive the encoded
9860 OCSP response via a callback. A server can query the supplied parameters
9861 and set the encoded OCSP response in the callback. Add simplified examples
9862 to s_client and s_server.
9866 ### Changes between 0.9.8f and 0.9.8g [19 Oct 2007]
9869 + Binary incompatibility of ssl_ctx_st structure
9870 + DTLS interoperation with non-compliant servers
9871 + Don't call get_session_cb() without proposed session
9872 + Fix ia64 assembler code
9874 *Andy Polyakov, Steve Henson*
9876 ### Changes between 0.9.8e and 0.9.8f [11 Oct 2007]
9878 * DTLS Handshake overhaul. There were longstanding issues with
9879 OpenSSL DTLS implementation, which were making it impossible for
9880 RFC 4347 compliant client to communicate with OpenSSL server.
9881 Unfortunately just fixing these incompatibilities would "cut off"
9882 pre-0.9.8f clients. To allow for hassle free upgrade post-0.9.8e
9883 server keeps tolerating non RFC compliant syntax. The opposite is
9884 not true, 0.9.8f client can not communicate with earlier server.
9885 This update even addresses CVE-2007-4995.
9889 * Changes to avoid need for function casts in OpenSSL: some compilers
9890 (gcc 4.2 and later) reject their use.
9891 *Kurt Roeckx <kurt@roeckx.be>, Peter Hartley <pdh@utter.chaos.org.uk>,
9894 * Add RFC4507 support to OpenSSL. This includes the corrections in
9895 RFC4507bis. The encrypted ticket format is an encrypted encoded
9896 SSL_SESSION structure, that way new session features are automatically
9899 If a client application caches session in an SSL_SESSION structure
9900 support is transparent because tickets are now stored in the encoded
9903 The SSL_CTX structure automatically generates keys for ticket
9904 protection in servers so again support should be possible
9905 with no application modification.
9907 If a client or server wishes to disable RFC4507 support then the option
9908 SSL_OP_NO_TICKET can be set.
9910 Add a TLS extension debugging callback to allow the contents of any client
9911 or server extensions to be examined.
9913 This work was sponsored by Google.
9917 * Add initial support for TLS extensions, specifically for the server_name
9918 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
9919 have new members for a host name. The SSL data structure has an
9920 additional member `SSL_CTX *initial_ctx` so that new sessions can be
9921 stored in that context to allow for session resumption, even after the
9922 SSL has been switched to a new SSL_CTX in reaction to a client's
9923 server_name extension.
9925 New functions (subject to change):
9927 SSL_get_servername()
9928 SSL_get_servername_type()
9931 New CTRL codes and macros (subject to change):
9933 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
9934 - SSL_CTX_set_tlsext_servername_callback()
9935 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
9936 - SSL_CTX_set_tlsext_servername_arg()
9937 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
9939 openssl s_client has a new '-servername ...' option.
9941 openssl s_server has new options '-servername_host ...', '-cert2 ...',
9942 '-key2 ...', '-servername_fatal' (subject to change). This allows
9943 testing the HostName extension for a specific single host name ('-cert'
9944 and '-key' remain fallbacks for handshakes without HostName
9945 negotiation). If the unrecognized_name alert has to be sent, this by
9946 default is a warning; it becomes fatal with the '-servername_fatal'
9949 *Peter Sylvester, Remy Allais, Christophe Renou, Steve Henson*
9951 * Add AES and SSE2 assembly language support to VC++ build.
9955 * Mitigate attack on final subtraction in Montgomery reduction.
9959 * Fix crypto/ec/ec_mult.c to work properly with scalars of value 0
9960 (which previously caused an internal error).
9964 * Squeeze another 10% out of IGE mode when in != out.
9968 * AES IGE mode speedup.
9970 *Dean Gaudet (Google)*
9972 * Add the Korean symmetric 128-bit cipher SEED (see
9973 <http://www.kisa.or.kr/kisa/seed/jsp/seed_eng.jsp>) and
9974 add SEED ciphersuites from RFC 4162:
9976 TLS_RSA_WITH_SEED_CBC_SHA = "SEED-SHA"
9977 TLS_DHE_DSS_WITH_SEED_CBC_SHA = "DHE-DSS-SEED-SHA"
9978 TLS_DHE_RSA_WITH_SEED_CBC_SHA = "DHE-RSA-SEED-SHA"
9979 TLS_DH_anon_WITH_SEED_CBC_SHA = "ADH-SEED-SHA"
9981 To minimize changes between patchlevels in the OpenSSL 0.9.8
9982 series, SEED remains excluded from compilation unless OpenSSL
9983 is configured with 'enable-seed'.
9985 *KISA, Bodo Moeller*
9987 * Mitigate branch prediction attacks, which can be practical if a
9988 single processor is shared, allowing a spy process to extract
9989 information. For detailed background information, see
9990 <http://eprint.iacr.org/2007/039> (O. Aciicmez, S. Gueron,
9991 J.-P. Seifert, "New Branch Prediction Vulnerabilities in OpenSSL
9992 and Necessary Software Countermeasures"). The core of the change
9993 are new versions BN_div_no_branch() and
9994 BN_mod_inverse_no_branch() of BN_div() and BN_mod_inverse(),
9995 respectively, which are slower, but avoid the security-relevant
9996 conditional branches. These are automatically called by BN_div()
9997 and BN_mod_inverse() if the flag BN_FLG_CONSTTIME is set for one
9998 of the input BIGNUMs. Also, BN_is_bit_set() has been changed to
9999 remove a conditional branch.
10001 BN_FLG_CONSTTIME is the new name for the previous
10002 BN_FLG_EXP_CONSTTIME flag, since it now affects more than just
10003 modular exponentiation. (Since OpenSSL 0.9.7h, setting this flag
10004 in the exponent causes BN_mod_exp_mont() to use the alternative
10005 implementation in BN_mod_exp_mont_consttime().) The old name
10006 remains as a deprecated alias.
10008 Similarly, RSA_FLAG_NO_EXP_CONSTTIME is replaced by a more general
10009 RSA_FLAG_NO_CONSTTIME flag since the RSA implementation now uses
10010 constant-time implementations for more than just exponentiation.
10011 Here too the old name is kept as a deprecated alias.
10013 BN_BLINDING_new() will now use BN_dup() for the modulus so that
10014 the BN_BLINDING structure gets an independent copy of the
10015 modulus. This means that the previous `BIGNUM *m` argument to
10016 BN_BLINDING_new() and to BN_BLINDING_create_param() now
10017 essentially becomes `const BIGNUM *m`, although we can't actually
10018 change this in the header file before 0.9.9. It allows
10019 RSA_setup_blinding() to use BN_with_flags() on the modulus to
10020 enable BN_FLG_CONSTTIME.
10022 *Matthew D Wood (Intel Corp)*
10024 * In the SSL/TLS server implementation, be strict about session ID
10025 context matching (which matters if an application uses a single
10026 external cache for different purposes). Previously,
10027 out-of-context reuse was forbidden only if SSL_VERIFY_PEER was
10028 set. This did ensure strict client verification, but meant that,
10029 with applications using a single external cache for quite
10030 different requirements, clients could circumvent ciphersuite
10031 restrictions for a given session ID context by starting a session
10032 in a different context.
10036 * Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
10037 a ciphersuite string such as "DEFAULT:RSA" cannot enable
10038 authentication-only ciphersuites.
10042 * Update the SSL_get_shared_ciphers() fix CVE-2006-3738 which was
10043 not complete and could lead to a possible single byte overflow
10044 ([CVE-2007-5135]) [Ben Laurie]
10046 ### Changes between 0.9.8d and 0.9.8e [23 Feb 2007]
10048 * Since AES128 and AES256 (and similarly Camellia128 and
10049 Camellia256) share a single mask bit in the logic of
10050 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
10051 kludge to work properly if AES128 is available and AES256 isn't
10052 (or if Camellia128 is available and Camellia256 isn't).
10056 * Fix the BIT STRING encoding generated by crypto/ec/ec_asn1.c
10057 (within i2d_ECPrivateKey, i2d_ECPKParameters, i2d_ECParameters):
10058 When a point or a seed is encoded in a BIT STRING, we need to
10059 prevent the removal of trailing zero bits to get the proper DER
10060 encoding. (By default, crypto/asn1/a_bitstr.c assumes the case
10061 of a NamedBitList, for which trailing 0 bits need to be removed.)
10065 * Have SSL/TLS server implementation tolerate "mismatched" record
10066 protocol version while receiving ClientHello even if the
10067 ClientHello is fragmented. (The server can't insist on the
10068 particular protocol version it has chosen before the ServerHello
10069 message has informed the client about his choice.)
10073 * Add RFC 3779 support.
10075 *Rob Austein for ARIN, Ben Laurie*
10077 * Load error codes if they are not already present instead of using a
10078 static variable. This allows them to be cleanly unloaded and reloaded.
10079 Improve header file function name parsing.
10083 * extend SMTP and IMAP protocol emulation in s_client to use EHLO
10084 or CAPABILITY handshake as required by RFCs.
10086 *Goetz Babin-Ebell*
10088 ### Changes between 0.9.8c and 0.9.8d [28 Sep 2006]
10090 * Introduce limits to prevent malicious keys being able to
10091 cause a denial of service. ([CVE-2006-2940])
10093 *Steve Henson, Bodo Moeller*
10095 * Fix ASN.1 parsing of certain invalid structures that can result
10096 in a denial of service. ([CVE-2006-2937]) [Steve Henson]
10098 * Fix buffer overflow in SSL_get_shared_ciphers() function.
10099 ([CVE-2006-3738]) [Tavis Ormandy and Will Drewry, Google Security Team]
10101 * Fix SSL client code which could crash if connecting to a
10102 malicious SSLv2 server. ([CVE-2006-4343])
10104 *Tavis Ormandy and Will Drewry, Google Security Team*
10106 * Since 0.9.8b, ciphersuite strings naming explicit ciphersuites
10107 match only those. Before that, "AES256-SHA" would be interpreted
10108 as a pattern and match "AES128-SHA" too (since AES128-SHA got
10109 the same strength classification in 0.9.7h) as we currently only
10110 have a single AES bit in the ciphersuite description bitmap.
10111 That change, however, also applied to ciphersuite strings such as
10112 "RC4-MD5" that intentionally matched multiple ciphersuites --
10113 namely, SSL 2.0 ciphersuites in addition to the more common ones
10114 from SSL 3.0/TLS 1.0.
10116 So we change the selection algorithm again: Naming an explicit
10117 ciphersuite selects this one ciphersuite, and any other similar
10118 ciphersuite (same bitmap) from *other* protocol versions.
10119 Thus, "RC4-MD5" again will properly select both the SSL 2.0
10120 ciphersuite and the SSL 3.0/TLS 1.0 ciphersuite.
10122 Since SSL 2.0 does not have any ciphersuites for which the
10123 128/256 bit distinction would be relevant, this works for now.
10124 The proper fix will be to use different bits for AES128 and
10125 AES256, which would have avoided the problems from the beginning;
10126 however, bits are scarce, so we can only do this in a new release
10127 (not just a patchlevel) when we can change the SSL_CIPHER
10128 definition to split the single 'unsigned long mask' bitmap into
10129 multiple values to extend the available space.
10133 ### Changes between 0.9.8b and 0.9.8c [05 Sep 2006]
10135 * Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
10136 ([CVE-2006-4339]) [Ben Laurie and Google Security Team]
10138 * Add AES IGE and biIGE modes.
10142 * Change the Unix randomness entropy gathering to use poll() when
10143 possible instead of select(), since the latter has some
10144 undesirable limitations.
10146 *Darryl Miles via Richard Levitte and Bodo Moeller*
10148 * Disable "ECCdraft" ciphersuites more thoroughly. Now special
10149 treatment in ssl/ssl_ciph.s makes sure that these ciphersuites
10150 cannot be implicitly activated as part of, e.g., the "AES" alias.
10151 However, please upgrade to OpenSSL 0.9.9[-dev] for
10152 non-experimental use of the ECC ciphersuites to get TLS extension
10153 support, which is required for curve and point format negotiation
10154 to avoid potential handshake problems.
10158 * Disable rogue ciphersuites:
10160 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
10161 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
10162 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
10164 The latter two were purportedly from
10165 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
10168 Also deactivate the remaining ciphersuites from
10169 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
10170 unofficial, and the ID has long expired.
10174 * Fix RSA blinding Heisenbug (problems sometimes occurred on
10175 dual-core machines) and other potential thread-safety issues.
10179 * Add the symmetric cipher Camellia (128-bit, 192-bit, 256-bit key
10180 versions), which is now available for royalty-free use
10181 (see <http://info.isl.ntt.co.jp/crypt/eng/info/chiteki.html>).
10182 Also, add Camellia TLS ciphersuites from RFC 4132.
10184 To minimize changes between patchlevels in the OpenSSL 0.9.8
10185 series, Camellia remains excluded from compilation unless OpenSSL
10186 is configured with 'enable-camellia'.
10190 * Disable the padding bug check when compression is in use. The padding
10191 bug check assumes the first packet is of even length, this is not
10192 necessarily true if compression is enabled and can result in false
10193 positives causing handshake failure. The actual bug test is ancient
10194 code so it is hoped that implementations will either have fixed it by
10195 now or any which still have the bug do not support compression.
10199 ### Changes between 0.9.8a and 0.9.8b [04 May 2006]
10201 * When applying a cipher rule check to see if string match is an explicit
10202 cipher suite and only match that one cipher suite if it is.
10206 * Link in manifests for VC++ if needed.
10208 *Austin Ziegler <halostatue@gmail.com>*
10210 * Update support for ECC-based TLS ciphersuites according to
10211 draft-ietf-tls-ecc-12.txt with proposed changes (but without
10212 TLS extensions, which are supported starting with the 0.9.9
10213 branch, not in the OpenSSL 0.9.8 branch).
10217 * New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support
10218 opaque EVP_CIPHER_CTX handling.
10222 * Fixes and enhancements to zlib compression code. We now only use
10223 "zlib1.dll" and use the default `__cdecl` calling convention on Win32
10224 to conform with the standards mentioned here:
10225 <http://www.zlib.net/DLL_FAQ.txt>
10226 Static zlib linking now works on Windows and the new --with-zlib-include
10227 --with-zlib-lib options to Configure can be used to supply the location
10228 of the headers and library. Gracefully handle case where zlib library
10233 * Several fixes and enhancements to the OID generation code. The old code
10234 sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't
10235 handle numbers larger than ULONG_MAX, truncated printing and had a
10236 non standard OBJ_obj2txt() behaviour.
10240 * Add support for building of engines under engine/ as shared libraries
10241 under VC++ build system.
10245 * Corrected the numerous bugs in the Win32 path splitter in DSO.
10246 Hopefully, we will not see any false combination of paths any more.
10250 ### Changes between 0.9.8 and 0.9.8a [11 Oct 2005]
10252 * Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
10253 (part of SSL_OP_ALL). This option used to disable the
10254 countermeasure against man-in-the-middle protocol-version
10255 rollback in the SSL 2.0 server implementation, which is a bad
10256 idea. ([CVE-2005-2969])
10258 *Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
10259 for Information Security, National Institute of Advanced Industrial
10260 Science and Technology [AIST], Japan)*
10262 * Add two function to clear and return the verify parameter flags.
10266 * Keep cipherlists sorted in the source instead of sorting them at
10267 runtime, thus removing the need for a lock.
10271 * Avoid some small subgroup attacks in Diffie-Hellman.
10273 *Nick Mathewson and Ben Laurie*
10275 * Add functions for well-known primes.
10279 * Extended Windows CE support.
10281 *Satoshi Nakamura and Andy Polyakov*
10283 * Initialize SSL_METHOD structures at compile time instead of during
10284 runtime, thus removing the need for a lock.
10288 * Make PKCS7_decrypt() work even if no certificate is supplied by
10289 attempting to decrypt each encrypted key in turn. Add support to
10294 ### Changes between 0.9.7h and 0.9.8 [05 Jul 2005]
10296 [NB: OpenSSL 0.9.7i and later 0.9.7 patch levels were released after
10299 * Add libcrypto.pc and libssl.pc for those who feel they need them.
10303 * Change CA.sh and CA.pl so they don't bundle the CSR and the private
10304 key into the same file any more.
10308 * Add initial support for Win64, both IA64 and AMD64/x64 flavors.
10312 * Add -utf8 command line and config file option to 'ca'.
10314 *Stefan <stf@udoma.org*
10316 * Removed the macro des_crypt(), as it seems to conflict with some
10317 libraries. Use DES_crypt().
10321 * Correct naming of the 'chil' and '4758cca' ENGINEs. This
10322 involves renaming the source and generated shared-libs for
10323 both. The engines will accept the corrected or legacy ids
10324 ('ncipher' and '4758_cca' respectively) when binding. NB,
10325 this only applies when building 'shared'.
10327 *Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe*
10329 * Add attribute functions to EVP_PKEY structure. Modify
10330 PKCS12_create() to recognize a CSP name attribute and
10331 use it. Make -CSP option work again in pkcs12 utility.
10335 * Add new functionality to the bn blinding code:
10336 - automatic re-creation of the BN_BLINDING parameters after
10337 a fixed number of uses (currently 32)
10338 - add new function for parameter creation
10339 - introduce flags to control the update behaviour of the
10340 BN_BLINDING parameters
10341 - hide BN_BLINDING structure
10342 Add a second BN_BLINDING slot to the RSA structure to improve
10343 performance when a single RSA object is shared among several
10348 * Add support for DTLS.
10350 *Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie*
10352 * Add support for DER encoded private keys (SSL_FILETYPE_ASN1)
10353 to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file()
10357 * Remove buggy and incomplete DH cert support from
10358 ssl/ssl_rsa.c and ssl/s3_both.c
10362 * Use SHA-1 instead of MD5 as the default digest algorithm for
10363 the `apps/openssl` commands.
10367 * Compile clean with "-Wall -Wmissing-prototypes
10368 -Wstrict-prototypes -Wmissing-declarations -Werror". Currently
10369 DEBUG_SAFESTACK must also be set.
10373 * Change ./Configure so that certain algorithms can be disabled by default.
10374 The new counterpiece to "no-xxx" is "enable-xxx".
10376 The patented RC5 and MDC2 algorithms will now be disabled unless
10377 "enable-rc5" and "enable-mdc2", respectively, are specified.
10379 (IDEA remains enabled despite being patented. This is because IDEA
10380 is frequently required for interoperability, and there is no license
10381 fee for non-commercial use. As before, "no-idea" can be used to
10382 avoid this algorithm.)
10386 * Add processing of proxy certificates (see RFC 3820). This work was
10387 sponsored by KTH (The Royal Institute of Technology in Stockholm) and
10388 EGEE (Enabling Grids for E-science in Europe).
10392 * RC4 performance overhaul on modern architectures/implementations, such
10393 as Intel P4, IA-64 and AMD64.
10397 * New utility extract-section.pl. This can be used specify an alternative
10398 section number in a pod file instead of having to treat each file as
10399 a separate case in Makefile. This can be done by adding two lines to the
10402 =for comment openssl_section:XXX
10404 The blank line is mandatory.
10408 * New arguments -certform, -keyform and -pass for s_client and s_server
10409 to allow alternative format key and certificate files and passphrase
10414 * New structure X509_VERIFY_PARAM which combines current verify parameters,
10415 update associated structures and add various utility functions.
10417 Add new policy related verify parameters, include policy checking in
10418 standard verify code. Enhance 'smime' application with extra parameters
10419 to support policy checking and print out.
10423 * Add a new engine to support VIA PadLock ACE extensions in the VIA C3
10424 Nehemiah processors. These extensions support AES encryption in hardware
10425 as well as RNG (though RNG support is currently disabled).
10427 *Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov*
10429 * Deprecate `BN_[get|set]_params()` functions (they were ignored internally).
10433 * New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented.
10435 *Andy Polyakov and a number of other people*
10437 * Improved PowerPC platform support. Most notably BIGNUM assembler
10438 implementation contributed by IBM.
10440 *Suresh Chari, Peter Waltenberg, Andy Polyakov*
10442 * The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public
10443 exponent rather than 'unsigned long'. There is a corresponding change to
10444 the new 'rsa_keygen' element of the RSA_METHOD structure.
10446 *Jelte Jansen, Geoff Thorpe*
10448 * Functionality for creating the initial serial number file is now
10449 moved from CA.pl to the 'ca' utility with a new option -create_serial.
10451 (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial
10452 number file to 1, which is bound to cause problems. To avoid
10453 the problems while respecting compatibility between different 0.9.7
10454 patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in
10455 CA.pl for serial number initialization. With the new release 0.9.8,
10456 we can fix the problem directly in the 'ca' utility.)
10460 * Reduced header interdependencies by declaring more opaque objects in
10461 ossl_typ.h. As a consequence, including some headers (eg. engine.h) will
10462 give fewer recursive includes, which could break lazy source code - so
10463 this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always,
10464 developers should define this symbol when building and using openssl to
10465 ensure they track the recommended behaviour, interfaces, [etc], but
10466 backwards-compatible behaviour prevails when this isn't defined.
10470 * New function X509_POLICY_NODE_print() which prints out policy nodes.
10474 * Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality.
10475 This will generate a random key of the appropriate length based on the
10476 cipher context. The EVP_CIPHER can provide its own random key generation
10477 routine to support keys of a specific form. This is used in the des and
10478 3des routines to generate a key of the correct parity. Update S/MIME
10479 code to use new functions and hence generate correct parity DES keys.
10480 Add EVP_CHECK_DES_KEY #define to return an error if the key is not
10481 valid (weak or incorrect parity).
10485 * Add a local set of CRLs that can be used by X509_verify_cert() as well
10486 as looking them up. This is useful when the verified structure may contain
10487 CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs
10488 present unless the new PKCS7_NO_CRL flag is asserted.
10492 * Extend ASN1 oid configuration module. It now additionally accepts the
10495 shortName = some long name, 1.2.3.4
10499 * Reimplemented the BN_CTX implementation. There is now no more static
10500 limitation on the number of variables it can handle nor the depth of the
10501 "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack
10502 information can now expand as required, and rather than having a single
10503 static array of bignums, BN_CTX now uses a linked-list of such arrays
10504 allowing it to expand on demand whilst maintaining the usefulness of
10505 BN_CTX's "bundling".
10509 * Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD
10510 to allow all RSA operations to function using a single BN_CTX.
10514 * Preliminary support for certificate policy evaluation and checking. This
10515 is initially intended to pass the tests outlined in "Conformance Testing
10516 of Relying Party Client Certificate Path Processing Logic" v1.07.
10520 * bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and
10521 remained unused and not that useful. A variety of other little bignum
10522 tweaks and fixes have also been made continuing on from the audit (see
10527 * Constify all or almost all d2i, c2i, s2i and r2i functions, along with
10528 associated ASN1, EVP and SSL functions and old ASN1 macros.
10532 * BN_zero() only needs to set 'top' and 'neg' to zero for correct results,
10533 and this should never fail. So the return value from the use of
10534 BN_set_word() (which can fail due to needless expansion) is now deprecated;
10535 if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro.
10539 * BN_CTX_get() should return zero-valued bignums, providing the same
10540 initialised value as BN_new().
10542 *Geoff Thorpe, suggested by Ulf Möller*
10544 * Support for inhibitAnyPolicy certificate extension.
10548 * An audit of the BIGNUM code is underway, for which debugging code is
10549 enabled when BN_DEBUG is defined. This makes stricter enforcements on what
10550 is considered valid when processing BIGNUMs, and causes execution to
10551 assert() when a problem is discovered. If BN_DEBUG_RAND is defined,
10552 further steps are taken to deliberately pollute unused data in BIGNUM
10553 structures to try and expose faulty code further on. For now, openssl will
10554 (in its default mode of operation) continue to tolerate the inconsistent
10555 forms that it has tolerated in the past, but authors and packagers should
10556 consider trying openssl and their own applications when compiled with
10557 these debugging symbols defined. It will help highlight potential bugs in
10558 their own code, and will improve the test coverage for OpenSSL itself. At
10559 some point, these tighter rules will become openssl's default to improve
10560 maintainability, though the assert()s and other overheads will remain only
10561 in debugging configurations. See bn.h for more details.
10563 *Geoff Thorpe, Nils Larsch, Ulf Möller*
10565 * BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure
10566 that can only be obtained through BN_CTX_new() (which implicitly
10567 initialises it). The presence of this function only made it possible
10568 to overwrite an existing structure (and cause memory leaks).
10572 * Because of the callback-based approach for implementing LHASH as a
10573 template type, lh_insert() adds opaque objects to hash-tables and
10574 lh_doall() or lh_doall_arg() are typically used with a destructor callback
10575 to clean up those corresponding objects before destroying the hash table
10576 (and losing the object pointers). So some over-zealous constifications in
10577 LHASH have been relaxed so that lh_insert() does not take (nor store) the
10578 objects as "const" and the `lh_doall[_arg]` callback wrappers are not
10579 prototyped to have "const" restrictions on the object pointers they are
10580 given (and so aren't required to cast them away any more).
10584 * The tmdiff.h API was so ugly and minimal that our own timing utility
10585 (speed) prefers to use its own implementation. The two implementations
10586 haven't been consolidated as yet (volunteers?) but the tmdiff API has had
10587 its object type properly exposed (MS_TM) instead of casting to/from
10588 `char *`. This may still change yet if someone realises MS_TM and
10590 aren't necessarily the greatest nomenclatures - but this is what was used
10591 internally to the implementation so I've used that for now.
10595 * Ensure that deprecated functions do not get compiled when
10596 OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of
10597 the self-tests were still using deprecated key-generation functions so
10598 these have been updated also.
10602 * Reorganise PKCS#7 code to separate the digest location functionality
10603 into PKCS7_find_digest(), digest addition into PKCS7_bio_add_digest().
10604 New function PKCS7_set_digest() to set the digest type for PKCS#7
10605 digestedData type. Add additional code to correctly generate the
10606 digestedData type and add support for this type in PKCS7 initialization
10611 * New function PKCS7_set0_type_other() this initializes a PKCS7
10612 structure of type "other".
10616 * Fix prime generation loop in crypto/bn/bn_prime.pl by making
10617 sure the loop does correctly stop and breaking ("division by zero")
10618 modulus operations are not performed. The (pre-generated) prime
10619 table crypto/bn/bn_prime.h was already correct, but it could not be
10620 re-generated on some platforms because of the "division by zero"
10621 situation in the script.
10623 *Ralf S. Engelschall*
10625 * Update support for ECC-based TLS ciphersuites according to
10626 draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with
10627 SHA-1 now is only used for "small" curves (where the
10628 representation of a field element takes up to 24 bytes); for
10629 larger curves, the field element resulting from ECDH is directly
10630 used as premaster secret.
10632 *Douglas Stebila (Sun Microsystems Laboratories)*
10634 * Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2
10635 curve secp160r1 to the tests.
10637 *Douglas Stebila (Sun Microsystems Laboratories)*
10639 * Add the possibility to load symbols globally with DSO.
10641 *Götz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte*
10643 * Add the functions ERR_set_mark() and ERR_pop_to_mark() for better
10644 control of the error stack.
10648 * Add support for STORE in ENGINE.
10652 * Add the STORE type. The intention is to provide a common interface
10653 to certificate and key stores, be they simple file-based stores, or
10654 HSM-type store, or LDAP stores, or...
10655 NOTE: The code is currently UNTESTED and isn't really used anywhere.
10659 * Add a generic structure called OPENSSL_ITEM. This can be used to
10660 pass a list of arguments to any function as well as provide a way
10661 for a function to pass data back to the caller.
10665 * Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup()
10666 works like BUF_strdup() but can be used to duplicate a portion of
10667 a string. The copy gets NUL-terminated. BUF_memdup() duplicates
10672 * Add the function sk_find_ex() which works like sk_find(), but will
10673 return an index to an element even if an exact match couldn't be
10674 found. The index is guaranteed to point at the element where the
10675 searched-for key would be inserted to preserve sorting order.
10679 * Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but
10680 takes an extra flags argument for optional functionality. Currently,
10681 the following flags are defined:
10683 OBJ_BSEARCH_VALUE_ON_NOMATCH
10684 This one gets OBJ_bsearch_ex() to return a pointer to the first
10685 element where the comparing function returns a negative or zero
10688 OBJ_BSEARCH_FIRST_VALUE_ON_MATCH
10689 This one gets OBJ_bsearch_ex() to return a pointer to the first
10690 element where the comparing function returns zero. This is useful
10691 if there are more than one element where the comparing function
10696 * Make it possible to create self-signed certificates with 'openssl ca'
10697 in such a way that the self-signed certificate becomes part of the
10698 CA database and uses the same mechanisms for serial number generation
10699 as all other certificate signing. The new flag '-selfsign' enables
10700 this functionality. Adapt CA.sh and CA.pl.in.
10704 * Add functionality to check the public key of a certificate request
10705 against a given private. This is useful to check that a certificate
10706 request can be signed by that key (self-signing).
10710 * Make it possible to have multiple active certificates with the same
10711 subject in the CA index file. This is done only if the keyword
10712 'unique_subject' is set to 'no' in the main CA section (default
10713 if 'CA_default') of the configuration file. The value is saved
10714 with the database itself in a separate index attribute file,
10715 named like the index file with '.attr' appended to the name.
10719 * Generate multi-valued AVAs using '+' notation in config files for
10724 * Support for nameConstraints certificate extension.
10728 * Support for policyConstraints certificate extension.
10732 * Support for policyMappings certificate extension.
10736 * Make sure the default DSA_METHOD implementation only uses its
10737 dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL,
10738 and change its own handlers to be NULL so as to remove unnecessary
10739 indirection. This lets alternative implementations fallback to the
10740 default implementation more easily.
10744 * Support for directoryName in GeneralName related extensions
10749 * Make it possible to link applications using Makefile.shared.
10750 Make that possible even when linking against static libraries!
10754 * Support for single pass processing for S/MIME signing. This now
10755 means that S/MIME signing can be done from a pipe, in addition
10756 cleartext signing (multipart/signed type) is effectively streaming
10757 and the signed data does not need to be all held in memory.
10759 This is done with a new flag PKCS7_STREAM. When this flag is set
10760 PKCS7_sign() only initializes the PKCS7 structure and the actual signing
10761 is done after the data is output (and digests calculated) in
10762 SMIME_write_PKCS7().
10766 * Add full support for -rpath/-R, both in shared libraries and
10767 applications, at least on the platforms where it's known how
10772 * In crypto/ec/ec_mult.c, implement fast point multiplication with
10773 precomputation, based on wNAF splitting: EC_GROUP_precompute_mult()
10774 will now compute a table of multiples of the generator that
10775 makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul()
10776 faster (notably in the case of a single point multiplication,
10777 scalar * generator).
10779 *Nils Larsch, Bodo Moeller*
10781 * IPv6 support for certificate extensions. The various extensions
10782 which use the IP:a.b.c.d can now take IPv6 addresses using the
10783 formats of RFC1884 2.2 . IPv6 addresses are now also displayed
10788 * Added an ENGINE that implements RSA by performing private key
10789 exponentiations with the GMP library. The conversions to and from
10790 GMP's mpz_t format aren't optimised nor are any montgomery forms
10791 cached, and on x86 it appears OpenSSL's own performance has caught up.
10792 However there are likely to be other architectures where GMP could
10793 provide a boost. This ENGINE is not built in by default, but it can be
10794 specified at Configure time and should be accompanied by the necessary
10795 linker additions, eg;
10796 ./config -DOPENSSL_USE_GMP -lgmp
10800 * "openssl engine" will not display ENGINE/DSO load failure errors when
10801 testing availability of engines with "-t" - the old behaviour is
10802 produced by increasing the feature's verbosity with "-tt".
10806 * ECDSA routines: under certain error conditions uninitialized BN objects
10807 could be freed. Solution: make sure initialization is performed early
10808 enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de>
10813 * Key-generation can now be implemented in RSA_METHOD, DSA_METHOD
10814 and DH_METHOD (eg. by ENGINE implementations) to override the normal
10815 software implementations. For DSA and DH, parameter generation can
10816 also be overridden by providing the appropriate method callbacks.
10820 * Change the "progress" mechanism used in key-generation and
10821 primality testing to functions that take a new BN_GENCB pointer in
10822 place of callback/argument pairs. The new API functions have `_ex`
10823 postfixes and the older functions are reimplemented as wrappers for
10824 the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide
10825 declarations of the old functions to help (graceful) attempts to
10826 migrate to the new functions. Also, the new key-generation API
10827 functions operate on a caller-supplied key-structure and return
10828 success/failure rather than returning a key or NULL - this is to
10829 help make "keygen" another member function of RSA_METHOD etc.
10831 Example for using the new callback interface:
10833 int (*my_callback)(int a, int b, BN_GENCB *cb) = ...;
10834 void *my_arg = ...;
10837 BN_GENCB_set(&my_cb, my_callback, my_arg);
10839 return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb);
10840 /* For the meaning of a, b in calls to my_callback(), see the
10841 * documentation of the function that calls the callback.
10842 * cb will point to my_cb; my_arg can be retrieved as cb->arg.
10843 * my_callback should return 1 if it wants BN_is_prime_ex()
10844 * to continue, or 0 to stop.
10849 * Change the ZLIB compression method to be stateful, and make it
10850 available to TLS with the number defined in
10851 draft-ietf-tls-compression-04.txt.
10855 * Add the ASN.1 structures and functions for CertificatePair, which
10856 is defined as follows (according to X.509_4thEditionDraftV6.pdf):
10858 CertificatePair ::= SEQUENCE {
10859 forward [0] Certificate OPTIONAL,
10860 reverse [1] Certificate OPTIONAL,
10861 -- at least one of the pair shall be present -- }
10863 Also implement the PEM functions to read and write certificate
10864 pairs, and defined the PEM tag as "CERTIFICATE PAIR".
10866 This needed to be defined, mostly for the sake of the LDAP
10867 attribute crossCertificatePair, but may prove useful elsewhere as
10872 * Make it possible to inhibit symlinking of shared libraries in
10873 Makefile.shared, for Cygwin's sake.
10877 * Extend the BIGNUM API by creating a function
10878 void BN_set_negative(BIGNUM *a, int neg);
10879 and a macro that behave like
10880 int BN_is_negative(const BIGNUM *a);
10882 to avoid the need to access 'a->neg' directly in applications.
10886 * Implement fast modular reduction for pseudo-Mersenne primes
10887 used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c).
10888 EC_GROUP_new_curve_GFp() will now automatically use this
10891 *Nils Larsch <nla@trustcenter.de>*
10893 * Add new lock type (CRYPTO_LOCK_BN).
10897 * Change the ENGINE framework to automatically load engines
10898 dynamically from specific directories unless they could be
10899 found to already be built in or loaded. Move all the
10900 current engines except for the cryptodev one to a new
10901 directory engines/.
10902 The engines in engines/ are built as shared libraries if
10903 the "shared" options was given to ./Configure or ./config.
10904 Otherwise, they are inserted in libcrypto.a.
10905 /usr/local/ssl/engines is the default directory for dynamic
10906 engines, but that can be overridden at configure time through
10907 the usual use of --prefix and/or --openssldir, and at run
10908 time with the environment variable OPENSSL_ENGINES.
10910 *Geoff Thorpe and Richard Levitte*
10912 * Add Makefile.shared, a helper makefile to build shared
10913 libraries. Adapt Makefile.org.
10917 * Add version info to Win32 DLLs.
10919 *Peter 'Luna' Runestig" <peter@runestig.com>*
10921 * Add new 'medium level' PKCS#12 API. Certificates and keys
10922 can be added using this API to created arbitrary PKCS#12
10923 files while avoiding the low-level API.
10925 New options to PKCS12_create(), key or cert can be NULL and
10926 will then be omitted from the output file. The encryption
10927 algorithm NIDs can be set to -1 for no encryption, the mac
10928 iteration count can be set to 0 to omit the mac.
10930 Enhance pkcs12 utility by making the -nokeys and -nocerts
10931 options work when creating a PKCS#12 file. New option -nomac
10932 to omit the mac, NONE can be set for an encryption algorithm.
10933 New code is modified to use the enhanced PKCS12_create()
10934 instead of the low-level API.
10938 * Extend ASN1 encoder to support indefinite length constructed
10939 encoding. This can output sequences tags and octet strings in
10940 this form. Modify pk7_asn1.c to support indefinite length
10941 encoding. This is experimental and needs additional code to
10942 be useful, such as an ASN1 bio and some enhanced streaming
10945 Extend template encode functionality so that tagging is passed
10946 down to the template encoder.
10950 * Let 'openssl req' fail if an argument to '-newkey' is not
10951 recognized instead of using RSA as a default.
10955 * Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt.
10956 As these are not official, they are not included in "ALL";
10957 the "ECCdraft" ciphersuite group alias can be used to select them.
10959 *Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)*
10961 * Add ECDH engine support.
10963 *Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)*
10965 * Add ECDH in new directory crypto/ecdh/.
10967 *Douglas Stebila (Sun Microsystems Laboratories)*
10969 * Let BN_rand_range() abort with an error after 100 iterations
10970 without success (which indicates a broken PRNG).
10974 * Change BN_mod_sqrt() so that it verifies that the input value
10975 is really the square of the return value. (Previously,
10976 BN_mod_sqrt would show GIGO behaviour.)
10980 * Add named elliptic curves over binary fields from X9.62, SECG,
10981 and WAP/WTLS; add OIDs that were still missing.
10983 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10985 * Extend the EC library for elliptic curves over binary fields
10986 (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/).
10989 EC_GF2m_simple_method
10993 EC_GROUP_new_curve_GF2m
10994 EC_GROUP_set_curve_GF2m
10995 EC_GROUP_get_curve_GF2m
10996 EC_POINT_set_affine_coordinates_GF2m
10997 EC_POINT_get_affine_coordinates_GF2m
10998 EC_POINT_set_compressed_coordinates_GF2m
11000 Point compression for binary fields is disabled by default for
11001 patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to
11004 As binary polynomials are represented as BIGNUMs, various members
11005 of the EC_GROUP and EC_POINT data structures can be shared
11006 between the implementations for prime fields and binary fields;
11007 the above `..._GF2m functions` (except for EX_GROUP_new_curve_GF2m)
11008 are essentially identical to their `..._GFp` counterparts.
11009 (For simplicity, the `..._GFp` prefix has been dropped from
11010 various internal method names.)
11012 An internal 'field_div' method (similar to 'field_mul' and
11013 'field_sqr') has been added; this is used only for binary fields.
11015 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
11017 * Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult()
11018 through methods ('mul', 'precompute_mult').
11020 The generic implementations (now internally called 'ec_wNAF_mul'
11021 and 'ec_wNAF_precomputed_mult') remain the default if these
11022 methods are undefined.
11024 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
11026 * New function EC_GROUP_get_degree, which is defined through
11027 EC_METHOD. For curves over prime fields, this returns the bit
11028 length of the modulus.
11030 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
11032 * New functions EC_GROUP_dup, EC_POINT_dup.
11033 (These simply call ..._new and ..._copy).
11035 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
11037 * Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c.
11038 Polynomials are represented as BIGNUMs (where the sign bit is not
11039 used) in the following functions [macros]:
11042 BN_GF2m_sub [= BN_GF2m_add]
11043 BN_GF2m_mod [wrapper for BN_GF2m_mod_arr]
11044 BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr]
11045 BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr]
11047 BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr]
11048 BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr]
11049 BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr]
11050 BN_GF2m_cmp [= BN_ucmp]
11052 (Note that only the 'mod' functions are actually for fields GF(2^m).
11053 BN_GF2m_add() is misnomer, but this is for the sake of consistency.)
11055 For some functions, an the irreducible polynomial defining a
11056 field can be given as an 'unsigned int[]' with strictly
11057 decreasing elements giving the indices of those bits that are set;
11058 i.e., p[] represents the polynomial
11059 f(t) = t^p[0] + t^p[1] + ... + t^p[k]
11061 p[0] > p[1] > ... > p[k] = 0.
11062 This applies to the following functions:
11065 BN_GF2m_mod_mul_arr
11066 BN_GF2m_mod_sqr_arr
11067 BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv]
11068 BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div]
11069 BN_GF2m_mod_exp_arr
11070 BN_GF2m_mod_sqrt_arr
11071 BN_GF2m_mod_solve_quad_arr
11075 Conversion can be performed by the following functions:
11080 bntest.c has additional tests for binary polynomial arithmetic.
11082 Two implementations for BN_GF2m_mod_div() are available.
11083 The default algorithm simply uses BN_GF2m_mod_inv() and
11084 BN_GF2m_mod_mul(). The alternative algorithm is compiled in only
11085 if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the
11086 copyright notice in crypto/bn/bn_gf2m.c before enabling it).
11088 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
11090 * Add new error code 'ERR_R_DISABLED' that can be used when some
11091 functionality is disabled at compile-time.
11093 *Douglas Stebila <douglas.stebila@sun.com>*
11095 * Change default behaviour of 'openssl asn1parse' so that more
11096 information is visible when viewing, e.g., a certificate:
11098 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump'
11099 mode the content of non-printable OCTET STRINGs is output in a
11100 style similar to INTEGERs, but with '[HEX DUMP]' prepended to
11101 avoid the appearance of a printable string.
11103 *Nils Larsch <nla@trustcenter.de>*
11105 * Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access
11107 EC_GROUP_set_asn1_flag()
11108 EC_GROUP_get_asn1_flag()
11109 EC_GROUP_set_point_conversion_form()
11110 EC_GROUP_get_point_conversion_form()
11111 These control ASN1 encoding details:
11112 - Curves (i.e., groups) are encoded explicitly unless asn1_flag
11113 has been set to OPENSSL_EC_NAMED_CURVE.
11114 - Points are encoded in uncompressed form by default; options for
11115 asn1_for are as for point2oct, namely
11116 POINT_CONVERSION_COMPRESSED
11117 POINT_CONVERSION_UNCOMPRESSED
11118 POINT_CONVERSION_HYBRID
11120 Also add 'seed' and 'seed_len' members to EC_GROUP with access
11122 EC_GROUP_set_seed()
11123 EC_GROUP_get0_seed()
11124 EC_GROUP_get_seed_len()
11125 This is used only for ASN1 purposes (so far).
11127 *Nils Larsch <nla@trustcenter.de>*
11129 * Add 'field_type' member to EC_METHOD, which holds the NID
11130 of the appropriate field type OID. The new function
11131 EC_METHOD_get_field_type() returns this value.
11133 *Nils Larsch <nla@trustcenter.de>*
11136 EC_POINT_point2bn()
11137 EC_POINT_bn2point()
11138 EC_POINT_point2hex()
11139 EC_POINT_hex2point()
11140 providing useful interfaces to EC_POINT_point2oct() and
11141 EC_POINT_oct2point().
11143 *Nils Larsch <nla@trustcenter.de>*
11145 * Change internals of the EC library so that the functions
11146 EC_GROUP_set_generator()
11147 EC_GROUP_get_generator()
11148 EC_GROUP_get_order()
11149 EC_GROUP_get_cofactor()
11150 are implemented directly in crypto/ec/ec_lib.c and not dispatched
11151 to methods, which would lead to unnecessary code duplication when
11152 adding different types of curves.
11154 *Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller*
11156 * Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM
11157 arithmetic, and such that modified wNAFs are generated
11158 (which avoid length expansion in many cases).
11162 * Add a function EC_GROUP_check_discriminant() (defined via
11163 EC_METHOD) that verifies that the curve discriminant is non-zero.
11165 Add a function EC_GROUP_check() that makes some sanity tests
11166 on a EC_GROUP, its generator and order. This includes
11167 EC_GROUP_check_discriminant().
11169 *Nils Larsch <nla@trustcenter.de>*
11171 * Add ECDSA in new directory crypto/ecdsa/.
11173 Add applications 'openssl ecparam' and 'openssl ecdsa'
11174 (these are based on 'openssl dsaparam' and 'openssl dsa').
11176 ECDSA support is also included in various other files across the
11177 library. Most notably,
11178 - 'openssl req' now has a '-newkey ecdsa:file' option;
11179 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA;
11180 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and
11181 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make
11182 them suitable for ECDSA where domain parameters must be
11183 extracted before the specific public key;
11184 - ECDSA engine support has been added.
11186 *Nils Larsch <nla@trustcenter.de>*
11188 * Include some named elliptic curves, and add OIDs from X9.62,
11189 SECG, and WAP/WTLS. Each curve can be obtained from the new
11191 EC_GROUP_new_by_curve_name(),
11192 and the list of available named curves can be obtained with
11193 EC_get_builtin_curves().
11194 Also add a 'curve_name' member to EC_GROUP objects, which can be
11196 EC_GROUP_set_curve_name()
11197 EC_GROUP_get_curve_name()
11199 *Nils Larsch <larsch@trustcenter.de, Bodo Moeller*
11201 * Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
11202 was actually never needed) and in BN_mul(). The removal in BN_mul()
11203 required a small change in bn_mul_part_recursive() and the addition
11204 of the functions bn_cmp_part_words(), bn_sub_part_words() and
11205 bn_add_part_words(), which do the same thing as bn_cmp_words(),
11206 bn_sub_words() and bn_add_words() except they take arrays with
11211 ### Changes between 0.9.7l and 0.9.7m [23 Feb 2007]
11213 * Cleanse PEM buffers before freeing them since they may contain
11216 *Benjamin Bennett <ben@psc.edu>*
11218 * Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
11219 a ciphersuite string such as "DEFAULT:RSA" cannot enable
11220 authentication-only ciphersuites.
11224 * Since AES128 and AES256 share a single mask bit in the logic of
11225 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
11226 kludge to work properly if AES128 is available and AES256 isn't.
11230 * Expand security boundary to match 1.1.1 module.
11234 * Remove redundant features: hash file source, editing of test vectors
11235 modify fipsld to use external fips_premain.c signature.
11239 * New perl script mkfipsscr.pl to create shell scripts or batch files to
11240 run algorithm test programs.
11244 * Make algorithm test programs more tolerant of whitespace.
11248 * Have SSL/TLS server implementation tolerate "mismatched" record
11249 protocol version while receiving ClientHello even if the
11250 ClientHello is fragmented. (The server can't insist on the
11251 particular protocol version it has chosen before the ServerHello
11252 message has informed the client about his choice.)
11256 * Load error codes if they are not already present instead of using a
11257 static variable. This allows them to be cleanly unloaded and reloaded.
11261 ### Changes between 0.9.7k and 0.9.7l [28 Sep 2006]
11263 * Introduce limits to prevent malicious keys being able to
11264 cause a denial of service. ([CVE-2006-2940])
11266 *Steve Henson, Bodo Moeller*
11268 * Fix ASN.1 parsing of certain invalid structures that can result
11269 in a denial of service. ([CVE-2006-2937]) [Steve Henson]
11271 * Fix buffer overflow in SSL_get_shared_ciphers() function.
11272 ([CVE-2006-3738]) [Tavis Ormandy and Will Drewry, Google Security Team]
11274 * Fix SSL client code which could crash if connecting to a
11275 malicious SSLv2 server. ([CVE-2006-4343])
11277 *Tavis Ormandy and Will Drewry, Google Security Team*
11279 * Change ciphersuite string processing so that an explicit
11280 ciphersuite selects this one ciphersuite (so that "AES256-SHA"
11281 will no longer include "AES128-SHA"), and any other similar
11282 ciphersuite (same bitmap) from *other* protocol versions (so that
11283 "RC4-MD5" will still include both the SSL 2.0 ciphersuite and the
11284 SSL 3.0/TLS 1.0 ciphersuite). This is a backport combining
11285 changes from 0.9.8b and 0.9.8d.
11289 ### Changes between 0.9.7j and 0.9.7k [05 Sep 2006]
11291 * Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
11292 ([CVE-2006-4339]) [Ben Laurie and Google Security Team]
11294 * Change the Unix randomness entropy gathering to use poll() when
11295 possible instead of select(), since the latter has some
11296 undesirable limitations.
11298 *Darryl Miles via Richard Levitte and Bodo Moeller*
11300 * Disable rogue ciphersuites:
11302 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
11303 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
11304 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
11306 The latter two were purportedly from
11307 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
11310 Also deactivate the remaining ciphersuites from
11311 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
11312 unofficial, and the ID has long expired.
11316 * Fix RSA blinding Heisenbug (problems sometimes occurred on
11317 dual-core machines) and other potential thread-safety issues.
11321 ### Changes between 0.9.7i and 0.9.7j [04 May 2006]
11323 * Adapt fipsld and the build system to link against the validated FIPS
11324 module in FIPS mode.
11328 * Fixes for VC++ 2005 build under Windows.
11332 * Add new Windows build target VC-32-GMAKE for VC++. This uses GNU make
11333 from a Windows bash shell such as MSYS. It is autodetected from the
11334 "config" script when run from a VC++ environment. Modify standard VC++
11335 build to use fipscanister.o from the GNU make build.
11339 ### Changes between 0.9.7h and 0.9.7i [14 Oct 2005]
11341 * Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS.
11342 The value now differs depending on if you build for FIPS or not.
11343 BEWARE! A program linked with a shared FIPSed libcrypto can't be
11344 safely run with a non-FIPSed libcrypto, as it may crash because of
11345 the difference induced by this change.
11349 ### Changes between 0.9.7g and 0.9.7h [11 Oct 2005]
11351 * Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
11352 (part of SSL_OP_ALL). This option used to disable the
11353 countermeasure against man-in-the-middle protocol-version
11354 rollback in the SSL 2.0 server implementation, which is a bad
11355 idea. ([CVE-2005-2969])
11357 *Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
11358 for Information Security, National Institute of Advanced Industrial
11359 Science and Technology [AIST, Japan)]*
11361 * Minimal support for X9.31 signatures and PSS padding modes. This is
11362 mainly for FIPS compliance and not fully integrated at this stage.
11366 * For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform
11367 the exponentiation using a fixed-length exponent. (Otherwise,
11368 the information leaked through timing could expose the secret key
11369 after many signatures; cf. Bleichenbacher's attack on DSA with
11374 * Make a new fixed-window mod_exp implementation the default for
11375 RSA, DSA, and DH private-key operations so that the sequence of
11376 squares and multiplies and the memory access pattern are
11377 independent of the particular secret key. This will mitigate
11378 cache-timing and potential related attacks.
11380 BN_mod_exp_mont_consttime() is the new exponentiation implementation,
11381 and this is automatically used by BN_mod_exp_mont() if the new flag
11382 BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH
11383 will use this BN flag for private exponents unless the flag
11384 RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or
11385 DH_FLAG_NO_EXP_CONSTTIME, respectively, is set.
11387 *Matthew D Wood (Intel Corp), with some changes by Bodo Moeller*
11389 * Change the client implementation for SSLv23_method() and
11390 SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0
11391 Client Hello message format if the SSL_OP_NO_SSLv2 option is set.
11392 (Previously, the SSL 2.0 backwards compatible Client Hello
11393 message format would be used even with SSL_OP_NO_SSLv2.)
11397 * Add support for smime-type MIME parameter in S/MIME messages which some
11402 * New function BN_MONT_CTX_set_locked() to set montgomery parameters in
11403 a threadsafe manner. Modify rsa code to use new function and add calls
11404 to dsa and dh code (which had race conditions before).
11408 * Include the fixed error library code in the C error file definitions
11409 instead of fixing them up at runtime. This keeps the error code
11410 structures constant.
11414 ### Changes between 0.9.7f and 0.9.7g [11 Apr 2005]
11416 [NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after
11419 * Fixes for newer kerberos headers. NB: the casts are needed because
11420 the 'length' field is signed on one version and unsigned on another
11421 with no (?) obvious way to tell the difference, without these VC++
11422 complains. Also the "definition" of FAR (blank) is no longer included
11423 nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up
11424 some needed definitions.
11428 * Undo Cygwin change.
11432 * Added support for proxy certificates according to RFC 3820.
11433 Because they may be a security thread to unaware applications,
11434 they must be explicitly allowed in run-time. See
11435 docs/HOWTO/proxy_certificates.txt for further information.
11439 ### Changes between 0.9.7e and 0.9.7f [22 Mar 2005]
11441 * Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating
11442 server and client random values. Previously
11443 (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in
11444 less random data when sizeof(time_t) > 4 (some 64 bit platforms).
11446 This change has negligible security impact because:
11448 1. Server and client random values still have 24 bytes of pseudo random
11451 2. Server and client random values are sent in the clear in the initial
11454 3. The master secret is derived using the premaster secret (48 bytes in
11455 size for static RSA ciphersuites) as well as client server and random
11458 The OpenSSL team would like to thank the UK NISCC for bringing this issue
11461 *Stephen Henson, reported by UK NISCC*
11463 * Use Windows randomness collection on Cygwin.
11467 * Fix hang in EGD/PRNGD query when communication socket is closed
11468 prematurely by EGD/PRNGD.
11470 *Darren Tucker <dtucker@zip.com.au> via Lutz Jänicke, resolves #1014*
11472 * Prompt for pass phrases when appropriate for PKCS12 input format.
11476 * Back-port of selected performance improvements from development
11477 branch, as well as improved support for PowerPC platforms.
11481 * Add lots of checks for memory allocation failure, error codes to indicate
11482 failure and freeing up memory if a failure occurs.
11484 *Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson*
11486 * Add new -passin argument to dgst.
11490 * Perform some character comparisons of different types in X509_NAME_cmp:
11491 this is needed for some certificates that re-encode DNs into UTF8Strings
11492 (in violation of RFC3280) and can't or won't issue name rollover
11497 * Make an explicit check during certificate validation to see that
11498 the CA setting in each certificate on the chain is correct. As a
11499 side effect always do the following basic checks on extensions,
11500 not just when there's an associated purpose to the check:
11502 - if there is an unhandled critical extension (unless the user
11503 has chosen to ignore this fault)
11504 - if the path length has been exceeded (if one is set at all)
11505 - that certain extensions fit the associated purpose (if one has
11510 ### Changes between 0.9.7d and 0.9.7e [25 Oct 2004]
11512 * Avoid a race condition when CRLs are checked in a multi threaded
11513 environment. This would happen due to the reordering of the revoked
11514 entries during signature checking and serial number lookup. Now the
11515 encoding is cached and the serial number sort performed under a lock.
11516 Add new STACK function sk_is_sorted().
11520 * Add Delta CRL to the extension code.
11524 * Various fixes to s3_pkt.c so alerts are sent properly.
11526 *David Holmes <d.holmes@f5.com>*
11528 * Reduce the chances of duplicate issuer name and serial numbers (in
11529 violation of RFC3280) using the OpenSSL certificate creation utilities.
11530 This is done by creating a random 64 bit value for the initial serial
11531 number when a serial number file is created or when a self signed
11532 certificate is created using 'openssl req -x509'. The initial serial
11533 number file is created using 'openssl x509 -next_serial' in CA.pl
11534 rather than being initialized to 1.
11538 ### Changes between 0.9.7c and 0.9.7d [17 Mar 2004]
11540 * Fix null-pointer assignment in do_change_cipher_spec() revealed
11541 by using the Codenomicon TLS Test Tool ([CVE-2004-0079])
11543 *Joe Orton, Steve Henson*
11545 * Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites
11548 *Joe Orton, Steve Henson*
11550 * Make it possible to have multiple active certificates with the same
11551 subject in the CA index file. This is done only if the keyword
11552 'unique_subject' is set to 'no' in the main CA section (default
11553 if 'CA_default') of the configuration file. The value is saved
11554 with the database itself in a separate index attribute file,
11555 named like the index file with '.attr' appended to the name.
11559 * X509 verify fixes. Disable broken certificate workarounds when
11560 X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if
11561 keyUsage extension present. Don't accept CRLs with unhandled critical
11562 extensions: since verify currently doesn't process CRL extensions this
11563 rejects a CRL with *any* critical extensions. Add new verify error codes
11568 * When creating an OCSP nonce use an OCTET STRING inside the extnValue.
11569 A clarification of RFC2560 will require the use of OCTET STRINGs and
11570 some implementations cannot handle the current raw format. Since OpenSSL
11571 copies and compares OCSP nonces as opaque blobs without any attempt at
11572 parsing them this should not create any compatibility issues.
11576 * New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when
11577 calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without
11578 this HMAC (and other) operations are several times slower than OpenSSL
11583 * Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex().
11585 *Peter Sylvester <Peter.Sylvester@EdelWeb.fr>*
11587 * Use the correct content when signing type "other".
11591 ### Changes between 0.9.7b and 0.9.7c [30 Sep 2003]
11593 * Fix various bugs revealed by running the NISCC test suite:
11595 Stop out of bounds reads in the ASN1 code when presented with
11596 invalid tags (CVE-2003-0543 and CVE-2003-0544).
11598 Free up ASN1_TYPE correctly if ANY type is invalid ([CVE-2003-0545]).
11600 If verify callback ignores invalid public key errors don't try to check
11601 certificate signature with the NULL public key.
11605 * New -ignore_err option in ocsp application to stop the server
11606 exiting on the first error in a request.
11610 * In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
11611 if the server requested one: as stated in TLS 1.0 and SSL 3.0
11616 * In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
11617 extra data after the compression methods not only for TLS 1.0
11618 but also for SSL 3.0 (as required by the specification).
11620 *Bodo Moeller; problem pointed out by Matthias Loepfe*
11622 * Change X509_certificate_type() to mark the key as exported/exportable
11623 when it's 512 *bits* long, not 512 bytes.
11627 * Change AES_cbc_encrypt() so it outputs exact multiple of
11628 blocks during encryption.
11632 * Various fixes to base64 BIO and non blocking I/O. On write
11633 flushes were not handled properly if the BIO retried. On read
11634 data was not being buffered properly and had various logic bugs.
11635 This also affects blocking I/O when the data being decoded is a
11640 * Various S/MIME bugfixes and compatibility changes:
11641 output correct application/pkcs7 MIME type if
11642 PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures.
11643 Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening
11644 of files as .eml work). Correctly handle very long lines in MIME
11649 ### Changes between 0.9.7a and 0.9.7b [10 Apr 2003]
11651 * Countermeasure against the Klima-Pokorny-Rosa extension of
11652 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
11653 a protocol version number mismatch like a decryption error
11654 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
11658 * Turn on RSA blinding by default in the default implementation
11659 to avoid a timing attack. Applications that don't want it can call
11660 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
11661 They would be ill-advised to do so in most cases.
11663 *Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller*
11665 * Change RSA blinding code so that it works when the PRNG is not
11666 seeded (in this case, the secret RSA exponent is abused as
11667 an unpredictable seed -- if it is not unpredictable, there
11668 is no point in blinding anyway). Make RSA blinding thread-safe
11669 by remembering the creator's thread ID in rsa->blinding and
11670 having all other threads use local one-time blinding factors
11671 (this requires more computation than sharing rsa->blinding, but
11672 avoids excessive locking; and if an RSA object is not shared
11673 between threads, blinding will still be very fast).
11677 * Fixed a typo bug that would cause ENGINE_set_default() to set an
11678 ENGINE as defaults for all supported algorithms irrespective of
11679 the 'flags' parameter. 'flags' is now honoured, so applications
11680 should make sure they are passing it correctly.
11684 * Target "mingw" now allows native Windows code to be generated in
11685 the Cygwin environment as well as with the MinGW compiler.
11689 ### Changes between 0.9.7 and 0.9.7a [19 Feb 2003]
11691 * In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
11692 via timing by performing a MAC computation even if incorrect
11693 block cipher padding has been found. This is a countermeasure
11694 against active attacks where the attacker has to distinguish
11695 between bad padding and a MAC verification error. ([CVE-2003-0078])
11697 *Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
11698 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
11699 Martin Vuagnoux (EPFL, Ilion)*
11701 * Make the no-err option work as intended. The intention with no-err
11702 is not to have the whole error stack handling routines removed from
11703 libcrypto, it's only intended to remove all the function name and
11704 reason texts, thereby removing some of the footprint that may not
11705 be interesting if those errors aren't displayed anyway.
11707 NOTE: it's still possible for any application or module to have its
11708 own set of error texts inserted. The routines are there, just not
11709 used by default when no-err is given.
11713 * Add support for FreeBSD on IA64.
11715 *dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454*
11717 * Adjust DES_cbc_cksum() so it returns the same value as the MIT
11718 Kerberos function mit_des_cbc_cksum(). Before this change,
11719 the value returned by DES_cbc_cksum() was like the one from
11720 mit_des_cbc_cksum(), except the bytes were swapped.
11722 *Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte*
11724 * Allow an application to disable the automatic SSL chain building.
11725 Before this a rather primitive chain build was always performed in
11726 ssl3_output_cert_chain(): an application had no way to send the
11727 correct chain if the automatic operation produced an incorrect result.
11729 Now the chain builder is disabled if either:
11731 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert().
11733 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set.
11735 The reasoning behind this is that an application would not want the
11736 auto chain building to take place if extra chain certificates are
11737 present and it might also want a means of sending no additional
11738 certificates (for example the chain has two certificates and the
11743 * Add the possibility to build without the ENGINE framework.
11745 *Steven Reddie <smr@essemer.com.au> via Richard Levitte*
11747 * Under Win32 gmtime() can return NULL: check return value in
11748 OPENSSL_gmtime(). Add error code for case where gmtime() fails.
11752 * DSA routines: under certain error conditions uninitialized BN objects
11753 could be freed. Solution: make sure initialization is performed early
11754 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>,
11755 Nils Larsch <nla@trustcenter.de> via PR#459)
11759 * Another fix for SSLv2 session ID handling: the session ID was incorrectly
11760 checked on reconnect on the client side, therefore session resumption
11761 could still fail with a "ssl session id is different" error. This
11762 behaviour is masked when SSL_OP_ALL is used due to
11763 SSL_OP_MICROSOFT_SESS_ID_BUG being set.
11764 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
11765 followup to PR #377.
11769 * IA-32 assembler support enhancements: unified ELF targets, support
11770 for SCO/Caldera platforms, fix for Cygwin shared build.
11774 * Add support for FreeBSD on sparc64. As a consequence, support for
11775 FreeBSD on non-x86 processors is separate from x86 processors on
11776 the config script, much like the NetBSD support.
11778 *Richard Levitte & Kris Kennaway <kris@obsecurity.org>*
11780 ### Changes between 0.9.6h and 0.9.7 [31 Dec 2002]
11782 [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after
11785 * Fix session ID handling in SSLv2 client code: the SERVER FINISHED
11786 code (06) was taken as the first octet of the session ID and the last
11787 octet was ignored consequently. As a result SSLv2 client side session
11788 caching could not have worked due to the session ID mismatch between
11790 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
11795 * Change the declaration of needed Kerberos libraries to use EX_LIBS
11796 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is
11801 * The hw_ncipher.c engine requires dynamic locks. Unfortunately, it
11802 seems that in spite of existing for more than a year, many application
11803 author have done nothing to provide the necessary callbacks, which
11804 means that this particular engine will not work properly anywhere.
11805 This is a very unfortunate situation which forces us, in the name
11806 of usability, to give the hw_ncipher.c a static lock, which is part
11808 NOTE: This is for the 0.9.7 series ONLY. This hack will never
11809 appear in 0.9.8 or later. We EXPECT application authors to have
11810 dealt properly with this when 0.9.8 is released (unless we actually
11811 make such changes in the libcrypto locking code that changes will
11812 have to be made anyway).
11816 * In asn1_d2i_read_bio() repeatedly call BIO_read() until all content
11817 octets have been read, EOF or an error occurs. Without this change
11818 some truncated ASN1 structures will not produce an error.
11822 * Disable Heimdal support, since it hasn't been fully implemented.
11823 Still give the possibility to force the use of Heimdal, but with
11824 warnings and a request that patches get sent to openssl-dev.
11828 * Add the VC-CE target, introduce the WINCE sysname, and add
11829 INSTALL.WCE and appropriate conditionals to make it build.
11831 *Steven Reddie <smr@essemer.com.au> via Richard Levitte*
11833 * Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and
11834 cygssl-x.y.z.dll, where x, y and z are the major, minor and
11835 edit numbers of the version.
11837 *Corinna Vinschen <vinschen@redhat.com> and Richard Levitte*
11839 * Introduce safe string copy and catenation functions
11840 (BUF_strlcpy() and BUF_strlcat()).
11842 *Ben Laurie (CHATS) and Richard Levitte*
11844 * Avoid using fixed-size buffers for one-line DNs.
11846 *Ben Laurie (CHATS)*
11848 * Add BUF_MEM_grow_clean() to avoid information leakage when
11849 resizing buffers containing secrets, and use where appropriate.
11851 *Ben Laurie (CHATS)*
11853 * Avoid using fixed size buffers for configuration file location.
11855 *Ben Laurie (CHATS)*
11857 * Avoid filename truncation for various CA files.
11859 *Ben Laurie (CHATS)*
11861 * Use sizeof in preference to magic numbers.
11863 *Ben Laurie (CHATS)*
11865 * Avoid filename truncation in cert requests.
11867 *Ben Laurie (CHATS)*
11869 * Add assertions to check for (supposedly impossible) buffer
11872 *Ben Laurie (CHATS)*
11874 * Don't cache truncated DNS entries in the local cache (this could
11875 potentially lead to a spoofing attack).
11877 *Ben Laurie (CHATS)*
11879 * Fix various buffers to be large enough for hex/decimal
11880 representations in a platform independent manner.
11882 *Ben Laurie (CHATS)*
11884 * Add CRYPTO_realloc_clean() to avoid information leakage when
11885 resizing buffers containing secrets, and use where appropriate.
11887 *Ben Laurie (CHATS)*
11889 * Add BIO_indent() to avoid much slightly worrying code to do
11892 *Ben Laurie (CHATS)*
11894 * Convert sprintf()/BIO_puts() to BIO_printf().
11896 *Ben Laurie (CHATS)*
11898 * buffer_gets() could terminate with the buffer only half
11901 *Ben Laurie (CHATS)*
11903 * Add assertions to prevent user-supplied crypto functions from
11904 overflowing internal buffers by having large block sizes, etc.
11906 *Ben Laurie (CHATS)*
11908 * New OPENSSL_assert() macro (similar to assert(), but enabled
11911 *Ben Laurie (CHATS)*
11913 * Eliminate unused copy of key in RC4.
11915 *Ben Laurie (CHATS)*
11917 * Eliminate unused and incorrectly sized buffers for IV in pem.h.
11919 *Ben Laurie (CHATS)*
11921 * Fix off-by-one error in EGD path.
11923 *Ben Laurie (CHATS)*
11925 * If RANDFILE path is too long, ignore instead of truncating.
11927 *Ben Laurie (CHATS)*
11929 * Eliminate unused and incorrectly sized X.509 structure
11932 *Ben Laurie (CHATS)*
11934 * Eliminate unused and dangerous function knumber().
11936 *Ben Laurie (CHATS)*
11938 * Eliminate unused and dangerous structure, KSSL_ERR.
11940 *Ben Laurie (CHATS)*
11942 * Protect against overlong session ID context length in an encoded
11943 session object. Since these are local, this does not appear to be
11946 *Ben Laurie (CHATS)*
11948 * Change from security patch (see 0.9.6e below) that did not affect
11949 the 0.9.6 release series:
11951 Remote buffer overflow in SSL3 protocol - an attacker could
11952 supply an oversized master key in Kerberos-enabled versions.
11955 *Ben Laurie (CHATS)*
11957 * Change the SSL kerb5 codes to match RFC 2712.
11961 * Make -nameopt work fully for req and add -reqopt switch.
11963 *Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson*
11965 * The "block size" for block ciphers in CFB and OFB mode should be 1.
11967 *Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>*
11969 * Make sure tests can be performed even if the corresponding algorithms
11970 have been removed entirely. This was also the last step to make
11971 OpenSSL compilable with DJGPP under all reasonable conditions.
11973 *Richard Levitte, Doug Kaufman <dkaufman@rahul.net>*
11975 * Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT
11976 to allow version independent disabling of normally unselected ciphers,
11977 which may be activated as a side-effect of selecting a single cipher.
11979 (E.g., cipher list string "RSA" enables ciphersuites that are left
11980 out of "ALL" because they do not provide symmetric encryption.
11981 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.)
11983 *Lutz Jaenicke, Bodo Moeller*
11985 * Add appropriate support for separate platform-dependent build
11986 directories. The recommended way to make a platform-dependent
11987 build directory is the following (tested on Linux), maybe with
11990 # Place yourself outside of the OpenSSL source tree. In
11991 # this example, the environment variable OPENSSL_SOURCE
11992 # is assumed to contain the absolute OpenSSL source directory.
11993 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`"
11994 cd objtree/"`uname -s`-`uname -r`-`uname -m`"
11995 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do
11996 mkdir -p `dirname $F`
11997 ln -s $OPENSSL_SOURCE/$F $F
12000 To be absolutely sure not to disturb the source tree, a "make clean"
12001 is a good thing. If it isn't successful, don't worry about it,
12002 it probably means the source directory is very clean.
12006 * Make sure any ENGINE control commands make local copies of string
12007 pointers passed to them whenever necessary. Otherwise it is possible
12008 the caller may have overwritten (or deallocated) the original string
12009 data when a later ENGINE operation tries to use the stored values.
12011 *Götz Babin-Ebell <babinebell@trustcenter.de>*
12013 * Improve diagnostics in file reading and command-line digests.
12015 *Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>*
12017 * Add AES modes CFB and OFB to the object database. Correct an
12018 error in AES-CFB decryption.
12022 * Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this
12023 allows existing EVP_CIPHER_CTX structures to be reused after
12024 calling `EVP_*Final()`. This behaviour is used by encryption
12025 BIOs and some applications. This has the side effect that
12026 applications must explicitly clean up cipher contexts with
12027 EVP_CIPHER_CTX_cleanup() or they will leak memory.
12031 * Check the values of dna and dnb in bn_mul_recursive before calling
12032 bn_mul_comba (a non zero value means the a or b arrays do not contain
12033 n2 elements) and fallback to bn_mul_normal if either is not zero.
12037 * Fix escaping of non-ASCII characters when using the -subj option
12038 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>)
12042 * Make object definitions compliant to LDAP (RFC2256): SN is the short
12043 form for "surname", serialNumber has no short form.
12044 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798;
12045 therefore remove "mail" short name for "internet 7".
12046 The OID for unique identifiers in X509 certificates is
12047 x500UniqueIdentifier, not uniqueIdentifier.
12048 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>)
12052 * Add an "init" command to the ENGINE config module and auto initialize
12053 ENGINEs. Without any "init" command the ENGINE will be initialized
12054 after all ctrl commands have been executed on it. If init=1 the
12055 ENGINE is initialized at that point (ctrls before that point are run
12056 on the uninitialized ENGINE and after on the initialized one). If
12057 init=0 then the ENGINE will not be initialized at all.
12061 * Fix the 'app_verify_callback' interface so that the user-defined
12062 argument is actually passed to the callback: In the
12063 SSL_CTX_set_cert_verify_callback() prototype, the callback
12064 declaration has been changed from
12067 int (*cb)(X509_STORE_CTX *,void *);
12068 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
12069 i=s->ctx->app_verify_callback(&ctx)
12070 has been changed into
12071 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
12073 To update applications using SSL_CTX_set_cert_verify_callback(),
12074 a dummy argument can be added to their callback functions.
12076 *D. K. Smetters <smetters@parc.xerox.com>*
12078 * Added the '4758cca' ENGINE to support IBM 4758 cards.
12080 *Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe*
12082 * Add and OPENSSL_LOAD_CONF define which will cause
12083 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
12084 This allows older applications to transparently support certain
12085 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
12086 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
12087 load the config file and OPENSSL_add_all_algorithms_conf() which will
12088 always load it have also been added.
12092 * Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
12093 Adjust NIDs and EVP layer.
12095 *Stephen Sprunk <stephen@sprunk.org> and Richard Levitte*
12097 * Config modules support in openssl utility.
12099 Most commands now load modules from the config file,
12100 though in a few (such as version) this isn't done
12101 because it couldn't be used for anything.
12103 In the case of ca and req the config file used is
12104 the same as the utility itself: that is the -config
12105 command line option can be used to specify an
12110 * Move default behaviour from OPENSSL_config(). If appname is NULL
12111 use "openssl_conf" if filename is NULL use default openssl config file.
12115 * Add an argument to OPENSSL_config() to allow the use of an alternative
12116 config section name. Add a new flag to tolerate a missing config file
12117 and move code to CONF_modules_load_file().
12121 * Support for crypto accelerator cards from Accelerated Encryption
12122 Processing, www.aep.ie. (Use engine 'aep')
12123 The support was copied from 0.9.6c [engine] and adapted/corrected
12124 to work with the new engine framework.
12126 *AEP Inc. and Richard Levitte*
12128 * Support for SureWare crypto accelerator cards from Baltimore
12129 Technologies. (Use engine 'sureware')
12130 The support was copied from 0.9.6c [engine] and adapted
12131 to work with the new engine framework.
12135 * Have the CHIL engine fork-safe (as defined by nCipher) and actually
12136 make the newer ENGINE framework commands for the CHIL engine work.
12138 *Toomas Kiisk <vix@cyber.ee> and Richard Levitte*
12140 * Make it possible to produce shared libraries on ReliantUNIX.
12142 *Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte*
12144 * Add the configuration target debug-linux-ppro.
12145 Make 'openssl rsa' use the general key loading routines
12146 implemented in `apps.c`, and make those routines able to
12147 handle the key format FORMAT_NETSCAPE and the variant
12150 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
12152 * Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
12154 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
12156 * Add -keyform to rsautl, and document -engine.
12158 *Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>*
12160 * Change BIO_new_file (crypto/bio/bss_file.c) to use new
12161 BIO_R_NO_SUCH_FILE error code rather than the generic
12162 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
12166 * Add new functions
12167 ERR_peek_last_error
12168 ERR_peek_last_error_line
12169 ERR_peek_last_error_line_data.
12170 These are similar to
12172 ERR_peek_error_line
12173 ERR_peek_error_line_data,
12174 but report on the latest error recorded rather than the first one
12175 still in the error queue.
12177 *Ben Laurie, Bodo Moeller*
12179 * default_algorithms option in ENGINE config module. This allows things
12181 default_algorithms = ALL
12182 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
12186 * Preliminary ENGINE config module.
12190 * New experimental application configuration code.
12194 * Change the AES code to follow the same name structure as all other
12195 symmetric ciphers, and behave the same way. Move everything to
12196 the directory crypto/aes, thereby obsoleting crypto/rijndael.
12198 *Stephen Sprunk <stephen@sprunk.org> and Richard Levitte*
12200 * SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
12202 *Ben Laurie and Theo de Raadt*
12204 * Add option to output public keys in req command.
12206 *Massimiliano Pala madwolf@openca.org*
12208 * Use wNAFs in EC_POINTs_mul() for improved efficiency
12209 (up to about 10% better than before for P-192 and P-224).
12213 * New functions/macros
12215 SSL_CTX_set_msg_callback(ctx, cb)
12216 SSL_CTX_set_msg_callback_arg(ctx, arg)
12217 SSL_set_msg_callback(ssl, cb)
12218 SSL_set_msg_callback_arg(ssl, arg)
12220 to request calling a callback function
12222 void cb(int write_p, int version, int content_type,
12223 const void *buf, size_t len, SSL *ssl, void *arg)
12225 whenever a protocol message has been completely received
12226 (write_p == 0) or sent (write_p == 1). Here 'version' is the
12227 protocol version according to which the SSL library interprets
12228 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
12229 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
12230 the content type as defined in the SSL 3.0/TLS 1.0 protocol
12231 specification (change_cipher_spec(20), alert(21), handshake(22)).
12232 'buf' and 'len' point to the actual message, 'ssl' to the
12233 SSL object, and 'arg' is the application-defined value set by
12234 SSL[_CTX]_set_msg_callback_arg().
12236 'openssl s_client' and 'openssl s_server' have new '-msg' options
12237 to enable a callback that displays all protocol messages.
12241 * Change the shared library support so shared libraries are built as
12242 soon as the corresponding static library is finished, and thereby get
12243 openssl and the test programs linked against the shared library.
12244 This still only happens when the keyword "shard" has been given to
12245 the configuration scripts.
12247 NOTE: shared library support is still an experimental thing, and
12248 backward binary compatibility is still not guaranteed.
12250 *"Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte*
12252 * Add support for Subject Information Access extension.
12254 *Peter Sylvester <Peter.Sylvester@EdelWeb.fr>*
12256 * Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
12257 additional bytes when new memory had to be allocated, not just
12258 when reusing an existing buffer.
12262 * New command line and configuration option 'utf8' for the req command.
12263 This allows field values to be specified as UTF8 strings.
12267 * Add -multi and -mr options to "openssl speed" - giving multiple parallel
12268 runs for the former and machine-readable output for the latter.
12272 * Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
12273 of the e-mail address in the DN (i.e., it will go into a certificate
12274 extension only). The new configuration file option 'email_in_dn = no'
12275 has the same effect.
12277 *Massimiliano Pala madwolf@openca.org*
12279 * Change all functions with names starting with `des_` to be starting
12280 with `DES_` instead. Add wrappers that are compatible with libdes,
12281 but are named `_ossl_old_des_*`. Finally, add macros that map the
12282 `des_*` symbols to the corresponding `_ossl_old_des_*` if libdes
12283 compatibility is desired. If OpenSSL 0.9.6c compatibility is
12284 desired, the `des_*` symbols will be mapped to `DES_*`, with one
12287 Since we provide two compatibility mappings, the user needs to
12288 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
12289 compatibility is desired. The default (i.e., when that macro
12290 isn't defined) is OpenSSL 0.9.6c compatibility.
12292 There are also macros that enable and disable the support of old
12293 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
12294 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
12295 are defined, the default will apply: to support the old des routines.
12297 In either case, one must include openssl/des.h to get the correct
12298 definitions. Do not try to just include openssl/des_old.h, that
12301 NOTE: This is a major break of an old API into a new one. Software
12302 authors are encouraged to switch to the `DES_` style functions. Some
12303 time in the future, des_old.h and the libdes compatibility functions
12304 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
12305 default), and then completely removed.
12309 * Test for certificates which contain unsupported critical extensions.
12310 If such a certificate is found during a verify operation it is
12311 rejected by default: this behaviour can be overridden by either
12312 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
12313 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
12314 X509_supported_extension() has also been added which returns 1 if a
12315 particular extension is supported.
12319 * Modify the behaviour of EVP cipher functions in similar way to digests
12320 to retain compatibility with existing code.
12324 * Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
12325 compatibility with existing code. In particular the 'ctx' parameter does
12326 not have to be to be initialized before the call to EVP_DigestInit() and
12327 it is tidied up after a call to EVP_DigestFinal(). New function
12328 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
12329 EVP_MD_CTX_copy() changed to not require the destination to be
12330 initialized valid and new function EVP_MD_CTX_copy_ex() added which
12331 requires the destination to be valid.
12333 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
12334 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
12338 * Change ssl3_get_message (ssl/s3_both.c) and the functions using it
12339 so that complete 'Handshake' protocol structures are kept in memory
12340 instead of overwriting 'msg_type' and 'length' with 'body' data.
12344 * Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
12346 *Massimo Santin via Richard Levitte*
12348 * Major restructuring to the underlying ENGINE code. This includes
12349 reduction of linker bloat, separation of pure "ENGINE" manipulation
12350 (initialisation, etc) from functionality dealing with implementations
12351 of specific crypto interfaces. This change also introduces integrated
12352 support for symmetric ciphers and digest implementations - so ENGINEs
12353 can now accelerate these by providing EVP_CIPHER and EVP_MD
12354 implementations of their own. This is detailed in
12355 [crypto/engine/README.md](crypto/engine/README.md)
12356 as it couldn't be adequately described here. However, there are a few
12357 API changes worth noting - some RSA, DSA, DH, and RAND functions that
12358 were changed in the original introduction of ENGINE code have now
12359 reverted back - the hooking from this code to ENGINE is now a good
12360 deal more passive and at run-time, operations deal directly with
12361 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
12362 dereferencing through an ENGINE pointer any more. Also, the ENGINE
12363 functions dealing with `BN_MOD_EXP[_CRT]` handlers have been removed -
12364 they were not being used by the framework as there is no concept of a
12365 BIGNUM_METHOD and they could not be generalised to the new
12366 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
12367 ENGINE_cpy() has been removed as it cannot be consistently defined in
12372 * Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
12376 * Change mkdef.pl to sort symbols that get the same entry number,
12377 and make sure the automatically generated functions `ERR_load_*`
12378 become part of libeay.num as well.
12382 * New function SSL_renegotiate_pending(). This returns true once
12383 renegotiation has been requested (either SSL_renegotiate() call
12384 or HelloRequest/ClientHello received from the peer) and becomes
12385 false once a handshake has been completed.
12386 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
12387 sends a HelloRequest, but does not ensure that a handshake takes
12388 place. SSL_renegotiate_pending() is useful for checking if the
12389 client has followed the request.)
12393 * New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
12394 By default, clients may request session resumption even during
12395 renegotiation (if session ID contexts permit); with this option,
12396 session resumption is possible only in the first handshake.
12398 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes
12399 more bits available for options that should not be part of
12400 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION).
12404 * Add some demos for certificate and certificate request creation.
12408 * Make maximum certificate chain size accepted from the peer application
12409 settable (`SSL*_get/set_max_cert_list()`), as proposed by
12410 "Douglas E. Engert" <deengert@anl.gov>.
12414 * Add support for shared libraries for Unixware-7
12415 (Boyd Lynn Gerber <gerberb@zenez.com>).
12419 * Add a "destroy" handler to ENGINEs that allows structural cleanup to
12420 be done prior to destruction. Use this to unload error strings from
12421 ENGINEs that load their own error strings. NB: This adds two new API
12422 functions to "get" and "set" this destroy handler in an ENGINE.
12426 * Alter all existing ENGINE implementations (except "openssl" and
12427 "openbsd") to dynamically instantiate their own error strings. This
12428 makes them more flexible to be built both as statically-linked ENGINEs
12429 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
12430 Also, add stub code to each that makes building them as self-contained
12431 shared-libraries easier (see [README-Engine.md](README-Engine.md)).
12435 * Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
12436 implementations into applications that are completely implemented in
12437 self-contained shared-libraries. The "dynamic" ENGINE exposes control
12438 commands that can be used to configure what shared-library to load and
12439 to control aspects of the way it is handled. Also, made an update to
12440 the [README-Engine.md](README-Engine.md) file
12441 that brings its information up-to-date and
12442 provides some information and instructions on the "dynamic" ENGINE
12443 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
12447 * Make it possible to unload ranges of ERR strings with a new
12448 "ERR_unload_strings" function.
12452 * Add a copy() function to EVP_MD.
12456 * Make EVP_MD routines take a context pointer instead of just the
12457 md_data void pointer.
12461 * Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
12462 that the digest can only process a single chunk of data
12463 (typically because it is provided by a piece of
12464 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
12465 is only going to provide a single chunk of data, and hence the
12466 framework needn't accumulate the data for oneshot drivers.
12470 * As with "ERR", make it possible to replace the underlying "ex_data"
12471 functions. This change also alters the storage and management of global
12472 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
12473 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
12474 index counters. The API functions that use this state have been changed
12475 to take a "class_index" rather than pointers to the class's local STACK
12476 and counter, and there is now an API function to dynamically create new
12477 classes. This centralisation allows us to (a) plug a lot of the
12478 thread-safety problems that existed, and (b) makes it possible to clean
12479 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
12480 such data would previously have always leaked in application code and
12481 workarounds were in place to make the memory debugging turn a blind eye
12482 to it. Application code that doesn't use this new function will still
12483 leak as before, but their memory debugging output will announce it now
12484 rather than letting it slide.
12486 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
12487 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
12488 has a return value to indicate success or failure.
12492 * Make it possible to replace the underlying "ERR" functions such that the
12493 global state (2 LHASH tables and 2 locks) is only used by the "default"
12494 implementation. This change also adds two functions to "get" and "set"
12495 the implementation prior to it being automatically set the first time
12496 any other ERR function takes place. Ie. an application can call "get",
12497 pass the return value to a module it has just loaded, and that module
12498 can call its own "set" function using that value. This means the
12499 module's "ERR" operations will use (and modify) the error state in the
12500 application and not in its own statically linked copy of OpenSSL code.
12504 * Give DH, DSA, and RSA types their own `*_up_ref()` function to increment
12505 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
12506 the operation, and provides a more encapsulated way for external code
12507 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
12508 to use these functions rather than manually incrementing the counts.
12510 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
12514 * Add EVP test program.
12518 * Add symmetric cipher support to ENGINE. Expect the API to change!
12522 * New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
12523 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
12524 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
12525 These allow a CRL to be built without having to access X509_CRL fields
12526 directly. Modify 'ca' application to use new functions.
12530 * Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
12531 bug workarounds. Rollback attack detection is a security feature.
12532 The problem will only arise on OpenSSL servers when TLSv1 is not
12533 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
12534 Software authors not wanting to support TLSv1 will have special reasons
12535 for their choice and can explicitly enable this option.
12537 *Bodo Moeller, Lutz Jaenicke*
12539 * Rationalise EVP so it can be extended: don't include a union of
12540 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX
12541 (similar to those existing for EVP_CIPHER_CTX).
12546 EVP_MD_CTX_init(&md); /* new function call */
12547 EVP_DigestInit(&md, EVP_sha1());
12548 EVP_DigestUpdate(&md, in, len);
12549 EVP_DigestFinal(&md, out, NULL);
12550 EVP_MD_CTX_cleanup(&md); /* new function call */
12554 * Make DES key schedule conform to the usual scheme, as well as
12555 correcting its structure. This means that calls to DES functions
12556 now have to pass a pointer to a des_key_schedule instead of a
12557 plain des_key_schedule (which was actually always a pointer
12560 des_key_schedule ks;
12562 des_set_key_checked(..., &ks);
12563 des_ncbc_encrypt(..., &ks, ...);
12565 (Note that a later change renames 'des_...' into 'DES_...'.)
12569 * Initial reduction of linker bloat: the use of some functions, such as
12570 PEM causes large amounts of unused functions to be linked in due to
12571 poor organisation. For example pem_all.c contains every PEM function
12572 which has a knock on effect of linking in large amounts of (unused)
12573 ASN1 code. Grouping together similar functions and splitting unrelated
12574 functions prevents this.
12578 * Cleanup of EVP macros.
12582 * Change historical references to `{NID,SN,LN}_des_ede` and ede3 to add the
12583 correct `_ecb suffix`.
12587 * Add initial OCSP responder support to ocsp application. The
12588 revocation information is handled using the text based index
12589 use by the ca application. The responder can either handle
12590 requests generated internally, supplied in files (for example
12591 via a CGI script) or using an internal minimal server.
12595 * Add configuration choices to get zlib compression for TLS.
12599 * Changes to Kerberos SSL for RFC 2712 compliance:
12600 1. Implemented real KerberosWrapper, instead of just using
12601 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
12602 2. Implemented optional authenticator field of KerberosWrapper.
12604 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
12605 and authenticator structs; see crypto/krb5/.
12607 Generalized Kerberos calls to support multiple Kerberos libraries.
12608 *Vern Staats <staatsvr@asc.hpc.mil>, Jeffrey Altman <jaltman@columbia.edu>
12609 via Richard Levitte*
12611 * Cause 'openssl speed' to use fully hard-coded DSA keys as it
12612 already does with RSA. testdsa.h now has 'priv_key/pub_key'
12613 values for each of the key sizes rather than having just
12614 parameters (and 'speed' generating keys each time).
12618 * Speed up EVP routines.
12621 pe 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
12622 s-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
12623 s-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
12624 s-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
12626 s-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
12627 s-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
12628 s-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
12631 s-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
12633 s-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
12637 * Added the OS2-EMX target.
12639 *"Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte*
12641 * Rewrite commands to use `NCONF` routines instead of the old `CONF`.
12642 New functions to support `NCONF` routines in extension code.
12643 New function `CONF_set_nconf()`
12644 to allow functions which take an `NCONF` to also handle the old `LHASH`
12645 structure: this means that the old `CONF` compatible routines can be
12646 retained (in particular w.rt. extensions) without having to duplicate the
12647 code. New function `X509V3_add_ext_nconf_sk()` to add extensions to a stack.
12651 * Enhance the general user interface with mechanisms for inner control
12652 and with possibilities to have yes/no kind of prompts.
12656 * Change all calls to low-level digest routines in the library and
12657 applications to use EVP. Add missing calls to HMAC_cleanup() and
12658 don't assume HMAC_CTX can be copied using memcpy().
12660 *Verdon Walker <VWalker@novell.com>, Steve Henson*
12662 * Add the possibility to control engines through control names but with
12663 arbitrary arguments instead of just a string.
12664 Change the key loaders to take a UI_METHOD instead of a callback
12665 function pointer. NOTE: this breaks binary compatibility with earlier
12666 versions of OpenSSL [engine].
12667 Adapt the nCipher code for these new conditions and add a card insertion
12672 * Enhance the general user interface with mechanisms to better support
12673 dialog box interfaces, application-defined prompts, the possibility
12674 to use defaults (for example default passwords from somewhere else)
12675 and interrupts/cancellations.
12679 * Tidy up PKCS#12 attribute handling. Add support for the CSP name
12680 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
12684 * Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
12685 tidy up some unnecessarily weird code in 'sk_new()').
12687 *Geoff, reported by Diego Tartara <dtartara@novamens.com>*
12689 * Change the key loading routines for ENGINEs to use the same kind
12690 callback (pem_password_cb) as all other routines that need this
12695 * Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
12696 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
12697 than this minimum value is recommended.
12701 * New random seeder for OpenVMS, using the system process statistics
12702 that are easily reachable.
12706 * Windows apparently can't transparently handle global
12707 variables defined in DLLs. Initialisations such as:
12709 const ASN1_ITEM *it = &ASN1_INTEGER_it;
12711 won't compile. This is used by the any applications that need to
12712 declare their own ASN1 modules. This was fixed by adding the option
12713 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
12714 needed for static libraries under Win32.
12718 * New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
12719 setting of purpose and trust fields. New X509_STORE trust and
12720 purpose functions and tidy up setting in other SSL functions.
12724 * Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
12725 structure. These are inherited by X509_STORE_CTX when it is
12726 initialised. This allows various defaults to be set in the
12727 X509_STORE structure (such as flags for CRL checking and custom
12728 purpose or trust settings) for functions which only use X509_STORE_CTX
12729 internally such as S/MIME.
12731 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
12732 trust settings if they are not set in X509_STORE. This allows X509_STORE
12733 purposes and trust (in S/MIME for example) to override any set by default.
12735 Add command line options for CRL checking to smime, s_client and s_server
12740 * Initial CRL based revocation checking. If the CRL checking flag(s)
12741 are set then the CRL is looked up in the X509_STORE structure and
12742 its validity and signature checked, then if the certificate is found
12743 in the CRL the verify fails with a revoked error.
12745 Various new CRL related callbacks added to X509_STORE_CTX structure.
12747 Command line options added to 'verify' application to support this.
12749 This needs some additional work, such as being able to handle multiple
12750 CRLs with different times, extension based lookup (rather than just
12751 by subject name) and ultimately more complete V2 CRL extension
12756 * Add a general user interface API (crypto/ui/). This is designed
12757 to replace things like des_read_password and friends (backward
12758 compatibility functions using this new API are provided).
12759 The purpose is to remove prompting functions from the DES code
12760 section as well as provide for prompting through dialog boxes in
12761 a window system and the like.
12765 * Add "ex_data" support to ENGINE so implementations can add state at a
12766 per-structure level rather than having to store it globally.
12770 * Make it possible for ENGINE structures to be copied when retrieved by
12771 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
12772 This causes the "original" ENGINE structure to act like a template,
12773 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
12774 operational state can be localised to each ENGINE structure, despite the
12775 fact they all share the same "methods". New ENGINE structures returned in
12776 this case have no functional references and the return value is the single
12777 structural reference. This matches the single structural reference returned
12778 by ENGINE_by_id() normally, when it is incremented on the pre-existing
12783 * Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
12784 needs to match any other type at all we need to manually clear the
12789 * Changes to the "openssl engine" utility to include;
12790 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
12791 about an ENGINE's available control commands.
12792 - executing control commands from command line arguments using the
12793 '-pre' and '-post' switches. '-post' is only used if '-t' is
12794 specified and the ENGINE is successfully initialised. The syntax for
12795 the individual commands are colon-separated, for example;
12796 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
12800 * New dynamic control command support for ENGINEs. ENGINEs can now
12801 declare their own commands (numbers), names (strings), descriptions,
12802 and input types for run-time discovery by calling applications. A
12803 subset of these commands are implicitly classed as "executable"
12804 depending on their input type, and only these can be invoked through
12805 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
12806 can be based on user input, config files, etc). The distinction is
12807 that "executable" commands cannot return anything other than a boolean
12808 result and can only support numeric or string input, whereas some
12809 discoverable commands may only be for direct use through
12810 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
12811 pointers, or other custom uses. The "executable" commands are to
12812 support parameterisations of ENGINE behaviour that can be
12813 unambiguously defined by ENGINEs and used consistently across any
12814 OpenSSL-based application. Commands have been added to all the
12815 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
12816 control over shared-library paths without source code alterations.
12820 * Changed all ENGINE implementations to dynamically allocate their
12821 ENGINEs rather than declaring them statically. Apart from this being
12822 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
12823 this also allows the implementations to compile without using the
12824 internal engine_int.h header.
12828 * Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
12829 'const' value. Any code that should be able to modify a RAND_METHOD
12830 should already have non-const pointers to it (ie. they should only
12831 modify their own ones).
12835 * Made a variety of little tweaks to the ENGINE code.
12836 - "atalla" and "ubsec" string definitions were moved from header files
12837 to C code. "nuron" string definitions were placed in variables
12838 rather than hard-coded - allowing parameterisation of these values
12839 later on via ctrl() commands.
12840 - Removed unused "#if 0"'d code.
12841 - Fixed engine list iteration code so it uses ENGINE_free() to release
12842 structural references.
12843 - Constified the RAND_METHOD element of ENGINE structures.
12844 - Constified various get/set functions as appropriate and added
12845 missing functions (including a catch-all ENGINE_cpy that duplicates
12846 all ENGINE values onto a new ENGINE except reference counts/state).
12847 - Removed NULL parameter checks in get/set functions. Setting a method
12848 or function to NULL is a way of cancelling out a previously set
12849 value. Passing a NULL ENGINE parameter is just plain stupid anyway
12850 and doesn't justify the extra error symbols and code.
12851 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
12852 flags from engine_int.h to engine.h.
12853 - Changed prototypes for ENGINE handler functions (init(), finish(),
12854 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
12858 * Implement binary inversion algorithm for BN_mod_inverse in addition
12859 to the algorithm using long division. The binary algorithm can be
12860 used only if the modulus is odd. On 32-bit systems, it is faster
12861 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
12862 roughly 5-15% for 256-bit moduli), so we use it only for moduli
12863 up to 450 bits. In 64-bit environments, the binary algorithm
12864 appears to be advantageous for much longer moduli; here we use it
12865 for moduli up to 2048 bits.
12869 * Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
12870 could not support the combine flag in choice fields.
12874 * Add a 'copy_extensions' option to the 'ca' utility. This copies
12875 extensions from a certificate request to the certificate.
12879 * Allow multiple 'certopt' and 'nameopt' options to be separated
12880 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
12881 file: this allows the display of the certificate about to be
12882 signed to be customised, to allow certain fields to be included
12883 or excluded and extension details. The old system didn't display
12884 multicharacter strings properly, omitted fields not in the policy
12885 and couldn't display additional details such as extensions.
12889 * Function EC_POINTs_mul for multiple scalar multiplication
12890 of an arbitrary number of elliptic curve points
12891 \sum scalars[i]*points[i],
12892 optionally including the generator defined for the EC_GROUP:
12893 scalar*generator + \sum scalars[i]*points[i].
12895 EC_POINT_mul is a simple wrapper function for the typical case
12896 that the point list has just one item (besides the optional
12901 * First EC_METHODs for curves over GF(p):
12903 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
12904 operations and provides various method functions that can also
12905 operate with faster implementations of modular arithmetic.
12907 EC_GFp_mont_method() reuses most functions that are part of
12908 EC_GFp_simple_method, but uses Montgomery arithmetic.
12910 *Bodo Moeller; point addition and point doubling
12911 implementation directly derived from source code provided by
12912 Lenka Fibikova <fibikova@exp-math.uni-essen.de>*
12914 * Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
12915 crypto/ec/ec_lib.c):
12917 Curves are EC_GROUP objects (with an optional group generator)
12918 based on EC_METHODs that are built into the library.
12920 Points are EC_POINT objects based on EC_GROUP objects.
12922 Most of the framework would be able to handle curves over arbitrary
12923 finite fields, but as there are no obvious types for fields other
12924 than GF(p), some functions are limited to that for now.
12928 * Add the -HTTP option to s_server. It is similar to -WWW, but requires
12929 that the file contains a complete HTTP response.
12933 * Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
12934 change the def and num file printf format specifier from "%-40sXXX"
12935 to "%-39s XXX". The latter will always guarantee a space after the
12936 field while the former will cause them to run together if the field
12937 is 40 of more characters long.
12941 * Constify the cipher and digest 'method' functions and structures
12942 and modify related functions to take constant EVP_MD and EVP_CIPHER
12947 * Hide BN_CTX structure details in bn_lcl.h instead of publishing them
12948 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
12952 * Modify `EVP_Digest*()` routines so they now return values. Although the
12953 internal software routines can never fail additional hardware versions
12958 * Clean up crypto/err/err.h and change some error codes to avoid conflicts:
12960 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
12961 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
12964 ERR_R_NESTED_ASN1_ERROR
12966 ERR_R_MISSING_ASN1_EOS
12967 were 4 .. 9, conflicting with
12968 ERR_LIB_RSA (= ERR_R_RSA_LIB)
12970 ERR_LIB_PEM (= ERR_R_PEM_LIB).
12971 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
12973 Add new error code 'ERR_R_INTERNAL_ERROR'.
12977 * Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
12982 * New option '-subj arg' for 'openssl req' and 'openssl ca'. This
12983 sets the subject name for a new request or supersedes the
12984 subject name in a given request. Formats that can be parsed are
12985 'CN=Some Name, OU=myOU, C=IT'
12987 'CN=Some Name/OU=myOU/C=IT'.
12989 Add options '-batch' and '-verbose' to 'openssl req'.
12991 *Massimiliano Pala <madwolf@hackmasters.net>*
12993 * Introduce the possibility to access global variables through
12994 functions on platform were that's the best way to handle exporting
12995 global variables in shared libraries. To enable this functionality,
12996 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
12997 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
12998 is normally done by Configure or something similar).
13000 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
13001 in the source file (foo.c) like this:
13003 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
13004 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
13006 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
13007 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
13009 OPENSSL_DECLARE_GLOBAL(int,foo);
13010 #define foo OPENSSL_GLOBAL_REF(foo)
13011 OPENSSL_DECLARE_GLOBAL(double,bar);
13012 #define bar OPENSSL_GLOBAL_REF(bar)
13014 The #defines are very important, and therefore so is including the
13015 header file everywhere where the defined globals are used.
13017 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
13018 of ASN.1 items, but that structure is a bit different.
13020 The largest change is in util/mkdef.pl which has been enhanced with
13021 better and easier to understand logic to choose which symbols should
13022 go into the Windows .def files as well as a number of fixes and code
13023 cleanup (among others, algorithm keywords are now sorted
13024 lexicographically to avoid constant rewrites).
13028 * In BN_div() keep a copy of the sign of 'num' before writing the
13029 result to 'rm' because if rm==num the value will be overwritten
13030 and produce the wrong result if 'num' is negative: this caused
13031 problems with BN_mod() and BN_nnmod().
13035 * Function OCSP_request_verify(). This checks the signature on an
13036 OCSP request and verifies the signer certificate. The signer
13037 certificate is just checked for a generic purpose and OCSP request
13042 * Add OCSP_check_validity() function to check the validity of OCSP
13043 responses. OCSP responses are prepared in real time and may only
13044 be a few seconds old. Simply checking that the current time lies
13045 between thisUpdate and nextUpdate max reject otherwise valid responses
13046 caused by either OCSP responder or client clock inaccuracy. Instead
13047 we allow thisUpdate and nextUpdate to fall within a certain period of
13048 the current time. The age of the response can also optionally be
13049 checked. Two new options -validity_period and -status_age added to
13054 * If signature or public key algorithm is unrecognized print out its
13055 OID rather that just UNKNOWN.
13059 * Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
13060 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
13061 ID to be generated from the issuer certificate alone which can then be
13062 passed to OCSP_id_issuer_cmp().
13066 * New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
13067 ASN1 modules to export functions returning ASN1_ITEM pointers
13068 instead of the ASN1_ITEM structures themselves. This adds several
13069 new macros which allow the underlying ASN1 function/structure to
13070 be accessed transparently. As a result code should not use ASN1_ITEM
13071 references directly (such as &X509_it) but instead use the relevant
13072 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
13073 use of the new ASN1 code on platforms where exporting structures
13074 is problematical (for example in shared libraries) but exporting
13075 functions returning pointers to structures is not.
13079 * Add support for overriding the generation of SSL/TLS session IDs.
13080 These callbacks can be registered either in an SSL_CTX or per SSL.
13081 The purpose of this is to allow applications to control, if they wish,
13082 the arbitrary values chosen for use as session IDs, particularly as it
13083 can be useful for session caching in multiple-server environments. A
13084 command-line switch for testing this (and any client code that wishes
13085 to use such a feature) has been added to "s_server".
13087 *Geoff Thorpe, Lutz Jaenicke*
13089 * Modify mkdef.pl to recognise and parse preprocessor conditionals
13090 of the form `#if defined(...) || defined(...) || ...` and
13091 `#if !defined(...) && !defined(...) && ...`. This also avoids
13092 the growing number of special cases it was previously handling.
13096 * Make all configuration macros available for application by making
13097 sure they are available in opensslconf.h, by giving them names starting
13098 with `OPENSSL_` to avoid conflicts with other packages and by making
13099 sure e_os2.h will cover all platform-specific cases together with
13101 Additionally, it is now possible to define configuration/platform-
13102 specific names (called "system identities"). In the C code, these
13103 are prefixed with `OPENSSL_SYSNAME_`. e_os2.h will create another
13104 macro with the name beginning with `OPENSSL_SYS_`, which is determined
13105 from `OPENSSL_SYSNAME_*` or compiler-specific macros depending on
13110 * New option -set_serial to 'req' and 'x509' this allows the serial
13111 number to use to be specified on the command line. Previously self
13112 signed certificates were hard coded with serial number 0 and the
13113 CA options of 'x509' had to use a serial number in a file which was
13118 * New options to 'ca' utility to support V2 CRL entry extensions.
13119 Currently CRL reason, invalidity date and hold instruction are
13120 supported. Add new CRL extensions to V3 code and some new objects.
13124 * New function EVP_CIPHER_CTX_set_padding() this is used to
13125 disable standard block padding (aka PKCS#5 padding) in the EVP
13126 API, which was previously mandatory. This means that the data is
13127 not padded in any way and so the total length much be a multiple
13128 of the block size, otherwise an error occurs.
13132 * Initial (incomplete) OCSP SSL support.
13136 * New function OCSP_parse_url(). This splits up a URL into its host,
13137 port and path components: primarily to parse OCSP URLs. New -url
13138 option to ocsp utility.
13142 * New nonce behavior. The return value of OCSP_check_nonce() now
13143 reflects the various checks performed. Applications can decide
13144 whether to tolerate certain situations such as an absent nonce
13145 in a response when one was present in a request: the ocsp application
13146 just prints out a warning. New function OCSP_add1_basic_nonce()
13147 this is to allow responders to include a nonce in a response even if
13148 the request is nonce-less.
13152 * Disable stdin buffering in `load_cert()` (`apps/apps.c`) so that no certs are
13153 skipped when using openssl x509 multiple times on a single input file,
13154 e.g. `(openssl x509 -out cert1; openssl x509 -out cert2) <certs`.
13158 * Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
13159 set string type: to handle setting ASN1_TIME structures. Fix ca
13160 utility to correctly initialize revocation date of CRLs.
13164 * New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
13165 the clients preferred ciphersuites and rather use its own preferences.
13166 Should help to work around M$ SGC (Server Gated Cryptography) bug in
13167 Internet Explorer by ensuring unchanged hash method during stepup.
13168 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
13172 * Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
13173 to aes and add a new 'exist' option to print out symbols that don't
13178 * Additional options to ocsp utility to allow flags to be set and
13179 additional certificates supplied.
13183 * Add the option -VAfile to 'openssl ocsp', so the user can give the
13184 OCSP client a number of certificate to only verify the response
13189 * Update Rijndael code to version 3.0 and change EVP AES ciphers to
13190 handle the new API. Currently only ECB, CBC modes supported. Add new
13193 Add TLS AES ciphersuites as described in RFC3268, "Advanced
13194 Encryption Standard (AES) Ciphersuites for Transport Layer
13195 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were
13196 not enabled by default and were not part of the "ALL" ciphersuite
13197 alias because they were not yet official; they could be
13198 explicitly requested by specifying the "AESdraft" ciphersuite
13199 group alias. In the final release of OpenSSL 0.9.7, the group
13200 alias is called "AES" and is part of "ALL".)
13202 *Ben Laurie, Steve Henson, Bodo Moeller*
13204 * New function OCSP_copy_nonce() to copy nonce value (if present) from
13205 request to response.
13209 * Functions for OCSP responders. OCSP_request_onereq_count(),
13210 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
13211 extract information from a certificate request. OCSP_response_create()
13212 creates a response and optionally adds a basic response structure.
13213 OCSP_basic_add1_status() adds a complete single response to a basic
13214 response and returns the OCSP_SINGLERESP structure just added (to allow
13215 extensions to be included for example). OCSP_basic_add1_cert() adds a
13216 certificate to a basic response and OCSP_basic_sign() signs a basic
13217 response with various flags. New helper functions ASN1_TIME_check()
13218 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
13219 (converts ASN1_TIME to GeneralizedTime).
13223 * Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
13224 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
13225 structure from a certificate. X509_pubkey_digest() digests the public_key
13226 contents: this is used in various key identifiers.
13230 * Make sk_sort() tolerate a NULL argument.
13232 *Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>*
13234 * New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
13235 passed by the function are trusted implicitly. If any of them signed the
13236 response then it is assumed to be valid and is not verified.
13240 * In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
13241 to data. This was previously part of the PKCS7 ASN1 code. This
13242 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
13243 *Steve Henson, reported by Kenneth R. Robinette
13244 <support@securenetterm.com>*
13246 * Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
13247 routines: without these tracing memory leaks is very painful.
13248 Fix leaks in PKCS12 and PKCS7 routines.
13252 * Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
13253 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
13254 effectively meant GeneralizedTime would never be used. Now it
13255 is initialised to -1 but X509_time_adj() now has to check the value
13256 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
13257 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
13258 *Steve Henson, reported by Kenneth R. Robinette
13259 <support@securenetterm.com>*
13261 * Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
13262 result in a zero length in the ASN1_INTEGER structure which was
13263 not consistent with the structure when d2i_ASN1_INTEGER() was used
13264 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
13265 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
13266 where it did not print out a minus for negative ASN1_INTEGER.
13270 * Add summary printout to ocsp utility. The various functions which
13271 convert status values to strings have been renamed to:
13272 OCSP_response_status_str(), OCSP_cert_status_str() and
13273 OCSP_crl_reason_str() and are no longer static. New options
13274 to verify nonce values and to disable verification. OCSP response
13275 printout format cleaned up.
13279 * Add additional OCSP certificate checks. These are those specified
13280 in RFC2560. This consists of two separate checks: the CA of the
13281 certificate being checked must either be the OCSP signer certificate
13282 or the issuer of the OCSP signer certificate. In the latter case the
13283 OCSP signer certificate must contain the OCSP signing extended key
13284 usage. This check is performed by attempting to match the OCSP
13285 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
13286 in the OCSP_CERTID structures of the response.
13290 * Initial OCSP certificate verification added to OCSP_basic_verify()
13291 and related routines. This uses the standard OpenSSL certificate
13292 verify routines to perform initial checks (just CA validity) and
13293 to obtain the certificate chain. Then additional checks will be
13294 performed on the chain. Currently the root CA is checked to see
13295 if it is explicitly trusted for OCSP signing. This is used to set
13296 a root CA as a global signing root: that is any certificate that
13297 chains to that CA is an acceptable OCSP signing certificate.
13301 * New '-extfile ...' option to 'openssl ca' for reading X.509v3
13302 extensions from a separate configuration file.
13303 As when reading extensions from the main configuration file,
13304 the '-extensions ...' option may be used for specifying the
13307 *Massimiliano Pala <madwolf@comune.modena.it>*
13309 * New OCSP utility. Allows OCSP requests to be generated or
13310 read. The request can be sent to a responder and the output
13311 parsed, outputted or printed in text form. Not complete yet:
13312 still needs to check the OCSP response validity.
13316 * New subcommands for 'openssl ca':
13317 `openssl ca -status <serial>` prints the status of the cert with
13318 the given serial number (according to the index file).
13319 `openssl ca -updatedb` updates the expiry status of certificates
13322 *Massimiliano Pala <madwolf@comune.modena.it>*
13324 * New '-newreq-nodes' command option to CA.pl. This is like
13325 '-newreq', but calls 'openssl req' with the '-nodes' option
13326 so that the resulting key is not encrypted.
13328 *Damien Miller <djm@mindrot.org>*
13330 * New configuration for the GNU Hurd.
13332 *Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte*
13334 * Initial code to implement OCSP basic response verify. This
13335 is currently incomplete. Currently just finds the signer's
13336 certificate and verifies the signature on the response.
13340 * New SSLeay_version code SSLEAY_DIR to determine the compiled-in
13341 value of OPENSSLDIR. This is available via the new '-d' option
13342 to 'openssl version', and is also included in 'openssl version -a'.
13346 * Allowing defining memory allocation callbacks that will be given
13347 file name and line number information in additional arguments
13348 (a `const char*` and an int). The basic functionality remains, as
13349 well as the original possibility to just replace malloc(),
13350 realloc() and free() by functions that do not know about these
13351 additional arguments. To register and find out the current
13352 settings for extended allocation functions, the following
13353 functions are provided:
13355 CRYPTO_set_mem_ex_functions
13356 CRYPTO_set_locked_mem_ex_functions
13357 CRYPTO_get_mem_ex_functions
13358 CRYPTO_get_locked_mem_ex_functions
13360 These work the same way as CRYPTO_set_mem_functions and friends.
13361 `CRYPTO_get_[locked_]mem_functions` now writes 0 where such an
13362 extended allocation function is enabled.
13363 Similarly, `CRYPTO_get_[locked_]mem_ex_functions` writes 0 where
13364 a conventional allocation function is enabled.
13366 *Richard Levitte, Bodo Moeller*
13368 * Finish off removing the remaining LHASH function pointer casts.
13369 There should no longer be any prototype-casting required when using
13370 the LHASH abstraction, and any casts that remain are "bugs". See
13371 the callback types and macros at the head of lhash.h for details
13372 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
13376 * Add automatic query of EGD sockets in RAND_poll() for the unix variant.
13377 If /dev/[u]random devices are not available or do not return enough
13378 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
13380 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
13381 /etc/entropy will be queried once each in this sequence, querying stops
13382 when enough entropy was collected without querying more sockets.
13386 * Change the Unix RAND_poll() variant to be able to poll several
13387 random devices, as specified by DEVRANDOM, until a sufficient amount
13388 of data has been collected. We spend at most 10 ms on each file
13389 (select timeout) and read in non-blocking mode. DEVRANDOM now
13390 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
13391 (previously it was just the string "/dev/urandom"), so on typical
13392 platforms the 10 ms delay will never occur.
13393 Also separate out the Unix variant to its own file, rand_unix.c.
13394 For VMS, there's a currently-empty rand_vms.c.
13398 * Move OCSP client related routines to ocsp_cl.c. These
13399 provide utility functions which an application needing
13400 to issue a request to an OCSP responder and analyse the
13401 response will typically need: as opposed to those which an
13402 OCSP responder itself would need which will be added later.
13404 OCSP_request_sign() signs an OCSP request with an API similar
13405 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
13406 response. OCSP_response_get1_basic() extracts basic response
13407 from response. OCSP_resp_find_status(): finds and extracts status
13408 information from an OCSP_CERTID structure (which will be created
13409 when the request structure is built). These are built from lower
13410 level functions which work on OCSP_SINGLERESP structures but
13411 won't normally be used unless the application wishes to examine
13412 extensions in the OCSP response for example.
13414 Replace nonce routines with a pair of functions.
13415 OCSP_request_add1_nonce() adds a nonce value and optionally
13416 generates a random value. OCSP_check_nonce() checks the
13417 validity of the nonce in an OCSP response.
13421 * Change function OCSP_request_add() to OCSP_request_add0_id().
13422 This doesn't copy the supplied OCSP_CERTID and avoids the
13423 need to free up the newly created id. Change return type
13424 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
13425 This can then be used to add extensions to the request.
13426 Deleted OCSP_request_new(), since most of its functionality
13427 is now in OCSP_REQUEST_new() (and the case insensitive name
13428 clash) apart from the ability to set the request name which
13429 will be added elsewhere.
13433 * Update OCSP API. Remove obsolete extensions argument from
13434 various functions. Extensions are now handled using the new
13435 OCSP extension code. New simple OCSP HTTP function which
13436 can be used to send requests and parse the response.
13440 * Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
13441 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
13442 uses the special reorder version of SET OF to sort the attributes
13443 and reorder them to match the encoded order. This resolves a long
13444 standing problem: a verify on a PKCS7 structure just after signing
13445 it used to fail because the attribute order did not match the
13446 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
13447 it uses the received order. This is necessary to tolerate some broken
13448 software that does not order SET OF. This is handled by encoding
13449 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
13450 to produce the required SET OF.
13454 * Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
13455 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
13456 files to get correct declarations of the ASN.1 item variables.
13460 * Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
13461 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
13462 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
13463 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
13464 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
13465 ASN1_ITEM and no wrapper functions.
13469 * New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
13470 replace the old function pointer based I/O routines. Change most of
13471 the `*_d2i_bio()` and `*_d2i_fp()` functions to use these.
13475 * Enhance mkdef.pl to be more accepting about spacing in C preprocessor
13476 lines, recognize more "algorithms" that can be deselected, and make
13477 it complain about algorithm deselection that isn't recognised.
13481 * New ASN1 functions to handle dup, sign, verify, digest, pack and
13482 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
13483 to use new functions. Add NO_ASN1_OLD which can be set to remove
13484 some old style ASN1 functions: this can be used to determine if old
13485 code will still work when these eventually go away.
13489 * New extension functions for OCSP structures, these follow the
13490 same conventions as certificates and CRLs.
13494 * New function X509V3_add1_i2d(). This automatically encodes and
13495 adds an extension. Its behaviour can be customised with various
13496 flags to append, replace or delete. Various wrappers added for
13497 certificates and CRLs.
13501 * Fix to avoid calling the underlying ASN1 print routine when
13502 an extension cannot be parsed. Correct a typo in the
13503 OCSP_SERVICELOC extension. Tidy up print OCSP format.
13507 * Make mkdef.pl parse some of the ASN1 macros and add appropriate
13508 entries for variables.
13512 * Add functionality to `apps/openssl.c` for detecting locking
13513 problems: As the program is single-threaded, all we have
13514 to do is register a locking callback using an array for
13515 storing which locks are currently held by the program.
13519 * Use a lock around the call to CRYPTO_get_ex_new_index() in
13520 SSL_get_ex_data_X509_STORE_idx(), which is used in
13521 ssl_verify_cert_chain() and thus can be called at any time
13522 during TLS/SSL handshakes so that thread-safety is essential.
13523 Unfortunately, the ex_data design is not at all suited
13524 for multi-threaded use, so it probably should be abolished.
13528 * Added Broadcom "ubsec" ENGINE to OpenSSL.
13530 *Broadcom, tweaked and integrated by Geoff Thorpe*
13532 * Move common extension printing code to new function
13533 X509V3_print_extensions(). Reorganise OCSP print routines and
13534 implement some needed OCSP ASN1 functions. Add OCSP extensions.
13538 * New function X509_signature_print() to remove duplication in some
13543 * Add a special meaning when SET OF and SEQUENCE OF flags are both
13544 set (this was treated exactly the same as SET OF previously). This
13545 is used to reorder the STACK representing the structure to match the
13546 encoding. This will be used to get round a problem where a PKCS7
13547 structure which was signed could not be verified because the STACK
13548 order did not reflect the encoded order.
13552 * Reimplement the OCSP ASN1 module using the new code.
13556 * Update the X509V3 code to permit the use of an ASN1_ITEM structure
13557 for its ASN1 operations. The old style function pointers still exist
13558 for now but they will eventually go away.
13562 * Merge in replacement ASN1 code from the ASN1 branch. This almost
13563 completely replaces the old ASN1 functionality with a table driven
13564 encoder and decoder which interprets an ASN1_ITEM structure describing
13565 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
13566 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
13567 has also been converted to the new form.
13571 * Change BN_mod_exp_recp so that negative moduli are tolerated
13572 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
13573 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
13574 for negative moduli.
13578 * Fix BN_uadd and BN_usub: Always return non-negative results instead
13579 of not touching the result's sign bit.
13583 * BN_div bugfix: If the result is 0, the sign (res->neg) must not be
13588 * Changed the LHASH code to use prototypes for callbacks, and created
13589 macros to declare and implement thin (optionally static) functions
13590 that provide type-safety and avoid function pointer casting for the
13591 type-specific callbacks.
13595 * Added Kerberos Cipher Suites to be used with TLS, as written in
13597 *Veers Staats <staatsvr@asc.hpc.mil>,
13598 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte*
13600 * Reformat the FAQ so the different questions and answers can be divided
13601 in sections depending on the subject.
13605 * Have the zlib compression code load ZLIB.DLL dynamically under
13610 * New function BN_mod_sqrt for computing square roots modulo a prime
13611 (using the probabilistic Tonelli-Shanks algorithm unless
13612 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
13613 be handled deterministically).
13615 *Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller*
13617 * Make BN_mod_inverse faster by explicitly handling small quotients
13618 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
13619 512 bits], about 30% for larger ones [1024 or 2048 bits].)
13623 * New function BN_kronecker.
13627 * Fix BN_gcd so that it works on negative inputs; the result is
13628 positive unless both parameters are zero.
13629 Previously something reasonably close to an infinite loop was
13630 possible because numbers could be growing instead of shrinking
13631 in the implementation of Euclid's algorithm.
13635 * Fix BN_is_word() and BN_is_one() macros to take into account the
13636 sign of the number in question.
13638 Fix BN_is_word(a,w) to work correctly for w == 0.
13640 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
13641 because its test if the absolute value of 'a' equals 'w'.
13642 Note that BN_abs_is_word does *not* handle w == 0 reliably;
13643 it exists mostly for use in the implementations of BN_is_zero(),
13644 BN_is_one(), and BN_is_word().
13648 * New function BN_swap.
13652 * Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
13653 the exponentiation functions are more likely to produce reasonable
13654 results on negative inputs.
13658 * Change BN_mod_mul so that the result is always non-negative.
13659 Previously, it could be negative if one of the factors was negative;
13660 I don't think anyone really wanted that behaviour.
13664 * Move `BN_mod_...` functions into new file `crypto/bn/bn_mod.c`
13665 (except for exponentiation, which stays in `crypto/bn/bn_exp.c`,
13666 and `BN_mod_mul_reciprocal`, which stays in `crypto/bn/bn_recp.c`)
13667 and add new functions:
13676 BN_mod_lshift1_quick
13678 BN_mod_lshift_quick
13680 These functions always generate non-negative results.
13682 `BN_nnmod` otherwise is `like BN_mod` (if `BN_mod` computes a remainder `r`
13683 such that `|m| < r < 0`, `BN_nnmod` will output `rem + |m|` instead).
13685 `BN_mod_XXX_quick(r, a, [b,] m)` generates the same result as
13686 `BN_mod_XXX(r, a, [b,] m, ctx)`, but requires that `a` [and `b`]
13687 be reduced modulo `m`.
13689 *Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller*
13692 The following entry accidentally appeared in the CHANGES file
13693 distributed with OpenSSL 0.9.7. The modifications described in
13694 it do *not* apply to OpenSSL 0.9.7.
13696 * Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
13697 was actually never needed) and in BN_mul(). The removal in BN_mul()
13698 required a small change in bn_mul_part_recursive() and the addition
13699 of the functions bn_cmp_part_words(), bn_sub_part_words() and
13700 bn_add_part_words(), which do the same thing as bn_cmp_words(),
13701 bn_sub_words() and bn_add_words() except they take arrays with
13707 * In 'openssl passwd', verify passwords read from the terminal
13708 unless the '-salt' option is used (which usually means that
13709 verification would just waste user's time since the resulting
13710 hash is going to be compared with some given password hash)
13711 or the new '-noverify' option is used.
13713 This is an incompatible change, but it does not affect
13714 non-interactive use of 'openssl passwd' (passwords on the command
13715 line, '-stdin' option, '-in ...' option) and thus should not
13716 cause any problems.
13720 * Remove all references to RSAref, since there's no more need for it.
13724 * Make DSO load along a path given through an environment variable
13725 (SHLIB_PATH) with shl_load().
13729 * Constify the ENGINE code as a result of BIGNUM constification.
13730 Also constify the RSA code and most things related to it. In a
13731 few places, most notable in the depth of the ASN.1 code, ugly
13732 casts back to non-const were required (to be solved at a later
13737 * Make it so the openssl application has all engines loaded by default.
13741 * Constify the BIGNUM routines a little more.
13745 * Add the following functions:
13747 ENGINE_load_cswift()
13749 ENGINE_load_atalla()
13750 ENGINE_load_nuron()
13751 ENGINE_load_builtin_engines()
13753 That way, an application can itself choose if external engines that
13754 are built-in in OpenSSL shall ever be used or not. The benefit is
13755 that applications won't have to be linked with libdl or other dso
13756 libraries unless it's really needed.
13758 Changed 'openssl engine' to load all engines on demand.
13759 Changed the engine header files to avoid the duplication of some
13760 declarations (they differed!).
13764 * 'openssl engine' can now list capabilities.
13768 * Better error reporting in 'openssl engine'.
13772 * Never call load_dh_param(NULL) in s_server.
13776 * Add engine application. It can currently list engines by name and
13777 identity, and test if they are actually available.
13781 * Improve RPM specification file by forcing symbolic linking and making
13782 sure the installed documentation is also owned by root.root.
13784 *Damien Miller <djm@mindrot.org>*
13786 * Give the OpenSSL applications more possibilities to make use of
13787 keys (public as well as private) handled by engines.
13791 * Add OCSP code that comes from CertCo.
13795 * Add VMS support for the Rijndael code.
13799 * Added untested support for Nuron crypto accelerator.
13803 * Add support for external cryptographic devices. This code was
13804 previously distributed separately as the "engine" branch.
13806 *Geoff Thorpe, Richard Levitte*
13808 * Rework the filename-translation in the DSO code. It is now possible to
13809 have far greater control over how a "name" is turned into a filename
13810 depending on the operating environment and any oddities about the
13811 different shared library filenames on each system.
13815 * Support threads on FreeBSD-elf in Configure.
13819 * Fix for SHA1 assembly problem with MASM: it produces
13820 warnings about corrupt line number information when assembling
13821 with debugging information. This is caused by the overlapping
13824 *Bernd Matthes <mainbug@celocom.de>, Steve Henson*
13827 NCONF_get_number() has no error checking at all. As a replacement,
13828 NCONF_get_number_e() is defined (`_e` for "error checking") and is
13829 promoted strongly. The old NCONF_get_number is kept around for
13830 binary backward compatibility.
13831 Make it possible for methods to load from something other than a BIO,
13832 by providing a function pointer that is given a name instead of a BIO.
13833 For example, this could be used to load configuration data from an
13838 * Fix for non blocking accept BIOs. Added new I/O special reason
13839 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
13840 with non blocking I/O was not possible because no retry code was
13841 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
13846 * Added the beginnings of Rijndael support.
13850 * Fix for bug in DirectoryString mask setting. Add support for
13851 X509_NAME_print_ex() in 'req' and X509_print_ex() function
13852 to allow certificate printing to more controllable, additional
13853 'certopt' option to 'x509' to allow new printing options to be
13858 * Clean old EAY MD5 hack from e_os.h.
13862 ### Changes between 0.9.6l and 0.9.6m [17 Mar 2004]
13864 * Fix null-pointer assignment in do_change_cipher_spec() revealed
13865 by using the Codenomicon TLS Test Tool ([CVE-2004-0079])
13867 *Joe Orton, Steve Henson*
13869 ### Changes between 0.9.6k and 0.9.6l [04 Nov 2003]
13871 * Fix additional bug revealed by the NISCC test suite:
13873 Stop bug triggering large recursion when presented with
13874 certain ASN.1 tags ([CVE-2003-0851])
13878 ### Changes between 0.9.6j and 0.9.6k [30 Sep 2003]
13880 * Fix various bugs revealed by running the NISCC test suite:
13882 Stop out of bounds reads in the ASN1 code when presented with
13883 invalid tags (CVE-2003-0543 and CVE-2003-0544).
13885 If verify callback ignores invalid public key errors don't try to check
13886 certificate signature with the NULL public key.
13890 * In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
13891 if the server requested one: as stated in TLS 1.0 and SSL 3.0
13896 * In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
13897 extra data after the compression methods not only for TLS 1.0
13898 but also for SSL 3.0 (as required by the specification).
13900 *Bodo Moeller; problem pointed out by Matthias Loepfe*
13902 * Change X509_certificate_type() to mark the key as exported/exportable
13903 when it's 512 *bits* long, not 512 bytes.
13907 ### Changes between 0.9.6i and 0.9.6j [10 Apr 2003]
13909 * Countermeasure against the Klima-Pokorny-Rosa extension of
13910 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
13911 a protocol version number mismatch like a decryption error
13912 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
13916 * Turn on RSA blinding by default in the default implementation
13917 to avoid a timing attack. Applications that don't want it can call
13918 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
13919 They would be ill-advised to do so in most cases.
13921 *Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller*
13923 * Change RSA blinding code so that it works when the PRNG is not
13924 seeded (in this case, the secret RSA exponent is abused as
13925 an unpredictable seed -- if it is not unpredictable, there
13926 is no point in blinding anyway). Make RSA blinding thread-safe
13927 by remembering the creator's thread ID in rsa->blinding and
13928 having all other threads use local one-time blinding factors
13929 (this requires more computation than sharing rsa->blinding, but
13930 avoids excessive locking; and if an RSA object is not shared
13931 between threads, blinding will still be very fast).
13935 ### Changes between 0.9.6h and 0.9.6i [19 Feb 2003]
13937 * In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
13938 via timing by performing a MAC computation even if incorrect
13939 block cipher padding has been found. This is a countermeasure
13940 against active attacks where the attacker has to distinguish
13941 between bad padding and a MAC verification error. ([CVE-2003-0078])
13943 *Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
13944 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
13945 Martin Vuagnoux (EPFL, Ilion)*
13947 ### Changes between 0.9.6g and 0.9.6h [5 Dec 2002]
13949 * New function OPENSSL_cleanse(), which is used to cleanse a section of
13950 memory from its contents. This is done with a counter that will
13951 place alternating values in each byte. This can be used to solve
13952 two issues: 1) the removal of calls to memset() by highly optimizing
13953 compilers, and 2) cleansing with other values than 0, since those can
13954 be read through on certain media, for example a swap space on disk.
13958 * Bugfix: client side session caching did not work with external caching,
13959 because the session->cipher setting was not restored when reloading
13960 from the external cache. This problem was masked, when
13961 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set.
13962 (Found by Steve Haslam <steve@araqnid.ddts.net>.)
13966 * Fix client_certificate (ssl/s2_clnt.c): The permissible total
13967 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33.
13969 *Zeev Lieber <zeev-l@yahoo.com>*
13971 * Undo an undocumented change introduced in 0.9.6e which caused
13972 repeated calls to OpenSSL_add_all_ciphers() and
13973 OpenSSL_add_all_digests() to be ignored, even after calling
13978 * Change the default configuration reader to deal with last line not
13979 being properly terminated.
13983 * Change X509_NAME_cmp() so it applies the special rules on handling
13984 DN values that are of type PrintableString, as well as RDNs of type
13985 emailAddress where the value has the type ia5String.
13987 *stefank@valicert.com via Richard Levitte*
13989 * Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half
13990 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently
13991 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be
13992 the bitwise-OR of the two for use by the majority of applications
13993 wanting this behaviour, and update the docs. The documented
13994 behaviour and actual behaviour were inconsistent and had been
13995 changing anyway, so this is more a bug-fix than a behavioural
13998 *Geoff Thorpe, diagnosed by Nadav Har'El*
14000 * Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c
14001 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes).
14005 * Fix initialization code race conditions in
14006 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(),
14007 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(),
14008 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(),
14009 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(),
14010 ssl2_get_cipher_by_char(),
14011 ssl3_get_cipher_by_char().
14013 *Patrick McCormick <patrick@tellme.com>, Bodo Moeller*
14015 * Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after
14016 the cached sessions are flushed, as the remove_cb() might use ex_data
14017 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com>
14018 (see [openssl.org #212]).
14020 *Geoff Thorpe, Lutz Jaenicke*
14022 * Fix typo in OBJ_txt2obj which incorrectly passed the content
14023 length, instead of the encoding length to d2i_ASN1_OBJECT.
14027 ### Changes between 0.9.6f and 0.9.6g [9 Aug 2002]
14029 * [In 0.9.6g-engine release:]
14030 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use `_stdcall`).
14032 *Lynn Gazis <lgazis@rainbow.com>*
14034 ### Changes between 0.9.6e and 0.9.6f [8 Aug 2002]
14036 * Fix ASN1 checks. Check for overflow by comparing with LONG_MAX
14037 and get fix the header length calculation.
14038 *Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>,
14039 Alon Kantor <alonk@checkpoint.com> (and others), Steve Henson*
14041 * Use proper error handling instead of 'assertions' in buffer
14042 overflow checks added in 0.9.6e. This prevents DoS (the
14043 assertions could call abort()).
14045 *Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller*
14047 ### Changes between 0.9.6d and 0.9.6e [30 Jul 2002]
14049 * Add various sanity checks to asn1_get_length() to reject
14050 the ASN1 length bytes if they exceed sizeof(long), will appear
14051 negative or the content length exceeds the length of the
14054 *Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>*
14056 * Fix cipher selection routines: ciphers without encryption had no flags
14057 for the cipher strength set and where therefore not handled correctly
14058 by the selection routines (PR #130).
14062 * Fix EVP_dsa_sha macro.
14067 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS
14068 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure
14069 that was added in OpenSSL 0.9.6d.
14071 As the countermeasure turned out to be incompatible with some
14072 broken SSL implementations, the new option is part of SSL_OP_ALL.
14073 SSL_OP_ALL is usually employed when compatibility with weird SSL
14074 implementations is desired (e.g. '-bugs' option to 's_client' and
14075 's_server'), so the new option is automatically set in many
14080 * Changes in security patch:
14082 Changes marked "(CHATS)" were sponsored by the Defense Advanced
14083 Research Projects Agency (DARPA) and Air Force Research Laboratory,
14084 Air Force Materiel Command, USAF, under agreement number
14087 * Add various sanity checks to asn1_get_length() to reject
14088 the ASN1 length bytes if they exceed sizeof(long), will appear
14089 negative or the content length exceeds the length of the
14090 supplied buffer. ([CVE-2002-0659])
14092 *Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>*
14094 * Assertions for various potential buffer overflows, not known to
14095 happen in practice.
14097 *Ben Laurie (CHATS)*
14099 * Various temporary buffers to hold ASCII versions of integers were
14100 too small for 64 bit platforms. ([CVE-2002-0655])
14101 *Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>*
14103 * Remote buffer overflow in SSL3 protocol - an attacker could
14104 supply an oversized session ID to a client. ([CVE-2002-0656])
14106 *Ben Laurie (CHATS)*
14108 * Remote buffer overflow in SSL2 protocol - an attacker could
14109 supply an oversized client master key. ([CVE-2002-0656])
14111 *Ben Laurie (CHATS)*
14113 ### Changes between 0.9.6c and 0.9.6d [9 May 2002]
14115 * Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
14116 encoded as NULL) with id-dsa-with-sha1.
14118 *Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller*
14120 * Check various `X509_...()` return values in `apps/req.c`.
14122 *Nils Larsch <nla@trustcenter.de>*
14124 * Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
14125 an end-of-file condition would erroneously be flagged, when the CRLF
14126 was just at the end of a processed block. The bug was discovered when
14127 processing data through a buffering memory BIO handing the data to a
14128 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
14129 <ptsekov@syntrex.com> and Nedelcho Stanev.
14133 * Implement a countermeasure against a vulnerability recently found
14134 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
14135 before application data chunks to avoid the use of known IVs
14136 with data potentially chosen by the attacker.
14140 * Fix length checks in ssl3_get_client_hello().
14144 * TLS/SSL library bugfix: use s->s3->in_read_app_data differently
14145 to prevent ssl3_read_internal() from incorrectly assuming that
14146 ssl3_read_bytes() found application data while handshake
14147 processing was enabled when in fact s->s3->in_read_app_data was
14148 merely automatically cleared during the initial handshake.
14150 *Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>*
14152 * Fix object definitions for Private and Enterprise: they were not
14153 recognized in their shortname (=lowercase) representation. Extend
14154 obj_dat.pl to issue an error when using undefined keywords instead
14155 of silently ignoring the problem (Svenning Sorensen
14156 <sss@sss.dnsalias.net>).
14160 * Fix DH_generate_parameters() so that it works for 'non-standard'
14161 generators, i.e. generators other than 2 and 5. (Previously, the
14162 code did not properly initialise the 'add' and 'rem' values to
14163 BN_generate_prime().)
14165 In the new general case, we do not insist that 'generator' is
14166 actually a primitive root: This requirement is rather pointless;
14167 a generator of the order-q subgroup is just as good, if not
14172 * Map new X509 verification errors to alerts. Discovered and submitted by
14173 Tom Wu <tom@arcot.com>.
14177 * Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
14178 returning non-zero before the data has been completely received
14179 when using non-blocking I/O.
14181 *Bodo Moeller; problem pointed out by John Hughes*
14183 * Some of the ciphers missed the strength entry (SSL_LOW etc).
14185 *Ben Laurie, Lutz Jaenicke*
14187 * Fix bug in SSL_clear(): bad sessions were not removed (found by
14188 Yoram Zahavi <YoramZ@gilian.com>).
14192 * Add information about CygWin 1.3 and on, and preserve proper
14193 configuration for the versions before that.
14195 *Corinna Vinschen <vinschen@redhat.com> and Richard Levitte*
14197 * Make removal from session cache (SSL_CTX_remove_session()) more robust:
14198 check whether we deal with a copy of a session and do not delete from
14199 the cache in this case. Problem reported by "Izhar Shoshani Levi"
14200 <izhar@checkpoint.com>.
14204 * Do not store session data into the internal session cache, if it
14205 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
14206 flag is set). Proposed by Aslam <aslam@funk.com>.
14210 * Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
14215 * [In 0.9.6d-engine release:]
14216 Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
14218 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
14220 * Add the configuration target linux-s390x.
14222 *Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte*
14224 * The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
14225 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
14226 variable as an indication that a ClientHello message has been
14227 received. As the flag value will be lost between multiple
14228 invocations of ssl3_accept when using non-blocking I/O, the
14229 function may not be aware that a handshake has actually taken
14230 place, thus preventing a new session from being added to the
14233 To avoid this problem, we now set s->new_session to 2 instead of
14234 using a local variable.
14236 *Lutz Jaenicke, Bodo Moeller*
14238 * Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
14239 if the SSL_R_LENGTH_MISMATCH error is detected.
14241 *Geoff Thorpe, Bodo Moeller*
14243 * New 'shared_ldflag' column in Configure platform table.
14247 * Fix EVP_CIPHER_mode macro.
14249 *"Dan S. Camper" <dan@bti.net>*
14251 * Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
14252 type, we must throw them away by setting rr->length to 0.
14254 *D P Chang <dpc@qualys.com>*
14256 ### Changes between 0.9.6b and 0.9.6c [21 dec 2001]
14258 * Fix BN_rand_range bug pointed out by Dominikus Scherkl
14259 <Dominikus.Scherkl@biodata.com>. (The previous implementation
14260 worked incorrectly for those cases where range = `10..._2` and
14261 `3*range` is two bits longer than range.)
14265 * Only add signing time to PKCS7 structures if it is not already
14270 * Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
14271 OBJ_ld_ce should be OBJ_id_ce.
14272 Also some ip-pda OIDs in crypto/objects/objects.txt were
14273 incorrect (cf. RFC 3039).
14275 *Matt Cooper, Frederic Giudicelli, Bodo Moeller*
14277 * Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
14278 returns early because it has nothing to do.
14280 *Andy Schneider <andy.schneider@bjss.co.uk>*
14282 * [In 0.9.6c-engine release:]
14283 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
14285 *Andy Schneider <andy.schneider@bjss.co.uk>*
14287 * [In 0.9.6c-engine release:]
14288 Add support for Cryptographic Appliance's keyserver technology.
14289 (Use engine 'keyclient')
14291 *Cryptographic Appliances and Geoff Thorpe*
14293 * Add a configuration entry for OS/390 Unix. The C compiler 'c89'
14294 is called via tools/c89.sh because arguments have to be
14295 rearranged (all '-L' options must appear before the first object
14298 *Richard Shapiro <rshapiro@abinitio.com>*
14300 * [In 0.9.6c-engine release:]
14301 Add support for Broadcom crypto accelerator cards, backported
14304 *Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox*
14306 * [In 0.9.6c-engine release:]
14307 Add support for SureWare crypto accelerator cards from
14308 Baltimore Technologies. (Use engine 'sureware')
14310 *Baltimore Technologies and Mark Cox*
14312 * [In 0.9.6c-engine release:]
14313 Add support for crypto accelerator cards from Accelerated
14314 Encryption Processing, www.aep.ie. (Use engine 'aep')
14316 *AEP Inc. and Mark Cox*
14318 * Add a configuration entry for gcc on UnixWare.
14320 *Gary Benson <gbenson@redhat.com>*
14322 * Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
14323 messages are stored in a single piece (fixed-length part and
14324 variable-length part combined) and fix various bugs found on the way.
14328 * Disable caching in BIO_gethostbyname(), directly use gethostbyname()
14329 instead. BIO_gethostbyname() does not know what timeouts are
14330 appropriate, so entries would stay in cache even when they have
14332 *Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>*
14334 * Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
14335 faced with a pathologically small ClientHello fragment that does
14336 not contain client_version: Instead of aborting with an error,
14337 simply choose the highest available protocol version (i.e.,
14338 TLS 1.0 unless it is disabled). In practice, ClientHello
14339 messages are never sent like this, but this change gives us
14340 strictly correct behaviour at least for TLS.
14344 * Fix SSL handshake functions and SSL_clear() such that SSL_clear()
14345 never resets s->method to s->ctx->method when called from within
14346 one of the SSL handshake functions.
14348 *Bodo Moeller; problem pointed out by Niko Baric*
14350 * In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
14351 (sent using the client's version number) if client_version is
14352 smaller than the protocol version in use. Also change
14353 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
14354 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
14355 the client will at least see that alert.
14359 * Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
14364 * Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
14365 client receives HelloRequest while in a handshake.
14367 *Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>*
14369 * Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
14370 should end in 'break', not 'goto end' which circumvents various
14371 cleanups done in state SSL_ST_OK. But session related stuff
14372 must be disabled for SSL_ST_OK in the case that we just sent a
14375 Also avoid some overhead by not calling ssl_init_wbio_buffer()
14376 before just sending a HelloRequest.
14378 *Bodo Moeller, Eric Rescorla <ekr@rtfm.com>*
14380 * Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
14381 reveal whether illegal block cipher padding was found or a MAC
14382 verification error occurred. (Neither SSLerr() codes nor alerts
14383 are directly visible to potential attackers, but the information
14384 may leak via logfiles.)
14386 Similar changes are not required for the SSL 2.0 implementation
14387 because the number of padding bytes is sent in clear for SSL 2.0,
14388 and the extra bytes are just ignored. However ssl/s2_pkt.c
14389 failed to verify that the purported number of padding bytes is in
14394 * Add OpenUNIX-8 support including shared libraries
14395 (Boyd Lynn Gerber <gerberb@zenez.com>).
14399 * Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
14400 'wristwatch attack' using huge encoding parameters (cf.
14401 James H. Manger's CRYPTO 2001 paper). Note that the
14402 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
14403 encoding parameters and hence was not vulnerable.
14407 * BN_sqr() bug fix.
14409 *Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>*
14411 * Rabin-Miller test analyses assume uniformly distributed witnesses,
14412 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
14413 followed by modular reduction.
14415 *Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>*
14417 * Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
14418 equivalent based on BN_pseudo_rand() instead of BN_rand().
14422 * s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
14423 This function was broken, as the check for a new client hello message
14424 to handle SGC did not allow these large messages.
14425 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
14429 * Add alert descriptions for TLSv1 to `SSL_alert_desc_string[_long]()`.
14433 * Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
14434 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
14438 * Rework the configuration and shared library support for Tru64 Unix.
14439 The configuration part makes use of modern compiler features and
14440 still retains old compiler behavior for those that run older versions
14441 of the OS. The shared library support part includes a variant that
14442 uses the RPATH feature, and is available through the special
14443 configuration target "alpha-cc-rpath", which will never be selected
14446 *Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte*
14448 * In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
14449 with the same message size as in ssl3_get_certificate_request().
14450 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
14451 messages might inadvertently be reject as too long.
14453 *Petr Lampa <lampa@fee.vutbr.cz>*
14455 * Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
14459 * Modified SSL library such that the verify_callback that has been set
14460 specifically for an SSL object with SSL_set_verify() is actually being
14461 used. Before the change, a verify_callback set with this function was
14462 ignored and the verify_callback() set in the SSL_CTX at the time of
14463 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
14464 to allow the necessary settings.
14468 * Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
14469 explicitly to NULL, as at least on Solaris 8 this seems not always to be
14470 done automatically (in contradiction to the requirements of the C
14471 standard). This made problems when used from OpenSSH.
14475 * In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
14476 dh->length and always used
14478 BN_rand_range(priv_key, dh->p).
14480 BN_rand_range() is not necessary for Diffie-Hellman, and this
14481 specific range makes Diffie-Hellman unnecessarily inefficient if
14482 dh->length (recommended exponent length) is much smaller than the
14483 length of dh->p. We could use BN_rand_range() if the order of
14484 the subgroup was stored in the DH structure, but we only have
14489 BN_rand(priv_key, l, ...)
14491 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
14498 RSA_eay_public_encrypt
14499 RSA_eay_private_decrypt
14500 RSA_eay_private_encrypt (signing)
14501 RSA_eay_public_decrypt (signature verification)
14503 (default implementations for RSA_public_encrypt,
14504 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
14505 always reject numbers >= n.
14509 * In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
14510 to synchronize access to 'locking_thread'. This is necessary on
14511 systems where access to 'locking_thread' (an 'unsigned long'
14512 variable) is not atomic.
14516 * In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
14517 *before* setting the 'crypto_lock_rand' flag. The previous code had
14518 a race condition if 0 is a valid thread ID.
14520 *Travis Vitek <vitek@roguewave.com>*
14522 * Add support for shared libraries under Irix.
14524 *Albert Chin-A-Young <china@thewrittenword.com>*
14526 * Add configuration option to build on Linux on both big-endian and
14527 little-endian MIPS.
14529 *Ralf Baechle <ralf@uni-koblenz.de>*
14531 * Add the possibility to create shared libraries on HP-UX.
14535 ### Changes between 0.9.6a and 0.9.6b [9 Jul 2001]
14537 * Change ssleay_rand_bytes (crypto/rand/md_rand.c)
14538 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
14539 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
14540 PRNG state recovery was possible based on the output of
14541 one PRNG request appropriately sized to gain knowledge on
14542 'md' followed by enough consecutive 1-byte PRNG requests
14543 to traverse all of 'state'.
14545 1. When updating 'md_local' (the current thread's copy of 'md')
14546 during PRNG output generation, hash all of the previous
14547 'md_local' value, not just the half used for PRNG output.
14549 2. Make the number of bytes from 'state' included into the hash
14550 independent from the number of PRNG bytes requested.
14552 The first measure alone would be sufficient to avoid
14553 Markku-Juhani's attack. (Actually it had never occurred
14554 to me that the half of 'md_local' used for chaining was the
14555 half from which PRNG output bytes were taken -- I had always
14556 assumed that the secret half would be used.) The second
14557 measure makes sure that additional data from 'state' is never
14558 mixed into 'md_local' in small portions; this heuristically
14559 further strengthens the PRNG.
14563 * Fix crypto/bn/asm/mips3.s.
14567 * When only the key is given to "enc", the IV is undefined. Print out
14568 an error message in this case.
14572 * Handle special case when X509_NAME is empty in X509 printing routines.
14576 * In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
14577 positive and less than q.
14581 * Don't change `*pointer` in CRYPTO_add_lock() is add_lock_callback is
14582 used: it isn't thread safe and the add_lock_callback should handle
14585 *Paul Rose <Paul.Rose@bridge.com>*
14587 * Verify that incoming data obeys the block size in
14588 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
14594 *Ulf Möller, Bodo Möller*
14596 * The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
14597 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
14598 when fixing the server behaviour for backwards-compatible 'client
14599 hello' messages. (Note that the attack is impractical against
14600 SSL 3.0 and TLS 1.0 anyway because length and version checking
14601 means that the probability of guessing a valid ciphertext is
14602 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
14605 Before 0.9.5, the countermeasure (hide the error by generating a
14606 random 'decryption result') did not work properly because
14607 ERR_clear_error() was missing, meaning that SSL_get_error() would
14608 detect the supposedly ignored error.
14610 Both problems are now fixed.
14614 * In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
14615 (previously it was 1024).
14619 * Fix for compatibility mode trust settings: ignore trust settings
14620 unless some valid trust or reject settings are present.
14624 * Fix for blowfish EVP: its a variable length cipher.
14628 * Fix various bugs related to DSA S/MIME verification. Handle missing
14629 parameters in DSA public key structures and return an error in the
14630 DSA routines if parameters are absent.
14634 * In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
14635 in the current directory if neither $RANDFILE nor $HOME was set.
14636 RAND_file_name() in 0.9.6a returned NULL in this case. This has
14637 caused some confusion to Windows users who haven't defined $HOME.
14638 Thus RAND_file_name() is changed again: e_os.h can define a
14639 DEFAULT_HOME, which will be used if $HOME is not set.
14640 For Windows, we use "C:"; on other platforms, we still require
14641 environment variables.
14643 * Move 'if (!initialized) RAND_poll()' into regions protected by
14644 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
14645 having multiple threads call RAND_poll() concurrently.
14649 * In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
14650 combination of a flag and a thread ID variable.
14651 Otherwise while one thread is in ssleay_rand_bytes (which sets the
14652 flag), *other* threads can enter ssleay_add_bytes without obeying
14653 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
14654 that they do not hold after the first thread unsets add_do_not_lock).
14658 * Change bctest again: '-x' expressions are not available in all
14659 versions of 'test'.
14663 ### Changes between 0.9.6 and 0.9.6a [5 Apr 2001]
14665 * Fix a couple of memory leaks in PKCS7_dataDecode()
14667 *Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>*
14669 * Change Configure and Makefiles to provide EXE_EXT, which will contain
14670 the default extension for executables, if any. Also, make the perl
14671 scripts that use symlink() to test if it really exists and use "cp"
14672 if it doesn't. All this made OpenSSL compilable and installable in
14677 * Fix for asn1_GetSequence() for indefinite length constructed data.
14678 If SEQUENCE is length is indefinite just set c->slen to the total
14679 amount of data available.
14681 *Steve Henson, reported by shige@FreeBSD.org*
14683 *This change does not apply to 0.9.7.*
14685 * Change bctest to avoid here-documents inside command substitution
14686 (workaround for FreeBSD /bin/sh bug).
14687 For compatibility with Ultrix, avoid shell functions (introduced
14688 in the bctest version that searches along $PATH).
14692 * Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
14693 with des_encrypt() defined on some operating systems, like Solaris
14698 * Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
14699 On the Importance of Eliminating Errors in Cryptographic
14700 Computations, J. Cryptology 14 (2001) 2, 101-119,
14701 <http://theory.stanford.edu/~dabo/papers/faults.ps.gz>).
14705 * MIPS assembler BIGNUM division bug fix.
14709 * Disabled incorrect Alpha assembler code.
14713 * Fix PKCS#7 decode routines so they correctly update the length
14714 after reading an EOC for the EXPLICIT tag.
14718 *This change does not apply to 0.9.7.*
14720 * Fix bug in PKCS#12 key generation routines. This was triggered
14721 if a 3DES key was generated with a 0 initial byte. Include
14722 PKCS12_BROKEN_KEYGEN compilation option to retain the old
14723 (but broken) behaviour.
14727 * Enhance bctest to search for a working bc along $PATH and print
14730 *Tim Rice <tim@multitalents.net> via Richard Levitte*
14732 * Fix memory leaks in err.c: free err_data string if necessary;
14733 don't write to the wrong index in ERR_set_error_data.
14737 * Implement ssl23_peek (analogous to ssl23_read), which previously
14742 * Replace rdtsc with `_emit` statements for VC++ version 5.
14744 *Jeremy Cooper <jeremy@baymoo.org>*
14746 * Make it possible to reuse SSLv2 sessions.
14750 * In copy_email() check for >= 0 as a return value for
14751 X509_NAME_get_index_by_NID() since 0 is a valid index.
14753 *Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>*
14755 * Avoid coredump with unsupported or invalid public keys by checking if
14756 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
14757 PKCS7_verify() fails with non detached data.
14761 * Don't use getenv in library functions when run as setuid/setgid.
14762 New function OPENSSL_issetugid().
14766 * Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
14767 due to incorrect handling of multi-threading:
14769 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
14771 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
14773 3. Count how many times MemCheck_off() has been called so that
14774 nested use can be treated correctly. This also avoids
14775 inband-signalling in the previous code (which relied on the
14776 assumption that thread ID 0 is impossible).
14780 * Add "-rand" option also to s_client and s_server.
14784 * Fix CPU detection on Irix 6.x.
14785 *Kurt Hockenbury <khockenb@stevens-tech.edu> and
14786 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>*
14788 * Fix X509_NAME bug which produced incorrect encoding if X509_NAME
14793 *This change does not apply to 0.9.7.*
14795 * Use the cached encoding of an X509_NAME structure rather than
14796 copying it. This is apparently the reason for the libsafe "errors"
14797 but the code is actually correct.
14801 * Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
14802 Bleichenbacher's DSA attack.
14803 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
14804 to be set and top=0 forces the highest bit to be set; top=-1 is new
14805 and leaves the highest bit random.
14807 *Ulf Moeller, Bodo Moeller*
14809 * In the `NCONF_...`-based implementations for `CONF_...` queries
14810 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
14811 a temporary CONF structure with the data component set to NULL
14812 (which gives segmentation faults in lh_retrieve).
14813 Instead, use NULL for the CONF pointer in CONF_get_string and
14814 CONF_get_number (which may use environment variables) and directly
14815 return NULL from CONF_get_section.
14819 * Fix potential buffer overrun for EBCDIC.
14823 * Tolerate nonRepudiation as being valid for S/MIME signing and certSign
14824 keyUsage if basicConstraints absent for a CA.
14828 * Make SMIME_write_PKCS7() write mail header values with a format that
14829 is more generally accepted (no spaces before the semicolon), since
14830 some programs can't parse those values properly otherwise. Also make
14831 sure BIO's that break lines after each write do not create invalid
14836 * Make the CRL encoding routines work with empty SEQUENCE OF. The
14837 macros previously used would not encode an empty SEQUENCE OF
14838 and break the signature.
14842 *This change does not apply to 0.9.7.*
14844 * Zero the premaster secret after deriving the master secret in
14849 * Add some EVP_add_digest_alias registrations (as found in
14850 OpenSSL_add_all_digests()) to SSL_library_init()
14851 aka OpenSSL_add_ssl_algorithms(). This provides improved
14852 compatibility with peers using X.509 certificates
14853 with unconventional AlgorithmIdentifier OIDs.
14857 * Fix for Irix with NO_ASM.
14859 *"Bruce W. Forsberg" <bruce.forsberg@baesystems.com>*
14861 * ./config script fixes.
14863 *Ulf Moeller, Richard Levitte*
14865 * Fix 'openssl passwd -1'.
14869 * Change PKCS12_key_gen_asc() so it can cope with non null
14870 terminated strings whose length is passed in the passlen
14871 parameter, for example from PEM callbacks. This was done
14872 by adding an extra length parameter to asc2uni().
14874 *Steve Henson, reported by <oddissey@samsung.co.kr>*
14876 * Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
14877 call failed, free the DSA structure.
14881 * Fix to uni2asc() to cope with zero length Unicode strings.
14882 These are present in some PKCS#12 files.
14886 * Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
14887 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
14888 when writing a 32767 byte record.
14890 *Bodo Moeller; problem reported by Eric Day <eday@concentric.net>*
14892 * In `RSA_eay_public_{en,ed}crypt` and RSA_eay_mod_exp (rsa_eay.c),
14893 obtain lock CRYPTO_LOCK_RSA before setting `rsa->_method_mod_{n,p,q}`.
14895 (RSA objects have a reference count access to which is protected
14896 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
14897 so they are meant to be shared between threads.)
14898 *Bodo Moeller, Geoff Thorpe; original patch submitted by
14899 "Reddie, Steven" <Steven.Reddie@ca.com>*
14901 * Fix a deadlock in CRYPTO_mem_leaks().
14905 * Use better test patterns in bntest.
14909 * rand_win.c fix for Borland C.
14913 * BN_rshift bugfix for n == 0.
14917 * Add a 'bctest' script that checks for some known 'bc' bugs
14918 so that 'make test' does not abort just because 'bc' is broken.
14922 * Store verify_result within SSL_SESSION also for client side to
14923 avoid potential security hole. (Re-used sessions on the client side
14924 always resulted in verify_result==X509_V_OK, not using the original
14925 result of the server certificate verification.)
14929 * Fix ssl3_pending: If the record in s->s3->rrec is not of type
14930 SSL3_RT_APPLICATION_DATA, return 0.
14931 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
14936 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
14937 releases, have been re-implemented by renaming the previous
14938 implementations of ssl2_read and ssl3_read to ssl2_read_internal
14939 and ssl3_read_internal, respectively, and adding 'peek' parameters
14940 to them. The new ssl[23]_{read,peek} functions are calls to
14941 ssl[23]_read_internal with the 'peek' flag set appropriately.
14942 A 'peek' parameter has also been added to ssl3_read_bytes, which
14943 does the actual work for ssl3_read_internal.
14947 * Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
14948 the method-specific "init()" handler. Also clean up ex_data after
14949 calling the method-specific "finish()" handler. Previously, this was
14950 happening the other way round.
14954 * Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
14955 The previous value, 12, was not always sufficient for BN_mod_exp().
14959 * Make sure that shared libraries get the internal name engine with
14960 the full version number and not just 0. This should mark the
14961 shared libraries as not backward compatible. Of course, this should
14962 be changed again when we can guarantee backward binary compatibility.
14966 * Fix typo in get_cert_by_subject() in by_dir.c
14968 *Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>*
14970 * Rework the system to generate shared libraries:
14972 - Make note of the expected extension for the shared libraries and
14973 if there is a need for symbolic links from for example libcrypto.so.0
14974 to libcrypto.so.0.9.7. There is extended info in Configure for
14977 - Make as few rebuilds of the shared libraries as possible.
14979 - Still avoid linking the OpenSSL programs with the shared libraries.
14981 - When installing, install the shared libraries separately from the
14986 * Fix SSL_CTX_set_read_ahead macro to actually use its argument.
14988 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
14989 and not in SSL_clear because the latter is also used by the
14990 accept/connect functions; previously, the settings made by
14991 SSL_set_read_ahead would be lost during the handshake.
14993 *Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>*
14995 * Correct util/mkdef.pl to be selective about disabled algorithms.
14996 Previously, it would create entries for disabled algorithms no
15001 * Added several new manual pages for SSL_* function.
15005 ### Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
15007 * In ssl23_get_client_hello, generate an error message when faced
15008 with an initial SSL 3.0/TLS record that is too small to contain the
15009 first two bytes of the ClientHello message, i.e. client_version.
15010 (Note that this is a pathologic case that probably has never happened
15011 in real life.) The previous approach was to use the version number
15012 from the record header as a substitute; but our protocol choice
15013 should not depend on that one because it is not authenticated
15014 by the Finished messages.
15018 * More robust randomness gathering functions for Windows.
15020 *Jeffrey Altman <jaltman@columbia.edu>*
15022 * For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
15023 not set then we don't setup the error code for issuer check errors
15024 to avoid possibly overwriting other errors which the callback does
15025 handle. If an application does set the flag then we assume it knows
15026 what it is doing and can handle the new informational codes
15031 * Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
15032 a general "ANY" type, as such it should be able to decode anything
15033 including tagged types. However it didn't check the class so it would
15034 wrongly interpret tagged types in the same way as their universal
15035 counterpart and unknown types were just rejected. Changed so that the
15036 tagged and unknown types are handled in the same way as a SEQUENCE:
15037 that is the encoding is stored intact. There is also a new type
15038 "V_ASN1_OTHER" which is used when the class is not universal, in this
15039 case we have no idea what the actual type is so we just lump them all
15044 * On VMS, stdout may very well lead to a file that is written to
15045 in a record-oriented fashion. That means that every write() will
15046 write a separate record, which will be read separately by the
15047 programs trying to read from it. This can be very confusing.
15049 The solution is to put a BIO filter in the way that will buffer
15050 text until a linefeed is reached, and then write everything a
15051 line at a time, so every record written will be an actual line,
15052 not chunks of lines and not (usually doesn't happen, but I've
15053 seen it once) several lines in one record. BIO_f_linebuffer() is
15056 Currently, it's a VMS-only method, because that's where it has
15057 been tested well enough.
15061 * Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
15062 it can return incorrect results.
15063 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
15064 but it was in 0.9.6-beta[12].)
15068 * Disable the check for content being present when verifying detached
15069 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
15070 include zero length content when signing messages.
15074 * New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
15075 BIO_ctrl (for BIO pairs).
15079 * Add DSO method for VMS.
15083 * Bug fix: Montgomery multiplication could produce results with the
15088 * Add RPM specification openssl.spec and modify it to build three
15089 packages. The default package contains applications, application
15090 documentation and run-time libraries. The devel package contains
15091 include files, static libraries and function documentation. The
15092 doc package contains the contents of the doc directory. The original
15093 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
15097 * Add a large number of documentation files for many SSL routines.
15099 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>*
15101 * Add a configuration entry for Sony News 4.
15103 *NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>*
15105 * Don't set the two most significant bits to one when generating a
15106 random number < q in the DSA library.
15110 * New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
15111 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
15112 the underlying transport is blocking) if a handshake took place.
15113 (The default behaviour is needed by applications such as s_client
15114 and s_server that use select() to determine when to use SSL_read;
15115 but for applications that know in advance when to expect data, it
15116 just makes things more complicated.)
15120 * Add RAND_egd_bytes(), which gives control over the number of bytes read
15125 * Add a few more EBCDIC conditionals that make `req` and `x509`
15126 work better on such systems.
15128 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
15130 * Add two demo programs for PKCS12_parse() and PKCS12_create().
15131 Update PKCS12_parse() so it copies the friendlyName and the
15132 keyid to the certificates aux info.
15136 * Fix bug in PKCS7_verify() which caused an infinite loop
15137 if there was more than one signature.
15139 *Sven Uszpelkat <su@celocom.de>*
15141 * Major change in util/mkdef.pl to include extra information
15142 about each symbol, as well as presenting variables as well
15143 as functions. This change means that there's n more need
15144 to rebuild the .num files when some algorithms are excluded.
15148 * Allow the verify time to be set by an application,
15149 rather than always using the current time.
15153 * Phase 2 verify code reorganisation. The certificate
15154 verify code now looks up an issuer certificate by a
15155 number of criteria: subject name, authority key id
15156 and key usage. It also verifies self signed certificates
15157 by the same criteria. The main comparison function is
15158 X509_check_issued() which performs these checks.
15160 Lot of changes were necessary in order to support this
15161 without completely rewriting the lookup code.
15163 Authority and subject key identifier are now cached.
15165 The LHASH 'certs' is X509_STORE has now been replaced
15166 by a STACK_OF(X509_OBJECT). This is mainly because an
15167 LHASH can't store or retrieve multiple objects with
15168 the same hash value.
15170 As a result various functions (which were all internal
15171 use only) have changed to handle the new X509_STORE
15172 structure. This will break anything that messed round
15173 with X509_STORE internally.
15175 The functions X509_STORE_add_cert() now checks for an
15176 exact match, rather than just subject name.
15178 The X509_STORE API doesn't directly support the retrieval
15179 of multiple certificates matching a given criteria, however
15180 this can be worked round by performing a lookup first
15181 (which will fill the cache with candidate certificates)
15182 and then examining the cache for matches. This is probably
15183 the best we can do without throwing out X509_LOOKUP
15184 entirely (maybe later...).
15186 The X509_VERIFY_CTX structure has been enhanced considerably.
15188 All certificate lookup operations now go via a get_issuer()
15189 callback. Although this currently uses an X509_STORE it
15190 can be replaced by custom lookups. This is a simple way
15191 to bypass the X509_STORE hackery necessary to make this
15192 work and makes it possible to use more efficient techniques
15193 in future. A very simple version which uses a simple
15194 STACK for its trusted certificate store is also provided
15195 using X509_STORE_CTX_trusted_stack().
15197 The verify_cb() and verify() callbacks now have equivalents
15198 in the X509_STORE_CTX structure.
15200 X509_STORE_CTX also has a 'flags' field which can be used
15201 to customise the verify behaviour.
15205 * Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
15206 excludes S/MIME capabilities.
15210 * When a certificate request is read in keep a copy of the
15211 original encoding of the signed data and use it when outputting
15212 again. Signatures then use the original encoding rather than
15213 a decoded, encoded version which may cause problems if the
15214 request is improperly encoded.
15218 * For consistency with other BIO_puts implementations, call
15219 buffer_write(b, ...) directly in buffer_puts instead of calling
15222 In BIO_puts, increment b->num_write as in BIO_write.
15224 *Peter.Sylvester@EdelWeb.fr*
15226 * Fix BN_mul_word for the case where the word is 0. (We have to use
15227 BN_zero, we may not return a BIGNUM with an array consisting of
15228 words set to zero.)
15232 * Avoid calling abort() from within the library when problems are
15233 detected, except if preprocessor symbols have been defined
15234 (such as REF_CHECK, BN_DEBUG etc.).
15238 * New openssl application 'rsautl'. This utility can be
15239 used for low-level RSA operations. DER public key
15240 BIO/fp routines also added.
15244 * New Configure entry and patches for compiling on QNX 4.
15246 *Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>*
15248 * A demo state-machine implementation was sponsored by
15249 Nuron (<http://www.nuron.com/>) and is now available in
15250 demos/state_machine.
15254 * New options added to the 'dgst' utility for signature
15255 generation and verification.
15259 * Unrecognized PKCS#7 content types are now handled via a
15260 catch all ASN1_TYPE structure. This allows unsupported
15261 types to be stored as a "blob" and an application can
15262 encode and decode it manually.
15266 * Fix various signed/unsigned issues to make a_strex.c
15267 compile under VC++.
15269 *Oscar Jacobsson <oscar.jacobsson@celocom.com>*
15271 * ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
15272 length if passed a buffer. ASN1_INTEGER_to_BN failed
15273 if passed a NULL BN and its argument was negative.
15275 *Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>*
15277 * Modification to PKCS#7 encoding routines to output definite
15278 length encoding. Since currently the whole structures are in
15279 memory there's not real point in using indefinite length
15280 constructed encoding. However if OpenSSL is compiled with
15281 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
15285 * Added BIO_vprintf() and BIO_vsnprintf().
15289 * Added more prefixes to parse for in the strings written
15290 through a logging bio, to cover all the levels that are available
15291 through syslog. The prefixes are now:
15293 PANIC, EMERG, EMR => LOG_EMERG
15294 ALERT, ALR => LOG_ALERT
15295 CRIT, CRI => LOG_CRIT
15296 ERROR, ERR => LOG_ERR
15297 WARNING, WARN, WAR => LOG_WARNING
15298 NOTICE, NOTE, NOT => LOG_NOTICE
15299 INFO, INF => LOG_INFO
15300 DEBUG, DBG => LOG_DEBUG
15302 and as before, if none of those prefixes are present at the
15303 beginning of the string, LOG_ERR is chosen.
15305 On Win32, the `LOG_*` levels are mapped according to this:
15307 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
15308 LOG_WARNING => EVENTLOG_WARNING_TYPE
15309 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
15313 * Made it possible to reconfigure with just the configuration
15314 argument "reconf" or "reconfigure". The command line arguments
15315 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
15316 and are retrieved from there when reconfiguring.
15322 *Assar Westerlund <assar@sics.se>, Richard Levitte*
15324 * Add the arguments -CAfile and -CApath to the pkcs12 utility.
15328 * The obj_dat.pl script was messing up the sorting of object
15329 names. The reason was that it compared the quoted version
15330 of strings as a result "OCSP" > "OCSP Signing" because
15331 " > SPACE. Changed script to store unquoted versions of
15332 names and add quotes on output. It was also omitting some
15333 names from the lookup table if they were given a default
15334 value (that is if SN is missing it is given the same
15335 value as LN and vice versa), these are now added on the
15336 grounds that if an object has a name we should be able to
15337 look it up. Finally added warning output when duplicate
15338 short or long names are found.
15342 * Changes needed for Tandem NSK.
15344 *Scott Uroff <scott@xypro.com>*
15346 * Fix SSL 2.0 rollback checking: Due to an off-by-one error in
15347 RSA_padding_check_SSLv23(), special padding was never detected
15348 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
15349 version rollback attacks was not effective.
15351 In s23_clnt.c, don't use special rollback-attack detection padding
15352 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
15353 client; similarly, in s23_srvr.c, don't do the rollback check if
15354 SSL 2.0 is the only protocol enabled in the server.
15358 * Make it possible to get hexdumps of unprintable data with 'openssl
15359 asn1parse'. By implication, the functions ASN1_parse_dump() and
15360 BIO_dump_indent() are added.
15364 * New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
15365 these print out strings and name structures based on various
15366 flags including RFC2253 support and proper handling of
15367 multibyte characters. Added options to the 'x509' utility
15368 to allow the various flags to be set.
15372 * Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
15373 Also change the functions X509_cmp_current_time() and
15374 X509_gmtime_adj() work with an ASN1_TIME structure,
15375 this will enable certificates using GeneralizedTime in validity
15376 dates to be checked.
15380 * Make the NEG_PUBKEY_BUG code (which tolerates invalid
15381 negative public key encodings) on by default,
15382 NO_NEG_PUBKEY_BUG can be set to disable it.
15386 * New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
15387 content octets. An i2c_ASN1_OBJECT is unnecessary because
15388 the encoding can be trivially obtained from the structure.
15392 * crypto/err.c locking bugfix: Use write locks (`CRYPTO_w_[un]lock`),
15393 not read locks (`CRYPTO_r_[un]lock`).
15397 * A first attempt at creating official support for shared
15398 libraries through configuration. I've kept it so the
15399 default is static libraries only, and the OpenSSL programs
15400 are always statically linked for now, but there are
15401 preparations for dynamic linking in place.
15402 This has been tested on Linux and Tru64.
15406 * Randomness polling function for Win9x, as described in:
15407 Peter Gutmann, Software Generation of Practically Strong
15412 * Fix so PRNG is seeded in req if using an already existing
15417 * New options to smime application. -inform and -outform
15418 allow alternative formats for the S/MIME message including
15419 PEM and DER. The -content option allows the content to be
15420 specified separately. This should allow things like Netscape
15421 form signing output easier to verify.
15425 * Fix the ASN1 encoding of tags using the 'long form'.
15429 * New ASN1 functions, `i2c_*` and `c2i_*` for INTEGER and BIT
15430 STRING types. These convert content octets to and from the
15431 underlying type. The actual tag and length octets are
15432 already assumed to have been read in and checked. These
15433 are needed because all other string types have virtually
15434 identical handling apart from the tag. By having versions
15435 of the ASN1 functions that just operate on content octets
15436 IMPLICIT tagging can be handled properly. It also allows
15437 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
15438 and ASN1_INTEGER are identical apart from the tag.
15442 * Change the handling of OID objects as follows:
15444 - New object identifiers are inserted in objects.txt, following
15445 the syntax given in [crypto/objects/README.md](crypto/objects/README.md).
15446 - objects.pl is used to process obj_mac.num and create a new
15448 - obj_dat.pl is used to create a new obj_dat.h, using the data in
15451 This is currently kind of a hack, and the perl code in objects.pl
15452 isn't very elegant, but it works as I intended. The simplest way
15453 to check that it worked correctly is to look in obj_dat.h and
15454 check the array nid_objs and make sure the objects haven't moved
15455 around (this is important!). Additions are OK, as well as
15456 consistent name changes.
15460 * Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
15464 * Addition of the command line parameter '-rand file' to 'openssl req'.
15465 The given file adds to whatever has already been seeded into the
15466 random pool through the RANDFILE configuration file option or
15467 environment variable, or the default random state file.
15471 * mkstack.pl now sorts each macro group into lexical order.
15472 Previously the output order depended on the order the files
15473 appeared in the directory, resulting in needless rewriting
15478 * Patches to make OpenSSL compile under Win32 again. Mostly
15479 work arounds for the VC++ problem that it treats func() as
15480 func(void). Also stripped out the parts of mkdef.pl that
15481 added extra typesafe functions: these no longer exist.
15485 * Reorganisation of the stack code. The macros are now all
15486 collected in safestack.h . Each macro is defined in terms of
15487 a "stack macro" of the form `SKM_<name>(type, a, b)`. The
15488 DEBUG_SAFESTACK is now handled in terms of function casts,
15489 this has the advantage of retaining type safety without the
15490 use of additional functions. If DEBUG_SAFESTACK is not defined
15491 then the non typesafe macros are used instead. Also modified the
15492 mkstack.pl script to handle the new form. Needs testing to see
15493 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
15494 the default if no major problems. Similar behaviour for ASN1_SET_OF
15495 and PKCS12_STACK_OF.
15499 * When some versions of IIS use the 'NET' form of private key the
15500 key derivation algorithm is different. Normally MD5(password) is
15501 used as a 128 bit RC4 key. In the modified case
15502 MD5(MD5(password) + "SGCKEYSALT") is used instead. Added some
15503 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
15504 as the old Netscape_RSA functions except they have an additional
15505 'sgckey' parameter which uses the modified algorithm. Also added
15506 an -sgckey command line option to the rsa utility. Thanks to
15507 Adrian Peck <bertie@ncipher.com> for posting details of the modified
15508 algorithm to openssl-dev.
15512 * The evp_local.h macros were using 'c.##kname' which resulted in
15513 invalid expansion on some systems (SCO 5.0.5 for example).
15514 Corrected to 'c.kname'.
15516 *Phillip Porch <root@theporch.com>*
15518 * New X509_get1_email() and X509_REQ_get1_email() functions that return
15519 a STACK of email addresses from a certificate or request, these look
15520 in the subject name and the subject alternative name extensions and
15521 omit any duplicate addresses.
15525 * Re-implement BN_mod_exp2_mont using independent (and larger) windows.
15526 This makes DSA verification about 2 % faster.
15530 * Increase maximum window size in `BN_mod_exp_...` to 6 bits instead of 5
15531 (meaning that now 2^5 values will be precomputed, which is only 4 KB
15532 plus overhead for 1024 bit moduli).
15533 This makes exponentiations about 0.5 % faster for 1024 bit
15534 exponents (as measured by "openssl speed rsa2048").
15538 * Rename memory handling macros to avoid conflicts with other
15540 Malloc => OPENSSL_malloc
15541 Malloc_locked => OPENSSL_malloc_locked
15542 Realloc => OPENSSL_realloc
15543 Free => OPENSSL_free
15547 * New function BN_mod_exp_mont_word for small bases (roughly 15%
15548 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
15552 * CygWin32 support.
15554 *John Jarvie <jjarvie@newsguy.com>*
15556 * The type-safe stack code has been rejigged. It is now only compiled
15557 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
15558 by default all type-specific stack functions are "#define"d back to
15559 standard stack functions. This results in more streamlined output
15560 but retains the type-safety checking possibilities of the original
15565 * The STACK code has been cleaned up, and certain type declarations
15566 that didn't make a lot of sense have been brought in line. This has
15567 also involved a cleanup of sorts in safestack.h to more correctly
15568 map type-safe stack functions onto their plain stack counterparts.
15569 This work has also resulted in a variety of "const"ifications of
15570 lots of the code, especially `_cmp` operations which should normally
15571 be prototyped with "const" parameters anyway.
15575 * When generating bytes for the first time in md_rand.c, 'stir the pool'
15576 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
15577 (The PRNG state consists of two parts, the large pool 'state' and 'md',
15578 where all of 'md' is used each time the PRNG is used, but 'state'
15579 is used only indexed by a cyclic counter. As entropy may not be
15580 well distributed from the beginning, 'md' is important as a
15581 chaining variable. However, the output function chains only half
15582 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
15583 all of 'md', and seeding with STATE_SIZE dummy bytes will result
15584 in all of 'state' being rewritten, with the new values depending
15585 on virtually all of 'md'. This overcomes the 80 bit limitation.)
15589 * In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
15590 the handshake is continued after ssl_verify_cert_chain();
15591 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
15592 can lead to 'unexplainable' connection aborts later.
15594 *Bodo Moeller; problem tracked down by Lutz Jaenicke*
15596 * Major EVP API cipher revision.
15597 Add hooks for extra EVP features. This allows various cipher
15598 parameters to be set in the EVP interface. Support added for variable
15599 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
15600 setting of RC2 and RC5 parameters.
15602 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
15605 Remove lots of duplicated code from the EVP library. For example *every*
15606 cipher init() function handles the 'iv' in the same way according to the
15607 cipher mode. They also all do nothing if the 'key' parameter is NULL and
15608 for CFB and OFB modes they zero ctx->num.
15610 New functionality allows removal of S/MIME code RC2 hack.
15612 Most of the routines have the same form and so can be declared in terms
15615 By shifting this to the top level EVP_CipherInit() it can be removed from
15616 all individual ciphers. If the cipher wants to handle IVs or keys
15617 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
15620 Change lots of functions like EVP_EncryptUpdate() to now return a
15621 value: although software versions of the algorithms cannot fail
15622 any installed hardware versions can.
15626 * Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
15627 this option is set, tolerate broken clients that send the negotiated
15628 protocol version number instead of the requested protocol version
15633 * Call dh_tmp_cb (set by `..._TMP_DH_CB`) with correct 'is_export' flag;
15634 i.e. non-zero for export ciphersuites, zero otherwise.
15635 Previous versions had this flag inverted, inconsistent with
15636 rsa_tmp_cb (..._TMP_RSA_CB).
15638 *Bodo Moeller; problem reported by Amit Chopra*
15640 * Add missing DSA library text string. Work around for some IIS
15641 key files with invalid SEQUENCE encoding.
15645 * Add a document (doc/standards.txt) that list all kinds of standards
15646 and so on that are implemented in OpenSSL.
15650 * Enhance c_rehash script. Old version would mishandle certificates
15651 with the same subject name hash and wouldn't handle CRLs at all.
15652 Added -fingerprint option to crl utility, to support new c_rehash
15657 * Eliminate non-ANSI declarations in crypto.h and stack.h.
15661 * Fix for SSL server purpose checking. Server checking was
15662 rejecting certificates which had extended key usage present
15663 but no ssl client purpose.
15665 *Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>*
15667 * Make PKCS#12 code work with no password. The PKCS#12 spec
15668 is a little unclear about how a blank password is handled.
15669 Since the password in encoded as a BMPString with terminating
15670 double NULL a zero length password would end up as just the
15671 double NULL. However no password at all is different and is
15672 handled differently in the PKCS#12 key generation code. NS
15673 treats a blank password as zero length. MSIE treats it as no
15674 password on export: but it will try both on import. We now do
15675 the same: PKCS12_parse() tries zero length and no password if
15676 the password is set to "" or NULL (NULL is now a valid password:
15677 it wasn't before) as does the pkcs12 application.
15681 * Bugfixes in `apps/x509.c`: Avoid a memory leak; and don't use
15682 perror when PEM_read_bio_X509_REQ fails, the error message must
15683 be obtained from the error queue.
15687 * Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
15688 it in ERR_remove_state if appropriate, and change ERR_get_state
15689 accordingly to avoid race conditions (this is necessary because
15690 thread_hash is no longer constant once set).
15694 * Bugfix for linux-elf makefile.one.
15698 * RSA_get_default_method() will now cause a default
15699 RSA_METHOD to be chosen if one doesn't exist already.
15700 Previously this was only set during a call to RSA_new()
15701 or RSA_new_method(NULL) meaning it was possible for
15702 RSA_get_default_method() to return NULL.
15706 * Added native name translation to the existing DSO code
15707 that will convert (if the flag to do so is set) filenames
15708 that are sufficiently small and have no path information
15709 into a canonical native form. Eg. "blah" converted to
15710 "libblah.so" or "blah.dll" etc.
15714 * New function ERR_error_string_n(e, buf, len) which is like
15715 ERR_error_string(e, buf), but writes at most 'len' bytes
15716 including the 0 terminator. For ERR_error_string_n, 'buf'
15719 *Damien Miller <djm@mindrot.org>, Bodo Moeller*
15721 * CONF library reworked to become more general. A new CONF
15722 configuration file reader "class" is implemented as well as a
15723 new functions (`NCONF_*`, for "New CONF") to handle it. The now
15724 old `CONF_*` functions are still there, but are reimplemented to
15725 work in terms of the new functions. Also, a set of functions
15726 to handle the internal storage of the configuration data is
15727 provided to make it easier to write new configuration file
15728 reader "classes" (I can definitely see something reading a
15729 configuration file in XML format, for example), called `_CONF_*`,
15730 or "the configuration storage API"...
15732 The new configuration file reading functions are:
15734 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
15735 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
15737 NCONF_default, NCONF_WIN32
15739 NCONF_dump_fp, NCONF_dump_bio
15741 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
15742 NCONF_new creates a new CONF object. This works in the same way
15743 as other interfaces in OpenSSL, like the BIO interface.
15744 `NCONF_dump_*` dump the internal storage of the configuration file,
15745 which is useful for debugging. All other functions take the same
15746 arguments as the old `CONF_*` functions with the exception of the
15747 first that must be a `CONF *` instead of a `LHASH *`.
15749 To make it easier to use the new classes with the old `CONF_*` functions,
15750 the function CONF_set_default_method is provided.
15754 * Add '-tls1' option to 'openssl ciphers', which was already
15755 mentioned in the documentation but had not been implemented.
15756 (This option is not yet really useful because even the additional
15757 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
15761 * Initial DSO code added into libcrypto for letting OpenSSL (and
15762 OpenSSL-based applications) load shared libraries and bind to
15763 them in a portable way.
15765 *Geoff Thorpe, with contributions from Richard Levitte*
15767 ### Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
15769 * Make sure _lrotl and _lrotr are only used with MSVC.
15771 * Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
15772 (the default implementation of RAND_status).
15774 * Rename openssl x509 option '-crlext', which was added in 0.9.5,
15775 to '-clrext' (= clear extensions), as intended and documented.
15776 *Bodo Moeller; inconsistency pointed out by Michael Attili
15777 <attili@amaxo.com>*
15779 * Fix for HMAC. It wasn't zeroing the rest of the block if the key length
15780 was larger than the MD block size.
15782 *Steve Henson, pointed out by Yost William <YostW@tce.com>*
15784 * Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
15785 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
15786 using the passed key: if the passed key was a private key the result
15787 of X509_print(), for example, would be to print out all the private key
15792 * des_quad_cksum() byte order bug fix.
15793 *Ulf Möller, using the problem description in krb4-0.9.7, where
15794 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>*
15796 * Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
15799 *Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>*
15801 * For easily testing in shell scripts whether some command
15802 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
15803 returns with exit code 0 iff no command of the given name is available.
15804 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
15805 the output goes to stdout and nothing is printed to stderr.
15806 Additional arguments are always ignored.
15808 Since for each cipher there is a command of the same name,
15809 the 'no-cipher' compilation switches can be tested this way.
15811 ('openssl no-XXX' is not able to detect pseudo-commands such
15812 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
15816 * Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
15820 * For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
15821 is set; it will be thrown away anyway because each handshake creates
15823 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
15824 to parameters -- in previous versions (since OpenSSL 0.9.3) the
15825 'default key' from SSL_CTX_set_tmp_dh would always be lost, meaning
15826 you effectively got SSL_OP_SINGLE_DH_USE when using this macro.
15830 * New s_client option -ign_eof: EOF at stdin is ignored, and
15831 'Q' and 'R' lose their special meanings (quit/renegotiate).
15832 This is part of what -quiet does; unlike -quiet, -ign_eof
15833 does not suppress any output.
15837 * Add compatibility options to the purpose and trust code. The
15838 purpose X509_PURPOSE_ANY is "any purpose" which automatically
15839 accepts a certificate or CA, this was the previous behaviour,
15840 with all the associated security issues.
15842 X509_TRUST_COMPAT is the old trust behaviour: only and
15843 automatically trust self signed roots in certificate store. A
15844 new trust setting X509_TRUST_DEFAULT is used to specify that
15845 a purpose has no associated trust setting and it should instead
15846 use the value in the default purpose.
15850 * Fix the PKCS#8 DSA private key code so it decodes keys again
15851 and fix a memory leak.
15855 * In util/mkerr.pl (which implements 'make errors'), preserve
15856 reason strings from the previous version of the .c file, as
15857 the default to have only downcase letters (and digits) in
15858 automatically generated reasons codes is not always appropriate.
15862 * In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
15863 using strerror. Previously, ERR_reason_error_string() returned
15864 library names as reason strings for SYSerr; but SYSerr is a special
15865 case where small numbers are errno values, not library numbers.
15869 * Add '-dsaparam' option to 'openssl dhparam' application. This
15870 converts DSA parameters into DH parameters. (When creating parameters,
15871 DSA_generate_parameters is used.)
15875 * Include 'length' (recommended exponent length) in C code generated
15876 by 'openssl dhparam -C'.
15880 * The second argument to set_label in perlasm was already being used
15881 so couldn't be used as a "file scope" flag. Moved to third argument
15886 * In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
15887 instead of RAND_bytes for encryption IVs and salts.
15891 * Include RAND_status() into RAND_METHOD instead of implementing
15892 it only for md_rand.c Otherwise replacing the PRNG by calling
15893 RAND_set_rand_method would be impossible.
15897 * Don't let DSA_generate_key() enter an infinite loop if the random
15898 number generation fails.
15902 * New 'rand' application for creating pseudo-random output.
15906 * Added configuration support for Linux/IA64
15908 *Rolf Haberrecker <rolf@suse.de>*
15910 * Assembler module support for Mingw32.
15914 * Shared library support for HPUX (in shlib/).
15916 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous*
15918 * Shared library support for Solaris gcc.
15920 *Lutz Behnke <behnke@trustcenter.de>*
15922 ### Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
15924 * PKCS7_encrypt() was adding text MIME headers twice because they
15925 were added manually and by SMIME_crlf_copy().
15929 * In bntest.c don't call BN_rand with zero bits argument.
15931 *Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>*
15933 * BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
15934 case was implemented. This caused BN_div_recp() to fail occasionally.
15938 * Add an optional second argument to the set_label() in the perl
15939 assembly language builder. If this argument exists and is set
15940 to 1 it signals that the assembler should use a symbol whose
15941 scope is the entire file, not just the current function. This
15942 is needed with MASM which uses the format label:: for this scope.
15944 *Steve Henson, pointed out by Peter Runestig <peter@runestig.com>*
15946 * Change the ASN1 types so they are typedefs by default. Before
15947 almost all types were #define'd to ASN1_STRING which was causing
15948 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
15953 * Change names of new functions to the new get1/get0 naming
15954 convention: After 'get1', the caller owns a reference count
15955 and has to call `..._free`; 'get0' returns a pointer to some
15956 data structure without incrementing reference counters.
15957 (Some of the existing 'get' functions increment a reference
15958 counter, some don't.)
15959 Similarly, 'set1' and 'add1' functions increase reference
15960 counters or duplicate objects.
15964 * Allow for the possibility of temp RSA key generation failure:
15965 the code used to assume it always worked and crashed on failure.
15969 * Fix potential buffer overrun problem in BIO_printf().
15970 *Ulf Möller, using public domain code by Patrick Powell; problem
15971 pointed out by David Sacerdote <das33@cornell.edu>*
15973 * Support EGD <http://www.lothar.com/tech/crypto/>. New functions
15974 RAND_egd() and RAND_status(). In the command line application,
15975 the EGD socket can be specified like a seed file using RANDFILE
15980 * Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
15981 Some CAs (e.g. Verisign) distribute certificates in this form.
15985 * Remove the SSL_ALLOW_ADH compile option and set the default cipher
15986 list to exclude them. This means that no special compilation option
15987 is needed to use anonymous DH: it just needs to be included in the
15992 * Change the EVP_MD_CTX_type macro so its meaning consistent with
15993 EVP_MD_type. The old functionality is available in a new macro called
15994 EVP_MD_md(). Change code that uses it and update docs.
15998 * `..._ctrl` functions now have corresponding `..._callback_ctrl` functions
15999 where the `void *` argument is replaced by a function pointer argument.
16000 Previously `void *` was abused to point to functions, which works on
16001 many platforms, but is not correct. As these functions are usually
16002 called by macros defined in OpenSSL header files, most source code
16003 should work without changes.
16007 * `<openssl/opensslconf.h>` (which is created by Configure) now contains
16008 sections with information on -D... compiler switches used for
16009 compiling the library so that applications can see them. To enable
16010 one of these sections, a pre-processor symbol `OPENSSL_..._DEFINES`
16011 must be defined. E.g.,
16012 #define OPENSSL_ALGORITHM_DEFINES
16013 #include <openssl/opensslconf.h>
16014 defines all pertinent `NO_<algo>` symbols, such as NO_IDEA, NO_RSA, etc.
16016 *Richard Levitte, Ulf and Bodo Möller*
16018 * Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
16023 * Change the 'other' type in certificate aux info to a STACK_OF
16024 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
16025 the required ASN1 format: arbitrary types determined by an OID.
16029 * Add some PEM_write_X509_REQ_NEW() functions and a command line
16030 argument to 'req'. This is not because the function is newer or
16031 better than others it just uses the work 'NEW' in the certificate
16032 request header lines. Some software needs this.
16036 * Reorganise password command line arguments: now passwords can be
16037 obtained from various sources. Delete the PEM_cb function and make
16038 it the default behaviour: i.e. if the callback is NULL and the
16039 usrdata argument is not NULL interpret it as a null terminated pass
16040 phrase. If usrdata and the callback are NULL then the pass phrase
16041 is prompted for as usual.
16045 * Add support for the Compaq Atalla crypto accelerator. If it is installed,
16046 the support is automatically enabled. The resulting binaries will
16047 autodetect the card and use it if present.
16049 *Ben Laurie and Compaq Inc.*
16051 * Work around for Netscape hang bug. This sends certificate request
16052 and server done in one record. Since this is perfectly legal in the
16053 SSL/TLS protocol it isn't a "bug" option and is on by default. See
16054 the bugs/SSLv3 entry for more info.
16058 * HP-UX tune-up: new unified configs, HP C compiler bug workaround.
16062 * Add -rand argument to smime and pkcs12 applications and read/write
16067 * New 'passwd' tool for crypt(3) and apr1 password hashes.
16071 * Add command line password options to the remaining applications.
16075 * Bug fix for BN_div_recp() for numerators with an even number of
16080 * More tests in bntest.c, and changed test_bn output.
16084 * ./config recognizes MacOS X now.
16088 * Bug fix for BN_div() when the first words of num and divisor are
16089 equal (it gave wrong results if `(rem=(n1-q*d0)&BN_MASK2) < d0)`.
16093 * Add support for various broken PKCS#8 formats, and command line
16094 options to produce them.
16098 * New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
16099 get temporary BIGNUMs from a BN_CTX.
16103 * Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
16108 * Change the `SSLeay_add_all_*()` functions to `OpenSSL_add_all_*()` and
16109 include a #define from the old name to the new. The original intent
16110 was that statically linked binaries could for example just call
16111 SSLeay_add_all_ciphers() to just add ciphers to the table and not
16112 link with digests. This never worked because SSLeay_add_all_digests()
16113 and SSLeay_add_all_ciphers() were in the same source file so calling
16114 one would link with the other. They are now in separate source files.
16118 * Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
16122 * Use a less unusual form of the Miller-Rabin primality test (it used
16123 a binary algorithm for exponentiation integrated into the Miller-Rabin
16124 loop, our standard modexp algorithms are faster).
16128 * Support for the EBCDIC character set completed.
16130 *Martin Kraemer <Martin.Kraemer@Mch.SNI.De>*
16132 * Source code cleanups: use const where appropriate, eliminate casts,
16133 use `void *` instead of `char *` in lhash.
16137 * Bugfix: ssl3_send_server_key_exchange was not restartable
16138 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
16139 this the server could overwrite ephemeral keys that the client
16144 * Turn DSA_is_prime into a macro that calls BN_is_prime,
16145 using 50 iterations of the Rabin-Miller test.
16147 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
16148 iterations of the Rabin-Miller test as required by the appendix
16149 to FIPS PUB 186[-1]) instead of DSA_is_prime.
16150 As BN_is_prime_fasttest includes trial division, DSA parameter
16151 generation becomes much faster.
16153 This implies a change for the callback functions in DSA_is_prime
16154 and DSA_generate_parameters: The callback function is called once
16155 for each positive witness in the Rabin-Miller test, not just
16156 occasionally in the inner loop; and the parameters to the
16157 callback function now provide an iteration count for the outer
16158 loop rather than for the current invocation of the inner loop.
16159 DSA_generate_parameters additionally can call the callback
16160 function with an 'iteration count' of -1, meaning that a
16161 candidate has passed the trial division test (when q is generated
16162 from an application-provided seed, trial division is skipped).
16166 * New function BN_is_prime_fasttest that optionally does trial
16167 division before starting the Rabin-Miller test and has
16168 an additional BN_CTX * argument (whereas BN_is_prime always
16169 has to allocate at least one BN_CTX).
16170 'callback(1, -1, cb_arg)' is called when a number has passed the
16171 trial division stage.
16175 * Fix for bug in CRL encoding. The validity dates weren't being handled
16180 * New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
16184 * New function BN_pseudo_rand().
16188 * Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
16189 bignum version of BN_from_montgomery() with the working code from
16190 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
16195 * Avoid a race condition in s2_clnt.c (function get_server_hello) that
16196 made it impossible to use the same SSL_SESSION data structure in
16197 SSL2 clients in multiple threads.
16201 * The return value of RAND_load_file() no longer counts bytes obtained
16202 by stat(). RAND_load_file(..., -1) is new and uses the complete file
16203 to seed the PRNG (previously an explicit byte count was required).
16205 *Ulf Möller, Bodo Möller*
16207 * Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
16208 used `char *` instead of `void *` and had casts all over the place.
16212 * Make BN_generate_prime() return NULL on error if ret!=NULL.
16216 * Retain source code compatibility for BN_prime_checks macro:
16217 BN_is_prime(..., BN_prime_checks, ...) now uses
16218 BN_prime_checks_for_size to determine the appropriate number of
16219 Rabin-Miller iterations.
16223 * Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
16224 DH_CHECK_P_NOT_SAFE_PRIME.
16225 (Check if this is true? OpenPGP calls them "strong".)
16229 * Merge the functionality of "dh" and "gendh" programs into a new program
16230 "dhparam". The old programs are retained for now but will handle DH keys
16231 (instead of parameters) in future.
16235 * Make the ciphers, s_server and s_client programs check the return values
16236 when a new cipher list is set.
16240 * Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
16241 ciphers. Before when the 56bit ciphers were enabled the sorting was
16244 The syntax for the cipher sorting has been extended to support sorting by
16245 cipher-strength (using the strength_bits hard coded in the tables).
16246 The new command is `@STRENGTH` (see also `doc/apps/ciphers.pod`).
16248 Fix a bug in the cipher-command parser: when supplying a cipher command
16249 string with an "undefined" symbol (neither command nor alphanumeric
16250 *A-Za-z0-9*, ssl_set_cipher_list used to hang in an endless loop. Now
16251 an error is flagged.
16253 Due to the strength-sorting extension, the code of the
16254 ssl_create_cipher_list() function was completely rearranged. I hope that
16255 the readability was also increased :-)
16257 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>*
16259 * Minor change to 'x509' utility. The -CAcreateserial option now uses 1
16260 for the first serial number and places 2 in the serial number file. This
16261 avoids problems when the root CA is created with serial number zero and
16262 the first user certificate has the same issuer name and serial number
16267 * Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
16268 the new code. Add documentation for this stuff.
16272 * Changes to X509_ATTRIBUTE utilities. These have been renamed from
16273 `X509_*()` to `X509at_*()` on the grounds that they don't handle X509
16274 structures and behave in an analogous way to the X509v3 functions:
16275 they shouldn't be called directly but wrapper functions should be used
16278 So we also now have some wrapper functions that call the X509at functions
16279 when passed certificate requests. (TO DO: similar things can be done with
16280 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
16281 things. Some of these need some d2i or i2d and print functionality
16282 because they handle more complex structures.)
16286 * Add missing #ifndefs that caused missing symbols when building libssl
16287 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
16288 NO_RSA in `ssl/s2*.c`.
16290 *Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller*
16292 * Precautions against using the PRNG uninitialized: RAND_bytes() now
16293 has a return value which indicates the quality of the random data
16294 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
16295 error queue. New function RAND_pseudo_bytes() generates output that is
16296 guaranteed to be unique but not unpredictable. RAND_add is like
16297 RAND_seed, but takes an extra argument for an entropy estimate
16298 (RAND_seed always assumes full entropy).
16302 * Do more iterations of Rabin-Miller probable prime test (specifically,
16303 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
16304 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
16305 in crypto/bn/bn_prime.c for the complete table). This guarantees a
16306 false-positive rate of at most 2^-80 for random input.
16310 * Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
16314 * New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
16315 in the 0.9.5 release), this returns the chain
16316 from an X509_CTX structure with a dup of the stack and all
16317 the X509 reference counts upped: so the stack will exist
16318 after X509_CTX_cleanup() has been called. Modify pkcs12.c
16321 Also make SSL_SESSION_print() print out the verify return
16326 * Add manpage for the pkcs12 command. Also change the default
16327 behaviour so MAC iteration counts are used unless the new
16328 -nomaciter option is used. This improves file security and
16329 only older versions of MSIE (4.0 for example) need it.
16333 * Honor the no-xxx Configure options when creating .DEF files.
16337 * Add PKCS#10 attributes to field table: challengePassword,
16338 unstructuredName and unstructuredAddress. These are taken from
16339 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
16340 international characters are used.
16342 More changes to X509_ATTRIBUTE code: allow the setting of types
16343 based on strings. Remove the 'loc' parameter when adding
16344 attributes because these will be a SET OF encoding which is sorted
16349 * Initial changes to the 'req' utility to allow request generation
16350 automation. This will allow an application to just generate a template
16351 file containing all the field values and have req construct the
16354 Initial support for X509_ATTRIBUTE handling. Stacks of these are
16355 used all over the place including certificate requests and PKCS#7
16356 structures. They are currently handled manually where necessary with
16357 some primitive wrappers for PKCS#7. The new functions behave in a
16358 manner analogous to the X509 extension functions: they allow
16359 attributes to be looked up by NID and added.
16361 Later something similar to the X509V3 code would be desirable to
16362 automatically handle the encoding, decoding and printing of the
16363 more complex types. The string types like challengePassword can
16364 be handled by the string table functions.
16366 Also modified the multi byte string table handling. Now there is
16367 a 'global mask' which masks out certain types. The table itself
16368 can use the flag STABLE_NO_MASK to ignore the mask setting: this
16369 is useful when for example there is only one permissible type
16370 (as in countryName) and using the mask might result in no valid
16375 * Clean up 'Finished' handling, and add functions SSL_get_finished and
16376 SSL_get_peer_finished to allow applications to obtain the latest
16377 Finished messages sent to the peer or expected from the peer,
16378 respectively. (SSL_get_peer_finished is usually the Finished message
16379 actually received from the peer, otherwise the protocol will be aborted.)
16381 As the Finished message are message digests of the complete handshake
16382 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
16383 be used for external authentication procedures when the authentication
16384 provided by SSL/TLS is not desired or is not enough.
16388 * Enhanced support for Alpha Linux is added. Now ./config checks if
16389 the host supports BWX extension and if Compaq C is present on the
16390 $PATH. Just exploiting of the BWX extension results in 20-30%
16391 performance kick for some algorithms, e.g. DES and RC4 to mention
16392 a couple. Compaq C in turn generates ~20% faster code for MD5 and
16397 * Add support for MS "fast SGC". This is arguably a violation of the
16398 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
16399 weak crypto and after checking the certificate is SGC a second one
16400 with strong crypto. MS SGC stops the first handshake after receiving
16401 the server certificate message and sends a second client hello. Since
16402 a server will typically do all the time consuming operations before
16403 expecting any further messages from the client (server key exchange
16404 is the most expensive) there is little difference between the two.
16406 To get OpenSSL to support MS SGC we have to permit a second client
16407 hello message after we have sent server done. In addition we have to
16408 reset the MAC if we do get this second client hello.
16412 * Add a function 'd2i_AutoPrivateKey()' this will automatically decide
16413 if a DER encoded private key is RSA or DSA traditional format. Changed
16414 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
16415 format DER encoded private key. Newer code should use PKCS#8 format which
16416 has the key type encoded in the ASN1 structure. Added DER private key
16417 support to pkcs8 application.
16421 * SSL 3/TLS 1 servers now don't request certificates when an anonymous
16422 ciphersuites has been selected (as required by the SSL 3/TLS 1
16423 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
16424 is set, we interpret this as a request to violate the specification
16425 (the worst that can happen is a handshake failure, and 'correct'
16426 behaviour would result in a handshake failure anyway).
16430 * In SSL_CTX_add_session, take into account that there might be multiple
16431 SSL_SESSION structures with the same session ID (e.g. when two threads
16432 concurrently obtain them from an external cache).
16433 The internal cache can handle only one SSL_SESSION with a given ID,
16434 so if there's a conflict, we now throw out the old one to achieve
16439 * Add OIDs for idea and blowfish in CBC mode. This will allow both
16440 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
16441 some routines that use cipher OIDs: some ciphers do not have OIDs
16442 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
16447 * Simplify the trust setting structure and code. Now we just have
16448 two sequences of OIDs for trusted and rejected settings. These will
16449 typically have values the same as the extended key usage extension
16450 and any application specific purposes.
16452 The trust checking code now has a default behaviour: it will just
16453 check for an object with the same NID as the passed id. Functions can
16454 be provided to override either the default behaviour or the behaviour
16455 for a given id. SSL client, server and email already have functions
16456 in place for compatibility: they check the NID and also return "trusted"
16457 if the certificate is self signed.
16461 * Add d2i,i2d bio/fp functions for PrivateKey: these convert the
16462 traditional format into an EVP_PKEY structure.
16466 * Add a password callback function PEM_cb() which either prompts for
16467 a password if usr_data is NULL or otherwise assumes it is a null
16468 terminated password. Allow passwords to be passed on command line
16469 environment or config files in a few more utilities.
16473 * Add a bunch of DER and PEM functions to handle PKCS#8 format private
16474 keys. Add some short names for PKCS#8 PBE algorithms and allow them
16475 to be specified on the command line for the pkcs8 and pkcs12 utilities.
16476 Update documentation.
16480 * Support for ASN1 "NULL" type. This could be handled before by using
16481 ASN1_TYPE but there wasn't any function that would try to read a NULL
16482 and produce an error if it couldn't. For compatibility we also have
16483 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
16484 don't allocate anything because they don't need to.
16488 * Initial support for MacOS is now provided. Examine INSTALL.MacOS
16491 *Andy Polyakov, Roy Woods <roy@centicsystems.ca>*
16493 * Rebuild of the memory allocation routines used by OpenSSL code and
16494 possibly others as well. The purpose is to make an interface that
16495 provide hooks so anyone can build a separate set of allocation and
16496 deallocation routines to be used by OpenSSL, for example memory
16497 pool implementations, or something else, which was previously hard
16498 since Malloc(), Realloc() and Free() were defined as macros having
16499 the values malloc, realloc and free, respectively (except for Win32
16500 compilations). The same is provided for memory debugging code.
16501 OpenSSL already comes with functionality to find memory leaks, but
16502 this gives people a chance to debug other memory problems.
16504 With these changes, a new set of functions and macros have appeared:
16506 CRYPTO_set_mem_debug_functions() [F]
16507 CRYPTO_get_mem_debug_functions() [F]
16508 CRYPTO_dbg_set_options() [F]
16509 CRYPTO_dbg_get_options() [F]
16510 CRYPTO_malloc_debug_init() [M]
16512 The memory debug functions are NULL by default, unless the library
16513 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
16514 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
16515 gives the standard debugging functions that come with OpenSSL) or
16516 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
16517 provided by the library user) must be used. When the standard
16518 debugging functions are used, CRYPTO_dbg_set_options can be used to
16519 request additional information:
16520 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
16521 the CRYPTO_MDEBUG_xxx macro when compiling the library.
16523 Also, things like CRYPTO_set_mem_functions will always give the
16524 expected result (the new set of functions is used for allocation
16525 and deallocation) at all times, regardless of platform and compiler
16528 To finish it up, some functions that were never use in any other
16529 way than through macros have a new API and new semantic:
16531 CRYPTO_dbg_malloc()
16532 CRYPTO_dbg_realloc()
16535 All macros of value have retained their old syntax.
16537 *Richard Levitte and Bodo Moeller*
16539 * Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
16540 ordering of SMIMECapabilities wasn't in "strength order" and there
16541 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
16546 * Some ASN1 types with illegal zero length encoding (INTEGER,
16547 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
16549 *Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson*
16551 * Merge in my S/MIME library for OpenSSL. This provides a simple
16552 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
16553 functionality to handle multipart/signed properly) and a utility
16554 called 'smime' to call all this stuff. This is based on code I
16555 originally wrote for Celo who have kindly allowed it to be
16556 included in OpenSSL.
16560 * Add variants des_set_key_checked and des_set_key_unchecked of
16561 des_set_key (aka des_key_sched). Global variable des_check_key
16562 decides which of these is called by des_set_key; this way
16563 des_check_key behaves as it always did, but applications and
16564 the library itself, which was buggy for des_check_key == 1,
16565 have a cleaner way to pick the version they need.
16569 * New function PKCS12_newpass() which changes the password of a
16574 * Modify X509_TRUST and X509_PURPOSE so it also uses a static and
16575 dynamic mix. In both cases the ids can be used as an index into the
16576 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
16577 functions so they accept a list of the field values and the
16578 application doesn't need to directly manipulate the X509_TRUST
16583 * Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
16588 * Modify the way the V3 extension code looks up extensions. This now
16589 works in a similar way to the object code: we have some "standard"
16590 extensions in a static table which is searched with OBJ_bsearch()
16591 and the application can add dynamic ones if needed. The file
16592 crypto/x509v3/ext_dat.h now has the info: this file needs to be
16593 updated whenever a new extension is added to the core code and kept
16594 in ext_nid order. There is a simple program 'tabtest.c' which checks
16595 this. New extensions are not added too often so this file can readily
16596 be maintained manually.
16598 There are two big advantages in doing things this way. The extensions
16599 can be looked up immediately and no longer need to be "added" using
16600 X509V3_add_standard_extensions(): this function now does nothing.
16601 Side note: I get *lots* of email saying the extension code doesn't
16602 work because people forget to call this function.
16603 Also no dynamic allocation is done unless new extensions are added:
16604 so if we don't add custom extensions there is no need to call
16605 X509V3_EXT_cleanup().
16609 * Modify enc utility's salting as follows: make salting the default. Add a
16610 magic header, so unsalted files fail gracefully instead of just decrypting
16611 to garbage. This is because not salting is a big security hole, so people
16612 should be discouraged from doing it.
16616 * Fixes and enhancements to the 'x509' utility. It allowed a message
16617 digest to be passed on the command line but it only used this
16618 parameter when signing a certificate. Modified so all relevant
16619 operations are affected by the digest parameter including the
16620 -fingerprint and -x509toreq options. Also -x509toreq choked if a
16621 DSA key was used because it didn't fix the digest.
16625 * Initial certificate chain verify code. Currently tests the untrusted
16626 certificates for consistency with the verify purpose (which is set
16627 when the X509_STORE_CTX structure is set up) and checks the pathlength.
16629 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
16630 this is because it will reject chains with invalid extensions whereas
16631 every previous version of OpenSSL and SSLeay made no checks at all.
16633 Trust code: checks the root CA for the relevant trust settings. Trust
16634 settings have an initial value consistent with the verify purpose: e.g.
16635 if the verify purpose is for SSL client use it expects the CA to be
16636 trusted for SSL client use. However the default value can be changed to
16637 permit custom trust settings: one example of this would be to only trust
16638 certificates from a specific "secure" set of CAs.
16640 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
16641 which should be used for version portability: especially since the
16642 verify structure is likely to change more often now.
16644 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
16645 to set them. If not set then assume SSL clients will verify SSL servers
16648 Two new options to the verify program: -untrusted allows a set of
16649 untrusted certificates to be passed in and -purpose which sets the
16650 intended purpose of the certificate. If a purpose is set then the
16651 new chain verify code is used to check extension consistency.
16655 * Support for the authority information access extension.
16659 * Modify RSA and DSA PEM read routines to transparently handle
16660 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
16661 public keys in a format compatible with certificate
16662 SubjectPublicKeyInfo structures. Unfortunately there were already
16663 functions called *_PublicKey_* which used various odd formats so
16664 these are retained for compatibility: however the DSA variants were
16665 never in a public release so they have been deleted. Changed dsa/rsa
16666 utilities to handle the new format: note no releases ever handled public
16667 keys so we should be OK.
16669 The primary motivation for this change is to avoid the same fiasco
16670 that dogs private keys: there are several incompatible private key
16671 formats some of which are standard and some OpenSSL specific and
16672 require various evil hacks to allow partial transparent handling and
16673 even then it doesn't work with DER formats. Given the option anything
16674 other than PKCS#8 should be dumped: but the other formats have to
16675 stay in the name of compatibility.
16677 With public keys and the benefit of hindsight one standard format
16678 is used which works with EVP_PKEY, RSA or DSA structures: though
16679 it clearly returns an error if you try to read the wrong kind of key.
16681 Added a -pubkey option to the 'x509' utility to output the public key.
16682 Also rename the `EVP_PKEY_get_*()` to `EVP_PKEY_rget_*()`
16683 (renamed to `EVP_PKEY_get1_*()` in the OpenSSL 0.9.5 release) and add
16684 `EVP_PKEY_rset_*()` functions (renamed to `EVP_PKEY_set1_*()`)
16685 that do the same as the `EVP_PKEY_assign_*()` except they up the
16686 reference count of the added key (they don't "swallow" the
16691 * Fixes to crypto/x509/by_file.c the code to read in certificates and
16692 CRLs would fail if the file contained no certificates or no CRLs:
16693 added a new function to read in both types and return the number
16694 read: this means that if none are read it will be an error. The
16695 DER versions of the certificate and CRL reader would always fail
16696 because it isn't possible to mix certificates and CRLs in DER format
16697 without choking one or the other routine. Changed this to just read
16698 a certificate: this is the best we can do. Also modified the code
16699 in `apps/verify.c` to take notice of return codes: it was previously
16700 attempting to read in certificates from NULL pointers and ignoring
16701 any errors: this is one reason why the cert and CRL reader seemed
16702 to work. It doesn't check return codes from the default certificate
16703 routines: these may well fail if the certificates aren't installed.
16707 * Code to support otherName option in GeneralName.
16711 * First update to verify code. Change the verify utility
16712 so it warns if it is passed a self signed certificate:
16713 for consistency with the normal behaviour. X509_verify
16714 has been modified to it will now verify a self signed
16715 certificate if *exactly* the same certificate appears
16716 in the store: it was previously impossible to trust a
16717 single self signed certificate. This means that:
16718 openssl verify ss.pem
16719 now gives a warning about a self signed certificate but
16720 openssl verify -CAfile ss.pem ss.pem
16725 * For servers, store verify_result in SSL_SESSION data structure
16726 (and add it to external session representation).
16727 This is needed when client certificate verifications fails,
16728 but an application-provided verification callback (set by
16729 SSL_CTX_set_cert_verify_callback) allows accepting the session
16730 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
16731 but returns 1): When the session is reused, we have to set
16732 ssl->verify_result to the appropriate error code to avoid
16735 *Bodo Moeller, problem pointed out by Lutz Jaenicke*
16737 * Fix a bug in the new PKCS#7 code: it didn't consider the
16738 case in PKCS7_dataInit() where the signed PKCS7 structure
16739 didn't contain any existing data because it was being created.
16741 *Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson*
16743 * Add a salt to the key derivation routines in enc.c. This
16744 forms the first 8 bytes of the encrypted file. Also add a
16745 -S option to allow a salt to be input on the command line.
16749 * New function X509_cmp(). Oddly enough there wasn't a function
16750 to compare two certificates. We do this by working out the SHA1
16751 hash and comparing that. X509_cmp() will be needed by the trust
16756 * SSL_get1_session() is like SSL_get_session(), but increments
16757 the reference count in the SSL_SESSION returned.
16759 *Geoff Thorpe <geoff@eu.c2.net>*
16761 * Fix for 'req': it was adding a null to request attributes.
16762 Also change the X509_LOOKUP and X509_INFO code to handle
16763 certificate auxiliary information.
16767 * Add support for 40 and 64 bit RC2 and RC4 algorithms: document
16772 * Add the possibility to add extra information to the memory leak
16773 detecting output, to form tracebacks, showing from where each
16774 allocation was originated: CRYPTO_push_info("constant string") adds
16775 the string plus current file name and line number to a per-thread
16776 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
16777 is like calling CYRPTO_pop_info() until the stack is empty.
16778 Also updated memory leak detection code to be multi-thread-safe.
16782 * Add options -text and -noout to pkcs7 utility and delete the
16783 encryption options which never did anything. Update docs.
16787 * Add options to some of the utilities to allow the pass phrase
16788 to be included on either the command line (not recommended on
16789 OSes like Unix) or read from the environment. Update the
16790 manpages and fix a few bugs.
16794 * Add a few manpages for some of the openssl commands.
16798 * Fix the -revoke option in ca. It was freeing up memory twice,
16799 leaking and not finding already revoked certificates.
16803 * Extensive changes to support certificate auxiliary information.
16804 This involves the use of X509_CERT_AUX structure and X509_AUX
16805 functions. An X509_AUX function such as PEM_read_X509_AUX()
16806 can still read in a certificate file in the usual way but it
16807 will also read in any additional "auxiliary information". By
16808 doing things this way a fair degree of compatibility can be
16809 retained: existing certificates can have this information added
16810 using the new 'x509' options.
16812 Current auxiliary information includes an "alias" and some trust
16813 settings. The trust settings will ultimately be used in enhanced
16814 certificate chain verification routines: currently a certificate
16815 can only be trusted if it is self signed and then it is trusted
16820 * Fix assembler for Alpha (tested only on DEC OSF not Linux or `*BSD`).
16821 The problem was that one of the replacement routines had not been working
16822 since SSLeay releases. For now the offending routine has been replaced
16823 with non-optimised assembler. Even so, this now gives around 95%
16824 performance improvement for 1024 bit RSA signs.
16828 * Hack to fix PKCS#7 decryption when used with some unorthodox RC2
16829 handling. Most clients have the effective key size in bits equal to
16830 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
16831 A few however don't do this and instead use the size of the decrypted key
16832 to determine the RC2 key length and the AlgorithmIdentifier to determine
16833 the effective key length. In this case the effective key length can still
16834 be 40 bits but the key length can be 168 bits for example. This is fixed
16835 by manually forcing an RC2 key into the EVP_PKEY structure because the
16836 EVP code can't currently handle unusual RC2 key sizes: it always assumes
16837 the key length and effective key length are equal.
16841 * Add a bunch of functions that should simplify the creation of
16842 X509_NAME structures. Now you should be able to do:
16843 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
16844 and have it automatically work out the correct field type and fill in
16845 the structures. The more adventurous can try:
16846 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
16847 and it will (hopefully) work out the correct multibyte encoding.
16851 * Change the 'req' utility to use the new field handling and multibyte
16852 copy routines. Before the DN field creation was handled in an ad hoc
16853 way in req, ca, and x509 which was rather broken and didn't support
16854 BMPStrings or UTF8Strings. Since some software doesn't implement
16855 BMPStrings or UTF8Strings yet, they can be enabled using the config file
16856 using the dirstring_type option. See the new comment in the default
16857 openssl.cnf for more info.
16861 * Make crypto/rand/md_rand.c more robust:
16862 - Assure unique random numbers after fork().
16863 - Make sure that concurrent threads access the global counter and
16864 md serializably so that we never lose entropy in them
16865 or use exactly the same state in multiple threads.
16866 Access to the large state is not always serializable because
16867 the additional locking could be a performance killer, and
16868 md should be large enough anyway.
16872 * New file `apps/app_rand.c` with commonly needed functionality
16873 for handling the random seed file.
16875 Use the random seed file in some applications that previously did not:
16877 dsaparam -genkey (which also ignored its '-rand' option),
16880 x509 (when signing).
16881 Except on systems with /dev/urandom, it is crucial to have a random
16882 seed file at least for key creation, DSA signing, and for DH exchanges;
16883 for RSA signatures we could do without one.
16885 gendh and gendsa (unlike genrsa) used to read only the first byte
16886 of each file listed in the '-rand' option. The function as previously
16887 found in genrsa is now in app_rand.c and is used by all programs
16888 that support '-rand'.
16892 * In RAND_write_file, use mode 0600 for creating files;
16893 don't just chmod when it may be too late.
16897 * Report an error from X509_STORE_load_locations
16898 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
16902 * New function ASN1_mbstring_copy() this copies a string in either
16903 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
16904 into an ASN1_STRING type. A mask of permissible types is passed
16905 and it chooses the "minimal" type to use or an error if not type
16910 * Add function equivalents to the various macros in asn1.h. The old
16911 macros are retained with an `M_` prefix. Code inside the library can
16912 use the `M_` macros. External code (including the openssl utility)
16913 should *NOT* in order to be "shared library friendly".
16917 * Add various functions that can check a certificate's extensions
16918 to see if it usable for various purposes such as SSL client,
16919 server or S/MIME and CAs of these types. This is currently
16920 VERY EXPERIMENTAL but will ultimately be used for certificate chain
16921 verification. Also added a -purpose flag to x509 utility to
16922 print out all the purposes.
16926 * Add a CRYPTO_EX_DATA to X509 certificate structure and associated
16931 * New `X509V3_{X509,CRL,REVOKED}_get_d2i()` functions. These will search
16932 for, obtain and decode and extension and obtain its critical flag.
16933 This allows all the necessary extension code to be handled in a
16934 single function call.
16938 * RC4 tune-up featuring 30-40% performance improvement on most RISC
16939 platforms. See crypto/rc4/rc4_enc.c for further details.
16943 * New -noout option to asn1parse. This causes no output to be produced
16944 its main use is when combined with -strparse and -out to extract data
16945 from a file (which may not be in ASN.1 format).
16949 * Fix for pkcs12 program. It was hashing an invalid certificate pointer
16950 when producing the local key id.
16952 *Richard Levitte <levitte@stacken.kth.se>*
16954 * New option -dhparam in s_server. This allows a DH parameter file to be
16955 stated explicitly. If it is not stated then it tries the first server
16956 certificate file. The previous behaviour hard coded the filename
16961 * Add -pubin and -pubout options to the rsa and dsa commands. These allow
16962 a public key to be input or output. For example:
16963 openssl rsa -in key.pem -pubout -out pubkey.pem
16964 Also added necessary DSA public key functions to handle this.
16968 * Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
16969 in the message. This was handled by allowing
16970 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
16972 *Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>*
16974 * Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
16975 to the end of the strings whereas this didn't. This would cause problems
16976 if strings read with d2i_ASN1_bytes() were later modified.
16978 *Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>*
16980 * Fix for base64 decode bug. When a base64 bio reads only one line of
16981 data and it contains EOF it will end up returning an error. This is
16982 caused by input 46 bytes long. The cause is due to the way base64
16983 BIOs find the start of base64 encoded data. They do this by trying a
16984 trial decode on each line until they find one that works. When they
16985 do a flag is set and it starts again knowing it can pass all the
16986 data directly through the decoder. Unfortunately it doesn't reset
16987 the context it uses. This means that if EOF is reached an attempt
16988 is made to pass two EOFs through the context and this causes the
16989 resulting error. This can also cause other problems as well. As is
16990 usual with these problems it takes *ages* to find and the fix is
16991 trivial: move one line.
16993 *Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer)*
16995 * Ugly workaround to get s_client and s_server working under Windows. The
16996 old code wouldn't work because it needed to select() on sockets and the
16997 tty (for keypresses and to see if data could be written). Win32 only
16998 supports select() on sockets so we select() with a 1s timeout on the
16999 sockets and then see if any characters are waiting to be read, if none
17000 are present then we retry, we also assume we can always write data to
17001 the tty. This isn't nice because the code then blocks until we've
17002 received a complete line of data and it is effectively polling the
17003 keyboard at 1s intervals: however it's quite a bit better than not
17004 working at all :-) A dedicated Windows application might handle this
17005 with an event loop for example.
17009 * Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
17010 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
17011 will be called when RSA_sign() and RSA_verify() are used. This is useful
17012 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
17013 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
17014 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
17015 This necessitated the support of an extra signature type NID_md5_sha1
17016 for SSL signatures and modifications to the SSL library to use it instead
17017 of calling RSA_public_decrypt() and RSA_private_encrypt().
17021 * Add new -verify -CAfile and -CApath options to the crl program, these
17022 will lookup a CRL issuers certificate and verify the signature in a
17023 similar way to the verify program. Tidy up the crl program so it
17024 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
17025 less strict. It will now permit CRL extensions even if it is not
17026 a V2 CRL: this will allow it to tolerate some broken CRLs.
17030 * Initialize all non-automatic variables each time one of the openssl
17031 sub-programs is started (this is necessary as they may be started
17032 multiple times from the "OpenSSL>" prompt).
17034 *Lennart Bang, Bodo Moeller*
17036 * Preliminary compilation option RSA_NULL which disables RSA crypto without
17037 removing all other RSA functionality (this is what NO_RSA does). This
17038 is so (for example) those in the US can disable those operations covered
17039 by the RSA patent while allowing storage and parsing of RSA keys and RSA
17044 * Non-copying interface to BIO pairs.
17045 (still largely untested)
17049 * New function ASN1_tag2str() to convert an ASN1 tag to a descriptive
17050 ASCII string. This was handled independently in various places before.
17054 * New functions UTF8_getc() and UTF8_putc() that parse and generate
17055 UTF8 strings a character at a time.
17059 * Use client_version from client hello to select the protocol
17060 (s23_srvr.c) and for RSA client key exchange verification
17061 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
17065 * Add various utility functions to handle SPKACs, these were previously
17066 handled by poking round in the structure internals. Added new function
17067 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
17068 print, verify and generate SPKACs. Based on an original idea from
17069 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
17073 * RIPEMD160 is operational on all platforms and is back in 'make test'.
17077 * Allow the config file extension section to be overwritten on the
17078 command line. Based on an original idea from Massimiliano Pala
17079 <madwolf@comune.modena.it>. The new option is called -extensions
17080 and can be applied to ca, req and x509. Also -reqexts to override
17081 the request extensions in req and -crlexts to override the crl extensions
17086 * Add new feature to the SPKAC handling in ca. Now you can include
17087 the same field multiple times by preceding it by "XXXX." for example:
17090 this is the same syntax as used in the req config file.
17094 * Allow certificate extensions to be added to certificate requests. These
17095 are specified in a 'req_extensions' option of the req section of the
17096 config file. They can be printed out with the -text option to req but
17097 are otherwise ignored at present.
17101 * Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
17102 data read consists of only the final block it would not decrypted because
17103 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
17104 A misplaced 'break' also meant the decrypted final block might not be
17105 copied until the next read.
17109 * Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
17110 a few extra parameters to the DH structure: these will be useful if
17111 for example we want the value of 'q' or implement X9.42 DH.
17115 * Initial support for DSA_METHOD. This is based on the RSA_METHOD and
17116 provides hooks that allow the default DSA functions or functions on a
17117 "per key" basis to be replaced. This allows hardware acceleration and
17118 hardware key storage to be handled without major modification to the
17119 library. Also added low-level modexp hooks and CRYPTO_EX structure and
17120 associated functions.
17124 * Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
17125 as "read only": it can't be written to and the buffer it points to will
17126 not be freed. Reading from a read only BIO is much more efficient than
17127 a normal memory BIO. This was added because there are several times when
17128 an area of memory needs to be read from a BIO. The previous method was
17129 to create a memory BIO and write the data to it, this results in two
17130 copies of the data and an O(n^2) reading algorithm. There is a new
17131 function BIO_new_mem_buf() which creates a read only memory BIO from
17132 an area of memory. Also modified the PKCS#7 routines to use read only
17137 * Bugfix: ssl23_get_client_hello did not work properly when called in
17138 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
17139 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
17140 but a retry condition occurred while trying to read the rest.
17144 * The PKCS7_ENC_CONTENT_new() function was setting the content type as
17145 NID_pkcs7_encrypted by default: this was wrong since this should almost
17146 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
17147 the encrypted data type: this is a more sensible place to put it and it
17148 allows the PKCS#12 code to be tidied up that duplicated this
17153 * Changed obj_dat.pl script so it takes its input and output files on
17154 the command line. This should avoid shell escape redirection problems
17159 * Initial support for certificate extension requests, these are included
17160 in things like Xenroll certificate requests. Included functions to allow
17161 extensions to be obtained and added.
17165 * -crlf option to s_client and s_server for sending newlines as
17166 CRLF (as required by many protocols).
17170 ### Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
17172 * Install libRSAglue.a when OpenSSL is built with RSAref.
17174 *Ralf S. Engelschall*
17176 * A few more `#ifndef NO_FP_API / #endif` pairs for consistency.
17178 *Andrija Antonijevic <TheAntony2@bigfoot.com>*
17180 * Fix -startdate and -enddate (which was missing) arguments to 'ca'
17185 * New function DSA_dup_DH, which duplicates DSA parameters/keys as
17186 DH parameters/keys (q is lost during that conversion, but the resulting
17187 DH parameters contain its length).
17189 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
17190 much faster than DH_generate_parameters (which creates parameters
17191 where `p = 2*q + 1`), and also the smaller q makes DH computations
17192 much more efficient (160-bit exponentiation instead of 1024-bit
17193 exponentiation); so this provides a convenient way to support DHE
17194 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
17195 utter importance to use
17196 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
17198 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
17199 when such DH parameters are used, because otherwise small subgroup
17200 attacks may become possible!
17204 * Avoid memory leak in i2d_DHparams.
17208 * Allow the -k option to be used more than once in the enc program:
17209 this allows the same encrypted message to be read by multiple recipients.
17213 * New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
17214 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
17215 it will always use the numerical form of the OID, even if it has a short
17220 * Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
17221 method only got called if p,q,dmp1,dmq1,iqmp components were present,
17222 otherwise bn_mod_exp was called. In the case of hardware keys for example
17223 no private key components need be present and it might store extra data
17224 in the RSA structure, which cannot be accessed from bn_mod_exp.
17225 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
17226 private key operations.
17230 * Added support for SPARC Linux.
17234 * pem_password_cb function type incompatibly changed from
17235 typedef int pem_password_cb(char *buf, int size, int rwflag);
17237 ....(char *buf, int size, int rwflag, void *userdata);
17238 so that applications can pass data to their callbacks:
17239 The `PEM[_ASN1]_{read,write}...` functions and macros now take an
17240 additional void * argument, which is just handed through whenever
17241 the password callback is called.
17243 *Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller*
17245 New function SSL_CTX_set_default_passwd_cb_userdata.
17247 Compatibility note: As many C implementations push function arguments
17248 onto the stack in reverse order, the new library version is likely to
17249 interoperate with programs that have been compiled with the old
17250 pem_password_cb definition (PEM_whatever takes some data that
17251 happens to be on the stack as its last argument, and the callback
17252 just ignores this garbage); but there is no guarantee whatsoever that
17255 * The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
17256 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
17257 problems not only on Windows, but also on some Unix platforms.
17258 To avoid problematic command lines, these definitions are now in an
17259 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
17260 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
17264 * MIPS III/IV assembler module is reimplemented.
17268 * More DES library cleanups: remove references to srand/rand and
17269 delete an unused file.
17273 * Add support for the free Netwide assembler (NASM) under Win32,
17274 since not many people have MASM (ml) and it can be hard to obtain.
17275 This is currently experimental but it seems to work OK and pass all
17276 the tests. Check out INSTALL.W32 for info.
17280 * Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
17281 without temporary keys kept an extra copy of the server key,
17282 and connections with temporary keys did not free everything in case
17287 * New function RSA_check_key and new openssl rsa option -check
17288 for verifying the consistency of RSA keys.
17290 *Ulf Moeller, Bodo Moeller*
17292 * Various changes to make Win32 compile work:
17293 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
17294 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
17295 comparison" warnings.
17296 3. Add `sk_<TYPE>_sort` to DEF file generator and do make update.
17300 * Add a debugging option to PKCS#5 v2 key generation function: when
17301 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
17302 derived keys are printed to stderr.
17306 * Copy the flags in ASN1_STRING_dup().
17308 *Roman E. Pavlov <pre@mo.msk.ru>*
17310 * The x509 application mishandled signing requests containing DSA
17311 keys when the signing key was also DSA and the parameters didn't match.
17313 It was supposed to omit the parameters when they matched the signing key:
17314 the verifying software was then supposed to automatically use the CA's
17315 parameters if they were absent from the end user certificate.
17317 Omitting parameters is no longer recommended. The test was also
17318 the wrong way round! This was probably due to unusual behaviour in
17319 EVP_cmp_parameters() which returns 1 if the parameters match.
17320 This meant that parameters were omitted when they *didn't* match and
17321 the certificate was useless. Certificates signed with 'ca' didn't have
17324 *Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>*
17326 * Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
17327 The interface is as follows:
17328 Applications can use
17329 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
17330 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
17331 "off" is now the default.
17332 The library internally uses
17333 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
17334 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
17335 to disable memory-checking temporarily.
17337 Some inconsistent states that previously were possible (and were
17338 even the default) are now avoided.
17340 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
17341 with each memory chunk allocated; this is occasionally more helpful
17342 than just having a counter.
17344 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
17346 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
17351 * Introduce "mode" for SSL structures (with defaults in SSL_CTX),
17352 which largely parallels "options", but is for changing API behaviour,
17353 whereas "options" are about protocol behaviour.
17354 Initial "mode" flags are:
17356 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
17357 a single record has been written.
17358 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
17359 retries use the same buffer location.
17360 (But all of the contents must be
17365 * Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
17368 * Fix problems with no-hmac etc.
17370 *Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>*
17372 * New functions RSA_get_default_method(), RSA_set_method() and
17373 RSA_get_method(). These allows replacement of RSA_METHODs without having
17374 to mess around with the internals of an RSA structure.
17378 * Fix memory leaks in DSA_do_sign and DSA_is_prime.
17379 Also really enable memory leak checks in openssl.c and in some
17382 *Chad C. Mulligan, Bodo Moeller*
17384 * Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
17385 up the length of negative integers. This has now been simplified to just
17386 store the length when it is first determined and use it later, rather
17387 than trying to keep track of where data is copied and updating it to
17389 *Steve Henson, reported by Brien Wheeler <bwheeler@authentica-security.com>*
17391 * Add a new function PKCS7_signatureVerify. This allows the verification
17392 of a PKCS#7 signature but with the signing certificate passed to the
17393 function itself. This contrasts with PKCS7_dataVerify which assumes the
17394 certificate is present in the PKCS#7 structure. This isn't always the
17395 case: certificates can be omitted from a PKCS#7 structure and be
17396 distributed by "out of band" means (such as a certificate database).
17400 * Complete the `PEM_*` macros with DECLARE_PEM versions to replace the
17401 function prototypes in pem.h, also change util/mkdef.pl to add the
17402 necessary function names.
17406 * mk1mf.pl (used by Windows builds) did not properly read the
17407 options set by Configure in the top level Makefile, and Configure
17408 was not even able to write more than one option correctly.
17409 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
17413 * New functions CONF_load_bio() and CONF_load_fp() to allow a config
17414 file to be loaded from a BIO or FILE pointer. The BIO version will
17415 for example allow memory BIOs to contain config info.
17419 * New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
17420 Whoever hopes to achieve shared-library compatibility across versions
17421 must use this, not the compile-time macro.
17422 (Exercise 0.9.4: Which is the minimum library version required by
17424 Note: All this applies only to multi-threaded programs, others don't
17429 * Add missing case to s3_clnt.c state machine -- one of the new SSL tests
17430 through a BIO pair triggered the default case, i.e.
17431 SSLerr(...,SSL_R_UNKNOWN_STATE).
17435 * New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
17436 can use the SSL library even if none of the specific BIOs is
17441 * Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
17442 for the encoded length.
17444 *Jeon KyoungHo <khjeon@sds.samsung.co.kr>*
17446 * Add initial documentation of the X509V3 functions.
17450 * Add a new pair of functions PEM_write_PKCS8PrivateKey() and
17451 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
17452 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
17453 secure PKCS#8 private key format with a high iteration count.
17457 * Fix determination of Perl interpreter: A perl or perl5
17458 *directory* in $PATH was also accepted as the interpreter.
17460 *Ralf S. Engelschall*
17462 * Fix demos/sign/sign.c: well there wasn't anything strictly speaking
17463 wrong with it but it was very old and did things like calling
17464 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
17465 unusual formatting.
17469 * Fix demos/selfsign.c: it used obsolete and deleted functions, changed
17470 to use the new extension code.
17474 * Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
17475 with macros. This should make it easier to change their form, add extra
17476 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
17481 * Add to configuration table a new entry that can specify an alternative
17482 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
17483 according to Mark Crispin <MRC@Panda.COM>.
17487 * DES CBC did not update the IV. Weird.
17491 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
17492 Changing the behaviour of the former might break existing programs --
17493 where IV updating is needed, des_ncbc_encrypt can be used.
17496 * When bntest is run from "make test" it drives bc to check its
17497 calculations, as well as internally checking them. If an internal check
17498 fails, it needs to cause bc to give a non-zero result or make test carries
17499 on without noticing the failure. Fixed.
17503 * DES library cleanups.
17507 * Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
17508 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
17509 ciphers. NOTE: although the key derivation function has been verified
17510 against some published test vectors it has not been extensively tested
17511 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
17516 * Instead of "mkdir -p", which is not fully portable, use new
17517 Perl script "util/mkdir-p.pl".
17521 * Rewrite the way password based encryption (PBE) is handled. It used to
17522 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
17523 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
17524 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
17525 the 'parameter' field of the AlgorithmIdentifier is passed to the
17526 underlying key generation function so it must do its own ASN1 parsing.
17527 This has also changed the EVP_PBE_CipherInit() function which now has a
17528 'parameter' argument instead of literal salt and iteration count values
17529 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
17533 * Support for PKCS#5 v1.5 compatible password based encryption algorithms
17534 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
17535 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
17536 KEY" because this clashed with PKCS#8 unencrypted string. Since this
17537 value was just used as a "magic string" and not used directly its
17538 value doesn't matter.
17542 * Introduce some semblance of const correctness to BN. Shame C doesn't
17547 * "linux-sparc64" configuration (ultrapenguin).
17549 *Ray Miller <ray.miller@oucs.ox.ac.uk>*
17550 "linux-sparc" configuration.
17552 *Christian Forster <fo@hawo.stw.uni-erlangen.de>*
17554 * config now generates no-xxx options for missing ciphers.
17558 * Support the EBCDIC character set (work in progress).
17559 File ebcdic.c not yet included because it has a different license.
17561 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
17563 * Support BS2000/OSD-POSIX.
17565 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
17567 * Make callbacks for key generation use `void *` instead of `char *`.
17571 * Make S/MIME samples compile (not yet tested).
17575 * Additional typesafe stacks.
17579 * New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
17583 ### Changes between 0.9.3 and 0.9.3a [29 May 1999]
17585 * New configuration variant "sco5-gcc".
17587 * Updated some demos.
17589 *Sean O Riordain, Wade Scholine*
17591 * Add missing BIO_free at exit of pkcs12 application.
17595 * Fix memory leak in conf.c.
17599 * Updates for Win32 to assembler version of MD5.
17603 * Set #! path to perl in `apps/der_chop` to where we found it
17604 instead of using a fixed path.
17608 * SHA library changes for irix64-mips4-cc.
17612 * Improvements for VMS support.
17616 ### Changes between 0.9.2b and 0.9.3 [24 May 1999]
17618 * Bignum library bug fix. IRIX 6 passes "make test" now!
17619 This also avoids the problems with SC4.2 and unpatched SC5.
17621 *Andy Polyakov <appro@fy.chalmers.se>*
17623 * New functions sk_num, sk_value and sk_set to replace the previous macros.
17624 These are required because of the typesafe stack would otherwise break
17625 existing code. If old code used a structure member which used to be STACK
17626 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
17627 sk_num or sk_value it would produce an error because the num, data members
17628 are not present in STACK_OF. Now it just produces a warning. sk_set
17629 replaces the old method of assigning a value to sk_value
17630 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
17631 that does this will no longer work (and should use sk_set instead) but
17632 this could be regarded as a "questionable" behaviour anyway.
17636 * Fix most of the other PKCS#7 bugs. The "experimental" code can now
17637 correctly handle encrypted S/MIME data.
17641 * Change type of various DES function arguments from des_cblock
17642 (which means, in function argument declarations, pointer to char)
17643 to des_cblock * (meaning pointer to array with 8 char elements),
17644 which allows the compiler to do more typechecking; it was like
17645 that back in SSLeay, but with lots of ugly casts.
17647 Introduce new type const_des_cblock.
17651 * Reorganise the PKCS#7 library and get rid of some of the more obvious
17652 problems: find RecipientInfo structure that matches recipient certificate
17653 and initialise the ASN1 structures properly based on passed cipher.
17657 * Belatedly make the BN tests actually check the results.
17661 * Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
17662 to and from BNs: it was completely broken. New compilation option
17663 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
17664 key elements as negative integers.
17668 * Reorganize and speed up MD5.
17670 *Andy Polyakov <appro@fy.chalmers.se>*
17674 *Richard Levitte <richard@levitte.org>*
17676 * New option -out to asn1parse to allow the parsed structure to be
17677 output to a file. This is most useful when combined with the -strparse
17678 option to examine the output of things like OCTET STRINGS.
17682 * Make SSL library a little more fool-proof by not requiring any longer
17683 that `SSL_set_{accept,connect}_state` be called before
17684 `SSL_{accept,connect}` may be used (`SSL_set_..._state` is omitted
17685 in many applications because usually everything *appeared* to work as
17686 intended anyway -- now it really works as intended).
17690 * Move openssl.cnf out of lib/.
17694 * Fix various things to let OpenSSL even pass "egcc -pipe -O2 -Wall
17695 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
17696 -Wmissing-declarations -Wnested-externs -Winline" with EGCS 1.1.2+
17698 *Ralf S. Engelschall*
17700 * Various fixes to the EVP and PKCS#7 code. It may now be able to
17701 handle PKCS#7 enveloped data properly.
17703 *Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve*
17705 * Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
17706 copying pointers. The cert_st handling is changed by this in
17707 various ways (and thus what used to be known as ctx->default_cert
17708 is now called ctx->cert, since we don't resort to `s->ctx->[default_]cert`
17709 any longer when s->cert does not give us what we need).
17710 ssl_cert_instantiate becomes obsolete by this change.
17711 As soon as we've got the new code right (possibly it already is?),
17712 we have solved a couple of bugs of the earlier code where s->cert
17713 was used as if it could not have been shared with other SSL structures.
17715 Note that using the SSL API in certain dirty ways now will result
17716 in different behaviour than observed with earlier library versions:
17717 Changing settings for an `SSL_CTX *ctx` after having done s = SSL_new(ctx)
17718 does not influence s as it used to.
17720 In order to clean up things more thoroughly, inside SSL_SESSION
17721 we don't use CERT any longer, but a new structure SESS_CERT
17722 that holds per-session data (if available); currently, this is
17723 the peer's certificate chain and, for clients, the server's certificate
17724 and temporary key. CERT holds only those values that can have
17725 meaningful defaults in an SSL_CTX.
17729 * New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
17730 from the internal representation. Various PKCS#7 fixes: remove some
17731 evil casts and set the enc_dig_alg field properly based on the signing
17736 * Allow PKCS#12 password to be set from the command line or the
17737 environment. Let 'ca' get its config file name from the environment
17738 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
17743 * Allow certificate policies extension to use an IA5STRING for the
17744 organization field. This is contrary to the PKIX definition but
17745 VeriSign uses it and IE5 only recognises this form. Document 'x509'
17750 * Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
17751 without disallowing inline assembler and the like for non-pedantic builds.
17755 * Support Borland C++ builder.
17757 *Janez Jere <jj@void.si>, modified by Ulf Möller*
17763 * SHA-1 cleanups and performance enhancements.
17765 *Andy Polyakov <appro@fy.chalmers.se>*
17767 * Sparc v8plus assembler for the bignum library.
17769 *Andy Polyakov <appro@fy.chalmers.se>*
17771 * Accept any -xxx and +xxx compiler options in Configure.
17775 * Update HPUX configuration.
17779 * Add missing `sk_<type>_unshift()` function to safestack.h
17781 *Ralf S. Engelschall*
17783 * New function SSL_CTX_use_certificate_chain_file that sets the
17784 "extra_cert"s in addition to the certificate. (This makes sense
17785 only for "PEM" format files, as chains as a whole are not
17790 * Support verify_depth from the SSL API.
17791 x509_vfy.c had what can be considered an off-by-one-error:
17792 Its depth (which was not part of the external interface)
17793 was actually counting the number of certificates in a chain;
17794 now it really counts the depth.
17798 * Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
17799 instead of X509err, which often resulted in confusing error
17800 messages since the error codes are not globally unique
17801 (e.g. an alleged error in ssl3_accept when a certificate
17802 didn't match the private key).
17804 * New function SSL_CTX_set_session_id_context that allows to set a default
17805 value (so that you don't need SSL_set_session_id_context for each
17806 connection using the SSL_CTX).
17810 * OAEP decoding bug fix.
17814 * Support INSTALL_PREFIX for package builders, as proposed by
17819 * New Configure options "threads" and "no-threads". For systems
17820 where the proper compiler options are known (currently Solaris
17821 and Linux), "threads" is the default.
17825 * New script util/mklink.pl as a faster substitute for util/mklink.sh.
17829 * Install various scripts to $(OPENSSLDIR)/misc, not to
17830 $(INSTALLTOP)/bin -- they shouldn't clutter directories
17831 such as /usr/local/bin.
17835 * "make linux-shared" to build shared libraries.
17837 *Niels Poppe <niels@netbox.org>*
17839 * New Configure option `no-<cipher>` (rsa, idea, rc5, ...).
17843 * Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
17844 extension adding in x509 utility.
17848 * Remove NOPROTO sections and error code comments.
17852 * Partial rewrite of the DEF file generator to now parse the ANSI
17857 * New Configure options --prefix=DIR and --openssldir=DIR.
17861 * Complete rewrite of the error code script(s). It is all now handled
17862 by one script at the top level which handles error code gathering,
17863 header rewriting and C source file generation. It should be much better
17864 than the old method: it now uses a modified version of Ulf's parser to
17865 read the ANSI prototypes in all header files (thus the old K&R definitions
17866 aren't needed for error creation any more) and do a better job of
17867 translating function codes into names. The old 'ASN1 error code embedded
17868 in a comment' is no longer necessary and it doesn't use .err files which
17869 have now been deleted. Also the error code call doesn't have to appear all
17870 on one line (which resulted in some large lines...).
17874 * Change #include filenames from `<foo.h>` to `<openssl/foo.h>`.
17878 * Change behaviour of ssl2_read when facing length-0 packets: Don't return
17879 0 (which usually indicates a closed connection), but continue reading.
17883 * Fix some race conditions.
17887 * Add support for CRL distribution points extension. Add Certificate
17888 Policies and CRL distribution points documentation.
17892 * Move the autogenerated header file parts to crypto/opensslconf.h.
17896 * Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
17897 8 of keying material. Merlin has also confirmed interop with this fix
17898 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
17900 *Merlin Hughes <merlin@baltimore.ie>*
17902 * Fix lots of warnings.
17904 *Richard Levitte <levitte@stacken.kth.se>*
17906 * In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
17907 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
17909 *Richard Levitte <levitte@stacken.kth.se>*
17911 * Fix problems with sizeof(long) == 8.
17913 *Andy Polyakov <appro@fy.chalmers.se>*
17915 * Change functions to ANSI C.
17919 * Fix typos in error codes.
17921 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller*
17923 * Remove defunct assembler files from Configure.
17927 * SPARC v8 assembler BIGNUM implementation.
17929 *Andy Polyakov <appro@fy.chalmers.se>*
17931 * Support for Certificate Policies extension: both print and set.
17932 Various additions to support the r2i method this uses.
17936 * A lot of constification, and fix a bug in X509_NAME_oneline() that could
17937 return a const string when you are expecting an allocated buffer.
17941 * Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
17942 types DirectoryString and DisplayText.
17946 * Add code to allow r2i extensions to access the configuration database,
17947 add an LHASH database driver and add several ctx helper functions.
17951 * Fix an evil bug in bn_expand2() which caused various BN functions to
17952 fail when they extended the size of a BIGNUM.
17956 * Various utility functions to handle SXNet extension. Modify mkdef.pl to
17957 support typesafe stack.
17961 * Fix typo in SSL_[gs]et_options().
17963 *Nils Frostberg <nils@medcom.se>*
17965 * Delete various functions and files that belonged to the (now obsolete)
17966 old X509V3 handling code.
17970 * New Configure option "rsaref".
17974 * Don't auto-generate pem.h.
17978 * Introduce type-safe ASN.1 SETs.
17982 * Convert various additional casted stacks to type-safe STACK_OF() variants.
17984 *Ben Laurie, Ralf S. Engelschall, Steve Henson*
17986 * Introduce type-safe STACKs. This will almost certainly break lots of code
17987 that links with OpenSSL (well at least cause lots of warnings), but fear
17988 not: the conversion is trivial, and it eliminates loads of evil casts. A
17989 few STACKed things have been converted already. Feel free to convert more.
17990 In the fullness of time, I'll do away with the STACK type altogether.
17994 * Add `openssl ca -revoke <certfile>` facility which revokes a certificate
17995 specified in `<certfile>` by updating the entry in the index.txt file.
17996 This way one no longer has to edit the index.txt file manually for
17997 revoking a certificate. The -revoke option does the gory details now.
17999 *Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall*
18001 * Fix `openssl crl -noout -text` combination where `-noout` killed the
18002 `-text` option at all and this way the `-noout -text` combination was
18003 inconsistent in `openssl crl` with the friends in `openssl x509|rsa|dsa`.
18005 *Ralf S. Engelschall*
18007 * Make sure a corresponding plain text error message exists for the
18008 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
18009 verify callback function determined that a certificate was revoked.
18011 *Ralf S. Engelschall*
18013 * Bugfix: In test/testenc, don't test `openssl <cipher>` for
18014 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
18015 all available ciphers including rc5, which was forgotten until now.
18016 In order to let the testing shell script know which algorithms
18017 are available, a new (up to now undocumented) command
18018 `openssl list-cipher-commands` is used.
18022 * Bugfix: s_client occasionally would sleep in select() when
18023 it should have checked SSL_pending() first.
18027 * New functions DSA_do_sign and DSA_do_verify to provide access to
18028 the raw DSA values prior to ASN.1 encoding.
18032 * Tweaks to Configure
18034 *Niels Poppe <niels@netbox.org>*
18036 * Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
18041 * New variables $(RANLIB) and $(PERL) in the Makefiles.
18045 * New config option to avoid instructions that are illegal on the 80386.
18046 The default code is faster, but requires at least a 486.
18050 * Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
18051 SSL2_SERVER_VERSION (not used at all) macros, which are now the
18052 same as SSL2_VERSION anyway.
18056 * New "-showcerts" option for s_client.
18060 * Still more PKCS#12 integration. Add pkcs12 application to openssl
18061 application. Various cleanups and fixes.
18065 * More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
18066 modify error routines to work internally. Add error codes and PBE init
18067 to library startup routines.
18071 * Further PKCS#12 integration. Added password based encryption, PKCS#8 and
18072 packing functions to asn1 and evp. Changed function names and error
18073 codes along the way.
18077 * PKCS12 integration: and so it begins... First of several patches to
18078 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
18079 objects to objects.h
18083 * Add a new 'indent' option to some X509V3 extension code. Initial ASN1
18084 and display support for Thawte strong extranet extension.
18088 * Add LinuxPPC support.
18090 *Jeff Dubrule <igor@pobox.org>*
18092 * Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
18093 bn_div_words in alpha.s.
18095 *Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie*
18097 * Make sure the RSA OAEP test is skipped under -DRSAref because
18098 OAEP isn't supported when OpenSSL is built with RSAref.
18100 *Ulf Moeller <ulf@fitug.de>*
18102 * Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
18103 so they no longer are missing under -DNOPROTO.
18105 *Soren S. Jorvang <soren@t.dk>*
18107 ### Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
18109 * Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
18110 doesn't work when the session is reused. Coming soon!
18114 * Fix a security hole, that allows sessions to be reused in the wrong
18115 context thus bypassing client cert protection! All software that uses
18116 client certs and session caches in multiple contexts NEEDS PATCHING to
18117 allow session reuse! A fuller solution is in the works.
18119 *Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)*
18121 * Some more source tree cleanups (removed obsolete files
18122 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
18123 permission on "config" script to be executable) and a fix for the INSTALL
18126 *Ulf Moeller <ulf@fitug.de>*
18128 * Remove some legacy and erroneous uses of malloc, free instead of
18131 *Lennart Bang <lob@netstream.se>, with minor changes by Steve*
18133 * Make rsa_oaep_test return non-zero on error.
18135 *Ulf Moeller <ulf@fitug.de>*
18137 * Add support for native Solaris shared libraries. Configure
18138 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
18139 if someone would make that last step automatic.
18141 *Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>*
18143 * ctx_size was not built with the right compiler during "make links". Fixed.
18147 * Change the meaning of 'ALL' in the cipher list. It now means "everything
18148 except NULL ciphers". This means the default cipher list will no longer
18149 enable NULL ciphers. They need to be specifically enabled e.g. with
18150 the string "DEFAULT:eNULL".
18154 * Fix to RSA private encryption routines: if p < q then it would
18155 occasionally produce an invalid result. This will only happen with
18156 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
18160 * Be less restrictive and allow also `perl util/perlpath.pl
18161 /path/to/bin/perl` in addition to `perl util/perlpath.pl /path/to/bin`,
18162 because this way one can also use an interpreter named `perl5` (which is
18163 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
18164 installed as `perl`).
18166 *Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
18168 * Let util/clean-depend.pl work also with older Perl 5.00x versions.
18170 *Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
18172 * Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
18173 advapi32.lib to Win32 build and change the pem test comparison
18174 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
18175 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
18176 and crypto/des/ede_cbcm_enc.c.
18180 * DES quad checksum was broken on big-endian architectures. Fixed.
18184 * Comment out two functions in bio.h that aren't implemented. Fix up the
18185 Win32 test batch file so it (might) work again. The Win32 test batch file
18186 is horrible: I feel ill....
18190 * Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
18191 in e_os.h. Audit of header files to check ANSI and non ANSI
18192 sections: 10 functions were absent from non ANSI section and not exported
18193 from Windows DLLs. Fixed up libeay.num for new functions.
18197 * Make `openssl version` output lines consistent.
18199 *Ralf S. Engelschall*
18201 * Fix Win32 symbol export lists for BIO functions: Added
18202 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
18203 to ms/libeay{16,32}.def.
18205 *Ralf S. Engelschall*
18207 * Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
18208 fine under Unix and passes some trivial tests I've now added. But the
18209 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
18210 added to make sure no one expects that this stuff really works in the
18211 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
18212 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
18215 *Ralf S. Engelschall*
18217 * Fix the generation of two part addresses in perl.
18219 *Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie*
18221 * Add config entry for Linux on MIPS.
18223 *John Tobey <jtobey@channel1.com>*
18225 * Make links whenever Configure is run, unless we are on Windoze.
18229 * Permit extensions to be added to CRLs using crl_section in openssl.cnf.
18230 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
18235 * Add a useful kludge to allow package maintainers to specify compiler and
18236 other platforms details on the command line without having to patch the
18237 Configure script every time: One now can use
18238 `perl Configure <id>:<details>`,
18239 i.e. platform ids are allowed to have details appended
18240 to them (separated by colons). This is treated as there would be a static
18241 pre-configured entry in Configure's %table under key `<id>` with value
18242 `<details>` and `perl Configure <id>` is called. So, when you want to
18243 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
18244 assembler stuff you can use `perl Configure "FreeBSD-elf:pgcc:-O6:::"`
18245 now, which overrides the FreeBSD-elf entry on-the-fly.
18247 *Ralf S. Engelschall*
18249 * Disable new TLS1 ciphersuites by default: they aren't official yet.
18253 * Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
18254 on the `perl Configure ...` command line. This way one can compile
18255 OpenSSL libraries with Position Independent Code (PIC) which is needed
18256 for linking it into DSOs.
18258 *Ralf S. Engelschall*
18260 * Remarkably, export ciphers were totally broken and no-one had noticed!
18265 * Cleaned up the LICENSE document: The official contact for any license
18266 questions now is the OpenSSL core team under openssl-core@openssl.org.
18267 And add a paragraph about the dual-license situation to make sure people
18268 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
18269 to the OpenSSL toolkit.
18271 *Ralf S. Engelschall*
18273 * General source tree makefile cleanups: Made `making xxx in yyy...`
18274 display consistent in the source tree and replaced `/bin/rm` by `rm`.
18275 Additionally cleaned up the `make links` target: Remove unnecessary
18276 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
18277 to speed processing and no longer clutter the display with confusing
18278 stuff. Instead only the actually done links are displayed.
18280 *Ralf S. Engelschall*
18282 * Permit null encryption ciphersuites, used for authentication only. It used
18283 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
18284 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
18289 * Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
18290 signed attributes when verifying signatures (this would break them),
18291 the detached data encoding was wrong and public keys obtained using
18292 X509_get_pubkey() weren't freed.
18296 * Add text documentation for the BUFFER functions. Also added a work around
18297 to a Win95 console bug. This was triggered by the password read stuff: the
18298 last character typed gets carried over to the next fread(). If you were
18299 generating a new cert request using 'req' for example then the last
18300 character of the passphrase would be CR which would then enter the first
18305 * Added the new 'Includes OpenSSL Cryptography Software' button as
18306 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
18307 button and can be used by applications based on OpenSSL to show the
18308 relationship to the OpenSSL project.
18310 *Ralf S. Engelschall*
18312 * Remove confusing variables in function signatures in files
18313 ssl/ssl_lib.c and ssl/ssl.h.
18315 *Lennart Bong <lob@kulthea.stacken.kth.se>*
18317 * Don't install bss_file.c under PREFIX/include/
18319 *Lennart Bong <lob@kulthea.stacken.kth.se>*
18321 * Get the Win32 compile working again. Modify mkdef.pl so it can handle
18322 functions that return function pointers and has support for NT specific
18323 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
18324 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
18325 unsigned to signed types: this was killing the Win32 compile.
18329 * Add new certificate file to stack functions,
18330 SSL_add_dir_cert_subjects_to_stack() and
18331 SSL_add_file_cert_subjects_to_stack(). These largely supplant
18332 SSL_load_client_CA_file(), and can be used to add multiple certs easily
18333 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
18334 This means that Apache-SSL and similar packages don't have to mess around
18335 to add as many CAs as they want to the preferred list.
18339 * Experiment with doxygen documentation. Currently only partially applied to
18341 See <http://www.stack.nl/~dimitri/doxygen/index.html>, and run doxygen with
18342 openssl.doxy as the configuration file.
18346 * Get rid of remaining C++-style comments which strict C compilers hate.
18348 *Ralf S. Engelschall, pointed out by Carlos Amengual*
18350 * Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
18351 compiled in by default: it has problems with large keys.
18355 * Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
18356 DH private keys and/or callback functions which directly correspond to
18357 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
18358 is needed for applications which have to configure certificates on a
18359 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
18361 For the RSA certificate situation is makes no difference, but
18362 for the DSA certificate situation this fixes the "no shared cipher"
18363 problem where the OpenSSL cipher selection procedure failed because the
18364 temporary keys were not overtaken from the context and the API provided
18365 no way to reconfigure them.
18366 The new functions now let applications reconfigure the stuff and they
18367 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
18368 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
18369 non-public-API function ssl_cert_instantiate() is used as a helper
18370 function and also to reduce code redundancy inside ssl_rsa.c.
18372 *Ralf S. Engelschall*
18374 * Move s_server -dcert and -dkey options out of the undocumented feature
18375 area because they are useful for the DSA situation and should be
18376 recognized by the users.
18378 *Ralf S. Engelschall*
18380 * Fix the cipher decision scheme for export ciphers: the export bits are
18381 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
18382 SSL_EXP_MASK. So, the original variable has to be used instead of the
18383 already masked variable.
18385 *Richard Levitte <levitte@stacken.kth.se>*
18387 * Fix `port` variable from `int` to `unsigned int` in crypto/bio/b_sock.c
18389 *Richard Levitte <levitte@stacken.kth.se>*
18391 * Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
18392 from `int` to `unsigned int` because it is a length and initialized by
18393 EVP_DigestFinal() which expects an `unsigned int *`.
18395 *Richard Levitte <levitte@stacken.kth.se>*
18397 * Don't hard-code path to Perl interpreter on shebang line of Configure
18398 script. Instead use the usual Shell->Perl transition trick.
18400 *Ralf S. Engelschall*
18402 * Make `openssl x509 -noout -modulus`' functional also for DSA certificates
18403 (in addition to RSA certificates) to match the behaviour of `openssl dsa
18404 -noout -modulus` as it's already the case for `openssl rsa -noout
18405 -modulus`. For RSA the -modulus is the real "modulus" while for DSA
18406 currently the public key is printed (a decision which was already done by
18407 `openssl dsa -modulus` in the past) which serves a similar purpose.
18408 Additionally the NO_RSA no longer completely removes the whole -modulus
18409 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
18412 *Ralf S. Engelschall*
18414 * Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
18415 BIO. See the source (crypto/evp/bio_ok.c) for more info.
18417 *Arne Ansper <arne@ats.cyber.ee>*
18419 * Dump the old yucky req code that tried (and failed) to allow raw OIDs
18420 to be added. Now both 'req' and 'ca' can use new objects defined in the
18425 * Add cool BIO that does syslog (or event log on NT).
18427 *Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie*
18429 * Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
18430 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
18431 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
18432 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
18436 * Add preliminary config info for new extension code.
18440 * Make RSA_NO_PADDING really use no padding.
18442 *Ulf Moeller <ulf@fitug.de>*
18444 * Generate errors when private/public key check is done.
18448 * Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
18449 for some CRL extensions and new objects added.
18453 * Really fix the ASN1 IMPLICIT bug this time... Partial support for private
18454 key usage extension and fuller support for authority key id.
18458 * Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
18459 padding method for RSA, which is recommended for new applications in PKCS
18460 #1 v2.0 (RFC 2437, October 1998).
18461 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
18462 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
18463 against Bleichbacher's attack on RSA.
18464 *Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
18467 * Updates to the new SSL compression code
18469 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
18471 * Fix so that the version number in the master secret, when passed
18472 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
18473 (because the server will not accept higher), that the version number
18474 is 0x03,0x01, not 0x03,0x00
18476 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
18478 * Run extensive memory leak checks on SSL commands. Fixed *lots* of memory
18479 leaks in `ssl/` relating to new `X509_get_pubkey()` behaviour. Also fixes
18480 in `apps/` and an unrelated leak in `crypto/dsa/dsa_vrf.c`.
18484 * Support for RAW extensions where an arbitrary extension can be
18485 created by including its DER encoding. See `apps/openssl.cnf` for
18490 * Make sure latest Perl versions don't interpret some generated C array
18491 code as Perl array code in the crypto/err/err_genc.pl script.
18493 *Lars Weber <3weber@informatik.uni-hamburg.de>*
18495 * Modify ms/do_ms.bat to not generate assembly language makefiles since
18496 not many people have the assembler. Various Win32 compilation fixes and
18497 update to the INSTALL.W32 file with (hopefully) more accurate Win32
18498 build instructions.
18502 * Modify configure script 'Configure' to automatically create crypto/date.h
18503 file under Win32 and also build pem.h from pem.org. New script
18504 util/mkfiles.pl to create the MINFO file on environments that can't do a
18505 'make files': perl util/mkfiles.pl >MINFO should work.
18509 * Major rework of DES function declarations, in the pursuit of correctness
18510 and purity. As a result, many evil casts evaporated, and some weirdness,
18511 too. You may find this causes warnings in your code. Zapping your evil
18512 casts will probably fix them. Mostly.
18516 * Fix for a typo in asn1.h. Bug fix to object creation script
18517 obj_dat.pl. It considered a zero in an object definition to mean
18518 "end of object": none of the objects in objects.h have any zeros
18519 so it wasn't spotted.
18521 *Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>*
18523 * Add support for Triple DES Cipher Block Chaining with Output Feedback
18524 Masking (CBCM). In the absence of test vectors, the best I have been able
18525 to do is check that the decrypt undoes the encrypt, so far. Send me test
18526 vectors if you have them.
18530 * Correct calculation of key length for export ciphers (too much space was
18531 allocated for null ciphers). This has not been tested!
18535 * Modifications to the mkdef.pl for Win32 DEF file creation. The usage
18536 message is now correct (it understands "crypto" and "ssl" on its
18537 command line). There is also now an "update" option. This will update
18538 the util/ssleay.num and util/libeay.num files with any new functions.
18540 perl util/mkdef.pl crypto ssl update
18541 it will update them.
18545 * Overhauled the Perl interface:
18546 - ported BN stuff to OpenSSL's different BN library
18547 - made the perl/ source tree CVS-aware
18548 - renamed the package from SSLeay to OpenSSL (the files still contain
18549 their history because I've copied them in the repository)
18550 - removed obsolete files (the test scripts will be replaced
18551 by better Test::Harness variants in the future)
18553 *Ralf S. Engelschall*
18555 * First cut for a very conservative source tree cleanup:
18556 1. merge various obsolete readme texts into doc/ssleay.txt
18557 where we collect the old documents and readme texts.
18558 2. remove the first part of files where I'm already sure that we no
18559 longer need them because of three reasons: either they are just temporary
18560 files which were left by Eric or they are preserved original files where
18561 I've verified that the diff is also available in the CVS via "cvs diff
18562 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
18563 the crypto/md/ stuff).
18565 *Ralf S. Engelschall*
18567 * More extension code. Incomplete support for subject and issuer alt
18568 name, issuer and authority key id. Change the i2v function parameters
18569 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
18570 what that's for :-) Fix to ASN1 macro which messed up
18571 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
18575 * Preliminary support for ENUMERATED type. This is largely copied from the
18580 * Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
18582 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
18584 * Make sure `make rehash` target really finds the `openssl` program.
18586 *Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
18588 * Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
18589 like to hear about it if this slows down other processors.
18593 * Add CygWin32 platform information to Configure script.
18595 *Alan Batie <batie@aahz.jf.intel.com>*
18597 * Fixed ms/32all.bat script: `no_asm` -> `no-asm`
18599 *Rainer W. Gerling <gerling@mpg-gv.mpg.de>*
18601 * New program nseq to manipulate netscape certificate sequences
18605 * Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
18610 * Fixes to BN code. Previously the default was to define BN_RECURSION
18611 but the BN code had some problems that would cause failures when
18612 doing certificate verification and some other functions.
18614 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
18616 * Add ASN1 and PEM code to support netscape certificate sequences.
18620 * Add ASN1 and PEM code to support netscape certificate sequences.
18624 * Add several PKIX and private extended key usage OIDs.
18628 * Modify the 'ca' program to handle the new extension code. Modify
18629 openssl.cnf for new extension format, add comments.
18633 * More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
18634 and add a sample to openssl.cnf so req -x509 now adds appropriate
18639 * Continued X509 V3 changes. Add to other makefiles, integrate with the
18640 error code, add initial support to X509_print() and x509 application.
18644 * Takes a deep breath and start adding X509 V3 extension support code. Add
18645 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
18646 stuff is currently isolated and isn't even compiled yet.
18650 * Continuing patches for GeneralizedTime. Fix up certificate and CRL
18651 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
18652 Removed the versions check from X509 routines when loading extensions:
18653 this allows certain broken certificates that don't set the version
18654 properly to be processed.
18658 * Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
18659 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
18660 can still be regenerated with "make depend".
18664 * Spelling mistake in C version of CAST-128.
18666 *Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>*
18668 * Changes to the error generation code. The perl script err-code.pl
18669 now reads in the old error codes and retains the old numbers, only
18670 adding new ones if necessary. It also only changes the .err files if new
18671 codes are added. The makefiles have been modified to only insert errors
18672 when needed (to avoid needlessly modifying header files). This is done
18673 by only inserting errors if the .err file is newer than the auto generated
18674 C file. To rebuild all the error codes from scratch (the old behaviour)
18675 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
18676 or delete all the .err files.
18680 * CAST-128 was incorrectly implemented for short keys. The C version has
18681 been fixed, but is untested. The assembler versions are also fixed, but
18682 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
18683 to regenerate it if needed.
18684 *Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
18685 Hagino <itojun@kame.net>*
18687 * File was opened incorrectly in randfile.c.
18689 *Ulf Möller <ulf@fitug.de>*
18691 * Beginning of support for GeneralizedTime. d2i, i2d, check and print
18692 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
18693 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
18694 al: it's just almost always a UTCTime. Note this patch adds new error
18695 codes so do a "make errors" if there are problems.
18699 * Correct Linux 1 recognition in config.
18701 *Ulf Möller <ulf@fitug.de>*
18703 * Remove pointless MD5 hash when using DSA keys in ca.
18705 *Anonymous <nobody@replay.com>*
18707 * Generate an error if given an empty string as a cert directory. Also
18708 generate an error if handed NULL (previously returned 0 to indicate an
18709 error, but didn't set one).
18711 *Ben Laurie, reported by Anonymous <nobody@replay.com>*
18713 * Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
18717 * Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
18718 parameters. This was causing a warning which killed off the Win32 compile.
18722 * Remove C++ style comments from crypto/bn/bn_local.h.
18724 *Neil Costigan <neil.costigan@celocom.com>*
18726 * The function OBJ_txt2nid was broken. It was supposed to return a nid
18727 based on a text string, looking up short and long names and finally
18728 "dot" format. The "dot" format stuff didn't work. Added new function
18729 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
18730 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
18731 OID is not part of the table.
18735 * Add prototypes to X509 lookup/verify methods, fixing a bug in
18736 X509_LOOKUP_by_alias().
18740 * Sort openssl functions by name.
18744 * Get the `gendsa` command working and add it to the `list` command. Remove
18745 encryption from sample DSA keys (in case anyone is interested the password
18750 * Make *all* `*_free` functions accept a NULL pointer.
18752 *Frans Heymans <fheymans@isaserver.be>*
18754 * If a DH key is generated in s3_srvr.c, don't blow it by trying to use
18757 *Anonymous <nobody@replay.com>*
18759 * s_server should send the CAfile as acceptable CAs, not its own cert.
18761 *Bodo Moeller <3moeller@informatik.uni-hamburg.de>*
18763 * Don't blow it for numeric `-newkey` arguments to `apps/req`.
18765 *Bodo Moeller <3moeller@informatik.uni-hamburg.de>*
18767 * Temp key "for export" tests were wrong in s3_srvr.c.
18769 *Anonymous <nobody@replay.com>*
18771 * Add prototype for temp key callback functions
18772 SSL_CTX_set_tmp_{rsa,dh}_callback().
18776 * Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
18777 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
18781 * X509_name_add_entry() freed the wrong thing after an error.
18783 *Arne Ansper <arne@ats.cyber.ee>*
18785 * rsa_eay.c would attempt to free a NULL context.
18787 *Arne Ansper <arne@ats.cyber.ee>*
18789 * BIO_s_socket() had a broken should_retry() on Windoze.
18791 *Arne Ansper <arne@ats.cyber.ee>*
18793 * BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
18795 *Arne Ansper <arne@ats.cyber.ee>*
18797 * Make sure the already existing X509_STORE->depth variable is initialized
18798 in X509_STORE_new(), but document the fact that this variable is still
18799 unused in the certificate verification process.
18801 *Ralf S. Engelschall*
18803 * Fix the various library and `apps/` files to free up pkeys obtained from
18804 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
18808 * Fix reference counting in X509_PUBKEY_get(). This makes
18809 demos/maurice/example2.c work, amongst others, probably.
18811 *Steve Henson and Ben Laurie*
18813 * First cut of a cleanup for `apps/`. First the `ssleay` program is now named
18814 `openssl` and second, the shortcut symlinks for the `openssl <command>`
18815 are no longer created. This way we have a single and consistent command
18816 line interface `openssl <command>`, similar to `cvs <command>`.
18818 *Ralf S. Engelschall, Paul Sutton and Ben Laurie*
18820 * ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
18821 BIT STRING wrapper always have zero unused bits.
18825 * Add CA.pl, perl version of CA.sh, add extended key usage OID.
18829 * Make the top-level INSTALL documentation easier to understand.
18833 * Makefiles updated to exit if an error occurs in a sub-directory
18834 make (including if user presses ^C) [Paul Sutton]
18836 * Make Montgomery context stuff explicit in RSA data structure.
18840 * Fix build order of pem and err to allow for generated pem.h.
18844 * Fix renumbering bug in X509_NAME_delete_entry().
18848 * Enhanced the err-ins.pl script so it makes the error library number
18849 global and can add a library name. This is needed for external ASN1 and
18850 other error libraries.
18854 * Fixed sk_insert which never worked properly.
18858 * Fix ASN1 macros so they can handle indefinite length constructed
18859 EXPLICIT tags. Some non standard certificates use these: they can now
18864 * Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
18865 into a single doc/ssleay.txt bundle. This way the information is still
18866 preserved but no longer messes up this directory. Now it's new room for
18867 the new set of documentation files.
18869 *Ralf S. Engelschall*
18871 * SETs were incorrectly DER encoded. This was a major pain, because they
18872 shared code with SEQUENCEs, which aren't coded the same. This means that
18873 almost everything to do with SETs or SEQUENCEs has either changed name or
18874 number of arguments.
18876 *Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>*
18878 * Fix test data to work with the above.
18882 * Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
18883 was already fixed by Eric for 0.9.1 it seems.
18885 *Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>*
18887 * Autodetect FreeBSD3.
18891 * Fix various bugs in Configure. This affects the following platforms:
18895 unixware-2.0-pentium
18900 * Eliminate generated files from CVS. Reorder tests to regenerate files
18901 before they are needed.
18905 * Generate Makefile.ssl from Makefile.org (to keep CVS happy).
18909 ### Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
18911 * Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
18912 changed SSLeay to OpenSSL in version strings.
18914 *Ralf S. Engelschall*
18916 * Some fixups to the top-level documents.
18920 * Fixed the nasty bug where rsaref.h was not found under compile-time
18921 because the symlink to include/ was missing.
18923 *Ralf S. Engelschall*
18925 * Incorporated the popular no-RSA/DSA-only patches
18926 which allow to compile a RSA-free SSLeay.
18928 *Andrew Cooke / Interrader Ldt., Ralf S. Engelschall*
18930 * Fixed nasty rehash problem under `make -f Makefile.ssl links`
18931 when "ssleay" is still not found.
18933 *Ralf S. Engelschall*
18935 * Added more platforms to Configure: Cray T3E, HPUX 11,
18937 *Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>*
18939 * Updated the README file.
18941 *Ralf S. Engelschall*
18943 * Added various .cvsignore files in the CVS repository subdirs
18944 to make a "cvs update" really silent.
18946 *Ralf S. Engelschall*
18948 * Recompiled the error-definition header files and added
18949 missing symbols to the Win32 linker tables.
18951 *Ralf S. Engelschall*
18953 * Cleaned up the top-level documents;
18954 o new files: CHANGES and LICENSE
18955 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
18956 o merged COPYRIGHT into LICENSE
18957 o removed obsolete TODO file
18958 o renamed MICROSOFT to INSTALL.W32
18960 *Ralf S. Engelschall*
18962 * Removed dummy files from the 0.9.1b source tree:
18963 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
18964 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
18965 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
18966 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
18967 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
18969 *Ralf S. Engelschall*
18971 * Added various platform portability fixes.
18975 * The Genesis of the OpenSSL rpject:
18976 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
18977 Young and Tim J. Hudson created while they were working for C2Net until
18980 *The OpenSSL Project*
18982 ### Changes between 0.9.0b and 0.9.1b [not released]
18984 * Updated a few CA certificates under certs/
18988 * Changed some BIGNUM api stuff.
18992 * Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
18993 DGUX x86, Linux Alpha, etc.
18997 * New COMP library [crypto/comp/] for SSL Record Layer Compression:
18998 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
19003 * Add -strparse option to asn1pars program which parses nested
19006 *Dr Stephen Henson <shenson@bigfoot.com>*
19008 * Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
19012 * DSA fix for "ca" program.
19016 * Added "-genkey" option to "dsaparam" program.
19020 * Added RIPE MD160 (rmd160) message digest.
19024 * Added -a (all) option to "ssleay version" command.
19028 * Added PLATFORM define which is the id given to Configure.
19032 * Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
19036 * Extended the ASN.1 parser routines.
19040 * Extended BIO routines to support REUSEADDR, seek, tell, etc.
19044 * Added a BN_CTX to the BN library.
19048 * Fixed the weak key values in DES library
19052 * Changed API in EVP library for cipher aliases.
19056 * Added support for RC2/64bit cipher.
19060 * Converted the lhash library to the crypto/mem.c functions.
19064 * Added more recognized ASN.1 object ids.
19068 * Added more RSA padding checks for SSL/TLS.
19072 * Added BIO proxy/filter functionality.
19076 * Added extra_certs to SSL_CTX which can be used
19077 send extra CA certificates to the client in the CA cert chain sending
19078 process. It can be configured with SSL_CTX_add_extra_chain_cert().
19082 * Now Fortezza is denied in the authentication phase because
19083 this is key exchange mechanism is not supported by SSLeay at all.
19087 * Additional PKCS1 checks.
19091 * Support the string "TLSv1" for all TLS v1 ciphers.
19095 * Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
19096 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
19100 * Fixed a few memory leaks.
19104 * Fixed various code and comment typos.
19108 * A minor bug in ssl/s3_clnt.c where there would always be 4 0
19109 bytes sent in the client random.
19111 *Edward Bishop <ebishop@spyglass.com>*
19115 [CVE-2020-1971]: https://www.openssl.org/news/vulnerabilities.html#CVE-2020-1971
19116 [CVE-2020-1967]: https://www.openssl.org/news/vulnerabilities.html#CVE-2020-1967
19117 [CVE-2019-1563]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1563
19118 [CVE-2019-1559]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1559
19119 [CVE-2019-1552]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1552
19120 [CVE-2019-1551]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1551
19121 [CVE-2019-1549]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1549
19122 [CVE-2019-1547]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1547
19123 [CVE-2019-1543]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1543
19124 [CVE-2018-5407]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-5407
19125 [CVE-2018-0739]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0739
19126 [CVE-2018-0737]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0737
19127 [CVE-2018-0735]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0735
19128 [CVE-2018-0734]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0734
19129 [CVE-2018-0733]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0733
19130 [CVE-2018-0732]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0732
19131 [CVE-2017-3738]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3738
19132 [CVE-2017-3737]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3737
19133 [CVE-2017-3736]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3736
19134 [CVE-2017-3735]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3735
19135 [CVE-2017-3733]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3733
19136 [CVE-2017-3732]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3732
19137 [CVE-2017-3731]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3731
19138 [CVE-2017-3730]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3730
19139 [CVE-2016-7055]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7055
19140 [CVE-2016-7054]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7054
19141 [CVE-2016-7053]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7053
19142 [CVE-2016-7052]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7052
19143 [CVE-2016-6309]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6309
19144 [CVE-2016-6308]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6308
19145 [CVE-2016-6307]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6307
19146 [CVE-2016-6306]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6306
19147 [CVE-2016-6305]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6305
19148 [CVE-2016-6304]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6304
19149 [CVE-2016-6303]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6303
19150 [CVE-2016-6302]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6302
19151 [CVE-2016-2183]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2183
19152 [CVE-2016-2182]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2182
19153 [CVE-2016-2181]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2181
19154 [CVE-2016-2180]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2180
19155 [CVE-2016-2179]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2179
19156 [CVE-2016-2178]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2178
19157 [CVE-2016-2177]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2177
19158 [CVE-2016-2176]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2176
19159 [CVE-2016-2109]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2109
19160 [CVE-2016-2107]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2107
19161 [CVE-2016-2106]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2106
19162 [CVE-2016-2105]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2105
19163 [CVE-2016-0800]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0800
19164 [CVE-2016-0799]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0799
19165 [CVE-2016-0798]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0798
19166 [CVE-2016-0797]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0797
19167 [CVE-2016-0705]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0705
19168 [CVE-2016-0702]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0702
19169 [CVE-2016-0701]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0701
19170 [CVE-2015-3197]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3197
19171 [CVE-2015-3196]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3196
19172 [CVE-2015-3195]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3195
19173 [CVE-2015-3194]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3194
19174 [CVE-2015-3193]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3193
19175 [CVE-2015-1793]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1793
19176 [CVE-2015-1792]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1792
19177 [CVE-2015-1791]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1791
19178 [CVE-2015-1790]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1790
19179 [CVE-2015-1789]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1789
19180 [CVE-2015-1788]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1788
19181 [CVE-2015-1787]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1787
19182 [CVE-2015-0293]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0293
19183 [CVE-2015-0291]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0291
19184 [CVE-2015-0290]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0290
19185 [CVE-2015-0289]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0289
19186 [CVE-2015-0288]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0288
19187 [CVE-2015-0287]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0287
19188 [CVE-2015-0286]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0286
19189 [CVE-2015-0285]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0285
19190 [CVE-2015-0209]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0209
19191 [CVE-2015-0208]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0208
19192 [CVE-2015-0207]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0207
19193 [CVE-2015-0206]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0206
19194 [CVE-2015-0205]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0205
19195 [CVE-2015-0204]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0204
19196 [CVE-2014-8275]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-8275
19197 [CVE-2014-5139]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-5139
19198 [CVE-2014-3572]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3572
19199 [CVE-2014-3571]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3571
19200 [CVE-2014-3570]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3570
19201 [CVE-2014-3569]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3569
19202 [CVE-2014-3568]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3568
19203 [CVE-2014-3567]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3567
19204 [CVE-2014-3566]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3566
19205 [CVE-2014-3513]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3513
19206 [CVE-2014-3512]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3512
19207 [CVE-2014-3511]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3511
19208 [CVE-2014-3510]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3510
19209 [CVE-2014-3509]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3509
19210 [CVE-2014-3508]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3508
19211 [CVE-2014-3507]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3507
19212 [CVE-2014-3506]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3506
19213 [CVE-2014-3505]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3505
19214 [CVE-2014-3470]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3470
19215 [CVE-2014-0224]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0224
19216 [CVE-2014-0221]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0221
19217 [CVE-2014-0195]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0195
19218 [CVE-2014-0160]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0160
19219 [CVE-2014-0076]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0076
19220 [CVE-2013-6450]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-6450
19221 [CVE-2013-4353]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-4353
19222 [CVE-2013-0169]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-0169
19223 [CVE-2013-0166]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-0166
19224 [CVE-2012-2686]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2686
19225 [CVE-2012-2333]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2333
19226 [CVE-2012-2110]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2110
19227 [CVE-2012-0884]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0884
19228 [CVE-2012-0050]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0050
19229 [CVE-2012-0027]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0027
19230 [CVE-2011-4619]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4619
19231 [CVE-2011-4577]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4577
19232 [CVE-2011-4576]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4576
19233 [CVE-2011-4109]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4109
19234 [CVE-2011-4108]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4108
19235 [CVE-2011-3210]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-3210
19236 [CVE-2011-3207]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-3207
19237 [CVE-2011-0014]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-0014
19238 [CVE-2010-4252]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-4252
19239 [CVE-2010-4180]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-4180
19240 [CVE-2010-3864]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-3864
19241 [CVE-2010-1633]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-1633
19242 [CVE-2010-0740]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-0740
19243 [CVE-2010-0433]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-0433
19244 [CVE-2009-4355]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-4355
19245 [CVE-2009-3555]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-3555
19246 [CVE-2009-3245]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-3245
19247 [CVE-2009-1386]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1386
19248 [CVE-2009-1379]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1379
19249 [CVE-2009-1378]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1378
19250 [CVE-2009-1377]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1377
19251 [CVE-2009-0789]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0789
19252 [CVE-2009-0591]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0591
19253 [CVE-2009-0590]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0590
19254 [CVE-2008-5077]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-5077
19255 [CVE-2008-1678]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-1678
19256 [CVE-2008-1672]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-1672
19257 [CVE-2008-0891]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-0891
19258 [CVE-2007-5135]: https://www.openssl.org/news/vulnerabilities.html#CVE-2007-5135
19259 [CVE-2007-4995]: https://www.openssl.org/news/vulnerabilities.html#CVE-2007-4995
19260 [CVE-2006-4343]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-4343
19261 [CVE-2006-4339]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-4339
19262 [CVE-2006-3738]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-3738
19263 [CVE-2006-2940]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-2940
19264 [CVE-2006-2937]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-2937
19265 [CVE-2005-2969]: https://www.openssl.org/news/vulnerabilities.html#CVE-2005-2969
19266 [CVE-2004-0112]: https://www.openssl.org/news/vulnerabilities.html#CVE-2004-0112
19267 [CVE-2004-0079]: https://www.openssl.org/news/vulnerabilities.html#CVE-2004-0079
19268 [CVE-2003-0851]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0851
19269 [CVE-2003-0545]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0545
19270 [CVE-2003-0544]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0544
19271 [CVE-2003-0543]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0543
19272 [CVE-2003-0078]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0078
19273 [CVE-2002-0659]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0659
19274 [CVE-2002-0657]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0657
19275 [CVE-2002-0656]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0656
19276 [CVE-2002-0655]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0655