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 support for TCP Fast Open (RFC7413) to macOS, Linux, and FreeBSD where
28 supported and enabled.
32 * Add ciphersuites based on DHE_PSK (RFC 4279) and ECDHE_PSK (RFC 5489)
33 to the list of ciphersuites providing Perfect Forward Secrecy as
34 required by SECLEVEL >= 3.
36 *Dmitry Belyavskiy, Nicola Tuveri*
38 * Add new SSL APIs to aid in efficiently implementing TLS/SSL fingerprinting. The
39 SSL_CTRL_GET_IANA_GROUPS control code, exposed as the SSL_get0_iana_groups()
40 function-like macro, retrieves the list of supported groups sent by the peer,
41 and the function SSL_client_hello_get_extension_order() populates a caller-supplied
42 array with the list of extension types present in the ClientHello, in order of
47 * Fixed PEM_write_bio_PKCS8PrivateKey() and PEM_write_bio_PKCS8PrivateKey_nid()
48 to make it possible to use empty passphrase strings.
52 * RNDR and RNDRRS support in provider functions to provide
53 random number generation for Arm CPUs (aarch64).
57 * s_client and s_server apps now explicitly say when the TLS version
58 does not include the renegotiation mechanism. This avoids confusion
59 between that scenario versus when the TLS version includes secure
60 renegotiation but the peer lacks support for it.
64 * AES-GCM enabled with AVX512 vAES and vPCLMULQDQ.
66 *Tomasz Kantecki, Andrey Matyukov*
68 * The default SSL/TLS security level has been changed from 1 to 2. RSA,
69 DSA and DH keys of 1024 bits and above and less than 2048 bits and ECC keys
70 of 160 bits and above and less than 224 bits were previously accepted by
71 default but are now no longer allowed. By default TLS compression was
72 already disabled in previous OpenSSL versions. At security level 2 it cannot
77 * The SSL_CTX_set_cipher_list family functions now accept ciphers using their
82 * The PVK key derivation function has been moved from b2i_PVK_bio_ex() into
83 the legacy crypto provider as an EVP_KDF. Applications requiring this KDF
84 will need to load the legacy crypto provider.
88 * The various OBJ_* functions have been made thread safe.
92 * CCM8 cipher suites in TLS have been downgraded to security level zero
93 because they use a short authentication tag which lowers their strength.
97 * Subject or issuer names in X.509 objects are now displayed as UTF-8 strings
102 * Parallel dual-prime 1536/2048-bit modular exponentiation for
103 AVX512_IFMA capable processors.
105 *Sergey Kirillov, Andrey Matyukov (Intel Corp)*
110 For OpenSSL 3.0 a [Migration guide][] has been added, so the CHANGES entries
111 listed here are only a brief description.
112 The migration guide contains more detailed information related to new features,
113 breaking changes, and mappings for the large list of deprecated functions.
115 [Migration guide]: https://github.com/openssl/openssl/tree/master/doc/man7/migration_guide.pod
117 ### Changes between 3.0.1 and 3.0.2 [xx XXX xxxx]
119 * Made the AES constant time code for no-asm configurations
120 optional due to the resulting 95% performance degradation.
121 The AES constant time code can be enabled, for no assembly
122 builds, with: ./config no-asm -DOPENSSL_AES_CONST_TIME
126 ### Changes between 3.0.0 and 3.0.1 [14 dec 2021]
128 * Fixed invalid handling of X509_verify_cert() internal errors in libssl
129 Internally libssl in OpenSSL calls X509_verify_cert() on the client side to
130 verify a certificate supplied by a server. That function may return a
131 negative return value to indicate an internal error (for example out of
132 memory). Such a negative return value is mishandled by OpenSSL and will cause
133 an IO function (such as SSL_connect() or SSL_do_handshake()) to not indicate
134 success and a subsequent call to SSL_get_error() to return the value
135 SSL_ERROR_WANT_RETRY_VERIFY. This return value is only supposed to be
136 returned by OpenSSL if the application has previously called
137 SSL_CTX_set_cert_verify_callback(). Since most applications do not do this
138 the SSL_ERROR_WANT_RETRY_VERIFY return value from SSL_get_error() will be
139 totally unexpected and applications may not behave correctly as a result. The
140 exact behaviour will depend on the application but it could result in
141 crashes, infinite loops or other similar incorrect responses.
143 This issue is made more serious in combination with a separate bug in OpenSSL
144 3.0 that will cause X509_verify_cert() to indicate an internal error when
145 processing a certificate chain. This will occur where a certificate does not
146 include the Subject Alternative Name extension but where a Certificate
147 Authority has enforced name constraints. This issue can occur even with valid
153 * Corrected a few file name and file reference bugs in the build,
154 installation and setup scripts, which lead to installation verification
155 failures. Slightly enhanced the installation verification script.
159 * Fixed EVP_PKEY_eq() to make it possible to use it with strictly private
164 * Fixed PVK encoder to properly query for the passphrase.
168 * Multiple fixes in the OSSL_HTTP API functions.
172 * Allow sign extension in OSSL_PARAM_allocate_from_text() for the
173 OSSL_PARAM_INTEGER data type and return error on negative numbers
174 used with the OSSL_PARAM_UNSIGNED_INTEGER data type. Make
175 OSSL_PARAM_BLD_push_BN{,_pad}() return an error on negative numbers.
179 * Allow copying uninitialized digest contexts with EVP_MD_CTX_copy_ex.
183 * Fixed detection of ARMv7 and ARM64 CPU features on FreeBSD.
187 * Multiple threading fixes.
191 * Added NULL digest implementation to keep compatibility with 1.1.1 version.
195 * Allow fetching an operation from the provider that owns an unexportable key
196 as a fallback if that is still allowed by the property query.
200 ### Changes between 1.1.1 and 3.0.0 [7 sep 2021]
202 * TLS_MAX_VERSION, DTLS_MAX_VERSION and DTLS_MIN_VERSION constants are now
207 * The `OPENSSL_s390xcap` environment variable can be used to set bits in the
208 S390X capability vector to zero. This simplifies testing of different code
209 paths on S390X architecture.
213 * Encrypting more than 2^64 TLS records with AES-GCM is disallowed
214 as per FIPS 140-2 IG A.5 "Key/IV Pair Uniqueness Requirements from
215 SP 800-38D". The communication will fail at this point.
219 * The EC_GROUP_clear_free() function is deprecated as there is nothing
220 confidential in EC_GROUP data.
224 * The byte order mark (BOM) character is ignored if encountered at the
225 beginning of a PEM-formatted file.
229 * Added CMS support for the Russian GOST algorithms.
233 * Due to move of the implementation of cryptographic operations
234 to the providers, validation of various operation parameters can
235 be postponed until the actual operation is executed where previously
236 it happened immediately when an operation parameter was set.
238 For example when setting an unsupported curve with
239 EVP_PKEY_CTX_set_ec_paramgen_curve_nid() this function call will not
240 fail but later keygen operations with the EVP_PKEY_CTX will fail.
242 *OpenSSL team members and many third party contributors*
244 * The EVP_get_cipherbyname() function will return NULL for algorithms such as
245 "AES-128-SIV", "AES-128-CBC-CTS" and "CAMELLIA-128-CBC-CTS" which were
246 previously only accessible via low level interfaces. Use EVP_CIPHER_fetch()
247 instead to retrieve these algorithms from a provider.
251 * On build targets where the multilib postfix is set in the build
252 configuration the libdir directory was changing based on whether
253 the lib directory with the multilib postfix exists on the system
254 or not. This unpredictable behavior was removed and eventual
255 multilib postfix is now always added to the default libdir. Use
256 `--libdir=lib` to override the libdir if adding the postfix is
261 * The triple DES key wrap functionality now conforms to RFC 3217 but is
262 no longer interoperable with OpenSSL 1.1.1.
266 * The ERR_GET_FUNC() function was removed. With the loss of meaningful
267 function codes, this function can only cause problems for calling
272 * Add a configurable flag to output date formats as ISO 8601. Does not
273 change the default date format.
277 * Version of MSVC earlier than 1300 could get link warnings, which could
278 be suppressed if the undocumented -DI_CAN_LIVE_WITH_LNK4049 was set.
279 Support for this flag has been removed.
283 * Rework and make DEBUG macros consistent. Remove unused -DCONF_DEBUG,
284 -DBN_CTX_DEBUG, and REF_PRINT. Add a new tracing category and use it for
285 printing reference counts. Rename -DDEBUG_UNUSED to -DUNUSED_RESULT_DEBUG
286 Fix BN_DEBUG_RAND so it compiles and, when set, force DEBUG_RAND to be set
287 also. Rename engine_debug_ref to be ENGINE_REF_PRINT also for consistency.
291 * The signatures of the functions to get and set options on SSL and
292 SSL_CTX objects changed from "unsigned long" to "uint64_t" type.
293 Some source code changes may be required.
297 * The public definitions of conf_method_st and conf_st have been
298 deprecated. They will be made opaque in a future release.
300 *Rich Salz and Tomáš Mráz*
302 * Client-initiated renegotiation is disabled by default. To allow it, use
303 the -client_renegotiation option, the SSL_OP_ALLOW_CLIENT_RENEGOTIATION
304 flag, or the "ClientRenegotiation" config parameter as appropriate.
308 * Add "abspath" and "includedir" pragma's to config files, to prevent,
309 or modify relative pathname inclusion.
313 * OpenSSL includes a cryptographic module that is intended to be FIPS 140-2
314 validated. Please consult the README-FIPS and
315 README-PROVIDERS files, as well as the migration guide.
317 *OpenSSL team members and many third party contributors*
319 * For the key types DH and DHX the allowed settable parameters are now different.
323 * The openssl commands that read keys, certificates, and CRLs now
324 automatically detect the PEM or DER format of the input files.
326 *David von Oheimb, Richard Levitte, and Tomáš Mráz*
328 * Added enhanced PKCS#12 APIs which accept a library context.
332 * The default manual page suffix ($MANSUFFIX) has been changed to "ossl"
336 * Added support for Kernel TLS (KTLS).
338 *Boris Pismenny, John Baldwin and Andrew Gallatin*
340 * Support for RFC 5746 secure renegotiation is now required by default for
341 SSL or TLS connections to succeed.
345 * The signature of the `copy` functional parameter of the
346 EVP_PKEY_meth_set_copy() function has changed so its `src` argument is
347 now `const EVP_PKEY_CTX *` instead of `EVP_PKEY_CTX *`. Similarly
348 the signature of the `pub_decode` functional parameter of the
349 EVP_PKEY_asn1_set_public() function has changed so its `pub` argument is
350 now `const X509_PUBKEY *` instead of `X509_PUBKEY *`.
354 * The error return values from some control calls (ctrl) have changed.
358 * A public key check is now performed during EVP_PKEY_derive_set_peer().
362 * Many functions in the EVP_ namespace that are getters of values from
363 implementations or contexts were renamed to include get or get0 in their
364 names. Old names are provided as macro aliases for compatibility and
369 * The EVP_PKEY_CTRL_PKCS7_ENCRYPT, EVP_PKEY_CTRL_PKCS7_DECRYPT,
370 EVP_PKEY_CTRL_PKCS7_SIGN, EVP_PKEY_CTRL_CMS_ENCRYPT,
371 EVP_PKEY_CTRL_CMS_DECRYPT, and EVP_PKEY_CTRL_CMS_SIGN control operations
376 * The EVP_PKEY_public_check() and EVP_PKEY_param_check() functions now work for
379 * The output from the command line applications may have minor
384 * The output from numerous "printing" may have minor changes.
388 * Windows thread synchronization uses read/write primitives (SRWLock) when
389 supported by the OS, otherwise CriticalSection continues to be used.
393 * Add filter BIO BIO_f_readbuffer() that allows BIO_tell() and BIO_seek() to
394 work on read only BIO source/sinks that do not support these functions.
395 This allows piping or redirection of a file BIO using stdin to be buffered
396 into memory. This is used internally in OSSL_DECODER_from_bio().
400 * OSSL_STORE_INFO_get_type() may now return an additional value. In 1.1.1
401 this function would return one of the values OSSL_STORE_INFO_NAME,
402 OSSL_STORE_INFO_PKEY, OSSL_STORE_INFO_PARAMS, OSSL_STORE_INFO_CERT or
403 OSSL_STORE_INFO_CRL. Decoded public keys would previously have been reported
404 as type OSSL_STORE_INFO_PKEY in 1.1.1. In 3.0 decoded public keys are now
405 reported as having the new type OSSL_STORE_INFO_PUBKEY. Applications
406 using this function should be amended to handle the changed return value.
410 * Improved adherence to Enhanced Security Services (ESS, RFC 2634 and RFC 5035)
411 for the TSP and CMS Advanced Electronic Signatures (CAdES) implementations.
412 As required by RFC 5035 check both ESSCertID and ESSCertIDv2 if both present.
413 Correct the semantics of checking the validation chain in case ESSCertID{,v2}
414 contains more than one certificate identifier: This means that all
415 certificates referenced there MUST be part of the validation chain.
419 * The implementation of older EVP ciphers related to CAST, IDEA, SEED, RC2, RC4,
420 RC5, DESX and DES have been moved to the legacy provider.
424 * The implementation of the EVP digests MD2, MD4, MDC2, WHIRLPOOL and
425 RIPEMD-160 have been moved to the legacy provider.
429 * The deprecated function EVP_PKEY_get0() now returns NULL being called for a
434 * The deprecated functions EVP_PKEY_get0_RSA(),
435 EVP_PKEY_get0_DSA(), EVP_PKEY_get0_EC_KEY(), EVP_PKEY_get0_DH(),
436 EVP_PKEY_get0_hmac(), EVP_PKEY_get0_poly1305() and EVP_PKEY_get0_siphash() as
437 well as the similarly named "get1" functions behave differently in
442 * A number of functions handling low-level keys or engines were deprecated
443 including EVP_PKEY_set1_engine(), EVP_PKEY_get0_engine(), EVP_PKEY_assign(),
444 EVP_PKEY_get0(), EVP_PKEY_get0_hmac(), EVP_PKEY_get0_poly1305() and
445 EVP_PKEY_get0_siphash().
449 * PKCS#5 PBKDF1 key derivation has been moved from PKCS5_PBE_keyivgen() into
450 the legacy crypto provider as an EVP_KDF. Applications requiring this KDF
451 will need to load the legacy crypto provider. This includes these PBE
452 algorithms which use this KDF:
453 - NID_pbeWithMD2AndDES_CBC
454 - NID_pbeWithMD5AndDES_CBC
455 - NID_pbeWithSHA1AndRC2_CBC
456 - NID_pbeWithMD2AndRC2_CBC
457 - NID_pbeWithMD5AndRC2_CBC
458 - NID_pbeWithSHA1AndDES_CBC
462 * Deprecated obsolete BIO_set_callback(), BIO_get_callback(), and
463 BIO_debug_callback() functions.
467 * Deprecated obsolete EVP_PKEY_CTX_get0_dh_kdf_ukm() and
468 EVP_PKEY_CTX_get0_ecdh_kdf_ukm() functions.
472 * The RAND_METHOD APIs have been deprecated.
476 * The SRP APIs have been deprecated.
480 * Add a compile time option to prevent the caching of provider fetched
481 algorithms. This is enabled by including the no-cached-fetch option
482 at configuration time.
486 * pkcs12 now uses defaults of PBKDF2, AES and SHA-256, with a MAC iteration
487 count of PKCS12_DEFAULT_ITER.
489 *Tomáš Mráz and Sahana Prasad*
491 * The openssl speed command does not use low-level API calls anymore.
495 * Parallel dual-prime 1024-bit modular exponentiation for AVX512_IFMA
498 *Ilya Albrekht, Sergey Kirillov, Andrey Matyukov (Intel Corp)*
500 * Combining the Configure options no-ec and no-dh no longer disables TLSv1.3.
504 * Implemented support for fully "pluggable" TLSv1.3 groups. This means that
505 providers may supply their own group implementations (using either the "key
506 exchange" or the "key encapsulation" methods) which will automatically be
507 detected and used by libssl.
509 *Matt Caswell, Nicola Tuveri*
511 * The undocumented function X509_certificate_type() has been deprecated;
515 * Deprecated the obsolete BN_pseudo_rand() and BN_pseudo_rand_range().
519 * Removed RSA padding mode for SSLv23 (which was only used for
520 SSLv2). This includes the functions RSA_padding_check_SSLv23() and
521 RSA_padding_add_SSLv23() and the `-ssl` option in the deprecated
526 * Deprecated the obsolete X9.31 RSA key generation related functions.
528 * While a callback function set via `SSL_CTX_set_cert_verify_callback()`
529 is not allowed to return a value > 1, this is no more taken as failure.
531 *Viktor Dukhovni and David von Oheimb*
533 * Deprecated the obsolete X9.31 RSA key generation related functions
534 BN_X931_generate_Xpq(), BN_X931_derive_prime_ex(), and
535 BN_X931_generate_prime_ex().
539 * The default key generation method for the regular 2-prime RSA keys was
540 changed to the FIPS 186-4 B.3.6 method.
544 * Deprecated the BN_is_prime_ex() and BN_is_prime_fasttest_ex() functions.
548 * Deprecated EVP_MD_CTX_set_update_fn() and EVP_MD_CTX_update_fn().
552 * Deprecated the type OCSP_REQ_CTX and the functions OCSP_REQ_CTX_*() and
553 replaced with OSSL_HTTP_REQ_CTX and the functions OSSL_HTTP_REQ_CTX_*().
555 *Rich Salz, Richard Levitte, and David von Oheimb*
557 * Deprecated `X509_http_nbio()` and `X509_CRL_http_nbio()`.
561 * Deprecated `OCSP_parse_url()`.
565 * Validation of SM2 keys has been separated from the validation of regular EC
570 * Behavior of the `pkey` app is changed, when using the `-check` or `-pubcheck`
571 switches: a validation failure triggers an early exit, returning a failure
572 exit status to the parent process.
576 * Changed behavior of SSL_CTX_set_ciphersuites() and SSL_set_ciphersuites()
577 to ignore unknown ciphers.
581 * The `-cipher-commands` and `-digest-commands` options
582 of the command line utility `list` have been deprecated.
583 Instead use the `-cipher-algorithms` and `-digest-algorithms` options.
587 * Added convenience functions for generating asymmetric key pairs:
588 The 'quick' one-shot (yet somewhat limited) function L<EVP_PKEY_Q_keygen(3)>
589 and macros for the most common cases: <EVP_RSA_gen(3)> and L<EVP_EC_gen(3)>.
593 * All of the low level EC_KEY functions have been deprecated.
595 *Shane Lontis, Paul Dale, Richard Levitte, and Tomáš Mráz*
597 * Deprecated all the libcrypto and libssl error string loading
602 * The functions SSL_CTX_set_tmp_dh_callback and SSL_set_tmp_dh_callback, as
603 well as the macros SSL_CTX_set_tmp_dh() and SSL_set_tmp_dh() have been
608 * The `-crypt` option to the `passwd` command line tool has been removed.
612 * The -C option to the `x509`, `dhparam`, `dsaparam`, and `ecparam` commands
617 * Add support for AES Key Wrap inverse ciphers to the EVP layer.
621 * Deprecated EVP_PKEY_set1_tls_encodedpoint() and
622 EVP_PKEY_get1_tls_encodedpoint().
626 * The security callback, which can be customised by application code, supports
627 the security operation SSL_SECOP_TMP_DH. One location of the "other" parameter
628 was incorrectly passing a DH object. It now passed an EVP_PKEY in all cases.
632 * Add PKCS7_get_octet_string() and PKCS7_type_is_other() to the public
633 interface. Their functionality remains unchanged.
637 * Added new option for 'openssl list', '-providers', which will display the
638 list of loaded providers, their names, version and status. It optionally
639 displays their gettable parameters.
643 * Removed EVP_PKEY_set_alias_type().
647 * Deprecated `EVP_PKEY_CTX_set_rsa_keygen_pubexp()` and introduced
648 `EVP_PKEY_CTX_set1_rsa_keygen_pubexp()`, which is now preferred.
652 * Changed all "STACK" functions to be macros instead of inline functions. Macro
653 parameters are still checked for type safety at compile time via helper
658 * Remove the RAND_DRBG API
660 *Paul Dale and Matthias St. Pierre*
662 * Allow `SSL_set1_host()` and `SSL_add1_host()` to take IP literal addresses
663 as well as actual hostnames.
667 * The 'MinProtocol' and 'MaxProtocol' configuration commands now silently
668 ignore TLS protocol version bounds when configuring DTLS-based contexts, and
669 conversely, silently ignore DTLS protocol version bounds when configuring
670 TLS-based contexts. The commands can be repeated to set bounds of both
671 types. The same applies with the corresponding "min_protocol" and
672 "max_protocol" command-line switches, in case some application uses both TLS
675 SSL_CTX instances that are created for a fixed protocol version (e.g.
676 `TLSv1_server_method()`) also silently ignore version bounds. Previously
677 attempts to apply bounds to these protocol versions would result in an
678 error. Now only the "version-flexible" SSL_CTX instances are subject to
679 limits in configuration files in command-line options.
683 * Deprecated the `ENGINE` API. Engines should be replaced with providers
688 * Reworked the recorded ERR codes to make better space for system errors.
689 To distinguish them, the macro `ERR_SYSTEM_ERROR()` indicates if the
690 given code is a system error (true) or an OpenSSL error (false).
694 * Reworked the test perl framework to better allow parallel testing.
696 *Nicola Tuveri and David von Oheimb*
698 * Added ciphertext stealing algorithms AES-128-CBC-CTS, AES-192-CBC-CTS and
699 AES-256-CBC-CTS to the providers. CS1, CS2 and CS3 variants are supported.
703 * 'Configure' has been changed to figure out the configuration target if
704 none is given on the command line. Consequently, the 'config' script is
705 now only a mere wrapper. All documentation is changed to only mention
708 *Rich Salz and Richard Levitte*
710 * Added a library context `OSSL_LIB_CTX` that applications as well as
711 other libraries can use to form a separate context within which
712 libcrypto operations are performed.
716 * Added various `_ex` functions to the OpenSSL API that support using
717 a non-default `OSSL_LIB_CTX`.
721 * Handshake now fails if Extended Master Secret extension is dropped
726 * Dropped interactive mode from the `openssl` program.
730 * Deprecated `EVP_PKEY_cmp()` and `EVP_PKEY_cmp_parameters()`.
732 *David von Oheimb and Shane Lontis*
734 * Deprecated `EC_METHOD_get_field_type()`.
738 * Deprecated EC_GFp_simple_method(), EC_GFp_mont_method(),
739 EC_GF2m_simple_method(), EC_GFp_nist_method(), EC_GFp_nistp224_method()
740 EC_GFp_nistp256_method(), and EC_GFp_nistp521_method().
744 * Deprecated EC_GROUP_new(), EC_GROUP_method_of(), and EC_POINT_method_of().
748 * Add CAdES-BES signature verification support, mostly derived
749 from ESSCertIDv2 TS (RFC 5816) contribution by Marek Klein.
751 *Filipe Raimundo da Silva*
753 * Add CAdES-BES signature scheme and attributes support (RFC 5126) to CMS API.
757 * Added the AuthEnvelopedData content type structure (RFC 5083) with AES-GCM
758 parameter (RFC 5084) for the Cryptographic Message Syntax (CMS).
762 * Deprecated EC_POINT_make_affine() and EC_POINTs_make_affine().
766 * Deprecated EC_GROUP_precompute_mult(), EC_GROUP_have_precompute_mult(), and
767 EC_KEY_precompute_mult().
771 * Deprecated EC_POINTs_mul().
775 * Removed FIPS_mode() and FIPS_mode_set().
779 * The SSL option SSL_OP_IGNORE_UNEXPECTED_EOF is introduced.
783 * Deprecated EC_POINT_set_Jprojective_coordinates_GFp() and
784 EC_POINT_get_Jprojective_coordinates_GFp().
788 * Added OSSL_PARAM_BLD to the public interface. This allows OSSL_PARAM
789 arrays to be more easily constructed via a series of utility functions.
790 Create a parameter builder using OSSL_PARAM_BLD_new(), add parameters using
791 the various push functions and finally convert to a passable OSSL_PARAM
792 array using OSSL_PARAM_BLD_to_param().
796 * The security strength of SHA1 and MD5 based signatures in TLS has been
801 * Added EVP_PKEY_set_type_by_keymgmt(), to initialise an EVP_PKEY to
802 contain a provider side internal key.
806 * ASN1_verify(), ASN1_digest() and ASN1_sign() have been deprecated.
810 * Project text documents not yet having a proper file name extension
811 (`HACKING`, `LICENSE`, `NOTES*`, `README*`, `VERSION`) have been renamed to
812 `*.md` as far as reasonable, else `*.txt`, for better use with file managers.
816 * The main project documents (README, NEWS, CHANGES, INSTALL, SUPPORT)
817 have been converted to Markdown with the goal to produce documents
818 which not only look pretty when viewed online in the browser, but
819 remain well readable inside a plain text editor.
821 To achieve this goal, a 'minimalistic' Markdown style has been applied
822 which avoids formatting elements that interfere too much with the
823 reading flow in the text file. For example, it
825 * avoids [ATX headings][] and uses [setext headings][] instead
826 (which works for `<h1>` and `<h2>` headings only).
827 * avoids [inline links][] and uses [reference links][] instead.
828 * avoids [fenced code blocks][] and uses [indented code blocks][] instead.
830 [ATX headings]: https://github.github.com/gfm/#atx-headings
831 [setext headings]: https://github.github.com/gfm/#setext-headings
832 [inline links]: https://github.github.com/gfm/#inline-link
833 [reference links]: https://github.github.com/gfm/#reference-link
834 [fenced code blocks]: https://github.github.com/gfm/#fenced-code-blocks
835 [indented code blocks]: https://github.github.com/gfm/#indented-code-blocks
837 *Matthias St. Pierre*
839 * The test suite is changed to preserve results of each test recipe.
840 A new directory test-runs/ with subdirectories named like the
841 test recipes are created in the build tree for this purpose.
845 * Added an implementation of CMP and CRMF (RFC 4210, RFC 4211 RFC 6712).
846 This adds `crypto/cmp/`, `crpyto/crmf/`, `apps/cmp.c`, and `test/cmp_*`.
847 See L<openssl-cmp(1)> and L<OSSL_CMP_exec_IR_ses(3)> as starting points.
849 *David von Oheimb, Martin Peylo*
851 * Generalized the HTTP client code from `crypto/ocsp/` into `crpyto/http/`.
852 It supports arbitrary request and response content types, GET redirection,
853 TLS, connections via HTTP(S) proxies, connections and exchange via
854 user-defined BIOs (allowing implicit connections), persistent connections,
855 and timeout checks. See L<OSSL_HTTP_transfer(3)> etc. for details.
856 The legacy OCSP-focused (and only partly documented) API
857 is retained for backward compatibility, while most of it is deprecated.
861 * Added `util/check-format.pl`, a tool for checking adherence to the
862 OpenSSL coding style <https://www.openssl.org/policies/codingstyle.html>.
863 The checks performed are incomplete and yield some false positives.
864 Still the tool should be useful for detecting most typical glitches.
868 * `BIO_do_connect()` and `BIO_do_handshake()` have been extended:
869 If domain name resolution yields multiple IP addresses all of them are tried
870 after `connect()` failures.
874 * All of the low level RSA functions have been deprecated.
878 * X509 certificates signed using SHA1 are no longer allowed at security
883 * The command line utilities dhparam, dsa, gendsa and dsaparam have been
884 modified to use PKEY APIs. These commands are now in maintenance mode
885 and no new features will be added to them.
889 * The command line utility rsautl has been deprecated.
893 * The command line utilities genrsa and rsa have been modified to use PKEY
894 APIs. They now write PKCS#8 keys by default. These commands are now in
895 maintenance mode and no new features will be added to them.
899 * All of the low level DH functions have been deprecated.
901 *Paul Dale and Matt Caswell*
903 * All of the low level DSA functions have been deprecated.
907 * Reworked the treatment of EC EVP_PKEYs with the SM2 curve to
908 automatically become EVP_PKEY_SM2 rather than EVP_PKEY_EC.
912 * Deprecated low level ECDH and ECDSA functions.
916 * Deprecated EVP_PKEY_decrypt_old() and EVP_PKEY_encrypt_old().
920 * Enhanced the documentation of EVP_PKEY_get_size(), EVP_PKEY_get_bits()
921 and EVP_PKEY_get_security_bits(). Especially EVP_PKEY_get_size() needed
922 a new formulation to include all the things it can be used for,
923 as well as words of caution.
927 * The SSL_CTX_set_tlsext_ticket_key_cb(3) function has been deprecated.
931 * All of the low level HMAC functions have been deprecated.
933 *Paul Dale and David von Oheimb*
935 * Over two thousand fixes were made to the documentation, including:
936 - Common options (such as -rand/-writerand, TLS version control, etc)
937 were refactored and point to newly-enhanced descriptions in openssl.pod.
938 - Added style conformance for all options (with help from Richard Levitte),
939 documented all reported missing options, added a CI build to check
940 that all options are documented and that no unimplemented options
942 - Documented some internals, such as all use of environment variables.
943 - Addressed all internal broken L<> references.
947 * All of the low level CMAC functions have been deprecated.
951 * The low-level MD2, MD4, MD5, MDC2, RIPEMD160 and Whirlpool digest
952 functions have been deprecated.
954 *Paul Dale and David von Oheimb*
956 * Corrected the documentation of the return values from the `EVP_DigestSign*`
957 set of functions. The documentation mentioned negative values for some
958 errors, but this was never the case, so the mention of negative values
961 Code that followed the documentation and thereby check with something
962 like `EVP_DigestSignInit(...) <= 0` will continue to work undisturbed.
966 * All of the low level cipher functions have been deprecated.
968 *Matt Caswell and Paul Dale*
970 * Removed include/openssl/opensslconf.h.in and replaced it with
971 include/openssl/configuration.h.in, which differs in not including
972 <openssl/macros.h>. A short header include/openssl/opensslconf.h
973 was added to include both.
975 This allows internal hacks where one might need to modify the set
976 of configured macros, for example this if deprecated symbols are
977 still supposed to be available internally:
979 #include <openssl/configuration.h>
981 #undef OPENSSL_NO_DEPRECATED
982 #define OPENSSL_SUPPRESS_DEPRECATED
984 #include <openssl/macros.h>
986 This should not be used by applications that use the exported
987 symbols, as that will lead to linking errors.
991 * Fixed an overflow bug in the x64_64 Montgomery squaring procedure
992 used in exponentiation with 512-bit moduli. No EC algorithms are
993 affected. Analysis suggests that attacks against 2-prime RSA1024,
994 3-prime RSA1536, and DSA1024 as a result of this defect would be very
995 difficult to perform and are not believed likely. Attacks against DH512
996 are considered just feasible. However, for an attack the target would
997 have to re-use the DH512 private key, which is not recommended anyway.
998 Also applications directly using the low-level API BN_mod_exp may be
999 affected if they use BN_FLG_CONSTTIME.
1004 * Most memory-debug features have been deprecated, and the functionality
1005 replaced with no-ops.
1009 * Added documentation for the STACK API.
1013 * Introduced a new method type and API, OSSL_ENCODER, to represent
1014 generic encoders. These do the same sort of job that PEM writers
1015 and d2i functions do, but with support for methods supplied by
1016 providers, and the possibility for providers to support other
1021 * Introduced a new method type and API, OSSL_DECODER, to represent
1022 generic decoders. These do the same sort of job that PEM readers
1023 and i2d functions do, but with support for methods supplied by
1024 providers, and the possibility for providers to support other
1029 * Added a .pragma directive to the syntax of configuration files, to
1030 allow varying behavior in a supported and predictable manner.
1031 Currently added pragma:
1035 This allows dollar signs to be a keyword character unless it's
1036 followed by a opening brace or parenthesis. This is useful for
1037 platforms where dollar signs are commonly used in names, such as
1038 volume names and system directory names on VMS.
1042 * Added functionality to create an EVP_PKEY from user data.
1046 * Change the interpretation of the '--api' configuration option to
1047 mean that this is a desired API compatibility level with no
1048 further meaning. The previous interpretation, that this would
1049 also mean to remove all deprecated symbols up to and including
1050 the given version, no requires that 'no-deprecated' is also used
1051 in the configuration.
1053 When building applications, the desired API compatibility level
1054 can be set with the OPENSSL_API_COMPAT macro like before. For
1055 API compatibility version below 3.0, the old style numerical
1056 value is valid as before, such as -DOPENSSL_API_COMPAT=0x10100000L.
1057 For version 3.0 and on, the value is expected to be the decimal
1058 value calculated from the major and minor version like this:
1060 MAJOR * 10000 + MINOR * 100
1064 -DOPENSSL_API_COMPAT=30000 For 3.0
1065 -DOPENSSL_API_COMPAT=30200 For 3.2
1067 To hide declarations that are deprecated up to and including the
1068 given API compatibility level, -DOPENSSL_NO_DEPRECATED must be
1069 given when building the application as well.
1073 * Added the X509_LOOKUP_METHOD called X509_LOOKUP_store, to allow
1074 access to certificate and CRL stores via URIs and OSSL_STORE
1077 This adds the following functions:
1079 - X509_LOOKUP_store()
1080 - X509_STORE_load_file()
1081 - X509_STORE_load_path()
1082 - X509_STORE_load_store()
1083 - SSL_add_store_cert_subjects_to_stack()
1084 - SSL_CTX_set_default_verify_store()
1085 - SSL_CTX_load_verify_file()
1086 - SSL_CTX_load_verify_dir()
1087 - SSL_CTX_load_verify_store()
1091 * Added a new method to gather entropy on VMS, based on SYS$GET_ENTROPY.
1092 The presence of this system service is determined at run-time.
1096 * Added functionality to create an EVP_PKEY context based on data
1097 for methods from providers. This takes an algorithm name and a
1098 property query string and simply stores them, with the intent
1099 that any operation that uses this context will use those strings
1100 to fetch the needed methods implicitly, thereby making the port
1101 of application written for pre-3.0 OpenSSL easier.
1105 * The undocumented function NCONF_WIN32() has been deprecated; for
1106 conversion details see the HISTORY section of doc/man5/config.pod
1110 * Introduced the new functions EVP_DigestSignInit_ex() and
1111 EVP_DigestVerifyInit_ex(). The macros EVP_DigestSignUpdate() and
1112 EVP_DigestVerifyUpdate() have been converted to functions. See the man
1113 pages for further details.
1117 * Over two thousand fixes were made to the documentation, including:
1118 adding missing command flags, better style conformance, documentation
1121 *Rich Salz, Richard Levitte*
1123 * s390x assembly pack: add hardware-support for P-256, P-384, P-521,
1124 X25519, X448, Ed25519 and Ed448.
1128 * Print all values for a PKCS#12 attribute with 'openssl pkcs12', not just
1133 * Deprecated the public definition of `ERR_STATE` as well as the function
1134 `ERR_get_state()`. This is done in preparation of making `ERR_STATE` an
1139 * Added ERR functionality to give callers access to the stored function
1140 names that have replaced the older function code based functions.
1142 New functions are ERR_peek_error_func(), ERR_peek_last_error_func(),
1143 ERR_peek_error_data(), ERR_peek_last_error_data(), ERR_get_error_all(),
1144 ERR_peek_error_all() and ERR_peek_last_error_all().
1146 Deprecate ERR functions ERR_get_error_line(), ERR_get_error_line_data(),
1147 ERR_peek_error_line_data(), ERR_peek_last_error_line_data() and
1148 ERR_func_error_string().
1152 * Extended testing to be verbose for failing tests only. The make variables
1153 VERBOSE_FAILURE or VF can be used to enable this:
1155 $ make VF=1 test # Unix
1156 $ mms /macro=(VF=1) test ! OpenVMS
1157 $ nmake VF=1 test # Windows
1161 * Added the `-copy_extensions` option to the `x509` command for use with
1162 `-req` and `-x509toreq`. When given with the `copy` or `copyall` argument,
1163 all extensions in the request are copied to the certificate or vice versa.
1165 *David von Oheimb*, *Kirill Stefanenkov <kirill_stefanenkov@rambler.ru>*
1167 * Added the `-copy_extensions` option to the `req` command for use with
1168 `-x509`. When given with the `copy` or `copyall` argument,
1169 all extensions in the certification request are copied to the certificate.
1173 * The `x509`, `req`, and `ca` commands now make sure that X.509v3 certificates
1174 they generate are by default RFC 5280 compliant in the following sense:
1175 There is a subjectKeyIdentifier extension with a hash value of the public key
1176 and for not self-signed certs there is an authorityKeyIdentifier extension
1177 with a keyIdentifier field or issuer information identifying the signing key.
1178 This is done unless some configuration overrides the new default behavior,
1179 such as `subjectKeyIdentifier = none` and `authorityKeyIdentifier = none`.
1183 * Added several checks to `X509_verify_cert()` according to requirements in
1184 RFC 5280 in case `X509_V_FLAG_X509_STRICT` is set
1185 (which may be done by using the CLI option `-x509_strict`):
1186 * The basicConstraints of CA certificates must be marked critical.
1187 * CA certificates must explicitly include the keyUsage extension.
1188 * If a pathlenConstraint is given the key usage keyCertSign must be allowed.
1189 * The issuer name of any certificate must not be empty.
1190 * The subject name of CA certs, certs with keyUsage crlSign,
1191 and certs without subjectAlternativeName must not be empty.
1192 * If a subjectAlternativeName extension is given it must not be empty.
1193 * The signatureAlgorithm field and the cert signature must be consistent.
1194 * Any given authorityKeyIdentifier and any given subjectKeyIdentifier
1195 must not be marked critical.
1196 * The authorityKeyIdentifier must be given for X.509v3 certs
1197 unless they are self-signed.
1198 * The subjectKeyIdentifier must be given for all X.509v3 CA certs.
1202 * Certificate verification using `X509_verify_cert()` meanwhile rejects EC keys
1203 with explicit curve parameters (specifiedCurve) as required by RFC 5480.
1207 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
1208 used even when parsing explicit parameters, when loading a encoded key
1209 or calling `EC_GROUP_new_from_ecpkparameters()`/
1210 `EC_GROUP_new_from_ecparameters()`.
1211 This prevents bypass of security hardening and performance gains,
1212 especially for curves with specialized EC_METHODs.
1213 By default, if a key encoded with explicit parameters is loaded and later
1214 encoded, the output is still encoded with explicit parameters, even if
1215 internally a "named" EC_GROUP is used for computation.
1219 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
1220 this change, EC_GROUP_set_generator would accept order and/or cofactor as
1221 NULL. After this change, only the cofactor parameter can be NULL. It also
1222 does some minimal sanity checks on the passed order.
1227 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
1228 An attack is simple, if the first CMS_recipientInfo is valid but the
1229 second CMS_recipientInfo is chosen ciphertext. If the second
1230 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
1231 encryption key will be replaced by garbage, and the message cannot be
1232 decoded, but if the RSA decryption fails, the correct encryption key is
1233 used and the recipient will not notice the attack.
1234 As a work around for this potential attack the length of the decrypted
1235 key must be equal to the cipher default key length, in case the
1236 certifiate is not given and all recipientInfo are tried out.
1237 The old behaviour can be re-enabled in the CMS code by setting the
1238 CMS_DEBUG_DECRYPT flag.
1242 * Early start up entropy quality from the DEVRANDOM seed source has been
1243 improved for older Linux systems. The RAND subsystem will wait for
1244 /dev/random to be producing output before seeding from /dev/urandom.
1245 The seeded state is stored for future library initialisations using
1246 a system global shared memory segment. The shared memory identifier
1247 can be configured by defining OPENSSL_RAND_SEED_DEVRANDOM_SHM_ID to
1248 the desired value. The default identifier is 114.
1252 * Revised BN_generate_prime_ex to not avoid factors 2..17863 in p-1
1253 when primes for RSA keys are computed.
1254 Since we previously always generated primes == 2 (mod 3) for RSA keys,
1255 the 2-prime and 3-prime RSA modules were easy to distinguish, since
1256 `N = p*q = 1 (mod 3)`, but `N = p*q*r = 2 (mod 3)`. Therefore fingerprinting
1257 2-prime vs. 3-prime RSA keys was possible by computing N mod 3.
1258 This avoids possible fingerprinting of newly generated RSA modules.
1262 * Correct the extended master secret constant on EBCDIC systems. Without this
1263 fix TLS connections between an EBCDIC system and a non-EBCDIC system that
1264 negotiate EMS will fail. Unfortunately this also means that TLS connections
1265 between EBCDIC systems with this fix, and EBCDIC systems without this
1266 fix will fail if they negotiate EMS.
1270 * Changed the library initialisation so that the config file is now loaded
1271 by default. This was already the case for libssl. It now occurs for both
1272 libcrypto and libssl. Use the OPENSSL_INIT_NO_LOAD_CONFIG option to
1273 `OPENSSL_init_crypto()` to suppress automatic loading of a config file.
1277 * Introduced new error raising macros, `ERR_raise()` and `ERR_raise_data()`,
1278 where the former acts as a replacement for `ERR_put_error()`, and the
1279 latter replaces the combination `ERR_put_error()` + `ERR_add_error_data()`.
1280 `ERR_raise_data()` adds more flexibility by taking a format string and
1281 an arbitrary number of arguments following it, to be processed with
1286 * Introduced a new function, `OSSL_PROVIDER_available()`, which can be used
1287 to check if a named provider is loaded and available. When called, it
1288 will also activate all fallback providers if such are still present.
1292 * Enforce a minimum DH modulus size of 512 bits.
1296 * Changed DH parameters to generate the order q subgroup instead of 2q.
1297 Previously generated DH parameters are still accepted by DH_check
1298 but DH_generate_key works around that by clearing bit 0 of the
1299 private key for those. This avoids leaking bit 0 of the private key.
1303 * Significantly reduce secure memory usage by the randomness pools.
1307 * `{CRYPTO,OPENSSL}_mem_debug_{push,pop}` are now no-ops and have been
1312 * A new type, EVP_KEYEXCH, has been introduced to represent key exchange
1313 algorithms. An implementation of a key exchange algorithm can be obtained
1314 by using the function EVP_KEYEXCH_fetch(). An EVP_KEYEXCH algorithm can be
1315 used in a call to EVP_PKEY_derive_init_ex() which works in a similar way to
1316 the older EVP_PKEY_derive_init() function. See the man pages for the new
1317 functions for further details.
1321 * The EVP_PKEY_CTX_set_dh_pad() macro has now been converted to a function.
1325 * Removed the function names from error messages and deprecated the
1330 * Removed NextStep support and the macro OPENSSL_UNISTD
1334 * Removed DES_check_key. Also removed OPENSSL_IMPLEMENT_GLOBAL,
1335 OPENSSL_GLOBAL_REF, OPENSSL_DECLARE_GLOBAL.
1336 Also removed "export var as function" capability; we do not export
1337 variables, only functions.
1341 * RC5_32_set_key has been changed to return an int type, with 0 indicating
1342 an error and 1 indicating success. In previous versions of OpenSSL this
1343 was a void type. If a key was set longer than the maximum possible this
1348 * Support SM2 signing and verification schemes with X509 certificate.
1352 * Use SHA256 as the default digest for TS query in the `ts` app.
1356 * Change PBKDF2 to conform to SP800-132 instead of the older PKCS5 RFC2898.
1360 * Default cipher lists/suites are now available via a function, the
1361 #defines are deprecated.
1365 * Add target VC-WIN32-UWP, VC-WIN64A-UWP, VC-WIN32-ARM-UWP and
1366 VC-WIN64-ARM-UWP in Windows OneCore target for making building libraries
1367 for Windows Store apps easier. Also, the "no-uplink" option has been added.
1371 * Join the directories crypto/x509 and crypto/x509v3
1375 * Added command 'openssl kdf' that uses the EVP_KDF API.
1379 * Added command 'openssl mac' that uses the EVP_MAC API.
1383 * Added OPENSSL_info() to get diverse built-in OpenSSL data, such
1384 as default directories. Also added the command 'openssl info'
1385 for scripting purposes.
1389 * The functions AES_ige_encrypt() and AES_bi_ige_encrypt() have been
1394 * Add prediction resistance to the DRBG reseeding process.
1398 * Limit the number of blocks in a data unit for AES-XTS to 2^20 as
1399 mandated by IEEE Std 1619-2018.
1403 * Added newline escaping functionality to a filename when using openssl dgst.
1404 This output format is to replicate the output format found in the `*sum`
1405 checksum programs. This aims to preserve backward compatibility.
1407 *Matt Eaton, Richard Levitte, and Paul Dale*
1409 * Removed the heartbeat message in DTLS feature, as it has very
1410 little usage and doesn't seem to fulfill a valuable purpose.
1411 The configuration option is now deprecated.
1415 * Changed the output of 'openssl {digestname} < file' to display the
1416 digest name in its output.
1420 * Added a new generic trace API which provides support for enabling
1421 instrumentation through trace output.
1423 *Richard Levitte & Matthias St. Pierre*
1425 * Added build tests for C++. These are generated files that only do one
1426 thing, to include one public OpenSSL head file each. This tests that
1427 the public header files can be usefully included in a C++ application.
1429 This test isn't enabled by default. It can be enabled with the option
1430 'enable-buildtest-c++'.
1434 * Added KB KDF (EVP_KDF_KB) to EVP_KDF.
1438 * Added SSH KDF (EVP_KDF_SSHKDF) and KRB5 KDF (EVP_KDF_KRB5KDF) to EVP_KDF.
1442 * Added Single Step KDF (EVP_KDF_SS), X963 KDF, and X942 KDF to EVP_KDF.
1446 * Added KMAC to EVP_MAC.
1450 * Added property based algorithm implementation selection framework to
1455 * Added SCA hardening for modular field inversion in EC_GROUP through
1456 a new dedicated field_inv() pointer in EC_METHOD.
1457 This also addresses a leakage affecting conversions from projective
1458 to affine coordinates.
1460 *Billy Bob Brumley, Nicola Tuveri*
1462 * Added EVP_KDF, an EVP layer KDF API, to simplify adding KDF and PRF
1463 implementations. This includes an EVP_PKEY to EVP_KDF bridge for
1464 those algorithms that were already supported through the EVP_PKEY API
1465 (scrypt, TLS1 PRF and HKDF). The low-level KDF functions for PBKDF2
1466 and scrypt are now wrappers that call EVP_KDF.
1470 * Build devcrypto engine as a dynamic engine.
1472 *Eneas U de Queiroz*
1474 * Add keyed BLAKE2 to EVP_MAC.
1478 * Fix a bug in the computation of the endpoint-pair shared secret used
1479 by DTLS over SCTP. This breaks interoperability with older versions
1480 of OpenSSL like OpenSSL 1.1.0 and OpenSSL 1.0.2. There is a runtime
1481 switch SSL_MODE_DTLS_SCTP_LABEL_LENGTH_BUG (off by default) enabling
1482 interoperability with such broken implementations. However, enabling
1483 this switch breaks interoperability with correct implementations.
1485 * Fix a use after free bug in d2i_X509_PUBKEY when overwriting a
1486 re-used X509_PUBKEY object if the second PUBKEY is malformed.
1490 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
1494 * Changed the license to the Apache License v2.0.
1498 * Switch to a new version scheme using three numbers MAJOR.MINOR.PATCH.
1500 - Major releases (indicated by incrementing the MAJOR release number)
1501 may introduce incompatible API/ABI changes.
1502 - Minor releases (indicated by incrementing the MINOR release number)
1503 may introduce new features but retain API/ABI compatibility.
1504 - Patch releases (indicated by incrementing the PATCH number)
1505 are intended for bug fixes and other improvements of existing
1506 features only (like improving performance or adding documentation)
1507 and retain API/ABI compatibility.
1511 * Add support for RFC5297 SIV mode (siv128), including AES-SIV.
1515 * Remove the 'dist' target and add a tarball building script. The
1516 'dist' target has fallen out of use, and it shouldn't be
1517 necessary to configure just to create a source distribution.
1521 * Recreate the OS390-Unix config target. It no longer relies on a
1522 special script like it did for OpenSSL pre-1.1.0.
1526 * Instead of having the source directories listed in Configure, add
1527 a 'build.info' keyword SUBDIRS to indicate what sub-directories to
1532 * Add GMAC to EVP_MAC.
1536 * Ported the HMAC, CMAC and SipHash EVP_PKEY_METHODs to EVP_MAC.
1540 * Added EVP_MAC, an EVP layer MAC API, to simplify adding MAC
1541 implementations. This includes a generic EVP_PKEY to EVP_MAC bridge,
1542 to facilitate the continued use of MACs through raw private keys in
1543 functionality such as `EVP_DigestSign*` and `EVP_DigestVerify*`.
1547 * Deprecate ECDH_KDF_X9_62().
1551 * Added EVP_PKEY_ECDH_KDF_X9_63 and ecdh_KDF_X9_63() as replacements for
1552 the EVP_PKEY_ECDH_KDF_X9_62 KDF type and ECDH_KDF_X9_62(). The old names
1553 are retained for backwards compatibility.
1557 * AES-XTS mode now enforces that its two keys are different to mitigate
1558 the attacked described in "Efficient Instantiations of Tweakable
1559 Blockciphers and Refinements to Modes OCB and PMAC" by Phillip Rogaway.
1560 Details of this attack can be obtained from:
1561 <http://web.cs.ucdavis.edu/%7Erogaway/papers/offsets.pdf>
1565 * Rename the object files, i.e. give them other names than in previous
1566 versions. Their names now include the name of the final product, as
1567 well as its type mnemonic (bin, lib, shlib).
1571 * Added new option for 'openssl list', '-objects', which will display the
1572 list of built in objects, i.e. OIDs with names.
1576 * Added the options `-crl_lastupdate` and `-crl_nextupdate` to `openssl ca`,
1577 allowing the `lastUpdate` and `nextUpdate` fields in the generated CRL to
1582 * Added support for Linux Kernel TLS data-path. The Linux Kernel data-path
1583 improves application performance by removing data copies and providing
1584 applications with zero-copy system calls such as sendfile and splice.
1588 * The SSL option SSL_OP_CLEANSE_PLAINTEXT is introduced.
1592 * `PKCS12_parse` now maintains the order of the parsed certificates
1593 when outputting them via `*ca` (rather than reversing it).
1597 * Deprecated pthread fork support methods.
1601 * Added support for FFDHE key exchange in TLS 1.3.
1605 * Added a new concept for OpenSSL plugability: providers. This
1606 functionality is designed to replace the ENGINE API and ENGINE
1607 implementations, and to be much more dynamic, allowing provider
1608 authors to introduce new algorithms among other things, as long as
1609 there's an API that supports the algorithm type.
1611 With this concept comes a new core API for interaction between
1612 libcrypto and provider implementations. Public libcrypto functions
1613 that want to use providers do so through this core API.
1615 The main documentation for this core API is found in
1616 doc/man7/provider.pod, doc/man7/provider-base.pod, and they in turn
1617 refer to other manuals describing the API specific for supported
1618 algorithm types (also called operations).
1625 ### Changes between 1.1.1m and 1.1.1n [xx XXX xxxx]
1627 ### Changes between 1.1.1l and 1.1.1m [14 Dec 2021]
1629 * Avoid loading of a dynamic engine twice.
1633 * Prioritise DANE TLSA issuer certs over peer certs
1637 * Fixed random API for MacOS prior to 10.12
1639 These MacOS versions don't support the CommonCrypto APIs
1643 ### Changes between 1.1.1k and 1.1.1l [24 Aug 2021]
1645 * Fixed an SM2 Decryption Buffer Overflow.
1647 In order to decrypt SM2 encrypted data an application is expected to
1648 call the API function EVP_PKEY_decrypt(). Typically an application will
1649 call this function twice. The first time, on entry, the "out" parameter
1650 can be NULL and, on exit, the "outlen" parameter is populated with the
1651 buffer size required to hold the decrypted plaintext. The application
1652 can then allocate a sufficiently sized buffer and call EVP_PKEY_decrypt()
1653 again, but this time passing a non-NULL value for the "out" parameter.
1655 A bug in the implementation of the SM2 decryption code means that the
1656 calculation of the buffer size required to hold the plaintext returned
1657 by the first call to EVP_PKEY_decrypt() can be smaller than the actual
1658 size required by the second call. This can lead to a buffer overflow
1659 when EVP_PKEY_decrypt() is called by the application a second time with
1660 a buffer that is too small.
1662 A malicious attacker who is able present SM2 content for decryption to
1663 an application could cause attacker chosen data to overflow the buffer
1664 by up to a maximum of 62 bytes altering the contents of other data held
1665 after the buffer, possibly changing application behaviour or causing
1666 the application to crash. The location of the buffer is application
1667 dependent but is typically heap allocated.
1672 * Fixed various read buffer overruns processing ASN.1 strings
1674 ASN.1 strings are represented internally within OpenSSL as an ASN1_STRING
1675 structure which contains a buffer holding the string data and a field
1676 holding the buffer length. This contrasts with normal C strings which
1677 are repesented as a buffer for the string data which is terminated
1678 with a NUL (0) byte.
1680 Although not a strict requirement, ASN.1 strings that are parsed using
1681 OpenSSL's own "d2i" functions (and other similar parsing functions) as
1682 well as any string whose value has been set with the ASN1_STRING_set()
1683 function will additionally NUL terminate the byte array in the
1684 ASN1_STRING structure.
1686 However, it is possible for applications to directly construct valid
1687 ASN1_STRING structures which do not NUL terminate the byte array by
1688 directly setting the "data" and "length" fields in the ASN1_STRING
1689 array. This can also happen by using the ASN1_STRING_set0() function.
1691 Numerous OpenSSL functions that print ASN.1 data have been found to
1692 assume that the ASN1_STRING byte array will be NUL terminated, even
1693 though this is not guaranteed for strings that have been directly
1694 constructed. Where an application requests an ASN.1 structure to be
1695 printed, and where that ASN.1 structure contains ASN1_STRINGs that have
1696 been directly constructed by the application without NUL terminating
1697 the "data" field, then a read buffer overrun can occur.
1699 The same thing can also occur during name constraints processing
1700 of certificates (for example if a certificate has been directly
1701 constructed by the application instead of loading it via the OpenSSL
1702 parsing functions, and the certificate contains non NUL terminated
1703 ASN1_STRING structures). It can also occur in the X509_get1_email(),
1704 X509_REQ_get1_email() and X509_get1_ocsp() functions.
1706 If a malicious actor can cause an application to directly construct an
1707 ASN1_STRING and then process it through one of the affected OpenSSL
1708 functions then this issue could be hit. This might result in a crash
1709 (causing a Denial of Service attack). It could also result in the
1710 disclosure of private memory contents (such as private keys, or
1711 sensitive plaintext).
1716 ### Changes between 1.1.1j and 1.1.1k [25 Mar 2021]
1718 * Fixed a problem with verifying a certificate chain when using the
1719 X509_V_FLAG_X509_STRICT flag. This flag enables additional security checks of
1720 the certificates present in a certificate chain. It is not set by default.
1722 Starting from OpenSSL version 1.1.1h a check to disallow certificates in
1723 the chain that have explicitly encoded elliptic curve parameters was added
1724 as an additional strict check.
1726 An error in the implementation of this check meant that the result of a
1727 previous check to confirm that certificates in the chain are valid CA
1728 certificates was overwritten. This effectively bypasses the check
1729 that non-CA certificates must not be able to issue other certificates.
1731 If a "purpose" has been configured then there is a subsequent opportunity
1732 for checks that the certificate is a valid CA. All of the named "purpose"
1733 values implemented in libcrypto perform this check. Therefore, where
1734 a purpose is set the certificate chain will still be rejected even when the
1735 strict flag has been used. A purpose is set by default in libssl client and
1736 server certificate verification routines, but it can be overridden or
1737 removed by an application.
1739 In order to be affected, an application must explicitly set the
1740 X509_V_FLAG_X509_STRICT verification flag and either not set a purpose
1741 for the certificate verification or, in the case of TLS client or server
1742 applications, override the default purpose.
1747 * Fixed an issue where an OpenSSL TLS server may crash if sent a maliciously
1748 crafted renegotiation ClientHello message from a client. If a TLSv1.2
1749 renegotiation ClientHello omits the signature_algorithms extension (where it
1750 was present in the initial ClientHello), but includes a
1751 signature_algorithms_cert extension then a NULL pointer dereference will
1752 result, leading to a crash and a denial of service attack.
1754 A server is only vulnerable if it has TLSv1.2 and renegotiation enabled
1755 (which is the default configuration). OpenSSL TLS clients are not impacted by
1759 *Peter Kästle and Samuel Sapalski*
1761 ### Changes between 1.1.1i and 1.1.1j [16 Feb 2021]
1763 * Fixed the X509_issuer_and_serial_hash() function. It attempts to
1764 create a unique hash value based on the issuer and serial number data
1765 contained within an X509 certificate. However it was failing to correctly
1766 handle any errors that may occur while parsing the issuer field (which might
1767 occur if the issuer field is maliciously constructed). This may subsequently
1768 result in a NULL pointer deref and a crash leading to a potential denial of
1774 * Fixed the RSA_padding_check_SSLv23() function and the RSA_SSLV23_PADDING
1775 padding mode to correctly check for rollback attacks. This is considered a
1776 bug in OpenSSL 1.1.1 because it does not support SSLv2. In 1.0.2 this is
1781 Fixed the EVP_CipherUpdate, EVP_EncryptUpdate and EVP_DecryptUpdate
1782 functions. Previously they could overflow the output length argument in some
1783 cases where the input length is close to the maximum permissable length for
1784 an integer on the platform. In such cases the return value from the function
1785 call would be 1 (indicating success), but the output length value would be
1786 negative. This could cause applications to behave incorrectly or crash.
1791 * Fixed SRP_Calc_client_key so that it runs in constant time. The previous
1792 implementation called BN_mod_exp without setting BN_FLG_CONSTTIME. This
1793 could be exploited in a side channel attack to recover the password. Since
1794 the attack is local host only this is outside of the current OpenSSL
1795 threat model and therefore no CVE is assigned.
1797 Thanks to Mohammed Sabt and Daniel De Almeida Braga for reporting this
1802 ### Changes between 1.1.1h and 1.1.1i [8 Dec 2020]
1804 * Fixed NULL pointer deref in the GENERAL_NAME_cmp function
1805 This function could crash if both GENERAL_NAMEs contain an EDIPARTYNAME.
1806 If an attacker can control both items being compared then this could lead
1807 to a possible denial of service attack. OpenSSL itself uses the
1808 GENERAL_NAME_cmp function for two purposes:
1809 1) Comparing CRL distribution point names between an available CRL and a
1810 CRL distribution point embedded in an X509 certificate
1811 2) When verifying that a timestamp response token signer matches the
1812 timestamp authority name (exposed via the API functions
1813 TS_RESP_verify_response and TS_RESP_verify_token)
1818 ### Changes between 1.1.1g and 1.1.1h [22 Sep 2020]
1820 * Certificates with explicit curve parameters are now disallowed in
1821 verification chains if the X509_V_FLAG_X509_STRICT flag is used.
1825 * The 'MinProtocol' and 'MaxProtocol' configuration commands now silently
1826 ignore TLS protocol version bounds when configuring DTLS-based contexts, and
1827 conversely, silently ignore DTLS protocol version bounds when configuring
1828 TLS-based contexts. The commands can be repeated to set bounds of both
1829 types. The same applies with the corresponding "min_protocol" and
1830 "max_protocol" command-line switches, in case some application uses both TLS
1833 SSL_CTX instances that are created for a fixed protocol version (e.g.
1834 TLSv1_server_method()) also silently ignore version bounds. Previously
1835 attempts to apply bounds to these protocol versions would result in an
1836 error. Now only the "version-flexible" SSL_CTX instances are subject to
1837 limits in configuration files in command-line options.
1841 * Handshake now fails if Extended Master Secret extension is dropped
1846 * The Oracle Developer Studio compiler will start reporting deprecated APIs
1848 ### Changes between 1.1.1f and 1.1.1g [21 Apr 2020]
1850 * Fixed segmentation fault in SSL_check_chain()
1851 Server or client applications that call the SSL_check_chain() function
1852 during or after a TLS 1.3 handshake may crash due to a NULL pointer
1853 dereference as a result of incorrect handling of the
1854 "signature_algorithms_cert" TLS extension. The crash occurs if an invalid
1855 or unrecognised signature algorithm is received from the peer. This could
1856 be exploited by a malicious peer in a Denial of Service attack.
1861 * Added AES consttime code for no-asm configurations
1862 an optional constant time support for AES was added
1863 when building openssl for no-asm.
1864 Enable with: ./config no-asm -DOPENSSL_AES_CONST_TIME
1865 Disable with: ./config no-asm -DOPENSSL_NO_AES_CONST_TIME
1866 At this time this feature is by default disabled.
1867 It will be enabled by default in 3.0.
1871 ### Changes between 1.1.1e and 1.1.1f [31 Mar 2020]
1873 * Revert the change of EOF detection while reading in libssl to avoid
1874 regressions in applications depending on the current way of reporting
1875 the EOF. As the existing method is not fully accurate the change to
1876 reporting the EOF via SSL_ERROR_SSL is kept on the current development
1877 branch and will be present in the 3.0 release.
1881 * Revised BN_generate_prime_ex to not avoid factors 3..17863 in p-1
1882 when primes for RSA keys are computed.
1883 Since we previously always generated primes == 2 (mod 3) for RSA keys,
1884 the 2-prime and 3-prime RSA modules were easy to distinguish, since
1885 N = p*q = 1 (mod 3), but N = p*q*r = 2 (mod 3). Therefore fingerprinting
1886 2-prime vs. 3-prime RSA keys was possible by computing N mod 3.
1887 This avoids possible fingerprinting of newly generated RSA modules.
1891 ### Changes between 1.1.1d and 1.1.1e [17 Mar 2020]
1893 * Properly detect EOF while reading in libssl. Previously if we hit an EOF
1894 while reading in libssl then we would report an error back to the
1895 application (SSL_ERROR_SYSCALL) but errno would be 0. We now add
1896 an error to the stack (which means we instead return SSL_ERROR_SSL) and
1897 therefore give a hint as to what went wrong.
1901 * Check that ed25519 and ed448 are allowed by the security level. Previously
1902 signature algorithms not using an MD were not being checked that they were
1903 allowed by the security level.
1907 * Fixed SSL_get_servername() behaviour. The behaviour of SSL_get_servername()
1908 was not quite right. The behaviour was not consistent between resumption
1909 and normal handshakes, and also not quite consistent with historical
1910 behaviour. The behaviour in various scenarios has been clarified and
1911 it has been updated to make it match historical behaviour as closely as
1916 * *[VMS only]* The header files that the VMS compilers include automatically,
1917 `__DECC_INCLUDE_PROLOGUE.H` and `__DECC_INCLUDE_EPILOGUE.H`, use pragmas
1918 that the C++ compiler doesn't understand. This is a shortcoming in the
1919 compiler, but can be worked around with `__cplusplus` guards.
1921 C++ applications that use OpenSSL libraries must be compiled using the
1922 qualifier `/NAMES=(AS_IS,SHORTENED)` to be able to use all the OpenSSL
1923 functions. Otherwise, only functions with symbols of less than 31
1924 characters can be used, as the linker will not be able to successfully
1925 resolve symbols with longer names.
1929 * Added a new method to gather entropy on VMS, based on SYS$GET_ENTROPY.
1930 The presence of this system service is determined at run-time.
1934 * Print all values for a PKCS#12 attribute with 'openssl pkcs12', not just
1939 ### Changes between 1.1.1c and 1.1.1d [10 Sep 2019]
1941 * Fixed a fork protection issue. OpenSSL 1.1.1 introduced a rewritten random
1942 number generator (RNG). This was intended to include protection in the
1943 event of a fork() system call in order to ensure that the parent and child
1944 processes did not share the same RNG state. However this protection was not
1945 being used in the default case.
1947 A partial mitigation for this issue is that the output from a high
1948 precision timer is mixed into the RNG state so the likelihood of a parent
1949 and child process sharing state is significantly reduced.
1951 If an application already calls OPENSSL_init_crypto() explicitly using
1952 OPENSSL_INIT_ATFORK then this problem does not occur at all.
1955 *Matthias St. Pierre*
1957 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
1958 used even when parsing explicit parameters, when loading a encoded key
1959 or calling `EC_GROUP_new_from_ecpkparameters()`/
1960 `EC_GROUP_new_from_ecparameters()`.
1961 This prevents bypass of security hardening and performance gains,
1962 especially for curves with specialized EC_METHODs.
1963 By default, if a key encoded with explicit parameters is loaded and later
1964 encoded, the output is still encoded with explicit parameters, even if
1965 internally a "named" EC_GROUP is used for computation.
1969 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
1970 this change, EC_GROUP_set_generator would accept order and/or cofactor as
1971 NULL. After this change, only the cofactor parameter can be NULL. It also
1972 does some minimal sanity checks on the passed order.
1977 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
1978 An attack is simple, if the first CMS_recipientInfo is valid but the
1979 second CMS_recipientInfo is chosen ciphertext. If the second
1980 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
1981 encryption key will be replaced by garbage, and the message cannot be
1982 decoded, but if the RSA decryption fails, the correct encryption key is
1983 used and the recipient will not notice the attack.
1984 As a work around for this potential attack the length of the decrypted
1985 key must be equal to the cipher default key length, in case the
1986 certifiate is not given and all recipientInfo are tried out.
1987 The old behaviour can be re-enabled in the CMS code by setting the
1988 CMS_DEBUG_DECRYPT flag.
1993 * Early start up entropy quality from the DEVRANDOM seed source has been
1994 improved for older Linux systems. The RAND subsystem will wait for
1995 /dev/random to be producing output before seeding from /dev/urandom.
1996 The seeded state is stored for future library initialisations using
1997 a system global shared memory segment. The shared memory identifier
1998 can be configured by defining OPENSSL_RAND_SEED_DEVRANDOM_SHM_ID to
1999 the desired value. The default identifier is 114.
2003 * Correct the extended master secret constant on EBCDIC systems. Without this
2004 fix TLS connections between an EBCDIC system and a non-EBCDIC system that
2005 negotiate EMS will fail. Unfortunately this also means that TLS connections
2006 between EBCDIC systems with this fix, and EBCDIC systems without this
2007 fix will fail if they negotiate EMS.
2011 * Use Windows installation paths in the mingw builds
2013 Mingw isn't a POSIX environment per se, which means that Windows
2014 paths should be used for installation.
2019 * Changed DH_check to accept parameters with order q and 2q subgroups.
2020 With order 2q subgroups the bit 0 of the private key is not secret
2021 but DH_generate_key works around that by clearing bit 0 of the
2022 private key for those. This avoids leaking bit 0 of the private key.
2026 * Significantly reduce secure memory usage by the randomness pools.
2030 * Revert the DEVRANDOM_WAIT feature for Linux systems
2032 The DEVRANDOM_WAIT feature added a select() call to wait for the
2033 /dev/random device to become readable before reading from the
2034 /dev/urandom device.
2036 It turned out that this change had negative side effects on
2037 performance which were not acceptable. After some discussion it
2038 was decided to revert this feature and leave it up to the OS
2039 resp. the platform maintainer to ensure a proper initialization
2040 during early boot time.
2042 *Matthias St. Pierre*
2044 ### Changes between 1.1.1b and 1.1.1c [28 May 2019]
2046 * Add build tests for C++. These are generated files that only do one
2047 thing, to include one public OpenSSL head file each. This tests that
2048 the public header files can be usefully included in a C++ application.
2050 This test isn't enabled by default. It can be enabled with the option
2051 'enable-buildtest-c++'.
2055 * Enable SHA3 pre-hashing for ECDSA and DSA.
2059 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
2060 This changes the size when using the `genpkey` command when no size is given.
2061 It fixes an omission in earlier changes that changed all RSA, DSA and DH
2062 generation commands to use 2048 bits by default.
2066 * Reorganize the manual pages to consistently have RETURN VALUES,
2067 EXAMPLES, SEE ALSO and HISTORY come in that order, and adjust
2068 util/fix-doc-nits accordingly.
2070 *Paul Yang, Joshua Lock*
2072 * Add the missing accessor EVP_PKEY_get0_engine()
2076 * Have commands like `s_client` and `s_server` output the signature scheme
2077 along with other cipher suite parameters when debugging.
2081 * Make OPENSSL_config() error agnostic again.
2085 * Do the error handling in RSA decryption constant time.
2089 * Prevent over long nonces in ChaCha20-Poly1305.
2091 ChaCha20-Poly1305 is an AEAD cipher, and requires a unique nonce input
2092 for every encryption operation. RFC 7539 specifies that the nonce value
2093 (IV) should be 96 bits (12 bytes). OpenSSL allows a variable nonce length
2094 and front pads the nonce with 0 bytes if it is less than 12
2095 bytes. However it also incorrectly allows a nonce to be set of up to 16
2096 bytes. In this case only the last 12 bytes are significant and any
2097 additional leading bytes are ignored.
2099 It is a requirement of using this cipher that nonce values are
2100 unique. Messages encrypted using a reused nonce value are susceptible to
2101 serious confidentiality and integrity attacks. If an application changes
2102 the default nonce length to be longer than 12 bytes and then makes a
2103 change to the leading bytes of the nonce expecting the new value to be a
2104 new unique nonce then such an application could inadvertently encrypt
2105 messages with a reused nonce.
2107 Additionally the ignored bytes in a long nonce are not covered by the
2108 integrity guarantee of this cipher. Any application that relies on the
2109 integrity of these ignored leading bytes of a long nonce may be further
2110 affected. Any OpenSSL internal use of this cipher, including in SSL/TLS,
2111 is safe because no such use sets such a long nonce value. However user
2112 applications that use this cipher directly and set a non-default nonce
2113 length to be longer than 12 bytes may be vulnerable.
2115 This issue was reported to OpenSSL on 16th of March 2019 by Joran Dirk
2121 * Add DEVRANDOM_WAIT feature for Linux systems
2123 On older Linux systems where the getrandom() system call is not available,
2124 OpenSSL normally uses the /dev/urandom device for seeding its CSPRNG.
2125 Contrary to getrandom(), the /dev/urandom device will not block during
2126 early boot when the kernel CSPRNG has not been seeded yet.
2128 To mitigate this known weakness, use select() to wait for /dev/random to
2129 become readable before reading from /dev/urandom.
2131 * Ensure that SM2 only uses SM3 as digest algorithm
2135 ### Changes between 1.1.1a and 1.1.1b [26 Feb 2019]
2137 * Change the info callback signals for the start and end of a post-handshake
2138 message exchange in TLSv1.3. In 1.1.1/1.1.1a we used SSL_CB_HANDSHAKE_START
2139 and SSL_CB_HANDSHAKE_DONE. Experience has shown that many applications get
2140 confused by this and assume that a TLSv1.2 renegotiation has started. This
2141 can break KeyUpdate handling. Instead we no longer signal the start and end
2142 of a post handshake message exchange (although the messages themselves are
2143 still signalled). This could break some applications that were expecting
2144 the old signals. However without this KeyUpdate is not usable for many
2149 ### Changes between 1.1.1 and 1.1.1a [20 Nov 2018]
2151 * Timing vulnerability in DSA signature generation
2153 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
2154 timing side channel attack. An attacker could use variations in the signing
2155 algorithm to recover the private key.
2157 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
2162 * Timing vulnerability in ECDSA signature generation
2164 The OpenSSL ECDSA signature algorithm has been shown to be vulnerable to a
2165 timing side channel attack. An attacker could use variations in the signing
2166 algorithm to recover the private key.
2168 This issue was reported to OpenSSL on 25th October 2018 by Samuel Weiser.
2173 * Fixed the issue that RAND_add()/RAND_seed() silently discards random input
2174 if its length exceeds 4096 bytes. The limit has been raised to a buffer size
2175 of two gigabytes and the error handling improved.
2177 This issue was reported to OpenSSL by Dr. Falko Strenzke. It has been
2178 categorized as a normal bug, not a security issue, because the DRBG reseeds
2179 automatically and is fully functional even without additional randomness
2180 provided by the application.
2182 ### Changes between 1.1.0i and 1.1.1 [11 Sep 2018]
2184 * Add a new ClientHello callback. Provides a callback interface that gives
2185 the application the ability to adjust the nascent SSL object at the
2186 earliest stage of ClientHello processing, immediately after extensions have
2187 been collected but before they have been processed. In particular, this
2188 callback can adjust the supported TLS versions in response to the contents
2193 * Add SM2 base algorithm support.
2197 * s390x assembly pack: add (improved) hardware-support for the following
2198 cryptographic primitives: sha3, shake, aes-gcm, aes-ccm, aes-ctr, aes-ofb,
2199 aes-cfb/cfb8, aes-ecb.
2203 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
2204 parameter is no longer accepted, as it leads to a corrupt table. NULL
2205 pem_str is reserved for alias entries only.
2209 * Use the new ec_scalar_mul_ladder scaffold to implement a specialized ladder
2210 step for prime curves. The new implementation is based on formulae from
2211 differential addition-and-doubling in homogeneous projective coordinates
2212 from Izu-Takagi "A fast parallel elliptic curve multiplication resistant
2213 against side channel attacks" and Brier-Joye "Weierstrass Elliptic Curves
2214 and Side-Channel Attacks" Eq. (8) for y-coordinate recovery, modified
2215 to work in projective coordinates.
2217 *Billy Bob Brumley, Nicola Tuveri*
2219 * Change generating and checking of primes so that the error rate of not
2220 being prime depends on the intended use based on the size of the input.
2221 For larger primes this will result in more rounds of Miller-Rabin.
2222 The maximal error rate for primes with more than 1080 bits is lowered
2225 *Kurt Roeckx, Annie Yousar*
2227 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
2231 * The 'tsget' script is renamed to 'tsget.pl', to avoid confusion when
2232 moving between systems, and to avoid confusion when a Windows build is
2233 done with mingw vs with MSVC. For POSIX installs, there's still a
2234 symlink or copy named 'tsget' to avoid that confusion as well.
2238 * Revert blinding in ECDSA sign and instead make problematic addition
2239 length-invariant. Switch even to fixed-length Montgomery multiplication.
2243 * Use the new ec_scalar_mul_ladder scaffold to implement a specialized ladder
2244 step for binary curves. The new implementation is based on formulae from
2245 differential addition-and-doubling in mixed Lopez-Dahab projective
2246 coordinates, modified to independently blind the operands.
2248 *Billy Bob Brumley, Sohaib ul Hassan, Nicola Tuveri*
2250 * Add a scaffold to optionally enhance the Montgomery ladder implementation
2251 for `ec_scalar_mul_ladder` (formerly `ec_mul_consttime`) allowing
2252 EC_METHODs to implement their own specialized "ladder step", to take
2253 advantage of more favorable coordinate systems or more efficient
2254 differential addition-and-doubling algorithms.
2256 *Billy Bob Brumley, Sohaib ul Hassan, Nicola Tuveri*
2258 * Modified the random device based seed sources to keep the relevant
2259 file descriptors open rather than reopening them on each access.
2260 This allows such sources to operate in a chroot() jail without
2261 the associated device nodes being available. This behaviour can be
2262 controlled using RAND_keep_random_devices_open().
2266 * Numerous side-channel attack mitigations have been applied. This may have
2267 performance impacts for some algorithms for the benefit of improved
2268 security. Specific changes are noted in this change log by their respective
2273 * AIX shared library support overhaul. Switch to AIX "natural" way of
2274 handling shared libraries, which means collecting shared objects of
2275 different versions and bitnesses in one common archive. This allows to
2276 mitigate conflict between 1.0 and 1.1 side-by-side installations. It
2277 doesn't affect the way 3rd party applications are linked, only how
2278 multi-version installation is managed.
2282 * Make ec_group_do_inverse_ord() more robust and available to other
2283 EC cryptosystems, so that irrespective of BN_FLG_CONSTTIME, SCA
2284 mitigations are applied to the fallback BN_mod_inverse().
2285 When using this function rather than BN_mod_inverse() directly, new
2286 EC cryptosystem implementations are then safer-by-default.
2290 * Add coordinate blinding for EC_POINT and implement projective
2291 coordinate blinding for generic prime curves as a countermeasure to
2292 chosen point SCA attacks.
2294 *Sohaib ul Hassan, Nicola Tuveri, Billy Bob Brumley*
2296 * Add blinding to ECDSA and DSA signatures to protect against side channel
2297 attacks discovered by Keegan Ryan (NCC Group).
2301 * Enforce checking in the `pkeyutl` command to ensure that the input
2302 length does not exceed the maximum supported digest length when performing
2303 a sign, verify or verifyrecover operation.
2307 * SSL_MODE_AUTO_RETRY is enabled by default. Applications that use blocking
2308 I/O in combination with something like select() or poll() will hang. This
2309 can be turned off again using SSL_CTX_clear_mode().
2310 Many applications do not properly handle non-application data records, and
2311 TLS 1.3 sends more of such records. Setting SSL_MODE_AUTO_RETRY works
2312 around the problems in those applications, but can also break some.
2313 It's recommended to read the manpages about SSL_read(), SSL_write(),
2314 SSL_get_error(), SSL_shutdown(), SSL_CTX_set_mode() and
2315 SSL_CTX_set_read_ahead() again.
2319 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
2320 now allow empty (zero character) pass phrases.
2324 * Apply blinding to binary field modular inversion and remove patent
2325 pending (OPENSSL_SUN_GF2M_DIV) BN_GF2m_mod_div implementation.
2329 * Deprecate ec2_mult.c and unify scalar multiplication code paths for
2330 binary and prime elliptic curves.
2334 * Remove ECDSA nonce padding: EC_POINT_mul is now responsible for
2335 constant time fixed point multiplication.
2339 * Revise elliptic curve scalar multiplication with timing attack
2340 defenses: ec_wNAF_mul redirects to a constant time implementation
2341 when computing fixed point and variable point multiplication (which
2342 in OpenSSL are mostly used with secret scalars in keygen, sign,
2343 ECDH derive operations).
2344 *Billy Bob Brumley, Nicola Tuveri, Cesar Pereida García,
2347 * Updated CONTRIBUTING
2351 * Updated DRBG / RAND to request nonce and additional low entropy
2352 randomness from the system.
2354 *Matthias St. Pierre*
2356 * Updated 'openssl rehash' to use OpenSSL consistent default.
2360 * Moved the load of the ssl_conf module to libcrypto, which helps
2361 loading engines that libssl uses before libssl is initialised.
2365 * Added EVP_PKEY_sign() and EVP_PKEY_verify() for EdDSA
2369 * Fixed X509_NAME_ENTRY_set to get multi-valued RDNs right in all cases.
2371 *Ingo Schwarze, Rich Salz*
2373 * Added output of accepting IP address and port for 'openssl s_server'
2377 * Added a new API for TLSv1.3 ciphersuites:
2378 SSL_CTX_set_ciphersuites()
2379 SSL_set_ciphersuites()
2383 * Memory allocation failures consistently add an error to the error
2388 * Don't use OPENSSL_ENGINES and OPENSSL_CONF environment values
2389 in libcrypto when run as setuid/setgid.
2393 * Load any config file by default when libssl is used.
2397 * Added new public header file <openssl/rand_drbg.h> and documentation
2398 for the RAND_DRBG API. See manual page RAND_DRBG(7) for an overview.
2400 *Matthias St. Pierre*
2402 * QNX support removed (cannot find contributors to get their approval
2403 for the license change).
2407 * TLSv1.3 replay protection for early data has been implemented. See the
2408 SSL_read_early_data() man page for further details.
2412 * Separated TLSv1.3 ciphersuite configuration out from TLSv1.2 ciphersuite
2413 configuration. TLSv1.3 ciphersuites are not compatible with TLSv1.2 and
2414 below. Similarly TLSv1.2 ciphersuites are not compatible with TLSv1.3.
2415 In order to avoid issues where legacy TLSv1.2 ciphersuite configuration
2416 would otherwise inadvertently disable all TLSv1.3 ciphersuites the
2417 configuration has been separated out. See the ciphers man page or the
2418 SSL_CTX_set_ciphersuites() man page for more information.
2422 * On POSIX (BSD, Linux, ...) systems the ocsp(1) command running
2423 in responder mode now supports the new "-multi" option, which
2424 spawns the specified number of child processes to handle OCSP
2425 requests. The "-timeout" option now also limits the OCSP
2426 responder's patience to wait to receive the full client request
2427 on a newly accepted connection. Child processes are respawned
2428 as needed, and the CA index file is automatically reloaded
2429 when changed. This makes it possible to run the "ocsp" responder
2430 as a long-running service, making the OpenSSL CA somewhat more
2431 feature-complete. In this mode, most diagnostic messages logged
2432 after entering the event loop are logged via syslog(3) rather than
2437 * Added support for X448 and Ed448. Heavily based on original work by
2442 * Extend OSSL_STORE with capabilities to search and to narrow the set of
2443 objects loaded. This adds the functions OSSL_STORE_expect() and
2444 OSSL_STORE_find() as well as needed tools to construct searches and
2445 get the search data out of them.
2449 * Support for TLSv1.3 added. Note that users upgrading from an earlier
2450 version of OpenSSL should review their configuration settings to ensure
2451 that they are still appropriate for TLSv1.3. For further information see:
2452 <https://wiki.openssl.org/index.php/TLS1.3>
2456 * Grand redesign of the OpenSSL random generator
2458 The default RAND method now utilizes an AES-CTR DRBG according to
2459 NIST standard SP 800-90Ar1. The new random generator is essentially
2460 a port of the default random generator from the OpenSSL FIPS 2.0
2461 object module. It is a hybrid deterministic random bit generator
2462 using an AES-CTR bit stream and which seeds and reseeds itself
2463 automatically using trusted system entropy sources.
2465 Some of its new features are:
2466 - Support for multiple DRBG instances with seed chaining.
2467 - The default RAND method makes use of a DRBG.
2468 - There is a public and private DRBG instance.
2469 - The DRBG instances are fork-safe.
2470 - Keep all global DRBG instances on the secure heap if it is enabled.
2471 - The public and private DRBG instance are per thread for lock free
2474 *Paul Dale, Benjamin Kaduk, Kurt Roeckx, Rich Salz, Matthias St. Pierre*
2476 * Changed Configure so it only says what it does and doesn't dump
2477 so much data. Instead, ./configdata.pm should be used as a script
2478 to display all sorts of configuration data.
2482 * Added processing of "make variables" to Configure.
2486 * Added SHA512/224 and SHA512/256 algorithm support.
2490 * The last traces of Netware support, first removed in 1.1.0, have
2495 * Get rid of Makefile.shared, and in the process, make the processing
2496 of certain files (rc.obj, or the .def/.map/.opt files produced from
2497 the ordinal files) more visible and hopefully easier to trace and
2498 debug (or make silent).
2502 * Make it possible to have environment variable assignments as
2503 arguments to config / Configure.
2507 * Add multi-prime RSA (RFC 8017) support.
2511 * Add SM3 implemented according to GB/T 32905-2016
2512 *Jack Lloyd <jack.lloyd@ribose.com>,*
2513 *Ronald Tse <ronald.tse@ribose.com>,*
2514 *Erick Borsboom <erick.borsboom@ribose.com>*
2516 * Add 'Maximum Fragment Length' TLS extension negotiation and support
2517 as documented in RFC6066.
2518 Based on a patch from Tomasz Moń
2520 *Filipe Raimundo da Silva*
2522 * Add SM4 implemented according to GB/T 32907-2016.
2523 *Jack Lloyd <jack.lloyd@ribose.com>,*
2524 *Ronald Tse <ronald.tse@ribose.com>,*
2525 *Erick Borsboom <erick.borsboom@ribose.com>*
2527 * Reimplement -newreq-nodes and ERR_error_string_n; the
2528 original author does not agree with the license change.
2532 * Add ARIA AEAD TLS support.
2536 * Some macro definitions to support VS6 have been removed. Visual
2537 Studio 6 has not worked since 1.1.0
2541 * Add ERR_clear_last_mark(), to allow callers to clear the last mark
2542 without clearing the errors.
2546 * Add "atfork" functions. If building on a system that without
2547 pthreads, see doc/man3/OPENSSL_fork_prepare.pod for application
2548 requirements. The RAND facility now uses/requires this.
2556 * The UI API becomes a permanent and integral part of libcrypto, i.e.
2557 not possible to disable entirely. However, it's still possible to
2558 disable the console reading UI method, UI_OpenSSL() (use UI_null()
2561 To disable, configure with 'no-ui-console'. 'no-ui' is still
2562 possible to use as an alias. Check at compile time with the
2563 macro OPENSSL_NO_UI_CONSOLE. The macro OPENSSL_NO_UI is still
2564 possible to check and is an alias for OPENSSL_NO_UI_CONSOLE.
2568 * Add a STORE module, which implements a uniform and URI based reader of
2569 stores that can contain keys, certificates, CRLs and numerous other
2570 objects. The main API is loosely based on a few stdio functions,
2571 and includes OSSL_STORE_open, OSSL_STORE_load, OSSL_STORE_eof,
2572 OSSL_STORE_error and OSSL_STORE_close.
2573 The implementation uses backends called "loaders" to implement arbitrary
2574 URI schemes. There is one built in "loader" for the 'file' scheme.
2578 * Add devcrypto engine. This has been implemented against cryptodev-linux,
2579 then adjusted to work on FreeBSD 8.4 as well.
2580 Enable by configuring with 'enable-devcryptoeng'. This is done by default
2581 on BSD implementations, as cryptodev.h is assumed to exist on all of them.
2585 * Module names can prefixed with OSSL_ or OPENSSL_. This affects
2586 util/mkerr.pl, which is adapted to allow those prefixes, leading to
2587 error code calls like this:
2589 OSSL_FOOerr(OSSL_FOO_F_SOMETHING, OSSL_FOO_R_WHATEVER);
2591 With this change, we claim the namespaces OSSL and OPENSSL in a manner
2592 that can be encoded in C. For the foreseeable future, this will only
2595 *Richard Levitte and Tim Hudson*
2597 * Removed BSD cryptodev engine.
2601 * Add a build target 'build_all_generated', to build all generated files
2602 and only that. This can be used to prepare everything that requires
2603 things like perl for a system that lacks perl and then move everything
2604 to that system and do the rest of the build there.
2608 * In the UI interface, make it possible to duplicate the user data. This
2609 can be used by engines that need to retain the data for a longer time
2610 than just the call where this user data is passed.
2614 * Ignore the '-named_curve auto' value for compatibility of applications
2617 *Tomáš Mráz <tmraz@fedoraproject.org>*
2619 * Fragmented SSL/TLS alerts are no longer accepted. An alert message is 2
2620 bytes long. In theory it is permissible in SSLv3 - TLSv1.2 to fragment such
2621 alerts across multiple records (some of which could be empty). In practice
2622 it make no sense to send an empty alert record, or to fragment one. TLSv1.3
2623 prohibits this altogether and other libraries (BoringSSL, NSS) do not
2624 support this at all. Supporting it adds significant complexity to the
2625 record layer, and its removal is unlikely to cause interoperability
2630 * Add the ASN.1 types INT32, UINT32, INT64, UINT64 and variants prefixed
2631 with Z. These are meant to replace LONG and ZLONG and to be size safe.
2632 The use of LONG and ZLONG is discouraged and scheduled for deprecation
2637 * Add the 'z' and 'j' modifiers to BIO_printf() et al formatting string,
2638 'z' is to be used for [s]size_t, and 'j' - with [u]int64_t.
2640 *Richard Levitte, Andy Polyakov*
2642 * Add EC_KEY_get0_engine(), which does for EC_KEY what RSA_get0_engine()
2647 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
2648 platform rather than 'mingw'.
2652 * The functions X509_STORE_add_cert and X509_STORE_add_crl return
2653 success if they are asked to add an object which already exists
2654 in the store. This change cascades to other functions which load
2655 certificates and CRLs.
2659 * x86_64 assembly pack: annotate code with DWARF CFI directives to
2660 facilitate stack unwinding even from assembly subroutines.
2664 * Remove VAX C specific definitions of OPENSSL_EXPORT, OPENSSL_EXTERN.
2665 Also remove OPENSSL_GLOBAL entirely, as it became a no-op.
2669 * Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
2670 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
2671 which is the minimum version we support.
2675 * Certificate time validation (X509_cmp_time) enforces stricter
2676 compliance with RFC 5280. Fractional seconds and timezone offsets
2677 are no longer allowed.
2681 * Add support for ARIA
2685 * s_client will now send the Server Name Indication (SNI) extension by
2686 default unless the new "-noservername" option is used. The server name is
2687 based on the host provided to the "-connect" option unless overridden by
2688 using "-servername".
2692 * Add support for SipHash
2696 * OpenSSL now fails if it receives an unrecognised record type in TLS1.0
2697 or TLS1.1. Previously this only happened in SSLv3 and TLS1.2. This is to
2698 prevent issues where no progress is being made and the peer continually
2699 sends unrecognised record types, using up resources processing them.
2703 * 'openssl passwd' can now produce SHA256 and SHA512 based output,
2704 using the algorithm defined in
2705 <https://www.akkadia.org/drepper/SHA-crypt.txt>
2709 * Heartbeat support has been removed; the ABI is changed for now.
2711 *Richard Levitte, Rich Salz*
2713 * Support for SSL_OP_NO_ENCRYPT_THEN_MAC in SSL_CONF_cmd.
2717 * The RSA "null" method, which was partially supported to avoid patent
2718 issues, has been replaced to always returns NULL.
2725 ### Changes between 1.1.0k and 1.1.0l [10 Sep 2019]
2727 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
2728 used even when parsing explicit parameters, when loading a encoded key
2729 or calling `EC_GROUP_new_from_ecpkparameters()`/
2730 `EC_GROUP_new_from_ecparameters()`.
2731 This prevents bypass of security hardening and performance gains,
2732 especially for curves with specialized EC_METHODs.
2733 By default, if a key encoded with explicit parameters is loaded and later
2734 encoded, the output is still encoded with explicit parameters, even if
2735 internally a "named" EC_GROUP is used for computation.
2739 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
2740 this change, EC_GROUP_set_generator would accept order and/or cofactor as
2741 NULL. After this change, only the cofactor parameter can be NULL. It also
2742 does some minimal sanity checks on the passed order.
2747 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
2748 An attack is simple, if the first CMS_recipientInfo is valid but the
2749 second CMS_recipientInfo is chosen ciphertext. If the second
2750 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
2751 encryption key will be replaced by garbage, and the message cannot be
2752 decoded, but if the RSA decryption fails, the correct encryption key is
2753 used and the recipient will not notice the attack.
2754 As a work around for this potential attack the length of the decrypted
2755 key must be equal to the cipher default key length, in case the
2756 certifiate is not given and all recipientInfo are tried out.
2757 The old behaviour can be re-enabled in the CMS code by setting the
2758 CMS_DEBUG_DECRYPT flag.
2763 * Use Windows installation paths in the mingw builds
2765 Mingw isn't a POSIX environment per se, which means that Windows
2766 paths should be used for installation.
2771 ### Changes between 1.1.0j and 1.1.0k [28 May 2019]
2773 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
2774 This changes the size when using the `genpkey` command when no size is given.
2775 It fixes an omission in earlier changes that changed all RSA, DSA and DH
2776 generation commands to use 2048 bits by default.
2780 * Prevent over long nonces in ChaCha20-Poly1305.
2782 ChaCha20-Poly1305 is an AEAD cipher, and requires a unique nonce input
2783 for every encryption operation. RFC 7539 specifies that the nonce value
2784 (IV) should be 96 bits (12 bytes). OpenSSL allows a variable nonce length
2785 and front pads the nonce with 0 bytes if it is less than 12
2786 bytes. However it also incorrectly allows a nonce to be set of up to 16
2787 bytes. In this case only the last 12 bytes are significant and any
2788 additional leading bytes are ignored.
2790 It is a requirement of using this cipher that nonce values are
2791 unique. Messages encrypted using a reused nonce value are susceptible to
2792 serious confidentiality and integrity attacks. If an application changes
2793 the default nonce length to be longer than 12 bytes and then makes a
2794 change to the leading bytes of the nonce expecting the new value to be a
2795 new unique nonce then such an application could inadvertently encrypt
2796 messages with a reused nonce.
2798 Additionally the ignored bytes in a long nonce are not covered by the
2799 integrity guarantee of this cipher. Any application that relies on the
2800 integrity of these ignored leading bytes of a long nonce may be further
2801 affected. Any OpenSSL internal use of this cipher, including in SSL/TLS,
2802 is safe because no such use sets such a long nonce value. However user
2803 applications that use this cipher directly and set a non-default nonce
2804 length to be longer than 12 bytes may be vulnerable.
2806 This issue was reported to OpenSSL on 16th of March 2019 by Joran Dirk
2812 * Added SCA hardening for modular field inversion in EC_GROUP through
2813 a new dedicated field_inv() pointer in EC_METHOD.
2814 This also addresses a leakage affecting conversions from projective
2815 to affine coordinates.
2817 *Billy Bob Brumley, Nicola Tuveri*
2819 * Fix a use after free bug in d2i_X509_PUBKEY when overwriting a
2820 re-used X509_PUBKEY object if the second PUBKEY is malformed.
2824 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
2828 * Remove the 'dist' target and add a tarball building script. The
2829 'dist' target has fallen out of use, and it shouldn't be
2830 necessary to configure just to create a source distribution.
2834 ### Changes between 1.1.0i and 1.1.0j [20 Nov 2018]
2836 * Timing vulnerability in DSA signature generation
2838 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
2839 timing side channel attack. An attacker could use variations in the signing
2840 algorithm to recover the private key.
2842 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
2847 * Timing vulnerability in ECDSA signature generation
2849 The OpenSSL ECDSA signature algorithm has been shown to be vulnerable to a
2850 timing side channel attack. An attacker could use variations in the signing
2851 algorithm to recover the private key.
2853 This issue was reported to OpenSSL on 25th October 2018 by Samuel Weiser.
2858 * Add coordinate blinding for EC_POINT and implement projective
2859 coordinate blinding for generic prime curves as a countermeasure to
2860 chosen point SCA attacks.
2862 *Sohaib ul Hassan, Nicola Tuveri, Billy Bob Brumley*
2864 ### Changes between 1.1.0h and 1.1.0i [14 Aug 2018]
2866 * Client DoS due to large DH parameter
2868 During key agreement in a TLS handshake using a DH(E) based ciphersuite a
2869 malicious server can send a very large prime value to the client. This will
2870 cause the client to spend an unreasonably long period of time generating a
2871 key for this prime resulting in a hang until the client has finished. This
2872 could be exploited in a Denial Of Service attack.
2874 This issue was reported to OpenSSL on 5th June 2018 by Guido Vranken
2879 * Cache timing vulnerability in RSA Key Generation
2881 The OpenSSL RSA Key generation algorithm has been shown to be vulnerable to
2882 a cache timing side channel attack. An attacker with sufficient access to
2883 mount cache timing attacks during the RSA key generation process could
2884 recover the private key.
2886 This issue was reported to OpenSSL on 4th April 2018 by Alejandro Cabrera
2887 Aldaya, Billy Brumley, Cesar Pereida Garcia and Luis Manuel Alvarez Tapia.
2892 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
2893 parameter is no longer accepted, as it leads to a corrupt table. NULL
2894 pem_str is reserved for alias entries only.
2898 * Revert blinding in ECDSA sign and instead make problematic addition
2899 length-invariant. Switch even to fixed-length Montgomery multiplication.
2903 * Change generating and checking of primes so that the error rate of not
2904 being prime depends on the intended use based on the size of the input.
2905 For larger primes this will result in more rounds of Miller-Rabin.
2906 The maximal error rate for primes with more than 1080 bits is lowered
2909 *Kurt Roeckx, Annie Yousar*
2911 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
2915 * Add blinding to ECDSA and DSA signatures to protect against side channel
2916 attacks discovered by Keegan Ryan (NCC Group).
2920 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
2921 now allow empty (zero character) pass phrases.
2925 * Certificate time validation (X509_cmp_time) enforces stricter
2926 compliance with RFC 5280. Fractional seconds and timezone offsets
2927 are no longer allowed.
2931 * Fixed a text canonicalisation bug in CMS
2933 Where a CMS detached signature is used with text content the text goes
2934 through a canonicalisation process first prior to signing or verifying a
2935 signature. This process strips trailing space at the end of lines, converts
2936 line terminators to CRLF and removes additional trailing line terminators
2937 at the end of a file. A bug in the canonicalisation process meant that
2938 some characters, such as form-feed, were incorrectly treated as whitespace
2939 and removed. This is contrary to the specification (RFC5485). This fix
2940 could mean that detached text data signed with an earlier version of
2941 OpenSSL 1.1.0 may fail to verify using the fixed version, or text data
2942 signed with a fixed OpenSSL may fail to verify with an earlier version of
2943 OpenSSL 1.1.0. A workaround is to only verify the canonicalised text data
2944 and use the "-binary" flag (for the "cms" command line application) or set
2945 the SMIME_BINARY/PKCS7_BINARY/CMS_BINARY flags (if using CMS_verify()).
2949 ### Changes between 1.1.0g and 1.1.0h [27 Mar 2018]
2951 * Constructed ASN.1 types with a recursive definition could exceed the stack
2953 Constructed ASN.1 types with a recursive definition (such as can be found
2954 in PKCS7) could eventually exceed the stack given malicious input with
2955 excessive recursion. This could result in a Denial Of Service attack. There
2956 are no such structures used within SSL/TLS that come from untrusted sources
2957 so this is considered safe.
2959 This issue was reported to OpenSSL on 4th January 2018 by the OSS-fuzz
2965 * Incorrect CRYPTO_memcmp on HP-UX PA-RISC
2967 Because of an implementation bug the PA-RISC CRYPTO_memcmp function is
2968 effectively reduced to only comparing the least significant bit of each
2969 byte. This allows an attacker to forge messages that would be considered as
2970 authenticated in an amount of tries lower than that guaranteed by the
2971 security claims of the scheme. The module can only be compiled by the
2972 HP-UX assembler, so that only HP-UX PA-RISC targets are affected.
2974 This issue was reported to OpenSSL on 2nd March 2018 by Peter Waltenberg
2980 * Add a build target 'build_all_generated', to build all generated files
2981 and only that. This can be used to prepare everything that requires
2982 things like perl for a system that lacks perl and then move everything
2983 to that system and do the rest of the build there.
2987 * Backport SSL_OP_NO_RENGOTIATION
2989 OpenSSL 1.0.2 and below had the ability to disable renegotiation using the
2990 (undocumented) SSL3_FLAGS_NO_RENEGOTIATE_CIPHERS flag. Due to the opacity
2991 changes this is no longer possible in 1.1.0. Therefore the new
2992 SSL_OP_NO_RENEGOTIATION option from 1.1.1-dev has been backported to
2993 1.1.0 to provide equivalent functionality.
2995 Note that if an application built against 1.1.0h headers (or above) is run
2996 using an older version of 1.1.0 (prior to 1.1.0h) then the option will be
2997 accepted but nothing will happen, i.e. renegotiation will not be prevented.
3001 * Removed the OS390-Unix config target. It relied on a script that doesn't
3006 * rsaz_1024_mul_avx2 overflow bug on x86_64
3008 There is an overflow bug in the AVX2 Montgomery multiplication procedure
3009 used in exponentiation with 1024-bit moduli. No EC algorithms are affected.
3010 Analysis suggests that attacks against RSA and DSA as a result of this
3011 defect would be very difficult to perform and are not believed likely.
3012 Attacks against DH1024 are considered just feasible, because most of the
3013 work necessary to deduce information about a private key may be performed
3014 offline. The amount of resources required for such an attack would be
3015 significant. However, for an attack on TLS to be meaningful, the server
3016 would have to share the DH1024 private key among multiple clients, which is
3017 no longer an option since CVE-2016-0701.
3019 This only affects processors that support the AVX2 but not ADX extensions
3020 like Intel Haswell (4th generation).
3022 This issue was reported to OpenSSL by David Benjamin (Google). The issue
3023 was originally found via the OSS-Fuzz project.
3028 ### Changes between 1.1.0f and 1.1.0g [2 Nov 2017]
3030 * bn_sqrx8x_internal carry bug on x86_64
3032 There is a carry propagating bug in the x86_64 Montgomery squaring
3033 procedure. No EC algorithms are affected. Analysis suggests that attacks
3034 against RSA and DSA as a result of this defect would be very difficult to
3035 perform and are not believed likely. Attacks against DH are considered just
3036 feasible (although very difficult) because most of the work necessary to
3037 deduce information about a private key may be performed offline. The amount
3038 of resources required for such an attack would be very significant and
3039 likely only accessible to a limited number of attackers. An attacker would
3040 additionally need online access to an unpatched system using the target
3041 private key in a scenario with persistent DH parameters and a private
3042 key that is shared between multiple clients.
3044 This only affects processors that support the BMI1, BMI2 and ADX extensions
3045 like Intel Broadwell (5th generation) and later or AMD Ryzen.
3047 This issue was reported to OpenSSL by the OSS-Fuzz project.
3052 * Malformed X.509 IPAddressFamily could cause OOB read
3054 If an X.509 certificate has a malformed IPAddressFamily extension,
3055 OpenSSL could do a one-byte buffer overread. The most likely result
3056 would be an erroneous display of the certificate in text format.
3058 This issue was reported to OpenSSL by the OSS-Fuzz project.
3063 ### Changes between 1.1.0e and 1.1.0f [25 May 2017]
3065 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
3066 platform rather than 'mingw'.
3070 * Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
3071 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
3072 which is the minimum version we support.
3076 ### Changes between 1.1.0d and 1.1.0e [16 Feb 2017]
3078 * Encrypt-Then-Mac renegotiation crash
3080 During a renegotiation handshake if the Encrypt-Then-Mac extension is
3081 negotiated where it was not in the original handshake (or vice-versa) then
3082 this can cause OpenSSL to crash (dependant on ciphersuite). Both clients
3083 and servers are affected.
3085 This issue was reported to OpenSSL by Joe Orton (Red Hat).
3090 ### Changes between 1.1.0c and 1.1.0d [26 Jan 2017]
3092 * Truncated packet could crash via OOB read
3094 If one side of an SSL/TLS path is running on a 32-bit host and a specific
3095 cipher is being used, then a truncated packet can cause that host to
3096 perform an out-of-bounds read, usually resulting in a crash.
3098 This issue was reported to OpenSSL by Robert Święcki of Google.
3103 * Bad (EC)DHE parameters cause a client crash
3105 If a malicious server supplies bad parameters for a DHE or ECDHE key
3106 exchange then this can result in the client attempting to dereference a
3107 NULL pointer leading to a client crash. This could be exploited in a Denial
3110 This issue was reported to OpenSSL by Guido Vranken.
3115 * BN_mod_exp may produce incorrect results on x86_64
3117 There is a carry propagating bug in the x86_64 Montgomery squaring
3118 procedure. No EC algorithms are affected. Analysis suggests that attacks
3119 against RSA and DSA as a result of this defect would be very difficult to
3120 perform and are not believed likely. Attacks against DH are considered just
3121 feasible (although very difficult) because most of the work necessary to
3122 deduce information about a private key may be performed offline. The amount
3123 of resources required for such an attack would be very significant and
3124 likely only accessible to a limited number of attackers. An attacker would
3125 additionally need online access to an unpatched system using the target
3126 private key in a scenario with persistent DH parameters and a private
3127 key that is shared between multiple clients. For example this can occur by
3128 default in OpenSSL DHE based SSL/TLS ciphersuites. Note: This issue is very
3129 similar to CVE-2015-3193 but must be treated as a separate problem.
3131 This issue was reported to OpenSSL by the OSS-Fuzz project.
3136 ### Changes between 1.1.0b and 1.1.0c [10 Nov 2016]
3138 * ChaCha20/Poly1305 heap-buffer-overflow
3140 TLS connections using `*-CHACHA20-POLY1305` ciphersuites are susceptible to
3141 a DoS attack by corrupting larger payloads. This can result in an OpenSSL
3142 crash. This issue is not considered to be exploitable beyond a DoS.
3144 This issue was reported to OpenSSL by Robert Święcki (Google Security Team)
3149 * CMS Null dereference
3151 Applications parsing invalid CMS structures can crash with a NULL pointer
3152 dereference. This is caused by a bug in the handling of the ASN.1 CHOICE
3153 type in OpenSSL 1.1.0 which can result in a NULL value being passed to the
3154 structure callback if an attempt is made to free certain invalid encodings.
3155 Only CHOICE structures using a callback which do not handle NULL value are
3158 This issue was reported to OpenSSL by Tyler Nighswander of ForAllSecure.
3163 * Montgomery multiplication may produce incorrect results
3165 There is a carry propagating bug in the Broadwell-specific Montgomery
3166 multiplication procedure that handles input lengths divisible by, but
3167 longer than 256 bits. Analysis suggests that attacks against RSA, DSA
3168 and DH private keys are impossible. This is because the subroutine in
3169 question is not used in operations with the private key itself and an input
3170 of the attacker's direct choice. Otherwise the bug can manifest itself as
3171 transient authentication and key negotiation failures or reproducible
3172 erroneous outcome of public-key operations with specially crafted input.
3173 Among EC algorithms only Brainpool P-512 curves are affected and one
3174 presumably can attack ECDH key negotiation. Impact was not analyzed in
3175 detail, because pre-requisites for attack are considered unlikely. Namely
3176 multiple clients have to choose the curve in question and the server has to
3177 share the private key among them, neither of which is default behaviour.
3178 Even then only clients that chose the curve will be affected.
3180 This issue was publicly reported as transient failures and was not
3181 initially recognized as a security issue. Thanks to Richard Morgan for
3182 providing reproducible case.
3187 * Removed automatic addition of RPATH in shared libraries and executables,
3188 as this was a remainder from OpenSSL 1.0.x and isn't needed any more.
3192 ### Changes between 1.1.0a and 1.1.0b [26 Sep 2016]
3194 * Fix Use After Free for large message sizes
3196 The patch applied to address CVE-2016-6307 resulted in an issue where if a
3197 message larger than approx 16k is received then the underlying buffer to
3198 store the incoming message is reallocated and moved. Unfortunately a
3199 dangling pointer to the old location is left which results in an attempt to
3200 write to the previously freed location. This is likely to result in a
3201 crash, however it could potentially lead to execution of arbitrary code.
3203 This issue only affects OpenSSL 1.1.0a.
3205 This issue was reported to OpenSSL by Robert Święcki.
3210 ### Changes between 1.1.0 and 1.1.0a [22 Sep 2016]
3212 * OCSP Status Request extension unbounded memory growth
3214 A malicious client can send an excessively large OCSP Status Request
3215 extension. If that client continually requests renegotiation, sending a
3216 large OCSP Status Request extension each time, then there will be unbounded
3217 memory growth on the server. This will eventually lead to a Denial Of
3218 Service attack through memory exhaustion. Servers with a default
3219 configuration are vulnerable even if they do not support OCSP. Builds using
3220 the "no-ocsp" build time option are not affected.
3222 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
3227 * SSL_peek() hang on empty record
3229 OpenSSL 1.1.0 SSL/TLS will hang during a call to SSL_peek() if the peer
3230 sends an empty record. This could be exploited by a malicious peer in a
3231 Denial Of Service attack.
3233 This issue was reported to OpenSSL by Alex Gaynor.
3238 * Excessive allocation of memory in tls_get_message_header() and
3239 dtls1_preprocess_fragment()
3241 A (D)TLS message includes 3 bytes for its length in the header for the
3242 message. This would allow for messages up to 16Mb in length. Messages of
3243 this length are excessive and OpenSSL includes a check to ensure that a
3244 peer is sending reasonably sized messages in order to avoid too much memory
3245 being consumed to service a connection. A flaw in the logic of version
3246 1.1.0 means that memory for the message is allocated too early, prior to
3247 the excessive message length check. Due to way memory is allocated in
3248 OpenSSL this could mean an attacker could force up to 21Mb to be allocated
3249 to service a connection. This could lead to a Denial of Service through
3250 memory exhaustion. However, the excessive message length check still takes
3251 place, and this would cause the connection to immediately fail. Assuming
3252 that the application calls SSL_free() on the failed connection in a timely
3253 manner then the 21Mb of allocated memory will then be immediately freed
3254 again. Therefore the excessive memory allocation will be transitory in
3255 nature. This then means that there is only a security impact if:
3257 1) The application does not call SSL_free() in a timely manner in the event
3258 that the connection fails
3260 2) The application is working in a constrained environment where there is
3261 very little free memory
3263 3) The attacker initiates multiple connection attempts such that there are
3264 multiple connections in a state where memory has been allocated for the
3265 connection; SSL_free() has not yet been called; and there is insufficient
3266 memory to service the multiple requests.
3268 Except in the instance of (1) above any Denial Of Service is likely to be
3269 transitory because as soon as the connection fails the memory is
3270 subsequently freed again in the SSL_free() call. However there is an
3271 increased risk during this period of application crashes due to the lack of
3272 memory - which would then mean a more serious Denial of Service.
3274 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
3275 (CVE-2016-6307 and CVE-2016-6308)
3279 * solaris-x86-cc, i.e. 32-bit configuration with vendor compiler,
3280 had to be removed. Primary reason is that vendor assembler can't
3281 assemble our modules with -KPIC flag. As result it, assembly
3282 support, was not even available as option. But its lack means
3283 lack of side-channel resistant code, which is incompatible with
3284 security by todays standards. Fortunately gcc is readily available
3285 prepackaged option, which we firmly point at...
3289 ### Changes between 1.0.2h and 1.1.0 [25 Aug 2016]
3291 * Windows command-line tool supports UTF-8 opt-in option for arguments
3292 and console input. Setting OPENSSL_WIN32_UTF8 environment variable
3293 (to any value) allows Windows user to access PKCS#12 file generated
3294 with Windows CryptoAPI and protected with non-ASCII password, as well
3295 as files generated under UTF-8 locale on Linux also protected with
3300 * To mitigate the SWEET32 attack ([CVE-2016-2183]), 3DES cipher suites
3301 have been disabled by default and removed from DEFAULT, just like RC4.
3302 See the RC4 item below to re-enable both.
3306 * The method for finding the storage location for the Windows RAND seed file
3307 has changed. First we check %RANDFILE%. If that is not set then we check
3308 the directories %HOME%, %USERPROFILE% and %SYSTEMROOT% in that order. If
3309 all else fails we fall back to C:\.
3313 * The EVP_EncryptUpdate() function has had its return type changed from void
3314 to int. A return of 0 indicates and error while a return of 1 indicates
3319 * The flags RSA_FLAG_NO_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME and
3320 DH_FLAG_NO_EXP_CONSTTIME which previously provided the ability to switch
3321 off the constant time implementation for RSA, DSA and DH have been made
3322 no-ops and deprecated.
3326 * Windows RAND implementation was simplified to only get entropy by
3327 calling CryptGenRandom(). Various other RAND-related tickets
3330 *Joseph Wylie Yandle, Rich Salz*
3332 * The stack and lhash API's were renamed to start with `OPENSSL_SK_`
3333 and `OPENSSL_LH_`, respectively. The old names are available
3334 with API compatibility. They new names are now completely documented.
3338 * Unify TYPE_up_ref(obj) methods signature.
3339 SSL_CTX_up_ref(), SSL_up_ref(), X509_up_ref(), EVP_PKEY_up_ref(),
3340 X509_CRL_up_ref(), X509_OBJECT_up_ref_count() methods are now returning an
3341 int (instead of void) like all others TYPE_up_ref() methods.
3342 So now these methods also check the return value of CRYPTO_atomic_add(),
3343 and the validity of object reference counter.
3345 *fdasilvayy@gmail.com*
3347 * With Windows Visual Studio builds, the .pdb files are installed
3348 alongside the installed libraries and executables. For a static
3349 library installation, ossl_static.pdb is the associate compiler
3350 generated .pdb file to be used when linking programs.
3354 * Remove openssl.spec. Packaging files belong with the packagers.
3358 * Automatic Darwin/OSX configuration has had a refresh, it will now
3359 recognise x86_64 architectures automatically. You can still decide
3360 to build for a different bitness with the environment variable
3361 KERNEL_BITS (can be 32 or 64), for example:
3363 KERNEL_BITS=32 ./config
3367 * Change default algorithms in pkcs8 utility to use PKCS#5 v2.0,
3368 256 bit AES and HMAC with SHA256.
3372 * Remove support for MIPS o32 ABI on IRIX (and IRIX only).
3376 * Triple-DES ciphers have been moved from HIGH to MEDIUM.
3380 * To enable users to have their own config files and build file templates,
3381 Configure looks in the directory indicated by the environment variable
3382 OPENSSL_LOCAL_CONFIG_DIR as well as the in-source Configurations/
3383 directory. On VMS, OPENSSL_LOCAL_CONFIG_DIR is expected to be a logical
3384 name and is used as is.
3388 * The following datatypes were made opaque: X509_OBJECT, X509_STORE_CTX,
3389 X509_STORE, X509_LOOKUP, and X509_LOOKUP_METHOD. The unused type
3390 X509_CERT_FILE_CTX was removed.
3394 * "shared" builds are now the default. To create only static libraries use
3395 the "no-shared" Configure option.
3399 * Remove the no-aes, no-hmac, no-rsa, no-sha and no-md5 Configure options.
3400 All of these option have not worked for some while and are fundamental
3405 * Make various cleanup routines no-ops and mark them as deprecated. Most
3406 global cleanup functions are no longer required because they are handled
3407 via auto-deinit (see OPENSSL_init_crypto and OPENSSL_init_ssl man pages).
3408 Explicitly de-initing can cause problems (e.g. where a library that uses
3409 OpenSSL de-inits, but an application is still using it). The affected
3410 functions are CONF_modules_free(), ENGINE_cleanup(), OBJ_cleanup(),
3411 EVP_cleanup(), BIO_sock_cleanup(), CRYPTO_cleanup_all_ex_data(),
3412 RAND_cleanup(), SSL_COMP_free_compression_methods(), ERR_free_strings() and
3413 COMP_zlib_cleanup().
3417 * --strict-warnings no longer enables runtime debugging options
3418 such as REF_DEBUG. Instead, debug options are automatically
3419 enabled with '--debug' builds.
3421 *Andy Polyakov, Emilia Käsper*
3423 * Made DH and DH_METHOD opaque. The structures for managing DH objects
3424 have been moved out of the public header files. New functions for managing
3425 these have been added.
3429 * Made RSA and RSA_METHOD opaque. The structures for managing RSA
3430 objects have been moved out of the public header files. New
3431 functions for managing these have been added.
3435 * Made DSA and DSA_METHOD opaque. The structures for managing DSA objects
3436 have been moved out of the public header files. New functions for managing
3437 these have been added.
3441 * Made BIO and BIO_METHOD opaque. The structures for managing BIOs have been
3442 moved out of the public header files. New functions for managing these
3447 * Removed no-rijndael as a config option. Rijndael is an old name for AES.
3451 * Removed the mk1mf build scripts.
3455 * Headers are now wrapped, if necessary, with OPENSSL_NO_xxx, so
3456 it is always safe to #include a header now.
3460 * Removed the aged BC-32 config and all its supporting scripts
3464 * Removed support for Ultrix, Netware, and OS/2.
3468 * Add support for HKDF.
3470 *Alessandro Ghedini*
3472 * Add support for blake2b and blake2s
3476 * Added support for "pipelining". Ciphers that have the
3477 EVP_CIPH_FLAG_PIPELINE flag set have a capability to process multiple
3478 encryptions/decryptions simultaneously. There are currently no built-in
3479 ciphers with this property but the expectation is that engines will be able
3480 to offer it to significantly improve throughput. Support has been extended
3481 into libssl so that multiple records for a single connection can be
3482 processed in one go (for >=TLS 1.1).
3486 * Added the AFALG engine. This is an async capable engine which is able to
3487 offload work to the Linux kernel. In this initial version it only supports
3488 AES128-CBC. The kernel must be version 4.1.0 or greater.
3492 * OpenSSL now uses a new threading API. It is no longer necessary to
3493 set locking callbacks to use OpenSSL in a multi-threaded environment. There
3494 are two supported threading models: pthreads and windows threads. It is
3495 also possible to configure OpenSSL at compile time for "no-threads". The
3496 old threading API should no longer be used. The functions have been
3497 replaced with "no-op" compatibility macros.
3499 *Alessandro Ghedini, Matt Caswell*
3501 * Modify behavior of ALPN to invoke callback after SNI/servername
3502 callback, such that updates to the SSL_CTX affect ALPN.
3506 * Add SSL_CIPHER queries for authentication and key-exchange.
3510 * Changes to the DEFAULT cipherlist:
3511 - Prefer (EC)DHE handshakes over plain RSA.
3512 - Prefer AEAD ciphers over legacy ciphers.
3513 - Prefer ECDSA over RSA when both certificates are available.
3514 - Prefer TLSv1.2 ciphers/PRF.
3515 - Remove DSS, SEED, IDEA, CAMELLIA, and AES-CCM from the
3520 * Change the ECC default curve list to be this, in order: x25519,
3521 secp256r1, secp521r1, secp384r1.
3525 * RC4 based libssl ciphersuites are now classed as "weak" ciphers and are
3526 disabled by default. They can be re-enabled using the
3527 enable-weak-ssl-ciphers option to Configure.
3531 * If the server has ALPN configured, but supports no protocols that the
3532 client advertises, send a fatal "no_application_protocol" alert.
3533 This behaviour is SHALL in RFC 7301, though it isn't universally
3534 implemented by other servers.
3538 * Add X25519 support.
3539 Add ASN.1 and EVP_PKEY methods for X25519. This includes support
3540 for public and private key encoding using the format documented in
3541 draft-ietf-curdle-pkix-02. The corresponding EVP_PKEY method supports
3542 key generation and key derivation.
3544 TLS support complies with draft-ietf-tls-rfc4492bis-08 and uses
3549 * Deprecate SRP_VBASE_get_by_user.
3550 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
3551 In order to fix an unavoidable memory leak ([CVE-2016-0798]),
3552 SRP_VBASE_get_by_user was changed to ignore the "fake user" SRP
3553 seed, even if the seed is configured.
3555 Users should use SRP_VBASE_get1_by_user instead. Note that in
3556 SRP_VBASE_get1_by_user, caller must free the returned value. Note
3557 also that even though configuring the SRP seed attempts to hide
3558 invalid usernames by continuing the handshake with fake
3559 credentials, this behaviour is not constant time and no strong
3560 guarantees are made that the handshake is indistinguishable from
3561 that of a valid user.
3565 * Configuration change; it's now possible to build dynamic engines
3566 without having to build shared libraries and vice versa. This
3567 only applies to the engines in `engines/`, those in `crypto/engine/`
3568 will always be built into libcrypto (i.e. "static").
3570 Building dynamic engines is enabled by default; to disable, use
3571 the configuration option "disable-dynamic-engine".
3573 The only requirements for building dynamic engines are the
3574 presence of the DSO module and building with position independent
3575 code, so they will also automatically be disabled if configuring
3576 with "disable-dso" or "disable-pic".
3578 The macros OPENSSL_NO_STATIC_ENGINE and OPENSSL_NO_DYNAMIC_ENGINE
3579 are also taken away from openssl/opensslconf.h, as they are
3584 * Configuration change; if there is a known flag to compile
3585 position independent code, it will always be applied on the
3586 libcrypto and libssl object files, and never on the application
3587 object files. This means other libraries that use routines from
3588 libcrypto / libssl can be made into shared libraries regardless
3589 of how OpenSSL was configured.
3591 If this isn't desirable, the configuration options "disable-pic"
3592 or "no-pic" can be used to disable the use of PIC. This will
3593 also disable building shared libraries and dynamic engines.
3597 * Removed JPAKE code. It was experimental and has no wide use.
3601 * The INSTALL_PREFIX Makefile variable has been renamed to
3602 DESTDIR. That makes for less confusion on what this variable
3603 is for. Also, the configuration option --install_prefix is
3608 * Heartbeat for TLS has been removed and is disabled by default
3609 for DTLS; configure with enable-heartbeats. Code that uses the
3610 old #define's might need to be updated.
3612 *Emilia Käsper, Rich Salz*
3614 * Rename REF_CHECK to REF_DEBUG.
3618 * New "unified" build system
3620 The "unified" build system is aimed to be a common system for all
3621 platforms we support. With it comes new support for VMS.
3623 This system builds supports building in a different directory tree
3624 than the source tree. It produces one Makefile (for unix family
3625 or lookalikes), or one descrip.mms (for VMS).
3627 The source of information to make the Makefile / descrip.mms is
3628 small files called 'build.info', holding the necessary
3629 information for each directory with source to compile, and a
3630 template in Configurations, like unix-Makefile.tmpl or
3633 With this change, the library names were also renamed on Windows
3634 and on VMS. They now have names that are closer to the standard
3635 on Unix, and include the major version number, and in certain
3636 cases, the architecture they are built for. See "Notes on shared
3637 libraries" in INSTALL.
3639 We rely heavily on the perl module Text::Template.
3643 * Added support for auto-initialisation and de-initialisation of the library.
3644 OpenSSL no longer requires explicit init or deinit routines to be called,
3645 except in certain circumstances. See the OPENSSL_init_crypto() and
3646 OPENSSL_init_ssl() man pages for further information.
3650 * The arguments to the DTLSv1_listen function have changed. Specifically the
3651 "peer" argument is now expected to be a BIO_ADDR object.
3653 * Rewrite of BIO networking library. The BIO library lacked consistent
3654 support of IPv6, and adding it required some more extensive
3655 modifications. This introduces the BIO_ADDR and BIO_ADDRINFO types,
3656 which hold all types of addresses and chains of address information.
3657 It also introduces a new API, with functions like BIO_socket,
3658 BIO_connect, BIO_listen, BIO_lookup and a rewrite of BIO_accept.
3659 The source/sink BIOs BIO_s_connect, BIO_s_accept and BIO_s_datagram
3660 have been adapted accordingly.
3664 * RSA_padding_check_PKCS1_type_1 now accepts inputs with and without
3669 * CRIME protection: disable compression by default, even if OpenSSL is
3670 compiled with zlib enabled. Applications can still enable compression
3671 by calling SSL_CTX_clear_options(ctx, SSL_OP_NO_COMPRESSION), or by
3672 using the SSL_CONF library to configure compression.
3676 * The signature of the session callback configured with
3677 SSL_CTX_sess_set_get_cb was changed. The read-only input buffer
3678 was explicitly marked as `const unsigned char*` instead of
3683 * Always DPURIFY. Remove the use of uninitialized memory in the
3684 RNG, and other conditional uses of DPURIFY. This makes -DPURIFY a no-op.
3688 * Removed many obsolete configuration items, including
3689 DES_PTR, DES_RISC1, DES_RISC2, DES_INT
3690 MD2_CHAR, MD2_INT, MD2_LONG
3692 IDEA_SHORT, IDEA_LONG
3693 RC2_SHORT, RC2_LONG, RC4_LONG, RC4_CHUNK, RC4_INDEX
3695 *Rich Salz, with advice from Andy Polyakov*
3697 * Many BN internals have been moved to an internal header file.
3699 *Rich Salz with help from Andy Polyakov*
3701 * Configuration and writing out the results from it has changed.
3702 Files such as Makefile include/openssl/opensslconf.h and are now
3703 produced through general templates, such as Makefile.in and
3704 crypto/opensslconf.h.in and some help from the perl module
3707 Also, the center of configuration information is no longer
3708 Makefile. Instead, Configure produces a perl module in
3709 configdata.pm which holds most of the config data (in the hash
3710 table %config), the target data that comes from the target
3711 configuration in one of the `Configurations/*.conf` files (in
3716 * To clarify their intended purposes, the Configure options
3717 --prefix and --openssldir change their semantics, and become more
3718 straightforward and less interdependent.
3720 --prefix shall be used exclusively to give the location INSTALLTOP
3721 where programs, scripts, libraries, include files and manuals are
3722 going to be installed. The default is now /usr/local.
3724 --openssldir shall be used exclusively to give the default
3725 location OPENSSLDIR where certificates, private keys, CRLs are
3726 managed. This is also where the default openssl.cnf gets
3728 If the directory given with this option is a relative path, the
3729 values of both the --prefix value and the --openssldir value will
3730 be combined to become OPENSSLDIR.
3731 The default for --openssldir is INSTALLTOP/ssl.
3733 Anyone who uses --openssldir to specify where OpenSSL is to be
3734 installed MUST change to use --prefix instead.
3738 * The GOST engine was out of date and therefore it has been removed. An up
3739 to date GOST engine is now being maintained in an external repository.
3740 See: <https://wiki.openssl.org/index.php/Binaries>. Libssl still retains
3741 support for GOST ciphersuites (these are only activated if a GOST engine
3746 * EGD is no longer supported by default; use enable-egd when
3749 *Ben Kaduk and Rich Salz*
3751 * The distribution now has Makefile.in files, which are used to
3752 create Makefile's when Configure is run. *Configure must be run
3753 before trying to build now.*
3757 * The return value for SSL_CIPHER_description() for error conditions
3762 * Support for RFC6698/RFC7671 DANE TLSA peer authentication.
3764 Obtaining and performing DNSSEC validation of TLSA records is
3765 the application's responsibility. The application provides
3766 the TLSA records of its choice to OpenSSL, and these are then
3767 used to authenticate the peer.
3769 The TLSA records need not even come from DNS. They can, for
3770 example, be used to implement local end-entity certificate or
3771 trust-anchor "pinning", where the "pin" data takes the form
3772 of TLSA records, which can augment or replace verification
3773 based on the usual WebPKI public certification authorities.
3777 * Revert default OPENSSL_NO_DEPRECATED setting. Instead OpenSSL
3778 continues to support deprecated interfaces in default builds.
3779 However, applications are strongly advised to compile their
3780 source files with -DOPENSSL_API_COMPAT=0x10100000L, which hides
3781 the declarations of all interfaces deprecated in 0.9.8, 1.0.0
3782 or the 1.1.0 releases.
3784 In environments in which all applications have been ported to
3785 not use any deprecated interfaces OpenSSL's Configure script
3786 should be used with the --api=1.1.0 option to entirely remove
3787 support for the deprecated features from the library and
3788 unconditionally disable them in the installed headers.
3789 Essentially the same effect can be achieved with the "no-deprecated"
3790 argument to Configure, except that this will always restrict
3791 the build to just the latest API, rather than a fixed API
3794 As applications are ported to future revisions of the API,
3795 they should update their compile-time OPENSSL_API_COMPAT define
3796 accordingly, but in most cases should be able to continue to
3797 compile with later releases.
3799 The OPENSSL_API_COMPAT versions for 1.0.0, and 0.9.8 are
3800 0x10000000L and 0x00908000L, respectively. However those
3801 versions did not support the OPENSSL_API_COMPAT feature, and
3802 so applications are not typically tested for explicit support
3803 of just the undeprecated features of either release.
3807 * Add support for setting the minimum and maximum supported protocol.
3808 It can bet set via the SSL_set_min_proto_version() and
3809 SSL_set_max_proto_version(), or via the SSL_CONF's MinProtocol and
3810 MaxProtocol. It's recommended to use the new APIs to disable
3811 protocols instead of disabling individual protocols using
3812 SSL_set_options() or SSL_CONF's Protocol. This change also
3813 removes support for disabling TLS 1.2 in the OpenSSL TLS
3814 client at compile time by defining OPENSSL_NO_TLS1_2_CLIENT.
3818 * Support for ChaCha20 and Poly1305 added to libcrypto and libssl.
3822 * New EC_KEY_METHOD, this replaces the older ECDSA_METHOD and ECDH_METHOD
3823 and integrates ECDSA and ECDH functionality into EC. Implementations can
3824 now redirect key generation and no longer need to convert to or from
3827 Note: the ecdsa.h and ecdh.h headers are now no longer needed and just
3828 include the ec.h header file instead.
3832 * Remove support for all 40 and 56 bit ciphers. This includes all the export
3833 ciphers who are no longer supported and drops support the ephemeral RSA key
3834 exchange. The LOW ciphers currently doesn't have any ciphers in it.
3838 * Made EVP_MD_CTX, EVP_MD, EVP_CIPHER_CTX, EVP_CIPHER and HMAC_CTX
3839 opaque. For HMAC_CTX, the following constructors and destructors
3842 HMAC_CTX *HMAC_CTX_new(void);
3843 void HMAC_CTX_free(HMAC_CTX *ctx);
3845 For EVP_MD and EVP_CIPHER, complete APIs to create, fill and
3846 destroy such methods has been added. See EVP_MD_meth_new(3) and
3847 EVP_CIPHER_meth_new(3) for documentation.
3850 1) `EVP_MD_CTX_cleanup()`, `EVP_CIPHER_CTX_cleanup()` and
3851 `HMAC_CTX_cleanup()` were removed. `HMAC_CTX_reset()` and
3852 `EVP_MD_CTX_reset()` should be called instead to reinitialise
3853 an already created structure.
3854 2) For consistency with the majority of our object creators and
3855 destructors, `EVP_MD_CTX_(create|destroy)` were renamed to
3856 `EVP_MD_CTX_(new|free)`. The old names are retained as macros
3857 for deprecated builds.
3861 * Added ASYNC support. Libcrypto now includes the async sub-library to enable
3862 cryptographic operations to be performed asynchronously as long as an
3863 asynchronous capable engine is used. See the ASYNC_start_job() man page for
3864 further details. Libssl has also had this capability integrated with the
3865 introduction of the new mode SSL_MODE_ASYNC and associated error
3866 SSL_ERROR_WANT_ASYNC. See the SSL_CTX_set_mode() and SSL_get_error() man
3867 pages. This work was developed in partnership with Intel Corp.
3871 * SSL_{CTX_}set_ecdh_auto() has been removed and ECDH is support is
3872 always enabled now. If you want to disable the support you should
3873 exclude it using the list of supported ciphers. This also means that the
3874 "-no_ecdhe" option has been removed from s_server.
3878 * SSL_{CTX}_set_tmp_ecdh() which can set 1 EC curve now internally calls
3879 SSL_{CTX_}set1_curves() which can set a list.
3883 * Remove support for SSL_{CTX_}set_tmp_ecdh_callback(). You should set the
3884 curve you want to support using SSL_{CTX_}set1_curves().
3888 * State machine rewrite. The state machine code has been significantly
3889 refactored in order to remove much duplication of code and solve issues
3890 with the old code (see [ssl/statem/README.md](ssl/statem/README.md) for
3891 further details). This change does have some associated API changes.
3892 Notably the SSL_state() function has been removed and replaced by
3893 SSL_get_state which now returns an "OSSL_HANDSHAKE_STATE" instead of an int.
3894 SSL_set_state() has been removed altogether. The previous handshake states
3895 defined in ssl.h and ssl3.h have also been removed.
3899 * All instances of the string "ssleay" in the public API were replaced
3900 with OpenSSL (case-matching; e.g., OPENSSL_VERSION for #define's)
3901 Some error codes related to internal RSA_eay API's were renamed.
3905 * The demo files in crypto/threads were moved to demo/threads.
3909 * Removed obsolete engines: 4758cca, aep, atalla, cswift, nuron, gmp,
3912 *Matt Caswell, Rich Salz*
3914 * New ASN.1 embed macro.
3916 New ASN.1 macro ASN1_EMBED. This is the same as ASN1_SIMPLE except the
3917 structure is not allocated: it is part of the parent. That is instead of
3925 This reduces memory fragmentation and make it impossible to accidentally
3926 set a mandatory field to NULL.
3928 This currently only works for some fields specifically a SEQUENCE, CHOICE,
3929 or ASN1_STRING type which is part of a parent SEQUENCE. Since it is
3930 equivalent to ASN1_SIMPLE it cannot be tagged, OPTIONAL, SET OF or
3935 * Remove EVP_CHECK_DES_KEY, a compile-time option that never compiled.
3939 * Removed DES and RC4 ciphersuites from DEFAULT. Also removed RC2 although
3940 in 1.0.2 EXPORT was already removed and the only RC2 ciphersuite is also
3941 an EXPORT one. COMPLEMENTOFDEFAULT has been updated accordingly to add
3942 DES and RC4 ciphersuites.
3946 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
3947 This changes the decoding behaviour for some invalid messages,
3948 though the change is mostly in the more lenient direction, and
3949 legacy behaviour is preserved as much as possible.
3953 * Fix no-stdio build.
3954 *David Woodhouse <David.Woodhouse@intel.com> and also*
3955 *Ivan Nestlerode <ivan.nestlerode@sonos.com>*
3957 * New testing framework
3958 The testing framework has been largely rewritten and is now using
3959 perl and the perl modules Test::Harness and an extended variant of
3960 Test::More called OpenSSL::Test to do its work. All test scripts in
3961 test/ have been rewritten into test recipes, and all direct calls to
3962 executables in test/Makefile have become individual recipes using the
3963 simplified testing OpenSSL::Test::Simple.
3965 For documentation on our testing modules, do:
3967 perldoc test/testlib/OpenSSL/Test/Simple.pm
3968 perldoc test/testlib/OpenSSL/Test.pm
3972 * Revamped memory debug; only -DCRYPTO_MDEBUG and -DCRYPTO_MDEBUG_ABORT
3973 are used; the latter aborts on memory leaks (usually checked on exit).
3974 Some undocumented "set malloc, etc., hooks" functions were removed
3975 and others were changed. All are now documented.
3979 * In DSA_generate_parameters_ex, if the provided seed is too short,
3982 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
3984 * Rewrite PSK to support ECDHE_PSK, DHE_PSK and RSA_PSK. Add ciphersuites
3985 from RFC4279, RFC4785, RFC5487, RFC5489.
3987 Thanks to Christian J. Dietrich and Giuseppe D'Angelo for the
3988 original RSA_PSK patch.
3992 * Dropped support for the SSL3_FLAGS_DELAY_CLIENT_FINISHED flag. This SSLeay
3993 era flag was never set throughout the codebase (only read). Also removed
3994 SSL3_FLAGS_POP_BUFFER which was only used if
3995 SSL3_FLAGS_DELAY_CLIENT_FINISHED was also set.
3999 * Changed the default name options in the "ca", "crl", "req" and "x509"
4000 to be "oneline" instead of "compat".
4004 * Remove SSL_OP_TLS_BLOCK_PADDING_BUG. This is SSLeay legacy, we're
4005 not aware of clients that still exhibit this bug, and the workaround
4006 hasn't been working properly for a while.
4010 * The return type of BIO_number_read() and BIO_number_written() as well as
4011 the corresponding num_read and num_write members in the BIO structure has
4012 changed from unsigned long to uint64_t. On platforms where an unsigned
4013 long is 32 bits (e.g. Windows) these counters could overflow if >4Gb is
4018 * Given the pervasive nature of TLS extensions it is inadvisable to run
4019 OpenSSL without support for them. It also means that maintaining
4020 the OPENSSL_NO_TLSEXT option within the code is very invasive (and probably
4021 not well tested). Therefore the OPENSSL_NO_TLSEXT option has been removed.
4025 * Removed support for the two export grade static DH ciphersuites
4026 EXP-DH-RSA-DES-CBC-SHA and EXP-DH-DSS-DES-CBC-SHA. These two ciphersuites
4027 were newly added (along with a number of other static DH ciphersuites) to
4028 1.0.2. However the two export ones have *never* worked since they were
4029 introduced. It seems strange in any case to be adding new export
4030 ciphersuites, and given "logjam" it also does not seem correct to fix them.
4034 * Version negotiation has been rewritten. In particular SSLv23_method(),
4035 SSLv23_client_method() and SSLv23_server_method() have been deprecated,
4036 and turned into macros which simply call the new preferred function names
4037 TLS_method(), TLS_client_method() and TLS_server_method(). All new code
4038 should use the new names instead. Also as part of this change the ssl23.h
4039 header file has been removed.
4043 * Support for Kerberos ciphersuites in TLS (RFC2712) has been removed. This
4044 code and the associated standard is no longer considered fit-for-purpose.
4048 * RT2547 was closed. When generating a private key, try to make the
4049 output file readable only by the owner. This behavior change might
4050 be noticeable when interacting with other software.
4052 * Documented all exdata functions. Added CRYPTO_free_ex_index.
4057 * Added HTTP GET support to the ocsp command.
4061 * Changed default digest for the dgst and enc commands from MD5 to
4066 * RAND_pseudo_bytes has been deprecated. Users should use RAND_bytes instead.
4070 * Added support for TLS extended master secret from
4071 draft-ietf-tls-session-hash-03.txt. Thanks for Alfredo Pironti for an
4072 initial patch which was a great help during development.
4076 * All libssl internal structures have been removed from the public header
4077 files, and the OPENSSL_NO_SSL_INTERN option has been removed (since it is
4078 now redundant). Users should not attempt to access internal structures
4079 directly. Instead they should use the provided API functions.
4083 * config has been changed so that by default OPENSSL_NO_DEPRECATED is used.
4084 Access to deprecated functions can be re-enabled by running config with
4085 "enable-deprecated". In addition applications wishing to use deprecated
4086 functions must define OPENSSL_USE_DEPRECATED. Note that this new behaviour
4087 will, by default, disable some transitive includes that previously existed
4088 in the header files (e.g. ec.h will no longer, by default, include bn.h)
4092 * Added support for OCB mode. OpenSSL has been granted a patent license
4093 compatible with the OpenSSL license for use of OCB. Details are available
4094 at <https://www.openssl.org/source/OCB-patent-grant-OpenSSL.pdf>. Support
4095 for OCB can be removed by calling config with no-ocb.
4099 * SSLv2 support has been removed. It still supports receiving a SSLv2
4100 compatible client hello.
4104 * Increased the minimal RSA keysize from 256 to 512 bits [Rich Salz],
4105 done while fixing the error code for the key-too-small case.
4107 *Annie Yousar <a.yousar@informatik.hu-berlin.de>*
4109 * CA.sh has been removed; use CA.pl instead.
4113 * Removed old DES API.
4117 * Remove various unsupported platforms:
4123 Sinix/ReliantUNIX RM400
4128 16-bit platforms such as WIN16
4132 * Clean up OPENSSL_NO_xxx #define's
4133 - Use setbuf() and remove OPENSSL_NO_SETVBUF_IONBF
4134 - Rename OPENSSL_SYSNAME_xxx to OPENSSL_SYS_xxx
4135 - OPENSSL_NO_EC{DH,DSA} merged into OPENSSL_NO_EC
4136 - OPENSSL_NO_RIPEMD160, OPENSSL_NO_RIPEMD merged into OPENSSL_NO_RMD160
4137 - OPENSSL_NO_FP_API merged into OPENSSL_NO_STDIO
4138 - Remove OPENSSL_NO_BIO OPENSSL_NO_BUFFER OPENSSL_NO_CHAIN_VERIFY
4139 OPENSSL_NO_EVP OPENSSL_NO_FIPS_ERR OPENSSL_NO_HASH_COMP
4140 OPENSSL_NO_LHASH OPENSSL_NO_OBJECT OPENSSL_NO_SPEED OPENSSL_NO_STACK
4141 OPENSSL_NO_X509 OPENSSL_NO_X509_VERIFY
4142 - Remove MS_STATIC; it's a relic from platforms <32 bits.
4146 * Cleaned up dead code
4147 Remove all but one '#ifdef undef' which is to be looked at.
4151 * Clean up calling of xxx_free routines.
4152 Just like free(), fix most of the xxx_free routines to accept
4153 NULL. Remove the non-null checks from callers. Save much code.
4157 * Add secure heap for storage of private keys (when possible).
4158 Add BIO_s_secmem(), CBIGNUM, etc.
4159 Contributed by Akamai Technologies under our Corporate CLA.
4163 * Experimental support for a new, fast, unbiased prime candidate generator,
4164 bn_probable_prime_dh_coprime(). Not currently used by any prime generator.
4166 *Felix Laurie von Massenbach <felix@erbridge.co.uk>*
4168 * New output format NSS in the sess_id command line tool. This allows
4169 exporting the session id and the master key in NSS keylog format.
4171 *Martin Kaiser <martin@kaiser.cx>*
4173 * Harmonize version and its documentation. -f flag is used to display
4176 *mancha <mancha1@zoho.com>*
4178 * Fix eckey_priv_encode so it immediately returns an error upon a failure
4179 in i2d_ECPrivateKey. Thanks to Ted Unangst for feedback on this issue.
4181 *mancha <mancha1@zoho.com>*
4183 * Fix some double frees. These are not thought to be exploitable.
4185 *mancha <mancha1@zoho.com>*
4187 * A missing bounds check in the handling of the TLS heartbeat extension
4188 can be used to reveal up to 64k of memory to a connected client or
4191 Thanks for Neel Mehta of Google Security for discovering this bug and to
4192 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
4193 preparing the fix ([CVE-2014-0160])
4195 *Adam Langley, Bodo Moeller*
4197 * Fix for the attack described in the paper "Recovering OpenSSL
4198 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
4199 by Yuval Yarom and Naomi Benger. Details can be obtained from:
4200 <http://eprint.iacr.org/2014/140>
4202 Thanks to Yuval Yarom and Naomi Benger for discovering this
4203 flaw and to Yuval Yarom for supplying a fix ([CVE-2014-0076])
4205 *Yuval Yarom and Naomi Benger*
4207 * Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
4208 this fixes a limitation in previous versions of OpenSSL.
4212 * Experimental encrypt-then-mac support.
4214 Experimental support for encrypt then mac from
4215 draft-gutmann-tls-encrypt-then-mac-02.txt
4217 To enable it set the appropriate extension number (0x42 for the test
4218 server) using e.g. -DTLSEXT_TYPE_encrypt_then_mac=0x42
4220 For non-compliant peers (i.e. just about everything) this should have no
4223 WARNING: EXPERIMENTAL, SUBJECT TO CHANGE.
4227 * Add EVP support for key wrapping algorithms, to avoid problems with
4228 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
4229 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
4230 algorithms and include tests cases.
4234 * Extend CMS code to support RSA-PSS signatures and RSA-OAEP for
4239 * Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
4240 MGF1 digest and OAEP label.
4244 * Make openssl verify return errors.
4246 *Chris Palmer <palmer@google.com> and Ben Laurie*
4248 * New function ASN1_TIME_diff to calculate the difference between two
4249 ASN1_TIME structures or one structure and the current time.
4253 * Update fips_test_suite to support multiple command line options. New
4254 test to induce all self test errors in sequence and check expected
4259 * Add FIPS_{rsa,dsa,ecdsa}_{sign,verify} functions which digest and
4260 sign or verify all in one operation.
4264 * Add fips_algvs: a multicall fips utility incorporating all the algorithm
4265 test programs and fips_test_suite. Includes functionality to parse
4266 the minimal script output of fipsalgest.pl directly.
4270 * Add authorisation parameter to FIPS_module_mode_set().
4274 * Add FIPS selftest for ECDH algorithm using P-224 and B-233 curves.
4278 * Use separate DRBG fields for internal and external flags. New function
4279 FIPS_drbg_health_check() to perform on demand health checking. Add
4280 generation tests to fips_test_suite with reduced health check interval to
4281 demonstrate periodic health checking. Add "nodh" option to
4282 fips_test_suite to skip very slow DH test.
4286 * New function FIPS_get_cipherbynid() to lookup FIPS supported ciphers
4291 * More extensive health check for DRBG checking many more failure modes.
4292 New function FIPS_selftest_drbg_all() to handle every possible DRBG
4293 combination: call this in fips_test_suite.
4297 * Add support for canonical generation of DSA parameter 'g'. See
4300 * Add support for HMAC DRBG from SP800-90. Update DRBG algorithm test and
4301 POST to handle HMAC cases.
4305 * Add functions FIPS_module_version() and FIPS_module_version_text()
4306 to return numerical and string versions of the FIPS module number.
4310 * Rename FIPS_mode_set and FIPS_mode to FIPS_module_mode_set and
4311 FIPS_module_mode. FIPS_mode and FIPS_mode_set will be implemented
4312 outside the validated module in the FIPS capable OpenSSL.
4316 * Minor change to DRBG entropy callback semantics. In some cases
4317 there is no multiple of the block length between min_len and
4318 max_len. Allow the callback to return more than max_len bytes
4319 of entropy but discard any extra: it is the callback's responsibility
4320 to ensure that the extra data discarded does not impact the
4321 requested amount of entropy.
4325 * Add PRNG security strength checks to RSA, DSA and ECDSA using
4326 information in FIPS186-3, SP800-57 and SP800-131A.
4330 * CCM support via EVP. Interface is very similar to GCM case except we
4331 must supply all data in one chunk (i.e. no update, final) and the
4332 message length must be supplied if AAD is used. Add algorithm test
4337 * Initial version of POST overhaul. Add POST callback to allow the status
4338 of POST to be monitored and/or failures induced. Modify fips_test_suite
4339 to use callback. Always run all selftests even if one fails.
4343 * XTS support including algorithm test driver in the fips_gcmtest program.
4344 Note: this does increase the maximum key length from 32 to 64 bytes but
4345 there should be no binary compatibility issues as existing applications
4346 will never use XTS mode.
4350 * Extensive reorganisation of FIPS PRNG behaviour. Remove all dependencies
4351 to OpenSSL RAND code and replace with a tiny FIPS RAND API which also
4352 performs algorithm blocking for unapproved PRNG types. Also do not
4353 set PRNG type in FIPS_mode_set(): leave this to the application.
4354 Add default OpenSSL DRBG handling: sets up FIPS PRNG and seeds with
4355 the standard OpenSSL PRNG: set additional data to a date time vector.
4359 * Rename old X9.31 PRNG functions of the form `FIPS_rand*` to `FIPS_x931*`.
4360 This shouldn't present any incompatibility problems because applications
4361 shouldn't be using these directly and any that are will need to rethink
4362 anyway as the X9.31 PRNG is now deprecated by FIPS 140-2
4366 * Extensive self tests and health checking required by SP800-90 DRBG.
4367 Remove strength parameter from FIPS_drbg_instantiate and always
4368 instantiate at maximum supported strength.
4372 * Add ECDH code to fips module and fips_ecdhvs for primitives only testing.
4376 * New algorithm test program fips_dhvs to handle DH primitives only testing.
4380 * New function DH_compute_key_padded() to compute a DH key and pad with
4381 leading zeroes if needed: this complies with SP800-56A et al.
4385 * Initial implementation of SP800-90 DRBGs for Hash and CTR. Not used by
4386 anything, incomplete, subject to change and largely untested at present.
4390 * Modify fipscanisteronly build option to only build the necessary object
4391 files by filtering FIPS_EX_OBJ through a perl script in crypto/Makefile.
4395 * Add experimental option FIPSSYMS to give all symbols in
4396 fipscanister.o and FIPS or fips prefix. This will avoid
4397 conflicts with future versions of OpenSSL. Add perl script
4398 util/fipsas.pl to preprocess assembly language source files
4399 and rename any affected symbols.
4403 * Add selftest checks and algorithm block of non-fips algorithms in
4404 FIPS mode. Remove DES2 from selftests.
4408 * Add ECDSA code to fips module. Add tiny fips_ecdsa_check to just
4409 return internal method without any ENGINE dependencies. Add new
4410 tiny fips sign and verify functions.
4414 * New build option no-ec2m to disable characteristic 2 code.
4418 * New build option "fipscanisteronly". This only builds fipscanister.o
4419 and (currently) associated fips utilities. Uses the file Makefile.fips
4420 instead of Makefile.org as the prototype.
4424 * Add some FIPS mode restrictions to GCM. Add internal IV generator.
4425 Update fips_gcmtest to use IV generator.
4429 * Initial, experimental EVP support for AES-GCM. AAD can be input by
4430 setting output buffer to NULL. The `*Final` function must be
4431 called although it will not retrieve any additional data. The tag
4432 can be set or retrieved with a ctrl. The IV length is by default 12
4433 bytes (96 bits) but can be set to an alternative value. If the IV
4434 length exceeds the maximum IV length (currently 16 bytes) it cannot be
4439 * New flag in ciphers: EVP_CIPH_FLAG_CUSTOM_CIPHER. This means the
4440 underlying do_cipher function handles all cipher semantics itself
4441 including padding and finalisation. This is useful if (for example)
4442 an ENGINE cipher handles block padding itself. The behaviour of
4443 do_cipher is subtly changed if this flag is set: the return value
4444 is the number of characters written to the output buffer (zero is
4445 no longer an error code) or a negative error code. Also if the
4446 input buffer is NULL and length 0 finalisation should be performed.
4450 * If a candidate issuer certificate is already part of the constructed
4451 path ignore it: new debug notification X509_V_ERR_PATH_LOOP for this case.
4455 * Improve forward-security support: add functions
4457 void SSL_CTX_set_not_resumable_session_callback(
4458 SSL_CTX *ctx, int (*cb)(SSL *ssl, int is_forward_secure))
4459 void SSL_set_not_resumable_session_callback(
4460 SSL *ssl, int (*cb)(SSL *ssl, int is_forward_secure))
4462 for use by SSL/TLS servers; the callback function will be called whenever a
4463 new session is created, and gets to decide whether the session may be
4464 cached to make it resumable (return 0) or not (return 1). (As by the
4465 SSL/TLS protocol specifications, the session_id sent by the server will be
4466 empty to indicate that the session is not resumable; also, the server will
4467 not generate RFC 4507 (RFC 5077) session tickets.)
4469 A simple reasonable callback implementation is to return is_forward_secure.
4470 This parameter will be set to 1 or 0 depending on the ciphersuite selected
4471 by the SSL/TLS server library, indicating whether it can provide forward
4474 *Emilia Käsper <emilia.kasper@esat.kuleuven.be> (Google)*
4476 * New -verify_name option in command line utilities to set verification
4481 * Initial CMAC implementation. WARNING: EXPERIMENTAL, API MAY CHANGE.
4482 Add CMAC pkey methods.
4486 * Experimental renegotiation in s_server -www mode. If the client
4487 browses /reneg connection is renegotiated. If /renegcert it is
4488 renegotiated requesting a certificate.
4492 * Add an "external" session cache for debugging purposes to s_server. This
4493 should help trace issues which normally are only apparent in deployed
4494 multi-process servers.
4498 * Extensive audit of libcrypto with DEBUG_UNUSED. Fix many cases where
4499 return value is ignored. NB. The functions RAND_add(), RAND_seed(),
4500 BIO_set_cipher() and some obscure PEM functions were changed so they
4501 can now return an error. The RAND changes required a change to the
4502 RAND_METHOD structure.
4506 * New macro `__owur` for "OpenSSL Warn Unused Result". This makes use of
4507 a gcc attribute to warn if the result of a function is ignored. This
4508 is enable if DEBUG_UNUSED is set. Add to several functions in evp.h
4509 whose return value is often ignored.
4513 * New -noct, -requestct, -requirect and -ctlogfile options for s_client.
4514 These allow SCTs (signed certificate timestamps) to be requested and
4515 validated when establishing a connection.
4517 *Rob Percival <robpercival@google.com>*
4522 ### Changes between 1.0.2s and 1.0.2t [10 Sep 2019]
4524 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
4525 used even when parsing explicit parameters, when loading a encoded key
4526 or calling `EC_GROUP_new_from_ecpkparameters()`/
4527 `EC_GROUP_new_from_ecparameters()`.
4528 This prevents bypass of security hardening and performance gains,
4529 especially for curves with specialized EC_METHODs.
4530 By default, if a key encoded with explicit parameters is loaded and later
4531 encoded, the output is still encoded with explicit parameters, even if
4532 internally a "named" EC_GROUP is used for computation.
4536 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
4537 this change, EC_GROUP_set_generator would accept order and/or cofactor as
4538 NULL. After this change, only the cofactor parameter can be NULL. It also
4539 does some minimal sanity checks on the passed order.
4544 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
4545 An attack is simple, if the first CMS_recipientInfo is valid but the
4546 second CMS_recipientInfo is chosen ciphertext. If the second
4547 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
4548 encryption key will be replaced by garbage, and the message cannot be
4549 decoded, but if the RSA decryption fails, the correct encryption key is
4550 used and the recipient will not notice the attack.
4551 As a work around for this potential attack the length of the decrypted
4552 key must be equal to the cipher default key length, in case the
4553 certifiate is not given and all recipientInfo are tried out.
4554 The old behaviour can be re-enabled in the CMS code by setting the
4555 CMS_DEBUG_DECRYPT flag.
4560 * Document issue with installation paths in diverse Windows builds
4562 '/usr/local/ssl' is an unsafe prefix for location to install OpenSSL
4563 binaries and run-time config file.
4568 ### Changes between 1.0.2r and 1.0.2s [28 May 2019]
4570 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
4571 This changes the size when using the `genpkey` command when no size is given.
4572 It fixes an omission in earlier changes that changed all RSA, DSA and DH
4573 generation commands to use 2048 bits by default.
4577 * Add FIPS support for Android Arm 64-bit
4579 Support for Android Arm 64-bit was added to the OpenSSL FIPS Object
4580 Module in Version 2.0.10. For some reason, the corresponding target
4581 'android64-aarch64' was missing OpenSSL 1.0.2, whence it could not be
4582 built with FIPS support on Android Arm 64-bit. This omission has been
4585 *Matthias St. Pierre*
4587 ### Changes between 1.0.2q and 1.0.2r [26 Feb 2019]
4589 * 0-byte record padding oracle
4591 If an application encounters a fatal protocol error and then calls
4592 SSL_shutdown() twice (once to send a close_notify, and once to receive one)
4593 then OpenSSL can respond differently to the calling application if a 0 byte
4594 record is received with invalid padding compared to if a 0 byte record is
4595 received with an invalid MAC. If the application then behaves differently
4596 based on that in a way that is detectable to the remote peer, then this
4597 amounts to a padding oracle that could be used to decrypt data.
4599 In order for this to be exploitable "non-stitched" ciphersuites must be in
4600 use. Stitched ciphersuites are optimised implementations of certain
4601 commonly used ciphersuites. Also the application must call SSL_shutdown()
4602 twice even if a protocol error has occurred (applications should not do
4603 this but some do anyway).
4605 This issue was discovered by Juraj Somorovsky, Robert Merget and Nimrod
4606 Aviram, with additional investigation by Steven Collison and Andrew
4607 Hourselt. It was reported to OpenSSL on 10th December 2018.
4612 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
4616 ### Changes between 1.0.2p and 1.0.2q [20 Nov 2018]
4618 * Microarchitecture timing vulnerability in ECC scalar multiplication
4620 OpenSSL ECC scalar multiplication, used in e.g. ECDSA and ECDH, has been
4621 shown to be vulnerable to a microarchitecture timing side channel attack.
4622 An attacker with sufficient access to mount local timing attacks during
4623 ECDSA signature generation could recover the private key.
4625 This issue was reported to OpenSSL on 26th October 2018 by Alejandro
4626 Cabrera Aldaya, Billy Brumley, Sohaib ul Hassan, Cesar Pereida Garcia and
4632 * Timing vulnerability in DSA signature generation
4634 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
4635 timing side channel attack. An attacker could use variations in the signing
4636 algorithm to recover the private key.
4638 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
4643 * Resolve a compatibility issue in EC_GROUP handling with the FIPS Object
4644 Module, accidentally introduced while backporting security fixes from the
4645 development branch and hindering the use of ECC in FIPS mode.
4649 ### Changes between 1.0.2o and 1.0.2p [14 Aug 2018]
4651 * Client DoS due to large DH parameter
4653 During key agreement in a TLS handshake using a DH(E) based ciphersuite a
4654 malicious server can send a very large prime value to the client. This will
4655 cause the client to spend an unreasonably long period of time generating a
4656 key for this prime resulting in a hang until the client has finished. This
4657 could be exploited in a Denial Of Service attack.
4659 This issue was reported to OpenSSL on 5th June 2018 by Guido Vranken
4664 * Cache timing vulnerability in RSA Key Generation
4666 The OpenSSL RSA Key generation algorithm has been shown to be vulnerable to
4667 a cache timing side channel attack. An attacker with sufficient access to
4668 mount cache timing attacks during the RSA key generation process could
4669 recover the private key.
4671 This issue was reported to OpenSSL on 4th April 2018 by Alejandro Cabrera
4672 Aldaya, Billy Brumley, Cesar Pereida Garcia and Luis Manuel Alvarez Tapia.
4677 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
4678 parameter is no longer accepted, as it leads to a corrupt table. NULL
4679 pem_str is reserved for alias entries only.
4683 * Revert blinding in ECDSA sign and instead make problematic addition
4684 length-invariant. Switch even to fixed-length Montgomery multiplication.
4688 * Change generating and checking of primes so that the error rate of not
4689 being prime depends on the intended use based on the size of the input.
4690 For larger primes this will result in more rounds of Miller-Rabin.
4691 The maximal error rate for primes with more than 1080 bits is lowered
4694 *Kurt Roeckx, Annie Yousar*
4696 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
4700 * Add blinding to ECDSA and DSA signatures to protect against side channel
4701 attacks discovered by Keegan Ryan (NCC Group).
4705 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
4706 now allow empty (zero character) pass phrases.
4710 * Certificate time validation (X509_cmp_time) enforces stricter
4711 compliance with RFC 5280. Fractional seconds and timezone offsets
4712 are no longer allowed.
4716 ### Changes between 1.0.2n and 1.0.2o [27 Mar 2018]
4718 * Constructed ASN.1 types with a recursive definition could exceed the stack
4720 Constructed ASN.1 types with a recursive definition (such as can be found
4721 in PKCS7) could eventually exceed the stack given malicious input with
4722 excessive recursion. This could result in a Denial Of Service attack. There
4723 are no such structures used within SSL/TLS that come from untrusted sources
4724 so this is considered safe.
4726 This issue was reported to OpenSSL on 4th January 2018 by the OSS-fuzz
4732 ### Changes between 1.0.2m and 1.0.2n [7 Dec 2017]
4734 * Read/write after SSL object in error state
4736 OpenSSL 1.0.2 (starting from version 1.0.2b) introduced an "error state"
4737 mechanism. The intent was that if a fatal error occurred during a handshake
4738 then OpenSSL would move into the error state and would immediately fail if
4739 you attempted to continue the handshake. This works as designed for the
4740 explicit handshake functions (SSL_do_handshake(), SSL_accept() and
4741 SSL_connect()), however due to a bug it does not work correctly if
4742 SSL_read() or SSL_write() is called directly. In that scenario, if the
4743 handshake fails then a fatal error will be returned in the initial function
4744 call. If SSL_read()/SSL_write() is subsequently called by the application
4745 for the same SSL object then it will succeed and the data is passed without
4746 being decrypted/encrypted directly from the SSL/TLS record layer.
4748 In order to exploit this issue an application bug would have to be present
4749 that resulted in a call to SSL_read()/SSL_write() being issued after having
4750 already received a fatal error.
4752 This issue was reported to OpenSSL by David Benjamin (Google).
4757 * rsaz_1024_mul_avx2 overflow bug on x86_64
4759 There is an overflow bug in the AVX2 Montgomery multiplication procedure
4760 used in exponentiation with 1024-bit moduli. No EC algorithms are affected.
4761 Analysis suggests that attacks against RSA and DSA as a result of this
4762 defect would be very difficult to perform and are not believed likely.
4763 Attacks against DH1024 are considered just feasible, because most of the
4764 work necessary to deduce information about a private key may be performed
4765 offline. The amount of resources required for such an attack would be
4766 significant. However, for an attack on TLS to be meaningful, the server
4767 would have to share the DH1024 private key among multiple clients, which is
4768 no longer an option since CVE-2016-0701.
4770 This only affects processors that support the AVX2 but not ADX extensions
4771 like Intel Haswell (4th generation).
4773 This issue was reported to OpenSSL by David Benjamin (Google). The issue
4774 was originally found via the OSS-Fuzz project.
4779 ### Changes between 1.0.2l and 1.0.2m [2 Nov 2017]
4781 * bn_sqrx8x_internal carry bug on x86_64
4783 There is a carry propagating bug in the x86_64 Montgomery squaring
4784 procedure. No EC algorithms are affected. Analysis suggests that attacks
4785 against RSA and DSA as a result of this defect would be very difficult to
4786 perform and are not believed likely. Attacks against DH are considered just
4787 feasible (although very difficult) because most of the work necessary to
4788 deduce information about a private key may be performed offline. The amount
4789 of resources required for such an attack would be very significant and
4790 likely only accessible to a limited number of attackers. An attacker would
4791 additionally need online access to an unpatched system using the target
4792 private key in a scenario with persistent DH parameters and a private
4793 key that is shared between multiple clients.
4795 This only affects processors that support the BMI1, BMI2 and ADX extensions
4796 like Intel Broadwell (5th generation) and later or AMD Ryzen.
4798 This issue was reported to OpenSSL by the OSS-Fuzz project.
4803 * Malformed X.509 IPAddressFamily could cause OOB read
4805 If an X.509 certificate has a malformed IPAddressFamily extension,
4806 OpenSSL could do a one-byte buffer overread. The most likely result
4807 would be an erroneous display of the certificate in text format.
4809 This issue was reported to OpenSSL by the OSS-Fuzz project.
4813 ### Changes between 1.0.2k and 1.0.2l [25 May 2017]
4815 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
4816 platform rather than 'mingw'.
4820 ### Changes between 1.0.2j and 1.0.2k [26 Jan 2017]
4822 * Truncated packet could crash via OOB read
4824 If one side of an SSL/TLS path is running on a 32-bit host and a specific
4825 cipher is being used, then a truncated packet can cause that host to
4826 perform an out-of-bounds read, usually resulting in a crash.
4828 This issue was reported to OpenSSL by Robert Święcki of Google.
4833 * BN_mod_exp may produce incorrect results on x86_64
4835 There is a carry propagating bug in the x86_64 Montgomery squaring
4836 procedure. No EC algorithms are affected. Analysis suggests that attacks
4837 against RSA and DSA as a result of this defect would be very difficult to
4838 perform and are not believed likely. Attacks against DH are considered just
4839 feasible (although very difficult) because most of the work necessary to
4840 deduce information about a private key may be performed offline. The amount
4841 of resources required for such an attack would be very significant and
4842 likely only accessible to a limited number of attackers. An attacker would
4843 additionally need online access to an unpatched system using the target
4844 private key in a scenario with persistent DH parameters and a private
4845 key that is shared between multiple clients. For example this can occur by
4846 default in OpenSSL DHE based SSL/TLS ciphersuites. Note: This issue is very
4847 similar to CVE-2015-3193 but must be treated as a separate problem.
4849 This issue was reported to OpenSSL by the OSS-Fuzz project.
4854 * Montgomery multiplication may produce incorrect results
4856 There is a carry propagating bug in the Broadwell-specific Montgomery
4857 multiplication procedure that handles input lengths divisible by, but
4858 longer than 256 bits. Analysis suggests that attacks against RSA, DSA
4859 and DH private keys are impossible. This is because the subroutine in
4860 question is not used in operations with the private key itself and an input
4861 of the attacker's direct choice. Otherwise the bug can manifest itself as
4862 transient authentication and key negotiation failures or reproducible
4863 erroneous outcome of public-key operations with specially crafted input.
4864 Among EC algorithms only Brainpool P-512 curves are affected and one
4865 presumably can attack ECDH key negotiation. Impact was not analyzed in
4866 detail, because pre-requisites for attack are considered unlikely. Namely
4867 multiple clients have to choose the curve in question and the server has to
4868 share the private key among them, neither of which is default behaviour.
4869 Even then only clients that chose the curve will be affected.
4871 This issue was publicly reported as transient failures and was not
4872 initially recognized as a security issue. Thanks to Richard Morgan for
4873 providing reproducible case.
4878 * OpenSSL now fails if it receives an unrecognised record type in TLS1.0
4879 or TLS1.1. Previously this only happened in SSLv3 and TLS1.2. This is to
4880 prevent issues where no progress is being made and the peer continually
4881 sends unrecognised record types, using up resources processing them.
4885 ### Changes between 1.0.2i and 1.0.2j [26 Sep 2016]
4887 * Missing CRL sanity check
4889 A bug fix which included a CRL sanity check was added to OpenSSL 1.1.0
4890 but was omitted from OpenSSL 1.0.2i. As a result any attempt to use
4891 CRLs in OpenSSL 1.0.2i will crash with a null pointer exception.
4893 This issue only affects the OpenSSL 1.0.2i
4898 ### Changes between 1.0.2h and 1.0.2i [22 Sep 2016]
4900 * OCSP Status Request extension unbounded memory growth
4902 A malicious client can send an excessively large OCSP Status Request
4903 extension. If that client continually requests renegotiation, sending a
4904 large OCSP Status Request extension each time, then there will be unbounded
4905 memory growth on the server. This will eventually lead to a Denial Of
4906 Service attack through memory exhaustion. Servers with a default
4907 configuration are vulnerable even if they do not support OCSP. Builds using
4908 the "no-ocsp" build time option are not affected.
4910 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4915 * In order to mitigate the SWEET32 attack, the DES ciphers were moved from
4918 This issue was reported to OpenSSL Karthikeyan Bhargavan and Gaetan
4924 * OOB write in MDC2_Update()
4926 An overflow can occur in MDC2_Update() either if called directly or
4927 through the EVP_DigestUpdate() function using MDC2. If an attacker
4928 is able to supply very large amounts of input data after a previous
4929 call to EVP_EncryptUpdate() with a partial block then a length check
4930 can overflow resulting in a heap corruption.
4932 The amount of data needed is comparable to SIZE_MAX which is impractical
4935 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4940 * Malformed SHA512 ticket DoS
4942 If a server uses SHA512 for TLS session ticket HMAC it is vulnerable to a
4943 DoS attack where a malformed ticket will result in an OOB read which will
4946 The use of SHA512 in TLS session tickets is comparatively rare as it requires
4947 a custom server callback and ticket lookup mechanism.
4949 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4954 * OOB write in BN_bn2dec()
4956 The function BN_bn2dec() does not check the return value of BN_div_word().
4957 This can cause an OOB write if an application uses this function with an
4958 overly large BIGNUM. This could be a problem if an overly large certificate
4959 or CRL is printed out from an untrusted source. TLS is not affected because
4960 record limits will reject an oversized certificate before it is parsed.
4962 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4967 * OOB read in TS_OBJ_print_bio()
4969 The function TS_OBJ_print_bio() misuses OBJ_obj2txt(): the return value is
4970 the total length the OID text representation would use and not the amount
4971 of data written. This will result in OOB reads when large OIDs are
4974 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4979 * Pointer arithmetic undefined behaviour
4981 Avoid some undefined pointer arithmetic
4983 A common idiom in the codebase is to check limits in the following manner:
4986 Where "p" points to some malloc'd data of SIZE bytes and
4989 "len" here could be from some externally supplied data (e.g. from a TLS
4992 The rules of C pointer arithmetic are such that "p + len" is only well
4993 defined where len <= SIZE. Therefore the above idiom is actually
4994 undefined behaviour.
4996 For example this could cause problems if some malloc implementation
4997 provides an address for "p" such that "p + len" actually overflows for
4998 values of len that are too big and therefore p + len < limit.
5000 This issue was reported to OpenSSL by Guido Vranken
5005 * Constant time flag not preserved in DSA signing
5007 Operations in the DSA signing algorithm should run in constant time in
5008 order to avoid side channel attacks. A flaw in the OpenSSL DSA
5009 implementation means that a non-constant time codepath is followed for
5010 certain operations. This has been demonstrated through a cache-timing
5011 attack to be sufficient for an attacker to recover the private DSA key.
5013 This issue was reported by César Pereida (Aalto University), Billy Brumley
5014 (Tampere University of Technology), and Yuval Yarom (The University of
5015 Adelaide and NICTA).
5020 * DTLS buffered message DoS
5022 In a DTLS connection where handshake messages are delivered out-of-order
5023 those messages that OpenSSL is not yet ready to process will be buffered
5024 for later use. Under certain circumstances, a flaw in the logic means that
5025 those messages do not get removed from the buffer even though the handshake
5026 has been completed. An attacker could force up to approx. 15 messages to
5027 remain in the buffer when they are no longer required. These messages will
5028 be cleared when the DTLS connection is closed. The default maximum size for
5029 a message is 100k. Therefore the attacker could force an additional 1500k
5030 to be consumed per connection. By opening many simulataneous connections an
5031 attacker could cause a DoS attack through memory exhaustion.
5033 This issue was reported to OpenSSL by Quan Luo.
5038 * DTLS replay protection DoS
5040 A flaw in the DTLS replay attack protection mechanism means that records
5041 that arrive for future epochs update the replay protection "window" before
5042 the MAC for the record has been validated. This could be exploited by an
5043 attacker by sending a record for the next epoch (which does not have to
5044 decrypt or have a valid MAC), with a very large sequence number. This means
5045 that all subsequent legitimate packets are dropped causing a denial of
5046 service for a specific DTLS connection.
5048 This issue was reported to OpenSSL by the OCAP audit team.
5053 * Certificate message OOB reads
5055 In OpenSSL 1.0.2 and earlier some missing message length checks can result
5056 in OOB reads of up to 2 bytes beyond an allocated buffer. There is a
5057 theoretical DoS risk but this has not been observed in practice on common
5060 The messages affected are client certificate, client certificate request
5061 and server certificate. As a result the attack can only be performed
5062 against a client or a server which enables client authentication.
5064 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5069 ### Changes between 1.0.2g and 1.0.2h [3 May 2016]
5071 * Prevent padding oracle in AES-NI CBC MAC check
5073 A MITM attacker can use a padding oracle attack to decrypt traffic
5074 when the connection uses an AES CBC cipher and the server support
5077 This issue was introduced as part of the fix for Lucky 13 padding
5078 attack ([CVE-2013-0169]). The padding check was rewritten to be in
5079 constant time by making sure that always the same bytes are read and
5080 compared against either the MAC or padding bytes. But it no longer
5081 checked that there was enough data to have both the MAC and padding
5084 This issue was reported by Juraj Somorovsky using TLS-Attacker.
5088 * Fix EVP_EncodeUpdate overflow
5090 An overflow can occur in the EVP_EncodeUpdate() function which is used for
5091 Base64 encoding of binary data. If an attacker is able to supply very large
5092 amounts of input data then a length check can overflow resulting in a heap
5095 Internally to OpenSSL the EVP_EncodeUpdate() function is primarily used by
5096 the `PEM_write_bio*` family of functions. These are mainly used within the
5097 OpenSSL command line applications, so any application which processes data
5098 from an untrusted source and outputs it as a PEM file should be considered
5099 vulnerable to this issue. User applications that call these APIs directly
5100 with large amounts of untrusted data may also be vulnerable.
5102 This issue was reported by Guido Vranken.
5107 * Fix EVP_EncryptUpdate overflow
5109 An overflow can occur in the EVP_EncryptUpdate() function. If an attacker
5110 is able to supply very large amounts of input data after a previous call to
5111 EVP_EncryptUpdate() with a partial block then a length check can overflow
5112 resulting in a heap corruption. Following an analysis of all OpenSSL
5113 internal usage of the EVP_EncryptUpdate() function all usage is one of two
5114 forms. The first form is where the EVP_EncryptUpdate() call is known to be
5115 the first called function after an EVP_EncryptInit(), and therefore that
5116 specific call must be safe. The second form is where the length passed to
5117 EVP_EncryptUpdate() can be seen from the code to be some small value and
5118 therefore there is no possibility of an overflow. Since all instances are
5119 one of these two forms, it is believed that there can be no overflows in
5120 internal code due to this problem. It should be noted that
5121 EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths.
5122 Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances
5123 of these calls have also been analysed too and it is believed there are no
5124 instances in internal usage where an overflow could occur.
5126 This issue was reported by Guido Vranken.
5131 * Prevent ASN.1 BIO excessive memory allocation
5133 When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio()
5134 a short invalid encoding can cause allocation of large amounts of memory
5135 potentially consuming excessive resources or exhausting memory.
5137 Any application parsing untrusted data through d2i BIO functions is
5138 affected. The memory based functions such as d2i_X509() are *not* affected.
5139 Since the memory based functions are used by the TLS library, TLS
5140 applications are not affected.
5142 This issue was reported by Brian Carpenter.
5149 ASN1 Strings that are over 1024 bytes can cause an overread in applications
5150 using the X509_NAME_oneline() function on EBCDIC systems. This could result
5151 in arbitrary stack data being returned in the buffer.
5153 This issue was reported by Guido Vranken.
5158 * Modify behavior of ALPN to invoke callback after SNI/servername
5159 callback, such that updates to the SSL_CTX affect ALPN.
5163 * Remove LOW from the DEFAULT cipher list. This removes singles DES from the
5168 * Only remove the SSLv2 methods with the no-ssl2-method option. When the
5169 methods are enabled and ssl2 is disabled the methods return NULL.
5173 ### Changes between 1.0.2f and 1.0.2g [1 Mar 2016]
5175 * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL.
5176 Builds that are not configured with "enable-weak-ssl-ciphers" will not
5177 provide any "EXPORT" or "LOW" strength ciphers.
5181 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2
5182 is by default disabled at build-time. Builds that are not configured with
5183 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used,
5184 users who want to negotiate SSLv2 via the version-flexible SSLv23_method()
5185 will need to explicitly call either of:
5187 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2);
5189 SSL_clear_options(ssl, SSL_OP_NO_SSLv2);
5191 as appropriate. Even if either of those is used, or the application
5192 explicitly uses the version-specific SSLv2_method() or its client and
5193 server variants, SSLv2 ciphers vulnerable to exhaustive search key
5194 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT
5195 ciphers, and SSLv2 56-bit DES are no longer available.
5200 * Fix a double-free in DSA code
5202 A double free bug was discovered when OpenSSL parses malformed DSA private
5203 keys and could lead to a DoS attack or memory corruption for applications
5204 that receive DSA private keys from untrusted sources. This scenario is
5207 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using
5213 * Disable SRP fake user seed to address a server memory leak.
5215 Add a new method SRP_VBASE_get1_by_user that handles the seed properly.
5217 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
5218 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user
5219 was changed to ignore the "fake user" SRP seed, even if the seed
5222 Users should use SRP_VBASE_get1_by_user instead. Note that in
5223 SRP_VBASE_get1_by_user, caller must free the returned value. Note
5224 also that even though configuring the SRP seed attempts to hide
5225 invalid usernames by continuing the handshake with fake
5226 credentials, this behaviour is not constant time and no strong
5227 guarantees are made that the handshake is indistinguishable from
5228 that of a valid user.
5233 * Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption
5235 In the BN_hex2bn function the number of hex digits is calculated using an
5236 int value `i`. Later `bn_expand` is called with a value of `i * 4`. For
5237 large values of `i` this can result in `bn_expand` not allocating any
5238 memory because `i * 4` is negative. This can leave the internal BIGNUM data
5239 field as NULL leading to a subsequent NULL ptr deref. For very large values
5240 of `i`, the calculation `i * 4` could be a positive value smaller than `i`.
5241 In this case memory is allocated to the internal BIGNUM data field, but it
5242 is insufficiently sized leading to heap corruption. A similar issue exists
5243 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn
5244 is ever called by user applications with very large untrusted hex/dec data.
5245 This is anticipated to be a rare occurrence.
5247 All OpenSSL internal usage of these functions use data that is not expected
5248 to be untrusted, e.g. config file data or application command line
5249 arguments. If user developed applications generate config file data based
5250 on untrusted data then it is possible that this could also lead to security
5251 consequences. This is also anticipated to be rare.
5253 This issue was reported to OpenSSL by Guido Vranken.
5258 * Fix memory issues in `BIO_*printf` functions
5260 The internal `fmtstr` function used in processing a "%s" format string in
5261 the `BIO_*printf` functions could overflow while calculating the length of a
5262 string and cause an OOB read when printing very long strings.
5264 Additionally the internal `doapr_outch` function can attempt to write to an
5265 OOB memory location (at an offset from the NULL pointer) in the event of a
5266 memory allocation failure. In 1.0.2 and below this could be caused where
5267 the size of a buffer to be allocated is greater than INT_MAX. E.g. this
5268 could be in processing a very long "%s" format string. Memory leaks can
5271 The first issue may mask the second issue dependent on compiler behaviour.
5272 These problems could enable attacks where large amounts of untrusted data
5273 is passed to the `BIO_*printf` functions. If applications use these functions
5274 in this way then they could be vulnerable. OpenSSL itself uses these
5275 functions when printing out human-readable dumps of ASN.1 data. Therefore
5276 applications that print this data could be vulnerable if the data is from
5277 untrusted sources. OpenSSL command line applications could also be
5278 vulnerable where they print out ASN.1 data, or if untrusted data is passed
5279 as command line arguments.
5281 Libssl is not considered directly vulnerable. Additionally certificates etc
5282 received via remote connections via libssl are also unlikely to be able to
5283 trigger these issues because of message size limits enforced within libssl.
5285 This issue was reported to OpenSSL Guido Vranken.
5290 * Side channel attack on modular exponentiation
5292 A side-channel attack was found which makes use of cache-bank conflicts on
5293 the Intel Sandy-Bridge microarchitecture which could lead to the recovery
5294 of RSA keys. The ability to exploit this issue is limited as it relies on
5295 an attacker who has control of code in a thread running on the same
5296 hyper-threaded core as the victim thread which is performing decryptions.
5298 This issue was reported to OpenSSL by Yuval Yarom, The University of
5299 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and
5300 Nadia Heninger, University of Pennsylvania with more information at
5301 <http://cachebleed.info>.
5306 * Change the `req` command to generate a 2048-bit RSA/DSA key by default,
5307 if no keysize is specified with default_bits. This fixes an
5308 omission in an earlier change that changed all RSA/DSA key generation
5309 commands to use 2048 bits by default.
5313 ### Changes between 1.0.2e and 1.0.2f [28 Jan 2016]
5315 * DH small subgroups
5317 Historically OpenSSL only ever generated DH parameters based on "safe"
5318 primes. More recently (in version 1.0.2) support was provided for
5319 generating X9.42 style parameter files such as those required for RFC 5114
5320 support. The primes used in such files may not be "safe". Where an
5321 application is using DH configured with parameters based on primes that are
5322 not "safe" then an attacker could use this fact to find a peer's private
5323 DH exponent. This attack requires that the attacker complete multiple
5324 handshakes in which the peer uses the same private DH exponent. For example
5325 this could be used to discover a TLS server's private DH exponent if it's
5326 reusing the private DH exponent or it's using a static DH ciphersuite.
5328 OpenSSL provides the option SSL_OP_SINGLE_DH_USE for ephemeral DH (DHE) in
5329 TLS. It is not on by default. If the option is not set then the server
5330 reuses the same private DH exponent for the life of the server process and
5331 would be vulnerable to this attack. It is believed that many popular
5332 applications do set this option and would therefore not be at risk.
5334 The fix for this issue adds an additional check where a "q" parameter is
5335 available (as is the case in X9.42 based parameters). This detects the
5336 only known attack, and is the only possible defense for static DH
5337 ciphersuites. This could have some performance impact.
5339 Additionally the SSL_OP_SINGLE_DH_USE option has been switched on by
5340 default and cannot be disabled. This could have some performance impact.
5342 This issue was reported to OpenSSL by Antonio Sanso (Adobe).
5347 * SSLv2 doesn't block disabled ciphers
5349 A malicious client can negotiate SSLv2 ciphers that have been disabled on
5350 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
5351 been disabled, provided that the SSLv2 protocol was not also disabled via
5354 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
5355 and Sebastian Schinzel.
5360 ### Changes between 1.0.2d and 1.0.2e [3 Dec 2015]
5362 * BN_mod_exp may produce incorrect results on x86_64
5364 There is a carry propagating bug in the x86_64 Montgomery squaring
5365 procedure. No EC algorithms are affected. Analysis suggests that attacks
5366 against RSA and DSA as a result of this defect would be very difficult to
5367 perform and are not believed likely. Attacks against DH are considered just
5368 feasible (although very difficult) because most of the work necessary to
5369 deduce information about a private key may be performed offline. The amount
5370 of resources required for such an attack would be very significant and
5371 likely only accessible to a limited number of attackers. An attacker would
5372 additionally need online access to an unpatched system using the target
5373 private key in a scenario with persistent DH parameters and a private
5374 key that is shared between multiple clients. For example this can occur by
5375 default in OpenSSL DHE based SSL/TLS ciphersuites.
5377 This issue was reported to OpenSSL by Hanno Böck.
5382 * Certificate verify crash with missing PSS parameter
5384 The signature verification routines will crash with a NULL pointer
5385 dereference if presented with an ASN.1 signature using the RSA PSS
5386 algorithm and absent mask generation function parameter. Since these
5387 routines are used to verify certificate signature algorithms this can be
5388 used to crash any certificate verification operation and exploited in a
5389 DoS attack. Any application which performs certificate verification is
5390 vulnerable including OpenSSL clients and servers which enable client
5393 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
5398 * X509_ATTRIBUTE memory leak
5400 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
5401 memory. This structure is used by the PKCS#7 and CMS routines so any
5402 application which reads PKCS#7 or CMS data from untrusted sources is
5403 affected. SSL/TLS is not affected.
5405 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
5411 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
5412 This changes the decoding behaviour for some invalid messages,
5413 though the change is mostly in the more lenient direction, and
5414 legacy behaviour is preserved as much as possible.
5418 * In DSA_generate_parameters_ex, if the provided seed is too short,
5421 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
5423 ### Changes between 1.0.2c and 1.0.2d [9 Jul 2015]
5425 * Alternate chains certificate forgery
5427 During certificate verification, OpenSSL will attempt to find an
5428 alternative certificate chain if the first attempt to build such a chain
5429 fails. An error in the implementation of this logic can mean that an
5430 attacker could cause certain checks on untrusted certificates to be
5431 bypassed, such as the CA flag, enabling them to use a valid leaf
5432 certificate to act as a CA and "issue" an invalid certificate.
5434 This issue was reported to OpenSSL by Adam Langley/David Benjamin
5439 ### Changes between 1.0.2b and 1.0.2c [12 Jun 2015]
5441 * Fix HMAC ABI incompatibility. The previous version introduced an ABI
5442 incompatibility in the handling of HMAC. The previous ABI has now been
5447 ### Changes between 1.0.2a and 1.0.2b [11 Jun 2015]
5449 * Malformed ECParameters causes infinite loop
5451 When processing an ECParameters structure OpenSSL enters an infinite loop
5452 if the curve specified is over a specially malformed binary polynomial
5455 This can be used to perform denial of service against any
5456 system which processes public keys, certificate requests or
5457 certificates. This includes TLS clients and TLS servers with
5458 client authentication enabled.
5460 This issue was reported to OpenSSL by Joseph Barr-Pixton.
5465 * Exploitable out-of-bounds read in X509_cmp_time
5467 X509_cmp_time does not properly check the length of the ASN1_TIME
5468 string and can read a few bytes out of bounds. In addition,
5469 X509_cmp_time accepts an arbitrary number of fractional seconds in the
5472 An attacker can use this to craft malformed certificates and CRLs of
5473 various sizes and potentially cause a segmentation fault, resulting in
5474 a DoS on applications that verify certificates or CRLs. TLS clients
5475 that verify CRLs are affected. TLS clients and servers with client
5476 authentication enabled may be affected if they use custom verification
5479 This issue was reported to OpenSSL by Robert Swiecki (Google), and
5480 independently by Hanno Böck.
5485 * PKCS7 crash with missing EnvelopedContent
5487 The PKCS#7 parsing code does not handle missing inner EncryptedContent
5488 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
5489 with missing content and trigger a NULL pointer dereference on parsing.
5491 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
5492 structures from untrusted sources are affected. OpenSSL clients and
5493 servers are not affected.
5495 This issue was reported to OpenSSL by Michal Zalewski (Google).
5500 * CMS verify infinite loop with unknown hash function
5502 When verifying a signedData message the CMS code can enter an infinite loop
5503 if presented with an unknown hash function OID. This can be used to perform
5504 denial of service against any system which verifies signedData messages using
5506 This issue was reported to OpenSSL by Johannes Bauer.
5511 * Race condition handling NewSessionTicket
5513 If a NewSessionTicket is received by a multi-threaded client when attempting to
5514 reuse a previous ticket then a race condition can occur potentially leading to
5515 a double free of the ticket data.
5520 * Only support 256-bit or stronger elliptic curves with the
5521 'ecdh_auto' setting (server) or by default (client). Of supported
5522 curves, prefer P-256 (both).
5526 ### Changes between 1.0.2 and 1.0.2a [19 Mar 2015]
5528 * ClientHello sigalgs DoS fix
5530 If a client connects to an OpenSSL 1.0.2 server and renegotiates with an
5531 invalid signature algorithms extension a NULL pointer dereference will
5532 occur. This can be exploited in a DoS attack against the server.
5534 This issue was was reported to OpenSSL by David Ramos of Stanford
5538 *Stephen Henson and Matt Caswell*
5540 * Multiblock corrupted pointer fix
5542 OpenSSL 1.0.2 introduced the "multiblock" performance improvement. This
5543 feature only applies on 64 bit x86 architecture platforms that support AES
5544 NI instructions. A defect in the implementation of "multiblock" can cause
5545 OpenSSL's internal write buffer to become incorrectly set to NULL when
5546 using non-blocking IO. Typically, when the user application is using a
5547 socket BIO for writing, this will only result in a failed connection.
5548 However if some other BIO is used then it is likely that a segmentation
5549 fault will be triggered, thus enabling a potential DoS attack.
5551 This issue was reported to OpenSSL by Daniel Danner and Rainer Mueller.
5556 * Segmentation fault in DTLSv1_listen fix
5558 The DTLSv1_listen function is intended to be stateless and processes the
5559 initial ClientHello from many peers. It is common for user code to loop
5560 over the call to DTLSv1_listen until a valid ClientHello is received with
5561 an associated cookie. A defect in the implementation of DTLSv1_listen means
5562 that state is preserved in the SSL object from one invocation to the next
5563 that can lead to a segmentation fault. Errors processing the initial
5564 ClientHello can trigger this scenario. An example of such an error could be
5565 that a DTLS1.0 only client is attempting to connect to a DTLS1.2 only
5568 This issue was reported to OpenSSL by Per Allansson.
5573 * Segmentation fault in ASN1_TYPE_cmp fix
5575 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
5576 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
5577 certificate signature algorithm consistency this can be used to crash any
5578 certificate verification operation and exploited in a DoS attack. Any
5579 application which performs certificate verification is vulnerable including
5580 OpenSSL clients and servers which enable client authentication.
5585 * Segmentation fault for invalid PSS parameters fix
5587 The signature verification routines will crash with a NULL pointer
5588 dereference if presented with an ASN.1 signature using the RSA PSS
5589 algorithm and invalid parameters. Since these routines are used to verify
5590 certificate signature algorithms this can be used to crash any
5591 certificate verification operation and exploited in a DoS attack. Any
5592 application which performs certificate verification is vulnerable including
5593 OpenSSL clients and servers which enable client authentication.
5595 This issue was was reported to OpenSSL by Brian Carpenter.
5600 * ASN.1 structure reuse memory corruption fix
5602 Reusing a structure in ASN.1 parsing may allow an attacker to cause
5603 memory corruption via an invalid write. Such reuse is and has been
5604 strongly discouraged and is believed to be rare.
5606 Applications that parse structures containing CHOICE or ANY DEFINED BY
5607 components may be affected. Certificate parsing (d2i_X509 and related
5608 functions) are however not affected. OpenSSL clients and servers are
5614 * PKCS7 NULL pointer dereferences fix
5616 The PKCS#7 parsing code does not handle missing outer ContentInfo
5617 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
5618 missing content and trigger a NULL pointer dereference on parsing.
5620 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
5621 otherwise parse PKCS#7 structures from untrusted sources are
5622 affected. OpenSSL clients and servers are not affected.
5624 This issue was reported to OpenSSL by Michal Zalewski (Google).
5629 * DoS via reachable assert in SSLv2 servers fix
5631 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
5632 servers that both support SSLv2 and enable export cipher suites by sending
5633 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
5635 This issue was discovered by Sean Burford (Google) and Emilia Käsper
5636 (OpenSSL development team).
5641 * Empty CKE with client auth and DHE fix
5643 If client auth is used then a server can seg fault in the event of a DHE
5644 ciphersuite being selected and a zero length ClientKeyExchange message
5645 being sent by the client. This could be exploited in a DoS attack.
5650 * Handshake with unseeded PRNG fix
5652 Under certain conditions an OpenSSL 1.0.2 client can complete a handshake
5653 with an unseeded PRNG. The conditions are:
5654 - The client is on a platform where the PRNG has not been seeded
5655 automatically, and the user has not seeded manually
5656 - A protocol specific client method version has been used (i.e. not
5657 SSL_client_methodv23)
5658 - A ciphersuite is used that does not require additional random data from
5659 the PRNG beyond the initial ClientHello client random (e.g. PSK-RC4-SHA).
5661 If the handshake succeeds then the client random that has been used will
5662 have been generated from a PRNG with insufficient entropy and therefore the
5663 output may be predictable.
5665 For example using the following command with an unseeded openssl will
5666 succeed on an unpatched platform:
5668 openssl s_client -psk 1a2b3c4d -tls1_2 -cipher PSK-RC4-SHA
5673 * Use After Free following d2i_ECPrivatekey error fix
5675 A malformed EC private key file consumed via the d2i_ECPrivateKey function
5676 could cause a use after free condition. This, in turn, could cause a double
5677 free in several private key parsing functions (such as d2i_PrivateKey
5678 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
5679 for applications that receive EC private keys from untrusted
5680 sources. This scenario is considered rare.
5682 This issue was discovered by the BoringSSL project and fixed in their
5688 * X509_to_X509_REQ NULL pointer deref fix
5690 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
5691 the certificate key is invalid. This function is rarely used in practice.
5693 This issue was discovered by Brian Carpenter.
5698 * Removed the export ciphers from the DEFAULT ciphers
5702 ### Changes between 1.0.1l and 1.0.2 [22 Jan 2015]
5704 * Facilitate "universal" ARM builds targeting range of ARM ISAs, e.g.
5705 ARMv5 through ARMv8, as opposite to "locking" it to single one.
5706 So far those who have to target multiple platforms would compromise
5707 and argue that binary targeting say ARMv5 would still execute on
5708 ARMv8. "Universal" build resolves this compromise by providing
5709 near-optimal performance even on newer platforms.
5713 * Accelerated NIST P-256 elliptic curve implementation for x86_64
5714 (other platforms pending).
5716 *Shay Gueron & Vlad Krasnov (Intel Corp), Andy Polyakov*
5718 * Add support for the SignedCertificateTimestampList certificate and
5719 OCSP response extensions from RFC6962.
5723 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
5724 for corner cases. (Certain input points at infinity could lead to
5725 bogus results, with non-infinity inputs mapped to infinity too.)
5729 * Initial support for PowerISA 2.0.7, first implemented in POWER8.
5730 This covers AES, SHA256/512 and GHASH. "Initial" means that most
5731 common cases are optimized and there still is room for further
5732 improvements. Vector Permutation AES for Altivec is also added.
5736 * Add support for little-endian ppc64 Linux target.
5738 *Marcelo Cerri (IBM)*
5740 * Initial support for AMRv8 ISA crypto extensions. This covers AES,
5741 SHA1, SHA256 and GHASH. "Initial" means that most common cases
5742 are optimized and there still is room for further improvements.
5743 Both 32- and 64-bit modes are supported.
5745 *Andy Polyakov, Ard Biesheuvel (Linaro)*
5747 * Improved ARMv7 NEON support.
5751 * Support for SPARC Architecture 2011 crypto extensions, first
5752 implemented in SPARC T4. This covers AES, DES, Camellia, SHA1,
5753 SHA256/512, MD5, GHASH and modular exponentiation.
5755 *Andy Polyakov, David Miller*
5757 * Accelerated modular exponentiation for Intel processors, a.k.a.
5760 *Shay Gueron & Vlad Krasnov (Intel Corp)*
5762 * Support for new and upcoming Intel processors, including AVX2,
5763 BMI and SHA ISA extensions. This includes additional "stitched"
5764 implementations, AESNI-SHA256 and GCM, and multi-buffer support
5767 This work was sponsored by Intel Corp.
5771 * Support for DTLS 1.2. This adds two sets of DTLS methods: DTLS_*_method()
5772 supports both DTLS 1.2 and 1.0 and should use whatever version the peer
5773 supports and DTLSv1_2_*_method() which supports DTLS 1.2 only.
5777 * Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
5778 this fixes a limitation in previous versions of OpenSSL.
5782 * Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
5783 MGF1 digest and OAEP label.
5787 * Add EVP support for key wrapping algorithms, to avoid problems with
5788 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
5789 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
5790 algorithms and include tests cases.
5794 * Add functions to allocate and set the fields of an ECDSA_METHOD
5797 *Douglas E. Engert, Steve Henson*
5799 * New functions OPENSSL_gmtime_diff and ASN1_TIME_diff to find the
5800 difference in days and seconds between two tm or ASN1_TIME structures.
5804 * Add -rev test option to s_server to just reverse order of characters
5805 received by client and send back to server. Also prints an abbreviated
5806 summary of the connection parameters.
5810 * New option -brief for s_client and s_server to print out a brief summary
5811 of connection parameters.
5815 * Add callbacks for arbitrary TLS extensions.
5817 *Trevor Perrin <trevp@trevp.net> and Ben Laurie*
5819 * New option -crl_download in several openssl utilities to download CRLs
5820 from CRLDP extension in certificates.
5824 * New options -CRL and -CRLform for s_client and s_server for CRLs.
5828 * New function X509_CRL_diff to generate a delta CRL from the difference
5829 of two full CRLs. Add support to "crl" utility.
5833 * New functions to set lookup_crls function and to retrieve
5834 X509_STORE from X509_STORE_CTX.
5838 * Print out deprecated issuer and subject unique ID fields in
5843 * Extend OCSP I/O functions so they can be used for simple general purpose
5844 HTTP as well as OCSP. New wrapper function which can be used to download
5845 CRLs using the OCSP API.
5849 * Delegate command line handling in s_client/s_server to SSL_CONF APIs.
5853 * `SSL_CONF*` functions. These provide a common framework for application
5854 configuration using configuration files or command lines.
5858 * SSL/TLS tracing code. This parses out SSL/TLS records using the
5859 message callback and prints the results. Needs compile time option
5860 "enable-ssl-trace". New options to s_client and s_server to enable
5865 * New ctrl and macro to retrieve supported points extensions.
5866 Print out extension in s_server and s_client.
5870 * New functions to retrieve certificate signature and signature
5875 * Add functions to retrieve and manipulate the raw cipherlist sent by a
5880 * New Suite B modes for TLS code. These use and enforce the requirements
5881 of RFC6460: restrict ciphersuites, only permit Suite B algorithms and
5882 only use Suite B curves. The Suite B modes can be set by using the
5883 strings "SUITEB128", "SUITEB192" or "SUITEB128ONLY" for the cipherstring.
5887 * New chain verification flags for Suite B levels of security. Check
5888 algorithms are acceptable when flags are set in X509_verify_cert.
5892 * Make tls1_check_chain return a set of flags indicating checks passed
5893 by a certificate chain. Add additional tests to handle client
5894 certificates: checks for matching certificate type and issuer name
5899 * If an attempt is made to use a signature algorithm not in the peer
5900 preference list abort the handshake. If client has no suitable
5901 signature algorithms in response to a certificate request do not
5902 use the certificate.
5906 * If server EC tmp key is not in client preference list abort handshake.
5910 * Add support for certificate stores in CERT structure. This makes it
5911 possible to have different stores per SSL structure or one store in
5912 the parent SSL_CTX. Include distinct stores for certificate chain
5913 verification and chain building. New ctrl SSL_CTRL_BUILD_CERT_CHAIN
5914 to build and store a certificate chain in CERT structure: returning
5915 an error if the chain cannot be built: this will allow applications
5916 to test if a chain is correctly configured.
5918 Note: if the CERT based stores are not set then the parent SSL_CTX
5919 store is used to retain compatibility with existing behaviour.
5923 * New function ssl_set_client_disabled to set a ciphersuite disabled
5924 mask based on the current session, check mask when sending client
5925 hello and checking the requested ciphersuite.
5929 * New ctrls to retrieve and set certificate types in a certificate
5930 request message. Print out received values in s_client. If certificate
5931 types is not set with custom values set sensible values based on
5932 supported signature algorithms.
5936 * Support for distinct client and server supported signature algorithms.
5940 * Add certificate callback. If set this is called whenever a certificate
5941 is required by client or server. An application can decide which
5942 certificate chain to present based on arbitrary criteria: for example
5943 supported signature algorithms. Add very simple example to s_server.
5944 This fixes many of the problems and restrictions of the existing client
5945 certificate callback: for example you can now clear an existing
5946 certificate and specify the whole chain.
5950 * Add new "valid_flags" field to CERT_PKEY structure which determines what
5951 the certificate can be used for (if anything). Set valid_flags field
5952 in new tls1_check_chain function. Simplify ssl_set_cert_masks which used
5953 to have similar checks in it.
5955 Add new "cert_flags" field to CERT structure and include a "strict mode".
5956 This enforces some TLS certificate requirements (such as only permitting
5957 certificate signature algorithms contained in the supported algorithms
5958 extension) which some implementations ignore: this option should be used
5959 with caution as it could cause interoperability issues.
5963 * Update and tidy signature algorithm extension processing. Work out
5964 shared signature algorithms based on preferences and peer algorithms
5965 and print them out in s_client and s_server. Abort handshake if no
5966 shared signature algorithms.
5970 * Add new functions to allow customised supported signature algorithms
5971 for SSL and SSL_CTX structures. Add options to s_client and s_server
5976 * New function SSL_certs_clear() to delete all references to certificates
5977 from an SSL structure. Before this once a certificate had been added
5978 it couldn't be removed.
5982 * Integrate hostname, email address and IP address checking with certificate
5983 verification. New verify options supporting checking in openssl utility.
5987 * Fixes and wildcard matching support to hostname and email checking
5988 functions. Add manual page.
5990 *Florian Weimer (Red Hat Product Security Team)*
5992 * New functions to check a hostname email or IP address against a
5993 certificate. Add options x509 utility to print results of checks against
5998 * Fix OCSP checking.
6000 *Rob Stradling <rob.stradling@comodo.com> and Ben Laurie*
6002 * Initial experimental support for explicitly trusted non-root CAs.
6003 OpenSSL still tries to build a complete chain to a root but if an
6004 intermediate CA has a trust setting included that is used. The first
6005 setting is used: whether to trust (e.g., -addtrust option to the x509
6010 * Add -trusted_first option which attempts to find certificates in the
6011 trusted store even if an untrusted chain is also supplied.
6015 * MIPS assembly pack updates: support for MIPS32r2 and SmartMIPS ASE,
6016 platform support for Linux and Android.
6020 * Support for linux-x32, ILP32 environment in x86_64 framework.
6024 * Experimental multi-implementation support for FIPS capable OpenSSL.
6025 When in FIPS mode the approved implementations are used as normal,
6026 when not in FIPS mode the internal unapproved versions are used instead.
6027 This means that the FIPS capable OpenSSL isn't forced to use the
6028 (often lower performance) FIPS implementations outside FIPS mode.
6032 * Transparently support X9.42 DH parameters when calling
6033 PEM_read_bio_DHparameters. This means existing applications can handle
6034 the new parameter format automatically.
6038 * Initial experimental support for X9.42 DH parameter format: mainly
6039 to support use of 'q' parameter for RFC5114 parameters.
6043 * Add DH parameters from RFC5114 including test data to dhtest.
6047 * Support for automatic EC temporary key parameter selection. If enabled
6048 the most preferred EC parameters are automatically used instead of
6049 hardcoded fixed parameters. Now a server just has to call:
6050 SSL_CTX_set_ecdh_auto(ctx, 1) and the server will automatically
6051 support ECDH and use the most appropriate parameters.
6055 * Enhance and tidy EC curve and point format TLS extension code. Use
6056 static structures instead of allocation if default values are used.
6057 New ctrls to set curves we wish to support and to retrieve shared curves.
6058 Print out shared curves in s_server. New options to s_server and s_client
6059 to set list of supported curves.
6063 * New ctrls to retrieve supported signature algorithms and
6064 supported curve values as an array of NIDs. Extend openssl utility
6065 to print out received values.
6069 * Add new APIs EC_curve_nist2nid and EC_curve_nid2nist which convert
6070 between NIDs and the more common NIST names such as "P-256". Enhance
6071 ecparam utility and ECC method to recognise the NIST names for curves.
6075 * Enhance SSL/TLS certificate chain handling to support different
6076 chains for each certificate instead of one chain in the parent SSL_CTX.
6080 * Support for fixed DH ciphersuite client authentication: where both
6081 server and client use DH certificates with common parameters.
6085 * Support for fixed DH ciphersuites: those requiring DH server
6090 * New function i2d_re_X509_tbs for re-encoding the TBS portion of
6092 Note: Related 1.0.2-beta specific macros X509_get_cert_info,
6093 X509_CINF_set_modified, X509_CINF_get_issuer, X509_CINF_get_extensions and
6094 X509_CINF_get_signature were reverted post internal team review.
6099 ### Changes between 1.0.1t and 1.0.1u [22 Sep 2016]
6101 * OCSP Status Request extension unbounded memory growth
6103 A malicious client can send an excessively large OCSP Status Request
6104 extension. If that client continually requests renegotiation, sending a
6105 large OCSP Status Request extension each time, then there will be unbounded
6106 memory growth on the server. This will eventually lead to a Denial Of
6107 Service attack through memory exhaustion. Servers with a default
6108 configuration are vulnerable even if they do not support OCSP. Builds using
6109 the "no-ocsp" build time option are not affected.
6111 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6116 * In order to mitigate the SWEET32 attack, the DES ciphers were moved from
6119 This issue was reported to OpenSSL Karthikeyan Bhargavan and Gaetan
6125 * OOB write in MDC2_Update()
6127 An overflow can occur in MDC2_Update() either if called directly or
6128 through the EVP_DigestUpdate() function using MDC2. If an attacker
6129 is able to supply very large amounts of input data after a previous
6130 call to EVP_EncryptUpdate() with a partial block then a length check
6131 can overflow resulting in a heap corruption.
6133 The amount of data needed is comparable to SIZE_MAX which is impractical
6136 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6141 * Malformed SHA512 ticket DoS
6143 If a server uses SHA512 for TLS session ticket HMAC it is vulnerable to a
6144 DoS attack where a malformed ticket will result in an OOB read which will
6147 The use of SHA512 in TLS session tickets is comparatively rare as it requires
6148 a custom server callback and ticket lookup mechanism.
6150 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6155 * OOB write in BN_bn2dec()
6157 The function BN_bn2dec() does not check the return value of BN_div_word().
6158 This can cause an OOB write if an application uses this function with an
6159 overly large BIGNUM. This could be a problem if an overly large certificate
6160 or CRL is printed out from an untrusted source. TLS is not affected because
6161 record limits will reject an oversized certificate before it is parsed.
6163 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6168 * OOB read in TS_OBJ_print_bio()
6170 The function TS_OBJ_print_bio() misuses OBJ_obj2txt(): the return value is
6171 the total length the OID text representation would use and not the amount
6172 of data written. This will result in OOB reads when large OIDs are
6175 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6180 * Pointer arithmetic undefined behaviour
6182 Avoid some undefined pointer arithmetic
6184 A common idiom in the codebase is to check limits in the following manner:
6187 Where "p" points to some malloc'd data of SIZE bytes and
6190 "len" here could be from some externally supplied data (e.g. from a TLS
6193 The rules of C pointer arithmetic are such that "p + len" is only well
6194 defined where len <= SIZE. Therefore the above idiom is actually
6195 undefined behaviour.
6197 For example this could cause problems if some malloc implementation
6198 provides an address for "p" such that "p + len" actually overflows for
6199 values of len that are too big and therefore p + len < limit.
6201 This issue was reported to OpenSSL by Guido Vranken
6206 * Constant time flag not preserved in DSA signing
6208 Operations in the DSA signing algorithm should run in constant time in
6209 order to avoid side channel attacks. A flaw in the OpenSSL DSA
6210 implementation means that a non-constant time codepath is followed for
6211 certain operations. This has been demonstrated through a cache-timing
6212 attack to be sufficient for an attacker to recover the private DSA key.
6214 This issue was reported by César Pereida (Aalto University), Billy Brumley
6215 (Tampere University of Technology), and Yuval Yarom (The University of
6216 Adelaide and NICTA).
6221 * DTLS buffered message DoS
6223 In a DTLS connection where handshake messages are delivered out-of-order
6224 those messages that OpenSSL is not yet ready to process will be buffered
6225 for later use. Under certain circumstances, a flaw in the logic means that
6226 those messages do not get removed from the buffer even though the handshake
6227 has been completed. An attacker could force up to approx. 15 messages to
6228 remain in the buffer when they are no longer required. These messages will
6229 be cleared when the DTLS connection is closed. The default maximum size for
6230 a message is 100k. Therefore the attacker could force an additional 1500k
6231 to be consumed per connection. By opening many simulataneous connections an
6232 attacker could cause a DoS attack through memory exhaustion.
6234 This issue was reported to OpenSSL by Quan Luo.
6239 * DTLS replay protection DoS
6241 A flaw in the DTLS replay attack protection mechanism means that records
6242 that arrive for future epochs update the replay protection "window" before
6243 the MAC for the record has been validated. This could be exploited by an
6244 attacker by sending a record for the next epoch (which does not have to
6245 decrypt or have a valid MAC), with a very large sequence number. This means
6246 that all subsequent legitimate packets are dropped causing a denial of
6247 service for a specific DTLS connection.
6249 This issue was reported to OpenSSL by the OCAP audit team.
6254 * Certificate message OOB reads
6256 In OpenSSL 1.0.2 and earlier some missing message length checks can result
6257 in OOB reads of up to 2 bytes beyond an allocated buffer. There is a
6258 theoretical DoS risk but this has not been observed in practice on common
6261 The messages affected are client certificate, client certificate request
6262 and server certificate. As a result the attack can only be performed
6263 against a client or a server which enables client authentication.
6265 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6270 ### Changes between 1.0.1s and 1.0.1t [3 May 2016]
6272 * Prevent padding oracle in AES-NI CBC MAC check
6274 A MITM attacker can use a padding oracle attack to decrypt traffic
6275 when the connection uses an AES CBC cipher and the server support
6278 This issue was introduced as part of the fix for Lucky 13 padding
6279 attack ([CVE-2013-0169]). The padding check was rewritten to be in
6280 constant time by making sure that always the same bytes are read and
6281 compared against either the MAC or padding bytes. But it no longer
6282 checked that there was enough data to have both the MAC and padding
6285 This issue was reported by Juraj Somorovsky using TLS-Attacker.
6290 * Fix EVP_EncodeUpdate overflow
6292 An overflow can occur in the EVP_EncodeUpdate() function which is used for
6293 Base64 encoding of binary data. If an attacker is able to supply very large
6294 amounts of input data then a length check can overflow resulting in a heap
6297 Internally to OpenSSL the EVP_EncodeUpdate() function is primarly used by
6298 the `PEM_write_bio*` family of functions. These are mainly used within the
6299 OpenSSL command line applications, so any application which processes data
6300 from an untrusted source and outputs it as a PEM file should be considered
6301 vulnerable to this issue. User applications that call these APIs directly
6302 with large amounts of untrusted data may also be vulnerable.
6304 This issue was reported by Guido Vranken.
6309 * Fix EVP_EncryptUpdate overflow
6311 An overflow can occur in the EVP_EncryptUpdate() function. If an attacker
6312 is able to supply very large amounts of input data after a previous call to
6313 EVP_EncryptUpdate() with a partial block then a length check can overflow
6314 resulting in a heap corruption. Following an analysis of all OpenSSL
6315 internal usage of the EVP_EncryptUpdate() function all usage is one of two
6316 forms. The first form is where the EVP_EncryptUpdate() call is known to be
6317 the first called function after an EVP_EncryptInit(), and therefore that
6318 specific call must be safe. The second form is where the length passed to
6319 EVP_EncryptUpdate() can be seen from the code to be some small value and
6320 therefore there is no possibility of an overflow. Since all instances are
6321 one of these two forms, it is believed that there can be no overflows in
6322 internal code due to this problem. It should be noted that
6323 EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths.
6324 Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances
6325 of these calls have also been analysed too and it is believed there are no
6326 instances in internal usage where an overflow could occur.
6328 This issue was reported by Guido Vranken.
6333 * Prevent ASN.1 BIO excessive memory allocation
6335 When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio()
6336 a short invalid encoding can casuse allocation of large amounts of memory
6337 potentially consuming excessive resources or exhausting memory.
6339 Any application parsing untrusted data through d2i BIO functions is
6340 affected. The memory based functions such as d2i_X509() are *not* affected.
6341 Since the memory based functions are used by the TLS library, TLS
6342 applications are not affected.
6344 This issue was reported by Brian Carpenter.
6351 ASN1 Strings that are over 1024 bytes can cause an overread in applications
6352 using the X509_NAME_oneline() function on EBCDIC systems. This could result
6353 in arbitrary stack data being returned in the buffer.
6355 This issue was reported by Guido Vranken.
6360 * Modify behavior of ALPN to invoke callback after SNI/servername
6361 callback, such that updates to the SSL_CTX affect ALPN.
6365 * Remove LOW from the DEFAULT cipher list. This removes singles DES from the
6370 * Only remove the SSLv2 methods with the no-ssl2-method option. When the
6371 methods are enabled and ssl2 is disabled the methods return NULL.
6375 ### Changes between 1.0.1r and 1.0.1s [1 Mar 2016]
6377 * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL.
6378 Builds that are not configured with "enable-weak-ssl-ciphers" will not
6379 provide any "EXPORT" or "LOW" strength ciphers.
6383 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2
6384 is by default disabled at build-time. Builds that are not configured with
6385 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used,
6386 users who want to negotiate SSLv2 via the version-flexible SSLv23_method()
6387 will need to explicitly call either of:
6389 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2);
6391 SSL_clear_options(ssl, SSL_OP_NO_SSLv2);
6393 as appropriate. Even if either of those is used, or the application
6394 explicitly uses the version-specific SSLv2_method() or its client and
6395 server variants, SSLv2 ciphers vulnerable to exhaustive search key
6396 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT
6397 ciphers, and SSLv2 56-bit DES are no longer available.
6402 * Fix a double-free in DSA code
6404 A double free bug was discovered when OpenSSL parses malformed DSA private
6405 keys and could lead to a DoS attack or memory corruption for applications
6406 that receive DSA private keys from untrusted sources. This scenario is
6409 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using
6415 * Disable SRP fake user seed to address a server memory leak.
6417 Add a new method SRP_VBASE_get1_by_user that handles the seed properly.
6419 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
6420 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user
6421 was changed to ignore the "fake user" SRP seed, even if the seed
6424 Users should use SRP_VBASE_get1_by_user instead. Note that in
6425 SRP_VBASE_get1_by_user, caller must free the returned value. Note
6426 also that even though configuring the SRP seed attempts to hide
6427 invalid usernames by continuing the handshake with fake
6428 credentials, this behaviour is not constant time and no strong
6429 guarantees are made that the handshake is indistinguishable from
6430 that of a valid user.
6435 * Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption
6437 In the BN_hex2bn function the number of hex digits is calculated using an
6438 int value `i`. Later `bn_expand` is called with a value of `i * 4`. For
6439 large values of `i` this can result in `bn_expand` not allocating any
6440 memory because `i * 4` is negative. This can leave the internal BIGNUM data
6441 field as NULL leading to a subsequent NULL ptr deref. For very large values
6442 of `i`, the calculation `i * 4` could be a positive value smaller than `i`.
6443 In this case memory is allocated to the internal BIGNUM data field, but it
6444 is insufficiently sized leading to heap corruption. A similar issue exists
6445 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn
6446 is ever called by user applications with very large untrusted hex/dec data.
6447 This is anticipated to be a rare occurrence.
6449 All OpenSSL internal usage of these functions use data that is not expected
6450 to be untrusted, e.g. config file data or application command line
6451 arguments. If user developed applications generate config file data based
6452 on untrusted data then it is possible that this could also lead to security
6453 consequences. This is also anticipated to be rare.
6455 This issue was reported to OpenSSL by Guido Vranken.
6460 * Fix memory issues in `BIO_*printf` functions
6462 The internal `fmtstr` function used in processing a "%s" format string in
6463 the `BIO_*printf` functions could overflow while calculating the length of a
6464 string and cause an OOB read when printing very long strings.
6466 Additionally the internal `doapr_outch` function can attempt to write to an
6467 OOB memory location (at an offset from the NULL pointer) in the event of a
6468 memory allocation failure. In 1.0.2 and below this could be caused where
6469 the size of a buffer to be allocated is greater than INT_MAX. E.g. this
6470 could be in processing a very long "%s" format string. Memory leaks can
6473 The first issue may mask the second issue dependent on compiler behaviour.
6474 These problems could enable attacks where large amounts of untrusted data
6475 is passed to the `BIO_*printf` functions. If applications use these functions
6476 in this way then they could be vulnerable. OpenSSL itself uses these
6477 functions when printing out human-readable dumps of ASN.1 data. Therefore
6478 applications that print this data could be vulnerable if the data is from
6479 untrusted sources. OpenSSL command line applications could also be
6480 vulnerable where they print out ASN.1 data, or if untrusted data is passed
6481 as command line arguments.
6483 Libssl is not considered directly vulnerable. Additionally certificates etc
6484 received via remote connections via libssl are also unlikely to be able to
6485 trigger these issues because of message size limits enforced within libssl.
6487 This issue was reported to OpenSSL Guido Vranken.
6492 * Side channel attack on modular exponentiation
6494 A side-channel attack was found which makes use of cache-bank conflicts on
6495 the Intel Sandy-Bridge microarchitecture which could lead to the recovery
6496 of RSA keys. The ability to exploit this issue is limited as it relies on
6497 an attacker who has control of code in a thread running on the same
6498 hyper-threaded core as the victim thread which is performing decryptions.
6500 This issue was reported to OpenSSL by Yuval Yarom, The University of
6501 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and
6502 Nadia Heninger, University of Pennsylvania with more information at
6503 <http://cachebleed.info>.
6508 * Change the req command to generate a 2048-bit RSA/DSA key by default,
6509 if no keysize is specified with default_bits. This fixes an
6510 omission in an earlier change that changed all RSA/DSA key generation
6511 commands to use 2048 bits by default.
6515 ### Changes between 1.0.1q and 1.0.1r [28 Jan 2016]
6517 * Protection for DH small subgroup attacks
6519 As a precautionary measure the SSL_OP_SINGLE_DH_USE option has been
6520 switched on by default and cannot be disabled. This could have some
6525 * SSLv2 doesn't block disabled ciphers
6527 A malicious client can negotiate SSLv2 ciphers that have been disabled on
6528 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
6529 been disabled, provided that the SSLv2 protocol was not also disabled via
6532 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
6533 and Sebastian Schinzel.
6538 * Reject DH handshakes with parameters shorter than 1024 bits.
6542 ### Changes between 1.0.1p and 1.0.1q [3 Dec 2015]
6544 * Certificate verify crash with missing PSS parameter
6546 The signature verification routines will crash with a NULL pointer
6547 dereference if presented with an ASN.1 signature using the RSA PSS
6548 algorithm and absent mask generation function parameter. Since these
6549 routines are used to verify certificate signature algorithms this can be
6550 used to crash any certificate verification operation and exploited in a
6551 DoS attack. Any application which performs certificate verification is
6552 vulnerable including OpenSSL clients and servers which enable client
6555 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
6560 * X509_ATTRIBUTE memory leak
6562 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
6563 memory. This structure is used by the PKCS#7 and CMS routines so any
6564 application which reads PKCS#7 or CMS data from untrusted sources is
6565 affected. SSL/TLS is not affected.
6567 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
6573 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
6574 This changes the decoding behaviour for some invalid messages,
6575 though the change is mostly in the more lenient direction, and
6576 legacy behaviour is preserved as much as possible.
6580 * In DSA_generate_parameters_ex, if the provided seed is too short,
6581 use a random seed, as already documented.
6583 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
6585 ### Changes between 1.0.1o and 1.0.1p [9 Jul 2015]
6587 * Alternate chains certificate forgery
6589 During certificate verfification, OpenSSL will attempt to find an
6590 alternative certificate chain if the first attempt to build such a chain
6591 fails. An error in the implementation of this logic can mean that an
6592 attacker could cause certain checks on untrusted certificates to be
6593 bypassed, such as the CA flag, enabling them to use a valid leaf
6594 certificate to act as a CA and "issue" an invalid certificate.
6596 This issue was reported to OpenSSL by Adam Langley/David Benjamin
6602 * Race condition handling PSK identify hint
6604 If PSK identity hints are received by a multi-threaded client then
6605 the values are wrongly updated in the parent SSL_CTX structure. This can
6606 result in a race condition potentially leading to a double free of the
6612 ### Changes between 1.0.1n and 1.0.1o [12 Jun 2015]
6614 * Fix HMAC ABI incompatibility. The previous version introduced an ABI
6615 incompatibility in the handling of HMAC. The previous ABI has now been
6618 ### Changes between 1.0.1m and 1.0.1n [11 Jun 2015]
6620 * Malformed ECParameters causes infinite loop
6622 When processing an ECParameters structure OpenSSL enters an infinite loop
6623 if the curve specified is over a specially malformed binary polynomial
6626 This can be used to perform denial of service against any
6627 system which processes public keys, certificate requests or
6628 certificates. This includes TLS clients and TLS servers with
6629 client authentication enabled.
6631 This issue was reported to OpenSSL by Joseph Barr-Pixton.
6636 * Exploitable out-of-bounds read in X509_cmp_time
6638 X509_cmp_time does not properly check the length of the ASN1_TIME
6639 string and can read a few bytes out of bounds. In addition,
6640 X509_cmp_time accepts an arbitrary number of fractional seconds in the
6643 An attacker can use this to craft malformed certificates and CRLs of
6644 various sizes and potentially cause a segmentation fault, resulting in
6645 a DoS on applications that verify certificates or CRLs. TLS clients
6646 that verify CRLs are affected. TLS clients and servers with client
6647 authentication enabled may be affected if they use custom verification
6650 This issue was reported to OpenSSL by Robert Swiecki (Google), and
6651 independently by Hanno Böck.
6656 * PKCS7 crash with missing EnvelopedContent
6658 The PKCS#7 parsing code does not handle missing inner EncryptedContent
6659 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
6660 with missing content and trigger a NULL pointer dereference on parsing.
6662 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
6663 structures from untrusted sources are affected. OpenSSL clients and
6664 servers are not affected.
6666 This issue was reported to OpenSSL by Michal Zalewski (Google).
6671 * CMS verify infinite loop with unknown hash function
6673 When verifying a signedData message the CMS code can enter an infinite loop
6674 if presented with an unknown hash function OID. This can be used to perform
6675 denial of service against any system which verifies signedData messages using
6677 This issue was reported to OpenSSL by Johannes Bauer.
6682 * Race condition handling NewSessionTicket
6684 If a NewSessionTicket is received by a multi-threaded client when attempting to
6685 reuse a previous ticket then a race condition can occur potentially leading to
6686 a double free of the ticket data.
6691 * Reject DH handshakes with parameters shorter than 768 bits.
6693 *Kurt Roeckx and Emilia Kasper*
6695 * dhparam: generate 2048-bit parameters by default.
6697 *Kurt Roeckx and Emilia Kasper*
6699 ### Changes between 1.0.1l and 1.0.1m [19 Mar 2015]
6701 * Segmentation fault in ASN1_TYPE_cmp fix
6703 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
6704 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
6705 certificate signature algorithm consistency this can be used to crash any
6706 certificate verification operation and exploited in a DoS attack. Any
6707 application which performs certificate verification is vulnerable including
6708 OpenSSL clients and servers which enable client authentication.
6713 * ASN.1 structure reuse memory corruption fix
6715 Reusing a structure in ASN.1 parsing may allow an attacker to cause
6716 memory corruption via an invalid write. Such reuse is and has been
6717 strongly discouraged and is believed to be rare.
6719 Applications that parse structures containing CHOICE or ANY DEFINED BY
6720 components may be affected. Certificate parsing (d2i_X509 and related
6721 functions) are however not affected. OpenSSL clients and servers are
6727 * PKCS7 NULL pointer dereferences fix
6729 The PKCS#7 parsing code does not handle missing outer ContentInfo
6730 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
6731 missing content and trigger a NULL pointer dereference on parsing.
6733 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
6734 otherwise parse PKCS#7 structures from untrusted sources are
6735 affected. OpenSSL clients and servers are not affected.
6737 This issue was reported to OpenSSL by Michal Zalewski (Google).
6742 * DoS via reachable assert in SSLv2 servers fix
6744 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
6745 servers that both support SSLv2 and enable export cipher suites by sending
6746 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
6748 This issue was discovered by Sean Burford (Google) and Emilia Käsper
6749 (OpenSSL development team).
6754 * Use After Free following d2i_ECPrivatekey error fix
6756 A malformed EC private key file consumed via the d2i_ECPrivateKey function
6757 could cause a use after free condition. This, in turn, could cause a double
6758 free in several private key parsing functions (such as d2i_PrivateKey
6759 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
6760 for applications that receive EC private keys from untrusted
6761 sources. This scenario is considered rare.
6763 This issue was discovered by the BoringSSL project and fixed in their
6769 * X509_to_X509_REQ NULL pointer deref fix
6771 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
6772 the certificate key is invalid. This function is rarely used in practice.
6774 This issue was discovered by Brian Carpenter.
6779 * Removed the export ciphers from the DEFAULT ciphers
6783 ### Changes between 1.0.1k and 1.0.1l [15 Jan 2015]
6785 * Build fixes for the Windows and OpenVMS platforms
6787 *Matt Caswell and Richard Levitte*
6789 ### Changes between 1.0.1j and 1.0.1k [8 Jan 2015]
6791 * Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
6792 message can cause a segmentation fault in OpenSSL due to a NULL pointer
6793 dereference. This could lead to a Denial Of Service attack. Thanks to
6794 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
6799 * Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
6800 dtls1_buffer_record function under certain conditions. In particular this
6801 could occur if an attacker sent repeated DTLS records with the same
6802 sequence number but for the next epoch. The memory leak could be exploited
6803 by an attacker in a Denial of Service attack through memory exhaustion.
6804 Thanks to Chris Mueller for reporting this issue.
6809 * Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
6810 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
6811 method would be set to NULL which could later result in a NULL pointer
6812 dereference. Thanks to Frank Schmirler for reporting this issue.
6817 * Abort handshake if server key exchange message is omitted for ephemeral
6820 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
6821 reporting this issue.
6826 * Remove non-export ephemeral RSA code on client and server. This code
6827 violated the TLS standard by allowing the use of temporary RSA keys in
6828 non-export ciphersuites and could be used by a server to effectively
6829 downgrade the RSA key length used to a value smaller than the server
6830 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
6831 INRIA or reporting this issue.
6836 * Fixed issue where DH client certificates are accepted without verification.
6837 An OpenSSL server will accept a DH certificate for client authentication
6838 without the certificate verify message. This effectively allows a client to
6839 authenticate without the use of a private key. This only affects servers
6840 which trust a client certificate authority which issues certificates
6841 containing DH keys: these are extremely rare and hardly ever encountered.
6842 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
6848 * Ensure that the session ID context of an SSL is updated when its
6849 SSL_CTX is updated via SSL_set_SSL_CTX.
6851 The session ID context is typically set from the parent SSL_CTX,
6852 and can vary with the CTX.
6856 * Fix various certificate fingerprint issues.
6858 By using non-DER or invalid encodings outside the signed portion of a
6859 certificate the fingerprint can be changed without breaking the signature.
6860 Although no details of the signed portion of the certificate can be changed
6861 this can cause problems with some applications: e.g. those using the
6862 certificate fingerprint for blacklists.
6864 1. Reject signatures with non zero unused bits.
6866 If the BIT STRING containing the signature has non zero unused bits reject
6867 the signature. All current signature algorithms require zero unused bits.
6869 2. Check certificate algorithm consistency.
6871 Check the AlgorithmIdentifier inside TBS matches the one in the
6872 certificate signature. NB: this will result in signature failure
6873 errors for some broken certificates.
6875 Thanks to Konrad Kraszewski from Google for reporting this issue.
6877 3. Check DSA/ECDSA signatures use DER.
6879 Re-encode DSA/ECDSA signatures and compare with the original received
6880 signature. Return an error if there is a mismatch.
6882 This will reject various cases including garbage after signature
6883 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
6884 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
6885 (negative or with leading zeroes).
6887 Further analysis was conducted and fixes were developed by Stephen Henson
6888 of the OpenSSL core team.
6894 * Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
6895 results on some platforms, including x86_64. This bug occurs at random
6896 with a very low probability, and is not known to be exploitable in any
6897 way, though its exact impact is difficult to determine. Thanks to Pieter
6898 Wuille (Blockstream) who reported this issue and also suggested an initial
6899 fix. Further analysis was conducted by the OpenSSL development team and
6900 Adam Langley of Google. The final fix was developed by Andy Polyakov of
6901 the OpenSSL core team.
6906 * Do not resume sessions on the server if the negotiated protocol
6907 version does not match the session's version. Resuming with a different
6908 version, while not strictly forbidden by the RFC, is of questionable
6909 sanity and breaks all known clients.
6911 *David Benjamin, Emilia Käsper*
6913 * Tighten handling of the ChangeCipherSpec (CCS) message: reject
6914 early CCS messages during renegotiation. (Note that because
6915 renegotiation is encrypted, this early CCS was not exploitable.)
6919 * Tighten client-side session ticket handling during renegotiation:
6920 ensure that the client only accepts a session ticket if the server sends
6921 the extension anew in the ServerHello. Previously, a TLS client would
6922 reuse the old extension state and thus accept a session ticket if one was
6923 announced in the initial ServerHello.
6925 Similarly, ensure that the client requires a session ticket if one
6926 was advertised in the ServerHello. Previously, a TLS client would
6927 ignore a missing NewSessionTicket message.
6931 ### Changes between 1.0.1i and 1.0.1j [15 Oct 2014]
6935 A flaw in the DTLS SRTP extension parsing code allows an attacker, who
6936 sends a carefully crafted handshake message, to cause OpenSSL to fail
6937 to free up to 64k of memory causing a memory leak. This could be
6938 exploited in a Denial Of Service attack. This issue affects OpenSSL
6939 1.0.1 server implementations for both SSL/TLS and DTLS regardless of
6940 whether SRTP is used or configured. Implementations of OpenSSL that
6941 have been compiled with OPENSSL_NO_SRTP defined are not affected.
6943 The fix was developed by the OpenSSL team.
6948 * Session Ticket Memory Leak.
6950 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
6951 integrity of that ticket is first verified. In the event of a session
6952 ticket integrity check failing, OpenSSL will fail to free memory
6953 causing a memory leak. By sending a large number of invalid session
6954 tickets an attacker could exploit this issue in a Denial Of Service
6960 * Build option no-ssl3 is incomplete.
6962 When OpenSSL is configured with "no-ssl3" as a build option, servers
6963 could accept and complete a SSL 3.0 handshake, and clients could be
6964 configured to send them.
6967 *Akamai and the OpenSSL team*
6969 * Add support for TLS_FALLBACK_SCSV.
6970 Client applications doing fallback retries should call
6971 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
6974 *Adam Langley, Bodo Moeller*
6976 * Add additional DigestInfo checks.
6978 Re-encode DigestInto in DER and check against the original when
6979 verifying RSA signature: this will reject any improperly encoded
6980 DigestInfo structures.
6982 Note: this is a precautionary measure and no attacks are currently known.
6986 ### Changes between 1.0.1h and 1.0.1i [6 Aug 2014]
6988 * Fix SRP buffer overrun vulnerability. Invalid parameters passed to the
6989 SRP code can be overrun an internal buffer. Add sanity check that
6990 g, A, B < N to SRP code.
6992 Thanks to Sean Devlin and Watson Ladd of Cryptography Services, NCC
6993 Group for discovering this issue.
6998 * A flaw in the OpenSSL SSL/TLS server code causes the server to negotiate
6999 TLS 1.0 instead of higher protocol versions when the ClientHello message
7000 is badly fragmented. This allows a man-in-the-middle attacker to force a
7001 downgrade to TLS 1.0 even if both the server and the client support a
7002 higher protocol version, by modifying the client's TLS records.
7004 Thanks to David Benjamin and Adam Langley (Google) for discovering and
7005 researching this issue.
7010 * OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
7011 to a denial of service attack. A malicious server can crash the client
7012 with a null pointer dereference (read) by specifying an anonymous (EC)DH
7013 ciphersuite and sending carefully crafted handshake messages.
7015 Thanks to Felix Gröbert (Google) for discovering and researching this
7021 * By sending carefully crafted DTLS packets an attacker could cause openssl
7022 to leak memory. This can be exploited through a Denial of Service attack.
7023 Thanks to Adam Langley for discovering and researching this issue.
7028 * An attacker can force openssl to consume large amounts of memory whilst
7029 processing DTLS handshake messages. This can be exploited through a
7030 Denial of Service attack.
7031 Thanks to Adam Langley for discovering and researching this issue.
7036 * An attacker can force an error condition which causes openssl to crash
7037 whilst processing DTLS packets due to memory being freed twice. This
7038 can be exploited through a Denial of Service attack.
7039 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
7045 * If a multithreaded client connects to a malicious server using a resumed
7046 session and the server sends an ec point format extension it could write
7047 up to 255 bytes to freed memory.
7049 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
7055 * A malicious server can crash an OpenSSL client with a null pointer
7056 dereference (read) by specifying an SRP ciphersuite even though it was not
7057 properly negotiated with the client. This can be exploited through a
7058 Denial of Service attack.
7060 Thanks to Joonas Kuorilehto and Riku Hietamäki (Codenomicon) for
7061 discovering and researching this issue.
7066 * A flaw in OBJ_obj2txt may cause pretty printing functions such as
7067 X509_name_oneline, X509_name_print_ex et al. to leak some information
7068 from the stack. Applications may be affected if they echo pretty printing
7069 output to the attacker.
7071 Thanks to Ivan Fratric (Google) for discovering this issue.
7074 *Emilia Käsper, and Steve Henson*
7076 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
7077 for corner cases. (Certain input points at infinity could lead to
7078 bogus results, with non-infinity inputs mapped to infinity too.)
7082 ### Changes between 1.0.1g and 1.0.1h [5 Jun 2014]
7084 * Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
7085 handshake can force the use of weak keying material in OpenSSL
7086 SSL/TLS clients and servers.
7088 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
7089 researching this issue. ([CVE-2014-0224])
7091 *KIKUCHI Masashi, Steve Henson*
7093 * Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
7094 OpenSSL DTLS client the code can be made to recurse eventually crashing
7097 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
7100 *Imre Rad, Steve Henson*
7102 * Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
7103 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
7104 client or server. This is potentially exploitable to run arbitrary
7105 code on a vulnerable client or server.
7107 Thanks to Jüri Aedla for reporting this issue. ([CVE-2014-0195])
7109 *Jüri Aedla, Steve Henson*
7111 * Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
7112 are subject to a denial of service attack.
7114 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
7115 this issue. ([CVE-2014-3470])
7117 *Felix Gröbert, Ivan Fratric, Steve Henson*
7119 * Harmonize version and its documentation. -f flag is used to display
7122 *mancha <mancha1@zoho.com>*
7124 * Fix eckey_priv_encode so it immediately returns an error upon a failure
7125 in i2d_ECPrivateKey.
7127 *mancha <mancha1@zoho.com>*
7129 * Fix some double frees. These are not thought to be exploitable.
7131 *mancha <mancha1@zoho.com>*
7133 ### Changes between 1.0.1f and 1.0.1g [7 Apr 2014]
7135 * A missing bounds check in the handling of the TLS heartbeat extension
7136 can be used to reveal up to 64k of memory to a connected client or
7139 Thanks for Neel Mehta of Google Security for discovering this bug and to
7140 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
7141 preparing the fix ([CVE-2014-0160])
7143 *Adam Langley, Bodo Moeller*
7145 * Fix for the attack described in the paper "Recovering OpenSSL
7146 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
7147 by Yuval Yarom and Naomi Benger. Details can be obtained from:
7148 <http://eprint.iacr.org/2014/140>
7150 Thanks to Yuval Yarom and Naomi Benger for discovering this
7151 flaw and to Yuval Yarom for supplying a fix ([CVE-2014-0076])
7153 *Yuval Yarom and Naomi Benger*
7155 * TLS pad extension: draft-agl-tls-padding-03
7157 Workaround for the "TLS hang bug" (see FAQ and PR#2771): if the
7158 TLS client Hello record length value would otherwise be > 255 and
7159 less that 512 pad with a dummy extension containing zeroes so it
7160 is at least 512 bytes long.
7162 *Adam Langley, Steve Henson*
7164 ### Changes between 1.0.1e and 1.0.1f [6 Jan 2014]
7166 * Fix for TLS record tampering bug. A carefully crafted invalid
7167 handshake could crash OpenSSL with a NULL pointer exception.
7168 Thanks to Anton Johansson for reporting this issues.
7171 * Keep original DTLS digest and encryption contexts in retransmission
7172 structures so we can use the previous session parameters if they need
7173 to be resent. ([CVE-2013-6450])
7177 * Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
7178 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
7179 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
7180 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
7181 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
7182 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
7184 *Rob Stradling, Adam Langley*
7186 ### Changes between 1.0.1d and 1.0.1e [11 Feb 2013]
7188 * Correct fix for CVE-2013-0169. The original didn't work on AES-NI
7189 supporting platforms or when small records were transferred.
7191 *Andy Polyakov, Steve Henson*
7193 ### Changes between 1.0.1c and 1.0.1d [5 Feb 2013]
7195 * Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
7197 This addresses the flaw in CBC record processing discovered by
7198 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
7199 at: <http://www.isg.rhul.ac.uk/tls/>
7201 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
7202 Security Group at Royal Holloway, University of London
7203 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
7204 Emilia Käsper for the initial patch.
7207 *Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson*
7209 * Fix flaw in AESNI handling of TLS 1.2 and 1.1 records for CBC mode
7210 ciphersuites which can be exploited in a denial of service attack.
7211 Thanks go to and to Adam Langley <agl@chromium.org> for discovering
7212 and detecting this bug and to Wolfgang Ettlinger
7213 <wolfgang.ettlinger@gmail.com> for independently discovering this issue.
7218 * Return an error when checking OCSP signatures when key is NULL.
7219 This fixes a DoS attack. ([CVE-2013-0166])
7223 * Make openssl verify return errors.
7225 *Chris Palmer <palmer@google.com> and Ben Laurie*
7227 * Call OCSP Stapling callback after ciphersuite has been chosen, so
7228 the right response is stapled. Also change SSL_get_certificate()
7229 so it returns the certificate actually sent.
7230 See <http://rt.openssl.org/Ticket/Display.html?id=2836>.
7232 *Rob Stradling <rob.stradling@comodo.com>*
7234 * Fix possible deadlock when decoding public keys.
7238 * Don't use TLS 1.0 record version number in initial client hello
7243 ### Changes between 1.0.1b and 1.0.1c [10 May 2012]
7245 * Sanity check record length before skipping explicit IV in TLS
7246 1.2, 1.1 and DTLS to fix DoS attack.
7248 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
7249 fuzzing as a service testing platform.
7254 * Initialise tkeylen properly when encrypting CMS messages.
7255 Thanks to Solar Designer of Openwall for reporting this issue.
7259 * In FIPS mode don't try to use composite ciphers as they are not
7264 ### Changes between 1.0.1a and 1.0.1b [26 Apr 2012]
7266 * OpenSSL 1.0.0 sets SSL_OP_ALL to 0x80000FFFL and OpenSSL 1.0.1 and
7267 1.0.1a set SSL_OP_NO_TLSv1_1 to 0x00000400L which would unfortunately
7268 mean any application compiled against OpenSSL 1.0.0 headers setting
7269 SSL_OP_ALL would also set SSL_OP_NO_TLSv1_1, unintentionally disabling
7270 TLS 1.1 also. Fix this by changing the value of SSL_OP_NO_TLSv1_1 to
7271 0x10000000L Any application which was previously compiled against
7272 OpenSSL 1.0.1 or 1.0.1a headers and which cares about SSL_OP_NO_TLSv1_1
7273 will need to be recompiled as a result. Letting be results in
7274 inability to disable specifically TLS 1.1 and in client context,
7275 in unlike event, limit maximum offered version to TLS 1.0 [see below].
7279 * In order to ensure interoperability SSL_OP_NO_protocolX does not
7280 disable just protocol X, but all protocols above X *if* there are
7281 protocols *below* X still enabled. In more practical terms it means
7282 that if application wants to disable TLS1.0 in favor of TLS1.1 and
7283 above, it's not sufficient to pass `SSL_OP_NO_TLSv1`, one has to pass
7284 `SSL_OP_NO_TLSv1|SSL_OP_NO_SSLv3|SSL_OP_NO_SSLv2`. This applies to
7289 ### Changes between 1.0.1 and 1.0.1a [19 Apr 2012]
7291 * Check for potentially exploitable overflows in asn1_d2i_read_bio
7292 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
7293 in CRYPTO_realloc_clean.
7295 Thanks to Tavis Ormandy, Google Security Team, for discovering this
7296 issue and to Adam Langley <agl@chromium.org> for fixing it.
7299 *Adam Langley (Google), Tavis Ormandy, Google Security Team*
7301 * Don't allow TLS 1.2 SHA-256 ciphersuites in TLS 1.0, 1.1 connections.
7305 * Workarounds for some broken servers that "hang" if a client hello
7306 record length exceeds 255 bytes.
7308 1. Do not use record version number > TLS 1.0 in initial client
7309 hello: some (but not all) hanging servers will now work.
7310 2. If we set OPENSSL_MAX_TLS1_2_CIPHER_LENGTH this will truncate
7311 the number of ciphers sent in the client hello. This should be
7312 set to an even number, such as 50, for example by passing:
7313 -DOPENSSL_MAX_TLS1_2_CIPHER_LENGTH=50 to config or Configure.
7314 Most broken servers should now work.
7315 3. If all else fails setting OPENSSL_NO_TLS1_2_CLIENT will disable
7316 TLS 1.2 client support entirely.
7320 * Fix SEGV in Vector Permutation AES module observed in OpenSSH.
7324 ### Changes between 1.0.0h and 1.0.1 [14 Mar 2012]
7326 * Add compatibility with old MDC2 signatures which use an ASN1 OCTET
7327 STRING form instead of a DigestInfo.
7331 * The format used for MDC2 RSA signatures is inconsistent between EVP
7332 and the RSA_sign/RSA_verify functions. This was made more apparent when
7333 OpenSSL used RSA_sign/RSA_verify for some RSA signatures in particular
7334 those which went through EVP_PKEY_METHOD in 1.0.0 and later. Detect
7335 the correct format in RSA_verify so both forms transparently work.
7339 * Some servers which support TLS 1.0 can choke if we initially indicate
7340 support for TLS 1.2 and later renegotiate using TLS 1.0 in the RSA
7341 encrypted premaster secret. As a workaround use the maximum permitted
7342 client version in client hello, this should keep such servers happy
7343 and still work with previous versions of OpenSSL.
7347 * Add support for TLS/DTLS heartbeats.
7349 *Robin Seggelmann <seggelmann@fh-muenster.de>*
7351 * Add support for SCTP.
7353 *Robin Seggelmann <seggelmann@fh-muenster.de>*
7355 * Improved PRNG seeding for VOS.
7357 *Paul Green <Paul.Green@stratus.com>*
7359 * Extensive assembler packs updates, most notably:
7361 - x86[_64]: AES-NI, PCLMULQDQ, RDRAND support;
7362 - x86[_64]: SSSE3 support (SHA1, vector-permutation AES);
7363 - x86_64: bit-sliced AES implementation;
7364 - ARM: NEON support, contemporary platforms optimizations;
7365 - s390x: z196 support;
7366 - `*`: GHASH and GF(2^m) multiplication implementations;
7370 * Make TLS-SRP code conformant with RFC 5054 API cleanup
7371 (removal of unnecessary code)
7373 *Peter Sylvester <peter.sylvester@edelweb.fr>*
7375 * Add TLS key material exporter from RFC 5705.
7379 * Add DTLS-SRTP negotiation from RFC 5764.
7383 * Add Next Protocol Negotiation,
7384 <http://tools.ietf.org/html/draft-agl-tls-nextprotoneg-00>. Can be
7385 disabled with a no-npn flag to config or Configure. Code donated
7388 *Adam Langley <agl@google.com> and Ben Laurie*
7390 * Add optional 64-bit optimized implementations of elliptic curves NIST-P224,
7391 NIST-P256, NIST-P521, with constant-time single point multiplication on
7392 typical inputs. Compiler support for the nonstandard type `__uint128_t` is
7393 required to use this (present in gcc 4.4 and later, for 64-bit builds).
7394 Code made available under Apache License version 2.0.
7396 Specify "enable-ec_nistp_64_gcc_128" on the Configure (or config) command
7397 line to include this in your build of OpenSSL, and run "make depend" (or
7398 "make update"). This enables the following EC_METHODs:
7400 EC_GFp_nistp224_method()
7401 EC_GFp_nistp256_method()
7402 EC_GFp_nistp521_method()
7404 EC_GROUP_new_by_curve_name() will automatically use these (while
7405 EC_GROUP_new_curve_GFp() currently prefers the more flexible
7408 *Emilia Käsper, Adam Langley, Bodo Moeller (Google)*
7410 * Use type ossl_ssize_t instead of ssize_t which isn't available on
7411 all platforms. Move ssize_t definition from e_os.h to the public
7412 header file e_os2.h as it now appears in public header file cms.h
7416 * New -sigopt option to the ca, req and x509 utilities. Additional
7417 signature parameters can be passed using this option and in
7422 * Add RSA PSS signing function. This will generate and set the
7423 appropriate AlgorithmIdentifiers for PSS based on those in the
7424 corresponding EVP_MD_CTX structure. No application support yet.
7428 * Support for companion algorithm specific ASN1 signing routines.
7429 New function ASN1_item_sign_ctx() signs a pre-initialised
7430 EVP_MD_CTX structure and sets AlgorithmIdentifiers based on
7431 the appropriate parameters.
7435 * Add new algorithm specific ASN1 verification initialisation function
7436 to EVP_PKEY_ASN1_METHOD: this is not in EVP_PKEY_METHOD since the ASN1
7437 handling will be the same no matter what EVP_PKEY_METHOD is used.
7438 Add a PSS handler to support verification of PSS signatures: checked
7439 against a number of sample certificates.
7443 * Add signature printing for PSS. Add PSS OIDs.
7445 *Steve Henson, Martin Kaiser <lists@kaiser.cx>*
7447 * Add algorithm specific signature printing. An individual ASN1 method
7448 can now print out signatures instead of the standard hex dump.
7450 More complex signatures (e.g. PSS) can print out more meaningful
7451 information. Include DSA version that prints out the signature
7456 * Password based recipient info support for CMS library: implementing
7461 * Split password based encryption into PBES2 and PBKDF2 functions. This
7462 neatly separates the code into cipher and PBE sections and is required
7463 for some algorithms that split PBES2 into separate pieces (such as
7464 password based CMS).
7468 * Session-handling fixes:
7469 - Fix handling of connections that are resuming with a session ID,
7470 but also support Session Tickets.
7471 - Fix a bug that suppressed issuing of a new ticket if the client
7472 presented a ticket with an expired session.
7473 - Try to set the ticket lifetime hint to something reasonable.
7474 - Make tickets shorter by excluding irrelevant information.
7475 - On the client side, don't ignore renewed tickets.
7477 *Adam Langley, Bodo Moeller (Google)*
7479 * Fix PSK session representation.
7483 * Add RC4-MD5 and AESNI-SHA1 "stitched" implementations.
7485 This work was sponsored by Intel.
7489 * Add GCM support to TLS library. Some custom code is needed to split
7490 the IV between the fixed (from PRF) and explicit (from TLS record)
7491 portions. This adds all GCM ciphersuites supported by RFC5288 and
7492 RFC5289. Generalise some `AES*` cipherstrings to include GCM and
7493 add a special AESGCM string for GCM only.
7497 * Expand range of ctrls for AES GCM. Permit setting invocation
7498 field on decrypt and retrieval of invocation field only on encrypt.
7502 * Add HMAC ECC ciphersuites from RFC5289. Include SHA384 PRF support.
7503 As required by RFC5289 these ciphersuites cannot be used if for
7504 versions of TLS earlier than 1.2.
7508 * For FIPS capable OpenSSL interpret a NULL default public key method
7509 as unset and return the appropriate default but do *not* set the default.
7510 This means we can return the appropriate method in applications that
7511 switch between FIPS and non-FIPS modes.
7515 * Redirect HMAC and CMAC operations to FIPS module in FIPS mode. If an
7516 ENGINE is used then we cannot handle that in the FIPS module so we
7517 keep original code iff non-FIPS operations are allowed.
7521 * Add -attime option to openssl utilities.
7523 *Peter Eckersley <pde@eff.org>, Ben Laurie and Steve Henson*
7525 * Redirect DSA and DH operations to FIPS module in FIPS mode.
7529 * Redirect ECDSA and ECDH operations to FIPS module in FIPS mode. Also use
7530 FIPS EC methods unconditionally for now.
7534 * New build option no-ec2m to disable characteristic 2 code.
7538 * Backport libcrypto audit of return value checking from 1.1.0-dev; not
7539 all cases can be covered as some introduce binary incompatibilities.
7543 * Redirect RSA operations to FIPS module including keygen,
7544 encrypt, decrypt, sign and verify. Block use of non FIPS RSA methods.
7548 * Add similar low-level API blocking to ciphers.
7552 * low-level digest APIs are not approved in FIPS mode: any attempt
7553 to use these will cause a fatal error. Applications that *really* want
7554 to use them can use the `private_*` version instead.
7558 * Redirect cipher operations to FIPS module for FIPS builds.
7562 * Redirect digest operations to FIPS module for FIPS builds.
7566 * Update build system to add "fips" flag which will link in fipscanister.o
7567 for static and shared library builds embedding a signature if needed.
7571 * Output TLS supported curves in preference order instead of numerical
7572 order. This is currently hardcoded for the highest order curves first.
7573 This should be configurable so applications can judge speed vs strength.
7577 * Add TLS v1.2 server support for client authentication.
7581 * Add support for FIPS mode in ssl library: disable SSLv3, non-FIPS ciphers
7586 * Functions FIPS_mode_set() and FIPS_mode() which call the underlying
7587 FIPS modules versions.
7591 * Add TLS v1.2 client side support for client authentication. Keep cache
7592 of handshake records longer as we don't know the hash algorithm to use
7593 until after the certificate request message is received.
7597 * Initial TLS v1.2 client support. Add a default signature algorithms
7598 extension including all the algorithms we support. Parse new signature
7599 format in client key exchange. Relax some ECC signing restrictions for
7600 TLS v1.2 as indicated in RFC5246.
7604 * Add server support for TLS v1.2 signature algorithms extension. Switch
7605 to new signature format when needed using client digest preference.
7606 All server ciphersuites should now work correctly in TLS v1.2. No client
7607 support yet and no support for client certificates.
7611 * Initial TLS v1.2 support. Add new SHA256 digest to ssl code, switch
7612 to SHA256 for PRF when using TLS v1.2 and later. Add new SHA256 based
7613 ciphersuites. At present only RSA key exchange ciphersuites work with
7614 TLS v1.2. Add new option for TLS v1.2 replacing the old and obsolete
7615 SSL_OP_PKCS1_CHECK flags with SSL_OP_NO_TLSv1_2. New TLSv1.2 methods
7616 and version checking.
7620 * New option OPENSSL_NO_SSL_INTERN. If an application can be compiled
7621 with this defined it will not be affected by any changes to ssl internal
7622 structures. Add several utility functions to allow openssl application
7623 to work with OPENSSL_NO_SSL_INTERN defined.
7627 * A long standing patch to add support for SRP from EdelWeb (Peter
7628 Sylvester and Christophe Renou) was integrated.
7629 *Christophe Renou <christophe.renou@edelweb.fr>, Peter Sylvester
7630 <peter.sylvester@edelweb.fr>, Tom Wu <tjw@cs.stanford.edu>, and
7633 * Add functions to copy EVP_PKEY_METHOD and retrieve flags and id.
7637 * Permit abbreviated handshakes when renegotiating using the function
7638 SSL_renegotiate_abbreviated().
7640 *Robin Seggelmann <seggelmann@fh-muenster.de>*
7642 * Add call to ENGINE_register_all_complete() to
7643 ENGINE_load_builtin_engines(), so some implementations get used
7644 automatically instead of needing explicit application support.
7648 * Add support for TLS key exporter as described in RFC5705.
7650 *Robin Seggelmann <seggelmann@fh-muenster.de>, Steve Henson*
7652 * Initial TLSv1.1 support. Since TLSv1.1 is very similar to TLS v1.0 only
7653 a few changes are required:
7655 Add SSL_OP_NO_TLSv1_1 flag.
7656 Add TLSv1_1 methods.
7657 Update version checking logic to handle version 1.1.
7658 Add explicit IV handling (ported from DTLS code).
7659 Add command line options to s_client/s_server.
7666 ### Changes between 1.0.0s and 1.0.0t [3 Dec 2015]
7668 * X509_ATTRIBUTE memory leak
7670 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
7671 memory. This structure is used by the PKCS#7 and CMS routines so any
7672 application which reads PKCS#7 or CMS data from untrusted sources is
7673 affected. SSL/TLS is not affected.
7675 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
7681 * Race condition handling PSK identify hint
7683 If PSK identity hints are received by a multi-threaded client then
7684 the values are wrongly updated in the parent SSL_CTX structure. This can
7685 result in a race condition potentially leading to a double free of the
7691 ### Changes between 1.0.0r and 1.0.0s [11 Jun 2015]
7693 * Malformed ECParameters causes infinite loop
7695 When processing an ECParameters structure OpenSSL enters an infinite loop
7696 if the curve specified is over a specially malformed binary polynomial
7699 This can be used to perform denial of service against any
7700 system which processes public keys, certificate requests or
7701 certificates. This includes TLS clients and TLS servers with
7702 client authentication enabled.
7704 This issue was reported to OpenSSL by Joseph Barr-Pixton.
7709 * Exploitable out-of-bounds read in X509_cmp_time
7711 X509_cmp_time does not properly check the length of the ASN1_TIME
7712 string and can read a few bytes out of bounds. In addition,
7713 X509_cmp_time accepts an arbitrary number of fractional seconds in the
7716 An attacker can use this to craft malformed certificates and CRLs of
7717 various sizes and potentially cause a segmentation fault, resulting in
7718 a DoS on applications that verify certificates or CRLs. TLS clients
7719 that verify CRLs are affected. TLS clients and servers with client
7720 authentication enabled may be affected if they use custom verification
7723 This issue was reported to OpenSSL by Robert Swiecki (Google), and
7724 independently by Hanno Böck.
7729 * PKCS7 crash with missing EnvelopedContent
7731 The PKCS#7 parsing code does not handle missing inner EncryptedContent
7732 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
7733 with missing content and trigger a NULL pointer dereference on parsing.
7735 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
7736 structures from untrusted sources are affected. OpenSSL clients and
7737 servers are not affected.
7739 This issue was reported to OpenSSL by Michal Zalewski (Google).
7744 * CMS verify infinite loop with unknown hash function
7746 When verifying a signedData message the CMS code can enter an infinite loop
7747 if presented with an unknown hash function OID. This can be used to perform
7748 denial of service against any system which verifies signedData messages using
7750 This issue was reported to OpenSSL by Johannes Bauer.
7755 * Race condition handling NewSessionTicket
7757 If a NewSessionTicket is received by a multi-threaded client when attempting to
7758 reuse a previous ticket then a race condition can occur potentially leading to
7759 a double free of the ticket data.
7764 ### Changes between 1.0.0q and 1.0.0r [19 Mar 2015]
7766 * Segmentation fault in ASN1_TYPE_cmp fix
7768 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
7769 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
7770 certificate signature algorithm consistency this can be used to crash any
7771 certificate verification operation and exploited in a DoS attack. Any
7772 application which performs certificate verification is vulnerable including
7773 OpenSSL clients and servers which enable client authentication.
7778 * ASN.1 structure reuse memory corruption fix
7780 Reusing a structure in ASN.1 parsing may allow an attacker to cause
7781 memory corruption via an invalid write. Such reuse is and has been
7782 strongly discouraged and is believed to be rare.
7784 Applications that parse structures containing CHOICE or ANY DEFINED BY
7785 components may be affected. Certificate parsing (d2i_X509 and related
7786 functions) are however not affected. OpenSSL clients and servers are
7792 * PKCS7 NULL pointer dereferences fix
7794 The PKCS#7 parsing code does not handle missing outer ContentInfo
7795 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
7796 missing content and trigger a NULL pointer dereference on parsing.
7798 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
7799 otherwise parse PKCS#7 structures from untrusted sources are
7800 affected. OpenSSL clients and servers are not affected.
7802 This issue was reported to OpenSSL by Michal Zalewski (Google).
7807 * DoS via reachable assert in SSLv2 servers fix
7809 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
7810 servers that both support SSLv2 and enable export cipher suites by sending
7811 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
7813 This issue was discovered by Sean Burford (Google) and Emilia Käsper
7814 (OpenSSL development team).
7819 * Use After Free following d2i_ECPrivatekey error fix
7821 A malformed EC private key file consumed via the d2i_ECPrivateKey function
7822 could cause a use after free condition. This, in turn, could cause a double
7823 free in several private key parsing functions (such as d2i_PrivateKey
7824 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
7825 for applications that receive EC private keys from untrusted
7826 sources. This scenario is considered rare.
7828 This issue was discovered by the BoringSSL project and fixed in their
7834 * X509_to_X509_REQ NULL pointer deref fix
7836 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
7837 the certificate key is invalid. This function is rarely used in practice.
7839 This issue was discovered by Brian Carpenter.
7844 * Removed the export ciphers from the DEFAULT ciphers
7848 ### Changes between 1.0.0p and 1.0.0q [15 Jan 2015]
7850 * Build fixes for the Windows and OpenVMS platforms
7852 *Matt Caswell and Richard Levitte*
7854 ### Changes between 1.0.0o and 1.0.0p [8 Jan 2015]
7856 * Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
7857 message can cause a segmentation fault in OpenSSL due to a NULL pointer
7858 dereference. This could lead to a Denial Of Service attack. Thanks to
7859 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
7864 * Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
7865 dtls1_buffer_record function under certain conditions. In particular this
7866 could occur if an attacker sent repeated DTLS records with the same
7867 sequence number but for the next epoch. The memory leak could be exploited
7868 by an attacker in a Denial of Service attack through memory exhaustion.
7869 Thanks to Chris Mueller for reporting this issue.
7874 * Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
7875 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
7876 method would be set to NULL which could later result in a NULL pointer
7877 dereference. Thanks to Frank Schmirler for reporting this issue.
7882 * Abort handshake if server key exchange message is omitted for ephemeral
7885 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
7886 reporting this issue.
7891 * Remove non-export ephemeral RSA code on client and server. This code
7892 violated the TLS standard by allowing the use of temporary RSA keys in
7893 non-export ciphersuites and could be used by a server to effectively
7894 downgrade the RSA key length used to a value smaller than the server
7895 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
7896 INRIA or reporting this issue.
7901 * Fixed issue where DH client certificates are accepted without verification.
7902 An OpenSSL server will accept a DH certificate for client authentication
7903 without the certificate verify message. This effectively allows a client to
7904 authenticate without the use of a private key. This only affects servers
7905 which trust a client certificate authority which issues certificates
7906 containing DH keys: these are extremely rare and hardly ever encountered.
7907 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
7913 * Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
7914 results on some platforms, including x86_64. This bug occurs at random
7915 with a very low probability, and is not known to be exploitable in any
7916 way, though its exact impact is difficult to determine. Thanks to Pieter
7917 Wuille (Blockstream) who reported this issue and also suggested an initial
7918 fix. Further analysis was conducted by the OpenSSL development team and
7919 Adam Langley of Google. The final fix was developed by Andy Polyakov of
7920 the OpenSSL core team.
7925 * Fix various certificate fingerprint issues.
7927 By using non-DER or invalid encodings outside the signed portion of a
7928 certificate the fingerprint can be changed without breaking the signature.
7929 Although no details of the signed portion of the certificate can be changed
7930 this can cause problems with some applications: e.g. those using the
7931 certificate fingerprint for blacklists.
7933 1. Reject signatures with non zero unused bits.
7935 If the BIT STRING containing the signature has non zero unused bits reject
7936 the signature. All current signature algorithms require zero unused bits.
7938 2. Check certificate algorithm consistency.
7940 Check the AlgorithmIdentifier inside TBS matches the one in the
7941 certificate signature. NB: this will result in signature failure
7942 errors for some broken certificates.
7944 Thanks to Konrad Kraszewski from Google for reporting this issue.
7946 3. Check DSA/ECDSA signatures use DER.
7948 Reencode DSA/ECDSA signatures and compare with the original received
7949 signature. Return an error if there is a mismatch.
7951 This will reject various cases including garbage after signature
7952 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
7953 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
7954 (negative or with leading zeroes).
7956 Further analysis was conducted and fixes were developed by Stephen Henson
7957 of the OpenSSL core team.
7963 ### Changes between 1.0.0n and 1.0.0o [15 Oct 2014]
7965 * Session Ticket Memory Leak.
7967 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
7968 integrity of that ticket is first verified. In the event of a session
7969 ticket integrity check failing, OpenSSL will fail to free memory
7970 causing a memory leak. By sending a large number of invalid session
7971 tickets an attacker could exploit this issue in a Denial Of Service
7977 * Build option no-ssl3 is incomplete.
7979 When OpenSSL is configured with "no-ssl3" as a build option, servers
7980 could accept and complete a SSL 3.0 handshake, and clients could be
7981 configured to send them.
7984 *Akamai and the OpenSSL team*
7986 * Add support for TLS_FALLBACK_SCSV.
7987 Client applications doing fallback retries should call
7988 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
7991 *Adam Langley, Bodo Moeller*
7993 * Add additional DigestInfo checks.
7995 Reencode DigestInto in DER and check against the original when
7996 verifying RSA signature: this will reject any improperly encoded
7997 DigestInfo structures.
7999 Note: this is a precautionary measure and no attacks are currently known.
8003 ### Changes between 1.0.0m and 1.0.0n [6 Aug 2014]
8005 * OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
8006 to a denial of service attack. A malicious server can crash the client
8007 with a null pointer dereference (read) by specifying an anonymous (EC)DH
8008 ciphersuite and sending carefully crafted handshake messages.
8010 Thanks to Felix Gröbert (Google) for discovering and researching this
8016 * By sending carefully crafted DTLS packets an attacker could cause openssl
8017 to leak memory. This can be exploited through a Denial of Service attack.
8018 Thanks to Adam Langley for discovering and researching this issue.
8023 * An attacker can force openssl to consume large amounts of memory whilst
8024 processing DTLS handshake messages. This can be exploited through a
8025 Denial of Service attack.
8026 Thanks to Adam Langley for discovering and researching this issue.
8031 * An attacker can force an error condition which causes openssl to crash
8032 whilst processing DTLS packets due to memory being freed twice. This
8033 can be exploited through a Denial of Service attack.
8034 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
8040 * If a multithreaded client connects to a malicious server using a resumed
8041 session and the server sends an ec point format extension it could write
8042 up to 255 bytes to freed memory.
8044 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
8050 * A flaw in OBJ_obj2txt may cause pretty printing functions such as
8051 X509_name_oneline, X509_name_print_ex et al. to leak some information
8052 from the stack. Applications may be affected if they echo pretty printing
8053 output to the attacker.
8055 Thanks to Ivan Fratric (Google) for discovering this issue.
8058 *Emilia Käsper, and Steve Henson*
8060 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
8061 for corner cases. (Certain input points at infinity could lead to
8062 bogus results, with non-infinity inputs mapped to infinity too.)
8066 ### Changes between 1.0.0l and 1.0.0m [5 Jun 2014]
8068 * Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
8069 handshake can force the use of weak keying material in OpenSSL
8070 SSL/TLS clients and servers.
8072 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
8073 researching this issue. ([CVE-2014-0224])
8075 *KIKUCHI Masashi, Steve Henson*
8077 * Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
8078 OpenSSL DTLS client the code can be made to recurse eventually crashing
8081 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
8084 *Imre Rad, Steve Henson*
8086 * Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
8087 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
8088 client or server. This is potentially exploitable to run arbitrary
8089 code on a vulnerable client or server.
8091 Thanks to Jüri Aedla for reporting this issue. ([CVE-2014-0195])
8093 *Jüri Aedla, Steve Henson*
8095 * Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
8096 are subject to a denial of service attack.
8098 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
8099 this issue. ([CVE-2014-3470])
8101 *Felix Gröbert, Ivan Fratric, Steve Henson*
8103 * Harmonize version and its documentation. -f flag is used to display
8106 *mancha <mancha1@zoho.com>*
8108 * Fix eckey_priv_encode so it immediately returns an error upon a failure
8109 in i2d_ECPrivateKey.
8111 *mancha <mancha1@zoho.com>*
8113 * Fix some double frees. These are not thought to be exploitable.
8115 *mancha <mancha1@zoho.com>*
8117 * Fix for the attack described in the paper "Recovering OpenSSL
8118 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
8119 by Yuval Yarom and Naomi Benger. Details can be obtained from:
8120 <http://eprint.iacr.org/2014/140>
8122 Thanks to Yuval Yarom and Naomi Benger for discovering this
8123 flaw and to Yuval Yarom for supplying a fix ([CVE-2014-0076])
8125 *Yuval Yarom and Naomi Benger*
8127 ### Changes between 1.0.0k and 1.0.0l [6 Jan 2014]
8129 * Keep original DTLS digest and encryption contexts in retransmission
8130 structures so we can use the previous session parameters if they need
8131 to be resent. ([CVE-2013-6450])
8135 * Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
8136 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
8137 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
8138 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
8139 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
8140 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
8142 *Rob Stradling, Adam Langley*
8144 ### Changes between 1.0.0j and 1.0.0k [5 Feb 2013]
8146 * Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
8148 This addresses the flaw in CBC record processing discovered by
8149 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
8150 at: <http://www.isg.rhul.ac.uk/tls/>
8152 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
8153 Security Group at Royal Holloway, University of London
8154 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
8155 Emilia Käsper for the initial patch.
8158 *Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson*
8160 * Return an error when checking OCSP signatures when key is NULL.
8161 This fixes a DoS attack. ([CVE-2013-0166])
8165 * Call OCSP Stapling callback after ciphersuite has been chosen, so
8166 the right response is stapled. Also change SSL_get_certificate()
8167 so it returns the certificate actually sent.
8168 See <http://rt.openssl.org/Ticket/Display.html?id=2836>.
8169 (This is a backport)
8171 *Rob Stradling <rob.stradling@comodo.com>*
8173 * Fix possible deadlock when decoding public keys.
8177 ### Changes between 1.0.0i and 1.0.0j [10 May 2012]
8179 [NB: OpenSSL 1.0.0i and later 1.0.0 patch levels were released after
8182 * Sanity check record length before skipping explicit IV in DTLS
8185 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
8186 fuzzing as a service testing platform.
8191 * Initialise tkeylen properly when encrypting CMS messages.
8192 Thanks to Solar Designer of Openwall for reporting this issue.
8196 ### Changes between 1.0.0h and 1.0.0i [19 Apr 2012]
8198 * Check for potentially exploitable overflows in asn1_d2i_read_bio
8199 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
8200 in CRYPTO_realloc_clean.
8202 Thanks to Tavis Ormandy, Google Security Team, for discovering this
8203 issue and to Adam Langley <agl@chromium.org> for fixing it.
8206 *Adam Langley (Google), Tavis Ormandy, Google Security Team*
8208 ### Changes between 1.0.0g and 1.0.0h [12 Mar 2012]
8210 * Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness
8211 in CMS and PKCS7 code. When RSA decryption fails use a random key for
8212 content decryption and always return the same error. Note: this attack
8213 needs on average 2^20 messages so it only affects automated senders. The
8214 old behaviour can be re-enabled in the CMS code by setting the
8215 CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where
8216 an MMA defence is not necessary.
8217 Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering
8218 this issue. ([CVE-2012-0884])
8222 * Fix CVE-2011-4619: make sure we really are receiving a
8223 client hello before rejecting multiple SGC restarts. Thanks to
8224 Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug.
8228 ### Changes between 1.0.0f and 1.0.0g [18 Jan 2012]
8230 * Fix for DTLS DoS issue introduced by fix for CVE-2011-4109.
8231 Thanks to Antonio Martin, Enterprise Secure Access Research and
8232 Development, Cisco Systems, Inc. for discovering this bug and
8233 preparing a fix. ([CVE-2012-0050])
8237 ### Changes between 1.0.0e and 1.0.0f [4 Jan 2012]
8239 * Nadhem Alfardan and Kenny Paterson have discovered an extension
8240 of the Vaudenay padding oracle attack on CBC mode encryption
8241 which enables an efficient plaintext recovery attack against
8242 the OpenSSL implementation of DTLS. Their attack exploits timing
8243 differences arising during decryption processing. A research
8244 paper describing this attack can be found at:
8245 <http://www.isg.rhul.ac.uk/~kp/dtls.pdf>
8246 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
8247 Security Group at Royal Holloway, University of London
8248 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
8249 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
8250 for preparing the fix. ([CVE-2011-4108])
8252 *Robin Seggelmann, Michael Tuexen*
8254 * Clear bytes used for block padding of SSL 3.0 records.
8257 *Adam Langley (Google)*
8259 * Only allow one SGC handshake restart for SSL/TLS. Thanks to George
8260 Kadianakis <desnacked@gmail.com> for discovering this issue and
8261 Adam Langley for preparing the fix. ([CVE-2011-4619])
8263 *Adam Langley (Google)*
8265 * Check parameters are not NULL in GOST ENGINE. ([CVE-2012-0027])
8267 *Andrey Kulikov <amdeich@gmail.com>*
8269 * Prevent malformed RFC3779 data triggering an assertion failure.
8270 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
8271 and Rob Austein <sra@hactrn.net> for fixing it. ([CVE-2011-4577])
8273 *Rob Austein <sra@hactrn.net>*
8275 * Improved PRNG seeding for VOS.
8277 *Paul Green <Paul.Green@stratus.com>*
8279 * Fix ssl_ciph.c set-up race.
8281 *Adam Langley (Google)*
8283 * Fix spurious failures in ecdsatest.c.
8285 *Emilia Käsper (Google)*
8287 * Fix the BIO_f_buffer() implementation (which was mixing different
8288 interpretations of the `..._len` fields).
8290 *Adam Langley (Google)*
8292 * Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
8293 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
8294 threads won't reuse the same blinding coefficients.
8296 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
8297 lock to call BN_BLINDING_invert_ex, and avoids one use of
8298 BN_BLINDING_update for each BN_BLINDING structure (previously,
8299 the last update always remained unused).
8301 *Emilia Käsper (Google)*
8303 * In ssl3_clear, preserve s3->init_extra along with s3->rbuf.
8305 *Bob Buckholz (Google)*
8307 ### Changes between 1.0.0d and 1.0.0e [6 Sep 2011]
8309 * Fix bug where CRLs with nextUpdate in the past are sometimes accepted
8310 by initialising X509_STORE_CTX properly. ([CVE-2011-3207])
8312 *Kaspar Brand <ossl@velox.ch>*
8314 * Fix SSL memory handling for (EC)DH ciphersuites, in particular
8315 for multi-threaded use of ECDH. ([CVE-2011-3210])
8317 *Adam Langley (Google)*
8319 * Fix x509_name_ex_d2i memory leak on bad inputs.
8323 * Remove hard coded ecdsaWithSHA1 signature tests in ssl code and check
8324 signature public key algorithm by using OID xref utilities instead.
8325 Before this you could only use some ECC ciphersuites with SHA1 only.
8329 * Add protection against ECDSA timing attacks as mentioned in the paper
8330 by Billy Bob Brumley and Nicola Tuveri, see:
8331 <http://eprint.iacr.org/2011/232.pdf>
8333 *Billy Bob Brumley and Nicola Tuveri*
8335 ### Changes between 1.0.0c and 1.0.0d [8 Feb 2011]
8337 * Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
8339 *Neel Mehta, Adam Langley, Bodo Moeller (Google)*
8341 * Fix bug in string printing code: if *any* escaping is enabled we must
8342 escape the escape character (backslash) or the resulting string is
8347 ### Changes between 1.0.0b and 1.0.0c [2 Dec 2010]
8349 * Disable code workaround for ancient and obsolete Netscape browsers
8350 and servers: an attacker can use it in a ciphersuite downgrade attack.
8351 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
8355 * Fixed J-PAKE implementation error, originally discovered by
8356 Sebastien Martini, further info and confirmation from Stefan
8357 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
8361 ### Changes between 1.0.0a and 1.0.0b [16 Nov 2010]
8363 * Fix extension code to avoid race conditions which can result in a buffer
8364 overrun vulnerability: resumed sessions must not be modified as they can
8365 be shared by multiple threads. CVE-2010-3864
8369 * Fix WIN32 build system to correctly link an ENGINE directory into
8374 ### Changes between 1.0.0 and 1.0.0a [01 Jun 2010]
8376 * Check return value of int_rsa_verify in pkey_rsa_verifyrecover
8379 *Steve Henson, Peter-Michael Hager <hager@dortmund.net>*
8381 ### Changes between 0.9.8n and 1.0.0 [29 Mar 2010]
8383 * Add "missing" function EVP_CIPHER_CTX_copy(). This copies a cipher
8384 context. The operation can be customised via the ctrl mechanism in
8385 case ENGINEs want to include additional functionality.
8389 * Tolerate yet another broken PKCS#8 key format: private key value negative.
8393 * Add new -subject_hash_old and -issuer_hash_old options to x509 utility to
8394 output hashes compatible with older versions of OpenSSL.
8396 *Willy Weisz <weisz@vcpc.univie.ac.at>*
8398 * Fix compression algorithm handling: if resuming a session use the
8399 compression algorithm of the resumed session instead of determining
8400 it from client hello again. Don't allow server to change algorithm.
8404 * Add load_crls() function to commands tidying load_certs() too. Add option
8405 to verify utility to allow additional CRLs to be included.
8409 * Update OCSP request code to permit adding custom headers to the request:
8410 some responders need this.
8414 * The function EVP_PKEY_sign() returns <=0 on error: check return code
8417 *Julia Lawall <julia@diku.dk>*
8419 * Update verify callback code in `apps/s_cb.c` and `apps/verify.c`, it
8420 needlessly dereferenced structures, used obsolete functions and
8421 didn't handle all updated verify codes correctly.
8425 * Disable MD2 in the default configuration.
8429 * In BIO_pop() and BIO_push() use the ctrl argument (which was NULL) to
8430 indicate the initial BIO being pushed or popped. This makes it possible
8431 to determine whether the BIO is the one explicitly called or as a result
8432 of the ctrl being passed down the chain. Fix BIO_pop() and SSL BIOs so
8433 it handles reference counts correctly and doesn't zero out the I/O bio
8434 when it is not being explicitly popped. WARNING: applications which
8435 included workarounds for the old buggy behaviour will need to be modified
8436 or they could free up already freed BIOs.
8440 * Extend the uni2asc/asc2uni => OPENSSL_uni2asc/OPENSSL_asc2uni
8441 renaming to all platforms (within the 0.9.8 branch, this was
8442 done conditionally on Netware platforms to avoid a name clash).
8444 *Guenter <lists@gknw.net>*
8446 * Add ECDHE and PSK support to DTLS.
8448 *Michael Tuexen <tuexen@fh-muenster.de>*
8450 * Add CHECKED_STACK_OF macro to safestack.h, otherwise safestack can't
8455 * Add "missing" function EVP_MD_flags() (without this the only way to
8456 retrieve a digest flags is by accessing the structure directly. Update
8457 `EVP_MD_do_all*()` and `EVP_CIPHER_do_all*()` to include the name a digest
8458 or cipher is registered as in the "from" argument. Print out all
8459 registered digests in the dgst usage message instead of manually
8460 attempting to work them out.
8464 * If no SSLv2 ciphers are used don't use an SSLv2 compatible client hello:
8465 this allows the use of compression and extensions. Change default cipher
8466 string to remove SSLv2 ciphersuites. This effectively avoids ancient SSLv2
8467 by default unless an application cipher string requests it.
8471 * Alter match criteria in PKCS12_parse(). It used to try to use local
8472 key ids to find matching certificates and keys but some PKCS#12 files
8473 don't follow the (somewhat unwritten) rules and this strategy fails.
8474 Now just gather all certificates together and the first private key
8475 then look for the first certificate that matches the key.
8479 * Support use of registered digest and cipher names for dgst and cipher
8480 commands instead of having to add each one as a special case. So now
8487 openssl dgst -sha256 foo
8489 and this works for ENGINE based algorithms too.
8493 * Update Gost ENGINE to support parameter files.
8495 *Victor B. Wagner <vitus@cryptocom.ru>*
8497 * Support GeneralizedTime in ca utility.
8499 *Oliver Martin <oliver@volatilevoid.net>, Steve Henson*
8501 * Enhance the hash format used for certificate directory links. The new
8502 form uses the canonical encoding (meaning equivalent names will work
8503 even if they aren't identical) and uses SHA1 instead of MD5. This form
8504 is incompatible with the older format and as a result c_rehash should
8505 be used to rebuild symbolic links.
8509 * Make PKCS#8 the default write format for private keys, replacing the
8510 traditional format. This form is standardised, more secure and doesn't
8511 include an implicit MD5 dependency.
8515 * Add a $gcc_devteam_warn option to Configure. The idea is that any code
8516 committed to OpenSSL should pass this lot as a minimum.
8520 * Add session ticket override functionality for use by EAP-FAST.
8522 *Jouni Malinen <j@w1.fi>*
8524 * Modify HMAC functions to return a value. Since these can be implemented
8525 in an ENGINE errors can occur.
8529 * Type-checked OBJ_bsearch_ex.
8533 * Type-checked OBJ_bsearch. Also some constification necessitated
8534 by type-checking. Still to come: TXT_DB, bsearch(?),
8535 OBJ_bsearch_ex, qsort, CRYPTO_EX_DATA, ASN1_VALUE, ASN1_STRING,
8540 * New function OPENSSL_gmtime_adj() to add a specific number of days and
8541 seconds to a tm structure directly, instead of going through OS
8542 specific date routines. This avoids any issues with OS routines such
8543 as the year 2038 bug. New `*_adj()` functions for ASN1 time structures
8544 and X509_time_adj_ex() to cover the extended range. The existing
8545 X509_time_adj() is still usable and will no longer have any date issues.
8549 * Delta CRL support. New use deltas option which will attempt to locate
8550 and search any appropriate delta CRLs available.
8552 This work was sponsored by Google.
8556 * Support for CRLs partitioned by reason code. Reorganise CRL processing
8557 code and add additional score elements. Validate alternate CRL paths
8558 as part of the CRL checking and indicate a new error "CRL path validation
8559 error" in this case. Applications wanting additional details can use
8560 the verify callback and check the new "parent" field. If this is not
8561 NULL CRL path validation is taking place. Existing applications won't
8562 see this because it requires extended CRL support which is off by
8565 This work was sponsored by Google.
8569 * Support for freshest CRL extension.
8571 This work was sponsored by Google.
8575 * Initial indirect CRL support. Currently only supported in the CRLs
8576 passed directly and not via lookup. Process certificate issuer
8577 CRL entry extension and lookup CRL entries by bother issuer name
8578 and serial number. Check and process CRL issuer entry in IDP extension.
8580 This work was sponsored by Google.
8584 * Add support for distinct certificate and CRL paths. The CRL issuer
8585 certificate is validated separately in this case. Only enabled if
8586 an extended CRL support flag is set: this flag will enable additional
8587 CRL functionality in future.
8589 This work was sponsored by Google.
8593 * Add support for policy mappings extension.
8595 This work was sponsored by Google.
8599 * Fixes to pathlength constraint, self issued certificate handling,
8600 policy processing to align with RFC3280 and PKITS tests.
8602 This work was sponsored by Google.
8606 * Support for name constraints certificate extension. DN, email, DNS
8607 and URI types are currently supported.
8609 This work was sponsored by Google.
8613 * To cater for systems that provide a pointer-based thread ID rather
8614 than numeric, deprecate the current numeric thread ID mechanism and
8615 replace it with a structure and associated callback type. This
8616 mechanism allows a numeric "hash" to be extracted from a thread ID in
8617 either case, and on platforms where pointers are larger than 'long',
8618 mixing is done to help ensure the numeric 'hash' is usable even if it
8619 can't be guaranteed unique. The default mechanism is to use "&errno"
8620 as a pointer-based thread ID to distinguish between threads.
8622 Applications that want to provide their own thread IDs should now use
8623 CRYPTO_THREADID_set_callback() to register a callback that will call
8624 either CRYPTO_THREADID_set_numeric() or CRYPTO_THREADID_set_pointer().
8626 Note that ERR_remove_state() is now deprecated, because it is tied
8627 to the assumption that thread IDs are numeric. ERR_remove_state(0)
8628 to free the current thread's error state should be replaced by
8629 ERR_remove_thread_state(NULL).
8631 (This new approach replaces the functions CRYPTO_set_idptr_callback(),
8632 CRYPTO_get_idptr_callback(), and CRYPTO_thread_idptr() that existed in
8633 OpenSSL 0.9.9-dev between June 2006 and August 2008. Also, if an
8634 application was previously providing a numeric thread callback that
8635 was inappropriate for distinguishing threads, then uniqueness might
8636 have been obtained with &errno that happened immediately in the
8637 intermediate development versions of OpenSSL; this is no longer the
8638 case, the numeric thread callback will now override the automatic use
8641 *Geoff Thorpe, with help from Bodo Moeller*
8643 * Initial support for different CRL issuing certificates. This covers a
8644 simple case where the self issued certificates in the chain exist and
8645 the real CRL issuer is higher in the existing chain.
8647 This work was sponsored by Google.
8651 * Removed effectively defunct crypto/store from the build.
8655 * Revamp of STACK to provide stronger type-checking. Still to come:
8656 TXT_DB, bsearch(?), OBJ_bsearch, qsort, CRYPTO_EX_DATA, ASN1_VALUE,
8657 ASN1_STRING, CONF_VALUE.
8661 * Add a new SSL_MODE_RELEASE_BUFFERS mode flag to release unused buffer
8662 RAM on SSL connections. This option can save about 34k per idle SSL.
8666 * Revamp of LHASH to provide stronger type-checking. Still to come:
8667 STACK, TXT_DB, bsearch, qsort.
8671 * Initial support for Cryptographic Message Syntax (aka CMS) based
8672 on RFC3850, RFC3851 and RFC3852. New cms directory and cms utility,
8673 support for data, signedData, compressedData, digestedData and
8674 encryptedData, envelopedData types included. Scripts to check against
8675 RFC4134 examples draft and interop and consistency checks of many
8676 content types and variants.
8680 * Add options to enc utility to support use of zlib compression BIO.
8684 * Extend mk1mf to support importing of options and assembly language
8685 files from Configure script, currently only included in VC-WIN32.
8686 The assembly language rules can now optionally generate the source
8687 files from the associated perl scripts.
8691 * Implement remaining functionality needed to support GOST ciphersuites.
8692 Interop testing has been performed using CryptoPro implementations.
8694 *Victor B. Wagner <vitus@cryptocom.ru>*
8696 * s390x assembler pack.
8700 * ARMv4 assembler pack. ARMv4 refers to v4 and later ISA, not CPU
8705 * Implement Opaque PRF Input TLS extension as specified in
8706 draft-rescorla-tls-opaque-prf-input-00.txt. Since this is not an
8707 official specification yet and no extension type assignment by
8708 IANA exists, this extension (for now) will have to be explicitly
8709 enabled when building OpenSSL by providing the extension number
8710 to use. For example, specify an option
8712 -DTLSEXT_TYPE_opaque_prf_input=0x9527
8714 to the "config" or "Configure" script to enable the extension,
8715 assuming extension number 0x9527 (which is a completely arbitrary
8716 and unofficial assignment based on the MD5 hash of the Internet
8717 Draft). Note that by doing so, you potentially lose
8718 interoperability with other TLS implementations since these might
8719 be using the same extension number for other purposes.
8721 SSL_set_tlsext_opaque_prf_input(ssl, src, len) is used to set the
8722 opaque PRF input value to use in the handshake. This will create
8723 an internal copy of the length-'len' string at 'src', and will
8724 return non-zero for success.
8726 To get more control and flexibility, provide a callback function
8729 SSL_CTX_set_tlsext_opaque_prf_input_callback(ctx, cb)
8730 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg(ctx, arg)
8734 int (*cb)(SSL *, void *peerinput, size_t len, void *arg);
8737 Callback function 'cb' will be called in handshakes, and is
8738 expected to use SSL_set_tlsext_opaque_prf_input() as appropriate.
8739 Argument 'arg' is for application purposes (the value as given to
8740 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg() will directly
8741 be provided to the callback function). The callback function
8742 has to return non-zero to report success: usually 1 to use opaque
8743 PRF input just if possible, or 2 to enforce use of the opaque PRF
8744 input. In the latter case, the library will abort the handshake
8745 if opaque PRF input is not successfully negotiated.
8747 Arguments 'peerinput' and 'len' given to the callback function
8748 will always be NULL and 0 in the case of a client. A server will
8749 see the client's opaque PRF input through these variables if
8750 available (NULL and 0 otherwise). Note that if the server
8751 provides an opaque PRF input, the length must be the same as the
8752 length of the client's opaque PRF input.
8754 Note that the callback function will only be called when creating
8755 a new session (session resumption can resume whatever was
8756 previously negotiated), and will not be called in SSL 2.0
8757 handshakes; thus, SSL_CTX_set_options(ctx, SSL_OP_NO_SSLv2) or
8758 SSL_set_options(ssl, SSL_OP_NO_SSLv2) is especially recommended
8759 for applications that need to enforce opaque PRF input.
8763 * Update ssl code to support digests other than SHA1+MD5 for handshake
8766 *Victor B. Wagner <vitus@cryptocom.ru>*
8768 * Add RFC4507 support to OpenSSL. This includes the corrections in
8769 RFC4507bis. The encrypted ticket format is an encrypted encoded
8770 SSL_SESSION structure, that way new session features are automatically
8773 If a client application caches session in an SSL_SESSION structure
8774 support is transparent because tickets are now stored in the encoded
8777 The SSL_CTX structure automatically generates keys for ticket
8778 protection in servers so again support should be possible
8779 with no application modification.
8781 If a client or server wishes to disable RFC4507 support then the option
8782 SSL_OP_NO_TICKET can be set.
8784 Add a TLS extension debugging callback to allow the contents of any client
8785 or server extensions to be examined.
8787 This work was sponsored by Google.
8791 * Final changes to avoid use of pointer pointer casts in OpenSSL.
8792 OpenSSL should now compile cleanly on gcc 4.2
8794 *Peter Hartley <pdh@utter.chaos.org.uk>, Steve Henson*
8796 * Update SSL library to use new EVP_PKEY MAC API. Include generic MAC
8797 support including streaming MAC support: this is required for GOST
8798 ciphersuite support.
8800 *Victor B. Wagner <vitus@cryptocom.ru>, Steve Henson*
8802 * Add option -stream to use PKCS#7 streaming in smime utility. New
8803 function i2d_PKCS7_bio_stream() and PEM_write_PKCS7_bio_stream()
8804 to output in BER and PEM format.
8808 * Experimental support for use of HMAC via EVP_PKEY interface. This
8809 allows HMAC to be handled via the `EVP_DigestSign*()` interface. The
8810 EVP_PKEY "key" in this case is the HMAC key, potentially allowing
8811 ENGINE support for HMAC keys which are unextractable. New -mac and
8812 -macopt options to dgst utility.
8816 * New option -sigopt to dgst utility. Update dgst to use
8817 `EVP_Digest{Sign,Verify}*`. These two changes make it possible to use
8818 alternative signing parameters such as X9.31 or PSS in the dgst
8823 * Change ssl_cipher_apply_rule(), the internal function that does
8824 the work each time a ciphersuite string requests enabling
8825 ("foo+bar"), moving ("+foo+bar"), disabling ("-foo+bar", or
8826 removing ("!foo+bar") a class of ciphersuites: Now it maintains
8827 the order of disabled ciphersuites such that those ciphersuites
8828 that most recently went from enabled to disabled not only stay
8829 in order with respect to each other, but also have higher priority
8830 than other disabled ciphersuites the next time ciphersuites are
8833 This means that you can now say, e.g., "PSK:-PSK:HIGH" to enable
8834 the same ciphersuites as with "HIGH" alone, but in a specific
8835 order where the PSK ciphersuites come first (since they are the
8836 most recently disabled ciphersuites when "HIGH" is parsed).
8838 Also, change ssl_create_cipher_list() (using this new
8839 functionality) such that between otherwise identical
8840 ciphersuites, ephemeral ECDH is preferred over ephemeral DH in
8845 * Change ssl_create_cipher_list() so that it automatically
8846 arranges the ciphersuites in reasonable order before starting
8847 to process the rule string. Thus, the definition for "DEFAULT"
8848 (SSL_DEFAULT_CIPHER_LIST) now is just "ALL:!aNULL:!eNULL", but
8849 remains equivalent to `"AES:ALL:!aNULL:!eNULL:+aECDH:+kRSA:+RC4:@STRENGTH"`.
8850 This makes it much easier to arrive at a reasonable default order
8851 in applications for which anonymous ciphers are OK (meaning
8852 that you can't actually use DEFAULT).
8854 *Bodo Moeller; suggested by Victor Duchovni*
8856 * Split the SSL/TLS algorithm mask (as used for ciphersuite string
8857 processing) into multiple integers instead of setting
8858 "SSL_MKEY_MASK" bits, "SSL_AUTH_MASK" bits, "SSL_ENC_MASK",
8859 "SSL_MAC_MASK", and "SSL_SSL_MASK" bits all in a single integer.
8860 (These masks as well as the individual bit definitions are hidden
8861 away into the non-exported interface ssl/ssl_locl.h, so this
8862 change to the definition of the SSL_CIPHER structure shouldn't
8863 affect applications.) This give us more bits for each of these
8864 categories, so there is no longer a need to coagulate AES128 and
8865 AES256 into a single algorithm bit, and to coagulate Camellia128
8866 and Camellia256 into a single algorithm bit, which has led to all
8869 Thus, among other things, the kludge introduced in 0.9.7m and
8870 0.9.8e for masking out AES256 independently of AES128 or masking
8871 out Camellia256 independently of AES256 is not needed here in 0.9.9.
8873 With the change, we also introduce new ciphersuite aliases that
8874 so far were missing: "AES128", "AES256", "CAMELLIA128", and
8879 * Add support for dsa-with-SHA224 and dsa-with-SHA256.
8880 Use the leftmost N bytes of the signature input if the input is
8881 larger than the prime q (with N being the size in bytes of q).
8885 * Very *very* experimental PKCS#7 streaming encoder support. Nothing uses
8886 it yet and it is largely untested.
8890 * Add support for the ecdsa-with-SHA224/256/384/512 signature types.
8894 * Initial incomplete changes to avoid need for function casts in OpenSSL
8895 some compilers (gcc 4.2 and later) reject their use. Safestack is
8896 reimplemented. Update ASN1 to avoid use of legacy functions.
8900 * Win32/64 targets are linked with Winsock2.
8904 * Add an X509_CRL_METHOD structure to allow CRL processing to be redirected
8905 to external functions. This can be used to increase CRL handling
8906 efficiency especially when CRLs are very large by (for example) storing
8907 the CRL revoked certificates in a database.
8911 * Overhaul of by_dir code. Add support for dynamic loading of CRLs so
8912 new CRLs added to a directory can be used. New command line option
8913 -verify_return_error to s_client and s_server. This causes real errors
8914 to be returned by the verify callback instead of carrying on no matter
8915 what. This reflects the way a "real world" verify callback would behave.
8919 * GOST engine, supporting several GOST algorithms and public key formats.
8920 Kindly donated by Cryptocom.
8924 * Partial support for Issuing Distribution Point CRL extension. CRLs
8925 partitioned by DP are handled but no indirect CRL or reason partitioning
8926 (yet). Complete overhaul of CRL handling: now the most suitable CRL is
8927 selected via a scoring technique which handles IDP and AKID in CRLs.
8931 * New X509_STORE_CTX callbacks lookup_crls() and lookup_certs() which
8932 will ultimately be used for all verify operations: this will remove the
8933 X509_STORE dependency on certificate verification and allow alternative
8934 lookup methods. X509_STORE based implementations of these two callbacks.
8938 * Allow multiple CRLs to exist in an X509_STORE with matching issuer names.
8939 Modify get_crl() to find a valid (unexpired) CRL if possible.
8943 * New function X509_CRL_match() to check if two CRLs are identical. Normally
8944 this would be called X509_CRL_cmp() but that name is already used by
8945 a function that just compares CRL issuer names. Cache several CRL
8946 extensions in X509_CRL structure and cache CRLDP in X509.
8950 * Store a "canonical" representation of X509_NAME structure (ASN1 Name)
8951 this maps equivalent X509_NAME structures into a consistent structure.
8952 Name comparison can then be performed rapidly using memcmp().
8956 * Non-blocking OCSP request processing. Add -timeout option to ocsp
8961 * Allow digests to supply their own micalg string for S/MIME type using
8962 the ctrl EVP_MD_CTRL_MICALG.
8966 * During PKCS7 signing pass the PKCS7 SignerInfo structure to the
8967 EVP_PKEY_METHOD before and after signing via the EVP_PKEY_CTRL_PKCS7_SIGN
8968 ctrl. It can then customise the structure before and/or after signing
8973 * New function OBJ_add_sigid() to allow application defined signature OIDs
8974 to be added to OpenSSLs internal tables. New function OBJ_sigid_free()
8975 to free up any added signature OIDs.
8979 * New functions EVP_CIPHER_do_all(), EVP_CIPHER_do_all_sorted(),
8980 EVP_MD_do_all() and EVP_MD_do_all_sorted() to enumerate internal
8981 digest and cipher tables. New options added to openssl utility:
8982 list-message-digest-algorithms and list-cipher-algorithms.
8986 * Change the array representation of binary polynomials: the list
8987 of degrees of non-zero coefficients is now terminated with -1.
8988 Previously it was terminated with 0, which was also part of the
8989 value; thus, the array representation was not applicable to
8990 polynomials where t^0 has coefficient zero. This change makes
8991 the array representation useful in a more general context.
8995 * Various modifications and fixes to SSL/TLS cipher string
8996 handling. For ECC, the code now distinguishes between fixed ECDH
8997 with RSA certificates on the one hand and with ECDSA certificates
8998 on the other hand, since these are separate ciphersuites. The
8999 unused code for Fortezza ciphersuites has been removed.
9001 For consistency with EDH, ephemeral ECDH is now called "EECDH"
9002 (not "ECDHE"). For consistency with the code for DH
9003 certificates, use of ECDH certificates is now considered ECDH
9004 authentication, not RSA or ECDSA authentication (the latter is
9005 merely the CA's signing algorithm and not actively used in the
9008 The temporary ciphersuite alias "ECCdraft" is no longer
9009 available, and ECC ciphersuites are no longer excluded from "ALL"
9010 and "DEFAULT". The following aliases now exist for RFC 4492
9011 ciphersuites, most of these by analogy with the DH case:
9013 kECDHr - ECDH cert, signed with RSA
9014 kECDHe - ECDH cert, signed with ECDSA
9015 kECDH - ECDH cert (signed with either RSA or ECDSA)
9016 kEECDH - ephemeral ECDH
9017 ECDH - ECDH cert or ephemeral ECDH
9023 AECDH - anonymous ECDH
9024 EECDH - non-anonymous ephemeral ECDH (equivalent to "kEECDH:-AECDH")
9028 * Add additional S/MIME capabilities for AES and GOST ciphers if supported.
9029 Use correct micalg parameters depending on digest(s) in signed message.
9033 * Add engine support for EVP_PKEY_ASN1_METHOD. Add functions to process
9034 an ENGINE asn1 method. Support ENGINE lookups in the ASN1 code.
9038 * Initial engine support for EVP_PKEY_METHOD. New functions to permit
9039 an engine to register a method. Add ENGINE lookups for methods and
9040 functional reference processing.
9044 * New functions `EVP_Digest{Sign,Verify)*`. These are enhanced versions of
9045 `EVP_{Sign,Verify}*` which allow an application to customise the signature
9050 * New -resign option to smime utility. This adds one or more signers
9051 to an existing PKCS#7 signedData structure. Also -md option to use an
9052 alternative message digest algorithm for signing.
9056 * Tidy up PKCS#7 routines and add new functions to make it easier to
9057 create PKCS7 structures containing multiple signers. Update smime
9058 application to support multiple signers.
9062 * New -macalg option to pkcs12 utility to allow setting of an alternative
9067 * Initial support for PKCS#5 v2.0 PRFs other than default SHA1 HMAC.
9068 Reorganize PBE internals to lookup from a static table using NIDs,
9069 add support for HMAC PBE OID translation. Add a EVP_CIPHER ctrl:
9070 EVP_CTRL_PBE_PRF_NID this allows a cipher to specify an alternative
9071 PRF which will be automatically used with PBES2.
9075 * Replace the algorithm specific calls to generate keys in "req" with the
9080 * Update PKCS#7 enveloped data routines to use new API. This is now
9081 supported by any public key method supporting the encrypt operation. A
9082 ctrl is added to allow the public key algorithm to examine or modify
9083 the PKCS#7 RecipientInfo structure if it needs to: for RSA this is
9088 * Add a ctrl to asn1 method to allow a public key algorithm to express
9089 a default digest type to use. In most cases this will be SHA1 but some
9090 algorithms (such as GOST) need to specify an alternative digest. The
9091 return value indicates how strong the preference is 1 means optional and
9092 2 is mandatory (that is it is the only supported type). Modify
9093 ASN1_item_sign() to accept a NULL digest argument to indicate it should
9094 use the default md. Update openssl utilities to use the default digest
9095 type for signing if it is not explicitly indicated.
9099 * Use OID cross reference table in ASN1_sign() and ASN1_verify(). New
9100 EVP_MD flag EVP_MD_FLAG_PKEY_METHOD_SIGNATURE. This uses the relevant
9101 signing method from the key type. This effectively removes the link
9102 between digests and public key types.
9106 * Add an OID cross reference table and utility functions. Its purpose is to
9107 translate between signature OIDs such as SHA1WithrsaEncryption and SHA1,
9108 rsaEncryption. This will allow some of the algorithm specific hackery
9109 needed to use the correct OID to be removed.
9113 * Remove algorithm specific dependencies when setting PKCS7_SIGNER_INFO
9114 structures for PKCS7_sign(). They are now set up by the relevant public
9119 * Add provisional EC pkey method with support for ECDSA and ECDH.
9123 * Add support for key derivation (agreement) in the API, DH method and
9128 * Add DSA pkey method and DH pkey methods, extend DH ASN1 method to support
9129 public and private key formats. As a side effect these add additional
9130 command line functionality not previously available: DSA signatures can be
9131 generated and verified using pkeyutl and DH key support and generation in
9138 *Oliver Tappe <zooey@hirschkaefer.de>*
9140 * New make target "install_html_docs" installs HTML renditions of the
9143 *Oliver Tappe <zooey@hirschkaefer.de>*
9145 * New utility "genpkey" this is analogous to "genrsa" etc except it can
9146 generate keys for any algorithm. Extend and update EVP_PKEY_METHOD to
9147 support key and parameter generation and add initial key generation
9148 functionality for RSA.
9152 * Add functions for main EVP_PKEY_method operations. The undocumented
9153 functions `EVP_PKEY_{encrypt,decrypt}` have been renamed to
9154 `EVP_PKEY_{encrypt,decrypt}_old`.
9158 * Initial definitions for EVP_PKEY_METHOD. This will be a high level public
9159 key API, doesn't do much yet.
9163 * New function EVP_PKEY_asn1_get0_info() to retrieve information about
9164 public key algorithms. New option to openssl utility:
9165 "list-public-key-algorithms" to print out info.
9169 * Implement the Supported Elliptic Curves Extension for
9170 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
9174 * Don't free up OIDs in OBJ_cleanup() if they are in use by EVP_MD or
9175 EVP_CIPHER structures to avoid later problems in EVP_cleanup().
9179 * New utilities pkey and pkeyparam. These are similar to algorithm specific
9180 utilities such as rsa, dsa, dsaparam etc except they process any key
9185 * Transfer public key printing routines to EVP_PKEY_ASN1_METHOD. New
9186 functions EVP_PKEY_print_public(), EVP_PKEY_print_private(),
9187 EVP_PKEY_print_param() to print public key data from an EVP_PKEY
9192 * Initial support for pluggable public key ASN1.
9193 De-spaghettify the public key ASN1 handling. Move public and private
9194 key ASN1 handling to a new EVP_PKEY_ASN1_METHOD structure. Relocate
9195 algorithm specific handling to a single module within the relevant
9196 algorithm directory. Add functions to allow (near) opaque processing
9197 of public and private key structures.
9201 * Implement the Supported Point Formats Extension for
9202 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
9206 * Add initial support for RFC 4279 PSK TLS ciphersuites. Add members
9207 for the psk identity [hint] and the psk callback functions to the
9208 SSL_SESSION, SSL and SSL_CTX structure.
9211 PSK-RC4-SHA, PSK-3DES-EDE-CBC-SHA, PSK-AES128-CBC-SHA,
9215 SSL_CTX_use_psk_identity_hint
9216 SSL_get_psk_identity_hint
9217 SSL_get_psk_identity
9218 SSL_use_psk_identity_hint
9220 *Mika Kousa and Pasi Eronen of Nokia Corporation*
9222 * Add RFC 3161 compliant time stamp request creation, response generation
9223 and response verification functionality.
9225 *Zoltán Glózik <zglozik@opentsa.org>, The OpenTSA Project*
9227 * Add initial support for TLS extensions, specifically for the server_name
9228 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
9229 have new members for a host name. The SSL data structure has an
9230 additional member `SSL_CTX *initial_ctx` so that new sessions can be
9231 stored in that context to allow for session resumption, even after the
9232 SSL has been switched to a new SSL_CTX in reaction to a client's
9233 server_name extension.
9235 New functions (subject to change):
9237 SSL_get_servername()
9238 SSL_get_servername_type()
9241 New CTRL codes and macros (subject to change):
9243 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
9244 - SSL_CTX_set_tlsext_servername_callback()
9245 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
9246 - SSL_CTX_set_tlsext_servername_arg()
9247 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
9249 openssl s_client has a new '-servername ...' option.
9251 openssl s_server has new options '-servername_host ...', '-cert2 ...',
9252 '-key2 ...', '-servername_fatal' (subject to change). This allows
9253 testing the HostName extension for a specific single host name ('-cert'
9254 and '-key' remain fallbacks for handshakes without HostName
9255 negotiation). If the unrecognized_name alert has to be sent, this by
9256 default is a warning; it becomes fatal with the '-servername_fatal'
9259 *Peter Sylvester, Remy Allais, Christophe Renou*
9261 * Whirlpool hash implementation is added.
9265 * BIGNUM code on 64-bit SPARCv9 targets is switched from bn(64,64) to
9266 bn(64,32). Because of instruction set limitations it doesn't have
9267 any negative impact on performance. This was done mostly in order
9268 to make it possible to share assembler modules, such as bn_mul_mont
9269 implementations, between 32- and 64-bit builds without hassle.
9273 * Move code previously exiled into file crypto/ec/ec2_smpt.c
9274 to ec2_smpl.c, and no longer require the OPENSSL_EC_BIN_PT_COMP
9279 * New candidate for BIGNUM assembler implementation, bn_mul_mont,
9280 dedicated Montgomery multiplication procedure, is introduced.
9281 BN_MONT_CTX is modified to allow bn_mul_mont to reach for higher
9282 "64-bit" performance on certain 32-bit targets.
9286 * New option SSL_OP_NO_COMP to disable use of compression selectively
9287 in SSL structures. New SSL ctrl to set maximum send fragment size.
9288 Save memory by setting the I/O buffer sizes dynamically instead of
9289 using the maximum available value.
9293 * New option -V for 'openssl ciphers'. This prints the ciphersuite code
9294 in addition to the text details.
9298 * Very, very preliminary EXPERIMENTAL support for printing of general
9299 ASN1 structures. This currently produces rather ugly output and doesn't
9300 handle several customised structures at all.
9304 * Integrated support for PVK file format and some related formats such
9305 as MS PUBLICKEYBLOB and PRIVATEKEYBLOB. Command line switches to support
9306 these in the 'rsa' and 'dsa' utilities.
9310 * Support for PKCS#1 RSAPublicKey format on rsa utility command line.
9314 * Remove the ancient ASN1_METHOD code. This was only ever used in one
9315 place for the (very old) "NETSCAPE" format certificates which are now
9316 handled using new ASN1 code equivalents.
9320 * Let the TLSv1_method() etc. functions return a 'const' SSL_METHOD
9321 pointer and make the SSL_METHOD parameter in SSL_CTX_new,
9322 SSL_CTX_set_ssl_version and SSL_set_ssl_method 'const'.
9326 * Modify CRL distribution points extension code to print out previously
9327 unsupported fields. Enhance extension setting code to allow setting of
9332 * Add print and set support for Issuing Distribution Point CRL extension.
9336 * Change 'Configure' script to enable Camellia by default.
9343 ### Changes between 0.9.8m and 0.9.8n [24 Mar 2010]
9345 * When rejecting SSL/TLS records due to an incorrect version number, never
9346 update s->server with a new major version number. As of
9347 - OpenSSL 0.9.8m if 'short' is a 16-bit type,
9348 - OpenSSL 0.9.8f if 'short' is longer than 16 bits,
9349 the previous behavior could result in a read attempt at NULL when
9350 receiving specific incorrect SSL/TLS records once record payload
9351 protection is active. ([CVE-2010-0740])
9353 *Bodo Moeller, Adam Langley <agl@chromium.org>*
9355 * Fix for CVE-2010-0433 where some kerberos enabled versions of OpenSSL
9356 could be crashed if the relevant tables were not present (e.g. chrooted).
9358 *Tomas Hoger <thoger@redhat.com>*
9360 ### Changes between 0.9.8l and 0.9.8m [25 Feb 2010]
9362 * Always check bn_wexpand() return values for failure. ([CVE-2009-3245])
9364 *Martin Olsson, Neel Mehta*
9366 * Fix X509_STORE locking: Every 'objs' access requires a lock (to
9367 accommodate for stack sorting, always a write lock!).
9371 * On some versions of WIN32 Heap32Next is very slow. This can cause
9372 excessive delays in the RAND_poll(): over a minute. As a workaround
9373 include a time check in the inner Heap32Next loop too.
9377 * The code that handled flushing of data in SSL/TLS originally used the
9378 BIO_CTRL_INFO ctrl to see if any data was pending first. This caused
9379 the problem outlined in PR#1949. The fix suggested there however can
9380 trigger problems with buggy BIO_CTRL_WPENDING (e.g. some versions
9381 of Apache). So instead simplify the code to flush unconditionally.
9382 This should be fine since flushing with no data to flush is a no op.
9386 * Handle TLS versions 2.0 and later properly and correctly use the
9387 highest version of TLS/SSL supported. Although TLS >= 2.0 is some way
9388 off ancient servers have a habit of sticking around for a while...
9392 * Modify compression code so it frees up structures without using the
9393 ex_data callbacks. This works around a problem where some applications
9394 call CRYPTO_cleanup_all_ex_data() before application exit (e.g. when
9395 restarting) then use compression (e.g. SSL with compression) later.
9396 This results in significant per-connection memory leaks and
9397 has caused some security issues including CVE-2008-1678 and
9402 * Constify crypto/cast (i.e., <openssl/cast.h>): a CAST_KEY doesn't
9403 change when encrypting or decrypting.
9407 * Add option SSL_OP_LEGACY_SERVER_CONNECT which will allow clients to
9408 connect and renegotiate with servers which do not support RI.
9409 Until RI is more widely deployed this option is enabled by default.
9413 * Add "missing" ssl ctrls to clear options and mode.
9417 * If client attempts to renegotiate and doesn't support RI respond with
9418 a no_renegotiation alert as required by RFC5746. Some renegotiating
9419 TLS clients will continue a connection gracefully when they receive
9420 the alert. Unfortunately OpenSSL mishandled this alert and would hang
9421 waiting for a server hello which it will never receive. Now we treat a
9422 received no_renegotiation alert as a fatal error. This is because
9423 applications requesting a renegotiation might well expect it to succeed
9424 and would have no code in place to handle the server denying it so the
9425 only safe thing to do is to terminate the connection.
9429 * Add ctrl macro SSL_get_secure_renegotiation_support() which returns 1 if
9430 peer supports secure renegotiation and 0 otherwise. Print out peer
9431 renegotiation support in s_client/s_server.
9435 * Replace the highly broken and deprecated SPKAC certification method with
9436 the updated NID creation version. This should correctly handle UTF8.
9440 * Implement RFC5746. Re-enable renegotiation but require the extension
9441 as needed. Unfortunately, SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION
9442 turns out to be a bad idea. It has been replaced by
9443 SSL_OP_ALLOW_UNSAFE_LEGACY_RENEGOTIATION which can be set with
9444 SSL_CTX_set_options(). This is really not recommended unless you
9445 know what you are doing.
9447 *Eric Rescorla <ekr@networkresonance.com>, Ben Laurie, Steve Henson*
9449 * Fixes to stateless session resumption handling. Use initial_ctx when
9450 issuing and attempting to decrypt tickets in case it has changed during
9451 servername handling. Use a non-zero length session ID when attempting
9452 stateless session resumption: this makes it possible to determine if
9453 a resumption has occurred immediately after receiving server hello
9454 (several places in OpenSSL subtly assume this) instead of later in
9459 * The functions ENGINE_ctrl(), OPENSSL_isservice(),
9460 CMS_get1_RecipientRequest() and RAND_bytes() can return <=0 on error
9461 fixes for a few places where the return code is not checked
9464 *Julia Lawall <julia@diku.dk>*
9466 * Add --strict-warnings option to Configure script to include devteam
9467 warnings in other configurations.
9471 * Add support for --libdir option and LIBDIR variable in makefiles. This
9472 makes it possible to install openssl libraries in locations which
9473 have names other than "lib", for example "/usr/lib64" which some
9476 *Steve Henson, based on patch from Jeremy Utley*
9478 * Don't allow the use of leading 0x80 in OIDs. This is a violation of
9479 X690 8.9.12 and can produce some misleading textual output of OIDs.
9481 *Steve Henson, reported by Dan Kaminsky*
9483 * Delete MD2 from algorithm tables. This follows the recommendation in
9484 several standards that it is not used in new applications due to
9485 several cryptographic weaknesses. For binary compatibility reasons
9486 the MD2 API is still compiled in by default.
9490 * Add compression id to {d2i,i2d}_SSL_SESSION so it is correctly saved
9495 * Rename uni2asc and asc2uni functions to OPENSSL_uni2asc and
9496 OPENSSL_asc2uni conditionally on Netware platforms to avoid a name
9499 *Guenter <lists@gknw.net>*
9501 * Fix the server certificate chain building code to use X509_verify_cert(),
9502 it used to have an ad-hoc builder which was unable to cope with anything
9503 other than a simple chain.
9505 *David Woodhouse <dwmw2@infradead.org>, Steve Henson*
9507 * Don't check self signed certificate signatures in X509_verify_cert()
9508 by default (a flag can override this): it just wastes time without
9509 adding any security. As a useful side effect self signed root CAs
9510 with non-FIPS digests are now usable in FIPS mode.
9514 * In dtls1_process_out_of_seq_message() the check if the current message
9515 is already buffered was missing. For every new message was memory
9516 allocated, allowing an attacker to perform an denial of service attack
9517 with sending out of seq handshake messages until there is no memory
9518 left. Additionally every future message was buffered, even if the
9519 sequence number made no sense and would be part of another handshake.
9520 So only messages with sequence numbers less than 10 in advance will be
9521 buffered. ([CVE-2009-1378])
9523 *Robin Seggelmann, discovered by Daniel Mentz*
9525 * Records are buffered if they arrive with a future epoch to be
9526 processed after finishing the corresponding handshake. There is
9527 currently no limitation to this buffer allowing an attacker to perform
9528 a DOS attack with sending records with future epochs until there is no
9529 memory left. This patch adds the pqueue_size() function to determine
9530 the size of a buffer and limits the record buffer to 100 entries.
9533 *Robin Seggelmann, discovered by Daniel Mentz*
9535 * Keep a copy of frag->msg_header.frag_len so it can be used after the
9536 parent structure is freed. ([CVE-2009-1379])
9540 * Handle non-blocking I/O properly in SSL_shutdown() call.
9542 *Darryl Miles <darryl-mailinglists@netbauds.net>*
9544 * Add `2.5.4.*` OIDs
9546 *Ilya O. <vrghost@gmail.com>*
9548 ### Changes between 0.9.8k and 0.9.8l [5 Nov 2009]
9550 * Disable renegotiation completely - this fixes a severe security
9551 problem ([CVE-2009-3555]) at the cost of breaking all
9552 renegotiation. Renegotiation can be re-enabled by setting
9553 SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION in s3->flags at
9554 run-time. This is really not recommended unless you know what
9559 ### Changes between 0.9.8j and 0.9.8k [25 Mar 2009]
9561 * Don't set val to NULL when freeing up structures, it is freed up by
9562 underlying code. If `sizeof(void *) > sizeof(long)` this can result in
9563 zeroing past the valid field. ([CVE-2009-0789])
9565 *Paolo Ganci <Paolo.Ganci@AdNovum.CH>*
9567 * Fix bug where return value of CMS_SignerInfo_verify_content() was not
9568 checked correctly. This would allow some invalid signed attributes to
9569 appear to verify correctly. ([CVE-2009-0591])
9571 *Ivan Nestlerode <inestlerode@us.ibm.com>*
9573 * Reject UniversalString and BMPString types with invalid lengths. This
9574 prevents a crash in ASN1_STRING_print_ex() which assumes the strings have
9575 a legal length. ([CVE-2009-0590])
9579 * Set S/MIME signing as the default purpose rather than setting it
9580 unconditionally. This allows applications to override it at the store
9585 * Permit restricted recursion of ASN1 strings. This is needed in practice
9586 to handle some structures.
9590 * Improve efficiency of mem_gets: don't search whole buffer each time
9593 *Jeremy Shapiro <jnshapir@us.ibm.com>*
9595 * New -hex option for openssl rand.
9599 * Print out UTF8String and NumericString when parsing ASN1.
9603 * Support NumericString type for name components.
9607 * Allow CC in the environment to override the automatically chosen
9608 compiler. Note that nothing is done to ensure flags work with the
9613 ### Changes between 0.9.8i and 0.9.8j [07 Jan 2009]
9615 * Properly check EVP_VerifyFinal() and similar return values
9618 *Ben Laurie, Bodo Moeller, Google Security Team*
9620 * Enable TLS extensions by default.
9624 * Allow the CHIL engine to be loaded, whether the application is
9625 multithreaded or not. (This does not release the developer from the
9626 obligation to set up the dynamic locking callbacks.)
9628 *Sander Temme <sander@temme.net>*
9630 * Use correct exit code if there is an error in dgst command.
9632 *Steve Henson; problem pointed out by Roland Dirlewanger*
9634 * Tweak Configure so that you need to say "experimental-jpake" to enable
9635 JPAKE, and need to use -DOPENSSL_EXPERIMENTAL_JPAKE in applications.
9639 * Add experimental JPAKE support, including demo authentication in
9640 s_client and s_server.
9644 * Set the comparison function in v3_addr_canonize().
9646 *Rob Austein <sra@hactrn.net>*
9648 * Add support for XMPP STARTTLS in s_client.
9650 *Philip Paeps <philip@freebsd.org>*
9652 * Change the server-side SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG behavior
9653 to ensure that even with this option, only ciphersuites in the
9654 server's preference list will be accepted. (Note that the option
9655 applies only when resuming a session, so the earlier behavior was
9656 just about the algorithm choice for symmetric cryptography.)
9660 ### Changes between 0.9.8h and 0.9.8i [15 Sep 2008]
9662 * Fix NULL pointer dereference if a DTLS server received
9663 ChangeCipherSpec as first record ([CVE-2009-1386]).
9667 * Fix a state transition in s3_srvr.c and d1_srvr.c
9668 (was using SSL3_ST_CW_CLNT_HELLO_B, should be `..._ST_SW_SRVR_...`).
9672 * The fix in 0.9.8c that supposedly got rid of unsafe
9673 double-checked locking was incomplete for RSA blinding,
9674 addressing just one layer of what turns out to have been
9675 doubly unsafe triple-checked locking.
9677 So now fix this for real by retiring the MONT_HELPER macro
9678 in crypto/rsa/rsa_eay.c.
9680 *Bodo Moeller; problem pointed out by Marius Schilder*
9682 * Various precautionary measures:
9684 - Avoid size_t integer overflow in HASH_UPDATE (md32_common.h).
9686 - Avoid a buffer overflow in d2i_SSL_SESSION() (ssl_asn1.c).
9687 (NB: This would require knowledge of the secret session ticket key
9688 to exploit, in which case you'd be SOL either way.)
9690 - Change bn_nist.c so that it will properly handle input BIGNUMs
9691 outside the expected range.
9693 - Enforce the 'num' check in BN_div() (bn_div.c) for non-BN_DEBUG
9696 *Neel Mehta, Bodo Moeller*
9698 * Allow engines to be "soft loaded" - i.e. optionally don't die if
9699 the load fails. Useful for distros.
9701 *Ben Laurie and the FreeBSD team*
9703 * Add support for Local Machine Keyset attribute in PKCS#12 files.
9707 * Fix BN_GF2m_mod_arr() top-bit cleanup code.
9711 * Expand ENGINE to support engine supplied SSL client certificate functions.
9713 This work was sponsored by Logica.
9717 * Add CryptoAPI ENGINE to support use of RSA and DSA keys held in Windows
9718 keystores. Support for SSL/TLS client authentication too.
9719 Not compiled unless enable-capieng specified to Configure.
9721 This work was sponsored by Logica.
9725 * Fix bug in X509_ATTRIBUTE creation: don't set attribute using
9726 ASN1_TYPE_set1 if MBSTRING flag set. This bug would crash certain
9727 attribute creation routines such as certificate requests and PKCS#12
9732 ### Changes between 0.9.8g and 0.9.8h [28 May 2008]
9734 * Fix flaw if 'Server Key exchange message' is omitted from a TLS
9735 handshake which could lead to a client crash as found using the
9736 Codenomicon TLS test suite ([CVE-2008-1672])
9738 *Steve Henson, Mark Cox*
9740 * Fix double free in TLS server name extensions which could lead to
9741 a remote crash found by Codenomicon TLS test suite ([CVE-2008-0891])
9745 * Clear error queue in SSL_CTX_use_certificate_chain_file()
9747 Clear the error queue to ensure that error entries left from
9748 older function calls do not interfere with the correct operation.
9750 *Lutz Jaenicke, Erik de Castro Lopo*
9752 * Remove root CA certificates of commercial CAs:
9754 The OpenSSL project does not recommend any specific CA and does not
9755 have any policy with respect to including or excluding any CA.
9756 Therefore it does not make any sense to ship an arbitrary selection
9757 of root CA certificates with the OpenSSL software.
9761 * RSA OAEP patches to fix two separate invalid memory reads.
9762 The first one involves inputs when 'lzero' is greater than
9763 'SHA_DIGEST_LENGTH' (it would read about SHA_DIGEST_LENGTH bytes
9764 before the beginning of from). The second one involves inputs where
9765 the 'db' section contains nothing but zeroes (there is a one-byte
9766 invalid read after the end of 'db').
9768 *Ivan Nestlerode <inestlerode@us.ibm.com>*
9770 * Partial backport from 0.9.9-dev:
9772 Introduce bn_mul_mont (dedicated Montgomery multiplication
9773 procedure) as a candidate for BIGNUM assembler implementation.
9774 While 0.9.9-dev uses assembler for various architectures, only
9775 x86_64 is available by default here in the 0.9.8 branch, and
9776 32-bit x86 is available through a compile-time setting.
9778 To try the 32-bit x86 assembler implementation, use Configure
9779 option "enable-montasm" (which exists only for this backport).
9781 As "enable-montasm" for 32-bit x86 disclaims code stability
9782 anyway, in this constellation we activate additional code
9783 backported from 0.9.9-dev for further performance improvements,
9784 namely BN_from_montgomery_word. (To enable this otherwise,
9785 e.g. x86_64, try `-DMONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD`.)
9787 *Andy Polyakov (backport partially by Bodo Moeller)*
9789 * Add TLS session ticket callback. This allows an application to set
9790 TLS ticket cipher and HMAC keys rather than relying on hardcoded fixed
9791 values. This is useful for key rollover for example where several key
9792 sets may exist with different names.
9796 * Reverse ENGINE-internal logic for caching default ENGINE handles.
9797 This was broken until now in 0.9.8 releases, such that the only way
9798 a registered ENGINE could be used (assuming it initialises
9799 successfully on the host) was to explicitly set it as the default
9800 for the relevant algorithms. This is in contradiction with 0.9.7
9801 behaviour and the documentation. With this fix, when an ENGINE is
9802 registered into a given algorithm's table of implementations, the
9803 'uptodate' flag is reset so that auto-discovery will be used next
9804 time a new context for that algorithm attempts to select an
9807 *Ian Lister (tweaked by Geoff Thorpe)*
9809 * Backport of CMS code to OpenSSL 0.9.8. This differs from the 0.9.9
9810 implementation in the following ways:
9812 Lack of EVP_PKEY_ASN1_METHOD means algorithm parameters have to be
9815 Lack of BER streaming support means one pass streaming processing is
9816 only supported if data is detached: setting the streaming flag is
9817 ignored for embedded content.
9819 CMS support is disabled by default and must be explicitly enabled
9820 with the enable-cms configuration option.
9824 * Update the GMP engine glue to do direct copies between BIGNUM and
9825 mpz_t when openssl and GMP use the same limb size. Otherwise the
9826 existing "conversion via a text string export" trick is still used.
9828 *Paul Sheer <paulsheer@gmail.com>*
9830 * Zlib compression BIO. This is a filter BIO which compressed and
9831 uncompresses any data passed through it.
9835 * Add AES_wrap_key() and AES_unwrap_key() functions to implement
9836 RFC3394 compatible AES key wrapping.
9840 * Add utility functions to handle ASN1 structures. ASN1_STRING_set0():
9841 sets string data without copying. X509_ALGOR_set0() and
9842 X509_ALGOR_get0(): set and retrieve X509_ALGOR (AlgorithmIdentifier)
9843 data. Attribute function X509at_get0_data_by_OBJ(): retrieves data
9844 from an X509_ATTRIBUTE structure optionally checking it occurs only
9845 once. ASN1_TYPE_set1(): set and ASN1_TYPE structure copying supplied
9850 * Fix BN flag handling in RSA_eay_mod_exp() and BN_MONT_CTX_set()
9851 to get the expected BN_FLG_CONSTTIME behavior.
9853 *Bodo Moeller (Google)*
9857 - fixed wrong usage of ioctlsocket() when build for LIBC BSD sockets
9858 - fixed do_tests.pl to run the test suite with CLIB builds too (CLIB_OPT)
9859 - added some more tests to do_tests.pl
9860 - fixed RunningProcess usage so that it works with newer LIBC NDKs too
9861 - removed usage of BN_LLONG for CLIB builds to avoid runtime dependency
9862 - added new Configure targets netware-clib-bsdsock, netware-clib-gcc,
9863 netware-clib-bsdsock-gcc, netware-libc-bsdsock-gcc
9864 - various changes to netware.pl to enable gcc-cross builds on Win32
9866 - changed crypto/bio/b_sock.c to work with macro functions (CLIB BSD)
9867 - various changes to fix missing prototype warnings
9868 - fixed x86nasm.pl to create correct asm files for NASM COFF output
9869 - added AES, WHIRLPOOL and CPUID assembler code to build files
9870 - added missing AES assembler make rules to mk1mf.pl
9871 - fixed order of includes in `apps/ocsp.c` so that `e_os.h` settings apply
9873 *Guenter Knauf <eflash@gmx.net>*
9875 * Implement certificate status request TLS extension defined in RFC3546.
9876 A client can set the appropriate parameters and receive the encoded
9877 OCSP response via a callback. A server can query the supplied parameters
9878 and set the encoded OCSP response in the callback. Add simplified examples
9879 to s_client and s_server.
9883 ### Changes between 0.9.8f and 0.9.8g [19 Oct 2007]
9886 + Binary incompatibility of ssl_ctx_st structure
9887 + DTLS interoperation with non-compliant servers
9888 + Don't call get_session_cb() without proposed session
9889 + Fix ia64 assembler code
9891 *Andy Polyakov, Steve Henson*
9893 ### Changes between 0.9.8e and 0.9.8f [11 Oct 2007]
9895 * DTLS Handshake overhaul. There were longstanding issues with
9896 OpenSSL DTLS implementation, which were making it impossible for
9897 RFC 4347 compliant client to communicate with OpenSSL server.
9898 Unfortunately just fixing these incompatibilities would "cut off"
9899 pre-0.9.8f clients. To allow for hassle free upgrade post-0.9.8e
9900 server keeps tolerating non RFC compliant syntax. The opposite is
9901 not true, 0.9.8f client can not communicate with earlier server.
9902 This update even addresses CVE-2007-4995.
9906 * Changes to avoid need for function casts in OpenSSL: some compilers
9907 (gcc 4.2 and later) reject their use.
9908 *Kurt Roeckx <kurt@roeckx.be>, Peter Hartley <pdh@utter.chaos.org.uk>,
9911 * Add RFC4507 support to OpenSSL. This includes the corrections in
9912 RFC4507bis. The encrypted ticket format is an encrypted encoded
9913 SSL_SESSION structure, that way new session features are automatically
9916 If a client application caches session in an SSL_SESSION structure
9917 support is transparent because tickets are now stored in the encoded
9920 The SSL_CTX structure automatically generates keys for ticket
9921 protection in servers so again support should be possible
9922 with no application modification.
9924 If a client or server wishes to disable RFC4507 support then the option
9925 SSL_OP_NO_TICKET can be set.
9927 Add a TLS extension debugging callback to allow the contents of any client
9928 or server extensions to be examined.
9930 This work was sponsored by Google.
9934 * Add initial support for TLS extensions, specifically for the server_name
9935 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
9936 have new members for a host name. The SSL data structure has an
9937 additional member `SSL_CTX *initial_ctx` so that new sessions can be
9938 stored in that context to allow for session resumption, even after the
9939 SSL has been switched to a new SSL_CTX in reaction to a client's
9940 server_name extension.
9942 New functions (subject to change):
9944 SSL_get_servername()
9945 SSL_get_servername_type()
9948 New CTRL codes and macros (subject to change):
9950 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
9951 - SSL_CTX_set_tlsext_servername_callback()
9952 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
9953 - SSL_CTX_set_tlsext_servername_arg()
9954 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
9956 openssl s_client has a new '-servername ...' option.
9958 openssl s_server has new options '-servername_host ...', '-cert2 ...',
9959 '-key2 ...', '-servername_fatal' (subject to change). This allows
9960 testing the HostName extension for a specific single host name ('-cert'
9961 and '-key' remain fallbacks for handshakes without HostName
9962 negotiation). If the unrecognized_name alert has to be sent, this by
9963 default is a warning; it becomes fatal with the '-servername_fatal'
9966 *Peter Sylvester, Remy Allais, Christophe Renou, Steve Henson*
9968 * Add AES and SSE2 assembly language support to VC++ build.
9972 * Mitigate attack on final subtraction in Montgomery reduction.
9976 * Fix crypto/ec/ec_mult.c to work properly with scalars of value 0
9977 (which previously caused an internal error).
9981 * Squeeze another 10% out of IGE mode when in != out.
9985 * AES IGE mode speedup.
9987 *Dean Gaudet (Google)*
9989 * Add the Korean symmetric 128-bit cipher SEED (see
9990 <http://www.kisa.or.kr/kisa/seed/jsp/seed_eng.jsp>) and
9991 add SEED ciphersuites from RFC 4162:
9993 TLS_RSA_WITH_SEED_CBC_SHA = "SEED-SHA"
9994 TLS_DHE_DSS_WITH_SEED_CBC_SHA = "DHE-DSS-SEED-SHA"
9995 TLS_DHE_RSA_WITH_SEED_CBC_SHA = "DHE-RSA-SEED-SHA"
9996 TLS_DH_anon_WITH_SEED_CBC_SHA = "ADH-SEED-SHA"
9998 To minimize changes between patchlevels in the OpenSSL 0.9.8
9999 series, SEED remains excluded from compilation unless OpenSSL
10000 is configured with 'enable-seed'.
10002 *KISA, Bodo Moeller*
10004 * Mitigate branch prediction attacks, which can be practical if a
10005 single processor is shared, allowing a spy process to extract
10006 information. For detailed background information, see
10007 <http://eprint.iacr.org/2007/039> (O. Aciicmez, S. Gueron,
10008 J.-P. Seifert, "New Branch Prediction Vulnerabilities in OpenSSL
10009 and Necessary Software Countermeasures"). The core of the change
10010 are new versions BN_div_no_branch() and
10011 BN_mod_inverse_no_branch() of BN_div() and BN_mod_inverse(),
10012 respectively, which are slower, but avoid the security-relevant
10013 conditional branches. These are automatically called by BN_div()
10014 and BN_mod_inverse() if the flag BN_FLG_CONSTTIME is set for one
10015 of the input BIGNUMs. Also, BN_is_bit_set() has been changed to
10016 remove a conditional branch.
10018 BN_FLG_CONSTTIME is the new name for the previous
10019 BN_FLG_EXP_CONSTTIME flag, since it now affects more than just
10020 modular exponentiation. (Since OpenSSL 0.9.7h, setting this flag
10021 in the exponent causes BN_mod_exp_mont() to use the alternative
10022 implementation in BN_mod_exp_mont_consttime().) The old name
10023 remains as a deprecated alias.
10025 Similarly, RSA_FLAG_NO_EXP_CONSTTIME is replaced by a more general
10026 RSA_FLAG_NO_CONSTTIME flag since the RSA implementation now uses
10027 constant-time implementations for more than just exponentiation.
10028 Here too the old name is kept as a deprecated alias.
10030 BN_BLINDING_new() will now use BN_dup() for the modulus so that
10031 the BN_BLINDING structure gets an independent copy of the
10032 modulus. This means that the previous `BIGNUM *m` argument to
10033 BN_BLINDING_new() and to BN_BLINDING_create_param() now
10034 essentially becomes `const BIGNUM *m`, although we can't actually
10035 change this in the header file before 0.9.9. It allows
10036 RSA_setup_blinding() to use BN_with_flags() on the modulus to
10037 enable BN_FLG_CONSTTIME.
10039 *Matthew D Wood (Intel Corp)*
10041 * In the SSL/TLS server implementation, be strict about session ID
10042 context matching (which matters if an application uses a single
10043 external cache for different purposes). Previously,
10044 out-of-context reuse was forbidden only if SSL_VERIFY_PEER was
10045 set. This did ensure strict client verification, but meant that,
10046 with applications using a single external cache for quite
10047 different requirements, clients could circumvent ciphersuite
10048 restrictions for a given session ID context by starting a session
10049 in a different context.
10053 * Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
10054 a ciphersuite string such as "DEFAULT:RSA" cannot enable
10055 authentication-only ciphersuites.
10059 * Update the SSL_get_shared_ciphers() fix CVE-2006-3738 which was
10060 not complete and could lead to a possible single byte overflow
10061 ([CVE-2007-5135]) [Ben Laurie]
10063 ### Changes between 0.9.8d and 0.9.8e [23 Feb 2007]
10065 * Since AES128 and AES256 (and similarly Camellia128 and
10066 Camellia256) share a single mask bit in the logic of
10067 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
10068 kludge to work properly if AES128 is available and AES256 isn't
10069 (or if Camellia128 is available and Camellia256 isn't).
10073 * Fix the BIT STRING encoding generated by crypto/ec/ec_asn1.c
10074 (within i2d_ECPrivateKey, i2d_ECPKParameters, i2d_ECParameters):
10075 When a point or a seed is encoded in a BIT STRING, we need to
10076 prevent the removal of trailing zero bits to get the proper DER
10077 encoding. (By default, crypto/asn1/a_bitstr.c assumes the case
10078 of a NamedBitList, for which trailing 0 bits need to be removed.)
10082 * Have SSL/TLS server implementation tolerate "mismatched" record
10083 protocol version while receiving ClientHello even if the
10084 ClientHello is fragmented. (The server can't insist on the
10085 particular protocol version it has chosen before the ServerHello
10086 message has informed the client about his choice.)
10090 * Add RFC 3779 support.
10092 *Rob Austein for ARIN, Ben Laurie*
10094 * Load error codes if they are not already present instead of using a
10095 static variable. This allows them to be cleanly unloaded and reloaded.
10096 Improve header file function name parsing.
10100 * extend SMTP and IMAP protocol emulation in s_client to use EHLO
10101 or CAPABILITY handshake as required by RFCs.
10103 *Goetz Babin-Ebell*
10105 ### Changes between 0.9.8c and 0.9.8d [28 Sep 2006]
10107 * Introduce limits to prevent malicious keys being able to
10108 cause a denial of service. ([CVE-2006-2940])
10110 *Steve Henson, Bodo Moeller*
10112 * Fix ASN.1 parsing of certain invalid structures that can result
10113 in a denial of service. ([CVE-2006-2937]) [Steve Henson]
10115 * Fix buffer overflow in SSL_get_shared_ciphers() function.
10116 ([CVE-2006-3738]) [Tavis Ormandy and Will Drewry, Google Security Team]
10118 * Fix SSL client code which could crash if connecting to a
10119 malicious SSLv2 server. ([CVE-2006-4343])
10121 *Tavis Ormandy and Will Drewry, Google Security Team*
10123 * Since 0.9.8b, ciphersuite strings naming explicit ciphersuites
10124 match only those. Before that, "AES256-SHA" would be interpreted
10125 as a pattern and match "AES128-SHA" too (since AES128-SHA got
10126 the same strength classification in 0.9.7h) as we currently only
10127 have a single AES bit in the ciphersuite description bitmap.
10128 That change, however, also applied to ciphersuite strings such as
10129 "RC4-MD5" that intentionally matched multiple ciphersuites --
10130 namely, SSL 2.0 ciphersuites in addition to the more common ones
10131 from SSL 3.0/TLS 1.0.
10133 So we change the selection algorithm again: Naming an explicit
10134 ciphersuite selects this one ciphersuite, and any other similar
10135 ciphersuite (same bitmap) from *other* protocol versions.
10136 Thus, "RC4-MD5" again will properly select both the SSL 2.0
10137 ciphersuite and the SSL 3.0/TLS 1.0 ciphersuite.
10139 Since SSL 2.0 does not have any ciphersuites for which the
10140 128/256 bit distinction would be relevant, this works for now.
10141 The proper fix will be to use different bits for AES128 and
10142 AES256, which would have avoided the problems from the beginning;
10143 however, bits are scarce, so we can only do this in a new release
10144 (not just a patchlevel) when we can change the SSL_CIPHER
10145 definition to split the single 'unsigned long mask' bitmap into
10146 multiple values to extend the available space.
10150 ### Changes between 0.9.8b and 0.9.8c [05 Sep 2006]
10152 * Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
10153 ([CVE-2006-4339]) [Ben Laurie and Google Security Team]
10155 * Add AES IGE and biIGE modes.
10159 * Change the Unix randomness entropy gathering to use poll() when
10160 possible instead of select(), since the latter has some
10161 undesirable limitations.
10163 *Darryl Miles via Richard Levitte and Bodo Moeller*
10165 * Disable "ECCdraft" ciphersuites more thoroughly. Now special
10166 treatment in ssl/ssl_ciph.s makes sure that these ciphersuites
10167 cannot be implicitly activated as part of, e.g., the "AES" alias.
10168 However, please upgrade to OpenSSL 0.9.9[-dev] for
10169 non-experimental use of the ECC ciphersuites to get TLS extension
10170 support, which is required for curve and point format negotiation
10171 to avoid potential handshake problems.
10175 * Disable rogue ciphersuites:
10177 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
10178 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
10179 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
10181 The latter two were purportedly from
10182 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
10185 Also deactivate the remaining ciphersuites from
10186 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
10187 unofficial, and the ID has long expired.
10191 * Fix RSA blinding Heisenbug (problems sometimes occurred on
10192 dual-core machines) and other potential thread-safety issues.
10196 * Add the symmetric cipher Camellia (128-bit, 192-bit, 256-bit key
10197 versions), which is now available for royalty-free use
10198 (see <http://info.isl.ntt.co.jp/crypt/eng/info/chiteki.html>).
10199 Also, add Camellia TLS ciphersuites from RFC 4132.
10201 To minimize changes between patchlevels in the OpenSSL 0.9.8
10202 series, Camellia remains excluded from compilation unless OpenSSL
10203 is configured with 'enable-camellia'.
10207 * Disable the padding bug check when compression is in use. The padding
10208 bug check assumes the first packet is of even length, this is not
10209 necessarily true if compression is enabled and can result in false
10210 positives causing handshake failure. The actual bug test is ancient
10211 code so it is hoped that implementations will either have fixed it by
10212 now or any which still have the bug do not support compression.
10216 ### Changes between 0.9.8a and 0.9.8b [04 May 2006]
10218 * When applying a cipher rule check to see if string match is an explicit
10219 cipher suite and only match that one cipher suite if it is.
10223 * Link in manifests for VC++ if needed.
10225 *Austin Ziegler <halostatue@gmail.com>*
10227 * Update support for ECC-based TLS ciphersuites according to
10228 draft-ietf-tls-ecc-12.txt with proposed changes (but without
10229 TLS extensions, which are supported starting with the 0.9.9
10230 branch, not in the OpenSSL 0.9.8 branch).
10234 * New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support
10235 opaque EVP_CIPHER_CTX handling.
10239 * Fixes and enhancements to zlib compression code. We now only use
10240 "zlib1.dll" and use the default `__cdecl` calling convention on Win32
10241 to conform with the standards mentioned here:
10242 <http://www.zlib.net/DLL_FAQ.txt>
10243 Static zlib linking now works on Windows and the new --with-zlib-include
10244 --with-zlib-lib options to Configure can be used to supply the location
10245 of the headers and library. Gracefully handle case where zlib library
10250 * Several fixes and enhancements to the OID generation code. The old code
10251 sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't
10252 handle numbers larger than ULONG_MAX, truncated printing and had a
10253 non standard OBJ_obj2txt() behaviour.
10257 * Add support for building of engines under engine/ as shared libraries
10258 under VC++ build system.
10262 * Corrected the numerous bugs in the Win32 path splitter in DSO.
10263 Hopefully, we will not see any false combination of paths any more.
10267 ### Changes between 0.9.8 and 0.9.8a [11 Oct 2005]
10269 * Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
10270 (part of SSL_OP_ALL). This option used to disable the
10271 countermeasure against man-in-the-middle protocol-version
10272 rollback in the SSL 2.0 server implementation, which is a bad
10273 idea. ([CVE-2005-2969])
10275 *Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
10276 for Information Security, National Institute of Advanced Industrial
10277 Science and Technology [AIST], Japan)*
10279 * Add two function to clear and return the verify parameter flags.
10283 * Keep cipherlists sorted in the source instead of sorting them at
10284 runtime, thus removing the need for a lock.
10288 * Avoid some small subgroup attacks in Diffie-Hellman.
10290 *Nick Mathewson and Ben Laurie*
10292 * Add functions for well-known primes.
10296 * Extended Windows CE support.
10298 *Satoshi Nakamura and Andy Polyakov*
10300 * Initialize SSL_METHOD structures at compile time instead of during
10301 runtime, thus removing the need for a lock.
10305 * Make PKCS7_decrypt() work even if no certificate is supplied by
10306 attempting to decrypt each encrypted key in turn. Add support to
10311 ### Changes between 0.9.7h and 0.9.8 [05 Jul 2005]
10313 [NB: OpenSSL 0.9.7i and later 0.9.7 patch levels were released after
10316 * Add libcrypto.pc and libssl.pc for those who feel they need them.
10320 * Change CA.sh and CA.pl so they don't bundle the CSR and the private
10321 key into the same file any more.
10325 * Add initial support for Win64, both IA64 and AMD64/x64 flavors.
10329 * Add -utf8 command line and config file option to 'ca'.
10331 *Stefan <stf@udoma.org*
10333 * Removed the macro des_crypt(), as it seems to conflict with some
10334 libraries. Use DES_crypt().
10338 * Correct naming of the 'chil' and '4758cca' ENGINEs. This
10339 involves renaming the source and generated shared-libs for
10340 both. The engines will accept the corrected or legacy ids
10341 ('ncipher' and '4758_cca' respectively) when binding. NB,
10342 this only applies when building 'shared'.
10344 *Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe*
10346 * Add attribute functions to EVP_PKEY structure. Modify
10347 PKCS12_create() to recognize a CSP name attribute and
10348 use it. Make -CSP option work again in pkcs12 utility.
10352 * Add new functionality to the bn blinding code:
10353 - automatic re-creation of the BN_BLINDING parameters after
10354 a fixed number of uses (currently 32)
10355 - add new function for parameter creation
10356 - introduce flags to control the update behaviour of the
10357 BN_BLINDING parameters
10358 - hide BN_BLINDING structure
10359 Add a second BN_BLINDING slot to the RSA structure to improve
10360 performance when a single RSA object is shared among several
10365 * Add support for DTLS.
10367 *Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie*
10369 * Add support for DER encoded private keys (SSL_FILETYPE_ASN1)
10370 to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file()
10374 * Remove buggy and incomplete DH cert support from
10375 ssl/ssl_rsa.c and ssl/s3_both.c
10379 * Use SHA-1 instead of MD5 as the default digest algorithm for
10380 the `apps/openssl` commands.
10384 * Compile clean with "-Wall -Wmissing-prototypes
10385 -Wstrict-prototypes -Wmissing-declarations -Werror". Currently
10386 DEBUG_SAFESTACK must also be set.
10390 * Change ./Configure so that certain algorithms can be disabled by default.
10391 The new counterpiece to "no-xxx" is "enable-xxx".
10393 The patented RC5 and MDC2 algorithms will now be disabled unless
10394 "enable-rc5" and "enable-mdc2", respectively, are specified.
10396 (IDEA remains enabled despite being patented. This is because IDEA
10397 is frequently required for interoperability, and there is no license
10398 fee for non-commercial use. As before, "no-idea" can be used to
10399 avoid this algorithm.)
10403 * Add processing of proxy certificates (see RFC 3820). This work was
10404 sponsored by KTH (The Royal Institute of Technology in Stockholm) and
10405 EGEE (Enabling Grids for E-science in Europe).
10409 * RC4 performance overhaul on modern architectures/implementations, such
10410 as Intel P4, IA-64 and AMD64.
10414 * New utility extract-section.pl. This can be used specify an alternative
10415 section number in a pod file instead of having to treat each file as
10416 a separate case in Makefile. This can be done by adding two lines to the
10419 =for comment openssl_section:XXX
10421 The blank line is mandatory.
10425 * New arguments -certform, -keyform and -pass for s_client and s_server
10426 to allow alternative format key and certificate files and passphrase
10431 * New structure X509_VERIFY_PARAM which combines current verify parameters,
10432 update associated structures and add various utility functions.
10434 Add new policy related verify parameters, include policy checking in
10435 standard verify code. Enhance 'smime' application with extra parameters
10436 to support policy checking and print out.
10440 * Add a new engine to support VIA PadLock ACE extensions in the VIA C3
10441 Nehemiah processors. These extensions support AES encryption in hardware
10442 as well as RNG (though RNG support is currently disabled).
10444 *Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov*
10446 * Deprecate `BN_[get|set]_params()` functions (they were ignored internally).
10450 * New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented.
10452 *Andy Polyakov and a number of other people*
10454 * Improved PowerPC platform support. Most notably BIGNUM assembler
10455 implementation contributed by IBM.
10457 *Suresh Chari, Peter Waltenberg, Andy Polyakov*
10459 * The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public
10460 exponent rather than 'unsigned long'. There is a corresponding change to
10461 the new 'rsa_keygen' element of the RSA_METHOD structure.
10463 *Jelte Jansen, Geoff Thorpe*
10465 * Functionality for creating the initial serial number file is now
10466 moved from CA.pl to the 'ca' utility with a new option -create_serial.
10468 (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial
10469 number file to 1, which is bound to cause problems. To avoid
10470 the problems while respecting compatibility between different 0.9.7
10471 patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in
10472 CA.pl for serial number initialization. With the new release 0.9.8,
10473 we can fix the problem directly in the 'ca' utility.)
10477 * Reduced header interdependencies by declaring more opaque objects in
10478 ossl_typ.h. As a consequence, including some headers (eg. engine.h) will
10479 give fewer recursive includes, which could break lazy source code - so
10480 this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always,
10481 developers should define this symbol when building and using openssl to
10482 ensure they track the recommended behaviour, interfaces, [etc], but
10483 backwards-compatible behaviour prevails when this isn't defined.
10487 * New function X509_POLICY_NODE_print() which prints out policy nodes.
10491 * Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality.
10492 This will generate a random key of the appropriate length based on the
10493 cipher context. The EVP_CIPHER can provide its own random key generation
10494 routine to support keys of a specific form. This is used in the des and
10495 3des routines to generate a key of the correct parity. Update S/MIME
10496 code to use new functions and hence generate correct parity DES keys.
10497 Add EVP_CHECK_DES_KEY #define to return an error if the key is not
10498 valid (weak or incorrect parity).
10502 * Add a local set of CRLs that can be used by X509_verify_cert() as well
10503 as looking them up. This is useful when the verified structure may contain
10504 CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs
10505 present unless the new PKCS7_NO_CRL flag is asserted.
10509 * Extend ASN1 oid configuration module. It now additionally accepts the
10512 shortName = some long name, 1.2.3.4
10516 * Reimplemented the BN_CTX implementation. There is now no more static
10517 limitation on the number of variables it can handle nor the depth of the
10518 "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack
10519 information can now expand as required, and rather than having a single
10520 static array of bignums, BN_CTX now uses a linked-list of such arrays
10521 allowing it to expand on demand whilst maintaining the usefulness of
10522 BN_CTX's "bundling".
10526 * Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD
10527 to allow all RSA operations to function using a single BN_CTX.
10531 * Preliminary support for certificate policy evaluation and checking. This
10532 is initially intended to pass the tests outlined in "Conformance Testing
10533 of Relying Party Client Certificate Path Processing Logic" v1.07.
10537 * bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and
10538 remained unused and not that useful. A variety of other little bignum
10539 tweaks and fixes have also been made continuing on from the audit (see
10544 * Constify all or almost all d2i, c2i, s2i and r2i functions, along with
10545 associated ASN1, EVP and SSL functions and old ASN1 macros.
10549 * BN_zero() only needs to set 'top' and 'neg' to zero for correct results,
10550 and this should never fail. So the return value from the use of
10551 BN_set_word() (which can fail due to needless expansion) is now deprecated;
10552 if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro.
10556 * BN_CTX_get() should return zero-valued bignums, providing the same
10557 initialised value as BN_new().
10559 *Geoff Thorpe, suggested by Ulf Möller*
10561 * Support for inhibitAnyPolicy certificate extension.
10565 * An audit of the BIGNUM code is underway, for which debugging code is
10566 enabled when BN_DEBUG is defined. This makes stricter enforcements on what
10567 is considered valid when processing BIGNUMs, and causes execution to
10568 assert() when a problem is discovered. If BN_DEBUG_RAND is defined,
10569 further steps are taken to deliberately pollute unused data in BIGNUM
10570 structures to try and expose faulty code further on. For now, openssl will
10571 (in its default mode of operation) continue to tolerate the inconsistent
10572 forms that it has tolerated in the past, but authors and packagers should
10573 consider trying openssl and their own applications when compiled with
10574 these debugging symbols defined. It will help highlight potential bugs in
10575 their own code, and will improve the test coverage for OpenSSL itself. At
10576 some point, these tighter rules will become openssl's default to improve
10577 maintainability, though the assert()s and other overheads will remain only
10578 in debugging configurations. See bn.h for more details.
10580 *Geoff Thorpe, Nils Larsch, Ulf Möller*
10582 * BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure
10583 that can only be obtained through BN_CTX_new() (which implicitly
10584 initialises it). The presence of this function only made it possible
10585 to overwrite an existing structure (and cause memory leaks).
10589 * Because of the callback-based approach for implementing LHASH as a
10590 template type, lh_insert() adds opaque objects to hash-tables and
10591 lh_doall() or lh_doall_arg() are typically used with a destructor callback
10592 to clean up those corresponding objects before destroying the hash table
10593 (and losing the object pointers). So some over-zealous constifications in
10594 LHASH have been relaxed so that lh_insert() does not take (nor store) the
10595 objects as "const" and the `lh_doall[_arg]` callback wrappers are not
10596 prototyped to have "const" restrictions on the object pointers they are
10597 given (and so aren't required to cast them away any more).
10601 * The tmdiff.h API was so ugly and minimal that our own timing utility
10602 (speed) prefers to use its own implementation. The two implementations
10603 haven't been consolidated as yet (volunteers?) but the tmdiff API has had
10604 its object type properly exposed (MS_TM) instead of casting to/from
10605 `char *`. This may still change yet if someone realises MS_TM and
10607 aren't necessarily the greatest nomenclatures - but this is what was used
10608 internally to the implementation so I've used that for now.
10612 * Ensure that deprecated functions do not get compiled when
10613 OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of
10614 the self-tests were still using deprecated key-generation functions so
10615 these have been updated also.
10619 * Reorganise PKCS#7 code to separate the digest location functionality
10620 into PKCS7_find_digest(), digest addition into PKCS7_bio_add_digest().
10621 New function PKCS7_set_digest() to set the digest type for PKCS#7
10622 digestedData type. Add additional code to correctly generate the
10623 digestedData type and add support for this type in PKCS7 initialization
10628 * New function PKCS7_set0_type_other() this initializes a PKCS7
10629 structure of type "other".
10633 * Fix prime generation loop in crypto/bn/bn_prime.pl by making
10634 sure the loop does correctly stop and breaking ("division by zero")
10635 modulus operations are not performed. The (pre-generated) prime
10636 table crypto/bn/bn_prime.h was already correct, but it could not be
10637 re-generated on some platforms because of the "division by zero"
10638 situation in the script.
10640 *Ralf S. Engelschall*
10642 * Update support for ECC-based TLS ciphersuites according to
10643 draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with
10644 SHA-1 now is only used for "small" curves (where the
10645 representation of a field element takes up to 24 bytes); for
10646 larger curves, the field element resulting from ECDH is directly
10647 used as premaster secret.
10649 *Douglas Stebila (Sun Microsystems Laboratories)*
10651 * Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2
10652 curve secp160r1 to the tests.
10654 *Douglas Stebila (Sun Microsystems Laboratories)*
10656 * Add the possibility to load symbols globally with DSO.
10658 *Götz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte*
10660 * Add the functions ERR_set_mark() and ERR_pop_to_mark() for better
10661 control of the error stack.
10665 * Add support for STORE in ENGINE.
10669 * Add the STORE type. The intention is to provide a common interface
10670 to certificate and key stores, be they simple file-based stores, or
10671 HSM-type store, or LDAP stores, or...
10672 NOTE: The code is currently UNTESTED and isn't really used anywhere.
10676 * Add a generic structure called OPENSSL_ITEM. This can be used to
10677 pass a list of arguments to any function as well as provide a way
10678 for a function to pass data back to the caller.
10682 * Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup()
10683 works like BUF_strdup() but can be used to duplicate a portion of
10684 a string. The copy gets NUL-terminated. BUF_memdup() duplicates
10689 * Add the function sk_find_ex() which works like sk_find(), but will
10690 return an index to an element even if an exact match couldn't be
10691 found. The index is guaranteed to point at the element where the
10692 searched-for key would be inserted to preserve sorting order.
10696 * Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but
10697 takes an extra flags argument for optional functionality. Currently,
10698 the following flags are defined:
10700 OBJ_BSEARCH_VALUE_ON_NOMATCH
10701 This one gets OBJ_bsearch_ex() to return a pointer to the first
10702 element where the comparing function returns a negative or zero
10705 OBJ_BSEARCH_FIRST_VALUE_ON_MATCH
10706 This one gets OBJ_bsearch_ex() to return a pointer to the first
10707 element where the comparing function returns zero. This is useful
10708 if there are more than one element where the comparing function
10713 * Make it possible to create self-signed certificates with 'openssl ca'
10714 in such a way that the self-signed certificate becomes part of the
10715 CA database and uses the same mechanisms for serial number generation
10716 as all other certificate signing. The new flag '-selfsign' enables
10717 this functionality. Adapt CA.sh and CA.pl.in.
10721 * Add functionality to check the public key of a certificate request
10722 against a given private. This is useful to check that a certificate
10723 request can be signed by that key (self-signing).
10727 * Make it possible to have multiple active certificates with the same
10728 subject in the CA index file. This is done only if the keyword
10729 'unique_subject' is set to 'no' in the main CA section (default
10730 if 'CA_default') of the configuration file. The value is saved
10731 with the database itself in a separate index attribute file,
10732 named like the index file with '.attr' appended to the name.
10736 * Generate multi-valued AVAs using '+' notation in config files for
10741 * Support for nameConstraints certificate extension.
10745 * Support for policyConstraints certificate extension.
10749 * Support for policyMappings certificate extension.
10753 * Make sure the default DSA_METHOD implementation only uses its
10754 dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL,
10755 and change its own handlers to be NULL so as to remove unnecessary
10756 indirection. This lets alternative implementations fallback to the
10757 default implementation more easily.
10761 * Support for directoryName in GeneralName related extensions
10766 * Make it possible to link applications using Makefile.shared.
10767 Make that possible even when linking against static libraries!
10771 * Support for single pass processing for S/MIME signing. This now
10772 means that S/MIME signing can be done from a pipe, in addition
10773 cleartext signing (multipart/signed type) is effectively streaming
10774 and the signed data does not need to be all held in memory.
10776 This is done with a new flag PKCS7_STREAM. When this flag is set
10777 PKCS7_sign() only initializes the PKCS7 structure and the actual signing
10778 is done after the data is output (and digests calculated) in
10779 SMIME_write_PKCS7().
10783 * Add full support for -rpath/-R, both in shared libraries and
10784 applications, at least on the platforms where it's known how
10789 * In crypto/ec/ec_mult.c, implement fast point multiplication with
10790 precomputation, based on wNAF splitting: EC_GROUP_precompute_mult()
10791 will now compute a table of multiples of the generator that
10792 makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul()
10793 faster (notably in the case of a single point multiplication,
10794 scalar * generator).
10796 *Nils Larsch, Bodo Moeller*
10798 * IPv6 support for certificate extensions. The various extensions
10799 which use the IP:a.b.c.d can now take IPv6 addresses using the
10800 formats of RFC1884 2.2 . IPv6 addresses are now also displayed
10805 * Added an ENGINE that implements RSA by performing private key
10806 exponentiations with the GMP library. The conversions to and from
10807 GMP's mpz_t format aren't optimised nor are any montgomery forms
10808 cached, and on x86 it appears OpenSSL's own performance has caught up.
10809 However there are likely to be other architectures where GMP could
10810 provide a boost. This ENGINE is not built in by default, but it can be
10811 specified at Configure time and should be accompanied by the necessary
10812 linker additions, eg;
10813 ./config -DOPENSSL_USE_GMP -lgmp
10817 * "openssl engine" will not display ENGINE/DSO load failure errors when
10818 testing availability of engines with "-t" - the old behaviour is
10819 produced by increasing the feature's verbosity with "-tt".
10823 * ECDSA routines: under certain error conditions uninitialized BN objects
10824 could be freed. Solution: make sure initialization is performed early
10825 enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de>
10830 * Key-generation can now be implemented in RSA_METHOD, DSA_METHOD
10831 and DH_METHOD (eg. by ENGINE implementations) to override the normal
10832 software implementations. For DSA and DH, parameter generation can
10833 also be overridden by providing the appropriate method callbacks.
10837 * Change the "progress" mechanism used in key-generation and
10838 primality testing to functions that take a new BN_GENCB pointer in
10839 place of callback/argument pairs. The new API functions have `_ex`
10840 postfixes and the older functions are reimplemented as wrappers for
10841 the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide
10842 declarations of the old functions to help (graceful) attempts to
10843 migrate to the new functions. Also, the new key-generation API
10844 functions operate on a caller-supplied key-structure and return
10845 success/failure rather than returning a key or NULL - this is to
10846 help make "keygen" another member function of RSA_METHOD etc.
10848 Example for using the new callback interface:
10850 int (*my_callback)(int a, int b, BN_GENCB *cb) = ...;
10851 void *my_arg = ...;
10854 BN_GENCB_set(&my_cb, my_callback, my_arg);
10856 return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb);
10857 /* For the meaning of a, b in calls to my_callback(), see the
10858 * documentation of the function that calls the callback.
10859 * cb will point to my_cb; my_arg can be retrieved as cb->arg.
10860 * my_callback should return 1 if it wants BN_is_prime_ex()
10861 * to continue, or 0 to stop.
10866 * Change the ZLIB compression method to be stateful, and make it
10867 available to TLS with the number defined in
10868 draft-ietf-tls-compression-04.txt.
10872 * Add the ASN.1 structures and functions for CertificatePair, which
10873 is defined as follows (according to X.509_4thEditionDraftV6.pdf):
10875 CertificatePair ::= SEQUENCE {
10876 forward [0] Certificate OPTIONAL,
10877 reverse [1] Certificate OPTIONAL,
10878 -- at least one of the pair shall be present -- }
10880 Also implement the PEM functions to read and write certificate
10881 pairs, and defined the PEM tag as "CERTIFICATE PAIR".
10883 This needed to be defined, mostly for the sake of the LDAP
10884 attribute crossCertificatePair, but may prove useful elsewhere as
10889 * Make it possible to inhibit symlinking of shared libraries in
10890 Makefile.shared, for Cygwin's sake.
10894 * Extend the BIGNUM API by creating a function
10895 void BN_set_negative(BIGNUM *a, int neg);
10896 and a macro that behave like
10897 int BN_is_negative(const BIGNUM *a);
10899 to avoid the need to access 'a->neg' directly in applications.
10903 * Implement fast modular reduction for pseudo-Mersenne primes
10904 used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c).
10905 EC_GROUP_new_curve_GFp() will now automatically use this
10908 *Nils Larsch <nla@trustcenter.de>*
10910 * Add new lock type (CRYPTO_LOCK_BN).
10914 * Change the ENGINE framework to automatically load engines
10915 dynamically from specific directories unless they could be
10916 found to already be built in or loaded. Move all the
10917 current engines except for the cryptodev one to a new
10918 directory engines/.
10919 The engines in engines/ are built as shared libraries if
10920 the "shared" options was given to ./Configure or ./config.
10921 Otherwise, they are inserted in libcrypto.a.
10922 /usr/local/ssl/engines is the default directory for dynamic
10923 engines, but that can be overridden at configure time through
10924 the usual use of --prefix and/or --openssldir, and at run
10925 time with the environment variable OPENSSL_ENGINES.
10927 *Geoff Thorpe and Richard Levitte*
10929 * Add Makefile.shared, a helper makefile to build shared
10930 libraries. Adapt Makefile.org.
10934 * Add version info to Win32 DLLs.
10936 *Peter 'Luna' Runestig" <peter@runestig.com>*
10938 * Add new 'medium level' PKCS#12 API. Certificates and keys
10939 can be added using this API to created arbitrary PKCS#12
10940 files while avoiding the low-level API.
10942 New options to PKCS12_create(), key or cert can be NULL and
10943 will then be omitted from the output file. The encryption
10944 algorithm NIDs can be set to -1 for no encryption, the mac
10945 iteration count can be set to 0 to omit the mac.
10947 Enhance pkcs12 utility by making the -nokeys and -nocerts
10948 options work when creating a PKCS#12 file. New option -nomac
10949 to omit the mac, NONE can be set for an encryption algorithm.
10950 New code is modified to use the enhanced PKCS12_create()
10951 instead of the low-level API.
10955 * Extend ASN1 encoder to support indefinite length constructed
10956 encoding. This can output sequences tags and octet strings in
10957 this form. Modify pk7_asn1.c to support indefinite length
10958 encoding. This is experimental and needs additional code to
10959 be useful, such as an ASN1 bio and some enhanced streaming
10962 Extend template encode functionality so that tagging is passed
10963 down to the template encoder.
10967 * Let 'openssl req' fail if an argument to '-newkey' is not
10968 recognized instead of using RSA as a default.
10972 * Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt.
10973 As these are not official, they are not included in "ALL";
10974 the "ECCdraft" ciphersuite group alias can be used to select them.
10976 *Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)*
10978 * Add ECDH engine support.
10980 *Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)*
10982 * Add ECDH in new directory crypto/ecdh/.
10984 *Douglas Stebila (Sun Microsystems Laboratories)*
10986 * Let BN_rand_range() abort with an error after 100 iterations
10987 without success (which indicates a broken PRNG).
10991 * Change BN_mod_sqrt() so that it verifies that the input value
10992 is really the square of the return value. (Previously,
10993 BN_mod_sqrt would show GIGO behaviour.)
10997 * Add named elliptic curves over binary fields from X9.62, SECG,
10998 and WAP/WTLS; add OIDs that were still missing.
11000 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
11002 * Extend the EC library for elliptic curves over binary fields
11003 (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/).
11006 EC_GF2m_simple_method
11010 EC_GROUP_new_curve_GF2m
11011 EC_GROUP_set_curve_GF2m
11012 EC_GROUP_get_curve_GF2m
11013 EC_POINT_set_affine_coordinates_GF2m
11014 EC_POINT_get_affine_coordinates_GF2m
11015 EC_POINT_set_compressed_coordinates_GF2m
11017 Point compression for binary fields is disabled by default for
11018 patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to
11021 As binary polynomials are represented as BIGNUMs, various members
11022 of the EC_GROUP and EC_POINT data structures can be shared
11023 between the implementations for prime fields and binary fields;
11024 the above `..._GF2m functions` (except for EX_GROUP_new_curve_GF2m)
11025 are essentially identical to their `..._GFp` counterparts.
11026 (For simplicity, the `..._GFp` prefix has been dropped from
11027 various internal method names.)
11029 An internal 'field_div' method (similar to 'field_mul' and
11030 'field_sqr') has been added; this is used only for binary fields.
11032 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
11034 * Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult()
11035 through methods ('mul', 'precompute_mult').
11037 The generic implementations (now internally called 'ec_wNAF_mul'
11038 and 'ec_wNAF_precomputed_mult') remain the default if these
11039 methods are undefined.
11041 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
11043 * New function EC_GROUP_get_degree, which is defined through
11044 EC_METHOD. For curves over prime fields, this returns the bit
11045 length of the modulus.
11047 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
11049 * New functions EC_GROUP_dup, EC_POINT_dup.
11050 (These simply call ..._new and ..._copy).
11052 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
11054 * Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c.
11055 Polynomials are represented as BIGNUMs (where the sign bit is not
11056 used) in the following functions [macros]:
11059 BN_GF2m_sub [= BN_GF2m_add]
11060 BN_GF2m_mod [wrapper for BN_GF2m_mod_arr]
11061 BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr]
11062 BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr]
11064 BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr]
11065 BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr]
11066 BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr]
11067 BN_GF2m_cmp [= BN_ucmp]
11069 (Note that only the 'mod' functions are actually for fields GF(2^m).
11070 BN_GF2m_add() is misnomer, but this is for the sake of consistency.)
11072 For some functions, an the irreducible polynomial defining a
11073 field can be given as an 'unsigned int[]' with strictly
11074 decreasing elements giving the indices of those bits that are set;
11075 i.e., p[] represents the polynomial
11076 f(t) = t^p[0] + t^p[1] + ... + t^p[k]
11078 p[0] > p[1] > ... > p[k] = 0.
11079 This applies to the following functions:
11082 BN_GF2m_mod_mul_arr
11083 BN_GF2m_mod_sqr_arr
11084 BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv]
11085 BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div]
11086 BN_GF2m_mod_exp_arr
11087 BN_GF2m_mod_sqrt_arr
11088 BN_GF2m_mod_solve_quad_arr
11092 Conversion can be performed by the following functions:
11097 bntest.c has additional tests for binary polynomial arithmetic.
11099 Two implementations for BN_GF2m_mod_div() are available.
11100 The default algorithm simply uses BN_GF2m_mod_inv() and
11101 BN_GF2m_mod_mul(). The alternative algorithm is compiled in only
11102 if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the
11103 copyright notice in crypto/bn/bn_gf2m.c before enabling it).
11105 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
11107 * Add new error code 'ERR_R_DISABLED' that can be used when some
11108 functionality is disabled at compile-time.
11110 *Douglas Stebila <douglas.stebila@sun.com>*
11112 * Change default behaviour of 'openssl asn1parse' so that more
11113 information is visible when viewing, e.g., a certificate:
11115 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump'
11116 mode the content of non-printable OCTET STRINGs is output in a
11117 style similar to INTEGERs, but with '[HEX DUMP]' prepended to
11118 avoid the appearance of a printable string.
11120 *Nils Larsch <nla@trustcenter.de>*
11122 * Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access
11124 EC_GROUP_set_asn1_flag()
11125 EC_GROUP_get_asn1_flag()
11126 EC_GROUP_set_point_conversion_form()
11127 EC_GROUP_get_point_conversion_form()
11128 These control ASN1 encoding details:
11129 - Curves (i.e., groups) are encoded explicitly unless asn1_flag
11130 has been set to OPENSSL_EC_NAMED_CURVE.
11131 - Points are encoded in uncompressed form by default; options for
11132 asn1_for are as for point2oct, namely
11133 POINT_CONVERSION_COMPRESSED
11134 POINT_CONVERSION_UNCOMPRESSED
11135 POINT_CONVERSION_HYBRID
11137 Also add 'seed' and 'seed_len' members to EC_GROUP with access
11139 EC_GROUP_set_seed()
11140 EC_GROUP_get0_seed()
11141 EC_GROUP_get_seed_len()
11142 This is used only for ASN1 purposes (so far).
11144 *Nils Larsch <nla@trustcenter.de>*
11146 * Add 'field_type' member to EC_METHOD, which holds the NID
11147 of the appropriate field type OID. The new function
11148 EC_METHOD_get_field_type() returns this value.
11150 *Nils Larsch <nla@trustcenter.de>*
11153 EC_POINT_point2bn()
11154 EC_POINT_bn2point()
11155 EC_POINT_point2hex()
11156 EC_POINT_hex2point()
11157 providing useful interfaces to EC_POINT_point2oct() and
11158 EC_POINT_oct2point().
11160 *Nils Larsch <nla@trustcenter.de>*
11162 * Change internals of the EC library so that the functions
11163 EC_GROUP_set_generator()
11164 EC_GROUP_get_generator()
11165 EC_GROUP_get_order()
11166 EC_GROUP_get_cofactor()
11167 are implemented directly in crypto/ec/ec_lib.c and not dispatched
11168 to methods, which would lead to unnecessary code duplication when
11169 adding different types of curves.
11171 *Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller*
11173 * Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM
11174 arithmetic, and such that modified wNAFs are generated
11175 (which avoid length expansion in many cases).
11179 * Add a function EC_GROUP_check_discriminant() (defined via
11180 EC_METHOD) that verifies that the curve discriminant is non-zero.
11182 Add a function EC_GROUP_check() that makes some sanity tests
11183 on a EC_GROUP, its generator and order. This includes
11184 EC_GROUP_check_discriminant().
11186 *Nils Larsch <nla@trustcenter.de>*
11188 * Add ECDSA in new directory crypto/ecdsa/.
11190 Add applications 'openssl ecparam' and 'openssl ecdsa'
11191 (these are based on 'openssl dsaparam' and 'openssl dsa').
11193 ECDSA support is also included in various other files across the
11194 library. Most notably,
11195 - 'openssl req' now has a '-newkey ecdsa:file' option;
11196 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA;
11197 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and
11198 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make
11199 them suitable for ECDSA where domain parameters must be
11200 extracted before the specific public key;
11201 - ECDSA engine support has been added.
11203 *Nils Larsch <nla@trustcenter.de>*
11205 * Include some named elliptic curves, and add OIDs from X9.62,
11206 SECG, and WAP/WTLS. Each curve can be obtained from the new
11208 EC_GROUP_new_by_curve_name(),
11209 and the list of available named curves can be obtained with
11210 EC_get_builtin_curves().
11211 Also add a 'curve_name' member to EC_GROUP objects, which can be
11213 EC_GROUP_set_curve_name()
11214 EC_GROUP_get_curve_name()
11216 *Nils Larsch <larsch@trustcenter.de, Bodo Moeller*
11218 * Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
11219 was actually never needed) and in BN_mul(). The removal in BN_mul()
11220 required a small change in bn_mul_part_recursive() and the addition
11221 of the functions bn_cmp_part_words(), bn_sub_part_words() and
11222 bn_add_part_words(), which do the same thing as bn_cmp_words(),
11223 bn_sub_words() and bn_add_words() except they take arrays with
11228 ### Changes between 0.9.7l and 0.9.7m [23 Feb 2007]
11230 * Cleanse PEM buffers before freeing them since they may contain
11233 *Benjamin Bennett <ben@psc.edu>*
11235 * Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
11236 a ciphersuite string such as "DEFAULT:RSA" cannot enable
11237 authentication-only ciphersuites.
11241 * Since AES128 and AES256 share a single mask bit in the logic of
11242 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
11243 kludge to work properly if AES128 is available and AES256 isn't.
11247 * Expand security boundary to match 1.1.1 module.
11251 * Remove redundant features: hash file source, editing of test vectors
11252 modify fipsld to use external fips_premain.c signature.
11256 * New perl script mkfipsscr.pl to create shell scripts or batch files to
11257 run algorithm test programs.
11261 * Make algorithm test programs more tolerant of whitespace.
11265 * Have SSL/TLS server implementation tolerate "mismatched" record
11266 protocol version while receiving ClientHello even if the
11267 ClientHello is fragmented. (The server can't insist on the
11268 particular protocol version it has chosen before the ServerHello
11269 message has informed the client about his choice.)
11273 * Load error codes if they are not already present instead of using a
11274 static variable. This allows them to be cleanly unloaded and reloaded.
11278 ### Changes between 0.9.7k and 0.9.7l [28 Sep 2006]
11280 * Introduce limits to prevent malicious keys being able to
11281 cause a denial of service. ([CVE-2006-2940])
11283 *Steve Henson, Bodo Moeller*
11285 * Fix ASN.1 parsing of certain invalid structures that can result
11286 in a denial of service. ([CVE-2006-2937]) [Steve Henson]
11288 * Fix buffer overflow in SSL_get_shared_ciphers() function.
11289 ([CVE-2006-3738]) [Tavis Ormandy and Will Drewry, Google Security Team]
11291 * Fix SSL client code which could crash if connecting to a
11292 malicious SSLv2 server. ([CVE-2006-4343])
11294 *Tavis Ormandy and Will Drewry, Google Security Team*
11296 * Change ciphersuite string processing so that an explicit
11297 ciphersuite selects this one ciphersuite (so that "AES256-SHA"
11298 will no longer include "AES128-SHA"), and any other similar
11299 ciphersuite (same bitmap) from *other* protocol versions (so that
11300 "RC4-MD5" will still include both the SSL 2.0 ciphersuite and the
11301 SSL 3.0/TLS 1.0 ciphersuite). This is a backport combining
11302 changes from 0.9.8b and 0.9.8d.
11306 ### Changes between 0.9.7j and 0.9.7k [05 Sep 2006]
11308 * Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
11309 ([CVE-2006-4339]) [Ben Laurie and Google Security Team]
11311 * Change the Unix randomness entropy gathering to use poll() when
11312 possible instead of select(), since the latter has some
11313 undesirable limitations.
11315 *Darryl Miles via Richard Levitte and Bodo Moeller*
11317 * Disable rogue ciphersuites:
11319 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
11320 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
11321 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
11323 The latter two were purportedly from
11324 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
11327 Also deactivate the remaining ciphersuites from
11328 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
11329 unofficial, and the ID has long expired.
11333 * Fix RSA blinding Heisenbug (problems sometimes occurred on
11334 dual-core machines) and other potential thread-safety issues.
11338 ### Changes between 0.9.7i and 0.9.7j [04 May 2006]
11340 * Adapt fipsld and the build system to link against the validated FIPS
11341 module in FIPS mode.
11345 * Fixes for VC++ 2005 build under Windows.
11349 * Add new Windows build target VC-32-GMAKE for VC++. This uses GNU make
11350 from a Windows bash shell such as MSYS. It is autodetected from the
11351 "config" script when run from a VC++ environment. Modify standard VC++
11352 build to use fipscanister.o from the GNU make build.
11356 ### Changes between 0.9.7h and 0.9.7i [14 Oct 2005]
11358 * Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS.
11359 The value now differs depending on if you build for FIPS or not.
11360 BEWARE! A program linked with a shared FIPSed libcrypto can't be
11361 safely run with a non-FIPSed libcrypto, as it may crash because of
11362 the difference induced by this change.
11366 ### Changes between 0.9.7g and 0.9.7h [11 Oct 2005]
11368 * Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
11369 (part of SSL_OP_ALL). This option used to disable the
11370 countermeasure against man-in-the-middle protocol-version
11371 rollback in the SSL 2.0 server implementation, which is a bad
11372 idea. ([CVE-2005-2969])
11374 *Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
11375 for Information Security, National Institute of Advanced Industrial
11376 Science and Technology [AIST, Japan)]*
11378 * Minimal support for X9.31 signatures and PSS padding modes. This is
11379 mainly for FIPS compliance and not fully integrated at this stage.
11383 * For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform
11384 the exponentiation using a fixed-length exponent. (Otherwise,
11385 the information leaked through timing could expose the secret key
11386 after many signatures; cf. Bleichenbacher's attack on DSA with
11391 * Make a new fixed-window mod_exp implementation the default for
11392 RSA, DSA, and DH private-key operations so that the sequence of
11393 squares and multiplies and the memory access pattern are
11394 independent of the particular secret key. This will mitigate
11395 cache-timing and potential related attacks.
11397 BN_mod_exp_mont_consttime() is the new exponentiation implementation,
11398 and this is automatically used by BN_mod_exp_mont() if the new flag
11399 BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH
11400 will use this BN flag for private exponents unless the flag
11401 RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or
11402 DH_FLAG_NO_EXP_CONSTTIME, respectively, is set.
11404 *Matthew D Wood (Intel Corp), with some changes by Bodo Moeller*
11406 * Change the client implementation for SSLv23_method() and
11407 SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0
11408 Client Hello message format if the SSL_OP_NO_SSLv2 option is set.
11409 (Previously, the SSL 2.0 backwards compatible Client Hello
11410 message format would be used even with SSL_OP_NO_SSLv2.)
11414 * Add support for smime-type MIME parameter in S/MIME messages which some
11419 * New function BN_MONT_CTX_set_locked() to set montgomery parameters in
11420 a threadsafe manner. Modify rsa code to use new function and add calls
11421 to dsa and dh code (which had race conditions before).
11425 * Include the fixed error library code in the C error file definitions
11426 instead of fixing them up at runtime. This keeps the error code
11427 structures constant.
11431 ### Changes between 0.9.7f and 0.9.7g [11 Apr 2005]
11433 [NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after
11436 * Fixes for newer kerberos headers. NB: the casts are needed because
11437 the 'length' field is signed on one version and unsigned on another
11438 with no (?) obvious way to tell the difference, without these VC++
11439 complains. Also the "definition" of FAR (blank) is no longer included
11440 nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up
11441 some needed definitions.
11445 * Undo Cygwin change.
11449 * Added support for proxy certificates according to RFC 3820.
11450 Because they may be a security thread to unaware applications,
11451 they must be explicitly allowed in run-time. See
11452 docs/HOWTO/proxy_certificates.txt for further information.
11456 ### Changes between 0.9.7e and 0.9.7f [22 Mar 2005]
11458 * Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating
11459 server and client random values. Previously
11460 (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in
11461 less random data when sizeof(time_t) > 4 (some 64 bit platforms).
11463 This change has negligible security impact because:
11465 1. Server and client random values still have 24 bytes of pseudo random
11468 2. Server and client random values are sent in the clear in the initial
11471 3. The master secret is derived using the premaster secret (48 bytes in
11472 size for static RSA ciphersuites) as well as client server and random
11475 The OpenSSL team would like to thank the UK NISCC for bringing this issue
11478 *Stephen Henson, reported by UK NISCC*
11480 * Use Windows randomness collection on Cygwin.
11484 * Fix hang in EGD/PRNGD query when communication socket is closed
11485 prematurely by EGD/PRNGD.
11487 *Darren Tucker <dtucker@zip.com.au> via Lutz Jänicke, resolves #1014*
11489 * Prompt for pass phrases when appropriate for PKCS12 input format.
11493 * Back-port of selected performance improvements from development
11494 branch, as well as improved support for PowerPC platforms.
11498 * Add lots of checks for memory allocation failure, error codes to indicate
11499 failure and freeing up memory if a failure occurs.
11501 *Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson*
11503 * Add new -passin argument to dgst.
11507 * Perform some character comparisons of different types in X509_NAME_cmp:
11508 this is needed for some certificates that re-encode DNs into UTF8Strings
11509 (in violation of RFC3280) and can't or won't issue name rollover
11514 * Make an explicit check during certificate validation to see that
11515 the CA setting in each certificate on the chain is correct. As a
11516 side effect always do the following basic checks on extensions,
11517 not just when there's an associated purpose to the check:
11519 - if there is an unhandled critical extension (unless the user
11520 has chosen to ignore this fault)
11521 - if the path length has been exceeded (if one is set at all)
11522 - that certain extensions fit the associated purpose (if one has
11527 ### Changes between 0.9.7d and 0.9.7e [25 Oct 2004]
11529 * Avoid a race condition when CRLs are checked in a multi threaded
11530 environment. This would happen due to the reordering of the revoked
11531 entries during signature checking and serial number lookup. Now the
11532 encoding is cached and the serial number sort performed under a lock.
11533 Add new STACK function sk_is_sorted().
11537 * Add Delta CRL to the extension code.
11541 * Various fixes to s3_pkt.c so alerts are sent properly.
11543 *David Holmes <d.holmes@f5.com>*
11545 * Reduce the chances of duplicate issuer name and serial numbers (in
11546 violation of RFC3280) using the OpenSSL certificate creation utilities.
11547 This is done by creating a random 64 bit value for the initial serial
11548 number when a serial number file is created or when a self signed
11549 certificate is created using 'openssl req -x509'. The initial serial
11550 number file is created using 'openssl x509 -next_serial' in CA.pl
11551 rather than being initialized to 1.
11555 ### Changes between 0.9.7c and 0.9.7d [17 Mar 2004]
11557 * Fix null-pointer assignment in do_change_cipher_spec() revealed
11558 by using the Codenomicon TLS Test Tool ([CVE-2004-0079])
11560 *Joe Orton, Steve Henson*
11562 * Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites
11565 *Joe Orton, Steve Henson*
11567 * Make it possible to have multiple active certificates with the same
11568 subject in the CA index file. This is done only if the keyword
11569 'unique_subject' is set to 'no' in the main CA section (default
11570 if 'CA_default') of the configuration file. The value is saved
11571 with the database itself in a separate index attribute file,
11572 named like the index file with '.attr' appended to the name.
11576 * X509 verify fixes. Disable broken certificate workarounds when
11577 X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if
11578 keyUsage extension present. Don't accept CRLs with unhandled critical
11579 extensions: since verify currently doesn't process CRL extensions this
11580 rejects a CRL with *any* critical extensions. Add new verify error codes
11585 * When creating an OCSP nonce use an OCTET STRING inside the extnValue.
11586 A clarification of RFC2560 will require the use of OCTET STRINGs and
11587 some implementations cannot handle the current raw format. Since OpenSSL
11588 copies and compares OCSP nonces as opaque blobs without any attempt at
11589 parsing them this should not create any compatibility issues.
11593 * New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when
11594 calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without
11595 this HMAC (and other) operations are several times slower than OpenSSL
11600 * Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex().
11602 *Peter Sylvester <Peter.Sylvester@EdelWeb.fr>*
11604 * Use the correct content when signing type "other".
11608 ### Changes between 0.9.7b and 0.9.7c [30 Sep 2003]
11610 * Fix various bugs revealed by running the NISCC test suite:
11612 Stop out of bounds reads in the ASN1 code when presented with
11613 invalid tags (CVE-2003-0543 and CVE-2003-0544).
11615 Free up ASN1_TYPE correctly if ANY type is invalid ([CVE-2003-0545]).
11617 If verify callback ignores invalid public key errors don't try to check
11618 certificate signature with the NULL public key.
11622 * New -ignore_err option in ocsp application to stop the server
11623 exiting on the first error in a request.
11627 * In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
11628 if the server requested one: as stated in TLS 1.0 and SSL 3.0
11633 * In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
11634 extra data after the compression methods not only for TLS 1.0
11635 but also for SSL 3.0 (as required by the specification).
11637 *Bodo Moeller; problem pointed out by Matthias Loepfe*
11639 * Change X509_certificate_type() to mark the key as exported/exportable
11640 when it's 512 *bits* long, not 512 bytes.
11644 * Change AES_cbc_encrypt() so it outputs exact multiple of
11645 blocks during encryption.
11649 * Various fixes to base64 BIO and non blocking I/O. On write
11650 flushes were not handled properly if the BIO retried. On read
11651 data was not being buffered properly and had various logic bugs.
11652 This also affects blocking I/O when the data being decoded is a
11657 * Various S/MIME bugfixes and compatibility changes:
11658 output correct application/pkcs7 MIME type if
11659 PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures.
11660 Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening
11661 of files as .eml work). Correctly handle very long lines in MIME
11666 ### Changes between 0.9.7a and 0.9.7b [10 Apr 2003]
11668 * Countermeasure against the Klima-Pokorny-Rosa extension of
11669 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
11670 a protocol version number mismatch like a decryption error
11671 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
11675 * Turn on RSA blinding by default in the default implementation
11676 to avoid a timing attack. Applications that don't want it can call
11677 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
11678 They would be ill-advised to do so in most cases.
11680 *Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller*
11682 * Change RSA blinding code so that it works when the PRNG is not
11683 seeded (in this case, the secret RSA exponent is abused as
11684 an unpredictable seed -- if it is not unpredictable, there
11685 is no point in blinding anyway). Make RSA blinding thread-safe
11686 by remembering the creator's thread ID in rsa->blinding and
11687 having all other threads use local one-time blinding factors
11688 (this requires more computation than sharing rsa->blinding, but
11689 avoids excessive locking; and if an RSA object is not shared
11690 between threads, blinding will still be very fast).
11694 * Fixed a typo bug that would cause ENGINE_set_default() to set an
11695 ENGINE as defaults for all supported algorithms irrespective of
11696 the 'flags' parameter. 'flags' is now honoured, so applications
11697 should make sure they are passing it correctly.
11701 * Target "mingw" now allows native Windows code to be generated in
11702 the Cygwin environment as well as with the MinGW compiler.
11706 ### Changes between 0.9.7 and 0.9.7a [19 Feb 2003]
11708 * In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
11709 via timing by performing a MAC computation even if incorrect
11710 block cipher padding has been found. This is a countermeasure
11711 against active attacks where the attacker has to distinguish
11712 between bad padding and a MAC verification error. ([CVE-2003-0078])
11714 *Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
11715 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
11716 Martin Vuagnoux (EPFL, Ilion)*
11718 * Make the no-err option work as intended. The intention with no-err
11719 is not to have the whole error stack handling routines removed from
11720 libcrypto, it's only intended to remove all the function name and
11721 reason texts, thereby removing some of the footprint that may not
11722 be interesting if those errors aren't displayed anyway.
11724 NOTE: it's still possible for any application or module to have its
11725 own set of error texts inserted. The routines are there, just not
11726 used by default when no-err is given.
11730 * Add support for FreeBSD on IA64.
11732 *dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454*
11734 * Adjust DES_cbc_cksum() so it returns the same value as the MIT
11735 Kerberos function mit_des_cbc_cksum(). Before this change,
11736 the value returned by DES_cbc_cksum() was like the one from
11737 mit_des_cbc_cksum(), except the bytes were swapped.
11739 *Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte*
11741 * Allow an application to disable the automatic SSL chain building.
11742 Before this a rather primitive chain build was always performed in
11743 ssl3_output_cert_chain(): an application had no way to send the
11744 correct chain if the automatic operation produced an incorrect result.
11746 Now the chain builder is disabled if either:
11748 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert().
11750 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set.
11752 The reasoning behind this is that an application would not want the
11753 auto chain building to take place if extra chain certificates are
11754 present and it might also want a means of sending no additional
11755 certificates (for example the chain has two certificates and the
11760 * Add the possibility to build without the ENGINE framework.
11762 *Steven Reddie <smr@essemer.com.au> via Richard Levitte*
11764 * Under Win32 gmtime() can return NULL: check return value in
11765 OPENSSL_gmtime(). Add error code for case where gmtime() fails.
11769 * DSA routines: under certain error conditions uninitialized BN objects
11770 could be freed. Solution: make sure initialization is performed early
11771 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>,
11772 Nils Larsch <nla@trustcenter.de> via PR#459)
11776 * Another fix for SSLv2 session ID handling: the session ID was incorrectly
11777 checked on reconnect on the client side, therefore session resumption
11778 could still fail with a "ssl session id is different" error. This
11779 behaviour is masked when SSL_OP_ALL is used due to
11780 SSL_OP_MICROSOFT_SESS_ID_BUG being set.
11781 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
11782 followup to PR #377.
11786 * IA-32 assembler support enhancements: unified ELF targets, support
11787 for SCO/Caldera platforms, fix for Cygwin shared build.
11791 * Add support for FreeBSD on sparc64. As a consequence, support for
11792 FreeBSD on non-x86 processors is separate from x86 processors on
11793 the config script, much like the NetBSD support.
11795 *Richard Levitte & Kris Kennaway <kris@obsecurity.org>*
11797 ### Changes between 0.9.6h and 0.9.7 [31 Dec 2002]
11799 [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after
11802 * Fix session ID handling in SSLv2 client code: the SERVER FINISHED
11803 code (06) was taken as the first octet of the session ID and the last
11804 octet was ignored consequently. As a result SSLv2 client side session
11805 caching could not have worked due to the session ID mismatch between
11807 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
11812 * Change the declaration of needed Kerberos libraries to use EX_LIBS
11813 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is
11818 * The hw_ncipher.c engine requires dynamic locks. Unfortunately, it
11819 seems that in spite of existing for more than a year, many application
11820 author have done nothing to provide the necessary callbacks, which
11821 means that this particular engine will not work properly anywhere.
11822 This is a very unfortunate situation which forces us, in the name
11823 of usability, to give the hw_ncipher.c a static lock, which is part
11825 NOTE: This is for the 0.9.7 series ONLY. This hack will never
11826 appear in 0.9.8 or later. We EXPECT application authors to have
11827 dealt properly with this when 0.9.8 is released (unless we actually
11828 make such changes in the libcrypto locking code that changes will
11829 have to be made anyway).
11833 * In asn1_d2i_read_bio() repeatedly call BIO_read() until all content
11834 octets have been read, EOF or an error occurs. Without this change
11835 some truncated ASN1 structures will not produce an error.
11839 * Disable Heimdal support, since it hasn't been fully implemented.
11840 Still give the possibility to force the use of Heimdal, but with
11841 warnings and a request that patches get sent to openssl-dev.
11845 * Add the VC-CE target, introduce the WINCE sysname, and add
11846 INSTALL.WCE and appropriate conditionals to make it build.
11848 *Steven Reddie <smr@essemer.com.au> via Richard Levitte*
11850 * Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and
11851 cygssl-x.y.z.dll, where x, y and z are the major, minor and
11852 edit numbers of the version.
11854 *Corinna Vinschen <vinschen@redhat.com> and Richard Levitte*
11856 * Introduce safe string copy and catenation functions
11857 (BUF_strlcpy() and BUF_strlcat()).
11859 *Ben Laurie (CHATS) and Richard Levitte*
11861 * Avoid using fixed-size buffers for one-line DNs.
11863 *Ben Laurie (CHATS)*
11865 * Add BUF_MEM_grow_clean() to avoid information leakage when
11866 resizing buffers containing secrets, and use where appropriate.
11868 *Ben Laurie (CHATS)*
11870 * Avoid using fixed size buffers for configuration file location.
11872 *Ben Laurie (CHATS)*
11874 * Avoid filename truncation for various CA files.
11876 *Ben Laurie (CHATS)*
11878 * Use sizeof in preference to magic numbers.
11880 *Ben Laurie (CHATS)*
11882 * Avoid filename truncation in cert requests.
11884 *Ben Laurie (CHATS)*
11886 * Add assertions to check for (supposedly impossible) buffer
11889 *Ben Laurie (CHATS)*
11891 * Don't cache truncated DNS entries in the local cache (this could
11892 potentially lead to a spoofing attack).
11894 *Ben Laurie (CHATS)*
11896 * Fix various buffers to be large enough for hex/decimal
11897 representations in a platform independent manner.
11899 *Ben Laurie (CHATS)*
11901 * Add CRYPTO_realloc_clean() to avoid information leakage when
11902 resizing buffers containing secrets, and use where appropriate.
11904 *Ben Laurie (CHATS)*
11906 * Add BIO_indent() to avoid much slightly worrying code to do
11909 *Ben Laurie (CHATS)*
11911 * Convert sprintf()/BIO_puts() to BIO_printf().
11913 *Ben Laurie (CHATS)*
11915 * buffer_gets() could terminate with the buffer only half
11918 *Ben Laurie (CHATS)*
11920 * Add assertions to prevent user-supplied crypto functions from
11921 overflowing internal buffers by having large block sizes, etc.
11923 *Ben Laurie (CHATS)*
11925 * New OPENSSL_assert() macro (similar to assert(), but enabled
11928 *Ben Laurie (CHATS)*
11930 * Eliminate unused copy of key in RC4.
11932 *Ben Laurie (CHATS)*
11934 * Eliminate unused and incorrectly sized buffers for IV in pem.h.
11936 *Ben Laurie (CHATS)*
11938 * Fix off-by-one error in EGD path.
11940 *Ben Laurie (CHATS)*
11942 * If RANDFILE path is too long, ignore instead of truncating.
11944 *Ben Laurie (CHATS)*
11946 * Eliminate unused and incorrectly sized X.509 structure
11949 *Ben Laurie (CHATS)*
11951 * Eliminate unused and dangerous function knumber().
11953 *Ben Laurie (CHATS)*
11955 * Eliminate unused and dangerous structure, KSSL_ERR.
11957 *Ben Laurie (CHATS)*
11959 * Protect against overlong session ID context length in an encoded
11960 session object. Since these are local, this does not appear to be
11963 *Ben Laurie (CHATS)*
11965 * Change from security patch (see 0.9.6e below) that did not affect
11966 the 0.9.6 release series:
11968 Remote buffer overflow in SSL3 protocol - an attacker could
11969 supply an oversized master key in Kerberos-enabled versions.
11972 *Ben Laurie (CHATS)*
11974 * Change the SSL kerb5 codes to match RFC 2712.
11978 * Make -nameopt work fully for req and add -reqopt switch.
11980 *Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson*
11982 * The "block size" for block ciphers in CFB and OFB mode should be 1.
11984 *Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>*
11986 * Make sure tests can be performed even if the corresponding algorithms
11987 have been removed entirely. This was also the last step to make
11988 OpenSSL compilable with DJGPP under all reasonable conditions.
11990 *Richard Levitte, Doug Kaufman <dkaufman@rahul.net>*
11992 * Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT
11993 to allow version independent disabling of normally unselected ciphers,
11994 which may be activated as a side-effect of selecting a single cipher.
11996 (E.g., cipher list string "RSA" enables ciphersuites that are left
11997 out of "ALL" because they do not provide symmetric encryption.
11998 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.)
12000 *Lutz Jaenicke, Bodo Moeller*
12002 * Add appropriate support for separate platform-dependent build
12003 directories. The recommended way to make a platform-dependent
12004 build directory is the following (tested on Linux), maybe with
12007 # Place yourself outside of the OpenSSL source tree. In
12008 # this example, the environment variable OPENSSL_SOURCE
12009 # is assumed to contain the absolute OpenSSL source directory.
12010 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`"
12011 cd objtree/"`uname -s`-`uname -r`-`uname -m`"
12012 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do
12013 mkdir -p `dirname $F`
12014 ln -s $OPENSSL_SOURCE/$F $F
12017 To be absolutely sure not to disturb the source tree, a "make clean"
12018 is a good thing. If it isn't successful, don't worry about it,
12019 it probably means the source directory is very clean.
12023 * Make sure any ENGINE control commands make local copies of string
12024 pointers passed to them whenever necessary. Otherwise it is possible
12025 the caller may have overwritten (or deallocated) the original string
12026 data when a later ENGINE operation tries to use the stored values.
12028 *Götz Babin-Ebell <babinebell@trustcenter.de>*
12030 * Improve diagnostics in file reading and command-line digests.
12032 *Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>*
12034 * Add AES modes CFB and OFB to the object database. Correct an
12035 error in AES-CFB decryption.
12039 * Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this
12040 allows existing EVP_CIPHER_CTX structures to be reused after
12041 calling `EVP_*Final()`. This behaviour is used by encryption
12042 BIOs and some applications. This has the side effect that
12043 applications must explicitly clean up cipher contexts with
12044 EVP_CIPHER_CTX_cleanup() or they will leak memory.
12048 * Check the values of dna and dnb in bn_mul_recursive before calling
12049 bn_mul_comba (a non zero value means the a or b arrays do not contain
12050 n2 elements) and fallback to bn_mul_normal if either is not zero.
12054 * Fix escaping of non-ASCII characters when using the -subj option
12055 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>)
12059 * Make object definitions compliant to LDAP (RFC2256): SN is the short
12060 form for "surname", serialNumber has no short form.
12061 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798;
12062 therefore remove "mail" short name for "internet 7".
12063 The OID for unique identifiers in X509 certificates is
12064 x500UniqueIdentifier, not uniqueIdentifier.
12065 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>)
12069 * Add an "init" command to the ENGINE config module and auto initialize
12070 ENGINEs. Without any "init" command the ENGINE will be initialized
12071 after all ctrl commands have been executed on it. If init=1 the
12072 ENGINE is initialized at that point (ctrls before that point are run
12073 on the uninitialized ENGINE and after on the initialized one). If
12074 init=0 then the ENGINE will not be initialized at all.
12078 * Fix the 'app_verify_callback' interface so that the user-defined
12079 argument is actually passed to the callback: In the
12080 SSL_CTX_set_cert_verify_callback() prototype, the callback
12081 declaration has been changed from
12084 int (*cb)(X509_STORE_CTX *,void *);
12085 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
12086 i=s->ctx->app_verify_callback(&ctx)
12087 has been changed into
12088 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
12090 To update applications using SSL_CTX_set_cert_verify_callback(),
12091 a dummy argument can be added to their callback functions.
12093 *D. K. Smetters <smetters@parc.xerox.com>*
12095 * Added the '4758cca' ENGINE to support IBM 4758 cards.
12097 *Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe*
12099 * Add and OPENSSL_LOAD_CONF define which will cause
12100 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
12101 This allows older applications to transparently support certain
12102 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
12103 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
12104 load the config file and OPENSSL_add_all_algorithms_conf() which will
12105 always load it have also been added.
12109 * Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
12110 Adjust NIDs and EVP layer.
12112 *Stephen Sprunk <stephen@sprunk.org> and Richard Levitte*
12114 * Config modules support in openssl utility.
12116 Most commands now load modules from the config file,
12117 though in a few (such as version) this isn't done
12118 because it couldn't be used for anything.
12120 In the case of ca and req the config file used is
12121 the same as the utility itself: that is the -config
12122 command line option can be used to specify an
12127 * Move default behaviour from OPENSSL_config(). If appname is NULL
12128 use "openssl_conf" if filename is NULL use default openssl config file.
12132 * Add an argument to OPENSSL_config() to allow the use of an alternative
12133 config section name. Add a new flag to tolerate a missing config file
12134 and move code to CONF_modules_load_file().
12138 * Support for crypto accelerator cards from Accelerated Encryption
12139 Processing, www.aep.ie. (Use engine 'aep')
12140 The support was copied from 0.9.6c [engine] and adapted/corrected
12141 to work with the new engine framework.
12143 *AEP Inc. and Richard Levitte*
12145 * Support for SureWare crypto accelerator cards from Baltimore
12146 Technologies. (Use engine 'sureware')
12147 The support was copied from 0.9.6c [engine] and adapted
12148 to work with the new engine framework.
12152 * Have the CHIL engine fork-safe (as defined by nCipher) and actually
12153 make the newer ENGINE framework commands for the CHIL engine work.
12155 *Toomas Kiisk <vix@cyber.ee> and Richard Levitte*
12157 * Make it possible to produce shared libraries on ReliantUNIX.
12159 *Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte*
12161 * Add the configuration target debug-linux-ppro.
12162 Make 'openssl rsa' use the general key loading routines
12163 implemented in `apps.c`, and make those routines able to
12164 handle the key format FORMAT_NETSCAPE and the variant
12167 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
12169 * Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
12171 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
12173 * Add -keyform to rsautl, and document -engine.
12175 *Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>*
12177 * Change BIO_new_file (crypto/bio/bss_file.c) to use new
12178 BIO_R_NO_SUCH_FILE error code rather than the generic
12179 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
12183 * Add new functions
12184 ERR_peek_last_error
12185 ERR_peek_last_error_line
12186 ERR_peek_last_error_line_data.
12187 These are similar to
12189 ERR_peek_error_line
12190 ERR_peek_error_line_data,
12191 but report on the latest error recorded rather than the first one
12192 still in the error queue.
12194 *Ben Laurie, Bodo Moeller*
12196 * default_algorithms option in ENGINE config module. This allows things
12198 default_algorithms = ALL
12199 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
12203 * Preliminary ENGINE config module.
12207 * New experimental application configuration code.
12211 * Change the AES code to follow the same name structure as all other
12212 symmetric ciphers, and behave the same way. Move everything to
12213 the directory crypto/aes, thereby obsoleting crypto/rijndael.
12215 *Stephen Sprunk <stephen@sprunk.org> and Richard Levitte*
12217 * SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
12219 *Ben Laurie and Theo de Raadt*
12221 * Add option to output public keys in req command.
12223 *Massimiliano Pala madwolf@openca.org*
12225 * Use wNAFs in EC_POINTs_mul() for improved efficiency
12226 (up to about 10% better than before for P-192 and P-224).
12230 * New functions/macros
12232 SSL_CTX_set_msg_callback(ctx, cb)
12233 SSL_CTX_set_msg_callback_arg(ctx, arg)
12234 SSL_set_msg_callback(ssl, cb)
12235 SSL_set_msg_callback_arg(ssl, arg)
12237 to request calling a callback function
12239 void cb(int write_p, int version, int content_type,
12240 const void *buf, size_t len, SSL *ssl, void *arg)
12242 whenever a protocol message has been completely received
12243 (write_p == 0) or sent (write_p == 1). Here 'version' is the
12244 protocol version according to which the SSL library interprets
12245 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
12246 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
12247 the content type as defined in the SSL 3.0/TLS 1.0 protocol
12248 specification (change_cipher_spec(20), alert(21), handshake(22)).
12249 'buf' and 'len' point to the actual message, 'ssl' to the
12250 SSL object, and 'arg' is the application-defined value set by
12251 SSL[_CTX]_set_msg_callback_arg().
12253 'openssl s_client' and 'openssl s_server' have new '-msg' options
12254 to enable a callback that displays all protocol messages.
12258 * Change the shared library support so shared libraries are built as
12259 soon as the corresponding static library is finished, and thereby get
12260 openssl and the test programs linked against the shared library.
12261 This still only happens when the keyword "shard" has been given to
12262 the configuration scripts.
12264 NOTE: shared library support is still an experimental thing, and
12265 backward binary compatibility is still not guaranteed.
12267 *"Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte*
12269 * Add support for Subject Information Access extension.
12271 *Peter Sylvester <Peter.Sylvester@EdelWeb.fr>*
12273 * Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
12274 additional bytes when new memory had to be allocated, not just
12275 when reusing an existing buffer.
12279 * New command line and configuration option 'utf8' for the req command.
12280 This allows field values to be specified as UTF8 strings.
12284 * Add -multi and -mr options to "openssl speed" - giving multiple parallel
12285 runs for the former and machine-readable output for the latter.
12289 * Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
12290 of the e-mail address in the DN (i.e., it will go into a certificate
12291 extension only). The new configuration file option 'email_in_dn = no'
12292 has the same effect.
12294 *Massimiliano Pala madwolf@openca.org*
12296 * Change all functions with names starting with `des_` to be starting
12297 with `DES_` instead. Add wrappers that are compatible with libdes,
12298 but are named `_ossl_old_des_*`. Finally, add macros that map the
12299 `des_*` symbols to the corresponding `_ossl_old_des_*` if libdes
12300 compatibility is desired. If OpenSSL 0.9.6c compatibility is
12301 desired, the `des_*` symbols will be mapped to `DES_*`, with one
12304 Since we provide two compatibility mappings, the user needs to
12305 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
12306 compatibility is desired. The default (i.e., when that macro
12307 isn't defined) is OpenSSL 0.9.6c compatibility.
12309 There are also macros that enable and disable the support of old
12310 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
12311 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
12312 are defined, the default will apply: to support the old des routines.
12314 In either case, one must include openssl/des.h to get the correct
12315 definitions. Do not try to just include openssl/des_old.h, that
12318 NOTE: This is a major break of an old API into a new one. Software
12319 authors are encouraged to switch to the `DES_` style functions. Some
12320 time in the future, des_old.h and the libdes compatibility functions
12321 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
12322 default), and then completely removed.
12326 * Test for certificates which contain unsupported critical extensions.
12327 If such a certificate is found during a verify operation it is
12328 rejected by default: this behaviour can be overridden by either
12329 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
12330 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
12331 X509_supported_extension() has also been added which returns 1 if a
12332 particular extension is supported.
12336 * Modify the behaviour of EVP cipher functions in similar way to digests
12337 to retain compatibility with existing code.
12341 * Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
12342 compatibility with existing code. In particular the 'ctx' parameter does
12343 not have to be to be initialized before the call to EVP_DigestInit() and
12344 it is tidied up after a call to EVP_DigestFinal(). New function
12345 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
12346 EVP_MD_CTX_copy() changed to not require the destination to be
12347 initialized valid and new function EVP_MD_CTX_copy_ex() added which
12348 requires the destination to be valid.
12350 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
12351 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
12355 * Change ssl3_get_message (ssl/s3_both.c) and the functions using it
12356 so that complete 'Handshake' protocol structures are kept in memory
12357 instead of overwriting 'msg_type' and 'length' with 'body' data.
12361 * Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
12363 *Massimo Santin via Richard Levitte*
12365 * Major restructuring to the underlying ENGINE code. This includes
12366 reduction of linker bloat, separation of pure "ENGINE" manipulation
12367 (initialisation, etc) from functionality dealing with implementations
12368 of specific crypto interfaces. This change also introduces integrated
12369 support for symmetric ciphers and digest implementations - so ENGINEs
12370 can now accelerate these by providing EVP_CIPHER and EVP_MD
12371 implementations of their own. This is detailed in
12372 [crypto/engine/README.md](crypto/engine/README.md)
12373 as it couldn't be adequately described here. However, there are a few
12374 API changes worth noting - some RSA, DSA, DH, and RAND functions that
12375 were changed in the original introduction of ENGINE code have now
12376 reverted back - the hooking from this code to ENGINE is now a good
12377 deal more passive and at run-time, operations deal directly with
12378 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
12379 dereferencing through an ENGINE pointer any more. Also, the ENGINE
12380 functions dealing with `BN_MOD_EXP[_CRT]` handlers have been removed -
12381 they were not being used by the framework as there is no concept of a
12382 BIGNUM_METHOD and they could not be generalised to the new
12383 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
12384 ENGINE_cpy() has been removed as it cannot be consistently defined in
12389 * Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
12393 * Change mkdef.pl to sort symbols that get the same entry number,
12394 and make sure the automatically generated functions `ERR_load_*`
12395 become part of libeay.num as well.
12399 * New function SSL_renegotiate_pending(). This returns true once
12400 renegotiation has been requested (either SSL_renegotiate() call
12401 or HelloRequest/ClientHello received from the peer) and becomes
12402 false once a handshake has been completed.
12403 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
12404 sends a HelloRequest, but does not ensure that a handshake takes
12405 place. SSL_renegotiate_pending() is useful for checking if the
12406 client has followed the request.)
12410 * New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
12411 By default, clients may request session resumption even during
12412 renegotiation (if session ID contexts permit); with this option,
12413 session resumption is possible only in the first handshake.
12415 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes
12416 more bits available for options that should not be part of
12417 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION).
12421 * Add some demos for certificate and certificate request creation.
12425 * Make maximum certificate chain size accepted from the peer application
12426 settable (`SSL*_get/set_max_cert_list()`), as proposed by
12427 "Douglas E. Engert" <deengert@anl.gov>.
12431 * Add support for shared libraries for Unixware-7
12432 (Boyd Lynn Gerber <gerberb@zenez.com>).
12436 * Add a "destroy" handler to ENGINEs that allows structural cleanup to
12437 be done prior to destruction. Use this to unload error strings from
12438 ENGINEs that load their own error strings. NB: This adds two new API
12439 functions to "get" and "set" this destroy handler in an ENGINE.
12443 * Alter all existing ENGINE implementations (except "openssl" and
12444 "openbsd") to dynamically instantiate their own error strings. This
12445 makes them more flexible to be built both as statically-linked ENGINEs
12446 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
12447 Also, add stub code to each that makes building them as self-contained
12448 shared-libraries easier (see [README-Engine.md](README-Engine.md)).
12452 * Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
12453 implementations into applications that are completely implemented in
12454 self-contained shared-libraries. The "dynamic" ENGINE exposes control
12455 commands that can be used to configure what shared-library to load and
12456 to control aspects of the way it is handled. Also, made an update to
12457 the [README-Engine.md](README-Engine.md) file
12458 that brings its information up-to-date and
12459 provides some information and instructions on the "dynamic" ENGINE
12460 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
12464 * Make it possible to unload ranges of ERR strings with a new
12465 "ERR_unload_strings" function.
12469 * Add a copy() function to EVP_MD.
12473 * Make EVP_MD routines take a context pointer instead of just the
12474 md_data void pointer.
12478 * Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
12479 that the digest can only process a single chunk of data
12480 (typically because it is provided by a piece of
12481 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
12482 is only going to provide a single chunk of data, and hence the
12483 framework needn't accumulate the data for oneshot drivers.
12487 * As with "ERR", make it possible to replace the underlying "ex_data"
12488 functions. This change also alters the storage and management of global
12489 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
12490 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
12491 index counters. The API functions that use this state have been changed
12492 to take a "class_index" rather than pointers to the class's local STACK
12493 and counter, and there is now an API function to dynamically create new
12494 classes. This centralisation allows us to (a) plug a lot of the
12495 thread-safety problems that existed, and (b) makes it possible to clean
12496 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
12497 such data would previously have always leaked in application code and
12498 workarounds were in place to make the memory debugging turn a blind eye
12499 to it. Application code that doesn't use this new function will still
12500 leak as before, but their memory debugging output will announce it now
12501 rather than letting it slide.
12503 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
12504 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
12505 has a return value to indicate success or failure.
12509 * Make it possible to replace the underlying "ERR" functions such that the
12510 global state (2 LHASH tables and 2 locks) is only used by the "default"
12511 implementation. This change also adds two functions to "get" and "set"
12512 the implementation prior to it being automatically set the first time
12513 any other ERR function takes place. Ie. an application can call "get",
12514 pass the return value to a module it has just loaded, and that module
12515 can call its own "set" function using that value. This means the
12516 module's "ERR" operations will use (and modify) the error state in the
12517 application and not in its own statically linked copy of OpenSSL code.
12521 * Give DH, DSA, and RSA types their own `*_up_ref()` function to increment
12522 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
12523 the operation, and provides a more encapsulated way for external code
12524 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
12525 to use these functions rather than manually incrementing the counts.
12527 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
12531 * Add EVP test program.
12535 * Add symmetric cipher support to ENGINE. Expect the API to change!
12539 * New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
12540 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
12541 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
12542 These allow a CRL to be built without having to access X509_CRL fields
12543 directly. Modify 'ca' application to use new functions.
12547 * Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
12548 bug workarounds. Rollback attack detection is a security feature.
12549 The problem will only arise on OpenSSL servers when TLSv1 is not
12550 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
12551 Software authors not wanting to support TLSv1 will have special reasons
12552 for their choice and can explicitly enable this option.
12554 *Bodo Moeller, Lutz Jaenicke*
12556 * Rationalise EVP so it can be extended: don't include a union of
12557 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX
12558 (similar to those existing for EVP_CIPHER_CTX).
12563 EVP_MD_CTX_init(&md); /* new function call */
12564 EVP_DigestInit(&md, EVP_sha1());
12565 EVP_DigestUpdate(&md, in, len);
12566 EVP_DigestFinal(&md, out, NULL);
12567 EVP_MD_CTX_cleanup(&md); /* new function call */
12571 * Make DES key schedule conform to the usual scheme, as well as
12572 correcting its structure. This means that calls to DES functions
12573 now have to pass a pointer to a des_key_schedule instead of a
12574 plain des_key_schedule (which was actually always a pointer
12577 des_key_schedule ks;
12579 des_set_key_checked(..., &ks);
12580 des_ncbc_encrypt(..., &ks, ...);
12582 (Note that a later change renames 'des_...' into 'DES_...'.)
12586 * Initial reduction of linker bloat: the use of some functions, such as
12587 PEM causes large amounts of unused functions to be linked in due to
12588 poor organisation. For example pem_all.c contains every PEM function
12589 which has a knock on effect of linking in large amounts of (unused)
12590 ASN1 code. Grouping together similar functions and splitting unrelated
12591 functions prevents this.
12595 * Cleanup of EVP macros.
12599 * Change historical references to `{NID,SN,LN}_des_ede` and ede3 to add the
12600 correct `_ecb suffix`.
12604 * Add initial OCSP responder support to ocsp application. The
12605 revocation information is handled using the text based index
12606 use by the ca application. The responder can either handle
12607 requests generated internally, supplied in files (for example
12608 via a CGI script) or using an internal minimal server.
12612 * Add configuration choices to get zlib compression for TLS.
12616 * Changes to Kerberos SSL for RFC 2712 compliance:
12617 1. Implemented real KerberosWrapper, instead of just using
12618 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
12619 2. Implemented optional authenticator field of KerberosWrapper.
12621 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
12622 and authenticator structs; see crypto/krb5/.
12624 Generalized Kerberos calls to support multiple Kerberos libraries.
12625 *Vern Staats <staatsvr@asc.hpc.mil>, Jeffrey Altman <jaltman@columbia.edu>
12626 via Richard Levitte*
12628 * Cause 'openssl speed' to use fully hard-coded DSA keys as it
12629 already does with RSA. testdsa.h now has 'priv_key/pub_key'
12630 values for each of the key sizes rather than having just
12631 parameters (and 'speed' generating keys each time).
12635 * Speed up EVP routines.
12638 pe 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
12639 s-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
12640 s-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
12641 s-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
12643 s-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
12644 s-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
12645 s-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
12648 s-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
12650 s-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
12654 * Added the OS2-EMX target.
12656 *"Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte*
12658 * Rewrite commands to use `NCONF` routines instead of the old `CONF`.
12659 New functions to support `NCONF` routines in extension code.
12660 New function `CONF_set_nconf()`
12661 to allow functions which take an `NCONF` to also handle the old `LHASH`
12662 structure: this means that the old `CONF` compatible routines can be
12663 retained (in particular w.rt. extensions) without having to duplicate the
12664 code. New function `X509V3_add_ext_nconf_sk()` to add extensions to a stack.
12668 * Enhance the general user interface with mechanisms for inner control
12669 and with possibilities to have yes/no kind of prompts.
12673 * Change all calls to low-level digest routines in the library and
12674 applications to use EVP. Add missing calls to HMAC_cleanup() and
12675 don't assume HMAC_CTX can be copied using memcpy().
12677 *Verdon Walker <VWalker@novell.com>, Steve Henson*
12679 * Add the possibility to control engines through control names but with
12680 arbitrary arguments instead of just a string.
12681 Change the key loaders to take a UI_METHOD instead of a callback
12682 function pointer. NOTE: this breaks binary compatibility with earlier
12683 versions of OpenSSL [engine].
12684 Adapt the nCipher code for these new conditions and add a card insertion
12689 * Enhance the general user interface with mechanisms to better support
12690 dialog box interfaces, application-defined prompts, the possibility
12691 to use defaults (for example default passwords from somewhere else)
12692 and interrupts/cancellations.
12696 * Tidy up PKCS#12 attribute handling. Add support for the CSP name
12697 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
12701 * Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
12702 tidy up some unnecessarily weird code in 'sk_new()').
12704 *Geoff, reported by Diego Tartara <dtartara@novamens.com>*
12706 * Change the key loading routines for ENGINEs to use the same kind
12707 callback (pem_password_cb) as all other routines that need this
12712 * Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
12713 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
12714 than this minimum value is recommended.
12718 * New random seeder for OpenVMS, using the system process statistics
12719 that are easily reachable.
12723 * Windows apparently can't transparently handle global
12724 variables defined in DLLs. Initialisations such as:
12726 const ASN1_ITEM *it = &ASN1_INTEGER_it;
12728 won't compile. This is used by the any applications that need to
12729 declare their own ASN1 modules. This was fixed by adding the option
12730 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
12731 needed for static libraries under Win32.
12735 * New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
12736 setting of purpose and trust fields. New X509_STORE trust and
12737 purpose functions and tidy up setting in other SSL functions.
12741 * Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
12742 structure. These are inherited by X509_STORE_CTX when it is
12743 initialised. This allows various defaults to be set in the
12744 X509_STORE structure (such as flags for CRL checking and custom
12745 purpose or trust settings) for functions which only use X509_STORE_CTX
12746 internally such as S/MIME.
12748 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
12749 trust settings if they are not set in X509_STORE. This allows X509_STORE
12750 purposes and trust (in S/MIME for example) to override any set by default.
12752 Add command line options for CRL checking to smime, s_client and s_server
12757 * Initial CRL based revocation checking. If the CRL checking flag(s)
12758 are set then the CRL is looked up in the X509_STORE structure and
12759 its validity and signature checked, then if the certificate is found
12760 in the CRL the verify fails with a revoked error.
12762 Various new CRL related callbacks added to X509_STORE_CTX structure.
12764 Command line options added to 'verify' application to support this.
12766 This needs some additional work, such as being able to handle multiple
12767 CRLs with different times, extension based lookup (rather than just
12768 by subject name) and ultimately more complete V2 CRL extension
12773 * Add a general user interface API (crypto/ui/). This is designed
12774 to replace things like des_read_password and friends (backward
12775 compatibility functions using this new API are provided).
12776 The purpose is to remove prompting functions from the DES code
12777 section as well as provide for prompting through dialog boxes in
12778 a window system and the like.
12782 * Add "ex_data" support to ENGINE so implementations can add state at a
12783 per-structure level rather than having to store it globally.
12787 * Make it possible for ENGINE structures to be copied when retrieved by
12788 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
12789 This causes the "original" ENGINE structure to act like a template,
12790 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
12791 operational state can be localised to each ENGINE structure, despite the
12792 fact they all share the same "methods". New ENGINE structures returned in
12793 this case have no functional references and the return value is the single
12794 structural reference. This matches the single structural reference returned
12795 by ENGINE_by_id() normally, when it is incremented on the pre-existing
12800 * Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
12801 needs to match any other type at all we need to manually clear the
12806 * Changes to the "openssl engine" utility to include;
12807 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
12808 about an ENGINE's available control commands.
12809 - executing control commands from command line arguments using the
12810 '-pre' and '-post' switches. '-post' is only used if '-t' is
12811 specified and the ENGINE is successfully initialised. The syntax for
12812 the individual commands are colon-separated, for example;
12813 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
12817 * New dynamic control command support for ENGINEs. ENGINEs can now
12818 declare their own commands (numbers), names (strings), descriptions,
12819 and input types for run-time discovery by calling applications. A
12820 subset of these commands are implicitly classed as "executable"
12821 depending on their input type, and only these can be invoked through
12822 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
12823 can be based on user input, config files, etc). The distinction is
12824 that "executable" commands cannot return anything other than a boolean
12825 result and can only support numeric or string input, whereas some
12826 discoverable commands may only be for direct use through
12827 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
12828 pointers, or other custom uses. The "executable" commands are to
12829 support parameterisations of ENGINE behaviour that can be
12830 unambiguously defined by ENGINEs and used consistently across any
12831 OpenSSL-based application. Commands have been added to all the
12832 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
12833 control over shared-library paths without source code alterations.
12837 * Changed all ENGINE implementations to dynamically allocate their
12838 ENGINEs rather than declaring them statically. Apart from this being
12839 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
12840 this also allows the implementations to compile without using the
12841 internal engine_int.h header.
12845 * Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
12846 'const' value. Any code that should be able to modify a RAND_METHOD
12847 should already have non-const pointers to it (ie. they should only
12848 modify their own ones).
12852 * Made a variety of little tweaks to the ENGINE code.
12853 - "atalla" and "ubsec" string definitions were moved from header files
12854 to C code. "nuron" string definitions were placed in variables
12855 rather than hard-coded - allowing parameterisation of these values
12856 later on via ctrl() commands.
12857 - Removed unused "#if 0"'d code.
12858 - Fixed engine list iteration code so it uses ENGINE_free() to release
12859 structural references.
12860 - Constified the RAND_METHOD element of ENGINE structures.
12861 - Constified various get/set functions as appropriate and added
12862 missing functions (including a catch-all ENGINE_cpy that duplicates
12863 all ENGINE values onto a new ENGINE except reference counts/state).
12864 - Removed NULL parameter checks in get/set functions. Setting a method
12865 or function to NULL is a way of cancelling out a previously set
12866 value. Passing a NULL ENGINE parameter is just plain stupid anyway
12867 and doesn't justify the extra error symbols and code.
12868 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
12869 flags from engine_int.h to engine.h.
12870 - Changed prototypes for ENGINE handler functions (init(), finish(),
12871 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
12875 * Implement binary inversion algorithm for BN_mod_inverse in addition
12876 to the algorithm using long division. The binary algorithm can be
12877 used only if the modulus is odd. On 32-bit systems, it is faster
12878 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
12879 roughly 5-15% for 256-bit moduli), so we use it only for moduli
12880 up to 450 bits. In 64-bit environments, the binary algorithm
12881 appears to be advantageous for much longer moduli; here we use it
12882 for moduli up to 2048 bits.
12886 * Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
12887 could not support the combine flag in choice fields.
12891 * Add a 'copy_extensions' option to the 'ca' utility. This copies
12892 extensions from a certificate request to the certificate.
12896 * Allow multiple 'certopt' and 'nameopt' options to be separated
12897 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
12898 file: this allows the display of the certificate about to be
12899 signed to be customised, to allow certain fields to be included
12900 or excluded and extension details. The old system didn't display
12901 multicharacter strings properly, omitted fields not in the policy
12902 and couldn't display additional details such as extensions.
12906 * Function EC_POINTs_mul for multiple scalar multiplication
12907 of an arbitrary number of elliptic curve points
12908 \sum scalars[i]*points[i],
12909 optionally including the generator defined for the EC_GROUP:
12910 scalar*generator + \sum scalars[i]*points[i].
12912 EC_POINT_mul is a simple wrapper function for the typical case
12913 that the point list has just one item (besides the optional
12918 * First EC_METHODs for curves over GF(p):
12920 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
12921 operations and provides various method functions that can also
12922 operate with faster implementations of modular arithmetic.
12924 EC_GFp_mont_method() reuses most functions that are part of
12925 EC_GFp_simple_method, but uses Montgomery arithmetic.
12927 *Bodo Moeller; point addition and point doubling
12928 implementation directly derived from source code provided by
12929 Lenka Fibikova <fibikova@exp-math.uni-essen.de>*
12931 * Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
12932 crypto/ec/ec_lib.c):
12934 Curves are EC_GROUP objects (with an optional group generator)
12935 based on EC_METHODs that are built into the library.
12937 Points are EC_POINT objects based on EC_GROUP objects.
12939 Most of the framework would be able to handle curves over arbitrary
12940 finite fields, but as there are no obvious types for fields other
12941 than GF(p), some functions are limited to that for now.
12945 * Add the -HTTP option to s_server. It is similar to -WWW, but requires
12946 that the file contains a complete HTTP response.
12950 * Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
12951 change the def and num file printf format specifier from "%-40sXXX"
12952 to "%-39s XXX". The latter will always guarantee a space after the
12953 field while the former will cause them to run together if the field
12954 is 40 of more characters long.
12958 * Constify the cipher and digest 'method' functions and structures
12959 and modify related functions to take constant EVP_MD and EVP_CIPHER
12964 * Hide BN_CTX structure details in bn_lcl.h instead of publishing them
12965 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
12969 * Modify `EVP_Digest*()` routines so they now return values. Although the
12970 internal software routines can never fail additional hardware versions
12975 * Clean up crypto/err/err.h and change some error codes to avoid conflicts:
12977 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
12978 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
12981 ERR_R_NESTED_ASN1_ERROR
12983 ERR_R_MISSING_ASN1_EOS
12984 were 4 .. 9, conflicting with
12985 ERR_LIB_RSA (= ERR_R_RSA_LIB)
12987 ERR_LIB_PEM (= ERR_R_PEM_LIB).
12988 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
12990 Add new error code 'ERR_R_INTERNAL_ERROR'.
12994 * Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
12999 * New option '-subj arg' for 'openssl req' and 'openssl ca'. This
13000 sets the subject name for a new request or supersedes the
13001 subject name in a given request. Formats that can be parsed are
13002 'CN=Some Name, OU=myOU, C=IT'
13004 'CN=Some Name/OU=myOU/C=IT'.
13006 Add options '-batch' and '-verbose' to 'openssl req'.
13008 *Massimiliano Pala <madwolf@hackmasters.net>*
13010 * Introduce the possibility to access global variables through
13011 functions on platform were that's the best way to handle exporting
13012 global variables in shared libraries. To enable this functionality,
13013 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
13014 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
13015 is normally done by Configure or something similar).
13017 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
13018 in the source file (foo.c) like this:
13020 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
13021 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
13023 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
13024 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
13026 OPENSSL_DECLARE_GLOBAL(int,foo);
13027 #define foo OPENSSL_GLOBAL_REF(foo)
13028 OPENSSL_DECLARE_GLOBAL(double,bar);
13029 #define bar OPENSSL_GLOBAL_REF(bar)
13031 The #defines are very important, and therefore so is including the
13032 header file everywhere where the defined globals are used.
13034 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
13035 of ASN.1 items, but that structure is a bit different.
13037 The largest change is in util/mkdef.pl which has been enhanced with
13038 better and easier to understand logic to choose which symbols should
13039 go into the Windows .def files as well as a number of fixes and code
13040 cleanup (among others, algorithm keywords are now sorted
13041 lexicographically to avoid constant rewrites).
13045 * In BN_div() keep a copy of the sign of 'num' before writing the
13046 result to 'rm' because if rm==num the value will be overwritten
13047 and produce the wrong result if 'num' is negative: this caused
13048 problems with BN_mod() and BN_nnmod().
13052 * Function OCSP_request_verify(). This checks the signature on an
13053 OCSP request and verifies the signer certificate. The signer
13054 certificate is just checked for a generic purpose and OCSP request
13059 * Add OCSP_check_validity() function to check the validity of OCSP
13060 responses. OCSP responses are prepared in real time and may only
13061 be a few seconds old. Simply checking that the current time lies
13062 between thisUpdate and nextUpdate max reject otherwise valid responses
13063 caused by either OCSP responder or client clock inaccuracy. Instead
13064 we allow thisUpdate and nextUpdate to fall within a certain period of
13065 the current time. The age of the response can also optionally be
13066 checked. Two new options -validity_period and -status_age added to
13071 * If signature or public key algorithm is unrecognized print out its
13072 OID rather that just UNKNOWN.
13076 * Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
13077 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
13078 ID to be generated from the issuer certificate alone which can then be
13079 passed to OCSP_id_issuer_cmp().
13083 * New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
13084 ASN1 modules to export functions returning ASN1_ITEM pointers
13085 instead of the ASN1_ITEM structures themselves. This adds several
13086 new macros which allow the underlying ASN1 function/structure to
13087 be accessed transparently. As a result code should not use ASN1_ITEM
13088 references directly (such as &X509_it) but instead use the relevant
13089 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
13090 use of the new ASN1 code on platforms where exporting structures
13091 is problematical (for example in shared libraries) but exporting
13092 functions returning pointers to structures is not.
13096 * Add support for overriding the generation of SSL/TLS session IDs.
13097 These callbacks can be registered either in an SSL_CTX or per SSL.
13098 The purpose of this is to allow applications to control, if they wish,
13099 the arbitrary values chosen for use as session IDs, particularly as it
13100 can be useful for session caching in multiple-server environments. A
13101 command-line switch for testing this (and any client code that wishes
13102 to use such a feature) has been added to "s_server".
13104 *Geoff Thorpe, Lutz Jaenicke*
13106 * Modify mkdef.pl to recognise and parse preprocessor conditionals
13107 of the form `#if defined(...) || defined(...) || ...` and
13108 `#if !defined(...) && !defined(...) && ...`. This also avoids
13109 the growing number of special cases it was previously handling.
13113 * Make all configuration macros available for application by making
13114 sure they are available in opensslconf.h, by giving them names starting
13115 with `OPENSSL_` to avoid conflicts with other packages and by making
13116 sure e_os2.h will cover all platform-specific cases together with
13118 Additionally, it is now possible to define configuration/platform-
13119 specific names (called "system identities"). In the C code, these
13120 are prefixed with `OPENSSL_SYSNAME_`. e_os2.h will create another
13121 macro with the name beginning with `OPENSSL_SYS_`, which is determined
13122 from `OPENSSL_SYSNAME_*` or compiler-specific macros depending on
13127 * New option -set_serial to 'req' and 'x509' this allows the serial
13128 number to use to be specified on the command line. Previously self
13129 signed certificates were hard coded with serial number 0 and the
13130 CA options of 'x509' had to use a serial number in a file which was
13135 * New options to 'ca' utility to support V2 CRL entry extensions.
13136 Currently CRL reason, invalidity date and hold instruction are
13137 supported. Add new CRL extensions to V3 code and some new objects.
13141 * New function EVP_CIPHER_CTX_set_padding() this is used to
13142 disable standard block padding (aka PKCS#5 padding) in the EVP
13143 API, which was previously mandatory. This means that the data is
13144 not padded in any way and so the total length much be a multiple
13145 of the block size, otherwise an error occurs.
13149 * Initial (incomplete) OCSP SSL support.
13153 * New function OCSP_parse_url(). This splits up a URL into its host,
13154 port and path components: primarily to parse OCSP URLs. New -url
13155 option to ocsp utility.
13159 * New nonce behavior. The return value of OCSP_check_nonce() now
13160 reflects the various checks performed. Applications can decide
13161 whether to tolerate certain situations such as an absent nonce
13162 in a response when one was present in a request: the ocsp application
13163 just prints out a warning. New function OCSP_add1_basic_nonce()
13164 this is to allow responders to include a nonce in a response even if
13165 the request is nonce-less.
13169 * Disable stdin buffering in `load_cert()` (`apps/apps.c`) so that no certs are
13170 skipped when using openssl x509 multiple times on a single input file,
13171 e.g. `(openssl x509 -out cert1; openssl x509 -out cert2) <certs`.
13175 * Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
13176 set string type: to handle setting ASN1_TIME structures. Fix ca
13177 utility to correctly initialize revocation date of CRLs.
13181 * New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
13182 the clients preferred ciphersuites and rather use its own preferences.
13183 Should help to work around M$ SGC (Server Gated Cryptography) bug in
13184 Internet Explorer by ensuring unchanged hash method during stepup.
13185 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
13189 * Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
13190 to aes and add a new 'exist' option to print out symbols that don't
13195 * Additional options to ocsp utility to allow flags to be set and
13196 additional certificates supplied.
13200 * Add the option -VAfile to 'openssl ocsp', so the user can give the
13201 OCSP client a number of certificate to only verify the response
13206 * Update Rijndael code to version 3.0 and change EVP AES ciphers to
13207 handle the new API. Currently only ECB, CBC modes supported. Add new
13210 Add TLS AES ciphersuites as described in RFC3268, "Advanced
13211 Encryption Standard (AES) Ciphersuites for Transport Layer
13212 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were
13213 not enabled by default and were not part of the "ALL" ciphersuite
13214 alias because they were not yet official; they could be
13215 explicitly requested by specifying the "AESdraft" ciphersuite
13216 group alias. In the final release of OpenSSL 0.9.7, the group
13217 alias is called "AES" and is part of "ALL".)
13219 *Ben Laurie, Steve Henson, Bodo Moeller*
13221 * New function OCSP_copy_nonce() to copy nonce value (if present) from
13222 request to response.
13226 * Functions for OCSP responders. OCSP_request_onereq_count(),
13227 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
13228 extract information from a certificate request. OCSP_response_create()
13229 creates a response and optionally adds a basic response structure.
13230 OCSP_basic_add1_status() adds a complete single response to a basic
13231 response and returns the OCSP_SINGLERESP structure just added (to allow
13232 extensions to be included for example). OCSP_basic_add1_cert() adds a
13233 certificate to a basic response and OCSP_basic_sign() signs a basic
13234 response with various flags. New helper functions ASN1_TIME_check()
13235 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
13236 (converts ASN1_TIME to GeneralizedTime).
13240 * Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
13241 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
13242 structure from a certificate. X509_pubkey_digest() digests the public_key
13243 contents: this is used in various key identifiers.
13247 * Make sk_sort() tolerate a NULL argument.
13249 *Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>*
13251 * New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
13252 passed by the function are trusted implicitly. If any of them signed the
13253 response then it is assumed to be valid and is not verified.
13257 * In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
13258 to data. This was previously part of the PKCS7 ASN1 code. This
13259 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
13260 *Steve Henson, reported by Kenneth R. Robinette
13261 <support@securenetterm.com>*
13263 * Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
13264 routines: without these tracing memory leaks is very painful.
13265 Fix leaks in PKCS12 and PKCS7 routines.
13269 * Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
13270 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
13271 effectively meant GeneralizedTime would never be used. Now it
13272 is initialised to -1 but X509_time_adj() now has to check the value
13273 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
13274 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
13275 *Steve Henson, reported by Kenneth R. Robinette
13276 <support@securenetterm.com>*
13278 * Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
13279 result in a zero length in the ASN1_INTEGER structure which was
13280 not consistent with the structure when d2i_ASN1_INTEGER() was used
13281 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
13282 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
13283 where it did not print out a minus for negative ASN1_INTEGER.
13287 * Add summary printout to ocsp utility. The various functions which
13288 convert status values to strings have been renamed to:
13289 OCSP_response_status_str(), OCSP_cert_status_str() and
13290 OCSP_crl_reason_str() and are no longer static. New options
13291 to verify nonce values and to disable verification. OCSP response
13292 printout format cleaned up.
13296 * Add additional OCSP certificate checks. These are those specified
13297 in RFC2560. This consists of two separate checks: the CA of the
13298 certificate being checked must either be the OCSP signer certificate
13299 or the issuer of the OCSP signer certificate. In the latter case the
13300 OCSP signer certificate must contain the OCSP signing extended key
13301 usage. This check is performed by attempting to match the OCSP
13302 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
13303 in the OCSP_CERTID structures of the response.
13307 * Initial OCSP certificate verification added to OCSP_basic_verify()
13308 and related routines. This uses the standard OpenSSL certificate
13309 verify routines to perform initial checks (just CA validity) and
13310 to obtain the certificate chain. Then additional checks will be
13311 performed on the chain. Currently the root CA is checked to see
13312 if it is explicitly trusted for OCSP signing. This is used to set
13313 a root CA as a global signing root: that is any certificate that
13314 chains to that CA is an acceptable OCSP signing certificate.
13318 * New '-extfile ...' option to 'openssl ca' for reading X.509v3
13319 extensions from a separate configuration file.
13320 As when reading extensions from the main configuration file,
13321 the '-extensions ...' option may be used for specifying the
13324 *Massimiliano Pala <madwolf@comune.modena.it>*
13326 * New OCSP utility. Allows OCSP requests to be generated or
13327 read. The request can be sent to a responder and the output
13328 parsed, outputted or printed in text form. Not complete yet:
13329 still needs to check the OCSP response validity.
13333 * New subcommands for 'openssl ca':
13334 `openssl ca -status <serial>` prints the status of the cert with
13335 the given serial number (according to the index file).
13336 `openssl ca -updatedb` updates the expiry status of certificates
13339 *Massimiliano Pala <madwolf@comune.modena.it>*
13341 * New '-newreq-nodes' command option to CA.pl. This is like
13342 '-newreq', but calls 'openssl req' with the '-nodes' option
13343 so that the resulting key is not encrypted.
13345 *Damien Miller <djm@mindrot.org>*
13347 * New configuration for the GNU Hurd.
13349 *Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte*
13351 * Initial code to implement OCSP basic response verify. This
13352 is currently incomplete. Currently just finds the signer's
13353 certificate and verifies the signature on the response.
13357 * New SSLeay_version code SSLEAY_DIR to determine the compiled-in
13358 value of OPENSSLDIR. This is available via the new '-d' option
13359 to 'openssl version', and is also included in 'openssl version -a'.
13363 * Allowing defining memory allocation callbacks that will be given
13364 file name and line number information in additional arguments
13365 (a `const char*` and an int). The basic functionality remains, as
13366 well as the original possibility to just replace malloc(),
13367 realloc() and free() by functions that do not know about these
13368 additional arguments. To register and find out the current
13369 settings for extended allocation functions, the following
13370 functions are provided:
13372 CRYPTO_set_mem_ex_functions
13373 CRYPTO_set_locked_mem_ex_functions
13374 CRYPTO_get_mem_ex_functions
13375 CRYPTO_get_locked_mem_ex_functions
13377 These work the same way as CRYPTO_set_mem_functions and friends.
13378 `CRYPTO_get_[locked_]mem_functions` now writes 0 where such an
13379 extended allocation function is enabled.
13380 Similarly, `CRYPTO_get_[locked_]mem_ex_functions` writes 0 where
13381 a conventional allocation function is enabled.
13383 *Richard Levitte, Bodo Moeller*
13385 * Finish off removing the remaining LHASH function pointer casts.
13386 There should no longer be any prototype-casting required when using
13387 the LHASH abstraction, and any casts that remain are "bugs". See
13388 the callback types and macros at the head of lhash.h for details
13389 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
13393 * Add automatic query of EGD sockets in RAND_poll() for the unix variant.
13394 If /dev/[u]random devices are not available or do not return enough
13395 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
13397 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
13398 /etc/entropy will be queried once each in this sequence, querying stops
13399 when enough entropy was collected without querying more sockets.
13403 * Change the Unix RAND_poll() variant to be able to poll several
13404 random devices, as specified by DEVRANDOM, until a sufficient amount
13405 of data has been collected. We spend at most 10 ms on each file
13406 (select timeout) and read in non-blocking mode. DEVRANDOM now
13407 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
13408 (previously it was just the string "/dev/urandom"), so on typical
13409 platforms the 10 ms delay will never occur.
13410 Also separate out the Unix variant to its own file, rand_unix.c.
13411 For VMS, there's a currently-empty rand_vms.c.
13415 * Move OCSP client related routines to ocsp_cl.c. These
13416 provide utility functions which an application needing
13417 to issue a request to an OCSP responder and analyse the
13418 response will typically need: as opposed to those which an
13419 OCSP responder itself would need which will be added later.
13421 OCSP_request_sign() signs an OCSP request with an API similar
13422 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
13423 response. OCSP_response_get1_basic() extracts basic response
13424 from response. OCSP_resp_find_status(): finds and extracts status
13425 information from an OCSP_CERTID structure (which will be created
13426 when the request structure is built). These are built from lower
13427 level functions which work on OCSP_SINGLERESP structures but
13428 won't normally be used unless the application wishes to examine
13429 extensions in the OCSP response for example.
13431 Replace nonce routines with a pair of functions.
13432 OCSP_request_add1_nonce() adds a nonce value and optionally
13433 generates a random value. OCSP_check_nonce() checks the
13434 validity of the nonce in an OCSP response.
13438 * Change function OCSP_request_add() to OCSP_request_add0_id().
13439 This doesn't copy the supplied OCSP_CERTID and avoids the
13440 need to free up the newly created id. Change return type
13441 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
13442 This can then be used to add extensions to the request.
13443 Deleted OCSP_request_new(), since most of its functionality
13444 is now in OCSP_REQUEST_new() (and the case insensitive name
13445 clash) apart from the ability to set the request name which
13446 will be added elsewhere.
13450 * Update OCSP API. Remove obsolete extensions argument from
13451 various functions. Extensions are now handled using the new
13452 OCSP extension code. New simple OCSP HTTP function which
13453 can be used to send requests and parse the response.
13457 * Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
13458 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
13459 uses the special reorder version of SET OF to sort the attributes
13460 and reorder them to match the encoded order. This resolves a long
13461 standing problem: a verify on a PKCS7 structure just after signing
13462 it used to fail because the attribute order did not match the
13463 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
13464 it uses the received order. This is necessary to tolerate some broken
13465 software that does not order SET OF. This is handled by encoding
13466 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
13467 to produce the required SET OF.
13471 * Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
13472 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
13473 files to get correct declarations of the ASN.1 item variables.
13477 * Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
13478 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
13479 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
13480 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
13481 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
13482 ASN1_ITEM and no wrapper functions.
13486 * New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
13487 replace the old function pointer based I/O routines. Change most of
13488 the `*_d2i_bio()` and `*_d2i_fp()` functions to use these.
13492 * Enhance mkdef.pl to be more accepting about spacing in C preprocessor
13493 lines, recognize more "algorithms" that can be deselected, and make
13494 it complain about algorithm deselection that isn't recognised.
13498 * New ASN1 functions to handle dup, sign, verify, digest, pack and
13499 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
13500 to use new functions. Add NO_ASN1_OLD which can be set to remove
13501 some old style ASN1 functions: this can be used to determine if old
13502 code will still work when these eventually go away.
13506 * New extension functions for OCSP structures, these follow the
13507 same conventions as certificates and CRLs.
13511 * New function X509V3_add1_i2d(). This automatically encodes and
13512 adds an extension. Its behaviour can be customised with various
13513 flags to append, replace or delete. Various wrappers added for
13514 certificates and CRLs.
13518 * Fix to avoid calling the underlying ASN1 print routine when
13519 an extension cannot be parsed. Correct a typo in the
13520 OCSP_SERVICELOC extension. Tidy up print OCSP format.
13524 * Make mkdef.pl parse some of the ASN1 macros and add appropriate
13525 entries for variables.
13529 * Add functionality to `apps/openssl.c` for detecting locking
13530 problems: As the program is single-threaded, all we have
13531 to do is register a locking callback using an array for
13532 storing which locks are currently held by the program.
13536 * Use a lock around the call to CRYPTO_get_ex_new_index() in
13537 SSL_get_ex_data_X509_STORE_idx(), which is used in
13538 ssl_verify_cert_chain() and thus can be called at any time
13539 during TLS/SSL handshakes so that thread-safety is essential.
13540 Unfortunately, the ex_data design is not at all suited
13541 for multi-threaded use, so it probably should be abolished.
13545 * Added Broadcom "ubsec" ENGINE to OpenSSL.
13547 *Broadcom, tweaked and integrated by Geoff Thorpe*
13549 * Move common extension printing code to new function
13550 X509V3_print_extensions(). Reorganise OCSP print routines and
13551 implement some needed OCSP ASN1 functions. Add OCSP extensions.
13555 * New function X509_signature_print() to remove duplication in some
13560 * Add a special meaning when SET OF and SEQUENCE OF flags are both
13561 set (this was treated exactly the same as SET OF previously). This
13562 is used to reorder the STACK representing the structure to match the
13563 encoding. This will be used to get round a problem where a PKCS7
13564 structure which was signed could not be verified because the STACK
13565 order did not reflect the encoded order.
13569 * Reimplement the OCSP ASN1 module using the new code.
13573 * Update the X509V3 code to permit the use of an ASN1_ITEM structure
13574 for its ASN1 operations. The old style function pointers still exist
13575 for now but they will eventually go away.
13579 * Merge in replacement ASN1 code from the ASN1 branch. This almost
13580 completely replaces the old ASN1 functionality with a table driven
13581 encoder and decoder which interprets an ASN1_ITEM structure describing
13582 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
13583 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
13584 has also been converted to the new form.
13588 * Change BN_mod_exp_recp so that negative moduli are tolerated
13589 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
13590 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
13591 for negative moduli.
13595 * Fix BN_uadd and BN_usub: Always return non-negative results instead
13596 of not touching the result's sign bit.
13600 * BN_div bugfix: If the result is 0, the sign (res->neg) must not be
13605 * Changed the LHASH code to use prototypes for callbacks, and created
13606 macros to declare and implement thin (optionally static) functions
13607 that provide type-safety and avoid function pointer casting for the
13608 type-specific callbacks.
13612 * Added Kerberos Cipher Suites to be used with TLS, as written in
13614 *Veers Staats <staatsvr@asc.hpc.mil>,
13615 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte*
13617 * Reformat the FAQ so the different questions and answers can be divided
13618 in sections depending on the subject.
13622 * Have the zlib compression code load ZLIB.DLL dynamically under
13627 * New function BN_mod_sqrt for computing square roots modulo a prime
13628 (using the probabilistic Tonelli-Shanks algorithm unless
13629 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
13630 be handled deterministically).
13632 *Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller*
13634 * Make BN_mod_inverse faster by explicitly handling small quotients
13635 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
13636 512 bits], about 30% for larger ones [1024 or 2048 bits].)
13640 * New function BN_kronecker.
13644 * Fix BN_gcd so that it works on negative inputs; the result is
13645 positive unless both parameters are zero.
13646 Previously something reasonably close to an infinite loop was
13647 possible because numbers could be growing instead of shrinking
13648 in the implementation of Euclid's algorithm.
13652 * Fix BN_is_word() and BN_is_one() macros to take into account the
13653 sign of the number in question.
13655 Fix BN_is_word(a,w) to work correctly for w == 0.
13657 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
13658 because its test if the absolute value of 'a' equals 'w'.
13659 Note that BN_abs_is_word does *not* handle w == 0 reliably;
13660 it exists mostly for use in the implementations of BN_is_zero(),
13661 BN_is_one(), and BN_is_word().
13665 * New function BN_swap.
13669 * Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
13670 the exponentiation functions are more likely to produce reasonable
13671 results on negative inputs.
13675 * Change BN_mod_mul so that the result is always non-negative.
13676 Previously, it could be negative if one of the factors was negative;
13677 I don't think anyone really wanted that behaviour.
13681 * Move `BN_mod_...` functions into new file `crypto/bn/bn_mod.c`
13682 (except for exponentiation, which stays in `crypto/bn/bn_exp.c`,
13683 and `BN_mod_mul_reciprocal`, which stays in `crypto/bn/bn_recp.c`)
13684 and add new functions:
13693 BN_mod_lshift1_quick
13695 BN_mod_lshift_quick
13697 These functions always generate non-negative results.
13699 `BN_nnmod` otherwise is `like BN_mod` (if `BN_mod` computes a remainder `r`
13700 such that `|m| < r < 0`, `BN_nnmod` will output `rem + |m|` instead).
13702 `BN_mod_XXX_quick(r, a, [b,] m)` generates the same result as
13703 `BN_mod_XXX(r, a, [b,] m, ctx)`, but requires that `a` [and `b`]
13704 be reduced modulo `m`.
13706 *Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller*
13709 The following entry accidentally appeared in the CHANGES file
13710 distributed with OpenSSL 0.9.7. The modifications described in
13711 it do *not* apply to OpenSSL 0.9.7.
13713 * Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
13714 was actually never needed) and in BN_mul(). The removal in BN_mul()
13715 required a small change in bn_mul_part_recursive() and the addition
13716 of the functions bn_cmp_part_words(), bn_sub_part_words() and
13717 bn_add_part_words(), which do the same thing as bn_cmp_words(),
13718 bn_sub_words() and bn_add_words() except they take arrays with
13724 * In 'openssl passwd', verify passwords read from the terminal
13725 unless the '-salt' option is used (which usually means that
13726 verification would just waste user's time since the resulting
13727 hash is going to be compared with some given password hash)
13728 or the new '-noverify' option is used.
13730 This is an incompatible change, but it does not affect
13731 non-interactive use of 'openssl passwd' (passwords on the command
13732 line, '-stdin' option, '-in ...' option) and thus should not
13733 cause any problems.
13737 * Remove all references to RSAref, since there's no more need for it.
13741 * Make DSO load along a path given through an environment variable
13742 (SHLIB_PATH) with shl_load().
13746 * Constify the ENGINE code as a result of BIGNUM constification.
13747 Also constify the RSA code and most things related to it. In a
13748 few places, most notable in the depth of the ASN.1 code, ugly
13749 casts back to non-const were required (to be solved at a later
13754 * Make it so the openssl application has all engines loaded by default.
13758 * Constify the BIGNUM routines a little more.
13762 * Add the following functions:
13764 ENGINE_load_cswift()
13766 ENGINE_load_atalla()
13767 ENGINE_load_nuron()
13768 ENGINE_load_builtin_engines()
13770 That way, an application can itself choose if external engines that
13771 are built-in in OpenSSL shall ever be used or not. The benefit is
13772 that applications won't have to be linked with libdl or other dso
13773 libraries unless it's really needed.
13775 Changed 'openssl engine' to load all engines on demand.
13776 Changed the engine header files to avoid the duplication of some
13777 declarations (they differed!).
13781 * 'openssl engine' can now list capabilities.
13785 * Better error reporting in 'openssl engine'.
13789 * Never call load_dh_param(NULL) in s_server.
13793 * Add engine application. It can currently list engines by name and
13794 identity, and test if they are actually available.
13798 * Improve RPM specification file by forcing symbolic linking and making
13799 sure the installed documentation is also owned by root.root.
13801 *Damien Miller <djm@mindrot.org>*
13803 * Give the OpenSSL applications more possibilities to make use of
13804 keys (public as well as private) handled by engines.
13808 * Add OCSP code that comes from CertCo.
13812 * Add VMS support for the Rijndael code.
13816 * Added untested support for Nuron crypto accelerator.
13820 * Add support for external cryptographic devices. This code was
13821 previously distributed separately as the "engine" branch.
13823 *Geoff Thorpe, Richard Levitte*
13825 * Rework the filename-translation in the DSO code. It is now possible to
13826 have far greater control over how a "name" is turned into a filename
13827 depending on the operating environment and any oddities about the
13828 different shared library filenames on each system.
13832 * Support threads on FreeBSD-elf in Configure.
13836 * Fix for SHA1 assembly problem with MASM: it produces
13837 warnings about corrupt line number information when assembling
13838 with debugging information. This is caused by the overlapping
13841 *Bernd Matthes <mainbug@celocom.de>, Steve Henson*
13844 NCONF_get_number() has no error checking at all. As a replacement,
13845 NCONF_get_number_e() is defined (`_e` for "error checking") and is
13846 promoted strongly. The old NCONF_get_number is kept around for
13847 binary backward compatibility.
13848 Make it possible for methods to load from something other than a BIO,
13849 by providing a function pointer that is given a name instead of a BIO.
13850 For example, this could be used to load configuration data from an
13855 * Fix for non blocking accept BIOs. Added new I/O special reason
13856 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
13857 with non blocking I/O was not possible because no retry code was
13858 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
13863 * Added the beginnings of Rijndael support.
13867 * Fix for bug in DirectoryString mask setting. Add support for
13868 X509_NAME_print_ex() in 'req' and X509_print_ex() function
13869 to allow certificate printing to more controllable, additional
13870 'certopt' option to 'x509' to allow new printing options to be
13875 * Clean old EAY MD5 hack from e_os.h.
13879 ### Changes between 0.9.6l and 0.9.6m [17 Mar 2004]
13881 * Fix null-pointer assignment in do_change_cipher_spec() revealed
13882 by using the Codenomicon TLS Test Tool ([CVE-2004-0079])
13884 *Joe Orton, Steve Henson*
13886 ### Changes between 0.9.6k and 0.9.6l [04 Nov 2003]
13888 * Fix additional bug revealed by the NISCC test suite:
13890 Stop bug triggering large recursion when presented with
13891 certain ASN.1 tags ([CVE-2003-0851])
13895 ### Changes between 0.9.6j and 0.9.6k [30 Sep 2003]
13897 * Fix various bugs revealed by running the NISCC test suite:
13899 Stop out of bounds reads in the ASN1 code when presented with
13900 invalid tags (CVE-2003-0543 and CVE-2003-0544).
13902 If verify callback ignores invalid public key errors don't try to check
13903 certificate signature with the NULL public key.
13907 * In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
13908 if the server requested one: as stated in TLS 1.0 and SSL 3.0
13913 * In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
13914 extra data after the compression methods not only for TLS 1.0
13915 but also for SSL 3.0 (as required by the specification).
13917 *Bodo Moeller; problem pointed out by Matthias Loepfe*
13919 * Change X509_certificate_type() to mark the key as exported/exportable
13920 when it's 512 *bits* long, not 512 bytes.
13924 ### Changes between 0.9.6i and 0.9.6j [10 Apr 2003]
13926 * Countermeasure against the Klima-Pokorny-Rosa extension of
13927 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
13928 a protocol version number mismatch like a decryption error
13929 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
13933 * Turn on RSA blinding by default in the default implementation
13934 to avoid a timing attack. Applications that don't want it can call
13935 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
13936 They would be ill-advised to do so in most cases.
13938 *Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller*
13940 * Change RSA blinding code so that it works when the PRNG is not
13941 seeded (in this case, the secret RSA exponent is abused as
13942 an unpredictable seed -- if it is not unpredictable, there
13943 is no point in blinding anyway). Make RSA blinding thread-safe
13944 by remembering the creator's thread ID in rsa->blinding and
13945 having all other threads use local one-time blinding factors
13946 (this requires more computation than sharing rsa->blinding, but
13947 avoids excessive locking; and if an RSA object is not shared
13948 between threads, blinding will still be very fast).
13952 ### Changes between 0.9.6h and 0.9.6i [19 Feb 2003]
13954 * In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
13955 via timing by performing a MAC computation even if incorrect
13956 block cipher padding has been found. This is a countermeasure
13957 against active attacks where the attacker has to distinguish
13958 between bad padding and a MAC verification error. ([CVE-2003-0078])
13960 *Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
13961 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
13962 Martin Vuagnoux (EPFL, Ilion)*
13964 ### Changes between 0.9.6g and 0.9.6h [5 Dec 2002]
13966 * New function OPENSSL_cleanse(), which is used to cleanse a section of
13967 memory from its contents. This is done with a counter that will
13968 place alternating values in each byte. This can be used to solve
13969 two issues: 1) the removal of calls to memset() by highly optimizing
13970 compilers, and 2) cleansing with other values than 0, since those can
13971 be read through on certain media, for example a swap space on disk.
13975 * Bugfix: client side session caching did not work with external caching,
13976 because the session->cipher setting was not restored when reloading
13977 from the external cache. This problem was masked, when
13978 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set.
13979 (Found by Steve Haslam <steve@araqnid.ddts.net>.)
13983 * Fix client_certificate (ssl/s2_clnt.c): The permissible total
13984 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33.
13986 *Zeev Lieber <zeev-l@yahoo.com>*
13988 * Undo an undocumented change introduced in 0.9.6e which caused
13989 repeated calls to OpenSSL_add_all_ciphers() and
13990 OpenSSL_add_all_digests() to be ignored, even after calling
13995 * Change the default configuration reader to deal with last line not
13996 being properly terminated.
14000 * Change X509_NAME_cmp() so it applies the special rules on handling
14001 DN values that are of type PrintableString, as well as RDNs of type
14002 emailAddress where the value has the type ia5String.
14004 *stefank@valicert.com via Richard Levitte*
14006 * Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half
14007 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently
14008 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be
14009 the bitwise-OR of the two for use by the majority of applications
14010 wanting this behaviour, and update the docs. The documented
14011 behaviour and actual behaviour were inconsistent and had been
14012 changing anyway, so this is more a bug-fix than a behavioural
14015 *Geoff Thorpe, diagnosed by Nadav Har'El*
14017 * Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c
14018 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes).
14022 * Fix initialization code race conditions in
14023 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(),
14024 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(),
14025 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(),
14026 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(),
14027 ssl2_get_cipher_by_char(),
14028 ssl3_get_cipher_by_char().
14030 *Patrick McCormick <patrick@tellme.com>, Bodo Moeller*
14032 * Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after
14033 the cached sessions are flushed, as the remove_cb() might use ex_data
14034 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com>
14035 (see [openssl.org #212]).
14037 *Geoff Thorpe, Lutz Jaenicke*
14039 * Fix typo in OBJ_txt2obj which incorrectly passed the content
14040 length, instead of the encoding length to d2i_ASN1_OBJECT.
14044 ### Changes between 0.9.6f and 0.9.6g [9 Aug 2002]
14046 * [In 0.9.6g-engine release:]
14047 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use `_stdcall`).
14049 *Lynn Gazis <lgazis@rainbow.com>*
14051 ### Changes between 0.9.6e and 0.9.6f [8 Aug 2002]
14053 * Fix ASN1 checks. Check for overflow by comparing with LONG_MAX
14054 and get fix the header length calculation.
14055 *Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>,
14056 Alon Kantor <alonk@checkpoint.com> (and others), Steve Henson*
14058 * Use proper error handling instead of 'assertions' in buffer
14059 overflow checks added in 0.9.6e. This prevents DoS (the
14060 assertions could call abort()).
14062 *Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller*
14064 ### Changes between 0.9.6d and 0.9.6e [30 Jul 2002]
14066 * Add various sanity checks to asn1_get_length() to reject
14067 the ASN1 length bytes if they exceed sizeof(long), will appear
14068 negative or the content length exceeds the length of the
14071 *Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>*
14073 * Fix cipher selection routines: ciphers without encryption had no flags
14074 for the cipher strength set and where therefore not handled correctly
14075 by the selection routines (PR #130).
14079 * Fix EVP_dsa_sha macro.
14084 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS
14085 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure
14086 that was added in OpenSSL 0.9.6d.
14088 As the countermeasure turned out to be incompatible with some
14089 broken SSL implementations, the new option is part of SSL_OP_ALL.
14090 SSL_OP_ALL is usually employed when compatibility with weird SSL
14091 implementations is desired (e.g. '-bugs' option to 's_client' and
14092 's_server'), so the new option is automatically set in many
14097 * Changes in security patch:
14099 Changes marked "(CHATS)" were sponsored by the Defense Advanced
14100 Research Projects Agency (DARPA) and Air Force Research Laboratory,
14101 Air Force Materiel Command, USAF, under agreement number
14104 * Add various sanity checks to asn1_get_length() to reject
14105 the ASN1 length bytes if they exceed sizeof(long), will appear
14106 negative or the content length exceeds the length of the
14107 supplied buffer. ([CVE-2002-0659])
14109 *Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>*
14111 * Assertions for various potential buffer overflows, not known to
14112 happen in practice.
14114 *Ben Laurie (CHATS)*
14116 * Various temporary buffers to hold ASCII versions of integers were
14117 too small for 64 bit platforms. ([CVE-2002-0655])
14118 *Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>*
14120 * Remote buffer overflow in SSL3 protocol - an attacker could
14121 supply an oversized session ID to a client. ([CVE-2002-0656])
14123 *Ben Laurie (CHATS)*
14125 * Remote buffer overflow in SSL2 protocol - an attacker could
14126 supply an oversized client master key. ([CVE-2002-0656])
14128 *Ben Laurie (CHATS)*
14130 ### Changes between 0.9.6c and 0.9.6d [9 May 2002]
14132 * Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
14133 encoded as NULL) with id-dsa-with-sha1.
14135 *Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller*
14137 * Check various `X509_...()` return values in `apps/req.c`.
14139 *Nils Larsch <nla@trustcenter.de>*
14141 * Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
14142 an end-of-file condition would erroneously be flagged, when the CRLF
14143 was just at the end of a processed block. The bug was discovered when
14144 processing data through a buffering memory BIO handing the data to a
14145 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
14146 <ptsekov@syntrex.com> and Nedelcho Stanev.
14150 * Implement a countermeasure against a vulnerability recently found
14151 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
14152 before application data chunks to avoid the use of known IVs
14153 with data potentially chosen by the attacker.
14157 * Fix length checks in ssl3_get_client_hello().
14161 * TLS/SSL library bugfix: use s->s3->in_read_app_data differently
14162 to prevent ssl3_read_internal() from incorrectly assuming that
14163 ssl3_read_bytes() found application data while handshake
14164 processing was enabled when in fact s->s3->in_read_app_data was
14165 merely automatically cleared during the initial handshake.
14167 *Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>*
14169 * Fix object definitions for Private and Enterprise: they were not
14170 recognized in their shortname (=lowercase) representation. Extend
14171 obj_dat.pl to issue an error when using undefined keywords instead
14172 of silently ignoring the problem (Svenning Sorensen
14173 <sss@sss.dnsalias.net>).
14177 * Fix DH_generate_parameters() so that it works for 'non-standard'
14178 generators, i.e. generators other than 2 and 5. (Previously, the
14179 code did not properly initialise the 'add' and 'rem' values to
14180 BN_generate_prime().)
14182 In the new general case, we do not insist that 'generator' is
14183 actually a primitive root: This requirement is rather pointless;
14184 a generator of the order-q subgroup is just as good, if not
14189 * Map new X509 verification errors to alerts. Discovered and submitted by
14190 Tom Wu <tom@arcot.com>.
14194 * Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
14195 returning non-zero before the data has been completely received
14196 when using non-blocking I/O.
14198 *Bodo Moeller; problem pointed out by John Hughes*
14200 * Some of the ciphers missed the strength entry (SSL_LOW etc).
14202 *Ben Laurie, Lutz Jaenicke*
14204 * Fix bug in SSL_clear(): bad sessions were not removed (found by
14205 Yoram Zahavi <YoramZ@gilian.com>).
14209 * Add information about CygWin 1.3 and on, and preserve proper
14210 configuration for the versions before that.
14212 *Corinna Vinschen <vinschen@redhat.com> and Richard Levitte*
14214 * Make removal from session cache (SSL_CTX_remove_session()) more robust:
14215 check whether we deal with a copy of a session and do not delete from
14216 the cache in this case. Problem reported by "Izhar Shoshani Levi"
14217 <izhar@checkpoint.com>.
14221 * Do not store session data into the internal session cache, if it
14222 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
14223 flag is set). Proposed by Aslam <aslam@funk.com>.
14227 * Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
14232 * [In 0.9.6d-engine release:]
14233 Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
14235 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
14237 * Add the configuration target linux-s390x.
14239 *Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte*
14241 * The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
14242 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
14243 variable as an indication that a ClientHello message has been
14244 received. As the flag value will be lost between multiple
14245 invocations of ssl3_accept when using non-blocking I/O, the
14246 function may not be aware that a handshake has actually taken
14247 place, thus preventing a new session from being added to the
14250 To avoid this problem, we now set s->new_session to 2 instead of
14251 using a local variable.
14253 *Lutz Jaenicke, Bodo Moeller*
14255 * Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
14256 if the SSL_R_LENGTH_MISMATCH error is detected.
14258 *Geoff Thorpe, Bodo Moeller*
14260 * New 'shared_ldflag' column in Configure platform table.
14264 * Fix EVP_CIPHER_mode macro.
14266 *"Dan S. Camper" <dan@bti.net>*
14268 * Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
14269 type, we must throw them away by setting rr->length to 0.
14271 *D P Chang <dpc@qualys.com>*
14273 ### Changes between 0.9.6b and 0.9.6c [21 dec 2001]
14275 * Fix BN_rand_range bug pointed out by Dominikus Scherkl
14276 <Dominikus.Scherkl@biodata.com>. (The previous implementation
14277 worked incorrectly for those cases where range = `10..._2` and
14278 `3*range` is two bits longer than range.)
14282 * Only add signing time to PKCS7 structures if it is not already
14287 * Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
14288 OBJ_ld_ce should be OBJ_id_ce.
14289 Also some ip-pda OIDs in crypto/objects/objects.txt were
14290 incorrect (cf. RFC 3039).
14292 *Matt Cooper, Frederic Giudicelli, Bodo Moeller*
14294 * Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
14295 returns early because it has nothing to do.
14297 *Andy Schneider <andy.schneider@bjss.co.uk>*
14299 * [In 0.9.6c-engine release:]
14300 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
14302 *Andy Schneider <andy.schneider@bjss.co.uk>*
14304 * [In 0.9.6c-engine release:]
14305 Add support for Cryptographic Appliance's keyserver technology.
14306 (Use engine 'keyclient')
14308 *Cryptographic Appliances and Geoff Thorpe*
14310 * Add a configuration entry for OS/390 Unix. The C compiler 'c89'
14311 is called via tools/c89.sh because arguments have to be
14312 rearranged (all '-L' options must appear before the first object
14315 *Richard Shapiro <rshapiro@abinitio.com>*
14317 * [In 0.9.6c-engine release:]
14318 Add support for Broadcom crypto accelerator cards, backported
14321 *Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox*
14323 * [In 0.9.6c-engine release:]
14324 Add support for SureWare crypto accelerator cards from
14325 Baltimore Technologies. (Use engine 'sureware')
14327 *Baltimore Technologies and Mark Cox*
14329 * [In 0.9.6c-engine release:]
14330 Add support for crypto accelerator cards from Accelerated
14331 Encryption Processing, www.aep.ie. (Use engine 'aep')
14333 *AEP Inc. and Mark Cox*
14335 * Add a configuration entry for gcc on UnixWare.
14337 *Gary Benson <gbenson@redhat.com>*
14339 * Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
14340 messages are stored in a single piece (fixed-length part and
14341 variable-length part combined) and fix various bugs found on the way.
14345 * Disable caching in BIO_gethostbyname(), directly use gethostbyname()
14346 instead. BIO_gethostbyname() does not know what timeouts are
14347 appropriate, so entries would stay in cache even when they have
14349 *Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>*
14351 * Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
14352 faced with a pathologically small ClientHello fragment that does
14353 not contain client_version: Instead of aborting with an error,
14354 simply choose the highest available protocol version (i.e.,
14355 TLS 1.0 unless it is disabled). In practice, ClientHello
14356 messages are never sent like this, but this change gives us
14357 strictly correct behaviour at least for TLS.
14361 * Fix SSL handshake functions and SSL_clear() such that SSL_clear()
14362 never resets s->method to s->ctx->method when called from within
14363 one of the SSL handshake functions.
14365 *Bodo Moeller; problem pointed out by Niko Baric*
14367 * In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
14368 (sent using the client's version number) if client_version is
14369 smaller than the protocol version in use. Also change
14370 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
14371 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
14372 the client will at least see that alert.
14376 * Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
14381 * Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
14382 client receives HelloRequest while in a handshake.
14384 *Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>*
14386 * Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
14387 should end in 'break', not 'goto end' which circumvents various
14388 cleanups done in state SSL_ST_OK. But session related stuff
14389 must be disabled for SSL_ST_OK in the case that we just sent a
14392 Also avoid some overhead by not calling ssl_init_wbio_buffer()
14393 before just sending a HelloRequest.
14395 *Bodo Moeller, Eric Rescorla <ekr@rtfm.com>*
14397 * Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
14398 reveal whether illegal block cipher padding was found or a MAC
14399 verification error occurred. (Neither SSLerr() codes nor alerts
14400 are directly visible to potential attackers, but the information
14401 may leak via logfiles.)
14403 Similar changes are not required for the SSL 2.0 implementation
14404 because the number of padding bytes is sent in clear for SSL 2.0,
14405 and the extra bytes are just ignored. However ssl/s2_pkt.c
14406 failed to verify that the purported number of padding bytes is in
14411 * Add OpenUNIX-8 support including shared libraries
14412 (Boyd Lynn Gerber <gerberb@zenez.com>).
14416 * Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
14417 'wristwatch attack' using huge encoding parameters (cf.
14418 James H. Manger's CRYPTO 2001 paper). Note that the
14419 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
14420 encoding parameters and hence was not vulnerable.
14424 * BN_sqr() bug fix.
14426 *Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>*
14428 * Rabin-Miller test analyses assume uniformly distributed witnesses,
14429 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
14430 followed by modular reduction.
14432 *Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>*
14434 * Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
14435 equivalent based on BN_pseudo_rand() instead of BN_rand().
14439 * s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
14440 This function was broken, as the check for a new client hello message
14441 to handle SGC did not allow these large messages.
14442 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
14446 * Add alert descriptions for TLSv1 to `SSL_alert_desc_string[_long]()`.
14450 * Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
14451 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
14455 * Rework the configuration and shared library support for Tru64 Unix.
14456 The configuration part makes use of modern compiler features and
14457 still retains old compiler behavior for those that run older versions
14458 of the OS. The shared library support part includes a variant that
14459 uses the RPATH feature, and is available through the special
14460 configuration target "alpha-cc-rpath", which will never be selected
14463 *Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte*
14465 * In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
14466 with the same message size as in ssl3_get_certificate_request().
14467 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
14468 messages might inadvertently be reject as too long.
14470 *Petr Lampa <lampa@fee.vutbr.cz>*
14472 * Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
14476 * Modified SSL library such that the verify_callback that has been set
14477 specifically for an SSL object with SSL_set_verify() is actually being
14478 used. Before the change, a verify_callback set with this function was
14479 ignored and the verify_callback() set in the SSL_CTX at the time of
14480 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
14481 to allow the necessary settings.
14485 * Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
14486 explicitly to NULL, as at least on Solaris 8 this seems not always to be
14487 done automatically (in contradiction to the requirements of the C
14488 standard). This made problems when used from OpenSSH.
14492 * In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
14493 dh->length and always used
14495 BN_rand_range(priv_key, dh->p).
14497 BN_rand_range() is not necessary for Diffie-Hellman, and this
14498 specific range makes Diffie-Hellman unnecessarily inefficient if
14499 dh->length (recommended exponent length) is much smaller than the
14500 length of dh->p. We could use BN_rand_range() if the order of
14501 the subgroup was stored in the DH structure, but we only have
14506 BN_rand(priv_key, l, ...)
14508 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
14515 RSA_eay_public_encrypt
14516 RSA_eay_private_decrypt
14517 RSA_eay_private_encrypt (signing)
14518 RSA_eay_public_decrypt (signature verification)
14520 (default implementations for RSA_public_encrypt,
14521 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
14522 always reject numbers >= n.
14526 * In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
14527 to synchronize access to 'locking_thread'. This is necessary on
14528 systems where access to 'locking_thread' (an 'unsigned long'
14529 variable) is not atomic.
14533 * In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
14534 *before* setting the 'crypto_lock_rand' flag. The previous code had
14535 a race condition if 0 is a valid thread ID.
14537 *Travis Vitek <vitek@roguewave.com>*
14539 * Add support for shared libraries under Irix.
14541 *Albert Chin-A-Young <china@thewrittenword.com>*
14543 * Add configuration option to build on Linux on both big-endian and
14544 little-endian MIPS.
14546 *Ralf Baechle <ralf@uni-koblenz.de>*
14548 * Add the possibility to create shared libraries on HP-UX.
14552 ### Changes between 0.9.6a and 0.9.6b [9 Jul 2001]
14554 * Change ssleay_rand_bytes (crypto/rand/md_rand.c)
14555 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
14556 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
14557 PRNG state recovery was possible based on the output of
14558 one PRNG request appropriately sized to gain knowledge on
14559 'md' followed by enough consecutive 1-byte PRNG requests
14560 to traverse all of 'state'.
14562 1. When updating 'md_local' (the current thread's copy of 'md')
14563 during PRNG output generation, hash all of the previous
14564 'md_local' value, not just the half used for PRNG output.
14566 2. Make the number of bytes from 'state' included into the hash
14567 independent from the number of PRNG bytes requested.
14569 The first measure alone would be sufficient to avoid
14570 Markku-Juhani's attack. (Actually it had never occurred
14571 to me that the half of 'md_local' used for chaining was the
14572 half from which PRNG output bytes were taken -- I had always
14573 assumed that the secret half would be used.) The second
14574 measure makes sure that additional data from 'state' is never
14575 mixed into 'md_local' in small portions; this heuristically
14576 further strengthens the PRNG.
14580 * Fix crypto/bn/asm/mips3.s.
14584 * When only the key is given to "enc", the IV is undefined. Print out
14585 an error message in this case.
14589 * Handle special case when X509_NAME is empty in X509 printing routines.
14593 * In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
14594 positive and less than q.
14598 * Don't change `*pointer` in CRYPTO_add_lock() is add_lock_callback is
14599 used: it isn't thread safe and the add_lock_callback should handle
14602 *Paul Rose <Paul.Rose@bridge.com>*
14604 * Verify that incoming data obeys the block size in
14605 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
14611 *Ulf Möller, Bodo Möller*
14613 * The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
14614 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
14615 when fixing the server behaviour for backwards-compatible 'client
14616 hello' messages. (Note that the attack is impractical against
14617 SSL 3.0 and TLS 1.0 anyway because length and version checking
14618 means that the probability of guessing a valid ciphertext is
14619 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
14622 Before 0.9.5, the countermeasure (hide the error by generating a
14623 random 'decryption result') did not work properly because
14624 ERR_clear_error() was missing, meaning that SSL_get_error() would
14625 detect the supposedly ignored error.
14627 Both problems are now fixed.
14631 * In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
14632 (previously it was 1024).
14636 * Fix for compatibility mode trust settings: ignore trust settings
14637 unless some valid trust or reject settings are present.
14641 * Fix for blowfish EVP: its a variable length cipher.
14645 * Fix various bugs related to DSA S/MIME verification. Handle missing
14646 parameters in DSA public key structures and return an error in the
14647 DSA routines if parameters are absent.
14651 * In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
14652 in the current directory if neither $RANDFILE nor $HOME was set.
14653 RAND_file_name() in 0.9.6a returned NULL in this case. This has
14654 caused some confusion to Windows users who haven't defined $HOME.
14655 Thus RAND_file_name() is changed again: e_os.h can define a
14656 DEFAULT_HOME, which will be used if $HOME is not set.
14657 For Windows, we use "C:"; on other platforms, we still require
14658 environment variables.
14660 * Move 'if (!initialized) RAND_poll()' into regions protected by
14661 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
14662 having multiple threads call RAND_poll() concurrently.
14666 * In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
14667 combination of a flag and a thread ID variable.
14668 Otherwise while one thread is in ssleay_rand_bytes (which sets the
14669 flag), *other* threads can enter ssleay_add_bytes without obeying
14670 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
14671 that they do not hold after the first thread unsets add_do_not_lock).
14675 * Change bctest again: '-x' expressions are not available in all
14676 versions of 'test'.
14680 ### Changes between 0.9.6 and 0.9.6a [5 Apr 2001]
14682 * Fix a couple of memory leaks in PKCS7_dataDecode()
14684 *Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>*
14686 * Change Configure and Makefiles to provide EXE_EXT, which will contain
14687 the default extension for executables, if any. Also, make the perl
14688 scripts that use symlink() to test if it really exists and use "cp"
14689 if it doesn't. All this made OpenSSL compilable and installable in
14694 * Fix for asn1_GetSequence() for indefinite length constructed data.
14695 If SEQUENCE is length is indefinite just set c->slen to the total
14696 amount of data available.
14698 *Steve Henson, reported by shige@FreeBSD.org*
14700 *This change does not apply to 0.9.7.*
14702 * Change bctest to avoid here-documents inside command substitution
14703 (workaround for FreeBSD /bin/sh bug).
14704 For compatibility with Ultrix, avoid shell functions (introduced
14705 in the bctest version that searches along $PATH).
14709 * Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
14710 with des_encrypt() defined on some operating systems, like Solaris
14715 * Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
14716 On the Importance of Eliminating Errors in Cryptographic
14717 Computations, J. Cryptology 14 (2001) 2, 101-119,
14718 <http://theory.stanford.edu/~dabo/papers/faults.ps.gz>).
14722 * MIPS assembler BIGNUM division bug fix.
14726 * Disabled incorrect Alpha assembler code.
14730 * Fix PKCS#7 decode routines so they correctly update the length
14731 after reading an EOC for the EXPLICIT tag.
14735 *This change does not apply to 0.9.7.*
14737 * Fix bug in PKCS#12 key generation routines. This was triggered
14738 if a 3DES key was generated with a 0 initial byte. Include
14739 PKCS12_BROKEN_KEYGEN compilation option to retain the old
14740 (but broken) behaviour.
14744 * Enhance bctest to search for a working bc along $PATH and print
14747 *Tim Rice <tim@multitalents.net> via Richard Levitte*
14749 * Fix memory leaks in err.c: free err_data string if necessary;
14750 don't write to the wrong index in ERR_set_error_data.
14754 * Implement ssl23_peek (analogous to ssl23_read), which previously
14759 * Replace rdtsc with `_emit` statements for VC++ version 5.
14761 *Jeremy Cooper <jeremy@baymoo.org>*
14763 * Make it possible to reuse SSLv2 sessions.
14767 * In copy_email() check for >= 0 as a return value for
14768 X509_NAME_get_index_by_NID() since 0 is a valid index.
14770 *Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>*
14772 * Avoid coredump with unsupported or invalid public keys by checking if
14773 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
14774 PKCS7_verify() fails with non detached data.
14778 * Don't use getenv in library functions when run as setuid/setgid.
14779 New function OPENSSL_issetugid().
14783 * Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
14784 due to incorrect handling of multi-threading:
14786 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
14788 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
14790 3. Count how many times MemCheck_off() has been called so that
14791 nested use can be treated correctly. This also avoids
14792 inband-signalling in the previous code (which relied on the
14793 assumption that thread ID 0 is impossible).
14797 * Add "-rand" option also to s_client and s_server.
14801 * Fix CPU detection on Irix 6.x.
14802 *Kurt Hockenbury <khockenb@stevens-tech.edu> and
14803 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>*
14805 * Fix X509_NAME bug which produced incorrect encoding if X509_NAME
14810 *This change does not apply to 0.9.7.*
14812 * Use the cached encoding of an X509_NAME structure rather than
14813 copying it. This is apparently the reason for the libsafe "errors"
14814 but the code is actually correct.
14818 * Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
14819 Bleichenbacher's DSA attack.
14820 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
14821 to be set and top=0 forces the highest bit to be set; top=-1 is new
14822 and leaves the highest bit random.
14824 *Ulf Moeller, Bodo Moeller*
14826 * In the `NCONF_...`-based implementations for `CONF_...` queries
14827 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
14828 a temporary CONF structure with the data component set to NULL
14829 (which gives segmentation faults in lh_retrieve).
14830 Instead, use NULL for the CONF pointer in CONF_get_string and
14831 CONF_get_number (which may use environment variables) and directly
14832 return NULL from CONF_get_section.
14836 * Fix potential buffer overrun for EBCDIC.
14840 * Tolerate nonRepudiation as being valid for S/MIME signing and certSign
14841 keyUsage if basicConstraints absent for a CA.
14845 * Make SMIME_write_PKCS7() write mail header values with a format that
14846 is more generally accepted (no spaces before the semicolon), since
14847 some programs can't parse those values properly otherwise. Also make
14848 sure BIO's that break lines after each write do not create invalid
14853 * Make the CRL encoding routines work with empty SEQUENCE OF. The
14854 macros previously used would not encode an empty SEQUENCE OF
14855 and break the signature.
14859 *This change does not apply to 0.9.7.*
14861 * Zero the premaster secret after deriving the master secret in
14866 * Add some EVP_add_digest_alias registrations (as found in
14867 OpenSSL_add_all_digests()) to SSL_library_init()
14868 aka OpenSSL_add_ssl_algorithms(). This provides improved
14869 compatibility with peers using X.509 certificates
14870 with unconventional AlgorithmIdentifier OIDs.
14874 * Fix for Irix with NO_ASM.
14876 *"Bruce W. Forsberg" <bruce.forsberg@baesystems.com>*
14878 * ./config script fixes.
14880 *Ulf Moeller, Richard Levitte*
14882 * Fix 'openssl passwd -1'.
14886 * Change PKCS12_key_gen_asc() so it can cope with non null
14887 terminated strings whose length is passed in the passlen
14888 parameter, for example from PEM callbacks. This was done
14889 by adding an extra length parameter to asc2uni().
14891 *Steve Henson, reported by <oddissey@samsung.co.kr>*
14893 * Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
14894 call failed, free the DSA structure.
14898 * Fix to uni2asc() to cope with zero length Unicode strings.
14899 These are present in some PKCS#12 files.
14903 * Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
14904 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
14905 when writing a 32767 byte record.
14907 *Bodo Moeller; problem reported by Eric Day <eday@concentric.net>*
14909 * In `RSA_eay_public_{en,ed}crypt` and RSA_eay_mod_exp (rsa_eay.c),
14910 obtain lock CRYPTO_LOCK_RSA before setting `rsa->_method_mod_{n,p,q}`.
14912 (RSA objects have a reference count access to which is protected
14913 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
14914 so they are meant to be shared between threads.)
14915 *Bodo Moeller, Geoff Thorpe; original patch submitted by
14916 "Reddie, Steven" <Steven.Reddie@ca.com>*
14918 * Fix a deadlock in CRYPTO_mem_leaks().
14922 * Use better test patterns in bntest.
14926 * rand_win.c fix for Borland C.
14930 * BN_rshift bugfix for n == 0.
14934 * Add a 'bctest' script that checks for some known 'bc' bugs
14935 so that 'make test' does not abort just because 'bc' is broken.
14939 * Store verify_result within SSL_SESSION also for client side to
14940 avoid potential security hole. (Re-used sessions on the client side
14941 always resulted in verify_result==X509_V_OK, not using the original
14942 result of the server certificate verification.)
14946 * Fix ssl3_pending: If the record in s->s3->rrec is not of type
14947 SSL3_RT_APPLICATION_DATA, return 0.
14948 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
14953 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
14954 releases, have been re-implemented by renaming the previous
14955 implementations of ssl2_read and ssl3_read to ssl2_read_internal
14956 and ssl3_read_internal, respectively, and adding 'peek' parameters
14957 to them. The new ssl[23]_{read,peek} functions are calls to
14958 ssl[23]_read_internal with the 'peek' flag set appropriately.
14959 A 'peek' parameter has also been added to ssl3_read_bytes, which
14960 does the actual work for ssl3_read_internal.
14964 * Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
14965 the method-specific "init()" handler. Also clean up ex_data after
14966 calling the method-specific "finish()" handler. Previously, this was
14967 happening the other way round.
14971 * Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
14972 The previous value, 12, was not always sufficient for BN_mod_exp().
14976 * Make sure that shared libraries get the internal name engine with
14977 the full version number and not just 0. This should mark the
14978 shared libraries as not backward compatible. Of course, this should
14979 be changed again when we can guarantee backward binary compatibility.
14983 * Fix typo in get_cert_by_subject() in by_dir.c
14985 *Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>*
14987 * Rework the system to generate shared libraries:
14989 - Make note of the expected extension for the shared libraries and
14990 if there is a need for symbolic links from for example libcrypto.so.0
14991 to libcrypto.so.0.9.7. There is extended info in Configure for
14994 - Make as few rebuilds of the shared libraries as possible.
14996 - Still avoid linking the OpenSSL programs with the shared libraries.
14998 - When installing, install the shared libraries separately from the
15003 * Fix SSL_CTX_set_read_ahead macro to actually use its argument.
15005 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
15006 and not in SSL_clear because the latter is also used by the
15007 accept/connect functions; previously, the settings made by
15008 SSL_set_read_ahead would be lost during the handshake.
15010 *Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>*
15012 * Correct util/mkdef.pl to be selective about disabled algorithms.
15013 Previously, it would create entries for disabled algorithms no
15018 * Added several new manual pages for SSL_* function.
15022 ### Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
15024 * In ssl23_get_client_hello, generate an error message when faced
15025 with an initial SSL 3.0/TLS record that is too small to contain the
15026 first two bytes of the ClientHello message, i.e. client_version.
15027 (Note that this is a pathologic case that probably has never happened
15028 in real life.) The previous approach was to use the version number
15029 from the record header as a substitute; but our protocol choice
15030 should not depend on that one because it is not authenticated
15031 by the Finished messages.
15035 * More robust randomness gathering functions for Windows.
15037 *Jeffrey Altman <jaltman@columbia.edu>*
15039 * For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
15040 not set then we don't setup the error code for issuer check errors
15041 to avoid possibly overwriting other errors which the callback does
15042 handle. If an application does set the flag then we assume it knows
15043 what it is doing and can handle the new informational codes
15048 * Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
15049 a general "ANY" type, as such it should be able to decode anything
15050 including tagged types. However it didn't check the class so it would
15051 wrongly interpret tagged types in the same way as their universal
15052 counterpart and unknown types were just rejected. Changed so that the
15053 tagged and unknown types are handled in the same way as a SEQUENCE:
15054 that is the encoding is stored intact. There is also a new type
15055 "V_ASN1_OTHER" which is used when the class is not universal, in this
15056 case we have no idea what the actual type is so we just lump them all
15061 * On VMS, stdout may very well lead to a file that is written to
15062 in a record-oriented fashion. That means that every write() will
15063 write a separate record, which will be read separately by the
15064 programs trying to read from it. This can be very confusing.
15066 The solution is to put a BIO filter in the way that will buffer
15067 text until a linefeed is reached, and then write everything a
15068 line at a time, so every record written will be an actual line,
15069 not chunks of lines and not (usually doesn't happen, but I've
15070 seen it once) several lines in one record. BIO_f_linebuffer() is
15073 Currently, it's a VMS-only method, because that's where it has
15074 been tested well enough.
15078 * Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
15079 it can return incorrect results.
15080 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
15081 but it was in 0.9.6-beta[12].)
15085 * Disable the check for content being present when verifying detached
15086 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
15087 include zero length content when signing messages.
15091 * New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
15092 BIO_ctrl (for BIO pairs).
15096 * Add DSO method for VMS.
15100 * Bug fix: Montgomery multiplication could produce results with the
15105 * Add RPM specification openssl.spec and modify it to build three
15106 packages. The default package contains applications, application
15107 documentation and run-time libraries. The devel package contains
15108 include files, static libraries and function documentation. The
15109 doc package contains the contents of the doc directory. The original
15110 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
15114 * Add a large number of documentation files for many SSL routines.
15116 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>*
15118 * Add a configuration entry for Sony News 4.
15120 *NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>*
15122 * Don't set the two most significant bits to one when generating a
15123 random number < q in the DSA library.
15127 * New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
15128 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
15129 the underlying transport is blocking) if a handshake took place.
15130 (The default behaviour is needed by applications such as s_client
15131 and s_server that use select() to determine when to use SSL_read;
15132 but for applications that know in advance when to expect data, it
15133 just makes things more complicated.)
15137 * Add RAND_egd_bytes(), which gives control over the number of bytes read
15142 * Add a few more EBCDIC conditionals that make `req` and `x509`
15143 work better on such systems.
15145 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
15147 * Add two demo programs for PKCS12_parse() and PKCS12_create().
15148 Update PKCS12_parse() so it copies the friendlyName and the
15149 keyid to the certificates aux info.
15153 * Fix bug in PKCS7_verify() which caused an infinite loop
15154 if there was more than one signature.
15156 *Sven Uszpelkat <su@celocom.de>*
15158 * Major change in util/mkdef.pl to include extra information
15159 about each symbol, as well as presenting variables as well
15160 as functions. This change means that there's n more need
15161 to rebuild the .num files when some algorithms are excluded.
15165 * Allow the verify time to be set by an application,
15166 rather than always using the current time.
15170 * Phase 2 verify code reorganisation. The certificate
15171 verify code now looks up an issuer certificate by a
15172 number of criteria: subject name, authority key id
15173 and key usage. It also verifies self signed certificates
15174 by the same criteria. The main comparison function is
15175 X509_check_issued() which performs these checks.
15177 Lot of changes were necessary in order to support this
15178 without completely rewriting the lookup code.
15180 Authority and subject key identifier are now cached.
15182 The LHASH 'certs' is X509_STORE has now been replaced
15183 by a STACK_OF(X509_OBJECT). This is mainly because an
15184 LHASH can't store or retrieve multiple objects with
15185 the same hash value.
15187 As a result various functions (which were all internal
15188 use only) have changed to handle the new X509_STORE
15189 structure. This will break anything that messed round
15190 with X509_STORE internally.
15192 The functions X509_STORE_add_cert() now checks for an
15193 exact match, rather than just subject name.
15195 The X509_STORE API doesn't directly support the retrieval
15196 of multiple certificates matching a given criteria, however
15197 this can be worked round by performing a lookup first
15198 (which will fill the cache with candidate certificates)
15199 and then examining the cache for matches. This is probably
15200 the best we can do without throwing out X509_LOOKUP
15201 entirely (maybe later...).
15203 The X509_VERIFY_CTX structure has been enhanced considerably.
15205 All certificate lookup operations now go via a get_issuer()
15206 callback. Although this currently uses an X509_STORE it
15207 can be replaced by custom lookups. This is a simple way
15208 to bypass the X509_STORE hackery necessary to make this
15209 work and makes it possible to use more efficient techniques
15210 in future. A very simple version which uses a simple
15211 STACK for its trusted certificate store is also provided
15212 using X509_STORE_CTX_trusted_stack().
15214 The verify_cb() and verify() callbacks now have equivalents
15215 in the X509_STORE_CTX structure.
15217 X509_STORE_CTX also has a 'flags' field which can be used
15218 to customise the verify behaviour.
15222 * Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
15223 excludes S/MIME capabilities.
15227 * When a certificate request is read in keep a copy of the
15228 original encoding of the signed data and use it when outputting
15229 again. Signatures then use the original encoding rather than
15230 a decoded, encoded version which may cause problems if the
15231 request is improperly encoded.
15235 * For consistency with other BIO_puts implementations, call
15236 buffer_write(b, ...) directly in buffer_puts instead of calling
15239 In BIO_puts, increment b->num_write as in BIO_write.
15241 *Peter.Sylvester@EdelWeb.fr*
15243 * Fix BN_mul_word for the case where the word is 0. (We have to use
15244 BN_zero, we may not return a BIGNUM with an array consisting of
15245 words set to zero.)
15249 * Avoid calling abort() from within the library when problems are
15250 detected, except if preprocessor symbols have been defined
15251 (such as REF_CHECK, BN_DEBUG etc.).
15255 * New openssl application 'rsautl'. This utility can be
15256 used for low-level RSA operations. DER public key
15257 BIO/fp routines also added.
15261 * New Configure entry and patches for compiling on QNX 4.
15263 *Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>*
15265 * A demo state-machine implementation was sponsored by
15266 Nuron (<http://www.nuron.com/>) and is now available in
15267 demos/state_machine.
15271 * New options added to the 'dgst' utility for signature
15272 generation and verification.
15276 * Unrecognized PKCS#7 content types are now handled via a
15277 catch all ASN1_TYPE structure. This allows unsupported
15278 types to be stored as a "blob" and an application can
15279 encode and decode it manually.
15283 * Fix various signed/unsigned issues to make a_strex.c
15284 compile under VC++.
15286 *Oscar Jacobsson <oscar.jacobsson@celocom.com>*
15288 * ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
15289 length if passed a buffer. ASN1_INTEGER_to_BN failed
15290 if passed a NULL BN and its argument was negative.
15292 *Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>*
15294 * Modification to PKCS#7 encoding routines to output definite
15295 length encoding. Since currently the whole structures are in
15296 memory there's not real point in using indefinite length
15297 constructed encoding. However if OpenSSL is compiled with
15298 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
15302 * Added BIO_vprintf() and BIO_vsnprintf().
15306 * Added more prefixes to parse for in the strings written
15307 through a logging bio, to cover all the levels that are available
15308 through syslog. The prefixes are now:
15310 PANIC, EMERG, EMR => LOG_EMERG
15311 ALERT, ALR => LOG_ALERT
15312 CRIT, CRI => LOG_CRIT
15313 ERROR, ERR => LOG_ERR
15314 WARNING, WARN, WAR => LOG_WARNING
15315 NOTICE, NOTE, NOT => LOG_NOTICE
15316 INFO, INF => LOG_INFO
15317 DEBUG, DBG => LOG_DEBUG
15319 and as before, if none of those prefixes are present at the
15320 beginning of the string, LOG_ERR is chosen.
15322 On Win32, the `LOG_*` levels are mapped according to this:
15324 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
15325 LOG_WARNING => EVENTLOG_WARNING_TYPE
15326 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
15330 * Made it possible to reconfigure with just the configuration
15331 argument "reconf" or "reconfigure". The command line arguments
15332 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
15333 and are retrieved from there when reconfiguring.
15339 *Assar Westerlund <assar@sics.se>, Richard Levitte*
15341 * Add the arguments -CAfile and -CApath to the pkcs12 utility.
15345 * The obj_dat.pl script was messing up the sorting of object
15346 names. The reason was that it compared the quoted version
15347 of strings as a result "OCSP" > "OCSP Signing" because
15348 " > SPACE. Changed script to store unquoted versions of
15349 names and add quotes on output. It was also omitting some
15350 names from the lookup table if they were given a default
15351 value (that is if SN is missing it is given the same
15352 value as LN and vice versa), these are now added on the
15353 grounds that if an object has a name we should be able to
15354 look it up. Finally added warning output when duplicate
15355 short or long names are found.
15359 * Changes needed for Tandem NSK.
15361 *Scott Uroff <scott@xypro.com>*
15363 * Fix SSL 2.0 rollback checking: Due to an off-by-one error in
15364 RSA_padding_check_SSLv23(), special padding was never detected
15365 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
15366 version rollback attacks was not effective.
15368 In s23_clnt.c, don't use special rollback-attack detection padding
15369 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
15370 client; similarly, in s23_srvr.c, don't do the rollback check if
15371 SSL 2.0 is the only protocol enabled in the server.
15375 * Make it possible to get hexdumps of unprintable data with 'openssl
15376 asn1parse'. By implication, the functions ASN1_parse_dump() and
15377 BIO_dump_indent() are added.
15381 * New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
15382 these print out strings and name structures based on various
15383 flags including RFC2253 support and proper handling of
15384 multibyte characters. Added options to the 'x509' utility
15385 to allow the various flags to be set.
15389 * Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
15390 Also change the functions X509_cmp_current_time() and
15391 X509_gmtime_adj() work with an ASN1_TIME structure,
15392 this will enable certificates using GeneralizedTime in validity
15393 dates to be checked.
15397 * Make the NEG_PUBKEY_BUG code (which tolerates invalid
15398 negative public key encodings) on by default,
15399 NO_NEG_PUBKEY_BUG can be set to disable it.
15403 * New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
15404 content octets. An i2c_ASN1_OBJECT is unnecessary because
15405 the encoding can be trivially obtained from the structure.
15409 * crypto/err.c locking bugfix: Use write locks (`CRYPTO_w_[un]lock`),
15410 not read locks (`CRYPTO_r_[un]lock`).
15414 * A first attempt at creating official support for shared
15415 libraries through configuration. I've kept it so the
15416 default is static libraries only, and the OpenSSL programs
15417 are always statically linked for now, but there are
15418 preparations for dynamic linking in place.
15419 This has been tested on Linux and Tru64.
15423 * Randomness polling function for Win9x, as described in:
15424 Peter Gutmann, Software Generation of Practically Strong
15429 * Fix so PRNG is seeded in req if using an already existing
15434 * New options to smime application. -inform and -outform
15435 allow alternative formats for the S/MIME message including
15436 PEM and DER. The -content option allows the content to be
15437 specified separately. This should allow things like Netscape
15438 form signing output easier to verify.
15442 * Fix the ASN1 encoding of tags using the 'long form'.
15446 * New ASN1 functions, `i2c_*` and `c2i_*` for INTEGER and BIT
15447 STRING types. These convert content octets to and from the
15448 underlying type. The actual tag and length octets are
15449 already assumed to have been read in and checked. These
15450 are needed because all other string types have virtually
15451 identical handling apart from the tag. By having versions
15452 of the ASN1 functions that just operate on content octets
15453 IMPLICIT tagging can be handled properly. It also allows
15454 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
15455 and ASN1_INTEGER are identical apart from the tag.
15459 * Change the handling of OID objects as follows:
15461 - New object identifiers are inserted in objects.txt, following
15462 the syntax given in [crypto/objects/README.md](crypto/objects/README.md).
15463 - objects.pl is used to process obj_mac.num and create a new
15465 - obj_dat.pl is used to create a new obj_dat.h, using the data in
15468 This is currently kind of a hack, and the perl code in objects.pl
15469 isn't very elegant, but it works as I intended. The simplest way
15470 to check that it worked correctly is to look in obj_dat.h and
15471 check the array nid_objs and make sure the objects haven't moved
15472 around (this is important!). Additions are OK, as well as
15473 consistent name changes.
15477 * Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
15481 * Addition of the command line parameter '-rand file' to 'openssl req'.
15482 The given file adds to whatever has already been seeded into the
15483 random pool through the RANDFILE configuration file option or
15484 environment variable, or the default random state file.
15488 * mkstack.pl now sorts each macro group into lexical order.
15489 Previously the output order depended on the order the files
15490 appeared in the directory, resulting in needless rewriting
15495 * Patches to make OpenSSL compile under Win32 again. Mostly
15496 work arounds for the VC++ problem that it treats func() as
15497 func(void). Also stripped out the parts of mkdef.pl that
15498 added extra typesafe functions: these no longer exist.
15502 * Reorganisation of the stack code. The macros are now all
15503 collected in safestack.h . Each macro is defined in terms of
15504 a "stack macro" of the form `SKM_<name>(type, a, b)`. The
15505 DEBUG_SAFESTACK is now handled in terms of function casts,
15506 this has the advantage of retaining type safety without the
15507 use of additional functions. If DEBUG_SAFESTACK is not defined
15508 then the non typesafe macros are used instead. Also modified the
15509 mkstack.pl script to handle the new form. Needs testing to see
15510 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
15511 the default if no major problems. Similar behaviour for ASN1_SET_OF
15512 and PKCS12_STACK_OF.
15516 * When some versions of IIS use the 'NET' form of private key the
15517 key derivation algorithm is different. Normally MD5(password) is
15518 used as a 128 bit RC4 key. In the modified case
15519 MD5(MD5(password) + "SGCKEYSALT") is used instead. Added some
15520 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
15521 as the old Netscape_RSA functions except they have an additional
15522 'sgckey' parameter which uses the modified algorithm. Also added
15523 an -sgckey command line option to the rsa utility. Thanks to
15524 Adrian Peck <bertie@ncipher.com> for posting details of the modified
15525 algorithm to openssl-dev.
15529 * The evp_local.h macros were using 'c.##kname' which resulted in
15530 invalid expansion on some systems (SCO 5.0.5 for example).
15531 Corrected to 'c.kname'.
15533 *Phillip Porch <root@theporch.com>*
15535 * New X509_get1_email() and X509_REQ_get1_email() functions that return
15536 a STACK of email addresses from a certificate or request, these look
15537 in the subject name and the subject alternative name extensions and
15538 omit any duplicate addresses.
15542 * Re-implement BN_mod_exp2_mont using independent (and larger) windows.
15543 This makes DSA verification about 2 % faster.
15547 * Increase maximum window size in `BN_mod_exp_...` to 6 bits instead of 5
15548 (meaning that now 2^5 values will be precomputed, which is only 4 KB
15549 plus overhead for 1024 bit moduli).
15550 This makes exponentiations about 0.5 % faster for 1024 bit
15551 exponents (as measured by "openssl speed rsa2048").
15555 * Rename memory handling macros to avoid conflicts with other
15557 Malloc => OPENSSL_malloc
15558 Malloc_locked => OPENSSL_malloc_locked
15559 Realloc => OPENSSL_realloc
15560 Free => OPENSSL_free
15564 * New function BN_mod_exp_mont_word for small bases (roughly 15%
15565 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
15569 * CygWin32 support.
15571 *John Jarvie <jjarvie@newsguy.com>*
15573 * The type-safe stack code has been rejigged. It is now only compiled
15574 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
15575 by default all type-specific stack functions are "#define"d back to
15576 standard stack functions. This results in more streamlined output
15577 but retains the type-safety checking possibilities of the original
15582 * The STACK code has been cleaned up, and certain type declarations
15583 that didn't make a lot of sense have been brought in line. This has
15584 also involved a cleanup of sorts in safestack.h to more correctly
15585 map type-safe stack functions onto their plain stack counterparts.
15586 This work has also resulted in a variety of "const"ifications of
15587 lots of the code, especially `_cmp` operations which should normally
15588 be prototyped with "const" parameters anyway.
15592 * When generating bytes for the first time in md_rand.c, 'stir the pool'
15593 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
15594 (The PRNG state consists of two parts, the large pool 'state' and 'md',
15595 where all of 'md' is used each time the PRNG is used, but 'state'
15596 is used only indexed by a cyclic counter. As entropy may not be
15597 well distributed from the beginning, 'md' is important as a
15598 chaining variable. However, the output function chains only half
15599 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
15600 all of 'md', and seeding with STATE_SIZE dummy bytes will result
15601 in all of 'state' being rewritten, with the new values depending
15602 on virtually all of 'md'. This overcomes the 80 bit limitation.)
15606 * In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
15607 the handshake is continued after ssl_verify_cert_chain();
15608 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
15609 can lead to 'unexplainable' connection aborts later.
15611 *Bodo Moeller; problem tracked down by Lutz Jaenicke*
15613 * Major EVP API cipher revision.
15614 Add hooks for extra EVP features. This allows various cipher
15615 parameters to be set in the EVP interface. Support added for variable
15616 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
15617 setting of RC2 and RC5 parameters.
15619 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
15622 Remove lots of duplicated code from the EVP library. For example *every*
15623 cipher init() function handles the 'iv' in the same way according to the
15624 cipher mode. They also all do nothing if the 'key' parameter is NULL and
15625 for CFB and OFB modes they zero ctx->num.
15627 New functionality allows removal of S/MIME code RC2 hack.
15629 Most of the routines have the same form and so can be declared in terms
15632 By shifting this to the top level EVP_CipherInit() it can be removed from
15633 all individual ciphers. If the cipher wants to handle IVs or keys
15634 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
15637 Change lots of functions like EVP_EncryptUpdate() to now return a
15638 value: although software versions of the algorithms cannot fail
15639 any installed hardware versions can.
15643 * Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
15644 this option is set, tolerate broken clients that send the negotiated
15645 protocol version number instead of the requested protocol version
15650 * Call dh_tmp_cb (set by `..._TMP_DH_CB`) with correct 'is_export' flag;
15651 i.e. non-zero for export ciphersuites, zero otherwise.
15652 Previous versions had this flag inverted, inconsistent with
15653 rsa_tmp_cb (..._TMP_RSA_CB).
15655 *Bodo Moeller; problem reported by Amit Chopra*
15657 * Add missing DSA library text string. Work around for some IIS
15658 key files with invalid SEQUENCE encoding.
15662 * Add a document (doc/standards.txt) that list all kinds of standards
15663 and so on that are implemented in OpenSSL.
15667 * Enhance c_rehash script. Old version would mishandle certificates
15668 with the same subject name hash and wouldn't handle CRLs at all.
15669 Added -fingerprint option to crl utility, to support new c_rehash
15674 * Eliminate non-ANSI declarations in crypto.h and stack.h.
15678 * Fix for SSL server purpose checking. Server checking was
15679 rejecting certificates which had extended key usage present
15680 but no ssl client purpose.
15682 *Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>*
15684 * Make PKCS#12 code work with no password. The PKCS#12 spec
15685 is a little unclear about how a blank password is handled.
15686 Since the password in encoded as a BMPString with terminating
15687 double NULL a zero length password would end up as just the
15688 double NULL. However no password at all is different and is
15689 handled differently in the PKCS#12 key generation code. NS
15690 treats a blank password as zero length. MSIE treats it as no
15691 password on export: but it will try both on import. We now do
15692 the same: PKCS12_parse() tries zero length and no password if
15693 the password is set to "" or NULL (NULL is now a valid password:
15694 it wasn't before) as does the pkcs12 application.
15698 * Bugfixes in `apps/x509.c`: Avoid a memory leak; and don't use
15699 perror when PEM_read_bio_X509_REQ fails, the error message must
15700 be obtained from the error queue.
15704 * Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
15705 it in ERR_remove_state if appropriate, and change ERR_get_state
15706 accordingly to avoid race conditions (this is necessary because
15707 thread_hash is no longer constant once set).
15711 * Bugfix for linux-elf makefile.one.
15715 * RSA_get_default_method() will now cause a default
15716 RSA_METHOD to be chosen if one doesn't exist already.
15717 Previously this was only set during a call to RSA_new()
15718 or RSA_new_method(NULL) meaning it was possible for
15719 RSA_get_default_method() to return NULL.
15723 * Added native name translation to the existing DSO code
15724 that will convert (if the flag to do so is set) filenames
15725 that are sufficiently small and have no path information
15726 into a canonical native form. Eg. "blah" converted to
15727 "libblah.so" or "blah.dll" etc.
15731 * New function ERR_error_string_n(e, buf, len) which is like
15732 ERR_error_string(e, buf), but writes at most 'len' bytes
15733 including the 0 terminator. For ERR_error_string_n, 'buf'
15736 *Damien Miller <djm@mindrot.org>, Bodo Moeller*
15738 * CONF library reworked to become more general. A new CONF
15739 configuration file reader "class" is implemented as well as a
15740 new functions (`NCONF_*`, for "New CONF") to handle it. The now
15741 old `CONF_*` functions are still there, but are reimplemented to
15742 work in terms of the new functions. Also, a set of functions
15743 to handle the internal storage of the configuration data is
15744 provided to make it easier to write new configuration file
15745 reader "classes" (I can definitely see something reading a
15746 configuration file in XML format, for example), called `_CONF_*`,
15747 or "the configuration storage API"...
15749 The new configuration file reading functions are:
15751 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
15752 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
15754 NCONF_default, NCONF_WIN32
15756 NCONF_dump_fp, NCONF_dump_bio
15758 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
15759 NCONF_new creates a new CONF object. This works in the same way
15760 as other interfaces in OpenSSL, like the BIO interface.
15761 `NCONF_dump_*` dump the internal storage of the configuration file,
15762 which is useful for debugging. All other functions take the same
15763 arguments as the old `CONF_*` functions with the exception of the
15764 first that must be a `CONF *` instead of a `LHASH *`.
15766 To make it easier to use the new classes with the old `CONF_*` functions,
15767 the function CONF_set_default_method is provided.
15771 * Add '-tls1' option to 'openssl ciphers', which was already
15772 mentioned in the documentation but had not been implemented.
15773 (This option is not yet really useful because even the additional
15774 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
15778 * Initial DSO code added into libcrypto for letting OpenSSL (and
15779 OpenSSL-based applications) load shared libraries and bind to
15780 them in a portable way.
15782 *Geoff Thorpe, with contributions from Richard Levitte*
15784 ### Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
15786 * Make sure _lrotl and _lrotr are only used with MSVC.
15788 * Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
15789 (the default implementation of RAND_status).
15791 * Rename openssl x509 option '-crlext', which was added in 0.9.5,
15792 to '-clrext' (= clear extensions), as intended and documented.
15793 *Bodo Moeller; inconsistency pointed out by Michael Attili
15794 <attili@amaxo.com>*
15796 * Fix for HMAC. It wasn't zeroing the rest of the block if the key length
15797 was larger than the MD block size.
15799 *Steve Henson, pointed out by Yost William <YostW@tce.com>*
15801 * Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
15802 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
15803 using the passed key: if the passed key was a private key the result
15804 of X509_print(), for example, would be to print out all the private key
15809 * des_quad_cksum() byte order bug fix.
15810 *Ulf Möller, using the problem description in krb4-0.9.7, where
15811 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>*
15813 * Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
15816 *Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>*
15818 * For easily testing in shell scripts whether some command
15819 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
15820 returns with exit code 0 iff no command of the given name is available.
15821 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
15822 the output goes to stdout and nothing is printed to stderr.
15823 Additional arguments are always ignored.
15825 Since for each cipher there is a command of the same name,
15826 the 'no-cipher' compilation switches can be tested this way.
15828 ('openssl no-XXX' is not able to detect pseudo-commands such
15829 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
15833 * Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
15837 * For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
15838 is set; it will be thrown away anyway because each handshake creates
15840 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
15841 to parameters -- in previous versions (since OpenSSL 0.9.3) the
15842 'default key' from SSL_CTX_set_tmp_dh would always be lost, meaning
15843 you effectively got SSL_OP_SINGLE_DH_USE when using this macro.
15847 * New s_client option -ign_eof: EOF at stdin is ignored, and
15848 'Q' and 'R' lose their special meanings (quit/renegotiate).
15849 This is part of what -quiet does; unlike -quiet, -ign_eof
15850 does not suppress any output.
15854 * Add compatibility options to the purpose and trust code. The
15855 purpose X509_PURPOSE_ANY is "any purpose" which automatically
15856 accepts a certificate or CA, this was the previous behaviour,
15857 with all the associated security issues.
15859 X509_TRUST_COMPAT is the old trust behaviour: only and
15860 automatically trust self signed roots in certificate store. A
15861 new trust setting X509_TRUST_DEFAULT is used to specify that
15862 a purpose has no associated trust setting and it should instead
15863 use the value in the default purpose.
15867 * Fix the PKCS#8 DSA private key code so it decodes keys again
15868 and fix a memory leak.
15872 * In util/mkerr.pl (which implements 'make errors'), preserve
15873 reason strings from the previous version of the .c file, as
15874 the default to have only downcase letters (and digits) in
15875 automatically generated reasons codes is not always appropriate.
15879 * In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
15880 using strerror. Previously, ERR_reason_error_string() returned
15881 library names as reason strings for SYSerr; but SYSerr is a special
15882 case where small numbers are errno values, not library numbers.
15886 * Add '-dsaparam' option to 'openssl dhparam' application. This
15887 converts DSA parameters into DH parameters. (When creating parameters,
15888 DSA_generate_parameters is used.)
15892 * Include 'length' (recommended exponent length) in C code generated
15893 by 'openssl dhparam -C'.
15897 * The second argument to set_label in perlasm was already being used
15898 so couldn't be used as a "file scope" flag. Moved to third argument
15903 * In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
15904 instead of RAND_bytes for encryption IVs and salts.
15908 * Include RAND_status() into RAND_METHOD instead of implementing
15909 it only for md_rand.c Otherwise replacing the PRNG by calling
15910 RAND_set_rand_method would be impossible.
15914 * Don't let DSA_generate_key() enter an infinite loop if the random
15915 number generation fails.
15919 * New 'rand' application for creating pseudo-random output.
15923 * Added configuration support for Linux/IA64
15925 *Rolf Haberrecker <rolf@suse.de>*
15927 * Assembler module support for Mingw32.
15931 * Shared library support for HPUX (in shlib/).
15933 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous*
15935 * Shared library support for Solaris gcc.
15937 *Lutz Behnke <behnke@trustcenter.de>*
15939 ### Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
15941 * PKCS7_encrypt() was adding text MIME headers twice because they
15942 were added manually and by SMIME_crlf_copy().
15946 * In bntest.c don't call BN_rand with zero bits argument.
15948 *Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>*
15950 * BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
15951 case was implemented. This caused BN_div_recp() to fail occasionally.
15955 * Add an optional second argument to the set_label() in the perl
15956 assembly language builder. If this argument exists and is set
15957 to 1 it signals that the assembler should use a symbol whose
15958 scope is the entire file, not just the current function. This
15959 is needed with MASM which uses the format label:: for this scope.
15961 *Steve Henson, pointed out by Peter Runestig <peter@runestig.com>*
15963 * Change the ASN1 types so they are typedefs by default. Before
15964 almost all types were #define'd to ASN1_STRING which was causing
15965 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
15970 * Change names of new functions to the new get1/get0 naming
15971 convention: After 'get1', the caller owns a reference count
15972 and has to call `..._free`; 'get0' returns a pointer to some
15973 data structure without incrementing reference counters.
15974 (Some of the existing 'get' functions increment a reference
15975 counter, some don't.)
15976 Similarly, 'set1' and 'add1' functions increase reference
15977 counters or duplicate objects.
15981 * Allow for the possibility of temp RSA key generation failure:
15982 the code used to assume it always worked and crashed on failure.
15986 * Fix potential buffer overrun problem in BIO_printf().
15987 *Ulf Möller, using public domain code by Patrick Powell; problem
15988 pointed out by David Sacerdote <das33@cornell.edu>*
15990 * Support EGD <http://www.lothar.com/tech/crypto/>. New functions
15991 RAND_egd() and RAND_status(). In the command line application,
15992 the EGD socket can be specified like a seed file using RANDFILE
15997 * Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
15998 Some CAs (e.g. Verisign) distribute certificates in this form.
16002 * Remove the SSL_ALLOW_ADH compile option and set the default cipher
16003 list to exclude them. This means that no special compilation option
16004 is needed to use anonymous DH: it just needs to be included in the
16009 * Change the EVP_MD_CTX_type macro so its meaning consistent with
16010 EVP_MD_type. The old functionality is available in a new macro called
16011 EVP_MD_md(). Change code that uses it and update docs.
16015 * `..._ctrl` functions now have corresponding `..._callback_ctrl` functions
16016 where the `void *` argument is replaced by a function pointer argument.
16017 Previously `void *` was abused to point to functions, which works on
16018 many platforms, but is not correct. As these functions are usually
16019 called by macros defined in OpenSSL header files, most source code
16020 should work without changes.
16024 * `<openssl/opensslconf.h>` (which is created by Configure) now contains
16025 sections with information on -D... compiler switches used for
16026 compiling the library so that applications can see them. To enable
16027 one of these sections, a pre-processor symbol `OPENSSL_..._DEFINES`
16028 must be defined. E.g.,
16029 #define OPENSSL_ALGORITHM_DEFINES
16030 #include <openssl/opensslconf.h>
16031 defines all pertinent `NO_<algo>` symbols, such as NO_IDEA, NO_RSA, etc.
16033 *Richard Levitte, Ulf and Bodo Möller*
16035 * Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
16040 * Change the 'other' type in certificate aux info to a STACK_OF
16041 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
16042 the required ASN1 format: arbitrary types determined by an OID.
16046 * Add some PEM_write_X509_REQ_NEW() functions and a command line
16047 argument to 'req'. This is not because the function is newer or
16048 better than others it just uses the work 'NEW' in the certificate
16049 request header lines. Some software needs this.
16053 * Reorganise password command line arguments: now passwords can be
16054 obtained from various sources. Delete the PEM_cb function and make
16055 it the default behaviour: i.e. if the callback is NULL and the
16056 usrdata argument is not NULL interpret it as a null terminated pass
16057 phrase. If usrdata and the callback are NULL then the pass phrase
16058 is prompted for as usual.
16062 * Add support for the Compaq Atalla crypto accelerator. If it is installed,
16063 the support is automatically enabled. The resulting binaries will
16064 autodetect the card and use it if present.
16066 *Ben Laurie and Compaq Inc.*
16068 * Work around for Netscape hang bug. This sends certificate request
16069 and server done in one record. Since this is perfectly legal in the
16070 SSL/TLS protocol it isn't a "bug" option and is on by default. See
16071 the bugs/SSLv3 entry for more info.
16075 * HP-UX tune-up: new unified configs, HP C compiler bug workaround.
16079 * Add -rand argument to smime and pkcs12 applications and read/write
16084 * New 'passwd' tool for crypt(3) and apr1 password hashes.
16088 * Add command line password options to the remaining applications.
16092 * Bug fix for BN_div_recp() for numerators with an even number of
16097 * More tests in bntest.c, and changed test_bn output.
16101 * ./config recognizes MacOS X now.
16105 * Bug fix for BN_div() when the first words of num and divisor are
16106 equal (it gave wrong results if `(rem=(n1-q*d0)&BN_MASK2) < d0)`.
16110 * Add support for various broken PKCS#8 formats, and command line
16111 options to produce them.
16115 * New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
16116 get temporary BIGNUMs from a BN_CTX.
16120 * Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
16125 * Change the `SSLeay_add_all_*()` functions to `OpenSSL_add_all_*()` and
16126 include a #define from the old name to the new. The original intent
16127 was that statically linked binaries could for example just call
16128 SSLeay_add_all_ciphers() to just add ciphers to the table and not
16129 link with digests. This never worked because SSLeay_add_all_digests()
16130 and SSLeay_add_all_ciphers() were in the same source file so calling
16131 one would link with the other. They are now in separate source files.
16135 * Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
16139 * Use a less unusual form of the Miller-Rabin primality test (it used
16140 a binary algorithm for exponentiation integrated into the Miller-Rabin
16141 loop, our standard modexp algorithms are faster).
16145 * Support for the EBCDIC character set completed.
16147 *Martin Kraemer <Martin.Kraemer@Mch.SNI.De>*
16149 * Source code cleanups: use const where appropriate, eliminate casts,
16150 use `void *` instead of `char *` in lhash.
16154 * Bugfix: ssl3_send_server_key_exchange was not restartable
16155 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
16156 this the server could overwrite ephemeral keys that the client
16161 * Turn DSA_is_prime into a macro that calls BN_is_prime,
16162 using 50 iterations of the Rabin-Miller test.
16164 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
16165 iterations of the Rabin-Miller test as required by the appendix
16166 to FIPS PUB 186[-1]) instead of DSA_is_prime.
16167 As BN_is_prime_fasttest includes trial division, DSA parameter
16168 generation becomes much faster.
16170 This implies a change for the callback functions in DSA_is_prime
16171 and DSA_generate_parameters: The callback function is called once
16172 for each positive witness in the Rabin-Miller test, not just
16173 occasionally in the inner loop; and the parameters to the
16174 callback function now provide an iteration count for the outer
16175 loop rather than for the current invocation of the inner loop.
16176 DSA_generate_parameters additionally can call the callback
16177 function with an 'iteration count' of -1, meaning that a
16178 candidate has passed the trial division test (when q is generated
16179 from an application-provided seed, trial division is skipped).
16183 * New function BN_is_prime_fasttest that optionally does trial
16184 division before starting the Rabin-Miller test and has
16185 an additional BN_CTX * argument (whereas BN_is_prime always
16186 has to allocate at least one BN_CTX).
16187 'callback(1, -1, cb_arg)' is called when a number has passed the
16188 trial division stage.
16192 * Fix for bug in CRL encoding. The validity dates weren't being handled
16197 * New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
16201 * New function BN_pseudo_rand().
16205 * Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
16206 bignum version of BN_from_montgomery() with the working code from
16207 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
16212 * Avoid a race condition in s2_clnt.c (function get_server_hello) that
16213 made it impossible to use the same SSL_SESSION data structure in
16214 SSL2 clients in multiple threads.
16218 * The return value of RAND_load_file() no longer counts bytes obtained
16219 by stat(). RAND_load_file(..., -1) is new and uses the complete file
16220 to seed the PRNG (previously an explicit byte count was required).
16222 *Ulf Möller, Bodo Möller*
16224 * Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
16225 used `char *` instead of `void *` and had casts all over the place.
16229 * Make BN_generate_prime() return NULL on error if ret!=NULL.
16233 * Retain source code compatibility for BN_prime_checks macro:
16234 BN_is_prime(..., BN_prime_checks, ...) now uses
16235 BN_prime_checks_for_size to determine the appropriate number of
16236 Rabin-Miller iterations.
16240 * Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
16241 DH_CHECK_P_NOT_SAFE_PRIME.
16242 (Check if this is true? OpenPGP calls them "strong".)
16246 * Merge the functionality of "dh" and "gendh" programs into a new program
16247 "dhparam". The old programs are retained for now but will handle DH keys
16248 (instead of parameters) in future.
16252 * Make the ciphers, s_server and s_client programs check the return values
16253 when a new cipher list is set.
16257 * Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
16258 ciphers. Before when the 56bit ciphers were enabled the sorting was
16261 The syntax for the cipher sorting has been extended to support sorting by
16262 cipher-strength (using the strength_bits hard coded in the tables).
16263 The new command is `@STRENGTH` (see also `doc/apps/ciphers.pod`).
16265 Fix a bug in the cipher-command parser: when supplying a cipher command
16266 string with an "undefined" symbol (neither command nor alphanumeric
16267 *A-Za-z0-9*, ssl_set_cipher_list used to hang in an endless loop. Now
16268 an error is flagged.
16270 Due to the strength-sorting extension, the code of the
16271 ssl_create_cipher_list() function was completely rearranged. I hope that
16272 the readability was also increased :-)
16274 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>*
16276 * Minor change to 'x509' utility. The -CAcreateserial option now uses 1
16277 for the first serial number and places 2 in the serial number file. This
16278 avoids problems when the root CA is created with serial number zero and
16279 the first user certificate has the same issuer name and serial number
16284 * Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
16285 the new code. Add documentation for this stuff.
16289 * Changes to X509_ATTRIBUTE utilities. These have been renamed from
16290 `X509_*()` to `X509at_*()` on the grounds that they don't handle X509
16291 structures and behave in an analogous way to the X509v3 functions:
16292 they shouldn't be called directly but wrapper functions should be used
16295 So we also now have some wrapper functions that call the X509at functions
16296 when passed certificate requests. (TO DO: similar things can be done with
16297 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
16298 things. Some of these need some d2i or i2d and print functionality
16299 because they handle more complex structures.)
16303 * Add missing #ifndefs that caused missing symbols when building libssl
16304 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
16305 NO_RSA in `ssl/s2*.c`.
16307 *Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller*
16309 * Precautions against using the PRNG uninitialized: RAND_bytes() now
16310 has a return value which indicates the quality of the random data
16311 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
16312 error queue. New function RAND_pseudo_bytes() generates output that is
16313 guaranteed to be unique but not unpredictable. RAND_add is like
16314 RAND_seed, but takes an extra argument for an entropy estimate
16315 (RAND_seed always assumes full entropy).
16319 * Do more iterations of Rabin-Miller probable prime test (specifically,
16320 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
16321 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
16322 in crypto/bn/bn_prime.c for the complete table). This guarantees a
16323 false-positive rate of at most 2^-80 for random input.
16327 * Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
16331 * New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
16332 in the 0.9.5 release), this returns the chain
16333 from an X509_CTX structure with a dup of the stack and all
16334 the X509 reference counts upped: so the stack will exist
16335 after X509_CTX_cleanup() has been called. Modify pkcs12.c
16338 Also make SSL_SESSION_print() print out the verify return
16343 * Add manpage for the pkcs12 command. Also change the default
16344 behaviour so MAC iteration counts are used unless the new
16345 -nomaciter option is used. This improves file security and
16346 only older versions of MSIE (4.0 for example) need it.
16350 * Honor the no-xxx Configure options when creating .DEF files.
16354 * Add PKCS#10 attributes to field table: challengePassword,
16355 unstructuredName and unstructuredAddress. These are taken from
16356 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
16357 international characters are used.
16359 More changes to X509_ATTRIBUTE code: allow the setting of types
16360 based on strings. Remove the 'loc' parameter when adding
16361 attributes because these will be a SET OF encoding which is sorted
16366 * Initial changes to the 'req' utility to allow request generation
16367 automation. This will allow an application to just generate a template
16368 file containing all the field values and have req construct the
16371 Initial support for X509_ATTRIBUTE handling. Stacks of these are
16372 used all over the place including certificate requests and PKCS#7
16373 structures. They are currently handled manually where necessary with
16374 some primitive wrappers for PKCS#7. The new functions behave in a
16375 manner analogous to the X509 extension functions: they allow
16376 attributes to be looked up by NID and added.
16378 Later something similar to the X509V3 code would be desirable to
16379 automatically handle the encoding, decoding and printing of the
16380 more complex types. The string types like challengePassword can
16381 be handled by the string table functions.
16383 Also modified the multi byte string table handling. Now there is
16384 a 'global mask' which masks out certain types. The table itself
16385 can use the flag STABLE_NO_MASK to ignore the mask setting: this
16386 is useful when for example there is only one permissible type
16387 (as in countryName) and using the mask might result in no valid
16392 * Clean up 'Finished' handling, and add functions SSL_get_finished and
16393 SSL_get_peer_finished to allow applications to obtain the latest
16394 Finished messages sent to the peer or expected from the peer,
16395 respectively. (SSL_get_peer_finished is usually the Finished message
16396 actually received from the peer, otherwise the protocol will be aborted.)
16398 As the Finished message are message digests of the complete handshake
16399 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
16400 be used for external authentication procedures when the authentication
16401 provided by SSL/TLS is not desired or is not enough.
16405 * Enhanced support for Alpha Linux is added. Now ./config checks if
16406 the host supports BWX extension and if Compaq C is present on the
16407 $PATH. Just exploiting of the BWX extension results in 20-30%
16408 performance kick for some algorithms, e.g. DES and RC4 to mention
16409 a couple. Compaq C in turn generates ~20% faster code for MD5 and
16414 * Add support for MS "fast SGC". This is arguably a violation of the
16415 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
16416 weak crypto and after checking the certificate is SGC a second one
16417 with strong crypto. MS SGC stops the first handshake after receiving
16418 the server certificate message and sends a second client hello. Since
16419 a server will typically do all the time consuming operations before
16420 expecting any further messages from the client (server key exchange
16421 is the most expensive) there is little difference between the two.
16423 To get OpenSSL to support MS SGC we have to permit a second client
16424 hello message after we have sent server done. In addition we have to
16425 reset the MAC if we do get this second client hello.
16429 * Add a function 'd2i_AutoPrivateKey()' this will automatically decide
16430 if a DER encoded private key is RSA or DSA traditional format. Changed
16431 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
16432 format DER encoded private key. Newer code should use PKCS#8 format which
16433 has the key type encoded in the ASN1 structure. Added DER private key
16434 support to pkcs8 application.
16438 * SSL 3/TLS 1 servers now don't request certificates when an anonymous
16439 ciphersuites has been selected (as required by the SSL 3/TLS 1
16440 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
16441 is set, we interpret this as a request to violate the specification
16442 (the worst that can happen is a handshake failure, and 'correct'
16443 behaviour would result in a handshake failure anyway).
16447 * In SSL_CTX_add_session, take into account that there might be multiple
16448 SSL_SESSION structures with the same session ID (e.g. when two threads
16449 concurrently obtain them from an external cache).
16450 The internal cache can handle only one SSL_SESSION with a given ID,
16451 so if there's a conflict, we now throw out the old one to achieve
16456 * Add OIDs for idea and blowfish in CBC mode. This will allow both
16457 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
16458 some routines that use cipher OIDs: some ciphers do not have OIDs
16459 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
16464 * Simplify the trust setting structure and code. Now we just have
16465 two sequences of OIDs for trusted and rejected settings. These will
16466 typically have values the same as the extended key usage extension
16467 and any application specific purposes.
16469 The trust checking code now has a default behaviour: it will just
16470 check for an object with the same NID as the passed id. Functions can
16471 be provided to override either the default behaviour or the behaviour
16472 for a given id. SSL client, server and email already have functions
16473 in place for compatibility: they check the NID and also return "trusted"
16474 if the certificate is self signed.
16478 * Add d2i,i2d bio/fp functions for PrivateKey: these convert the
16479 traditional format into an EVP_PKEY structure.
16483 * Add a password callback function PEM_cb() which either prompts for
16484 a password if usr_data is NULL or otherwise assumes it is a null
16485 terminated password. Allow passwords to be passed on command line
16486 environment or config files in a few more utilities.
16490 * Add a bunch of DER and PEM functions to handle PKCS#8 format private
16491 keys. Add some short names for PKCS#8 PBE algorithms and allow them
16492 to be specified on the command line for the pkcs8 and pkcs12 utilities.
16493 Update documentation.
16497 * Support for ASN1 "NULL" type. This could be handled before by using
16498 ASN1_TYPE but there wasn't any function that would try to read a NULL
16499 and produce an error if it couldn't. For compatibility we also have
16500 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
16501 don't allocate anything because they don't need to.
16505 * Initial support for MacOS is now provided. Examine INSTALL.MacOS
16508 *Andy Polyakov, Roy Woods <roy@centicsystems.ca>*
16510 * Rebuild of the memory allocation routines used by OpenSSL code and
16511 possibly others as well. The purpose is to make an interface that
16512 provide hooks so anyone can build a separate set of allocation and
16513 deallocation routines to be used by OpenSSL, for example memory
16514 pool implementations, or something else, which was previously hard
16515 since Malloc(), Realloc() and Free() were defined as macros having
16516 the values malloc, realloc and free, respectively (except for Win32
16517 compilations). The same is provided for memory debugging code.
16518 OpenSSL already comes with functionality to find memory leaks, but
16519 this gives people a chance to debug other memory problems.
16521 With these changes, a new set of functions and macros have appeared:
16523 CRYPTO_set_mem_debug_functions() [F]
16524 CRYPTO_get_mem_debug_functions() [F]
16525 CRYPTO_dbg_set_options() [F]
16526 CRYPTO_dbg_get_options() [F]
16527 CRYPTO_malloc_debug_init() [M]
16529 The memory debug functions are NULL by default, unless the library
16530 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
16531 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
16532 gives the standard debugging functions that come with OpenSSL) or
16533 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
16534 provided by the library user) must be used. When the standard
16535 debugging functions are used, CRYPTO_dbg_set_options can be used to
16536 request additional information:
16537 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
16538 the CRYPTO_MDEBUG_xxx macro when compiling the library.
16540 Also, things like CRYPTO_set_mem_functions will always give the
16541 expected result (the new set of functions is used for allocation
16542 and deallocation) at all times, regardless of platform and compiler
16545 To finish it up, some functions that were never use in any other
16546 way than through macros have a new API and new semantic:
16548 CRYPTO_dbg_malloc()
16549 CRYPTO_dbg_realloc()
16552 All macros of value have retained their old syntax.
16554 *Richard Levitte and Bodo Moeller*
16556 * Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
16557 ordering of SMIMECapabilities wasn't in "strength order" and there
16558 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
16563 * Some ASN1 types with illegal zero length encoding (INTEGER,
16564 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
16566 *Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson*
16568 * Merge in my S/MIME library for OpenSSL. This provides a simple
16569 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
16570 functionality to handle multipart/signed properly) and a utility
16571 called 'smime' to call all this stuff. This is based on code I
16572 originally wrote for Celo who have kindly allowed it to be
16573 included in OpenSSL.
16577 * Add variants des_set_key_checked and des_set_key_unchecked of
16578 des_set_key (aka des_key_sched). Global variable des_check_key
16579 decides which of these is called by des_set_key; this way
16580 des_check_key behaves as it always did, but applications and
16581 the library itself, which was buggy for des_check_key == 1,
16582 have a cleaner way to pick the version they need.
16586 * New function PKCS12_newpass() which changes the password of a
16591 * Modify X509_TRUST and X509_PURPOSE so it also uses a static and
16592 dynamic mix. In both cases the ids can be used as an index into the
16593 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
16594 functions so they accept a list of the field values and the
16595 application doesn't need to directly manipulate the X509_TRUST
16600 * Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
16605 * Modify the way the V3 extension code looks up extensions. This now
16606 works in a similar way to the object code: we have some "standard"
16607 extensions in a static table which is searched with OBJ_bsearch()
16608 and the application can add dynamic ones if needed. The file
16609 crypto/x509v3/ext_dat.h now has the info: this file needs to be
16610 updated whenever a new extension is added to the core code and kept
16611 in ext_nid order. There is a simple program 'tabtest.c' which checks
16612 this. New extensions are not added too often so this file can readily
16613 be maintained manually.
16615 There are two big advantages in doing things this way. The extensions
16616 can be looked up immediately and no longer need to be "added" using
16617 X509V3_add_standard_extensions(): this function now does nothing.
16618 Side note: I get *lots* of email saying the extension code doesn't
16619 work because people forget to call this function.
16620 Also no dynamic allocation is done unless new extensions are added:
16621 so if we don't add custom extensions there is no need to call
16622 X509V3_EXT_cleanup().
16626 * Modify enc utility's salting as follows: make salting the default. Add a
16627 magic header, so unsalted files fail gracefully instead of just decrypting
16628 to garbage. This is because not salting is a big security hole, so people
16629 should be discouraged from doing it.
16633 * Fixes and enhancements to the 'x509' utility. It allowed a message
16634 digest to be passed on the command line but it only used this
16635 parameter when signing a certificate. Modified so all relevant
16636 operations are affected by the digest parameter including the
16637 -fingerprint and -x509toreq options. Also -x509toreq choked if a
16638 DSA key was used because it didn't fix the digest.
16642 * Initial certificate chain verify code. Currently tests the untrusted
16643 certificates for consistency with the verify purpose (which is set
16644 when the X509_STORE_CTX structure is set up) and checks the pathlength.
16646 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
16647 this is because it will reject chains with invalid extensions whereas
16648 every previous version of OpenSSL and SSLeay made no checks at all.
16650 Trust code: checks the root CA for the relevant trust settings. Trust
16651 settings have an initial value consistent with the verify purpose: e.g.
16652 if the verify purpose is for SSL client use it expects the CA to be
16653 trusted for SSL client use. However the default value can be changed to
16654 permit custom trust settings: one example of this would be to only trust
16655 certificates from a specific "secure" set of CAs.
16657 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
16658 which should be used for version portability: especially since the
16659 verify structure is likely to change more often now.
16661 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
16662 to set them. If not set then assume SSL clients will verify SSL servers
16665 Two new options to the verify program: -untrusted allows a set of
16666 untrusted certificates to be passed in and -purpose which sets the
16667 intended purpose of the certificate. If a purpose is set then the
16668 new chain verify code is used to check extension consistency.
16672 * Support for the authority information access extension.
16676 * Modify RSA and DSA PEM read routines to transparently handle
16677 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
16678 public keys in a format compatible with certificate
16679 SubjectPublicKeyInfo structures. Unfortunately there were already
16680 functions called *_PublicKey_* which used various odd formats so
16681 these are retained for compatibility: however the DSA variants were
16682 never in a public release so they have been deleted. Changed dsa/rsa
16683 utilities to handle the new format: note no releases ever handled public
16684 keys so we should be OK.
16686 The primary motivation for this change is to avoid the same fiasco
16687 that dogs private keys: there are several incompatible private key
16688 formats some of which are standard and some OpenSSL specific and
16689 require various evil hacks to allow partial transparent handling and
16690 even then it doesn't work with DER formats. Given the option anything
16691 other than PKCS#8 should be dumped: but the other formats have to
16692 stay in the name of compatibility.
16694 With public keys and the benefit of hindsight one standard format
16695 is used which works with EVP_PKEY, RSA or DSA structures: though
16696 it clearly returns an error if you try to read the wrong kind of key.
16698 Added a -pubkey option to the 'x509' utility to output the public key.
16699 Also rename the `EVP_PKEY_get_*()` to `EVP_PKEY_rget_*()`
16700 (renamed to `EVP_PKEY_get1_*()` in the OpenSSL 0.9.5 release) and add
16701 `EVP_PKEY_rset_*()` functions (renamed to `EVP_PKEY_set1_*()`)
16702 that do the same as the `EVP_PKEY_assign_*()` except they up the
16703 reference count of the added key (they don't "swallow" the
16708 * Fixes to crypto/x509/by_file.c the code to read in certificates and
16709 CRLs would fail if the file contained no certificates or no CRLs:
16710 added a new function to read in both types and return the number
16711 read: this means that if none are read it will be an error. The
16712 DER versions of the certificate and CRL reader would always fail
16713 because it isn't possible to mix certificates and CRLs in DER format
16714 without choking one or the other routine. Changed this to just read
16715 a certificate: this is the best we can do. Also modified the code
16716 in `apps/verify.c` to take notice of return codes: it was previously
16717 attempting to read in certificates from NULL pointers and ignoring
16718 any errors: this is one reason why the cert and CRL reader seemed
16719 to work. It doesn't check return codes from the default certificate
16720 routines: these may well fail if the certificates aren't installed.
16724 * Code to support otherName option in GeneralName.
16728 * First update to verify code. Change the verify utility
16729 so it warns if it is passed a self signed certificate:
16730 for consistency with the normal behaviour. X509_verify
16731 has been modified to it will now verify a self signed
16732 certificate if *exactly* the same certificate appears
16733 in the store: it was previously impossible to trust a
16734 single self signed certificate. This means that:
16735 openssl verify ss.pem
16736 now gives a warning about a self signed certificate but
16737 openssl verify -CAfile ss.pem ss.pem
16742 * For servers, store verify_result in SSL_SESSION data structure
16743 (and add it to external session representation).
16744 This is needed when client certificate verifications fails,
16745 but an application-provided verification callback (set by
16746 SSL_CTX_set_cert_verify_callback) allows accepting the session
16747 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
16748 but returns 1): When the session is reused, we have to set
16749 ssl->verify_result to the appropriate error code to avoid
16752 *Bodo Moeller, problem pointed out by Lutz Jaenicke*
16754 * Fix a bug in the new PKCS#7 code: it didn't consider the
16755 case in PKCS7_dataInit() where the signed PKCS7 structure
16756 didn't contain any existing data because it was being created.
16758 *Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson*
16760 * Add a salt to the key derivation routines in enc.c. This
16761 forms the first 8 bytes of the encrypted file. Also add a
16762 -S option to allow a salt to be input on the command line.
16766 * New function X509_cmp(). Oddly enough there wasn't a function
16767 to compare two certificates. We do this by working out the SHA1
16768 hash and comparing that. X509_cmp() will be needed by the trust
16773 * SSL_get1_session() is like SSL_get_session(), but increments
16774 the reference count in the SSL_SESSION returned.
16776 *Geoff Thorpe <geoff@eu.c2.net>*
16778 * Fix for 'req': it was adding a null to request attributes.
16779 Also change the X509_LOOKUP and X509_INFO code to handle
16780 certificate auxiliary information.
16784 * Add support for 40 and 64 bit RC2 and RC4 algorithms: document
16789 * Add the possibility to add extra information to the memory leak
16790 detecting output, to form tracebacks, showing from where each
16791 allocation was originated: CRYPTO_push_info("constant string") adds
16792 the string plus current file name and line number to a per-thread
16793 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
16794 is like calling CYRPTO_pop_info() until the stack is empty.
16795 Also updated memory leak detection code to be multi-thread-safe.
16799 * Add options -text and -noout to pkcs7 utility and delete the
16800 encryption options which never did anything. Update docs.
16804 * Add options to some of the utilities to allow the pass phrase
16805 to be included on either the command line (not recommended on
16806 OSes like Unix) or read from the environment. Update the
16807 manpages and fix a few bugs.
16811 * Add a few manpages for some of the openssl commands.
16815 * Fix the -revoke option in ca. It was freeing up memory twice,
16816 leaking and not finding already revoked certificates.
16820 * Extensive changes to support certificate auxiliary information.
16821 This involves the use of X509_CERT_AUX structure and X509_AUX
16822 functions. An X509_AUX function such as PEM_read_X509_AUX()
16823 can still read in a certificate file in the usual way but it
16824 will also read in any additional "auxiliary information". By
16825 doing things this way a fair degree of compatibility can be
16826 retained: existing certificates can have this information added
16827 using the new 'x509' options.
16829 Current auxiliary information includes an "alias" and some trust
16830 settings. The trust settings will ultimately be used in enhanced
16831 certificate chain verification routines: currently a certificate
16832 can only be trusted if it is self signed and then it is trusted
16837 * Fix assembler for Alpha (tested only on DEC OSF not Linux or `*BSD`).
16838 The problem was that one of the replacement routines had not been working
16839 since SSLeay releases. For now the offending routine has been replaced
16840 with non-optimised assembler. Even so, this now gives around 95%
16841 performance improvement for 1024 bit RSA signs.
16845 * Hack to fix PKCS#7 decryption when used with some unorthodox RC2
16846 handling. Most clients have the effective key size in bits equal to
16847 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
16848 A few however don't do this and instead use the size of the decrypted key
16849 to determine the RC2 key length and the AlgorithmIdentifier to determine
16850 the effective key length. In this case the effective key length can still
16851 be 40 bits but the key length can be 168 bits for example. This is fixed
16852 by manually forcing an RC2 key into the EVP_PKEY structure because the
16853 EVP code can't currently handle unusual RC2 key sizes: it always assumes
16854 the key length and effective key length are equal.
16858 * Add a bunch of functions that should simplify the creation of
16859 X509_NAME structures. Now you should be able to do:
16860 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
16861 and have it automatically work out the correct field type and fill in
16862 the structures. The more adventurous can try:
16863 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
16864 and it will (hopefully) work out the correct multibyte encoding.
16868 * Change the 'req' utility to use the new field handling and multibyte
16869 copy routines. Before the DN field creation was handled in an ad hoc
16870 way in req, ca, and x509 which was rather broken and didn't support
16871 BMPStrings or UTF8Strings. Since some software doesn't implement
16872 BMPStrings or UTF8Strings yet, they can be enabled using the config file
16873 using the dirstring_type option. See the new comment in the default
16874 openssl.cnf for more info.
16878 * Make crypto/rand/md_rand.c more robust:
16879 - Assure unique random numbers after fork().
16880 - Make sure that concurrent threads access the global counter and
16881 md serializably so that we never lose entropy in them
16882 or use exactly the same state in multiple threads.
16883 Access to the large state is not always serializable because
16884 the additional locking could be a performance killer, and
16885 md should be large enough anyway.
16889 * New file `apps/app_rand.c` with commonly needed functionality
16890 for handling the random seed file.
16892 Use the random seed file in some applications that previously did not:
16894 dsaparam -genkey (which also ignored its '-rand' option),
16897 x509 (when signing).
16898 Except on systems with /dev/urandom, it is crucial to have a random
16899 seed file at least for key creation, DSA signing, and for DH exchanges;
16900 for RSA signatures we could do without one.
16902 gendh and gendsa (unlike genrsa) used to read only the first byte
16903 of each file listed in the '-rand' option. The function as previously
16904 found in genrsa is now in app_rand.c and is used by all programs
16905 that support '-rand'.
16909 * In RAND_write_file, use mode 0600 for creating files;
16910 don't just chmod when it may be too late.
16914 * Report an error from X509_STORE_load_locations
16915 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
16919 * New function ASN1_mbstring_copy() this copies a string in either
16920 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
16921 into an ASN1_STRING type. A mask of permissible types is passed
16922 and it chooses the "minimal" type to use or an error if not type
16927 * Add function equivalents to the various macros in asn1.h. The old
16928 macros are retained with an `M_` prefix. Code inside the library can
16929 use the `M_` macros. External code (including the openssl utility)
16930 should *NOT* in order to be "shared library friendly".
16934 * Add various functions that can check a certificate's extensions
16935 to see if it usable for various purposes such as SSL client,
16936 server or S/MIME and CAs of these types. This is currently
16937 VERY EXPERIMENTAL but will ultimately be used for certificate chain
16938 verification. Also added a -purpose flag to x509 utility to
16939 print out all the purposes.
16943 * Add a CRYPTO_EX_DATA to X509 certificate structure and associated
16948 * New `X509V3_{X509,CRL,REVOKED}_get_d2i()` functions. These will search
16949 for, obtain and decode and extension and obtain its critical flag.
16950 This allows all the necessary extension code to be handled in a
16951 single function call.
16955 * RC4 tune-up featuring 30-40% performance improvement on most RISC
16956 platforms. See crypto/rc4/rc4_enc.c for further details.
16960 * New -noout option to asn1parse. This causes no output to be produced
16961 its main use is when combined with -strparse and -out to extract data
16962 from a file (which may not be in ASN.1 format).
16966 * Fix for pkcs12 program. It was hashing an invalid certificate pointer
16967 when producing the local key id.
16969 *Richard Levitte <levitte@stacken.kth.se>*
16971 * New option -dhparam in s_server. This allows a DH parameter file to be
16972 stated explicitly. If it is not stated then it tries the first server
16973 certificate file. The previous behaviour hard coded the filename
16978 * Add -pubin and -pubout options to the rsa and dsa commands. These allow
16979 a public key to be input or output. For example:
16980 openssl rsa -in key.pem -pubout -out pubkey.pem
16981 Also added necessary DSA public key functions to handle this.
16985 * Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
16986 in the message. This was handled by allowing
16987 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
16989 *Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>*
16991 * Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
16992 to the end of the strings whereas this didn't. This would cause problems
16993 if strings read with d2i_ASN1_bytes() were later modified.
16995 *Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>*
16997 * Fix for base64 decode bug. When a base64 bio reads only one line of
16998 data and it contains EOF it will end up returning an error. This is
16999 caused by input 46 bytes long. The cause is due to the way base64
17000 BIOs find the start of base64 encoded data. They do this by trying a
17001 trial decode on each line until they find one that works. When they
17002 do a flag is set and it starts again knowing it can pass all the
17003 data directly through the decoder. Unfortunately it doesn't reset
17004 the context it uses. This means that if EOF is reached an attempt
17005 is made to pass two EOFs through the context and this causes the
17006 resulting error. This can also cause other problems as well. As is
17007 usual with these problems it takes *ages* to find and the fix is
17008 trivial: move one line.
17010 *Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer)*
17012 * Ugly workaround to get s_client and s_server working under Windows. The
17013 old code wouldn't work because it needed to select() on sockets and the
17014 tty (for keypresses and to see if data could be written). Win32 only
17015 supports select() on sockets so we select() with a 1s timeout on the
17016 sockets and then see if any characters are waiting to be read, if none
17017 are present then we retry, we also assume we can always write data to
17018 the tty. This isn't nice because the code then blocks until we've
17019 received a complete line of data and it is effectively polling the
17020 keyboard at 1s intervals: however it's quite a bit better than not
17021 working at all :-) A dedicated Windows application might handle this
17022 with an event loop for example.
17026 * Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
17027 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
17028 will be called when RSA_sign() and RSA_verify() are used. This is useful
17029 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
17030 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
17031 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
17032 This necessitated the support of an extra signature type NID_md5_sha1
17033 for SSL signatures and modifications to the SSL library to use it instead
17034 of calling RSA_public_decrypt() and RSA_private_encrypt().
17038 * Add new -verify -CAfile and -CApath options to the crl program, these
17039 will lookup a CRL issuers certificate and verify the signature in a
17040 similar way to the verify program. Tidy up the crl program so it
17041 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
17042 less strict. It will now permit CRL extensions even if it is not
17043 a V2 CRL: this will allow it to tolerate some broken CRLs.
17047 * Initialize all non-automatic variables each time one of the openssl
17048 sub-programs is started (this is necessary as they may be started
17049 multiple times from the "OpenSSL>" prompt).
17051 *Lennart Bang, Bodo Moeller*
17053 * Preliminary compilation option RSA_NULL which disables RSA crypto without
17054 removing all other RSA functionality (this is what NO_RSA does). This
17055 is so (for example) those in the US can disable those operations covered
17056 by the RSA patent while allowing storage and parsing of RSA keys and RSA
17061 * Non-copying interface to BIO pairs.
17062 (still largely untested)
17066 * New function ASN1_tag2str() to convert an ASN1 tag to a descriptive
17067 ASCII string. This was handled independently in various places before.
17071 * New functions UTF8_getc() and UTF8_putc() that parse and generate
17072 UTF8 strings a character at a time.
17076 * Use client_version from client hello to select the protocol
17077 (s23_srvr.c) and for RSA client key exchange verification
17078 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
17082 * Add various utility functions to handle SPKACs, these were previously
17083 handled by poking round in the structure internals. Added new function
17084 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
17085 print, verify and generate SPKACs. Based on an original idea from
17086 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
17090 * RIPEMD160 is operational on all platforms and is back in 'make test'.
17094 * Allow the config file extension section to be overwritten on the
17095 command line. Based on an original idea from Massimiliano Pala
17096 <madwolf@comune.modena.it>. The new option is called -extensions
17097 and can be applied to ca, req and x509. Also -reqexts to override
17098 the request extensions in req and -crlexts to override the crl extensions
17103 * Add new feature to the SPKAC handling in ca. Now you can include
17104 the same field multiple times by preceding it by "XXXX." for example:
17107 this is the same syntax as used in the req config file.
17111 * Allow certificate extensions to be added to certificate requests. These
17112 are specified in a 'req_extensions' option of the req section of the
17113 config file. They can be printed out with the -text option to req but
17114 are otherwise ignored at present.
17118 * Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
17119 data read consists of only the final block it would not decrypted because
17120 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
17121 A misplaced 'break' also meant the decrypted final block might not be
17122 copied until the next read.
17126 * Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
17127 a few extra parameters to the DH structure: these will be useful if
17128 for example we want the value of 'q' or implement X9.42 DH.
17132 * Initial support for DSA_METHOD. This is based on the RSA_METHOD and
17133 provides hooks that allow the default DSA functions or functions on a
17134 "per key" basis to be replaced. This allows hardware acceleration and
17135 hardware key storage to be handled without major modification to the
17136 library. Also added low-level modexp hooks and CRYPTO_EX structure and
17137 associated functions.
17141 * Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
17142 as "read only": it can't be written to and the buffer it points to will
17143 not be freed. Reading from a read only BIO is much more efficient than
17144 a normal memory BIO. This was added because there are several times when
17145 an area of memory needs to be read from a BIO. The previous method was
17146 to create a memory BIO and write the data to it, this results in two
17147 copies of the data and an O(n^2) reading algorithm. There is a new
17148 function BIO_new_mem_buf() which creates a read only memory BIO from
17149 an area of memory. Also modified the PKCS#7 routines to use read only
17154 * Bugfix: ssl23_get_client_hello did not work properly when called in
17155 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
17156 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
17157 but a retry condition occurred while trying to read the rest.
17161 * The PKCS7_ENC_CONTENT_new() function was setting the content type as
17162 NID_pkcs7_encrypted by default: this was wrong since this should almost
17163 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
17164 the encrypted data type: this is a more sensible place to put it and it
17165 allows the PKCS#12 code to be tidied up that duplicated this
17170 * Changed obj_dat.pl script so it takes its input and output files on
17171 the command line. This should avoid shell escape redirection problems
17176 * Initial support for certificate extension requests, these are included
17177 in things like Xenroll certificate requests. Included functions to allow
17178 extensions to be obtained and added.
17182 * -crlf option to s_client and s_server for sending newlines as
17183 CRLF (as required by many protocols).
17187 ### Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
17189 * Install libRSAglue.a when OpenSSL is built with RSAref.
17191 *Ralf S. Engelschall*
17193 * A few more `#ifndef NO_FP_API / #endif` pairs for consistency.
17195 *Andrija Antonijevic <TheAntony2@bigfoot.com>*
17197 * Fix -startdate and -enddate (which was missing) arguments to 'ca'
17202 * New function DSA_dup_DH, which duplicates DSA parameters/keys as
17203 DH parameters/keys (q is lost during that conversion, but the resulting
17204 DH parameters contain its length).
17206 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
17207 much faster than DH_generate_parameters (which creates parameters
17208 where `p = 2*q + 1`), and also the smaller q makes DH computations
17209 much more efficient (160-bit exponentiation instead of 1024-bit
17210 exponentiation); so this provides a convenient way to support DHE
17211 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
17212 utter importance to use
17213 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
17215 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
17216 when such DH parameters are used, because otherwise small subgroup
17217 attacks may become possible!
17221 * Avoid memory leak in i2d_DHparams.
17225 * Allow the -k option to be used more than once in the enc program:
17226 this allows the same encrypted message to be read by multiple recipients.
17230 * New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
17231 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
17232 it will always use the numerical form of the OID, even if it has a short
17237 * Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
17238 method only got called if p,q,dmp1,dmq1,iqmp components were present,
17239 otherwise bn_mod_exp was called. In the case of hardware keys for example
17240 no private key components need be present and it might store extra data
17241 in the RSA structure, which cannot be accessed from bn_mod_exp.
17242 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
17243 private key operations.
17247 * Added support for SPARC Linux.
17251 * pem_password_cb function type incompatibly changed from
17252 typedef int pem_password_cb(char *buf, int size, int rwflag);
17254 ....(char *buf, int size, int rwflag, void *userdata);
17255 so that applications can pass data to their callbacks:
17256 The `PEM[_ASN1]_{read,write}...` functions and macros now take an
17257 additional void * argument, which is just handed through whenever
17258 the password callback is called.
17260 *Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller*
17262 New function SSL_CTX_set_default_passwd_cb_userdata.
17264 Compatibility note: As many C implementations push function arguments
17265 onto the stack in reverse order, the new library version is likely to
17266 interoperate with programs that have been compiled with the old
17267 pem_password_cb definition (PEM_whatever takes some data that
17268 happens to be on the stack as its last argument, and the callback
17269 just ignores this garbage); but there is no guarantee whatsoever that
17272 * The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
17273 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
17274 problems not only on Windows, but also on some Unix platforms.
17275 To avoid problematic command lines, these definitions are now in an
17276 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
17277 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
17281 * MIPS III/IV assembler module is reimplemented.
17285 * More DES library cleanups: remove references to srand/rand and
17286 delete an unused file.
17290 * Add support for the free Netwide assembler (NASM) under Win32,
17291 since not many people have MASM (ml) and it can be hard to obtain.
17292 This is currently experimental but it seems to work OK and pass all
17293 the tests. Check out INSTALL.W32 for info.
17297 * Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
17298 without temporary keys kept an extra copy of the server key,
17299 and connections with temporary keys did not free everything in case
17304 * New function RSA_check_key and new openssl rsa option -check
17305 for verifying the consistency of RSA keys.
17307 *Ulf Moeller, Bodo Moeller*
17309 * Various changes to make Win32 compile work:
17310 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
17311 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
17312 comparison" warnings.
17313 3. Add `sk_<TYPE>_sort` to DEF file generator and do make update.
17317 * Add a debugging option to PKCS#5 v2 key generation function: when
17318 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
17319 derived keys are printed to stderr.
17323 * Copy the flags in ASN1_STRING_dup().
17325 *Roman E. Pavlov <pre@mo.msk.ru>*
17327 * The x509 application mishandled signing requests containing DSA
17328 keys when the signing key was also DSA and the parameters didn't match.
17330 It was supposed to omit the parameters when they matched the signing key:
17331 the verifying software was then supposed to automatically use the CA's
17332 parameters if they were absent from the end user certificate.
17334 Omitting parameters is no longer recommended. The test was also
17335 the wrong way round! This was probably due to unusual behaviour in
17336 EVP_cmp_parameters() which returns 1 if the parameters match.
17337 This meant that parameters were omitted when they *didn't* match and
17338 the certificate was useless. Certificates signed with 'ca' didn't have
17341 *Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>*
17343 * Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
17344 The interface is as follows:
17345 Applications can use
17346 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
17347 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
17348 "off" is now the default.
17349 The library internally uses
17350 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
17351 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
17352 to disable memory-checking temporarily.
17354 Some inconsistent states that previously were possible (and were
17355 even the default) are now avoided.
17357 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
17358 with each memory chunk allocated; this is occasionally more helpful
17359 than just having a counter.
17361 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
17363 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
17368 * Introduce "mode" for SSL structures (with defaults in SSL_CTX),
17369 which largely parallels "options", but is for changing API behaviour,
17370 whereas "options" are about protocol behaviour.
17371 Initial "mode" flags are:
17373 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
17374 a single record has been written.
17375 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
17376 retries use the same buffer location.
17377 (But all of the contents must be
17382 * Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
17385 * Fix problems with no-hmac etc.
17387 *Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>*
17389 * New functions RSA_get_default_method(), RSA_set_method() and
17390 RSA_get_method(). These allows replacement of RSA_METHODs without having
17391 to mess around with the internals of an RSA structure.
17395 * Fix memory leaks in DSA_do_sign and DSA_is_prime.
17396 Also really enable memory leak checks in openssl.c and in some
17399 *Chad C. Mulligan, Bodo Moeller*
17401 * Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
17402 up the length of negative integers. This has now been simplified to just
17403 store the length when it is first determined and use it later, rather
17404 than trying to keep track of where data is copied and updating it to
17406 *Steve Henson, reported by Brien Wheeler <bwheeler@authentica-security.com>*
17408 * Add a new function PKCS7_signatureVerify. This allows the verification
17409 of a PKCS#7 signature but with the signing certificate passed to the
17410 function itself. This contrasts with PKCS7_dataVerify which assumes the
17411 certificate is present in the PKCS#7 structure. This isn't always the
17412 case: certificates can be omitted from a PKCS#7 structure and be
17413 distributed by "out of band" means (such as a certificate database).
17417 * Complete the `PEM_*` macros with DECLARE_PEM versions to replace the
17418 function prototypes in pem.h, also change util/mkdef.pl to add the
17419 necessary function names.
17423 * mk1mf.pl (used by Windows builds) did not properly read the
17424 options set by Configure in the top level Makefile, and Configure
17425 was not even able to write more than one option correctly.
17426 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
17430 * New functions CONF_load_bio() and CONF_load_fp() to allow a config
17431 file to be loaded from a BIO or FILE pointer. The BIO version will
17432 for example allow memory BIOs to contain config info.
17436 * New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
17437 Whoever hopes to achieve shared-library compatibility across versions
17438 must use this, not the compile-time macro.
17439 (Exercise 0.9.4: Which is the minimum library version required by
17441 Note: All this applies only to multi-threaded programs, others don't
17446 * Add missing case to s3_clnt.c state machine -- one of the new SSL tests
17447 through a BIO pair triggered the default case, i.e.
17448 SSLerr(...,SSL_R_UNKNOWN_STATE).
17452 * New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
17453 can use the SSL library even if none of the specific BIOs is
17458 * Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
17459 for the encoded length.
17461 *Jeon KyoungHo <khjeon@sds.samsung.co.kr>*
17463 * Add initial documentation of the X509V3 functions.
17467 * Add a new pair of functions PEM_write_PKCS8PrivateKey() and
17468 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
17469 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
17470 secure PKCS#8 private key format with a high iteration count.
17474 * Fix determination of Perl interpreter: A perl or perl5
17475 *directory* in $PATH was also accepted as the interpreter.
17477 *Ralf S. Engelschall*
17479 * Fix demos/sign/sign.c: well there wasn't anything strictly speaking
17480 wrong with it but it was very old and did things like calling
17481 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
17482 unusual formatting.
17486 * Fix demos/selfsign.c: it used obsolete and deleted functions, changed
17487 to use the new extension code.
17491 * Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
17492 with macros. This should make it easier to change their form, add extra
17493 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
17498 * Add to configuration table a new entry that can specify an alternative
17499 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
17500 according to Mark Crispin <MRC@Panda.COM>.
17504 * DES CBC did not update the IV. Weird.
17508 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
17509 Changing the behaviour of the former might break existing programs --
17510 where IV updating is needed, des_ncbc_encrypt can be used.
17513 * When bntest is run from "make test" it drives bc to check its
17514 calculations, as well as internally checking them. If an internal check
17515 fails, it needs to cause bc to give a non-zero result or make test carries
17516 on without noticing the failure. Fixed.
17520 * DES library cleanups.
17524 * Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
17525 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
17526 ciphers. NOTE: although the key derivation function has been verified
17527 against some published test vectors it has not been extensively tested
17528 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
17533 * Instead of "mkdir -p", which is not fully portable, use new
17534 Perl script "util/mkdir-p.pl".
17538 * Rewrite the way password based encryption (PBE) is handled. It used to
17539 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
17540 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
17541 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
17542 the 'parameter' field of the AlgorithmIdentifier is passed to the
17543 underlying key generation function so it must do its own ASN1 parsing.
17544 This has also changed the EVP_PBE_CipherInit() function which now has a
17545 'parameter' argument instead of literal salt and iteration count values
17546 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
17550 * Support for PKCS#5 v1.5 compatible password based encryption algorithms
17551 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
17552 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
17553 KEY" because this clashed with PKCS#8 unencrypted string. Since this
17554 value was just used as a "magic string" and not used directly its
17555 value doesn't matter.
17559 * Introduce some semblance of const correctness to BN. Shame C doesn't
17564 * "linux-sparc64" configuration (ultrapenguin).
17566 *Ray Miller <ray.miller@oucs.ox.ac.uk>*
17567 "linux-sparc" configuration.
17569 *Christian Forster <fo@hawo.stw.uni-erlangen.de>*
17571 * config now generates no-xxx options for missing ciphers.
17575 * Support the EBCDIC character set (work in progress).
17576 File ebcdic.c not yet included because it has a different license.
17578 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
17580 * Support BS2000/OSD-POSIX.
17582 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
17584 * Make callbacks for key generation use `void *` instead of `char *`.
17588 * Make S/MIME samples compile (not yet tested).
17592 * Additional typesafe stacks.
17596 * New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
17600 ### Changes between 0.9.3 and 0.9.3a [29 May 1999]
17602 * New configuration variant "sco5-gcc".
17604 * Updated some demos.
17606 *Sean O Riordain, Wade Scholine*
17608 * Add missing BIO_free at exit of pkcs12 application.
17612 * Fix memory leak in conf.c.
17616 * Updates for Win32 to assembler version of MD5.
17620 * Set #! path to perl in `apps/der_chop` to where we found it
17621 instead of using a fixed path.
17625 * SHA library changes for irix64-mips4-cc.
17629 * Improvements for VMS support.
17633 ### Changes between 0.9.2b and 0.9.3 [24 May 1999]
17635 * Bignum library bug fix. IRIX 6 passes "make test" now!
17636 This also avoids the problems with SC4.2 and unpatched SC5.
17638 *Andy Polyakov <appro@fy.chalmers.se>*
17640 * New functions sk_num, sk_value and sk_set to replace the previous macros.
17641 These are required because of the typesafe stack would otherwise break
17642 existing code. If old code used a structure member which used to be STACK
17643 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
17644 sk_num or sk_value it would produce an error because the num, data members
17645 are not present in STACK_OF. Now it just produces a warning. sk_set
17646 replaces the old method of assigning a value to sk_value
17647 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
17648 that does this will no longer work (and should use sk_set instead) but
17649 this could be regarded as a "questionable" behaviour anyway.
17653 * Fix most of the other PKCS#7 bugs. The "experimental" code can now
17654 correctly handle encrypted S/MIME data.
17658 * Change type of various DES function arguments from des_cblock
17659 (which means, in function argument declarations, pointer to char)
17660 to des_cblock * (meaning pointer to array with 8 char elements),
17661 which allows the compiler to do more typechecking; it was like
17662 that back in SSLeay, but with lots of ugly casts.
17664 Introduce new type const_des_cblock.
17668 * Reorganise the PKCS#7 library and get rid of some of the more obvious
17669 problems: find RecipientInfo structure that matches recipient certificate
17670 and initialise the ASN1 structures properly based on passed cipher.
17674 * Belatedly make the BN tests actually check the results.
17678 * Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
17679 to and from BNs: it was completely broken. New compilation option
17680 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
17681 key elements as negative integers.
17685 * Reorganize and speed up MD5.
17687 *Andy Polyakov <appro@fy.chalmers.se>*
17691 *Richard Levitte <richard@levitte.org>*
17693 * New option -out to asn1parse to allow the parsed structure to be
17694 output to a file. This is most useful when combined with the -strparse
17695 option to examine the output of things like OCTET STRINGS.
17699 * Make SSL library a little more fool-proof by not requiring any longer
17700 that `SSL_set_{accept,connect}_state` be called before
17701 `SSL_{accept,connect}` may be used (`SSL_set_..._state` is omitted
17702 in many applications because usually everything *appeared* to work as
17703 intended anyway -- now it really works as intended).
17707 * Move openssl.cnf out of lib/.
17711 * Fix various things to let OpenSSL even pass "egcc -pipe -O2 -Wall
17712 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
17713 -Wmissing-declarations -Wnested-externs -Winline" with EGCS 1.1.2+
17715 *Ralf S. Engelschall*
17717 * Various fixes to the EVP and PKCS#7 code. It may now be able to
17718 handle PKCS#7 enveloped data properly.
17720 *Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve*
17722 * Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
17723 copying pointers. The cert_st handling is changed by this in
17724 various ways (and thus what used to be known as ctx->default_cert
17725 is now called ctx->cert, since we don't resort to `s->ctx->[default_]cert`
17726 any longer when s->cert does not give us what we need).
17727 ssl_cert_instantiate becomes obsolete by this change.
17728 As soon as we've got the new code right (possibly it already is?),
17729 we have solved a couple of bugs of the earlier code where s->cert
17730 was used as if it could not have been shared with other SSL structures.
17732 Note that using the SSL API in certain dirty ways now will result
17733 in different behaviour than observed with earlier library versions:
17734 Changing settings for an `SSL_CTX *ctx` after having done s = SSL_new(ctx)
17735 does not influence s as it used to.
17737 In order to clean up things more thoroughly, inside SSL_SESSION
17738 we don't use CERT any longer, but a new structure SESS_CERT
17739 that holds per-session data (if available); currently, this is
17740 the peer's certificate chain and, for clients, the server's certificate
17741 and temporary key. CERT holds only those values that can have
17742 meaningful defaults in an SSL_CTX.
17746 * New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
17747 from the internal representation. Various PKCS#7 fixes: remove some
17748 evil casts and set the enc_dig_alg field properly based on the signing
17753 * Allow PKCS#12 password to be set from the command line or the
17754 environment. Let 'ca' get its config file name from the environment
17755 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
17760 * Allow certificate policies extension to use an IA5STRING for the
17761 organization field. This is contrary to the PKIX definition but
17762 VeriSign uses it and IE5 only recognises this form. Document 'x509'
17767 * Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
17768 without disallowing inline assembler and the like for non-pedantic builds.
17772 * Support Borland C++ builder.
17774 *Janez Jere <jj@void.si>, modified by Ulf Möller*
17780 * SHA-1 cleanups and performance enhancements.
17782 *Andy Polyakov <appro@fy.chalmers.se>*
17784 * Sparc v8plus assembler for the bignum library.
17786 *Andy Polyakov <appro@fy.chalmers.se>*
17788 * Accept any -xxx and +xxx compiler options in Configure.
17792 * Update HPUX configuration.
17796 * Add missing `sk_<type>_unshift()` function to safestack.h
17798 *Ralf S. Engelschall*
17800 * New function SSL_CTX_use_certificate_chain_file that sets the
17801 "extra_cert"s in addition to the certificate. (This makes sense
17802 only for "PEM" format files, as chains as a whole are not
17807 * Support verify_depth from the SSL API.
17808 x509_vfy.c had what can be considered an off-by-one-error:
17809 Its depth (which was not part of the external interface)
17810 was actually counting the number of certificates in a chain;
17811 now it really counts the depth.
17815 * Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
17816 instead of X509err, which often resulted in confusing error
17817 messages since the error codes are not globally unique
17818 (e.g. an alleged error in ssl3_accept when a certificate
17819 didn't match the private key).
17821 * New function SSL_CTX_set_session_id_context that allows to set a default
17822 value (so that you don't need SSL_set_session_id_context for each
17823 connection using the SSL_CTX).
17827 * OAEP decoding bug fix.
17831 * Support INSTALL_PREFIX for package builders, as proposed by
17836 * New Configure options "threads" and "no-threads". For systems
17837 where the proper compiler options are known (currently Solaris
17838 and Linux), "threads" is the default.
17842 * New script util/mklink.pl as a faster substitute for util/mklink.sh.
17846 * Install various scripts to $(OPENSSLDIR)/misc, not to
17847 $(INSTALLTOP)/bin -- they shouldn't clutter directories
17848 such as /usr/local/bin.
17852 * "make linux-shared" to build shared libraries.
17854 *Niels Poppe <niels@netbox.org>*
17856 * New Configure option `no-<cipher>` (rsa, idea, rc5, ...).
17860 * Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
17861 extension adding in x509 utility.
17865 * Remove NOPROTO sections and error code comments.
17869 * Partial rewrite of the DEF file generator to now parse the ANSI
17874 * New Configure options --prefix=DIR and --openssldir=DIR.
17878 * Complete rewrite of the error code script(s). It is all now handled
17879 by one script at the top level which handles error code gathering,
17880 header rewriting and C source file generation. It should be much better
17881 than the old method: it now uses a modified version of Ulf's parser to
17882 read the ANSI prototypes in all header files (thus the old K&R definitions
17883 aren't needed for error creation any more) and do a better job of
17884 translating function codes into names. The old 'ASN1 error code embedded
17885 in a comment' is no longer necessary and it doesn't use .err files which
17886 have now been deleted. Also the error code call doesn't have to appear all
17887 on one line (which resulted in some large lines...).
17891 * Change #include filenames from `<foo.h>` to `<openssl/foo.h>`.
17895 * Change behaviour of ssl2_read when facing length-0 packets: Don't return
17896 0 (which usually indicates a closed connection), but continue reading.
17900 * Fix some race conditions.
17904 * Add support for CRL distribution points extension. Add Certificate
17905 Policies and CRL distribution points documentation.
17909 * Move the autogenerated header file parts to crypto/opensslconf.h.
17913 * Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
17914 8 of keying material. Merlin has also confirmed interop with this fix
17915 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
17917 *Merlin Hughes <merlin@baltimore.ie>*
17919 * Fix lots of warnings.
17921 *Richard Levitte <levitte@stacken.kth.se>*
17923 * In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
17924 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
17926 *Richard Levitte <levitte@stacken.kth.se>*
17928 * Fix problems with sizeof(long) == 8.
17930 *Andy Polyakov <appro@fy.chalmers.se>*
17932 * Change functions to ANSI C.
17936 * Fix typos in error codes.
17938 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller*
17940 * Remove defunct assembler files from Configure.
17944 * SPARC v8 assembler BIGNUM implementation.
17946 *Andy Polyakov <appro@fy.chalmers.se>*
17948 * Support for Certificate Policies extension: both print and set.
17949 Various additions to support the r2i method this uses.
17953 * A lot of constification, and fix a bug in X509_NAME_oneline() that could
17954 return a const string when you are expecting an allocated buffer.
17958 * Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
17959 types DirectoryString and DisplayText.
17963 * Add code to allow r2i extensions to access the configuration database,
17964 add an LHASH database driver and add several ctx helper functions.
17968 * Fix an evil bug in bn_expand2() which caused various BN functions to
17969 fail when they extended the size of a BIGNUM.
17973 * Various utility functions to handle SXNet extension. Modify mkdef.pl to
17974 support typesafe stack.
17978 * Fix typo in SSL_[gs]et_options().
17980 *Nils Frostberg <nils@medcom.se>*
17982 * Delete various functions and files that belonged to the (now obsolete)
17983 old X509V3 handling code.
17987 * New Configure option "rsaref".
17991 * Don't auto-generate pem.h.
17995 * Introduce type-safe ASN.1 SETs.
17999 * Convert various additional casted stacks to type-safe STACK_OF() variants.
18001 *Ben Laurie, Ralf S. Engelschall, Steve Henson*
18003 * Introduce type-safe STACKs. This will almost certainly break lots of code
18004 that links with OpenSSL (well at least cause lots of warnings), but fear
18005 not: the conversion is trivial, and it eliminates loads of evil casts. A
18006 few STACKed things have been converted already. Feel free to convert more.
18007 In the fullness of time, I'll do away with the STACK type altogether.
18011 * Add `openssl ca -revoke <certfile>` facility which revokes a certificate
18012 specified in `<certfile>` by updating the entry in the index.txt file.
18013 This way one no longer has to edit the index.txt file manually for
18014 revoking a certificate. The -revoke option does the gory details now.
18016 *Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall*
18018 * Fix `openssl crl -noout -text` combination where `-noout` killed the
18019 `-text` option at all and this way the `-noout -text` combination was
18020 inconsistent in `openssl crl` with the friends in `openssl x509|rsa|dsa`.
18022 *Ralf S. Engelschall*
18024 * Make sure a corresponding plain text error message exists for the
18025 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
18026 verify callback function determined that a certificate was revoked.
18028 *Ralf S. Engelschall*
18030 * Bugfix: In test/testenc, don't test `openssl <cipher>` for
18031 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
18032 all available ciphers including rc5, which was forgotten until now.
18033 In order to let the testing shell script know which algorithms
18034 are available, a new (up to now undocumented) command
18035 `openssl list-cipher-commands` is used.
18039 * Bugfix: s_client occasionally would sleep in select() when
18040 it should have checked SSL_pending() first.
18044 * New functions DSA_do_sign and DSA_do_verify to provide access to
18045 the raw DSA values prior to ASN.1 encoding.
18049 * Tweaks to Configure
18051 *Niels Poppe <niels@netbox.org>*
18053 * Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
18058 * New variables $(RANLIB) and $(PERL) in the Makefiles.
18062 * New config option to avoid instructions that are illegal on the 80386.
18063 The default code is faster, but requires at least a 486.
18067 * Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
18068 SSL2_SERVER_VERSION (not used at all) macros, which are now the
18069 same as SSL2_VERSION anyway.
18073 * New "-showcerts" option for s_client.
18077 * Still more PKCS#12 integration. Add pkcs12 application to openssl
18078 application. Various cleanups and fixes.
18082 * More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
18083 modify error routines to work internally. Add error codes and PBE init
18084 to library startup routines.
18088 * Further PKCS#12 integration. Added password based encryption, PKCS#8 and
18089 packing functions to asn1 and evp. Changed function names and error
18090 codes along the way.
18094 * PKCS12 integration: and so it begins... First of several patches to
18095 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
18096 objects to objects.h
18100 * Add a new 'indent' option to some X509V3 extension code. Initial ASN1
18101 and display support for Thawte strong extranet extension.
18105 * Add LinuxPPC support.
18107 *Jeff Dubrule <igor@pobox.org>*
18109 * Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
18110 bn_div_words in alpha.s.
18112 *Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie*
18114 * Make sure the RSA OAEP test is skipped under -DRSAref because
18115 OAEP isn't supported when OpenSSL is built with RSAref.
18117 *Ulf Moeller <ulf@fitug.de>*
18119 * Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
18120 so they no longer are missing under -DNOPROTO.
18122 *Soren S. Jorvang <soren@t.dk>*
18124 ### Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
18126 * Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
18127 doesn't work when the session is reused. Coming soon!
18131 * Fix a security hole, that allows sessions to be reused in the wrong
18132 context thus bypassing client cert protection! All software that uses
18133 client certs and session caches in multiple contexts NEEDS PATCHING to
18134 allow session reuse! A fuller solution is in the works.
18136 *Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)*
18138 * Some more source tree cleanups (removed obsolete files
18139 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
18140 permission on "config" script to be executable) and a fix for the INSTALL
18143 *Ulf Moeller <ulf@fitug.de>*
18145 * Remove some legacy and erroneous uses of malloc, free instead of
18148 *Lennart Bang <lob@netstream.se>, with minor changes by Steve*
18150 * Make rsa_oaep_test return non-zero on error.
18152 *Ulf Moeller <ulf@fitug.de>*
18154 * Add support for native Solaris shared libraries. Configure
18155 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
18156 if someone would make that last step automatic.
18158 *Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>*
18160 * ctx_size was not built with the right compiler during "make links". Fixed.
18164 * Change the meaning of 'ALL' in the cipher list. It now means "everything
18165 except NULL ciphers". This means the default cipher list will no longer
18166 enable NULL ciphers. They need to be specifically enabled e.g. with
18167 the string "DEFAULT:eNULL".
18171 * Fix to RSA private encryption routines: if p < q then it would
18172 occasionally produce an invalid result. This will only happen with
18173 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
18177 * Be less restrictive and allow also `perl util/perlpath.pl
18178 /path/to/bin/perl` in addition to `perl util/perlpath.pl /path/to/bin`,
18179 because this way one can also use an interpreter named `perl5` (which is
18180 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
18181 installed as `perl`).
18183 *Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
18185 * Let util/clean-depend.pl work also with older Perl 5.00x versions.
18187 *Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
18189 * Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
18190 advapi32.lib to Win32 build and change the pem test comparison
18191 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
18192 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
18193 and crypto/des/ede_cbcm_enc.c.
18197 * DES quad checksum was broken on big-endian architectures. Fixed.
18201 * Comment out two functions in bio.h that aren't implemented. Fix up the
18202 Win32 test batch file so it (might) work again. The Win32 test batch file
18203 is horrible: I feel ill....
18207 * Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
18208 in e_os.h. Audit of header files to check ANSI and non ANSI
18209 sections: 10 functions were absent from non ANSI section and not exported
18210 from Windows DLLs. Fixed up libeay.num for new functions.
18214 * Make `openssl version` output lines consistent.
18216 *Ralf S. Engelschall*
18218 * Fix Win32 symbol export lists for BIO functions: Added
18219 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
18220 to ms/libeay{16,32}.def.
18222 *Ralf S. Engelschall*
18224 * Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
18225 fine under Unix and passes some trivial tests I've now added. But the
18226 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
18227 added to make sure no one expects that this stuff really works in the
18228 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
18229 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
18232 *Ralf S. Engelschall*
18234 * Fix the generation of two part addresses in perl.
18236 *Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie*
18238 * Add config entry for Linux on MIPS.
18240 *John Tobey <jtobey@channel1.com>*
18242 * Make links whenever Configure is run, unless we are on Windoze.
18246 * Permit extensions to be added to CRLs using crl_section in openssl.cnf.
18247 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
18252 * Add a useful kludge to allow package maintainers to specify compiler and
18253 other platforms details on the command line without having to patch the
18254 Configure script every time: One now can use
18255 `perl Configure <id>:<details>`,
18256 i.e. platform ids are allowed to have details appended
18257 to them (separated by colons). This is treated as there would be a static
18258 pre-configured entry in Configure's %table under key `<id>` with value
18259 `<details>` and `perl Configure <id>` is called. So, when you want to
18260 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
18261 assembler stuff you can use `perl Configure "FreeBSD-elf:pgcc:-O6:::"`
18262 now, which overrides the FreeBSD-elf entry on-the-fly.
18264 *Ralf S. Engelschall*
18266 * Disable new TLS1 ciphersuites by default: they aren't official yet.
18270 * Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
18271 on the `perl Configure ...` command line. This way one can compile
18272 OpenSSL libraries with Position Independent Code (PIC) which is needed
18273 for linking it into DSOs.
18275 *Ralf S. Engelschall*
18277 * Remarkably, export ciphers were totally broken and no-one had noticed!
18282 * Cleaned up the LICENSE document: The official contact for any license
18283 questions now is the OpenSSL core team under openssl-core@openssl.org.
18284 And add a paragraph about the dual-license situation to make sure people
18285 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
18286 to the OpenSSL toolkit.
18288 *Ralf S. Engelschall*
18290 * General source tree makefile cleanups: Made `making xxx in yyy...`
18291 display consistent in the source tree and replaced `/bin/rm` by `rm`.
18292 Additionally cleaned up the `make links` target: Remove unnecessary
18293 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
18294 to speed processing and no longer clutter the display with confusing
18295 stuff. Instead only the actually done links are displayed.
18297 *Ralf S. Engelschall*
18299 * Permit null encryption ciphersuites, used for authentication only. It used
18300 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
18301 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
18306 * Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
18307 signed attributes when verifying signatures (this would break them),
18308 the detached data encoding was wrong and public keys obtained using
18309 X509_get_pubkey() weren't freed.
18313 * Add text documentation for the BUFFER functions. Also added a work around
18314 to a Win95 console bug. This was triggered by the password read stuff: the
18315 last character typed gets carried over to the next fread(). If you were
18316 generating a new cert request using 'req' for example then the last
18317 character of the passphrase would be CR which would then enter the first
18322 * Added the new 'Includes OpenSSL Cryptography Software' button as
18323 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
18324 button and can be used by applications based on OpenSSL to show the
18325 relationship to the OpenSSL project.
18327 *Ralf S. Engelschall*
18329 * Remove confusing variables in function signatures in files
18330 ssl/ssl_lib.c and ssl/ssl.h.
18332 *Lennart Bong <lob@kulthea.stacken.kth.se>*
18334 * Don't install bss_file.c under PREFIX/include/
18336 *Lennart Bong <lob@kulthea.stacken.kth.se>*
18338 * Get the Win32 compile working again. Modify mkdef.pl so it can handle
18339 functions that return function pointers and has support for NT specific
18340 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
18341 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
18342 unsigned to signed types: this was killing the Win32 compile.
18346 * Add new certificate file to stack functions,
18347 SSL_add_dir_cert_subjects_to_stack() and
18348 SSL_add_file_cert_subjects_to_stack(). These largely supplant
18349 SSL_load_client_CA_file(), and can be used to add multiple certs easily
18350 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
18351 This means that Apache-SSL and similar packages don't have to mess around
18352 to add as many CAs as they want to the preferred list.
18356 * Experiment with doxygen documentation. Currently only partially applied to
18358 See <http://www.stack.nl/~dimitri/doxygen/index.html>, and run doxygen with
18359 openssl.doxy as the configuration file.
18363 * Get rid of remaining C++-style comments which strict C compilers hate.
18365 *Ralf S. Engelschall, pointed out by Carlos Amengual*
18367 * Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
18368 compiled in by default: it has problems with large keys.
18372 * Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
18373 DH private keys and/or callback functions which directly correspond to
18374 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
18375 is needed for applications which have to configure certificates on a
18376 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
18378 For the RSA certificate situation is makes no difference, but
18379 for the DSA certificate situation this fixes the "no shared cipher"
18380 problem where the OpenSSL cipher selection procedure failed because the
18381 temporary keys were not overtaken from the context and the API provided
18382 no way to reconfigure them.
18383 The new functions now let applications reconfigure the stuff and they
18384 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
18385 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
18386 non-public-API function ssl_cert_instantiate() is used as a helper
18387 function and also to reduce code redundancy inside ssl_rsa.c.
18389 *Ralf S. Engelschall*
18391 * Move s_server -dcert and -dkey options out of the undocumented feature
18392 area because they are useful for the DSA situation and should be
18393 recognized by the users.
18395 *Ralf S. Engelschall*
18397 * Fix the cipher decision scheme for export ciphers: the export bits are
18398 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
18399 SSL_EXP_MASK. So, the original variable has to be used instead of the
18400 already masked variable.
18402 *Richard Levitte <levitte@stacken.kth.se>*
18404 * Fix `port` variable from `int` to `unsigned int` in crypto/bio/b_sock.c
18406 *Richard Levitte <levitte@stacken.kth.se>*
18408 * Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
18409 from `int` to `unsigned int` because it is a length and initialized by
18410 EVP_DigestFinal() which expects an `unsigned int *`.
18412 *Richard Levitte <levitte@stacken.kth.se>*
18414 * Don't hard-code path to Perl interpreter on shebang line of Configure
18415 script. Instead use the usual Shell->Perl transition trick.
18417 *Ralf S. Engelschall*
18419 * Make `openssl x509 -noout -modulus`' functional also for DSA certificates
18420 (in addition to RSA certificates) to match the behaviour of `openssl dsa
18421 -noout -modulus` as it's already the case for `openssl rsa -noout
18422 -modulus`. For RSA the -modulus is the real "modulus" while for DSA
18423 currently the public key is printed (a decision which was already done by
18424 `openssl dsa -modulus` in the past) which serves a similar purpose.
18425 Additionally the NO_RSA no longer completely removes the whole -modulus
18426 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
18429 *Ralf S. Engelschall*
18431 * Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
18432 BIO. See the source (crypto/evp/bio_ok.c) for more info.
18434 *Arne Ansper <arne@ats.cyber.ee>*
18436 * Dump the old yucky req code that tried (and failed) to allow raw OIDs
18437 to be added. Now both 'req' and 'ca' can use new objects defined in the
18442 * Add cool BIO that does syslog (or event log on NT).
18444 *Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie*
18446 * Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
18447 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
18448 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
18449 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
18453 * Add preliminary config info for new extension code.
18457 * Make RSA_NO_PADDING really use no padding.
18459 *Ulf Moeller <ulf@fitug.de>*
18461 * Generate errors when private/public key check is done.
18465 * Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
18466 for some CRL extensions and new objects added.
18470 * Really fix the ASN1 IMPLICIT bug this time... Partial support for private
18471 key usage extension and fuller support for authority key id.
18475 * Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
18476 padding method for RSA, which is recommended for new applications in PKCS
18477 #1 v2.0 (RFC 2437, October 1998).
18478 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
18479 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
18480 against Bleichbacher's attack on RSA.
18481 *Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
18484 * Updates to the new SSL compression code
18486 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
18488 * Fix so that the version number in the master secret, when passed
18489 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
18490 (because the server will not accept higher), that the version number
18491 is 0x03,0x01, not 0x03,0x00
18493 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
18495 * Run extensive memory leak checks on SSL commands. Fixed *lots* of memory
18496 leaks in `ssl/` relating to new `X509_get_pubkey()` behaviour. Also fixes
18497 in `apps/` and an unrelated leak in `crypto/dsa/dsa_vrf.c`.
18501 * Support for RAW extensions where an arbitrary extension can be
18502 created by including its DER encoding. See `apps/openssl.cnf` for
18507 * Make sure latest Perl versions don't interpret some generated C array
18508 code as Perl array code in the crypto/err/err_genc.pl script.
18510 *Lars Weber <3weber@informatik.uni-hamburg.de>*
18512 * Modify ms/do_ms.bat to not generate assembly language makefiles since
18513 not many people have the assembler. Various Win32 compilation fixes and
18514 update to the INSTALL.W32 file with (hopefully) more accurate Win32
18515 build instructions.
18519 * Modify configure script 'Configure' to automatically create crypto/date.h
18520 file under Win32 and also build pem.h from pem.org. New script
18521 util/mkfiles.pl to create the MINFO file on environments that can't do a
18522 'make files': perl util/mkfiles.pl >MINFO should work.
18526 * Major rework of DES function declarations, in the pursuit of correctness
18527 and purity. As a result, many evil casts evaporated, and some weirdness,
18528 too. You may find this causes warnings in your code. Zapping your evil
18529 casts will probably fix them. Mostly.
18533 * Fix for a typo in asn1.h. Bug fix to object creation script
18534 obj_dat.pl. It considered a zero in an object definition to mean
18535 "end of object": none of the objects in objects.h have any zeros
18536 so it wasn't spotted.
18538 *Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>*
18540 * Add support for Triple DES Cipher Block Chaining with Output Feedback
18541 Masking (CBCM). In the absence of test vectors, the best I have been able
18542 to do is check that the decrypt undoes the encrypt, so far. Send me test
18543 vectors if you have them.
18547 * Correct calculation of key length for export ciphers (too much space was
18548 allocated for null ciphers). This has not been tested!
18552 * Modifications to the mkdef.pl for Win32 DEF file creation. The usage
18553 message is now correct (it understands "crypto" and "ssl" on its
18554 command line). There is also now an "update" option. This will update
18555 the util/ssleay.num and util/libeay.num files with any new functions.
18557 perl util/mkdef.pl crypto ssl update
18558 it will update them.
18562 * Overhauled the Perl interface:
18563 - ported BN stuff to OpenSSL's different BN library
18564 - made the perl/ source tree CVS-aware
18565 - renamed the package from SSLeay to OpenSSL (the files still contain
18566 their history because I've copied them in the repository)
18567 - removed obsolete files (the test scripts will be replaced
18568 by better Test::Harness variants in the future)
18570 *Ralf S. Engelschall*
18572 * First cut for a very conservative source tree cleanup:
18573 1. merge various obsolete readme texts into doc/ssleay.txt
18574 where we collect the old documents and readme texts.
18575 2. remove the first part of files where I'm already sure that we no
18576 longer need them because of three reasons: either they are just temporary
18577 files which were left by Eric or they are preserved original files where
18578 I've verified that the diff is also available in the CVS via "cvs diff
18579 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
18580 the crypto/md/ stuff).
18582 *Ralf S. Engelschall*
18584 * More extension code. Incomplete support for subject and issuer alt
18585 name, issuer and authority key id. Change the i2v function parameters
18586 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
18587 what that's for :-) Fix to ASN1 macro which messed up
18588 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
18592 * Preliminary support for ENUMERATED type. This is largely copied from the
18597 * Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
18599 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
18601 * Make sure `make rehash` target really finds the `openssl` program.
18603 *Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
18605 * Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
18606 like to hear about it if this slows down other processors.
18610 * Add CygWin32 platform information to Configure script.
18612 *Alan Batie <batie@aahz.jf.intel.com>*
18614 * Fixed ms/32all.bat script: `no_asm` -> `no-asm`
18616 *Rainer W. Gerling <gerling@mpg-gv.mpg.de>*
18618 * New program nseq to manipulate netscape certificate sequences
18622 * Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
18627 * Fixes to BN code. Previously the default was to define BN_RECURSION
18628 but the BN code had some problems that would cause failures when
18629 doing certificate verification and some other functions.
18631 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
18633 * Add ASN1 and PEM code to support netscape certificate sequences.
18637 * Add ASN1 and PEM code to support netscape certificate sequences.
18641 * Add several PKIX and private extended key usage OIDs.
18645 * Modify the 'ca' program to handle the new extension code. Modify
18646 openssl.cnf for new extension format, add comments.
18650 * More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
18651 and add a sample to openssl.cnf so req -x509 now adds appropriate
18656 * Continued X509 V3 changes. Add to other makefiles, integrate with the
18657 error code, add initial support to X509_print() and x509 application.
18661 * Takes a deep breath and start adding X509 V3 extension support code. Add
18662 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
18663 stuff is currently isolated and isn't even compiled yet.
18667 * Continuing patches for GeneralizedTime. Fix up certificate and CRL
18668 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
18669 Removed the versions check from X509 routines when loading extensions:
18670 this allows certain broken certificates that don't set the version
18671 properly to be processed.
18675 * Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
18676 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
18677 can still be regenerated with "make depend".
18681 * Spelling mistake in C version of CAST-128.
18683 *Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>*
18685 * Changes to the error generation code. The perl script err-code.pl
18686 now reads in the old error codes and retains the old numbers, only
18687 adding new ones if necessary. It also only changes the .err files if new
18688 codes are added. The makefiles have been modified to only insert errors
18689 when needed (to avoid needlessly modifying header files). This is done
18690 by only inserting errors if the .err file is newer than the auto generated
18691 C file. To rebuild all the error codes from scratch (the old behaviour)
18692 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
18693 or delete all the .err files.
18697 * CAST-128 was incorrectly implemented for short keys. The C version has
18698 been fixed, but is untested. The assembler versions are also fixed, but
18699 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
18700 to regenerate it if needed.
18701 *Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
18702 Hagino <itojun@kame.net>*
18704 * File was opened incorrectly in randfile.c.
18706 *Ulf Möller <ulf@fitug.de>*
18708 * Beginning of support for GeneralizedTime. d2i, i2d, check and print
18709 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
18710 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
18711 al: it's just almost always a UTCTime. Note this patch adds new error
18712 codes so do a "make errors" if there are problems.
18716 * Correct Linux 1 recognition in config.
18718 *Ulf Möller <ulf@fitug.de>*
18720 * Remove pointless MD5 hash when using DSA keys in ca.
18722 *Anonymous <nobody@replay.com>*
18724 * Generate an error if given an empty string as a cert directory. Also
18725 generate an error if handed NULL (previously returned 0 to indicate an
18726 error, but didn't set one).
18728 *Ben Laurie, reported by Anonymous <nobody@replay.com>*
18730 * Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
18734 * Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
18735 parameters. This was causing a warning which killed off the Win32 compile.
18739 * Remove C++ style comments from crypto/bn/bn_local.h.
18741 *Neil Costigan <neil.costigan@celocom.com>*
18743 * The function OBJ_txt2nid was broken. It was supposed to return a nid
18744 based on a text string, looking up short and long names and finally
18745 "dot" format. The "dot" format stuff didn't work. Added new function
18746 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
18747 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
18748 OID is not part of the table.
18752 * Add prototypes to X509 lookup/verify methods, fixing a bug in
18753 X509_LOOKUP_by_alias().
18757 * Sort openssl functions by name.
18761 * Get the `gendsa` command working and add it to the `list` command. Remove
18762 encryption from sample DSA keys (in case anyone is interested the password
18767 * Make *all* `*_free` functions accept a NULL pointer.
18769 *Frans Heymans <fheymans@isaserver.be>*
18771 * If a DH key is generated in s3_srvr.c, don't blow it by trying to use
18774 *Anonymous <nobody@replay.com>*
18776 * s_server should send the CAfile as acceptable CAs, not its own cert.
18778 *Bodo Moeller <3moeller@informatik.uni-hamburg.de>*
18780 * Don't blow it for numeric `-newkey` arguments to `apps/req`.
18782 *Bodo Moeller <3moeller@informatik.uni-hamburg.de>*
18784 * Temp key "for export" tests were wrong in s3_srvr.c.
18786 *Anonymous <nobody@replay.com>*
18788 * Add prototype for temp key callback functions
18789 SSL_CTX_set_tmp_{rsa,dh}_callback().
18793 * Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
18794 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
18798 * X509_name_add_entry() freed the wrong thing after an error.
18800 *Arne Ansper <arne@ats.cyber.ee>*
18802 * rsa_eay.c would attempt to free a NULL context.
18804 *Arne Ansper <arne@ats.cyber.ee>*
18806 * BIO_s_socket() had a broken should_retry() on Windoze.
18808 *Arne Ansper <arne@ats.cyber.ee>*
18810 * BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
18812 *Arne Ansper <arne@ats.cyber.ee>*
18814 * Make sure the already existing X509_STORE->depth variable is initialized
18815 in X509_STORE_new(), but document the fact that this variable is still
18816 unused in the certificate verification process.
18818 *Ralf S. Engelschall*
18820 * Fix the various library and `apps/` files to free up pkeys obtained from
18821 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
18825 * Fix reference counting in X509_PUBKEY_get(). This makes
18826 demos/maurice/example2.c work, amongst others, probably.
18828 *Steve Henson and Ben Laurie*
18830 * First cut of a cleanup for `apps/`. First the `ssleay` program is now named
18831 `openssl` and second, the shortcut symlinks for the `openssl <command>`
18832 are no longer created. This way we have a single and consistent command
18833 line interface `openssl <command>`, similar to `cvs <command>`.
18835 *Ralf S. Engelschall, Paul Sutton and Ben Laurie*
18837 * ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
18838 BIT STRING wrapper always have zero unused bits.
18842 * Add CA.pl, perl version of CA.sh, add extended key usage OID.
18846 * Make the top-level INSTALL documentation easier to understand.
18850 * Makefiles updated to exit if an error occurs in a sub-directory
18851 make (including if user presses ^C) [Paul Sutton]
18853 * Make Montgomery context stuff explicit in RSA data structure.
18857 * Fix build order of pem and err to allow for generated pem.h.
18861 * Fix renumbering bug in X509_NAME_delete_entry().
18865 * Enhanced the err-ins.pl script so it makes the error library number
18866 global and can add a library name. This is needed for external ASN1 and
18867 other error libraries.
18871 * Fixed sk_insert which never worked properly.
18875 * Fix ASN1 macros so they can handle indefinite length constructed
18876 EXPLICIT tags. Some non standard certificates use these: they can now
18881 * Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
18882 into a single doc/ssleay.txt bundle. This way the information is still
18883 preserved but no longer messes up this directory. Now it's new room for
18884 the new set of documentation files.
18886 *Ralf S. Engelschall*
18888 * SETs were incorrectly DER encoded. This was a major pain, because they
18889 shared code with SEQUENCEs, which aren't coded the same. This means that
18890 almost everything to do with SETs or SEQUENCEs has either changed name or
18891 number of arguments.
18893 *Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>*
18895 * Fix test data to work with the above.
18899 * Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
18900 was already fixed by Eric for 0.9.1 it seems.
18902 *Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>*
18904 * Autodetect FreeBSD3.
18908 * Fix various bugs in Configure. This affects the following platforms:
18912 unixware-2.0-pentium
18917 * Eliminate generated files from CVS. Reorder tests to regenerate files
18918 before they are needed.
18922 * Generate Makefile.ssl from Makefile.org (to keep CVS happy).
18926 ### Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
18928 * Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
18929 changed SSLeay to OpenSSL in version strings.
18931 *Ralf S. Engelschall*
18933 * Some fixups to the top-level documents.
18937 * Fixed the nasty bug where rsaref.h was not found under compile-time
18938 because the symlink to include/ was missing.
18940 *Ralf S. Engelschall*
18942 * Incorporated the popular no-RSA/DSA-only patches
18943 which allow to compile a RSA-free SSLeay.
18945 *Andrew Cooke / Interrader Ldt., Ralf S. Engelschall*
18947 * Fixed nasty rehash problem under `make -f Makefile.ssl links`
18948 when "ssleay" is still not found.
18950 *Ralf S. Engelschall*
18952 * Added more platforms to Configure: Cray T3E, HPUX 11,
18954 *Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>*
18956 * Updated the README file.
18958 *Ralf S. Engelschall*
18960 * Added various .cvsignore files in the CVS repository subdirs
18961 to make a "cvs update" really silent.
18963 *Ralf S. Engelschall*
18965 * Recompiled the error-definition header files and added
18966 missing symbols to the Win32 linker tables.
18968 *Ralf S. Engelschall*
18970 * Cleaned up the top-level documents;
18971 o new files: CHANGES and LICENSE
18972 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
18973 o merged COPYRIGHT into LICENSE
18974 o removed obsolete TODO file
18975 o renamed MICROSOFT to INSTALL.W32
18977 *Ralf S. Engelschall*
18979 * Removed dummy files from the 0.9.1b source tree:
18980 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
18981 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
18982 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
18983 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
18984 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
18986 *Ralf S. Engelschall*
18988 * Added various platform portability fixes.
18992 * The Genesis of the OpenSSL rpject:
18993 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
18994 Young and Tim J. Hudson created while they were working for C2Net until
18997 *The OpenSSL Project*
18999 ### Changes between 0.9.0b and 0.9.1b [not released]
19001 * Updated a few CA certificates under certs/
19005 * Changed some BIGNUM api stuff.
19009 * Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
19010 DGUX x86, Linux Alpha, etc.
19014 * New COMP library [crypto/comp/] for SSL Record Layer Compression:
19015 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
19020 * Add -strparse option to asn1pars program which parses nested
19023 *Dr Stephen Henson <shenson@bigfoot.com>*
19025 * Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
19029 * DSA fix for "ca" program.
19033 * Added "-genkey" option to "dsaparam" program.
19037 * Added RIPE MD160 (rmd160) message digest.
19041 * Added -a (all) option to "ssleay version" command.
19045 * Added PLATFORM define which is the id given to Configure.
19049 * Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
19053 * Extended the ASN.1 parser routines.
19057 * Extended BIO routines to support REUSEADDR, seek, tell, etc.
19061 * Added a BN_CTX to the BN library.
19065 * Fixed the weak key values in DES library
19069 * Changed API in EVP library for cipher aliases.
19073 * Added support for RC2/64bit cipher.
19077 * Converted the lhash library to the crypto/mem.c functions.
19081 * Added more recognized ASN.1 object ids.
19085 * Added more RSA padding checks for SSL/TLS.
19089 * Added BIO proxy/filter functionality.
19093 * Added extra_certs to SSL_CTX which can be used
19094 send extra CA certificates to the client in the CA cert chain sending
19095 process. It can be configured with SSL_CTX_add_extra_chain_cert().
19099 * Now Fortezza is denied in the authentication phase because
19100 this is key exchange mechanism is not supported by SSLeay at all.
19104 * Additional PKCS1 checks.
19108 * Support the string "TLSv1" for all TLS v1 ciphers.
19112 * Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
19113 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
19117 * Fixed a few memory leaks.
19121 * Fixed various code and comment typos.
19125 * A minor bug in ssl/s3_clnt.c where there would always be 4 0
19126 bytes sent in the client random.
19128 *Edward Bishop <ebishop@spyglass.com>*
19132 [CVE-2020-1971]: https://www.openssl.org/news/vulnerabilities.html#CVE-2020-1971
19133 [CVE-2020-1967]: https://www.openssl.org/news/vulnerabilities.html#CVE-2020-1967
19134 [CVE-2019-1563]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1563
19135 [CVE-2019-1559]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1559
19136 [CVE-2019-1552]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1552
19137 [CVE-2019-1551]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1551
19138 [CVE-2019-1549]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1549
19139 [CVE-2019-1547]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1547
19140 [CVE-2019-1543]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1543
19141 [CVE-2018-5407]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-5407
19142 [CVE-2018-0739]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0739
19143 [CVE-2018-0737]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0737
19144 [CVE-2018-0735]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0735
19145 [CVE-2018-0734]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0734
19146 [CVE-2018-0733]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0733
19147 [CVE-2018-0732]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0732
19148 [CVE-2017-3738]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3738
19149 [CVE-2017-3737]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3737
19150 [CVE-2017-3736]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3736
19151 [CVE-2017-3735]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3735
19152 [CVE-2017-3733]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3733
19153 [CVE-2017-3732]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3732
19154 [CVE-2017-3731]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3731
19155 [CVE-2017-3730]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3730
19156 [CVE-2016-7055]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7055
19157 [CVE-2016-7054]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7054
19158 [CVE-2016-7053]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7053
19159 [CVE-2016-7052]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7052
19160 [CVE-2016-6309]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6309
19161 [CVE-2016-6308]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6308
19162 [CVE-2016-6307]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6307
19163 [CVE-2016-6306]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6306
19164 [CVE-2016-6305]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6305
19165 [CVE-2016-6304]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6304
19166 [CVE-2016-6303]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6303
19167 [CVE-2016-6302]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6302
19168 [CVE-2016-2183]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2183
19169 [CVE-2016-2182]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2182
19170 [CVE-2016-2181]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2181
19171 [CVE-2016-2180]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2180
19172 [CVE-2016-2179]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2179
19173 [CVE-2016-2178]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2178
19174 [CVE-2016-2177]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2177
19175 [CVE-2016-2176]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2176
19176 [CVE-2016-2109]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2109
19177 [CVE-2016-2107]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2107
19178 [CVE-2016-2106]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2106
19179 [CVE-2016-2105]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2105
19180 [CVE-2016-0800]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0800
19181 [CVE-2016-0799]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0799
19182 [CVE-2016-0798]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0798
19183 [CVE-2016-0797]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0797
19184 [CVE-2016-0705]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0705
19185 [CVE-2016-0702]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0702
19186 [CVE-2016-0701]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0701
19187 [CVE-2015-3197]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3197
19188 [CVE-2015-3196]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3196
19189 [CVE-2015-3195]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3195
19190 [CVE-2015-3194]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3194
19191 [CVE-2015-3193]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3193
19192 [CVE-2015-1793]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1793
19193 [CVE-2015-1792]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1792
19194 [CVE-2015-1791]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1791
19195 [CVE-2015-1790]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1790
19196 [CVE-2015-1789]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1789
19197 [CVE-2015-1788]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1788
19198 [CVE-2015-1787]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1787
19199 [CVE-2015-0293]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0293
19200 [CVE-2015-0291]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0291
19201 [CVE-2015-0290]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0290
19202 [CVE-2015-0289]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0289
19203 [CVE-2015-0288]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0288
19204 [CVE-2015-0287]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0287
19205 [CVE-2015-0286]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0286
19206 [CVE-2015-0285]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0285
19207 [CVE-2015-0209]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0209
19208 [CVE-2015-0208]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0208
19209 [CVE-2015-0207]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0207
19210 [CVE-2015-0206]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0206
19211 [CVE-2015-0205]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0205
19212 [CVE-2015-0204]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0204
19213 [CVE-2014-8275]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-8275
19214 [CVE-2014-5139]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-5139
19215 [CVE-2014-3572]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3572
19216 [CVE-2014-3571]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3571
19217 [CVE-2014-3570]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3570
19218 [CVE-2014-3569]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3569
19219 [CVE-2014-3568]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3568
19220 [CVE-2014-3567]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3567
19221 [CVE-2014-3566]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3566
19222 [CVE-2014-3513]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3513
19223 [CVE-2014-3512]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3512
19224 [CVE-2014-3511]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3511
19225 [CVE-2014-3510]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3510
19226 [CVE-2014-3509]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3509
19227 [CVE-2014-3508]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3508
19228 [CVE-2014-3507]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3507
19229 [CVE-2014-3506]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3506
19230 [CVE-2014-3505]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3505
19231 [CVE-2014-3470]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3470
19232 [CVE-2014-0224]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0224
19233 [CVE-2014-0221]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0221
19234 [CVE-2014-0195]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0195
19235 [CVE-2014-0160]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0160
19236 [CVE-2014-0076]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0076
19237 [CVE-2013-6450]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-6450
19238 [CVE-2013-4353]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-4353
19239 [CVE-2013-0169]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-0169
19240 [CVE-2013-0166]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-0166
19241 [CVE-2012-2686]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2686
19242 [CVE-2012-2333]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2333
19243 [CVE-2012-2110]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2110
19244 [CVE-2012-0884]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0884
19245 [CVE-2012-0050]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0050
19246 [CVE-2012-0027]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0027
19247 [CVE-2011-4619]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4619
19248 [CVE-2011-4577]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4577
19249 [CVE-2011-4576]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4576
19250 [CVE-2011-4109]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4109
19251 [CVE-2011-4108]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4108
19252 [CVE-2011-3210]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-3210
19253 [CVE-2011-3207]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-3207
19254 [CVE-2011-0014]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-0014
19255 [CVE-2010-4252]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-4252
19256 [CVE-2010-4180]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-4180
19257 [CVE-2010-3864]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-3864
19258 [CVE-2010-1633]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-1633
19259 [CVE-2010-0740]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-0740
19260 [CVE-2010-0433]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-0433
19261 [CVE-2009-4355]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-4355
19262 [CVE-2009-3555]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-3555
19263 [CVE-2009-3245]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-3245
19264 [CVE-2009-1386]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1386
19265 [CVE-2009-1379]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1379
19266 [CVE-2009-1378]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1378
19267 [CVE-2009-1377]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1377
19268 [CVE-2009-0789]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0789
19269 [CVE-2009-0591]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0591
19270 [CVE-2009-0590]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0590
19271 [CVE-2008-5077]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-5077
19272 [CVE-2008-1678]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-1678
19273 [CVE-2008-1672]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-1672
19274 [CVE-2008-0891]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-0891
19275 [CVE-2007-5135]: https://www.openssl.org/news/vulnerabilities.html#CVE-2007-5135
19276 [CVE-2007-4995]: https://www.openssl.org/news/vulnerabilities.html#CVE-2007-4995
19277 [CVE-2006-4343]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-4343
19278 [CVE-2006-4339]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-4339
19279 [CVE-2006-3738]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-3738
19280 [CVE-2006-2940]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-2940
19281 [CVE-2006-2937]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-2937
19282 [CVE-2005-2969]: https://www.openssl.org/news/vulnerabilities.html#CVE-2005-2969
19283 [CVE-2004-0112]: https://www.openssl.org/news/vulnerabilities.html#CVE-2004-0112
19284 [CVE-2004-0079]: https://www.openssl.org/news/vulnerabilities.html#CVE-2004-0079
19285 [CVE-2003-0851]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0851
19286 [CVE-2003-0545]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0545
19287 [CVE-2003-0544]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0544
19288 [CVE-2003-0543]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0543
19289 [CVE-2003-0078]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0078
19290 [CVE-2002-0659]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0659
19291 [CVE-2002-0657]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0657
19292 [CVE-2002-0656]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0656
19293 [CVE-2002-0655]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0655