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 ciphersuites based on DHE_PSK (RFC 4279) and ECDHE_PSK (RFC 5489)
28 to the list of ciphersuites providing Perfect Forward Secrecy as
29 required by SECLEVEL >= 3.
31 *Dmitry Belyavskiy, Nicola Tuveri*
33 * Add new SSL APIs to aid in efficiently implementing TLS/SSL fingerprinting. The
34 SSL_CTRL_GET_IANA_GROUPS control code, exposed as the SSL_get0_iana_groups()
35 function-like macro, retrieves the list of supported groups sent by the peer,
36 and the function SSL_client_hello_get_extension_order() populates a caller-supplied
37 array with the list of extension types present in the ClientHello, in order of
42 * Fixed PEM_write_bio_PKCS8PrivateKey() and PEM_write_bio_PKCS8PrivateKey_nid()
43 to make it possible to use empty passphrase strings.
47 * RNDR and RNDRRS support in provider functions to provide
48 random number generation for Arm CPUs (aarch64).
52 * s_client and s_server apps now explicitly say when the TLS version
53 does not include the renegotiation mechanism. This avoids confusion
54 between that scenario versus when the TLS version includes secure
55 renegotiation but the peer lacks support for it.
59 * AES-GCM enabled with AVX512 vAES and vPCLMULQDQ.
61 *Tomasz Kantecki, Andrey Matyukov*
63 * The default SSL/TLS security level has been changed from 1 to 2. RSA,
64 DSA and DH keys of 1024 bits and above and less than 2048 bits and ECC keys
65 of 160 bits and above and less than 224 bits were previously accepted by
66 default but are now no longer allowed. By default TLS compression was
67 already disabled in previous OpenSSL versions. At security level 2 it cannot
72 * The SSL_CTX_set_cipher_list family functions now accept ciphers using their
77 * The PVK key derivation function has been moved from b2i_PVK_bio_ex() into
78 the legacy crypto provider as an EVP_KDF. Applications requiring this KDF
79 will need to load the legacy crypto provider.
83 * The various OBJ_* functions have been made thread safe.
87 * CCM8 cipher suites in TLS have been downgraded to security level zero
88 because they use a short authentication tag which lowers their strength.
92 * Subject or issuer names in X.509 objects are now displayed as UTF-8 strings
97 * Parallel dual-prime 1536/2048-bit modular exponentiation for
98 AVX512_IFMA capable processors.
100 *Sergey Kirillov, Andrey Matyukov (Intel Corp)*
105 For OpenSSL 3.0 a [Migration guide][] has been added, so the CHANGES entries
106 listed here are only a brief description.
107 The migration guide contains more detailed information related to new features,
108 breaking changes, and mappings for the large list of deprecated functions.
110 [Migration guide]: https://github.com/openssl/openssl/tree/master/doc/man7/migration_guide.pod
112 ### Changes between 3.0.1 and 3.0.2 [xx XXX xxxx]
114 * Made the AES constant time code for no-asm configurations
115 optional due to the resulting 95% performance degradation.
116 The AES constant time code can be enabled, for no assembly
117 builds, with: ./config no-asm -DOPENSSL_AES_CONST_TIME
121 ### Changes between 3.0.0 and 3.0.1 [14 dec 2021]
123 * Fixed invalid handling of X509_verify_cert() internal errors in libssl
124 Internally libssl in OpenSSL calls X509_verify_cert() on the client side to
125 verify a certificate supplied by a server. That function may return a
126 negative return value to indicate an internal error (for example out of
127 memory). Such a negative return value is mishandled by OpenSSL and will cause
128 an IO function (such as SSL_connect() or SSL_do_handshake()) to not indicate
129 success and a subsequent call to SSL_get_error() to return the value
130 SSL_ERROR_WANT_RETRY_VERIFY. This return value is only supposed to be
131 returned by OpenSSL if the application has previously called
132 SSL_CTX_set_cert_verify_callback(). Since most applications do not do this
133 the SSL_ERROR_WANT_RETRY_VERIFY return value from SSL_get_error() will be
134 totally unexpected and applications may not behave correctly as a result. The
135 exact behaviour will depend on the application but it could result in
136 crashes, infinite loops or other similar incorrect responses.
138 This issue is made more serious in combination with a separate bug in OpenSSL
139 3.0 that will cause X509_verify_cert() to indicate an internal error when
140 processing a certificate chain. This will occur where a certificate does not
141 include the Subject Alternative Name extension but where a Certificate
142 Authority has enforced name constraints. This issue can occur even with valid
148 * Corrected a few file name and file reference bugs in the build,
149 installation and setup scripts, which lead to installation verification
150 failures. Slightly enhanced the installation verification script.
154 * Fixed EVP_PKEY_eq() to make it possible to use it with strictly private
159 * Fixed PVK encoder to properly query for the passphrase.
163 * Multiple fixes in the OSSL_HTTP API functions.
167 * Allow sign extension in OSSL_PARAM_allocate_from_text() for the
168 OSSL_PARAM_INTEGER data type and return error on negative numbers
169 used with the OSSL_PARAM_UNSIGNED_INTEGER data type. Make
170 OSSL_PARAM_BLD_push_BN{,_pad}() return an error on negative numbers.
174 * Allow copying uninitialized digest contexts with EVP_MD_CTX_copy_ex.
178 * Fixed detection of ARMv7 and ARM64 CPU features on FreeBSD.
182 * Multiple threading fixes.
186 * Added NULL digest implementation to keep compatibility with 1.1.1 version.
190 * Allow fetching an operation from the provider that owns an unexportable key
191 as a fallback if that is still allowed by the property query.
195 ### Changes between 1.1.1 and 3.0.0 [7 sep 2021]
197 * TLS_MAX_VERSION, DTLS_MAX_VERSION and DTLS_MIN_VERSION constants are now
202 * The `OPENSSL_s390xcap` environment variable can be used to set bits in the
203 S390X capability vector to zero. This simplifies testing of different code
204 paths on S390X architecture.
208 * Encrypting more than 2^64 TLS records with AES-GCM is disallowed
209 as per FIPS 140-2 IG A.5 "Key/IV Pair Uniqueness Requirements from
210 SP 800-38D". The communication will fail at this point.
214 * The EC_GROUP_clear_free() function is deprecated as there is nothing
215 confidential in EC_GROUP data.
219 * The byte order mark (BOM) character is ignored if encountered at the
220 beginning of a PEM-formatted file.
224 * Added CMS support for the Russian GOST algorithms.
228 * Due to move of the implementation of cryptographic operations
229 to the providers, validation of various operation parameters can
230 be postponed until the actual operation is executed where previously
231 it happened immediately when an operation parameter was set.
233 For example when setting an unsupported curve with
234 EVP_PKEY_CTX_set_ec_paramgen_curve_nid() this function call will not
235 fail but later keygen operations with the EVP_PKEY_CTX will fail.
237 *OpenSSL team members and many third party contributors*
239 * The EVP_get_cipherbyname() function will return NULL for algorithms such as
240 "AES-128-SIV", "AES-128-CBC-CTS" and "CAMELLIA-128-CBC-CTS" which were
241 previously only accessible via low level interfaces. Use EVP_CIPHER_fetch()
242 instead to retrieve these algorithms from a provider.
246 * On build targets where the multilib postfix is set in the build
247 configuration the libdir directory was changing based on whether
248 the lib directory with the multilib postfix exists on the system
249 or not. This unpredictable behavior was removed and eventual
250 multilib postfix is now always added to the default libdir. Use
251 `--libdir=lib` to override the libdir if adding the postfix is
256 * The triple DES key wrap functionality now conforms to RFC 3217 but is
257 no longer interoperable with OpenSSL 1.1.1.
261 * The ERR_GET_FUNC() function was removed. With the loss of meaningful
262 function codes, this function can only cause problems for calling
267 * Add a configurable flag to output date formats as ISO 8601. Does not
268 change the default date format.
272 * Version of MSVC earlier than 1300 could get link warnings, which could
273 be suppressed if the undocumented -DI_CAN_LIVE_WITH_LNK4049 was set.
274 Support for this flag has been removed.
278 * Rework and make DEBUG macros consistent. Remove unused -DCONF_DEBUG,
279 -DBN_CTX_DEBUG, and REF_PRINT. Add a new tracing category and use it for
280 printing reference counts. Rename -DDEBUG_UNUSED to -DUNUSED_RESULT_DEBUG
281 Fix BN_DEBUG_RAND so it compiles and, when set, force DEBUG_RAND to be set
282 also. Rename engine_debug_ref to be ENGINE_REF_PRINT also for consistency.
286 * The signatures of the functions to get and set options on SSL and
287 SSL_CTX objects changed from "unsigned long" to "uint64_t" type.
288 Some source code changes may be required.
292 * The public definitions of conf_method_st and conf_st have been
293 deprecated. They will be made opaque in a future release.
295 *Rich Salz and Tomáš Mráz*
297 * Client-initiated renegotiation is disabled by default. To allow it, use
298 the -client_renegotiation option, the SSL_OP_ALLOW_CLIENT_RENEGOTIATION
299 flag, or the "ClientRenegotiation" config parameter as appropriate.
303 * Add "abspath" and "includedir" pragma's to config files, to prevent,
304 or modify relative pathname inclusion.
308 * OpenSSL includes a cryptographic module that is intended to be FIPS 140-2
309 validated. Please consult the README-FIPS and
310 README-PROVIDERS files, as well as the migration guide.
312 *OpenSSL team members and many third party contributors*
314 * For the key types DH and DHX the allowed settable parameters are now different.
318 * The openssl commands that read keys, certificates, and CRLs now
319 automatically detect the PEM or DER format of the input files.
321 *David von Oheimb, Richard Levitte, and Tomáš Mráz*
323 * Added enhanced PKCS#12 APIs which accept a library context.
327 * The default manual page suffix ($MANSUFFIX) has been changed to "ossl"
331 * Added support for Kernel TLS (KTLS).
333 *Boris Pismenny, John Baldwin and Andrew Gallatin*
335 * Support for RFC 5746 secure renegotiation is now required by default for
336 SSL or TLS connections to succeed.
340 * The signature of the `copy` functional parameter of the
341 EVP_PKEY_meth_set_copy() function has changed so its `src` argument is
342 now `const EVP_PKEY_CTX *` instead of `EVP_PKEY_CTX *`. Similarly
343 the signature of the `pub_decode` functional parameter of the
344 EVP_PKEY_asn1_set_public() function has changed so its `pub` argument is
345 now `const X509_PUBKEY *` instead of `X509_PUBKEY *`.
349 * The error return values from some control calls (ctrl) have changed.
353 * A public key check is now performed during EVP_PKEY_derive_set_peer().
357 * Many functions in the EVP_ namespace that are getters of values from
358 implementations or contexts were renamed to include get or get0 in their
359 names. Old names are provided as macro aliases for compatibility and
364 * The EVP_PKEY_CTRL_PKCS7_ENCRYPT, EVP_PKEY_CTRL_PKCS7_DECRYPT,
365 EVP_PKEY_CTRL_PKCS7_SIGN, EVP_PKEY_CTRL_CMS_ENCRYPT,
366 EVP_PKEY_CTRL_CMS_DECRYPT, and EVP_PKEY_CTRL_CMS_SIGN control operations
371 * The EVP_PKEY_public_check() and EVP_PKEY_param_check() functions now work for
374 * The output from the command line applications may have minor
379 * The output from numerous "printing" may have minor changes.
383 * Windows thread synchronization uses read/write primitives (SRWLock) when
384 supported by the OS, otherwise CriticalSection continues to be used.
388 * Add filter BIO BIO_f_readbuffer() that allows BIO_tell() and BIO_seek() to
389 work on read only BIO source/sinks that do not support these functions.
390 This allows piping or redirection of a file BIO using stdin to be buffered
391 into memory. This is used internally in OSSL_DECODER_from_bio().
395 * OSSL_STORE_INFO_get_type() may now return an additional value. In 1.1.1
396 this function would return one of the values OSSL_STORE_INFO_NAME,
397 OSSL_STORE_INFO_PKEY, OSSL_STORE_INFO_PARAMS, OSSL_STORE_INFO_CERT or
398 OSSL_STORE_INFO_CRL. Decoded public keys would previously have been reported
399 as type OSSL_STORE_INFO_PKEY in 1.1.1. In 3.0 decoded public keys are now
400 reported as having the new type OSSL_STORE_INFO_PUBKEY. Applications
401 using this function should be amended to handle the changed return value.
405 * Improved adherence to Enhanced Security Services (ESS, RFC 2634 and RFC 5035)
406 for the TSP and CMS Advanced Electronic Signatures (CAdES) implementations.
407 As required by RFC 5035 check both ESSCertID and ESSCertIDv2 if both present.
408 Correct the semantics of checking the validation chain in case ESSCertID{,v2}
409 contains more than one certificate identifier: This means that all
410 certificates referenced there MUST be part of the validation chain.
414 * The implementation of older EVP ciphers related to CAST, IDEA, SEED, RC2, RC4,
415 RC5, DESX and DES have been moved to the legacy provider.
419 * The implementation of the EVP digests MD2, MD4, MDC2, WHIRLPOOL and
420 RIPEMD-160 have been moved to the legacy provider.
424 * The deprecated function EVP_PKEY_get0() now returns NULL being called for a
429 * The deprecated functions EVP_PKEY_get0_RSA(),
430 EVP_PKEY_get0_DSA(), EVP_PKEY_get0_EC_KEY(), EVP_PKEY_get0_DH(),
431 EVP_PKEY_get0_hmac(), EVP_PKEY_get0_poly1305() and EVP_PKEY_get0_siphash() as
432 well as the similarly named "get1" functions behave differently in
437 * A number of functions handling low-level keys or engines were deprecated
438 including EVP_PKEY_set1_engine(), EVP_PKEY_get0_engine(), EVP_PKEY_assign(),
439 EVP_PKEY_get0(), EVP_PKEY_get0_hmac(), EVP_PKEY_get0_poly1305() and
440 EVP_PKEY_get0_siphash().
444 * PKCS#5 PBKDF1 key derivation has been moved from PKCS5_PBE_keyivgen() into
445 the legacy crypto provider as an EVP_KDF. Applications requiring this KDF
446 will need to load the legacy crypto provider. This includes these PBE
447 algorithms which use this KDF:
448 - NID_pbeWithMD2AndDES_CBC
449 - NID_pbeWithMD5AndDES_CBC
450 - NID_pbeWithSHA1AndRC2_CBC
451 - NID_pbeWithMD2AndRC2_CBC
452 - NID_pbeWithMD5AndRC2_CBC
453 - NID_pbeWithSHA1AndDES_CBC
457 * Deprecated obsolete BIO_set_callback(), BIO_get_callback(), and
458 BIO_debug_callback() functions.
462 * Deprecated obsolete EVP_PKEY_CTX_get0_dh_kdf_ukm() and
463 EVP_PKEY_CTX_get0_ecdh_kdf_ukm() functions.
467 * The RAND_METHOD APIs have been deprecated.
471 * The SRP APIs have been deprecated.
475 * Add a compile time option to prevent the caching of provider fetched
476 algorithms. This is enabled by including the no-cached-fetch option
477 at configuration time.
481 * pkcs12 now uses defaults of PBKDF2, AES and SHA-256, with a MAC iteration
482 count of PKCS12_DEFAULT_ITER.
484 *Tomáš Mráz and Sahana Prasad*
486 * The openssl speed command does not use low-level API calls anymore.
490 * Parallel dual-prime 1024-bit modular exponentiation for AVX512_IFMA
493 *Ilya Albrekht, Sergey Kirillov, Andrey Matyukov (Intel Corp)*
495 * Combining the Configure options no-ec and no-dh no longer disables TLSv1.3.
499 * Implemented support for fully "pluggable" TLSv1.3 groups. This means that
500 providers may supply their own group implementations (using either the "key
501 exchange" or the "key encapsulation" methods) which will automatically be
502 detected and used by libssl.
504 *Matt Caswell, Nicola Tuveri*
506 * The undocumented function X509_certificate_type() has been deprecated;
510 * Deprecated the obsolete BN_pseudo_rand() and BN_pseudo_rand_range().
514 * Removed RSA padding mode for SSLv23 (which was only used for
515 SSLv2). This includes the functions RSA_padding_check_SSLv23() and
516 RSA_padding_add_SSLv23() and the `-ssl` option in the deprecated
521 * Deprecated the obsolete X9.31 RSA key generation related functions.
523 * While a callback function set via `SSL_CTX_set_cert_verify_callback()`
524 is not allowed to return a value > 1, this is no more taken as failure.
526 *Viktor Dukhovni and David von Oheimb*
528 * Deprecated the obsolete X9.31 RSA key generation related functions
529 BN_X931_generate_Xpq(), BN_X931_derive_prime_ex(), and
530 BN_X931_generate_prime_ex().
534 * The default key generation method for the regular 2-prime RSA keys was
535 changed to the FIPS 186-4 B.3.6 method.
539 * Deprecated the BN_is_prime_ex() and BN_is_prime_fasttest_ex() functions.
543 * Deprecated EVP_MD_CTX_set_update_fn() and EVP_MD_CTX_update_fn().
547 * Deprecated the type OCSP_REQ_CTX and the functions OCSP_REQ_CTX_*() and
548 replaced with OSSL_HTTP_REQ_CTX and the functions OSSL_HTTP_REQ_CTX_*().
550 *Rich Salz, Richard Levitte, and David von Oheimb*
552 * Deprecated `X509_http_nbio()` and `X509_CRL_http_nbio()`.
556 * Deprecated `OCSP_parse_url()`.
560 * Validation of SM2 keys has been separated from the validation of regular EC
565 * Behavior of the `pkey` app is changed, when using the `-check` or `-pubcheck`
566 switches: a validation failure triggers an early exit, returning a failure
567 exit status to the parent process.
571 * Changed behavior of SSL_CTX_set_ciphersuites() and SSL_set_ciphersuites()
572 to ignore unknown ciphers.
576 * The `-cipher-commands` and `-digest-commands` options
577 of the command line utility `list` have been deprecated.
578 Instead use the `-cipher-algorithms` and `-digest-algorithms` options.
582 * Added convenience functions for generating asymmetric key pairs:
583 The 'quick' one-shot (yet somewhat limited) function L<EVP_PKEY_Q_keygen(3)>
584 and macros for the most common cases: <EVP_RSA_gen(3)> and L<EVP_EC_gen(3)>.
588 * All of the low level EC_KEY functions have been deprecated.
590 *Shane Lontis, Paul Dale, Richard Levitte, and Tomáš Mráz*
592 * Deprecated all the libcrypto and libssl error string loading
597 * The functions SSL_CTX_set_tmp_dh_callback and SSL_set_tmp_dh_callback, as
598 well as the macros SSL_CTX_set_tmp_dh() and SSL_set_tmp_dh() have been
603 * The `-crypt` option to the `passwd` command line tool has been removed.
607 * The -C option to the `x509`, `dhparam`, `dsaparam`, and `ecparam` commands
612 * Add support for AES Key Wrap inverse ciphers to the EVP layer.
616 * Deprecated EVP_PKEY_set1_tls_encodedpoint() and
617 EVP_PKEY_get1_tls_encodedpoint().
621 * The security callback, which can be customised by application code, supports
622 the security operation SSL_SECOP_TMP_DH. One location of the "other" parameter
623 was incorrectly passing a DH object. It now passed an EVP_PKEY in all cases.
627 * Add PKCS7_get_octet_string() and PKCS7_type_is_other() to the public
628 interface. Their functionality remains unchanged.
632 * Added new option for 'openssl list', '-providers', which will display the
633 list of loaded providers, their names, version and status. It optionally
634 displays their gettable parameters.
638 * Removed EVP_PKEY_set_alias_type().
642 * Deprecated `EVP_PKEY_CTX_set_rsa_keygen_pubexp()` and introduced
643 `EVP_PKEY_CTX_set1_rsa_keygen_pubexp()`, which is now preferred.
647 * Changed all "STACK" functions to be macros instead of inline functions. Macro
648 parameters are still checked for type safety at compile time via helper
653 * Remove the RAND_DRBG API
655 *Paul Dale and Matthias St. Pierre*
657 * Allow `SSL_set1_host()` and `SSL_add1_host()` to take IP literal addresses
658 as well as actual hostnames.
662 * The 'MinProtocol' and 'MaxProtocol' configuration commands now silently
663 ignore TLS protocol version bounds when configuring DTLS-based contexts, and
664 conversely, silently ignore DTLS protocol version bounds when configuring
665 TLS-based contexts. The commands can be repeated to set bounds of both
666 types. The same applies with the corresponding "min_protocol" and
667 "max_protocol" command-line switches, in case some application uses both TLS
670 SSL_CTX instances that are created for a fixed protocol version (e.g.
671 `TLSv1_server_method()`) also silently ignore version bounds. Previously
672 attempts to apply bounds to these protocol versions would result in an
673 error. Now only the "version-flexible" SSL_CTX instances are subject to
674 limits in configuration files in command-line options.
678 * Deprecated the `ENGINE` API. Engines should be replaced with providers
683 * Reworked the recorded ERR codes to make better space for system errors.
684 To distinguish them, the macro `ERR_SYSTEM_ERROR()` indicates if the
685 given code is a system error (true) or an OpenSSL error (false).
689 * Reworked the test perl framework to better allow parallel testing.
691 *Nicola Tuveri and David von Oheimb*
693 * Added ciphertext stealing algorithms AES-128-CBC-CTS, AES-192-CBC-CTS and
694 AES-256-CBC-CTS to the providers. CS1, CS2 and CS3 variants are supported.
698 * 'Configure' has been changed to figure out the configuration target if
699 none is given on the command line. Consequently, the 'config' script is
700 now only a mere wrapper. All documentation is changed to only mention
703 *Rich Salz and Richard Levitte*
705 * Added a library context `OSSL_LIB_CTX` that applications as well as
706 other libraries can use to form a separate context within which
707 libcrypto operations are performed.
711 * Added various `_ex` functions to the OpenSSL API that support using
712 a non-default `OSSL_LIB_CTX`.
716 * Handshake now fails if Extended Master Secret extension is dropped
721 * Dropped interactive mode from the `openssl` program.
725 * Deprecated `EVP_PKEY_cmp()` and `EVP_PKEY_cmp_parameters()`.
727 *David von Oheimb and Shane Lontis*
729 * Deprecated `EC_METHOD_get_field_type()`.
733 * Deprecated EC_GFp_simple_method(), EC_GFp_mont_method(),
734 EC_GF2m_simple_method(), EC_GFp_nist_method(), EC_GFp_nistp224_method()
735 EC_GFp_nistp256_method(), and EC_GFp_nistp521_method().
739 * Deprecated EC_GROUP_new(), EC_GROUP_method_of(), and EC_POINT_method_of().
743 * Add CAdES-BES signature verification support, mostly derived
744 from ESSCertIDv2 TS (RFC 5816) contribution by Marek Klein.
746 *Filipe Raimundo da Silva*
748 * Add CAdES-BES signature scheme and attributes support (RFC 5126) to CMS API.
752 * Added the AuthEnvelopedData content type structure (RFC 5083) with AES-GCM
753 parameter (RFC 5084) for the Cryptographic Message Syntax (CMS).
757 * Deprecated EC_POINT_make_affine() and EC_POINTs_make_affine().
761 * Deprecated EC_GROUP_precompute_mult(), EC_GROUP_have_precompute_mult(), and
762 EC_KEY_precompute_mult().
766 * Deprecated EC_POINTs_mul().
770 * Removed FIPS_mode() and FIPS_mode_set().
774 * The SSL option SSL_OP_IGNORE_UNEXPECTED_EOF is introduced.
778 * Deprecated EC_POINT_set_Jprojective_coordinates_GFp() and
779 EC_POINT_get_Jprojective_coordinates_GFp().
783 * Added OSSL_PARAM_BLD to the public interface. This allows OSSL_PARAM
784 arrays to be more easily constructed via a series of utility functions.
785 Create a parameter builder using OSSL_PARAM_BLD_new(), add parameters using
786 the various push functions and finally convert to a passable OSSL_PARAM
787 array using OSSL_PARAM_BLD_to_param().
791 * The security strength of SHA1 and MD5 based signatures in TLS has been
796 * Added EVP_PKEY_set_type_by_keymgmt(), to initialise an EVP_PKEY to
797 contain a provider side internal key.
801 * ASN1_verify(), ASN1_digest() and ASN1_sign() have been deprecated.
805 * Project text documents not yet having a proper file name extension
806 (`HACKING`, `LICENSE`, `NOTES*`, `README*`, `VERSION`) have been renamed to
807 `*.md` as far as reasonable, else `*.txt`, for better use with file managers.
811 * The main project documents (README, NEWS, CHANGES, INSTALL, SUPPORT)
812 have been converted to Markdown with the goal to produce documents
813 which not only look pretty when viewed online in the browser, but
814 remain well readable inside a plain text editor.
816 To achieve this goal, a 'minimalistic' Markdown style has been applied
817 which avoids formatting elements that interfere too much with the
818 reading flow in the text file. For example, it
820 * avoids [ATX headings][] and uses [setext headings][] instead
821 (which works for `<h1>` and `<h2>` headings only).
822 * avoids [inline links][] and uses [reference links][] instead.
823 * avoids [fenced code blocks][] and uses [indented code blocks][] instead.
825 [ATX headings]: https://github.github.com/gfm/#atx-headings
826 [setext headings]: https://github.github.com/gfm/#setext-headings
827 [inline links]: https://github.github.com/gfm/#inline-link
828 [reference links]: https://github.github.com/gfm/#reference-link
829 [fenced code blocks]: https://github.github.com/gfm/#fenced-code-blocks
830 [indented code blocks]: https://github.github.com/gfm/#indented-code-blocks
832 *Matthias St. Pierre*
834 * The test suite is changed to preserve results of each test recipe.
835 A new directory test-runs/ with subdirectories named like the
836 test recipes are created in the build tree for this purpose.
840 * Added an implementation of CMP and CRMF (RFC 4210, RFC 4211 RFC 6712).
841 This adds `crypto/cmp/`, `crpyto/crmf/`, `apps/cmp.c`, and `test/cmp_*`.
842 See L<openssl-cmp(1)> and L<OSSL_CMP_exec_IR_ses(3)> as starting points.
844 *David von Oheimb, Martin Peylo*
846 * Generalized the HTTP client code from `crypto/ocsp/` into `crpyto/http/`.
847 It supports arbitrary request and response content types, GET redirection,
848 TLS, connections via HTTP(S) proxies, connections and exchange via
849 user-defined BIOs (allowing implicit connections), persistent connections,
850 and timeout checks. See L<OSSL_HTTP_transfer(3)> etc. for details.
851 The legacy OCSP-focused (and only partly documented) API
852 is retained for backward compatibility, while most of it is deprecated.
856 * Added `util/check-format.pl`, a tool for checking adherence to the
857 OpenSSL coding style <https://www.openssl.org/policies/codingstyle.html>.
858 The checks performed are incomplete and yield some false positives.
859 Still the tool should be useful for detecting most typical glitches.
863 * `BIO_do_connect()` and `BIO_do_handshake()` have been extended:
864 If domain name resolution yields multiple IP addresses all of them are tried
865 after `connect()` failures.
869 * All of the low level RSA functions have been deprecated.
873 * X509 certificates signed using SHA1 are no longer allowed at security
878 * The command line utilities dhparam, dsa, gendsa and dsaparam have been
879 modified to use PKEY APIs. These commands are now in maintenance mode
880 and no new features will be added to them.
884 * The command line utility rsautl has been deprecated.
888 * The command line utilities genrsa and rsa have been modified to use PKEY
889 APIs. They now write PKCS#8 keys by default. These commands are now in
890 maintenance mode and no new features will be added to them.
894 * All of the low level DH functions have been deprecated.
896 *Paul Dale and Matt Caswell*
898 * All of the low level DSA functions have been deprecated.
902 * Reworked the treatment of EC EVP_PKEYs with the SM2 curve to
903 automatically become EVP_PKEY_SM2 rather than EVP_PKEY_EC.
907 * Deprecated low level ECDH and ECDSA functions.
911 * Deprecated EVP_PKEY_decrypt_old() and EVP_PKEY_encrypt_old().
915 * Enhanced the documentation of EVP_PKEY_get_size(), EVP_PKEY_get_bits()
916 and EVP_PKEY_get_security_bits(). Especially EVP_PKEY_get_size() needed
917 a new formulation to include all the things it can be used for,
918 as well as words of caution.
922 * The SSL_CTX_set_tlsext_ticket_key_cb(3) function has been deprecated.
926 * All of the low level HMAC functions have been deprecated.
928 *Paul Dale and David von Oheimb*
930 * Over two thousand fixes were made to the documentation, including:
931 - Common options (such as -rand/-writerand, TLS version control, etc)
932 were refactored and point to newly-enhanced descriptions in openssl.pod.
933 - Added style conformance for all options (with help from Richard Levitte),
934 documented all reported missing options, added a CI build to check
935 that all options are documented and that no unimplemented options
937 - Documented some internals, such as all use of environment variables.
938 - Addressed all internal broken L<> references.
942 * All of the low level CMAC functions have been deprecated.
946 * The low-level MD2, MD4, MD5, MDC2, RIPEMD160 and Whirlpool digest
947 functions have been deprecated.
949 *Paul Dale and David von Oheimb*
951 * Corrected the documentation of the return values from the `EVP_DigestSign*`
952 set of functions. The documentation mentioned negative values for some
953 errors, but this was never the case, so the mention of negative values
956 Code that followed the documentation and thereby check with something
957 like `EVP_DigestSignInit(...) <= 0` will continue to work undisturbed.
961 * All of the low level cipher functions have been deprecated.
963 *Matt Caswell and Paul Dale*
965 * Removed include/openssl/opensslconf.h.in and replaced it with
966 include/openssl/configuration.h.in, which differs in not including
967 <openssl/macros.h>. A short header include/openssl/opensslconf.h
968 was added to include both.
970 This allows internal hacks where one might need to modify the set
971 of configured macros, for example this if deprecated symbols are
972 still supposed to be available internally:
974 #include <openssl/configuration.h>
976 #undef OPENSSL_NO_DEPRECATED
977 #define OPENSSL_SUPPRESS_DEPRECATED
979 #include <openssl/macros.h>
981 This should not be used by applications that use the exported
982 symbols, as that will lead to linking errors.
986 * Fixed an overflow bug in the x64_64 Montgomery squaring procedure
987 used in exponentiation with 512-bit moduli. No EC algorithms are
988 affected. Analysis suggests that attacks against 2-prime RSA1024,
989 3-prime RSA1536, and DSA1024 as a result of this defect would be very
990 difficult to perform and are not believed likely. Attacks against DH512
991 are considered just feasible. However, for an attack the target would
992 have to re-use the DH512 private key, which is not recommended anyway.
993 Also applications directly using the low-level API BN_mod_exp may be
994 affected if they use BN_FLG_CONSTTIME.
999 * Most memory-debug features have been deprecated, and the functionality
1000 replaced with no-ops.
1004 * Added documentation for the STACK API.
1008 * Introduced a new method type and API, OSSL_ENCODER, to represent
1009 generic encoders. These do the same sort of job that PEM writers
1010 and d2i functions do, but with support for methods supplied by
1011 providers, and the possibility for providers to support other
1016 * Introduced a new method type and API, OSSL_DECODER, to represent
1017 generic decoders. These do the same sort of job that PEM readers
1018 and i2d functions do, but with support for methods supplied by
1019 providers, and the possibility for providers to support other
1024 * Added a .pragma directive to the syntax of configuration files, to
1025 allow varying behavior in a supported and predictable manner.
1026 Currently added pragma:
1030 This allows dollar signs to be a keyword character unless it's
1031 followed by a opening brace or parenthesis. This is useful for
1032 platforms where dollar signs are commonly used in names, such as
1033 volume names and system directory names on VMS.
1037 * Added functionality to create an EVP_PKEY from user data.
1041 * Change the interpretation of the '--api' configuration option to
1042 mean that this is a desired API compatibility level with no
1043 further meaning. The previous interpretation, that this would
1044 also mean to remove all deprecated symbols up to and including
1045 the given version, no requires that 'no-deprecated' is also used
1046 in the configuration.
1048 When building applications, the desired API compatibility level
1049 can be set with the OPENSSL_API_COMPAT macro like before. For
1050 API compatibility version below 3.0, the old style numerical
1051 value is valid as before, such as -DOPENSSL_API_COMPAT=0x10100000L.
1052 For version 3.0 and on, the value is expected to be the decimal
1053 value calculated from the major and minor version like this:
1055 MAJOR * 10000 + MINOR * 100
1059 -DOPENSSL_API_COMPAT=30000 For 3.0
1060 -DOPENSSL_API_COMPAT=30200 For 3.2
1062 To hide declarations that are deprecated up to and including the
1063 given API compatibility level, -DOPENSSL_NO_DEPRECATED must be
1064 given when building the application as well.
1068 * Added the X509_LOOKUP_METHOD called X509_LOOKUP_store, to allow
1069 access to certificate and CRL stores via URIs and OSSL_STORE
1072 This adds the following functions:
1074 - X509_LOOKUP_store()
1075 - X509_STORE_load_file()
1076 - X509_STORE_load_path()
1077 - X509_STORE_load_store()
1078 - SSL_add_store_cert_subjects_to_stack()
1079 - SSL_CTX_set_default_verify_store()
1080 - SSL_CTX_load_verify_file()
1081 - SSL_CTX_load_verify_dir()
1082 - SSL_CTX_load_verify_store()
1086 * Added a new method to gather entropy on VMS, based on SYS$GET_ENTROPY.
1087 The presence of this system service is determined at run-time.
1091 * Added functionality to create an EVP_PKEY context based on data
1092 for methods from providers. This takes an algorithm name and a
1093 property query string and simply stores them, with the intent
1094 that any operation that uses this context will use those strings
1095 to fetch the needed methods implicitly, thereby making the port
1096 of application written for pre-3.0 OpenSSL easier.
1100 * The undocumented function NCONF_WIN32() has been deprecated; for
1101 conversion details see the HISTORY section of doc/man5/config.pod
1105 * Introduced the new functions EVP_DigestSignInit_ex() and
1106 EVP_DigestVerifyInit_ex(). The macros EVP_DigestSignUpdate() and
1107 EVP_DigestVerifyUpdate() have been converted to functions. See the man
1108 pages for further details.
1112 * Over two thousand fixes were made to the documentation, including:
1113 adding missing command flags, better style conformance, documentation
1116 *Rich Salz, Richard Levitte*
1118 * s390x assembly pack: add hardware-support for P-256, P-384, P-521,
1119 X25519, X448, Ed25519 and Ed448.
1123 * Print all values for a PKCS#12 attribute with 'openssl pkcs12', not just
1128 * Deprecated the public definition of `ERR_STATE` as well as the function
1129 `ERR_get_state()`. This is done in preparation of making `ERR_STATE` an
1134 * Added ERR functionality to give callers access to the stored function
1135 names that have replaced the older function code based functions.
1137 New functions are ERR_peek_error_func(), ERR_peek_last_error_func(),
1138 ERR_peek_error_data(), ERR_peek_last_error_data(), ERR_get_error_all(),
1139 ERR_peek_error_all() and ERR_peek_last_error_all().
1141 Deprecate ERR functions ERR_get_error_line(), ERR_get_error_line_data(),
1142 ERR_peek_error_line_data(), ERR_peek_last_error_line_data() and
1143 ERR_func_error_string().
1147 * Extended testing to be verbose for failing tests only. The make variables
1148 VERBOSE_FAILURE or VF can be used to enable this:
1150 $ make VF=1 test # Unix
1151 $ mms /macro=(VF=1) test ! OpenVMS
1152 $ nmake VF=1 test # Windows
1156 * Added the `-copy_extensions` option to the `x509` command for use with
1157 `-req` and `-x509toreq`. When given with the `copy` or `copyall` argument,
1158 all extensions in the request are copied to the certificate or vice versa.
1160 *David von Oheimb*, *Kirill Stefanenkov <kirill_stefanenkov@rambler.ru>*
1162 * Added the `-copy_extensions` option to the `req` command for use with
1163 `-x509`. When given with the `copy` or `copyall` argument,
1164 all extensions in the certification request are copied to the certificate.
1168 * The `x509`, `req`, and `ca` commands now make sure that X.509v3 certificates
1169 they generate are by default RFC 5280 compliant in the following sense:
1170 There is a subjectKeyIdentifier extension with a hash value of the public key
1171 and for not self-signed certs there is an authorityKeyIdentifier extension
1172 with a keyIdentifier field or issuer information identifying the signing key.
1173 This is done unless some configuration overrides the new default behavior,
1174 such as `subjectKeyIdentifier = none` and `authorityKeyIdentifier = none`.
1178 * Added several checks to `X509_verify_cert()` according to requirements in
1179 RFC 5280 in case `X509_V_FLAG_X509_STRICT` is set
1180 (which may be done by using the CLI option `-x509_strict`):
1181 * The basicConstraints of CA certificates must be marked critical.
1182 * CA certificates must explicitly include the keyUsage extension.
1183 * If a pathlenConstraint is given the key usage keyCertSign must be allowed.
1184 * The issuer name of any certificate must not be empty.
1185 * The subject name of CA certs, certs with keyUsage crlSign,
1186 and certs without subjectAlternativeName must not be empty.
1187 * If a subjectAlternativeName extension is given it must not be empty.
1188 * The signatureAlgorithm field and the cert signature must be consistent.
1189 * Any given authorityKeyIdentifier and any given subjectKeyIdentifier
1190 must not be marked critical.
1191 * The authorityKeyIdentifier must be given for X.509v3 certs
1192 unless they are self-signed.
1193 * The subjectKeyIdentifier must be given for all X.509v3 CA certs.
1197 * Certificate verification using `X509_verify_cert()` meanwhile rejects EC keys
1198 with explicit curve parameters (specifiedCurve) as required by RFC 5480.
1202 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
1203 used even when parsing explicit parameters, when loading a encoded key
1204 or calling `EC_GROUP_new_from_ecpkparameters()`/
1205 `EC_GROUP_new_from_ecparameters()`.
1206 This prevents bypass of security hardening and performance gains,
1207 especially for curves with specialized EC_METHODs.
1208 By default, if a key encoded with explicit parameters is loaded and later
1209 encoded, the output is still encoded with explicit parameters, even if
1210 internally a "named" EC_GROUP is used for computation.
1214 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
1215 this change, EC_GROUP_set_generator would accept order and/or cofactor as
1216 NULL. After this change, only the cofactor parameter can be NULL. It also
1217 does some minimal sanity checks on the passed order.
1222 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
1223 An attack is simple, if the first CMS_recipientInfo is valid but the
1224 second CMS_recipientInfo is chosen ciphertext. If the second
1225 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
1226 encryption key will be replaced by garbage, and the message cannot be
1227 decoded, but if the RSA decryption fails, the correct encryption key is
1228 used and the recipient will not notice the attack.
1229 As a work around for this potential attack the length of the decrypted
1230 key must be equal to the cipher default key length, in case the
1231 certifiate is not given and all recipientInfo are tried out.
1232 The old behaviour can be re-enabled in the CMS code by setting the
1233 CMS_DEBUG_DECRYPT flag.
1237 * Early start up entropy quality from the DEVRANDOM seed source has been
1238 improved for older Linux systems. The RAND subsystem will wait for
1239 /dev/random to be producing output before seeding from /dev/urandom.
1240 The seeded state is stored for future library initialisations using
1241 a system global shared memory segment. The shared memory identifier
1242 can be configured by defining OPENSSL_RAND_SEED_DEVRANDOM_SHM_ID to
1243 the desired value. The default identifier is 114.
1247 * Revised BN_generate_prime_ex to not avoid factors 2..17863 in p-1
1248 when primes for RSA keys are computed.
1249 Since we previously always generated primes == 2 (mod 3) for RSA keys,
1250 the 2-prime and 3-prime RSA modules were easy to distinguish, since
1251 `N = p*q = 1 (mod 3)`, but `N = p*q*r = 2 (mod 3)`. Therefore fingerprinting
1252 2-prime vs. 3-prime RSA keys was possible by computing N mod 3.
1253 This avoids possible fingerprinting of newly generated RSA modules.
1257 * Correct the extended master secret constant on EBCDIC systems. Without this
1258 fix TLS connections between an EBCDIC system and a non-EBCDIC system that
1259 negotiate EMS will fail. Unfortunately this also means that TLS connections
1260 between EBCDIC systems with this fix, and EBCDIC systems without this
1261 fix will fail if they negotiate EMS.
1265 * Changed the library initialisation so that the config file is now loaded
1266 by default. This was already the case for libssl. It now occurs for both
1267 libcrypto and libssl. Use the OPENSSL_INIT_NO_LOAD_CONFIG option to
1268 `OPENSSL_init_crypto()` to suppress automatic loading of a config file.
1272 * Introduced new error raising macros, `ERR_raise()` and `ERR_raise_data()`,
1273 where the former acts as a replacement for `ERR_put_error()`, and the
1274 latter replaces the combination `ERR_put_error()` + `ERR_add_error_data()`.
1275 `ERR_raise_data()` adds more flexibility by taking a format string and
1276 an arbitrary number of arguments following it, to be processed with
1281 * Introduced a new function, `OSSL_PROVIDER_available()`, which can be used
1282 to check if a named provider is loaded and available. When called, it
1283 will also activate all fallback providers if such are still present.
1287 * Enforce a minimum DH modulus size of 512 bits.
1291 * Changed DH parameters to generate the order q subgroup instead of 2q.
1292 Previously generated DH parameters are still accepted by DH_check
1293 but DH_generate_key works around that by clearing bit 0 of the
1294 private key for those. This avoids leaking bit 0 of the private key.
1298 * Significantly reduce secure memory usage by the randomness pools.
1302 * `{CRYPTO,OPENSSL}_mem_debug_{push,pop}` are now no-ops and have been
1307 * A new type, EVP_KEYEXCH, has been introduced to represent key exchange
1308 algorithms. An implementation of a key exchange algorithm can be obtained
1309 by using the function EVP_KEYEXCH_fetch(). An EVP_KEYEXCH algorithm can be
1310 used in a call to EVP_PKEY_derive_init_ex() which works in a similar way to
1311 the older EVP_PKEY_derive_init() function. See the man pages for the new
1312 functions for further details.
1316 * The EVP_PKEY_CTX_set_dh_pad() macro has now been converted to a function.
1320 * Removed the function names from error messages and deprecated the
1325 * Removed NextStep support and the macro OPENSSL_UNISTD
1329 * Removed DES_check_key. Also removed OPENSSL_IMPLEMENT_GLOBAL,
1330 OPENSSL_GLOBAL_REF, OPENSSL_DECLARE_GLOBAL.
1331 Also removed "export var as function" capability; we do not export
1332 variables, only functions.
1336 * RC5_32_set_key has been changed to return an int type, with 0 indicating
1337 an error and 1 indicating success. In previous versions of OpenSSL this
1338 was a void type. If a key was set longer than the maximum possible this
1343 * Support SM2 signing and verification schemes with X509 certificate.
1347 * Use SHA256 as the default digest for TS query in the `ts` app.
1351 * Change PBKDF2 to conform to SP800-132 instead of the older PKCS5 RFC2898.
1355 * Default cipher lists/suites are now available via a function, the
1356 #defines are deprecated.
1360 * Add target VC-WIN32-UWP, VC-WIN64A-UWP, VC-WIN32-ARM-UWP and
1361 VC-WIN64-ARM-UWP in Windows OneCore target for making building libraries
1362 for Windows Store apps easier. Also, the "no-uplink" option has been added.
1366 * Join the directories crypto/x509 and crypto/x509v3
1370 * Added command 'openssl kdf' that uses the EVP_KDF API.
1374 * Added command 'openssl mac' that uses the EVP_MAC API.
1378 * Added OPENSSL_info() to get diverse built-in OpenSSL data, such
1379 as default directories. Also added the command 'openssl info'
1380 for scripting purposes.
1384 * The functions AES_ige_encrypt() and AES_bi_ige_encrypt() have been
1389 * Add prediction resistance to the DRBG reseeding process.
1393 * Limit the number of blocks in a data unit for AES-XTS to 2^20 as
1394 mandated by IEEE Std 1619-2018.
1398 * Added newline escaping functionality to a filename when using openssl dgst.
1399 This output format is to replicate the output format found in the `*sum`
1400 checksum programs. This aims to preserve backward compatibility.
1402 *Matt Eaton, Richard Levitte, and Paul Dale*
1404 * Removed the heartbeat message in DTLS feature, as it has very
1405 little usage and doesn't seem to fulfill a valuable purpose.
1406 The configuration option is now deprecated.
1410 * Changed the output of 'openssl {digestname} < file' to display the
1411 digest name in its output.
1415 * Added a new generic trace API which provides support for enabling
1416 instrumentation through trace output.
1418 *Richard Levitte & Matthias St. Pierre*
1420 * Added build tests for C++. These are generated files that only do one
1421 thing, to include one public OpenSSL head file each. This tests that
1422 the public header files can be usefully included in a C++ application.
1424 This test isn't enabled by default. It can be enabled with the option
1425 'enable-buildtest-c++'.
1429 * Added KB KDF (EVP_KDF_KB) to EVP_KDF.
1433 * Added SSH KDF (EVP_KDF_SSHKDF) and KRB5 KDF (EVP_KDF_KRB5KDF) to EVP_KDF.
1437 * Added Single Step KDF (EVP_KDF_SS), X963 KDF, and X942 KDF to EVP_KDF.
1441 * Added KMAC to EVP_MAC.
1445 * Added property based algorithm implementation selection framework to
1450 * Added SCA hardening for modular field inversion in EC_GROUP through
1451 a new dedicated field_inv() pointer in EC_METHOD.
1452 This also addresses a leakage affecting conversions from projective
1453 to affine coordinates.
1455 *Billy Bob Brumley, Nicola Tuveri*
1457 * Added EVP_KDF, an EVP layer KDF API, to simplify adding KDF and PRF
1458 implementations. This includes an EVP_PKEY to EVP_KDF bridge for
1459 those algorithms that were already supported through the EVP_PKEY API
1460 (scrypt, TLS1 PRF and HKDF). The low-level KDF functions for PBKDF2
1461 and scrypt are now wrappers that call EVP_KDF.
1465 * Build devcrypto engine as a dynamic engine.
1467 *Eneas U de Queiroz*
1469 * Add keyed BLAKE2 to EVP_MAC.
1473 * Fix a bug in the computation of the endpoint-pair shared secret used
1474 by DTLS over SCTP. This breaks interoperability with older versions
1475 of OpenSSL like OpenSSL 1.1.0 and OpenSSL 1.0.2. There is a runtime
1476 switch SSL_MODE_DTLS_SCTP_LABEL_LENGTH_BUG (off by default) enabling
1477 interoperability with such broken implementations. However, enabling
1478 this switch breaks interoperability with correct implementations.
1480 * Fix a use after free bug in d2i_X509_PUBKEY when overwriting a
1481 re-used X509_PUBKEY object if the second PUBKEY is malformed.
1485 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
1489 * Changed the license to the Apache License v2.0.
1493 * Switch to a new version scheme using three numbers MAJOR.MINOR.PATCH.
1495 - Major releases (indicated by incrementing the MAJOR release number)
1496 may introduce incompatible API/ABI changes.
1497 - Minor releases (indicated by incrementing the MINOR release number)
1498 may introduce new features but retain API/ABI compatibility.
1499 - Patch releases (indicated by incrementing the PATCH number)
1500 are intended for bug fixes and other improvements of existing
1501 features only (like improving performance or adding documentation)
1502 and retain API/ABI compatibility.
1506 * Add support for RFC5297 SIV mode (siv128), including AES-SIV.
1510 * Remove the 'dist' target and add a tarball building script. The
1511 'dist' target has fallen out of use, and it shouldn't be
1512 necessary to configure just to create a source distribution.
1516 * Recreate the OS390-Unix config target. It no longer relies on a
1517 special script like it did for OpenSSL pre-1.1.0.
1521 * Instead of having the source directories listed in Configure, add
1522 a 'build.info' keyword SUBDIRS to indicate what sub-directories to
1527 * Add GMAC to EVP_MAC.
1531 * Ported the HMAC, CMAC and SipHash EVP_PKEY_METHODs to EVP_MAC.
1535 * Added EVP_MAC, an EVP layer MAC API, to simplify adding MAC
1536 implementations. This includes a generic EVP_PKEY to EVP_MAC bridge,
1537 to facilitate the continued use of MACs through raw private keys in
1538 functionality such as `EVP_DigestSign*` and `EVP_DigestVerify*`.
1542 * Deprecate ECDH_KDF_X9_62().
1546 * Added EVP_PKEY_ECDH_KDF_X9_63 and ecdh_KDF_X9_63() as replacements for
1547 the EVP_PKEY_ECDH_KDF_X9_62 KDF type and ECDH_KDF_X9_62(). The old names
1548 are retained for backwards compatibility.
1552 * AES-XTS mode now enforces that its two keys are different to mitigate
1553 the attacked described in "Efficient Instantiations of Tweakable
1554 Blockciphers and Refinements to Modes OCB and PMAC" by Phillip Rogaway.
1555 Details of this attack can be obtained from:
1556 <http://web.cs.ucdavis.edu/%7Erogaway/papers/offsets.pdf>
1560 * Rename the object files, i.e. give them other names than in previous
1561 versions. Their names now include the name of the final product, as
1562 well as its type mnemonic (bin, lib, shlib).
1566 * Added new option for 'openssl list', '-objects', which will display the
1567 list of built in objects, i.e. OIDs with names.
1571 * Added the options `-crl_lastupdate` and `-crl_nextupdate` to `openssl ca`,
1572 allowing the `lastUpdate` and `nextUpdate` fields in the generated CRL to
1577 * Added support for Linux Kernel TLS data-path. The Linux Kernel data-path
1578 improves application performance by removing data copies and providing
1579 applications with zero-copy system calls such as sendfile and splice.
1583 * The SSL option SSL_OP_CLEANSE_PLAINTEXT is introduced.
1587 * `PKCS12_parse` now maintains the order of the parsed certificates
1588 when outputting them via `*ca` (rather than reversing it).
1592 * Deprecated pthread fork support methods.
1596 * Added support for FFDHE key exchange in TLS 1.3.
1600 * Added a new concept for OpenSSL plugability: providers. This
1601 functionality is designed to replace the ENGINE API and ENGINE
1602 implementations, and to be much more dynamic, allowing provider
1603 authors to introduce new algorithms among other things, as long as
1604 there's an API that supports the algorithm type.
1606 With this concept comes a new core API for interaction between
1607 libcrypto and provider implementations. Public libcrypto functions
1608 that want to use providers do so through this core API.
1610 The main documentation for this core API is found in
1611 doc/man7/provider.pod, doc/man7/provider-base.pod, and they in turn
1612 refer to other manuals describing the API specific for supported
1613 algorithm types (also called operations).
1620 ### Changes between 1.1.1m and 1.1.1n [xx XXX xxxx]
1622 ### Changes between 1.1.1l and 1.1.1m [14 Dec 2021]
1624 * Avoid loading of a dynamic engine twice.
1628 * Prioritise DANE TLSA issuer certs over peer certs
1632 * Fixed random API for MacOS prior to 10.12
1634 These MacOS versions don't support the CommonCrypto APIs
1638 ### Changes between 1.1.1k and 1.1.1l [24 Aug 2021]
1640 * Fixed an SM2 Decryption Buffer Overflow.
1642 In order to decrypt SM2 encrypted data an application is expected to
1643 call the API function EVP_PKEY_decrypt(). Typically an application will
1644 call this function twice. The first time, on entry, the "out" parameter
1645 can be NULL and, on exit, the "outlen" parameter is populated with the
1646 buffer size required to hold the decrypted plaintext. The application
1647 can then allocate a sufficiently sized buffer and call EVP_PKEY_decrypt()
1648 again, but this time passing a non-NULL value for the "out" parameter.
1650 A bug in the implementation of the SM2 decryption code means that the
1651 calculation of the buffer size required to hold the plaintext returned
1652 by the first call to EVP_PKEY_decrypt() can be smaller than the actual
1653 size required by the second call. This can lead to a buffer overflow
1654 when EVP_PKEY_decrypt() is called by the application a second time with
1655 a buffer that is too small.
1657 A malicious attacker who is able present SM2 content for decryption to
1658 an application could cause attacker chosen data to overflow the buffer
1659 by up to a maximum of 62 bytes altering the contents of other data held
1660 after the buffer, possibly changing application behaviour or causing
1661 the application to crash. The location of the buffer is application
1662 dependent but is typically heap allocated.
1667 * Fixed various read buffer overruns processing ASN.1 strings
1669 ASN.1 strings are represented internally within OpenSSL as an ASN1_STRING
1670 structure which contains a buffer holding the string data and a field
1671 holding the buffer length. This contrasts with normal C strings which
1672 are repesented as a buffer for the string data which is terminated
1673 with a NUL (0) byte.
1675 Although not a strict requirement, ASN.1 strings that are parsed using
1676 OpenSSL's own "d2i" functions (and other similar parsing functions) as
1677 well as any string whose value has been set with the ASN1_STRING_set()
1678 function will additionally NUL terminate the byte array in the
1679 ASN1_STRING structure.
1681 However, it is possible for applications to directly construct valid
1682 ASN1_STRING structures which do not NUL terminate the byte array by
1683 directly setting the "data" and "length" fields in the ASN1_STRING
1684 array. This can also happen by using the ASN1_STRING_set0() function.
1686 Numerous OpenSSL functions that print ASN.1 data have been found to
1687 assume that the ASN1_STRING byte array will be NUL terminated, even
1688 though this is not guaranteed for strings that have been directly
1689 constructed. Where an application requests an ASN.1 structure to be
1690 printed, and where that ASN.1 structure contains ASN1_STRINGs that have
1691 been directly constructed by the application without NUL terminating
1692 the "data" field, then a read buffer overrun can occur.
1694 The same thing can also occur during name constraints processing
1695 of certificates (for example if a certificate has been directly
1696 constructed by the application instead of loading it via the OpenSSL
1697 parsing functions, and the certificate contains non NUL terminated
1698 ASN1_STRING structures). It can also occur in the X509_get1_email(),
1699 X509_REQ_get1_email() and X509_get1_ocsp() functions.
1701 If a malicious actor can cause an application to directly construct an
1702 ASN1_STRING and then process it through one of the affected OpenSSL
1703 functions then this issue could be hit. This might result in a crash
1704 (causing a Denial of Service attack). It could also result in the
1705 disclosure of private memory contents (such as private keys, or
1706 sensitive plaintext).
1711 ### Changes between 1.1.1j and 1.1.1k [25 Mar 2021]
1713 * Fixed a problem with verifying a certificate chain when using the
1714 X509_V_FLAG_X509_STRICT flag. This flag enables additional security checks of
1715 the certificates present in a certificate chain. It is not set by default.
1717 Starting from OpenSSL version 1.1.1h a check to disallow certificates in
1718 the chain that have explicitly encoded elliptic curve parameters was added
1719 as an additional strict check.
1721 An error in the implementation of this check meant that the result of a
1722 previous check to confirm that certificates in the chain are valid CA
1723 certificates was overwritten. This effectively bypasses the check
1724 that non-CA certificates must not be able to issue other certificates.
1726 If a "purpose" has been configured then there is a subsequent opportunity
1727 for checks that the certificate is a valid CA. All of the named "purpose"
1728 values implemented in libcrypto perform this check. Therefore, where
1729 a purpose is set the certificate chain will still be rejected even when the
1730 strict flag has been used. A purpose is set by default in libssl client and
1731 server certificate verification routines, but it can be overridden or
1732 removed by an application.
1734 In order to be affected, an application must explicitly set the
1735 X509_V_FLAG_X509_STRICT verification flag and either not set a purpose
1736 for the certificate verification or, in the case of TLS client or server
1737 applications, override the default purpose.
1742 * Fixed an issue where an OpenSSL TLS server may crash if sent a maliciously
1743 crafted renegotiation ClientHello message from a client. If a TLSv1.2
1744 renegotiation ClientHello omits the signature_algorithms extension (where it
1745 was present in the initial ClientHello), but includes a
1746 signature_algorithms_cert extension then a NULL pointer dereference will
1747 result, leading to a crash and a denial of service attack.
1749 A server is only vulnerable if it has TLSv1.2 and renegotiation enabled
1750 (which is the default configuration). OpenSSL TLS clients are not impacted by
1754 *Peter Kästle and Samuel Sapalski*
1756 ### Changes between 1.1.1i and 1.1.1j [16 Feb 2021]
1758 * Fixed the X509_issuer_and_serial_hash() function. It attempts to
1759 create a unique hash value based on the issuer and serial number data
1760 contained within an X509 certificate. However it was failing to correctly
1761 handle any errors that may occur while parsing the issuer field (which might
1762 occur if the issuer field is maliciously constructed). This may subsequently
1763 result in a NULL pointer deref and a crash leading to a potential denial of
1769 * Fixed the RSA_padding_check_SSLv23() function and the RSA_SSLV23_PADDING
1770 padding mode to correctly check for rollback attacks. This is considered a
1771 bug in OpenSSL 1.1.1 because it does not support SSLv2. In 1.0.2 this is
1776 Fixed the EVP_CipherUpdate, EVP_EncryptUpdate and EVP_DecryptUpdate
1777 functions. Previously they could overflow the output length argument in some
1778 cases where the input length is close to the maximum permissable length for
1779 an integer on the platform. In such cases the return value from the function
1780 call would be 1 (indicating success), but the output length value would be
1781 negative. This could cause applications to behave incorrectly or crash.
1786 * Fixed SRP_Calc_client_key so that it runs in constant time. The previous
1787 implementation called BN_mod_exp without setting BN_FLG_CONSTTIME. This
1788 could be exploited in a side channel attack to recover the password. Since
1789 the attack is local host only this is outside of the current OpenSSL
1790 threat model and therefore no CVE is assigned.
1792 Thanks to Mohammed Sabt and Daniel De Almeida Braga for reporting this
1797 ### Changes between 1.1.1h and 1.1.1i [8 Dec 2020]
1799 * Fixed NULL pointer deref in the GENERAL_NAME_cmp function
1800 This function could crash if both GENERAL_NAMEs contain an EDIPARTYNAME.
1801 If an attacker can control both items being compared then this could lead
1802 to a possible denial of service attack. OpenSSL itself uses the
1803 GENERAL_NAME_cmp function for two purposes:
1804 1) Comparing CRL distribution point names between an available CRL and a
1805 CRL distribution point embedded in an X509 certificate
1806 2) When verifying that a timestamp response token signer matches the
1807 timestamp authority name (exposed via the API functions
1808 TS_RESP_verify_response and TS_RESP_verify_token)
1813 ### Changes between 1.1.1g and 1.1.1h [22 Sep 2020]
1815 * Certificates with explicit curve parameters are now disallowed in
1816 verification chains if the X509_V_FLAG_X509_STRICT flag is used.
1820 * The 'MinProtocol' and 'MaxProtocol' configuration commands now silently
1821 ignore TLS protocol version bounds when configuring DTLS-based contexts, and
1822 conversely, silently ignore DTLS protocol version bounds when configuring
1823 TLS-based contexts. The commands can be repeated to set bounds of both
1824 types. The same applies with the corresponding "min_protocol" and
1825 "max_protocol" command-line switches, in case some application uses both TLS
1828 SSL_CTX instances that are created for a fixed protocol version (e.g.
1829 TLSv1_server_method()) also silently ignore version bounds. Previously
1830 attempts to apply bounds to these protocol versions would result in an
1831 error. Now only the "version-flexible" SSL_CTX instances are subject to
1832 limits in configuration files in command-line options.
1836 * Handshake now fails if Extended Master Secret extension is dropped
1841 * The Oracle Developer Studio compiler will start reporting deprecated APIs
1843 ### Changes between 1.1.1f and 1.1.1g [21 Apr 2020]
1845 * Fixed segmentation fault in SSL_check_chain()
1846 Server or client applications that call the SSL_check_chain() function
1847 during or after a TLS 1.3 handshake may crash due to a NULL pointer
1848 dereference as a result of incorrect handling of the
1849 "signature_algorithms_cert" TLS extension. The crash occurs if an invalid
1850 or unrecognised signature algorithm is received from the peer. This could
1851 be exploited by a malicious peer in a Denial of Service attack.
1856 * Added AES consttime code for no-asm configurations
1857 an optional constant time support for AES was added
1858 when building openssl for no-asm.
1859 Enable with: ./config no-asm -DOPENSSL_AES_CONST_TIME
1860 Disable with: ./config no-asm -DOPENSSL_NO_AES_CONST_TIME
1861 At this time this feature is by default disabled.
1862 It will be enabled by default in 3.0.
1866 ### Changes between 1.1.1e and 1.1.1f [31 Mar 2020]
1868 * Revert the change of EOF detection while reading in libssl to avoid
1869 regressions in applications depending on the current way of reporting
1870 the EOF. As the existing method is not fully accurate the change to
1871 reporting the EOF via SSL_ERROR_SSL is kept on the current development
1872 branch and will be present in the 3.0 release.
1876 * Revised BN_generate_prime_ex to not avoid factors 3..17863 in p-1
1877 when primes for RSA keys are computed.
1878 Since we previously always generated primes == 2 (mod 3) for RSA keys,
1879 the 2-prime and 3-prime RSA modules were easy to distinguish, since
1880 N = p*q = 1 (mod 3), but N = p*q*r = 2 (mod 3). Therefore fingerprinting
1881 2-prime vs. 3-prime RSA keys was possible by computing N mod 3.
1882 This avoids possible fingerprinting of newly generated RSA modules.
1886 ### Changes between 1.1.1d and 1.1.1e [17 Mar 2020]
1888 * Properly detect EOF while reading in libssl. Previously if we hit an EOF
1889 while reading in libssl then we would report an error back to the
1890 application (SSL_ERROR_SYSCALL) but errno would be 0. We now add
1891 an error to the stack (which means we instead return SSL_ERROR_SSL) and
1892 therefore give a hint as to what went wrong.
1896 * Check that ed25519 and ed448 are allowed by the security level. Previously
1897 signature algorithms not using an MD were not being checked that they were
1898 allowed by the security level.
1902 * Fixed SSL_get_servername() behaviour. The behaviour of SSL_get_servername()
1903 was not quite right. The behaviour was not consistent between resumption
1904 and normal handshakes, and also not quite consistent with historical
1905 behaviour. The behaviour in various scenarios has been clarified and
1906 it has been updated to make it match historical behaviour as closely as
1911 * *[VMS only]* The header files that the VMS compilers include automatically,
1912 `__DECC_INCLUDE_PROLOGUE.H` and `__DECC_INCLUDE_EPILOGUE.H`, use pragmas
1913 that the C++ compiler doesn't understand. This is a shortcoming in the
1914 compiler, but can be worked around with `__cplusplus` guards.
1916 C++ applications that use OpenSSL libraries must be compiled using the
1917 qualifier `/NAMES=(AS_IS,SHORTENED)` to be able to use all the OpenSSL
1918 functions. Otherwise, only functions with symbols of less than 31
1919 characters can be used, as the linker will not be able to successfully
1920 resolve symbols with longer names.
1924 * Added a new method to gather entropy on VMS, based on SYS$GET_ENTROPY.
1925 The presence of this system service is determined at run-time.
1929 * Print all values for a PKCS#12 attribute with 'openssl pkcs12', not just
1934 ### Changes between 1.1.1c and 1.1.1d [10 Sep 2019]
1936 * Fixed a fork protection issue. OpenSSL 1.1.1 introduced a rewritten random
1937 number generator (RNG). This was intended to include protection in the
1938 event of a fork() system call in order to ensure that the parent and child
1939 processes did not share the same RNG state. However this protection was not
1940 being used in the default case.
1942 A partial mitigation for this issue is that the output from a high
1943 precision timer is mixed into the RNG state so the likelihood of a parent
1944 and child process sharing state is significantly reduced.
1946 If an application already calls OPENSSL_init_crypto() explicitly using
1947 OPENSSL_INIT_ATFORK then this problem does not occur at all.
1950 *Matthias St. Pierre*
1952 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
1953 used even when parsing explicit parameters, when loading a encoded key
1954 or calling `EC_GROUP_new_from_ecpkparameters()`/
1955 `EC_GROUP_new_from_ecparameters()`.
1956 This prevents bypass of security hardening and performance gains,
1957 especially for curves with specialized EC_METHODs.
1958 By default, if a key encoded with explicit parameters is loaded and later
1959 encoded, the output is still encoded with explicit parameters, even if
1960 internally a "named" EC_GROUP is used for computation.
1964 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
1965 this change, EC_GROUP_set_generator would accept order and/or cofactor as
1966 NULL. After this change, only the cofactor parameter can be NULL. It also
1967 does some minimal sanity checks on the passed order.
1972 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
1973 An attack is simple, if the first CMS_recipientInfo is valid but the
1974 second CMS_recipientInfo is chosen ciphertext. If the second
1975 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
1976 encryption key will be replaced by garbage, and the message cannot be
1977 decoded, but if the RSA decryption fails, the correct encryption key is
1978 used and the recipient will not notice the attack.
1979 As a work around for this potential attack the length of the decrypted
1980 key must be equal to the cipher default key length, in case the
1981 certifiate is not given and all recipientInfo are tried out.
1982 The old behaviour can be re-enabled in the CMS code by setting the
1983 CMS_DEBUG_DECRYPT flag.
1988 * Early start up entropy quality from the DEVRANDOM seed source has been
1989 improved for older Linux systems. The RAND subsystem will wait for
1990 /dev/random to be producing output before seeding from /dev/urandom.
1991 The seeded state is stored for future library initialisations using
1992 a system global shared memory segment. The shared memory identifier
1993 can be configured by defining OPENSSL_RAND_SEED_DEVRANDOM_SHM_ID to
1994 the desired value. The default identifier is 114.
1998 * Correct the extended master secret constant on EBCDIC systems. Without this
1999 fix TLS connections between an EBCDIC system and a non-EBCDIC system that
2000 negotiate EMS will fail. Unfortunately this also means that TLS connections
2001 between EBCDIC systems with this fix, and EBCDIC systems without this
2002 fix will fail if they negotiate EMS.
2006 * Use Windows installation paths in the mingw builds
2008 Mingw isn't a POSIX environment per se, which means that Windows
2009 paths should be used for installation.
2014 * Changed DH_check to accept parameters with order q and 2q subgroups.
2015 With order 2q subgroups the bit 0 of the private key is not secret
2016 but DH_generate_key works around that by clearing bit 0 of the
2017 private key for those. This avoids leaking bit 0 of the private key.
2021 * Significantly reduce secure memory usage by the randomness pools.
2025 * Revert the DEVRANDOM_WAIT feature for Linux systems
2027 The DEVRANDOM_WAIT feature added a select() call to wait for the
2028 /dev/random device to become readable before reading from the
2029 /dev/urandom device.
2031 It turned out that this change had negative side effects on
2032 performance which were not acceptable. After some discussion it
2033 was decided to revert this feature and leave it up to the OS
2034 resp. the platform maintainer to ensure a proper initialization
2035 during early boot time.
2037 *Matthias St. Pierre*
2039 ### Changes between 1.1.1b and 1.1.1c [28 May 2019]
2041 * Add build tests for C++. These are generated files that only do one
2042 thing, to include one public OpenSSL head file each. This tests that
2043 the public header files can be usefully included in a C++ application.
2045 This test isn't enabled by default. It can be enabled with the option
2046 'enable-buildtest-c++'.
2050 * Enable SHA3 pre-hashing for ECDSA and DSA.
2054 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
2055 This changes the size when using the `genpkey` command when no size is given.
2056 It fixes an omission in earlier changes that changed all RSA, DSA and DH
2057 generation commands to use 2048 bits by default.
2061 * Reorganize the manual pages to consistently have RETURN VALUES,
2062 EXAMPLES, SEE ALSO and HISTORY come in that order, and adjust
2063 util/fix-doc-nits accordingly.
2065 *Paul Yang, Joshua Lock*
2067 * Add the missing accessor EVP_PKEY_get0_engine()
2071 * Have commands like `s_client` and `s_server` output the signature scheme
2072 along with other cipher suite parameters when debugging.
2076 * Make OPENSSL_config() error agnostic again.
2080 * Do the error handling in RSA decryption constant time.
2084 * Prevent over long nonces in ChaCha20-Poly1305.
2086 ChaCha20-Poly1305 is an AEAD cipher, and requires a unique nonce input
2087 for every encryption operation. RFC 7539 specifies that the nonce value
2088 (IV) should be 96 bits (12 bytes). OpenSSL allows a variable nonce length
2089 and front pads the nonce with 0 bytes if it is less than 12
2090 bytes. However it also incorrectly allows a nonce to be set of up to 16
2091 bytes. In this case only the last 12 bytes are significant and any
2092 additional leading bytes are ignored.
2094 It is a requirement of using this cipher that nonce values are
2095 unique. Messages encrypted using a reused nonce value are susceptible to
2096 serious confidentiality and integrity attacks. If an application changes
2097 the default nonce length to be longer than 12 bytes and then makes a
2098 change to the leading bytes of the nonce expecting the new value to be a
2099 new unique nonce then such an application could inadvertently encrypt
2100 messages with a reused nonce.
2102 Additionally the ignored bytes in a long nonce are not covered by the
2103 integrity guarantee of this cipher. Any application that relies on the
2104 integrity of these ignored leading bytes of a long nonce may be further
2105 affected. Any OpenSSL internal use of this cipher, including in SSL/TLS,
2106 is safe because no such use sets such a long nonce value. However user
2107 applications that use this cipher directly and set a non-default nonce
2108 length to be longer than 12 bytes may be vulnerable.
2110 This issue was reported to OpenSSL on 16th of March 2019 by Joran Dirk
2116 * Add DEVRANDOM_WAIT feature for Linux systems
2118 On older Linux systems where the getrandom() system call is not available,
2119 OpenSSL normally uses the /dev/urandom device for seeding its CSPRNG.
2120 Contrary to getrandom(), the /dev/urandom device will not block during
2121 early boot when the kernel CSPRNG has not been seeded yet.
2123 To mitigate this known weakness, use select() to wait for /dev/random to
2124 become readable before reading from /dev/urandom.
2126 * Ensure that SM2 only uses SM3 as digest algorithm
2130 ### Changes between 1.1.1a and 1.1.1b [26 Feb 2019]
2132 * Change the info callback signals for the start and end of a post-handshake
2133 message exchange in TLSv1.3. In 1.1.1/1.1.1a we used SSL_CB_HANDSHAKE_START
2134 and SSL_CB_HANDSHAKE_DONE. Experience has shown that many applications get
2135 confused by this and assume that a TLSv1.2 renegotiation has started. This
2136 can break KeyUpdate handling. Instead we no longer signal the start and end
2137 of a post handshake message exchange (although the messages themselves are
2138 still signalled). This could break some applications that were expecting
2139 the old signals. However without this KeyUpdate is not usable for many
2144 ### Changes between 1.1.1 and 1.1.1a [20 Nov 2018]
2146 * Timing vulnerability in DSA signature generation
2148 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
2149 timing side channel attack. An attacker could use variations in the signing
2150 algorithm to recover the private key.
2152 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
2157 * Timing vulnerability in ECDSA signature generation
2159 The OpenSSL ECDSA signature algorithm has been shown to be vulnerable to a
2160 timing side channel attack. An attacker could use variations in the signing
2161 algorithm to recover the private key.
2163 This issue was reported to OpenSSL on 25th October 2018 by Samuel Weiser.
2168 * Fixed the issue that RAND_add()/RAND_seed() silently discards random input
2169 if its length exceeds 4096 bytes. The limit has been raised to a buffer size
2170 of two gigabytes and the error handling improved.
2172 This issue was reported to OpenSSL by Dr. Falko Strenzke. It has been
2173 categorized as a normal bug, not a security issue, because the DRBG reseeds
2174 automatically and is fully functional even without additional randomness
2175 provided by the application.
2177 ### Changes between 1.1.0i and 1.1.1 [11 Sep 2018]
2179 * Add a new ClientHello callback. Provides a callback interface that gives
2180 the application the ability to adjust the nascent SSL object at the
2181 earliest stage of ClientHello processing, immediately after extensions have
2182 been collected but before they have been processed. In particular, this
2183 callback can adjust the supported TLS versions in response to the contents
2188 * Add SM2 base algorithm support.
2192 * s390x assembly pack: add (improved) hardware-support for the following
2193 cryptographic primitives: sha3, shake, aes-gcm, aes-ccm, aes-ctr, aes-ofb,
2194 aes-cfb/cfb8, aes-ecb.
2198 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
2199 parameter is no longer accepted, as it leads to a corrupt table. NULL
2200 pem_str is reserved for alias entries only.
2204 * Use the new ec_scalar_mul_ladder scaffold to implement a specialized ladder
2205 step for prime curves. The new implementation is based on formulae from
2206 differential addition-and-doubling in homogeneous projective coordinates
2207 from Izu-Takagi "A fast parallel elliptic curve multiplication resistant
2208 against side channel attacks" and Brier-Joye "Weierstrass Elliptic Curves
2209 and Side-Channel Attacks" Eq. (8) for y-coordinate recovery, modified
2210 to work in projective coordinates.
2212 *Billy Bob Brumley, Nicola Tuveri*
2214 * Change generating and checking of primes so that the error rate of not
2215 being prime depends on the intended use based on the size of the input.
2216 For larger primes this will result in more rounds of Miller-Rabin.
2217 The maximal error rate for primes with more than 1080 bits is lowered
2220 *Kurt Roeckx, Annie Yousar*
2222 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
2226 * The 'tsget' script is renamed to 'tsget.pl', to avoid confusion when
2227 moving between systems, and to avoid confusion when a Windows build is
2228 done with mingw vs with MSVC. For POSIX installs, there's still a
2229 symlink or copy named 'tsget' to avoid that confusion as well.
2233 * Revert blinding in ECDSA sign and instead make problematic addition
2234 length-invariant. Switch even to fixed-length Montgomery multiplication.
2238 * Use the new ec_scalar_mul_ladder scaffold to implement a specialized ladder
2239 step for binary curves. The new implementation is based on formulae from
2240 differential addition-and-doubling in mixed Lopez-Dahab projective
2241 coordinates, modified to independently blind the operands.
2243 *Billy Bob Brumley, Sohaib ul Hassan, Nicola Tuveri*
2245 * Add a scaffold to optionally enhance the Montgomery ladder implementation
2246 for `ec_scalar_mul_ladder` (formerly `ec_mul_consttime`) allowing
2247 EC_METHODs to implement their own specialized "ladder step", to take
2248 advantage of more favorable coordinate systems or more efficient
2249 differential addition-and-doubling algorithms.
2251 *Billy Bob Brumley, Sohaib ul Hassan, Nicola Tuveri*
2253 * Modified the random device based seed sources to keep the relevant
2254 file descriptors open rather than reopening them on each access.
2255 This allows such sources to operate in a chroot() jail without
2256 the associated device nodes being available. This behaviour can be
2257 controlled using RAND_keep_random_devices_open().
2261 * Numerous side-channel attack mitigations have been applied. This may have
2262 performance impacts for some algorithms for the benefit of improved
2263 security. Specific changes are noted in this change log by their respective
2268 * AIX shared library support overhaul. Switch to AIX "natural" way of
2269 handling shared libraries, which means collecting shared objects of
2270 different versions and bitnesses in one common archive. This allows to
2271 mitigate conflict between 1.0 and 1.1 side-by-side installations. It
2272 doesn't affect the way 3rd party applications are linked, only how
2273 multi-version installation is managed.
2277 * Make ec_group_do_inverse_ord() more robust and available to other
2278 EC cryptosystems, so that irrespective of BN_FLG_CONSTTIME, SCA
2279 mitigations are applied to the fallback BN_mod_inverse().
2280 When using this function rather than BN_mod_inverse() directly, new
2281 EC cryptosystem implementations are then safer-by-default.
2285 * Add coordinate blinding for EC_POINT and implement projective
2286 coordinate blinding for generic prime curves as a countermeasure to
2287 chosen point SCA attacks.
2289 *Sohaib ul Hassan, Nicola Tuveri, Billy Bob Brumley*
2291 * Add blinding to ECDSA and DSA signatures to protect against side channel
2292 attacks discovered by Keegan Ryan (NCC Group).
2296 * Enforce checking in the `pkeyutl` command to ensure that the input
2297 length does not exceed the maximum supported digest length when performing
2298 a sign, verify or verifyrecover operation.
2302 * SSL_MODE_AUTO_RETRY is enabled by default. Applications that use blocking
2303 I/O in combination with something like select() or poll() will hang. This
2304 can be turned off again using SSL_CTX_clear_mode().
2305 Many applications do not properly handle non-application data records, and
2306 TLS 1.3 sends more of such records. Setting SSL_MODE_AUTO_RETRY works
2307 around the problems in those applications, but can also break some.
2308 It's recommended to read the manpages about SSL_read(), SSL_write(),
2309 SSL_get_error(), SSL_shutdown(), SSL_CTX_set_mode() and
2310 SSL_CTX_set_read_ahead() again.
2314 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
2315 now allow empty (zero character) pass phrases.
2319 * Apply blinding to binary field modular inversion and remove patent
2320 pending (OPENSSL_SUN_GF2M_DIV) BN_GF2m_mod_div implementation.
2324 * Deprecate ec2_mult.c and unify scalar multiplication code paths for
2325 binary and prime elliptic curves.
2329 * Remove ECDSA nonce padding: EC_POINT_mul is now responsible for
2330 constant time fixed point multiplication.
2334 * Revise elliptic curve scalar multiplication with timing attack
2335 defenses: ec_wNAF_mul redirects to a constant time implementation
2336 when computing fixed point and variable point multiplication (which
2337 in OpenSSL are mostly used with secret scalars in keygen, sign,
2338 ECDH derive operations).
2339 *Billy Bob Brumley, Nicola Tuveri, Cesar Pereida García,
2342 * Updated CONTRIBUTING
2346 * Updated DRBG / RAND to request nonce and additional low entropy
2347 randomness from the system.
2349 *Matthias St. Pierre*
2351 * Updated 'openssl rehash' to use OpenSSL consistent default.
2355 * Moved the load of the ssl_conf module to libcrypto, which helps
2356 loading engines that libssl uses before libssl is initialised.
2360 * Added EVP_PKEY_sign() and EVP_PKEY_verify() for EdDSA
2364 * Fixed X509_NAME_ENTRY_set to get multi-valued RDNs right in all cases.
2366 *Ingo Schwarze, Rich Salz*
2368 * Added output of accepting IP address and port for 'openssl s_server'
2372 * Added a new API for TLSv1.3 ciphersuites:
2373 SSL_CTX_set_ciphersuites()
2374 SSL_set_ciphersuites()
2378 * Memory allocation failures consistently add an error to the error
2383 * Don't use OPENSSL_ENGINES and OPENSSL_CONF environment values
2384 in libcrypto when run as setuid/setgid.
2388 * Load any config file by default when libssl is used.
2392 * Added new public header file <openssl/rand_drbg.h> and documentation
2393 for the RAND_DRBG API. See manual page RAND_DRBG(7) for an overview.
2395 *Matthias St. Pierre*
2397 * QNX support removed (cannot find contributors to get their approval
2398 for the license change).
2402 * TLSv1.3 replay protection for early data has been implemented. See the
2403 SSL_read_early_data() man page for further details.
2407 * Separated TLSv1.3 ciphersuite configuration out from TLSv1.2 ciphersuite
2408 configuration. TLSv1.3 ciphersuites are not compatible with TLSv1.2 and
2409 below. Similarly TLSv1.2 ciphersuites are not compatible with TLSv1.3.
2410 In order to avoid issues where legacy TLSv1.2 ciphersuite configuration
2411 would otherwise inadvertently disable all TLSv1.3 ciphersuites the
2412 configuration has been separated out. See the ciphers man page or the
2413 SSL_CTX_set_ciphersuites() man page for more information.
2417 * On POSIX (BSD, Linux, ...) systems the ocsp(1) command running
2418 in responder mode now supports the new "-multi" option, which
2419 spawns the specified number of child processes to handle OCSP
2420 requests. The "-timeout" option now also limits the OCSP
2421 responder's patience to wait to receive the full client request
2422 on a newly accepted connection. Child processes are respawned
2423 as needed, and the CA index file is automatically reloaded
2424 when changed. This makes it possible to run the "ocsp" responder
2425 as a long-running service, making the OpenSSL CA somewhat more
2426 feature-complete. In this mode, most diagnostic messages logged
2427 after entering the event loop are logged via syslog(3) rather than
2432 * Added support for X448 and Ed448. Heavily based on original work by
2437 * Extend OSSL_STORE with capabilities to search and to narrow the set of
2438 objects loaded. This adds the functions OSSL_STORE_expect() and
2439 OSSL_STORE_find() as well as needed tools to construct searches and
2440 get the search data out of them.
2444 * Support for TLSv1.3 added. Note that users upgrading from an earlier
2445 version of OpenSSL should review their configuration settings to ensure
2446 that they are still appropriate for TLSv1.3. For further information see:
2447 <https://wiki.openssl.org/index.php/TLS1.3>
2451 * Grand redesign of the OpenSSL random generator
2453 The default RAND method now utilizes an AES-CTR DRBG according to
2454 NIST standard SP 800-90Ar1. The new random generator is essentially
2455 a port of the default random generator from the OpenSSL FIPS 2.0
2456 object module. It is a hybrid deterministic random bit generator
2457 using an AES-CTR bit stream and which seeds and reseeds itself
2458 automatically using trusted system entropy sources.
2460 Some of its new features are:
2461 - Support for multiple DRBG instances with seed chaining.
2462 - The default RAND method makes use of a DRBG.
2463 - There is a public and private DRBG instance.
2464 - The DRBG instances are fork-safe.
2465 - Keep all global DRBG instances on the secure heap if it is enabled.
2466 - The public and private DRBG instance are per thread for lock free
2469 *Paul Dale, Benjamin Kaduk, Kurt Roeckx, Rich Salz, Matthias St. Pierre*
2471 * Changed Configure so it only says what it does and doesn't dump
2472 so much data. Instead, ./configdata.pm should be used as a script
2473 to display all sorts of configuration data.
2477 * Added processing of "make variables" to Configure.
2481 * Added SHA512/224 and SHA512/256 algorithm support.
2485 * The last traces of Netware support, first removed in 1.1.0, have
2490 * Get rid of Makefile.shared, and in the process, make the processing
2491 of certain files (rc.obj, or the .def/.map/.opt files produced from
2492 the ordinal files) more visible and hopefully easier to trace and
2493 debug (or make silent).
2497 * Make it possible to have environment variable assignments as
2498 arguments to config / Configure.
2502 * Add multi-prime RSA (RFC 8017) support.
2506 * Add SM3 implemented according to GB/T 32905-2016
2507 *Jack Lloyd <jack.lloyd@ribose.com>,*
2508 *Ronald Tse <ronald.tse@ribose.com>,*
2509 *Erick Borsboom <erick.borsboom@ribose.com>*
2511 * Add 'Maximum Fragment Length' TLS extension negotiation and support
2512 as documented in RFC6066.
2513 Based on a patch from Tomasz Moń
2515 *Filipe Raimundo da Silva*
2517 * Add SM4 implemented according to GB/T 32907-2016.
2518 *Jack Lloyd <jack.lloyd@ribose.com>,*
2519 *Ronald Tse <ronald.tse@ribose.com>,*
2520 *Erick Borsboom <erick.borsboom@ribose.com>*
2522 * Reimplement -newreq-nodes and ERR_error_string_n; the
2523 original author does not agree with the license change.
2527 * Add ARIA AEAD TLS support.
2531 * Some macro definitions to support VS6 have been removed. Visual
2532 Studio 6 has not worked since 1.1.0
2536 * Add ERR_clear_last_mark(), to allow callers to clear the last mark
2537 without clearing the errors.
2541 * Add "atfork" functions. If building on a system that without
2542 pthreads, see doc/man3/OPENSSL_fork_prepare.pod for application
2543 requirements. The RAND facility now uses/requires this.
2551 * The UI API becomes a permanent and integral part of libcrypto, i.e.
2552 not possible to disable entirely. However, it's still possible to
2553 disable the console reading UI method, UI_OpenSSL() (use UI_null()
2556 To disable, configure with 'no-ui-console'. 'no-ui' is still
2557 possible to use as an alias. Check at compile time with the
2558 macro OPENSSL_NO_UI_CONSOLE. The macro OPENSSL_NO_UI is still
2559 possible to check and is an alias for OPENSSL_NO_UI_CONSOLE.
2563 * Add a STORE module, which implements a uniform and URI based reader of
2564 stores that can contain keys, certificates, CRLs and numerous other
2565 objects. The main API is loosely based on a few stdio functions,
2566 and includes OSSL_STORE_open, OSSL_STORE_load, OSSL_STORE_eof,
2567 OSSL_STORE_error and OSSL_STORE_close.
2568 The implementation uses backends called "loaders" to implement arbitrary
2569 URI schemes. There is one built in "loader" for the 'file' scheme.
2573 * Add devcrypto engine. This has been implemented against cryptodev-linux,
2574 then adjusted to work on FreeBSD 8.4 as well.
2575 Enable by configuring with 'enable-devcryptoeng'. This is done by default
2576 on BSD implementations, as cryptodev.h is assumed to exist on all of them.
2580 * Module names can prefixed with OSSL_ or OPENSSL_. This affects
2581 util/mkerr.pl, which is adapted to allow those prefixes, leading to
2582 error code calls like this:
2584 OSSL_FOOerr(OSSL_FOO_F_SOMETHING, OSSL_FOO_R_WHATEVER);
2586 With this change, we claim the namespaces OSSL and OPENSSL in a manner
2587 that can be encoded in C. For the foreseeable future, this will only
2590 *Richard Levitte and Tim Hudson*
2592 * Removed BSD cryptodev engine.
2596 * Add a build target 'build_all_generated', to build all generated files
2597 and only that. This can be used to prepare everything that requires
2598 things like perl for a system that lacks perl and then move everything
2599 to that system and do the rest of the build there.
2603 * In the UI interface, make it possible to duplicate the user data. This
2604 can be used by engines that need to retain the data for a longer time
2605 than just the call where this user data is passed.
2609 * Ignore the '-named_curve auto' value for compatibility of applications
2612 *Tomáš Mráz <tmraz@fedoraproject.org>*
2614 * Fragmented SSL/TLS alerts are no longer accepted. An alert message is 2
2615 bytes long. In theory it is permissible in SSLv3 - TLSv1.2 to fragment such
2616 alerts across multiple records (some of which could be empty). In practice
2617 it make no sense to send an empty alert record, or to fragment one. TLSv1.3
2618 prohibits this altogether and other libraries (BoringSSL, NSS) do not
2619 support this at all. Supporting it adds significant complexity to the
2620 record layer, and its removal is unlikely to cause interoperability
2625 * Add the ASN.1 types INT32, UINT32, INT64, UINT64 and variants prefixed
2626 with Z. These are meant to replace LONG and ZLONG and to be size safe.
2627 The use of LONG and ZLONG is discouraged and scheduled for deprecation
2632 * Add the 'z' and 'j' modifiers to BIO_printf() et al formatting string,
2633 'z' is to be used for [s]size_t, and 'j' - with [u]int64_t.
2635 *Richard Levitte, Andy Polyakov*
2637 * Add EC_KEY_get0_engine(), which does for EC_KEY what RSA_get0_engine()
2642 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
2643 platform rather than 'mingw'.
2647 * The functions X509_STORE_add_cert and X509_STORE_add_crl return
2648 success if they are asked to add an object which already exists
2649 in the store. This change cascades to other functions which load
2650 certificates and CRLs.
2654 * x86_64 assembly pack: annotate code with DWARF CFI directives to
2655 facilitate stack unwinding even from assembly subroutines.
2659 * Remove VAX C specific definitions of OPENSSL_EXPORT, OPENSSL_EXTERN.
2660 Also remove OPENSSL_GLOBAL entirely, as it became a no-op.
2664 * Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
2665 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
2666 which is the minimum version we support.
2670 * Certificate time validation (X509_cmp_time) enforces stricter
2671 compliance with RFC 5280. Fractional seconds and timezone offsets
2672 are no longer allowed.
2676 * Add support for ARIA
2680 * s_client will now send the Server Name Indication (SNI) extension by
2681 default unless the new "-noservername" option is used. The server name is
2682 based on the host provided to the "-connect" option unless overridden by
2683 using "-servername".
2687 * Add support for SipHash
2691 * OpenSSL now fails if it receives an unrecognised record type in TLS1.0
2692 or TLS1.1. Previously this only happened in SSLv3 and TLS1.2. This is to
2693 prevent issues where no progress is being made and the peer continually
2694 sends unrecognised record types, using up resources processing them.
2698 * 'openssl passwd' can now produce SHA256 and SHA512 based output,
2699 using the algorithm defined in
2700 <https://www.akkadia.org/drepper/SHA-crypt.txt>
2704 * Heartbeat support has been removed; the ABI is changed for now.
2706 *Richard Levitte, Rich Salz*
2708 * Support for SSL_OP_NO_ENCRYPT_THEN_MAC in SSL_CONF_cmd.
2712 * The RSA "null" method, which was partially supported to avoid patent
2713 issues, has been replaced to always returns NULL.
2720 ### Changes between 1.1.0k and 1.1.0l [10 Sep 2019]
2722 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
2723 used even when parsing explicit parameters, when loading a encoded key
2724 or calling `EC_GROUP_new_from_ecpkparameters()`/
2725 `EC_GROUP_new_from_ecparameters()`.
2726 This prevents bypass of security hardening and performance gains,
2727 especially for curves with specialized EC_METHODs.
2728 By default, if a key encoded with explicit parameters is loaded and later
2729 encoded, the output is still encoded with explicit parameters, even if
2730 internally a "named" EC_GROUP is used for computation.
2734 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
2735 this change, EC_GROUP_set_generator would accept order and/or cofactor as
2736 NULL. After this change, only the cofactor parameter can be NULL. It also
2737 does some minimal sanity checks on the passed order.
2742 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
2743 An attack is simple, if the first CMS_recipientInfo is valid but the
2744 second CMS_recipientInfo is chosen ciphertext. If the second
2745 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
2746 encryption key will be replaced by garbage, and the message cannot be
2747 decoded, but if the RSA decryption fails, the correct encryption key is
2748 used and the recipient will not notice the attack.
2749 As a work around for this potential attack the length of the decrypted
2750 key must be equal to the cipher default key length, in case the
2751 certifiate is not given and all recipientInfo are tried out.
2752 The old behaviour can be re-enabled in the CMS code by setting the
2753 CMS_DEBUG_DECRYPT flag.
2758 * Use Windows installation paths in the mingw builds
2760 Mingw isn't a POSIX environment per se, which means that Windows
2761 paths should be used for installation.
2766 ### Changes between 1.1.0j and 1.1.0k [28 May 2019]
2768 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
2769 This changes the size when using the `genpkey` command when no size is given.
2770 It fixes an omission in earlier changes that changed all RSA, DSA and DH
2771 generation commands to use 2048 bits by default.
2775 * Prevent over long nonces in ChaCha20-Poly1305.
2777 ChaCha20-Poly1305 is an AEAD cipher, and requires a unique nonce input
2778 for every encryption operation. RFC 7539 specifies that the nonce value
2779 (IV) should be 96 bits (12 bytes). OpenSSL allows a variable nonce length
2780 and front pads the nonce with 0 bytes if it is less than 12
2781 bytes. However it also incorrectly allows a nonce to be set of up to 16
2782 bytes. In this case only the last 12 bytes are significant and any
2783 additional leading bytes are ignored.
2785 It is a requirement of using this cipher that nonce values are
2786 unique. Messages encrypted using a reused nonce value are susceptible to
2787 serious confidentiality and integrity attacks. If an application changes
2788 the default nonce length to be longer than 12 bytes and then makes a
2789 change to the leading bytes of the nonce expecting the new value to be a
2790 new unique nonce then such an application could inadvertently encrypt
2791 messages with a reused nonce.
2793 Additionally the ignored bytes in a long nonce are not covered by the
2794 integrity guarantee of this cipher. Any application that relies on the
2795 integrity of these ignored leading bytes of a long nonce may be further
2796 affected. Any OpenSSL internal use of this cipher, including in SSL/TLS,
2797 is safe because no such use sets such a long nonce value. However user
2798 applications that use this cipher directly and set a non-default nonce
2799 length to be longer than 12 bytes may be vulnerable.
2801 This issue was reported to OpenSSL on 16th of March 2019 by Joran Dirk
2807 * Added SCA hardening for modular field inversion in EC_GROUP through
2808 a new dedicated field_inv() pointer in EC_METHOD.
2809 This also addresses a leakage affecting conversions from projective
2810 to affine coordinates.
2812 *Billy Bob Brumley, Nicola Tuveri*
2814 * Fix a use after free bug in d2i_X509_PUBKEY when overwriting a
2815 re-used X509_PUBKEY object if the second PUBKEY is malformed.
2819 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
2823 * Remove the 'dist' target and add a tarball building script. The
2824 'dist' target has fallen out of use, and it shouldn't be
2825 necessary to configure just to create a source distribution.
2829 ### Changes between 1.1.0i and 1.1.0j [20 Nov 2018]
2831 * Timing vulnerability in DSA signature generation
2833 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
2834 timing side channel attack. An attacker could use variations in the signing
2835 algorithm to recover the private key.
2837 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
2842 * Timing vulnerability in ECDSA signature generation
2844 The OpenSSL ECDSA signature algorithm has been shown to be vulnerable to a
2845 timing side channel attack. An attacker could use variations in the signing
2846 algorithm to recover the private key.
2848 This issue was reported to OpenSSL on 25th October 2018 by Samuel Weiser.
2853 * Add coordinate blinding for EC_POINT and implement projective
2854 coordinate blinding for generic prime curves as a countermeasure to
2855 chosen point SCA attacks.
2857 *Sohaib ul Hassan, Nicola Tuveri, Billy Bob Brumley*
2859 ### Changes between 1.1.0h and 1.1.0i [14 Aug 2018]
2861 * Client DoS due to large DH parameter
2863 During key agreement in a TLS handshake using a DH(E) based ciphersuite a
2864 malicious server can send a very large prime value to the client. This will
2865 cause the client to spend an unreasonably long period of time generating a
2866 key for this prime resulting in a hang until the client has finished. This
2867 could be exploited in a Denial Of Service attack.
2869 This issue was reported to OpenSSL on 5th June 2018 by Guido Vranken
2874 * Cache timing vulnerability in RSA Key Generation
2876 The OpenSSL RSA Key generation algorithm has been shown to be vulnerable to
2877 a cache timing side channel attack. An attacker with sufficient access to
2878 mount cache timing attacks during the RSA key generation process could
2879 recover the private key.
2881 This issue was reported to OpenSSL on 4th April 2018 by Alejandro Cabrera
2882 Aldaya, Billy Brumley, Cesar Pereida Garcia and Luis Manuel Alvarez Tapia.
2887 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
2888 parameter is no longer accepted, as it leads to a corrupt table. NULL
2889 pem_str is reserved for alias entries only.
2893 * Revert blinding in ECDSA sign and instead make problematic addition
2894 length-invariant. Switch even to fixed-length Montgomery multiplication.
2898 * Change generating and checking of primes so that the error rate of not
2899 being prime depends on the intended use based on the size of the input.
2900 For larger primes this will result in more rounds of Miller-Rabin.
2901 The maximal error rate for primes with more than 1080 bits is lowered
2904 *Kurt Roeckx, Annie Yousar*
2906 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
2910 * Add blinding to ECDSA and DSA signatures to protect against side channel
2911 attacks discovered by Keegan Ryan (NCC Group).
2915 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
2916 now allow empty (zero character) pass phrases.
2920 * Certificate time validation (X509_cmp_time) enforces stricter
2921 compliance with RFC 5280. Fractional seconds and timezone offsets
2922 are no longer allowed.
2926 * Fixed a text canonicalisation bug in CMS
2928 Where a CMS detached signature is used with text content the text goes
2929 through a canonicalisation process first prior to signing or verifying a
2930 signature. This process strips trailing space at the end of lines, converts
2931 line terminators to CRLF and removes additional trailing line terminators
2932 at the end of a file. A bug in the canonicalisation process meant that
2933 some characters, such as form-feed, were incorrectly treated as whitespace
2934 and removed. This is contrary to the specification (RFC5485). This fix
2935 could mean that detached text data signed with an earlier version of
2936 OpenSSL 1.1.0 may fail to verify using the fixed version, or text data
2937 signed with a fixed OpenSSL may fail to verify with an earlier version of
2938 OpenSSL 1.1.0. A workaround is to only verify the canonicalised text data
2939 and use the "-binary" flag (for the "cms" command line application) or set
2940 the SMIME_BINARY/PKCS7_BINARY/CMS_BINARY flags (if using CMS_verify()).
2944 ### Changes between 1.1.0g and 1.1.0h [27 Mar 2018]
2946 * Constructed ASN.1 types with a recursive definition could exceed the stack
2948 Constructed ASN.1 types with a recursive definition (such as can be found
2949 in PKCS7) could eventually exceed the stack given malicious input with
2950 excessive recursion. This could result in a Denial Of Service attack. There
2951 are no such structures used within SSL/TLS that come from untrusted sources
2952 so this is considered safe.
2954 This issue was reported to OpenSSL on 4th January 2018 by the OSS-fuzz
2960 * Incorrect CRYPTO_memcmp on HP-UX PA-RISC
2962 Because of an implementation bug the PA-RISC CRYPTO_memcmp function is
2963 effectively reduced to only comparing the least significant bit of each
2964 byte. This allows an attacker to forge messages that would be considered as
2965 authenticated in an amount of tries lower than that guaranteed by the
2966 security claims of the scheme. The module can only be compiled by the
2967 HP-UX assembler, so that only HP-UX PA-RISC targets are affected.
2969 This issue was reported to OpenSSL on 2nd March 2018 by Peter Waltenberg
2975 * Add a build target 'build_all_generated', to build all generated files
2976 and only that. This can be used to prepare everything that requires
2977 things like perl for a system that lacks perl and then move everything
2978 to that system and do the rest of the build there.
2982 * Backport SSL_OP_NO_RENGOTIATION
2984 OpenSSL 1.0.2 and below had the ability to disable renegotiation using the
2985 (undocumented) SSL3_FLAGS_NO_RENEGOTIATE_CIPHERS flag. Due to the opacity
2986 changes this is no longer possible in 1.1.0. Therefore the new
2987 SSL_OP_NO_RENEGOTIATION option from 1.1.1-dev has been backported to
2988 1.1.0 to provide equivalent functionality.
2990 Note that if an application built against 1.1.0h headers (or above) is run
2991 using an older version of 1.1.0 (prior to 1.1.0h) then the option will be
2992 accepted but nothing will happen, i.e. renegotiation will not be prevented.
2996 * Removed the OS390-Unix config target. It relied on a script that doesn't
3001 * rsaz_1024_mul_avx2 overflow bug on x86_64
3003 There is an overflow bug in the AVX2 Montgomery multiplication procedure
3004 used in exponentiation with 1024-bit moduli. No EC algorithms are affected.
3005 Analysis suggests that attacks against RSA and DSA as a result of this
3006 defect would be very difficult to perform and are not believed likely.
3007 Attacks against DH1024 are considered just feasible, because most of the
3008 work necessary to deduce information about a private key may be performed
3009 offline. The amount of resources required for such an attack would be
3010 significant. However, for an attack on TLS to be meaningful, the server
3011 would have to share the DH1024 private key among multiple clients, which is
3012 no longer an option since CVE-2016-0701.
3014 This only affects processors that support the AVX2 but not ADX extensions
3015 like Intel Haswell (4th generation).
3017 This issue was reported to OpenSSL by David Benjamin (Google). The issue
3018 was originally found via the OSS-Fuzz project.
3023 ### Changes between 1.1.0f and 1.1.0g [2 Nov 2017]
3025 * bn_sqrx8x_internal carry bug on x86_64
3027 There is a carry propagating bug in the x86_64 Montgomery squaring
3028 procedure. No EC algorithms are affected. Analysis suggests that attacks
3029 against RSA and DSA as a result of this defect would be very difficult to
3030 perform and are not believed likely. Attacks against DH are considered just
3031 feasible (although very difficult) because most of the work necessary to
3032 deduce information about a private key may be performed offline. The amount
3033 of resources required for such an attack would be very significant and
3034 likely only accessible to a limited number of attackers. An attacker would
3035 additionally need online access to an unpatched system using the target
3036 private key in a scenario with persistent DH parameters and a private
3037 key that is shared between multiple clients.
3039 This only affects processors that support the BMI1, BMI2 and ADX extensions
3040 like Intel Broadwell (5th generation) and later or AMD Ryzen.
3042 This issue was reported to OpenSSL by the OSS-Fuzz project.
3047 * Malformed X.509 IPAddressFamily could cause OOB read
3049 If an X.509 certificate has a malformed IPAddressFamily extension,
3050 OpenSSL could do a one-byte buffer overread. The most likely result
3051 would be an erroneous display of the certificate in text format.
3053 This issue was reported to OpenSSL by the OSS-Fuzz project.
3058 ### Changes between 1.1.0e and 1.1.0f [25 May 2017]
3060 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
3061 platform rather than 'mingw'.
3065 * Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
3066 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
3067 which is the minimum version we support.
3071 ### Changes between 1.1.0d and 1.1.0e [16 Feb 2017]
3073 * Encrypt-Then-Mac renegotiation crash
3075 During a renegotiation handshake if the Encrypt-Then-Mac extension is
3076 negotiated where it was not in the original handshake (or vice-versa) then
3077 this can cause OpenSSL to crash (dependant on ciphersuite). Both clients
3078 and servers are affected.
3080 This issue was reported to OpenSSL by Joe Orton (Red Hat).
3085 ### Changes between 1.1.0c and 1.1.0d [26 Jan 2017]
3087 * Truncated packet could crash via OOB read
3089 If one side of an SSL/TLS path is running on a 32-bit host and a specific
3090 cipher is being used, then a truncated packet can cause that host to
3091 perform an out-of-bounds read, usually resulting in a crash.
3093 This issue was reported to OpenSSL by Robert Święcki of Google.
3098 * Bad (EC)DHE parameters cause a client crash
3100 If a malicious server supplies bad parameters for a DHE or ECDHE key
3101 exchange then this can result in the client attempting to dereference a
3102 NULL pointer leading to a client crash. This could be exploited in a Denial
3105 This issue was reported to OpenSSL by Guido Vranken.
3110 * BN_mod_exp may produce incorrect results on x86_64
3112 There is a carry propagating bug in the x86_64 Montgomery squaring
3113 procedure. No EC algorithms are affected. Analysis suggests that attacks
3114 against RSA and DSA as a result of this defect would be very difficult to
3115 perform and are not believed likely. Attacks against DH are considered just
3116 feasible (although very difficult) because most of the work necessary to
3117 deduce information about a private key may be performed offline. The amount
3118 of resources required for such an attack would be very significant and
3119 likely only accessible to a limited number of attackers. An attacker would
3120 additionally need online access to an unpatched system using the target
3121 private key in a scenario with persistent DH parameters and a private
3122 key that is shared between multiple clients. For example this can occur by
3123 default in OpenSSL DHE based SSL/TLS ciphersuites. Note: This issue is very
3124 similar to CVE-2015-3193 but must be treated as a separate problem.
3126 This issue was reported to OpenSSL by the OSS-Fuzz project.
3131 ### Changes between 1.1.0b and 1.1.0c [10 Nov 2016]
3133 * ChaCha20/Poly1305 heap-buffer-overflow
3135 TLS connections using `*-CHACHA20-POLY1305` ciphersuites are susceptible to
3136 a DoS attack by corrupting larger payloads. This can result in an OpenSSL
3137 crash. This issue is not considered to be exploitable beyond a DoS.
3139 This issue was reported to OpenSSL by Robert Święcki (Google Security Team)
3144 * CMS Null dereference
3146 Applications parsing invalid CMS structures can crash with a NULL pointer
3147 dereference. This is caused by a bug in the handling of the ASN.1 CHOICE
3148 type in OpenSSL 1.1.0 which can result in a NULL value being passed to the
3149 structure callback if an attempt is made to free certain invalid encodings.
3150 Only CHOICE structures using a callback which do not handle NULL value are
3153 This issue was reported to OpenSSL by Tyler Nighswander of ForAllSecure.
3158 * Montgomery multiplication may produce incorrect results
3160 There is a carry propagating bug in the Broadwell-specific Montgomery
3161 multiplication procedure that handles input lengths divisible by, but
3162 longer than 256 bits. Analysis suggests that attacks against RSA, DSA
3163 and DH private keys are impossible. This is because the subroutine in
3164 question is not used in operations with the private key itself and an input
3165 of the attacker's direct choice. Otherwise the bug can manifest itself as
3166 transient authentication and key negotiation failures or reproducible
3167 erroneous outcome of public-key operations with specially crafted input.
3168 Among EC algorithms only Brainpool P-512 curves are affected and one
3169 presumably can attack ECDH key negotiation. Impact was not analyzed in
3170 detail, because pre-requisites for attack are considered unlikely. Namely
3171 multiple clients have to choose the curve in question and the server has to
3172 share the private key among them, neither of which is default behaviour.
3173 Even then only clients that chose the curve will be affected.
3175 This issue was publicly reported as transient failures and was not
3176 initially recognized as a security issue. Thanks to Richard Morgan for
3177 providing reproducible case.
3182 * Removed automatic addition of RPATH in shared libraries and executables,
3183 as this was a remainder from OpenSSL 1.0.x and isn't needed any more.
3187 ### Changes between 1.1.0a and 1.1.0b [26 Sep 2016]
3189 * Fix Use After Free for large message sizes
3191 The patch applied to address CVE-2016-6307 resulted in an issue where if a
3192 message larger than approx 16k is received then the underlying buffer to
3193 store the incoming message is reallocated and moved. Unfortunately a
3194 dangling pointer to the old location is left which results in an attempt to
3195 write to the previously freed location. This is likely to result in a
3196 crash, however it could potentially lead to execution of arbitrary code.
3198 This issue only affects OpenSSL 1.1.0a.
3200 This issue was reported to OpenSSL by Robert Święcki.
3205 ### Changes between 1.1.0 and 1.1.0a [22 Sep 2016]
3207 * OCSP Status Request extension unbounded memory growth
3209 A malicious client can send an excessively large OCSP Status Request
3210 extension. If that client continually requests renegotiation, sending a
3211 large OCSP Status Request extension each time, then there will be unbounded
3212 memory growth on the server. This will eventually lead to a Denial Of
3213 Service attack through memory exhaustion. Servers with a default
3214 configuration are vulnerable even if they do not support OCSP. Builds using
3215 the "no-ocsp" build time option are not affected.
3217 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
3222 * SSL_peek() hang on empty record
3224 OpenSSL 1.1.0 SSL/TLS will hang during a call to SSL_peek() if the peer
3225 sends an empty record. This could be exploited by a malicious peer in a
3226 Denial Of Service attack.
3228 This issue was reported to OpenSSL by Alex Gaynor.
3233 * Excessive allocation of memory in tls_get_message_header() and
3234 dtls1_preprocess_fragment()
3236 A (D)TLS message includes 3 bytes for its length in the header for the
3237 message. This would allow for messages up to 16Mb in length. Messages of
3238 this length are excessive and OpenSSL includes a check to ensure that a
3239 peer is sending reasonably sized messages in order to avoid too much memory
3240 being consumed to service a connection. A flaw in the logic of version
3241 1.1.0 means that memory for the message is allocated too early, prior to
3242 the excessive message length check. Due to way memory is allocated in
3243 OpenSSL this could mean an attacker could force up to 21Mb to be allocated
3244 to service a connection. This could lead to a Denial of Service through
3245 memory exhaustion. However, the excessive message length check still takes
3246 place, and this would cause the connection to immediately fail. Assuming
3247 that the application calls SSL_free() on the failed connection in a timely
3248 manner then the 21Mb of allocated memory will then be immediately freed
3249 again. Therefore the excessive memory allocation will be transitory in
3250 nature. This then means that there is only a security impact if:
3252 1) The application does not call SSL_free() in a timely manner in the event
3253 that the connection fails
3255 2) The application is working in a constrained environment where there is
3256 very little free memory
3258 3) The attacker initiates multiple connection attempts such that there are
3259 multiple connections in a state where memory has been allocated for the
3260 connection; SSL_free() has not yet been called; and there is insufficient
3261 memory to service the multiple requests.
3263 Except in the instance of (1) above any Denial Of Service is likely to be
3264 transitory because as soon as the connection fails the memory is
3265 subsequently freed again in the SSL_free() call. However there is an
3266 increased risk during this period of application crashes due to the lack of
3267 memory - which would then mean a more serious Denial of Service.
3269 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
3270 (CVE-2016-6307 and CVE-2016-6308)
3274 * solaris-x86-cc, i.e. 32-bit configuration with vendor compiler,
3275 had to be removed. Primary reason is that vendor assembler can't
3276 assemble our modules with -KPIC flag. As result it, assembly
3277 support, was not even available as option. But its lack means
3278 lack of side-channel resistant code, which is incompatible with
3279 security by todays standards. Fortunately gcc is readily available
3280 prepackaged option, which we firmly point at...
3284 ### Changes between 1.0.2h and 1.1.0 [25 Aug 2016]
3286 * Windows command-line tool supports UTF-8 opt-in option for arguments
3287 and console input. Setting OPENSSL_WIN32_UTF8 environment variable
3288 (to any value) allows Windows user to access PKCS#12 file generated
3289 with Windows CryptoAPI and protected with non-ASCII password, as well
3290 as files generated under UTF-8 locale on Linux also protected with
3295 * To mitigate the SWEET32 attack ([CVE-2016-2183]), 3DES cipher suites
3296 have been disabled by default and removed from DEFAULT, just like RC4.
3297 See the RC4 item below to re-enable both.
3301 * The method for finding the storage location for the Windows RAND seed file
3302 has changed. First we check %RANDFILE%. If that is not set then we check
3303 the directories %HOME%, %USERPROFILE% and %SYSTEMROOT% in that order. If
3304 all else fails we fall back to C:\.
3308 * The EVP_EncryptUpdate() function has had its return type changed from void
3309 to int. A return of 0 indicates and error while a return of 1 indicates
3314 * The flags RSA_FLAG_NO_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME and
3315 DH_FLAG_NO_EXP_CONSTTIME which previously provided the ability to switch
3316 off the constant time implementation for RSA, DSA and DH have been made
3317 no-ops and deprecated.
3321 * Windows RAND implementation was simplified to only get entropy by
3322 calling CryptGenRandom(). Various other RAND-related tickets
3325 *Joseph Wylie Yandle, Rich Salz*
3327 * The stack and lhash API's were renamed to start with `OPENSSL_SK_`
3328 and `OPENSSL_LH_`, respectively. The old names are available
3329 with API compatibility. They new names are now completely documented.
3333 * Unify TYPE_up_ref(obj) methods signature.
3334 SSL_CTX_up_ref(), SSL_up_ref(), X509_up_ref(), EVP_PKEY_up_ref(),
3335 X509_CRL_up_ref(), X509_OBJECT_up_ref_count() methods are now returning an
3336 int (instead of void) like all others TYPE_up_ref() methods.
3337 So now these methods also check the return value of CRYPTO_atomic_add(),
3338 and the validity of object reference counter.
3340 *fdasilvayy@gmail.com*
3342 * With Windows Visual Studio builds, the .pdb files are installed
3343 alongside the installed libraries and executables. For a static
3344 library installation, ossl_static.pdb is the associate compiler
3345 generated .pdb file to be used when linking programs.
3349 * Remove openssl.spec. Packaging files belong with the packagers.
3353 * Automatic Darwin/OSX configuration has had a refresh, it will now
3354 recognise x86_64 architectures automatically. You can still decide
3355 to build for a different bitness with the environment variable
3356 KERNEL_BITS (can be 32 or 64), for example:
3358 KERNEL_BITS=32 ./config
3362 * Change default algorithms in pkcs8 utility to use PKCS#5 v2.0,
3363 256 bit AES and HMAC with SHA256.
3367 * Remove support for MIPS o32 ABI on IRIX (and IRIX only).
3371 * Triple-DES ciphers have been moved from HIGH to MEDIUM.
3375 * To enable users to have their own config files and build file templates,
3376 Configure looks in the directory indicated by the environment variable
3377 OPENSSL_LOCAL_CONFIG_DIR as well as the in-source Configurations/
3378 directory. On VMS, OPENSSL_LOCAL_CONFIG_DIR is expected to be a logical
3379 name and is used as is.
3383 * The following datatypes were made opaque: X509_OBJECT, X509_STORE_CTX,
3384 X509_STORE, X509_LOOKUP, and X509_LOOKUP_METHOD. The unused type
3385 X509_CERT_FILE_CTX was removed.
3389 * "shared" builds are now the default. To create only static libraries use
3390 the "no-shared" Configure option.
3394 * Remove the no-aes, no-hmac, no-rsa, no-sha and no-md5 Configure options.
3395 All of these option have not worked for some while and are fundamental
3400 * Make various cleanup routines no-ops and mark them as deprecated. Most
3401 global cleanup functions are no longer required because they are handled
3402 via auto-deinit (see OPENSSL_init_crypto and OPENSSL_init_ssl man pages).
3403 Explicitly de-initing can cause problems (e.g. where a library that uses
3404 OpenSSL de-inits, but an application is still using it). The affected
3405 functions are CONF_modules_free(), ENGINE_cleanup(), OBJ_cleanup(),
3406 EVP_cleanup(), BIO_sock_cleanup(), CRYPTO_cleanup_all_ex_data(),
3407 RAND_cleanup(), SSL_COMP_free_compression_methods(), ERR_free_strings() and
3408 COMP_zlib_cleanup().
3412 * --strict-warnings no longer enables runtime debugging options
3413 such as REF_DEBUG. Instead, debug options are automatically
3414 enabled with '--debug' builds.
3416 *Andy Polyakov, Emilia Käsper*
3418 * Made DH and DH_METHOD opaque. The structures for managing DH objects
3419 have been moved out of the public header files. New functions for managing
3420 these have been added.
3424 * Made RSA and RSA_METHOD opaque. The structures for managing RSA
3425 objects have been moved out of the public header files. New
3426 functions for managing these have been added.
3430 * Made DSA and DSA_METHOD opaque. The structures for managing DSA objects
3431 have been moved out of the public header files. New functions for managing
3432 these have been added.
3436 * Made BIO and BIO_METHOD opaque. The structures for managing BIOs have been
3437 moved out of the public header files. New functions for managing these
3442 * Removed no-rijndael as a config option. Rijndael is an old name for AES.
3446 * Removed the mk1mf build scripts.
3450 * Headers are now wrapped, if necessary, with OPENSSL_NO_xxx, so
3451 it is always safe to #include a header now.
3455 * Removed the aged BC-32 config and all its supporting scripts
3459 * Removed support for Ultrix, Netware, and OS/2.
3463 * Add support for HKDF.
3465 *Alessandro Ghedini*
3467 * Add support for blake2b and blake2s
3471 * Added support for "pipelining". Ciphers that have the
3472 EVP_CIPH_FLAG_PIPELINE flag set have a capability to process multiple
3473 encryptions/decryptions simultaneously. There are currently no built-in
3474 ciphers with this property but the expectation is that engines will be able
3475 to offer it to significantly improve throughput. Support has been extended
3476 into libssl so that multiple records for a single connection can be
3477 processed in one go (for >=TLS 1.1).
3481 * Added the AFALG engine. This is an async capable engine which is able to
3482 offload work to the Linux kernel. In this initial version it only supports
3483 AES128-CBC. The kernel must be version 4.1.0 or greater.
3487 * OpenSSL now uses a new threading API. It is no longer necessary to
3488 set locking callbacks to use OpenSSL in a multi-threaded environment. There
3489 are two supported threading models: pthreads and windows threads. It is
3490 also possible to configure OpenSSL at compile time for "no-threads". The
3491 old threading API should no longer be used. The functions have been
3492 replaced with "no-op" compatibility macros.
3494 *Alessandro Ghedini, Matt Caswell*
3496 * Modify behavior of ALPN to invoke callback after SNI/servername
3497 callback, such that updates to the SSL_CTX affect ALPN.
3501 * Add SSL_CIPHER queries for authentication and key-exchange.
3505 * Changes to the DEFAULT cipherlist:
3506 - Prefer (EC)DHE handshakes over plain RSA.
3507 - Prefer AEAD ciphers over legacy ciphers.
3508 - Prefer ECDSA over RSA when both certificates are available.
3509 - Prefer TLSv1.2 ciphers/PRF.
3510 - Remove DSS, SEED, IDEA, CAMELLIA, and AES-CCM from the
3515 * Change the ECC default curve list to be this, in order: x25519,
3516 secp256r1, secp521r1, secp384r1.
3520 * RC4 based libssl ciphersuites are now classed as "weak" ciphers and are
3521 disabled by default. They can be re-enabled using the
3522 enable-weak-ssl-ciphers option to Configure.
3526 * If the server has ALPN configured, but supports no protocols that the
3527 client advertises, send a fatal "no_application_protocol" alert.
3528 This behaviour is SHALL in RFC 7301, though it isn't universally
3529 implemented by other servers.
3533 * Add X25519 support.
3534 Add ASN.1 and EVP_PKEY methods for X25519. This includes support
3535 for public and private key encoding using the format documented in
3536 draft-ietf-curdle-pkix-02. The corresponding EVP_PKEY method supports
3537 key generation and key derivation.
3539 TLS support complies with draft-ietf-tls-rfc4492bis-08 and uses
3544 * Deprecate SRP_VBASE_get_by_user.
3545 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
3546 In order to fix an unavoidable memory leak ([CVE-2016-0798]),
3547 SRP_VBASE_get_by_user was changed to ignore the "fake user" SRP
3548 seed, even if the seed is configured.
3550 Users should use SRP_VBASE_get1_by_user instead. Note that in
3551 SRP_VBASE_get1_by_user, caller must free the returned value. Note
3552 also that even though configuring the SRP seed attempts to hide
3553 invalid usernames by continuing the handshake with fake
3554 credentials, this behaviour is not constant time and no strong
3555 guarantees are made that the handshake is indistinguishable from
3556 that of a valid user.
3560 * Configuration change; it's now possible to build dynamic engines
3561 without having to build shared libraries and vice versa. This
3562 only applies to the engines in `engines/`, those in `crypto/engine/`
3563 will always be built into libcrypto (i.e. "static").
3565 Building dynamic engines is enabled by default; to disable, use
3566 the configuration option "disable-dynamic-engine".
3568 The only requirements for building dynamic engines are the
3569 presence of the DSO module and building with position independent
3570 code, so they will also automatically be disabled if configuring
3571 with "disable-dso" or "disable-pic".
3573 The macros OPENSSL_NO_STATIC_ENGINE and OPENSSL_NO_DYNAMIC_ENGINE
3574 are also taken away from openssl/opensslconf.h, as they are
3579 * Configuration change; if there is a known flag to compile
3580 position independent code, it will always be applied on the
3581 libcrypto and libssl object files, and never on the application
3582 object files. This means other libraries that use routines from
3583 libcrypto / libssl can be made into shared libraries regardless
3584 of how OpenSSL was configured.
3586 If this isn't desirable, the configuration options "disable-pic"
3587 or "no-pic" can be used to disable the use of PIC. This will
3588 also disable building shared libraries and dynamic engines.
3592 * Removed JPAKE code. It was experimental and has no wide use.
3596 * The INSTALL_PREFIX Makefile variable has been renamed to
3597 DESTDIR. That makes for less confusion on what this variable
3598 is for. Also, the configuration option --install_prefix is
3603 * Heartbeat for TLS has been removed and is disabled by default
3604 for DTLS; configure with enable-heartbeats. Code that uses the
3605 old #define's might need to be updated.
3607 *Emilia Käsper, Rich Salz*
3609 * Rename REF_CHECK to REF_DEBUG.
3613 * New "unified" build system
3615 The "unified" build system is aimed to be a common system for all
3616 platforms we support. With it comes new support for VMS.
3618 This system builds supports building in a different directory tree
3619 than the source tree. It produces one Makefile (for unix family
3620 or lookalikes), or one descrip.mms (for VMS).
3622 The source of information to make the Makefile / descrip.mms is
3623 small files called 'build.info', holding the necessary
3624 information for each directory with source to compile, and a
3625 template in Configurations, like unix-Makefile.tmpl or
3628 With this change, the library names were also renamed on Windows
3629 and on VMS. They now have names that are closer to the standard
3630 on Unix, and include the major version number, and in certain
3631 cases, the architecture they are built for. See "Notes on shared
3632 libraries" in INSTALL.
3634 We rely heavily on the perl module Text::Template.
3638 * Added support for auto-initialisation and de-initialisation of the library.
3639 OpenSSL no longer requires explicit init or deinit routines to be called,
3640 except in certain circumstances. See the OPENSSL_init_crypto() and
3641 OPENSSL_init_ssl() man pages for further information.
3645 * The arguments to the DTLSv1_listen function have changed. Specifically the
3646 "peer" argument is now expected to be a BIO_ADDR object.
3648 * Rewrite of BIO networking library. The BIO library lacked consistent
3649 support of IPv6, and adding it required some more extensive
3650 modifications. This introduces the BIO_ADDR and BIO_ADDRINFO types,
3651 which hold all types of addresses and chains of address information.
3652 It also introduces a new API, with functions like BIO_socket,
3653 BIO_connect, BIO_listen, BIO_lookup and a rewrite of BIO_accept.
3654 The source/sink BIOs BIO_s_connect, BIO_s_accept and BIO_s_datagram
3655 have been adapted accordingly.
3659 * RSA_padding_check_PKCS1_type_1 now accepts inputs with and without
3664 * CRIME protection: disable compression by default, even if OpenSSL is
3665 compiled with zlib enabled. Applications can still enable compression
3666 by calling SSL_CTX_clear_options(ctx, SSL_OP_NO_COMPRESSION), or by
3667 using the SSL_CONF library to configure compression.
3671 * The signature of the session callback configured with
3672 SSL_CTX_sess_set_get_cb was changed. The read-only input buffer
3673 was explicitly marked as `const unsigned char*` instead of
3678 * Always DPURIFY. Remove the use of uninitialized memory in the
3679 RNG, and other conditional uses of DPURIFY. This makes -DPURIFY a no-op.
3683 * Removed many obsolete configuration items, including
3684 DES_PTR, DES_RISC1, DES_RISC2, DES_INT
3685 MD2_CHAR, MD2_INT, MD2_LONG
3687 IDEA_SHORT, IDEA_LONG
3688 RC2_SHORT, RC2_LONG, RC4_LONG, RC4_CHUNK, RC4_INDEX
3690 *Rich Salz, with advice from Andy Polyakov*
3692 * Many BN internals have been moved to an internal header file.
3694 *Rich Salz with help from Andy Polyakov*
3696 * Configuration and writing out the results from it has changed.
3697 Files such as Makefile include/openssl/opensslconf.h and are now
3698 produced through general templates, such as Makefile.in and
3699 crypto/opensslconf.h.in and some help from the perl module
3702 Also, the center of configuration information is no longer
3703 Makefile. Instead, Configure produces a perl module in
3704 configdata.pm which holds most of the config data (in the hash
3705 table %config), the target data that comes from the target
3706 configuration in one of the `Configurations/*.conf` files (in
3711 * To clarify their intended purposes, the Configure options
3712 --prefix and --openssldir change their semantics, and become more
3713 straightforward and less interdependent.
3715 --prefix shall be used exclusively to give the location INSTALLTOP
3716 where programs, scripts, libraries, include files and manuals are
3717 going to be installed. The default is now /usr/local.
3719 --openssldir shall be used exclusively to give the default
3720 location OPENSSLDIR where certificates, private keys, CRLs are
3721 managed. This is also where the default openssl.cnf gets
3723 If the directory given with this option is a relative path, the
3724 values of both the --prefix value and the --openssldir value will
3725 be combined to become OPENSSLDIR.
3726 The default for --openssldir is INSTALLTOP/ssl.
3728 Anyone who uses --openssldir to specify where OpenSSL is to be
3729 installed MUST change to use --prefix instead.
3733 * The GOST engine was out of date and therefore it has been removed. An up
3734 to date GOST engine is now being maintained in an external repository.
3735 See: <https://wiki.openssl.org/index.php/Binaries>. Libssl still retains
3736 support for GOST ciphersuites (these are only activated if a GOST engine
3741 * EGD is no longer supported by default; use enable-egd when
3744 *Ben Kaduk and Rich Salz*
3746 * The distribution now has Makefile.in files, which are used to
3747 create Makefile's when Configure is run. *Configure must be run
3748 before trying to build now.*
3752 * The return value for SSL_CIPHER_description() for error conditions
3757 * Support for RFC6698/RFC7671 DANE TLSA peer authentication.
3759 Obtaining and performing DNSSEC validation of TLSA records is
3760 the application's responsibility. The application provides
3761 the TLSA records of its choice to OpenSSL, and these are then
3762 used to authenticate the peer.
3764 The TLSA records need not even come from DNS. They can, for
3765 example, be used to implement local end-entity certificate or
3766 trust-anchor "pinning", where the "pin" data takes the form
3767 of TLSA records, which can augment or replace verification
3768 based on the usual WebPKI public certification authorities.
3772 * Revert default OPENSSL_NO_DEPRECATED setting. Instead OpenSSL
3773 continues to support deprecated interfaces in default builds.
3774 However, applications are strongly advised to compile their
3775 source files with -DOPENSSL_API_COMPAT=0x10100000L, which hides
3776 the declarations of all interfaces deprecated in 0.9.8, 1.0.0
3777 or the 1.1.0 releases.
3779 In environments in which all applications have been ported to
3780 not use any deprecated interfaces OpenSSL's Configure script
3781 should be used with the --api=1.1.0 option to entirely remove
3782 support for the deprecated features from the library and
3783 unconditionally disable them in the installed headers.
3784 Essentially the same effect can be achieved with the "no-deprecated"
3785 argument to Configure, except that this will always restrict
3786 the build to just the latest API, rather than a fixed API
3789 As applications are ported to future revisions of the API,
3790 they should update their compile-time OPENSSL_API_COMPAT define
3791 accordingly, but in most cases should be able to continue to
3792 compile with later releases.
3794 The OPENSSL_API_COMPAT versions for 1.0.0, and 0.9.8 are
3795 0x10000000L and 0x00908000L, respectively. However those
3796 versions did not support the OPENSSL_API_COMPAT feature, and
3797 so applications are not typically tested for explicit support
3798 of just the undeprecated features of either release.
3802 * Add support for setting the minimum and maximum supported protocol.
3803 It can bet set via the SSL_set_min_proto_version() and
3804 SSL_set_max_proto_version(), or via the SSL_CONF's MinProtocol and
3805 MaxProtocol. It's recommended to use the new APIs to disable
3806 protocols instead of disabling individual protocols using
3807 SSL_set_options() or SSL_CONF's Protocol. This change also
3808 removes support for disabling TLS 1.2 in the OpenSSL TLS
3809 client at compile time by defining OPENSSL_NO_TLS1_2_CLIENT.
3813 * Support for ChaCha20 and Poly1305 added to libcrypto and libssl.
3817 * New EC_KEY_METHOD, this replaces the older ECDSA_METHOD and ECDH_METHOD
3818 and integrates ECDSA and ECDH functionality into EC. Implementations can
3819 now redirect key generation and no longer need to convert to or from
3822 Note: the ecdsa.h and ecdh.h headers are now no longer needed and just
3823 include the ec.h header file instead.
3827 * Remove support for all 40 and 56 bit ciphers. This includes all the export
3828 ciphers who are no longer supported and drops support the ephemeral RSA key
3829 exchange. The LOW ciphers currently doesn't have any ciphers in it.
3833 * Made EVP_MD_CTX, EVP_MD, EVP_CIPHER_CTX, EVP_CIPHER and HMAC_CTX
3834 opaque. For HMAC_CTX, the following constructors and destructors
3837 HMAC_CTX *HMAC_CTX_new(void);
3838 void HMAC_CTX_free(HMAC_CTX *ctx);
3840 For EVP_MD and EVP_CIPHER, complete APIs to create, fill and
3841 destroy such methods has been added. See EVP_MD_meth_new(3) and
3842 EVP_CIPHER_meth_new(3) for documentation.
3845 1) `EVP_MD_CTX_cleanup()`, `EVP_CIPHER_CTX_cleanup()` and
3846 `HMAC_CTX_cleanup()` were removed. `HMAC_CTX_reset()` and
3847 `EVP_MD_CTX_reset()` should be called instead to reinitialise
3848 an already created structure.
3849 2) For consistency with the majority of our object creators and
3850 destructors, `EVP_MD_CTX_(create|destroy)` were renamed to
3851 `EVP_MD_CTX_(new|free)`. The old names are retained as macros
3852 for deprecated builds.
3856 * Added ASYNC support. Libcrypto now includes the async sub-library to enable
3857 cryptographic operations to be performed asynchronously as long as an
3858 asynchronous capable engine is used. See the ASYNC_start_job() man page for
3859 further details. Libssl has also had this capability integrated with the
3860 introduction of the new mode SSL_MODE_ASYNC and associated error
3861 SSL_ERROR_WANT_ASYNC. See the SSL_CTX_set_mode() and SSL_get_error() man
3862 pages. This work was developed in partnership with Intel Corp.
3866 * SSL_{CTX_}set_ecdh_auto() has been removed and ECDH is support is
3867 always enabled now. If you want to disable the support you should
3868 exclude it using the list of supported ciphers. This also means that the
3869 "-no_ecdhe" option has been removed from s_server.
3873 * SSL_{CTX}_set_tmp_ecdh() which can set 1 EC curve now internally calls
3874 SSL_{CTX_}set1_curves() which can set a list.
3878 * Remove support for SSL_{CTX_}set_tmp_ecdh_callback(). You should set the
3879 curve you want to support using SSL_{CTX_}set1_curves().
3883 * State machine rewrite. The state machine code has been significantly
3884 refactored in order to remove much duplication of code and solve issues
3885 with the old code (see [ssl/statem/README.md](ssl/statem/README.md) for
3886 further details). This change does have some associated API changes.
3887 Notably the SSL_state() function has been removed and replaced by
3888 SSL_get_state which now returns an "OSSL_HANDSHAKE_STATE" instead of an int.
3889 SSL_set_state() has been removed altogether. The previous handshake states
3890 defined in ssl.h and ssl3.h have also been removed.
3894 * All instances of the string "ssleay" in the public API were replaced
3895 with OpenSSL (case-matching; e.g., OPENSSL_VERSION for #define's)
3896 Some error codes related to internal RSA_eay API's were renamed.
3900 * The demo files in crypto/threads were moved to demo/threads.
3904 * Removed obsolete engines: 4758cca, aep, atalla, cswift, nuron, gmp,
3907 *Matt Caswell, Rich Salz*
3909 * New ASN.1 embed macro.
3911 New ASN.1 macro ASN1_EMBED. This is the same as ASN1_SIMPLE except the
3912 structure is not allocated: it is part of the parent. That is instead of
3920 This reduces memory fragmentation and make it impossible to accidentally
3921 set a mandatory field to NULL.
3923 This currently only works for some fields specifically a SEQUENCE, CHOICE,
3924 or ASN1_STRING type which is part of a parent SEQUENCE. Since it is
3925 equivalent to ASN1_SIMPLE it cannot be tagged, OPTIONAL, SET OF or
3930 * Remove EVP_CHECK_DES_KEY, a compile-time option that never compiled.
3934 * Removed DES and RC4 ciphersuites from DEFAULT. Also removed RC2 although
3935 in 1.0.2 EXPORT was already removed and the only RC2 ciphersuite is also
3936 an EXPORT one. COMPLEMENTOFDEFAULT has been updated accordingly to add
3937 DES and RC4 ciphersuites.
3941 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
3942 This changes the decoding behaviour for some invalid messages,
3943 though the change is mostly in the more lenient direction, and
3944 legacy behaviour is preserved as much as possible.
3948 * Fix no-stdio build.
3949 *David Woodhouse <David.Woodhouse@intel.com> and also*
3950 *Ivan Nestlerode <ivan.nestlerode@sonos.com>*
3952 * New testing framework
3953 The testing framework has been largely rewritten and is now using
3954 perl and the perl modules Test::Harness and an extended variant of
3955 Test::More called OpenSSL::Test to do its work. All test scripts in
3956 test/ have been rewritten into test recipes, and all direct calls to
3957 executables in test/Makefile have become individual recipes using the
3958 simplified testing OpenSSL::Test::Simple.
3960 For documentation on our testing modules, do:
3962 perldoc test/testlib/OpenSSL/Test/Simple.pm
3963 perldoc test/testlib/OpenSSL/Test.pm
3967 * Revamped memory debug; only -DCRYPTO_MDEBUG and -DCRYPTO_MDEBUG_ABORT
3968 are used; the latter aborts on memory leaks (usually checked on exit).
3969 Some undocumented "set malloc, etc., hooks" functions were removed
3970 and others were changed. All are now documented.
3974 * In DSA_generate_parameters_ex, if the provided seed is too short,
3977 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
3979 * Rewrite PSK to support ECDHE_PSK, DHE_PSK and RSA_PSK. Add ciphersuites
3980 from RFC4279, RFC4785, RFC5487, RFC5489.
3982 Thanks to Christian J. Dietrich and Giuseppe D'Angelo for the
3983 original RSA_PSK patch.
3987 * Dropped support for the SSL3_FLAGS_DELAY_CLIENT_FINISHED flag. This SSLeay
3988 era flag was never set throughout the codebase (only read). Also removed
3989 SSL3_FLAGS_POP_BUFFER which was only used if
3990 SSL3_FLAGS_DELAY_CLIENT_FINISHED was also set.
3994 * Changed the default name options in the "ca", "crl", "req" and "x509"
3995 to be "oneline" instead of "compat".
3999 * Remove SSL_OP_TLS_BLOCK_PADDING_BUG. This is SSLeay legacy, we're
4000 not aware of clients that still exhibit this bug, and the workaround
4001 hasn't been working properly for a while.
4005 * The return type of BIO_number_read() and BIO_number_written() as well as
4006 the corresponding num_read and num_write members in the BIO structure has
4007 changed from unsigned long to uint64_t. On platforms where an unsigned
4008 long is 32 bits (e.g. Windows) these counters could overflow if >4Gb is
4013 * Given the pervasive nature of TLS extensions it is inadvisable to run
4014 OpenSSL without support for them. It also means that maintaining
4015 the OPENSSL_NO_TLSEXT option within the code is very invasive (and probably
4016 not well tested). Therefore the OPENSSL_NO_TLSEXT option has been removed.
4020 * Removed support for the two export grade static DH ciphersuites
4021 EXP-DH-RSA-DES-CBC-SHA and EXP-DH-DSS-DES-CBC-SHA. These two ciphersuites
4022 were newly added (along with a number of other static DH ciphersuites) to
4023 1.0.2. However the two export ones have *never* worked since they were
4024 introduced. It seems strange in any case to be adding new export
4025 ciphersuites, and given "logjam" it also does not seem correct to fix them.
4029 * Version negotiation has been rewritten. In particular SSLv23_method(),
4030 SSLv23_client_method() and SSLv23_server_method() have been deprecated,
4031 and turned into macros which simply call the new preferred function names
4032 TLS_method(), TLS_client_method() and TLS_server_method(). All new code
4033 should use the new names instead. Also as part of this change the ssl23.h
4034 header file has been removed.
4038 * Support for Kerberos ciphersuites in TLS (RFC2712) has been removed. This
4039 code and the associated standard is no longer considered fit-for-purpose.
4043 * RT2547 was closed. When generating a private key, try to make the
4044 output file readable only by the owner. This behavior change might
4045 be noticeable when interacting with other software.
4047 * Documented all exdata functions. Added CRYPTO_free_ex_index.
4052 * Added HTTP GET support to the ocsp command.
4056 * Changed default digest for the dgst and enc commands from MD5 to
4061 * RAND_pseudo_bytes has been deprecated. Users should use RAND_bytes instead.
4065 * Added support for TLS extended master secret from
4066 draft-ietf-tls-session-hash-03.txt. Thanks for Alfredo Pironti for an
4067 initial patch which was a great help during development.
4071 * All libssl internal structures have been removed from the public header
4072 files, and the OPENSSL_NO_SSL_INTERN option has been removed (since it is
4073 now redundant). Users should not attempt to access internal structures
4074 directly. Instead they should use the provided API functions.
4078 * config has been changed so that by default OPENSSL_NO_DEPRECATED is used.
4079 Access to deprecated functions can be re-enabled by running config with
4080 "enable-deprecated". In addition applications wishing to use deprecated
4081 functions must define OPENSSL_USE_DEPRECATED. Note that this new behaviour
4082 will, by default, disable some transitive includes that previously existed
4083 in the header files (e.g. ec.h will no longer, by default, include bn.h)
4087 * Added support for OCB mode. OpenSSL has been granted a patent license
4088 compatible with the OpenSSL license for use of OCB. Details are available
4089 at <https://www.openssl.org/source/OCB-patent-grant-OpenSSL.pdf>. Support
4090 for OCB can be removed by calling config with no-ocb.
4094 * SSLv2 support has been removed. It still supports receiving a SSLv2
4095 compatible client hello.
4099 * Increased the minimal RSA keysize from 256 to 512 bits [Rich Salz],
4100 done while fixing the error code for the key-too-small case.
4102 *Annie Yousar <a.yousar@informatik.hu-berlin.de>*
4104 * CA.sh has been removed; use CA.pl instead.
4108 * Removed old DES API.
4112 * Remove various unsupported platforms:
4118 Sinix/ReliantUNIX RM400
4123 16-bit platforms such as WIN16
4127 * Clean up OPENSSL_NO_xxx #define's
4128 - Use setbuf() and remove OPENSSL_NO_SETVBUF_IONBF
4129 - Rename OPENSSL_SYSNAME_xxx to OPENSSL_SYS_xxx
4130 - OPENSSL_NO_EC{DH,DSA} merged into OPENSSL_NO_EC
4131 - OPENSSL_NO_RIPEMD160, OPENSSL_NO_RIPEMD merged into OPENSSL_NO_RMD160
4132 - OPENSSL_NO_FP_API merged into OPENSSL_NO_STDIO
4133 - Remove OPENSSL_NO_BIO OPENSSL_NO_BUFFER OPENSSL_NO_CHAIN_VERIFY
4134 OPENSSL_NO_EVP OPENSSL_NO_FIPS_ERR OPENSSL_NO_HASH_COMP
4135 OPENSSL_NO_LHASH OPENSSL_NO_OBJECT OPENSSL_NO_SPEED OPENSSL_NO_STACK
4136 OPENSSL_NO_X509 OPENSSL_NO_X509_VERIFY
4137 - Remove MS_STATIC; it's a relic from platforms <32 bits.
4141 * Cleaned up dead code
4142 Remove all but one '#ifdef undef' which is to be looked at.
4146 * Clean up calling of xxx_free routines.
4147 Just like free(), fix most of the xxx_free routines to accept
4148 NULL. Remove the non-null checks from callers. Save much code.
4152 * Add secure heap for storage of private keys (when possible).
4153 Add BIO_s_secmem(), CBIGNUM, etc.
4154 Contributed by Akamai Technologies under our Corporate CLA.
4158 * Experimental support for a new, fast, unbiased prime candidate generator,
4159 bn_probable_prime_dh_coprime(). Not currently used by any prime generator.
4161 *Felix Laurie von Massenbach <felix@erbridge.co.uk>*
4163 * New output format NSS in the sess_id command line tool. This allows
4164 exporting the session id and the master key in NSS keylog format.
4166 *Martin Kaiser <martin@kaiser.cx>*
4168 * Harmonize version and its documentation. -f flag is used to display
4171 *mancha <mancha1@zoho.com>*
4173 * Fix eckey_priv_encode so it immediately returns an error upon a failure
4174 in i2d_ECPrivateKey. Thanks to Ted Unangst for feedback on this issue.
4176 *mancha <mancha1@zoho.com>*
4178 * Fix some double frees. These are not thought to be exploitable.
4180 *mancha <mancha1@zoho.com>*
4182 * A missing bounds check in the handling of the TLS heartbeat extension
4183 can be used to reveal up to 64k of memory to a connected client or
4186 Thanks for Neel Mehta of Google Security for discovering this bug and to
4187 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
4188 preparing the fix ([CVE-2014-0160])
4190 *Adam Langley, Bodo Moeller*
4192 * Fix for the attack described in the paper "Recovering OpenSSL
4193 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
4194 by Yuval Yarom and Naomi Benger. Details can be obtained from:
4195 <http://eprint.iacr.org/2014/140>
4197 Thanks to Yuval Yarom and Naomi Benger for discovering this
4198 flaw and to Yuval Yarom for supplying a fix ([CVE-2014-0076])
4200 *Yuval Yarom and Naomi Benger*
4202 * Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
4203 this fixes a limitation in previous versions of OpenSSL.
4207 * Experimental encrypt-then-mac support.
4209 Experimental support for encrypt then mac from
4210 draft-gutmann-tls-encrypt-then-mac-02.txt
4212 To enable it set the appropriate extension number (0x42 for the test
4213 server) using e.g. -DTLSEXT_TYPE_encrypt_then_mac=0x42
4215 For non-compliant peers (i.e. just about everything) this should have no
4218 WARNING: EXPERIMENTAL, SUBJECT TO CHANGE.
4222 * Add EVP support for key wrapping algorithms, to avoid problems with
4223 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
4224 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
4225 algorithms and include tests cases.
4229 * Extend CMS code to support RSA-PSS signatures and RSA-OAEP for
4234 * Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
4235 MGF1 digest and OAEP label.
4239 * Make openssl verify return errors.
4241 *Chris Palmer <palmer@google.com> and Ben Laurie*
4243 * New function ASN1_TIME_diff to calculate the difference between two
4244 ASN1_TIME structures or one structure and the current time.
4248 * Update fips_test_suite to support multiple command line options. New
4249 test to induce all self test errors in sequence and check expected
4254 * Add FIPS_{rsa,dsa,ecdsa}_{sign,verify} functions which digest and
4255 sign or verify all in one operation.
4259 * Add fips_algvs: a multicall fips utility incorporating all the algorithm
4260 test programs and fips_test_suite. Includes functionality to parse
4261 the minimal script output of fipsalgest.pl directly.
4265 * Add authorisation parameter to FIPS_module_mode_set().
4269 * Add FIPS selftest for ECDH algorithm using P-224 and B-233 curves.
4273 * Use separate DRBG fields for internal and external flags. New function
4274 FIPS_drbg_health_check() to perform on demand health checking. Add
4275 generation tests to fips_test_suite with reduced health check interval to
4276 demonstrate periodic health checking. Add "nodh" option to
4277 fips_test_suite to skip very slow DH test.
4281 * New function FIPS_get_cipherbynid() to lookup FIPS supported ciphers
4286 * More extensive health check for DRBG checking many more failure modes.
4287 New function FIPS_selftest_drbg_all() to handle every possible DRBG
4288 combination: call this in fips_test_suite.
4292 * Add support for canonical generation of DSA parameter 'g'. See
4295 * Add support for HMAC DRBG from SP800-90. Update DRBG algorithm test and
4296 POST to handle HMAC cases.
4300 * Add functions FIPS_module_version() and FIPS_module_version_text()
4301 to return numerical and string versions of the FIPS module number.
4305 * Rename FIPS_mode_set and FIPS_mode to FIPS_module_mode_set and
4306 FIPS_module_mode. FIPS_mode and FIPS_mode_set will be implemented
4307 outside the validated module in the FIPS capable OpenSSL.
4311 * Minor change to DRBG entropy callback semantics. In some cases
4312 there is no multiple of the block length between min_len and
4313 max_len. Allow the callback to return more than max_len bytes
4314 of entropy but discard any extra: it is the callback's responsibility
4315 to ensure that the extra data discarded does not impact the
4316 requested amount of entropy.
4320 * Add PRNG security strength checks to RSA, DSA and ECDSA using
4321 information in FIPS186-3, SP800-57 and SP800-131A.
4325 * CCM support via EVP. Interface is very similar to GCM case except we
4326 must supply all data in one chunk (i.e. no update, final) and the
4327 message length must be supplied if AAD is used. Add algorithm test
4332 * Initial version of POST overhaul. Add POST callback to allow the status
4333 of POST to be monitored and/or failures induced. Modify fips_test_suite
4334 to use callback. Always run all selftests even if one fails.
4338 * XTS support including algorithm test driver in the fips_gcmtest program.
4339 Note: this does increase the maximum key length from 32 to 64 bytes but
4340 there should be no binary compatibility issues as existing applications
4341 will never use XTS mode.
4345 * Extensive reorganisation of FIPS PRNG behaviour. Remove all dependencies
4346 to OpenSSL RAND code and replace with a tiny FIPS RAND API which also
4347 performs algorithm blocking for unapproved PRNG types. Also do not
4348 set PRNG type in FIPS_mode_set(): leave this to the application.
4349 Add default OpenSSL DRBG handling: sets up FIPS PRNG and seeds with
4350 the standard OpenSSL PRNG: set additional data to a date time vector.
4354 * Rename old X9.31 PRNG functions of the form `FIPS_rand*` to `FIPS_x931*`.
4355 This shouldn't present any incompatibility problems because applications
4356 shouldn't be using these directly and any that are will need to rethink
4357 anyway as the X9.31 PRNG is now deprecated by FIPS 140-2
4361 * Extensive self tests and health checking required by SP800-90 DRBG.
4362 Remove strength parameter from FIPS_drbg_instantiate and always
4363 instantiate at maximum supported strength.
4367 * Add ECDH code to fips module and fips_ecdhvs for primitives only testing.
4371 * New algorithm test program fips_dhvs to handle DH primitives only testing.
4375 * New function DH_compute_key_padded() to compute a DH key and pad with
4376 leading zeroes if needed: this complies with SP800-56A et al.
4380 * Initial implementation of SP800-90 DRBGs for Hash and CTR. Not used by
4381 anything, incomplete, subject to change and largely untested at present.
4385 * Modify fipscanisteronly build option to only build the necessary object
4386 files by filtering FIPS_EX_OBJ through a perl script in crypto/Makefile.
4390 * Add experimental option FIPSSYMS to give all symbols in
4391 fipscanister.o and FIPS or fips prefix. This will avoid
4392 conflicts with future versions of OpenSSL. Add perl script
4393 util/fipsas.pl to preprocess assembly language source files
4394 and rename any affected symbols.
4398 * Add selftest checks and algorithm block of non-fips algorithms in
4399 FIPS mode. Remove DES2 from selftests.
4403 * Add ECDSA code to fips module. Add tiny fips_ecdsa_check to just
4404 return internal method without any ENGINE dependencies. Add new
4405 tiny fips sign and verify functions.
4409 * New build option no-ec2m to disable characteristic 2 code.
4413 * New build option "fipscanisteronly". This only builds fipscanister.o
4414 and (currently) associated fips utilities. Uses the file Makefile.fips
4415 instead of Makefile.org as the prototype.
4419 * Add some FIPS mode restrictions to GCM. Add internal IV generator.
4420 Update fips_gcmtest to use IV generator.
4424 * Initial, experimental EVP support for AES-GCM. AAD can be input by
4425 setting output buffer to NULL. The `*Final` function must be
4426 called although it will not retrieve any additional data. The tag
4427 can be set or retrieved with a ctrl. The IV length is by default 12
4428 bytes (96 bits) but can be set to an alternative value. If the IV
4429 length exceeds the maximum IV length (currently 16 bytes) it cannot be
4434 * New flag in ciphers: EVP_CIPH_FLAG_CUSTOM_CIPHER. This means the
4435 underlying do_cipher function handles all cipher semantics itself
4436 including padding and finalisation. This is useful if (for example)
4437 an ENGINE cipher handles block padding itself. The behaviour of
4438 do_cipher is subtly changed if this flag is set: the return value
4439 is the number of characters written to the output buffer (zero is
4440 no longer an error code) or a negative error code. Also if the
4441 input buffer is NULL and length 0 finalisation should be performed.
4445 * If a candidate issuer certificate is already part of the constructed
4446 path ignore it: new debug notification X509_V_ERR_PATH_LOOP for this case.
4450 * Improve forward-security support: add functions
4452 void SSL_CTX_set_not_resumable_session_callback(
4453 SSL_CTX *ctx, int (*cb)(SSL *ssl, int is_forward_secure))
4454 void SSL_set_not_resumable_session_callback(
4455 SSL *ssl, int (*cb)(SSL *ssl, int is_forward_secure))
4457 for use by SSL/TLS servers; the callback function will be called whenever a
4458 new session is created, and gets to decide whether the session may be
4459 cached to make it resumable (return 0) or not (return 1). (As by the
4460 SSL/TLS protocol specifications, the session_id sent by the server will be
4461 empty to indicate that the session is not resumable; also, the server will
4462 not generate RFC 4507 (RFC 5077) session tickets.)
4464 A simple reasonable callback implementation is to return is_forward_secure.
4465 This parameter will be set to 1 or 0 depending on the ciphersuite selected
4466 by the SSL/TLS server library, indicating whether it can provide forward
4469 *Emilia Käsper <emilia.kasper@esat.kuleuven.be> (Google)*
4471 * New -verify_name option in command line utilities to set verification
4476 * Initial CMAC implementation. WARNING: EXPERIMENTAL, API MAY CHANGE.
4477 Add CMAC pkey methods.
4481 * Experimental renegotiation in s_server -www mode. If the client
4482 browses /reneg connection is renegotiated. If /renegcert it is
4483 renegotiated requesting a certificate.
4487 * Add an "external" session cache for debugging purposes to s_server. This
4488 should help trace issues which normally are only apparent in deployed
4489 multi-process servers.
4493 * Extensive audit of libcrypto with DEBUG_UNUSED. Fix many cases where
4494 return value is ignored. NB. The functions RAND_add(), RAND_seed(),
4495 BIO_set_cipher() and some obscure PEM functions were changed so they
4496 can now return an error. The RAND changes required a change to the
4497 RAND_METHOD structure.
4501 * New macro `__owur` for "OpenSSL Warn Unused Result". This makes use of
4502 a gcc attribute to warn if the result of a function is ignored. This
4503 is enable if DEBUG_UNUSED is set. Add to several functions in evp.h
4504 whose return value is often ignored.
4508 * New -noct, -requestct, -requirect and -ctlogfile options for s_client.
4509 These allow SCTs (signed certificate timestamps) to be requested and
4510 validated when establishing a connection.
4512 *Rob Percival <robpercival@google.com>*
4517 ### Changes between 1.0.2s and 1.0.2t [10 Sep 2019]
4519 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
4520 used even when parsing explicit parameters, when loading a encoded key
4521 or calling `EC_GROUP_new_from_ecpkparameters()`/
4522 `EC_GROUP_new_from_ecparameters()`.
4523 This prevents bypass of security hardening and performance gains,
4524 especially for curves with specialized EC_METHODs.
4525 By default, if a key encoded with explicit parameters is loaded and later
4526 encoded, the output is still encoded with explicit parameters, even if
4527 internally a "named" EC_GROUP is used for computation.
4531 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
4532 this change, EC_GROUP_set_generator would accept order and/or cofactor as
4533 NULL. After this change, only the cofactor parameter can be NULL. It also
4534 does some minimal sanity checks on the passed order.
4539 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
4540 An attack is simple, if the first CMS_recipientInfo is valid but the
4541 second CMS_recipientInfo is chosen ciphertext. If the second
4542 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
4543 encryption key will be replaced by garbage, and the message cannot be
4544 decoded, but if the RSA decryption fails, the correct encryption key is
4545 used and the recipient will not notice the attack.
4546 As a work around for this potential attack the length of the decrypted
4547 key must be equal to the cipher default key length, in case the
4548 certifiate is not given and all recipientInfo are tried out.
4549 The old behaviour can be re-enabled in the CMS code by setting the
4550 CMS_DEBUG_DECRYPT flag.
4555 * Document issue with installation paths in diverse Windows builds
4557 '/usr/local/ssl' is an unsafe prefix for location to install OpenSSL
4558 binaries and run-time config file.
4563 ### Changes between 1.0.2r and 1.0.2s [28 May 2019]
4565 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
4566 This changes the size when using the `genpkey` command when no size is given.
4567 It fixes an omission in earlier changes that changed all RSA, DSA and DH
4568 generation commands to use 2048 bits by default.
4572 * Add FIPS support for Android Arm 64-bit
4574 Support for Android Arm 64-bit was added to the OpenSSL FIPS Object
4575 Module in Version 2.0.10. For some reason, the corresponding target
4576 'android64-aarch64' was missing OpenSSL 1.0.2, whence it could not be
4577 built with FIPS support on Android Arm 64-bit. This omission has been
4580 *Matthias St. Pierre*
4582 ### Changes between 1.0.2q and 1.0.2r [26 Feb 2019]
4584 * 0-byte record padding oracle
4586 If an application encounters a fatal protocol error and then calls
4587 SSL_shutdown() twice (once to send a close_notify, and once to receive one)
4588 then OpenSSL can respond differently to the calling application if a 0 byte
4589 record is received with invalid padding compared to if a 0 byte record is
4590 received with an invalid MAC. If the application then behaves differently
4591 based on that in a way that is detectable to the remote peer, then this
4592 amounts to a padding oracle that could be used to decrypt data.
4594 In order for this to be exploitable "non-stitched" ciphersuites must be in
4595 use. Stitched ciphersuites are optimised implementations of certain
4596 commonly used ciphersuites. Also the application must call SSL_shutdown()
4597 twice even if a protocol error has occurred (applications should not do
4598 this but some do anyway).
4600 This issue was discovered by Juraj Somorovsky, Robert Merget and Nimrod
4601 Aviram, with additional investigation by Steven Collison and Andrew
4602 Hourselt. It was reported to OpenSSL on 10th December 2018.
4607 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
4611 ### Changes between 1.0.2p and 1.0.2q [20 Nov 2018]
4613 * Microarchitecture timing vulnerability in ECC scalar multiplication
4615 OpenSSL ECC scalar multiplication, used in e.g. ECDSA and ECDH, has been
4616 shown to be vulnerable to a microarchitecture timing side channel attack.
4617 An attacker with sufficient access to mount local timing attacks during
4618 ECDSA signature generation could recover the private key.
4620 This issue was reported to OpenSSL on 26th October 2018 by Alejandro
4621 Cabrera Aldaya, Billy Brumley, Sohaib ul Hassan, Cesar Pereida Garcia and
4627 * Timing vulnerability in DSA signature generation
4629 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
4630 timing side channel attack. An attacker could use variations in the signing
4631 algorithm to recover the private key.
4633 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
4638 * Resolve a compatibility issue in EC_GROUP handling with the FIPS Object
4639 Module, accidentally introduced while backporting security fixes from the
4640 development branch and hindering the use of ECC in FIPS mode.
4644 ### Changes between 1.0.2o and 1.0.2p [14 Aug 2018]
4646 * Client DoS due to large DH parameter
4648 During key agreement in a TLS handshake using a DH(E) based ciphersuite a
4649 malicious server can send a very large prime value to the client. This will
4650 cause the client to spend an unreasonably long period of time generating a
4651 key for this prime resulting in a hang until the client has finished. This
4652 could be exploited in a Denial Of Service attack.
4654 This issue was reported to OpenSSL on 5th June 2018 by Guido Vranken
4659 * Cache timing vulnerability in RSA Key Generation
4661 The OpenSSL RSA Key generation algorithm has been shown to be vulnerable to
4662 a cache timing side channel attack. An attacker with sufficient access to
4663 mount cache timing attacks during the RSA key generation process could
4664 recover the private key.
4666 This issue was reported to OpenSSL on 4th April 2018 by Alejandro Cabrera
4667 Aldaya, Billy Brumley, Cesar Pereida Garcia and Luis Manuel Alvarez Tapia.
4672 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
4673 parameter is no longer accepted, as it leads to a corrupt table. NULL
4674 pem_str is reserved for alias entries only.
4678 * Revert blinding in ECDSA sign and instead make problematic addition
4679 length-invariant. Switch even to fixed-length Montgomery multiplication.
4683 * Change generating and checking of primes so that the error rate of not
4684 being prime depends on the intended use based on the size of the input.
4685 For larger primes this will result in more rounds of Miller-Rabin.
4686 The maximal error rate for primes with more than 1080 bits is lowered
4689 *Kurt Roeckx, Annie Yousar*
4691 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
4695 * Add blinding to ECDSA and DSA signatures to protect against side channel
4696 attacks discovered by Keegan Ryan (NCC Group).
4700 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
4701 now allow empty (zero character) pass phrases.
4705 * Certificate time validation (X509_cmp_time) enforces stricter
4706 compliance with RFC 5280. Fractional seconds and timezone offsets
4707 are no longer allowed.
4711 ### Changes between 1.0.2n and 1.0.2o [27 Mar 2018]
4713 * Constructed ASN.1 types with a recursive definition could exceed the stack
4715 Constructed ASN.1 types with a recursive definition (such as can be found
4716 in PKCS7) could eventually exceed the stack given malicious input with
4717 excessive recursion. This could result in a Denial Of Service attack. There
4718 are no such structures used within SSL/TLS that come from untrusted sources
4719 so this is considered safe.
4721 This issue was reported to OpenSSL on 4th January 2018 by the OSS-fuzz
4727 ### Changes between 1.0.2m and 1.0.2n [7 Dec 2017]
4729 * Read/write after SSL object in error state
4731 OpenSSL 1.0.2 (starting from version 1.0.2b) introduced an "error state"
4732 mechanism. The intent was that if a fatal error occurred during a handshake
4733 then OpenSSL would move into the error state and would immediately fail if
4734 you attempted to continue the handshake. This works as designed for the
4735 explicit handshake functions (SSL_do_handshake(), SSL_accept() and
4736 SSL_connect()), however due to a bug it does not work correctly if
4737 SSL_read() or SSL_write() is called directly. In that scenario, if the
4738 handshake fails then a fatal error will be returned in the initial function
4739 call. If SSL_read()/SSL_write() is subsequently called by the application
4740 for the same SSL object then it will succeed and the data is passed without
4741 being decrypted/encrypted directly from the SSL/TLS record layer.
4743 In order to exploit this issue an application bug would have to be present
4744 that resulted in a call to SSL_read()/SSL_write() being issued after having
4745 already received a fatal error.
4747 This issue was reported to OpenSSL by David Benjamin (Google).
4752 * rsaz_1024_mul_avx2 overflow bug on x86_64
4754 There is an overflow bug in the AVX2 Montgomery multiplication procedure
4755 used in exponentiation with 1024-bit moduli. No EC algorithms are affected.
4756 Analysis suggests that attacks against RSA and DSA as a result of this
4757 defect would be very difficult to perform and are not believed likely.
4758 Attacks against DH1024 are considered just feasible, because most of the
4759 work necessary to deduce information about a private key may be performed
4760 offline. The amount of resources required for such an attack would be
4761 significant. However, for an attack on TLS to be meaningful, the server
4762 would have to share the DH1024 private key among multiple clients, which is
4763 no longer an option since CVE-2016-0701.
4765 This only affects processors that support the AVX2 but not ADX extensions
4766 like Intel Haswell (4th generation).
4768 This issue was reported to OpenSSL by David Benjamin (Google). The issue
4769 was originally found via the OSS-Fuzz project.
4774 ### Changes between 1.0.2l and 1.0.2m [2 Nov 2017]
4776 * bn_sqrx8x_internal carry bug on x86_64
4778 There is a carry propagating bug in the x86_64 Montgomery squaring
4779 procedure. No EC algorithms are affected. Analysis suggests that attacks
4780 against RSA and DSA as a result of this defect would be very difficult to
4781 perform and are not believed likely. Attacks against DH are considered just
4782 feasible (although very difficult) because most of the work necessary to
4783 deduce information about a private key may be performed offline. The amount
4784 of resources required for such an attack would be very significant and
4785 likely only accessible to a limited number of attackers. An attacker would
4786 additionally need online access to an unpatched system using the target
4787 private key in a scenario with persistent DH parameters and a private
4788 key that is shared between multiple clients.
4790 This only affects processors that support the BMI1, BMI2 and ADX extensions
4791 like Intel Broadwell (5th generation) and later or AMD Ryzen.
4793 This issue was reported to OpenSSL by the OSS-Fuzz project.
4798 * Malformed X.509 IPAddressFamily could cause OOB read
4800 If an X.509 certificate has a malformed IPAddressFamily extension,
4801 OpenSSL could do a one-byte buffer overread. The most likely result
4802 would be an erroneous display of the certificate in text format.
4804 This issue was reported to OpenSSL by the OSS-Fuzz project.
4808 ### Changes between 1.0.2k and 1.0.2l [25 May 2017]
4810 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
4811 platform rather than 'mingw'.
4815 ### Changes between 1.0.2j and 1.0.2k [26 Jan 2017]
4817 * Truncated packet could crash via OOB read
4819 If one side of an SSL/TLS path is running on a 32-bit host and a specific
4820 cipher is being used, then a truncated packet can cause that host to
4821 perform an out-of-bounds read, usually resulting in a crash.
4823 This issue was reported to OpenSSL by Robert Święcki of Google.
4828 * BN_mod_exp may produce incorrect results on x86_64
4830 There is a carry propagating bug in the x86_64 Montgomery squaring
4831 procedure. No EC algorithms are affected. Analysis suggests that attacks
4832 against RSA and DSA as a result of this defect would be very difficult to
4833 perform and are not believed likely. Attacks against DH are considered just
4834 feasible (although very difficult) because most of the work necessary to
4835 deduce information about a private key may be performed offline. The amount
4836 of resources required for such an attack would be very significant and
4837 likely only accessible to a limited number of attackers. An attacker would
4838 additionally need online access to an unpatched system using the target
4839 private key in a scenario with persistent DH parameters and a private
4840 key that is shared between multiple clients. For example this can occur by
4841 default in OpenSSL DHE based SSL/TLS ciphersuites. Note: This issue is very
4842 similar to CVE-2015-3193 but must be treated as a separate problem.
4844 This issue was reported to OpenSSL by the OSS-Fuzz project.
4849 * Montgomery multiplication may produce incorrect results
4851 There is a carry propagating bug in the Broadwell-specific Montgomery
4852 multiplication procedure that handles input lengths divisible by, but
4853 longer than 256 bits. Analysis suggests that attacks against RSA, DSA
4854 and DH private keys are impossible. This is because the subroutine in
4855 question is not used in operations with the private key itself and an input
4856 of the attacker's direct choice. Otherwise the bug can manifest itself as
4857 transient authentication and key negotiation failures or reproducible
4858 erroneous outcome of public-key operations with specially crafted input.
4859 Among EC algorithms only Brainpool P-512 curves are affected and one
4860 presumably can attack ECDH key negotiation. Impact was not analyzed in
4861 detail, because pre-requisites for attack are considered unlikely. Namely
4862 multiple clients have to choose the curve in question and the server has to
4863 share the private key among them, neither of which is default behaviour.
4864 Even then only clients that chose the curve will be affected.
4866 This issue was publicly reported as transient failures and was not
4867 initially recognized as a security issue. Thanks to Richard Morgan for
4868 providing reproducible case.
4873 * OpenSSL now fails if it receives an unrecognised record type in TLS1.0
4874 or TLS1.1. Previously this only happened in SSLv3 and TLS1.2. This is to
4875 prevent issues where no progress is being made and the peer continually
4876 sends unrecognised record types, using up resources processing them.
4880 ### Changes between 1.0.2i and 1.0.2j [26 Sep 2016]
4882 * Missing CRL sanity check
4884 A bug fix which included a CRL sanity check was added to OpenSSL 1.1.0
4885 but was omitted from OpenSSL 1.0.2i. As a result any attempt to use
4886 CRLs in OpenSSL 1.0.2i will crash with a null pointer exception.
4888 This issue only affects the OpenSSL 1.0.2i
4893 ### Changes between 1.0.2h and 1.0.2i [22 Sep 2016]
4895 * OCSP Status Request extension unbounded memory growth
4897 A malicious client can send an excessively large OCSP Status Request
4898 extension. If that client continually requests renegotiation, sending a
4899 large OCSP Status Request extension each time, then there will be unbounded
4900 memory growth on the server. This will eventually lead to a Denial Of
4901 Service attack through memory exhaustion. Servers with a default
4902 configuration are vulnerable even if they do not support OCSP. Builds using
4903 the "no-ocsp" build time option are not affected.
4905 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4910 * In order to mitigate the SWEET32 attack, the DES ciphers were moved from
4913 This issue was reported to OpenSSL Karthikeyan Bhargavan and Gaetan
4919 * OOB write in MDC2_Update()
4921 An overflow can occur in MDC2_Update() either if called directly or
4922 through the EVP_DigestUpdate() function using MDC2. If an attacker
4923 is able to supply very large amounts of input data after a previous
4924 call to EVP_EncryptUpdate() with a partial block then a length check
4925 can overflow resulting in a heap corruption.
4927 The amount of data needed is comparable to SIZE_MAX which is impractical
4930 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4935 * Malformed SHA512 ticket DoS
4937 If a server uses SHA512 for TLS session ticket HMAC it is vulnerable to a
4938 DoS attack where a malformed ticket will result in an OOB read which will
4941 The use of SHA512 in TLS session tickets is comparatively rare as it requires
4942 a custom server callback and ticket lookup mechanism.
4944 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4949 * OOB write in BN_bn2dec()
4951 The function BN_bn2dec() does not check the return value of BN_div_word().
4952 This can cause an OOB write if an application uses this function with an
4953 overly large BIGNUM. This could be a problem if an overly large certificate
4954 or CRL is printed out from an untrusted source. TLS is not affected because
4955 record limits will reject an oversized certificate before it is parsed.
4957 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4962 * OOB read in TS_OBJ_print_bio()
4964 The function TS_OBJ_print_bio() misuses OBJ_obj2txt(): the return value is
4965 the total length the OID text representation would use and not the amount
4966 of data written. This will result in OOB reads when large OIDs are
4969 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4974 * Pointer arithmetic undefined behaviour
4976 Avoid some undefined pointer arithmetic
4978 A common idiom in the codebase is to check limits in the following manner:
4981 Where "p" points to some malloc'd data of SIZE bytes and
4984 "len" here could be from some externally supplied data (e.g. from a TLS
4987 The rules of C pointer arithmetic are such that "p + len" is only well
4988 defined where len <= SIZE. Therefore the above idiom is actually
4989 undefined behaviour.
4991 For example this could cause problems if some malloc implementation
4992 provides an address for "p" such that "p + len" actually overflows for
4993 values of len that are too big and therefore p + len < limit.
4995 This issue was reported to OpenSSL by Guido Vranken
5000 * Constant time flag not preserved in DSA signing
5002 Operations in the DSA signing algorithm should run in constant time in
5003 order to avoid side channel attacks. A flaw in the OpenSSL DSA
5004 implementation means that a non-constant time codepath is followed for
5005 certain operations. This has been demonstrated through a cache-timing
5006 attack to be sufficient for an attacker to recover the private DSA key.
5008 This issue was reported by César Pereida (Aalto University), Billy Brumley
5009 (Tampere University of Technology), and Yuval Yarom (The University of
5010 Adelaide and NICTA).
5015 * DTLS buffered message DoS
5017 In a DTLS connection where handshake messages are delivered out-of-order
5018 those messages that OpenSSL is not yet ready to process will be buffered
5019 for later use. Under certain circumstances, a flaw in the logic means that
5020 those messages do not get removed from the buffer even though the handshake
5021 has been completed. An attacker could force up to approx. 15 messages to
5022 remain in the buffer when they are no longer required. These messages will
5023 be cleared when the DTLS connection is closed. The default maximum size for
5024 a message is 100k. Therefore the attacker could force an additional 1500k
5025 to be consumed per connection. By opening many simulataneous connections an
5026 attacker could cause a DoS attack through memory exhaustion.
5028 This issue was reported to OpenSSL by Quan Luo.
5033 * DTLS replay protection DoS
5035 A flaw in the DTLS replay attack protection mechanism means that records
5036 that arrive for future epochs update the replay protection "window" before
5037 the MAC for the record has been validated. This could be exploited by an
5038 attacker by sending a record for the next epoch (which does not have to
5039 decrypt or have a valid MAC), with a very large sequence number. This means
5040 that all subsequent legitimate packets are dropped causing a denial of
5041 service for a specific DTLS connection.
5043 This issue was reported to OpenSSL by the OCAP audit team.
5048 * Certificate message OOB reads
5050 In OpenSSL 1.0.2 and earlier some missing message length checks can result
5051 in OOB reads of up to 2 bytes beyond an allocated buffer. There is a
5052 theoretical DoS risk but this has not been observed in practice on common
5055 The messages affected are client certificate, client certificate request
5056 and server certificate. As a result the attack can only be performed
5057 against a client or a server which enables client authentication.
5059 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5064 ### Changes between 1.0.2g and 1.0.2h [3 May 2016]
5066 * Prevent padding oracle in AES-NI CBC MAC check
5068 A MITM attacker can use a padding oracle attack to decrypt traffic
5069 when the connection uses an AES CBC cipher and the server support
5072 This issue was introduced as part of the fix for Lucky 13 padding
5073 attack ([CVE-2013-0169]). The padding check was rewritten to be in
5074 constant time by making sure that always the same bytes are read and
5075 compared against either the MAC or padding bytes. But it no longer
5076 checked that there was enough data to have both the MAC and padding
5079 This issue was reported by Juraj Somorovsky using TLS-Attacker.
5083 * Fix EVP_EncodeUpdate overflow
5085 An overflow can occur in the EVP_EncodeUpdate() function which is used for
5086 Base64 encoding of binary data. If an attacker is able to supply very large
5087 amounts of input data then a length check can overflow resulting in a heap
5090 Internally to OpenSSL the EVP_EncodeUpdate() function is primarily used by
5091 the `PEM_write_bio*` family of functions. These are mainly used within the
5092 OpenSSL command line applications, so any application which processes data
5093 from an untrusted source and outputs it as a PEM file should be considered
5094 vulnerable to this issue. User applications that call these APIs directly
5095 with large amounts of untrusted data may also be vulnerable.
5097 This issue was reported by Guido Vranken.
5102 * Fix EVP_EncryptUpdate overflow
5104 An overflow can occur in the EVP_EncryptUpdate() function. If an attacker
5105 is able to supply very large amounts of input data after a previous call to
5106 EVP_EncryptUpdate() with a partial block then a length check can overflow
5107 resulting in a heap corruption. Following an analysis of all OpenSSL
5108 internal usage of the EVP_EncryptUpdate() function all usage is one of two
5109 forms. The first form is where the EVP_EncryptUpdate() call is known to be
5110 the first called function after an EVP_EncryptInit(), and therefore that
5111 specific call must be safe. The second form is where the length passed to
5112 EVP_EncryptUpdate() can be seen from the code to be some small value and
5113 therefore there is no possibility of an overflow. Since all instances are
5114 one of these two forms, it is believed that there can be no overflows in
5115 internal code due to this problem. It should be noted that
5116 EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths.
5117 Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances
5118 of these calls have also been analysed too and it is believed there are no
5119 instances in internal usage where an overflow could occur.
5121 This issue was reported by Guido Vranken.
5126 * Prevent ASN.1 BIO excessive memory allocation
5128 When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio()
5129 a short invalid encoding can cause allocation of large amounts of memory
5130 potentially consuming excessive resources or exhausting memory.
5132 Any application parsing untrusted data through d2i BIO functions is
5133 affected. The memory based functions such as d2i_X509() are *not* affected.
5134 Since the memory based functions are used by the TLS library, TLS
5135 applications are not affected.
5137 This issue was reported by Brian Carpenter.
5144 ASN1 Strings that are over 1024 bytes can cause an overread in applications
5145 using the X509_NAME_oneline() function on EBCDIC systems. This could result
5146 in arbitrary stack data being returned in the buffer.
5148 This issue was reported by Guido Vranken.
5153 * Modify behavior of ALPN to invoke callback after SNI/servername
5154 callback, such that updates to the SSL_CTX affect ALPN.
5158 * Remove LOW from the DEFAULT cipher list. This removes singles DES from the
5163 * Only remove the SSLv2 methods with the no-ssl2-method option. When the
5164 methods are enabled and ssl2 is disabled the methods return NULL.
5168 ### Changes between 1.0.2f and 1.0.2g [1 Mar 2016]
5170 * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL.
5171 Builds that are not configured with "enable-weak-ssl-ciphers" will not
5172 provide any "EXPORT" or "LOW" strength ciphers.
5176 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2
5177 is by default disabled at build-time. Builds that are not configured with
5178 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used,
5179 users who want to negotiate SSLv2 via the version-flexible SSLv23_method()
5180 will need to explicitly call either of:
5182 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2);
5184 SSL_clear_options(ssl, SSL_OP_NO_SSLv2);
5186 as appropriate. Even if either of those is used, or the application
5187 explicitly uses the version-specific SSLv2_method() or its client and
5188 server variants, SSLv2 ciphers vulnerable to exhaustive search key
5189 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT
5190 ciphers, and SSLv2 56-bit DES are no longer available.
5195 * Fix a double-free in DSA code
5197 A double free bug was discovered when OpenSSL parses malformed DSA private
5198 keys and could lead to a DoS attack or memory corruption for applications
5199 that receive DSA private keys from untrusted sources. This scenario is
5202 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using
5208 * Disable SRP fake user seed to address a server memory leak.
5210 Add a new method SRP_VBASE_get1_by_user that handles the seed properly.
5212 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
5213 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user
5214 was changed to ignore the "fake user" SRP seed, even if the seed
5217 Users should use SRP_VBASE_get1_by_user instead. Note that in
5218 SRP_VBASE_get1_by_user, caller must free the returned value. Note
5219 also that even though configuring the SRP seed attempts to hide
5220 invalid usernames by continuing the handshake with fake
5221 credentials, this behaviour is not constant time and no strong
5222 guarantees are made that the handshake is indistinguishable from
5223 that of a valid user.
5228 * Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption
5230 In the BN_hex2bn function the number of hex digits is calculated using an
5231 int value `i`. Later `bn_expand` is called with a value of `i * 4`. For
5232 large values of `i` this can result in `bn_expand` not allocating any
5233 memory because `i * 4` is negative. This can leave the internal BIGNUM data
5234 field as NULL leading to a subsequent NULL ptr deref. For very large values
5235 of `i`, the calculation `i * 4` could be a positive value smaller than `i`.
5236 In this case memory is allocated to the internal BIGNUM data field, but it
5237 is insufficiently sized leading to heap corruption. A similar issue exists
5238 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn
5239 is ever called by user applications with very large untrusted hex/dec data.
5240 This is anticipated to be a rare occurrence.
5242 All OpenSSL internal usage of these functions use data that is not expected
5243 to be untrusted, e.g. config file data or application command line
5244 arguments. If user developed applications generate config file data based
5245 on untrusted data then it is possible that this could also lead to security
5246 consequences. This is also anticipated to be rare.
5248 This issue was reported to OpenSSL by Guido Vranken.
5253 * Fix memory issues in `BIO_*printf` functions
5255 The internal `fmtstr` function used in processing a "%s" format string in
5256 the `BIO_*printf` functions could overflow while calculating the length of a
5257 string and cause an OOB read when printing very long strings.
5259 Additionally the internal `doapr_outch` function can attempt to write to an
5260 OOB memory location (at an offset from the NULL pointer) in the event of a
5261 memory allocation failure. In 1.0.2 and below this could be caused where
5262 the size of a buffer to be allocated is greater than INT_MAX. E.g. this
5263 could be in processing a very long "%s" format string. Memory leaks can
5266 The first issue may mask the second issue dependent on compiler behaviour.
5267 These problems could enable attacks where large amounts of untrusted data
5268 is passed to the `BIO_*printf` functions. If applications use these functions
5269 in this way then they could be vulnerable. OpenSSL itself uses these
5270 functions when printing out human-readable dumps of ASN.1 data. Therefore
5271 applications that print this data could be vulnerable if the data is from
5272 untrusted sources. OpenSSL command line applications could also be
5273 vulnerable where they print out ASN.1 data, or if untrusted data is passed
5274 as command line arguments.
5276 Libssl is not considered directly vulnerable. Additionally certificates etc
5277 received via remote connections via libssl are also unlikely to be able to
5278 trigger these issues because of message size limits enforced within libssl.
5280 This issue was reported to OpenSSL Guido Vranken.
5285 * Side channel attack on modular exponentiation
5287 A side-channel attack was found which makes use of cache-bank conflicts on
5288 the Intel Sandy-Bridge microarchitecture which could lead to the recovery
5289 of RSA keys. The ability to exploit this issue is limited as it relies on
5290 an attacker who has control of code in a thread running on the same
5291 hyper-threaded core as the victim thread which is performing decryptions.
5293 This issue was reported to OpenSSL by Yuval Yarom, The University of
5294 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and
5295 Nadia Heninger, University of Pennsylvania with more information at
5296 <http://cachebleed.info>.
5301 * Change the `req` command to generate a 2048-bit RSA/DSA key by default,
5302 if no keysize is specified with default_bits. This fixes an
5303 omission in an earlier change that changed all RSA/DSA key generation
5304 commands to use 2048 bits by default.
5308 ### Changes between 1.0.2e and 1.0.2f [28 Jan 2016]
5310 * DH small subgroups
5312 Historically OpenSSL only ever generated DH parameters based on "safe"
5313 primes. More recently (in version 1.0.2) support was provided for
5314 generating X9.42 style parameter files such as those required for RFC 5114
5315 support. The primes used in such files may not be "safe". Where an
5316 application is using DH configured with parameters based on primes that are
5317 not "safe" then an attacker could use this fact to find a peer's private
5318 DH exponent. This attack requires that the attacker complete multiple
5319 handshakes in which the peer uses the same private DH exponent. For example
5320 this could be used to discover a TLS server's private DH exponent if it's
5321 reusing the private DH exponent or it's using a static DH ciphersuite.
5323 OpenSSL provides the option SSL_OP_SINGLE_DH_USE for ephemeral DH (DHE) in
5324 TLS. It is not on by default. If the option is not set then the server
5325 reuses the same private DH exponent for the life of the server process and
5326 would be vulnerable to this attack. It is believed that many popular
5327 applications do set this option and would therefore not be at risk.
5329 The fix for this issue adds an additional check where a "q" parameter is
5330 available (as is the case in X9.42 based parameters). This detects the
5331 only known attack, and is the only possible defense for static DH
5332 ciphersuites. This could have some performance impact.
5334 Additionally the SSL_OP_SINGLE_DH_USE option has been switched on by
5335 default and cannot be disabled. This could have some performance impact.
5337 This issue was reported to OpenSSL by Antonio Sanso (Adobe).
5342 * SSLv2 doesn't block disabled ciphers
5344 A malicious client can negotiate SSLv2 ciphers that have been disabled on
5345 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
5346 been disabled, provided that the SSLv2 protocol was not also disabled via
5349 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
5350 and Sebastian Schinzel.
5355 ### Changes between 1.0.2d and 1.0.2e [3 Dec 2015]
5357 * BN_mod_exp may produce incorrect results on x86_64
5359 There is a carry propagating bug in the x86_64 Montgomery squaring
5360 procedure. No EC algorithms are affected. Analysis suggests that attacks
5361 against RSA and DSA as a result of this defect would be very difficult to
5362 perform and are not believed likely. Attacks against DH are considered just
5363 feasible (although very difficult) because most of the work necessary to
5364 deduce information about a private key may be performed offline. The amount
5365 of resources required for such an attack would be very significant and
5366 likely only accessible to a limited number of attackers. An attacker would
5367 additionally need online access to an unpatched system using the target
5368 private key in a scenario with persistent DH parameters and a private
5369 key that is shared between multiple clients. For example this can occur by
5370 default in OpenSSL DHE based SSL/TLS ciphersuites.
5372 This issue was reported to OpenSSL by Hanno Böck.
5377 * Certificate verify crash with missing PSS parameter
5379 The signature verification routines will crash with a NULL pointer
5380 dereference if presented with an ASN.1 signature using the RSA PSS
5381 algorithm and absent mask generation function parameter. Since these
5382 routines are used to verify certificate signature algorithms this can be
5383 used to crash any certificate verification operation and exploited in a
5384 DoS attack. Any application which performs certificate verification is
5385 vulnerable including OpenSSL clients and servers which enable client
5388 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
5393 * X509_ATTRIBUTE memory leak
5395 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
5396 memory. This structure is used by the PKCS#7 and CMS routines so any
5397 application which reads PKCS#7 or CMS data from untrusted sources is
5398 affected. SSL/TLS is not affected.
5400 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
5406 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
5407 This changes the decoding behaviour for some invalid messages,
5408 though the change is mostly in the more lenient direction, and
5409 legacy behaviour is preserved as much as possible.
5413 * In DSA_generate_parameters_ex, if the provided seed is too short,
5416 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
5418 ### Changes between 1.0.2c and 1.0.2d [9 Jul 2015]
5420 * Alternate chains certificate forgery
5422 During certificate verification, OpenSSL will attempt to find an
5423 alternative certificate chain if the first attempt to build such a chain
5424 fails. An error in the implementation of this logic can mean that an
5425 attacker could cause certain checks on untrusted certificates to be
5426 bypassed, such as the CA flag, enabling them to use a valid leaf
5427 certificate to act as a CA and "issue" an invalid certificate.
5429 This issue was reported to OpenSSL by Adam Langley/David Benjamin
5434 ### Changes between 1.0.2b and 1.0.2c [12 Jun 2015]
5436 * Fix HMAC ABI incompatibility. The previous version introduced an ABI
5437 incompatibility in the handling of HMAC. The previous ABI has now been
5442 ### Changes between 1.0.2a and 1.0.2b [11 Jun 2015]
5444 * Malformed ECParameters causes infinite loop
5446 When processing an ECParameters structure OpenSSL enters an infinite loop
5447 if the curve specified is over a specially malformed binary polynomial
5450 This can be used to perform denial of service against any
5451 system which processes public keys, certificate requests or
5452 certificates. This includes TLS clients and TLS servers with
5453 client authentication enabled.
5455 This issue was reported to OpenSSL by Joseph Barr-Pixton.
5460 * Exploitable out-of-bounds read in X509_cmp_time
5462 X509_cmp_time does not properly check the length of the ASN1_TIME
5463 string and can read a few bytes out of bounds. In addition,
5464 X509_cmp_time accepts an arbitrary number of fractional seconds in the
5467 An attacker can use this to craft malformed certificates and CRLs of
5468 various sizes and potentially cause a segmentation fault, resulting in
5469 a DoS on applications that verify certificates or CRLs. TLS clients
5470 that verify CRLs are affected. TLS clients and servers with client
5471 authentication enabled may be affected if they use custom verification
5474 This issue was reported to OpenSSL by Robert Swiecki (Google), and
5475 independently by Hanno Böck.
5480 * PKCS7 crash with missing EnvelopedContent
5482 The PKCS#7 parsing code does not handle missing inner EncryptedContent
5483 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
5484 with missing content and trigger a NULL pointer dereference on parsing.
5486 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
5487 structures from untrusted sources are affected. OpenSSL clients and
5488 servers are not affected.
5490 This issue was reported to OpenSSL by Michal Zalewski (Google).
5495 * CMS verify infinite loop with unknown hash function
5497 When verifying a signedData message the CMS code can enter an infinite loop
5498 if presented with an unknown hash function OID. This can be used to perform
5499 denial of service against any system which verifies signedData messages using
5501 This issue was reported to OpenSSL by Johannes Bauer.
5506 * Race condition handling NewSessionTicket
5508 If a NewSessionTicket is received by a multi-threaded client when attempting to
5509 reuse a previous ticket then a race condition can occur potentially leading to
5510 a double free of the ticket data.
5515 * Only support 256-bit or stronger elliptic curves with the
5516 'ecdh_auto' setting (server) or by default (client). Of supported
5517 curves, prefer P-256 (both).
5521 ### Changes between 1.0.2 and 1.0.2a [19 Mar 2015]
5523 * ClientHello sigalgs DoS fix
5525 If a client connects to an OpenSSL 1.0.2 server and renegotiates with an
5526 invalid signature algorithms extension a NULL pointer dereference will
5527 occur. This can be exploited in a DoS attack against the server.
5529 This issue was was reported to OpenSSL by David Ramos of Stanford
5533 *Stephen Henson and Matt Caswell*
5535 * Multiblock corrupted pointer fix
5537 OpenSSL 1.0.2 introduced the "multiblock" performance improvement. This
5538 feature only applies on 64 bit x86 architecture platforms that support AES
5539 NI instructions. A defect in the implementation of "multiblock" can cause
5540 OpenSSL's internal write buffer to become incorrectly set to NULL when
5541 using non-blocking IO. Typically, when the user application is using a
5542 socket BIO for writing, this will only result in a failed connection.
5543 However if some other BIO is used then it is likely that a segmentation
5544 fault will be triggered, thus enabling a potential DoS attack.
5546 This issue was reported to OpenSSL by Daniel Danner and Rainer Mueller.
5551 * Segmentation fault in DTLSv1_listen fix
5553 The DTLSv1_listen function is intended to be stateless and processes the
5554 initial ClientHello from many peers. It is common for user code to loop
5555 over the call to DTLSv1_listen until a valid ClientHello is received with
5556 an associated cookie. A defect in the implementation of DTLSv1_listen means
5557 that state is preserved in the SSL object from one invocation to the next
5558 that can lead to a segmentation fault. Errors processing the initial
5559 ClientHello can trigger this scenario. An example of such an error could be
5560 that a DTLS1.0 only client is attempting to connect to a DTLS1.2 only
5563 This issue was reported to OpenSSL by Per Allansson.
5568 * Segmentation fault in ASN1_TYPE_cmp fix
5570 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
5571 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
5572 certificate signature algorithm consistency this can be used to crash any
5573 certificate verification operation and exploited in a DoS attack. Any
5574 application which performs certificate verification is vulnerable including
5575 OpenSSL clients and servers which enable client authentication.
5580 * Segmentation fault for invalid PSS parameters fix
5582 The signature verification routines will crash with a NULL pointer
5583 dereference if presented with an ASN.1 signature using the RSA PSS
5584 algorithm and invalid parameters. Since these routines are used to verify
5585 certificate signature algorithms this can be used to crash any
5586 certificate verification operation and exploited in a DoS attack. Any
5587 application which performs certificate verification is vulnerable including
5588 OpenSSL clients and servers which enable client authentication.
5590 This issue was was reported to OpenSSL by Brian Carpenter.
5595 * ASN.1 structure reuse memory corruption fix
5597 Reusing a structure in ASN.1 parsing may allow an attacker to cause
5598 memory corruption via an invalid write. Such reuse is and has been
5599 strongly discouraged and is believed to be rare.
5601 Applications that parse structures containing CHOICE or ANY DEFINED BY
5602 components may be affected. Certificate parsing (d2i_X509 and related
5603 functions) are however not affected. OpenSSL clients and servers are
5609 * PKCS7 NULL pointer dereferences fix
5611 The PKCS#7 parsing code does not handle missing outer ContentInfo
5612 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
5613 missing content and trigger a NULL pointer dereference on parsing.
5615 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
5616 otherwise parse PKCS#7 structures from untrusted sources are
5617 affected. OpenSSL clients and servers are not affected.
5619 This issue was reported to OpenSSL by Michal Zalewski (Google).
5624 * DoS via reachable assert in SSLv2 servers fix
5626 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
5627 servers that both support SSLv2 and enable export cipher suites by sending
5628 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
5630 This issue was discovered by Sean Burford (Google) and Emilia Käsper
5631 (OpenSSL development team).
5636 * Empty CKE with client auth and DHE fix
5638 If client auth is used then a server can seg fault in the event of a DHE
5639 ciphersuite being selected and a zero length ClientKeyExchange message
5640 being sent by the client. This could be exploited in a DoS attack.
5645 * Handshake with unseeded PRNG fix
5647 Under certain conditions an OpenSSL 1.0.2 client can complete a handshake
5648 with an unseeded PRNG. The conditions are:
5649 - The client is on a platform where the PRNG has not been seeded
5650 automatically, and the user has not seeded manually
5651 - A protocol specific client method version has been used (i.e. not
5652 SSL_client_methodv23)
5653 - A ciphersuite is used that does not require additional random data from
5654 the PRNG beyond the initial ClientHello client random (e.g. PSK-RC4-SHA).
5656 If the handshake succeeds then the client random that has been used will
5657 have been generated from a PRNG with insufficient entropy and therefore the
5658 output may be predictable.
5660 For example using the following command with an unseeded openssl will
5661 succeed on an unpatched platform:
5663 openssl s_client -psk 1a2b3c4d -tls1_2 -cipher PSK-RC4-SHA
5668 * Use After Free following d2i_ECPrivatekey error fix
5670 A malformed EC private key file consumed via the d2i_ECPrivateKey function
5671 could cause a use after free condition. This, in turn, could cause a double
5672 free in several private key parsing functions (such as d2i_PrivateKey
5673 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
5674 for applications that receive EC private keys from untrusted
5675 sources. This scenario is considered rare.
5677 This issue was discovered by the BoringSSL project and fixed in their
5683 * X509_to_X509_REQ NULL pointer deref fix
5685 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
5686 the certificate key is invalid. This function is rarely used in practice.
5688 This issue was discovered by Brian Carpenter.
5693 * Removed the export ciphers from the DEFAULT ciphers
5697 ### Changes between 1.0.1l and 1.0.2 [22 Jan 2015]
5699 * Facilitate "universal" ARM builds targeting range of ARM ISAs, e.g.
5700 ARMv5 through ARMv8, as opposite to "locking" it to single one.
5701 So far those who have to target multiple platforms would compromise
5702 and argue that binary targeting say ARMv5 would still execute on
5703 ARMv8. "Universal" build resolves this compromise by providing
5704 near-optimal performance even on newer platforms.
5708 * Accelerated NIST P-256 elliptic curve implementation for x86_64
5709 (other platforms pending).
5711 *Shay Gueron & Vlad Krasnov (Intel Corp), Andy Polyakov*
5713 * Add support for the SignedCertificateTimestampList certificate and
5714 OCSP response extensions from RFC6962.
5718 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
5719 for corner cases. (Certain input points at infinity could lead to
5720 bogus results, with non-infinity inputs mapped to infinity too.)
5724 * Initial support for PowerISA 2.0.7, first implemented in POWER8.
5725 This covers AES, SHA256/512 and GHASH. "Initial" means that most
5726 common cases are optimized and there still is room for further
5727 improvements. Vector Permutation AES for Altivec is also added.
5731 * Add support for little-endian ppc64 Linux target.
5733 *Marcelo Cerri (IBM)*
5735 * Initial support for AMRv8 ISA crypto extensions. This covers AES,
5736 SHA1, SHA256 and GHASH. "Initial" means that most common cases
5737 are optimized and there still is room for further improvements.
5738 Both 32- and 64-bit modes are supported.
5740 *Andy Polyakov, Ard Biesheuvel (Linaro)*
5742 * Improved ARMv7 NEON support.
5746 * Support for SPARC Architecture 2011 crypto extensions, first
5747 implemented in SPARC T4. This covers AES, DES, Camellia, SHA1,
5748 SHA256/512, MD5, GHASH and modular exponentiation.
5750 *Andy Polyakov, David Miller*
5752 * Accelerated modular exponentiation for Intel processors, a.k.a.
5755 *Shay Gueron & Vlad Krasnov (Intel Corp)*
5757 * Support for new and upcoming Intel processors, including AVX2,
5758 BMI and SHA ISA extensions. This includes additional "stitched"
5759 implementations, AESNI-SHA256 and GCM, and multi-buffer support
5762 This work was sponsored by Intel Corp.
5766 * Support for DTLS 1.2. This adds two sets of DTLS methods: DTLS_*_method()
5767 supports both DTLS 1.2 and 1.0 and should use whatever version the peer
5768 supports and DTLSv1_2_*_method() which supports DTLS 1.2 only.
5772 * Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
5773 this fixes a limitation in previous versions of OpenSSL.
5777 * Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
5778 MGF1 digest and OAEP label.
5782 * Add EVP support for key wrapping algorithms, to avoid problems with
5783 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
5784 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
5785 algorithms and include tests cases.
5789 * Add functions to allocate and set the fields of an ECDSA_METHOD
5792 *Douglas E. Engert, Steve Henson*
5794 * New functions OPENSSL_gmtime_diff and ASN1_TIME_diff to find the
5795 difference in days and seconds between two tm or ASN1_TIME structures.
5799 * Add -rev test option to s_server to just reverse order of characters
5800 received by client and send back to server. Also prints an abbreviated
5801 summary of the connection parameters.
5805 * New option -brief for s_client and s_server to print out a brief summary
5806 of connection parameters.
5810 * Add callbacks for arbitrary TLS extensions.
5812 *Trevor Perrin <trevp@trevp.net> and Ben Laurie*
5814 * New option -crl_download in several openssl utilities to download CRLs
5815 from CRLDP extension in certificates.
5819 * New options -CRL and -CRLform for s_client and s_server for CRLs.
5823 * New function X509_CRL_diff to generate a delta CRL from the difference
5824 of two full CRLs. Add support to "crl" utility.
5828 * New functions to set lookup_crls function and to retrieve
5829 X509_STORE from X509_STORE_CTX.
5833 * Print out deprecated issuer and subject unique ID fields in
5838 * Extend OCSP I/O functions so they can be used for simple general purpose
5839 HTTP as well as OCSP. New wrapper function which can be used to download
5840 CRLs using the OCSP API.
5844 * Delegate command line handling in s_client/s_server to SSL_CONF APIs.
5848 * `SSL_CONF*` functions. These provide a common framework for application
5849 configuration using configuration files or command lines.
5853 * SSL/TLS tracing code. This parses out SSL/TLS records using the
5854 message callback and prints the results. Needs compile time option
5855 "enable-ssl-trace". New options to s_client and s_server to enable
5860 * New ctrl and macro to retrieve supported points extensions.
5861 Print out extension in s_server and s_client.
5865 * New functions to retrieve certificate signature and signature
5870 * Add functions to retrieve and manipulate the raw cipherlist sent by a
5875 * New Suite B modes for TLS code. These use and enforce the requirements
5876 of RFC6460: restrict ciphersuites, only permit Suite B algorithms and
5877 only use Suite B curves. The Suite B modes can be set by using the
5878 strings "SUITEB128", "SUITEB192" or "SUITEB128ONLY" for the cipherstring.
5882 * New chain verification flags for Suite B levels of security. Check
5883 algorithms are acceptable when flags are set in X509_verify_cert.
5887 * Make tls1_check_chain return a set of flags indicating checks passed
5888 by a certificate chain. Add additional tests to handle client
5889 certificates: checks for matching certificate type and issuer name
5894 * If an attempt is made to use a signature algorithm not in the peer
5895 preference list abort the handshake. If client has no suitable
5896 signature algorithms in response to a certificate request do not
5897 use the certificate.
5901 * If server EC tmp key is not in client preference list abort handshake.
5905 * Add support for certificate stores in CERT structure. This makes it
5906 possible to have different stores per SSL structure or one store in
5907 the parent SSL_CTX. Include distinct stores for certificate chain
5908 verification and chain building. New ctrl SSL_CTRL_BUILD_CERT_CHAIN
5909 to build and store a certificate chain in CERT structure: returning
5910 an error if the chain cannot be built: this will allow applications
5911 to test if a chain is correctly configured.
5913 Note: if the CERT based stores are not set then the parent SSL_CTX
5914 store is used to retain compatibility with existing behaviour.
5918 * New function ssl_set_client_disabled to set a ciphersuite disabled
5919 mask based on the current session, check mask when sending client
5920 hello and checking the requested ciphersuite.
5924 * New ctrls to retrieve and set certificate types in a certificate
5925 request message. Print out received values in s_client. If certificate
5926 types is not set with custom values set sensible values based on
5927 supported signature algorithms.
5931 * Support for distinct client and server supported signature algorithms.
5935 * Add certificate callback. If set this is called whenever a certificate
5936 is required by client or server. An application can decide which
5937 certificate chain to present based on arbitrary criteria: for example
5938 supported signature algorithms. Add very simple example to s_server.
5939 This fixes many of the problems and restrictions of the existing client
5940 certificate callback: for example you can now clear an existing
5941 certificate and specify the whole chain.
5945 * Add new "valid_flags" field to CERT_PKEY structure which determines what
5946 the certificate can be used for (if anything). Set valid_flags field
5947 in new tls1_check_chain function. Simplify ssl_set_cert_masks which used
5948 to have similar checks in it.
5950 Add new "cert_flags" field to CERT structure and include a "strict mode".
5951 This enforces some TLS certificate requirements (such as only permitting
5952 certificate signature algorithms contained in the supported algorithms
5953 extension) which some implementations ignore: this option should be used
5954 with caution as it could cause interoperability issues.
5958 * Update and tidy signature algorithm extension processing. Work out
5959 shared signature algorithms based on preferences and peer algorithms
5960 and print them out in s_client and s_server. Abort handshake if no
5961 shared signature algorithms.
5965 * Add new functions to allow customised supported signature algorithms
5966 for SSL and SSL_CTX structures. Add options to s_client and s_server
5971 * New function SSL_certs_clear() to delete all references to certificates
5972 from an SSL structure. Before this once a certificate had been added
5973 it couldn't be removed.
5977 * Integrate hostname, email address and IP address checking with certificate
5978 verification. New verify options supporting checking in openssl utility.
5982 * Fixes and wildcard matching support to hostname and email checking
5983 functions. Add manual page.
5985 *Florian Weimer (Red Hat Product Security Team)*
5987 * New functions to check a hostname email or IP address against a
5988 certificate. Add options x509 utility to print results of checks against
5993 * Fix OCSP checking.
5995 *Rob Stradling <rob.stradling@comodo.com> and Ben Laurie*
5997 * Initial experimental support for explicitly trusted non-root CAs.
5998 OpenSSL still tries to build a complete chain to a root but if an
5999 intermediate CA has a trust setting included that is used. The first
6000 setting is used: whether to trust (e.g., -addtrust option to the x509
6005 * Add -trusted_first option which attempts to find certificates in the
6006 trusted store even if an untrusted chain is also supplied.
6010 * MIPS assembly pack updates: support for MIPS32r2 and SmartMIPS ASE,
6011 platform support for Linux and Android.
6015 * Support for linux-x32, ILP32 environment in x86_64 framework.
6019 * Experimental multi-implementation support for FIPS capable OpenSSL.
6020 When in FIPS mode the approved implementations are used as normal,
6021 when not in FIPS mode the internal unapproved versions are used instead.
6022 This means that the FIPS capable OpenSSL isn't forced to use the
6023 (often lower performance) FIPS implementations outside FIPS mode.
6027 * Transparently support X9.42 DH parameters when calling
6028 PEM_read_bio_DHparameters. This means existing applications can handle
6029 the new parameter format automatically.
6033 * Initial experimental support for X9.42 DH parameter format: mainly
6034 to support use of 'q' parameter for RFC5114 parameters.
6038 * Add DH parameters from RFC5114 including test data to dhtest.
6042 * Support for automatic EC temporary key parameter selection. If enabled
6043 the most preferred EC parameters are automatically used instead of
6044 hardcoded fixed parameters. Now a server just has to call:
6045 SSL_CTX_set_ecdh_auto(ctx, 1) and the server will automatically
6046 support ECDH and use the most appropriate parameters.
6050 * Enhance and tidy EC curve and point format TLS extension code. Use
6051 static structures instead of allocation if default values are used.
6052 New ctrls to set curves we wish to support and to retrieve shared curves.
6053 Print out shared curves in s_server. New options to s_server and s_client
6054 to set list of supported curves.
6058 * New ctrls to retrieve supported signature algorithms and
6059 supported curve values as an array of NIDs. Extend openssl utility
6060 to print out received values.
6064 * Add new APIs EC_curve_nist2nid and EC_curve_nid2nist which convert
6065 between NIDs and the more common NIST names such as "P-256". Enhance
6066 ecparam utility and ECC method to recognise the NIST names for curves.
6070 * Enhance SSL/TLS certificate chain handling to support different
6071 chains for each certificate instead of one chain in the parent SSL_CTX.
6075 * Support for fixed DH ciphersuite client authentication: where both
6076 server and client use DH certificates with common parameters.
6080 * Support for fixed DH ciphersuites: those requiring DH server
6085 * New function i2d_re_X509_tbs for re-encoding the TBS portion of
6087 Note: Related 1.0.2-beta specific macros X509_get_cert_info,
6088 X509_CINF_set_modified, X509_CINF_get_issuer, X509_CINF_get_extensions and
6089 X509_CINF_get_signature were reverted post internal team review.
6094 ### Changes between 1.0.1t and 1.0.1u [22 Sep 2016]
6096 * OCSP Status Request extension unbounded memory growth
6098 A malicious client can send an excessively large OCSP Status Request
6099 extension. If that client continually requests renegotiation, sending a
6100 large OCSP Status Request extension each time, then there will be unbounded
6101 memory growth on the server. This will eventually lead to a Denial Of
6102 Service attack through memory exhaustion. Servers with a default
6103 configuration are vulnerable even if they do not support OCSP. Builds using
6104 the "no-ocsp" build time option are not affected.
6106 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6111 * In order to mitigate the SWEET32 attack, the DES ciphers were moved from
6114 This issue was reported to OpenSSL Karthikeyan Bhargavan and Gaetan
6120 * OOB write in MDC2_Update()
6122 An overflow can occur in MDC2_Update() either if called directly or
6123 through the EVP_DigestUpdate() function using MDC2. If an attacker
6124 is able to supply very large amounts of input data after a previous
6125 call to EVP_EncryptUpdate() with a partial block then a length check
6126 can overflow resulting in a heap corruption.
6128 The amount of data needed is comparable to SIZE_MAX which is impractical
6131 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6136 * Malformed SHA512 ticket DoS
6138 If a server uses SHA512 for TLS session ticket HMAC it is vulnerable to a
6139 DoS attack where a malformed ticket will result in an OOB read which will
6142 The use of SHA512 in TLS session tickets is comparatively rare as it requires
6143 a custom server callback and ticket lookup mechanism.
6145 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6150 * OOB write in BN_bn2dec()
6152 The function BN_bn2dec() does not check the return value of BN_div_word().
6153 This can cause an OOB write if an application uses this function with an
6154 overly large BIGNUM. This could be a problem if an overly large certificate
6155 or CRL is printed out from an untrusted source. TLS is not affected because
6156 record limits will reject an oversized certificate before it is parsed.
6158 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6163 * OOB read in TS_OBJ_print_bio()
6165 The function TS_OBJ_print_bio() misuses OBJ_obj2txt(): the return value is
6166 the total length the OID text representation would use and not the amount
6167 of data written. This will result in OOB reads when large OIDs are
6170 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6175 * Pointer arithmetic undefined behaviour
6177 Avoid some undefined pointer arithmetic
6179 A common idiom in the codebase is to check limits in the following manner:
6182 Where "p" points to some malloc'd data of SIZE bytes and
6185 "len" here could be from some externally supplied data (e.g. from a TLS
6188 The rules of C pointer arithmetic are such that "p + len" is only well
6189 defined where len <= SIZE. Therefore the above idiom is actually
6190 undefined behaviour.
6192 For example this could cause problems if some malloc implementation
6193 provides an address for "p" such that "p + len" actually overflows for
6194 values of len that are too big and therefore p + len < limit.
6196 This issue was reported to OpenSSL by Guido Vranken
6201 * Constant time flag not preserved in DSA signing
6203 Operations in the DSA signing algorithm should run in constant time in
6204 order to avoid side channel attacks. A flaw in the OpenSSL DSA
6205 implementation means that a non-constant time codepath is followed for
6206 certain operations. This has been demonstrated through a cache-timing
6207 attack to be sufficient for an attacker to recover the private DSA key.
6209 This issue was reported by César Pereida (Aalto University), Billy Brumley
6210 (Tampere University of Technology), and Yuval Yarom (The University of
6211 Adelaide and NICTA).
6216 * DTLS buffered message DoS
6218 In a DTLS connection where handshake messages are delivered out-of-order
6219 those messages that OpenSSL is not yet ready to process will be buffered
6220 for later use. Under certain circumstances, a flaw in the logic means that
6221 those messages do not get removed from the buffer even though the handshake
6222 has been completed. An attacker could force up to approx. 15 messages to
6223 remain in the buffer when they are no longer required. These messages will
6224 be cleared when the DTLS connection is closed. The default maximum size for
6225 a message is 100k. Therefore the attacker could force an additional 1500k
6226 to be consumed per connection. By opening many simulataneous connections an
6227 attacker could cause a DoS attack through memory exhaustion.
6229 This issue was reported to OpenSSL by Quan Luo.
6234 * DTLS replay protection DoS
6236 A flaw in the DTLS replay attack protection mechanism means that records
6237 that arrive for future epochs update the replay protection "window" before
6238 the MAC for the record has been validated. This could be exploited by an
6239 attacker by sending a record for the next epoch (which does not have to
6240 decrypt or have a valid MAC), with a very large sequence number. This means
6241 that all subsequent legitimate packets are dropped causing a denial of
6242 service for a specific DTLS connection.
6244 This issue was reported to OpenSSL by the OCAP audit team.
6249 * Certificate message OOB reads
6251 In OpenSSL 1.0.2 and earlier some missing message length checks can result
6252 in OOB reads of up to 2 bytes beyond an allocated buffer. There is a
6253 theoretical DoS risk but this has not been observed in practice on common
6256 The messages affected are client certificate, client certificate request
6257 and server certificate. As a result the attack can only be performed
6258 against a client or a server which enables client authentication.
6260 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6265 ### Changes between 1.0.1s and 1.0.1t [3 May 2016]
6267 * Prevent padding oracle in AES-NI CBC MAC check
6269 A MITM attacker can use a padding oracle attack to decrypt traffic
6270 when the connection uses an AES CBC cipher and the server support
6273 This issue was introduced as part of the fix for Lucky 13 padding
6274 attack ([CVE-2013-0169]). The padding check was rewritten to be in
6275 constant time by making sure that always the same bytes are read and
6276 compared against either the MAC or padding bytes. But it no longer
6277 checked that there was enough data to have both the MAC and padding
6280 This issue was reported by Juraj Somorovsky using TLS-Attacker.
6285 * Fix EVP_EncodeUpdate overflow
6287 An overflow can occur in the EVP_EncodeUpdate() function which is used for
6288 Base64 encoding of binary data. If an attacker is able to supply very large
6289 amounts of input data then a length check can overflow resulting in a heap
6292 Internally to OpenSSL the EVP_EncodeUpdate() function is primarly used by
6293 the `PEM_write_bio*` family of functions. These are mainly used within the
6294 OpenSSL command line applications, so any application which processes data
6295 from an untrusted source and outputs it as a PEM file should be considered
6296 vulnerable to this issue. User applications that call these APIs directly
6297 with large amounts of untrusted data may also be vulnerable.
6299 This issue was reported by Guido Vranken.
6304 * Fix EVP_EncryptUpdate overflow
6306 An overflow can occur in the EVP_EncryptUpdate() function. If an attacker
6307 is able to supply very large amounts of input data after a previous call to
6308 EVP_EncryptUpdate() with a partial block then a length check can overflow
6309 resulting in a heap corruption. Following an analysis of all OpenSSL
6310 internal usage of the EVP_EncryptUpdate() function all usage is one of two
6311 forms. The first form is where the EVP_EncryptUpdate() call is known to be
6312 the first called function after an EVP_EncryptInit(), and therefore that
6313 specific call must be safe. The second form is where the length passed to
6314 EVP_EncryptUpdate() can be seen from the code to be some small value and
6315 therefore there is no possibility of an overflow. Since all instances are
6316 one of these two forms, it is believed that there can be no overflows in
6317 internal code due to this problem. It should be noted that
6318 EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths.
6319 Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances
6320 of these calls have also been analysed too and it is believed there are no
6321 instances in internal usage where an overflow could occur.
6323 This issue was reported by Guido Vranken.
6328 * Prevent ASN.1 BIO excessive memory allocation
6330 When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio()
6331 a short invalid encoding can casuse allocation of large amounts of memory
6332 potentially consuming excessive resources or exhausting memory.
6334 Any application parsing untrusted data through d2i BIO functions is
6335 affected. The memory based functions such as d2i_X509() are *not* affected.
6336 Since the memory based functions are used by the TLS library, TLS
6337 applications are not affected.
6339 This issue was reported by Brian Carpenter.
6346 ASN1 Strings that are over 1024 bytes can cause an overread in applications
6347 using the X509_NAME_oneline() function on EBCDIC systems. This could result
6348 in arbitrary stack data being returned in the buffer.
6350 This issue was reported by Guido Vranken.
6355 * Modify behavior of ALPN to invoke callback after SNI/servername
6356 callback, such that updates to the SSL_CTX affect ALPN.
6360 * Remove LOW from the DEFAULT cipher list. This removes singles DES from the
6365 * Only remove the SSLv2 methods with the no-ssl2-method option. When the
6366 methods are enabled and ssl2 is disabled the methods return NULL.
6370 ### Changes between 1.0.1r and 1.0.1s [1 Mar 2016]
6372 * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL.
6373 Builds that are not configured with "enable-weak-ssl-ciphers" will not
6374 provide any "EXPORT" or "LOW" strength ciphers.
6378 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2
6379 is by default disabled at build-time. Builds that are not configured with
6380 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used,
6381 users who want to negotiate SSLv2 via the version-flexible SSLv23_method()
6382 will need to explicitly call either of:
6384 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2);
6386 SSL_clear_options(ssl, SSL_OP_NO_SSLv2);
6388 as appropriate. Even if either of those is used, or the application
6389 explicitly uses the version-specific SSLv2_method() or its client and
6390 server variants, SSLv2 ciphers vulnerable to exhaustive search key
6391 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT
6392 ciphers, and SSLv2 56-bit DES are no longer available.
6397 * Fix a double-free in DSA code
6399 A double free bug was discovered when OpenSSL parses malformed DSA private
6400 keys and could lead to a DoS attack or memory corruption for applications
6401 that receive DSA private keys from untrusted sources. This scenario is
6404 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using
6410 * Disable SRP fake user seed to address a server memory leak.
6412 Add a new method SRP_VBASE_get1_by_user that handles the seed properly.
6414 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
6415 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user
6416 was changed to ignore the "fake user" SRP seed, even if the seed
6419 Users should use SRP_VBASE_get1_by_user instead. Note that in
6420 SRP_VBASE_get1_by_user, caller must free the returned value. Note
6421 also that even though configuring the SRP seed attempts to hide
6422 invalid usernames by continuing the handshake with fake
6423 credentials, this behaviour is not constant time and no strong
6424 guarantees are made that the handshake is indistinguishable from
6425 that of a valid user.
6430 * Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption
6432 In the BN_hex2bn function the number of hex digits is calculated using an
6433 int value `i`. Later `bn_expand` is called with a value of `i * 4`. For
6434 large values of `i` this can result in `bn_expand` not allocating any
6435 memory because `i * 4` is negative. This can leave the internal BIGNUM data
6436 field as NULL leading to a subsequent NULL ptr deref. For very large values
6437 of `i`, the calculation `i * 4` could be a positive value smaller than `i`.
6438 In this case memory is allocated to the internal BIGNUM data field, but it
6439 is insufficiently sized leading to heap corruption. A similar issue exists
6440 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn
6441 is ever called by user applications with very large untrusted hex/dec data.
6442 This is anticipated to be a rare occurrence.
6444 All OpenSSL internal usage of these functions use data that is not expected
6445 to be untrusted, e.g. config file data or application command line
6446 arguments. If user developed applications generate config file data based
6447 on untrusted data then it is possible that this could also lead to security
6448 consequences. This is also anticipated to be rare.
6450 This issue was reported to OpenSSL by Guido Vranken.
6455 * Fix memory issues in `BIO_*printf` functions
6457 The internal `fmtstr` function used in processing a "%s" format string in
6458 the `BIO_*printf` functions could overflow while calculating the length of a
6459 string and cause an OOB read when printing very long strings.
6461 Additionally the internal `doapr_outch` function can attempt to write to an
6462 OOB memory location (at an offset from the NULL pointer) in the event of a
6463 memory allocation failure. In 1.0.2 and below this could be caused where
6464 the size of a buffer to be allocated is greater than INT_MAX. E.g. this
6465 could be in processing a very long "%s" format string. Memory leaks can
6468 The first issue may mask the second issue dependent on compiler behaviour.
6469 These problems could enable attacks where large amounts of untrusted data
6470 is passed to the `BIO_*printf` functions. If applications use these functions
6471 in this way then they could be vulnerable. OpenSSL itself uses these
6472 functions when printing out human-readable dumps of ASN.1 data. Therefore
6473 applications that print this data could be vulnerable if the data is from
6474 untrusted sources. OpenSSL command line applications could also be
6475 vulnerable where they print out ASN.1 data, or if untrusted data is passed
6476 as command line arguments.
6478 Libssl is not considered directly vulnerable. Additionally certificates etc
6479 received via remote connections via libssl are also unlikely to be able to
6480 trigger these issues because of message size limits enforced within libssl.
6482 This issue was reported to OpenSSL Guido Vranken.
6487 * Side channel attack on modular exponentiation
6489 A side-channel attack was found which makes use of cache-bank conflicts on
6490 the Intel Sandy-Bridge microarchitecture which could lead to the recovery
6491 of RSA keys. The ability to exploit this issue is limited as it relies on
6492 an attacker who has control of code in a thread running on the same
6493 hyper-threaded core as the victim thread which is performing decryptions.
6495 This issue was reported to OpenSSL by Yuval Yarom, The University of
6496 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and
6497 Nadia Heninger, University of Pennsylvania with more information at
6498 <http://cachebleed.info>.
6503 * Change the req command to generate a 2048-bit RSA/DSA key by default,
6504 if no keysize is specified with default_bits. This fixes an
6505 omission in an earlier change that changed all RSA/DSA key generation
6506 commands to use 2048 bits by default.
6510 ### Changes between 1.0.1q and 1.0.1r [28 Jan 2016]
6512 * Protection for DH small subgroup attacks
6514 As a precautionary measure the SSL_OP_SINGLE_DH_USE option has been
6515 switched on by default and cannot be disabled. This could have some
6520 * SSLv2 doesn't block disabled ciphers
6522 A malicious client can negotiate SSLv2 ciphers that have been disabled on
6523 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
6524 been disabled, provided that the SSLv2 protocol was not also disabled via
6527 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
6528 and Sebastian Schinzel.
6533 * Reject DH handshakes with parameters shorter than 1024 bits.
6537 ### Changes between 1.0.1p and 1.0.1q [3 Dec 2015]
6539 * Certificate verify crash with missing PSS parameter
6541 The signature verification routines will crash with a NULL pointer
6542 dereference if presented with an ASN.1 signature using the RSA PSS
6543 algorithm and absent mask generation function parameter. Since these
6544 routines are used to verify certificate signature algorithms this can be
6545 used to crash any certificate verification operation and exploited in a
6546 DoS attack. Any application which performs certificate verification is
6547 vulnerable including OpenSSL clients and servers which enable client
6550 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
6555 * X509_ATTRIBUTE memory leak
6557 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
6558 memory. This structure is used by the PKCS#7 and CMS routines so any
6559 application which reads PKCS#7 or CMS data from untrusted sources is
6560 affected. SSL/TLS is not affected.
6562 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
6568 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
6569 This changes the decoding behaviour for some invalid messages,
6570 though the change is mostly in the more lenient direction, and
6571 legacy behaviour is preserved as much as possible.
6575 * In DSA_generate_parameters_ex, if the provided seed is too short,
6576 use a random seed, as already documented.
6578 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
6580 ### Changes between 1.0.1o and 1.0.1p [9 Jul 2015]
6582 * Alternate chains certificate forgery
6584 During certificate verfification, OpenSSL will attempt to find an
6585 alternative certificate chain if the first attempt to build such a chain
6586 fails. An error in the implementation of this logic can mean that an
6587 attacker could cause certain checks on untrusted certificates to be
6588 bypassed, such as the CA flag, enabling them to use a valid leaf
6589 certificate to act as a CA and "issue" an invalid certificate.
6591 This issue was reported to OpenSSL by Adam Langley/David Benjamin
6597 * Race condition handling PSK identify hint
6599 If PSK identity hints are received by a multi-threaded client then
6600 the values are wrongly updated in the parent SSL_CTX structure. This can
6601 result in a race condition potentially leading to a double free of the
6607 ### Changes between 1.0.1n and 1.0.1o [12 Jun 2015]
6609 * Fix HMAC ABI incompatibility. The previous version introduced an ABI
6610 incompatibility in the handling of HMAC. The previous ABI has now been
6613 ### Changes between 1.0.1m and 1.0.1n [11 Jun 2015]
6615 * Malformed ECParameters causes infinite loop
6617 When processing an ECParameters structure OpenSSL enters an infinite loop
6618 if the curve specified is over a specially malformed binary polynomial
6621 This can be used to perform denial of service against any
6622 system which processes public keys, certificate requests or
6623 certificates. This includes TLS clients and TLS servers with
6624 client authentication enabled.
6626 This issue was reported to OpenSSL by Joseph Barr-Pixton.
6631 * Exploitable out-of-bounds read in X509_cmp_time
6633 X509_cmp_time does not properly check the length of the ASN1_TIME
6634 string and can read a few bytes out of bounds. In addition,
6635 X509_cmp_time accepts an arbitrary number of fractional seconds in the
6638 An attacker can use this to craft malformed certificates and CRLs of
6639 various sizes and potentially cause a segmentation fault, resulting in
6640 a DoS on applications that verify certificates or CRLs. TLS clients
6641 that verify CRLs are affected. TLS clients and servers with client
6642 authentication enabled may be affected if they use custom verification
6645 This issue was reported to OpenSSL by Robert Swiecki (Google), and
6646 independently by Hanno Böck.
6651 * PKCS7 crash with missing EnvelopedContent
6653 The PKCS#7 parsing code does not handle missing inner EncryptedContent
6654 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
6655 with missing content and trigger a NULL pointer dereference on parsing.
6657 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
6658 structures from untrusted sources are affected. OpenSSL clients and
6659 servers are not affected.
6661 This issue was reported to OpenSSL by Michal Zalewski (Google).
6666 * CMS verify infinite loop with unknown hash function
6668 When verifying a signedData message the CMS code can enter an infinite loop
6669 if presented with an unknown hash function OID. This can be used to perform
6670 denial of service against any system which verifies signedData messages using
6672 This issue was reported to OpenSSL by Johannes Bauer.
6677 * Race condition handling NewSessionTicket
6679 If a NewSessionTicket is received by a multi-threaded client when attempting to
6680 reuse a previous ticket then a race condition can occur potentially leading to
6681 a double free of the ticket data.
6686 * Reject DH handshakes with parameters shorter than 768 bits.
6688 *Kurt Roeckx and Emilia Kasper*
6690 * dhparam: generate 2048-bit parameters by default.
6692 *Kurt Roeckx and Emilia Kasper*
6694 ### Changes between 1.0.1l and 1.0.1m [19 Mar 2015]
6696 * Segmentation fault in ASN1_TYPE_cmp fix
6698 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
6699 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
6700 certificate signature algorithm consistency this can be used to crash any
6701 certificate verification operation and exploited in a DoS attack. Any
6702 application which performs certificate verification is vulnerable including
6703 OpenSSL clients and servers which enable client authentication.
6708 * ASN.1 structure reuse memory corruption fix
6710 Reusing a structure in ASN.1 parsing may allow an attacker to cause
6711 memory corruption via an invalid write. Such reuse is and has been
6712 strongly discouraged and is believed to be rare.
6714 Applications that parse structures containing CHOICE or ANY DEFINED BY
6715 components may be affected. Certificate parsing (d2i_X509 and related
6716 functions) are however not affected. OpenSSL clients and servers are
6722 * PKCS7 NULL pointer dereferences fix
6724 The PKCS#7 parsing code does not handle missing outer ContentInfo
6725 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
6726 missing content and trigger a NULL pointer dereference on parsing.
6728 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
6729 otherwise parse PKCS#7 structures from untrusted sources are
6730 affected. OpenSSL clients and servers are not affected.
6732 This issue was reported to OpenSSL by Michal Zalewski (Google).
6737 * DoS via reachable assert in SSLv2 servers fix
6739 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
6740 servers that both support SSLv2 and enable export cipher suites by sending
6741 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
6743 This issue was discovered by Sean Burford (Google) and Emilia Käsper
6744 (OpenSSL development team).
6749 * Use After Free following d2i_ECPrivatekey error fix
6751 A malformed EC private key file consumed via the d2i_ECPrivateKey function
6752 could cause a use after free condition. This, in turn, could cause a double
6753 free in several private key parsing functions (such as d2i_PrivateKey
6754 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
6755 for applications that receive EC private keys from untrusted
6756 sources. This scenario is considered rare.
6758 This issue was discovered by the BoringSSL project and fixed in their
6764 * X509_to_X509_REQ NULL pointer deref fix
6766 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
6767 the certificate key is invalid. This function is rarely used in practice.
6769 This issue was discovered by Brian Carpenter.
6774 * Removed the export ciphers from the DEFAULT ciphers
6778 ### Changes between 1.0.1k and 1.0.1l [15 Jan 2015]
6780 * Build fixes for the Windows and OpenVMS platforms
6782 *Matt Caswell and Richard Levitte*
6784 ### Changes between 1.0.1j and 1.0.1k [8 Jan 2015]
6786 * Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
6787 message can cause a segmentation fault in OpenSSL due to a NULL pointer
6788 dereference. This could lead to a Denial Of Service attack. Thanks to
6789 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
6794 * Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
6795 dtls1_buffer_record function under certain conditions. In particular this
6796 could occur if an attacker sent repeated DTLS records with the same
6797 sequence number but for the next epoch. The memory leak could be exploited
6798 by an attacker in a Denial of Service attack through memory exhaustion.
6799 Thanks to Chris Mueller for reporting this issue.
6804 * Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
6805 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
6806 method would be set to NULL which could later result in a NULL pointer
6807 dereference. Thanks to Frank Schmirler for reporting this issue.
6812 * Abort handshake if server key exchange message is omitted for ephemeral
6815 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
6816 reporting this issue.
6821 * Remove non-export ephemeral RSA code on client and server. This code
6822 violated the TLS standard by allowing the use of temporary RSA keys in
6823 non-export ciphersuites and could be used by a server to effectively
6824 downgrade the RSA key length used to a value smaller than the server
6825 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
6826 INRIA or reporting this issue.
6831 * Fixed issue where DH client certificates are accepted without verification.
6832 An OpenSSL server will accept a DH certificate for client authentication
6833 without the certificate verify message. This effectively allows a client to
6834 authenticate without the use of a private key. This only affects servers
6835 which trust a client certificate authority which issues certificates
6836 containing DH keys: these are extremely rare and hardly ever encountered.
6837 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
6843 * Ensure that the session ID context of an SSL is updated when its
6844 SSL_CTX is updated via SSL_set_SSL_CTX.
6846 The session ID context is typically set from the parent SSL_CTX,
6847 and can vary with the CTX.
6851 * Fix various certificate fingerprint issues.
6853 By using non-DER or invalid encodings outside the signed portion of a
6854 certificate the fingerprint can be changed without breaking the signature.
6855 Although no details of the signed portion of the certificate can be changed
6856 this can cause problems with some applications: e.g. those using the
6857 certificate fingerprint for blacklists.
6859 1. Reject signatures with non zero unused bits.
6861 If the BIT STRING containing the signature has non zero unused bits reject
6862 the signature. All current signature algorithms require zero unused bits.
6864 2. Check certificate algorithm consistency.
6866 Check the AlgorithmIdentifier inside TBS matches the one in the
6867 certificate signature. NB: this will result in signature failure
6868 errors for some broken certificates.
6870 Thanks to Konrad Kraszewski from Google for reporting this issue.
6872 3. Check DSA/ECDSA signatures use DER.
6874 Re-encode DSA/ECDSA signatures and compare with the original received
6875 signature. Return an error if there is a mismatch.
6877 This will reject various cases including garbage after signature
6878 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
6879 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
6880 (negative or with leading zeroes).
6882 Further analysis was conducted and fixes were developed by Stephen Henson
6883 of the OpenSSL core team.
6889 * Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
6890 results on some platforms, including x86_64. This bug occurs at random
6891 with a very low probability, and is not known to be exploitable in any
6892 way, though its exact impact is difficult to determine. Thanks to Pieter
6893 Wuille (Blockstream) who reported this issue and also suggested an initial
6894 fix. Further analysis was conducted by the OpenSSL development team and
6895 Adam Langley of Google. The final fix was developed by Andy Polyakov of
6896 the OpenSSL core team.
6901 * Do not resume sessions on the server if the negotiated protocol
6902 version does not match the session's version. Resuming with a different
6903 version, while not strictly forbidden by the RFC, is of questionable
6904 sanity and breaks all known clients.
6906 *David Benjamin, Emilia Käsper*
6908 * Tighten handling of the ChangeCipherSpec (CCS) message: reject
6909 early CCS messages during renegotiation. (Note that because
6910 renegotiation is encrypted, this early CCS was not exploitable.)
6914 * Tighten client-side session ticket handling during renegotiation:
6915 ensure that the client only accepts a session ticket if the server sends
6916 the extension anew in the ServerHello. Previously, a TLS client would
6917 reuse the old extension state and thus accept a session ticket if one was
6918 announced in the initial ServerHello.
6920 Similarly, ensure that the client requires a session ticket if one
6921 was advertised in the ServerHello. Previously, a TLS client would
6922 ignore a missing NewSessionTicket message.
6926 ### Changes between 1.0.1i and 1.0.1j [15 Oct 2014]
6930 A flaw in the DTLS SRTP extension parsing code allows an attacker, who
6931 sends a carefully crafted handshake message, to cause OpenSSL to fail
6932 to free up to 64k of memory causing a memory leak. This could be
6933 exploited in a Denial Of Service attack. This issue affects OpenSSL
6934 1.0.1 server implementations for both SSL/TLS and DTLS regardless of
6935 whether SRTP is used or configured. Implementations of OpenSSL that
6936 have been compiled with OPENSSL_NO_SRTP defined are not affected.
6938 The fix was developed by the OpenSSL team.
6943 * Session Ticket Memory Leak.
6945 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
6946 integrity of that ticket is first verified. In the event of a session
6947 ticket integrity check failing, OpenSSL will fail to free memory
6948 causing a memory leak. By sending a large number of invalid session
6949 tickets an attacker could exploit this issue in a Denial Of Service
6955 * Build option no-ssl3 is incomplete.
6957 When OpenSSL is configured with "no-ssl3" as a build option, servers
6958 could accept and complete a SSL 3.0 handshake, and clients could be
6959 configured to send them.
6962 *Akamai and the OpenSSL team*
6964 * Add support for TLS_FALLBACK_SCSV.
6965 Client applications doing fallback retries should call
6966 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
6969 *Adam Langley, Bodo Moeller*
6971 * Add additional DigestInfo checks.
6973 Re-encode DigestInto in DER and check against the original when
6974 verifying RSA signature: this will reject any improperly encoded
6975 DigestInfo structures.
6977 Note: this is a precautionary measure and no attacks are currently known.
6981 ### Changes between 1.0.1h and 1.0.1i [6 Aug 2014]
6983 * Fix SRP buffer overrun vulnerability. Invalid parameters passed to the
6984 SRP code can be overrun an internal buffer. Add sanity check that
6985 g, A, B < N to SRP code.
6987 Thanks to Sean Devlin and Watson Ladd of Cryptography Services, NCC
6988 Group for discovering this issue.
6993 * A flaw in the OpenSSL SSL/TLS server code causes the server to negotiate
6994 TLS 1.0 instead of higher protocol versions when the ClientHello message
6995 is badly fragmented. This allows a man-in-the-middle attacker to force a
6996 downgrade to TLS 1.0 even if both the server and the client support a
6997 higher protocol version, by modifying the client's TLS records.
6999 Thanks to David Benjamin and Adam Langley (Google) for discovering and
7000 researching this issue.
7005 * OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
7006 to a denial of service attack. A malicious server can crash the client
7007 with a null pointer dereference (read) by specifying an anonymous (EC)DH
7008 ciphersuite and sending carefully crafted handshake messages.
7010 Thanks to Felix Gröbert (Google) for discovering and researching this
7016 * By sending carefully crafted DTLS packets an attacker could cause openssl
7017 to leak memory. This can be exploited through a Denial of Service attack.
7018 Thanks to Adam Langley for discovering and researching this issue.
7023 * An attacker can force openssl to consume large amounts of memory whilst
7024 processing DTLS handshake messages. This can be exploited through a
7025 Denial of Service attack.
7026 Thanks to Adam Langley for discovering and researching this issue.
7031 * An attacker can force an error condition which causes openssl to crash
7032 whilst processing DTLS packets due to memory being freed twice. This
7033 can be exploited through a Denial of Service attack.
7034 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
7040 * If a multithreaded client connects to a malicious server using a resumed
7041 session and the server sends an ec point format extension it could write
7042 up to 255 bytes to freed memory.
7044 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
7050 * A malicious server can crash an OpenSSL client with a null pointer
7051 dereference (read) by specifying an SRP ciphersuite even though it was not
7052 properly negotiated with the client. This can be exploited through a
7053 Denial of Service attack.
7055 Thanks to Joonas Kuorilehto and Riku Hietamäki (Codenomicon) for
7056 discovering and researching this issue.
7061 * A flaw in OBJ_obj2txt may cause pretty printing functions such as
7062 X509_name_oneline, X509_name_print_ex et al. to leak some information
7063 from the stack. Applications may be affected if they echo pretty printing
7064 output to the attacker.
7066 Thanks to Ivan Fratric (Google) for discovering this issue.
7069 *Emilia Käsper, and Steve Henson*
7071 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
7072 for corner cases. (Certain input points at infinity could lead to
7073 bogus results, with non-infinity inputs mapped to infinity too.)
7077 ### Changes between 1.0.1g and 1.0.1h [5 Jun 2014]
7079 * Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
7080 handshake can force the use of weak keying material in OpenSSL
7081 SSL/TLS clients and servers.
7083 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
7084 researching this issue. ([CVE-2014-0224])
7086 *KIKUCHI Masashi, Steve Henson*
7088 * Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
7089 OpenSSL DTLS client the code can be made to recurse eventually crashing
7092 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
7095 *Imre Rad, Steve Henson*
7097 * Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
7098 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
7099 client or server. This is potentially exploitable to run arbitrary
7100 code on a vulnerable client or server.
7102 Thanks to Jüri Aedla for reporting this issue. ([CVE-2014-0195])
7104 *Jüri Aedla, Steve Henson*
7106 * Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
7107 are subject to a denial of service attack.
7109 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
7110 this issue. ([CVE-2014-3470])
7112 *Felix Gröbert, Ivan Fratric, Steve Henson*
7114 * Harmonize version and its documentation. -f flag is used to display
7117 *mancha <mancha1@zoho.com>*
7119 * Fix eckey_priv_encode so it immediately returns an error upon a failure
7120 in i2d_ECPrivateKey.
7122 *mancha <mancha1@zoho.com>*
7124 * Fix some double frees. These are not thought to be exploitable.
7126 *mancha <mancha1@zoho.com>*
7128 ### Changes between 1.0.1f and 1.0.1g [7 Apr 2014]
7130 * A missing bounds check in the handling of the TLS heartbeat extension
7131 can be used to reveal up to 64k of memory to a connected client or
7134 Thanks for Neel Mehta of Google Security for discovering this bug and to
7135 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
7136 preparing the fix ([CVE-2014-0160])
7138 *Adam Langley, Bodo Moeller*
7140 * Fix for the attack described in the paper "Recovering OpenSSL
7141 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
7142 by Yuval Yarom and Naomi Benger. Details can be obtained from:
7143 <http://eprint.iacr.org/2014/140>
7145 Thanks to Yuval Yarom and Naomi Benger for discovering this
7146 flaw and to Yuval Yarom for supplying a fix ([CVE-2014-0076])
7148 *Yuval Yarom and Naomi Benger*
7150 * TLS pad extension: draft-agl-tls-padding-03
7152 Workaround for the "TLS hang bug" (see FAQ and PR#2771): if the
7153 TLS client Hello record length value would otherwise be > 255 and
7154 less that 512 pad with a dummy extension containing zeroes so it
7155 is at least 512 bytes long.
7157 *Adam Langley, Steve Henson*
7159 ### Changes between 1.0.1e and 1.0.1f [6 Jan 2014]
7161 * Fix for TLS record tampering bug. A carefully crafted invalid
7162 handshake could crash OpenSSL with a NULL pointer exception.
7163 Thanks to Anton Johansson for reporting this issues.
7166 * Keep original DTLS digest and encryption contexts in retransmission
7167 structures so we can use the previous session parameters if they need
7168 to be resent. ([CVE-2013-6450])
7172 * Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
7173 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
7174 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
7175 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
7176 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
7177 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
7179 *Rob Stradling, Adam Langley*
7181 ### Changes between 1.0.1d and 1.0.1e [11 Feb 2013]
7183 * Correct fix for CVE-2013-0169. The original didn't work on AES-NI
7184 supporting platforms or when small records were transferred.
7186 *Andy Polyakov, Steve Henson*
7188 ### Changes between 1.0.1c and 1.0.1d [5 Feb 2013]
7190 * Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
7192 This addresses the flaw in CBC record processing discovered by
7193 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
7194 at: <http://www.isg.rhul.ac.uk/tls/>
7196 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
7197 Security Group at Royal Holloway, University of London
7198 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
7199 Emilia Käsper for the initial patch.
7202 *Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson*
7204 * Fix flaw in AESNI handling of TLS 1.2 and 1.1 records for CBC mode
7205 ciphersuites which can be exploited in a denial of service attack.
7206 Thanks go to and to Adam Langley <agl@chromium.org> for discovering
7207 and detecting this bug and to Wolfgang Ettlinger
7208 <wolfgang.ettlinger@gmail.com> for independently discovering this issue.
7213 * Return an error when checking OCSP signatures when key is NULL.
7214 This fixes a DoS attack. ([CVE-2013-0166])
7218 * Make openssl verify return errors.
7220 *Chris Palmer <palmer@google.com> and Ben Laurie*
7222 * Call OCSP Stapling callback after ciphersuite has been chosen, so
7223 the right response is stapled. Also change SSL_get_certificate()
7224 so it returns the certificate actually sent.
7225 See <http://rt.openssl.org/Ticket/Display.html?id=2836>.
7227 *Rob Stradling <rob.stradling@comodo.com>*
7229 * Fix possible deadlock when decoding public keys.
7233 * Don't use TLS 1.0 record version number in initial client hello
7238 ### Changes between 1.0.1b and 1.0.1c [10 May 2012]
7240 * Sanity check record length before skipping explicit IV in TLS
7241 1.2, 1.1 and DTLS to fix DoS attack.
7243 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
7244 fuzzing as a service testing platform.
7249 * Initialise tkeylen properly when encrypting CMS messages.
7250 Thanks to Solar Designer of Openwall for reporting this issue.
7254 * In FIPS mode don't try to use composite ciphers as they are not
7259 ### Changes between 1.0.1a and 1.0.1b [26 Apr 2012]
7261 * OpenSSL 1.0.0 sets SSL_OP_ALL to 0x80000FFFL and OpenSSL 1.0.1 and
7262 1.0.1a set SSL_OP_NO_TLSv1_1 to 0x00000400L which would unfortunately
7263 mean any application compiled against OpenSSL 1.0.0 headers setting
7264 SSL_OP_ALL would also set SSL_OP_NO_TLSv1_1, unintentionally disabling
7265 TLS 1.1 also. Fix this by changing the value of SSL_OP_NO_TLSv1_1 to
7266 0x10000000L Any application which was previously compiled against
7267 OpenSSL 1.0.1 or 1.0.1a headers and which cares about SSL_OP_NO_TLSv1_1
7268 will need to be recompiled as a result. Letting be results in
7269 inability to disable specifically TLS 1.1 and in client context,
7270 in unlike event, limit maximum offered version to TLS 1.0 [see below].
7274 * In order to ensure interoperability SSL_OP_NO_protocolX does not
7275 disable just protocol X, but all protocols above X *if* there are
7276 protocols *below* X still enabled. In more practical terms it means
7277 that if application wants to disable TLS1.0 in favor of TLS1.1 and
7278 above, it's not sufficient to pass `SSL_OP_NO_TLSv1`, one has to pass
7279 `SSL_OP_NO_TLSv1|SSL_OP_NO_SSLv3|SSL_OP_NO_SSLv2`. This applies to
7284 ### Changes between 1.0.1 and 1.0.1a [19 Apr 2012]
7286 * Check for potentially exploitable overflows in asn1_d2i_read_bio
7287 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
7288 in CRYPTO_realloc_clean.
7290 Thanks to Tavis Ormandy, Google Security Team, for discovering this
7291 issue and to Adam Langley <agl@chromium.org> for fixing it.
7294 *Adam Langley (Google), Tavis Ormandy, Google Security Team*
7296 * Don't allow TLS 1.2 SHA-256 ciphersuites in TLS 1.0, 1.1 connections.
7300 * Workarounds for some broken servers that "hang" if a client hello
7301 record length exceeds 255 bytes.
7303 1. Do not use record version number > TLS 1.0 in initial client
7304 hello: some (but not all) hanging servers will now work.
7305 2. If we set OPENSSL_MAX_TLS1_2_CIPHER_LENGTH this will truncate
7306 the number of ciphers sent in the client hello. This should be
7307 set to an even number, such as 50, for example by passing:
7308 -DOPENSSL_MAX_TLS1_2_CIPHER_LENGTH=50 to config or Configure.
7309 Most broken servers should now work.
7310 3. If all else fails setting OPENSSL_NO_TLS1_2_CLIENT will disable
7311 TLS 1.2 client support entirely.
7315 * Fix SEGV in Vector Permutation AES module observed in OpenSSH.
7319 ### Changes between 1.0.0h and 1.0.1 [14 Mar 2012]
7321 * Add compatibility with old MDC2 signatures which use an ASN1 OCTET
7322 STRING form instead of a DigestInfo.
7326 * The format used for MDC2 RSA signatures is inconsistent between EVP
7327 and the RSA_sign/RSA_verify functions. This was made more apparent when
7328 OpenSSL used RSA_sign/RSA_verify for some RSA signatures in particular
7329 those which went through EVP_PKEY_METHOD in 1.0.0 and later. Detect
7330 the correct format in RSA_verify so both forms transparently work.
7334 * Some servers which support TLS 1.0 can choke if we initially indicate
7335 support for TLS 1.2 and later renegotiate using TLS 1.0 in the RSA
7336 encrypted premaster secret. As a workaround use the maximum permitted
7337 client version in client hello, this should keep such servers happy
7338 and still work with previous versions of OpenSSL.
7342 * Add support for TLS/DTLS heartbeats.
7344 *Robin Seggelmann <seggelmann@fh-muenster.de>*
7346 * Add support for SCTP.
7348 *Robin Seggelmann <seggelmann@fh-muenster.de>*
7350 * Improved PRNG seeding for VOS.
7352 *Paul Green <Paul.Green@stratus.com>*
7354 * Extensive assembler packs updates, most notably:
7356 - x86[_64]: AES-NI, PCLMULQDQ, RDRAND support;
7357 - x86[_64]: SSSE3 support (SHA1, vector-permutation AES);
7358 - x86_64: bit-sliced AES implementation;
7359 - ARM: NEON support, contemporary platforms optimizations;
7360 - s390x: z196 support;
7361 - `*`: GHASH and GF(2^m) multiplication implementations;
7365 * Make TLS-SRP code conformant with RFC 5054 API cleanup
7366 (removal of unnecessary code)
7368 *Peter Sylvester <peter.sylvester@edelweb.fr>*
7370 * Add TLS key material exporter from RFC 5705.
7374 * Add DTLS-SRTP negotiation from RFC 5764.
7378 * Add Next Protocol Negotiation,
7379 <http://tools.ietf.org/html/draft-agl-tls-nextprotoneg-00>. Can be
7380 disabled with a no-npn flag to config or Configure. Code donated
7383 *Adam Langley <agl@google.com> and Ben Laurie*
7385 * Add optional 64-bit optimized implementations of elliptic curves NIST-P224,
7386 NIST-P256, NIST-P521, with constant-time single point multiplication on
7387 typical inputs. Compiler support for the nonstandard type `__uint128_t` is
7388 required to use this (present in gcc 4.4 and later, for 64-bit builds).
7389 Code made available under Apache License version 2.0.
7391 Specify "enable-ec_nistp_64_gcc_128" on the Configure (or config) command
7392 line to include this in your build of OpenSSL, and run "make depend" (or
7393 "make update"). This enables the following EC_METHODs:
7395 EC_GFp_nistp224_method()
7396 EC_GFp_nistp256_method()
7397 EC_GFp_nistp521_method()
7399 EC_GROUP_new_by_curve_name() will automatically use these (while
7400 EC_GROUP_new_curve_GFp() currently prefers the more flexible
7403 *Emilia Käsper, Adam Langley, Bodo Moeller (Google)*
7405 * Use type ossl_ssize_t instead of ssize_t which isn't available on
7406 all platforms. Move ssize_t definition from e_os.h to the public
7407 header file e_os2.h as it now appears in public header file cms.h
7411 * New -sigopt option to the ca, req and x509 utilities. Additional
7412 signature parameters can be passed using this option and in
7417 * Add RSA PSS signing function. This will generate and set the
7418 appropriate AlgorithmIdentifiers for PSS based on those in the
7419 corresponding EVP_MD_CTX structure. No application support yet.
7423 * Support for companion algorithm specific ASN1 signing routines.
7424 New function ASN1_item_sign_ctx() signs a pre-initialised
7425 EVP_MD_CTX structure and sets AlgorithmIdentifiers based on
7426 the appropriate parameters.
7430 * Add new algorithm specific ASN1 verification initialisation function
7431 to EVP_PKEY_ASN1_METHOD: this is not in EVP_PKEY_METHOD since the ASN1
7432 handling will be the same no matter what EVP_PKEY_METHOD is used.
7433 Add a PSS handler to support verification of PSS signatures: checked
7434 against a number of sample certificates.
7438 * Add signature printing for PSS. Add PSS OIDs.
7440 *Steve Henson, Martin Kaiser <lists@kaiser.cx>*
7442 * Add algorithm specific signature printing. An individual ASN1 method
7443 can now print out signatures instead of the standard hex dump.
7445 More complex signatures (e.g. PSS) can print out more meaningful
7446 information. Include DSA version that prints out the signature
7451 * Password based recipient info support for CMS library: implementing
7456 * Split password based encryption into PBES2 and PBKDF2 functions. This
7457 neatly separates the code into cipher and PBE sections and is required
7458 for some algorithms that split PBES2 into separate pieces (such as
7459 password based CMS).
7463 * Session-handling fixes:
7464 - Fix handling of connections that are resuming with a session ID,
7465 but also support Session Tickets.
7466 - Fix a bug that suppressed issuing of a new ticket if the client
7467 presented a ticket with an expired session.
7468 - Try to set the ticket lifetime hint to something reasonable.
7469 - Make tickets shorter by excluding irrelevant information.
7470 - On the client side, don't ignore renewed tickets.
7472 *Adam Langley, Bodo Moeller (Google)*
7474 * Fix PSK session representation.
7478 * Add RC4-MD5 and AESNI-SHA1 "stitched" implementations.
7480 This work was sponsored by Intel.
7484 * Add GCM support to TLS library. Some custom code is needed to split
7485 the IV between the fixed (from PRF) and explicit (from TLS record)
7486 portions. This adds all GCM ciphersuites supported by RFC5288 and
7487 RFC5289. Generalise some `AES*` cipherstrings to include GCM and
7488 add a special AESGCM string for GCM only.
7492 * Expand range of ctrls for AES GCM. Permit setting invocation
7493 field on decrypt and retrieval of invocation field only on encrypt.
7497 * Add HMAC ECC ciphersuites from RFC5289. Include SHA384 PRF support.
7498 As required by RFC5289 these ciphersuites cannot be used if for
7499 versions of TLS earlier than 1.2.
7503 * For FIPS capable OpenSSL interpret a NULL default public key method
7504 as unset and return the appropriate default but do *not* set the default.
7505 This means we can return the appropriate method in applications that
7506 switch between FIPS and non-FIPS modes.
7510 * Redirect HMAC and CMAC operations to FIPS module in FIPS mode. If an
7511 ENGINE is used then we cannot handle that in the FIPS module so we
7512 keep original code iff non-FIPS operations are allowed.
7516 * Add -attime option to openssl utilities.
7518 *Peter Eckersley <pde@eff.org>, Ben Laurie and Steve Henson*
7520 * Redirect DSA and DH operations to FIPS module in FIPS mode.
7524 * Redirect ECDSA and ECDH operations to FIPS module in FIPS mode. Also use
7525 FIPS EC methods unconditionally for now.
7529 * New build option no-ec2m to disable characteristic 2 code.
7533 * Backport libcrypto audit of return value checking from 1.1.0-dev; not
7534 all cases can be covered as some introduce binary incompatibilities.
7538 * Redirect RSA operations to FIPS module including keygen,
7539 encrypt, decrypt, sign and verify. Block use of non FIPS RSA methods.
7543 * Add similar low-level API blocking to ciphers.
7547 * low-level digest APIs are not approved in FIPS mode: any attempt
7548 to use these will cause a fatal error. Applications that *really* want
7549 to use them can use the `private_*` version instead.
7553 * Redirect cipher operations to FIPS module for FIPS builds.
7557 * Redirect digest operations to FIPS module for FIPS builds.
7561 * Update build system to add "fips" flag which will link in fipscanister.o
7562 for static and shared library builds embedding a signature if needed.
7566 * Output TLS supported curves in preference order instead of numerical
7567 order. This is currently hardcoded for the highest order curves first.
7568 This should be configurable so applications can judge speed vs strength.
7572 * Add TLS v1.2 server support for client authentication.
7576 * Add support for FIPS mode in ssl library: disable SSLv3, non-FIPS ciphers
7581 * Functions FIPS_mode_set() and FIPS_mode() which call the underlying
7582 FIPS modules versions.
7586 * Add TLS v1.2 client side support for client authentication. Keep cache
7587 of handshake records longer as we don't know the hash algorithm to use
7588 until after the certificate request message is received.
7592 * Initial TLS v1.2 client support. Add a default signature algorithms
7593 extension including all the algorithms we support. Parse new signature
7594 format in client key exchange. Relax some ECC signing restrictions for
7595 TLS v1.2 as indicated in RFC5246.
7599 * Add server support for TLS v1.2 signature algorithms extension. Switch
7600 to new signature format when needed using client digest preference.
7601 All server ciphersuites should now work correctly in TLS v1.2. No client
7602 support yet and no support for client certificates.
7606 * Initial TLS v1.2 support. Add new SHA256 digest to ssl code, switch
7607 to SHA256 for PRF when using TLS v1.2 and later. Add new SHA256 based
7608 ciphersuites. At present only RSA key exchange ciphersuites work with
7609 TLS v1.2. Add new option for TLS v1.2 replacing the old and obsolete
7610 SSL_OP_PKCS1_CHECK flags with SSL_OP_NO_TLSv1_2. New TLSv1.2 methods
7611 and version checking.
7615 * New option OPENSSL_NO_SSL_INTERN. If an application can be compiled
7616 with this defined it will not be affected by any changes to ssl internal
7617 structures. Add several utility functions to allow openssl application
7618 to work with OPENSSL_NO_SSL_INTERN defined.
7622 * A long standing patch to add support for SRP from EdelWeb (Peter
7623 Sylvester and Christophe Renou) was integrated.
7624 *Christophe Renou <christophe.renou@edelweb.fr>, Peter Sylvester
7625 <peter.sylvester@edelweb.fr>, Tom Wu <tjw@cs.stanford.edu>, and
7628 * Add functions to copy EVP_PKEY_METHOD and retrieve flags and id.
7632 * Permit abbreviated handshakes when renegotiating using the function
7633 SSL_renegotiate_abbreviated().
7635 *Robin Seggelmann <seggelmann@fh-muenster.de>*
7637 * Add call to ENGINE_register_all_complete() to
7638 ENGINE_load_builtin_engines(), so some implementations get used
7639 automatically instead of needing explicit application support.
7643 * Add support for TLS key exporter as described in RFC5705.
7645 *Robin Seggelmann <seggelmann@fh-muenster.de>, Steve Henson*
7647 * Initial TLSv1.1 support. Since TLSv1.1 is very similar to TLS v1.0 only
7648 a few changes are required:
7650 Add SSL_OP_NO_TLSv1_1 flag.
7651 Add TLSv1_1 methods.
7652 Update version checking logic to handle version 1.1.
7653 Add explicit IV handling (ported from DTLS code).
7654 Add command line options to s_client/s_server.
7661 ### Changes between 1.0.0s and 1.0.0t [3 Dec 2015]
7663 * X509_ATTRIBUTE memory leak
7665 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
7666 memory. This structure is used by the PKCS#7 and CMS routines so any
7667 application which reads PKCS#7 or CMS data from untrusted sources is
7668 affected. SSL/TLS is not affected.
7670 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
7676 * Race condition handling PSK identify hint
7678 If PSK identity hints are received by a multi-threaded client then
7679 the values are wrongly updated in the parent SSL_CTX structure. This can
7680 result in a race condition potentially leading to a double free of the
7686 ### Changes between 1.0.0r and 1.0.0s [11 Jun 2015]
7688 * Malformed ECParameters causes infinite loop
7690 When processing an ECParameters structure OpenSSL enters an infinite loop
7691 if the curve specified is over a specially malformed binary polynomial
7694 This can be used to perform denial of service against any
7695 system which processes public keys, certificate requests or
7696 certificates. This includes TLS clients and TLS servers with
7697 client authentication enabled.
7699 This issue was reported to OpenSSL by Joseph Barr-Pixton.
7704 * Exploitable out-of-bounds read in X509_cmp_time
7706 X509_cmp_time does not properly check the length of the ASN1_TIME
7707 string and can read a few bytes out of bounds. In addition,
7708 X509_cmp_time accepts an arbitrary number of fractional seconds in the
7711 An attacker can use this to craft malformed certificates and CRLs of
7712 various sizes and potentially cause a segmentation fault, resulting in
7713 a DoS on applications that verify certificates or CRLs. TLS clients
7714 that verify CRLs are affected. TLS clients and servers with client
7715 authentication enabled may be affected if they use custom verification
7718 This issue was reported to OpenSSL by Robert Swiecki (Google), and
7719 independently by Hanno Böck.
7724 * PKCS7 crash with missing EnvelopedContent
7726 The PKCS#7 parsing code does not handle missing inner EncryptedContent
7727 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
7728 with missing content and trigger a NULL pointer dereference on parsing.
7730 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
7731 structures from untrusted sources are affected. OpenSSL clients and
7732 servers are not affected.
7734 This issue was reported to OpenSSL by Michal Zalewski (Google).
7739 * CMS verify infinite loop with unknown hash function
7741 When verifying a signedData message the CMS code can enter an infinite loop
7742 if presented with an unknown hash function OID. This can be used to perform
7743 denial of service against any system which verifies signedData messages using
7745 This issue was reported to OpenSSL by Johannes Bauer.
7750 * Race condition handling NewSessionTicket
7752 If a NewSessionTicket is received by a multi-threaded client when attempting to
7753 reuse a previous ticket then a race condition can occur potentially leading to
7754 a double free of the ticket data.
7759 ### Changes between 1.0.0q and 1.0.0r [19 Mar 2015]
7761 * Segmentation fault in ASN1_TYPE_cmp fix
7763 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
7764 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
7765 certificate signature algorithm consistency this can be used to crash any
7766 certificate verification operation and exploited in a DoS attack. Any
7767 application which performs certificate verification is vulnerable including
7768 OpenSSL clients and servers which enable client authentication.
7773 * ASN.1 structure reuse memory corruption fix
7775 Reusing a structure in ASN.1 parsing may allow an attacker to cause
7776 memory corruption via an invalid write. Such reuse is and has been
7777 strongly discouraged and is believed to be rare.
7779 Applications that parse structures containing CHOICE or ANY DEFINED BY
7780 components may be affected. Certificate parsing (d2i_X509 and related
7781 functions) are however not affected. OpenSSL clients and servers are
7787 * PKCS7 NULL pointer dereferences fix
7789 The PKCS#7 parsing code does not handle missing outer ContentInfo
7790 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
7791 missing content and trigger a NULL pointer dereference on parsing.
7793 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
7794 otherwise parse PKCS#7 structures from untrusted sources are
7795 affected. OpenSSL clients and servers are not affected.
7797 This issue was reported to OpenSSL by Michal Zalewski (Google).
7802 * DoS via reachable assert in SSLv2 servers fix
7804 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
7805 servers that both support SSLv2 and enable export cipher suites by sending
7806 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
7808 This issue was discovered by Sean Burford (Google) and Emilia Käsper
7809 (OpenSSL development team).
7814 * Use After Free following d2i_ECPrivatekey error fix
7816 A malformed EC private key file consumed via the d2i_ECPrivateKey function
7817 could cause a use after free condition. This, in turn, could cause a double
7818 free in several private key parsing functions (such as d2i_PrivateKey
7819 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
7820 for applications that receive EC private keys from untrusted
7821 sources. This scenario is considered rare.
7823 This issue was discovered by the BoringSSL project and fixed in their
7829 * X509_to_X509_REQ NULL pointer deref fix
7831 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
7832 the certificate key is invalid. This function is rarely used in practice.
7834 This issue was discovered by Brian Carpenter.
7839 * Removed the export ciphers from the DEFAULT ciphers
7843 ### Changes between 1.0.0p and 1.0.0q [15 Jan 2015]
7845 * Build fixes for the Windows and OpenVMS platforms
7847 *Matt Caswell and Richard Levitte*
7849 ### Changes between 1.0.0o and 1.0.0p [8 Jan 2015]
7851 * Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
7852 message can cause a segmentation fault in OpenSSL due to a NULL pointer
7853 dereference. This could lead to a Denial Of Service attack. Thanks to
7854 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
7859 * Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
7860 dtls1_buffer_record function under certain conditions. In particular this
7861 could occur if an attacker sent repeated DTLS records with the same
7862 sequence number but for the next epoch. The memory leak could be exploited
7863 by an attacker in a Denial of Service attack through memory exhaustion.
7864 Thanks to Chris Mueller for reporting this issue.
7869 * Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
7870 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
7871 method would be set to NULL which could later result in a NULL pointer
7872 dereference. Thanks to Frank Schmirler for reporting this issue.
7877 * Abort handshake if server key exchange message is omitted for ephemeral
7880 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
7881 reporting this issue.
7886 * Remove non-export ephemeral RSA code on client and server. This code
7887 violated the TLS standard by allowing the use of temporary RSA keys in
7888 non-export ciphersuites and could be used by a server to effectively
7889 downgrade the RSA key length used to a value smaller than the server
7890 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
7891 INRIA or reporting this issue.
7896 * Fixed issue where DH client certificates are accepted without verification.
7897 An OpenSSL server will accept a DH certificate for client authentication
7898 without the certificate verify message. This effectively allows a client to
7899 authenticate without the use of a private key. This only affects servers
7900 which trust a client certificate authority which issues certificates
7901 containing DH keys: these are extremely rare and hardly ever encountered.
7902 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
7908 * Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
7909 results on some platforms, including x86_64. This bug occurs at random
7910 with a very low probability, and is not known to be exploitable in any
7911 way, though its exact impact is difficult to determine. Thanks to Pieter
7912 Wuille (Blockstream) who reported this issue and also suggested an initial
7913 fix. Further analysis was conducted by the OpenSSL development team and
7914 Adam Langley of Google. The final fix was developed by Andy Polyakov of
7915 the OpenSSL core team.
7920 * Fix various certificate fingerprint issues.
7922 By using non-DER or invalid encodings outside the signed portion of a
7923 certificate the fingerprint can be changed without breaking the signature.
7924 Although no details of the signed portion of the certificate can be changed
7925 this can cause problems with some applications: e.g. those using the
7926 certificate fingerprint for blacklists.
7928 1. Reject signatures with non zero unused bits.
7930 If the BIT STRING containing the signature has non zero unused bits reject
7931 the signature. All current signature algorithms require zero unused bits.
7933 2. Check certificate algorithm consistency.
7935 Check the AlgorithmIdentifier inside TBS matches the one in the
7936 certificate signature. NB: this will result in signature failure
7937 errors for some broken certificates.
7939 Thanks to Konrad Kraszewski from Google for reporting this issue.
7941 3. Check DSA/ECDSA signatures use DER.
7943 Reencode DSA/ECDSA signatures and compare with the original received
7944 signature. Return an error if there is a mismatch.
7946 This will reject various cases including garbage after signature
7947 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
7948 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
7949 (negative or with leading zeroes).
7951 Further analysis was conducted and fixes were developed by Stephen Henson
7952 of the OpenSSL core team.
7958 ### Changes between 1.0.0n and 1.0.0o [15 Oct 2014]
7960 * Session Ticket Memory Leak.
7962 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
7963 integrity of that ticket is first verified. In the event of a session
7964 ticket integrity check failing, OpenSSL will fail to free memory
7965 causing a memory leak. By sending a large number of invalid session
7966 tickets an attacker could exploit this issue in a Denial Of Service
7972 * Build option no-ssl3 is incomplete.
7974 When OpenSSL is configured with "no-ssl3" as a build option, servers
7975 could accept and complete a SSL 3.0 handshake, and clients could be
7976 configured to send them.
7979 *Akamai and the OpenSSL team*
7981 * Add support for TLS_FALLBACK_SCSV.
7982 Client applications doing fallback retries should call
7983 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
7986 *Adam Langley, Bodo Moeller*
7988 * Add additional DigestInfo checks.
7990 Reencode DigestInto in DER and check against the original when
7991 verifying RSA signature: this will reject any improperly encoded
7992 DigestInfo structures.
7994 Note: this is a precautionary measure and no attacks are currently known.
7998 ### Changes between 1.0.0m and 1.0.0n [6 Aug 2014]
8000 * OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
8001 to a denial of service attack. A malicious server can crash the client
8002 with a null pointer dereference (read) by specifying an anonymous (EC)DH
8003 ciphersuite and sending carefully crafted handshake messages.
8005 Thanks to Felix Gröbert (Google) for discovering and researching this
8011 * By sending carefully crafted DTLS packets an attacker could cause openssl
8012 to leak memory. This can be exploited through a Denial of Service attack.
8013 Thanks to Adam Langley for discovering and researching this issue.
8018 * An attacker can force openssl to consume large amounts of memory whilst
8019 processing DTLS handshake messages. This can be exploited through a
8020 Denial of Service attack.
8021 Thanks to Adam Langley for discovering and researching this issue.
8026 * An attacker can force an error condition which causes openssl to crash
8027 whilst processing DTLS packets due to memory being freed twice. This
8028 can be exploited through a Denial of Service attack.
8029 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
8035 * If a multithreaded client connects to a malicious server using a resumed
8036 session and the server sends an ec point format extension it could write
8037 up to 255 bytes to freed memory.
8039 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
8045 * A flaw in OBJ_obj2txt may cause pretty printing functions such as
8046 X509_name_oneline, X509_name_print_ex et al. to leak some information
8047 from the stack. Applications may be affected if they echo pretty printing
8048 output to the attacker.
8050 Thanks to Ivan Fratric (Google) for discovering this issue.
8053 *Emilia Käsper, and Steve Henson*
8055 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
8056 for corner cases. (Certain input points at infinity could lead to
8057 bogus results, with non-infinity inputs mapped to infinity too.)
8061 ### Changes between 1.0.0l and 1.0.0m [5 Jun 2014]
8063 * Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
8064 handshake can force the use of weak keying material in OpenSSL
8065 SSL/TLS clients and servers.
8067 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
8068 researching this issue. ([CVE-2014-0224])
8070 *KIKUCHI Masashi, Steve Henson*
8072 * Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
8073 OpenSSL DTLS client the code can be made to recurse eventually crashing
8076 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
8079 *Imre Rad, Steve Henson*
8081 * Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
8082 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
8083 client or server. This is potentially exploitable to run arbitrary
8084 code on a vulnerable client or server.
8086 Thanks to Jüri Aedla for reporting this issue. ([CVE-2014-0195])
8088 *Jüri Aedla, Steve Henson*
8090 * Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
8091 are subject to a denial of service attack.
8093 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
8094 this issue. ([CVE-2014-3470])
8096 *Felix Gröbert, Ivan Fratric, Steve Henson*
8098 * Harmonize version and its documentation. -f flag is used to display
8101 *mancha <mancha1@zoho.com>*
8103 * Fix eckey_priv_encode so it immediately returns an error upon a failure
8104 in i2d_ECPrivateKey.
8106 *mancha <mancha1@zoho.com>*
8108 * Fix some double frees. These are not thought to be exploitable.
8110 *mancha <mancha1@zoho.com>*
8112 * Fix for the attack described in the paper "Recovering OpenSSL
8113 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
8114 by Yuval Yarom and Naomi Benger. Details can be obtained from:
8115 <http://eprint.iacr.org/2014/140>
8117 Thanks to Yuval Yarom and Naomi Benger for discovering this
8118 flaw and to Yuval Yarom for supplying a fix ([CVE-2014-0076])
8120 *Yuval Yarom and Naomi Benger*
8122 ### Changes between 1.0.0k and 1.0.0l [6 Jan 2014]
8124 * Keep original DTLS digest and encryption contexts in retransmission
8125 structures so we can use the previous session parameters if they need
8126 to be resent. ([CVE-2013-6450])
8130 * Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
8131 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
8132 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
8133 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
8134 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
8135 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
8137 *Rob Stradling, Adam Langley*
8139 ### Changes between 1.0.0j and 1.0.0k [5 Feb 2013]
8141 * Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
8143 This addresses the flaw in CBC record processing discovered by
8144 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
8145 at: <http://www.isg.rhul.ac.uk/tls/>
8147 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
8148 Security Group at Royal Holloway, University of London
8149 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
8150 Emilia Käsper for the initial patch.
8153 *Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson*
8155 * Return an error when checking OCSP signatures when key is NULL.
8156 This fixes a DoS attack. ([CVE-2013-0166])
8160 * Call OCSP Stapling callback after ciphersuite has been chosen, so
8161 the right response is stapled. Also change SSL_get_certificate()
8162 so it returns the certificate actually sent.
8163 See <http://rt.openssl.org/Ticket/Display.html?id=2836>.
8164 (This is a backport)
8166 *Rob Stradling <rob.stradling@comodo.com>*
8168 * Fix possible deadlock when decoding public keys.
8172 ### Changes between 1.0.0i and 1.0.0j [10 May 2012]
8174 [NB: OpenSSL 1.0.0i and later 1.0.0 patch levels were released after
8177 * Sanity check record length before skipping explicit IV in DTLS
8180 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
8181 fuzzing as a service testing platform.
8186 * Initialise tkeylen properly when encrypting CMS messages.
8187 Thanks to Solar Designer of Openwall for reporting this issue.
8191 ### Changes between 1.0.0h and 1.0.0i [19 Apr 2012]
8193 * Check for potentially exploitable overflows in asn1_d2i_read_bio
8194 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
8195 in CRYPTO_realloc_clean.
8197 Thanks to Tavis Ormandy, Google Security Team, for discovering this
8198 issue and to Adam Langley <agl@chromium.org> for fixing it.
8201 *Adam Langley (Google), Tavis Ormandy, Google Security Team*
8203 ### Changes between 1.0.0g and 1.0.0h [12 Mar 2012]
8205 * Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness
8206 in CMS and PKCS7 code. When RSA decryption fails use a random key for
8207 content decryption and always return the same error. Note: this attack
8208 needs on average 2^20 messages so it only affects automated senders. The
8209 old behaviour can be re-enabled in the CMS code by setting the
8210 CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where
8211 an MMA defence is not necessary.
8212 Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering
8213 this issue. ([CVE-2012-0884])
8217 * Fix CVE-2011-4619: make sure we really are receiving a
8218 client hello before rejecting multiple SGC restarts. Thanks to
8219 Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug.
8223 ### Changes between 1.0.0f and 1.0.0g [18 Jan 2012]
8225 * Fix for DTLS DoS issue introduced by fix for CVE-2011-4109.
8226 Thanks to Antonio Martin, Enterprise Secure Access Research and
8227 Development, Cisco Systems, Inc. for discovering this bug and
8228 preparing a fix. ([CVE-2012-0050])
8232 ### Changes between 1.0.0e and 1.0.0f [4 Jan 2012]
8234 * Nadhem Alfardan and Kenny Paterson have discovered an extension
8235 of the Vaudenay padding oracle attack on CBC mode encryption
8236 which enables an efficient plaintext recovery attack against
8237 the OpenSSL implementation of DTLS. Their attack exploits timing
8238 differences arising during decryption processing. A research
8239 paper describing this attack can be found at:
8240 <http://www.isg.rhul.ac.uk/~kp/dtls.pdf>
8241 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
8242 Security Group at Royal Holloway, University of London
8243 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
8244 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
8245 for preparing the fix. ([CVE-2011-4108])
8247 *Robin Seggelmann, Michael Tuexen*
8249 * Clear bytes used for block padding of SSL 3.0 records.
8252 *Adam Langley (Google)*
8254 * Only allow one SGC handshake restart for SSL/TLS. Thanks to George
8255 Kadianakis <desnacked@gmail.com> for discovering this issue and
8256 Adam Langley for preparing the fix. ([CVE-2011-4619])
8258 *Adam Langley (Google)*
8260 * Check parameters are not NULL in GOST ENGINE. ([CVE-2012-0027])
8262 *Andrey Kulikov <amdeich@gmail.com>*
8264 * Prevent malformed RFC3779 data triggering an assertion failure.
8265 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
8266 and Rob Austein <sra@hactrn.net> for fixing it. ([CVE-2011-4577])
8268 *Rob Austein <sra@hactrn.net>*
8270 * Improved PRNG seeding for VOS.
8272 *Paul Green <Paul.Green@stratus.com>*
8274 * Fix ssl_ciph.c set-up race.
8276 *Adam Langley (Google)*
8278 * Fix spurious failures in ecdsatest.c.
8280 *Emilia Käsper (Google)*
8282 * Fix the BIO_f_buffer() implementation (which was mixing different
8283 interpretations of the `..._len` fields).
8285 *Adam Langley (Google)*
8287 * Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
8288 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
8289 threads won't reuse the same blinding coefficients.
8291 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
8292 lock to call BN_BLINDING_invert_ex, and avoids one use of
8293 BN_BLINDING_update for each BN_BLINDING structure (previously,
8294 the last update always remained unused).
8296 *Emilia Käsper (Google)*
8298 * In ssl3_clear, preserve s3->init_extra along with s3->rbuf.
8300 *Bob Buckholz (Google)*
8302 ### Changes between 1.0.0d and 1.0.0e [6 Sep 2011]
8304 * Fix bug where CRLs with nextUpdate in the past are sometimes accepted
8305 by initialising X509_STORE_CTX properly. ([CVE-2011-3207])
8307 *Kaspar Brand <ossl@velox.ch>*
8309 * Fix SSL memory handling for (EC)DH ciphersuites, in particular
8310 for multi-threaded use of ECDH. ([CVE-2011-3210])
8312 *Adam Langley (Google)*
8314 * Fix x509_name_ex_d2i memory leak on bad inputs.
8318 * Remove hard coded ecdsaWithSHA1 signature tests in ssl code and check
8319 signature public key algorithm by using OID xref utilities instead.
8320 Before this you could only use some ECC ciphersuites with SHA1 only.
8324 * Add protection against ECDSA timing attacks as mentioned in the paper
8325 by Billy Bob Brumley and Nicola Tuveri, see:
8326 <http://eprint.iacr.org/2011/232.pdf>
8328 *Billy Bob Brumley and Nicola Tuveri*
8330 ### Changes between 1.0.0c and 1.0.0d [8 Feb 2011]
8332 * Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
8334 *Neel Mehta, Adam Langley, Bodo Moeller (Google)*
8336 * Fix bug in string printing code: if *any* escaping is enabled we must
8337 escape the escape character (backslash) or the resulting string is
8342 ### Changes between 1.0.0b and 1.0.0c [2 Dec 2010]
8344 * Disable code workaround for ancient and obsolete Netscape browsers
8345 and servers: an attacker can use it in a ciphersuite downgrade attack.
8346 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
8350 * Fixed J-PAKE implementation error, originally discovered by
8351 Sebastien Martini, further info and confirmation from Stefan
8352 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
8356 ### Changes between 1.0.0a and 1.0.0b [16 Nov 2010]
8358 * Fix extension code to avoid race conditions which can result in a buffer
8359 overrun vulnerability: resumed sessions must not be modified as they can
8360 be shared by multiple threads. CVE-2010-3864
8364 * Fix WIN32 build system to correctly link an ENGINE directory into
8369 ### Changes between 1.0.0 and 1.0.0a [01 Jun 2010]
8371 * Check return value of int_rsa_verify in pkey_rsa_verifyrecover
8374 *Steve Henson, Peter-Michael Hager <hager@dortmund.net>*
8376 ### Changes between 0.9.8n and 1.0.0 [29 Mar 2010]
8378 * Add "missing" function EVP_CIPHER_CTX_copy(). This copies a cipher
8379 context. The operation can be customised via the ctrl mechanism in
8380 case ENGINEs want to include additional functionality.
8384 * Tolerate yet another broken PKCS#8 key format: private key value negative.
8388 * Add new -subject_hash_old and -issuer_hash_old options to x509 utility to
8389 output hashes compatible with older versions of OpenSSL.
8391 *Willy Weisz <weisz@vcpc.univie.ac.at>*
8393 * Fix compression algorithm handling: if resuming a session use the
8394 compression algorithm of the resumed session instead of determining
8395 it from client hello again. Don't allow server to change algorithm.
8399 * Add load_crls() function to commands tidying load_certs() too. Add option
8400 to verify utility to allow additional CRLs to be included.
8404 * Update OCSP request code to permit adding custom headers to the request:
8405 some responders need this.
8409 * The function EVP_PKEY_sign() returns <=0 on error: check return code
8412 *Julia Lawall <julia@diku.dk>*
8414 * Update verify callback code in `apps/s_cb.c` and `apps/verify.c`, it
8415 needlessly dereferenced structures, used obsolete functions and
8416 didn't handle all updated verify codes correctly.
8420 * Disable MD2 in the default configuration.
8424 * In BIO_pop() and BIO_push() use the ctrl argument (which was NULL) to
8425 indicate the initial BIO being pushed or popped. This makes it possible
8426 to determine whether the BIO is the one explicitly called or as a result
8427 of the ctrl being passed down the chain. Fix BIO_pop() and SSL BIOs so
8428 it handles reference counts correctly and doesn't zero out the I/O bio
8429 when it is not being explicitly popped. WARNING: applications which
8430 included workarounds for the old buggy behaviour will need to be modified
8431 or they could free up already freed BIOs.
8435 * Extend the uni2asc/asc2uni => OPENSSL_uni2asc/OPENSSL_asc2uni
8436 renaming to all platforms (within the 0.9.8 branch, this was
8437 done conditionally on Netware platforms to avoid a name clash).
8439 *Guenter <lists@gknw.net>*
8441 * Add ECDHE and PSK support to DTLS.
8443 *Michael Tuexen <tuexen@fh-muenster.de>*
8445 * Add CHECKED_STACK_OF macro to safestack.h, otherwise safestack can't
8450 * Add "missing" function EVP_MD_flags() (without this the only way to
8451 retrieve a digest flags is by accessing the structure directly. Update
8452 `EVP_MD_do_all*()` and `EVP_CIPHER_do_all*()` to include the name a digest
8453 or cipher is registered as in the "from" argument. Print out all
8454 registered digests in the dgst usage message instead of manually
8455 attempting to work them out.
8459 * If no SSLv2 ciphers are used don't use an SSLv2 compatible client hello:
8460 this allows the use of compression and extensions. Change default cipher
8461 string to remove SSLv2 ciphersuites. This effectively avoids ancient SSLv2
8462 by default unless an application cipher string requests it.
8466 * Alter match criteria in PKCS12_parse(). It used to try to use local
8467 key ids to find matching certificates and keys but some PKCS#12 files
8468 don't follow the (somewhat unwritten) rules and this strategy fails.
8469 Now just gather all certificates together and the first private key
8470 then look for the first certificate that matches the key.
8474 * Support use of registered digest and cipher names for dgst and cipher
8475 commands instead of having to add each one as a special case. So now
8482 openssl dgst -sha256 foo
8484 and this works for ENGINE based algorithms too.
8488 * Update Gost ENGINE to support parameter files.
8490 *Victor B. Wagner <vitus@cryptocom.ru>*
8492 * Support GeneralizedTime in ca utility.
8494 *Oliver Martin <oliver@volatilevoid.net>, Steve Henson*
8496 * Enhance the hash format used for certificate directory links. The new
8497 form uses the canonical encoding (meaning equivalent names will work
8498 even if they aren't identical) and uses SHA1 instead of MD5. This form
8499 is incompatible with the older format and as a result c_rehash should
8500 be used to rebuild symbolic links.
8504 * Make PKCS#8 the default write format for private keys, replacing the
8505 traditional format. This form is standardised, more secure and doesn't
8506 include an implicit MD5 dependency.
8510 * Add a $gcc_devteam_warn option to Configure. The idea is that any code
8511 committed to OpenSSL should pass this lot as a minimum.
8515 * Add session ticket override functionality for use by EAP-FAST.
8517 *Jouni Malinen <j@w1.fi>*
8519 * Modify HMAC functions to return a value. Since these can be implemented
8520 in an ENGINE errors can occur.
8524 * Type-checked OBJ_bsearch_ex.
8528 * Type-checked OBJ_bsearch. Also some constification necessitated
8529 by type-checking. Still to come: TXT_DB, bsearch(?),
8530 OBJ_bsearch_ex, qsort, CRYPTO_EX_DATA, ASN1_VALUE, ASN1_STRING,
8535 * New function OPENSSL_gmtime_adj() to add a specific number of days and
8536 seconds to a tm structure directly, instead of going through OS
8537 specific date routines. This avoids any issues with OS routines such
8538 as the year 2038 bug. New `*_adj()` functions for ASN1 time structures
8539 and X509_time_adj_ex() to cover the extended range. The existing
8540 X509_time_adj() is still usable and will no longer have any date issues.
8544 * Delta CRL support. New use deltas option which will attempt to locate
8545 and search any appropriate delta CRLs available.
8547 This work was sponsored by Google.
8551 * Support for CRLs partitioned by reason code. Reorganise CRL processing
8552 code and add additional score elements. Validate alternate CRL paths
8553 as part of the CRL checking and indicate a new error "CRL path validation
8554 error" in this case. Applications wanting additional details can use
8555 the verify callback and check the new "parent" field. If this is not
8556 NULL CRL path validation is taking place. Existing applications won't
8557 see this because it requires extended CRL support which is off by
8560 This work was sponsored by Google.
8564 * Support for freshest CRL extension.
8566 This work was sponsored by Google.
8570 * Initial indirect CRL support. Currently only supported in the CRLs
8571 passed directly and not via lookup. Process certificate issuer
8572 CRL entry extension and lookup CRL entries by bother issuer name
8573 and serial number. Check and process CRL issuer entry in IDP extension.
8575 This work was sponsored by Google.
8579 * Add support for distinct certificate and CRL paths. The CRL issuer
8580 certificate is validated separately in this case. Only enabled if
8581 an extended CRL support flag is set: this flag will enable additional
8582 CRL functionality in future.
8584 This work was sponsored by Google.
8588 * Add support for policy mappings extension.
8590 This work was sponsored by Google.
8594 * Fixes to pathlength constraint, self issued certificate handling,
8595 policy processing to align with RFC3280 and PKITS tests.
8597 This work was sponsored by Google.
8601 * Support for name constraints certificate extension. DN, email, DNS
8602 and URI types are currently supported.
8604 This work was sponsored by Google.
8608 * To cater for systems that provide a pointer-based thread ID rather
8609 than numeric, deprecate the current numeric thread ID mechanism and
8610 replace it with a structure and associated callback type. This
8611 mechanism allows a numeric "hash" to be extracted from a thread ID in
8612 either case, and on platforms where pointers are larger than 'long',
8613 mixing is done to help ensure the numeric 'hash' is usable even if it
8614 can't be guaranteed unique. The default mechanism is to use "&errno"
8615 as a pointer-based thread ID to distinguish between threads.
8617 Applications that want to provide their own thread IDs should now use
8618 CRYPTO_THREADID_set_callback() to register a callback that will call
8619 either CRYPTO_THREADID_set_numeric() or CRYPTO_THREADID_set_pointer().
8621 Note that ERR_remove_state() is now deprecated, because it is tied
8622 to the assumption that thread IDs are numeric. ERR_remove_state(0)
8623 to free the current thread's error state should be replaced by
8624 ERR_remove_thread_state(NULL).
8626 (This new approach replaces the functions CRYPTO_set_idptr_callback(),
8627 CRYPTO_get_idptr_callback(), and CRYPTO_thread_idptr() that existed in
8628 OpenSSL 0.9.9-dev between June 2006 and August 2008. Also, if an
8629 application was previously providing a numeric thread callback that
8630 was inappropriate for distinguishing threads, then uniqueness might
8631 have been obtained with &errno that happened immediately in the
8632 intermediate development versions of OpenSSL; this is no longer the
8633 case, the numeric thread callback will now override the automatic use
8636 *Geoff Thorpe, with help from Bodo Moeller*
8638 * Initial support for different CRL issuing certificates. This covers a
8639 simple case where the self issued certificates in the chain exist and
8640 the real CRL issuer is higher in the existing chain.
8642 This work was sponsored by Google.
8646 * Removed effectively defunct crypto/store from the build.
8650 * Revamp of STACK to provide stronger type-checking. Still to come:
8651 TXT_DB, bsearch(?), OBJ_bsearch, qsort, CRYPTO_EX_DATA, ASN1_VALUE,
8652 ASN1_STRING, CONF_VALUE.
8656 * Add a new SSL_MODE_RELEASE_BUFFERS mode flag to release unused buffer
8657 RAM on SSL connections. This option can save about 34k per idle SSL.
8661 * Revamp of LHASH to provide stronger type-checking. Still to come:
8662 STACK, TXT_DB, bsearch, qsort.
8666 * Initial support for Cryptographic Message Syntax (aka CMS) based
8667 on RFC3850, RFC3851 and RFC3852. New cms directory and cms utility,
8668 support for data, signedData, compressedData, digestedData and
8669 encryptedData, envelopedData types included. Scripts to check against
8670 RFC4134 examples draft and interop and consistency checks of many
8671 content types and variants.
8675 * Add options to enc utility to support use of zlib compression BIO.
8679 * Extend mk1mf to support importing of options and assembly language
8680 files from Configure script, currently only included in VC-WIN32.
8681 The assembly language rules can now optionally generate the source
8682 files from the associated perl scripts.
8686 * Implement remaining functionality needed to support GOST ciphersuites.
8687 Interop testing has been performed using CryptoPro implementations.
8689 *Victor B. Wagner <vitus@cryptocom.ru>*
8691 * s390x assembler pack.
8695 * ARMv4 assembler pack. ARMv4 refers to v4 and later ISA, not CPU
8700 * Implement Opaque PRF Input TLS extension as specified in
8701 draft-rescorla-tls-opaque-prf-input-00.txt. Since this is not an
8702 official specification yet and no extension type assignment by
8703 IANA exists, this extension (for now) will have to be explicitly
8704 enabled when building OpenSSL by providing the extension number
8705 to use. For example, specify an option
8707 -DTLSEXT_TYPE_opaque_prf_input=0x9527
8709 to the "config" or "Configure" script to enable the extension,
8710 assuming extension number 0x9527 (which is a completely arbitrary
8711 and unofficial assignment based on the MD5 hash of the Internet
8712 Draft). Note that by doing so, you potentially lose
8713 interoperability with other TLS implementations since these might
8714 be using the same extension number for other purposes.
8716 SSL_set_tlsext_opaque_prf_input(ssl, src, len) is used to set the
8717 opaque PRF input value to use in the handshake. This will create
8718 an internal copy of the length-'len' string at 'src', and will
8719 return non-zero for success.
8721 To get more control and flexibility, provide a callback function
8724 SSL_CTX_set_tlsext_opaque_prf_input_callback(ctx, cb)
8725 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg(ctx, arg)
8729 int (*cb)(SSL *, void *peerinput, size_t len, void *arg);
8732 Callback function 'cb' will be called in handshakes, and is
8733 expected to use SSL_set_tlsext_opaque_prf_input() as appropriate.
8734 Argument 'arg' is for application purposes (the value as given to
8735 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg() will directly
8736 be provided to the callback function). The callback function
8737 has to return non-zero to report success: usually 1 to use opaque
8738 PRF input just if possible, or 2 to enforce use of the opaque PRF
8739 input. In the latter case, the library will abort the handshake
8740 if opaque PRF input is not successfully negotiated.
8742 Arguments 'peerinput' and 'len' given to the callback function
8743 will always be NULL and 0 in the case of a client. A server will
8744 see the client's opaque PRF input through these variables if
8745 available (NULL and 0 otherwise). Note that if the server
8746 provides an opaque PRF input, the length must be the same as the
8747 length of the client's opaque PRF input.
8749 Note that the callback function will only be called when creating
8750 a new session (session resumption can resume whatever was
8751 previously negotiated), and will not be called in SSL 2.0
8752 handshakes; thus, SSL_CTX_set_options(ctx, SSL_OP_NO_SSLv2) or
8753 SSL_set_options(ssl, SSL_OP_NO_SSLv2) is especially recommended
8754 for applications that need to enforce opaque PRF input.
8758 * Update ssl code to support digests other than SHA1+MD5 for handshake
8761 *Victor B. Wagner <vitus@cryptocom.ru>*
8763 * Add RFC4507 support to OpenSSL. This includes the corrections in
8764 RFC4507bis. The encrypted ticket format is an encrypted encoded
8765 SSL_SESSION structure, that way new session features are automatically
8768 If a client application caches session in an SSL_SESSION structure
8769 support is transparent because tickets are now stored in the encoded
8772 The SSL_CTX structure automatically generates keys for ticket
8773 protection in servers so again support should be possible
8774 with no application modification.
8776 If a client or server wishes to disable RFC4507 support then the option
8777 SSL_OP_NO_TICKET can be set.
8779 Add a TLS extension debugging callback to allow the contents of any client
8780 or server extensions to be examined.
8782 This work was sponsored by Google.
8786 * Final changes to avoid use of pointer pointer casts in OpenSSL.
8787 OpenSSL should now compile cleanly on gcc 4.2
8789 *Peter Hartley <pdh@utter.chaos.org.uk>, Steve Henson*
8791 * Update SSL library to use new EVP_PKEY MAC API. Include generic MAC
8792 support including streaming MAC support: this is required for GOST
8793 ciphersuite support.
8795 *Victor B. Wagner <vitus@cryptocom.ru>, Steve Henson*
8797 * Add option -stream to use PKCS#7 streaming in smime utility. New
8798 function i2d_PKCS7_bio_stream() and PEM_write_PKCS7_bio_stream()
8799 to output in BER and PEM format.
8803 * Experimental support for use of HMAC via EVP_PKEY interface. This
8804 allows HMAC to be handled via the `EVP_DigestSign*()` interface. The
8805 EVP_PKEY "key" in this case is the HMAC key, potentially allowing
8806 ENGINE support for HMAC keys which are unextractable. New -mac and
8807 -macopt options to dgst utility.
8811 * New option -sigopt to dgst utility. Update dgst to use
8812 `EVP_Digest{Sign,Verify}*`. These two changes make it possible to use
8813 alternative signing parameters such as X9.31 or PSS in the dgst
8818 * Change ssl_cipher_apply_rule(), the internal function that does
8819 the work each time a ciphersuite string requests enabling
8820 ("foo+bar"), moving ("+foo+bar"), disabling ("-foo+bar", or
8821 removing ("!foo+bar") a class of ciphersuites: Now it maintains
8822 the order of disabled ciphersuites such that those ciphersuites
8823 that most recently went from enabled to disabled not only stay
8824 in order with respect to each other, but also have higher priority
8825 than other disabled ciphersuites the next time ciphersuites are
8828 This means that you can now say, e.g., "PSK:-PSK:HIGH" to enable
8829 the same ciphersuites as with "HIGH" alone, but in a specific
8830 order where the PSK ciphersuites come first (since they are the
8831 most recently disabled ciphersuites when "HIGH" is parsed).
8833 Also, change ssl_create_cipher_list() (using this new
8834 functionality) such that between otherwise identical
8835 ciphersuites, ephemeral ECDH is preferred over ephemeral DH in
8840 * Change ssl_create_cipher_list() so that it automatically
8841 arranges the ciphersuites in reasonable order before starting
8842 to process the rule string. Thus, the definition for "DEFAULT"
8843 (SSL_DEFAULT_CIPHER_LIST) now is just "ALL:!aNULL:!eNULL", but
8844 remains equivalent to `"AES:ALL:!aNULL:!eNULL:+aECDH:+kRSA:+RC4:@STRENGTH"`.
8845 This makes it much easier to arrive at a reasonable default order
8846 in applications for which anonymous ciphers are OK (meaning
8847 that you can't actually use DEFAULT).
8849 *Bodo Moeller; suggested by Victor Duchovni*
8851 * Split the SSL/TLS algorithm mask (as used for ciphersuite string
8852 processing) into multiple integers instead of setting
8853 "SSL_MKEY_MASK" bits, "SSL_AUTH_MASK" bits, "SSL_ENC_MASK",
8854 "SSL_MAC_MASK", and "SSL_SSL_MASK" bits all in a single integer.
8855 (These masks as well as the individual bit definitions are hidden
8856 away into the non-exported interface ssl/ssl_locl.h, so this
8857 change to the definition of the SSL_CIPHER structure shouldn't
8858 affect applications.) This give us more bits for each of these
8859 categories, so there is no longer a need to coagulate AES128 and
8860 AES256 into a single algorithm bit, and to coagulate Camellia128
8861 and Camellia256 into a single algorithm bit, which has led to all
8864 Thus, among other things, the kludge introduced in 0.9.7m and
8865 0.9.8e for masking out AES256 independently of AES128 or masking
8866 out Camellia256 independently of AES256 is not needed here in 0.9.9.
8868 With the change, we also introduce new ciphersuite aliases that
8869 so far were missing: "AES128", "AES256", "CAMELLIA128", and
8874 * Add support for dsa-with-SHA224 and dsa-with-SHA256.
8875 Use the leftmost N bytes of the signature input if the input is
8876 larger than the prime q (with N being the size in bytes of q).
8880 * Very *very* experimental PKCS#7 streaming encoder support. Nothing uses
8881 it yet and it is largely untested.
8885 * Add support for the ecdsa-with-SHA224/256/384/512 signature types.
8889 * Initial incomplete changes to avoid need for function casts in OpenSSL
8890 some compilers (gcc 4.2 and later) reject their use. Safestack is
8891 reimplemented. Update ASN1 to avoid use of legacy functions.
8895 * Win32/64 targets are linked with Winsock2.
8899 * Add an X509_CRL_METHOD structure to allow CRL processing to be redirected
8900 to external functions. This can be used to increase CRL handling
8901 efficiency especially when CRLs are very large by (for example) storing
8902 the CRL revoked certificates in a database.
8906 * Overhaul of by_dir code. Add support for dynamic loading of CRLs so
8907 new CRLs added to a directory can be used. New command line option
8908 -verify_return_error to s_client and s_server. This causes real errors
8909 to be returned by the verify callback instead of carrying on no matter
8910 what. This reflects the way a "real world" verify callback would behave.
8914 * GOST engine, supporting several GOST algorithms and public key formats.
8915 Kindly donated by Cryptocom.
8919 * Partial support for Issuing Distribution Point CRL extension. CRLs
8920 partitioned by DP are handled but no indirect CRL or reason partitioning
8921 (yet). Complete overhaul of CRL handling: now the most suitable CRL is
8922 selected via a scoring technique which handles IDP and AKID in CRLs.
8926 * New X509_STORE_CTX callbacks lookup_crls() and lookup_certs() which
8927 will ultimately be used for all verify operations: this will remove the
8928 X509_STORE dependency on certificate verification and allow alternative
8929 lookup methods. X509_STORE based implementations of these two callbacks.
8933 * Allow multiple CRLs to exist in an X509_STORE with matching issuer names.
8934 Modify get_crl() to find a valid (unexpired) CRL if possible.
8938 * New function X509_CRL_match() to check if two CRLs are identical. Normally
8939 this would be called X509_CRL_cmp() but that name is already used by
8940 a function that just compares CRL issuer names. Cache several CRL
8941 extensions in X509_CRL structure and cache CRLDP in X509.
8945 * Store a "canonical" representation of X509_NAME structure (ASN1 Name)
8946 this maps equivalent X509_NAME structures into a consistent structure.
8947 Name comparison can then be performed rapidly using memcmp().
8951 * Non-blocking OCSP request processing. Add -timeout option to ocsp
8956 * Allow digests to supply their own micalg string for S/MIME type using
8957 the ctrl EVP_MD_CTRL_MICALG.
8961 * During PKCS7 signing pass the PKCS7 SignerInfo structure to the
8962 EVP_PKEY_METHOD before and after signing via the EVP_PKEY_CTRL_PKCS7_SIGN
8963 ctrl. It can then customise the structure before and/or after signing
8968 * New function OBJ_add_sigid() to allow application defined signature OIDs
8969 to be added to OpenSSLs internal tables. New function OBJ_sigid_free()
8970 to free up any added signature OIDs.
8974 * New functions EVP_CIPHER_do_all(), EVP_CIPHER_do_all_sorted(),
8975 EVP_MD_do_all() and EVP_MD_do_all_sorted() to enumerate internal
8976 digest and cipher tables. New options added to openssl utility:
8977 list-message-digest-algorithms and list-cipher-algorithms.
8981 * Change the array representation of binary polynomials: the list
8982 of degrees of non-zero coefficients is now terminated with -1.
8983 Previously it was terminated with 0, which was also part of the
8984 value; thus, the array representation was not applicable to
8985 polynomials where t^0 has coefficient zero. This change makes
8986 the array representation useful in a more general context.
8990 * Various modifications and fixes to SSL/TLS cipher string
8991 handling. For ECC, the code now distinguishes between fixed ECDH
8992 with RSA certificates on the one hand and with ECDSA certificates
8993 on the other hand, since these are separate ciphersuites. The
8994 unused code for Fortezza ciphersuites has been removed.
8996 For consistency with EDH, ephemeral ECDH is now called "EECDH"
8997 (not "ECDHE"). For consistency with the code for DH
8998 certificates, use of ECDH certificates is now considered ECDH
8999 authentication, not RSA or ECDSA authentication (the latter is
9000 merely the CA's signing algorithm and not actively used in the
9003 The temporary ciphersuite alias "ECCdraft" is no longer
9004 available, and ECC ciphersuites are no longer excluded from "ALL"
9005 and "DEFAULT". The following aliases now exist for RFC 4492
9006 ciphersuites, most of these by analogy with the DH case:
9008 kECDHr - ECDH cert, signed with RSA
9009 kECDHe - ECDH cert, signed with ECDSA
9010 kECDH - ECDH cert (signed with either RSA or ECDSA)
9011 kEECDH - ephemeral ECDH
9012 ECDH - ECDH cert or ephemeral ECDH
9018 AECDH - anonymous ECDH
9019 EECDH - non-anonymous ephemeral ECDH (equivalent to "kEECDH:-AECDH")
9023 * Add additional S/MIME capabilities for AES and GOST ciphers if supported.
9024 Use correct micalg parameters depending on digest(s) in signed message.
9028 * Add engine support for EVP_PKEY_ASN1_METHOD. Add functions to process
9029 an ENGINE asn1 method. Support ENGINE lookups in the ASN1 code.
9033 * Initial engine support for EVP_PKEY_METHOD. New functions to permit
9034 an engine to register a method. Add ENGINE lookups for methods and
9035 functional reference processing.
9039 * New functions `EVP_Digest{Sign,Verify)*`. These are enhanced versions of
9040 `EVP_{Sign,Verify}*` which allow an application to customise the signature
9045 * New -resign option to smime utility. This adds one or more signers
9046 to an existing PKCS#7 signedData structure. Also -md option to use an
9047 alternative message digest algorithm for signing.
9051 * Tidy up PKCS#7 routines and add new functions to make it easier to
9052 create PKCS7 structures containing multiple signers. Update smime
9053 application to support multiple signers.
9057 * New -macalg option to pkcs12 utility to allow setting of an alternative
9062 * Initial support for PKCS#5 v2.0 PRFs other than default SHA1 HMAC.
9063 Reorganize PBE internals to lookup from a static table using NIDs,
9064 add support for HMAC PBE OID translation. Add a EVP_CIPHER ctrl:
9065 EVP_CTRL_PBE_PRF_NID this allows a cipher to specify an alternative
9066 PRF which will be automatically used with PBES2.
9070 * Replace the algorithm specific calls to generate keys in "req" with the
9075 * Update PKCS#7 enveloped data routines to use new API. This is now
9076 supported by any public key method supporting the encrypt operation. A
9077 ctrl is added to allow the public key algorithm to examine or modify
9078 the PKCS#7 RecipientInfo structure if it needs to: for RSA this is
9083 * Add a ctrl to asn1 method to allow a public key algorithm to express
9084 a default digest type to use. In most cases this will be SHA1 but some
9085 algorithms (such as GOST) need to specify an alternative digest. The
9086 return value indicates how strong the preference is 1 means optional and
9087 2 is mandatory (that is it is the only supported type). Modify
9088 ASN1_item_sign() to accept a NULL digest argument to indicate it should
9089 use the default md. Update openssl utilities to use the default digest
9090 type for signing if it is not explicitly indicated.
9094 * Use OID cross reference table in ASN1_sign() and ASN1_verify(). New
9095 EVP_MD flag EVP_MD_FLAG_PKEY_METHOD_SIGNATURE. This uses the relevant
9096 signing method from the key type. This effectively removes the link
9097 between digests and public key types.
9101 * Add an OID cross reference table and utility functions. Its purpose is to
9102 translate between signature OIDs such as SHA1WithrsaEncryption and SHA1,
9103 rsaEncryption. This will allow some of the algorithm specific hackery
9104 needed to use the correct OID to be removed.
9108 * Remove algorithm specific dependencies when setting PKCS7_SIGNER_INFO
9109 structures for PKCS7_sign(). They are now set up by the relevant public
9114 * Add provisional EC pkey method with support for ECDSA and ECDH.
9118 * Add support for key derivation (agreement) in the API, DH method and
9123 * Add DSA pkey method and DH pkey methods, extend DH ASN1 method to support
9124 public and private key formats. As a side effect these add additional
9125 command line functionality not previously available: DSA signatures can be
9126 generated and verified using pkeyutl and DH key support and generation in
9133 *Oliver Tappe <zooey@hirschkaefer.de>*
9135 * New make target "install_html_docs" installs HTML renditions of the
9138 *Oliver Tappe <zooey@hirschkaefer.de>*
9140 * New utility "genpkey" this is analogous to "genrsa" etc except it can
9141 generate keys for any algorithm. Extend and update EVP_PKEY_METHOD to
9142 support key and parameter generation and add initial key generation
9143 functionality for RSA.
9147 * Add functions for main EVP_PKEY_method operations. The undocumented
9148 functions `EVP_PKEY_{encrypt,decrypt}` have been renamed to
9149 `EVP_PKEY_{encrypt,decrypt}_old`.
9153 * Initial definitions for EVP_PKEY_METHOD. This will be a high level public
9154 key API, doesn't do much yet.
9158 * New function EVP_PKEY_asn1_get0_info() to retrieve information about
9159 public key algorithms. New option to openssl utility:
9160 "list-public-key-algorithms" to print out info.
9164 * Implement the Supported Elliptic Curves Extension for
9165 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
9169 * Don't free up OIDs in OBJ_cleanup() if they are in use by EVP_MD or
9170 EVP_CIPHER structures to avoid later problems in EVP_cleanup().
9174 * New utilities pkey and pkeyparam. These are similar to algorithm specific
9175 utilities such as rsa, dsa, dsaparam etc except they process any key
9180 * Transfer public key printing routines to EVP_PKEY_ASN1_METHOD. New
9181 functions EVP_PKEY_print_public(), EVP_PKEY_print_private(),
9182 EVP_PKEY_print_param() to print public key data from an EVP_PKEY
9187 * Initial support for pluggable public key ASN1.
9188 De-spaghettify the public key ASN1 handling. Move public and private
9189 key ASN1 handling to a new EVP_PKEY_ASN1_METHOD structure. Relocate
9190 algorithm specific handling to a single module within the relevant
9191 algorithm directory. Add functions to allow (near) opaque processing
9192 of public and private key structures.
9196 * Implement the Supported Point Formats Extension for
9197 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
9201 * Add initial support for RFC 4279 PSK TLS ciphersuites. Add members
9202 for the psk identity [hint] and the psk callback functions to the
9203 SSL_SESSION, SSL and SSL_CTX structure.
9206 PSK-RC4-SHA, PSK-3DES-EDE-CBC-SHA, PSK-AES128-CBC-SHA,
9210 SSL_CTX_use_psk_identity_hint
9211 SSL_get_psk_identity_hint
9212 SSL_get_psk_identity
9213 SSL_use_psk_identity_hint
9215 *Mika Kousa and Pasi Eronen of Nokia Corporation*
9217 * Add RFC 3161 compliant time stamp request creation, response generation
9218 and response verification functionality.
9220 *Zoltán Glózik <zglozik@opentsa.org>, The OpenTSA Project*
9222 * Add initial support for TLS extensions, specifically for the server_name
9223 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
9224 have new members for a host name. The SSL data structure has an
9225 additional member `SSL_CTX *initial_ctx` so that new sessions can be
9226 stored in that context to allow for session resumption, even after the
9227 SSL has been switched to a new SSL_CTX in reaction to a client's
9228 server_name extension.
9230 New functions (subject to change):
9232 SSL_get_servername()
9233 SSL_get_servername_type()
9236 New CTRL codes and macros (subject to change):
9238 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
9239 - SSL_CTX_set_tlsext_servername_callback()
9240 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
9241 - SSL_CTX_set_tlsext_servername_arg()
9242 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
9244 openssl s_client has a new '-servername ...' option.
9246 openssl s_server has new options '-servername_host ...', '-cert2 ...',
9247 '-key2 ...', '-servername_fatal' (subject to change). This allows
9248 testing the HostName extension for a specific single host name ('-cert'
9249 and '-key' remain fallbacks for handshakes without HostName
9250 negotiation). If the unrecognized_name alert has to be sent, this by
9251 default is a warning; it becomes fatal with the '-servername_fatal'
9254 *Peter Sylvester, Remy Allais, Christophe Renou*
9256 * Whirlpool hash implementation is added.
9260 * BIGNUM code on 64-bit SPARCv9 targets is switched from bn(64,64) to
9261 bn(64,32). Because of instruction set limitations it doesn't have
9262 any negative impact on performance. This was done mostly in order
9263 to make it possible to share assembler modules, such as bn_mul_mont
9264 implementations, between 32- and 64-bit builds without hassle.
9268 * Move code previously exiled into file crypto/ec/ec2_smpt.c
9269 to ec2_smpl.c, and no longer require the OPENSSL_EC_BIN_PT_COMP
9274 * New candidate for BIGNUM assembler implementation, bn_mul_mont,
9275 dedicated Montgomery multiplication procedure, is introduced.
9276 BN_MONT_CTX is modified to allow bn_mul_mont to reach for higher
9277 "64-bit" performance on certain 32-bit targets.
9281 * New option SSL_OP_NO_COMP to disable use of compression selectively
9282 in SSL structures. New SSL ctrl to set maximum send fragment size.
9283 Save memory by setting the I/O buffer sizes dynamically instead of
9284 using the maximum available value.
9288 * New option -V for 'openssl ciphers'. This prints the ciphersuite code
9289 in addition to the text details.
9293 * Very, very preliminary EXPERIMENTAL support for printing of general
9294 ASN1 structures. This currently produces rather ugly output and doesn't
9295 handle several customised structures at all.
9299 * Integrated support for PVK file format and some related formats such
9300 as MS PUBLICKEYBLOB and PRIVATEKEYBLOB. Command line switches to support
9301 these in the 'rsa' and 'dsa' utilities.
9305 * Support for PKCS#1 RSAPublicKey format on rsa utility command line.
9309 * Remove the ancient ASN1_METHOD code. This was only ever used in one
9310 place for the (very old) "NETSCAPE" format certificates which are now
9311 handled using new ASN1 code equivalents.
9315 * Let the TLSv1_method() etc. functions return a 'const' SSL_METHOD
9316 pointer and make the SSL_METHOD parameter in SSL_CTX_new,
9317 SSL_CTX_set_ssl_version and SSL_set_ssl_method 'const'.
9321 * Modify CRL distribution points extension code to print out previously
9322 unsupported fields. Enhance extension setting code to allow setting of
9327 * Add print and set support for Issuing Distribution Point CRL extension.
9331 * Change 'Configure' script to enable Camellia by default.
9338 ### Changes between 0.9.8m and 0.9.8n [24 Mar 2010]
9340 * When rejecting SSL/TLS records due to an incorrect version number, never
9341 update s->server with a new major version number. As of
9342 - OpenSSL 0.9.8m if 'short' is a 16-bit type,
9343 - OpenSSL 0.9.8f if 'short' is longer than 16 bits,
9344 the previous behavior could result in a read attempt at NULL when
9345 receiving specific incorrect SSL/TLS records once record payload
9346 protection is active. ([CVE-2010-0740])
9348 *Bodo Moeller, Adam Langley <agl@chromium.org>*
9350 * Fix for CVE-2010-0433 where some kerberos enabled versions of OpenSSL
9351 could be crashed if the relevant tables were not present (e.g. chrooted).
9353 *Tomas Hoger <thoger@redhat.com>*
9355 ### Changes between 0.9.8l and 0.9.8m [25 Feb 2010]
9357 * Always check bn_wexpand() return values for failure. ([CVE-2009-3245])
9359 *Martin Olsson, Neel Mehta*
9361 * Fix X509_STORE locking: Every 'objs' access requires a lock (to
9362 accommodate for stack sorting, always a write lock!).
9366 * On some versions of WIN32 Heap32Next is very slow. This can cause
9367 excessive delays in the RAND_poll(): over a minute. As a workaround
9368 include a time check in the inner Heap32Next loop too.
9372 * The code that handled flushing of data in SSL/TLS originally used the
9373 BIO_CTRL_INFO ctrl to see if any data was pending first. This caused
9374 the problem outlined in PR#1949. The fix suggested there however can
9375 trigger problems with buggy BIO_CTRL_WPENDING (e.g. some versions
9376 of Apache). So instead simplify the code to flush unconditionally.
9377 This should be fine since flushing with no data to flush is a no op.
9381 * Handle TLS versions 2.0 and later properly and correctly use the
9382 highest version of TLS/SSL supported. Although TLS >= 2.0 is some way
9383 off ancient servers have a habit of sticking around for a while...
9387 * Modify compression code so it frees up structures without using the
9388 ex_data callbacks. This works around a problem where some applications
9389 call CRYPTO_cleanup_all_ex_data() before application exit (e.g. when
9390 restarting) then use compression (e.g. SSL with compression) later.
9391 This results in significant per-connection memory leaks and
9392 has caused some security issues including CVE-2008-1678 and
9397 * Constify crypto/cast (i.e., <openssl/cast.h>): a CAST_KEY doesn't
9398 change when encrypting or decrypting.
9402 * Add option SSL_OP_LEGACY_SERVER_CONNECT which will allow clients to
9403 connect and renegotiate with servers which do not support RI.
9404 Until RI is more widely deployed this option is enabled by default.
9408 * Add "missing" ssl ctrls to clear options and mode.
9412 * If client attempts to renegotiate and doesn't support RI respond with
9413 a no_renegotiation alert as required by RFC5746. Some renegotiating
9414 TLS clients will continue a connection gracefully when they receive
9415 the alert. Unfortunately OpenSSL mishandled this alert and would hang
9416 waiting for a server hello which it will never receive. Now we treat a
9417 received no_renegotiation alert as a fatal error. This is because
9418 applications requesting a renegotiation might well expect it to succeed
9419 and would have no code in place to handle the server denying it so the
9420 only safe thing to do is to terminate the connection.
9424 * Add ctrl macro SSL_get_secure_renegotiation_support() which returns 1 if
9425 peer supports secure renegotiation and 0 otherwise. Print out peer
9426 renegotiation support in s_client/s_server.
9430 * Replace the highly broken and deprecated SPKAC certification method with
9431 the updated NID creation version. This should correctly handle UTF8.
9435 * Implement RFC5746. Re-enable renegotiation but require the extension
9436 as needed. Unfortunately, SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION
9437 turns out to be a bad idea. It has been replaced by
9438 SSL_OP_ALLOW_UNSAFE_LEGACY_RENEGOTIATION which can be set with
9439 SSL_CTX_set_options(). This is really not recommended unless you
9440 know what you are doing.
9442 *Eric Rescorla <ekr@networkresonance.com>, Ben Laurie, Steve Henson*
9444 * Fixes to stateless session resumption handling. Use initial_ctx when
9445 issuing and attempting to decrypt tickets in case it has changed during
9446 servername handling. Use a non-zero length session ID when attempting
9447 stateless session resumption: this makes it possible to determine if
9448 a resumption has occurred immediately after receiving server hello
9449 (several places in OpenSSL subtly assume this) instead of later in
9454 * The functions ENGINE_ctrl(), OPENSSL_isservice(),
9455 CMS_get1_RecipientRequest() and RAND_bytes() can return <=0 on error
9456 fixes for a few places where the return code is not checked
9459 *Julia Lawall <julia@diku.dk>*
9461 * Add --strict-warnings option to Configure script to include devteam
9462 warnings in other configurations.
9466 * Add support for --libdir option and LIBDIR variable in makefiles. This
9467 makes it possible to install openssl libraries in locations which
9468 have names other than "lib", for example "/usr/lib64" which some
9471 *Steve Henson, based on patch from Jeremy Utley*
9473 * Don't allow the use of leading 0x80 in OIDs. This is a violation of
9474 X690 8.9.12 and can produce some misleading textual output of OIDs.
9476 *Steve Henson, reported by Dan Kaminsky*
9478 * Delete MD2 from algorithm tables. This follows the recommendation in
9479 several standards that it is not used in new applications due to
9480 several cryptographic weaknesses. For binary compatibility reasons
9481 the MD2 API is still compiled in by default.
9485 * Add compression id to {d2i,i2d}_SSL_SESSION so it is correctly saved
9490 * Rename uni2asc and asc2uni functions to OPENSSL_uni2asc and
9491 OPENSSL_asc2uni conditionally on Netware platforms to avoid a name
9494 *Guenter <lists@gknw.net>*
9496 * Fix the server certificate chain building code to use X509_verify_cert(),
9497 it used to have an ad-hoc builder which was unable to cope with anything
9498 other than a simple chain.
9500 *David Woodhouse <dwmw2@infradead.org>, Steve Henson*
9502 * Don't check self signed certificate signatures in X509_verify_cert()
9503 by default (a flag can override this): it just wastes time without
9504 adding any security. As a useful side effect self signed root CAs
9505 with non-FIPS digests are now usable in FIPS mode.
9509 * In dtls1_process_out_of_seq_message() the check if the current message
9510 is already buffered was missing. For every new message was memory
9511 allocated, allowing an attacker to perform an denial of service attack
9512 with sending out of seq handshake messages until there is no memory
9513 left. Additionally every future message was buffered, even if the
9514 sequence number made no sense and would be part of another handshake.
9515 So only messages with sequence numbers less than 10 in advance will be
9516 buffered. ([CVE-2009-1378])
9518 *Robin Seggelmann, discovered by Daniel Mentz*
9520 * Records are buffered if they arrive with a future epoch to be
9521 processed after finishing the corresponding handshake. There is
9522 currently no limitation to this buffer allowing an attacker to perform
9523 a DOS attack with sending records with future epochs until there is no
9524 memory left. This patch adds the pqueue_size() function to determine
9525 the size of a buffer and limits the record buffer to 100 entries.
9528 *Robin Seggelmann, discovered by Daniel Mentz*
9530 * Keep a copy of frag->msg_header.frag_len so it can be used after the
9531 parent structure is freed. ([CVE-2009-1379])
9535 * Handle non-blocking I/O properly in SSL_shutdown() call.
9537 *Darryl Miles <darryl-mailinglists@netbauds.net>*
9539 * Add `2.5.4.*` OIDs
9541 *Ilya O. <vrghost@gmail.com>*
9543 ### Changes between 0.9.8k and 0.9.8l [5 Nov 2009]
9545 * Disable renegotiation completely - this fixes a severe security
9546 problem ([CVE-2009-3555]) at the cost of breaking all
9547 renegotiation. Renegotiation can be re-enabled by setting
9548 SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION in s3->flags at
9549 run-time. This is really not recommended unless you know what
9554 ### Changes between 0.9.8j and 0.9.8k [25 Mar 2009]
9556 * Don't set val to NULL when freeing up structures, it is freed up by
9557 underlying code. If `sizeof(void *) > sizeof(long)` this can result in
9558 zeroing past the valid field. ([CVE-2009-0789])
9560 *Paolo Ganci <Paolo.Ganci@AdNovum.CH>*
9562 * Fix bug where return value of CMS_SignerInfo_verify_content() was not
9563 checked correctly. This would allow some invalid signed attributes to
9564 appear to verify correctly. ([CVE-2009-0591])
9566 *Ivan Nestlerode <inestlerode@us.ibm.com>*
9568 * Reject UniversalString and BMPString types with invalid lengths. This
9569 prevents a crash in ASN1_STRING_print_ex() which assumes the strings have
9570 a legal length. ([CVE-2009-0590])
9574 * Set S/MIME signing as the default purpose rather than setting it
9575 unconditionally. This allows applications to override it at the store
9580 * Permit restricted recursion of ASN1 strings. This is needed in practice
9581 to handle some structures.
9585 * Improve efficiency of mem_gets: don't search whole buffer each time
9588 *Jeremy Shapiro <jnshapir@us.ibm.com>*
9590 * New -hex option for openssl rand.
9594 * Print out UTF8String and NumericString when parsing ASN1.
9598 * Support NumericString type for name components.
9602 * Allow CC in the environment to override the automatically chosen
9603 compiler. Note that nothing is done to ensure flags work with the
9608 ### Changes between 0.9.8i and 0.9.8j [07 Jan 2009]
9610 * Properly check EVP_VerifyFinal() and similar return values
9613 *Ben Laurie, Bodo Moeller, Google Security Team*
9615 * Enable TLS extensions by default.
9619 * Allow the CHIL engine to be loaded, whether the application is
9620 multithreaded or not. (This does not release the developer from the
9621 obligation to set up the dynamic locking callbacks.)
9623 *Sander Temme <sander@temme.net>*
9625 * Use correct exit code if there is an error in dgst command.
9627 *Steve Henson; problem pointed out by Roland Dirlewanger*
9629 * Tweak Configure so that you need to say "experimental-jpake" to enable
9630 JPAKE, and need to use -DOPENSSL_EXPERIMENTAL_JPAKE in applications.
9634 * Add experimental JPAKE support, including demo authentication in
9635 s_client and s_server.
9639 * Set the comparison function in v3_addr_canonize().
9641 *Rob Austein <sra@hactrn.net>*
9643 * Add support for XMPP STARTTLS in s_client.
9645 *Philip Paeps <philip@freebsd.org>*
9647 * Change the server-side SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG behavior
9648 to ensure that even with this option, only ciphersuites in the
9649 server's preference list will be accepted. (Note that the option
9650 applies only when resuming a session, so the earlier behavior was
9651 just about the algorithm choice for symmetric cryptography.)
9655 ### Changes between 0.9.8h and 0.9.8i [15 Sep 2008]
9657 * Fix NULL pointer dereference if a DTLS server received
9658 ChangeCipherSpec as first record ([CVE-2009-1386]).
9662 * Fix a state transition in s3_srvr.c and d1_srvr.c
9663 (was using SSL3_ST_CW_CLNT_HELLO_B, should be `..._ST_SW_SRVR_...`).
9667 * The fix in 0.9.8c that supposedly got rid of unsafe
9668 double-checked locking was incomplete for RSA blinding,
9669 addressing just one layer of what turns out to have been
9670 doubly unsafe triple-checked locking.
9672 So now fix this for real by retiring the MONT_HELPER macro
9673 in crypto/rsa/rsa_eay.c.
9675 *Bodo Moeller; problem pointed out by Marius Schilder*
9677 * Various precautionary measures:
9679 - Avoid size_t integer overflow in HASH_UPDATE (md32_common.h).
9681 - Avoid a buffer overflow in d2i_SSL_SESSION() (ssl_asn1.c).
9682 (NB: This would require knowledge of the secret session ticket key
9683 to exploit, in which case you'd be SOL either way.)
9685 - Change bn_nist.c so that it will properly handle input BIGNUMs
9686 outside the expected range.
9688 - Enforce the 'num' check in BN_div() (bn_div.c) for non-BN_DEBUG
9691 *Neel Mehta, Bodo Moeller*
9693 * Allow engines to be "soft loaded" - i.e. optionally don't die if
9694 the load fails. Useful for distros.
9696 *Ben Laurie and the FreeBSD team*
9698 * Add support for Local Machine Keyset attribute in PKCS#12 files.
9702 * Fix BN_GF2m_mod_arr() top-bit cleanup code.
9706 * Expand ENGINE to support engine supplied SSL client certificate functions.
9708 This work was sponsored by Logica.
9712 * Add CryptoAPI ENGINE to support use of RSA and DSA keys held in Windows
9713 keystores. Support for SSL/TLS client authentication too.
9714 Not compiled unless enable-capieng specified to Configure.
9716 This work was sponsored by Logica.
9720 * Fix bug in X509_ATTRIBUTE creation: don't set attribute using
9721 ASN1_TYPE_set1 if MBSTRING flag set. This bug would crash certain
9722 attribute creation routines such as certificate requests and PKCS#12
9727 ### Changes between 0.9.8g and 0.9.8h [28 May 2008]
9729 * Fix flaw if 'Server Key exchange message' is omitted from a TLS
9730 handshake which could lead to a client crash as found using the
9731 Codenomicon TLS test suite ([CVE-2008-1672])
9733 *Steve Henson, Mark Cox*
9735 * Fix double free in TLS server name extensions which could lead to
9736 a remote crash found by Codenomicon TLS test suite ([CVE-2008-0891])
9740 * Clear error queue in SSL_CTX_use_certificate_chain_file()
9742 Clear the error queue to ensure that error entries left from
9743 older function calls do not interfere with the correct operation.
9745 *Lutz Jaenicke, Erik de Castro Lopo*
9747 * Remove root CA certificates of commercial CAs:
9749 The OpenSSL project does not recommend any specific CA and does not
9750 have any policy with respect to including or excluding any CA.
9751 Therefore it does not make any sense to ship an arbitrary selection
9752 of root CA certificates with the OpenSSL software.
9756 * RSA OAEP patches to fix two separate invalid memory reads.
9757 The first one involves inputs when 'lzero' is greater than
9758 'SHA_DIGEST_LENGTH' (it would read about SHA_DIGEST_LENGTH bytes
9759 before the beginning of from). The second one involves inputs where
9760 the 'db' section contains nothing but zeroes (there is a one-byte
9761 invalid read after the end of 'db').
9763 *Ivan Nestlerode <inestlerode@us.ibm.com>*
9765 * Partial backport from 0.9.9-dev:
9767 Introduce bn_mul_mont (dedicated Montgomery multiplication
9768 procedure) as a candidate for BIGNUM assembler implementation.
9769 While 0.9.9-dev uses assembler for various architectures, only
9770 x86_64 is available by default here in the 0.9.8 branch, and
9771 32-bit x86 is available through a compile-time setting.
9773 To try the 32-bit x86 assembler implementation, use Configure
9774 option "enable-montasm" (which exists only for this backport).
9776 As "enable-montasm" for 32-bit x86 disclaims code stability
9777 anyway, in this constellation we activate additional code
9778 backported from 0.9.9-dev for further performance improvements,
9779 namely BN_from_montgomery_word. (To enable this otherwise,
9780 e.g. x86_64, try `-DMONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD`.)
9782 *Andy Polyakov (backport partially by Bodo Moeller)*
9784 * Add TLS session ticket callback. This allows an application to set
9785 TLS ticket cipher and HMAC keys rather than relying on hardcoded fixed
9786 values. This is useful for key rollover for example where several key
9787 sets may exist with different names.
9791 * Reverse ENGINE-internal logic for caching default ENGINE handles.
9792 This was broken until now in 0.9.8 releases, such that the only way
9793 a registered ENGINE could be used (assuming it initialises
9794 successfully on the host) was to explicitly set it as the default
9795 for the relevant algorithms. This is in contradiction with 0.9.7
9796 behaviour and the documentation. With this fix, when an ENGINE is
9797 registered into a given algorithm's table of implementations, the
9798 'uptodate' flag is reset so that auto-discovery will be used next
9799 time a new context for that algorithm attempts to select an
9802 *Ian Lister (tweaked by Geoff Thorpe)*
9804 * Backport of CMS code to OpenSSL 0.9.8. This differs from the 0.9.9
9805 implementation in the following ways:
9807 Lack of EVP_PKEY_ASN1_METHOD means algorithm parameters have to be
9810 Lack of BER streaming support means one pass streaming processing is
9811 only supported if data is detached: setting the streaming flag is
9812 ignored for embedded content.
9814 CMS support is disabled by default and must be explicitly enabled
9815 with the enable-cms configuration option.
9819 * Update the GMP engine glue to do direct copies between BIGNUM and
9820 mpz_t when openssl and GMP use the same limb size. Otherwise the
9821 existing "conversion via a text string export" trick is still used.
9823 *Paul Sheer <paulsheer@gmail.com>*
9825 * Zlib compression BIO. This is a filter BIO which compressed and
9826 uncompresses any data passed through it.
9830 * Add AES_wrap_key() and AES_unwrap_key() functions to implement
9831 RFC3394 compatible AES key wrapping.
9835 * Add utility functions to handle ASN1 structures. ASN1_STRING_set0():
9836 sets string data without copying. X509_ALGOR_set0() and
9837 X509_ALGOR_get0(): set and retrieve X509_ALGOR (AlgorithmIdentifier)
9838 data. Attribute function X509at_get0_data_by_OBJ(): retrieves data
9839 from an X509_ATTRIBUTE structure optionally checking it occurs only
9840 once. ASN1_TYPE_set1(): set and ASN1_TYPE structure copying supplied
9845 * Fix BN flag handling in RSA_eay_mod_exp() and BN_MONT_CTX_set()
9846 to get the expected BN_FLG_CONSTTIME behavior.
9848 *Bodo Moeller (Google)*
9852 - fixed wrong usage of ioctlsocket() when build for LIBC BSD sockets
9853 - fixed do_tests.pl to run the test suite with CLIB builds too (CLIB_OPT)
9854 - added some more tests to do_tests.pl
9855 - fixed RunningProcess usage so that it works with newer LIBC NDKs too
9856 - removed usage of BN_LLONG for CLIB builds to avoid runtime dependency
9857 - added new Configure targets netware-clib-bsdsock, netware-clib-gcc,
9858 netware-clib-bsdsock-gcc, netware-libc-bsdsock-gcc
9859 - various changes to netware.pl to enable gcc-cross builds on Win32
9861 - changed crypto/bio/b_sock.c to work with macro functions (CLIB BSD)
9862 - various changes to fix missing prototype warnings
9863 - fixed x86nasm.pl to create correct asm files for NASM COFF output
9864 - added AES, WHIRLPOOL and CPUID assembler code to build files
9865 - added missing AES assembler make rules to mk1mf.pl
9866 - fixed order of includes in `apps/ocsp.c` so that `e_os.h` settings apply
9868 *Guenter Knauf <eflash@gmx.net>*
9870 * Implement certificate status request TLS extension defined in RFC3546.
9871 A client can set the appropriate parameters and receive the encoded
9872 OCSP response via a callback. A server can query the supplied parameters
9873 and set the encoded OCSP response in the callback. Add simplified examples
9874 to s_client and s_server.
9878 ### Changes between 0.9.8f and 0.9.8g [19 Oct 2007]
9881 + Binary incompatibility of ssl_ctx_st structure
9882 + DTLS interoperation with non-compliant servers
9883 + Don't call get_session_cb() without proposed session
9884 + Fix ia64 assembler code
9886 *Andy Polyakov, Steve Henson*
9888 ### Changes between 0.9.8e and 0.9.8f [11 Oct 2007]
9890 * DTLS Handshake overhaul. There were longstanding issues with
9891 OpenSSL DTLS implementation, which were making it impossible for
9892 RFC 4347 compliant client to communicate with OpenSSL server.
9893 Unfortunately just fixing these incompatibilities would "cut off"
9894 pre-0.9.8f clients. To allow for hassle free upgrade post-0.9.8e
9895 server keeps tolerating non RFC compliant syntax. The opposite is
9896 not true, 0.9.8f client can not communicate with earlier server.
9897 This update even addresses CVE-2007-4995.
9901 * Changes to avoid need for function casts in OpenSSL: some compilers
9902 (gcc 4.2 and later) reject their use.
9903 *Kurt Roeckx <kurt@roeckx.be>, Peter Hartley <pdh@utter.chaos.org.uk>,
9906 * Add RFC4507 support to OpenSSL. This includes the corrections in
9907 RFC4507bis. The encrypted ticket format is an encrypted encoded
9908 SSL_SESSION structure, that way new session features are automatically
9911 If a client application caches session in an SSL_SESSION structure
9912 support is transparent because tickets are now stored in the encoded
9915 The SSL_CTX structure automatically generates keys for ticket
9916 protection in servers so again support should be possible
9917 with no application modification.
9919 If a client or server wishes to disable RFC4507 support then the option
9920 SSL_OP_NO_TICKET can be set.
9922 Add a TLS extension debugging callback to allow the contents of any client
9923 or server extensions to be examined.
9925 This work was sponsored by Google.
9929 * Add initial support for TLS extensions, specifically for the server_name
9930 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
9931 have new members for a host name. The SSL data structure has an
9932 additional member `SSL_CTX *initial_ctx` so that new sessions can be
9933 stored in that context to allow for session resumption, even after the
9934 SSL has been switched to a new SSL_CTX in reaction to a client's
9935 server_name extension.
9937 New functions (subject to change):
9939 SSL_get_servername()
9940 SSL_get_servername_type()
9943 New CTRL codes and macros (subject to change):
9945 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
9946 - SSL_CTX_set_tlsext_servername_callback()
9947 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
9948 - SSL_CTX_set_tlsext_servername_arg()
9949 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
9951 openssl s_client has a new '-servername ...' option.
9953 openssl s_server has new options '-servername_host ...', '-cert2 ...',
9954 '-key2 ...', '-servername_fatal' (subject to change). This allows
9955 testing the HostName extension for a specific single host name ('-cert'
9956 and '-key' remain fallbacks for handshakes without HostName
9957 negotiation). If the unrecognized_name alert has to be sent, this by
9958 default is a warning; it becomes fatal with the '-servername_fatal'
9961 *Peter Sylvester, Remy Allais, Christophe Renou, Steve Henson*
9963 * Add AES and SSE2 assembly language support to VC++ build.
9967 * Mitigate attack on final subtraction in Montgomery reduction.
9971 * Fix crypto/ec/ec_mult.c to work properly with scalars of value 0
9972 (which previously caused an internal error).
9976 * Squeeze another 10% out of IGE mode when in != out.
9980 * AES IGE mode speedup.
9982 *Dean Gaudet (Google)*
9984 * Add the Korean symmetric 128-bit cipher SEED (see
9985 <http://www.kisa.or.kr/kisa/seed/jsp/seed_eng.jsp>) and
9986 add SEED ciphersuites from RFC 4162:
9988 TLS_RSA_WITH_SEED_CBC_SHA = "SEED-SHA"
9989 TLS_DHE_DSS_WITH_SEED_CBC_SHA = "DHE-DSS-SEED-SHA"
9990 TLS_DHE_RSA_WITH_SEED_CBC_SHA = "DHE-RSA-SEED-SHA"
9991 TLS_DH_anon_WITH_SEED_CBC_SHA = "ADH-SEED-SHA"
9993 To minimize changes between patchlevels in the OpenSSL 0.9.8
9994 series, SEED remains excluded from compilation unless OpenSSL
9995 is configured with 'enable-seed'.
9997 *KISA, Bodo Moeller*
9999 * Mitigate branch prediction attacks, which can be practical if a
10000 single processor is shared, allowing a spy process to extract
10001 information. For detailed background information, see
10002 <http://eprint.iacr.org/2007/039> (O. Aciicmez, S. Gueron,
10003 J.-P. Seifert, "New Branch Prediction Vulnerabilities in OpenSSL
10004 and Necessary Software Countermeasures"). The core of the change
10005 are new versions BN_div_no_branch() and
10006 BN_mod_inverse_no_branch() of BN_div() and BN_mod_inverse(),
10007 respectively, which are slower, but avoid the security-relevant
10008 conditional branches. These are automatically called by BN_div()
10009 and BN_mod_inverse() if the flag BN_FLG_CONSTTIME is set for one
10010 of the input BIGNUMs. Also, BN_is_bit_set() has been changed to
10011 remove a conditional branch.
10013 BN_FLG_CONSTTIME is the new name for the previous
10014 BN_FLG_EXP_CONSTTIME flag, since it now affects more than just
10015 modular exponentiation. (Since OpenSSL 0.9.7h, setting this flag
10016 in the exponent causes BN_mod_exp_mont() to use the alternative
10017 implementation in BN_mod_exp_mont_consttime().) The old name
10018 remains as a deprecated alias.
10020 Similarly, RSA_FLAG_NO_EXP_CONSTTIME is replaced by a more general
10021 RSA_FLAG_NO_CONSTTIME flag since the RSA implementation now uses
10022 constant-time implementations for more than just exponentiation.
10023 Here too the old name is kept as a deprecated alias.
10025 BN_BLINDING_new() will now use BN_dup() for the modulus so that
10026 the BN_BLINDING structure gets an independent copy of the
10027 modulus. This means that the previous `BIGNUM *m` argument to
10028 BN_BLINDING_new() and to BN_BLINDING_create_param() now
10029 essentially becomes `const BIGNUM *m`, although we can't actually
10030 change this in the header file before 0.9.9. It allows
10031 RSA_setup_blinding() to use BN_with_flags() on the modulus to
10032 enable BN_FLG_CONSTTIME.
10034 *Matthew D Wood (Intel Corp)*
10036 * In the SSL/TLS server implementation, be strict about session ID
10037 context matching (which matters if an application uses a single
10038 external cache for different purposes). Previously,
10039 out-of-context reuse was forbidden only if SSL_VERIFY_PEER was
10040 set. This did ensure strict client verification, but meant that,
10041 with applications using a single external cache for quite
10042 different requirements, clients could circumvent ciphersuite
10043 restrictions for a given session ID context by starting a session
10044 in a different context.
10048 * Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
10049 a ciphersuite string such as "DEFAULT:RSA" cannot enable
10050 authentication-only ciphersuites.
10054 * Update the SSL_get_shared_ciphers() fix CVE-2006-3738 which was
10055 not complete and could lead to a possible single byte overflow
10056 ([CVE-2007-5135]) [Ben Laurie]
10058 ### Changes between 0.9.8d and 0.9.8e [23 Feb 2007]
10060 * Since AES128 and AES256 (and similarly Camellia128 and
10061 Camellia256) share a single mask bit in the logic of
10062 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
10063 kludge to work properly if AES128 is available and AES256 isn't
10064 (or if Camellia128 is available and Camellia256 isn't).
10068 * Fix the BIT STRING encoding generated by crypto/ec/ec_asn1.c
10069 (within i2d_ECPrivateKey, i2d_ECPKParameters, i2d_ECParameters):
10070 When a point or a seed is encoded in a BIT STRING, we need to
10071 prevent the removal of trailing zero bits to get the proper DER
10072 encoding. (By default, crypto/asn1/a_bitstr.c assumes the case
10073 of a NamedBitList, for which trailing 0 bits need to be removed.)
10077 * Have SSL/TLS server implementation tolerate "mismatched" record
10078 protocol version while receiving ClientHello even if the
10079 ClientHello is fragmented. (The server can't insist on the
10080 particular protocol version it has chosen before the ServerHello
10081 message has informed the client about his choice.)
10085 * Add RFC 3779 support.
10087 *Rob Austein for ARIN, Ben Laurie*
10089 * Load error codes if they are not already present instead of using a
10090 static variable. This allows them to be cleanly unloaded and reloaded.
10091 Improve header file function name parsing.
10095 * extend SMTP and IMAP protocol emulation in s_client to use EHLO
10096 or CAPABILITY handshake as required by RFCs.
10098 *Goetz Babin-Ebell*
10100 ### Changes between 0.9.8c and 0.9.8d [28 Sep 2006]
10102 * Introduce limits to prevent malicious keys being able to
10103 cause a denial of service. ([CVE-2006-2940])
10105 *Steve Henson, Bodo Moeller*
10107 * Fix ASN.1 parsing of certain invalid structures that can result
10108 in a denial of service. ([CVE-2006-2937]) [Steve Henson]
10110 * Fix buffer overflow in SSL_get_shared_ciphers() function.
10111 ([CVE-2006-3738]) [Tavis Ormandy and Will Drewry, Google Security Team]
10113 * Fix SSL client code which could crash if connecting to a
10114 malicious SSLv2 server. ([CVE-2006-4343])
10116 *Tavis Ormandy and Will Drewry, Google Security Team*
10118 * Since 0.9.8b, ciphersuite strings naming explicit ciphersuites
10119 match only those. Before that, "AES256-SHA" would be interpreted
10120 as a pattern and match "AES128-SHA" too (since AES128-SHA got
10121 the same strength classification in 0.9.7h) as we currently only
10122 have a single AES bit in the ciphersuite description bitmap.
10123 That change, however, also applied to ciphersuite strings such as
10124 "RC4-MD5" that intentionally matched multiple ciphersuites --
10125 namely, SSL 2.0 ciphersuites in addition to the more common ones
10126 from SSL 3.0/TLS 1.0.
10128 So we change the selection algorithm again: Naming an explicit
10129 ciphersuite selects this one ciphersuite, and any other similar
10130 ciphersuite (same bitmap) from *other* protocol versions.
10131 Thus, "RC4-MD5" again will properly select both the SSL 2.0
10132 ciphersuite and the SSL 3.0/TLS 1.0 ciphersuite.
10134 Since SSL 2.0 does not have any ciphersuites for which the
10135 128/256 bit distinction would be relevant, this works for now.
10136 The proper fix will be to use different bits for AES128 and
10137 AES256, which would have avoided the problems from the beginning;
10138 however, bits are scarce, so we can only do this in a new release
10139 (not just a patchlevel) when we can change the SSL_CIPHER
10140 definition to split the single 'unsigned long mask' bitmap into
10141 multiple values to extend the available space.
10145 ### Changes between 0.9.8b and 0.9.8c [05 Sep 2006]
10147 * Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
10148 ([CVE-2006-4339]) [Ben Laurie and Google Security Team]
10150 * Add AES IGE and biIGE modes.
10154 * Change the Unix randomness entropy gathering to use poll() when
10155 possible instead of select(), since the latter has some
10156 undesirable limitations.
10158 *Darryl Miles via Richard Levitte and Bodo Moeller*
10160 * Disable "ECCdraft" ciphersuites more thoroughly. Now special
10161 treatment in ssl/ssl_ciph.s makes sure that these ciphersuites
10162 cannot be implicitly activated as part of, e.g., the "AES" alias.
10163 However, please upgrade to OpenSSL 0.9.9[-dev] for
10164 non-experimental use of the ECC ciphersuites to get TLS extension
10165 support, which is required for curve and point format negotiation
10166 to avoid potential handshake problems.
10170 * Disable rogue ciphersuites:
10172 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
10173 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
10174 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
10176 The latter two were purportedly from
10177 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
10180 Also deactivate the remaining ciphersuites from
10181 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
10182 unofficial, and the ID has long expired.
10186 * Fix RSA blinding Heisenbug (problems sometimes occurred on
10187 dual-core machines) and other potential thread-safety issues.
10191 * Add the symmetric cipher Camellia (128-bit, 192-bit, 256-bit key
10192 versions), which is now available for royalty-free use
10193 (see <http://info.isl.ntt.co.jp/crypt/eng/info/chiteki.html>).
10194 Also, add Camellia TLS ciphersuites from RFC 4132.
10196 To minimize changes between patchlevels in the OpenSSL 0.9.8
10197 series, Camellia remains excluded from compilation unless OpenSSL
10198 is configured with 'enable-camellia'.
10202 * Disable the padding bug check when compression is in use. The padding
10203 bug check assumes the first packet is of even length, this is not
10204 necessarily true if compression is enabled and can result in false
10205 positives causing handshake failure. The actual bug test is ancient
10206 code so it is hoped that implementations will either have fixed it by
10207 now or any which still have the bug do not support compression.
10211 ### Changes between 0.9.8a and 0.9.8b [04 May 2006]
10213 * When applying a cipher rule check to see if string match is an explicit
10214 cipher suite and only match that one cipher suite if it is.
10218 * Link in manifests for VC++ if needed.
10220 *Austin Ziegler <halostatue@gmail.com>*
10222 * Update support for ECC-based TLS ciphersuites according to
10223 draft-ietf-tls-ecc-12.txt with proposed changes (but without
10224 TLS extensions, which are supported starting with the 0.9.9
10225 branch, not in the OpenSSL 0.9.8 branch).
10229 * New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support
10230 opaque EVP_CIPHER_CTX handling.
10234 * Fixes and enhancements to zlib compression code. We now only use
10235 "zlib1.dll" and use the default `__cdecl` calling convention on Win32
10236 to conform with the standards mentioned here:
10237 <http://www.zlib.net/DLL_FAQ.txt>
10238 Static zlib linking now works on Windows and the new --with-zlib-include
10239 --with-zlib-lib options to Configure can be used to supply the location
10240 of the headers and library. Gracefully handle case where zlib library
10245 * Several fixes and enhancements to the OID generation code. The old code
10246 sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't
10247 handle numbers larger than ULONG_MAX, truncated printing and had a
10248 non standard OBJ_obj2txt() behaviour.
10252 * Add support for building of engines under engine/ as shared libraries
10253 under VC++ build system.
10257 * Corrected the numerous bugs in the Win32 path splitter in DSO.
10258 Hopefully, we will not see any false combination of paths any more.
10262 ### Changes between 0.9.8 and 0.9.8a [11 Oct 2005]
10264 * Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
10265 (part of SSL_OP_ALL). This option used to disable the
10266 countermeasure against man-in-the-middle protocol-version
10267 rollback in the SSL 2.0 server implementation, which is a bad
10268 idea. ([CVE-2005-2969])
10270 *Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
10271 for Information Security, National Institute of Advanced Industrial
10272 Science and Technology [AIST], Japan)*
10274 * Add two function to clear and return the verify parameter flags.
10278 * Keep cipherlists sorted in the source instead of sorting them at
10279 runtime, thus removing the need for a lock.
10283 * Avoid some small subgroup attacks in Diffie-Hellman.
10285 *Nick Mathewson and Ben Laurie*
10287 * Add functions for well-known primes.
10291 * Extended Windows CE support.
10293 *Satoshi Nakamura and Andy Polyakov*
10295 * Initialize SSL_METHOD structures at compile time instead of during
10296 runtime, thus removing the need for a lock.
10300 * Make PKCS7_decrypt() work even if no certificate is supplied by
10301 attempting to decrypt each encrypted key in turn. Add support to
10306 ### Changes between 0.9.7h and 0.9.8 [05 Jul 2005]
10308 [NB: OpenSSL 0.9.7i and later 0.9.7 patch levels were released after
10311 * Add libcrypto.pc and libssl.pc for those who feel they need them.
10315 * Change CA.sh and CA.pl so they don't bundle the CSR and the private
10316 key into the same file any more.
10320 * Add initial support for Win64, both IA64 and AMD64/x64 flavors.
10324 * Add -utf8 command line and config file option to 'ca'.
10326 *Stefan <stf@udoma.org*
10328 * Removed the macro des_crypt(), as it seems to conflict with some
10329 libraries. Use DES_crypt().
10333 * Correct naming of the 'chil' and '4758cca' ENGINEs. This
10334 involves renaming the source and generated shared-libs for
10335 both. The engines will accept the corrected or legacy ids
10336 ('ncipher' and '4758_cca' respectively) when binding. NB,
10337 this only applies when building 'shared'.
10339 *Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe*
10341 * Add attribute functions to EVP_PKEY structure. Modify
10342 PKCS12_create() to recognize a CSP name attribute and
10343 use it. Make -CSP option work again in pkcs12 utility.
10347 * Add new functionality to the bn blinding code:
10348 - automatic re-creation of the BN_BLINDING parameters after
10349 a fixed number of uses (currently 32)
10350 - add new function for parameter creation
10351 - introduce flags to control the update behaviour of the
10352 BN_BLINDING parameters
10353 - hide BN_BLINDING structure
10354 Add a second BN_BLINDING slot to the RSA structure to improve
10355 performance when a single RSA object is shared among several
10360 * Add support for DTLS.
10362 *Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie*
10364 * Add support for DER encoded private keys (SSL_FILETYPE_ASN1)
10365 to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file()
10369 * Remove buggy and incomplete DH cert support from
10370 ssl/ssl_rsa.c and ssl/s3_both.c
10374 * Use SHA-1 instead of MD5 as the default digest algorithm for
10375 the `apps/openssl` commands.
10379 * Compile clean with "-Wall -Wmissing-prototypes
10380 -Wstrict-prototypes -Wmissing-declarations -Werror". Currently
10381 DEBUG_SAFESTACK must also be set.
10385 * Change ./Configure so that certain algorithms can be disabled by default.
10386 The new counterpiece to "no-xxx" is "enable-xxx".
10388 The patented RC5 and MDC2 algorithms will now be disabled unless
10389 "enable-rc5" and "enable-mdc2", respectively, are specified.
10391 (IDEA remains enabled despite being patented. This is because IDEA
10392 is frequently required for interoperability, and there is no license
10393 fee for non-commercial use. As before, "no-idea" can be used to
10394 avoid this algorithm.)
10398 * Add processing of proxy certificates (see RFC 3820). This work was
10399 sponsored by KTH (The Royal Institute of Technology in Stockholm) and
10400 EGEE (Enabling Grids for E-science in Europe).
10404 * RC4 performance overhaul on modern architectures/implementations, such
10405 as Intel P4, IA-64 and AMD64.
10409 * New utility extract-section.pl. This can be used specify an alternative
10410 section number in a pod file instead of having to treat each file as
10411 a separate case in Makefile. This can be done by adding two lines to the
10414 =for comment openssl_section:XXX
10416 The blank line is mandatory.
10420 * New arguments -certform, -keyform and -pass for s_client and s_server
10421 to allow alternative format key and certificate files and passphrase
10426 * New structure X509_VERIFY_PARAM which combines current verify parameters,
10427 update associated structures and add various utility functions.
10429 Add new policy related verify parameters, include policy checking in
10430 standard verify code. Enhance 'smime' application with extra parameters
10431 to support policy checking and print out.
10435 * Add a new engine to support VIA PadLock ACE extensions in the VIA C3
10436 Nehemiah processors. These extensions support AES encryption in hardware
10437 as well as RNG (though RNG support is currently disabled).
10439 *Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov*
10441 * Deprecate `BN_[get|set]_params()` functions (they were ignored internally).
10445 * New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented.
10447 *Andy Polyakov and a number of other people*
10449 * Improved PowerPC platform support. Most notably BIGNUM assembler
10450 implementation contributed by IBM.
10452 *Suresh Chari, Peter Waltenberg, Andy Polyakov*
10454 * The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public
10455 exponent rather than 'unsigned long'. There is a corresponding change to
10456 the new 'rsa_keygen' element of the RSA_METHOD structure.
10458 *Jelte Jansen, Geoff Thorpe*
10460 * Functionality for creating the initial serial number file is now
10461 moved from CA.pl to the 'ca' utility with a new option -create_serial.
10463 (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial
10464 number file to 1, which is bound to cause problems. To avoid
10465 the problems while respecting compatibility between different 0.9.7
10466 patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in
10467 CA.pl for serial number initialization. With the new release 0.9.8,
10468 we can fix the problem directly in the 'ca' utility.)
10472 * Reduced header interdependencies by declaring more opaque objects in
10473 ossl_typ.h. As a consequence, including some headers (eg. engine.h) will
10474 give fewer recursive includes, which could break lazy source code - so
10475 this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always,
10476 developers should define this symbol when building and using openssl to
10477 ensure they track the recommended behaviour, interfaces, [etc], but
10478 backwards-compatible behaviour prevails when this isn't defined.
10482 * New function X509_POLICY_NODE_print() which prints out policy nodes.
10486 * Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality.
10487 This will generate a random key of the appropriate length based on the
10488 cipher context. The EVP_CIPHER can provide its own random key generation
10489 routine to support keys of a specific form. This is used in the des and
10490 3des routines to generate a key of the correct parity. Update S/MIME
10491 code to use new functions and hence generate correct parity DES keys.
10492 Add EVP_CHECK_DES_KEY #define to return an error if the key is not
10493 valid (weak or incorrect parity).
10497 * Add a local set of CRLs that can be used by X509_verify_cert() as well
10498 as looking them up. This is useful when the verified structure may contain
10499 CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs
10500 present unless the new PKCS7_NO_CRL flag is asserted.
10504 * Extend ASN1 oid configuration module. It now additionally accepts the
10507 shortName = some long name, 1.2.3.4
10511 * Reimplemented the BN_CTX implementation. There is now no more static
10512 limitation on the number of variables it can handle nor the depth of the
10513 "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack
10514 information can now expand as required, and rather than having a single
10515 static array of bignums, BN_CTX now uses a linked-list of such arrays
10516 allowing it to expand on demand whilst maintaining the usefulness of
10517 BN_CTX's "bundling".
10521 * Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD
10522 to allow all RSA operations to function using a single BN_CTX.
10526 * Preliminary support for certificate policy evaluation and checking. This
10527 is initially intended to pass the tests outlined in "Conformance Testing
10528 of Relying Party Client Certificate Path Processing Logic" v1.07.
10532 * bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and
10533 remained unused and not that useful. A variety of other little bignum
10534 tweaks and fixes have also been made continuing on from the audit (see
10539 * Constify all or almost all d2i, c2i, s2i and r2i functions, along with
10540 associated ASN1, EVP and SSL functions and old ASN1 macros.
10544 * BN_zero() only needs to set 'top' and 'neg' to zero for correct results,
10545 and this should never fail. So the return value from the use of
10546 BN_set_word() (which can fail due to needless expansion) is now deprecated;
10547 if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro.
10551 * BN_CTX_get() should return zero-valued bignums, providing the same
10552 initialised value as BN_new().
10554 *Geoff Thorpe, suggested by Ulf Möller*
10556 * Support for inhibitAnyPolicy certificate extension.
10560 * An audit of the BIGNUM code is underway, for which debugging code is
10561 enabled when BN_DEBUG is defined. This makes stricter enforcements on what
10562 is considered valid when processing BIGNUMs, and causes execution to
10563 assert() when a problem is discovered. If BN_DEBUG_RAND is defined,
10564 further steps are taken to deliberately pollute unused data in BIGNUM
10565 structures to try and expose faulty code further on. For now, openssl will
10566 (in its default mode of operation) continue to tolerate the inconsistent
10567 forms that it has tolerated in the past, but authors and packagers should
10568 consider trying openssl and their own applications when compiled with
10569 these debugging symbols defined. It will help highlight potential bugs in
10570 their own code, and will improve the test coverage for OpenSSL itself. At
10571 some point, these tighter rules will become openssl's default to improve
10572 maintainability, though the assert()s and other overheads will remain only
10573 in debugging configurations. See bn.h for more details.
10575 *Geoff Thorpe, Nils Larsch, Ulf Möller*
10577 * BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure
10578 that can only be obtained through BN_CTX_new() (which implicitly
10579 initialises it). The presence of this function only made it possible
10580 to overwrite an existing structure (and cause memory leaks).
10584 * Because of the callback-based approach for implementing LHASH as a
10585 template type, lh_insert() adds opaque objects to hash-tables and
10586 lh_doall() or lh_doall_arg() are typically used with a destructor callback
10587 to clean up those corresponding objects before destroying the hash table
10588 (and losing the object pointers). So some over-zealous constifications in
10589 LHASH have been relaxed so that lh_insert() does not take (nor store) the
10590 objects as "const" and the `lh_doall[_arg]` callback wrappers are not
10591 prototyped to have "const" restrictions on the object pointers they are
10592 given (and so aren't required to cast them away any more).
10596 * The tmdiff.h API was so ugly and minimal that our own timing utility
10597 (speed) prefers to use its own implementation. The two implementations
10598 haven't been consolidated as yet (volunteers?) but the tmdiff API has had
10599 its object type properly exposed (MS_TM) instead of casting to/from
10600 `char *`. This may still change yet if someone realises MS_TM and
10602 aren't necessarily the greatest nomenclatures - but this is what was used
10603 internally to the implementation so I've used that for now.
10607 * Ensure that deprecated functions do not get compiled when
10608 OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of
10609 the self-tests were still using deprecated key-generation functions so
10610 these have been updated also.
10614 * Reorganise PKCS#7 code to separate the digest location functionality
10615 into PKCS7_find_digest(), digest addition into PKCS7_bio_add_digest().
10616 New function PKCS7_set_digest() to set the digest type for PKCS#7
10617 digestedData type. Add additional code to correctly generate the
10618 digestedData type and add support for this type in PKCS7 initialization
10623 * New function PKCS7_set0_type_other() this initializes a PKCS7
10624 structure of type "other".
10628 * Fix prime generation loop in crypto/bn/bn_prime.pl by making
10629 sure the loop does correctly stop and breaking ("division by zero")
10630 modulus operations are not performed. The (pre-generated) prime
10631 table crypto/bn/bn_prime.h was already correct, but it could not be
10632 re-generated on some platforms because of the "division by zero"
10633 situation in the script.
10635 *Ralf S. Engelschall*
10637 * Update support for ECC-based TLS ciphersuites according to
10638 draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with
10639 SHA-1 now is only used for "small" curves (where the
10640 representation of a field element takes up to 24 bytes); for
10641 larger curves, the field element resulting from ECDH is directly
10642 used as premaster secret.
10644 *Douglas Stebila (Sun Microsystems Laboratories)*
10646 * Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2
10647 curve secp160r1 to the tests.
10649 *Douglas Stebila (Sun Microsystems Laboratories)*
10651 * Add the possibility to load symbols globally with DSO.
10653 *Götz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte*
10655 * Add the functions ERR_set_mark() and ERR_pop_to_mark() for better
10656 control of the error stack.
10660 * Add support for STORE in ENGINE.
10664 * Add the STORE type. The intention is to provide a common interface
10665 to certificate and key stores, be they simple file-based stores, or
10666 HSM-type store, or LDAP stores, or...
10667 NOTE: The code is currently UNTESTED and isn't really used anywhere.
10671 * Add a generic structure called OPENSSL_ITEM. This can be used to
10672 pass a list of arguments to any function as well as provide a way
10673 for a function to pass data back to the caller.
10677 * Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup()
10678 works like BUF_strdup() but can be used to duplicate a portion of
10679 a string. The copy gets NUL-terminated. BUF_memdup() duplicates
10684 * Add the function sk_find_ex() which works like sk_find(), but will
10685 return an index to an element even if an exact match couldn't be
10686 found. The index is guaranteed to point at the element where the
10687 searched-for key would be inserted to preserve sorting order.
10691 * Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but
10692 takes an extra flags argument for optional functionality. Currently,
10693 the following flags are defined:
10695 OBJ_BSEARCH_VALUE_ON_NOMATCH
10696 This one gets OBJ_bsearch_ex() to return a pointer to the first
10697 element where the comparing function returns a negative or zero
10700 OBJ_BSEARCH_FIRST_VALUE_ON_MATCH
10701 This one gets OBJ_bsearch_ex() to return a pointer to the first
10702 element where the comparing function returns zero. This is useful
10703 if there are more than one element where the comparing function
10708 * Make it possible to create self-signed certificates with 'openssl ca'
10709 in such a way that the self-signed certificate becomes part of the
10710 CA database and uses the same mechanisms for serial number generation
10711 as all other certificate signing. The new flag '-selfsign' enables
10712 this functionality. Adapt CA.sh and CA.pl.in.
10716 * Add functionality to check the public key of a certificate request
10717 against a given private. This is useful to check that a certificate
10718 request can be signed by that key (self-signing).
10722 * Make it possible to have multiple active certificates with the same
10723 subject in the CA index file. This is done only if the keyword
10724 'unique_subject' is set to 'no' in the main CA section (default
10725 if 'CA_default') of the configuration file. The value is saved
10726 with the database itself in a separate index attribute file,
10727 named like the index file with '.attr' appended to the name.
10731 * Generate multi-valued AVAs using '+' notation in config files for
10736 * Support for nameConstraints certificate extension.
10740 * Support for policyConstraints certificate extension.
10744 * Support for policyMappings certificate extension.
10748 * Make sure the default DSA_METHOD implementation only uses its
10749 dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL,
10750 and change its own handlers to be NULL so as to remove unnecessary
10751 indirection. This lets alternative implementations fallback to the
10752 default implementation more easily.
10756 * Support for directoryName in GeneralName related extensions
10761 * Make it possible to link applications using Makefile.shared.
10762 Make that possible even when linking against static libraries!
10766 * Support for single pass processing for S/MIME signing. This now
10767 means that S/MIME signing can be done from a pipe, in addition
10768 cleartext signing (multipart/signed type) is effectively streaming
10769 and the signed data does not need to be all held in memory.
10771 This is done with a new flag PKCS7_STREAM. When this flag is set
10772 PKCS7_sign() only initializes the PKCS7 structure and the actual signing
10773 is done after the data is output (and digests calculated) in
10774 SMIME_write_PKCS7().
10778 * Add full support for -rpath/-R, both in shared libraries and
10779 applications, at least on the platforms where it's known how
10784 * In crypto/ec/ec_mult.c, implement fast point multiplication with
10785 precomputation, based on wNAF splitting: EC_GROUP_precompute_mult()
10786 will now compute a table of multiples of the generator that
10787 makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul()
10788 faster (notably in the case of a single point multiplication,
10789 scalar * generator).
10791 *Nils Larsch, Bodo Moeller*
10793 * IPv6 support for certificate extensions. The various extensions
10794 which use the IP:a.b.c.d can now take IPv6 addresses using the
10795 formats of RFC1884 2.2 . IPv6 addresses are now also displayed
10800 * Added an ENGINE that implements RSA by performing private key
10801 exponentiations with the GMP library. The conversions to and from
10802 GMP's mpz_t format aren't optimised nor are any montgomery forms
10803 cached, and on x86 it appears OpenSSL's own performance has caught up.
10804 However there are likely to be other architectures where GMP could
10805 provide a boost. This ENGINE is not built in by default, but it can be
10806 specified at Configure time and should be accompanied by the necessary
10807 linker additions, eg;
10808 ./config -DOPENSSL_USE_GMP -lgmp
10812 * "openssl engine" will not display ENGINE/DSO load failure errors when
10813 testing availability of engines with "-t" - the old behaviour is
10814 produced by increasing the feature's verbosity with "-tt".
10818 * ECDSA routines: under certain error conditions uninitialized BN objects
10819 could be freed. Solution: make sure initialization is performed early
10820 enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de>
10825 * Key-generation can now be implemented in RSA_METHOD, DSA_METHOD
10826 and DH_METHOD (eg. by ENGINE implementations) to override the normal
10827 software implementations. For DSA and DH, parameter generation can
10828 also be overridden by providing the appropriate method callbacks.
10832 * Change the "progress" mechanism used in key-generation and
10833 primality testing to functions that take a new BN_GENCB pointer in
10834 place of callback/argument pairs. The new API functions have `_ex`
10835 postfixes and the older functions are reimplemented as wrappers for
10836 the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide
10837 declarations of the old functions to help (graceful) attempts to
10838 migrate to the new functions. Also, the new key-generation API
10839 functions operate on a caller-supplied key-structure and return
10840 success/failure rather than returning a key or NULL - this is to
10841 help make "keygen" another member function of RSA_METHOD etc.
10843 Example for using the new callback interface:
10845 int (*my_callback)(int a, int b, BN_GENCB *cb) = ...;
10846 void *my_arg = ...;
10849 BN_GENCB_set(&my_cb, my_callback, my_arg);
10851 return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb);
10852 /* For the meaning of a, b in calls to my_callback(), see the
10853 * documentation of the function that calls the callback.
10854 * cb will point to my_cb; my_arg can be retrieved as cb->arg.
10855 * my_callback should return 1 if it wants BN_is_prime_ex()
10856 * to continue, or 0 to stop.
10861 * Change the ZLIB compression method to be stateful, and make it
10862 available to TLS with the number defined in
10863 draft-ietf-tls-compression-04.txt.
10867 * Add the ASN.1 structures and functions for CertificatePair, which
10868 is defined as follows (according to X.509_4thEditionDraftV6.pdf):
10870 CertificatePair ::= SEQUENCE {
10871 forward [0] Certificate OPTIONAL,
10872 reverse [1] Certificate OPTIONAL,
10873 -- at least one of the pair shall be present -- }
10875 Also implement the PEM functions to read and write certificate
10876 pairs, and defined the PEM tag as "CERTIFICATE PAIR".
10878 This needed to be defined, mostly for the sake of the LDAP
10879 attribute crossCertificatePair, but may prove useful elsewhere as
10884 * Make it possible to inhibit symlinking of shared libraries in
10885 Makefile.shared, for Cygwin's sake.
10889 * Extend the BIGNUM API by creating a function
10890 void BN_set_negative(BIGNUM *a, int neg);
10891 and a macro that behave like
10892 int BN_is_negative(const BIGNUM *a);
10894 to avoid the need to access 'a->neg' directly in applications.
10898 * Implement fast modular reduction for pseudo-Mersenne primes
10899 used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c).
10900 EC_GROUP_new_curve_GFp() will now automatically use this
10903 *Nils Larsch <nla@trustcenter.de>*
10905 * Add new lock type (CRYPTO_LOCK_BN).
10909 * Change the ENGINE framework to automatically load engines
10910 dynamically from specific directories unless they could be
10911 found to already be built in or loaded. Move all the
10912 current engines except for the cryptodev one to a new
10913 directory engines/.
10914 The engines in engines/ are built as shared libraries if
10915 the "shared" options was given to ./Configure or ./config.
10916 Otherwise, they are inserted in libcrypto.a.
10917 /usr/local/ssl/engines is the default directory for dynamic
10918 engines, but that can be overridden at configure time through
10919 the usual use of --prefix and/or --openssldir, and at run
10920 time with the environment variable OPENSSL_ENGINES.
10922 *Geoff Thorpe and Richard Levitte*
10924 * Add Makefile.shared, a helper makefile to build shared
10925 libraries. Adapt Makefile.org.
10929 * Add version info to Win32 DLLs.
10931 *Peter 'Luna' Runestig" <peter@runestig.com>*
10933 * Add new 'medium level' PKCS#12 API. Certificates and keys
10934 can be added using this API to created arbitrary PKCS#12
10935 files while avoiding the low-level API.
10937 New options to PKCS12_create(), key or cert can be NULL and
10938 will then be omitted from the output file. The encryption
10939 algorithm NIDs can be set to -1 for no encryption, the mac
10940 iteration count can be set to 0 to omit the mac.
10942 Enhance pkcs12 utility by making the -nokeys and -nocerts
10943 options work when creating a PKCS#12 file. New option -nomac
10944 to omit the mac, NONE can be set for an encryption algorithm.
10945 New code is modified to use the enhanced PKCS12_create()
10946 instead of the low-level API.
10950 * Extend ASN1 encoder to support indefinite length constructed
10951 encoding. This can output sequences tags and octet strings in
10952 this form. Modify pk7_asn1.c to support indefinite length
10953 encoding. This is experimental and needs additional code to
10954 be useful, such as an ASN1 bio and some enhanced streaming
10957 Extend template encode functionality so that tagging is passed
10958 down to the template encoder.
10962 * Let 'openssl req' fail if an argument to '-newkey' is not
10963 recognized instead of using RSA as a default.
10967 * Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt.
10968 As these are not official, they are not included in "ALL";
10969 the "ECCdraft" ciphersuite group alias can be used to select them.
10971 *Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)*
10973 * Add ECDH engine support.
10975 *Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)*
10977 * Add ECDH in new directory crypto/ecdh/.
10979 *Douglas Stebila (Sun Microsystems Laboratories)*
10981 * Let BN_rand_range() abort with an error after 100 iterations
10982 without success (which indicates a broken PRNG).
10986 * Change BN_mod_sqrt() so that it verifies that the input value
10987 is really the square of the return value. (Previously,
10988 BN_mod_sqrt would show GIGO behaviour.)
10992 * Add named elliptic curves over binary fields from X9.62, SECG,
10993 and WAP/WTLS; add OIDs that were still missing.
10995 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10997 * Extend the EC library for elliptic curves over binary fields
10998 (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/).
11001 EC_GF2m_simple_method
11005 EC_GROUP_new_curve_GF2m
11006 EC_GROUP_set_curve_GF2m
11007 EC_GROUP_get_curve_GF2m
11008 EC_POINT_set_affine_coordinates_GF2m
11009 EC_POINT_get_affine_coordinates_GF2m
11010 EC_POINT_set_compressed_coordinates_GF2m
11012 Point compression for binary fields is disabled by default for
11013 patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to
11016 As binary polynomials are represented as BIGNUMs, various members
11017 of the EC_GROUP and EC_POINT data structures can be shared
11018 between the implementations for prime fields and binary fields;
11019 the above `..._GF2m functions` (except for EX_GROUP_new_curve_GF2m)
11020 are essentially identical to their `..._GFp` counterparts.
11021 (For simplicity, the `..._GFp` prefix has been dropped from
11022 various internal method names.)
11024 An internal 'field_div' method (similar to 'field_mul' and
11025 'field_sqr') has been added; this is used only for binary fields.
11027 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
11029 * Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult()
11030 through methods ('mul', 'precompute_mult').
11032 The generic implementations (now internally called 'ec_wNAF_mul'
11033 and 'ec_wNAF_precomputed_mult') remain the default if these
11034 methods are undefined.
11036 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
11038 * New function EC_GROUP_get_degree, which is defined through
11039 EC_METHOD. For curves over prime fields, this returns the bit
11040 length of the modulus.
11042 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
11044 * New functions EC_GROUP_dup, EC_POINT_dup.
11045 (These simply call ..._new and ..._copy).
11047 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
11049 * Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c.
11050 Polynomials are represented as BIGNUMs (where the sign bit is not
11051 used) in the following functions [macros]:
11054 BN_GF2m_sub [= BN_GF2m_add]
11055 BN_GF2m_mod [wrapper for BN_GF2m_mod_arr]
11056 BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr]
11057 BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr]
11059 BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr]
11060 BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr]
11061 BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr]
11062 BN_GF2m_cmp [= BN_ucmp]
11064 (Note that only the 'mod' functions are actually for fields GF(2^m).
11065 BN_GF2m_add() is misnomer, but this is for the sake of consistency.)
11067 For some functions, an the irreducible polynomial defining a
11068 field can be given as an 'unsigned int[]' with strictly
11069 decreasing elements giving the indices of those bits that are set;
11070 i.e., p[] represents the polynomial
11071 f(t) = t^p[0] + t^p[1] + ... + t^p[k]
11073 p[0] > p[1] > ... > p[k] = 0.
11074 This applies to the following functions:
11077 BN_GF2m_mod_mul_arr
11078 BN_GF2m_mod_sqr_arr
11079 BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv]
11080 BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div]
11081 BN_GF2m_mod_exp_arr
11082 BN_GF2m_mod_sqrt_arr
11083 BN_GF2m_mod_solve_quad_arr
11087 Conversion can be performed by the following functions:
11092 bntest.c has additional tests for binary polynomial arithmetic.
11094 Two implementations for BN_GF2m_mod_div() are available.
11095 The default algorithm simply uses BN_GF2m_mod_inv() and
11096 BN_GF2m_mod_mul(). The alternative algorithm is compiled in only
11097 if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the
11098 copyright notice in crypto/bn/bn_gf2m.c before enabling it).
11100 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
11102 * Add new error code 'ERR_R_DISABLED' that can be used when some
11103 functionality is disabled at compile-time.
11105 *Douglas Stebila <douglas.stebila@sun.com>*
11107 * Change default behaviour of 'openssl asn1parse' so that more
11108 information is visible when viewing, e.g., a certificate:
11110 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump'
11111 mode the content of non-printable OCTET STRINGs is output in a
11112 style similar to INTEGERs, but with '[HEX DUMP]' prepended to
11113 avoid the appearance of a printable string.
11115 *Nils Larsch <nla@trustcenter.de>*
11117 * Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access
11119 EC_GROUP_set_asn1_flag()
11120 EC_GROUP_get_asn1_flag()
11121 EC_GROUP_set_point_conversion_form()
11122 EC_GROUP_get_point_conversion_form()
11123 These control ASN1 encoding details:
11124 - Curves (i.e., groups) are encoded explicitly unless asn1_flag
11125 has been set to OPENSSL_EC_NAMED_CURVE.
11126 - Points are encoded in uncompressed form by default; options for
11127 asn1_for are as for point2oct, namely
11128 POINT_CONVERSION_COMPRESSED
11129 POINT_CONVERSION_UNCOMPRESSED
11130 POINT_CONVERSION_HYBRID
11132 Also add 'seed' and 'seed_len' members to EC_GROUP with access
11134 EC_GROUP_set_seed()
11135 EC_GROUP_get0_seed()
11136 EC_GROUP_get_seed_len()
11137 This is used only for ASN1 purposes (so far).
11139 *Nils Larsch <nla@trustcenter.de>*
11141 * Add 'field_type' member to EC_METHOD, which holds the NID
11142 of the appropriate field type OID. The new function
11143 EC_METHOD_get_field_type() returns this value.
11145 *Nils Larsch <nla@trustcenter.de>*
11148 EC_POINT_point2bn()
11149 EC_POINT_bn2point()
11150 EC_POINT_point2hex()
11151 EC_POINT_hex2point()
11152 providing useful interfaces to EC_POINT_point2oct() and
11153 EC_POINT_oct2point().
11155 *Nils Larsch <nla@trustcenter.de>*
11157 * Change internals of the EC library so that the functions
11158 EC_GROUP_set_generator()
11159 EC_GROUP_get_generator()
11160 EC_GROUP_get_order()
11161 EC_GROUP_get_cofactor()
11162 are implemented directly in crypto/ec/ec_lib.c and not dispatched
11163 to methods, which would lead to unnecessary code duplication when
11164 adding different types of curves.
11166 *Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller*
11168 * Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM
11169 arithmetic, and such that modified wNAFs are generated
11170 (which avoid length expansion in many cases).
11174 * Add a function EC_GROUP_check_discriminant() (defined via
11175 EC_METHOD) that verifies that the curve discriminant is non-zero.
11177 Add a function EC_GROUP_check() that makes some sanity tests
11178 on a EC_GROUP, its generator and order. This includes
11179 EC_GROUP_check_discriminant().
11181 *Nils Larsch <nla@trustcenter.de>*
11183 * Add ECDSA in new directory crypto/ecdsa/.
11185 Add applications 'openssl ecparam' and 'openssl ecdsa'
11186 (these are based on 'openssl dsaparam' and 'openssl dsa').
11188 ECDSA support is also included in various other files across the
11189 library. Most notably,
11190 - 'openssl req' now has a '-newkey ecdsa:file' option;
11191 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA;
11192 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and
11193 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make
11194 them suitable for ECDSA where domain parameters must be
11195 extracted before the specific public key;
11196 - ECDSA engine support has been added.
11198 *Nils Larsch <nla@trustcenter.de>*
11200 * Include some named elliptic curves, and add OIDs from X9.62,
11201 SECG, and WAP/WTLS. Each curve can be obtained from the new
11203 EC_GROUP_new_by_curve_name(),
11204 and the list of available named curves can be obtained with
11205 EC_get_builtin_curves().
11206 Also add a 'curve_name' member to EC_GROUP objects, which can be
11208 EC_GROUP_set_curve_name()
11209 EC_GROUP_get_curve_name()
11211 *Nils Larsch <larsch@trustcenter.de, Bodo Moeller*
11213 * Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
11214 was actually never needed) and in BN_mul(). The removal in BN_mul()
11215 required a small change in bn_mul_part_recursive() and the addition
11216 of the functions bn_cmp_part_words(), bn_sub_part_words() and
11217 bn_add_part_words(), which do the same thing as bn_cmp_words(),
11218 bn_sub_words() and bn_add_words() except they take arrays with
11223 ### Changes between 0.9.7l and 0.9.7m [23 Feb 2007]
11225 * Cleanse PEM buffers before freeing them since they may contain
11228 *Benjamin Bennett <ben@psc.edu>*
11230 * Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
11231 a ciphersuite string such as "DEFAULT:RSA" cannot enable
11232 authentication-only ciphersuites.
11236 * Since AES128 and AES256 share a single mask bit in the logic of
11237 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
11238 kludge to work properly if AES128 is available and AES256 isn't.
11242 * Expand security boundary to match 1.1.1 module.
11246 * Remove redundant features: hash file source, editing of test vectors
11247 modify fipsld to use external fips_premain.c signature.
11251 * New perl script mkfipsscr.pl to create shell scripts or batch files to
11252 run algorithm test programs.
11256 * Make algorithm test programs more tolerant of whitespace.
11260 * Have SSL/TLS server implementation tolerate "mismatched" record
11261 protocol version while receiving ClientHello even if the
11262 ClientHello is fragmented. (The server can't insist on the
11263 particular protocol version it has chosen before the ServerHello
11264 message has informed the client about his choice.)
11268 * Load error codes if they are not already present instead of using a
11269 static variable. This allows them to be cleanly unloaded and reloaded.
11273 ### Changes between 0.9.7k and 0.9.7l [28 Sep 2006]
11275 * Introduce limits to prevent malicious keys being able to
11276 cause a denial of service. ([CVE-2006-2940])
11278 *Steve Henson, Bodo Moeller*
11280 * Fix ASN.1 parsing of certain invalid structures that can result
11281 in a denial of service. ([CVE-2006-2937]) [Steve Henson]
11283 * Fix buffer overflow in SSL_get_shared_ciphers() function.
11284 ([CVE-2006-3738]) [Tavis Ormandy and Will Drewry, Google Security Team]
11286 * Fix SSL client code which could crash if connecting to a
11287 malicious SSLv2 server. ([CVE-2006-4343])
11289 *Tavis Ormandy and Will Drewry, Google Security Team*
11291 * Change ciphersuite string processing so that an explicit
11292 ciphersuite selects this one ciphersuite (so that "AES256-SHA"
11293 will no longer include "AES128-SHA"), and any other similar
11294 ciphersuite (same bitmap) from *other* protocol versions (so that
11295 "RC4-MD5" will still include both the SSL 2.0 ciphersuite and the
11296 SSL 3.0/TLS 1.0 ciphersuite). This is a backport combining
11297 changes from 0.9.8b and 0.9.8d.
11301 ### Changes between 0.9.7j and 0.9.7k [05 Sep 2006]
11303 * Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
11304 ([CVE-2006-4339]) [Ben Laurie and Google Security Team]
11306 * Change the Unix randomness entropy gathering to use poll() when
11307 possible instead of select(), since the latter has some
11308 undesirable limitations.
11310 *Darryl Miles via Richard Levitte and Bodo Moeller*
11312 * Disable rogue ciphersuites:
11314 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
11315 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
11316 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
11318 The latter two were purportedly from
11319 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
11322 Also deactivate the remaining ciphersuites from
11323 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
11324 unofficial, and the ID has long expired.
11328 * Fix RSA blinding Heisenbug (problems sometimes occurred on
11329 dual-core machines) and other potential thread-safety issues.
11333 ### Changes between 0.9.7i and 0.9.7j [04 May 2006]
11335 * Adapt fipsld and the build system to link against the validated FIPS
11336 module in FIPS mode.
11340 * Fixes for VC++ 2005 build under Windows.
11344 * Add new Windows build target VC-32-GMAKE for VC++. This uses GNU make
11345 from a Windows bash shell such as MSYS. It is autodetected from the
11346 "config" script when run from a VC++ environment. Modify standard VC++
11347 build to use fipscanister.o from the GNU make build.
11351 ### Changes between 0.9.7h and 0.9.7i [14 Oct 2005]
11353 * Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS.
11354 The value now differs depending on if you build for FIPS or not.
11355 BEWARE! A program linked with a shared FIPSed libcrypto can't be
11356 safely run with a non-FIPSed libcrypto, as it may crash because of
11357 the difference induced by this change.
11361 ### Changes between 0.9.7g and 0.9.7h [11 Oct 2005]
11363 * Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
11364 (part of SSL_OP_ALL). This option used to disable the
11365 countermeasure against man-in-the-middle protocol-version
11366 rollback in the SSL 2.0 server implementation, which is a bad
11367 idea. ([CVE-2005-2969])
11369 *Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
11370 for Information Security, National Institute of Advanced Industrial
11371 Science and Technology [AIST, Japan)]*
11373 * Minimal support for X9.31 signatures and PSS padding modes. This is
11374 mainly for FIPS compliance and not fully integrated at this stage.
11378 * For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform
11379 the exponentiation using a fixed-length exponent. (Otherwise,
11380 the information leaked through timing could expose the secret key
11381 after many signatures; cf. Bleichenbacher's attack on DSA with
11386 * Make a new fixed-window mod_exp implementation the default for
11387 RSA, DSA, and DH private-key operations so that the sequence of
11388 squares and multiplies and the memory access pattern are
11389 independent of the particular secret key. This will mitigate
11390 cache-timing and potential related attacks.
11392 BN_mod_exp_mont_consttime() is the new exponentiation implementation,
11393 and this is automatically used by BN_mod_exp_mont() if the new flag
11394 BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH
11395 will use this BN flag for private exponents unless the flag
11396 RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or
11397 DH_FLAG_NO_EXP_CONSTTIME, respectively, is set.
11399 *Matthew D Wood (Intel Corp), with some changes by Bodo Moeller*
11401 * Change the client implementation for SSLv23_method() and
11402 SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0
11403 Client Hello message format if the SSL_OP_NO_SSLv2 option is set.
11404 (Previously, the SSL 2.0 backwards compatible Client Hello
11405 message format would be used even with SSL_OP_NO_SSLv2.)
11409 * Add support for smime-type MIME parameter in S/MIME messages which some
11414 * New function BN_MONT_CTX_set_locked() to set montgomery parameters in
11415 a threadsafe manner. Modify rsa code to use new function and add calls
11416 to dsa and dh code (which had race conditions before).
11420 * Include the fixed error library code in the C error file definitions
11421 instead of fixing them up at runtime. This keeps the error code
11422 structures constant.
11426 ### Changes between 0.9.7f and 0.9.7g [11 Apr 2005]
11428 [NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after
11431 * Fixes for newer kerberos headers. NB: the casts are needed because
11432 the 'length' field is signed on one version and unsigned on another
11433 with no (?) obvious way to tell the difference, without these VC++
11434 complains. Also the "definition" of FAR (blank) is no longer included
11435 nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up
11436 some needed definitions.
11440 * Undo Cygwin change.
11444 * Added support for proxy certificates according to RFC 3820.
11445 Because they may be a security thread to unaware applications,
11446 they must be explicitly allowed in run-time. See
11447 docs/HOWTO/proxy_certificates.txt for further information.
11451 ### Changes between 0.9.7e and 0.9.7f [22 Mar 2005]
11453 * Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating
11454 server and client random values. Previously
11455 (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in
11456 less random data when sizeof(time_t) > 4 (some 64 bit platforms).
11458 This change has negligible security impact because:
11460 1. Server and client random values still have 24 bytes of pseudo random
11463 2. Server and client random values are sent in the clear in the initial
11466 3. The master secret is derived using the premaster secret (48 bytes in
11467 size for static RSA ciphersuites) as well as client server and random
11470 The OpenSSL team would like to thank the UK NISCC for bringing this issue
11473 *Stephen Henson, reported by UK NISCC*
11475 * Use Windows randomness collection on Cygwin.
11479 * Fix hang in EGD/PRNGD query when communication socket is closed
11480 prematurely by EGD/PRNGD.
11482 *Darren Tucker <dtucker@zip.com.au> via Lutz Jänicke, resolves #1014*
11484 * Prompt for pass phrases when appropriate for PKCS12 input format.
11488 * Back-port of selected performance improvements from development
11489 branch, as well as improved support for PowerPC platforms.
11493 * Add lots of checks for memory allocation failure, error codes to indicate
11494 failure and freeing up memory if a failure occurs.
11496 *Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson*
11498 * Add new -passin argument to dgst.
11502 * Perform some character comparisons of different types in X509_NAME_cmp:
11503 this is needed for some certificates that re-encode DNs into UTF8Strings
11504 (in violation of RFC3280) and can't or won't issue name rollover
11509 * Make an explicit check during certificate validation to see that
11510 the CA setting in each certificate on the chain is correct. As a
11511 side effect always do the following basic checks on extensions,
11512 not just when there's an associated purpose to the check:
11514 - if there is an unhandled critical extension (unless the user
11515 has chosen to ignore this fault)
11516 - if the path length has been exceeded (if one is set at all)
11517 - that certain extensions fit the associated purpose (if one has
11522 ### Changes between 0.9.7d and 0.9.7e [25 Oct 2004]
11524 * Avoid a race condition when CRLs are checked in a multi threaded
11525 environment. This would happen due to the reordering of the revoked
11526 entries during signature checking and serial number lookup. Now the
11527 encoding is cached and the serial number sort performed under a lock.
11528 Add new STACK function sk_is_sorted().
11532 * Add Delta CRL to the extension code.
11536 * Various fixes to s3_pkt.c so alerts are sent properly.
11538 *David Holmes <d.holmes@f5.com>*
11540 * Reduce the chances of duplicate issuer name and serial numbers (in
11541 violation of RFC3280) using the OpenSSL certificate creation utilities.
11542 This is done by creating a random 64 bit value for the initial serial
11543 number when a serial number file is created or when a self signed
11544 certificate is created using 'openssl req -x509'. The initial serial
11545 number file is created using 'openssl x509 -next_serial' in CA.pl
11546 rather than being initialized to 1.
11550 ### Changes between 0.9.7c and 0.9.7d [17 Mar 2004]
11552 * Fix null-pointer assignment in do_change_cipher_spec() revealed
11553 by using the Codenomicon TLS Test Tool ([CVE-2004-0079])
11555 *Joe Orton, Steve Henson*
11557 * Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites
11560 *Joe Orton, Steve Henson*
11562 * Make it possible to have multiple active certificates with the same
11563 subject in the CA index file. This is done only if the keyword
11564 'unique_subject' is set to 'no' in the main CA section (default
11565 if 'CA_default') of the configuration file. The value is saved
11566 with the database itself in a separate index attribute file,
11567 named like the index file with '.attr' appended to the name.
11571 * X509 verify fixes. Disable broken certificate workarounds when
11572 X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if
11573 keyUsage extension present. Don't accept CRLs with unhandled critical
11574 extensions: since verify currently doesn't process CRL extensions this
11575 rejects a CRL with *any* critical extensions. Add new verify error codes
11580 * When creating an OCSP nonce use an OCTET STRING inside the extnValue.
11581 A clarification of RFC2560 will require the use of OCTET STRINGs and
11582 some implementations cannot handle the current raw format. Since OpenSSL
11583 copies and compares OCSP nonces as opaque blobs without any attempt at
11584 parsing them this should not create any compatibility issues.
11588 * New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when
11589 calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without
11590 this HMAC (and other) operations are several times slower than OpenSSL
11595 * Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex().
11597 *Peter Sylvester <Peter.Sylvester@EdelWeb.fr>*
11599 * Use the correct content when signing type "other".
11603 ### Changes between 0.9.7b and 0.9.7c [30 Sep 2003]
11605 * Fix various bugs revealed by running the NISCC test suite:
11607 Stop out of bounds reads in the ASN1 code when presented with
11608 invalid tags (CVE-2003-0543 and CVE-2003-0544).
11610 Free up ASN1_TYPE correctly if ANY type is invalid ([CVE-2003-0545]).
11612 If verify callback ignores invalid public key errors don't try to check
11613 certificate signature with the NULL public key.
11617 * New -ignore_err option in ocsp application to stop the server
11618 exiting on the first error in a request.
11622 * In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
11623 if the server requested one: as stated in TLS 1.0 and SSL 3.0
11628 * In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
11629 extra data after the compression methods not only for TLS 1.0
11630 but also for SSL 3.0 (as required by the specification).
11632 *Bodo Moeller; problem pointed out by Matthias Loepfe*
11634 * Change X509_certificate_type() to mark the key as exported/exportable
11635 when it's 512 *bits* long, not 512 bytes.
11639 * Change AES_cbc_encrypt() so it outputs exact multiple of
11640 blocks during encryption.
11644 * Various fixes to base64 BIO and non blocking I/O. On write
11645 flushes were not handled properly if the BIO retried. On read
11646 data was not being buffered properly and had various logic bugs.
11647 This also affects blocking I/O when the data being decoded is a
11652 * Various S/MIME bugfixes and compatibility changes:
11653 output correct application/pkcs7 MIME type if
11654 PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures.
11655 Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening
11656 of files as .eml work). Correctly handle very long lines in MIME
11661 ### Changes between 0.9.7a and 0.9.7b [10 Apr 2003]
11663 * Countermeasure against the Klima-Pokorny-Rosa extension of
11664 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
11665 a protocol version number mismatch like a decryption error
11666 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
11670 * Turn on RSA blinding by default in the default implementation
11671 to avoid a timing attack. Applications that don't want it can call
11672 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
11673 They would be ill-advised to do so in most cases.
11675 *Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller*
11677 * Change RSA blinding code so that it works when the PRNG is not
11678 seeded (in this case, the secret RSA exponent is abused as
11679 an unpredictable seed -- if it is not unpredictable, there
11680 is no point in blinding anyway). Make RSA blinding thread-safe
11681 by remembering the creator's thread ID in rsa->blinding and
11682 having all other threads use local one-time blinding factors
11683 (this requires more computation than sharing rsa->blinding, but
11684 avoids excessive locking; and if an RSA object is not shared
11685 between threads, blinding will still be very fast).
11689 * Fixed a typo bug that would cause ENGINE_set_default() to set an
11690 ENGINE as defaults for all supported algorithms irrespective of
11691 the 'flags' parameter. 'flags' is now honoured, so applications
11692 should make sure they are passing it correctly.
11696 * Target "mingw" now allows native Windows code to be generated in
11697 the Cygwin environment as well as with the MinGW compiler.
11701 ### Changes between 0.9.7 and 0.9.7a [19 Feb 2003]
11703 * In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
11704 via timing by performing a MAC computation even if incorrect
11705 block cipher padding has been found. This is a countermeasure
11706 against active attacks where the attacker has to distinguish
11707 between bad padding and a MAC verification error. ([CVE-2003-0078])
11709 *Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
11710 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
11711 Martin Vuagnoux (EPFL, Ilion)*
11713 * Make the no-err option work as intended. The intention with no-err
11714 is not to have the whole error stack handling routines removed from
11715 libcrypto, it's only intended to remove all the function name and
11716 reason texts, thereby removing some of the footprint that may not
11717 be interesting if those errors aren't displayed anyway.
11719 NOTE: it's still possible for any application or module to have its
11720 own set of error texts inserted. The routines are there, just not
11721 used by default when no-err is given.
11725 * Add support for FreeBSD on IA64.
11727 *dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454*
11729 * Adjust DES_cbc_cksum() so it returns the same value as the MIT
11730 Kerberos function mit_des_cbc_cksum(). Before this change,
11731 the value returned by DES_cbc_cksum() was like the one from
11732 mit_des_cbc_cksum(), except the bytes were swapped.
11734 *Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte*
11736 * Allow an application to disable the automatic SSL chain building.
11737 Before this a rather primitive chain build was always performed in
11738 ssl3_output_cert_chain(): an application had no way to send the
11739 correct chain if the automatic operation produced an incorrect result.
11741 Now the chain builder is disabled if either:
11743 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert().
11745 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set.
11747 The reasoning behind this is that an application would not want the
11748 auto chain building to take place if extra chain certificates are
11749 present and it might also want a means of sending no additional
11750 certificates (for example the chain has two certificates and the
11755 * Add the possibility to build without the ENGINE framework.
11757 *Steven Reddie <smr@essemer.com.au> via Richard Levitte*
11759 * Under Win32 gmtime() can return NULL: check return value in
11760 OPENSSL_gmtime(). Add error code for case where gmtime() fails.
11764 * DSA routines: under certain error conditions uninitialized BN objects
11765 could be freed. Solution: make sure initialization is performed early
11766 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>,
11767 Nils Larsch <nla@trustcenter.de> via PR#459)
11771 * Another fix for SSLv2 session ID handling: the session ID was incorrectly
11772 checked on reconnect on the client side, therefore session resumption
11773 could still fail with a "ssl session id is different" error. This
11774 behaviour is masked when SSL_OP_ALL is used due to
11775 SSL_OP_MICROSOFT_SESS_ID_BUG being set.
11776 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
11777 followup to PR #377.
11781 * IA-32 assembler support enhancements: unified ELF targets, support
11782 for SCO/Caldera platforms, fix for Cygwin shared build.
11786 * Add support for FreeBSD on sparc64. As a consequence, support for
11787 FreeBSD on non-x86 processors is separate from x86 processors on
11788 the config script, much like the NetBSD support.
11790 *Richard Levitte & Kris Kennaway <kris@obsecurity.org>*
11792 ### Changes between 0.9.6h and 0.9.7 [31 Dec 2002]
11794 [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after
11797 * Fix session ID handling in SSLv2 client code: the SERVER FINISHED
11798 code (06) was taken as the first octet of the session ID and the last
11799 octet was ignored consequently. As a result SSLv2 client side session
11800 caching could not have worked due to the session ID mismatch between
11802 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
11807 * Change the declaration of needed Kerberos libraries to use EX_LIBS
11808 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is
11813 * The hw_ncipher.c engine requires dynamic locks. Unfortunately, it
11814 seems that in spite of existing for more than a year, many application
11815 author have done nothing to provide the necessary callbacks, which
11816 means that this particular engine will not work properly anywhere.
11817 This is a very unfortunate situation which forces us, in the name
11818 of usability, to give the hw_ncipher.c a static lock, which is part
11820 NOTE: This is for the 0.9.7 series ONLY. This hack will never
11821 appear in 0.9.8 or later. We EXPECT application authors to have
11822 dealt properly with this when 0.9.8 is released (unless we actually
11823 make such changes in the libcrypto locking code that changes will
11824 have to be made anyway).
11828 * In asn1_d2i_read_bio() repeatedly call BIO_read() until all content
11829 octets have been read, EOF or an error occurs. Without this change
11830 some truncated ASN1 structures will not produce an error.
11834 * Disable Heimdal support, since it hasn't been fully implemented.
11835 Still give the possibility to force the use of Heimdal, but with
11836 warnings and a request that patches get sent to openssl-dev.
11840 * Add the VC-CE target, introduce the WINCE sysname, and add
11841 INSTALL.WCE and appropriate conditionals to make it build.
11843 *Steven Reddie <smr@essemer.com.au> via Richard Levitte*
11845 * Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and
11846 cygssl-x.y.z.dll, where x, y and z are the major, minor and
11847 edit numbers of the version.
11849 *Corinna Vinschen <vinschen@redhat.com> and Richard Levitte*
11851 * Introduce safe string copy and catenation functions
11852 (BUF_strlcpy() and BUF_strlcat()).
11854 *Ben Laurie (CHATS) and Richard Levitte*
11856 * Avoid using fixed-size buffers for one-line DNs.
11858 *Ben Laurie (CHATS)*
11860 * Add BUF_MEM_grow_clean() to avoid information leakage when
11861 resizing buffers containing secrets, and use where appropriate.
11863 *Ben Laurie (CHATS)*
11865 * Avoid using fixed size buffers for configuration file location.
11867 *Ben Laurie (CHATS)*
11869 * Avoid filename truncation for various CA files.
11871 *Ben Laurie (CHATS)*
11873 * Use sizeof in preference to magic numbers.
11875 *Ben Laurie (CHATS)*
11877 * Avoid filename truncation in cert requests.
11879 *Ben Laurie (CHATS)*
11881 * Add assertions to check for (supposedly impossible) buffer
11884 *Ben Laurie (CHATS)*
11886 * Don't cache truncated DNS entries in the local cache (this could
11887 potentially lead to a spoofing attack).
11889 *Ben Laurie (CHATS)*
11891 * Fix various buffers to be large enough for hex/decimal
11892 representations in a platform independent manner.
11894 *Ben Laurie (CHATS)*
11896 * Add CRYPTO_realloc_clean() to avoid information leakage when
11897 resizing buffers containing secrets, and use where appropriate.
11899 *Ben Laurie (CHATS)*
11901 * Add BIO_indent() to avoid much slightly worrying code to do
11904 *Ben Laurie (CHATS)*
11906 * Convert sprintf()/BIO_puts() to BIO_printf().
11908 *Ben Laurie (CHATS)*
11910 * buffer_gets() could terminate with the buffer only half
11913 *Ben Laurie (CHATS)*
11915 * Add assertions to prevent user-supplied crypto functions from
11916 overflowing internal buffers by having large block sizes, etc.
11918 *Ben Laurie (CHATS)*
11920 * New OPENSSL_assert() macro (similar to assert(), but enabled
11923 *Ben Laurie (CHATS)*
11925 * Eliminate unused copy of key in RC4.
11927 *Ben Laurie (CHATS)*
11929 * Eliminate unused and incorrectly sized buffers for IV in pem.h.
11931 *Ben Laurie (CHATS)*
11933 * Fix off-by-one error in EGD path.
11935 *Ben Laurie (CHATS)*
11937 * If RANDFILE path is too long, ignore instead of truncating.
11939 *Ben Laurie (CHATS)*
11941 * Eliminate unused and incorrectly sized X.509 structure
11944 *Ben Laurie (CHATS)*
11946 * Eliminate unused and dangerous function knumber().
11948 *Ben Laurie (CHATS)*
11950 * Eliminate unused and dangerous structure, KSSL_ERR.
11952 *Ben Laurie (CHATS)*
11954 * Protect against overlong session ID context length in an encoded
11955 session object. Since these are local, this does not appear to be
11958 *Ben Laurie (CHATS)*
11960 * Change from security patch (see 0.9.6e below) that did not affect
11961 the 0.9.6 release series:
11963 Remote buffer overflow in SSL3 protocol - an attacker could
11964 supply an oversized master key in Kerberos-enabled versions.
11967 *Ben Laurie (CHATS)*
11969 * Change the SSL kerb5 codes to match RFC 2712.
11973 * Make -nameopt work fully for req and add -reqopt switch.
11975 *Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson*
11977 * The "block size" for block ciphers in CFB and OFB mode should be 1.
11979 *Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>*
11981 * Make sure tests can be performed even if the corresponding algorithms
11982 have been removed entirely. This was also the last step to make
11983 OpenSSL compilable with DJGPP under all reasonable conditions.
11985 *Richard Levitte, Doug Kaufman <dkaufman@rahul.net>*
11987 * Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT
11988 to allow version independent disabling of normally unselected ciphers,
11989 which may be activated as a side-effect of selecting a single cipher.
11991 (E.g., cipher list string "RSA" enables ciphersuites that are left
11992 out of "ALL" because they do not provide symmetric encryption.
11993 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.)
11995 *Lutz Jaenicke, Bodo Moeller*
11997 * Add appropriate support for separate platform-dependent build
11998 directories. The recommended way to make a platform-dependent
11999 build directory is the following (tested on Linux), maybe with
12002 # Place yourself outside of the OpenSSL source tree. In
12003 # this example, the environment variable OPENSSL_SOURCE
12004 # is assumed to contain the absolute OpenSSL source directory.
12005 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`"
12006 cd objtree/"`uname -s`-`uname -r`-`uname -m`"
12007 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do
12008 mkdir -p `dirname $F`
12009 ln -s $OPENSSL_SOURCE/$F $F
12012 To be absolutely sure not to disturb the source tree, a "make clean"
12013 is a good thing. If it isn't successful, don't worry about it,
12014 it probably means the source directory is very clean.
12018 * Make sure any ENGINE control commands make local copies of string
12019 pointers passed to them whenever necessary. Otherwise it is possible
12020 the caller may have overwritten (or deallocated) the original string
12021 data when a later ENGINE operation tries to use the stored values.
12023 *Götz Babin-Ebell <babinebell@trustcenter.de>*
12025 * Improve diagnostics in file reading and command-line digests.
12027 *Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>*
12029 * Add AES modes CFB and OFB to the object database. Correct an
12030 error in AES-CFB decryption.
12034 * Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this
12035 allows existing EVP_CIPHER_CTX structures to be reused after
12036 calling `EVP_*Final()`. This behaviour is used by encryption
12037 BIOs and some applications. This has the side effect that
12038 applications must explicitly clean up cipher contexts with
12039 EVP_CIPHER_CTX_cleanup() or they will leak memory.
12043 * Check the values of dna and dnb in bn_mul_recursive before calling
12044 bn_mul_comba (a non zero value means the a or b arrays do not contain
12045 n2 elements) and fallback to bn_mul_normal if either is not zero.
12049 * Fix escaping of non-ASCII characters when using the -subj option
12050 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>)
12054 * Make object definitions compliant to LDAP (RFC2256): SN is the short
12055 form for "surname", serialNumber has no short form.
12056 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798;
12057 therefore remove "mail" short name for "internet 7".
12058 The OID for unique identifiers in X509 certificates is
12059 x500UniqueIdentifier, not uniqueIdentifier.
12060 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>)
12064 * Add an "init" command to the ENGINE config module and auto initialize
12065 ENGINEs. Without any "init" command the ENGINE will be initialized
12066 after all ctrl commands have been executed on it. If init=1 the
12067 ENGINE is initialized at that point (ctrls before that point are run
12068 on the uninitialized ENGINE and after on the initialized one). If
12069 init=0 then the ENGINE will not be initialized at all.
12073 * Fix the 'app_verify_callback' interface so that the user-defined
12074 argument is actually passed to the callback: In the
12075 SSL_CTX_set_cert_verify_callback() prototype, the callback
12076 declaration has been changed from
12079 int (*cb)(X509_STORE_CTX *,void *);
12080 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
12081 i=s->ctx->app_verify_callback(&ctx)
12082 has been changed into
12083 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
12085 To update applications using SSL_CTX_set_cert_verify_callback(),
12086 a dummy argument can be added to their callback functions.
12088 *D. K. Smetters <smetters@parc.xerox.com>*
12090 * Added the '4758cca' ENGINE to support IBM 4758 cards.
12092 *Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe*
12094 * Add and OPENSSL_LOAD_CONF define which will cause
12095 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
12096 This allows older applications to transparently support certain
12097 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
12098 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
12099 load the config file and OPENSSL_add_all_algorithms_conf() which will
12100 always load it have also been added.
12104 * Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
12105 Adjust NIDs and EVP layer.
12107 *Stephen Sprunk <stephen@sprunk.org> and Richard Levitte*
12109 * Config modules support in openssl utility.
12111 Most commands now load modules from the config file,
12112 though in a few (such as version) this isn't done
12113 because it couldn't be used for anything.
12115 In the case of ca and req the config file used is
12116 the same as the utility itself: that is the -config
12117 command line option can be used to specify an
12122 * Move default behaviour from OPENSSL_config(). If appname is NULL
12123 use "openssl_conf" if filename is NULL use default openssl config file.
12127 * Add an argument to OPENSSL_config() to allow the use of an alternative
12128 config section name. Add a new flag to tolerate a missing config file
12129 and move code to CONF_modules_load_file().
12133 * Support for crypto accelerator cards from Accelerated Encryption
12134 Processing, www.aep.ie. (Use engine 'aep')
12135 The support was copied from 0.9.6c [engine] and adapted/corrected
12136 to work with the new engine framework.
12138 *AEP Inc. and Richard Levitte*
12140 * Support for SureWare crypto accelerator cards from Baltimore
12141 Technologies. (Use engine 'sureware')
12142 The support was copied from 0.9.6c [engine] and adapted
12143 to work with the new engine framework.
12147 * Have the CHIL engine fork-safe (as defined by nCipher) and actually
12148 make the newer ENGINE framework commands for the CHIL engine work.
12150 *Toomas Kiisk <vix@cyber.ee> and Richard Levitte*
12152 * Make it possible to produce shared libraries on ReliantUNIX.
12154 *Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte*
12156 * Add the configuration target debug-linux-ppro.
12157 Make 'openssl rsa' use the general key loading routines
12158 implemented in `apps.c`, and make those routines able to
12159 handle the key format FORMAT_NETSCAPE and the variant
12162 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
12164 * Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
12166 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
12168 * Add -keyform to rsautl, and document -engine.
12170 *Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>*
12172 * Change BIO_new_file (crypto/bio/bss_file.c) to use new
12173 BIO_R_NO_SUCH_FILE error code rather than the generic
12174 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
12178 * Add new functions
12179 ERR_peek_last_error
12180 ERR_peek_last_error_line
12181 ERR_peek_last_error_line_data.
12182 These are similar to
12184 ERR_peek_error_line
12185 ERR_peek_error_line_data,
12186 but report on the latest error recorded rather than the first one
12187 still in the error queue.
12189 *Ben Laurie, Bodo Moeller*
12191 * default_algorithms option in ENGINE config module. This allows things
12193 default_algorithms = ALL
12194 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
12198 * Preliminary ENGINE config module.
12202 * New experimental application configuration code.
12206 * Change the AES code to follow the same name structure as all other
12207 symmetric ciphers, and behave the same way. Move everything to
12208 the directory crypto/aes, thereby obsoleting crypto/rijndael.
12210 *Stephen Sprunk <stephen@sprunk.org> and Richard Levitte*
12212 * SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
12214 *Ben Laurie and Theo de Raadt*
12216 * Add option to output public keys in req command.
12218 *Massimiliano Pala madwolf@openca.org*
12220 * Use wNAFs in EC_POINTs_mul() for improved efficiency
12221 (up to about 10% better than before for P-192 and P-224).
12225 * New functions/macros
12227 SSL_CTX_set_msg_callback(ctx, cb)
12228 SSL_CTX_set_msg_callback_arg(ctx, arg)
12229 SSL_set_msg_callback(ssl, cb)
12230 SSL_set_msg_callback_arg(ssl, arg)
12232 to request calling a callback function
12234 void cb(int write_p, int version, int content_type,
12235 const void *buf, size_t len, SSL *ssl, void *arg)
12237 whenever a protocol message has been completely received
12238 (write_p == 0) or sent (write_p == 1). Here 'version' is the
12239 protocol version according to which the SSL library interprets
12240 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
12241 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
12242 the content type as defined in the SSL 3.0/TLS 1.0 protocol
12243 specification (change_cipher_spec(20), alert(21), handshake(22)).
12244 'buf' and 'len' point to the actual message, 'ssl' to the
12245 SSL object, and 'arg' is the application-defined value set by
12246 SSL[_CTX]_set_msg_callback_arg().
12248 'openssl s_client' and 'openssl s_server' have new '-msg' options
12249 to enable a callback that displays all protocol messages.
12253 * Change the shared library support so shared libraries are built as
12254 soon as the corresponding static library is finished, and thereby get
12255 openssl and the test programs linked against the shared library.
12256 This still only happens when the keyword "shard" has been given to
12257 the configuration scripts.
12259 NOTE: shared library support is still an experimental thing, and
12260 backward binary compatibility is still not guaranteed.
12262 *"Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte*
12264 * Add support for Subject Information Access extension.
12266 *Peter Sylvester <Peter.Sylvester@EdelWeb.fr>*
12268 * Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
12269 additional bytes when new memory had to be allocated, not just
12270 when reusing an existing buffer.
12274 * New command line and configuration option 'utf8' for the req command.
12275 This allows field values to be specified as UTF8 strings.
12279 * Add -multi and -mr options to "openssl speed" - giving multiple parallel
12280 runs for the former and machine-readable output for the latter.
12284 * Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
12285 of the e-mail address in the DN (i.e., it will go into a certificate
12286 extension only). The new configuration file option 'email_in_dn = no'
12287 has the same effect.
12289 *Massimiliano Pala madwolf@openca.org*
12291 * Change all functions with names starting with `des_` to be starting
12292 with `DES_` instead. Add wrappers that are compatible with libdes,
12293 but are named `_ossl_old_des_*`. Finally, add macros that map the
12294 `des_*` symbols to the corresponding `_ossl_old_des_*` if libdes
12295 compatibility is desired. If OpenSSL 0.9.6c compatibility is
12296 desired, the `des_*` symbols will be mapped to `DES_*`, with one
12299 Since we provide two compatibility mappings, the user needs to
12300 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
12301 compatibility is desired. The default (i.e., when that macro
12302 isn't defined) is OpenSSL 0.9.6c compatibility.
12304 There are also macros that enable and disable the support of old
12305 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
12306 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
12307 are defined, the default will apply: to support the old des routines.
12309 In either case, one must include openssl/des.h to get the correct
12310 definitions. Do not try to just include openssl/des_old.h, that
12313 NOTE: This is a major break of an old API into a new one. Software
12314 authors are encouraged to switch to the `DES_` style functions. Some
12315 time in the future, des_old.h and the libdes compatibility functions
12316 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
12317 default), and then completely removed.
12321 * Test for certificates which contain unsupported critical extensions.
12322 If such a certificate is found during a verify operation it is
12323 rejected by default: this behaviour can be overridden by either
12324 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
12325 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
12326 X509_supported_extension() has also been added which returns 1 if a
12327 particular extension is supported.
12331 * Modify the behaviour of EVP cipher functions in similar way to digests
12332 to retain compatibility with existing code.
12336 * Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
12337 compatibility with existing code. In particular the 'ctx' parameter does
12338 not have to be to be initialized before the call to EVP_DigestInit() and
12339 it is tidied up after a call to EVP_DigestFinal(). New function
12340 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
12341 EVP_MD_CTX_copy() changed to not require the destination to be
12342 initialized valid and new function EVP_MD_CTX_copy_ex() added which
12343 requires the destination to be valid.
12345 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
12346 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
12350 * Change ssl3_get_message (ssl/s3_both.c) and the functions using it
12351 so that complete 'Handshake' protocol structures are kept in memory
12352 instead of overwriting 'msg_type' and 'length' with 'body' data.
12356 * Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
12358 *Massimo Santin via Richard Levitte*
12360 * Major restructuring to the underlying ENGINE code. This includes
12361 reduction of linker bloat, separation of pure "ENGINE" manipulation
12362 (initialisation, etc) from functionality dealing with implementations
12363 of specific crypto interfaces. This change also introduces integrated
12364 support for symmetric ciphers and digest implementations - so ENGINEs
12365 can now accelerate these by providing EVP_CIPHER and EVP_MD
12366 implementations of their own. This is detailed in
12367 [crypto/engine/README.md](crypto/engine/README.md)
12368 as it couldn't be adequately described here. However, there are a few
12369 API changes worth noting - some RSA, DSA, DH, and RAND functions that
12370 were changed in the original introduction of ENGINE code have now
12371 reverted back - the hooking from this code to ENGINE is now a good
12372 deal more passive and at run-time, operations deal directly with
12373 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
12374 dereferencing through an ENGINE pointer any more. Also, the ENGINE
12375 functions dealing with `BN_MOD_EXP[_CRT]` handlers have been removed -
12376 they were not being used by the framework as there is no concept of a
12377 BIGNUM_METHOD and they could not be generalised to the new
12378 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
12379 ENGINE_cpy() has been removed as it cannot be consistently defined in
12384 * Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
12388 * Change mkdef.pl to sort symbols that get the same entry number,
12389 and make sure the automatically generated functions `ERR_load_*`
12390 become part of libeay.num as well.
12394 * New function SSL_renegotiate_pending(). This returns true once
12395 renegotiation has been requested (either SSL_renegotiate() call
12396 or HelloRequest/ClientHello received from the peer) and becomes
12397 false once a handshake has been completed.
12398 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
12399 sends a HelloRequest, but does not ensure that a handshake takes
12400 place. SSL_renegotiate_pending() is useful for checking if the
12401 client has followed the request.)
12405 * New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
12406 By default, clients may request session resumption even during
12407 renegotiation (if session ID contexts permit); with this option,
12408 session resumption is possible only in the first handshake.
12410 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes
12411 more bits available for options that should not be part of
12412 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION).
12416 * Add some demos for certificate and certificate request creation.
12420 * Make maximum certificate chain size accepted from the peer application
12421 settable (`SSL*_get/set_max_cert_list()`), as proposed by
12422 "Douglas E. Engert" <deengert@anl.gov>.
12426 * Add support for shared libraries for Unixware-7
12427 (Boyd Lynn Gerber <gerberb@zenez.com>).
12431 * Add a "destroy" handler to ENGINEs that allows structural cleanup to
12432 be done prior to destruction. Use this to unload error strings from
12433 ENGINEs that load their own error strings. NB: This adds two new API
12434 functions to "get" and "set" this destroy handler in an ENGINE.
12438 * Alter all existing ENGINE implementations (except "openssl" and
12439 "openbsd") to dynamically instantiate their own error strings. This
12440 makes them more flexible to be built both as statically-linked ENGINEs
12441 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
12442 Also, add stub code to each that makes building them as self-contained
12443 shared-libraries easier (see [README-Engine.md](README-Engine.md)).
12447 * Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
12448 implementations into applications that are completely implemented in
12449 self-contained shared-libraries. The "dynamic" ENGINE exposes control
12450 commands that can be used to configure what shared-library to load and
12451 to control aspects of the way it is handled. Also, made an update to
12452 the [README-Engine.md](README-Engine.md) file
12453 that brings its information up-to-date and
12454 provides some information and instructions on the "dynamic" ENGINE
12455 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
12459 * Make it possible to unload ranges of ERR strings with a new
12460 "ERR_unload_strings" function.
12464 * Add a copy() function to EVP_MD.
12468 * Make EVP_MD routines take a context pointer instead of just the
12469 md_data void pointer.
12473 * Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
12474 that the digest can only process a single chunk of data
12475 (typically because it is provided by a piece of
12476 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
12477 is only going to provide a single chunk of data, and hence the
12478 framework needn't accumulate the data for oneshot drivers.
12482 * As with "ERR", make it possible to replace the underlying "ex_data"
12483 functions. This change also alters the storage and management of global
12484 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
12485 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
12486 index counters. The API functions that use this state have been changed
12487 to take a "class_index" rather than pointers to the class's local STACK
12488 and counter, and there is now an API function to dynamically create new
12489 classes. This centralisation allows us to (a) plug a lot of the
12490 thread-safety problems that existed, and (b) makes it possible to clean
12491 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
12492 such data would previously have always leaked in application code and
12493 workarounds were in place to make the memory debugging turn a blind eye
12494 to it. Application code that doesn't use this new function will still
12495 leak as before, but their memory debugging output will announce it now
12496 rather than letting it slide.
12498 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
12499 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
12500 has a return value to indicate success or failure.
12504 * Make it possible to replace the underlying "ERR" functions such that the
12505 global state (2 LHASH tables and 2 locks) is only used by the "default"
12506 implementation. This change also adds two functions to "get" and "set"
12507 the implementation prior to it being automatically set the first time
12508 any other ERR function takes place. Ie. an application can call "get",
12509 pass the return value to a module it has just loaded, and that module
12510 can call its own "set" function using that value. This means the
12511 module's "ERR" operations will use (and modify) the error state in the
12512 application and not in its own statically linked copy of OpenSSL code.
12516 * Give DH, DSA, and RSA types their own `*_up_ref()` function to increment
12517 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
12518 the operation, and provides a more encapsulated way for external code
12519 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
12520 to use these functions rather than manually incrementing the counts.
12522 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
12526 * Add EVP test program.
12530 * Add symmetric cipher support to ENGINE. Expect the API to change!
12534 * New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
12535 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
12536 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
12537 These allow a CRL to be built without having to access X509_CRL fields
12538 directly. Modify 'ca' application to use new functions.
12542 * Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
12543 bug workarounds. Rollback attack detection is a security feature.
12544 The problem will only arise on OpenSSL servers when TLSv1 is not
12545 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
12546 Software authors not wanting to support TLSv1 will have special reasons
12547 for their choice and can explicitly enable this option.
12549 *Bodo Moeller, Lutz Jaenicke*
12551 * Rationalise EVP so it can be extended: don't include a union of
12552 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX
12553 (similar to those existing for EVP_CIPHER_CTX).
12558 EVP_MD_CTX_init(&md); /* new function call */
12559 EVP_DigestInit(&md, EVP_sha1());
12560 EVP_DigestUpdate(&md, in, len);
12561 EVP_DigestFinal(&md, out, NULL);
12562 EVP_MD_CTX_cleanup(&md); /* new function call */
12566 * Make DES key schedule conform to the usual scheme, as well as
12567 correcting its structure. This means that calls to DES functions
12568 now have to pass a pointer to a des_key_schedule instead of a
12569 plain des_key_schedule (which was actually always a pointer
12572 des_key_schedule ks;
12574 des_set_key_checked(..., &ks);
12575 des_ncbc_encrypt(..., &ks, ...);
12577 (Note that a later change renames 'des_...' into 'DES_...'.)
12581 * Initial reduction of linker bloat: the use of some functions, such as
12582 PEM causes large amounts of unused functions to be linked in due to
12583 poor organisation. For example pem_all.c contains every PEM function
12584 which has a knock on effect of linking in large amounts of (unused)
12585 ASN1 code. Grouping together similar functions and splitting unrelated
12586 functions prevents this.
12590 * Cleanup of EVP macros.
12594 * Change historical references to `{NID,SN,LN}_des_ede` and ede3 to add the
12595 correct `_ecb suffix`.
12599 * Add initial OCSP responder support to ocsp application. The
12600 revocation information is handled using the text based index
12601 use by the ca application. The responder can either handle
12602 requests generated internally, supplied in files (for example
12603 via a CGI script) or using an internal minimal server.
12607 * Add configuration choices to get zlib compression for TLS.
12611 * Changes to Kerberos SSL for RFC 2712 compliance:
12612 1. Implemented real KerberosWrapper, instead of just using
12613 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
12614 2. Implemented optional authenticator field of KerberosWrapper.
12616 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
12617 and authenticator structs; see crypto/krb5/.
12619 Generalized Kerberos calls to support multiple Kerberos libraries.
12620 *Vern Staats <staatsvr@asc.hpc.mil>, Jeffrey Altman <jaltman@columbia.edu>
12621 via Richard Levitte*
12623 * Cause 'openssl speed' to use fully hard-coded DSA keys as it
12624 already does with RSA. testdsa.h now has 'priv_key/pub_key'
12625 values for each of the key sizes rather than having just
12626 parameters (and 'speed' generating keys each time).
12630 * Speed up EVP routines.
12633 pe 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
12634 s-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
12635 s-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
12636 s-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
12638 s-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
12639 s-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
12640 s-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
12643 s-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
12645 s-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
12649 * Added the OS2-EMX target.
12651 *"Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte*
12653 * Rewrite commands to use `NCONF` routines instead of the old `CONF`.
12654 New functions to support `NCONF` routines in extension code.
12655 New function `CONF_set_nconf()`
12656 to allow functions which take an `NCONF` to also handle the old `LHASH`
12657 structure: this means that the old `CONF` compatible routines can be
12658 retained (in particular w.rt. extensions) without having to duplicate the
12659 code. New function `X509V3_add_ext_nconf_sk()` to add extensions to a stack.
12663 * Enhance the general user interface with mechanisms for inner control
12664 and with possibilities to have yes/no kind of prompts.
12668 * Change all calls to low-level digest routines in the library and
12669 applications to use EVP. Add missing calls to HMAC_cleanup() and
12670 don't assume HMAC_CTX can be copied using memcpy().
12672 *Verdon Walker <VWalker@novell.com>, Steve Henson*
12674 * Add the possibility to control engines through control names but with
12675 arbitrary arguments instead of just a string.
12676 Change the key loaders to take a UI_METHOD instead of a callback
12677 function pointer. NOTE: this breaks binary compatibility with earlier
12678 versions of OpenSSL [engine].
12679 Adapt the nCipher code for these new conditions and add a card insertion
12684 * Enhance the general user interface with mechanisms to better support
12685 dialog box interfaces, application-defined prompts, the possibility
12686 to use defaults (for example default passwords from somewhere else)
12687 and interrupts/cancellations.
12691 * Tidy up PKCS#12 attribute handling. Add support for the CSP name
12692 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
12696 * Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
12697 tidy up some unnecessarily weird code in 'sk_new()').
12699 *Geoff, reported by Diego Tartara <dtartara@novamens.com>*
12701 * Change the key loading routines for ENGINEs to use the same kind
12702 callback (pem_password_cb) as all other routines that need this
12707 * Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
12708 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
12709 than this minimum value is recommended.
12713 * New random seeder for OpenVMS, using the system process statistics
12714 that are easily reachable.
12718 * Windows apparently can't transparently handle global
12719 variables defined in DLLs. Initialisations such as:
12721 const ASN1_ITEM *it = &ASN1_INTEGER_it;
12723 won't compile. This is used by the any applications that need to
12724 declare their own ASN1 modules. This was fixed by adding the option
12725 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
12726 needed for static libraries under Win32.
12730 * New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
12731 setting of purpose and trust fields. New X509_STORE trust and
12732 purpose functions and tidy up setting in other SSL functions.
12736 * Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
12737 structure. These are inherited by X509_STORE_CTX when it is
12738 initialised. This allows various defaults to be set in the
12739 X509_STORE structure (such as flags for CRL checking and custom
12740 purpose or trust settings) for functions which only use X509_STORE_CTX
12741 internally such as S/MIME.
12743 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
12744 trust settings if they are not set in X509_STORE. This allows X509_STORE
12745 purposes and trust (in S/MIME for example) to override any set by default.
12747 Add command line options for CRL checking to smime, s_client and s_server
12752 * Initial CRL based revocation checking. If the CRL checking flag(s)
12753 are set then the CRL is looked up in the X509_STORE structure and
12754 its validity and signature checked, then if the certificate is found
12755 in the CRL the verify fails with a revoked error.
12757 Various new CRL related callbacks added to X509_STORE_CTX structure.
12759 Command line options added to 'verify' application to support this.
12761 This needs some additional work, such as being able to handle multiple
12762 CRLs with different times, extension based lookup (rather than just
12763 by subject name) and ultimately more complete V2 CRL extension
12768 * Add a general user interface API (crypto/ui/). This is designed
12769 to replace things like des_read_password and friends (backward
12770 compatibility functions using this new API are provided).
12771 The purpose is to remove prompting functions from the DES code
12772 section as well as provide for prompting through dialog boxes in
12773 a window system and the like.
12777 * Add "ex_data" support to ENGINE so implementations can add state at a
12778 per-structure level rather than having to store it globally.
12782 * Make it possible for ENGINE structures to be copied when retrieved by
12783 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
12784 This causes the "original" ENGINE structure to act like a template,
12785 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
12786 operational state can be localised to each ENGINE structure, despite the
12787 fact they all share the same "methods". New ENGINE structures returned in
12788 this case have no functional references and the return value is the single
12789 structural reference. This matches the single structural reference returned
12790 by ENGINE_by_id() normally, when it is incremented on the pre-existing
12795 * Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
12796 needs to match any other type at all we need to manually clear the
12801 * Changes to the "openssl engine" utility to include;
12802 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
12803 about an ENGINE's available control commands.
12804 - executing control commands from command line arguments using the
12805 '-pre' and '-post' switches. '-post' is only used if '-t' is
12806 specified and the ENGINE is successfully initialised. The syntax for
12807 the individual commands are colon-separated, for example;
12808 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
12812 * New dynamic control command support for ENGINEs. ENGINEs can now
12813 declare their own commands (numbers), names (strings), descriptions,
12814 and input types for run-time discovery by calling applications. A
12815 subset of these commands are implicitly classed as "executable"
12816 depending on their input type, and only these can be invoked through
12817 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
12818 can be based on user input, config files, etc). The distinction is
12819 that "executable" commands cannot return anything other than a boolean
12820 result and can only support numeric or string input, whereas some
12821 discoverable commands may only be for direct use through
12822 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
12823 pointers, or other custom uses. The "executable" commands are to
12824 support parameterisations of ENGINE behaviour that can be
12825 unambiguously defined by ENGINEs and used consistently across any
12826 OpenSSL-based application. Commands have been added to all the
12827 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
12828 control over shared-library paths without source code alterations.
12832 * Changed all ENGINE implementations to dynamically allocate their
12833 ENGINEs rather than declaring them statically. Apart from this being
12834 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
12835 this also allows the implementations to compile without using the
12836 internal engine_int.h header.
12840 * Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
12841 'const' value. Any code that should be able to modify a RAND_METHOD
12842 should already have non-const pointers to it (ie. they should only
12843 modify their own ones).
12847 * Made a variety of little tweaks to the ENGINE code.
12848 - "atalla" and "ubsec" string definitions were moved from header files
12849 to C code. "nuron" string definitions were placed in variables
12850 rather than hard-coded - allowing parameterisation of these values
12851 later on via ctrl() commands.
12852 - Removed unused "#if 0"'d code.
12853 - Fixed engine list iteration code so it uses ENGINE_free() to release
12854 structural references.
12855 - Constified the RAND_METHOD element of ENGINE structures.
12856 - Constified various get/set functions as appropriate and added
12857 missing functions (including a catch-all ENGINE_cpy that duplicates
12858 all ENGINE values onto a new ENGINE except reference counts/state).
12859 - Removed NULL parameter checks in get/set functions. Setting a method
12860 or function to NULL is a way of cancelling out a previously set
12861 value. Passing a NULL ENGINE parameter is just plain stupid anyway
12862 and doesn't justify the extra error symbols and code.
12863 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
12864 flags from engine_int.h to engine.h.
12865 - Changed prototypes for ENGINE handler functions (init(), finish(),
12866 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
12870 * Implement binary inversion algorithm for BN_mod_inverse in addition
12871 to the algorithm using long division. The binary algorithm can be
12872 used only if the modulus is odd. On 32-bit systems, it is faster
12873 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
12874 roughly 5-15% for 256-bit moduli), so we use it only for moduli
12875 up to 450 bits. In 64-bit environments, the binary algorithm
12876 appears to be advantageous for much longer moduli; here we use it
12877 for moduli up to 2048 bits.
12881 * Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
12882 could not support the combine flag in choice fields.
12886 * Add a 'copy_extensions' option to the 'ca' utility. This copies
12887 extensions from a certificate request to the certificate.
12891 * Allow multiple 'certopt' and 'nameopt' options to be separated
12892 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
12893 file: this allows the display of the certificate about to be
12894 signed to be customised, to allow certain fields to be included
12895 or excluded and extension details. The old system didn't display
12896 multicharacter strings properly, omitted fields not in the policy
12897 and couldn't display additional details such as extensions.
12901 * Function EC_POINTs_mul for multiple scalar multiplication
12902 of an arbitrary number of elliptic curve points
12903 \sum scalars[i]*points[i],
12904 optionally including the generator defined for the EC_GROUP:
12905 scalar*generator + \sum scalars[i]*points[i].
12907 EC_POINT_mul is a simple wrapper function for the typical case
12908 that the point list has just one item (besides the optional
12913 * First EC_METHODs for curves over GF(p):
12915 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
12916 operations and provides various method functions that can also
12917 operate with faster implementations of modular arithmetic.
12919 EC_GFp_mont_method() reuses most functions that are part of
12920 EC_GFp_simple_method, but uses Montgomery arithmetic.
12922 *Bodo Moeller; point addition and point doubling
12923 implementation directly derived from source code provided by
12924 Lenka Fibikova <fibikova@exp-math.uni-essen.de>*
12926 * Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
12927 crypto/ec/ec_lib.c):
12929 Curves are EC_GROUP objects (with an optional group generator)
12930 based on EC_METHODs that are built into the library.
12932 Points are EC_POINT objects based on EC_GROUP objects.
12934 Most of the framework would be able to handle curves over arbitrary
12935 finite fields, but as there are no obvious types for fields other
12936 than GF(p), some functions are limited to that for now.
12940 * Add the -HTTP option to s_server. It is similar to -WWW, but requires
12941 that the file contains a complete HTTP response.
12945 * Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
12946 change the def and num file printf format specifier from "%-40sXXX"
12947 to "%-39s XXX". The latter will always guarantee a space after the
12948 field while the former will cause them to run together if the field
12949 is 40 of more characters long.
12953 * Constify the cipher and digest 'method' functions and structures
12954 and modify related functions to take constant EVP_MD and EVP_CIPHER
12959 * Hide BN_CTX structure details in bn_lcl.h instead of publishing them
12960 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
12964 * Modify `EVP_Digest*()` routines so they now return values. Although the
12965 internal software routines can never fail additional hardware versions
12970 * Clean up crypto/err/err.h and change some error codes to avoid conflicts:
12972 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
12973 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
12976 ERR_R_NESTED_ASN1_ERROR
12978 ERR_R_MISSING_ASN1_EOS
12979 were 4 .. 9, conflicting with
12980 ERR_LIB_RSA (= ERR_R_RSA_LIB)
12982 ERR_LIB_PEM (= ERR_R_PEM_LIB).
12983 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
12985 Add new error code 'ERR_R_INTERNAL_ERROR'.
12989 * Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
12994 * New option '-subj arg' for 'openssl req' and 'openssl ca'. This
12995 sets the subject name for a new request or supersedes the
12996 subject name in a given request. Formats that can be parsed are
12997 'CN=Some Name, OU=myOU, C=IT'
12999 'CN=Some Name/OU=myOU/C=IT'.
13001 Add options '-batch' and '-verbose' to 'openssl req'.
13003 *Massimiliano Pala <madwolf@hackmasters.net>*
13005 * Introduce the possibility to access global variables through
13006 functions on platform were that's the best way to handle exporting
13007 global variables in shared libraries. To enable this functionality,
13008 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
13009 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
13010 is normally done by Configure or something similar).
13012 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
13013 in the source file (foo.c) like this:
13015 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
13016 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
13018 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
13019 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
13021 OPENSSL_DECLARE_GLOBAL(int,foo);
13022 #define foo OPENSSL_GLOBAL_REF(foo)
13023 OPENSSL_DECLARE_GLOBAL(double,bar);
13024 #define bar OPENSSL_GLOBAL_REF(bar)
13026 The #defines are very important, and therefore so is including the
13027 header file everywhere where the defined globals are used.
13029 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
13030 of ASN.1 items, but that structure is a bit different.
13032 The largest change is in util/mkdef.pl which has been enhanced with
13033 better and easier to understand logic to choose which symbols should
13034 go into the Windows .def files as well as a number of fixes and code
13035 cleanup (among others, algorithm keywords are now sorted
13036 lexicographically to avoid constant rewrites).
13040 * In BN_div() keep a copy of the sign of 'num' before writing the
13041 result to 'rm' because if rm==num the value will be overwritten
13042 and produce the wrong result if 'num' is negative: this caused
13043 problems with BN_mod() and BN_nnmod().
13047 * Function OCSP_request_verify(). This checks the signature on an
13048 OCSP request and verifies the signer certificate. The signer
13049 certificate is just checked for a generic purpose and OCSP request
13054 * Add OCSP_check_validity() function to check the validity of OCSP
13055 responses. OCSP responses are prepared in real time and may only
13056 be a few seconds old. Simply checking that the current time lies
13057 between thisUpdate and nextUpdate max reject otherwise valid responses
13058 caused by either OCSP responder or client clock inaccuracy. Instead
13059 we allow thisUpdate and nextUpdate to fall within a certain period of
13060 the current time. The age of the response can also optionally be
13061 checked. Two new options -validity_period and -status_age added to
13066 * If signature or public key algorithm is unrecognized print out its
13067 OID rather that just UNKNOWN.
13071 * Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
13072 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
13073 ID to be generated from the issuer certificate alone which can then be
13074 passed to OCSP_id_issuer_cmp().
13078 * New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
13079 ASN1 modules to export functions returning ASN1_ITEM pointers
13080 instead of the ASN1_ITEM structures themselves. This adds several
13081 new macros which allow the underlying ASN1 function/structure to
13082 be accessed transparently. As a result code should not use ASN1_ITEM
13083 references directly (such as &X509_it) but instead use the relevant
13084 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
13085 use of the new ASN1 code on platforms where exporting structures
13086 is problematical (for example in shared libraries) but exporting
13087 functions returning pointers to structures is not.
13091 * Add support for overriding the generation of SSL/TLS session IDs.
13092 These callbacks can be registered either in an SSL_CTX or per SSL.
13093 The purpose of this is to allow applications to control, if they wish,
13094 the arbitrary values chosen for use as session IDs, particularly as it
13095 can be useful for session caching in multiple-server environments. A
13096 command-line switch for testing this (and any client code that wishes
13097 to use such a feature) has been added to "s_server".
13099 *Geoff Thorpe, Lutz Jaenicke*
13101 * Modify mkdef.pl to recognise and parse preprocessor conditionals
13102 of the form `#if defined(...) || defined(...) || ...` and
13103 `#if !defined(...) && !defined(...) && ...`. This also avoids
13104 the growing number of special cases it was previously handling.
13108 * Make all configuration macros available for application by making
13109 sure they are available in opensslconf.h, by giving them names starting
13110 with `OPENSSL_` to avoid conflicts with other packages and by making
13111 sure e_os2.h will cover all platform-specific cases together with
13113 Additionally, it is now possible to define configuration/platform-
13114 specific names (called "system identities"). In the C code, these
13115 are prefixed with `OPENSSL_SYSNAME_`. e_os2.h will create another
13116 macro with the name beginning with `OPENSSL_SYS_`, which is determined
13117 from `OPENSSL_SYSNAME_*` or compiler-specific macros depending on
13122 * New option -set_serial to 'req' and 'x509' this allows the serial
13123 number to use to be specified on the command line. Previously self
13124 signed certificates were hard coded with serial number 0 and the
13125 CA options of 'x509' had to use a serial number in a file which was
13130 * New options to 'ca' utility to support V2 CRL entry extensions.
13131 Currently CRL reason, invalidity date and hold instruction are
13132 supported. Add new CRL extensions to V3 code and some new objects.
13136 * New function EVP_CIPHER_CTX_set_padding() this is used to
13137 disable standard block padding (aka PKCS#5 padding) in the EVP
13138 API, which was previously mandatory. This means that the data is
13139 not padded in any way and so the total length much be a multiple
13140 of the block size, otherwise an error occurs.
13144 * Initial (incomplete) OCSP SSL support.
13148 * New function OCSP_parse_url(). This splits up a URL into its host,
13149 port and path components: primarily to parse OCSP URLs. New -url
13150 option to ocsp utility.
13154 * New nonce behavior. The return value of OCSP_check_nonce() now
13155 reflects the various checks performed. Applications can decide
13156 whether to tolerate certain situations such as an absent nonce
13157 in a response when one was present in a request: the ocsp application
13158 just prints out a warning. New function OCSP_add1_basic_nonce()
13159 this is to allow responders to include a nonce in a response even if
13160 the request is nonce-less.
13164 * Disable stdin buffering in `load_cert()` (`apps/apps.c`) so that no certs are
13165 skipped when using openssl x509 multiple times on a single input file,
13166 e.g. `(openssl x509 -out cert1; openssl x509 -out cert2) <certs`.
13170 * Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
13171 set string type: to handle setting ASN1_TIME structures. Fix ca
13172 utility to correctly initialize revocation date of CRLs.
13176 * New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
13177 the clients preferred ciphersuites and rather use its own preferences.
13178 Should help to work around M$ SGC (Server Gated Cryptography) bug in
13179 Internet Explorer by ensuring unchanged hash method during stepup.
13180 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
13184 * Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
13185 to aes and add a new 'exist' option to print out symbols that don't
13190 * Additional options to ocsp utility to allow flags to be set and
13191 additional certificates supplied.
13195 * Add the option -VAfile to 'openssl ocsp', so the user can give the
13196 OCSP client a number of certificate to only verify the response
13201 * Update Rijndael code to version 3.0 and change EVP AES ciphers to
13202 handle the new API. Currently only ECB, CBC modes supported. Add new
13205 Add TLS AES ciphersuites as described in RFC3268, "Advanced
13206 Encryption Standard (AES) Ciphersuites for Transport Layer
13207 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were
13208 not enabled by default and were not part of the "ALL" ciphersuite
13209 alias because they were not yet official; they could be
13210 explicitly requested by specifying the "AESdraft" ciphersuite
13211 group alias. In the final release of OpenSSL 0.9.7, the group
13212 alias is called "AES" and is part of "ALL".)
13214 *Ben Laurie, Steve Henson, Bodo Moeller*
13216 * New function OCSP_copy_nonce() to copy nonce value (if present) from
13217 request to response.
13221 * Functions for OCSP responders. OCSP_request_onereq_count(),
13222 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
13223 extract information from a certificate request. OCSP_response_create()
13224 creates a response and optionally adds a basic response structure.
13225 OCSP_basic_add1_status() adds a complete single response to a basic
13226 response and returns the OCSP_SINGLERESP structure just added (to allow
13227 extensions to be included for example). OCSP_basic_add1_cert() adds a
13228 certificate to a basic response and OCSP_basic_sign() signs a basic
13229 response with various flags. New helper functions ASN1_TIME_check()
13230 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
13231 (converts ASN1_TIME to GeneralizedTime).
13235 * Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
13236 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
13237 structure from a certificate. X509_pubkey_digest() digests the public_key
13238 contents: this is used in various key identifiers.
13242 * Make sk_sort() tolerate a NULL argument.
13244 *Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>*
13246 * New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
13247 passed by the function are trusted implicitly. If any of them signed the
13248 response then it is assumed to be valid and is not verified.
13252 * In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
13253 to data. This was previously part of the PKCS7 ASN1 code. This
13254 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
13255 *Steve Henson, reported by Kenneth R. Robinette
13256 <support@securenetterm.com>*
13258 * Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
13259 routines: without these tracing memory leaks is very painful.
13260 Fix leaks in PKCS12 and PKCS7 routines.
13264 * Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
13265 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
13266 effectively meant GeneralizedTime would never be used. Now it
13267 is initialised to -1 but X509_time_adj() now has to check the value
13268 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
13269 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
13270 *Steve Henson, reported by Kenneth R. Robinette
13271 <support@securenetterm.com>*
13273 * Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
13274 result in a zero length in the ASN1_INTEGER structure which was
13275 not consistent with the structure when d2i_ASN1_INTEGER() was used
13276 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
13277 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
13278 where it did not print out a minus for negative ASN1_INTEGER.
13282 * Add summary printout to ocsp utility. The various functions which
13283 convert status values to strings have been renamed to:
13284 OCSP_response_status_str(), OCSP_cert_status_str() and
13285 OCSP_crl_reason_str() and are no longer static. New options
13286 to verify nonce values and to disable verification. OCSP response
13287 printout format cleaned up.
13291 * Add additional OCSP certificate checks. These are those specified
13292 in RFC2560. This consists of two separate checks: the CA of the
13293 certificate being checked must either be the OCSP signer certificate
13294 or the issuer of the OCSP signer certificate. In the latter case the
13295 OCSP signer certificate must contain the OCSP signing extended key
13296 usage. This check is performed by attempting to match the OCSP
13297 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
13298 in the OCSP_CERTID structures of the response.
13302 * Initial OCSP certificate verification added to OCSP_basic_verify()
13303 and related routines. This uses the standard OpenSSL certificate
13304 verify routines to perform initial checks (just CA validity) and
13305 to obtain the certificate chain. Then additional checks will be
13306 performed on the chain. Currently the root CA is checked to see
13307 if it is explicitly trusted for OCSP signing. This is used to set
13308 a root CA as a global signing root: that is any certificate that
13309 chains to that CA is an acceptable OCSP signing certificate.
13313 * New '-extfile ...' option to 'openssl ca' for reading X.509v3
13314 extensions from a separate configuration file.
13315 As when reading extensions from the main configuration file,
13316 the '-extensions ...' option may be used for specifying the
13319 *Massimiliano Pala <madwolf@comune.modena.it>*
13321 * New OCSP utility. Allows OCSP requests to be generated or
13322 read. The request can be sent to a responder and the output
13323 parsed, outputted or printed in text form. Not complete yet:
13324 still needs to check the OCSP response validity.
13328 * New subcommands for 'openssl ca':
13329 `openssl ca -status <serial>` prints the status of the cert with
13330 the given serial number (according to the index file).
13331 `openssl ca -updatedb` updates the expiry status of certificates
13334 *Massimiliano Pala <madwolf@comune.modena.it>*
13336 * New '-newreq-nodes' command option to CA.pl. This is like
13337 '-newreq', but calls 'openssl req' with the '-nodes' option
13338 so that the resulting key is not encrypted.
13340 *Damien Miller <djm@mindrot.org>*
13342 * New configuration for the GNU Hurd.
13344 *Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte*
13346 * Initial code to implement OCSP basic response verify. This
13347 is currently incomplete. Currently just finds the signer's
13348 certificate and verifies the signature on the response.
13352 * New SSLeay_version code SSLEAY_DIR to determine the compiled-in
13353 value of OPENSSLDIR. This is available via the new '-d' option
13354 to 'openssl version', and is also included in 'openssl version -a'.
13358 * Allowing defining memory allocation callbacks that will be given
13359 file name and line number information in additional arguments
13360 (a `const char*` and an int). The basic functionality remains, as
13361 well as the original possibility to just replace malloc(),
13362 realloc() and free() by functions that do not know about these
13363 additional arguments. To register and find out the current
13364 settings for extended allocation functions, the following
13365 functions are provided:
13367 CRYPTO_set_mem_ex_functions
13368 CRYPTO_set_locked_mem_ex_functions
13369 CRYPTO_get_mem_ex_functions
13370 CRYPTO_get_locked_mem_ex_functions
13372 These work the same way as CRYPTO_set_mem_functions and friends.
13373 `CRYPTO_get_[locked_]mem_functions` now writes 0 where such an
13374 extended allocation function is enabled.
13375 Similarly, `CRYPTO_get_[locked_]mem_ex_functions` writes 0 where
13376 a conventional allocation function is enabled.
13378 *Richard Levitte, Bodo Moeller*
13380 * Finish off removing the remaining LHASH function pointer casts.
13381 There should no longer be any prototype-casting required when using
13382 the LHASH abstraction, and any casts that remain are "bugs". See
13383 the callback types and macros at the head of lhash.h for details
13384 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
13388 * Add automatic query of EGD sockets in RAND_poll() for the unix variant.
13389 If /dev/[u]random devices are not available or do not return enough
13390 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
13392 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
13393 /etc/entropy will be queried once each in this sequence, querying stops
13394 when enough entropy was collected without querying more sockets.
13398 * Change the Unix RAND_poll() variant to be able to poll several
13399 random devices, as specified by DEVRANDOM, until a sufficient amount
13400 of data has been collected. We spend at most 10 ms on each file
13401 (select timeout) and read in non-blocking mode. DEVRANDOM now
13402 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
13403 (previously it was just the string "/dev/urandom"), so on typical
13404 platforms the 10 ms delay will never occur.
13405 Also separate out the Unix variant to its own file, rand_unix.c.
13406 For VMS, there's a currently-empty rand_vms.c.
13410 * Move OCSP client related routines to ocsp_cl.c. These
13411 provide utility functions which an application needing
13412 to issue a request to an OCSP responder and analyse the
13413 response will typically need: as opposed to those which an
13414 OCSP responder itself would need which will be added later.
13416 OCSP_request_sign() signs an OCSP request with an API similar
13417 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
13418 response. OCSP_response_get1_basic() extracts basic response
13419 from response. OCSP_resp_find_status(): finds and extracts status
13420 information from an OCSP_CERTID structure (which will be created
13421 when the request structure is built). These are built from lower
13422 level functions which work on OCSP_SINGLERESP structures but
13423 won't normally be used unless the application wishes to examine
13424 extensions in the OCSP response for example.
13426 Replace nonce routines with a pair of functions.
13427 OCSP_request_add1_nonce() adds a nonce value and optionally
13428 generates a random value. OCSP_check_nonce() checks the
13429 validity of the nonce in an OCSP response.
13433 * Change function OCSP_request_add() to OCSP_request_add0_id().
13434 This doesn't copy the supplied OCSP_CERTID and avoids the
13435 need to free up the newly created id. Change return type
13436 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
13437 This can then be used to add extensions to the request.
13438 Deleted OCSP_request_new(), since most of its functionality
13439 is now in OCSP_REQUEST_new() (and the case insensitive name
13440 clash) apart from the ability to set the request name which
13441 will be added elsewhere.
13445 * Update OCSP API. Remove obsolete extensions argument from
13446 various functions. Extensions are now handled using the new
13447 OCSP extension code. New simple OCSP HTTP function which
13448 can be used to send requests and parse the response.
13452 * Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
13453 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
13454 uses the special reorder version of SET OF to sort the attributes
13455 and reorder them to match the encoded order. This resolves a long
13456 standing problem: a verify on a PKCS7 structure just after signing
13457 it used to fail because the attribute order did not match the
13458 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
13459 it uses the received order. This is necessary to tolerate some broken
13460 software that does not order SET OF. This is handled by encoding
13461 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
13462 to produce the required SET OF.
13466 * Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
13467 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
13468 files to get correct declarations of the ASN.1 item variables.
13472 * Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
13473 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
13474 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
13475 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
13476 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
13477 ASN1_ITEM and no wrapper functions.
13481 * New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
13482 replace the old function pointer based I/O routines. Change most of
13483 the `*_d2i_bio()` and `*_d2i_fp()` functions to use these.
13487 * Enhance mkdef.pl to be more accepting about spacing in C preprocessor
13488 lines, recognize more "algorithms" that can be deselected, and make
13489 it complain about algorithm deselection that isn't recognised.
13493 * New ASN1 functions to handle dup, sign, verify, digest, pack and
13494 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
13495 to use new functions. Add NO_ASN1_OLD which can be set to remove
13496 some old style ASN1 functions: this can be used to determine if old
13497 code will still work when these eventually go away.
13501 * New extension functions for OCSP structures, these follow the
13502 same conventions as certificates and CRLs.
13506 * New function X509V3_add1_i2d(). This automatically encodes and
13507 adds an extension. Its behaviour can be customised with various
13508 flags to append, replace or delete. Various wrappers added for
13509 certificates and CRLs.
13513 * Fix to avoid calling the underlying ASN1 print routine when
13514 an extension cannot be parsed. Correct a typo in the
13515 OCSP_SERVICELOC extension. Tidy up print OCSP format.
13519 * Make mkdef.pl parse some of the ASN1 macros and add appropriate
13520 entries for variables.
13524 * Add functionality to `apps/openssl.c` for detecting locking
13525 problems: As the program is single-threaded, all we have
13526 to do is register a locking callback using an array for
13527 storing which locks are currently held by the program.
13531 * Use a lock around the call to CRYPTO_get_ex_new_index() in
13532 SSL_get_ex_data_X509_STORE_idx(), which is used in
13533 ssl_verify_cert_chain() and thus can be called at any time
13534 during TLS/SSL handshakes so that thread-safety is essential.
13535 Unfortunately, the ex_data design is not at all suited
13536 for multi-threaded use, so it probably should be abolished.
13540 * Added Broadcom "ubsec" ENGINE to OpenSSL.
13542 *Broadcom, tweaked and integrated by Geoff Thorpe*
13544 * Move common extension printing code to new function
13545 X509V3_print_extensions(). Reorganise OCSP print routines and
13546 implement some needed OCSP ASN1 functions. Add OCSP extensions.
13550 * New function X509_signature_print() to remove duplication in some
13555 * Add a special meaning when SET OF and SEQUENCE OF flags are both
13556 set (this was treated exactly the same as SET OF previously). This
13557 is used to reorder the STACK representing the structure to match the
13558 encoding. This will be used to get round a problem where a PKCS7
13559 structure which was signed could not be verified because the STACK
13560 order did not reflect the encoded order.
13564 * Reimplement the OCSP ASN1 module using the new code.
13568 * Update the X509V3 code to permit the use of an ASN1_ITEM structure
13569 for its ASN1 operations. The old style function pointers still exist
13570 for now but they will eventually go away.
13574 * Merge in replacement ASN1 code from the ASN1 branch. This almost
13575 completely replaces the old ASN1 functionality with a table driven
13576 encoder and decoder which interprets an ASN1_ITEM structure describing
13577 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
13578 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
13579 has also been converted to the new form.
13583 * Change BN_mod_exp_recp so that negative moduli are tolerated
13584 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
13585 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
13586 for negative moduli.
13590 * Fix BN_uadd and BN_usub: Always return non-negative results instead
13591 of not touching the result's sign bit.
13595 * BN_div bugfix: If the result is 0, the sign (res->neg) must not be
13600 * Changed the LHASH code to use prototypes for callbacks, and created
13601 macros to declare and implement thin (optionally static) functions
13602 that provide type-safety and avoid function pointer casting for the
13603 type-specific callbacks.
13607 * Added Kerberos Cipher Suites to be used with TLS, as written in
13609 *Veers Staats <staatsvr@asc.hpc.mil>,
13610 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte*
13612 * Reformat the FAQ so the different questions and answers can be divided
13613 in sections depending on the subject.
13617 * Have the zlib compression code load ZLIB.DLL dynamically under
13622 * New function BN_mod_sqrt for computing square roots modulo a prime
13623 (using the probabilistic Tonelli-Shanks algorithm unless
13624 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
13625 be handled deterministically).
13627 *Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller*
13629 * Make BN_mod_inverse faster by explicitly handling small quotients
13630 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
13631 512 bits], about 30% for larger ones [1024 or 2048 bits].)
13635 * New function BN_kronecker.
13639 * Fix BN_gcd so that it works on negative inputs; the result is
13640 positive unless both parameters are zero.
13641 Previously something reasonably close to an infinite loop was
13642 possible because numbers could be growing instead of shrinking
13643 in the implementation of Euclid's algorithm.
13647 * Fix BN_is_word() and BN_is_one() macros to take into account the
13648 sign of the number in question.
13650 Fix BN_is_word(a,w) to work correctly for w == 0.
13652 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
13653 because its test if the absolute value of 'a' equals 'w'.
13654 Note that BN_abs_is_word does *not* handle w == 0 reliably;
13655 it exists mostly for use in the implementations of BN_is_zero(),
13656 BN_is_one(), and BN_is_word().
13660 * New function BN_swap.
13664 * Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
13665 the exponentiation functions are more likely to produce reasonable
13666 results on negative inputs.
13670 * Change BN_mod_mul so that the result is always non-negative.
13671 Previously, it could be negative if one of the factors was negative;
13672 I don't think anyone really wanted that behaviour.
13676 * Move `BN_mod_...` functions into new file `crypto/bn/bn_mod.c`
13677 (except for exponentiation, which stays in `crypto/bn/bn_exp.c`,
13678 and `BN_mod_mul_reciprocal`, which stays in `crypto/bn/bn_recp.c`)
13679 and add new functions:
13688 BN_mod_lshift1_quick
13690 BN_mod_lshift_quick
13692 These functions always generate non-negative results.
13694 `BN_nnmod` otherwise is `like BN_mod` (if `BN_mod` computes a remainder `r`
13695 such that `|m| < r < 0`, `BN_nnmod` will output `rem + |m|` instead).
13697 `BN_mod_XXX_quick(r, a, [b,] m)` generates the same result as
13698 `BN_mod_XXX(r, a, [b,] m, ctx)`, but requires that `a` [and `b`]
13699 be reduced modulo `m`.
13701 *Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller*
13704 The following entry accidentally appeared in the CHANGES file
13705 distributed with OpenSSL 0.9.7. The modifications described in
13706 it do *not* apply to OpenSSL 0.9.7.
13708 * Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
13709 was actually never needed) and in BN_mul(). The removal in BN_mul()
13710 required a small change in bn_mul_part_recursive() and the addition
13711 of the functions bn_cmp_part_words(), bn_sub_part_words() and
13712 bn_add_part_words(), which do the same thing as bn_cmp_words(),
13713 bn_sub_words() and bn_add_words() except they take arrays with
13719 * In 'openssl passwd', verify passwords read from the terminal
13720 unless the '-salt' option is used (which usually means that
13721 verification would just waste user's time since the resulting
13722 hash is going to be compared with some given password hash)
13723 or the new '-noverify' option is used.
13725 This is an incompatible change, but it does not affect
13726 non-interactive use of 'openssl passwd' (passwords on the command
13727 line, '-stdin' option, '-in ...' option) and thus should not
13728 cause any problems.
13732 * Remove all references to RSAref, since there's no more need for it.
13736 * Make DSO load along a path given through an environment variable
13737 (SHLIB_PATH) with shl_load().
13741 * Constify the ENGINE code as a result of BIGNUM constification.
13742 Also constify the RSA code and most things related to it. In a
13743 few places, most notable in the depth of the ASN.1 code, ugly
13744 casts back to non-const were required (to be solved at a later
13749 * Make it so the openssl application has all engines loaded by default.
13753 * Constify the BIGNUM routines a little more.
13757 * Add the following functions:
13759 ENGINE_load_cswift()
13761 ENGINE_load_atalla()
13762 ENGINE_load_nuron()
13763 ENGINE_load_builtin_engines()
13765 That way, an application can itself choose if external engines that
13766 are built-in in OpenSSL shall ever be used or not. The benefit is
13767 that applications won't have to be linked with libdl or other dso
13768 libraries unless it's really needed.
13770 Changed 'openssl engine' to load all engines on demand.
13771 Changed the engine header files to avoid the duplication of some
13772 declarations (they differed!).
13776 * 'openssl engine' can now list capabilities.
13780 * Better error reporting in 'openssl engine'.
13784 * Never call load_dh_param(NULL) in s_server.
13788 * Add engine application. It can currently list engines by name and
13789 identity, and test if they are actually available.
13793 * Improve RPM specification file by forcing symbolic linking and making
13794 sure the installed documentation is also owned by root.root.
13796 *Damien Miller <djm@mindrot.org>*
13798 * Give the OpenSSL applications more possibilities to make use of
13799 keys (public as well as private) handled by engines.
13803 * Add OCSP code that comes from CertCo.
13807 * Add VMS support for the Rijndael code.
13811 * Added untested support for Nuron crypto accelerator.
13815 * Add support for external cryptographic devices. This code was
13816 previously distributed separately as the "engine" branch.
13818 *Geoff Thorpe, Richard Levitte*
13820 * Rework the filename-translation in the DSO code. It is now possible to
13821 have far greater control over how a "name" is turned into a filename
13822 depending on the operating environment and any oddities about the
13823 different shared library filenames on each system.
13827 * Support threads on FreeBSD-elf in Configure.
13831 * Fix for SHA1 assembly problem with MASM: it produces
13832 warnings about corrupt line number information when assembling
13833 with debugging information. This is caused by the overlapping
13836 *Bernd Matthes <mainbug@celocom.de>, Steve Henson*
13839 NCONF_get_number() has no error checking at all. As a replacement,
13840 NCONF_get_number_e() is defined (`_e` for "error checking") and is
13841 promoted strongly. The old NCONF_get_number is kept around for
13842 binary backward compatibility.
13843 Make it possible for methods to load from something other than a BIO,
13844 by providing a function pointer that is given a name instead of a BIO.
13845 For example, this could be used to load configuration data from an
13850 * Fix for non blocking accept BIOs. Added new I/O special reason
13851 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
13852 with non blocking I/O was not possible because no retry code was
13853 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
13858 * Added the beginnings of Rijndael support.
13862 * Fix for bug in DirectoryString mask setting. Add support for
13863 X509_NAME_print_ex() in 'req' and X509_print_ex() function
13864 to allow certificate printing to more controllable, additional
13865 'certopt' option to 'x509' to allow new printing options to be
13870 * Clean old EAY MD5 hack from e_os.h.
13874 ### Changes between 0.9.6l and 0.9.6m [17 Mar 2004]
13876 * Fix null-pointer assignment in do_change_cipher_spec() revealed
13877 by using the Codenomicon TLS Test Tool ([CVE-2004-0079])
13879 *Joe Orton, Steve Henson*
13881 ### Changes between 0.9.6k and 0.9.6l [04 Nov 2003]
13883 * Fix additional bug revealed by the NISCC test suite:
13885 Stop bug triggering large recursion when presented with
13886 certain ASN.1 tags ([CVE-2003-0851])
13890 ### Changes between 0.9.6j and 0.9.6k [30 Sep 2003]
13892 * Fix various bugs revealed by running the NISCC test suite:
13894 Stop out of bounds reads in the ASN1 code when presented with
13895 invalid tags (CVE-2003-0543 and CVE-2003-0544).
13897 If verify callback ignores invalid public key errors don't try to check
13898 certificate signature with the NULL public key.
13902 * In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
13903 if the server requested one: as stated in TLS 1.0 and SSL 3.0
13908 * In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
13909 extra data after the compression methods not only for TLS 1.0
13910 but also for SSL 3.0 (as required by the specification).
13912 *Bodo Moeller; problem pointed out by Matthias Loepfe*
13914 * Change X509_certificate_type() to mark the key as exported/exportable
13915 when it's 512 *bits* long, not 512 bytes.
13919 ### Changes between 0.9.6i and 0.9.6j [10 Apr 2003]
13921 * Countermeasure against the Klima-Pokorny-Rosa extension of
13922 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
13923 a protocol version number mismatch like a decryption error
13924 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
13928 * Turn on RSA blinding by default in the default implementation
13929 to avoid a timing attack. Applications that don't want it can call
13930 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
13931 They would be ill-advised to do so in most cases.
13933 *Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller*
13935 * Change RSA blinding code so that it works when the PRNG is not
13936 seeded (in this case, the secret RSA exponent is abused as
13937 an unpredictable seed -- if it is not unpredictable, there
13938 is no point in blinding anyway). Make RSA blinding thread-safe
13939 by remembering the creator's thread ID in rsa->blinding and
13940 having all other threads use local one-time blinding factors
13941 (this requires more computation than sharing rsa->blinding, but
13942 avoids excessive locking; and if an RSA object is not shared
13943 between threads, blinding will still be very fast).
13947 ### Changes between 0.9.6h and 0.9.6i [19 Feb 2003]
13949 * In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
13950 via timing by performing a MAC computation even if incorrect
13951 block cipher padding has been found. This is a countermeasure
13952 against active attacks where the attacker has to distinguish
13953 between bad padding and a MAC verification error. ([CVE-2003-0078])
13955 *Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
13956 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
13957 Martin Vuagnoux (EPFL, Ilion)*
13959 ### Changes between 0.9.6g and 0.9.6h [5 Dec 2002]
13961 * New function OPENSSL_cleanse(), which is used to cleanse a section of
13962 memory from its contents. This is done with a counter that will
13963 place alternating values in each byte. This can be used to solve
13964 two issues: 1) the removal of calls to memset() by highly optimizing
13965 compilers, and 2) cleansing with other values than 0, since those can
13966 be read through on certain media, for example a swap space on disk.
13970 * Bugfix: client side session caching did not work with external caching,
13971 because the session->cipher setting was not restored when reloading
13972 from the external cache. This problem was masked, when
13973 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set.
13974 (Found by Steve Haslam <steve@araqnid.ddts.net>.)
13978 * Fix client_certificate (ssl/s2_clnt.c): The permissible total
13979 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33.
13981 *Zeev Lieber <zeev-l@yahoo.com>*
13983 * Undo an undocumented change introduced in 0.9.6e which caused
13984 repeated calls to OpenSSL_add_all_ciphers() and
13985 OpenSSL_add_all_digests() to be ignored, even after calling
13990 * Change the default configuration reader to deal with last line not
13991 being properly terminated.
13995 * Change X509_NAME_cmp() so it applies the special rules on handling
13996 DN values that are of type PrintableString, as well as RDNs of type
13997 emailAddress where the value has the type ia5String.
13999 *stefank@valicert.com via Richard Levitte*
14001 * Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half
14002 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently
14003 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be
14004 the bitwise-OR of the two for use by the majority of applications
14005 wanting this behaviour, and update the docs. The documented
14006 behaviour and actual behaviour were inconsistent and had been
14007 changing anyway, so this is more a bug-fix than a behavioural
14010 *Geoff Thorpe, diagnosed by Nadav Har'El*
14012 * Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c
14013 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes).
14017 * Fix initialization code race conditions in
14018 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(),
14019 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(),
14020 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(),
14021 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(),
14022 ssl2_get_cipher_by_char(),
14023 ssl3_get_cipher_by_char().
14025 *Patrick McCormick <patrick@tellme.com>, Bodo Moeller*
14027 * Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after
14028 the cached sessions are flushed, as the remove_cb() might use ex_data
14029 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com>
14030 (see [openssl.org #212]).
14032 *Geoff Thorpe, Lutz Jaenicke*
14034 * Fix typo in OBJ_txt2obj which incorrectly passed the content
14035 length, instead of the encoding length to d2i_ASN1_OBJECT.
14039 ### Changes between 0.9.6f and 0.9.6g [9 Aug 2002]
14041 * [In 0.9.6g-engine release:]
14042 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use `_stdcall`).
14044 *Lynn Gazis <lgazis@rainbow.com>*
14046 ### Changes between 0.9.6e and 0.9.6f [8 Aug 2002]
14048 * Fix ASN1 checks. Check for overflow by comparing with LONG_MAX
14049 and get fix the header length calculation.
14050 *Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>,
14051 Alon Kantor <alonk@checkpoint.com> (and others), Steve Henson*
14053 * Use proper error handling instead of 'assertions' in buffer
14054 overflow checks added in 0.9.6e. This prevents DoS (the
14055 assertions could call abort()).
14057 *Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller*
14059 ### Changes between 0.9.6d and 0.9.6e [30 Jul 2002]
14061 * Add various sanity checks to asn1_get_length() to reject
14062 the ASN1 length bytes if they exceed sizeof(long), will appear
14063 negative or the content length exceeds the length of the
14066 *Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>*
14068 * Fix cipher selection routines: ciphers without encryption had no flags
14069 for the cipher strength set and where therefore not handled correctly
14070 by the selection routines (PR #130).
14074 * Fix EVP_dsa_sha macro.
14079 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS
14080 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure
14081 that was added in OpenSSL 0.9.6d.
14083 As the countermeasure turned out to be incompatible with some
14084 broken SSL implementations, the new option is part of SSL_OP_ALL.
14085 SSL_OP_ALL is usually employed when compatibility with weird SSL
14086 implementations is desired (e.g. '-bugs' option to 's_client' and
14087 's_server'), so the new option is automatically set in many
14092 * Changes in security patch:
14094 Changes marked "(CHATS)" were sponsored by the Defense Advanced
14095 Research Projects Agency (DARPA) and Air Force Research Laboratory,
14096 Air Force Materiel Command, USAF, under agreement number
14099 * Add various sanity checks to asn1_get_length() to reject
14100 the ASN1 length bytes if they exceed sizeof(long), will appear
14101 negative or the content length exceeds the length of the
14102 supplied buffer. ([CVE-2002-0659])
14104 *Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>*
14106 * Assertions for various potential buffer overflows, not known to
14107 happen in practice.
14109 *Ben Laurie (CHATS)*
14111 * Various temporary buffers to hold ASCII versions of integers were
14112 too small for 64 bit platforms. ([CVE-2002-0655])
14113 *Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>*
14115 * Remote buffer overflow in SSL3 protocol - an attacker could
14116 supply an oversized session ID to a client. ([CVE-2002-0656])
14118 *Ben Laurie (CHATS)*
14120 * Remote buffer overflow in SSL2 protocol - an attacker could
14121 supply an oversized client master key. ([CVE-2002-0656])
14123 *Ben Laurie (CHATS)*
14125 ### Changes between 0.9.6c and 0.9.6d [9 May 2002]
14127 * Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
14128 encoded as NULL) with id-dsa-with-sha1.
14130 *Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller*
14132 * Check various `X509_...()` return values in `apps/req.c`.
14134 *Nils Larsch <nla@trustcenter.de>*
14136 * Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
14137 an end-of-file condition would erroneously be flagged, when the CRLF
14138 was just at the end of a processed block. The bug was discovered when
14139 processing data through a buffering memory BIO handing the data to a
14140 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
14141 <ptsekov@syntrex.com> and Nedelcho Stanev.
14145 * Implement a countermeasure against a vulnerability recently found
14146 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
14147 before application data chunks to avoid the use of known IVs
14148 with data potentially chosen by the attacker.
14152 * Fix length checks in ssl3_get_client_hello().
14156 * TLS/SSL library bugfix: use s->s3->in_read_app_data differently
14157 to prevent ssl3_read_internal() from incorrectly assuming that
14158 ssl3_read_bytes() found application data while handshake
14159 processing was enabled when in fact s->s3->in_read_app_data was
14160 merely automatically cleared during the initial handshake.
14162 *Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>*
14164 * Fix object definitions for Private and Enterprise: they were not
14165 recognized in their shortname (=lowercase) representation. Extend
14166 obj_dat.pl to issue an error when using undefined keywords instead
14167 of silently ignoring the problem (Svenning Sorensen
14168 <sss@sss.dnsalias.net>).
14172 * Fix DH_generate_parameters() so that it works for 'non-standard'
14173 generators, i.e. generators other than 2 and 5. (Previously, the
14174 code did not properly initialise the 'add' and 'rem' values to
14175 BN_generate_prime().)
14177 In the new general case, we do not insist that 'generator' is
14178 actually a primitive root: This requirement is rather pointless;
14179 a generator of the order-q subgroup is just as good, if not
14184 * Map new X509 verification errors to alerts. Discovered and submitted by
14185 Tom Wu <tom@arcot.com>.
14189 * Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
14190 returning non-zero before the data has been completely received
14191 when using non-blocking I/O.
14193 *Bodo Moeller; problem pointed out by John Hughes*
14195 * Some of the ciphers missed the strength entry (SSL_LOW etc).
14197 *Ben Laurie, Lutz Jaenicke*
14199 * Fix bug in SSL_clear(): bad sessions were not removed (found by
14200 Yoram Zahavi <YoramZ@gilian.com>).
14204 * Add information about CygWin 1.3 and on, and preserve proper
14205 configuration for the versions before that.
14207 *Corinna Vinschen <vinschen@redhat.com> and Richard Levitte*
14209 * Make removal from session cache (SSL_CTX_remove_session()) more robust:
14210 check whether we deal with a copy of a session and do not delete from
14211 the cache in this case. Problem reported by "Izhar Shoshani Levi"
14212 <izhar@checkpoint.com>.
14216 * Do not store session data into the internal session cache, if it
14217 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
14218 flag is set). Proposed by Aslam <aslam@funk.com>.
14222 * Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
14227 * [In 0.9.6d-engine release:]
14228 Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
14230 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
14232 * Add the configuration target linux-s390x.
14234 *Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte*
14236 * The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
14237 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
14238 variable as an indication that a ClientHello message has been
14239 received. As the flag value will be lost between multiple
14240 invocations of ssl3_accept when using non-blocking I/O, the
14241 function may not be aware that a handshake has actually taken
14242 place, thus preventing a new session from being added to the
14245 To avoid this problem, we now set s->new_session to 2 instead of
14246 using a local variable.
14248 *Lutz Jaenicke, Bodo Moeller*
14250 * Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
14251 if the SSL_R_LENGTH_MISMATCH error is detected.
14253 *Geoff Thorpe, Bodo Moeller*
14255 * New 'shared_ldflag' column in Configure platform table.
14259 * Fix EVP_CIPHER_mode macro.
14261 *"Dan S. Camper" <dan@bti.net>*
14263 * Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
14264 type, we must throw them away by setting rr->length to 0.
14266 *D P Chang <dpc@qualys.com>*
14268 ### Changes between 0.9.6b and 0.9.6c [21 dec 2001]
14270 * Fix BN_rand_range bug pointed out by Dominikus Scherkl
14271 <Dominikus.Scherkl@biodata.com>. (The previous implementation
14272 worked incorrectly for those cases where range = `10..._2` and
14273 `3*range` is two bits longer than range.)
14277 * Only add signing time to PKCS7 structures if it is not already
14282 * Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
14283 OBJ_ld_ce should be OBJ_id_ce.
14284 Also some ip-pda OIDs in crypto/objects/objects.txt were
14285 incorrect (cf. RFC 3039).
14287 *Matt Cooper, Frederic Giudicelli, Bodo Moeller*
14289 * Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
14290 returns early because it has nothing to do.
14292 *Andy Schneider <andy.schneider@bjss.co.uk>*
14294 * [In 0.9.6c-engine release:]
14295 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
14297 *Andy Schneider <andy.schneider@bjss.co.uk>*
14299 * [In 0.9.6c-engine release:]
14300 Add support for Cryptographic Appliance's keyserver technology.
14301 (Use engine 'keyclient')
14303 *Cryptographic Appliances and Geoff Thorpe*
14305 * Add a configuration entry for OS/390 Unix. The C compiler 'c89'
14306 is called via tools/c89.sh because arguments have to be
14307 rearranged (all '-L' options must appear before the first object
14310 *Richard Shapiro <rshapiro@abinitio.com>*
14312 * [In 0.9.6c-engine release:]
14313 Add support for Broadcom crypto accelerator cards, backported
14316 *Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox*
14318 * [In 0.9.6c-engine release:]
14319 Add support for SureWare crypto accelerator cards from
14320 Baltimore Technologies. (Use engine 'sureware')
14322 *Baltimore Technologies and Mark Cox*
14324 * [In 0.9.6c-engine release:]
14325 Add support for crypto accelerator cards from Accelerated
14326 Encryption Processing, www.aep.ie. (Use engine 'aep')
14328 *AEP Inc. and Mark Cox*
14330 * Add a configuration entry for gcc on UnixWare.
14332 *Gary Benson <gbenson@redhat.com>*
14334 * Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
14335 messages are stored in a single piece (fixed-length part and
14336 variable-length part combined) and fix various bugs found on the way.
14340 * Disable caching in BIO_gethostbyname(), directly use gethostbyname()
14341 instead. BIO_gethostbyname() does not know what timeouts are
14342 appropriate, so entries would stay in cache even when they have
14344 *Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>*
14346 * Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
14347 faced with a pathologically small ClientHello fragment that does
14348 not contain client_version: Instead of aborting with an error,
14349 simply choose the highest available protocol version (i.e.,
14350 TLS 1.0 unless it is disabled). In practice, ClientHello
14351 messages are never sent like this, but this change gives us
14352 strictly correct behaviour at least for TLS.
14356 * Fix SSL handshake functions and SSL_clear() such that SSL_clear()
14357 never resets s->method to s->ctx->method when called from within
14358 one of the SSL handshake functions.
14360 *Bodo Moeller; problem pointed out by Niko Baric*
14362 * In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
14363 (sent using the client's version number) if client_version is
14364 smaller than the protocol version in use. Also change
14365 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
14366 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
14367 the client will at least see that alert.
14371 * Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
14376 * Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
14377 client receives HelloRequest while in a handshake.
14379 *Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>*
14381 * Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
14382 should end in 'break', not 'goto end' which circumvents various
14383 cleanups done in state SSL_ST_OK. But session related stuff
14384 must be disabled for SSL_ST_OK in the case that we just sent a
14387 Also avoid some overhead by not calling ssl_init_wbio_buffer()
14388 before just sending a HelloRequest.
14390 *Bodo Moeller, Eric Rescorla <ekr@rtfm.com>*
14392 * Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
14393 reveal whether illegal block cipher padding was found or a MAC
14394 verification error occurred. (Neither SSLerr() codes nor alerts
14395 are directly visible to potential attackers, but the information
14396 may leak via logfiles.)
14398 Similar changes are not required for the SSL 2.0 implementation
14399 because the number of padding bytes is sent in clear for SSL 2.0,
14400 and the extra bytes are just ignored. However ssl/s2_pkt.c
14401 failed to verify that the purported number of padding bytes is in
14406 * Add OpenUNIX-8 support including shared libraries
14407 (Boyd Lynn Gerber <gerberb@zenez.com>).
14411 * Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
14412 'wristwatch attack' using huge encoding parameters (cf.
14413 James H. Manger's CRYPTO 2001 paper). Note that the
14414 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
14415 encoding parameters and hence was not vulnerable.
14419 * BN_sqr() bug fix.
14421 *Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>*
14423 * Rabin-Miller test analyses assume uniformly distributed witnesses,
14424 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
14425 followed by modular reduction.
14427 *Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>*
14429 * Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
14430 equivalent based on BN_pseudo_rand() instead of BN_rand().
14434 * s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
14435 This function was broken, as the check for a new client hello message
14436 to handle SGC did not allow these large messages.
14437 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
14441 * Add alert descriptions for TLSv1 to `SSL_alert_desc_string[_long]()`.
14445 * Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
14446 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
14450 * Rework the configuration and shared library support for Tru64 Unix.
14451 The configuration part makes use of modern compiler features and
14452 still retains old compiler behavior for those that run older versions
14453 of the OS. The shared library support part includes a variant that
14454 uses the RPATH feature, and is available through the special
14455 configuration target "alpha-cc-rpath", which will never be selected
14458 *Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte*
14460 * In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
14461 with the same message size as in ssl3_get_certificate_request().
14462 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
14463 messages might inadvertently be reject as too long.
14465 *Petr Lampa <lampa@fee.vutbr.cz>*
14467 * Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
14471 * Modified SSL library such that the verify_callback that has been set
14472 specifically for an SSL object with SSL_set_verify() is actually being
14473 used. Before the change, a verify_callback set with this function was
14474 ignored and the verify_callback() set in the SSL_CTX at the time of
14475 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
14476 to allow the necessary settings.
14480 * Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
14481 explicitly to NULL, as at least on Solaris 8 this seems not always to be
14482 done automatically (in contradiction to the requirements of the C
14483 standard). This made problems when used from OpenSSH.
14487 * In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
14488 dh->length and always used
14490 BN_rand_range(priv_key, dh->p).
14492 BN_rand_range() is not necessary for Diffie-Hellman, and this
14493 specific range makes Diffie-Hellman unnecessarily inefficient if
14494 dh->length (recommended exponent length) is much smaller than the
14495 length of dh->p. We could use BN_rand_range() if the order of
14496 the subgroup was stored in the DH structure, but we only have
14501 BN_rand(priv_key, l, ...)
14503 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
14510 RSA_eay_public_encrypt
14511 RSA_eay_private_decrypt
14512 RSA_eay_private_encrypt (signing)
14513 RSA_eay_public_decrypt (signature verification)
14515 (default implementations for RSA_public_encrypt,
14516 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
14517 always reject numbers >= n.
14521 * In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
14522 to synchronize access to 'locking_thread'. This is necessary on
14523 systems where access to 'locking_thread' (an 'unsigned long'
14524 variable) is not atomic.
14528 * In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
14529 *before* setting the 'crypto_lock_rand' flag. The previous code had
14530 a race condition if 0 is a valid thread ID.
14532 *Travis Vitek <vitek@roguewave.com>*
14534 * Add support for shared libraries under Irix.
14536 *Albert Chin-A-Young <china@thewrittenword.com>*
14538 * Add configuration option to build on Linux on both big-endian and
14539 little-endian MIPS.
14541 *Ralf Baechle <ralf@uni-koblenz.de>*
14543 * Add the possibility to create shared libraries on HP-UX.
14547 ### Changes between 0.9.6a and 0.9.6b [9 Jul 2001]
14549 * Change ssleay_rand_bytes (crypto/rand/md_rand.c)
14550 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
14551 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
14552 PRNG state recovery was possible based on the output of
14553 one PRNG request appropriately sized to gain knowledge on
14554 'md' followed by enough consecutive 1-byte PRNG requests
14555 to traverse all of 'state'.
14557 1. When updating 'md_local' (the current thread's copy of 'md')
14558 during PRNG output generation, hash all of the previous
14559 'md_local' value, not just the half used for PRNG output.
14561 2. Make the number of bytes from 'state' included into the hash
14562 independent from the number of PRNG bytes requested.
14564 The first measure alone would be sufficient to avoid
14565 Markku-Juhani's attack. (Actually it had never occurred
14566 to me that the half of 'md_local' used for chaining was the
14567 half from which PRNG output bytes were taken -- I had always
14568 assumed that the secret half would be used.) The second
14569 measure makes sure that additional data from 'state' is never
14570 mixed into 'md_local' in small portions; this heuristically
14571 further strengthens the PRNG.
14575 * Fix crypto/bn/asm/mips3.s.
14579 * When only the key is given to "enc", the IV is undefined. Print out
14580 an error message in this case.
14584 * Handle special case when X509_NAME is empty in X509 printing routines.
14588 * In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
14589 positive and less than q.
14593 * Don't change `*pointer` in CRYPTO_add_lock() is add_lock_callback is
14594 used: it isn't thread safe and the add_lock_callback should handle
14597 *Paul Rose <Paul.Rose@bridge.com>*
14599 * Verify that incoming data obeys the block size in
14600 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
14606 *Ulf Möller, Bodo Möller*
14608 * The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
14609 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
14610 when fixing the server behaviour for backwards-compatible 'client
14611 hello' messages. (Note that the attack is impractical against
14612 SSL 3.0 and TLS 1.0 anyway because length and version checking
14613 means that the probability of guessing a valid ciphertext is
14614 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
14617 Before 0.9.5, the countermeasure (hide the error by generating a
14618 random 'decryption result') did not work properly because
14619 ERR_clear_error() was missing, meaning that SSL_get_error() would
14620 detect the supposedly ignored error.
14622 Both problems are now fixed.
14626 * In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
14627 (previously it was 1024).
14631 * Fix for compatibility mode trust settings: ignore trust settings
14632 unless some valid trust or reject settings are present.
14636 * Fix for blowfish EVP: its a variable length cipher.
14640 * Fix various bugs related to DSA S/MIME verification. Handle missing
14641 parameters in DSA public key structures and return an error in the
14642 DSA routines if parameters are absent.
14646 * In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
14647 in the current directory if neither $RANDFILE nor $HOME was set.
14648 RAND_file_name() in 0.9.6a returned NULL in this case. This has
14649 caused some confusion to Windows users who haven't defined $HOME.
14650 Thus RAND_file_name() is changed again: e_os.h can define a
14651 DEFAULT_HOME, which will be used if $HOME is not set.
14652 For Windows, we use "C:"; on other platforms, we still require
14653 environment variables.
14655 * Move 'if (!initialized) RAND_poll()' into regions protected by
14656 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
14657 having multiple threads call RAND_poll() concurrently.
14661 * In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
14662 combination of a flag and a thread ID variable.
14663 Otherwise while one thread is in ssleay_rand_bytes (which sets the
14664 flag), *other* threads can enter ssleay_add_bytes without obeying
14665 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
14666 that they do not hold after the first thread unsets add_do_not_lock).
14670 * Change bctest again: '-x' expressions are not available in all
14671 versions of 'test'.
14675 ### Changes between 0.9.6 and 0.9.6a [5 Apr 2001]
14677 * Fix a couple of memory leaks in PKCS7_dataDecode()
14679 *Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>*
14681 * Change Configure and Makefiles to provide EXE_EXT, which will contain
14682 the default extension for executables, if any. Also, make the perl
14683 scripts that use symlink() to test if it really exists and use "cp"
14684 if it doesn't. All this made OpenSSL compilable and installable in
14689 * Fix for asn1_GetSequence() for indefinite length constructed data.
14690 If SEQUENCE is length is indefinite just set c->slen to the total
14691 amount of data available.
14693 *Steve Henson, reported by shige@FreeBSD.org*
14695 *This change does not apply to 0.9.7.*
14697 * Change bctest to avoid here-documents inside command substitution
14698 (workaround for FreeBSD /bin/sh bug).
14699 For compatibility with Ultrix, avoid shell functions (introduced
14700 in the bctest version that searches along $PATH).
14704 * Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
14705 with des_encrypt() defined on some operating systems, like Solaris
14710 * Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
14711 On the Importance of Eliminating Errors in Cryptographic
14712 Computations, J. Cryptology 14 (2001) 2, 101-119,
14713 <http://theory.stanford.edu/~dabo/papers/faults.ps.gz>).
14717 * MIPS assembler BIGNUM division bug fix.
14721 * Disabled incorrect Alpha assembler code.
14725 * Fix PKCS#7 decode routines so they correctly update the length
14726 after reading an EOC for the EXPLICIT tag.
14730 *This change does not apply to 0.9.7.*
14732 * Fix bug in PKCS#12 key generation routines. This was triggered
14733 if a 3DES key was generated with a 0 initial byte. Include
14734 PKCS12_BROKEN_KEYGEN compilation option to retain the old
14735 (but broken) behaviour.
14739 * Enhance bctest to search for a working bc along $PATH and print
14742 *Tim Rice <tim@multitalents.net> via Richard Levitte*
14744 * Fix memory leaks in err.c: free err_data string if necessary;
14745 don't write to the wrong index in ERR_set_error_data.
14749 * Implement ssl23_peek (analogous to ssl23_read), which previously
14754 * Replace rdtsc with `_emit` statements for VC++ version 5.
14756 *Jeremy Cooper <jeremy@baymoo.org>*
14758 * Make it possible to reuse SSLv2 sessions.
14762 * In copy_email() check for >= 0 as a return value for
14763 X509_NAME_get_index_by_NID() since 0 is a valid index.
14765 *Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>*
14767 * Avoid coredump with unsupported or invalid public keys by checking if
14768 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
14769 PKCS7_verify() fails with non detached data.
14773 * Don't use getenv in library functions when run as setuid/setgid.
14774 New function OPENSSL_issetugid().
14778 * Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
14779 due to incorrect handling of multi-threading:
14781 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
14783 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
14785 3. Count how many times MemCheck_off() has been called so that
14786 nested use can be treated correctly. This also avoids
14787 inband-signalling in the previous code (which relied on the
14788 assumption that thread ID 0 is impossible).
14792 * Add "-rand" option also to s_client and s_server.
14796 * Fix CPU detection on Irix 6.x.
14797 *Kurt Hockenbury <khockenb@stevens-tech.edu> and
14798 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>*
14800 * Fix X509_NAME bug which produced incorrect encoding if X509_NAME
14805 *This change does not apply to 0.9.7.*
14807 * Use the cached encoding of an X509_NAME structure rather than
14808 copying it. This is apparently the reason for the libsafe "errors"
14809 but the code is actually correct.
14813 * Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
14814 Bleichenbacher's DSA attack.
14815 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
14816 to be set and top=0 forces the highest bit to be set; top=-1 is new
14817 and leaves the highest bit random.
14819 *Ulf Moeller, Bodo Moeller*
14821 * In the `NCONF_...`-based implementations for `CONF_...` queries
14822 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
14823 a temporary CONF structure with the data component set to NULL
14824 (which gives segmentation faults in lh_retrieve).
14825 Instead, use NULL for the CONF pointer in CONF_get_string and
14826 CONF_get_number (which may use environment variables) and directly
14827 return NULL from CONF_get_section.
14831 * Fix potential buffer overrun for EBCDIC.
14835 * Tolerate nonRepudiation as being valid for S/MIME signing and certSign
14836 keyUsage if basicConstraints absent for a CA.
14840 * Make SMIME_write_PKCS7() write mail header values with a format that
14841 is more generally accepted (no spaces before the semicolon), since
14842 some programs can't parse those values properly otherwise. Also make
14843 sure BIO's that break lines after each write do not create invalid
14848 * Make the CRL encoding routines work with empty SEQUENCE OF. The
14849 macros previously used would not encode an empty SEQUENCE OF
14850 and break the signature.
14854 *This change does not apply to 0.9.7.*
14856 * Zero the premaster secret after deriving the master secret in
14861 * Add some EVP_add_digest_alias registrations (as found in
14862 OpenSSL_add_all_digests()) to SSL_library_init()
14863 aka OpenSSL_add_ssl_algorithms(). This provides improved
14864 compatibility with peers using X.509 certificates
14865 with unconventional AlgorithmIdentifier OIDs.
14869 * Fix for Irix with NO_ASM.
14871 *"Bruce W. Forsberg" <bruce.forsberg@baesystems.com>*
14873 * ./config script fixes.
14875 *Ulf Moeller, Richard Levitte*
14877 * Fix 'openssl passwd -1'.
14881 * Change PKCS12_key_gen_asc() so it can cope with non null
14882 terminated strings whose length is passed in the passlen
14883 parameter, for example from PEM callbacks. This was done
14884 by adding an extra length parameter to asc2uni().
14886 *Steve Henson, reported by <oddissey@samsung.co.kr>*
14888 * Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
14889 call failed, free the DSA structure.
14893 * Fix to uni2asc() to cope with zero length Unicode strings.
14894 These are present in some PKCS#12 files.
14898 * Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
14899 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
14900 when writing a 32767 byte record.
14902 *Bodo Moeller; problem reported by Eric Day <eday@concentric.net>*
14904 * In `RSA_eay_public_{en,ed}crypt` and RSA_eay_mod_exp (rsa_eay.c),
14905 obtain lock CRYPTO_LOCK_RSA before setting `rsa->_method_mod_{n,p,q}`.
14907 (RSA objects have a reference count access to which is protected
14908 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
14909 so they are meant to be shared between threads.)
14910 *Bodo Moeller, Geoff Thorpe; original patch submitted by
14911 "Reddie, Steven" <Steven.Reddie@ca.com>*
14913 * Fix a deadlock in CRYPTO_mem_leaks().
14917 * Use better test patterns in bntest.
14921 * rand_win.c fix for Borland C.
14925 * BN_rshift bugfix for n == 0.
14929 * Add a 'bctest' script that checks for some known 'bc' bugs
14930 so that 'make test' does not abort just because 'bc' is broken.
14934 * Store verify_result within SSL_SESSION also for client side to
14935 avoid potential security hole. (Re-used sessions on the client side
14936 always resulted in verify_result==X509_V_OK, not using the original
14937 result of the server certificate verification.)
14941 * Fix ssl3_pending: If the record in s->s3->rrec is not of type
14942 SSL3_RT_APPLICATION_DATA, return 0.
14943 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
14948 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
14949 releases, have been re-implemented by renaming the previous
14950 implementations of ssl2_read and ssl3_read to ssl2_read_internal
14951 and ssl3_read_internal, respectively, and adding 'peek' parameters
14952 to them. The new ssl[23]_{read,peek} functions are calls to
14953 ssl[23]_read_internal with the 'peek' flag set appropriately.
14954 A 'peek' parameter has also been added to ssl3_read_bytes, which
14955 does the actual work for ssl3_read_internal.
14959 * Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
14960 the method-specific "init()" handler. Also clean up ex_data after
14961 calling the method-specific "finish()" handler. Previously, this was
14962 happening the other way round.
14966 * Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
14967 The previous value, 12, was not always sufficient for BN_mod_exp().
14971 * Make sure that shared libraries get the internal name engine with
14972 the full version number and not just 0. This should mark the
14973 shared libraries as not backward compatible. Of course, this should
14974 be changed again when we can guarantee backward binary compatibility.
14978 * Fix typo in get_cert_by_subject() in by_dir.c
14980 *Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>*
14982 * Rework the system to generate shared libraries:
14984 - Make note of the expected extension for the shared libraries and
14985 if there is a need for symbolic links from for example libcrypto.so.0
14986 to libcrypto.so.0.9.7. There is extended info in Configure for
14989 - Make as few rebuilds of the shared libraries as possible.
14991 - Still avoid linking the OpenSSL programs with the shared libraries.
14993 - When installing, install the shared libraries separately from the
14998 * Fix SSL_CTX_set_read_ahead macro to actually use its argument.
15000 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
15001 and not in SSL_clear because the latter is also used by the
15002 accept/connect functions; previously, the settings made by
15003 SSL_set_read_ahead would be lost during the handshake.
15005 *Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>*
15007 * Correct util/mkdef.pl to be selective about disabled algorithms.
15008 Previously, it would create entries for disabled algorithms no
15013 * Added several new manual pages for SSL_* function.
15017 ### Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
15019 * In ssl23_get_client_hello, generate an error message when faced
15020 with an initial SSL 3.0/TLS record that is too small to contain the
15021 first two bytes of the ClientHello message, i.e. client_version.
15022 (Note that this is a pathologic case that probably has never happened
15023 in real life.) The previous approach was to use the version number
15024 from the record header as a substitute; but our protocol choice
15025 should not depend on that one because it is not authenticated
15026 by the Finished messages.
15030 * More robust randomness gathering functions for Windows.
15032 *Jeffrey Altman <jaltman@columbia.edu>*
15034 * For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
15035 not set then we don't setup the error code for issuer check errors
15036 to avoid possibly overwriting other errors which the callback does
15037 handle. If an application does set the flag then we assume it knows
15038 what it is doing and can handle the new informational codes
15043 * Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
15044 a general "ANY" type, as such it should be able to decode anything
15045 including tagged types. However it didn't check the class so it would
15046 wrongly interpret tagged types in the same way as their universal
15047 counterpart and unknown types were just rejected. Changed so that the
15048 tagged and unknown types are handled in the same way as a SEQUENCE:
15049 that is the encoding is stored intact. There is also a new type
15050 "V_ASN1_OTHER" which is used when the class is not universal, in this
15051 case we have no idea what the actual type is so we just lump them all
15056 * On VMS, stdout may very well lead to a file that is written to
15057 in a record-oriented fashion. That means that every write() will
15058 write a separate record, which will be read separately by the
15059 programs trying to read from it. This can be very confusing.
15061 The solution is to put a BIO filter in the way that will buffer
15062 text until a linefeed is reached, and then write everything a
15063 line at a time, so every record written will be an actual line,
15064 not chunks of lines and not (usually doesn't happen, but I've
15065 seen it once) several lines in one record. BIO_f_linebuffer() is
15068 Currently, it's a VMS-only method, because that's where it has
15069 been tested well enough.
15073 * Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
15074 it can return incorrect results.
15075 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
15076 but it was in 0.9.6-beta[12].)
15080 * Disable the check for content being present when verifying detached
15081 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
15082 include zero length content when signing messages.
15086 * New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
15087 BIO_ctrl (for BIO pairs).
15091 * Add DSO method for VMS.
15095 * Bug fix: Montgomery multiplication could produce results with the
15100 * Add RPM specification openssl.spec and modify it to build three
15101 packages. The default package contains applications, application
15102 documentation and run-time libraries. The devel package contains
15103 include files, static libraries and function documentation. The
15104 doc package contains the contents of the doc directory. The original
15105 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
15109 * Add a large number of documentation files for many SSL routines.
15111 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>*
15113 * Add a configuration entry for Sony News 4.
15115 *NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>*
15117 * Don't set the two most significant bits to one when generating a
15118 random number < q in the DSA library.
15122 * New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
15123 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
15124 the underlying transport is blocking) if a handshake took place.
15125 (The default behaviour is needed by applications such as s_client
15126 and s_server that use select() to determine when to use SSL_read;
15127 but for applications that know in advance when to expect data, it
15128 just makes things more complicated.)
15132 * Add RAND_egd_bytes(), which gives control over the number of bytes read
15137 * Add a few more EBCDIC conditionals that make `req` and `x509`
15138 work better on such systems.
15140 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
15142 * Add two demo programs for PKCS12_parse() and PKCS12_create().
15143 Update PKCS12_parse() so it copies the friendlyName and the
15144 keyid to the certificates aux info.
15148 * Fix bug in PKCS7_verify() which caused an infinite loop
15149 if there was more than one signature.
15151 *Sven Uszpelkat <su@celocom.de>*
15153 * Major change in util/mkdef.pl to include extra information
15154 about each symbol, as well as presenting variables as well
15155 as functions. This change means that there's n more need
15156 to rebuild the .num files when some algorithms are excluded.
15160 * Allow the verify time to be set by an application,
15161 rather than always using the current time.
15165 * Phase 2 verify code reorganisation. The certificate
15166 verify code now looks up an issuer certificate by a
15167 number of criteria: subject name, authority key id
15168 and key usage. It also verifies self signed certificates
15169 by the same criteria. The main comparison function is
15170 X509_check_issued() which performs these checks.
15172 Lot of changes were necessary in order to support this
15173 without completely rewriting the lookup code.
15175 Authority and subject key identifier are now cached.
15177 The LHASH 'certs' is X509_STORE has now been replaced
15178 by a STACK_OF(X509_OBJECT). This is mainly because an
15179 LHASH can't store or retrieve multiple objects with
15180 the same hash value.
15182 As a result various functions (which were all internal
15183 use only) have changed to handle the new X509_STORE
15184 structure. This will break anything that messed round
15185 with X509_STORE internally.
15187 The functions X509_STORE_add_cert() now checks for an
15188 exact match, rather than just subject name.
15190 The X509_STORE API doesn't directly support the retrieval
15191 of multiple certificates matching a given criteria, however
15192 this can be worked round by performing a lookup first
15193 (which will fill the cache with candidate certificates)
15194 and then examining the cache for matches. This is probably
15195 the best we can do without throwing out X509_LOOKUP
15196 entirely (maybe later...).
15198 The X509_VERIFY_CTX structure has been enhanced considerably.
15200 All certificate lookup operations now go via a get_issuer()
15201 callback. Although this currently uses an X509_STORE it
15202 can be replaced by custom lookups. This is a simple way
15203 to bypass the X509_STORE hackery necessary to make this
15204 work and makes it possible to use more efficient techniques
15205 in future. A very simple version which uses a simple
15206 STACK for its trusted certificate store is also provided
15207 using X509_STORE_CTX_trusted_stack().
15209 The verify_cb() and verify() callbacks now have equivalents
15210 in the X509_STORE_CTX structure.
15212 X509_STORE_CTX also has a 'flags' field which can be used
15213 to customise the verify behaviour.
15217 * Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
15218 excludes S/MIME capabilities.
15222 * When a certificate request is read in keep a copy of the
15223 original encoding of the signed data and use it when outputting
15224 again. Signatures then use the original encoding rather than
15225 a decoded, encoded version which may cause problems if the
15226 request is improperly encoded.
15230 * For consistency with other BIO_puts implementations, call
15231 buffer_write(b, ...) directly in buffer_puts instead of calling
15234 In BIO_puts, increment b->num_write as in BIO_write.
15236 *Peter.Sylvester@EdelWeb.fr*
15238 * Fix BN_mul_word for the case where the word is 0. (We have to use
15239 BN_zero, we may not return a BIGNUM with an array consisting of
15240 words set to zero.)
15244 * Avoid calling abort() from within the library when problems are
15245 detected, except if preprocessor symbols have been defined
15246 (such as REF_CHECK, BN_DEBUG etc.).
15250 * New openssl application 'rsautl'. This utility can be
15251 used for low-level RSA operations. DER public key
15252 BIO/fp routines also added.
15256 * New Configure entry and patches for compiling on QNX 4.
15258 *Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>*
15260 * A demo state-machine implementation was sponsored by
15261 Nuron (<http://www.nuron.com/>) and is now available in
15262 demos/state_machine.
15266 * New options added to the 'dgst' utility for signature
15267 generation and verification.
15271 * Unrecognized PKCS#7 content types are now handled via a
15272 catch all ASN1_TYPE structure. This allows unsupported
15273 types to be stored as a "blob" and an application can
15274 encode and decode it manually.
15278 * Fix various signed/unsigned issues to make a_strex.c
15279 compile under VC++.
15281 *Oscar Jacobsson <oscar.jacobsson@celocom.com>*
15283 * ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
15284 length if passed a buffer. ASN1_INTEGER_to_BN failed
15285 if passed a NULL BN and its argument was negative.
15287 *Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>*
15289 * Modification to PKCS#7 encoding routines to output definite
15290 length encoding. Since currently the whole structures are in
15291 memory there's not real point in using indefinite length
15292 constructed encoding. However if OpenSSL is compiled with
15293 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
15297 * Added BIO_vprintf() and BIO_vsnprintf().
15301 * Added more prefixes to parse for in the strings written
15302 through a logging bio, to cover all the levels that are available
15303 through syslog. The prefixes are now:
15305 PANIC, EMERG, EMR => LOG_EMERG
15306 ALERT, ALR => LOG_ALERT
15307 CRIT, CRI => LOG_CRIT
15308 ERROR, ERR => LOG_ERR
15309 WARNING, WARN, WAR => LOG_WARNING
15310 NOTICE, NOTE, NOT => LOG_NOTICE
15311 INFO, INF => LOG_INFO
15312 DEBUG, DBG => LOG_DEBUG
15314 and as before, if none of those prefixes are present at the
15315 beginning of the string, LOG_ERR is chosen.
15317 On Win32, the `LOG_*` levels are mapped according to this:
15319 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
15320 LOG_WARNING => EVENTLOG_WARNING_TYPE
15321 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
15325 * Made it possible to reconfigure with just the configuration
15326 argument "reconf" or "reconfigure". The command line arguments
15327 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
15328 and are retrieved from there when reconfiguring.
15334 *Assar Westerlund <assar@sics.se>, Richard Levitte*
15336 * Add the arguments -CAfile and -CApath to the pkcs12 utility.
15340 * The obj_dat.pl script was messing up the sorting of object
15341 names. The reason was that it compared the quoted version
15342 of strings as a result "OCSP" > "OCSP Signing" because
15343 " > SPACE. Changed script to store unquoted versions of
15344 names and add quotes on output. It was also omitting some
15345 names from the lookup table if they were given a default
15346 value (that is if SN is missing it is given the same
15347 value as LN and vice versa), these are now added on the
15348 grounds that if an object has a name we should be able to
15349 look it up. Finally added warning output when duplicate
15350 short or long names are found.
15354 * Changes needed for Tandem NSK.
15356 *Scott Uroff <scott@xypro.com>*
15358 * Fix SSL 2.0 rollback checking: Due to an off-by-one error in
15359 RSA_padding_check_SSLv23(), special padding was never detected
15360 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
15361 version rollback attacks was not effective.
15363 In s23_clnt.c, don't use special rollback-attack detection padding
15364 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
15365 client; similarly, in s23_srvr.c, don't do the rollback check if
15366 SSL 2.0 is the only protocol enabled in the server.
15370 * Make it possible to get hexdumps of unprintable data with 'openssl
15371 asn1parse'. By implication, the functions ASN1_parse_dump() and
15372 BIO_dump_indent() are added.
15376 * New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
15377 these print out strings and name structures based on various
15378 flags including RFC2253 support and proper handling of
15379 multibyte characters. Added options to the 'x509' utility
15380 to allow the various flags to be set.
15384 * Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
15385 Also change the functions X509_cmp_current_time() and
15386 X509_gmtime_adj() work with an ASN1_TIME structure,
15387 this will enable certificates using GeneralizedTime in validity
15388 dates to be checked.
15392 * Make the NEG_PUBKEY_BUG code (which tolerates invalid
15393 negative public key encodings) on by default,
15394 NO_NEG_PUBKEY_BUG can be set to disable it.
15398 * New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
15399 content octets. An i2c_ASN1_OBJECT is unnecessary because
15400 the encoding can be trivially obtained from the structure.
15404 * crypto/err.c locking bugfix: Use write locks (`CRYPTO_w_[un]lock`),
15405 not read locks (`CRYPTO_r_[un]lock`).
15409 * A first attempt at creating official support for shared
15410 libraries through configuration. I've kept it so the
15411 default is static libraries only, and the OpenSSL programs
15412 are always statically linked for now, but there are
15413 preparations for dynamic linking in place.
15414 This has been tested on Linux and Tru64.
15418 * Randomness polling function for Win9x, as described in:
15419 Peter Gutmann, Software Generation of Practically Strong
15424 * Fix so PRNG is seeded in req if using an already existing
15429 * New options to smime application. -inform and -outform
15430 allow alternative formats for the S/MIME message including
15431 PEM and DER. The -content option allows the content to be
15432 specified separately. This should allow things like Netscape
15433 form signing output easier to verify.
15437 * Fix the ASN1 encoding of tags using the 'long form'.
15441 * New ASN1 functions, `i2c_*` and `c2i_*` for INTEGER and BIT
15442 STRING types. These convert content octets to and from the
15443 underlying type. The actual tag and length octets are
15444 already assumed to have been read in and checked. These
15445 are needed because all other string types have virtually
15446 identical handling apart from the tag. By having versions
15447 of the ASN1 functions that just operate on content octets
15448 IMPLICIT tagging can be handled properly. It also allows
15449 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
15450 and ASN1_INTEGER are identical apart from the tag.
15454 * Change the handling of OID objects as follows:
15456 - New object identifiers are inserted in objects.txt, following
15457 the syntax given in [crypto/objects/README.md](crypto/objects/README.md).
15458 - objects.pl is used to process obj_mac.num and create a new
15460 - obj_dat.pl is used to create a new obj_dat.h, using the data in
15463 This is currently kind of a hack, and the perl code in objects.pl
15464 isn't very elegant, but it works as I intended. The simplest way
15465 to check that it worked correctly is to look in obj_dat.h and
15466 check the array nid_objs and make sure the objects haven't moved
15467 around (this is important!). Additions are OK, as well as
15468 consistent name changes.
15472 * Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
15476 * Addition of the command line parameter '-rand file' to 'openssl req'.
15477 The given file adds to whatever has already been seeded into the
15478 random pool through the RANDFILE configuration file option or
15479 environment variable, or the default random state file.
15483 * mkstack.pl now sorts each macro group into lexical order.
15484 Previously the output order depended on the order the files
15485 appeared in the directory, resulting in needless rewriting
15490 * Patches to make OpenSSL compile under Win32 again. Mostly
15491 work arounds for the VC++ problem that it treats func() as
15492 func(void). Also stripped out the parts of mkdef.pl that
15493 added extra typesafe functions: these no longer exist.
15497 * Reorganisation of the stack code. The macros are now all
15498 collected in safestack.h . Each macro is defined in terms of
15499 a "stack macro" of the form `SKM_<name>(type, a, b)`. The
15500 DEBUG_SAFESTACK is now handled in terms of function casts,
15501 this has the advantage of retaining type safety without the
15502 use of additional functions. If DEBUG_SAFESTACK is not defined
15503 then the non typesafe macros are used instead. Also modified the
15504 mkstack.pl script to handle the new form. Needs testing to see
15505 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
15506 the default if no major problems. Similar behaviour for ASN1_SET_OF
15507 and PKCS12_STACK_OF.
15511 * When some versions of IIS use the 'NET' form of private key the
15512 key derivation algorithm is different. Normally MD5(password) is
15513 used as a 128 bit RC4 key. In the modified case
15514 MD5(MD5(password) + "SGCKEYSALT") is used instead. Added some
15515 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
15516 as the old Netscape_RSA functions except they have an additional
15517 'sgckey' parameter which uses the modified algorithm. Also added
15518 an -sgckey command line option to the rsa utility. Thanks to
15519 Adrian Peck <bertie@ncipher.com> for posting details of the modified
15520 algorithm to openssl-dev.
15524 * The evp_local.h macros were using 'c.##kname' which resulted in
15525 invalid expansion on some systems (SCO 5.0.5 for example).
15526 Corrected to 'c.kname'.
15528 *Phillip Porch <root@theporch.com>*
15530 * New X509_get1_email() and X509_REQ_get1_email() functions that return
15531 a STACK of email addresses from a certificate or request, these look
15532 in the subject name and the subject alternative name extensions and
15533 omit any duplicate addresses.
15537 * Re-implement BN_mod_exp2_mont using independent (and larger) windows.
15538 This makes DSA verification about 2 % faster.
15542 * Increase maximum window size in `BN_mod_exp_...` to 6 bits instead of 5
15543 (meaning that now 2^5 values will be precomputed, which is only 4 KB
15544 plus overhead for 1024 bit moduli).
15545 This makes exponentiations about 0.5 % faster for 1024 bit
15546 exponents (as measured by "openssl speed rsa2048").
15550 * Rename memory handling macros to avoid conflicts with other
15552 Malloc => OPENSSL_malloc
15553 Malloc_locked => OPENSSL_malloc_locked
15554 Realloc => OPENSSL_realloc
15555 Free => OPENSSL_free
15559 * New function BN_mod_exp_mont_word for small bases (roughly 15%
15560 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
15564 * CygWin32 support.
15566 *John Jarvie <jjarvie@newsguy.com>*
15568 * The type-safe stack code has been rejigged. It is now only compiled
15569 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
15570 by default all type-specific stack functions are "#define"d back to
15571 standard stack functions. This results in more streamlined output
15572 but retains the type-safety checking possibilities of the original
15577 * The STACK code has been cleaned up, and certain type declarations
15578 that didn't make a lot of sense have been brought in line. This has
15579 also involved a cleanup of sorts in safestack.h to more correctly
15580 map type-safe stack functions onto their plain stack counterparts.
15581 This work has also resulted in a variety of "const"ifications of
15582 lots of the code, especially `_cmp` operations which should normally
15583 be prototyped with "const" parameters anyway.
15587 * When generating bytes for the first time in md_rand.c, 'stir the pool'
15588 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
15589 (The PRNG state consists of two parts, the large pool 'state' and 'md',
15590 where all of 'md' is used each time the PRNG is used, but 'state'
15591 is used only indexed by a cyclic counter. As entropy may not be
15592 well distributed from the beginning, 'md' is important as a
15593 chaining variable. However, the output function chains only half
15594 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
15595 all of 'md', and seeding with STATE_SIZE dummy bytes will result
15596 in all of 'state' being rewritten, with the new values depending
15597 on virtually all of 'md'. This overcomes the 80 bit limitation.)
15601 * In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
15602 the handshake is continued after ssl_verify_cert_chain();
15603 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
15604 can lead to 'unexplainable' connection aborts later.
15606 *Bodo Moeller; problem tracked down by Lutz Jaenicke*
15608 * Major EVP API cipher revision.
15609 Add hooks for extra EVP features. This allows various cipher
15610 parameters to be set in the EVP interface. Support added for variable
15611 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
15612 setting of RC2 and RC5 parameters.
15614 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
15617 Remove lots of duplicated code from the EVP library. For example *every*
15618 cipher init() function handles the 'iv' in the same way according to the
15619 cipher mode. They also all do nothing if the 'key' parameter is NULL and
15620 for CFB and OFB modes they zero ctx->num.
15622 New functionality allows removal of S/MIME code RC2 hack.
15624 Most of the routines have the same form and so can be declared in terms
15627 By shifting this to the top level EVP_CipherInit() it can be removed from
15628 all individual ciphers. If the cipher wants to handle IVs or keys
15629 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
15632 Change lots of functions like EVP_EncryptUpdate() to now return a
15633 value: although software versions of the algorithms cannot fail
15634 any installed hardware versions can.
15638 * Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
15639 this option is set, tolerate broken clients that send the negotiated
15640 protocol version number instead of the requested protocol version
15645 * Call dh_tmp_cb (set by `..._TMP_DH_CB`) with correct 'is_export' flag;
15646 i.e. non-zero for export ciphersuites, zero otherwise.
15647 Previous versions had this flag inverted, inconsistent with
15648 rsa_tmp_cb (..._TMP_RSA_CB).
15650 *Bodo Moeller; problem reported by Amit Chopra*
15652 * Add missing DSA library text string. Work around for some IIS
15653 key files with invalid SEQUENCE encoding.
15657 * Add a document (doc/standards.txt) that list all kinds of standards
15658 and so on that are implemented in OpenSSL.
15662 * Enhance c_rehash script. Old version would mishandle certificates
15663 with the same subject name hash and wouldn't handle CRLs at all.
15664 Added -fingerprint option to crl utility, to support new c_rehash
15669 * Eliminate non-ANSI declarations in crypto.h and stack.h.
15673 * Fix for SSL server purpose checking. Server checking was
15674 rejecting certificates which had extended key usage present
15675 but no ssl client purpose.
15677 *Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>*
15679 * Make PKCS#12 code work with no password. The PKCS#12 spec
15680 is a little unclear about how a blank password is handled.
15681 Since the password in encoded as a BMPString with terminating
15682 double NULL a zero length password would end up as just the
15683 double NULL. However no password at all is different and is
15684 handled differently in the PKCS#12 key generation code. NS
15685 treats a blank password as zero length. MSIE treats it as no
15686 password on export: but it will try both on import. We now do
15687 the same: PKCS12_parse() tries zero length and no password if
15688 the password is set to "" or NULL (NULL is now a valid password:
15689 it wasn't before) as does the pkcs12 application.
15693 * Bugfixes in `apps/x509.c`: Avoid a memory leak; and don't use
15694 perror when PEM_read_bio_X509_REQ fails, the error message must
15695 be obtained from the error queue.
15699 * Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
15700 it in ERR_remove_state if appropriate, and change ERR_get_state
15701 accordingly to avoid race conditions (this is necessary because
15702 thread_hash is no longer constant once set).
15706 * Bugfix for linux-elf makefile.one.
15710 * RSA_get_default_method() will now cause a default
15711 RSA_METHOD to be chosen if one doesn't exist already.
15712 Previously this was only set during a call to RSA_new()
15713 or RSA_new_method(NULL) meaning it was possible for
15714 RSA_get_default_method() to return NULL.
15718 * Added native name translation to the existing DSO code
15719 that will convert (if the flag to do so is set) filenames
15720 that are sufficiently small and have no path information
15721 into a canonical native form. Eg. "blah" converted to
15722 "libblah.so" or "blah.dll" etc.
15726 * New function ERR_error_string_n(e, buf, len) which is like
15727 ERR_error_string(e, buf), but writes at most 'len' bytes
15728 including the 0 terminator. For ERR_error_string_n, 'buf'
15731 *Damien Miller <djm@mindrot.org>, Bodo Moeller*
15733 * CONF library reworked to become more general. A new CONF
15734 configuration file reader "class" is implemented as well as a
15735 new functions (`NCONF_*`, for "New CONF") to handle it. The now
15736 old `CONF_*` functions are still there, but are reimplemented to
15737 work in terms of the new functions. Also, a set of functions
15738 to handle the internal storage of the configuration data is
15739 provided to make it easier to write new configuration file
15740 reader "classes" (I can definitely see something reading a
15741 configuration file in XML format, for example), called `_CONF_*`,
15742 or "the configuration storage API"...
15744 The new configuration file reading functions are:
15746 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
15747 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
15749 NCONF_default, NCONF_WIN32
15751 NCONF_dump_fp, NCONF_dump_bio
15753 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
15754 NCONF_new creates a new CONF object. This works in the same way
15755 as other interfaces in OpenSSL, like the BIO interface.
15756 `NCONF_dump_*` dump the internal storage of the configuration file,
15757 which is useful for debugging. All other functions take the same
15758 arguments as the old `CONF_*` functions with the exception of the
15759 first that must be a `CONF *` instead of a `LHASH *`.
15761 To make it easier to use the new classes with the old `CONF_*` functions,
15762 the function CONF_set_default_method is provided.
15766 * Add '-tls1' option to 'openssl ciphers', which was already
15767 mentioned in the documentation but had not been implemented.
15768 (This option is not yet really useful because even the additional
15769 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
15773 * Initial DSO code added into libcrypto for letting OpenSSL (and
15774 OpenSSL-based applications) load shared libraries and bind to
15775 them in a portable way.
15777 *Geoff Thorpe, with contributions from Richard Levitte*
15779 ### Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
15781 * Make sure _lrotl and _lrotr are only used with MSVC.
15783 * Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
15784 (the default implementation of RAND_status).
15786 * Rename openssl x509 option '-crlext', which was added in 0.9.5,
15787 to '-clrext' (= clear extensions), as intended and documented.
15788 *Bodo Moeller; inconsistency pointed out by Michael Attili
15789 <attili@amaxo.com>*
15791 * Fix for HMAC. It wasn't zeroing the rest of the block if the key length
15792 was larger than the MD block size.
15794 *Steve Henson, pointed out by Yost William <YostW@tce.com>*
15796 * Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
15797 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
15798 using the passed key: if the passed key was a private key the result
15799 of X509_print(), for example, would be to print out all the private key
15804 * des_quad_cksum() byte order bug fix.
15805 *Ulf Möller, using the problem description in krb4-0.9.7, where
15806 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>*
15808 * Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
15811 *Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>*
15813 * For easily testing in shell scripts whether some command
15814 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
15815 returns with exit code 0 iff no command of the given name is available.
15816 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
15817 the output goes to stdout and nothing is printed to stderr.
15818 Additional arguments are always ignored.
15820 Since for each cipher there is a command of the same name,
15821 the 'no-cipher' compilation switches can be tested this way.
15823 ('openssl no-XXX' is not able to detect pseudo-commands such
15824 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
15828 * Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
15832 * For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
15833 is set; it will be thrown away anyway because each handshake creates
15835 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
15836 to parameters -- in previous versions (since OpenSSL 0.9.3) the
15837 'default key' from SSL_CTX_set_tmp_dh would always be lost, meaning
15838 you effectively got SSL_OP_SINGLE_DH_USE when using this macro.
15842 * New s_client option -ign_eof: EOF at stdin is ignored, and
15843 'Q' and 'R' lose their special meanings (quit/renegotiate).
15844 This is part of what -quiet does; unlike -quiet, -ign_eof
15845 does not suppress any output.
15849 * Add compatibility options to the purpose and trust code. The
15850 purpose X509_PURPOSE_ANY is "any purpose" which automatically
15851 accepts a certificate or CA, this was the previous behaviour,
15852 with all the associated security issues.
15854 X509_TRUST_COMPAT is the old trust behaviour: only and
15855 automatically trust self signed roots in certificate store. A
15856 new trust setting X509_TRUST_DEFAULT is used to specify that
15857 a purpose has no associated trust setting and it should instead
15858 use the value in the default purpose.
15862 * Fix the PKCS#8 DSA private key code so it decodes keys again
15863 and fix a memory leak.
15867 * In util/mkerr.pl (which implements 'make errors'), preserve
15868 reason strings from the previous version of the .c file, as
15869 the default to have only downcase letters (and digits) in
15870 automatically generated reasons codes is not always appropriate.
15874 * In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
15875 using strerror. Previously, ERR_reason_error_string() returned
15876 library names as reason strings for SYSerr; but SYSerr is a special
15877 case where small numbers are errno values, not library numbers.
15881 * Add '-dsaparam' option to 'openssl dhparam' application. This
15882 converts DSA parameters into DH parameters. (When creating parameters,
15883 DSA_generate_parameters is used.)
15887 * Include 'length' (recommended exponent length) in C code generated
15888 by 'openssl dhparam -C'.
15892 * The second argument to set_label in perlasm was already being used
15893 so couldn't be used as a "file scope" flag. Moved to third argument
15898 * In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
15899 instead of RAND_bytes for encryption IVs and salts.
15903 * Include RAND_status() into RAND_METHOD instead of implementing
15904 it only for md_rand.c Otherwise replacing the PRNG by calling
15905 RAND_set_rand_method would be impossible.
15909 * Don't let DSA_generate_key() enter an infinite loop if the random
15910 number generation fails.
15914 * New 'rand' application for creating pseudo-random output.
15918 * Added configuration support for Linux/IA64
15920 *Rolf Haberrecker <rolf@suse.de>*
15922 * Assembler module support for Mingw32.
15926 * Shared library support for HPUX (in shlib/).
15928 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous*
15930 * Shared library support for Solaris gcc.
15932 *Lutz Behnke <behnke@trustcenter.de>*
15934 ### Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
15936 * PKCS7_encrypt() was adding text MIME headers twice because they
15937 were added manually and by SMIME_crlf_copy().
15941 * In bntest.c don't call BN_rand with zero bits argument.
15943 *Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>*
15945 * BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
15946 case was implemented. This caused BN_div_recp() to fail occasionally.
15950 * Add an optional second argument to the set_label() in the perl
15951 assembly language builder. If this argument exists and is set
15952 to 1 it signals that the assembler should use a symbol whose
15953 scope is the entire file, not just the current function. This
15954 is needed with MASM which uses the format label:: for this scope.
15956 *Steve Henson, pointed out by Peter Runestig <peter@runestig.com>*
15958 * Change the ASN1 types so they are typedefs by default. Before
15959 almost all types were #define'd to ASN1_STRING which was causing
15960 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
15965 * Change names of new functions to the new get1/get0 naming
15966 convention: After 'get1', the caller owns a reference count
15967 and has to call `..._free`; 'get0' returns a pointer to some
15968 data structure without incrementing reference counters.
15969 (Some of the existing 'get' functions increment a reference
15970 counter, some don't.)
15971 Similarly, 'set1' and 'add1' functions increase reference
15972 counters or duplicate objects.
15976 * Allow for the possibility of temp RSA key generation failure:
15977 the code used to assume it always worked and crashed on failure.
15981 * Fix potential buffer overrun problem in BIO_printf().
15982 *Ulf Möller, using public domain code by Patrick Powell; problem
15983 pointed out by David Sacerdote <das33@cornell.edu>*
15985 * Support EGD <http://www.lothar.com/tech/crypto/>. New functions
15986 RAND_egd() and RAND_status(). In the command line application,
15987 the EGD socket can be specified like a seed file using RANDFILE
15992 * Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
15993 Some CAs (e.g. Verisign) distribute certificates in this form.
15997 * Remove the SSL_ALLOW_ADH compile option and set the default cipher
15998 list to exclude them. This means that no special compilation option
15999 is needed to use anonymous DH: it just needs to be included in the
16004 * Change the EVP_MD_CTX_type macro so its meaning consistent with
16005 EVP_MD_type. The old functionality is available in a new macro called
16006 EVP_MD_md(). Change code that uses it and update docs.
16010 * `..._ctrl` functions now have corresponding `..._callback_ctrl` functions
16011 where the `void *` argument is replaced by a function pointer argument.
16012 Previously `void *` was abused to point to functions, which works on
16013 many platforms, but is not correct. As these functions are usually
16014 called by macros defined in OpenSSL header files, most source code
16015 should work without changes.
16019 * `<openssl/opensslconf.h>` (which is created by Configure) now contains
16020 sections with information on -D... compiler switches used for
16021 compiling the library so that applications can see them. To enable
16022 one of these sections, a pre-processor symbol `OPENSSL_..._DEFINES`
16023 must be defined. E.g.,
16024 #define OPENSSL_ALGORITHM_DEFINES
16025 #include <openssl/opensslconf.h>
16026 defines all pertinent `NO_<algo>` symbols, such as NO_IDEA, NO_RSA, etc.
16028 *Richard Levitte, Ulf and Bodo Möller*
16030 * Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
16035 * Change the 'other' type in certificate aux info to a STACK_OF
16036 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
16037 the required ASN1 format: arbitrary types determined by an OID.
16041 * Add some PEM_write_X509_REQ_NEW() functions and a command line
16042 argument to 'req'. This is not because the function is newer or
16043 better than others it just uses the work 'NEW' in the certificate
16044 request header lines. Some software needs this.
16048 * Reorganise password command line arguments: now passwords can be
16049 obtained from various sources. Delete the PEM_cb function and make
16050 it the default behaviour: i.e. if the callback is NULL and the
16051 usrdata argument is not NULL interpret it as a null terminated pass
16052 phrase. If usrdata and the callback are NULL then the pass phrase
16053 is prompted for as usual.
16057 * Add support for the Compaq Atalla crypto accelerator. If it is installed,
16058 the support is automatically enabled. The resulting binaries will
16059 autodetect the card and use it if present.
16061 *Ben Laurie and Compaq Inc.*
16063 * Work around for Netscape hang bug. This sends certificate request
16064 and server done in one record. Since this is perfectly legal in the
16065 SSL/TLS protocol it isn't a "bug" option and is on by default. See
16066 the bugs/SSLv3 entry for more info.
16070 * HP-UX tune-up: new unified configs, HP C compiler bug workaround.
16074 * Add -rand argument to smime and pkcs12 applications and read/write
16079 * New 'passwd' tool for crypt(3) and apr1 password hashes.
16083 * Add command line password options to the remaining applications.
16087 * Bug fix for BN_div_recp() for numerators with an even number of
16092 * More tests in bntest.c, and changed test_bn output.
16096 * ./config recognizes MacOS X now.
16100 * Bug fix for BN_div() when the first words of num and divisor are
16101 equal (it gave wrong results if `(rem=(n1-q*d0)&BN_MASK2) < d0)`.
16105 * Add support for various broken PKCS#8 formats, and command line
16106 options to produce them.
16110 * New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
16111 get temporary BIGNUMs from a BN_CTX.
16115 * Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
16120 * Change the `SSLeay_add_all_*()` functions to `OpenSSL_add_all_*()` and
16121 include a #define from the old name to the new. The original intent
16122 was that statically linked binaries could for example just call
16123 SSLeay_add_all_ciphers() to just add ciphers to the table and not
16124 link with digests. This never worked because SSLeay_add_all_digests()
16125 and SSLeay_add_all_ciphers() were in the same source file so calling
16126 one would link with the other. They are now in separate source files.
16130 * Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
16134 * Use a less unusual form of the Miller-Rabin primality test (it used
16135 a binary algorithm for exponentiation integrated into the Miller-Rabin
16136 loop, our standard modexp algorithms are faster).
16140 * Support for the EBCDIC character set completed.
16142 *Martin Kraemer <Martin.Kraemer@Mch.SNI.De>*
16144 * Source code cleanups: use const where appropriate, eliminate casts,
16145 use `void *` instead of `char *` in lhash.
16149 * Bugfix: ssl3_send_server_key_exchange was not restartable
16150 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
16151 this the server could overwrite ephemeral keys that the client
16156 * Turn DSA_is_prime into a macro that calls BN_is_prime,
16157 using 50 iterations of the Rabin-Miller test.
16159 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
16160 iterations of the Rabin-Miller test as required by the appendix
16161 to FIPS PUB 186[-1]) instead of DSA_is_prime.
16162 As BN_is_prime_fasttest includes trial division, DSA parameter
16163 generation becomes much faster.
16165 This implies a change for the callback functions in DSA_is_prime
16166 and DSA_generate_parameters: The callback function is called once
16167 for each positive witness in the Rabin-Miller test, not just
16168 occasionally in the inner loop; and the parameters to the
16169 callback function now provide an iteration count for the outer
16170 loop rather than for the current invocation of the inner loop.
16171 DSA_generate_parameters additionally can call the callback
16172 function with an 'iteration count' of -1, meaning that a
16173 candidate has passed the trial division test (when q is generated
16174 from an application-provided seed, trial division is skipped).
16178 * New function BN_is_prime_fasttest that optionally does trial
16179 division before starting the Rabin-Miller test and has
16180 an additional BN_CTX * argument (whereas BN_is_prime always
16181 has to allocate at least one BN_CTX).
16182 'callback(1, -1, cb_arg)' is called when a number has passed the
16183 trial division stage.
16187 * Fix for bug in CRL encoding. The validity dates weren't being handled
16192 * New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
16196 * New function BN_pseudo_rand().
16200 * Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
16201 bignum version of BN_from_montgomery() with the working code from
16202 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
16207 * Avoid a race condition in s2_clnt.c (function get_server_hello) that
16208 made it impossible to use the same SSL_SESSION data structure in
16209 SSL2 clients in multiple threads.
16213 * The return value of RAND_load_file() no longer counts bytes obtained
16214 by stat(). RAND_load_file(..., -1) is new and uses the complete file
16215 to seed the PRNG (previously an explicit byte count was required).
16217 *Ulf Möller, Bodo Möller*
16219 * Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
16220 used `char *` instead of `void *` and had casts all over the place.
16224 * Make BN_generate_prime() return NULL on error if ret!=NULL.
16228 * Retain source code compatibility for BN_prime_checks macro:
16229 BN_is_prime(..., BN_prime_checks, ...) now uses
16230 BN_prime_checks_for_size to determine the appropriate number of
16231 Rabin-Miller iterations.
16235 * Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
16236 DH_CHECK_P_NOT_SAFE_PRIME.
16237 (Check if this is true? OpenPGP calls them "strong".)
16241 * Merge the functionality of "dh" and "gendh" programs into a new program
16242 "dhparam". The old programs are retained for now but will handle DH keys
16243 (instead of parameters) in future.
16247 * Make the ciphers, s_server and s_client programs check the return values
16248 when a new cipher list is set.
16252 * Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
16253 ciphers. Before when the 56bit ciphers were enabled the sorting was
16256 The syntax for the cipher sorting has been extended to support sorting by
16257 cipher-strength (using the strength_bits hard coded in the tables).
16258 The new command is `@STRENGTH` (see also `doc/apps/ciphers.pod`).
16260 Fix a bug in the cipher-command parser: when supplying a cipher command
16261 string with an "undefined" symbol (neither command nor alphanumeric
16262 *A-Za-z0-9*, ssl_set_cipher_list used to hang in an endless loop. Now
16263 an error is flagged.
16265 Due to the strength-sorting extension, the code of the
16266 ssl_create_cipher_list() function was completely rearranged. I hope that
16267 the readability was also increased :-)
16269 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>*
16271 * Minor change to 'x509' utility. The -CAcreateserial option now uses 1
16272 for the first serial number and places 2 in the serial number file. This
16273 avoids problems when the root CA is created with serial number zero and
16274 the first user certificate has the same issuer name and serial number
16279 * Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
16280 the new code. Add documentation for this stuff.
16284 * Changes to X509_ATTRIBUTE utilities. These have been renamed from
16285 `X509_*()` to `X509at_*()` on the grounds that they don't handle X509
16286 structures and behave in an analogous way to the X509v3 functions:
16287 they shouldn't be called directly but wrapper functions should be used
16290 So we also now have some wrapper functions that call the X509at functions
16291 when passed certificate requests. (TO DO: similar things can be done with
16292 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
16293 things. Some of these need some d2i or i2d and print functionality
16294 because they handle more complex structures.)
16298 * Add missing #ifndefs that caused missing symbols when building libssl
16299 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
16300 NO_RSA in `ssl/s2*.c`.
16302 *Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller*
16304 * Precautions against using the PRNG uninitialized: RAND_bytes() now
16305 has a return value which indicates the quality of the random data
16306 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
16307 error queue. New function RAND_pseudo_bytes() generates output that is
16308 guaranteed to be unique but not unpredictable. RAND_add is like
16309 RAND_seed, but takes an extra argument for an entropy estimate
16310 (RAND_seed always assumes full entropy).
16314 * Do more iterations of Rabin-Miller probable prime test (specifically,
16315 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
16316 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
16317 in crypto/bn/bn_prime.c for the complete table). This guarantees a
16318 false-positive rate of at most 2^-80 for random input.
16322 * Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
16326 * New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
16327 in the 0.9.5 release), this returns the chain
16328 from an X509_CTX structure with a dup of the stack and all
16329 the X509 reference counts upped: so the stack will exist
16330 after X509_CTX_cleanup() has been called. Modify pkcs12.c
16333 Also make SSL_SESSION_print() print out the verify return
16338 * Add manpage for the pkcs12 command. Also change the default
16339 behaviour so MAC iteration counts are used unless the new
16340 -nomaciter option is used. This improves file security and
16341 only older versions of MSIE (4.0 for example) need it.
16345 * Honor the no-xxx Configure options when creating .DEF files.
16349 * Add PKCS#10 attributes to field table: challengePassword,
16350 unstructuredName and unstructuredAddress. These are taken from
16351 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
16352 international characters are used.
16354 More changes to X509_ATTRIBUTE code: allow the setting of types
16355 based on strings. Remove the 'loc' parameter when adding
16356 attributes because these will be a SET OF encoding which is sorted
16361 * Initial changes to the 'req' utility to allow request generation
16362 automation. This will allow an application to just generate a template
16363 file containing all the field values and have req construct the
16366 Initial support for X509_ATTRIBUTE handling. Stacks of these are
16367 used all over the place including certificate requests and PKCS#7
16368 structures. They are currently handled manually where necessary with
16369 some primitive wrappers for PKCS#7. The new functions behave in a
16370 manner analogous to the X509 extension functions: they allow
16371 attributes to be looked up by NID and added.
16373 Later something similar to the X509V3 code would be desirable to
16374 automatically handle the encoding, decoding and printing of the
16375 more complex types. The string types like challengePassword can
16376 be handled by the string table functions.
16378 Also modified the multi byte string table handling. Now there is
16379 a 'global mask' which masks out certain types. The table itself
16380 can use the flag STABLE_NO_MASK to ignore the mask setting: this
16381 is useful when for example there is only one permissible type
16382 (as in countryName) and using the mask might result in no valid
16387 * Clean up 'Finished' handling, and add functions SSL_get_finished and
16388 SSL_get_peer_finished to allow applications to obtain the latest
16389 Finished messages sent to the peer or expected from the peer,
16390 respectively. (SSL_get_peer_finished is usually the Finished message
16391 actually received from the peer, otherwise the protocol will be aborted.)
16393 As the Finished message are message digests of the complete handshake
16394 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
16395 be used for external authentication procedures when the authentication
16396 provided by SSL/TLS is not desired or is not enough.
16400 * Enhanced support for Alpha Linux is added. Now ./config checks if
16401 the host supports BWX extension and if Compaq C is present on the
16402 $PATH. Just exploiting of the BWX extension results in 20-30%
16403 performance kick for some algorithms, e.g. DES and RC4 to mention
16404 a couple. Compaq C in turn generates ~20% faster code for MD5 and
16409 * Add support for MS "fast SGC". This is arguably a violation of the
16410 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
16411 weak crypto and after checking the certificate is SGC a second one
16412 with strong crypto. MS SGC stops the first handshake after receiving
16413 the server certificate message and sends a second client hello. Since
16414 a server will typically do all the time consuming operations before
16415 expecting any further messages from the client (server key exchange
16416 is the most expensive) there is little difference between the two.
16418 To get OpenSSL to support MS SGC we have to permit a second client
16419 hello message after we have sent server done. In addition we have to
16420 reset the MAC if we do get this second client hello.
16424 * Add a function 'd2i_AutoPrivateKey()' this will automatically decide
16425 if a DER encoded private key is RSA or DSA traditional format. Changed
16426 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
16427 format DER encoded private key. Newer code should use PKCS#8 format which
16428 has the key type encoded in the ASN1 structure. Added DER private key
16429 support to pkcs8 application.
16433 * SSL 3/TLS 1 servers now don't request certificates when an anonymous
16434 ciphersuites has been selected (as required by the SSL 3/TLS 1
16435 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
16436 is set, we interpret this as a request to violate the specification
16437 (the worst that can happen is a handshake failure, and 'correct'
16438 behaviour would result in a handshake failure anyway).
16442 * In SSL_CTX_add_session, take into account that there might be multiple
16443 SSL_SESSION structures with the same session ID (e.g. when two threads
16444 concurrently obtain them from an external cache).
16445 The internal cache can handle only one SSL_SESSION with a given ID,
16446 so if there's a conflict, we now throw out the old one to achieve
16451 * Add OIDs for idea and blowfish in CBC mode. This will allow both
16452 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
16453 some routines that use cipher OIDs: some ciphers do not have OIDs
16454 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
16459 * Simplify the trust setting structure and code. Now we just have
16460 two sequences of OIDs for trusted and rejected settings. These will
16461 typically have values the same as the extended key usage extension
16462 and any application specific purposes.
16464 The trust checking code now has a default behaviour: it will just
16465 check for an object with the same NID as the passed id. Functions can
16466 be provided to override either the default behaviour or the behaviour
16467 for a given id. SSL client, server and email already have functions
16468 in place for compatibility: they check the NID and also return "trusted"
16469 if the certificate is self signed.
16473 * Add d2i,i2d bio/fp functions for PrivateKey: these convert the
16474 traditional format into an EVP_PKEY structure.
16478 * Add a password callback function PEM_cb() which either prompts for
16479 a password if usr_data is NULL or otherwise assumes it is a null
16480 terminated password. Allow passwords to be passed on command line
16481 environment or config files in a few more utilities.
16485 * Add a bunch of DER and PEM functions to handle PKCS#8 format private
16486 keys. Add some short names for PKCS#8 PBE algorithms and allow them
16487 to be specified on the command line for the pkcs8 and pkcs12 utilities.
16488 Update documentation.
16492 * Support for ASN1 "NULL" type. This could be handled before by using
16493 ASN1_TYPE but there wasn't any function that would try to read a NULL
16494 and produce an error if it couldn't. For compatibility we also have
16495 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
16496 don't allocate anything because they don't need to.
16500 * Initial support for MacOS is now provided. Examine INSTALL.MacOS
16503 *Andy Polyakov, Roy Woods <roy@centicsystems.ca>*
16505 * Rebuild of the memory allocation routines used by OpenSSL code and
16506 possibly others as well. The purpose is to make an interface that
16507 provide hooks so anyone can build a separate set of allocation and
16508 deallocation routines to be used by OpenSSL, for example memory
16509 pool implementations, or something else, which was previously hard
16510 since Malloc(), Realloc() and Free() were defined as macros having
16511 the values malloc, realloc and free, respectively (except for Win32
16512 compilations). The same is provided for memory debugging code.
16513 OpenSSL already comes with functionality to find memory leaks, but
16514 this gives people a chance to debug other memory problems.
16516 With these changes, a new set of functions and macros have appeared:
16518 CRYPTO_set_mem_debug_functions() [F]
16519 CRYPTO_get_mem_debug_functions() [F]
16520 CRYPTO_dbg_set_options() [F]
16521 CRYPTO_dbg_get_options() [F]
16522 CRYPTO_malloc_debug_init() [M]
16524 The memory debug functions are NULL by default, unless the library
16525 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
16526 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
16527 gives the standard debugging functions that come with OpenSSL) or
16528 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
16529 provided by the library user) must be used. When the standard
16530 debugging functions are used, CRYPTO_dbg_set_options can be used to
16531 request additional information:
16532 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
16533 the CRYPTO_MDEBUG_xxx macro when compiling the library.
16535 Also, things like CRYPTO_set_mem_functions will always give the
16536 expected result (the new set of functions is used for allocation
16537 and deallocation) at all times, regardless of platform and compiler
16540 To finish it up, some functions that were never use in any other
16541 way than through macros have a new API and new semantic:
16543 CRYPTO_dbg_malloc()
16544 CRYPTO_dbg_realloc()
16547 All macros of value have retained their old syntax.
16549 *Richard Levitte and Bodo Moeller*
16551 * Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
16552 ordering of SMIMECapabilities wasn't in "strength order" and there
16553 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
16558 * Some ASN1 types with illegal zero length encoding (INTEGER,
16559 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
16561 *Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson*
16563 * Merge in my S/MIME library for OpenSSL. This provides a simple
16564 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
16565 functionality to handle multipart/signed properly) and a utility
16566 called 'smime' to call all this stuff. This is based on code I
16567 originally wrote for Celo who have kindly allowed it to be
16568 included in OpenSSL.
16572 * Add variants des_set_key_checked and des_set_key_unchecked of
16573 des_set_key (aka des_key_sched). Global variable des_check_key
16574 decides which of these is called by des_set_key; this way
16575 des_check_key behaves as it always did, but applications and
16576 the library itself, which was buggy for des_check_key == 1,
16577 have a cleaner way to pick the version they need.
16581 * New function PKCS12_newpass() which changes the password of a
16586 * Modify X509_TRUST and X509_PURPOSE so it also uses a static and
16587 dynamic mix. In both cases the ids can be used as an index into the
16588 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
16589 functions so they accept a list of the field values and the
16590 application doesn't need to directly manipulate the X509_TRUST
16595 * Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
16600 * Modify the way the V3 extension code looks up extensions. This now
16601 works in a similar way to the object code: we have some "standard"
16602 extensions in a static table which is searched with OBJ_bsearch()
16603 and the application can add dynamic ones if needed. The file
16604 crypto/x509v3/ext_dat.h now has the info: this file needs to be
16605 updated whenever a new extension is added to the core code and kept
16606 in ext_nid order. There is a simple program 'tabtest.c' which checks
16607 this. New extensions are not added too often so this file can readily
16608 be maintained manually.
16610 There are two big advantages in doing things this way. The extensions
16611 can be looked up immediately and no longer need to be "added" using
16612 X509V3_add_standard_extensions(): this function now does nothing.
16613 Side note: I get *lots* of email saying the extension code doesn't
16614 work because people forget to call this function.
16615 Also no dynamic allocation is done unless new extensions are added:
16616 so if we don't add custom extensions there is no need to call
16617 X509V3_EXT_cleanup().
16621 * Modify enc utility's salting as follows: make salting the default. Add a
16622 magic header, so unsalted files fail gracefully instead of just decrypting
16623 to garbage. This is because not salting is a big security hole, so people
16624 should be discouraged from doing it.
16628 * Fixes and enhancements to the 'x509' utility. It allowed a message
16629 digest to be passed on the command line but it only used this
16630 parameter when signing a certificate. Modified so all relevant
16631 operations are affected by the digest parameter including the
16632 -fingerprint and -x509toreq options. Also -x509toreq choked if a
16633 DSA key was used because it didn't fix the digest.
16637 * Initial certificate chain verify code. Currently tests the untrusted
16638 certificates for consistency with the verify purpose (which is set
16639 when the X509_STORE_CTX structure is set up) and checks the pathlength.
16641 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
16642 this is because it will reject chains with invalid extensions whereas
16643 every previous version of OpenSSL and SSLeay made no checks at all.
16645 Trust code: checks the root CA for the relevant trust settings. Trust
16646 settings have an initial value consistent with the verify purpose: e.g.
16647 if the verify purpose is for SSL client use it expects the CA to be
16648 trusted for SSL client use. However the default value can be changed to
16649 permit custom trust settings: one example of this would be to only trust
16650 certificates from a specific "secure" set of CAs.
16652 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
16653 which should be used for version portability: especially since the
16654 verify structure is likely to change more often now.
16656 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
16657 to set them. If not set then assume SSL clients will verify SSL servers
16660 Two new options to the verify program: -untrusted allows a set of
16661 untrusted certificates to be passed in and -purpose which sets the
16662 intended purpose of the certificate. If a purpose is set then the
16663 new chain verify code is used to check extension consistency.
16667 * Support for the authority information access extension.
16671 * Modify RSA and DSA PEM read routines to transparently handle
16672 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
16673 public keys in a format compatible with certificate
16674 SubjectPublicKeyInfo structures. Unfortunately there were already
16675 functions called *_PublicKey_* which used various odd formats so
16676 these are retained for compatibility: however the DSA variants were
16677 never in a public release so they have been deleted. Changed dsa/rsa
16678 utilities to handle the new format: note no releases ever handled public
16679 keys so we should be OK.
16681 The primary motivation for this change is to avoid the same fiasco
16682 that dogs private keys: there are several incompatible private key
16683 formats some of which are standard and some OpenSSL specific and
16684 require various evil hacks to allow partial transparent handling and
16685 even then it doesn't work with DER formats. Given the option anything
16686 other than PKCS#8 should be dumped: but the other formats have to
16687 stay in the name of compatibility.
16689 With public keys and the benefit of hindsight one standard format
16690 is used which works with EVP_PKEY, RSA or DSA structures: though
16691 it clearly returns an error if you try to read the wrong kind of key.
16693 Added a -pubkey option to the 'x509' utility to output the public key.
16694 Also rename the `EVP_PKEY_get_*()` to `EVP_PKEY_rget_*()`
16695 (renamed to `EVP_PKEY_get1_*()` in the OpenSSL 0.9.5 release) and add
16696 `EVP_PKEY_rset_*()` functions (renamed to `EVP_PKEY_set1_*()`)
16697 that do the same as the `EVP_PKEY_assign_*()` except they up the
16698 reference count of the added key (they don't "swallow" the
16703 * Fixes to crypto/x509/by_file.c the code to read in certificates and
16704 CRLs would fail if the file contained no certificates or no CRLs:
16705 added a new function to read in both types and return the number
16706 read: this means that if none are read it will be an error. The
16707 DER versions of the certificate and CRL reader would always fail
16708 because it isn't possible to mix certificates and CRLs in DER format
16709 without choking one or the other routine. Changed this to just read
16710 a certificate: this is the best we can do. Also modified the code
16711 in `apps/verify.c` to take notice of return codes: it was previously
16712 attempting to read in certificates from NULL pointers and ignoring
16713 any errors: this is one reason why the cert and CRL reader seemed
16714 to work. It doesn't check return codes from the default certificate
16715 routines: these may well fail if the certificates aren't installed.
16719 * Code to support otherName option in GeneralName.
16723 * First update to verify code. Change the verify utility
16724 so it warns if it is passed a self signed certificate:
16725 for consistency with the normal behaviour. X509_verify
16726 has been modified to it will now verify a self signed
16727 certificate if *exactly* the same certificate appears
16728 in the store: it was previously impossible to trust a
16729 single self signed certificate. This means that:
16730 openssl verify ss.pem
16731 now gives a warning about a self signed certificate but
16732 openssl verify -CAfile ss.pem ss.pem
16737 * For servers, store verify_result in SSL_SESSION data structure
16738 (and add it to external session representation).
16739 This is needed when client certificate verifications fails,
16740 but an application-provided verification callback (set by
16741 SSL_CTX_set_cert_verify_callback) allows accepting the session
16742 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
16743 but returns 1): When the session is reused, we have to set
16744 ssl->verify_result to the appropriate error code to avoid
16747 *Bodo Moeller, problem pointed out by Lutz Jaenicke*
16749 * Fix a bug in the new PKCS#7 code: it didn't consider the
16750 case in PKCS7_dataInit() where the signed PKCS7 structure
16751 didn't contain any existing data because it was being created.
16753 *Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson*
16755 * Add a salt to the key derivation routines in enc.c. This
16756 forms the first 8 bytes of the encrypted file. Also add a
16757 -S option to allow a salt to be input on the command line.
16761 * New function X509_cmp(). Oddly enough there wasn't a function
16762 to compare two certificates. We do this by working out the SHA1
16763 hash and comparing that. X509_cmp() will be needed by the trust
16768 * SSL_get1_session() is like SSL_get_session(), but increments
16769 the reference count in the SSL_SESSION returned.
16771 *Geoff Thorpe <geoff@eu.c2.net>*
16773 * Fix for 'req': it was adding a null to request attributes.
16774 Also change the X509_LOOKUP and X509_INFO code to handle
16775 certificate auxiliary information.
16779 * Add support for 40 and 64 bit RC2 and RC4 algorithms: document
16784 * Add the possibility to add extra information to the memory leak
16785 detecting output, to form tracebacks, showing from where each
16786 allocation was originated: CRYPTO_push_info("constant string") adds
16787 the string plus current file name and line number to a per-thread
16788 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
16789 is like calling CYRPTO_pop_info() until the stack is empty.
16790 Also updated memory leak detection code to be multi-thread-safe.
16794 * Add options -text and -noout to pkcs7 utility and delete the
16795 encryption options which never did anything. Update docs.
16799 * Add options to some of the utilities to allow the pass phrase
16800 to be included on either the command line (not recommended on
16801 OSes like Unix) or read from the environment. Update the
16802 manpages and fix a few bugs.
16806 * Add a few manpages for some of the openssl commands.
16810 * Fix the -revoke option in ca. It was freeing up memory twice,
16811 leaking and not finding already revoked certificates.
16815 * Extensive changes to support certificate auxiliary information.
16816 This involves the use of X509_CERT_AUX structure and X509_AUX
16817 functions. An X509_AUX function such as PEM_read_X509_AUX()
16818 can still read in a certificate file in the usual way but it
16819 will also read in any additional "auxiliary information". By
16820 doing things this way a fair degree of compatibility can be
16821 retained: existing certificates can have this information added
16822 using the new 'x509' options.
16824 Current auxiliary information includes an "alias" and some trust
16825 settings. The trust settings will ultimately be used in enhanced
16826 certificate chain verification routines: currently a certificate
16827 can only be trusted if it is self signed and then it is trusted
16832 * Fix assembler for Alpha (tested only on DEC OSF not Linux or `*BSD`).
16833 The problem was that one of the replacement routines had not been working
16834 since SSLeay releases. For now the offending routine has been replaced
16835 with non-optimised assembler. Even so, this now gives around 95%
16836 performance improvement for 1024 bit RSA signs.
16840 * Hack to fix PKCS#7 decryption when used with some unorthodox RC2
16841 handling. Most clients have the effective key size in bits equal to
16842 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
16843 A few however don't do this and instead use the size of the decrypted key
16844 to determine the RC2 key length and the AlgorithmIdentifier to determine
16845 the effective key length. In this case the effective key length can still
16846 be 40 bits but the key length can be 168 bits for example. This is fixed
16847 by manually forcing an RC2 key into the EVP_PKEY structure because the
16848 EVP code can't currently handle unusual RC2 key sizes: it always assumes
16849 the key length and effective key length are equal.
16853 * Add a bunch of functions that should simplify the creation of
16854 X509_NAME structures. Now you should be able to do:
16855 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
16856 and have it automatically work out the correct field type and fill in
16857 the structures. The more adventurous can try:
16858 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
16859 and it will (hopefully) work out the correct multibyte encoding.
16863 * Change the 'req' utility to use the new field handling and multibyte
16864 copy routines. Before the DN field creation was handled in an ad hoc
16865 way in req, ca, and x509 which was rather broken and didn't support
16866 BMPStrings or UTF8Strings. Since some software doesn't implement
16867 BMPStrings or UTF8Strings yet, they can be enabled using the config file
16868 using the dirstring_type option. See the new comment in the default
16869 openssl.cnf for more info.
16873 * Make crypto/rand/md_rand.c more robust:
16874 - Assure unique random numbers after fork().
16875 - Make sure that concurrent threads access the global counter and
16876 md serializably so that we never lose entropy in them
16877 or use exactly the same state in multiple threads.
16878 Access to the large state is not always serializable because
16879 the additional locking could be a performance killer, and
16880 md should be large enough anyway.
16884 * New file `apps/app_rand.c` with commonly needed functionality
16885 for handling the random seed file.
16887 Use the random seed file in some applications that previously did not:
16889 dsaparam -genkey (which also ignored its '-rand' option),
16892 x509 (when signing).
16893 Except on systems with /dev/urandom, it is crucial to have a random
16894 seed file at least for key creation, DSA signing, and for DH exchanges;
16895 for RSA signatures we could do without one.
16897 gendh and gendsa (unlike genrsa) used to read only the first byte
16898 of each file listed in the '-rand' option. The function as previously
16899 found in genrsa is now in app_rand.c and is used by all programs
16900 that support '-rand'.
16904 * In RAND_write_file, use mode 0600 for creating files;
16905 don't just chmod when it may be too late.
16909 * Report an error from X509_STORE_load_locations
16910 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
16914 * New function ASN1_mbstring_copy() this copies a string in either
16915 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
16916 into an ASN1_STRING type. A mask of permissible types is passed
16917 and it chooses the "minimal" type to use or an error if not type
16922 * Add function equivalents to the various macros in asn1.h. The old
16923 macros are retained with an `M_` prefix. Code inside the library can
16924 use the `M_` macros. External code (including the openssl utility)
16925 should *NOT* in order to be "shared library friendly".
16929 * Add various functions that can check a certificate's extensions
16930 to see if it usable for various purposes such as SSL client,
16931 server or S/MIME and CAs of these types. This is currently
16932 VERY EXPERIMENTAL but will ultimately be used for certificate chain
16933 verification. Also added a -purpose flag to x509 utility to
16934 print out all the purposes.
16938 * Add a CRYPTO_EX_DATA to X509 certificate structure and associated
16943 * New `X509V3_{X509,CRL,REVOKED}_get_d2i()` functions. These will search
16944 for, obtain and decode and extension and obtain its critical flag.
16945 This allows all the necessary extension code to be handled in a
16946 single function call.
16950 * RC4 tune-up featuring 30-40% performance improvement on most RISC
16951 platforms. See crypto/rc4/rc4_enc.c for further details.
16955 * New -noout option to asn1parse. This causes no output to be produced
16956 its main use is when combined with -strparse and -out to extract data
16957 from a file (which may not be in ASN.1 format).
16961 * Fix for pkcs12 program. It was hashing an invalid certificate pointer
16962 when producing the local key id.
16964 *Richard Levitte <levitte@stacken.kth.se>*
16966 * New option -dhparam in s_server. This allows a DH parameter file to be
16967 stated explicitly. If it is not stated then it tries the first server
16968 certificate file. The previous behaviour hard coded the filename
16973 * Add -pubin and -pubout options to the rsa and dsa commands. These allow
16974 a public key to be input or output. For example:
16975 openssl rsa -in key.pem -pubout -out pubkey.pem
16976 Also added necessary DSA public key functions to handle this.
16980 * Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
16981 in the message. This was handled by allowing
16982 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
16984 *Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>*
16986 * Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
16987 to the end of the strings whereas this didn't. This would cause problems
16988 if strings read with d2i_ASN1_bytes() were later modified.
16990 *Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>*
16992 * Fix for base64 decode bug. When a base64 bio reads only one line of
16993 data and it contains EOF it will end up returning an error. This is
16994 caused by input 46 bytes long. The cause is due to the way base64
16995 BIOs find the start of base64 encoded data. They do this by trying a
16996 trial decode on each line until they find one that works. When they
16997 do a flag is set and it starts again knowing it can pass all the
16998 data directly through the decoder. Unfortunately it doesn't reset
16999 the context it uses. This means that if EOF is reached an attempt
17000 is made to pass two EOFs through the context and this causes the
17001 resulting error. This can also cause other problems as well. As is
17002 usual with these problems it takes *ages* to find and the fix is
17003 trivial: move one line.
17005 *Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer)*
17007 * Ugly workaround to get s_client and s_server working under Windows. The
17008 old code wouldn't work because it needed to select() on sockets and the
17009 tty (for keypresses and to see if data could be written). Win32 only
17010 supports select() on sockets so we select() with a 1s timeout on the
17011 sockets and then see if any characters are waiting to be read, if none
17012 are present then we retry, we also assume we can always write data to
17013 the tty. This isn't nice because the code then blocks until we've
17014 received a complete line of data and it is effectively polling the
17015 keyboard at 1s intervals: however it's quite a bit better than not
17016 working at all :-) A dedicated Windows application might handle this
17017 with an event loop for example.
17021 * Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
17022 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
17023 will be called when RSA_sign() and RSA_verify() are used. This is useful
17024 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
17025 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
17026 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
17027 This necessitated the support of an extra signature type NID_md5_sha1
17028 for SSL signatures and modifications to the SSL library to use it instead
17029 of calling RSA_public_decrypt() and RSA_private_encrypt().
17033 * Add new -verify -CAfile and -CApath options to the crl program, these
17034 will lookup a CRL issuers certificate and verify the signature in a
17035 similar way to the verify program. Tidy up the crl program so it
17036 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
17037 less strict. It will now permit CRL extensions even if it is not
17038 a V2 CRL: this will allow it to tolerate some broken CRLs.
17042 * Initialize all non-automatic variables each time one of the openssl
17043 sub-programs is started (this is necessary as they may be started
17044 multiple times from the "OpenSSL>" prompt).
17046 *Lennart Bang, Bodo Moeller*
17048 * Preliminary compilation option RSA_NULL which disables RSA crypto without
17049 removing all other RSA functionality (this is what NO_RSA does). This
17050 is so (for example) those in the US can disable those operations covered
17051 by the RSA patent while allowing storage and parsing of RSA keys and RSA
17056 * Non-copying interface to BIO pairs.
17057 (still largely untested)
17061 * New function ASN1_tag2str() to convert an ASN1 tag to a descriptive
17062 ASCII string. This was handled independently in various places before.
17066 * New functions UTF8_getc() and UTF8_putc() that parse and generate
17067 UTF8 strings a character at a time.
17071 * Use client_version from client hello to select the protocol
17072 (s23_srvr.c) and for RSA client key exchange verification
17073 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
17077 * Add various utility functions to handle SPKACs, these were previously
17078 handled by poking round in the structure internals. Added new function
17079 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
17080 print, verify and generate SPKACs. Based on an original idea from
17081 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
17085 * RIPEMD160 is operational on all platforms and is back in 'make test'.
17089 * Allow the config file extension section to be overwritten on the
17090 command line. Based on an original idea from Massimiliano Pala
17091 <madwolf@comune.modena.it>. The new option is called -extensions
17092 and can be applied to ca, req and x509. Also -reqexts to override
17093 the request extensions in req and -crlexts to override the crl extensions
17098 * Add new feature to the SPKAC handling in ca. Now you can include
17099 the same field multiple times by preceding it by "XXXX." for example:
17102 this is the same syntax as used in the req config file.
17106 * Allow certificate extensions to be added to certificate requests. These
17107 are specified in a 'req_extensions' option of the req section of the
17108 config file. They can be printed out with the -text option to req but
17109 are otherwise ignored at present.
17113 * Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
17114 data read consists of only the final block it would not decrypted because
17115 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
17116 A misplaced 'break' also meant the decrypted final block might not be
17117 copied until the next read.
17121 * Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
17122 a few extra parameters to the DH structure: these will be useful if
17123 for example we want the value of 'q' or implement X9.42 DH.
17127 * Initial support for DSA_METHOD. This is based on the RSA_METHOD and
17128 provides hooks that allow the default DSA functions or functions on a
17129 "per key" basis to be replaced. This allows hardware acceleration and
17130 hardware key storage to be handled without major modification to the
17131 library. Also added low-level modexp hooks and CRYPTO_EX structure and
17132 associated functions.
17136 * Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
17137 as "read only": it can't be written to and the buffer it points to will
17138 not be freed. Reading from a read only BIO is much more efficient than
17139 a normal memory BIO. This was added because there are several times when
17140 an area of memory needs to be read from a BIO. The previous method was
17141 to create a memory BIO and write the data to it, this results in two
17142 copies of the data and an O(n^2) reading algorithm. There is a new
17143 function BIO_new_mem_buf() which creates a read only memory BIO from
17144 an area of memory. Also modified the PKCS#7 routines to use read only
17149 * Bugfix: ssl23_get_client_hello did not work properly when called in
17150 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
17151 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
17152 but a retry condition occurred while trying to read the rest.
17156 * The PKCS7_ENC_CONTENT_new() function was setting the content type as
17157 NID_pkcs7_encrypted by default: this was wrong since this should almost
17158 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
17159 the encrypted data type: this is a more sensible place to put it and it
17160 allows the PKCS#12 code to be tidied up that duplicated this
17165 * Changed obj_dat.pl script so it takes its input and output files on
17166 the command line. This should avoid shell escape redirection problems
17171 * Initial support for certificate extension requests, these are included
17172 in things like Xenroll certificate requests. Included functions to allow
17173 extensions to be obtained and added.
17177 * -crlf option to s_client and s_server for sending newlines as
17178 CRLF (as required by many protocols).
17182 ### Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
17184 * Install libRSAglue.a when OpenSSL is built with RSAref.
17186 *Ralf S. Engelschall*
17188 * A few more `#ifndef NO_FP_API / #endif` pairs for consistency.
17190 *Andrija Antonijevic <TheAntony2@bigfoot.com>*
17192 * Fix -startdate and -enddate (which was missing) arguments to 'ca'
17197 * New function DSA_dup_DH, which duplicates DSA parameters/keys as
17198 DH parameters/keys (q is lost during that conversion, but the resulting
17199 DH parameters contain its length).
17201 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
17202 much faster than DH_generate_parameters (which creates parameters
17203 where `p = 2*q + 1`), and also the smaller q makes DH computations
17204 much more efficient (160-bit exponentiation instead of 1024-bit
17205 exponentiation); so this provides a convenient way to support DHE
17206 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
17207 utter importance to use
17208 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
17210 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
17211 when such DH parameters are used, because otherwise small subgroup
17212 attacks may become possible!
17216 * Avoid memory leak in i2d_DHparams.
17220 * Allow the -k option to be used more than once in the enc program:
17221 this allows the same encrypted message to be read by multiple recipients.
17225 * New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
17226 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
17227 it will always use the numerical form of the OID, even if it has a short
17232 * Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
17233 method only got called if p,q,dmp1,dmq1,iqmp components were present,
17234 otherwise bn_mod_exp was called. In the case of hardware keys for example
17235 no private key components need be present and it might store extra data
17236 in the RSA structure, which cannot be accessed from bn_mod_exp.
17237 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
17238 private key operations.
17242 * Added support for SPARC Linux.
17246 * pem_password_cb function type incompatibly changed from
17247 typedef int pem_password_cb(char *buf, int size, int rwflag);
17249 ....(char *buf, int size, int rwflag, void *userdata);
17250 so that applications can pass data to their callbacks:
17251 The `PEM[_ASN1]_{read,write}...` functions and macros now take an
17252 additional void * argument, which is just handed through whenever
17253 the password callback is called.
17255 *Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller*
17257 New function SSL_CTX_set_default_passwd_cb_userdata.
17259 Compatibility note: As many C implementations push function arguments
17260 onto the stack in reverse order, the new library version is likely to
17261 interoperate with programs that have been compiled with the old
17262 pem_password_cb definition (PEM_whatever takes some data that
17263 happens to be on the stack as its last argument, and the callback
17264 just ignores this garbage); but there is no guarantee whatsoever that
17267 * The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
17268 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
17269 problems not only on Windows, but also on some Unix platforms.
17270 To avoid problematic command lines, these definitions are now in an
17271 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
17272 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
17276 * MIPS III/IV assembler module is reimplemented.
17280 * More DES library cleanups: remove references to srand/rand and
17281 delete an unused file.
17285 * Add support for the free Netwide assembler (NASM) under Win32,
17286 since not many people have MASM (ml) and it can be hard to obtain.
17287 This is currently experimental but it seems to work OK and pass all
17288 the tests. Check out INSTALL.W32 for info.
17292 * Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
17293 without temporary keys kept an extra copy of the server key,
17294 and connections with temporary keys did not free everything in case
17299 * New function RSA_check_key and new openssl rsa option -check
17300 for verifying the consistency of RSA keys.
17302 *Ulf Moeller, Bodo Moeller*
17304 * Various changes to make Win32 compile work:
17305 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
17306 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
17307 comparison" warnings.
17308 3. Add `sk_<TYPE>_sort` to DEF file generator and do make update.
17312 * Add a debugging option to PKCS#5 v2 key generation function: when
17313 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
17314 derived keys are printed to stderr.
17318 * Copy the flags in ASN1_STRING_dup().
17320 *Roman E. Pavlov <pre@mo.msk.ru>*
17322 * The x509 application mishandled signing requests containing DSA
17323 keys when the signing key was also DSA and the parameters didn't match.
17325 It was supposed to omit the parameters when they matched the signing key:
17326 the verifying software was then supposed to automatically use the CA's
17327 parameters if they were absent from the end user certificate.
17329 Omitting parameters is no longer recommended. The test was also
17330 the wrong way round! This was probably due to unusual behaviour in
17331 EVP_cmp_parameters() which returns 1 if the parameters match.
17332 This meant that parameters were omitted when they *didn't* match and
17333 the certificate was useless. Certificates signed with 'ca' didn't have
17336 *Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>*
17338 * Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
17339 The interface is as follows:
17340 Applications can use
17341 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
17342 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
17343 "off" is now the default.
17344 The library internally uses
17345 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
17346 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
17347 to disable memory-checking temporarily.
17349 Some inconsistent states that previously were possible (and were
17350 even the default) are now avoided.
17352 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
17353 with each memory chunk allocated; this is occasionally more helpful
17354 than just having a counter.
17356 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
17358 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
17363 * Introduce "mode" for SSL structures (with defaults in SSL_CTX),
17364 which largely parallels "options", but is for changing API behaviour,
17365 whereas "options" are about protocol behaviour.
17366 Initial "mode" flags are:
17368 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
17369 a single record has been written.
17370 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
17371 retries use the same buffer location.
17372 (But all of the contents must be
17377 * Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
17380 * Fix problems with no-hmac etc.
17382 *Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>*
17384 * New functions RSA_get_default_method(), RSA_set_method() and
17385 RSA_get_method(). These allows replacement of RSA_METHODs without having
17386 to mess around with the internals of an RSA structure.
17390 * Fix memory leaks in DSA_do_sign and DSA_is_prime.
17391 Also really enable memory leak checks in openssl.c and in some
17394 *Chad C. Mulligan, Bodo Moeller*
17396 * Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
17397 up the length of negative integers. This has now been simplified to just
17398 store the length when it is first determined and use it later, rather
17399 than trying to keep track of where data is copied and updating it to
17401 *Steve Henson, reported by Brien Wheeler <bwheeler@authentica-security.com>*
17403 * Add a new function PKCS7_signatureVerify. This allows the verification
17404 of a PKCS#7 signature but with the signing certificate passed to the
17405 function itself. This contrasts with PKCS7_dataVerify which assumes the
17406 certificate is present in the PKCS#7 structure. This isn't always the
17407 case: certificates can be omitted from a PKCS#7 structure and be
17408 distributed by "out of band" means (such as a certificate database).
17412 * Complete the `PEM_*` macros with DECLARE_PEM versions to replace the
17413 function prototypes in pem.h, also change util/mkdef.pl to add the
17414 necessary function names.
17418 * mk1mf.pl (used by Windows builds) did not properly read the
17419 options set by Configure in the top level Makefile, and Configure
17420 was not even able to write more than one option correctly.
17421 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
17425 * New functions CONF_load_bio() and CONF_load_fp() to allow a config
17426 file to be loaded from a BIO or FILE pointer. The BIO version will
17427 for example allow memory BIOs to contain config info.
17431 * New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
17432 Whoever hopes to achieve shared-library compatibility across versions
17433 must use this, not the compile-time macro.
17434 (Exercise 0.9.4: Which is the minimum library version required by
17436 Note: All this applies only to multi-threaded programs, others don't
17441 * Add missing case to s3_clnt.c state machine -- one of the new SSL tests
17442 through a BIO pair triggered the default case, i.e.
17443 SSLerr(...,SSL_R_UNKNOWN_STATE).
17447 * New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
17448 can use the SSL library even if none of the specific BIOs is
17453 * Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
17454 for the encoded length.
17456 *Jeon KyoungHo <khjeon@sds.samsung.co.kr>*
17458 * Add initial documentation of the X509V3 functions.
17462 * Add a new pair of functions PEM_write_PKCS8PrivateKey() and
17463 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
17464 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
17465 secure PKCS#8 private key format with a high iteration count.
17469 * Fix determination of Perl interpreter: A perl or perl5
17470 *directory* in $PATH was also accepted as the interpreter.
17472 *Ralf S. Engelschall*
17474 * Fix demos/sign/sign.c: well there wasn't anything strictly speaking
17475 wrong with it but it was very old and did things like calling
17476 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
17477 unusual formatting.
17481 * Fix demos/selfsign.c: it used obsolete and deleted functions, changed
17482 to use the new extension code.
17486 * Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
17487 with macros. This should make it easier to change their form, add extra
17488 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
17493 * Add to configuration table a new entry that can specify an alternative
17494 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
17495 according to Mark Crispin <MRC@Panda.COM>.
17499 * DES CBC did not update the IV. Weird.
17503 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
17504 Changing the behaviour of the former might break existing programs --
17505 where IV updating is needed, des_ncbc_encrypt can be used.
17508 * When bntest is run from "make test" it drives bc to check its
17509 calculations, as well as internally checking them. If an internal check
17510 fails, it needs to cause bc to give a non-zero result or make test carries
17511 on without noticing the failure. Fixed.
17515 * DES library cleanups.
17519 * Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
17520 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
17521 ciphers. NOTE: although the key derivation function has been verified
17522 against some published test vectors it has not been extensively tested
17523 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
17528 * Instead of "mkdir -p", which is not fully portable, use new
17529 Perl script "util/mkdir-p.pl".
17533 * Rewrite the way password based encryption (PBE) is handled. It used to
17534 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
17535 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
17536 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
17537 the 'parameter' field of the AlgorithmIdentifier is passed to the
17538 underlying key generation function so it must do its own ASN1 parsing.
17539 This has also changed the EVP_PBE_CipherInit() function which now has a
17540 'parameter' argument instead of literal salt and iteration count values
17541 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
17545 * Support for PKCS#5 v1.5 compatible password based encryption algorithms
17546 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
17547 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
17548 KEY" because this clashed with PKCS#8 unencrypted string. Since this
17549 value was just used as a "magic string" and not used directly its
17550 value doesn't matter.
17554 * Introduce some semblance of const correctness to BN. Shame C doesn't
17559 * "linux-sparc64" configuration (ultrapenguin).
17561 *Ray Miller <ray.miller@oucs.ox.ac.uk>*
17562 "linux-sparc" configuration.
17564 *Christian Forster <fo@hawo.stw.uni-erlangen.de>*
17566 * config now generates no-xxx options for missing ciphers.
17570 * Support the EBCDIC character set (work in progress).
17571 File ebcdic.c not yet included because it has a different license.
17573 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
17575 * Support BS2000/OSD-POSIX.
17577 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
17579 * Make callbacks for key generation use `void *` instead of `char *`.
17583 * Make S/MIME samples compile (not yet tested).
17587 * Additional typesafe stacks.
17591 * New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
17595 ### Changes between 0.9.3 and 0.9.3a [29 May 1999]
17597 * New configuration variant "sco5-gcc".
17599 * Updated some demos.
17601 *Sean O Riordain, Wade Scholine*
17603 * Add missing BIO_free at exit of pkcs12 application.
17607 * Fix memory leak in conf.c.
17611 * Updates for Win32 to assembler version of MD5.
17615 * Set #! path to perl in `apps/der_chop` to where we found it
17616 instead of using a fixed path.
17620 * SHA library changes for irix64-mips4-cc.
17624 * Improvements for VMS support.
17628 ### Changes between 0.9.2b and 0.9.3 [24 May 1999]
17630 * Bignum library bug fix. IRIX 6 passes "make test" now!
17631 This also avoids the problems with SC4.2 and unpatched SC5.
17633 *Andy Polyakov <appro@fy.chalmers.se>*
17635 * New functions sk_num, sk_value and sk_set to replace the previous macros.
17636 These are required because of the typesafe stack would otherwise break
17637 existing code. If old code used a structure member which used to be STACK
17638 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
17639 sk_num or sk_value it would produce an error because the num, data members
17640 are not present in STACK_OF. Now it just produces a warning. sk_set
17641 replaces the old method of assigning a value to sk_value
17642 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
17643 that does this will no longer work (and should use sk_set instead) but
17644 this could be regarded as a "questionable" behaviour anyway.
17648 * Fix most of the other PKCS#7 bugs. The "experimental" code can now
17649 correctly handle encrypted S/MIME data.
17653 * Change type of various DES function arguments from des_cblock
17654 (which means, in function argument declarations, pointer to char)
17655 to des_cblock * (meaning pointer to array with 8 char elements),
17656 which allows the compiler to do more typechecking; it was like
17657 that back in SSLeay, but with lots of ugly casts.
17659 Introduce new type const_des_cblock.
17663 * Reorganise the PKCS#7 library and get rid of some of the more obvious
17664 problems: find RecipientInfo structure that matches recipient certificate
17665 and initialise the ASN1 structures properly based on passed cipher.
17669 * Belatedly make the BN tests actually check the results.
17673 * Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
17674 to and from BNs: it was completely broken. New compilation option
17675 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
17676 key elements as negative integers.
17680 * Reorganize and speed up MD5.
17682 *Andy Polyakov <appro@fy.chalmers.se>*
17686 *Richard Levitte <richard@levitte.org>*
17688 * New option -out to asn1parse to allow the parsed structure to be
17689 output to a file. This is most useful when combined with the -strparse
17690 option to examine the output of things like OCTET STRINGS.
17694 * Make SSL library a little more fool-proof by not requiring any longer
17695 that `SSL_set_{accept,connect}_state` be called before
17696 `SSL_{accept,connect}` may be used (`SSL_set_..._state` is omitted
17697 in many applications because usually everything *appeared* to work as
17698 intended anyway -- now it really works as intended).
17702 * Move openssl.cnf out of lib/.
17706 * Fix various things to let OpenSSL even pass "egcc -pipe -O2 -Wall
17707 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
17708 -Wmissing-declarations -Wnested-externs -Winline" with EGCS 1.1.2+
17710 *Ralf S. Engelschall*
17712 * Various fixes to the EVP and PKCS#7 code. It may now be able to
17713 handle PKCS#7 enveloped data properly.
17715 *Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve*
17717 * Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
17718 copying pointers. The cert_st handling is changed by this in
17719 various ways (and thus what used to be known as ctx->default_cert
17720 is now called ctx->cert, since we don't resort to `s->ctx->[default_]cert`
17721 any longer when s->cert does not give us what we need).
17722 ssl_cert_instantiate becomes obsolete by this change.
17723 As soon as we've got the new code right (possibly it already is?),
17724 we have solved a couple of bugs of the earlier code where s->cert
17725 was used as if it could not have been shared with other SSL structures.
17727 Note that using the SSL API in certain dirty ways now will result
17728 in different behaviour than observed with earlier library versions:
17729 Changing settings for an `SSL_CTX *ctx` after having done s = SSL_new(ctx)
17730 does not influence s as it used to.
17732 In order to clean up things more thoroughly, inside SSL_SESSION
17733 we don't use CERT any longer, but a new structure SESS_CERT
17734 that holds per-session data (if available); currently, this is
17735 the peer's certificate chain and, for clients, the server's certificate
17736 and temporary key. CERT holds only those values that can have
17737 meaningful defaults in an SSL_CTX.
17741 * New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
17742 from the internal representation. Various PKCS#7 fixes: remove some
17743 evil casts and set the enc_dig_alg field properly based on the signing
17748 * Allow PKCS#12 password to be set from the command line or the
17749 environment. Let 'ca' get its config file name from the environment
17750 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
17755 * Allow certificate policies extension to use an IA5STRING for the
17756 organization field. This is contrary to the PKIX definition but
17757 VeriSign uses it and IE5 only recognises this form. Document 'x509'
17762 * Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
17763 without disallowing inline assembler and the like for non-pedantic builds.
17767 * Support Borland C++ builder.
17769 *Janez Jere <jj@void.si>, modified by Ulf Möller*
17775 * SHA-1 cleanups and performance enhancements.
17777 *Andy Polyakov <appro@fy.chalmers.se>*
17779 * Sparc v8plus assembler for the bignum library.
17781 *Andy Polyakov <appro@fy.chalmers.se>*
17783 * Accept any -xxx and +xxx compiler options in Configure.
17787 * Update HPUX configuration.
17791 * Add missing `sk_<type>_unshift()` function to safestack.h
17793 *Ralf S. Engelschall*
17795 * New function SSL_CTX_use_certificate_chain_file that sets the
17796 "extra_cert"s in addition to the certificate. (This makes sense
17797 only for "PEM" format files, as chains as a whole are not
17802 * Support verify_depth from the SSL API.
17803 x509_vfy.c had what can be considered an off-by-one-error:
17804 Its depth (which was not part of the external interface)
17805 was actually counting the number of certificates in a chain;
17806 now it really counts the depth.
17810 * Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
17811 instead of X509err, which often resulted in confusing error
17812 messages since the error codes are not globally unique
17813 (e.g. an alleged error in ssl3_accept when a certificate
17814 didn't match the private key).
17816 * New function SSL_CTX_set_session_id_context that allows to set a default
17817 value (so that you don't need SSL_set_session_id_context for each
17818 connection using the SSL_CTX).
17822 * OAEP decoding bug fix.
17826 * Support INSTALL_PREFIX for package builders, as proposed by
17831 * New Configure options "threads" and "no-threads". For systems
17832 where the proper compiler options are known (currently Solaris
17833 and Linux), "threads" is the default.
17837 * New script util/mklink.pl as a faster substitute for util/mklink.sh.
17841 * Install various scripts to $(OPENSSLDIR)/misc, not to
17842 $(INSTALLTOP)/bin -- they shouldn't clutter directories
17843 such as /usr/local/bin.
17847 * "make linux-shared" to build shared libraries.
17849 *Niels Poppe <niels@netbox.org>*
17851 * New Configure option `no-<cipher>` (rsa, idea, rc5, ...).
17855 * Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
17856 extension adding in x509 utility.
17860 * Remove NOPROTO sections and error code comments.
17864 * Partial rewrite of the DEF file generator to now parse the ANSI
17869 * New Configure options --prefix=DIR and --openssldir=DIR.
17873 * Complete rewrite of the error code script(s). It is all now handled
17874 by one script at the top level which handles error code gathering,
17875 header rewriting and C source file generation. It should be much better
17876 than the old method: it now uses a modified version of Ulf's parser to
17877 read the ANSI prototypes in all header files (thus the old K&R definitions
17878 aren't needed for error creation any more) and do a better job of
17879 translating function codes into names. The old 'ASN1 error code embedded
17880 in a comment' is no longer necessary and it doesn't use .err files which
17881 have now been deleted. Also the error code call doesn't have to appear all
17882 on one line (which resulted in some large lines...).
17886 * Change #include filenames from `<foo.h>` to `<openssl/foo.h>`.
17890 * Change behaviour of ssl2_read when facing length-0 packets: Don't return
17891 0 (which usually indicates a closed connection), but continue reading.
17895 * Fix some race conditions.
17899 * Add support for CRL distribution points extension. Add Certificate
17900 Policies and CRL distribution points documentation.
17904 * Move the autogenerated header file parts to crypto/opensslconf.h.
17908 * Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
17909 8 of keying material. Merlin has also confirmed interop with this fix
17910 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
17912 *Merlin Hughes <merlin@baltimore.ie>*
17914 * Fix lots of warnings.
17916 *Richard Levitte <levitte@stacken.kth.se>*
17918 * In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
17919 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
17921 *Richard Levitte <levitte@stacken.kth.se>*
17923 * Fix problems with sizeof(long) == 8.
17925 *Andy Polyakov <appro@fy.chalmers.se>*
17927 * Change functions to ANSI C.
17931 * Fix typos in error codes.
17933 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller*
17935 * Remove defunct assembler files from Configure.
17939 * SPARC v8 assembler BIGNUM implementation.
17941 *Andy Polyakov <appro@fy.chalmers.se>*
17943 * Support for Certificate Policies extension: both print and set.
17944 Various additions to support the r2i method this uses.
17948 * A lot of constification, and fix a bug in X509_NAME_oneline() that could
17949 return a const string when you are expecting an allocated buffer.
17953 * Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
17954 types DirectoryString and DisplayText.
17958 * Add code to allow r2i extensions to access the configuration database,
17959 add an LHASH database driver and add several ctx helper functions.
17963 * Fix an evil bug in bn_expand2() which caused various BN functions to
17964 fail when they extended the size of a BIGNUM.
17968 * Various utility functions to handle SXNet extension. Modify mkdef.pl to
17969 support typesafe stack.
17973 * Fix typo in SSL_[gs]et_options().
17975 *Nils Frostberg <nils@medcom.se>*
17977 * Delete various functions and files that belonged to the (now obsolete)
17978 old X509V3 handling code.
17982 * New Configure option "rsaref".
17986 * Don't auto-generate pem.h.
17990 * Introduce type-safe ASN.1 SETs.
17994 * Convert various additional casted stacks to type-safe STACK_OF() variants.
17996 *Ben Laurie, Ralf S. Engelschall, Steve Henson*
17998 * Introduce type-safe STACKs. This will almost certainly break lots of code
17999 that links with OpenSSL (well at least cause lots of warnings), but fear
18000 not: the conversion is trivial, and it eliminates loads of evil casts. A
18001 few STACKed things have been converted already. Feel free to convert more.
18002 In the fullness of time, I'll do away with the STACK type altogether.
18006 * Add `openssl ca -revoke <certfile>` facility which revokes a certificate
18007 specified in `<certfile>` by updating the entry in the index.txt file.
18008 This way one no longer has to edit the index.txt file manually for
18009 revoking a certificate. The -revoke option does the gory details now.
18011 *Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall*
18013 * Fix `openssl crl -noout -text` combination where `-noout` killed the
18014 `-text` option at all and this way the `-noout -text` combination was
18015 inconsistent in `openssl crl` with the friends in `openssl x509|rsa|dsa`.
18017 *Ralf S. Engelschall*
18019 * Make sure a corresponding plain text error message exists for the
18020 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
18021 verify callback function determined that a certificate was revoked.
18023 *Ralf S. Engelschall*
18025 * Bugfix: In test/testenc, don't test `openssl <cipher>` for
18026 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
18027 all available ciphers including rc5, which was forgotten until now.
18028 In order to let the testing shell script know which algorithms
18029 are available, a new (up to now undocumented) command
18030 `openssl list-cipher-commands` is used.
18034 * Bugfix: s_client occasionally would sleep in select() when
18035 it should have checked SSL_pending() first.
18039 * New functions DSA_do_sign and DSA_do_verify to provide access to
18040 the raw DSA values prior to ASN.1 encoding.
18044 * Tweaks to Configure
18046 *Niels Poppe <niels@netbox.org>*
18048 * Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
18053 * New variables $(RANLIB) and $(PERL) in the Makefiles.
18057 * New config option to avoid instructions that are illegal on the 80386.
18058 The default code is faster, but requires at least a 486.
18062 * Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
18063 SSL2_SERVER_VERSION (not used at all) macros, which are now the
18064 same as SSL2_VERSION anyway.
18068 * New "-showcerts" option for s_client.
18072 * Still more PKCS#12 integration. Add pkcs12 application to openssl
18073 application. Various cleanups and fixes.
18077 * More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
18078 modify error routines to work internally. Add error codes and PBE init
18079 to library startup routines.
18083 * Further PKCS#12 integration. Added password based encryption, PKCS#8 and
18084 packing functions to asn1 and evp. Changed function names and error
18085 codes along the way.
18089 * PKCS12 integration: and so it begins... First of several patches to
18090 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
18091 objects to objects.h
18095 * Add a new 'indent' option to some X509V3 extension code. Initial ASN1
18096 and display support for Thawte strong extranet extension.
18100 * Add LinuxPPC support.
18102 *Jeff Dubrule <igor@pobox.org>*
18104 * Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
18105 bn_div_words in alpha.s.
18107 *Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie*
18109 * Make sure the RSA OAEP test is skipped under -DRSAref because
18110 OAEP isn't supported when OpenSSL is built with RSAref.
18112 *Ulf Moeller <ulf@fitug.de>*
18114 * Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
18115 so they no longer are missing under -DNOPROTO.
18117 *Soren S. Jorvang <soren@t.dk>*
18119 ### Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
18121 * Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
18122 doesn't work when the session is reused. Coming soon!
18126 * Fix a security hole, that allows sessions to be reused in the wrong
18127 context thus bypassing client cert protection! All software that uses
18128 client certs and session caches in multiple contexts NEEDS PATCHING to
18129 allow session reuse! A fuller solution is in the works.
18131 *Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)*
18133 * Some more source tree cleanups (removed obsolete files
18134 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
18135 permission on "config" script to be executable) and a fix for the INSTALL
18138 *Ulf Moeller <ulf@fitug.de>*
18140 * Remove some legacy and erroneous uses of malloc, free instead of
18143 *Lennart Bang <lob@netstream.se>, with minor changes by Steve*
18145 * Make rsa_oaep_test return non-zero on error.
18147 *Ulf Moeller <ulf@fitug.de>*
18149 * Add support for native Solaris shared libraries. Configure
18150 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
18151 if someone would make that last step automatic.
18153 *Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>*
18155 * ctx_size was not built with the right compiler during "make links". Fixed.
18159 * Change the meaning of 'ALL' in the cipher list. It now means "everything
18160 except NULL ciphers". This means the default cipher list will no longer
18161 enable NULL ciphers. They need to be specifically enabled e.g. with
18162 the string "DEFAULT:eNULL".
18166 * Fix to RSA private encryption routines: if p < q then it would
18167 occasionally produce an invalid result. This will only happen with
18168 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
18172 * Be less restrictive and allow also `perl util/perlpath.pl
18173 /path/to/bin/perl` in addition to `perl util/perlpath.pl /path/to/bin`,
18174 because this way one can also use an interpreter named `perl5` (which is
18175 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
18176 installed as `perl`).
18178 *Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
18180 * Let util/clean-depend.pl work also with older Perl 5.00x versions.
18182 *Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
18184 * Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
18185 advapi32.lib to Win32 build and change the pem test comparison
18186 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
18187 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
18188 and crypto/des/ede_cbcm_enc.c.
18192 * DES quad checksum was broken on big-endian architectures. Fixed.
18196 * Comment out two functions in bio.h that aren't implemented. Fix up the
18197 Win32 test batch file so it (might) work again. The Win32 test batch file
18198 is horrible: I feel ill....
18202 * Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
18203 in e_os.h. Audit of header files to check ANSI and non ANSI
18204 sections: 10 functions were absent from non ANSI section and not exported
18205 from Windows DLLs. Fixed up libeay.num for new functions.
18209 * Make `openssl version` output lines consistent.
18211 *Ralf S. Engelschall*
18213 * Fix Win32 symbol export lists for BIO functions: Added
18214 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
18215 to ms/libeay{16,32}.def.
18217 *Ralf S. Engelschall*
18219 * Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
18220 fine under Unix and passes some trivial tests I've now added. But the
18221 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
18222 added to make sure no one expects that this stuff really works in the
18223 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
18224 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
18227 *Ralf S. Engelschall*
18229 * Fix the generation of two part addresses in perl.
18231 *Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie*
18233 * Add config entry for Linux on MIPS.
18235 *John Tobey <jtobey@channel1.com>*
18237 * Make links whenever Configure is run, unless we are on Windoze.
18241 * Permit extensions to be added to CRLs using crl_section in openssl.cnf.
18242 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
18247 * Add a useful kludge to allow package maintainers to specify compiler and
18248 other platforms details on the command line without having to patch the
18249 Configure script every time: One now can use
18250 `perl Configure <id>:<details>`,
18251 i.e. platform ids are allowed to have details appended
18252 to them (separated by colons). This is treated as there would be a static
18253 pre-configured entry in Configure's %table under key `<id>` with value
18254 `<details>` and `perl Configure <id>` is called. So, when you want to
18255 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
18256 assembler stuff you can use `perl Configure "FreeBSD-elf:pgcc:-O6:::"`
18257 now, which overrides the FreeBSD-elf entry on-the-fly.
18259 *Ralf S. Engelschall*
18261 * Disable new TLS1 ciphersuites by default: they aren't official yet.
18265 * Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
18266 on the `perl Configure ...` command line. This way one can compile
18267 OpenSSL libraries with Position Independent Code (PIC) which is needed
18268 for linking it into DSOs.
18270 *Ralf S. Engelschall*
18272 * Remarkably, export ciphers were totally broken and no-one had noticed!
18277 * Cleaned up the LICENSE document: The official contact for any license
18278 questions now is the OpenSSL core team under openssl-core@openssl.org.
18279 And add a paragraph about the dual-license situation to make sure people
18280 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
18281 to the OpenSSL toolkit.
18283 *Ralf S. Engelschall*
18285 * General source tree makefile cleanups: Made `making xxx in yyy...`
18286 display consistent in the source tree and replaced `/bin/rm` by `rm`.
18287 Additionally cleaned up the `make links` target: Remove unnecessary
18288 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
18289 to speed processing and no longer clutter the display with confusing
18290 stuff. Instead only the actually done links are displayed.
18292 *Ralf S. Engelschall*
18294 * Permit null encryption ciphersuites, used for authentication only. It used
18295 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
18296 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
18301 * Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
18302 signed attributes when verifying signatures (this would break them),
18303 the detached data encoding was wrong and public keys obtained using
18304 X509_get_pubkey() weren't freed.
18308 * Add text documentation for the BUFFER functions. Also added a work around
18309 to a Win95 console bug. This was triggered by the password read stuff: the
18310 last character typed gets carried over to the next fread(). If you were
18311 generating a new cert request using 'req' for example then the last
18312 character of the passphrase would be CR which would then enter the first
18317 * Added the new 'Includes OpenSSL Cryptography Software' button as
18318 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
18319 button and can be used by applications based on OpenSSL to show the
18320 relationship to the OpenSSL project.
18322 *Ralf S. Engelschall*
18324 * Remove confusing variables in function signatures in files
18325 ssl/ssl_lib.c and ssl/ssl.h.
18327 *Lennart Bong <lob@kulthea.stacken.kth.se>*
18329 * Don't install bss_file.c under PREFIX/include/
18331 *Lennart Bong <lob@kulthea.stacken.kth.se>*
18333 * Get the Win32 compile working again. Modify mkdef.pl so it can handle
18334 functions that return function pointers and has support for NT specific
18335 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
18336 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
18337 unsigned to signed types: this was killing the Win32 compile.
18341 * Add new certificate file to stack functions,
18342 SSL_add_dir_cert_subjects_to_stack() and
18343 SSL_add_file_cert_subjects_to_stack(). These largely supplant
18344 SSL_load_client_CA_file(), and can be used to add multiple certs easily
18345 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
18346 This means that Apache-SSL and similar packages don't have to mess around
18347 to add as many CAs as they want to the preferred list.
18351 * Experiment with doxygen documentation. Currently only partially applied to
18353 See <http://www.stack.nl/~dimitri/doxygen/index.html>, and run doxygen with
18354 openssl.doxy as the configuration file.
18358 * Get rid of remaining C++-style comments which strict C compilers hate.
18360 *Ralf S. Engelschall, pointed out by Carlos Amengual*
18362 * Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
18363 compiled in by default: it has problems with large keys.
18367 * Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
18368 DH private keys and/or callback functions which directly correspond to
18369 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
18370 is needed for applications which have to configure certificates on a
18371 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
18373 For the RSA certificate situation is makes no difference, but
18374 for the DSA certificate situation this fixes the "no shared cipher"
18375 problem where the OpenSSL cipher selection procedure failed because the
18376 temporary keys were not overtaken from the context and the API provided
18377 no way to reconfigure them.
18378 The new functions now let applications reconfigure the stuff and they
18379 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
18380 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
18381 non-public-API function ssl_cert_instantiate() is used as a helper
18382 function and also to reduce code redundancy inside ssl_rsa.c.
18384 *Ralf S. Engelschall*
18386 * Move s_server -dcert and -dkey options out of the undocumented feature
18387 area because they are useful for the DSA situation and should be
18388 recognized by the users.
18390 *Ralf S. Engelschall*
18392 * Fix the cipher decision scheme for export ciphers: the export bits are
18393 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
18394 SSL_EXP_MASK. So, the original variable has to be used instead of the
18395 already masked variable.
18397 *Richard Levitte <levitte@stacken.kth.se>*
18399 * Fix `port` variable from `int` to `unsigned int` in crypto/bio/b_sock.c
18401 *Richard Levitte <levitte@stacken.kth.se>*
18403 * Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
18404 from `int` to `unsigned int` because it is a length and initialized by
18405 EVP_DigestFinal() which expects an `unsigned int *`.
18407 *Richard Levitte <levitte@stacken.kth.se>*
18409 * Don't hard-code path to Perl interpreter on shebang line of Configure
18410 script. Instead use the usual Shell->Perl transition trick.
18412 *Ralf S. Engelschall*
18414 * Make `openssl x509 -noout -modulus`' functional also for DSA certificates
18415 (in addition to RSA certificates) to match the behaviour of `openssl dsa
18416 -noout -modulus` as it's already the case for `openssl rsa -noout
18417 -modulus`. For RSA the -modulus is the real "modulus" while for DSA
18418 currently the public key is printed (a decision which was already done by
18419 `openssl dsa -modulus` in the past) which serves a similar purpose.
18420 Additionally the NO_RSA no longer completely removes the whole -modulus
18421 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
18424 *Ralf S. Engelschall*
18426 * Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
18427 BIO. See the source (crypto/evp/bio_ok.c) for more info.
18429 *Arne Ansper <arne@ats.cyber.ee>*
18431 * Dump the old yucky req code that tried (and failed) to allow raw OIDs
18432 to be added. Now both 'req' and 'ca' can use new objects defined in the
18437 * Add cool BIO that does syslog (or event log on NT).
18439 *Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie*
18441 * Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
18442 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
18443 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
18444 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
18448 * Add preliminary config info for new extension code.
18452 * Make RSA_NO_PADDING really use no padding.
18454 *Ulf Moeller <ulf@fitug.de>*
18456 * Generate errors when private/public key check is done.
18460 * Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
18461 for some CRL extensions and new objects added.
18465 * Really fix the ASN1 IMPLICIT bug this time... Partial support for private
18466 key usage extension and fuller support for authority key id.
18470 * Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
18471 padding method for RSA, which is recommended for new applications in PKCS
18472 #1 v2.0 (RFC 2437, October 1998).
18473 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
18474 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
18475 against Bleichbacher's attack on RSA.
18476 *Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
18479 * Updates to the new SSL compression code
18481 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
18483 * Fix so that the version number in the master secret, when passed
18484 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
18485 (because the server will not accept higher), that the version number
18486 is 0x03,0x01, not 0x03,0x00
18488 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
18490 * Run extensive memory leak checks on SSL commands. Fixed *lots* of memory
18491 leaks in `ssl/` relating to new `X509_get_pubkey()` behaviour. Also fixes
18492 in `apps/` and an unrelated leak in `crypto/dsa/dsa_vrf.c`.
18496 * Support for RAW extensions where an arbitrary extension can be
18497 created by including its DER encoding. See `apps/openssl.cnf` for
18502 * Make sure latest Perl versions don't interpret some generated C array
18503 code as Perl array code in the crypto/err/err_genc.pl script.
18505 *Lars Weber <3weber@informatik.uni-hamburg.de>*
18507 * Modify ms/do_ms.bat to not generate assembly language makefiles since
18508 not many people have the assembler. Various Win32 compilation fixes and
18509 update to the INSTALL.W32 file with (hopefully) more accurate Win32
18510 build instructions.
18514 * Modify configure script 'Configure' to automatically create crypto/date.h
18515 file under Win32 and also build pem.h from pem.org. New script
18516 util/mkfiles.pl to create the MINFO file on environments that can't do a
18517 'make files': perl util/mkfiles.pl >MINFO should work.
18521 * Major rework of DES function declarations, in the pursuit of correctness
18522 and purity. As a result, many evil casts evaporated, and some weirdness,
18523 too. You may find this causes warnings in your code. Zapping your evil
18524 casts will probably fix them. Mostly.
18528 * Fix for a typo in asn1.h. Bug fix to object creation script
18529 obj_dat.pl. It considered a zero in an object definition to mean
18530 "end of object": none of the objects in objects.h have any zeros
18531 so it wasn't spotted.
18533 *Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>*
18535 * Add support for Triple DES Cipher Block Chaining with Output Feedback
18536 Masking (CBCM). In the absence of test vectors, the best I have been able
18537 to do is check that the decrypt undoes the encrypt, so far. Send me test
18538 vectors if you have them.
18542 * Correct calculation of key length for export ciphers (too much space was
18543 allocated for null ciphers). This has not been tested!
18547 * Modifications to the mkdef.pl for Win32 DEF file creation. The usage
18548 message is now correct (it understands "crypto" and "ssl" on its
18549 command line). There is also now an "update" option. This will update
18550 the util/ssleay.num and util/libeay.num files with any new functions.
18552 perl util/mkdef.pl crypto ssl update
18553 it will update them.
18557 * Overhauled the Perl interface:
18558 - ported BN stuff to OpenSSL's different BN library
18559 - made the perl/ source tree CVS-aware
18560 - renamed the package from SSLeay to OpenSSL (the files still contain
18561 their history because I've copied them in the repository)
18562 - removed obsolete files (the test scripts will be replaced
18563 by better Test::Harness variants in the future)
18565 *Ralf S. Engelschall*
18567 * First cut for a very conservative source tree cleanup:
18568 1. merge various obsolete readme texts into doc/ssleay.txt
18569 where we collect the old documents and readme texts.
18570 2. remove the first part of files where I'm already sure that we no
18571 longer need them because of three reasons: either they are just temporary
18572 files which were left by Eric or they are preserved original files where
18573 I've verified that the diff is also available in the CVS via "cvs diff
18574 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
18575 the crypto/md/ stuff).
18577 *Ralf S. Engelschall*
18579 * More extension code. Incomplete support for subject and issuer alt
18580 name, issuer and authority key id. Change the i2v function parameters
18581 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
18582 what that's for :-) Fix to ASN1 macro which messed up
18583 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
18587 * Preliminary support for ENUMERATED type. This is largely copied from the
18592 * Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
18594 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
18596 * Make sure `make rehash` target really finds the `openssl` program.
18598 *Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
18600 * Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
18601 like to hear about it if this slows down other processors.
18605 * Add CygWin32 platform information to Configure script.
18607 *Alan Batie <batie@aahz.jf.intel.com>*
18609 * Fixed ms/32all.bat script: `no_asm` -> `no-asm`
18611 *Rainer W. Gerling <gerling@mpg-gv.mpg.de>*
18613 * New program nseq to manipulate netscape certificate sequences
18617 * Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
18622 * Fixes to BN code. Previously the default was to define BN_RECURSION
18623 but the BN code had some problems that would cause failures when
18624 doing certificate verification and some other functions.
18626 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
18628 * Add ASN1 and PEM code to support netscape certificate sequences.
18632 * Add ASN1 and PEM code to support netscape certificate sequences.
18636 * Add several PKIX and private extended key usage OIDs.
18640 * Modify the 'ca' program to handle the new extension code. Modify
18641 openssl.cnf for new extension format, add comments.
18645 * More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
18646 and add a sample to openssl.cnf so req -x509 now adds appropriate
18651 * Continued X509 V3 changes. Add to other makefiles, integrate with the
18652 error code, add initial support to X509_print() and x509 application.
18656 * Takes a deep breath and start adding X509 V3 extension support code. Add
18657 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
18658 stuff is currently isolated and isn't even compiled yet.
18662 * Continuing patches for GeneralizedTime. Fix up certificate and CRL
18663 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
18664 Removed the versions check from X509 routines when loading extensions:
18665 this allows certain broken certificates that don't set the version
18666 properly to be processed.
18670 * Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
18671 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
18672 can still be regenerated with "make depend".
18676 * Spelling mistake in C version of CAST-128.
18678 *Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>*
18680 * Changes to the error generation code. The perl script err-code.pl
18681 now reads in the old error codes and retains the old numbers, only
18682 adding new ones if necessary. It also only changes the .err files if new
18683 codes are added. The makefiles have been modified to only insert errors
18684 when needed (to avoid needlessly modifying header files). This is done
18685 by only inserting errors if the .err file is newer than the auto generated
18686 C file. To rebuild all the error codes from scratch (the old behaviour)
18687 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
18688 or delete all the .err files.
18692 * CAST-128 was incorrectly implemented for short keys. The C version has
18693 been fixed, but is untested. The assembler versions are also fixed, but
18694 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
18695 to regenerate it if needed.
18696 *Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
18697 Hagino <itojun@kame.net>*
18699 * File was opened incorrectly in randfile.c.
18701 *Ulf Möller <ulf@fitug.de>*
18703 * Beginning of support for GeneralizedTime. d2i, i2d, check and print
18704 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
18705 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
18706 al: it's just almost always a UTCTime. Note this patch adds new error
18707 codes so do a "make errors" if there are problems.
18711 * Correct Linux 1 recognition in config.
18713 *Ulf Möller <ulf@fitug.de>*
18715 * Remove pointless MD5 hash when using DSA keys in ca.
18717 *Anonymous <nobody@replay.com>*
18719 * Generate an error if given an empty string as a cert directory. Also
18720 generate an error if handed NULL (previously returned 0 to indicate an
18721 error, but didn't set one).
18723 *Ben Laurie, reported by Anonymous <nobody@replay.com>*
18725 * Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
18729 * Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
18730 parameters. This was causing a warning which killed off the Win32 compile.
18734 * Remove C++ style comments from crypto/bn/bn_local.h.
18736 *Neil Costigan <neil.costigan@celocom.com>*
18738 * The function OBJ_txt2nid was broken. It was supposed to return a nid
18739 based on a text string, looking up short and long names and finally
18740 "dot" format. The "dot" format stuff didn't work. Added new function
18741 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
18742 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
18743 OID is not part of the table.
18747 * Add prototypes to X509 lookup/verify methods, fixing a bug in
18748 X509_LOOKUP_by_alias().
18752 * Sort openssl functions by name.
18756 * Get the `gendsa` command working and add it to the `list` command. Remove
18757 encryption from sample DSA keys (in case anyone is interested the password
18762 * Make *all* `*_free` functions accept a NULL pointer.
18764 *Frans Heymans <fheymans@isaserver.be>*
18766 * If a DH key is generated in s3_srvr.c, don't blow it by trying to use
18769 *Anonymous <nobody@replay.com>*
18771 * s_server should send the CAfile as acceptable CAs, not its own cert.
18773 *Bodo Moeller <3moeller@informatik.uni-hamburg.de>*
18775 * Don't blow it for numeric `-newkey` arguments to `apps/req`.
18777 *Bodo Moeller <3moeller@informatik.uni-hamburg.de>*
18779 * Temp key "for export" tests were wrong in s3_srvr.c.
18781 *Anonymous <nobody@replay.com>*
18783 * Add prototype for temp key callback functions
18784 SSL_CTX_set_tmp_{rsa,dh}_callback().
18788 * Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
18789 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
18793 * X509_name_add_entry() freed the wrong thing after an error.
18795 *Arne Ansper <arne@ats.cyber.ee>*
18797 * rsa_eay.c would attempt to free a NULL context.
18799 *Arne Ansper <arne@ats.cyber.ee>*
18801 * BIO_s_socket() had a broken should_retry() on Windoze.
18803 *Arne Ansper <arne@ats.cyber.ee>*
18805 * BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
18807 *Arne Ansper <arne@ats.cyber.ee>*
18809 * Make sure the already existing X509_STORE->depth variable is initialized
18810 in X509_STORE_new(), but document the fact that this variable is still
18811 unused in the certificate verification process.
18813 *Ralf S. Engelschall*
18815 * Fix the various library and `apps/` files to free up pkeys obtained from
18816 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
18820 * Fix reference counting in X509_PUBKEY_get(). This makes
18821 demos/maurice/example2.c work, amongst others, probably.
18823 *Steve Henson and Ben Laurie*
18825 * First cut of a cleanup for `apps/`. First the `ssleay` program is now named
18826 `openssl` and second, the shortcut symlinks for the `openssl <command>`
18827 are no longer created. This way we have a single and consistent command
18828 line interface `openssl <command>`, similar to `cvs <command>`.
18830 *Ralf S. Engelschall, Paul Sutton and Ben Laurie*
18832 * ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
18833 BIT STRING wrapper always have zero unused bits.
18837 * Add CA.pl, perl version of CA.sh, add extended key usage OID.
18841 * Make the top-level INSTALL documentation easier to understand.
18845 * Makefiles updated to exit if an error occurs in a sub-directory
18846 make (including if user presses ^C) [Paul Sutton]
18848 * Make Montgomery context stuff explicit in RSA data structure.
18852 * Fix build order of pem and err to allow for generated pem.h.
18856 * Fix renumbering bug in X509_NAME_delete_entry().
18860 * Enhanced the err-ins.pl script so it makes the error library number
18861 global and can add a library name. This is needed for external ASN1 and
18862 other error libraries.
18866 * Fixed sk_insert which never worked properly.
18870 * Fix ASN1 macros so they can handle indefinite length constructed
18871 EXPLICIT tags. Some non standard certificates use these: they can now
18876 * Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
18877 into a single doc/ssleay.txt bundle. This way the information is still
18878 preserved but no longer messes up this directory. Now it's new room for
18879 the new set of documentation files.
18881 *Ralf S. Engelschall*
18883 * SETs were incorrectly DER encoded. This was a major pain, because they
18884 shared code with SEQUENCEs, which aren't coded the same. This means that
18885 almost everything to do with SETs or SEQUENCEs has either changed name or
18886 number of arguments.
18888 *Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>*
18890 * Fix test data to work with the above.
18894 * Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
18895 was already fixed by Eric for 0.9.1 it seems.
18897 *Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>*
18899 * Autodetect FreeBSD3.
18903 * Fix various bugs in Configure. This affects the following platforms:
18907 unixware-2.0-pentium
18912 * Eliminate generated files from CVS. Reorder tests to regenerate files
18913 before they are needed.
18917 * Generate Makefile.ssl from Makefile.org (to keep CVS happy).
18921 ### Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
18923 * Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
18924 changed SSLeay to OpenSSL in version strings.
18926 *Ralf S. Engelschall*
18928 * Some fixups to the top-level documents.
18932 * Fixed the nasty bug where rsaref.h was not found under compile-time
18933 because the symlink to include/ was missing.
18935 *Ralf S. Engelschall*
18937 * Incorporated the popular no-RSA/DSA-only patches
18938 which allow to compile a RSA-free SSLeay.
18940 *Andrew Cooke / Interrader Ldt., Ralf S. Engelschall*
18942 * Fixed nasty rehash problem under `make -f Makefile.ssl links`
18943 when "ssleay" is still not found.
18945 *Ralf S. Engelschall*
18947 * Added more platforms to Configure: Cray T3E, HPUX 11,
18949 *Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>*
18951 * Updated the README file.
18953 *Ralf S. Engelschall*
18955 * Added various .cvsignore files in the CVS repository subdirs
18956 to make a "cvs update" really silent.
18958 *Ralf S. Engelschall*
18960 * Recompiled the error-definition header files and added
18961 missing symbols to the Win32 linker tables.
18963 *Ralf S. Engelschall*
18965 * Cleaned up the top-level documents;
18966 o new files: CHANGES and LICENSE
18967 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
18968 o merged COPYRIGHT into LICENSE
18969 o removed obsolete TODO file
18970 o renamed MICROSOFT to INSTALL.W32
18972 *Ralf S. Engelschall*
18974 * Removed dummy files from the 0.9.1b source tree:
18975 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
18976 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
18977 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
18978 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
18979 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
18981 *Ralf S. Engelschall*
18983 * Added various platform portability fixes.
18987 * The Genesis of the OpenSSL rpject:
18988 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
18989 Young and Tim J. Hudson created while they were working for C2Net until
18992 *The OpenSSL Project*
18994 ### Changes between 0.9.0b and 0.9.1b [not released]
18996 * Updated a few CA certificates under certs/
19000 * Changed some BIGNUM api stuff.
19004 * Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
19005 DGUX x86, Linux Alpha, etc.
19009 * New COMP library [crypto/comp/] for SSL Record Layer Compression:
19010 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
19015 * Add -strparse option to asn1pars program which parses nested
19018 *Dr Stephen Henson <shenson@bigfoot.com>*
19020 * Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
19024 * DSA fix for "ca" program.
19028 * Added "-genkey" option to "dsaparam" program.
19032 * Added RIPE MD160 (rmd160) message digest.
19036 * Added -a (all) option to "ssleay version" command.
19040 * Added PLATFORM define which is the id given to Configure.
19044 * Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
19048 * Extended the ASN.1 parser routines.
19052 * Extended BIO routines to support REUSEADDR, seek, tell, etc.
19056 * Added a BN_CTX to the BN library.
19060 * Fixed the weak key values in DES library
19064 * Changed API in EVP library for cipher aliases.
19068 * Added support for RC2/64bit cipher.
19072 * Converted the lhash library to the crypto/mem.c functions.
19076 * Added more recognized ASN.1 object ids.
19080 * Added more RSA padding checks for SSL/TLS.
19084 * Added BIO proxy/filter functionality.
19088 * Added extra_certs to SSL_CTX which can be used
19089 send extra CA certificates to the client in the CA cert chain sending
19090 process. It can be configured with SSL_CTX_add_extra_chain_cert().
19094 * Now Fortezza is denied in the authentication phase because
19095 this is key exchange mechanism is not supported by SSLeay at all.
19099 * Additional PKCS1 checks.
19103 * Support the string "TLSv1" for all TLS v1 ciphers.
19107 * Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
19108 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
19112 * Fixed a few memory leaks.
19116 * Fixed various code and comment typos.
19120 * A minor bug in ssl/s3_clnt.c where there would always be 4 0
19121 bytes sent in the client random.
19123 *Edward Bishop <ebishop@spyglass.com>*
19127 [CVE-2020-1971]: https://www.openssl.org/news/vulnerabilities.html#CVE-2020-1971
19128 [CVE-2020-1967]: https://www.openssl.org/news/vulnerabilities.html#CVE-2020-1967
19129 [CVE-2019-1563]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1563
19130 [CVE-2019-1559]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1559
19131 [CVE-2019-1552]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1552
19132 [CVE-2019-1551]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1551
19133 [CVE-2019-1549]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1549
19134 [CVE-2019-1547]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1547
19135 [CVE-2019-1543]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1543
19136 [CVE-2018-5407]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-5407
19137 [CVE-2018-0739]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0739
19138 [CVE-2018-0737]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0737
19139 [CVE-2018-0735]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0735
19140 [CVE-2018-0734]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0734
19141 [CVE-2018-0733]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0733
19142 [CVE-2018-0732]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0732
19143 [CVE-2017-3738]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3738
19144 [CVE-2017-3737]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3737
19145 [CVE-2017-3736]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3736
19146 [CVE-2017-3735]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3735
19147 [CVE-2017-3733]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3733
19148 [CVE-2017-3732]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3732
19149 [CVE-2017-3731]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3731
19150 [CVE-2017-3730]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3730
19151 [CVE-2016-7055]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7055
19152 [CVE-2016-7054]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7054
19153 [CVE-2016-7053]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7053
19154 [CVE-2016-7052]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7052
19155 [CVE-2016-6309]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6309
19156 [CVE-2016-6308]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6308
19157 [CVE-2016-6307]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6307
19158 [CVE-2016-6306]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6306
19159 [CVE-2016-6305]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6305
19160 [CVE-2016-6304]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6304
19161 [CVE-2016-6303]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6303
19162 [CVE-2016-6302]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6302
19163 [CVE-2016-2183]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2183
19164 [CVE-2016-2182]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2182
19165 [CVE-2016-2181]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2181
19166 [CVE-2016-2180]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2180
19167 [CVE-2016-2179]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2179
19168 [CVE-2016-2178]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2178
19169 [CVE-2016-2177]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2177
19170 [CVE-2016-2176]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2176
19171 [CVE-2016-2109]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2109
19172 [CVE-2016-2107]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2107
19173 [CVE-2016-2106]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2106
19174 [CVE-2016-2105]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2105
19175 [CVE-2016-0800]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0800
19176 [CVE-2016-0799]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0799
19177 [CVE-2016-0798]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0798
19178 [CVE-2016-0797]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0797
19179 [CVE-2016-0705]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0705
19180 [CVE-2016-0702]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0702
19181 [CVE-2016-0701]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0701
19182 [CVE-2015-3197]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3197
19183 [CVE-2015-3196]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3196
19184 [CVE-2015-3195]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3195
19185 [CVE-2015-3194]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3194
19186 [CVE-2015-3193]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3193
19187 [CVE-2015-1793]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1793
19188 [CVE-2015-1792]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1792
19189 [CVE-2015-1791]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1791
19190 [CVE-2015-1790]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1790
19191 [CVE-2015-1789]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1789
19192 [CVE-2015-1788]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1788
19193 [CVE-2015-1787]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1787
19194 [CVE-2015-0293]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0293
19195 [CVE-2015-0291]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0291
19196 [CVE-2015-0290]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0290
19197 [CVE-2015-0289]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0289
19198 [CVE-2015-0288]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0288
19199 [CVE-2015-0287]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0287
19200 [CVE-2015-0286]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0286
19201 [CVE-2015-0285]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0285
19202 [CVE-2015-0209]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0209
19203 [CVE-2015-0208]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0208
19204 [CVE-2015-0207]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0207
19205 [CVE-2015-0206]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0206
19206 [CVE-2015-0205]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0205
19207 [CVE-2015-0204]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0204
19208 [CVE-2014-8275]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-8275
19209 [CVE-2014-5139]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-5139
19210 [CVE-2014-3572]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3572
19211 [CVE-2014-3571]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3571
19212 [CVE-2014-3570]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3570
19213 [CVE-2014-3569]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3569
19214 [CVE-2014-3568]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3568
19215 [CVE-2014-3567]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3567
19216 [CVE-2014-3566]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3566
19217 [CVE-2014-3513]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3513
19218 [CVE-2014-3512]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3512
19219 [CVE-2014-3511]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3511
19220 [CVE-2014-3510]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3510
19221 [CVE-2014-3509]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3509
19222 [CVE-2014-3508]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3508
19223 [CVE-2014-3507]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3507
19224 [CVE-2014-3506]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3506
19225 [CVE-2014-3505]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3505
19226 [CVE-2014-3470]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3470
19227 [CVE-2014-0224]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0224
19228 [CVE-2014-0221]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0221
19229 [CVE-2014-0195]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0195
19230 [CVE-2014-0160]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0160
19231 [CVE-2014-0076]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0076
19232 [CVE-2013-6450]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-6450
19233 [CVE-2013-4353]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-4353
19234 [CVE-2013-0169]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-0169
19235 [CVE-2013-0166]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-0166
19236 [CVE-2012-2686]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2686
19237 [CVE-2012-2333]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2333
19238 [CVE-2012-2110]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2110
19239 [CVE-2012-0884]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0884
19240 [CVE-2012-0050]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0050
19241 [CVE-2012-0027]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0027
19242 [CVE-2011-4619]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4619
19243 [CVE-2011-4577]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4577
19244 [CVE-2011-4576]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4576
19245 [CVE-2011-4109]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4109
19246 [CVE-2011-4108]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4108
19247 [CVE-2011-3210]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-3210
19248 [CVE-2011-3207]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-3207
19249 [CVE-2011-0014]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-0014
19250 [CVE-2010-4252]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-4252
19251 [CVE-2010-4180]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-4180
19252 [CVE-2010-3864]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-3864
19253 [CVE-2010-1633]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-1633
19254 [CVE-2010-0740]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-0740
19255 [CVE-2010-0433]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-0433
19256 [CVE-2009-4355]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-4355
19257 [CVE-2009-3555]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-3555
19258 [CVE-2009-3245]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-3245
19259 [CVE-2009-1386]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1386
19260 [CVE-2009-1379]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1379
19261 [CVE-2009-1378]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1378
19262 [CVE-2009-1377]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1377
19263 [CVE-2009-0789]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0789
19264 [CVE-2009-0591]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0591
19265 [CVE-2009-0590]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0590
19266 [CVE-2008-5077]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-5077
19267 [CVE-2008-1678]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-1678
19268 [CVE-2008-1672]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-1672
19269 [CVE-2008-0891]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-0891
19270 [CVE-2007-5135]: https://www.openssl.org/news/vulnerabilities.html#CVE-2007-5135
19271 [CVE-2007-4995]: https://www.openssl.org/news/vulnerabilities.html#CVE-2007-4995
19272 [CVE-2006-4343]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-4343
19273 [CVE-2006-4339]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-4339
19274 [CVE-2006-3738]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-3738
19275 [CVE-2006-2940]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-2940
19276 [CVE-2006-2937]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-2937
19277 [CVE-2005-2969]: https://www.openssl.org/news/vulnerabilities.html#CVE-2005-2969
19278 [CVE-2004-0112]: https://www.openssl.org/news/vulnerabilities.html#CVE-2004-0112
19279 [CVE-2004-0079]: https://www.openssl.org/news/vulnerabilities.html#CVE-2004-0079
19280 [CVE-2003-0851]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0851
19281 [CVE-2003-0545]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0545
19282 [CVE-2003-0544]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0544
19283 [CVE-2003-0543]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0543
19284 [CVE-2003-0078]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0078
19285 [CVE-2002-0659]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0659
19286 [CVE-2002-0657]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0657
19287 [CVE-2002-0656]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0656
19288 [CVE-2002-0655]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0655