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 3.0](#openssl-30)
14 - [OpenSSL 1.1.1](#openssl-111)
15 - [OpenSSL 1.1.0](#openssl-110)
16 - [OpenSSL 1.0.2](#openssl-102)
17 - [OpenSSL 1.0.1](#openssl-101)
18 - [OpenSSL 1.0.0](#openssl-100)
19 - [OpenSSL 0.9.x](#openssl-09x)
24 For OpenSSL 3.0 a [Migration guide][] has been added, so the CHANGES entries
25 listed here are only a brief description.
26 The migration guide contains more detailed information related to new features,
27 breaking changes, and mappings for the large list of deprecated functions.
29 [Migration guide]: https://github.com/openssl/openssl/tree/master/doc/man7/migration_guide.pod
31 ### Changes between 3.0.1 and 3.0.2 [15 Mar 2022]
33 * Fixed a bug in the BN_mod_sqrt() function that can cause it to loop forever
36 Internally this function is used when parsing certificates that contain
37 elliptic curve public keys in compressed form or explicit elliptic curve
38 parameters with a base point encoded in compressed form.
40 It is possible to trigger the infinite loop by crafting a certificate that
41 has invalid explicit curve parameters.
43 Since certificate parsing happens prior to verification of the certificate
44 signature, any process that parses an externally supplied certificate may thus
45 be subject to a denial of service attack. The infinite loop can also be
46 reached when parsing crafted private keys as they can contain explicit
47 elliptic curve parameters.
49 Thus vulnerable situations include:
51 - TLS clients consuming server certificates
52 - TLS servers consuming client certificates
53 - Hosting providers taking certificates or private keys from customers
54 - Certificate authorities parsing certification requests from subscribers
55 - Anything else which parses ASN.1 elliptic curve parameters
57 Also any other applications that use the BN_mod_sqrt() where the attacker
58 can control the parameter values are vulnerable to this DoS issue.
63 * Add ciphersuites based on DHE_PSK (RFC 4279) and ECDHE_PSK (RFC 5489)
64 to the list of ciphersuites providing Perfect Forward Secrecy as
65 required by SECLEVEL >= 3.
67 *Dmitry Belyavskiy, Nicola Tuveri*
69 * Made the AES constant time code for no-asm configurations
70 optional due to the resulting 95% performance degradation.
71 The AES constant time code can be enabled, for no assembly
72 builds, with: ./config no-asm -DOPENSSL_AES_CONST_TIME
76 * Fixed PEM_write_bio_PKCS8PrivateKey() to make it possible to use empty
81 * The negative return value handling of the certificate verification callback
82 was reverted. The replacement is to set the verification retry state with
83 the SSL_set_retry_verify() function.
87 ### Changes between 3.0.0 and 3.0.1 [14 Dec 2021]
89 * Fixed invalid handling of X509_verify_cert() internal errors in libssl
90 Internally libssl in OpenSSL calls X509_verify_cert() on the client side to
91 verify a certificate supplied by a server. That function may return a
92 negative return value to indicate an internal error (for example out of
93 memory). Such a negative return value is mishandled by OpenSSL and will cause
94 an IO function (such as SSL_connect() or SSL_do_handshake()) to not indicate
95 success and a subsequent call to SSL_get_error() to return the value
96 SSL_ERROR_WANT_RETRY_VERIFY. This return value is only supposed to be
97 returned by OpenSSL if the application has previously called
98 SSL_CTX_set_cert_verify_callback(). Since most applications do not do this
99 the SSL_ERROR_WANT_RETRY_VERIFY return value from SSL_get_error() will be
100 totally unexpected and applications may not behave correctly as a result. The
101 exact behaviour will depend on the application but it could result in
102 crashes, infinite loops or other similar incorrect responses.
104 This issue is made more serious in combination with a separate bug in OpenSSL
105 3.0 that will cause X509_verify_cert() to indicate an internal error when
106 processing a certificate chain. This will occur where a certificate does not
107 include the Subject Alternative Name extension but where a Certificate
108 Authority has enforced name constraints. This issue can occur even with valid
114 * Corrected a few file name and file reference bugs in the build,
115 installation and setup scripts, which lead to installation verification
116 failures. Slightly enhanced the installation verification script.
120 * Fixed EVP_PKEY_eq() to make it possible to use it with strictly private
125 * Fixed PVK encoder to properly query for the passphrase.
129 * Multiple fixes in the OSSL_HTTP API functions.
133 * Allow sign extension in OSSL_PARAM_allocate_from_text() for the
134 OSSL_PARAM_INTEGER data type and return error on negative numbers
135 used with the OSSL_PARAM_UNSIGNED_INTEGER data type. Make
136 OSSL_PARAM_BLD_push_BN{,_pad}() return an error on negative numbers.
140 * Allow copying uninitialized digest contexts with EVP_MD_CTX_copy_ex.
144 * Fixed detection of ARMv7 and ARM64 CPU features on FreeBSD.
148 * Multiple threading fixes.
152 * Added NULL digest implementation to keep compatibility with 1.1.1 version.
156 * Allow fetching an operation from the provider that owns an unexportable key
157 as a fallback if that is still allowed by the property query.
161 ### Changes between 1.1.1 and 3.0.0 [7 sep 2021]
163 * TLS_MAX_VERSION, DTLS_MAX_VERSION and DTLS_MIN_VERSION constants are now
168 * The `OPENSSL_s390xcap` environment variable can be used to set bits in the
169 S390X capability vector to zero. This simplifies testing of different code
170 paths on S390X architecture.
174 * Encrypting more than 2^64 TLS records with AES-GCM is disallowed
175 as per FIPS 140-2 IG A.5 "Key/IV Pair Uniqueness Requirements from
176 SP 800-38D". The communication will fail at this point.
180 * The EC_GROUP_clear_free() function is deprecated as there is nothing
181 confidential in EC_GROUP data.
185 * The byte order mark (BOM) character is ignored if encountered at the
186 beginning of a PEM-formatted file.
190 * Added CMS support for the Russian GOST algorithms.
194 * Due to move of the implementation of cryptographic operations
195 to the providers, validation of various operation parameters can
196 be postponed until the actual operation is executed where previously
197 it happened immediately when an operation parameter was set.
199 For example when setting an unsupported curve with
200 EVP_PKEY_CTX_set_ec_paramgen_curve_nid() this function call will not
201 fail but later keygen operations with the EVP_PKEY_CTX will fail.
203 *OpenSSL team members and many third party contributors*
205 * The EVP_get_cipherbyname() function will return NULL for algorithms such as
206 "AES-128-SIV", "AES-128-CBC-CTS" and "CAMELLIA-128-CBC-CTS" which were
207 previously only accessible via low level interfaces. Use EVP_CIPHER_fetch()
208 instead to retrieve these algorithms from a provider.
212 * On build targets where the multilib postfix is set in the build
213 configuration the libdir directory was changing based on whether
214 the lib directory with the multilib postfix exists on the system
215 or not. This unpredictable behavior was removed and eventual
216 multilib postfix is now always added to the default libdir. Use
217 `--libdir=lib` to override the libdir if adding the postfix is
222 * The triple DES key wrap functionality now conforms to RFC 3217 but is
223 no longer interoperable with OpenSSL 1.1.1.
227 * The ERR_GET_FUNC() function was removed. With the loss of meaningful
228 function codes, this function can only cause problems for calling
233 * Add a configurable flag to output date formats as ISO 8601. Does not
234 change the default date format.
238 * Version of MSVC earlier than 1300 could get link warnings, which could
239 be suppressed if the undocumented -DI_CAN_LIVE_WITH_LNK4049 was set.
240 Support for this flag has been removed.
244 * Rework and make DEBUG macros consistent. Remove unused -DCONF_DEBUG,
245 -DBN_CTX_DEBUG, and REF_PRINT. Add a new tracing category and use it for
246 printing reference counts. Rename -DDEBUG_UNUSED to -DUNUSED_RESULT_DEBUG
247 Fix BN_DEBUG_RAND so it compiles and, when set, force DEBUG_RAND to be set
248 also. Rename engine_debug_ref to be ENGINE_REF_PRINT also for consistency.
252 * The signatures of the functions to get and set options on SSL and
253 SSL_CTX objects changed from "unsigned long" to "uint64_t" type.
254 Some source code changes may be required.
258 * The public definitions of conf_method_st and conf_st have been
259 deprecated. They will be made opaque in a future release.
261 *Rich Salz and Tomáš Mráz*
263 * Client-initiated renegotiation is disabled by default. To allow it, use
264 the -client_renegotiation option, the SSL_OP_ALLOW_CLIENT_RENEGOTIATION
265 flag, or the "ClientRenegotiation" config parameter as appropriate.
269 * Add "abspath" and "includedir" pragma's to config files, to prevent,
270 or modify relative pathname inclusion.
274 * OpenSSL includes a cryptographic module that is intended to be FIPS 140-2
275 validated. Please consult the README-FIPS and
276 README-PROVIDERS files, as well as the migration guide.
278 *OpenSSL team members and many third party contributors*
280 * For the key types DH and DHX the allowed settable parameters are now different.
284 * The openssl commands that read keys, certificates, and CRLs now
285 automatically detect the PEM or DER format of the input files.
287 *David von Oheimb, Richard Levitte, and Tomáš Mráz*
289 * Added enhanced PKCS#12 APIs which accept a library context.
293 * The default manual page suffix ($MANSUFFIX) has been changed to "ossl"
297 * Added support for Kernel TLS (KTLS).
299 *Boris Pismenny, John Baldwin and Andrew Gallatin*
301 * Support for RFC 5746 secure renegotiation is now required by default for
302 SSL or TLS connections to succeed.
306 * The signature of the `copy` functional parameter of the
307 EVP_PKEY_meth_set_copy() function has changed so its `src` argument is
308 now `const EVP_PKEY_CTX *` instead of `EVP_PKEY_CTX *`. Similarly
309 the signature of the `pub_decode` functional parameter of the
310 EVP_PKEY_asn1_set_public() function has changed so its `pub` argument is
311 now `const X509_PUBKEY *` instead of `X509_PUBKEY *`.
315 * The error return values from some control calls (ctrl) have changed.
319 * A public key check is now performed during EVP_PKEY_derive_set_peer().
323 * Many functions in the EVP_ namespace that are getters of values from
324 implementations or contexts were renamed to include get or get0 in their
325 names. Old names are provided as macro aliases for compatibility and
330 * The EVP_PKEY_CTRL_PKCS7_ENCRYPT, EVP_PKEY_CTRL_PKCS7_DECRYPT,
331 EVP_PKEY_CTRL_PKCS7_SIGN, EVP_PKEY_CTRL_CMS_ENCRYPT,
332 EVP_PKEY_CTRL_CMS_DECRYPT, and EVP_PKEY_CTRL_CMS_SIGN control operations
337 * The EVP_PKEY_public_check() and EVP_PKEY_param_check() functions now work for
340 * The output from the command line applications may have minor
345 * The output from numerous "printing" may have minor changes.
349 * Windows thread synchronization uses read/write primitives (SRWLock) when
350 supported by the OS, otherwise CriticalSection continues to be used.
354 * Add filter BIO BIO_f_readbuffer() that allows BIO_tell() and BIO_seek() to
355 work on read only BIO source/sinks that do not support these functions.
356 This allows piping or redirection of a file BIO using stdin to be buffered
357 into memory. This is used internally in OSSL_DECODER_from_bio().
361 * OSSL_STORE_INFO_get_type() may now return an additional value. In 1.1.1
362 this function would return one of the values OSSL_STORE_INFO_NAME,
363 OSSL_STORE_INFO_PKEY, OSSL_STORE_INFO_PARAMS, OSSL_STORE_INFO_CERT or
364 OSSL_STORE_INFO_CRL. Decoded public keys would previously have been reported
365 as type OSSL_STORE_INFO_PKEY in 1.1.1. In 3.0 decoded public keys are now
366 reported as having the new type OSSL_STORE_INFO_PUBKEY. Applications
367 using this function should be amended to handle the changed return value.
371 * Improved adherence to Enhanced Security Services (ESS, RFC 2634 and RFC 5035)
372 for the TSP and CMS Advanced Electronic Signatures (CAdES) implementations.
373 As required by RFC 5035 check both ESSCertID and ESSCertIDv2 if both present.
374 Correct the semantics of checking the validation chain in case ESSCertID{,v2}
375 contains more than one certificate identifier: This means that all
376 certificates referenced there MUST be part of the validation chain.
380 * The implementation of older EVP ciphers related to CAST, IDEA, SEED, RC2, RC4,
381 RC5, DESX and DES have been moved to the legacy provider.
385 * The implementation of the EVP digests MD2, MD4, MDC2, WHIRLPOOL and
386 RIPEMD-160 have been moved to the legacy provider.
390 * The deprecated function EVP_PKEY_get0() now returns NULL being called for a
395 * The deprecated functions EVP_PKEY_get0_RSA(),
396 EVP_PKEY_get0_DSA(), EVP_PKEY_get0_EC_KEY(), EVP_PKEY_get0_DH(),
397 EVP_PKEY_get0_hmac(), EVP_PKEY_get0_poly1305() and EVP_PKEY_get0_siphash() as
398 well as the similarly named "get1" functions behave differently in
403 * A number of functions handling low-level keys or engines were deprecated
404 including EVP_PKEY_set1_engine(), EVP_PKEY_get0_engine(), EVP_PKEY_assign(),
405 EVP_PKEY_get0(), EVP_PKEY_get0_hmac(), EVP_PKEY_get0_poly1305() and
406 EVP_PKEY_get0_siphash().
410 * PKCS#5 PBKDF1 key derivation has been moved from PKCS5_PBE_keyivgen() into
411 the legacy crypto provider as an EVP_KDF. Applications requiring this KDF
412 will need to load the legacy crypto provider. This includes these PBE
413 algorithms which use this KDF:
414 - NID_pbeWithMD2AndDES_CBC
415 - NID_pbeWithMD5AndDES_CBC
416 - NID_pbeWithSHA1AndRC2_CBC
417 - NID_pbeWithMD2AndRC2_CBC
418 - NID_pbeWithMD5AndRC2_CBC
419 - NID_pbeWithSHA1AndDES_CBC
423 * Deprecated obsolete BIO_set_callback(), BIO_get_callback(), and
424 BIO_debug_callback() functions.
428 * Deprecated obsolete EVP_PKEY_CTX_get0_dh_kdf_ukm() and
429 EVP_PKEY_CTX_get0_ecdh_kdf_ukm() functions.
433 * The RAND_METHOD APIs have been deprecated.
437 * The SRP APIs have been deprecated.
441 * Add a compile time option to prevent the caching of provider fetched
442 algorithms. This is enabled by including the no-cached-fetch option
443 at configuration time.
447 * pkcs12 now uses defaults of PBKDF2, AES and SHA-256, with a MAC iteration
448 count of PKCS12_DEFAULT_ITER.
450 *Tomáš Mráz and Sahana Prasad*
452 * The openssl speed command does not use low-level API calls anymore.
456 * Parallel dual-prime 1024-bit modular exponentiation for AVX512_IFMA
459 *Ilya Albrekht, Sergey Kirillov, Andrey Matyukov (Intel Corp)*
461 * Combining the Configure options no-ec and no-dh no longer disables TLSv1.3.
465 * Implemented support for fully "pluggable" TLSv1.3 groups. This means that
466 providers may supply their own group implementations (using either the "key
467 exchange" or the "key encapsulation" methods) which will automatically be
468 detected and used by libssl.
470 *Matt Caswell, Nicola Tuveri*
472 * The undocumented function X509_certificate_type() has been deprecated;
476 * Deprecated the obsolete BN_pseudo_rand() and BN_pseudo_rand_range().
480 * Removed RSA padding mode for SSLv23 (which was only used for
481 SSLv2). This includes the functions RSA_padding_check_SSLv23() and
482 RSA_padding_add_SSLv23() and the `-ssl` option in the deprecated
487 * Deprecated the obsolete X9.31 RSA key generation related functions.
489 * While a callback function set via `SSL_CTX_set_cert_verify_callback()`
490 is not allowed to return a value > 1, this is no more taken as failure.
492 *Viktor Dukhovni and David von Oheimb*
494 * Deprecated the obsolete X9.31 RSA key generation related functions
495 BN_X931_generate_Xpq(), BN_X931_derive_prime_ex(), and
496 BN_X931_generate_prime_ex().
500 * The default key generation method for the regular 2-prime RSA keys was
501 changed to the FIPS 186-4 B.3.6 method.
505 * Deprecated the BN_is_prime_ex() and BN_is_prime_fasttest_ex() functions.
509 * Deprecated EVP_MD_CTX_set_update_fn() and EVP_MD_CTX_update_fn().
513 * Deprecated the type OCSP_REQ_CTX and the functions OCSP_REQ_CTX_*() and
514 replaced with OSSL_HTTP_REQ_CTX and the functions OSSL_HTTP_REQ_CTX_*().
516 *Rich Salz, Richard Levitte, and David von Oheimb*
518 * Deprecated `X509_http_nbio()` and `X509_CRL_http_nbio()`.
522 * Deprecated `OCSP_parse_url()`.
526 * Validation of SM2 keys has been separated from the validation of regular EC
531 * Behavior of the `pkey` app is changed, when using the `-check` or `-pubcheck`
532 switches: a validation failure triggers an early exit, returning a failure
533 exit status to the parent process.
537 * Changed behavior of SSL_CTX_set_ciphersuites() and SSL_set_ciphersuites()
538 to ignore unknown ciphers.
542 * The `-cipher-commands` and `-digest-commands` options
543 of the command line utility `list` have been deprecated.
544 Instead use the `-cipher-algorithms` and `-digest-algorithms` options.
548 * Added convenience functions for generating asymmetric key pairs:
549 The 'quick' one-shot (yet somewhat limited) function L<EVP_PKEY_Q_keygen(3)>
550 and macros for the most common cases: <EVP_RSA_gen(3)> and L<EVP_EC_gen(3)>.
554 * All of the low level EC_KEY functions have been deprecated.
556 *Shane Lontis, Paul Dale, Richard Levitte, and Tomáš Mráz*
558 * Deprecated all the libcrypto and libssl error string loading
563 * The functions SSL_CTX_set_tmp_dh_callback and SSL_set_tmp_dh_callback, as
564 well as the macros SSL_CTX_set_tmp_dh() and SSL_set_tmp_dh() have been
569 * The `-crypt` option to the `passwd` command line tool has been removed.
573 * The -C option to the `x509`, `dhparam`, `dsaparam`, and `ecparam` commands
578 * Add support for AES Key Wrap inverse ciphers to the EVP layer.
582 * Deprecated EVP_PKEY_set1_tls_encodedpoint() and
583 EVP_PKEY_get1_tls_encodedpoint().
587 * The security callback, which can be customised by application code, supports
588 the security operation SSL_SECOP_TMP_DH. One location of the "other" parameter
589 was incorrectly passing a DH object. It now passed an EVP_PKEY in all cases.
593 * Add PKCS7_get_octet_string() and PKCS7_type_is_other() to the public
594 interface. Their functionality remains unchanged.
598 * Added new option for 'openssl list', '-providers', which will display the
599 list of loaded providers, their names, version and status. It optionally
600 displays their gettable parameters.
604 * Removed EVP_PKEY_set_alias_type().
608 * Deprecated `EVP_PKEY_CTX_set_rsa_keygen_pubexp()` and introduced
609 `EVP_PKEY_CTX_set1_rsa_keygen_pubexp()`, which is now preferred.
613 * Changed all "STACK" functions to be macros instead of inline functions. Macro
614 parameters are still checked for type safety at compile time via helper
619 * Remove the RAND_DRBG API
621 *Paul Dale and Matthias St. Pierre*
623 * Allow `SSL_set1_host()` and `SSL_add1_host()` to take IP literal addresses
624 as well as actual hostnames.
628 * The 'MinProtocol' and 'MaxProtocol' configuration commands now silently
629 ignore TLS protocol version bounds when configuring DTLS-based contexts, and
630 conversely, silently ignore DTLS protocol version bounds when configuring
631 TLS-based contexts. The commands can be repeated to set bounds of both
632 types. The same applies with the corresponding "min_protocol" and
633 "max_protocol" command-line switches, in case some application uses both TLS
636 SSL_CTX instances that are created for a fixed protocol version (e.g.
637 `TLSv1_server_method()`) also silently ignore version bounds. Previously
638 attempts to apply bounds to these protocol versions would result in an
639 error. Now only the "version-flexible" SSL_CTX instances are subject to
640 limits in configuration files in command-line options.
644 * Deprecated the `ENGINE` API. Engines should be replaced with providers
649 * Reworked the recorded ERR codes to make better space for system errors.
650 To distinguish them, the macro `ERR_SYSTEM_ERROR()` indicates if the
651 given code is a system error (true) or an OpenSSL error (false).
655 * Reworked the test perl framework to better allow parallel testing.
657 *Nicola Tuveri and David von Oheimb*
659 * Added ciphertext stealing algorithms AES-128-CBC-CTS, AES-192-CBC-CTS and
660 AES-256-CBC-CTS to the providers. CS1, CS2 and CS3 variants are supported.
664 * 'Configure' has been changed to figure out the configuration target if
665 none is given on the command line. Consequently, the 'config' script is
666 now only a mere wrapper. All documentation is changed to only mention
669 *Rich Salz and Richard Levitte*
671 * Added a library context `OSSL_LIB_CTX` that applications as well as
672 other libraries can use to form a separate context within which
673 libcrypto operations are performed.
677 * Added various `_ex` functions to the OpenSSL API that support using
678 a non-default `OSSL_LIB_CTX`.
682 * Handshake now fails if Extended Master Secret extension is dropped
687 * Dropped interactive mode from the `openssl` program.
691 * Deprecated `EVP_PKEY_cmp()` and `EVP_PKEY_cmp_parameters()`.
693 *David von Oheimb and Shane Lontis*
695 * Deprecated `EC_METHOD_get_field_type()`.
699 * Deprecated EC_GFp_simple_method(), EC_GFp_mont_method(),
700 EC_GF2m_simple_method(), EC_GFp_nist_method(), EC_GFp_nistp224_method()
701 EC_GFp_nistp256_method(), and EC_GFp_nistp521_method().
705 * Deprecated EC_GROUP_new(), EC_GROUP_method_of(), and EC_POINT_method_of().
709 * Add CAdES-BES signature verification support, mostly derived
710 from ESSCertIDv2 TS (RFC 5816) contribution by Marek Klein.
712 *Filipe Raimundo da Silva*
714 * Add CAdES-BES signature scheme and attributes support (RFC 5126) to CMS API.
718 * Added the AuthEnvelopedData content type structure (RFC 5083) with AES-GCM
719 parameter (RFC 5084) for the Cryptographic Message Syntax (CMS).
723 * Deprecated EC_POINT_make_affine() and EC_POINTs_make_affine().
727 * Deprecated EC_GROUP_precompute_mult(), EC_GROUP_have_precompute_mult(), and
728 EC_KEY_precompute_mult().
732 * Deprecated EC_POINTs_mul().
736 * Removed FIPS_mode() and FIPS_mode_set().
740 * The SSL option SSL_OP_IGNORE_UNEXPECTED_EOF is introduced.
744 * Deprecated EC_POINT_set_Jprojective_coordinates_GFp() and
745 EC_POINT_get_Jprojective_coordinates_GFp().
749 * Added OSSL_PARAM_BLD to the public interface. This allows OSSL_PARAM
750 arrays to be more easily constructed via a series of utility functions.
751 Create a parameter builder using OSSL_PARAM_BLD_new(), add parameters using
752 the various push functions and finally convert to a passable OSSL_PARAM
753 array using OSSL_PARAM_BLD_to_param().
757 * The security strength of SHA1 and MD5 based signatures in TLS has been
762 * Added EVP_PKEY_set_type_by_keymgmt(), to initialise an EVP_PKEY to
763 contain a provider side internal key.
767 * ASN1_verify(), ASN1_digest() and ASN1_sign() have been deprecated.
771 * Project text documents not yet having a proper file name extension
772 (`HACKING`, `LICENSE`, `NOTES*`, `README*`, `VERSION`) have been renamed to
773 `*.md` as far as reasonable, else `*.txt`, for better use with file managers.
777 * The main project documents (README, NEWS, CHANGES, INSTALL, SUPPORT)
778 have been converted to Markdown with the goal to produce documents
779 which not only look pretty when viewed online in the browser, but
780 remain well readable inside a plain text editor.
782 To achieve this goal, a 'minimalistic' Markdown style has been applied
783 which avoids formatting elements that interfere too much with the
784 reading flow in the text file. For example, it
786 * avoids [ATX headings][] and uses [setext headings][] instead
787 (which works for `<h1>` and `<h2>` headings only).
788 * avoids [inline links][] and uses [reference links][] instead.
789 * avoids [fenced code blocks][] and uses [indented code blocks][] instead.
791 [ATX headings]: https://github.github.com/gfm/#atx-headings
792 [setext headings]: https://github.github.com/gfm/#setext-headings
793 [inline links]: https://github.github.com/gfm/#inline-link
794 [reference links]: https://github.github.com/gfm/#reference-link
795 [fenced code blocks]: https://github.github.com/gfm/#fenced-code-blocks
796 [indented code blocks]: https://github.github.com/gfm/#indented-code-blocks
798 *Matthias St. Pierre*
800 * The test suite is changed to preserve results of each test recipe.
801 A new directory test-runs/ with subdirectories named like the
802 test recipes are created in the build tree for this purpose.
806 * Added an implementation of CMP and CRMF (RFC 4210, RFC 4211 RFC 6712).
807 This adds `crypto/cmp/`, `crpyto/crmf/`, `apps/cmp.c`, and `test/cmp_*`.
808 See L<openssl-cmp(1)> and L<OSSL_CMP_exec_IR_ses(3)> as starting points.
810 *David von Oheimb, Martin Peylo*
812 * Generalized the HTTP client code from `crypto/ocsp/` into `crpyto/http/`.
813 It supports arbitrary request and response content types, GET redirection,
814 TLS, connections via HTTP(S) proxies, connections and exchange via
815 user-defined BIOs (allowing implicit connections), persistent connections,
816 and timeout checks. See L<OSSL_HTTP_transfer(3)> etc. for details.
817 The legacy OCSP-focused (and only partly documented) API
818 is retained for backward compatibility, while most of it is deprecated.
822 * Added `util/check-format.pl`, a tool for checking adherence to the
823 OpenSSL coding style <https://www.openssl.org/policies/codingstyle.html>.
824 The checks performed are incomplete and yield some false positives.
825 Still the tool should be useful for detecting most typical glitches.
829 * `BIO_do_connect()` and `BIO_do_handshake()` have been extended:
830 If domain name resolution yields multiple IP addresses all of them are tried
831 after `connect()` failures.
835 * All of the low level RSA functions have been deprecated.
839 * X509 certificates signed using SHA1 are no longer allowed at security
844 * The command line utilities dhparam, dsa, gendsa and dsaparam have been
845 modified to use PKEY APIs. These commands are now in maintenance mode
846 and no new features will be added to them.
850 * The command line utility rsautl has been deprecated.
854 * The command line utilities genrsa and rsa have been modified to use PKEY
855 APIs. They now write PKCS#8 keys by default. These commands are now in
856 maintenance mode and no new features will be added to them.
860 * All of the low level DH functions have been deprecated.
862 *Paul Dale and Matt Caswell*
864 * All of the low level DSA functions have been deprecated.
868 * Reworked the treatment of EC EVP_PKEYs with the SM2 curve to
869 automatically become EVP_PKEY_SM2 rather than EVP_PKEY_EC.
873 * Deprecated low level ECDH and ECDSA functions.
877 * Deprecated EVP_PKEY_decrypt_old() and EVP_PKEY_encrypt_old().
881 * Enhanced the documentation of EVP_PKEY_get_size(), EVP_PKEY_get_bits()
882 and EVP_PKEY_get_security_bits(). Especially EVP_PKEY_get_size() needed
883 a new formulation to include all the things it can be used for,
884 as well as words of caution.
888 * The SSL_CTX_set_tlsext_ticket_key_cb(3) function has been deprecated.
892 * All of the low level HMAC functions have been deprecated.
894 *Paul Dale and David von Oheimb*
896 * Over two thousand fixes were made to the documentation, including:
897 - Common options (such as -rand/-writerand, TLS version control, etc)
898 were refactored and point to newly-enhanced descriptions in openssl.pod.
899 - Added style conformance for all options (with help from Richard Levitte),
900 documented all reported missing options, added a CI build to check
901 that all options are documented and that no unimplemented options
903 - Documented some internals, such as all use of environment variables.
904 - Addressed all internal broken L<> references.
908 * All of the low level CMAC functions have been deprecated.
912 * The low-level MD2, MD4, MD5, MDC2, RIPEMD160 and Whirlpool digest
913 functions have been deprecated.
915 *Paul Dale and David von Oheimb*
917 * Corrected the documentation of the return values from the `EVP_DigestSign*`
918 set of functions. The documentation mentioned negative values for some
919 errors, but this was never the case, so the mention of negative values
922 Code that followed the documentation and thereby check with something
923 like `EVP_DigestSignInit(...) <= 0` will continue to work undisturbed.
927 * All of the low level cipher functions have been deprecated.
929 *Matt Caswell and Paul Dale*
931 * Removed include/openssl/opensslconf.h.in and replaced it with
932 include/openssl/configuration.h.in, which differs in not including
933 <openssl/macros.h>. A short header include/openssl/opensslconf.h
934 was added to include both.
936 This allows internal hacks where one might need to modify the set
937 of configured macros, for example this if deprecated symbols are
938 still supposed to be available internally:
940 #include <openssl/configuration.h>
942 #undef OPENSSL_NO_DEPRECATED
943 #define OPENSSL_SUPPRESS_DEPRECATED
945 #include <openssl/macros.h>
947 This should not be used by applications that use the exported
948 symbols, as that will lead to linking errors.
952 * Fixed an overflow bug in the x64_64 Montgomery squaring procedure
953 used in exponentiation with 512-bit moduli. No EC algorithms are
954 affected. Analysis suggests that attacks against 2-prime RSA1024,
955 3-prime RSA1536, and DSA1024 as a result of this defect would be very
956 difficult to perform and are not believed likely. Attacks against DH512
957 are considered just feasible. However, for an attack the target would
958 have to re-use the DH512 private key, which is not recommended anyway.
959 Also applications directly using the low-level API BN_mod_exp may be
960 affected if they use BN_FLG_CONSTTIME.
965 * Most memory-debug features have been deprecated, and the functionality
966 replaced with no-ops.
970 * Added documentation for the STACK API.
974 * Introduced a new method type and API, OSSL_ENCODER, to represent
975 generic encoders. These do the same sort of job that PEM writers
976 and d2i functions do, but with support for methods supplied by
977 providers, and the possibility for providers to support other
982 * Introduced a new method type and API, OSSL_DECODER, to represent
983 generic decoders. These do the same sort of job that PEM readers
984 and i2d functions do, but with support for methods supplied by
985 providers, and the possibility for providers to support other
990 * Added a .pragma directive to the syntax of configuration files, to
991 allow varying behavior in a supported and predictable manner.
992 Currently added pragma:
996 This allows dollar signs to be a keyword character unless it's
997 followed by a opening brace or parenthesis. This is useful for
998 platforms where dollar signs are commonly used in names, such as
999 volume names and system directory names on VMS.
1003 * Added functionality to create an EVP_PKEY from user data.
1007 * Change the interpretation of the '--api' configuration option to
1008 mean that this is a desired API compatibility level with no
1009 further meaning. The previous interpretation, that this would
1010 also mean to remove all deprecated symbols up to and including
1011 the given version, no requires that 'no-deprecated' is also used
1012 in the configuration.
1014 When building applications, the desired API compatibility level
1015 can be set with the OPENSSL_API_COMPAT macro like before. For
1016 API compatibility version below 3.0, the old style numerical
1017 value is valid as before, such as -DOPENSSL_API_COMPAT=0x10100000L.
1018 For version 3.0 and on, the value is expected to be the decimal
1019 value calculated from the major and minor version like this:
1021 MAJOR * 10000 + MINOR * 100
1025 -DOPENSSL_API_COMPAT=30000 For 3.0
1026 -DOPENSSL_API_COMPAT=30200 For 3.2
1028 To hide declarations that are deprecated up to and including the
1029 given API compatibility level, -DOPENSSL_NO_DEPRECATED must be
1030 given when building the application as well.
1034 * Added the X509_LOOKUP_METHOD called X509_LOOKUP_store, to allow
1035 access to certificate and CRL stores via URIs and OSSL_STORE
1038 This adds the following functions:
1040 - X509_LOOKUP_store()
1041 - X509_STORE_load_file()
1042 - X509_STORE_load_path()
1043 - X509_STORE_load_store()
1044 - SSL_add_store_cert_subjects_to_stack()
1045 - SSL_CTX_set_default_verify_store()
1046 - SSL_CTX_load_verify_file()
1047 - SSL_CTX_load_verify_dir()
1048 - SSL_CTX_load_verify_store()
1052 * Added a new method to gather entropy on VMS, based on SYS$GET_ENTROPY.
1053 The presence of this system service is determined at run-time.
1057 * Added functionality to create an EVP_PKEY context based on data
1058 for methods from providers. This takes an algorithm name and a
1059 property query string and simply stores them, with the intent
1060 that any operation that uses this context will use those strings
1061 to fetch the needed methods implicitly, thereby making the port
1062 of application written for pre-3.0 OpenSSL easier.
1066 * The undocumented function NCONF_WIN32() has been deprecated; for
1067 conversion details see the HISTORY section of doc/man5/config.pod
1071 * Introduced the new functions EVP_DigestSignInit_ex() and
1072 EVP_DigestVerifyInit_ex(). The macros EVP_DigestSignUpdate() and
1073 EVP_DigestVerifyUpdate() have been converted to functions. See the man
1074 pages for further details.
1078 * Over two thousand fixes were made to the documentation, including:
1079 adding missing command flags, better style conformance, documentation
1082 *Rich Salz, Richard Levitte*
1084 * s390x assembly pack: add hardware-support for P-256, P-384, P-521,
1085 X25519, X448, Ed25519 and Ed448.
1089 * Print all values for a PKCS#12 attribute with 'openssl pkcs12', not just
1094 * Deprecated the public definition of `ERR_STATE` as well as the function
1095 `ERR_get_state()`. This is done in preparation of making `ERR_STATE` an
1100 * Added ERR functionality to give callers access to the stored function
1101 names that have replaced the older function code based functions.
1103 New functions are ERR_peek_error_func(), ERR_peek_last_error_func(),
1104 ERR_peek_error_data(), ERR_peek_last_error_data(), ERR_get_error_all(),
1105 ERR_peek_error_all() and ERR_peek_last_error_all().
1107 Deprecate ERR functions ERR_get_error_line(), ERR_get_error_line_data(),
1108 ERR_peek_error_line_data(), ERR_peek_last_error_line_data() and
1109 ERR_func_error_string().
1113 * Extended testing to be verbose for failing tests only. The make variables
1114 VERBOSE_FAILURE or VF can be used to enable this:
1116 $ make VF=1 test # Unix
1117 $ mms /macro=(VF=1) test ! OpenVMS
1118 $ nmake VF=1 test # Windows
1122 * Added the `-copy_extensions` option to the `x509` command for use with
1123 `-req` and `-x509toreq`. When given with the `copy` or `copyall` argument,
1124 all extensions in the request are copied to the certificate or vice versa.
1126 *David von Oheimb*, *Kirill Stefanenkov <kirill_stefanenkov@rambler.ru>*
1128 * Added the `-copy_extensions` option to the `req` command for use with
1129 `-x509`. When given with the `copy` or `copyall` argument,
1130 all extensions in the certification request are copied to the certificate.
1134 * The `x509`, `req`, and `ca` commands now make sure that X.509v3 certificates
1135 they generate are by default RFC 5280 compliant in the following sense:
1136 There is a subjectKeyIdentifier extension with a hash value of the public key
1137 and for not self-signed certs there is an authorityKeyIdentifier extension
1138 with a keyIdentifier field or issuer information identifying the signing key.
1139 This is done unless some configuration overrides the new default behavior,
1140 such as `subjectKeyIdentifier = none` and `authorityKeyIdentifier = none`.
1144 * Added several checks to `X509_verify_cert()` according to requirements in
1145 RFC 5280 in case `X509_V_FLAG_X509_STRICT` is set
1146 (which may be done by using the CLI option `-x509_strict`):
1147 * The basicConstraints of CA certificates must be marked critical.
1148 * CA certificates must explicitly include the keyUsage extension.
1149 * If a pathlenConstraint is given the key usage keyCertSign must be allowed.
1150 * The issuer name of any certificate must not be empty.
1151 * The subject name of CA certs, certs with keyUsage crlSign,
1152 and certs without subjectAlternativeName must not be empty.
1153 * If a subjectAlternativeName extension is given it must not be empty.
1154 * The signatureAlgorithm field and the cert signature must be consistent.
1155 * Any given authorityKeyIdentifier and any given subjectKeyIdentifier
1156 must not be marked critical.
1157 * The authorityKeyIdentifier must be given for X.509v3 certs
1158 unless they are self-signed.
1159 * The subjectKeyIdentifier must be given for all X.509v3 CA certs.
1163 * Certificate verification using `X509_verify_cert()` meanwhile rejects EC keys
1164 with explicit curve parameters (specifiedCurve) as required by RFC 5480.
1168 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
1169 used even when parsing explicit parameters, when loading a encoded key
1170 or calling `EC_GROUP_new_from_ecpkparameters()`/
1171 `EC_GROUP_new_from_ecparameters()`.
1172 This prevents bypass of security hardening and performance gains,
1173 especially for curves with specialized EC_METHODs.
1174 By default, if a key encoded with explicit parameters is loaded and later
1175 encoded, the output is still encoded with explicit parameters, even if
1176 internally a "named" EC_GROUP is used for computation.
1180 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
1181 this change, EC_GROUP_set_generator would accept order and/or cofactor as
1182 NULL. After this change, only the cofactor parameter can be NULL. It also
1183 does some minimal sanity checks on the passed order.
1188 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
1189 An attack is simple, if the first CMS_recipientInfo is valid but the
1190 second CMS_recipientInfo is chosen ciphertext. If the second
1191 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
1192 encryption key will be replaced by garbage, and the message cannot be
1193 decoded, but if the RSA decryption fails, the correct encryption key is
1194 used and the recipient will not notice the attack.
1195 As a work around for this potential attack the length of the decrypted
1196 key must be equal to the cipher default key length, in case the
1197 certifiate is not given and all recipientInfo are tried out.
1198 The old behaviour can be re-enabled in the CMS code by setting the
1199 CMS_DEBUG_DECRYPT flag.
1203 * Early start up entropy quality from the DEVRANDOM seed source has been
1204 improved for older Linux systems. The RAND subsystem will wait for
1205 /dev/random to be producing output before seeding from /dev/urandom.
1206 The seeded state is stored for future library initialisations using
1207 a system global shared memory segment. The shared memory identifier
1208 can be configured by defining OPENSSL_RAND_SEED_DEVRANDOM_SHM_ID to
1209 the desired value. The default identifier is 114.
1213 * Revised BN_generate_prime_ex to not avoid factors 2..17863 in p-1
1214 when primes for RSA keys are computed.
1215 Since we previously always generated primes == 2 (mod 3) for RSA keys,
1216 the 2-prime and 3-prime RSA modules were easy to distinguish, since
1217 `N = p*q = 1 (mod 3)`, but `N = p*q*r = 2 (mod 3)`. Therefore fingerprinting
1218 2-prime vs. 3-prime RSA keys was possible by computing N mod 3.
1219 This avoids possible fingerprinting of newly generated RSA modules.
1223 * Correct the extended master secret constant on EBCDIC systems. Without this
1224 fix TLS connections between an EBCDIC system and a non-EBCDIC system that
1225 negotiate EMS will fail. Unfortunately this also means that TLS connections
1226 between EBCDIC systems with this fix, and EBCDIC systems without this
1227 fix will fail if they negotiate EMS.
1231 * Changed the library initialisation so that the config file is now loaded
1232 by default. This was already the case for libssl. It now occurs for both
1233 libcrypto and libssl. Use the OPENSSL_INIT_NO_LOAD_CONFIG option to
1234 `OPENSSL_init_crypto()` to suppress automatic loading of a config file.
1238 * Introduced new error raising macros, `ERR_raise()` and `ERR_raise_data()`,
1239 where the former acts as a replacement for `ERR_put_error()`, and the
1240 latter replaces the combination `ERR_put_error()` + `ERR_add_error_data()`.
1241 `ERR_raise_data()` adds more flexibility by taking a format string and
1242 an arbitrary number of arguments following it, to be processed with
1247 * Introduced a new function, `OSSL_PROVIDER_available()`, which can be used
1248 to check if a named provider is loaded and available. When called, it
1249 will also activate all fallback providers if such are still present.
1253 * Enforce a minimum DH modulus size of 512 bits.
1257 * Changed DH parameters to generate the order q subgroup instead of 2q.
1258 Previously generated DH parameters are still accepted by DH_check
1259 but DH_generate_key works around that by clearing bit 0 of the
1260 private key for those. This avoids leaking bit 0 of the private key.
1264 * Significantly reduce secure memory usage by the randomness pools.
1268 * `{CRYPTO,OPENSSL}_mem_debug_{push,pop}` are now no-ops and have been
1273 * A new type, EVP_KEYEXCH, has been introduced to represent key exchange
1274 algorithms. An implementation of a key exchange algorithm can be obtained
1275 by using the function EVP_KEYEXCH_fetch(). An EVP_KEYEXCH algorithm can be
1276 used in a call to EVP_PKEY_derive_init_ex() which works in a similar way to
1277 the older EVP_PKEY_derive_init() function. See the man pages for the new
1278 functions for further details.
1282 * The EVP_PKEY_CTX_set_dh_pad() macro has now been converted to a function.
1286 * Removed the function names from error messages and deprecated the
1291 * Removed NextStep support and the macro OPENSSL_UNISTD
1295 * Removed DES_check_key. Also removed OPENSSL_IMPLEMENT_GLOBAL,
1296 OPENSSL_GLOBAL_REF, OPENSSL_DECLARE_GLOBAL.
1297 Also removed "export var as function" capability; we do not export
1298 variables, only functions.
1302 * RC5_32_set_key has been changed to return an int type, with 0 indicating
1303 an error and 1 indicating success. In previous versions of OpenSSL this
1304 was a void type. If a key was set longer than the maximum possible this
1309 * Support SM2 signing and verification schemes with X509 certificate.
1313 * Use SHA256 as the default digest for TS query in the `ts` app.
1317 * Change PBKDF2 to conform to SP800-132 instead of the older PKCS5 RFC2898.
1321 * Default cipher lists/suites are now available via a function, the
1322 #defines are deprecated.
1326 * Add target VC-WIN32-UWP, VC-WIN64A-UWP, VC-WIN32-ARM-UWP and
1327 VC-WIN64-ARM-UWP in Windows OneCore target for making building libraries
1328 for Windows Store apps easier. Also, the "no-uplink" option has been added.
1332 * Join the directories crypto/x509 and crypto/x509v3
1336 * Added command 'openssl kdf' that uses the EVP_KDF API.
1340 * Added command 'openssl mac' that uses the EVP_MAC API.
1344 * Added OPENSSL_info() to get diverse built-in OpenSSL data, such
1345 as default directories. Also added the command 'openssl info'
1346 for scripting purposes.
1350 * The functions AES_ige_encrypt() and AES_bi_ige_encrypt() have been
1355 * Add prediction resistance to the DRBG reseeding process.
1359 * Limit the number of blocks in a data unit for AES-XTS to 2^20 as
1360 mandated by IEEE Std 1619-2018.
1364 * Added newline escaping functionality to a filename when using openssl dgst.
1365 This output format is to replicate the output format found in the `*sum`
1366 checksum programs. This aims to preserve backward compatibility.
1368 *Matt Eaton, Richard Levitte, and Paul Dale*
1370 * Removed the heartbeat message in DTLS feature, as it has very
1371 little usage and doesn't seem to fulfill a valuable purpose.
1372 The configuration option is now deprecated.
1376 * Changed the output of 'openssl {digestname} < file' to display the
1377 digest name in its output.
1381 * Added a new generic trace API which provides support for enabling
1382 instrumentation through trace output.
1384 *Richard Levitte & Matthias St. Pierre*
1386 * Added build tests for C++. These are generated files that only do one
1387 thing, to include one public OpenSSL head file each. This tests that
1388 the public header files can be usefully included in a C++ application.
1390 This test isn't enabled by default. It can be enabled with the option
1391 'enable-buildtest-c++'.
1395 * Added KB KDF (EVP_KDF_KB) to EVP_KDF.
1399 * Added SSH KDF (EVP_KDF_SSHKDF) and KRB5 KDF (EVP_KDF_KRB5KDF) to EVP_KDF.
1403 * Added Single Step KDF (EVP_KDF_SS), X963 KDF, and X942 KDF to EVP_KDF.
1407 * Added KMAC to EVP_MAC.
1411 * Added property based algorithm implementation selection framework to
1416 * Added SCA hardening for modular field inversion in EC_GROUP through
1417 a new dedicated field_inv() pointer in EC_METHOD.
1418 This also addresses a leakage affecting conversions from projective
1419 to affine coordinates.
1421 *Billy Bob Brumley, Nicola Tuveri*
1423 * Added EVP_KDF, an EVP layer KDF API, to simplify adding KDF and PRF
1424 implementations. This includes an EVP_PKEY to EVP_KDF bridge for
1425 those algorithms that were already supported through the EVP_PKEY API
1426 (scrypt, TLS1 PRF and HKDF). The low-level KDF functions for PBKDF2
1427 and scrypt are now wrappers that call EVP_KDF.
1431 * Build devcrypto engine as a dynamic engine.
1433 *Eneas U de Queiroz*
1435 * Add keyed BLAKE2 to EVP_MAC.
1439 * Fix a bug in the computation of the endpoint-pair shared secret used
1440 by DTLS over SCTP. This breaks interoperability with older versions
1441 of OpenSSL like OpenSSL 1.1.0 and OpenSSL 1.0.2. There is a runtime
1442 switch SSL_MODE_DTLS_SCTP_LABEL_LENGTH_BUG (off by default) enabling
1443 interoperability with such broken implementations. However, enabling
1444 this switch breaks interoperability with correct implementations.
1446 * Fix a use after free bug in d2i_X509_PUBKEY when overwriting a
1447 re-used X509_PUBKEY object if the second PUBKEY is malformed.
1451 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
1455 * Changed the license to the Apache License v2.0.
1459 * Switch to a new version scheme using three numbers MAJOR.MINOR.PATCH.
1461 - Major releases (indicated by incrementing the MAJOR release number)
1462 may introduce incompatible API/ABI changes.
1463 - Minor releases (indicated by incrementing the MINOR release number)
1464 may introduce new features but retain API/ABI compatibility.
1465 - Patch releases (indicated by incrementing the PATCH number)
1466 are intended for bug fixes and other improvements of existing
1467 features only (like improving performance or adding documentation)
1468 and retain API/ABI compatibility.
1472 * Add support for RFC5297 SIV mode (siv128), including AES-SIV.
1476 * Remove the 'dist' target and add a tarball building script. The
1477 'dist' target has fallen out of use, and it shouldn't be
1478 necessary to configure just to create a source distribution.
1482 * Recreate the OS390-Unix config target. It no longer relies on a
1483 special script like it did for OpenSSL pre-1.1.0.
1487 * Instead of having the source directories listed in Configure, add
1488 a 'build.info' keyword SUBDIRS to indicate what sub-directories to
1493 * Add GMAC to EVP_MAC.
1497 * Ported the HMAC, CMAC and SipHash EVP_PKEY_METHODs to EVP_MAC.
1501 * Added EVP_MAC, an EVP layer MAC API, to simplify adding MAC
1502 implementations. This includes a generic EVP_PKEY to EVP_MAC bridge,
1503 to facilitate the continued use of MACs through raw private keys in
1504 functionality such as `EVP_DigestSign*` and `EVP_DigestVerify*`.
1508 * Deprecate ECDH_KDF_X9_62().
1512 * Added EVP_PKEY_ECDH_KDF_X9_63 and ecdh_KDF_X9_63() as replacements for
1513 the EVP_PKEY_ECDH_KDF_X9_62 KDF type and ECDH_KDF_X9_62(). The old names
1514 are retained for backwards compatibility.
1518 * AES-XTS mode now enforces that its two keys are different to mitigate
1519 the attacked described in "Efficient Instantiations of Tweakable
1520 Blockciphers and Refinements to Modes OCB and PMAC" by Phillip Rogaway.
1521 Details of this attack can be obtained from:
1522 <http://web.cs.ucdavis.edu/%7Erogaway/papers/offsets.pdf>
1526 * Rename the object files, i.e. give them other names than in previous
1527 versions. Their names now include the name of the final product, as
1528 well as its type mnemonic (bin, lib, shlib).
1532 * Added new option for 'openssl list', '-objects', which will display the
1533 list of built in objects, i.e. OIDs with names.
1537 * Added the options `-crl_lastupdate` and `-crl_nextupdate` to `openssl ca`,
1538 allowing the `lastUpdate` and `nextUpdate` fields in the generated CRL to
1543 * Added support for Linux Kernel TLS data-path. The Linux Kernel data-path
1544 improves application performance by removing data copies and providing
1545 applications with zero-copy system calls such as sendfile and splice.
1549 * The SSL option SSL_OP_CLEANSE_PLAINTEXT is introduced.
1553 * `PKCS12_parse` now maintains the order of the parsed certificates
1554 when outputting them via `*ca` (rather than reversing it).
1558 * Deprecated pthread fork support methods.
1562 * Added support for FFDHE key exchange in TLS 1.3.
1566 * Added a new concept for OpenSSL plugability: providers. This
1567 functionality is designed to replace the ENGINE API and ENGINE
1568 implementations, and to be much more dynamic, allowing provider
1569 authors to introduce new algorithms among other things, as long as
1570 there's an API that supports the algorithm type.
1572 With this concept comes a new core API for interaction between
1573 libcrypto and provider implementations. Public libcrypto functions
1574 that want to use providers do so through this core API.
1576 The main documentation for this core API is found in
1577 doc/man7/provider.pod, doc/man7/provider-base.pod, and they in turn
1578 refer to other manuals describing the API specific for supported
1579 algorithm types (also called operations).
1586 ### Changes between 1.1.1l and 1.1.1m [xx XXX xxxx]
1588 * Avoid loading of a dynamic engine twice.
1592 * Prioritise DANE TLSA issuer certs over peer certs
1596 * Fixed random API for MacOS prior to 10.12
1598 These MacOS versions don't support the CommonCrypto APIs
1602 ### Changes between 1.1.1k and 1.1.1l [24 Aug 2021]
1604 * Fixed an SM2 Decryption Buffer Overflow.
1606 In order to decrypt SM2 encrypted data an application is expected to
1607 call the API function EVP_PKEY_decrypt(). Typically an application will
1608 call this function twice. The first time, on entry, the "out" parameter
1609 can be NULL and, on exit, the "outlen" parameter is populated with the
1610 buffer size required to hold the decrypted plaintext. The application
1611 can then allocate a sufficiently sized buffer and call EVP_PKEY_decrypt()
1612 again, but this time passing a non-NULL value for the "out" parameter.
1614 A bug in the implementation of the SM2 decryption code means that the
1615 calculation of the buffer size required to hold the plaintext returned
1616 by the first call to EVP_PKEY_decrypt() can be smaller than the actual
1617 size required by the second call. This can lead to a buffer overflow
1618 when EVP_PKEY_decrypt() is called by the application a second time with
1619 a buffer that is too small.
1621 A malicious attacker who is able present SM2 content for decryption to
1622 an application could cause attacker chosen data to overflow the buffer
1623 by up to a maximum of 62 bytes altering the contents of other data held
1624 after the buffer, possibly changing application behaviour or causing
1625 the application to crash. The location of the buffer is application
1626 dependent but is typically heap allocated.
1631 * Fixed various read buffer overruns processing ASN.1 strings
1633 ASN.1 strings are represented internally within OpenSSL as an ASN1_STRING
1634 structure which contains a buffer holding the string data and a field
1635 holding the buffer length. This contrasts with normal C strings which
1636 are repesented as a buffer for the string data which is terminated
1637 with a NUL (0) byte.
1639 Although not a strict requirement, ASN.1 strings that are parsed using
1640 OpenSSL's own "d2i" functions (and other similar parsing functions) as
1641 well as any string whose value has been set with the ASN1_STRING_set()
1642 function will additionally NUL terminate the byte array in the
1643 ASN1_STRING structure.
1645 However, it is possible for applications to directly construct valid
1646 ASN1_STRING structures which do not NUL terminate the byte array by
1647 directly setting the "data" and "length" fields in the ASN1_STRING
1648 array. This can also happen by using the ASN1_STRING_set0() function.
1650 Numerous OpenSSL functions that print ASN.1 data have been found to
1651 assume that the ASN1_STRING byte array will be NUL terminated, even
1652 though this is not guaranteed for strings that have been directly
1653 constructed. Where an application requests an ASN.1 structure to be
1654 printed, and where that ASN.1 structure contains ASN1_STRINGs that have
1655 been directly constructed by the application without NUL terminating
1656 the "data" field, then a read buffer overrun can occur.
1658 The same thing can also occur during name constraints processing
1659 of certificates (for example if a certificate has been directly
1660 constructed by the application instead of loading it via the OpenSSL
1661 parsing functions, and the certificate contains non NUL terminated
1662 ASN1_STRING structures). It can also occur in the X509_get1_email(),
1663 X509_REQ_get1_email() and X509_get1_ocsp() functions.
1665 If a malicious actor can cause an application to directly construct an
1666 ASN1_STRING and then process it through one of the affected OpenSSL
1667 functions then this issue could be hit. This might result in a crash
1668 (causing a Denial of Service attack). It could also result in the
1669 disclosure of private memory contents (such as private keys, or
1670 sensitive plaintext).
1675 ### Changes between 1.1.1j and 1.1.1k [25 Mar 2021]
1677 * Fixed a problem with verifying a certificate chain when using the
1678 X509_V_FLAG_X509_STRICT flag. This flag enables additional security checks of
1679 the certificates present in a certificate chain. It is not set by default.
1681 Starting from OpenSSL version 1.1.1h a check to disallow certificates in
1682 the chain that have explicitly encoded elliptic curve parameters was added
1683 as an additional strict check.
1685 An error in the implementation of this check meant that the result of a
1686 previous check to confirm that certificates in the chain are valid CA
1687 certificates was overwritten. This effectively bypasses the check
1688 that non-CA certificates must not be able to issue other certificates.
1690 If a "purpose" has been configured then there is a subsequent opportunity
1691 for checks that the certificate is a valid CA. All of the named "purpose"
1692 values implemented in libcrypto perform this check. Therefore, where
1693 a purpose is set the certificate chain will still be rejected even when the
1694 strict flag has been used. A purpose is set by default in libssl client and
1695 server certificate verification routines, but it can be overridden or
1696 removed by an application.
1698 In order to be affected, an application must explicitly set the
1699 X509_V_FLAG_X509_STRICT verification flag and either not set a purpose
1700 for the certificate verification or, in the case of TLS client or server
1701 applications, override the default purpose.
1706 * Fixed an issue where an OpenSSL TLS server may crash if sent a maliciously
1707 crafted renegotiation ClientHello message from a client. If a TLSv1.2
1708 renegotiation ClientHello omits the signature_algorithms extension (where it
1709 was present in the initial ClientHello), but includes a
1710 signature_algorithms_cert extension then a NULL pointer dereference will
1711 result, leading to a crash and a denial of service attack.
1713 A server is only vulnerable if it has TLSv1.2 and renegotiation enabled
1714 (which is the default configuration). OpenSSL TLS clients are not impacted by
1718 *Peter Kästle and Samuel Sapalski*
1720 ### Changes between 1.1.1i and 1.1.1j [16 Feb 2021]
1722 * Fixed the X509_issuer_and_serial_hash() function. It attempts to
1723 create a unique hash value based on the issuer and serial number data
1724 contained within an X509 certificate. However it was failing to correctly
1725 handle any errors that may occur while parsing the issuer field (which might
1726 occur if the issuer field is maliciously constructed). This may subsequently
1727 result in a NULL pointer deref and a crash leading to a potential denial of
1733 * Fixed the RSA_padding_check_SSLv23() function and the RSA_SSLV23_PADDING
1734 padding mode to correctly check for rollback attacks. This is considered a
1735 bug in OpenSSL 1.1.1 because it does not support SSLv2. In 1.0.2 this is
1740 Fixed the EVP_CipherUpdate, EVP_EncryptUpdate and EVP_DecryptUpdate
1741 functions. Previously they could overflow the output length argument in some
1742 cases where the input length is close to the maximum permissable length for
1743 an integer on the platform. In such cases the return value from the function
1744 call would be 1 (indicating success), but the output length value would be
1745 negative. This could cause applications to behave incorrectly or crash.
1750 * Fixed SRP_Calc_client_key so that it runs in constant time. The previous
1751 implementation called BN_mod_exp without setting BN_FLG_CONSTTIME. This
1752 could be exploited in a side channel attack to recover the password. Since
1753 the attack is local host only this is outside of the current OpenSSL
1754 threat model and therefore no CVE is assigned.
1756 Thanks to Mohammed Sabt and Daniel De Almeida Braga for reporting this
1761 ### Changes between 1.1.1h and 1.1.1i [8 Dec 2020]
1763 * Fixed NULL pointer deref in the GENERAL_NAME_cmp function
1764 This function could crash if both GENERAL_NAMEs contain an EDIPARTYNAME.
1765 If an attacker can control both items being compared then this could lead
1766 to a possible denial of service attack. OpenSSL itself uses the
1767 GENERAL_NAME_cmp function for two purposes:
1768 1) Comparing CRL distribution point names between an available CRL and a
1769 CRL distribution point embedded in an X509 certificate
1770 2) When verifying that a timestamp response token signer matches the
1771 timestamp authority name (exposed via the API functions
1772 TS_RESP_verify_response and TS_RESP_verify_token)
1777 ### Changes between 1.1.1g and 1.1.1h [22 Sep 2020]
1779 * Certificates with explicit curve parameters are now disallowed in
1780 verification chains if the X509_V_FLAG_X509_STRICT flag is used.
1784 * The 'MinProtocol' and 'MaxProtocol' configuration commands now silently
1785 ignore TLS protocol version bounds when configuring DTLS-based contexts, and
1786 conversely, silently ignore DTLS protocol version bounds when configuring
1787 TLS-based contexts. The commands can be repeated to set bounds of both
1788 types. The same applies with the corresponding "min_protocol" and
1789 "max_protocol" command-line switches, in case some application uses both TLS
1792 SSL_CTX instances that are created for a fixed protocol version (e.g.
1793 TLSv1_server_method()) also silently ignore version bounds. Previously
1794 attempts to apply bounds to these protocol versions would result in an
1795 error. Now only the "version-flexible" SSL_CTX instances are subject to
1796 limits in configuration files in command-line options.
1800 * Handshake now fails if Extended Master Secret extension is dropped
1805 * The Oracle Developer Studio compiler will start reporting deprecated APIs
1807 ### Changes between 1.1.1f and 1.1.1g [21 Apr 2020]
1809 * Fixed segmentation fault in SSL_check_chain()
1810 Server or client applications that call the SSL_check_chain() function
1811 during or after a TLS 1.3 handshake may crash due to a NULL pointer
1812 dereference as a result of incorrect handling of the
1813 "signature_algorithms_cert" TLS extension. The crash occurs if an invalid
1814 or unrecognised signature algorithm is received from the peer. This could
1815 be exploited by a malicious peer in a Denial of Service attack.
1820 * Added AES consttime code for no-asm configurations
1821 an optional constant time support for AES was added
1822 when building openssl for no-asm.
1823 Enable with: ./config no-asm -DOPENSSL_AES_CONST_TIME
1824 Disable with: ./config no-asm -DOPENSSL_NO_AES_CONST_TIME
1825 At this time this feature is by default disabled.
1826 It will be enabled by default in 3.0.
1830 ### Changes between 1.1.1e and 1.1.1f [31 Mar 2020]
1832 * Revert the change of EOF detection while reading in libssl to avoid
1833 regressions in applications depending on the current way of reporting
1834 the EOF. As the existing method is not fully accurate the change to
1835 reporting the EOF via SSL_ERROR_SSL is kept on the current development
1836 branch and will be present in the 3.0 release.
1840 * Revised BN_generate_prime_ex to not avoid factors 3..17863 in p-1
1841 when primes for RSA keys are computed.
1842 Since we previously always generated primes == 2 (mod 3) for RSA keys,
1843 the 2-prime and 3-prime RSA modules were easy to distinguish, since
1844 N = p*q = 1 (mod 3), but N = p*q*r = 2 (mod 3). Therefore fingerprinting
1845 2-prime vs. 3-prime RSA keys was possible by computing N mod 3.
1846 This avoids possible fingerprinting of newly generated RSA modules.
1850 ### Changes between 1.1.1d and 1.1.1e [17 Mar 2020]
1852 * Properly detect EOF while reading in libssl. Previously if we hit an EOF
1853 while reading in libssl then we would report an error back to the
1854 application (SSL_ERROR_SYSCALL) but errno would be 0. We now add
1855 an error to the stack (which means we instead return SSL_ERROR_SSL) and
1856 therefore give a hint as to what went wrong.
1860 * Check that ed25519 and ed448 are allowed by the security level. Previously
1861 signature algorithms not using an MD were not being checked that they were
1862 allowed by the security level.
1866 * Fixed SSL_get_servername() behaviour. The behaviour of SSL_get_servername()
1867 was not quite right. The behaviour was not consistent between resumption
1868 and normal handshakes, and also not quite consistent with historical
1869 behaviour. The behaviour in various scenarios has been clarified and
1870 it has been updated to make it match historical behaviour as closely as
1875 * *[VMS only]* The header files that the VMS compilers include automatically,
1876 `__DECC_INCLUDE_PROLOGUE.H` and `__DECC_INCLUDE_EPILOGUE.H`, use pragmas
1877 that the C++ compiler doesn't understand. This is a shortcoming in the
1878 compiler, but can be worked around with `__cplusplus` guards.
1880 C++ applications that use OpenSSL libraries must be compiled using the
1881 qualifier `/NAMES=(AS_IS,SHORTENED)` to be able to use all the OpenSSL
1882 functions. Otherwise, only functions with symbols of less than 31
1883 characters can be used, as the linker will not be able to successfully
1884 resolve symbols with longer names.
1888 * Added a new method to gather entropy on VMS, based on SYS$GET_ENTROPY.
1889 The presence of this system service is determined at run-time.
1893 * Print all values for a PKCS#12 attribute with 'openssl pkcs12', not just
1898 ### Changes between 1.1.1c and 1.1.1d [10 Sep 2019]
1900 * Fixed a fork protection issue. OpenSSL 1.1.1 introduced a rewritten random
1901 number generator (RNG). This was intended to include protection in the
1902 event of a fork() system call in order to ensure that the parent and child
1903 processes did not share the same RNG state. However this protection was not
1904 being used in the default case.
1906 A partial mitigation for this issue is that the output from a high
1907 precision timer is mixed into the RNG state so the likelihood of a parent
1908 and child process sharing state is significantly reduced.
1910 If an application already calls OPENSSL_init_crypto() explicitly using
1911 OPENSSL_INIT_ATFORK then this problem does not occur at all.
1914 *Matthias St. Pierre*
1916 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
1917 used even when parsing explicit parameters, when loading a encoded key
1918 or calling `EC_GROUP_new_from_ecpkparameters()`/
1919 `EC_GROUP_new_from_ecparameters()`.
1920 This prevents bypass of security hardening and performance gains,
1921 especially for curves with specialized EC_METHODs.
1922 By default, if a key encoded with explicit parameters is loaded and later
1923 encoded, the output is still encoded with explicit parameters, even if
1924 internally a "named" EC_GROUP is used for computation.
1928 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
1929 this change, EC_GROUP_set_generator would accept order and/or cofactor as
1930 NULL. After this change, only the cofactor parameter can be NULL. It also
1931 does some minimal sanity checks on the passed order.
1936 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
1937 An attack is simple, if the first CMS_recipientInfo is valid but the
1938 second CMS_recipientInfo is chosen ciphertext. If the second
1939 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
1940 encryption key will be replaced by garbage, and the message cannot be
1941 decoded, but if the RSA decryption fails, the correct encryption key is
1942 used and the recipient will not notice the attack.
1943 As a work around for this potential attack the length of the decrypted
1944 key must be equal to the cipher default key length, in case the
1945 certifiate is not given and all recipientInfo are tried out.
1946 The old behaviour can be re-enabled in the CMS code by setting the
1947 CMS_DEBUG_DECRYPT flag.
1952 * Early start up entropy quality from the DEVRANDOM seed source has been
1953 improved for older Linux systems. The RAND subsystem will wait for
1954 /dev/random to be producing output before seeding from /dev/urandom.
1955 The seeded state is stored for future library initialisations using
1956 a system global shared memory segment. The shared memory identifier
1957 can be configured by defining OPENSSL_RAND_SEED_DEVRANDOM_SHM_ID to
1958 the desired value. The default identifier is 114.
1962 * Correct the extended master secret constant on EBCDIC systems. Without this
1963 fix TLS connections between an EBCDIC system and a non-EBCDIC system that
1964 negotiate EMS will fail. Unfortunately this also means that TLS connections
1965 between EBCDIC systems with this fix, and EBCDIC systems without this
1966 fix will fail if they negotiate EMS.
1970 * Use Windows installation paths in the mingw builds
1972 Mingw isn't a POSIX environment per se, which means that Windows
1973 paths should be used for installation.
1978 * Changed DH_check to accept parameters with order q and 2q subgroups.
1979 With order 2q subgroups the bit 0 of the private key is not secret
1980 but DH_generate_key works around that by clearing bit 0 of the
1981 private key for those. This avoids leaking bit 0 of the private key.
1985 * Significantly reduce secure memory usage by the randomness pools.
1989 * Revert the DEVRANDOM_WAIT feature for Linux systems
1991 The DEVRANDOM_WAIT feature added a select() call to wait for the
1992 /dev/random device to become readable before reading from the
1993 /dev/urandom device.
1995 It turned out that this change had negative side effects on
1996 performance which were not acceptable. After some discussion it
1997 was decided to revert this feature and leave it up to the OS
1998 resp. the platform maintainer to ensure a proper initialization
1999 during early boot time.
2001 *Matthias St. Pierre*
2003 ### Changes between 1.1.1b and 1.1.1c [28 May 2019]
2005 * Add build tests for C++. These are generated files that only do one
2006 thing, to include one public OpenSSL head file each. This tests that
2007 the public header files can be usefully included in a C++ application.
2009 This test isn't enabled by default. It can be enabled with the option
2010 'enable-buildtest-c++'.
2014 * Enable SHA3 pre-hashing for ECDSA and DSA.
2018 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
2019 This changes the size when using the `genpkey` command when no size is given.
2020 It fixes an omission in earlier changes that changed all RSA, DSA and DH
2021 generation commands to use 2048 bits by default.
2025 * Reorganize the manual pages to consistently have RETURN VALUES,
2026 EXAMPLES, SEE ALSO and HISTORY come in that order, and adjust
2027 util/fix-doc-nits accordingly.
2029 *Paul Yang, Joshua Lock*
2031 * Add the missing accessor EVP_PKEY_get0_engine()
2035 * Have commands like `s_client` and `s_server` output the signature scheme
2036 along with other cipher suite parameters when debugging.
2040 * Make OPENSSL_config() error agnostic again.
2044 * Do the error handling in RSA decryption constant time.
2048 * Prevent over long nonces in ChaCha20-Poly1305.
2050 ChaCha20-Poly1305 is an AEAD cipher, and requires a unique nonce input
2051 for every encryption operation. RFC 7539 specifies that the nonce value
2052 (IV) should be 96 bits (12 bytes). OpenSSL allows a variable nonce length
2053 and front pads the nonce with 0 bytes if it is less than 12
2054 bytes. However it also incorrectly allows a nonce to be set of up to 16
2055 bytes. In this case only the last 12 bytes are significant and any
2056 additional leading bytes are ignored.
2058 It is a requirement of using this cipher that nonce values are
2059 unique. Messages encrypted using a reused nonce value are susceptible to
2060 serious confidentiality and integrity attacks. If an application changes
2061 the default nonce length to be longer than 12 bytes and then makes a
2062 change to the leading bytes of the nonce expecting the new value to be a
2063 new unique nonce then such an application could inadvertently encrypt
2064 messages with a reused nonce.
2066 Additionally the ignored bytes in a long nonce are not covered by the
2067 integrity guarantee of this cipher. Any application that relies on the
2068 integrity of these ignored leading bytes of a long nonce may be further
2069 affected. Any OpenSSL internal use of this cipher, including in SSL/TLS,
2070 is safe because no such use sets such a long nonce value. However user
2071 applications that use this cipher directly and set a non-default nonce
2072 length to be longer than 12 bytes may be vulnerable.
2074 This issue was reported to OpenSSL on 16th of March 2019 by Joran Dirk
2080 * Add DEVRANDOM_WAIT feature for Linux systems
2082 On older Linux systems where the getrandom() system call is not available,
2083 OpenSSL normally uses the /dev/urandom device for seeding its CSPRNG.
2084 Contrary to getrandom(), the /dev/urandom device will not block during
2085 early boot when the kernel CSPRNG has not been seeded yet.
2087 To mitigate this known weakness, use select() to wait for /dev/random to
2088 become readable before reading from /dev/urandom.
2090 * Ensure that SM2 only uses SM3 as digest algorithm
2094 ### Changes between 1.1.1a and 1.1.1b [26 Feb 2019]
2096 * Change the info callback signals for the start and end of a post-handshake
2097 message exchange in TLSv1.3. In 1.1.1/1.1.1a we used SSL_CB_HANDSHAKE_START
2098 and SSL_CB_HANDSHAKE_DONE. Experience has shown that many applications get
2099 confused by this and assume that a TLSv1.2 renegotiation has started. This
2100 can break KeyUpdate handling. Instead we no longer signal the start and end
2101 of a post handshake message exchange (although the messages themselves are
2102 still signalled). This could break some applications that were expecting
2103 the old signals. However without this KeyUpdate is not usable for many
2108 ### Changes between 1.1.1 and 1.1.1a [20 Nov 2018]
2110 * Timing vulnerability in DSA signature generation
2112 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
2113 timing side channel attack. An attacker could use variations in the signing
2114 algorithm to recover the private key.
2116 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
2121 * Timing vulnerability in ECDSA signature generation
2123 The OpenSSL ECDSA signature algorithm has been shown to be vulnerable to a
2124 timing side channel attack. An attacker could use variations in the signing
2125 algorithm to recover the private key.
2127 This issue was reported to OpenSSL on 25th October 2018 by Samuel Weiser.
2132 * Fixed the issue that RAND_add()/RAND_seed() silently discards random input
2133 if its length exceeds 4096 bytes. The limit has been raised to a buffer size
2134 of two gigabytes and the error handling improved.
2136 This issue was reported to OpenSSL by Dr. Falko Strenzke. It has been
2137 categorized as a normal bug, not a security issue, because the DRBG reseeds
2138 automatically and is fully functional even without additional randomness
2139 provided by the application.
2141 ### Changes between 1.1.0i and 1.1.1 [11 Sep 2018]
2143 * Add a new ClientHello callback. Provides a callback interface that gives
2144 the application the ability to adjust the nascent SSL object at the
2145 earliest stage of ClientHello processing, immediately after extensions have
2146 been collected but before they have been processed. In particular, this
2147 callback can adjust the supported TLS versions in response to the contents
2152 * Add SM2 base algorithm support.
2156 * s390x assembly pack: add (improved) hardware-support for the following
2157 cryptographic primitives: sha3, shake, aes-gcm, aes-ccm, aes-ctr, aes-ofb,
2158 aes-cfb/cfb8, aes-ecb.
2162 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
2163 parameter is no longer accepted, as it leads to a corrupt table. NULL
2164 pem_str is reserved for alias entries only.
2168 * Use the new ec_scalar_mul_ladder scaffold to implement a specialized ladder
2169 step for prime curves. The new implementation is based on formulae from
2170 differential addition-and-doubling in homogeneous projective coordinates
2171 from Izu-Takagi "A fast parallel elliptic curve multiplication resistant
2172 against side channel attacks" and Brier-Joye "Weierstrass Elliptic Curves
2173 and Side-Channel Attacks" Eq. (8) for y-coordinate recovery, modified
2174 to work in projective coordinates.
2176 *Billy Bob Brumley, Nicola Tuveri*
2178 * Change generating and checking of primes so that the error rate of not
2179 being prime depends on the intended use based on the size of the input.
2180 For larger primes this will result in more rounds of Miller-Rabin.
2181 The maximal error rate for primes with more than 1080 bits is lowered
2184 *Kurt Roeckx, Annie Yousar*
2186 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
2190 * The 'tsget' script is renamed to 'tsget.pl', to avoid confusion when
2191 moving between systems, and to avoid confusion when a Windows build is
2192 done with mingw vs with MSVC. For POSIX installs, there's still a
2193 symlink or copy named 'tsget' to avoid that confusion as well.
2197 * Revert blinding in ECDSA sign and instead make problematic addition
2198 length-invariant. Switch even to fixed-length Montgomery multiplication.
2202 * Use the new ec_scalar_mul_ladder scaffold to implement a specialized ladder
2203 step for binary curves. The new implementation is based on formulae from
2204 differential addition-and-doubling in mixed Lopez-Dahab projective
2205 coordinates, modified to independently blind the operands.
2207 *Billy Bob Brumley, Sohaib ul Hassan, Nicola Tuveri*
2209 * Add a scaffold to optionally enhance the Montgomery ladder implementation
2210 for `ec_scalar_mul_ladder` (formerly `ec_mul_consttime`) allowing
2211 EC_METHODs to implement their own specialized "ladder step", to take
2212 advantage of more favorable coordinate systems or more efficient
2213 differential addition-and-doubling algorithms.
2215 *Billy Bob Brumley, Sohaib ul Hassan, Nicola Tuveri*
2217 * Modified the random device based seed sources to keep the relevant
2218 file descriptors open rather than reopening them on each access.
2219 This allows such sources to operate in a chroot() jail without
2220 the associated device nodes being available. This behaviour can be
2221 controlled using RAND_keep_random_devices_open().
2225 * Numerous side-channel attack mitigations have been applied. This may have
2226 performance impacts for some algorithms for the benefit of improved
2227 security. Specific changes are noted in this change log by their respective
2232 * AIX shared library support overhaul. Switch to AIX "natural" way of
2233 handling shared libraries, which means collecting shared objects of
2234 different versions and bitnesses in one common archive. This allows to
2235 mitigate conflict between 1.0 and 1.1 side-by-side installations. It
2236 doesn't affect the way 3rd party applications are linked, only how
2237 multi-version installation is managed.
2241 * Make ec_group_do_inverse_ord() more robust and available to other
2242 EC cryptosystems, so that irrespective of BN_FLG_CONSTTIME, SCA
2243 mitigations are applied to the fallback BN_mod_inverse().
2244 When using this function rather than BN_mod_inverse() directly, new
2245 EC cryptosystem implementations are then safer-by-default.
2249 * Add coordinate blinding for EC_POINT and implement projective
2250 coordinate blinding for generic prime curves as a countermeasure to
2251 chosen point SCA attacks.
2253 *Sohaib ul Hassan, Nicola Tuveri, Billy Bob Brumley*
2255 * Add blinding to ECDSA and DSA signatures to protect against side channel
2256 attacks discovered by Keegan Ryan (NCC Group).
2260 * Enforce checking in the `pkeyutl` command to ensure that the input
2261 length does not exceed the maximum supported digest length when performing
2262 a sign, verify or verifyrecover operation.
2266 * SSL_MODE_AUTO_RETRY is enabled by default. Applications that use blocking
2267 I/O in combination with something like select() or poll() will hang. This
2268 can be turned off again using SSL_CTX_clear_mode().
2269 Many applications do not properly handle non-application data records, and
2270 TLS 1.3 sends more of such records. Setting SSL_MODE_AUTO_RETRY works
2271 around the problems in those applications, but can also break some.
2272 It's recommended to read the manpages about SSL_read(), SSL_write(),
2273 SSL_get_error(), SSL_shutdown(), SSL_CTX_set_mode() and
2274 SSL_CTX_set_read_ahead() again.
2278 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
2279 now allow empty (zero character) pass phrases.
2283 * Apply blinding to binary field modular inversion and remove patent
2284 pending (OPENSSL_SUN_GF2M_DIV) BN_GF2m_mod_div implementation.
2288 * Deprecate ec2_mult.c and unify scalar multiplication code paths for
2289 binary and prime elliptic curves.
2293 * Remove ECDSA nonce padding: EC_POINT_mul is now responsible for
2294 constant time fixed point multiplication.
2298 * Revise elliptic curve scalar multiplication with timing attack
2299 defenses: ec_wNAF_mul redirects to a constant time implementation
2300 when computing fixed point and variable point multiplication (which
2301 in OpenSSL are mostly used with secret scalars in keygen, sign,
2302 ECDH derive operations).
2303 *Billy Bob Brumley, Nicola Tuveri, Cesar Pereida García,
2306 * Updated CONTRIBUTING
2310 * Updated DRBG / RAND to request nonce and additional low entropy
2311 randomness from the system.
2313 *Matthias St. Pierre*
2315 * Updated 'openssl rehash' to use OpenSSL consistent default.
2319 * Moved the load of the ssl_conf module to libcrypto, which helps
2320 loading engines that libssl uses before libssl is initialised.
2324 * Added EVP_PKEY_sign() and EVP_PKEY_verify() for EdDSA
2328 * Fixed X509_NAME_ENTRY_set to get multi-valued RDNs right in all cases.
2330 *Ingo Schwarze, Rich Salz*
2332 * Added output of accepting IP address and port for 'openssl s_server'
2336 * Added a new API for TLSv1.3 ciphersuites:
2337 SSL_CTX_set_ciphersuites()
2338 SSL_set_ciphersuites()
2342 * Memory allocation failures consistently add an error to the error
2347 * Don't use OPENSSL_ENGINES and OPENSSL_CONF environment values
2348 in libcrypto when run as setuid/setgid.
2352 * Load any config file by default when libssl is used.
2356 * Added new public header file <openssl/rand_drbg.h> and documentation
2357 for the RAND_DRBG API. See manual page RAND_DRBG(7) for an overview.
2359 *Matthias St. Pierre*
2361 * QNX support removed (cannot find contributors to get their approval
2362 for the license change).
2366 * TLSv1.3 replay protection for early data has been implemented. See the
2367 SSL_read_early_data() man page for further details.
2371 * Separated TLSv1.3 ciphersuite configuration out from TLSv1.2 ciphersuite
2372 configuration. TLSv1.3 ciphersuites are not compatible with TLSv1.2 and
2373 below. Similarly TLSv1.2 ciphersuites are not compatible with TLSv1.3.
2374 In order to avoid issues where legacy TLSv1.2 ciphersuite configuration
2375 would otherwise inadvertently disable all TLSv1.3 ciphersuites the
2376 configuration has been separated out. See the ciphers man page or the
2377 SSL_CTX_set_ciphersuites() man page for more information.
2381 * On POSIX (BSD, Linux, ...) systems the ocsp(1) command running
2382 in responder mode now supports the new "-multi" option, which
2383 spawns the specified number of child processes to handle OCSP
2384 requests. The "-timeout" option now also limits the OCSP
2385 responder's patience to wait to receive the full client request
2386 on a newly accepted connection. Child processes are respawned
2387 as needed, and the CA index file is automatically reloaded
2388 when changed. This makes it possible to run the "ocsp" responder
2389 as a long-running service, making the OpenSSL CA somewhat more
2390 feature-complete. In this mode, most diagnostic messages logged
2391 after entering the event loop are logged via syslog(3) rather than
2396 * Added support for X448 and Ed448. Heavily based on original work by
2401 * Extend OSSL_STORE with capabilities to search and to narrow the set of
2402 objects loaded. This adds the functions OSSL_STORE_expect() and
2403 OSSL_STORE_find() as well as needed tools to construct searches and
2404 get the search data out of them.
2408 * Support for TLSv1.3 added. Note that users upgrading from an earlier
2409 version of OpenSSL should review their configuration settings to ensure
2410 that they are still appropriate for TLSv1.3. For further information see:
2411 <https://wiki.openssl.org/index.php/TLS1.3>
2415 * Grand redesign of the OpenSSL random generator
2417 The default RAND method now utilizes an AES-CTR DRBG according to
2418 NIST standard SP 800-90Ar1. The new random generator is essentially
2419 a port of the default random generator from the OpenSSL FIPS 2.0
2420 object module. It is a hybrid deterministic random bit generator
2421 using an AES-CTR bit stream and which seeds and reseeds itself
2422 automatically using trusted system entropy sources.
2424 Some of its new features are:
2425 - Support for multiple DRBG instances with seed chaining.
2426 - The default RAND method makes use of a DRBG.
2427 - There is a public and private DRBG instance.
2428 - The DRBG instances are fork-safe.
2429 - Keep all global DRBG instances on the secure heap if it is enabled.
2430 - The public and private DRBG instance are per thread for lock free
2433 *Paul Dale, Benjamin Kaduk, Kurt Roeckx, Rich Salz, Matthias St. Pierre*
2435 * Changed Configure so it only says what it does and doesn't dump
2436 so much data. Instead, ./configdata.pm should be used as a script
2437 to display all sorts of configuration data.
2441 * Added processing of "make variables" to Configure.
2445 * Added SHA512/224 and SHA512/256 algorithm support.
2449 * The last traces of Netware support, first removed in 1.1.0, have
2454 * Get rid of Makefile.shared, and in the process, make the processing
2455 of certain files (rc.obj, or the .def/.map/.opt files produced from
2456 the ordinal files) more visible and hopefully easier to trace and
2457 debug (or make silent).
2461 * Make it possible to have environment variable assignments as
2462 arguments to config / Configure.
2466 * Add multi-prime RSA (RFC 8017) support.
2470 * Add SM3 implemented according to GB/T 32905-2016
2471 *Jack Lloyd <jack.lloyd@ribose.com>,*
2472 *Ronald Tse <ronald.tse@ribose.com>,*
2473 *Erick Borsboom <erick.borsboom@ribose.com>*
2475 * Add 'Maximum Fragment Length' TLS extension negotiation and support
2476 as documented in RFC6066.
2477 Based on a patch from Tomasz Moń
2479 *Filipe Raimundo da Silva*
2481 * Add SM4 implemented according to GB/T 32907-2016.
2482 *Jack Lloyd <jack.lloyd@ribose.com>,*
2483 *Ronald Tse <ronald.tse@ribose.com>,*
2484 *Erick Borsboom <erick.borsboom@ribose.com>*
2486 * Reimplement -newreq-nodes and ERR_error_string_n; the
2487 original author does not agree with the license change.
2491 * Add ARIA AEAD TLS support.
2495 * Some macro definitions to support VS6 have been removed. Visual
2496 Studio 6 has not worked since 1.1.0
2500 * Add ERR_clear_last_mark(), to allow callers to clear the last mark
2501 without clearing the errors.
2505 * Add "atfork" functions. If building on a system that without
2506 pthreads, see doc/man3/OPENSSL_fork_prepare.pod for application
2507 requirements. The RAND facility now uses/requires this.
2515 * The UI API becomes a permanent and integral part of libcrypto, i.e.
2516 not possible to disable entirely. However, it's still possible to
2517 disable the console reading UI method, UI_OpenSSL() (use UI_null()
2520 To disable, configure with 'no-ui-console'. 'no-ui' is still
2521 possible to use as an alias. Check at compile time with the
2522 macro OPENSSL_NO_UI_CONSOLE. The macro OPENSSL_NO_UI is still
2523 possible to check and is an alias for OPENSSL_NO_UI_CONSOLE.
2527 * Add a STORE module, which implements a uniform and URI based reader of
2528 stores that can contain keys, certificates, CRLs and numerous other
2529 objects. The main API is loosely based on a few stdio functions,
2530 and includes OSSL_STORE_open, OSSL_STORE_load, OSSL_STORE_eof,
2531 OSSL_STORE_error and OSSL_STORE_close.
2532 The implementation uses backends called "loaders" to implement arbitrary
2533 URI schemes. There is one built in "loader" for the 'file' scheme.
2537 * Add devcrypto engine. This has been implemented against cryptodev-linux,
2538 then adjusted to work on FreeBSD 8.4 as well.
2539 Enable by configuring with 'enable-devcryptoeng'. This is done by default
2540 on BSD implementations, as cryptodev.h is assumed to exist on all of them.
2544 * Module names can prefixed with OSSL_ or OPENSSL_. This affects
2545 util/mkerr.pl, which is adapted to allow those prefixes, leading to
2546 error code calls like this:
2548 OSSL_FOOerr(OSSL_FOO_F_SOMETHING, OSSL_FOO_R_WHATEVER);
2550 With this change, we claim the namespaces OSSL and OPENSSL in a manner
2551 that can be encoded in C. For the foreseeable future, this will only
2554 *Richard Levitte and Tim Hudson*
2556 * Removed BSD cryptodev engine.
2560 * Add a build target 'build_all_generated', to build all generated files
2561 and only that. This can be used to prepare everything that requires
2562 things like perl for a system that lacks perl and then move everything
2563 to that system and do the rest of the build there.
2567 * In the UI interface, make it possible to duplicate the user data. This
2568 can be used by engines that need to retain the data for a longer time
2569 than just the call where this user data is passed.
2573 * Ignore the '-named_curve auto' value for compatibility of applications
2576 *Tomáš Mráz <tmraz@fedoraproject.org>*
2578 * Fragmented SSL/TLS alerts are no longer accepted. An alert message is 2
2579 bytes long. In theory it is permissible in SSLv3 - TLSv1.2 to fragment such
2580 alerts across multiple records (some of which could be empty). In practice
2581 it make no sense to send an empty alert record, or to fragment one. TLSv1.3
2582 prohibits this altogether and other libraries (BoringSSL, NSS) do not
2583 support this at all. Supporting it adds significant complexity to the
2584 record layer, and its removal is unlikely to cause interoperability
2589 * Add the ASN.1 types INT32, UINT32, INT64, UINT64 and variants prefixed
2590 with Z. These are meant to replace LONG and ZLONG and to be size safe.
2591 The use of LONG and ZLONG is discouraged and scheduled for deprecation
2596 * Add the 'z' and 'j' modifiers to BIO_printf() et al formatting string,
2597 'z' is to be used for [s]size_t, and 'j' - with [u]int64_t.
2599 *Richard Levitte, Andy Polyakov*
2601 * Add EC_KEY_get0_engine(), which does for EC_KEY what RSA_get0_engine()
2606 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
2607 platform rather than 'mingw'.
2611 * The functions X509_STORE_add_cert and X509_STORE_add_crl return
2612 success if they are asked to add an object which already exists
2613 in the store. This change cascades to other functions which load
2614 certificates and CRLs.
2618 * x86_64 assembly pack: annotate code with DWARF CFI directives to
2619 facilitate stack unwinding even from assembly subroutines.
2623 * Remove VAX C specific definitions of OPENSSL_EXPORT, OPENSSL_EXTERN.
2624 Also remove OPENSSL_GLOBAL entirely, as it became a no-op.
2628 * Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
2629 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
2630 which is the minimum version we support.
2634 * Certificate time validation (X509_cmp_time) enforces stricter
2635 compliance with RFC 5280. Fractional seconds and timezone offsets
2636 are no longer allowed.
2640 * Add support for ARIA
2644 * s_client will now send the Server Name Indication (SNI) extension by
2645 default unless the new "-noservername" option is used. The server name is
2646 based on the host provided to the "-connect" option unless overridden by
2647 using "-servername".
2651 * Add support for SipHash
2655 * OpenSSL now fails if it receives an unrecognised record type in TLS1.0
2656 or TLS1.1. Previously this only happened in SSLv3 and TLS1.2. This is to
2657 prevent issues where no progress is being made and the peer continually
2658 sends unrecognised record types, using up resources processing them.
2662 * 'openssl passwd' can now produce SHA256 and SHA512 based output,
2663 using the algorithm defined in
2664 <https://www.akkadia.org/drepper/SHA-crypt.txt>
2668 * Heartbeat support has been removed; the ABI is changed for now.
2670 *Richard Levitte, Rich Salz*
2672 * Support for SSL_OP_NO_ENCRYPT_THEN_MAC in SSL_CONF_cmd.
2676 * The RSA "null" method, which was partially supported to avoid patent
2677 issues, has been replaced to always returns NULL.
2684 ### Changes between 1.1.0k and 1.1.0l [10 Sep 2019]
2686 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
2687 used even when parsing explicit parameters, when loading a encoded key
2688 or calling `EC_GROUP_new_from_ecpkparameters()`/
2689 `EC_GROUP_new_from_ecparameters()`.
2690 This prevents bypass of security hardening and performance gains,
2691 especially for curves with specialized EC_METHODs.
2692 By default, if a key encoded with explicit parameters is loaded and later
2693 encoded, the output is still encoded with explicit parameters, even if
2694 internally a "named" EC_GROUP is used for computation.
2698 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
2699 this change, EC_GROUP_set_generator would accept order and/or cofactor as
2700 NULL. After this change, only the cofactor parameter can be NULL. It also
2701 does some minimal sanity checks on the passed order.
2706 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
2707 An attack is simple, if the first CMS_recipientInfo is valid but the
2708 second CMS_recipientInfo is chosen ciphertext. If the second
2709 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
2710 encryption key will be replaced by garbage, and the message cannot be
2711 decoded, but if the RSA decryption fails, the correct encryption key is
2712 used and the recipient will not notice the attack.
2713 As a work around for this potential attack the length of the decrypted
2714 key must be equal to the cipher default key length, in case the
2715 certifiate is not given and all recipientInfo are tried out.
2716 The old behaviour can be re-enabled in the CMS code by setting the
2717 CMS_DEBUG_DECRYPT flag.
2722 * Use Windows installation paths in the mingw builds
2724 Mingw isn't a POSIX environment per se, which means that Windows
2725 paths should be used for installation.
2730 ### Changes between 1.1.0j and 1.1.0k [28 May 2019]
2732 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
2733 This changes the size when using the `genpkey` command when no size is given.
2734 It fixes an omission in earlier changes that changed all RSA, DSA and DH
2735 generation commands to use 2048 bits by default.
2739 * Prevent over long nonces in ChaCha20-Poly1305.
2741 ChaCha20-Poly1305 is an AEAD cipher, and requires a unique nonce input
2742 for every encryption operation. RFC 7539 specifies that the nonce value
2743 (IV) should be 96 bits (12 bytes). OpenSSL allows a variable nonce length
2744 and front pads the nonce with 0 bytes if it is less than 12
2745 bytes. However it also incorrectly allows a nonce to be set of up to 16
2746 bytes. In this case only the last 12 bytes are significant and any
2747 additional leading bytes are ignored.
2749 It is a requirement of using this cipher that nonce values are
2750 unique. Messages encrypted using a reused nonce value are susceptible to
2751 serious confidentiality and integrity attacks. If an application changes
2752 the default nonce length to be longer than 12 bytes and then makes a
2753 change to the leading bytes of the nonce expecting the new value to be a
2754 new unique nonce then such an application could inadvertently encrypt
2755 messages with a reused nonce.
2757 Additionally the ignored bytes in a long nonce are not covered by the
2758 integrity guarantee of this cipher. Any application that relies on the
2759 integrity of these ignored leading bytes of a long nonce may be further
2760 affected. Any OpenSSL internal use of this cipher, including in SSL/TLS,
2761 is safe because no such use sets such a long nonce value. However user
2762 applications that use this cipher directly and set a non-default nonce
2763 length to be longer than 12 bytes may be vulnerable.
2765 This issue was reported to OpenSSL on 16th of March 2019 by Joran Dirk
2771 * Added SCA hardening for modular field inversion in EC_GROUP through
2772 a new dedicated field_inv() pointer in EC_METHOD.
2773 This also addresses a leakage affecting conversions from projective
2774 to affine coordinates.
2776 *Billy Bob Brumley, Nicola Tuveri*
2778 * Fix a use after free bug in d2i_X509_PUBKEY when overwriting a
2779 re-used X509_PUBKEY object if the second PUBKEY is malformed.
2783 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
2787 * Remove the 'dist' target and add a tarball building script. The
2788 'dist' target has fallen out of use, and it shouldn't be
2789 necessary to configure just to create a source distribution.
2793 ### Changes between 1.1.0i and 1.1.0j [20 Nov 2018]
2795 * Timing vulnerability in DSA signature generation
2797 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
2798 timing side channel attack. An attacker could use variations in the signing
2799 algorithm to recover the private key.
2801 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
2806 * Timing vulnerability in ECDSA signature generation
2808 The OpenSSL ECDSA signature algorithm has been shown to be vulnerable to a
2809 timing side channel attack. An attacker could use variations in the signing
2810 algorithm to recover the private key.
2812 This issue was reported to OpenSSL on 25th October 2018 by Samuel Weiser.
2817 * Add coordinate blinding for EC_POINT and implement projective
2818 coordinate blinding for generic prime curves as a countermeasure to
2819 chosen point SCA attacks.
2821 *Sohaib ul Hassan, Nicola Tuveri, Billy Bob Brumley*
2823 ### Changes between 1.1.0h and 1.1.0i [14 Aug 2018]
2825 * Client DoS due to large DH parameter
2827 During key agreement in a TLS handshake using a DH(E) based ciphersuite a
2828 malicious server can send a very large prime value to the client. This will
2829 cause the client to spend an unreasonably long period of time generating a
2830 key for this prime resulting in a hang until the client has finished. This
2831 could be exploited in a Denial Of Service attack.
2833 This issue was reported to OpenSSL on 5th June 2018 by Guido Vranken
2838 * Cache timing vulnerability in RSA Key Generation
2840 The OpenSSL RSA Key generation algorithm has been shown to be vulnerable to
2841 a cache timing side channel attack. An attacker with sufficient access to
2842 mount cache timing attacks during the RSA key generation process could
2843 recover the private key.
2845 This issue was reported to OpenSSL on 4th April 2018 by Alejandro Cabrera
2846 Aldaya, Billy Brumley, Cesar Pereida Garcia and Luis Manuel Alvarez Tapia.
2851 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
2852 parameter is no longer accepted, as it leads to a corrupt table. NULL
2853 pem_str is reserved for alias entries only.
2857 * Revert blinding in ECDSA sign and instead make problematic addition
2858 length-invariant. Switch even to fixed-length Montgomery multiplication.
2862 * Change generating and checking of primes so that the error rate of not
2863 being prime depends on the intended use based on the size of the input.
2864 For larger primes this will result in more rounds of Miller-Rabin.
2865 The maximal error rate for primes with more than 1080 bits is lowered
2868 *Kurt Roeckx, Annie Yousar*
2870 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
2874 * Add blinding to ECDSA and DSA signatures to protect against side channel
2875 attacks discovered by Keegan Ryan (NCC Group).
2879 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
2880 now allow empty (zero character) pass phrases.
2884 * Certificate time validation (X509_cmp_time) enforces stricter
2885 compliance with RFC 5280. Fractional seconds and timezone offsets
2886 are no longer allowed.
2890 * Fixed a text canonicalisation bug in CMS
2892 Where a CMS detached signature is used with text content the text goes
2893 through a canonicalisation process first prior to signing or verifying a
2894 signature. This process strips trailing space at the end of lines, converts
2895 line terminators to CRLF and removes additional trailing line terminators
2896 at the end of a file. A bug in the canonicalisation process meant that
2897 some characters, such as form-feed, were incorrectly treated as whitespace
2898 and removed. This is contrary to the specification (RFC5485). This fix
2899 could mean that detached text data signed with an earlier version of
2900 OpenSSL 1.1.0 may fail to verify using the fixed version, or text data
2901 signed with a fixed OpenSSL may fail to verify with an earlier version of
2902 OpenSSL 1.1.0. A workaround is to only verify the canonicalised text data
2903 and use the "-binary" flag (for the "cms" command line application) or set
2904 the SMIME_BINARY/PKCS7_BINARY/CMS_BINARY flags (if using CMS_verify()).
2908 ### Changes between 1.1.0g and 1.1.0h [27 Mar 2018]
2910 * Constructed ASN.1 types with a recursive definition could exceed the stack
2912 Constructed ASN.1 types with a recursive definition (such as can be found
2913 in PKCS7) could eventually exceed the stack given malicious input with
2914 excessive recursion. This could result in a Denial Of Service attack. There
2915 are no such structures used within SSL/TLS that come from untrusted sources
2916 so this is considered safe.
2918 This issue was reported to OpenSSL on 4th January 2018 by the OSS-fuzz
2924 * Incorrect CRYPTO_memcmp on HP-UX PA-RISC
2926 Because of an implementation bug the PA-RISC CRYPTO_memcmp function is
2927 effectively reduced to only comparing the least significant bit of each
2928 byte. This allows an attacker to forge messages that would be considered as
2929 authenticated in an amount of tries lower than that guaranteed by the
2930 security claims of the scheme. The module can only be compiled by the
2931 HP-UX assembler, so that only HP-UX PA-RISC targets are affected.
2933 This issue was reported to OpenSSL on 2nd March 2018 by Peter Waltenberg
2939 * Add a build target 'build_all_generated', to build all generated files
2940 and only that. This can be used to prepare everything that requires
2941 things like perl for a system that lacks perl and then move everything
2942 to that system and do the rest of the build there.
2946 * Backport SSL_OP_NO_RENGOTIATION
2948 OpenSSL 1.0.2 and below had the ability to disable renegotiation using the
2949 (undocumented) SSL3_FLAGS_NO_RENEGOTIATE_CIPHERS flag. Due to the opacity
2950 changes this is no longer possible in 1.1.0. Therefore the new
2951 SSL_OP_NO_RENEGOTIATION option from 1.1.1-dev has been backported to
2952 1.1.0 to provide equivalent functionality.
2954 Note that if an application built against 1.1.0h headers (or above) is run
2955 using an older version of 1.1.0 (prior to 1.1.0h) then the option will be
2956 accepted but nothing will happen, i.e. renegotiation will not be prevented.
2960 * Removed the OS390-Unix config target. It relied on a script that doesn't
2965 * rsaz_1024_mul_avx2 overflow bug on x86_64
2967 There is an overflow bug in the AVX2 Montgomery multiplication procedure
2968 used in exponentiation with 1024-bit moduli. No EC algorithms are affected.
2969 Analysis suggests that attacks against RSA and DSA as a result of this
2970 defect would be very difficult to perform and are not believed likely.
2971 Attacks against DH1024 are considered just feasible, because most of the
2972 work necessary to deduce information about a private key may be performed
2973 offline. The amount of resources required for such an attack would be
2974 significant. However, for an attack on TLS to be meaningful, the server
2975 would have to share the DH1024 private key among multiple clients, which is
2976 no longer an option since CVE-2016-0701.
2978 This only affects processors that support the AVX2 but not ADX extensions
2979 like Intel Haswell (4th generation).
2981 This issue was reported to OpenSSL by David Benjamin (Google). The issue
2982 was originally found via the OSS-Fuzz project.
2987 ### Changes between 1.1.0f and 1.1.0g [2 Nov 2017]
2989 * bn_sqrx8x_internal carry bug on x86_64
2991 There is a carry propagating bug in the x86_64 Montgomery squaring
2992 procedure. No EC algorithms are affected. Analysis suggests that attacks
2993 against RSA and DSA as a result of this defect would be very difficult to
2994 perform and are not believed likely. Attacks against DH are considered just
2995 feasible (although very difficult) because most of the work necessary to
2996 deduce information about a private key may be performed offline. The amount
2997 of resources required for such an attack would be very significant and
2998 likely only accessible to a limited number of attackers. An attacker would
2999 additionally need online access to an unpatched system using the target
3000 private key in a scenario with persistent DH parameters and a private
3001 key that is shared between multiple clients.
3003 This only affects processors that support the BMI1, BMI2 and ADX extensions
3004 like Intel Broadwell (5th generation) and later or AMD Ryzen.
3006 This issue was reported to OpenSSL by the OSS-Fuzz project.
3011 * Malformed X.509 IPAddressFamily could cause OOB read
3013 If an X.509 certificate has a malformed IPAddressFamily extension,
3014 OpenSSL could do a one-byte buffer overread. The most likely result
3015 would be an erroneous display of the certificate in text format.
3017 This issue was reported to OpenSSL by the OSS-Fuzz project.
3022 ### Changes between 1.1.0e and 1.1.0f [25 May 2017]
3024 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
3025 platform rather than 'mingw'.
3029 * Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
3030 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
3031 which is the minimum version we support.
3035 ### Changes between 1.1.0d and 1.1.0e [16 Feb 2017]
3037 * Encrypt-Then-Mac renegotiation crash
3039 During a renegotiation handshake if the Encrypt-Then-Mac extension is
3040 negotiated where it was not in the original handshake (or vice-versa) then
3041 this can cause OpenSSL to crash (dependant on ciphersuite). Both clients
3042 and servers are affected.
3044 This issue was reported to OpenSSL by Joe Orton (Red Hat).
3049 ### Changes between 1.1.0c and 1.1.0d [26 Jan 2017]
3051 * Truncated packet could crash via OOB read
3053 If one side of an SSL/TLS path is running on a 32-bit host and a specific
3054 cipher is being used, then a truncated packet can cause that host to
3055 perform an out-of-bounds read, usually resulting in a crash.
3057 This issue was reported to OpenSSL by Robert Święcki of Google.
3062 * Bad (EC)DHE parameters cause a client crash
3064 If a malicious server supplies bad parameters for a DHE or ECDHE key
3065 exchange then this can result in the client attempting to dereference a
3066 NULL pointer leading to a client crash. This could be exploited in a Denial
3069 This issue was reported to OpenSSL by Guido Vranken.
3074 * BN_mod_exp may produce incorrect results on x86_64
3076 There is a carry propagating bug in the x86_64 Montgomery squaring
3077 procedure. No EC algorithms are affected. Analysis suggests that attacks
3078 against RSA and DSA as a result of this defect would be very difficult to
3079 perform and are not believed likely. Attacks against DH are considered just
3080 feasible (although very difficult) because most of the work necessary to
3081 deduce information about a private key may be performed offline. The amount
3082 of resources required for such an attack would be very significant and
3083 likely only accessible to a limited number of attackers. An attacker would
3084 additionally need online access to an unpatched system using the target
3085 private key in a scenario with persistent DH parameters and a private
3086 key that is shared between multiple clients. For example this can occur by
3087 default in OpenSSL DHE based SSL/TLS ciphersuites. Note: This issue is very
3088 similar to CVE-2015-3193 but must be treated as a separate problem.
3090 This issue was reported to OpenSSL by the OSS-Fuzz project.
3095 ### Changes between 1.1.0b and 1.1.0c [10 Nov 2016]
3097 * ChaCha20/Poly1305 heap-buffer-overflow
3099 TLS connections using `*-CHACHA20-POLY1305` ciphersuites are susceptible to
3100 a DoS attack by corrupting larger payloads. This can result in an OpenSSL
3101 crash. This issue is not considered to be exploitable beyond a DoS.
3103 This issue was reported to OpenSSL by Robert Święcki (Google Security Team)
3108 * CMS Null dereference
3110 Applications parsing invalid CMS structures can crash with a NULL pointer
3111 dereference. This is caused by a bug in the handling of the ASN.1 CHOICE
3112 type in OpenSSL 1.1.0 which can result in a NULL value being passed to the
3113 structure callback if an attempt is made to free certain invalid encodings.
3114 Only CHOICE structures using a callback which do not handle NULL value are
3117 This issue was reported to OpenSSL by Tyler Nighswander of ForAllSecure.
3122 * Montgomery multiplication may produce incorrect results
3124 There is a carry propagating bug in the Broadwell-specific Montgomery
3125 multiplication procedure that handles input lengths divisible by, but
3126 longer than 256 bits. Analysis suggests that attacks against RSA, DSA
3127 and DH private keys are impossible. This is because the subroutine in
3128 question is not used in operations with the private key itself and an input
3129 of the attacker's direct choice. Otherwise the bug can manifest itself as
3130 transient authentication and key negotiation failures or reproducible
3131 erroneous outcome of public-key operations with specially crafted input.
3132 Among EC algorithms only Brainpool P-512 curves are affected and one
3133 presumably can attack ECDH key negotiation. Impact was not analyzed in
3134 detail, because pre-requisites for attack are considered unlikely. Namely
3135 multiple clients have to choose the curve in question and the server has to
3136 share the private key among them, neither of which is default behaviour.
3137 Even then only clients that chose the curve will be affected.
3139 This issue was publicly reported as transient failures and was not
3140 initially recognized as a security issue. Thanks to Richard Morgan for
3141 providing reproducible case.
3146 * Removed automatic addition of RPATH in shared libraries and executables,
3147 as this was a remainder from OpenSSL 1.0.x and isn't needed any more.
3151 ### Changes between 1.1.0a and 1.1.0b [26 Sep 2016]
3153 * Fix Use After Free for large message sizes
3155 The patch applied to address CVE-2016-6307 resulted in an issue where if a
3156 message larger than approx 16k is received then the underlying buffer to
3157 store the incoming message is reallocated and moved. Unfortunately a
3158 dangling pointer to the old location is left which results in an attempt to
3159 write to the previously freed location. This is likely to result in a
3160 crash, however it could potentially lead to execution of arbitrary code.
3162 This issue only affects OpenSSL 1.1.0a.
3164 This issue was reported to OpenSSL by Robert Święcki.
3169 ### Changes between 1.1.0 and 1.1.0a [22 Sep 2016]
3171 * OCSP Status Request extension unbounded memory growth
3173 A malicious client can send an excessively large OCSP Status Request
3174 extension. If that client continually requests renegotiation, sending a
3175 large OCSP Status Request extension each time, then there will be unbounded
3176 memory growth on the server. This will eventually lead to a Denial Of
3177 Service attack through memory exhaustion. Servers with a default
3178 configuration are vulnerable even if they do not support OCSP. Builds using
3179 the "no-ocsp" build time option are not affected.
3181 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
3186 * SSL_peek() hang on empty record
3188 OpenSSL 1.1.0 SSL/TLS will hang during a call to SSL_peek() if the peer
3189 sends an empty record. This could be exploited by a malicious peer in a
3190 Denial Of Service attack.
3192 This issue was reported to OpenSSL by Alex Gaynor.
3197 * Excessive allocation of memory in tls_get_message_header() and
3198 dtls1_preprocess_fragment()
3200 A (D)TLS message includes 3 bytes for its length in the header for the
3201 message. This would allow for messages up to 16Mb in length. Messages of
3202 this length are excessive and OpenSSL includes a check to ensure that a
3203 peer is sending reasonably sized messages in order to avoid too much memory
3204 being consumed to service a connection. A flaw in the logic of version
3205 1.1.0 means that memory for the message is allocated too early, prior to
3206 the excessive message length check. Due to way memory is allocated in
3207 OpenSSL this could mean an attacker could force up to 21Mb to be allocated
3208 to service a connection. This could lead to a Denial of Service through
3209 memory exhaustion. However, the excessive message length check still takes
3210 place, and this would cause the connection to immediately fail. Assuming
3211 that the application calls SSL_free() on the failed connection in a timely
3212 manner then the 21Mb of allocated memory will then be immediately freed
3213 again. Therefore the excessive memory allocation will be transitory in
3214 nature. This then means that there is only a security impact if:
3216 1) The application does not call SSL_free() in a timely manner in the event
3217 that the connection fails
3219 2) The application is working in a constrained environment where there is
3220 very little free memory
3222 3) The attacker initiates multiple connection attempts such that there are
3223 multiple connections in a state where memory has been allocated for the
3224 connection; SSL_free() has not yet been called; and there is insufficient
3225 memory to service the multiple requests.
3227 Except in the instance of (1) above any Denial Of Service is likely to be
3228 transitory because as soon as the connection fails the memory is
3229 subsequently freed again in the SSL_free() call. However there is an
3230 increased risk during this period of application crashes due to the lack of
3231 memory - which would then mean a more serious Denial of Service.
3233 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
3234 (CVE-2016-6307 and CVE-2016-6308)
3238 * solaris-x86-cc, i.e. 32-bit configuration with vendor compiler,
3239 had to be removed. Primary reason is that vendor assembler can't
3240 assemble our modules with -KPIC flag. As result it, assembly
3241 support, was not even available as option. But its lack means
3242 lack of side-channel resistant code, which is incompatible with
3243 security by todays standards. Fortunately gcc is readily available
3244 prepackaged option, which we firmly point at...
3248 ### Changes between 1.0.2h and 1.1.0 [25 Aug 2016]
3250 * Windows command-line tool supports UTF-8 opt-in option for arguments
3251 and console input. Setting OPENSSL_WIN32_UTF8 environment variable
3252 (to any value) allows Windows user to access PKCS#12 file generated
3253 with Windows CryptoAPI and protected with non-ASCII password, as well
3254 as files generated under UTF-8 locale on Linux also protected with
3259 * To mitigate the SWEET32 attack ([CVE-2016-2183]), 3DES cipher suites
3260 have been disabled by default and removed from DEFAULT, just like RC4.
3261 See the RC4 item below to re-enable both.
3265 * The method for finding the storage location for the Windows RAND seed file
3266 has changed. First we check %RANDFILE%. If that is not set then we check
3267 the directories %HOME%, %USERPROFILE% and %SYSTEMROOT% in that order. If
3268 all else fails we fall back to C:\.
3272 * The EVP_EncryptUpdate() function has had its return type changed from void
3273 to int. A return of 0 indicates and error while a return of 1 indicates
3278 * The flags RSA_FLAG_NO_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME and
3279 DH_FLAG_NO_EXP_CONSTTIME which previously provided the ability to switch
3280 off the constant time implementation for RSA, DSA and DH have been made
3281 no-ops and deprecated.
3285 * Windows RAND implementation was simplified to only get entropy by
3286 calling CryptGenRandom(). Various other RAND-related tickets
3289 *Joseph Wylie Yandle, Rich Salz*
3291 * The stack and lhash API's were renamed to start with `OPENSSL_SK_`
3292 and `OPENSSL_LH_`, respectively. The old names are available
3293 with API compatibility. They new names are now completely documented.
3297 * Unify TYPE_up_ref(obj) methods signature.
3298 SSL_CTX_up_ref(), SSL_up_ref(), X509_up_ref(), EVP_PKEY_up_ref(),
3299 X509_CRL_up_ref(), X509_OBJECT_up_ref_count() methods are now returning an
3300 int (instead of void) like all others TYPE_up_ref() methods.
3301 So now these methods also check the return value of CRYPTO_atomic_add(),
3302 and the validity of object reference counter.
3304 *fdasilvayy@gmail.com*
3306 * With Windows Visual Studio builds, the .pdb files are installed
3307 alongside the installed libraries and executables. For a static
3308 library installation, ossl_static.pdb is the associate compiler
3309 generated .pdb file to be used when linking programs.
3313 * Remove openssl.spec. Packaging files belong with the packagers.
3317 * Automatic Darwin/OSX configuration has had a refresh, it will now
3318 recognise x86_64 architectures automatically. You can still decide
3319 to build for a different bitness with the environment variable
3320 KERNEL_BITS (can be 32 or 64), for example:
3322 KERNEL_BITS=32 ./config
3326 * Change default algorithms in pkcs8 utility to use PKCS#5 v2.0,
3327 256 bit AES and HMAC with SHA256.
3331 * Remove support for MIPS o32 ABI on IRIX (and IRIX only).
3335 * Triple-DES ciphers have been moved from HIGH to MEDIUM.
3339 * To enable users to have their own config files and build file templates,
3340 Configure looks in the directory indicated by the environment variable
3341 OPENSSL_LOCAL_CONFIG_DIR as well as the in-source Configurations/
3342 directory. On VMS, OPENSSL_LOCAL_CONFIG_DIR is expected to be a logical
3343 name and is used as is.
3347 * The following datatypes were made opaque: X509_OBJECT, X509_STORE_CTX,
3348 X509_STORE, X509_LOOKUP, and X509_LOOKUP_METHOD. The unused type
3349 X509_CERT_FILE_CTX was removed.
3353 * "shared" builds are now the default. To create only static libraries use
3354 the "no-shared" Configure option.
3358 * Remove the no-aes, no-hmac, no-rsa, no-sha and no-md5 Configure options.
3359 All of these option have not worked for some while and are fundamental
3364 * Make various cleanup routines no-ops and mark them as deprecated. Most
3365 global cleanup functions are no longer required because they are handled
3366 via auto-deinit (see OPENSSL_init_crypto and OPENSSL_init_ssl man pages).
3367 Explicitly de-initing can cause problems (e.g. where a library that uses
3368 OpenSSL de-inits, but an application is still using it). The affected
3369 functions are CONF_modules_free(), ENGINE_cleanup(), OBJ_cleanup(),
3370 EVP_cleanup(), BIO_sock_cleanup(), CRYPTO_cleanup_all_ex_data(),
3371 RAND_cleanup(), SSL_COMP_free_compression_methods(), ERR_free_strings() and
3372 COMP_zlib_cleanup().
3376 * --strict-warnings no longer enables runtime debugging options
3377 such as REF_DEBUG. Instead, debug options are automatically
3378 enabled with '--debug' builds.
3380 *Andy Polyakov, Emilia Käsper*
3382 * Made DH and DH_METHOD opaque. The structures for managing DH objects
3383 have been moved out of the public header files. New functions for managing
3384 these have been added.
3388 * Made RSA and RSA_METHOD opaque. The structures for managing RSA
3389 objects have been moved out of the public header files. New
3390 functions for managing these have been added.
3394 * Made DSA and DSA_METHOD opaque. The structures for managing DSA objects
3395 have been moved out of the public header files. New functions for managing
3396 these have been added.
3400 * Made BIO and BIO_METHOD opaque. The structures for managing BIOs have been
3401 moved out of the public header files. New functions for managing these
3406 * Removed no-rijndael as a config option. Rijndael is an old name for AES.
3410 * Removed the mk1mf build scripts.
3414 * Headers are now wrapped, if necessary, with OPENSSL_NO_xxx, so
3415 it is always safe to #include a header now.
3419 * Removed the aged BC-32 config and all its supporting scripts
3423 * Removed support for Ultrix, Netware, and OS/2.
3427 * Add support for HKDF.
3429 *Alessandro Ghedini*
3431 * Add support for blake2b and blake2s
3435 * Added support for "pipelining". Ciphers that have the
3436 EVP_CIPH_FLAG_PIPELINE flag set have a capability to process multiple
3437 encryptions/decryptions simultaneously. There are currently no built-in
3438 ciphers with this property but the expectation is that engines will be able
3439 to offer it to significantly improve throughput. Support has been extended
3440 into libssl so that multiple records for a single connection can be
3441 processed in one go (for >=TLS 1.1).
3445 * Added the AFALG engine. This is an async capable engine which is able to
3446 offload work to the Linux kernel. In this initial version it only supports
3447 AES128-CBC. The kernel must be version 4.1.0 or greater.
3451 * OpenSSL now uses a new threading API. It is no longer necessary to
3452 set locking callbacks to use OpenSSL in a multi-threaded environment. There
3453 are two supported threading models: pthreads and windows threads. It is
3454 also possible to configure OpenSSL at compile time for "no-threads". The
3455 old threading API should no longer be used. The functions have been
3456 replaced with "no-op" compatibility macros.
3458 *Alessandro Ghedini, Matt Caswell*
3460 * Modify behavior of ALPN to invoke callback after SNI/servername
3461 callback, such that updates to the SSL_CTX affect ALPN.
3465 * Add SSL_CIPHER queries for authentication and key-exchange.
3469 * Changes to the DEFAULT cipherlist:
3470 - Prefer (EC)DHE handshakes over plain RSA.
3471 - Prefer AEAD ciphers over legacy ciphers.
3472 - Prefer ECDSA over RSA when both certificates are available.
3473 - Prefer TLSv1.2 ciphers/PRF.
3474 - Remove DSS, SEED, IDEA, CAMELLIA, and AES-CCM from the
3479 * Change the ECC default curve list to be this, in order: x25519,
3480 secp256r1, secp521r1, secp384r1.
3484 * RC4 based libssl ciphersuites are now classed as "weak" ciphers and are
3485 disabled by default. They can be re-enabled using the
3486 enable-weak-ssl-ciphers option to Configure.
3490 * If the server has ALPN configured, but supports no protocols that the
3491 client advertises, send a fatal "no_application_protocol" alert.
3492 This behaviour is SHALL in RFC 7301, though it isn't universally
3493 implemented by other servers.
3497 * Add X25519 support.
3498 Add ASN.1 and EVP_PKEY methods for X25519. This includes support
3499 for public and private key encoding using the format documented in
3500 draft-ietf-curdle-pkix-02. The corresponding EVP_PKEY method supports
3501 key generation and key derivation.
3503 TLS support complies with draft-ietf-tls-rfc4492bis-08 and uses
3508 * Deprecate SRP_VBASE_get_by_user.
3509 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
3510 In order to fix an unavoidable memory leak ([CVE-2016-0798]),
3511 SRP_VBASE_get_by_user was changed to ignore the "fake user" SRP
3512 seed, even if the seed is configured.
3514 Users should use SRP_VBASE_get1_by_user instead. Note that in
3515 SRP_VBASE_get1_by_user, caller must free the returned value. Note
3516 also that even though configuring the SRP seed attempts to hide
3517 invalid usernames by continuing the handshake with fake
3518 credentials, this behaviour is not constant time and no strong
3519 guarantees are made that the handshake is indistinguishable from
3520 that of a valid user.
3524 * Configuration change; it's now possible to build dynamic engines
3525 without having to build shared libraries and vice versa. This
3526 only applies to the engines in `engines/`, those in `crypto/engine/`
3527 will always be built into libcrypto (i.e. "static").
3529 Building dynamic engines is enabled by default; to disable, use
3530 the configuration option "disable-dynamic-engine".
3532 The only requirements for building dynamic engines are the
3533 presence of the DSO module and building with position independent
3534 code, so they will also automatically be disabled if configuring
3535 with "disable-dso" or "disable-pic".
3537 The macros OPENSSL_NO_STATIC_ENGINE and OPENSSL_NO_DYNAMIC_ENGINE
3538 are also taken away from openssl/opensslconf.h, as they are
3543 * Configuration change; if there is a known flag to compile
3544 position independent code, it will always be applied on the
3545 libcrypto and libssl object files, and never on the application
3546 object files. This means other libraries that use routines from
3547 libcrypto / libssl can be made into shared libraries regardless
3548 of how OpenSSL was configured.
3550 If this isn't desirable, the configuration options "disable-pic"
3551 or "no-pic" can be used to disable the use of PIC. This will
3552 also disable building shared libraries and dynamic engines.
3556 * Removed JPAKE code. It was experimental and has no wide use.
3560 * The INSTALL_PREFIX Makefile variable has been renamed to
3561 DESTDIR. That makes for less confusion on what this variable
3562 is for. Also, the configuration option --install_prefix is
3567 * Heartbeat for TLS has been removed and is disabled by default
3568 for DTLS; configure with enable-heartbeats. Code that uses the
3569 old #define's might need to be updated.
3571 *Emilia Käsper, Rich Salz*
3573 * Rename REF_CHECK to REF_DEBUG.
3577 * New "unified" build system
3579 The "unified" build system is aimed to be a common system for all
3580 platforms we support. With it comes new support for VMS.
3582 This system builds supports building in a different directory tree
3583 than the source tree. It produces one Makefile (for unix family
3584 or lookalikes), or one descrip.mms (for VMS).
3586 The source of information to make the Makefile / descrip.mms is
3587 small files called 'build.info', holding the necessary
3588 information for each directory with source to compile, and a
3589 template in Configurations, like unix-Makefile.tmpl or
3592 With this change, the library names were also renamed on Windows
3593 and on VMS. They now have names that are closer to the standard
3594 on Unix, and include the major version number, and in certain
3595 cases, the architecture they are built for. See "Notes on shared
3596 libraries" in INSTALL.
3598 We rely heavily on the perl module Text::Template.
3602 * Added support for auto-initialisation and de-initialisation of the library.
3603 OpenSSL no longer requires explicit init or deinit routines to be called,
3604 except in certain circumstances. See the OPENSSL_init_crypto() and
3605 OPENSSL_init_ssl() man pages for further information.
3609 * The arguments to the DTLSv1_listen function have changed. Specifically the
3610 "peer" argument is now expected to be a BIO_ADDR object.
3612 * Rewrite of BIO networking library. The BIO library lacked consistent
3613 support of IPv6, and adding it required some more extensive
3614 modifications. This introduces the BIO_ADDR and BIO_ADDRINFO types,
3615 which hold all types of addresses and chains of address information.
3616 It also introduces a new API, with functions like BIO_socket,
3617 BIO_connect, BIO_listen, BIO_lookup and a rewrite of BIO_accept.
3618 The source/sink BIOs BIO_s_connect, BIO_s_accept and BIO_s_datagram
3619 have been adapted accordingly.
3623 * RSA_padding_check_PKCS1_type_1 now accepts inputs with and without
3628 * CRIME protection: disable compression by default, even if OpenSSL is
3629 compiled with zlib enabled. Applications can still enable compression
3630 by calling SSL_CTX_clear_options(ctx, SSL_OP_NO_COMPRESSION), or by
3631 using the SSL_CONF library to configure compression.
3635 * The signature of the session callback configured with
3636 SSL_CTX_sess_set_get_cb was changed. The read-only input buffer
3637 was explicitly marked as `const unsigned char*` instead of
3642 * Always DPURIFY. Remove the use of uninitialized memory in the
3643 RNG, and other conditional uses of DPURIFY. This makes -DPURIFY a no-op.
3647 * Removed many obsolete configuration items, including
3648 DES_PTR, DES_RISC1, DES_RISC2, DES_INT
3649 MD2_CHAR, MD2_INT, MD2_LONG
3651 IDEA_SHORT, IDEA_LONG
3652 RC2_SHORT, RC2_LONG, RC4_LONG, RC4_CHUNK, RC4_INDEX
3654 *Rich Salz, with advice from Andy Polyakov*
3656 * Many BN internals have been moved to an internal header file.
3658 *Rich Salz with help from Andy Polyakov*
3660 * Configuration and writing out the results from it has changed.
3661 Files such as Makefile include/openssl/opensslconf.h and are now
3662 produced through general templates, such as Makefile.in and
3663 crypto/opensslconf.h.in and some help from the perl module
3666 Also, the center of configuration information is no longer
3667 Makefile. Instead, Configure produces a perl module in
3668 configdata.pm which holds most of the config data (in the hash
3669 table %config), the target data that comes from the target
3670 configuration in one of the `Configurations/*.conf` files (in
3675 * To clarify their intended purposes, the Configure options
3676 --prefix and --openssldir change their semantics, and become more
3677 straightforward and less interdependent.
3679 --prefix shall be used exclusively to give the location INSTALLTOP
3680 where programs, scripts, libraries, include files and manuals are
3681 going to be installed. The default is now /usr/local.
3683 --openssldir shall be used exclusively to give the default
3684 location OPENSSLDIR where certificates, private keys, CRLs are
3685 managed. This is also where the default openssl.cnf gets
3687 If the directory given with this option is a relative path, the
3688 values of both the --prefix value and the --openssldir value will
3689 be combined to become OPENSSLDIR.
3690 The default for --openssldir is INSTALLTOP/ssl.
3692 Anyone who uses --openssldir to specify where OpenSSL is to be
3693 installed MUST change to use --prefix instead.
3697 * The GOST engine was out of date and therefore it has been removed. An up
3698 to date GOST engine is now being maintained in an external repository.
3699 See: <https://wiki.openssl.org/index.php/Binaries>. Libssl still retains
3700 support for GOST ciphersuites (these are only activated if a GOST engine
3705 * EGD is no longer supported by default; use enable-egd when
3708 *Ben Kaduk and Rich Salz*
3710 * The distribution now has Makefile.in files, which are used to
3711 create Makefile's when Configure is run. *Configure must be run
3712 before trying to build now.*
3716 * The return value for SSL_CIPHER_description() for error conditions
3721 * Support for RFC6698/RFC7671 DANE TLSA peer authentication.
3723 Obtaining and performing DNSSEC validation of TLSA records is
3724 the application's responsibility. The application provides
3725 the TLSA records of its choice to OpenSSL, and these are then
3726 used to authenticate the peer.
3728 The TLSA records need not even come from DNS. They can, for
3729 example, be used to implement local end-entity certificate or
3730 trust-anchor "pinning", where the "pin" data takes the form
3731 of TLSA records, which can augment or replace verification
3732 based on the usual WebPKI public certification authorities.
3736 * Revert default OPENSSL_NO_DEPRECATED setting. Instead OpenSSL
3737 continues to support deprecated interfaces in default builds.
3738 However, applications are strongly advised to compile their
3739 source files with -DOPENSSL_API_COMPAT=0x10100000L, which hides
3740 the declarations of all interfaces deprecated in 0.9.8, 1.0.0
3741 or the 1.1.0 releases.
3743 In environments in which all applications have been ported to
3744 not use any deprecated interfaces OpenSSL's Configure script
3745 should be used with the --api=1.1.0 option to entirely remove
3746 support for the deprecated features from the library and
3747 unconditionally disable them in the installed headers.
3748 Essentially the same effect can be achieved with the "no-deprecated"
3749 argument to Configure, except that this will always restrict
3750 the build to just the latest API, rather than a fixed API
3753 As applications are ported to future revisions of the API,
3754 they should update their compile-time OPENSSL_API_COMPAT define
3755 accordingly, but in most cases should be able to continue to
3756 compile with later releases.
3758 The OPENSSL_API_COMPAT versions for 1.0.0, and 0.9.8 are
3759 0x10000000L and 0x00908000L, respectively. However those
3760 versions did not support the OPENSSL_API_COMPAT feature, and
3761 so applications are not typically tested for explicit support
3762 of just the undeprecated features of either release.
3766 * Add support for setting the minimum and maximum supported protocol.
3767 It can bet set via the SSL_set_min_proto_version() and
3768 SSL_set_max_proto_version(), or via the SSL_CONF's MinProtocol and
3769 MaxProtocol. It's recommended to use the new APIs to disable
3770 protocols instead of disabling individual protocols using
3771 SSL_set_options() or SSL_CONF's Protocol. This change also
3772 removes support for disabling TLS 1.2 in the OpenSSL TLS
3773 client at compile time by defining OPENSSL_NO_TLS1_2_CLIENT.
3777 * Support for ChaCha20 and Poly1305 added to libcrypto and libssl.
3781 * New EC_KEY_METHOD, this replaces the older ECDSA_METHOD and ECDH_METHOD
3782 and integrates ECDSA and ECDH functionality into EC. Implementations can
3783 now redirect key generation and no longer need to convert to or from
3786 Note: the ecdsa.h and ecdh.h headers are now no longer needed and just
3787 include the ec.h header file instead.
3791 * Remove support for all 40 and 56 bit ciphers. This includes all the export
3792 ciphers who are no longer supported and drops support the ephemeral RSA key
3793 exchange. The LOW ciphers currently doesn't have any ciphers in it.
3797 * Made EVP_MD_CTX, EVP_MD, EVP_CIPHER_CTX, EVP_CIPHER and HMAC_CTX
3798 opaque. For HMAC_CTX, the following constructors and destructors
3801 HMAC_CTX *HMAC_CTX_new(void);
3802 void HMAC_CTX_free(HMAC_CTX *ctx);
3804 For EVP_MD and EVP_CIPHER, complete APIs to create, fill and
3805 destroy such methods has been added. See EVP_MD_meth_new(3) and
3806 EVP_CIPHER_meth_new(3) for documentation.
3809 1) `EVP_MD_CTX_cleanup()`, `EVP_CIPHER_CTX_cleanup()` and
3810 `HMAC_CTX_cleanup()` were removed. `HMAC_CTX_reset()` and
3811 `EVP_MD_CTX_reset()` should be called instead to reinitialise
3812 an already created structure.
3813 2) For consistency with the majority of our object creators and
3814 destructors, `EVP_MD_CTX_(create|destroy)` were renamed to
3815 `EVP_MD_CTX_(new|free)`. The old names are retained as macros
3816 for deprecated builds.
3820 * Added ASYNC support. Libcrypto now includes the async sub-library to enable
3821 cryptographic operations to be performed asynchronously as long as an
3822 asynchronous capable engine is used. See the ASYNC_start_job() man page for
3823 further details. Libssl has also had this capability integrated with the
3824 introduction of the new mode SSL_MODE_ASYNC and associated error
3825 SSL_ERROR_WANT_ASYNC. See the SSL_CTX_set_mode() and SSL_get_error() man
3826 pages. This work was developed in partnership with Intel Corp.
3830 * SSL_{CTX_}set_ecdh_auto() has been removed and ECDH is support is
3831 always enabled now. If you want to disable the support you should
3832 exclude it using the list of supported ciphers. This also means that the
3833 "-no_ecdhe" option has been removed from s_server.
3837 * SSL_{CTX}_set_tmp_ecdh() which can set 1 EC curve now internally calls
3838 SSL_{CTX_}set1_curves() which can set a list.
3842 * Remove support for SSL_{CTX_}set_tmp_ecdh_callback(). You should set the
3843 curve you want to support using SSL_{CTX_}set1_curves().
3847 * State machine rewrite. The state machine code has been significantly
3848 refactored in order to remove much duplication of code and solve issues
3849 with the old code (see [ssl/statem/README.md](ssl/statem/README.md) for
3850 further details). This change does have some associated API changes.
3851 Notably the SSL_state() function has been removed and replaced by
3852 SSL_get_state which now returns an "OSSL_HANDSHAKE_STATE" instead of an int.
3853 SSL_set_state() has been removed altogether. The previous handshake states
3854 defined in ssl.h and ssl3.h have also been removed.
3858 * All instances of the string "ssleay" in the public API were replaced
3859 with OpenSSL (case-matching; e.g., OPENSSL_VERSION for #define's)
3860 Some error codes related to internal RSA_eay API's were renamed.
3864 * The demo files in crypto/threads were moved to demo/threads.
3868 * Removed obsolete engines: 4758cca, aep, atalla, cswift, nuron, gmp,
3871 *Matt Caswell, Rich Salz*
3873 * New ASN.1 embed macro.
3875 New ASN.1 macro ASN1_EMBED. This is the same as ASN1_SIMPLE except the
3876 structure is not allocated: it is part of the parent. That is instead of
3884 This reduces memory fragmentation and make it impossible to accidentally
3885 set a mandatory field to NULL.
3887 This currently only works for some fields specifically a SEQUENCE, CHOICE,
3888 or ASN1_STRING type which is part of a parent SEQUENCE. Since it is
3889 equivalent to ASN1_SIMPLE it cannot be tagged, OPTIONAL, SET OF or
3894 * Remove EVP_CHECK_DES_KEY, a compile-time option that never compiled.
3898 * Removed DES and RC4 ciphersuites from DEFAULT. Also removed RC2 although
3899 in 1.0.2 EXPORT was already removed and the only RC2 ciphersuite is also
3900 an EXPORT one. COMPLEMENTOFDEFAULT has been updated accordingly to add
3901 DES and RC4 ciphersuites.
3905 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
3906 This changes the decoding behaviour for some invalid messages,
3907 though the change is mostly in the more lenient direction, and
3908 legacy behaviour is preserved as much as possible.
3912 * Fix no-stdio build.
3913 *David Woodhouse <David.Woodhouse@intel.com> and also*
3914 *Ivan Nestlerode <ivan.nestlerode@sonos.com>*
3916 * New testing framework
3917 The testing framework has been largely rewritten and is now using
3918 perl and the perl modules Test::Harness and an extended variant of
3919 Test::More called OpenSSL::Test to do its work. All test scripts in
3920 test/ have been rewritten into test recipes, and all direct calls to
3921 executables in test/Makefile have become individual recipes using the
3922 simplified testing OpenSSL::Test::Simple.
3924 For documentation on our testing modules, do:
3926 perldoc test/testlib/OpenSSL/Test/Simple.pm
3927 perldoc test/testlib/OpenSSL/Test.pm
3931 * Revamped memory debug; only -DCRYPTO_MDEBUG and -DCRYPTO_MDEBUG_ABORT
3932 are used; the latter aborts on memory leaks (usually checked on exit).
3933 Some undocumented "set malloc, etc., hooks" functions were removed
3934 and others were changed. All are now documented.
3938 * In DSA_generate_parameters_ex, if the provided seed is too short,
3941 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
3943 * Rewrite PSK to support ECDHE_PSK, DHE_PSK and RSA_PSK. Add ciphersuites
3944 from RFC4279, RFC4785, RFC5487, RFC5489.
3946 Thanks to Christian J. Dietrich and Giuseppe D'Angelo for the
3947 original RSA_PSK patch.
3951 * Dropped support for the SSL3_FLAGS_DELAY_CLIENT_FINISHED flag. This SSLeay
3952 era flag was never set throughout the codebase (only read). Also removed
3953 SSL3_FLAGS_POP_BUFFER which was only used if
3954 SSL3_FLAGS_DELAY_CLIENT_FINISHED was also set.
3958 * Changed the default name options in the "ca", "crl", "req" and "x509"
3959 to be "oneline" instead of "compat".
3963 * Remove SSL_OP_TLS_BLOCK_PADDING_BUG. This is SSLeay legacy, we're
3964 not aware of clients that still exhibit this bug, and the workaround
3965 hasn't been working properly for a while.
3969 * The return type of BIO_number_read() and BIO_number_written() as well as
3970 the corresponding num_read and num_write members in the BIO structure has
3971 changed from unsigned long to uint64_t. On platforms where an unsigned
3972 long is 32 bits (e.g. Windows) these counters could overflow if >4Gb is
3977 * Given the pervasive nature of TLS extensions it is inadvisable to run
3978 OpenSSL without support for them. It also means that maintaining
3979 the OPENSSL_NO_TLSEXT option within the code is very invasive (and probably
3980 not well tested). Therefore the OPENSSL_NO_TLSEXT option has been removed.
3984 * Removed support for the two export grade static DH ciphersuites
3985 EXP-DH-RSA-DES-CBC-SHA and EXP-DH-DSS-DES-CBC-SHA. These two ciphersuites
3986 were newly added (along with a number of other static DH ciphersuites) to
3987 1.0.2. However the two export ones have *never* worked since they were
3988 introduced. It seems strange in any case to be adding new export
3989 ciphersuites, and given "logjam" it also does not seem correct to fix them.
3993 * Version negotiation has been rewritten. In particular SSLv23_method(),
3994 SSLv23_client_method() and SSLv23_server_method() have been deprecated,
3995 and turned into macros which simply call the new preferred function names
3996 TLS_method(), TLS_client_method() and TLS_server_method(). All new code
3997 should use the new names instead. Also as part of this change the ssl23.h
3998 header file has been removed.
4002 * Support for Kerberos ciphersuites in TLS (RFC2712) has been removed. This
4003 code and the associated standard is no longer considered fit-for-purpose.
4007 * RT2547 was closed. When generating a private key, try to make the
4008 output file readable only by the owner. This behavior change might
4009 be noticeable when interacting with other software.
4011 * Documented all exdata functions. Added CRYPTO_free_ex_index.
4016 * Added HTTP GET support to the ocsp command.
4020 * Changed default digest for the dgst and enc commands from MD5 to
4025 * RAND_pseudo_bytes has been deprecated. Users should use RAND_bytes instead.
4029 * Added support for TLS extended master secret from
4030 draft-ietf-tls-session-hash-03.txt. Thanks for Alfredo Pironti for an
4031 initial patch which was a great help during development.
4035 * All libssl internal structures have been removed from the public header
4036 files, and the OPENSSL_NO_SSL_INTERN option has been removed (since it is
4037 now redundant). Users should not attempt to access internal structures
4038 directly. Instead they should use the provided API functions.
4042 * config has been changed so that by default OPENSSL_NO_DEPRECATED is used.
4043 Access to deprecated functions can be re-enabled by running config with
4044 "enable-deprecated". In addition applications wishing to use deprecated
4045 functions must define OPENSSL_USE_DEPRECATED. Note that this new behaviour
4046 will, by default, disable some transitive includes that previously existed
4047 in the header files (e.g. ec.h will no longer, by default, include bn.h)
4051 * Added support for OCB mode. OpenSSL has been granted a patent license
4052 compatible with the OpenSSL license for use of OCB. Details are available
4053 at <https://www.openssl.org/source/OCB-patent-grant-OpenSSL.pdf>. Support
4054 for OCB can be removed by calling config with no-ocb.
4058 * SSLv2 support has been removed. It still supports receiving a SSLv2
4059 compatible client hello.
4063 * Increased the minimal RSA keysize from 256 to 512 bits [Rich Salz],
4064 done while fixing the error code for the key-too-small case.
4066 *Annie Yousar <a.yousar@informatik.hu-berlin.de>*
4068 * CA.sh has been removed; use CA.pl instead.
4072 * Removed old DES API.
4076 * Remove various unsupported platforms:
4082 Sinix/ReliantUNIX RM400
4087 16-bit platforms such as WIN16
4091 * Clean up OPENSSL_NO_xxx #define's
4092 - Use setbuf() and remove OPENSSL_NO_SETVBUF_IONBF
4093 - Rename OPENSSL_SYSNAME_xxx to OPENSSL_SYS_xxx
4094 - OPENSSL_NO_EC{DH,DSA} merged into OPENSSL_NO_EC
4095 - OPENSSL_NO_RIPEMD160, OPENSSL_NO_RIPEMD merged into OPENSSL_NO_RMD160
4096 - OPENSSL_NO_FP_API merged into OPENSSL_NO_STDIO
4097 - Remove OPENSSL_NO_BIO OPENSSL_NO_BUFFER OPENSSL_NO_CHAIN_VERIFY
4098 OPENSSL_NO_EVP OPENSSL_NO_FIPS_ERR OPENSSL_NO_HASH_COMP
4099 OPENSSL_NO_LHASH OPENSSL_NO_OBJECT OPENSSL_NO_SPEED OPENSSL_NO_STACK
4100 OPENSSL_NO_X509 OPENSSL_NO_X509_VERIFY
4101 - Remove MS_STATIC; it's a relic from platforms <32 bits.
4105 * Cleaned up dead code
4106 Remove all but one '#ifdef undef' which is to be looked at.
4110 * Clean up calling of xxx_free routines.
4111 Just like free(), fix most of the xxx_free routines to accept
4112 NULL. Remove the non-null checks from callers. Save much code.
4116 * Add secure heap for storage of private keys (when possible).
4117 Add BIO_s_secmem(), CBIGNUM, etc.
4118 Contributed by Akamai Technologies under our Corporate CLA.
4122 * Experimental support for a new, fast, unbiased prime candidate generator,
4123 bn_probable_prime_dh_coprime(). Not currently used by any prime generator.
4125 *Felix Laurie von Massenbach <felix@erbridge.co.uk>*
4127 * New output format NSS in the sess_id command line tool. This allows
4128 exporting the session id and the master key in NSS keylog format.
4130 *Martin Kaiser <martin@kaiser.cx>*
4132 * Harmonize version and its documentation. -f flag is used to display
4135 *mancha <mancha1@zoho.com>*
4137 * Fix eckey_priv_encode so it immediately returns an error upon a failure
4138 in i2d_ECPrivateKey. Thanks to Ted Unangst for feedback on this issue.
4140 *mancha <mancha1@zoho.com>*
4142 * Fix some double frees. These are not thought to be exploitable.
4144 *mancha <mancha1@zoho.com>*
4146 * A missing bounds check in the handling of the TLS heartbeat extension
4147 can be used to reveal up to 64k of memory to a connected client or
4150 Thanks for Neel Mehta of Google Security for discovering this bug and to
4151 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
4152 preparing the fix ([CVE-2014-0160])
4154 *Adam Langley, Bodo Moeller*
4156 * Fix for the attack described in the paper "Recovering OpenSSL
4157 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
4158 by Yuval Yarom and Naomi Benger. Details can be obtained from:
4159 <http://eprint.iacr.org/2014/140>
4161 Thanks to Yuval Yarom and Naomi Benger for discovering this
4162 flaw and to Yuval Yarom for supplying a fix ([CVE-2014-0076])
4164 *Yuval Yarom and Naomi Benger*
4166 * Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
4167 this fixes a limitation in previous versions of OpenSSL.
4171 * Experimental encrypt-then-mac support.
4173 Experimental support for encrypt then mac from
4174 draft-gutmann-tls-encrypt-then-mac-02.txt
4176 To enable it set the appropriate extension number (0x42 for the test
4177 server) using e.g. -DTLSEXT_TYPE_encrypt_then_mac=0x42
4179 For non-compliant peers (i.e. just about everything) this should have no
4182 WARNING: EXPERIMENTAL, SUBJECT TO CHANGE.
4186 * Add EVP support for key wrapping algorithms, to avoid problems with
4187 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
4188 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
4189 algorithms and include tests cases.
4193 * Extend CMS code to support RSA-PSS signatures and RSA-OAEP for
4198 * Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
4199 MGF1 digest and OAEP label.
4203 * Make openssl verify return errors.
4205 *Chris Palmer <palmer@google.com> and Ben Laurie*
4207 * New function ASN1_TIME_diff to calculate the difference between two
4208 ASN1_TIME structures or one structure and the current time.
4212 * Update fips_test_suite to support multiple command line options. New
4213 test to induce all self test errors in sequence and check expected
4218 * Add FIPS_{rsa,dsa,ecdsa}_{sign,verify} functions which digest and
4219 sign or verify all in one operation.
4223 * Add fips_algvs: a multicall fips utility incorporating all the algorithm
4224 test programs and fips_test_suite. Includes functionality to parse
4225 the minimal script output of fipsalgest.pl directly.
4229 * Add authorisation parameter to FIPS_module_mode_set().
4233 * Add FIPS selftest for ECDH algorithm using P-224 and B-233 curves.
4237 * Use separate DRBG fields for internal and external flags. New function
4238 FIPS_drbg_health_check() to perform on demand health checking. Add
4239 generation tests to fips_test_suite with reduced health check interval to
4240 demonstrate periodic health checking. Add "nodh" option to
4241 fips_test_suite to skip very slow DH test.
4245 * New function FIPS_get_cipherbynid() to lookup FIPS supported ciphers
4250 * More extensive health check for DRBG checking many more failure modes.
4251 New function FIPS_selftest_drbg_all() to handle every possible DRBG
4252 combination: call this in fips_test_suite.
4256 * Add support for canonical generation of DSA parameter 'g'. See
4259 * Add support for HMAC DRBG from SP800-90. Update DRBG algorithm test and
4260 POST to handle HMAC cases.
4264 * Add functions FIPS_module_version() and FIPS_module_version_text()
4265 to return numerical and string versions of the FIPS module number.
4269 * Rename FIPS_mode_set and FIPS_mode to FIPS_module_mode_set and
4270 FIPS_module_mode. FIPS_mode and FIPS_mode_set will be implemented
4271 outside the validated module in the FIPS capable OpenSSL.
4275 * Minor change to DRBG entropy callback semantics. In some cases
4276 there is no multiple of the block length between min_len and
4277 max_len. Allow the callback to return more than max_len bytes
4278 of entropy but discard any extra: it is the callback's responsibility
4279 to ensure that the extra data discarded does not impact the
4280 requested amount of entropy.
4284 * Add PRNG security strength checks to RSA, DSA and ECDSA using
4285 information in FIPS186-3, SP800-57 and SP800-131A.
4289 * CCM support via EVP. Interface is very similar to GCM case except we
4290 must supply all data in one chunk (i.e. no update, final) and the
4291 message length must be supplied if AAD is used. Add algorithm test
4296 * Initial version of POST overhaul. Add POST callback to allow the status
4297 of POST to be monitored and/or failures induced. Modify fips_test_suite
4298 to use callback. Always run all selftests even if one fails.
4302 * XTS support including algorithm test driver in the fips_gcmtest program.
4303 Note: this does increase the maximum key length from 32 to 64 bytes but
4304 there should be no binary compatibility issues as existing applications
4305 will never use XTS mode.
4309 * Extensive reorganisation of FIPS PRNG behaviour. Remove all dependencies
4310 to OpenSSL RAND code and replace with a tiny FIPS RAND API which also
4311 performs algorithm blocking for unapproved PRNG types. Also do not
4312 set PRNG type in FIPS_mode_set(): leave this to the application.
4313 Add default OpenSSL DRBG handling: sets up FIPS PRNG and seeds with
4314 the standard OpenSSL PRNG: set additional data to a date time vector.
4318 * Rename old X9.31 PRNG functions of the form `FIPS_rand*` to `FIPS_x931*`.
4319 This shouldn't present any incompatibility problems because applications
4320 shouldn't be using these directly and any that are will need to rethink
4321 anyway as the X9.31 PRNG is now deprecated by FIPS 140-2
4325 * Extensive self tests and health checking required by SP800-90 DRBG.
4326 Remove strength parameter from FIPS_drbg_instantiate and always
4327 instantiate at maximum supported strength.
4331 * Add ECDH code to fips module and fips_ecdhvs for primitives only testing.
4335 * New algorithm test program fips_dhvs to handle DH primitives only testing.
4339 * New function DH_compute_key_padded() to compute a DH key and pad with
4340 leading zeroes if needed: this complies with SP800-56A et al.
4344 * Initial implementation of SP800-90 DRBGs for Hash and CTR. Not used by
4345 anything, incomplete, subject to change and largely untested at present.
4349 * Modify fipscanisteronly build option to only build the necessary object
4350 files by filtering FIPS_EX_OBJ through a perl script in crypto/Makefile.
4354 * Add experimental option FIPSSYMS to give all symbols in
4355 fipscanister.o and FIPS or fips prefix. This will avoid
4356 conflicts with future versions of OpenSSL. Add perl script
4357 util/fipsas.pl to preprocess assembly language source files
4358 and rename any affected symbols.
4362 * Add selftest checks and algorithm block of non-fips algorithms in
4363 FIPS mode. Remove DES2 from selftests.
4367 * Add ECDSA code to fips module. Add tiny fips_ecdsa_check to just
4368 return internal method without any ENGINE dependencies. Add new
4369 tiny fips sign and verify functions.
4373 * New build option no-ec2m to disable characteristic 2 code.
4377 * New build option "fipscanisteronly". This only builds fipscanister.o
4378 and (currently) associated fips utilities. Uses the file Makefile.fips
4379 instead of Makefile.org as the prototype.
4383 * Add some FIPS mode restrictions to GCM. Add internal IV generator.
4384 Update fips_gcmtest to use IV generator.
4388 * Initial, experimental EVP support for AES-GCM. AAD can be input by
4389 setting output buffer to NULL. The `*Final` function must be
4390 called although it will not retrieve any additional data. The tag
4391 can be set or retrieved with a ctrl. The IV length is by default 12
4392 bytes (96 bits) but can be set to an alternative value. If the IV
4393 length exceeds the maximum IV length (currently 16 bytes) it cannot be
4398 * New flag in ciphers: EVP_CIPH_FLAG_CUSTOM_CIPHER. This means the
4399 underlying do_cipher function handles all cipher semantics itself
4400 including padding and finalisation. This is useful if (for example)
4401 an ENGINE cipher handles block padding itself. The behaviour of
4402 do_cipher is subtly changed if this flag is set: the return value
4403 is the number of characters written to the output buffer (zero is
4404 no longer an error code) or a negative error code. Also if the
4405 input buffer is NULL and length 0 finalisation should be performed.
4409 * If a candidate issuer certificate is already part of the constructed
4410 path ignore it: new debug notification X509_V_ERR_PATH_LOOP for this case.
4414 * Improve forward-security support: add functions
4416 void SSL_CTX_set_not_resumable_session_callback(
4417 SSL_CTX *ctx, int (*cb)(SSL *ssl, int is_forward_secure))
4418 void SSL_set_not_resumable_session_callback(
4419 SSL *ssl, int (*cb)(SSL *ssl, int is_forward_secure))
4421 for use by SSL/TLS servers; the callback function will be called whenever a
4422 new session is created, and gets to decide whether the session may be
4423 cached to make it resumable (return 0) or not (return 1). (As by the
4424 SSL/TLS protocol specifications, the session_id sent by the server will be
4425 empty to indicate that the session is not resumable; also, the server will
4426 not generate RFC 4507 (RFC 5077) session tickets.)
4428 A simple reasonable callback implementation is to return is_forward_secure.
4429 This parameter will be set to 1 or 0 depending on the ciphersuite selected
4430 by the SSL/TLS server library, indicating whether it can provide forward
4433 *Emilia Käsper <emilia.kasper@esat.kuleuven.be> (Google)*
4435 * New -verify_name option in command line utilities to set verification
4440 * Initial CMAC implementation. WARNING: EXPERIMENTAL, API MAY CHANGE.
4441 Add CMAC pkey methods.
4445 * Experimental renegotiation in s_server -www mode. If the client
4446 browses /reneg connection is renegotiated. If /renegcert it is
4447 renegotiated requesting a certificate.
4451 * Add an "external" session cache for debugging purposes to s_server. This
4452 should help trace issues which normally are only apparent in deployed
4453 multi-process servers.
4457 * Extensive audit of libcrypto with DEBUG_UNUSED. Fix many cases where
4458 return value is ignored. NB. The functions RAND_add(), RAND_seed(),
4459 BIO_set_cipher() and some obscure PEM functions were changed so they
4460 can now return an error. The RAND changes required a change to the
4461 RAND_METHOD structure.
4465 * New macro `__owur` for "OpenSSL Warn Unused Result". This makes use of
4466 a gcc attribute to warn if the result of a function is ignored. This
4467 is enable if DEBUG_UNUSED is set. Add to several functions in evp.h
4468 whose return value is often ignored.
4472 * New -noct, -requestct, -requirect and -ctlogfile options for s_client.
4473 These allow SCTs (signed certificate timestamps) to be requested and
4474 validated when establishing a connection.
4476 *Rob Percival <robpercival@google.com>*
4481 ### Changes between 1.0.2s and 1.0.2t [10 Sep 2019]
4483 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
4484 used even when parsing explicit parameters, when loading a encoded key
4485 or calling `EC_GROUP_new_from_ecpkparameters()`/
4486 `EC_GROUP_new_from_ecparameters()`.
4487 This prevents bypass of security hardening and performance gains,
4488 especially for curves with specialized EC_METHODs.
4489 By default, if a key encoded with explicit parameters is loaded and later
4490 encoded, the output is still encoded with explicit parameters, even if
4491 internally a "named" EC_GROUP is used for computation.
4495 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
4496 this change, EC_GROUP_set_generator would accept order and/or cofactor as
4497 NULL. After this change, only the cofactor parameter can be NULL. It also
4498 does some minimal sanity checks on the passed order.
4503 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
4504 An attack is simple, if the first CMS_recipientInfo is valid but the
4505 second CMS_recipientInfo is chosen ciphertext. If the second
4506 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
4507 encryption key will be replaced by garbage, and the message cannot be
4508 decoded, but if the RSA decryption fails, the correct encryption key is
4509 used and the recipient will not notice the attack.
4510 As a work around for this potential attack the length of the decrypted
4511 key must be equal to the cipher default key length, in case the
4512 certifiate is not given and all recipientInfo are tried out.
4513 The old behaviour can be re-enabled in the CMS code by setting the
4514 CMS_DEBUG_DECRYPT flag.
4519 * Document issue with installation paths in diverse Windows builds
4521 '/usr/local/ssl' is an unsafe prefix for location to install OpenSSL
4522 binaries and run-time config file.
4527 ### Changes between 1.0.2r and 1.0.2s [28 May 2019]
4529 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
4530 This changes the size when using the `genpkey` command when no size is given.
4531 It fixes an omission in earlier changes that changed all RSA, DSA and DH
4532 generation commands to use 2048 bits by default.
4536 * Add FIPS support for Android Arm 64-bit
4538 Support for Android Arm 64-bit was added to the OpenSSL FIPS Object
4539 Module in Version 2.0.10. For some reason, the corresponding target
4540 'android64-aarch64' was missing OpenSSL 1.0.2, whence it could not be
4541 built with FIPS support on Android Arm 64-bit. This omission has been
4544 *Matthias St. Pierre*
4546 ### Changes between 1.0.2q and 1.0.2r [26 Feb 2019]
4548 * 0-byte record padding oracle
4550 If an application encounters a fatal protocol error and then calls
4551 SSL_shutdown() twice (once to send a close_notify, and once to receive one)
4552 then OpenSSL can respond differently to the calling application if a 0 byte
4553 record is received with invalid padding compared to if a 0 byte record is
4554 received with an invalid MAC. If the application then behaves differently
4555 based on that in a way that is detectable to the remote peer, then this
4556 amounts to a padding oracle that could be used to decrypt data.
4558 In order for this to be exploitable "non-stitched" ciphersuites must be in
4559 use. Stitched ciphersuites are optimised implementations of certain
4560 commonly used ciphersuites. Also the application must call SSL_shutdown()
4561 twice even if a protocol error has occurred (applications should not do
4562 this but some do anyway).
4564 This issue was discovered by Juraj Somorovsky, Robert Merget and Nimrod
4565 Aviram, with additional investigation by Steven Collison and Andrew
4566 Hourselt. It was reported to OpenSSL on 10th December 2018.
4571 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
4575 ### Changes between 1.0.2p and 1.0.2q [20 Nov 2018]
4577 * Microarchitecture timing vulnerability in ECC scalar multiplication
4579 OpenSSL ECC scalar multiplication, used in e.g. ECDSA and ECDH, has been
4580 shown to be vulnerable to a microarchitecture timing side channel attack.
4581 An attacker with sufficient access to mount local timing attacks during
4582 ECDSA signature generation could recover the private key.
4584 This issue was reported to OpenSSL on 26th October 2018 by Alejandro
4585 Cabrera Aldaya, Billy Brumley, Sohaib ul Hassan, Cesar Pereida Garcia and
4591 * Timing vulnerability in DSA signature generation
4593 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
4594 timing side channel attack. An attacker could use variations in the signing
4595 algorithm to recover the private key.
4597 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
4602 * Resolve a compatibility issue in EC_GROUP handling with the FIPS Object
4603 Module, accidentally introduced while backporting security fixes from the
4604 development branch and hindering the use of ECC in FIPS mode.
4608 ### Changes between 1.0.2o and 1.0.2p [14 Aug 2018]
4610 * Client DoS due to large DH parameter
4612 During key agreement in a TLS handshake using a DH(E) based ciphersuite a
4613 malicious server can send a very large prime value to the client. This will
4614 cause the client to spend an unreasonably long period of time generating a
4615 key for this prime resulting in a hang until the client has finished. This
4616 could be exploited in a Denial Of Service attack.
4618 This issue was reported to OpenSSL on 5th June 2018 by Guido Vranken
4623 * Cache timing vulnerability in RSA Key Generation
4625 The OpenSSL RSA Key generation algorithm has been shown to be vulnerable to
4626 a cache timing side channel attack. An attacker with sufficient access to
4627 mount cache timing attacks during the RSA key generation process could
4628 recover the private key.
4630 This issue was reported to OpenSSL on 4th April 2018 by Alejandro Cabrera
4631 Aldaya, Billy Brumley, Cesar Pereida Garcia and Luis Manuel Alvarez Tapia.
4636 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
4637 parameter is no longer accepted, as it leads to a corrupt table. NULL
4638 pem_str is reserved for alias entries only.
4642 * Revert blinding in ECDSA sign and instead make problematic addition
4643 length-invariant. Switch even to fixed-length Montgomery multiplication.
4647 * Change generating and checking of primes so that the error rate of not
4648 being prime depends on the intended use based on the size of the input.
4649 For larger primes this will result in more rounds of Miller-Rabin.
4650 The maximal error rate for primes with more than 1080 bits is lowered
4653 *Kurt Roeckx, Annie Yousar*
4655 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
4659 * Add blinding to ECDSA and DSA signatures to protect against side channel
4660 attacks discovered by Keegan Ryan (NCC Group).
4664 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
4665 now allow empty (zero character) pass phrases.
4669 * Certificate time validation (X509_cmp_time) enforces stricter
4670 compliance with RFC 5280. Fractional seconds and timezone offsets
4671 are no longer allowed.
4675 ### Changes between 1.0.2n and 1.0.2o [27 Mar 2018]
4677 * Constructed ASN.1 types with a recursive definition could exceed the stack
4679 Constructed ASN.1 types with a recursive definition (such as can be found
4680 in PKCS7) could eventually exceed the stack given malicious input with
4681 excessive recursion. This could result in a Denial Of Service attack. There
4682 are no such structures used within SSL/TLS that come from untrusted sources
4683 so this is considered safe.
4685 This issue was reported to OpenSSL on 4th January 2018 by the OSS-fuzz
4691 ### Changes between 1.0.2m and 1.0.2n [7 Dec 2017]
4693 * Read/write after SSL object in error state
4695 OpenSSL 1.0.2 (starting from version 1.0.2b) introduced an "error state"
4696 mechanism. The intent was that if a fatal error occurred during a handshake
4697 then OpenSSL would move into the error state and would immediately fail if
4698 you attempted to continue the handshake. This works as designed for the
4699 explicit handshake functions (SSL_do_handshake(), SSL_accept() and
4700 SSL_connect()), however due to a bug it does not work correctly if
4701 SSL_read() or SSL_write() is called directly. In that scenario, if the
4702 handshake fails then a fatal error will be returned in the initial function
4703 call. If SSL_read()/SSL_write() is subsequently called by the application
4704 for the same SSL object then it will succeed and the data is passed without
4705 being decrypted/encrypted directly from the SSL/TLS record layer.
4707 In order to exploit this issue an application bug would have to be present
4708 that resulted in a call to SSL_read()/SSL_write() being issued after having
4709 already received a fatal error.
4711 This issue was reported to OpenSSL by David Benjamin (Google).
4716 * rsaz_1024_mul_avx2 overflow bug on x86_64
4718 There is an overflow bug in the AVX2 Montgomery multiplication procedure
4719 used in exponentiation with 1024-bit moduli. No EC algorithms are affected.
4720 Analysis suggests that attacks against RSA and DSA as a result of this
4721 defect would be very difficult to perform and are not believed likely.
4722 Attacks against DH1024 are considered just feasible, because most of the
4723 work necessary to deduce information about a private key may be performed
4724 offline. The amount of resources required for such an attack would be
4725 significant. However, for an attack on TLS to be meaningful, the server
4726 would have to share the DH1024 private key among multiple clients, which is
4727 no longer an option since CVE-2016-0701.
4729 This only affects processors that support the AVX2 but not ADX extensions
4730 like Intel Haswell (4th generation).
4732 This issue was reported to OpenSSL by David Benjamin (Google). The issue
4733 was originally found via the OSS-Fuzz project.
4738 ### Changes between 1.0.2l and 1.0.2m [2 Nov 2017]
4740 * bn_sqrx8x_internal carry bug on x86_64
4742 There is a carry propagating bug in the x86_64 Montgomery squaring
4743 procedure. No EC algorithms are affected. Analysis suggests that attacks
4744 against RSA and DSA as a result of this defect would be very difficult to
4745 perform and are not believed likely. Attacks against DH are considered just
4746 feasible (although very difficult) because most of the work necessary to
4747 deduce information about a private key may be performed offline. The amount
4748 of resources required for such an attack would be very significant and
4749 likely only accessible to a limited number of attackers. An attacker would
4750 additionally need online access to an unpatched system using the target
4751 private key in a scenario with persistent DH parameters and a private
4752 key that is shared between multiple clients.
4754 This only affects processors that support the BMI1, BMI2 and ADX extensions
4755 like Intel Broadwell (5th generation) and later or AMD Ryzen.
4757 This issue was reported to OpenSSL by the OSS-Fuzz project.
4762 * Malformed X.509 IPAddressFamily could cause OOB read
4764 If an X.509 certificate has a malformed IPAddressFamily extension,
4765 OpenSSL could do a one-byte buffer overread. The most likely result
4766 would be an erroneous display of the certificate in text format.
4768 This issue was reported to OpenSSL by the OSS-Fuzz project.
4772 ### Changes between 1.0.2k and 1.0.2l [25 May 2017]
4774 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
4775 platform rather than 'mingw'.
4779 ### Changes between 1.0.2j and 1.0.2k [26 Jan 2017]
4781 * Truncated packet could crash via OOB read
4783 If one side of an SSL/TLS path is running on a 32-bit host and a specific
4784 cipher is being used, then a truncated packet can cause that host to
4785 perform an out-of-bounds read, usually resulting in a crash.
4787 This issue was reported to OpenSSL by Robert Święcki of Google.
4792 * BN_mod_exp may produce incorrect results on x86_64
4794 There is a carry propagating bug in the x86_64 Montgomery squaring
4795 procedure. No EC algorithms are affected. Analysis suggests that attacks
4796 against RSA and DSA as a result of this defect would be very difficult to
4797 perform and are not believed likely. Attacks against DH are considered just
4798 feasible (although very difficult) because most of the work necessary to
4799 deduce information about a private key may be performed offline. The amount
4800 of resources required for such an attack would be very significant and
4801 likely only accessible to a limited number of attackers. An attacker would
4802 additionally need online access to an unpatched system using the target
4803 private key in a scenario with persistent DH parameters and a private
4804 key that is shared between multiple clients. For example this can occur by
4805 default in OpenSSL DHE based SSL/TLS ciphersuites. Note: This issue is very
4806 similar to CVE-2015-3193 but must be treated as a separate problem.
4808 This issue was reported to OpenSSL by the OSS-Fuzz project.
4813 * Montgomery multiplication may produce incorrect results
4815 There is a carry propagating bug in the Broadwell-specific Montgomery
4816 multiplication procedure that handles input lengths divisible by, but
4817 longer than 256 bits. Analysis suggests that attacks against RSA, DSA
4818 and DH private keys are impossible. This is because the subroutine in
4819 question is not used in operations with the private key itself and an input
4820 of the attacker's direct choice. Otherwise the bug can manifest itself as
4821 transient authentication and key negotiation failures or reproducible
4822 erroneous outcome of public-key operations with specially crafted input.
4823 Among EC algorithms only Brainpool P-512 curves are affected and one
4824 presumably can attack ECDH key negotiation. Impact was not analyzed in
4825 detail, because pre-requisites for attack are considered unlikely. Namely
4826 multiple clients have to choose the curve in question and the server has to
4827 share the private key among them, neither of which is default behaviour.
4828 Even then only clients that chose the curve will be affected.
4830 This issue was publicly reported as transient failures and was not
4831 initially recognized as a security issue. Thanks to Richard Morgan for
4832 providing reproducible case.
4837 * OpenSSL now fails if it receives an unrecognised record type in TLS1.0
4838 or TLS1.1. Previously this only happened in SSLv3 and TLS1.2. This is to
4839 prevent issues where no progress is being made and the peer continually
4840 sends unrecognised record types, using up resources processing them.
4844 ### Changes between 1.0.2i and 1.0.2j [26 Sep 2016]
4846 * Missing CRL sanity check
4848 A bug fix which included a CRL sanity check was added to OpenSSL 1.1.0
4849 but was omitted from OpenSSL 1.0.2i. As a result any attempt to use
4850 CRLs in OpenSSL 1.0.2i will crash with a null pointer exception.
4852 This issue only affects the OpenSSL 1.0.2i
4857 ### Changes between 1.0.2h and 1.0.2i [22 Sep 2016]
4859 * OCSP Status Request extension unbounded memory growth
4861 A malicious client can send an excessively large OCSP Status Request
4862 extension. If that client continually requests renegotiation, sending a
4863 large OCSP Status Request extension each time, then there will be unbounded
4864 memory growth on the server. This will eventually lead to a Denial Of
4865 Service attack through memory exhaustion. Servers with a default
4866 configuration are vulnerable even if they do not support OCSP. Builds using
4867 the "no-ocsp" build time option are not affected.
4869 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4874 * In order to mitigate the SWEET32 attack, the DES ciphers were moved from
4877 This issue was reported to OpenSSL Karthikeyan Bhargavan and Gaetan
4883 * OOB write in MDC2_Update()
4885 An overflow can occur in MDC2_Update() either if called directly or
4886 through the EVP_DigestUpdate() function using MDC2. If an attacker
4887 is able to supply very large amounts of input data after a previous
4888 call to EVP_EncryptUpdate() with a partial block then a length check
4889 can overflow resulting in a heap corruption.
4891 The amount of data needed is comparable to SIZE_MAX which is impractical
4894 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4899 * Malformed SHA512 ticket DoS
4901 If a server uses SHA512 for TLS session ticket HMAC it is vulnerable to a
4902 DoS attack where a malformed ticket will result in an OOB read which will
4905 The use of SHA512 in TLS session tickets is comparatively rare as it requires
4906 a custom server callback and ticket lookup mechanism.
4908 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4913 * OOB write in BN_bn2dec()
4915 The function BN_bn2dec() does not check the return value of BN_div_word().
4916 This can cause an OOB write if an application uses this function with an
4917 overly large BIGNUM. This could be a problem if an overly large certificate
4918 or CRL is printed out from an untrusted source. TLS is not affected because
4919 record limits will reject an oversized certificate before it is parsed.
4921 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4926 * OOB read in TS_OBJ_print_bio()
4928 The function TS_OBJ_print_bio() misuses OBJ_obj2txt(): the return value is
4929 the total length the OID text representation would use and not the amount
4930 of data written. This will result in OOB reads when large OIDs are
4933 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4938 * Pointer arithmetic undefined behaviour
4940 Avoid some undefined pointer arithmetic
4942 A common idiom in the codebase is to check limits in the following manner:
4945 Where "p" points to some malloc'd data of SIZE bytes and
4948 "len" here could be from some externally supplied data (e.g. from a TLS
4951 The rules of C pointer arithmetic are such that "p + len" is only well
4952 defined where len <= SIZE. Therefore the above idiom is actually
4953 undefined behaviour.
4955 For example this could cause problems if some malloc implementation
4956 provides an address for "p" such that "p + len" actually overflows for
4957 values of len that are too big and therefore p + len < limit.
4959 This issue was reported to OpenSSL by Guido Vranken
4964 * Constant time flag not preserved in DSA signing
4966 Operations in the DSA signing algorithm should run in constant time in
4967 order to avoid side channel attacks. A flaw in the OpenSSL DSA
4968 implementation means that a non-constant time codepath is followed for
4969 certain operations. This has been demonstrated through a cache-timing
4970 attack to be sufficient for an attacker to recover the private DSA key.
4972 This issue was reported by César Pereida (Aalto University), Billy Brumley
4973 (Tampere University of Technology), and Yuval Yarom (The University of
4974 Adelaide and NICTA).
4979 * DTLS buffered message DoS
4981 In a DTLS connection where handshake messages are delivered out-of-order
4982 those messages that OpenSSL is not yet ready to process will be buffered
4983 for later use. Under certain circumstances, a flaw in the logic means that
4984 those messages do not get removed from the buffer even though the handshake
4985 has been completed. An attacker could force up to approx. 15 messages to
4986 remain in the buffer when they are no longer required. These messages will
4987 be cleared when the DTLS connection is closed. The default maximum size for
4988 a message is 100k. Therefore the attacker could force an additional 1500k
4989 to be consumed per connection. By opening many simulataneous connections an
4990 attacker could cause a DoS attack through memory exhaustion.
4992 This issue was reported to OpenSSL by Quan Luo.
4997 * DTLS replay protection DoS
4999 A flaw in the DTLS replay attack protection mechanism means that records
5000 that arrive for future epochs update the replay protection "window" before
5001 the MAC for the record has been validated. This could be exploited by an
5002 attacker by sending a record for the next epoch (which does not have to
5003 decrypt or have a valid MAC), with a very large sequence number. This means
5004 that all subsequent legitimate packets are dropped causing a denial of
5005 service for a specific DTLS connection.
5007 This issue was reported to OpenSSL by the OCAP audit team.
5012 * Certificate message OOB reads
5014 In OpenSSL 1.0.2 and earlier some missing message length checks can result
5015 in OOB reads of up to 2 bytes beyond an allocated buffer. There is a
5016 theoretical DoS risk but this has not been observed in practice on common
5019 The messages affected are client certificate, client certificate request
5020 and server certificate. As a result the attack can only be performed
5021 against a client or a server which enables client authentication.
5023 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5028 ### Changes between 1.0.2g and 1.0.2h [3 May 2016]
5030 * Prevent padding oracle in AES-NI CBC MAC check
5032 A MITM attacker can use a padding oracle attack to decrypt traffic
5033 when the connection uses an AES CBC cipher and the server support
5036 This issue was introduced as part of the fix for Lucky 13 padding
5037 attack ([CVE-2013-0169]). The padding check was rewritten to be in
5038 constant time by making sure that always the same bytes are read and
5039 compared against either the MAC or padding bytes. But it no longer
5040 checked that there was enough data to have both the MAC and padding
5043 This issue was reported by Juraj Somorovsky using TLS-Attacker.
5047 * Fix EVP_EncodeUpdate overflow
5049 An overflow can occur in the EVP_EncodeUpdate() function which is used for
5050 Base64 encoding of binary data. If an attacker is able to supply very large
5051 amounts of input data then a length check can overflow resulting in a heap
5054 Internally to OpenSSL the EVP_EncodeUpdate() function is primarily used by
5055 the `PEM_write_bio*` family of functions. These are mainly used within the
5056 OpenSSL command line applications, so any application which processes data
5057 from an untrusted source and outputs it as a PEM file should be considered
5058 vulnerable to this issue. User applications that call these APIs directly
5059 with large amounts of untrusted data may also be vulnerable.
5061 This issue was reported by Guido Vranken.
5066 * Fix EVP_EncryptUpdate overflow
5068 An overflow can occur in the EVP_EncryptUpdate() function. If an attacker
5069 is able to supply very large amounts of input data after a previous call to
5070 EVP_EncryptUpdate() with a partial block then a length check can overflow
5071 resulting in a heap corruption. Following an analysis of all OpenSSL
5072 internal usage of the EVP_EncryptUpdate() function all usage is one of two
5073 forms. The first form is where the EVP_EncryptUpdate() call is known to be
5074 the first called function after an EVP_EncryptInit(), and therefore that
5075 specific call must be safe. The second form is where the length passed to
5076 EVP_EncryptUpdate() can be seen from the code to be some small value and
5077 therefore there is no possibility of an overflow. Since all instances are
5078 one of these two forms, it is believed that there can be no overflows in
5079 internal code due to this problem. It should be noted that
5080 EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths.
5081 Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances
5082 of these calls have also been analysed too and it is believed there are no
5083 instances in internal usage where an overflow could occur.
5085 This issue was reported by Guido Vranken.
5090 * Prevent ASN.1 BIO excessive memory allocation
5092 When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio()
5093 a short invalid encoding can cause allocation of large amounts of memory
5094 potentially consuming excessive resources or exhausting memory.
5096 Any application parsing untrusted data through d2i BIO functions is
5097 affected. The memory based functions such as d2i_X509() are *not* affected.
5098 Since the memory based functions are used by the TLS library, TLS
5099 applications are not affected.
5101 This issue was reported by Brian Carpenter.
5108 ASN1 Strings that are over 1024 bytes can cause an overread in applications
5109 using the X509_NAME_oneline() function on EBCDIC systems. This could result
5110 in arbitrary stack data being returned in the buffer.
5112 This issue was reported by Guido Vranken.
5117 * Modify behavior of ALPN to invoke callback after SNI/servername
5118 callback, such that updates to the SSL_CTX affect ALPN.
5122 * Remove LOW from the DEFAULT cipher list. This removes singles DES from the
5127 * Only remove the SSLv2 methods with the no-ssl2-method option. When the
5128 methods are enabled and ssl2 is disabled the methods return NULL.
5132 ### Changes between 1.0.2f and 1.0.2g [1 Mar 2016]
5134 * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL.
5135 Builds that are not configured with "enable-weak-ssl-ciphers" will not
5136 provide any "EXPORT" or "LOW" strength ciphers.
5140 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2
5141 is by default disabled at build-time. Builds that are not configured with
5142 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used,
5143 users who want to negotiate SSLv2 via the version-flexible SSLv23_method()
5144 will need to explicitly call either of:
5146 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2);
5148 SSL_clear_options(ssl, SSL_OP_NO_SSLv2);
5150 as appropriate. Even if either of those is used, or the application
5151 explicitly uses the version-specific SSLv2_method() or its client and
5152 server variants, SSLv2 ciphers vulnerable to exhaustive search key
5153 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT
5154 ciphers, and SSLv2 56-bit DES are no longer available.
5159 * Fix a double-free in DSA code
5161 A double free bug was discovered when OpenSSL parses malformed DSA private
5162 keys and could lead to a DoS attack or memory corruption for applications
5163 that receive DSA private keys from untrusted sources. This scenario is
5166 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using
5172 * Disable SRP fake user seed to address a server memory leak.
5174 Add a new method SRP_VBASE_get1_by_user that handles the seed properly.
5176 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
5177 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user
5178 was changed to ignore the "fake user" SRP seed, even if the seed
5181 Users should use SRP_VBASE_get1_by_user instead. Note that in
5182 SRP_VBASE_get1_by_user, caller must free the returned value. Note
5183 also that even though configuring the SRP seed attempts to hide
5184 invalid usernames by continuing the handshake with fake
5185 credentials, this behaviour is not constant time and no strong
5186 guarantees are made that the handshake is indistinguishable from
5187 that of a valid user.
5192 * Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption
5194 In the BN_hex2bn function the number of hex digits is calculated using an
5195 int value `i`. Later `bn_expand` is called with a value of `i * 4`. For
5196 large values of `i` this can result in `bn_expand` not allocating any
5197 memory because `i * 4` is negative. This can leave the internal BIGNUM data
5198 field as NULL leading to a subsequent NULL ptr deref. For very large values
5199 of `i`, the calculation `i * 4` could be a positive value smaller than `i`.
5200 In this case memory is allocated to the internal BIGNUM data field, but it
5201 is insufficiently sized leading to heap corruption. A similar issue exists
5202 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn
5203 is ever called by user applications with very large untrusted hex/dec data.
5204 This is anticipated to be a rare occurrence.
5206 All OpenSSL internal usage of these functions use data that is not expected
5207 to be untrusted, e.g. config file data or application command line
5208 arguments. If user developed applications generate config file data based
5209 on untrusted data then it is possible that this could also lead to security
5210 consequences. This is also anticipated to be rare.
5212 This issue was reported to OpenSSL by Guido Vranken.
5217 * Fix memory issues in `BIO_*printf` functions
5219 The internal `fmtstr` function used in processing a "%s" format string in
5220 the `BIO_*printf` functions could overflow while calculating the length of a
5221 string and cause an OOB read when printing very long strings.
5223 Additionally the internal `doapr_outch` function can attempt to write to an
5224 OOB memory location (at an offset from the NULL pointer) in the event of a
5225 memory allocation failure. In 1.0.2 and below this could be caused where
5226 the size of a buffer to be allocated is greater than INT_MAX. E.g. this
5227 could be in processing a very long "%s" format string. Memory leaks can
5230 The first issue may mask the second issue dependent on compiler behaviour.
5231 These problems could enable attacks where large amounts of untrusted data
5232 is passed to the `BIO_*printf` functions. If applications use these functions
5233 in this way then they could be vulnerable. OpenSSL itself uses these
5234 functions when printing out human-readable dumps of ASN.1 data. Therefore
5235 applications that print this data could be vulnerable if the data is from
5236 untrusted sources. OpenSSL command line applications could also be
5237 vulnerable where they print out ASN.1 data, or if untrusted data is passed
5238 as command line arguments.
5240 Libssl is not considered directly vulnerable. Additionally certificates etc
5241 received via remote connections via libssl are also unlikely to be able to
5242 trigger these issues because of message size limits enforced within libssl.
5244 This issue was reported to OpenSSL Guido Vranken.
5249 * Side channel attack on modular exponentiation
5251 A side-channel attack was found which makes use of cache-bank conflicts on
5252 the Intel Sandy-Bridge microarchitecture which could lead to the recovery
5253 of RSA keys. The ability to exploit this issue is limited as it relies on
5254 an attacker who has control of code in a thread running on the same
5255 hyper-threaded core as the victim thread which is performing decryptions.
5257 This issue was reported to OpenSSL by Yuval Yarom, The University of
5258 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and
5259 Nadia Heninger, University of Pennsylvania with more information at
5260 <http://cachebleed.info>.
5265 * Change the `req` command to generate a 2048-bit RSA/DSA key by default,
5266 if no keysize is specified with default_bits. This fixes an
5267 omission in an earlier change that changed all RSA/DSA key generation
5268 commands to use 2048 bits by default.
5272 ### Changes between 1.0.2e and 1.0.2f [28 Jan 2016]
5274 * DH small subgroups
5276 Historically OpenSSL only ever generated DH parameters based on "safe"
5277 primes. More recently (in version 1.0.2) support was provided for
5278 generating X9.42 style parameter files such as those required for RFC 5114
5279 support. The primes used in such files may not be "safe". Where an
5280 application is using DH configured with parameters based on primes that are
5281 not "safe" then an attacker could use this fact to find a peer's private
5282 DH exponent. This attack requires that the attacker complete multiple
5283 handshakes in which the peer uses the same private DH exponent. For example
5284 this could be used to discover a TLS server's private DH exponent if it's
5285 reusing the private DH exponent or it's using a static DH ciphersuite.
5287 OpenSSL provides the option SSL_OP_SINGLE_DH_USE for ephemeral DH (DHE) in
5288 TLS. It is not on by default. If the option is not set then the server
5289 reuses the same private DH exponent for the life of the server process and
5290 would be vulnerable to this attack. It is believed that many popular
5291 applications do set this option and would therefore not be at risk.
5293 The fix for this issue adds an additional check where a "q" parameter is
5294 available (as is the case in X9.42 based parameters). This detects the
5295 only known attack, and is the only possible defense for static DH
5296 ciphersuites. This could have some performance impact.
5298 Additionally the SSL_OP_SINGLE_DH_USE option has been switched on by
5299 default and cannot be disabled. This could have some performance impact.
5301 This issue was reported to OpenSSL by Antonio Sanso (Adobe).
5306 * SSLv2 doesn't block disabled ciphers
5308 A malicious client can negotiate SSLv2 ciphers that have been disabled on
5309 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
5310 been disabled, provided that the SSLv2 protocol was not also disabled via
5313 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
5314 and Sebastian Schinzel.
5319 ### Changes between 1.0.2d and 1.0.2e [3 Dec 2015]
5321 * BN_mod_exp may produce incorrect results on x86_64
5323 There is a carry propagating bug in the x86_64 Montgomery squaring
5324 procedure. No EC algorithms are affected. Analysis suggests that attacks
5325 against RSA and DSA as a result of this defect would be very difficult to
5326 perform and are not believed likely. Attacks against DH are considered just
5327 feasible (although very difficult) because most of the work necessary to
5328 deduce information about a private key may be performed offline. The amount
5329 of resources required for such an attack would be very significant and
5330 likely only accessible to a limited number of attackers. An attacker would
5331 additionally need online access to an unpatched system using the target
5332 private key in a scenario with persistent DH parameters and a private
5333 key that is shared between multiple clients. For example this can occur by
5334 default in OpenSSL DHE based SSL/TLS ciphersuites.
5336 This issue was reported to OpenSSL by Hanno Böck.
5341 * Certificate verify crash with missing PSS parameter
5343 The signature verification routines will crash with a NULL pointer
5344 dereference if presented with an ASN.1 signature using the RSA PSS
5345 algorithm and absent mask generation function parameter. Since these
5346 routines are used to verify certificate signature algorithms this can be
5347 used to crash any certificate verification operation and exploited in a
5348 DoS attack. Any application which performs certificate verification is
5349 vulnerable including OpenSSL clients and servers which enable client
5352 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
5357 * X509_ATTRIBUTE memory leak
5359 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
5360 memory. This structure is used by the PKCS#7 and CMS routines so any
5361 application which reads PKCS#7 or CMS data from untrusted sources is
5362 affected. SSL/TLS is not affected.
5364 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
5370 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
5371 This changes the decoding behaviour for some invalid messages,
5372 though the change is mostly in the more lenient direction, and
5373 legacy behaviour is preserved as much as possible.
5377 * In DSA_generate_parameters_ex, if the provided seed is too short,
5380 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
5382 ### Changes between 1.0.2c and 1.0.2d [9 Jul 2015]
5384 * Alternate chains certificate forgery
5386 During certificate verification, OpenSSL will attempt to find an
5387 alternative certificate chain if the first attempt to build such a chain
5388 fails. An error in the implementation of this logic can mean that an
5389 attacker could cause certain checks on untrusted certificates to be
5390 bypassed, such as the CA flag, enabling them to use a valid leaf
5391 certificate to act as a CA and "issue" an invalid certificate.
5393 This issue was reported to OpenSSL by Adam Langley/David Benjamin
5398 ### Changes between 1.0.2b and 1.0.2c [12 Jun 2015]
5400 * Fix HMAC ABI incompatibility. The previous version introduced an ABI
5401 incompatibility in the handling of HMAC. The previous ABI has now been
5406 ### Changes between 1.0.2a and 1.0.2b [11 Jun 2015]
5408 * Malformed ECParameters causes infinite loop
5410 When processing an ECParameters structure OpenSSL enters an infinite loop
5411 if the curve specified is over a specially malformed binary polynomial
5414 This can be used to perform denial of service against any
5415 system which processes public keys, certificate requests or
5416 certificates. This includes TLS clients and TLS servers with
5417 client authentication enabled.
5419 This issue was reported to OpenSSL by Joseph Barr-Pixton.
5424 * Exploitable out-of-bounds read in X509_cmp_time
5426 X509_cmp_time does not properly check the length of the ASN1_TIME
5427 string and can read a few bytes out of bounds. In addition,
5428 X509_cmp_time accepts an arbitrary number of fractional seconds in the
5431 An attacker can use this to craft malformed certificates and CRLs of
5432 various sizes and potentially cause a segmentation fault, resulting in
5433 a DoS on applications that verify certificates or CRLs. TLS clients
5434 that verify CRLs are affected. TLS clients and servers with client
5435 authentication enabled may be affected if they use custom verification
5438 This issue was reported to OpenSSL by Robert Swiecki (Google), and
5439 independently by Hanno Böck.
5444 * PKCS7 crash with missing EnvelopedContent
5446 The PKCS#7 parsing code does not handle missing inner EncryptedContent
5447 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
5448 with missing content and trigger a NULL pointer dereference on parsing.
5450 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
5451 structures from untrusted sources are affected. OpenSSL clients and
5452 servers are not affected.
5454 This issue was reported to OpenSSL by Michal Zalewski (Google).
5459 * CMS verify infinite loop with unknown hash function
5461 When verifying a signedData message the CMS code can enter an infinite loop
5462 if presented with an unknown hash function OID. This can be used to perform
5463 denial of service against any system which verifies signedData messages using
5465 This issue was reported to OpenSSL by Johannes Bauer.
5470 * Race condition handling NewSessionTicket
5472 If a NewSessionTicket is received by a multi-threaded client when attempting to
5473 reuse a previous ticket then a race condition can occur potentially leading to
5474 a double free of the ticket data.
5479 * Only support 256-bit or stronger elliptic curves with the
5480 'ecdh_auto' setting (server) or by default (client). Of supported
5481 curves, prefer P-256 (both).
5485 ### Changes between 1.0.2 and 1.0.2a [19 Mar 2015]
5487 * ClientHello sigalgs DoS fix
5489 If a client connects to an OpenSSL 1.0.2 server and renegotiates with an
5490 invalid signature algorithms extension a NULL pointer dereference will
5491 occur. This can be exploited in a DoS attack against the server.
5493 This issue was was reported to OpenSSL by David Ramos of Stanford
5497 *Stephen Henson and Matt Caswell*
5499 * Multiblock corrupted pointer fix
5501 OpenSSL 1.0.2 introduced the "multiblock" performance improvement. This
5502 feature only applies on 64 bit x86 architecture platforms that support AES
5503 NI instructions. A defect in the implementation of "multiblock" can cause
5504 OpenSSL's internal write buffer to become incorrectly set to NULL when
5505 using non-blocking IO. Typically, when the user application is using a
5506 socket BIO for writing, this will only result in a failed connection.
5507 However if some other BIO is used then it is likely that a segmentation
5508 fault will be triggered, thus enabling a potential DoS attack.
5510 This issue was reported to OpenSSL by Daniel Danner and Rainer Mueller.
5515 * Segmentation fault in DTLSv1_listen fix
5517 The DTLSv1_listen function is intended to be stateless and processes the
5518 initial ClientHello from many peers. It is common for user code to loop
5519 over the call to DTLSv1_listen until a valid ClientHello is received with
5520 an associated cookie. A defect in the implementation of DTLSv1_listen means
5521 that state is preserved in the SSL object from one invocation to the next
5522 that can lead to a segmentation fault. Errors processing the initial
5523 ClientHello can trigger this scenario. An example of such an error could be
5524 that a DTLS1.0 only client is attempting to connect to a DTLS1.2 only
5527 This issue was reported to OpenSSL by Per Allansson.
5532 * Segmentation fault in ASN1_TYPE_cmp fix
5534 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
5535 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
5536 certificate signature algorithm consistency this can be used to crash any
5537 certificate verification operation and exploited in a DoS attack. Any
5538 application which performs certificate verification is vulnerable including
5539 OpenSSL clients and servers which enable client authentication.
5544 * Segmentation fault for invalid PSS parameters fix
5546 The signature verification routines will crash with a NULL pointer
5547 dereference if presented with an ASN.1 signature using the RSA PSS
5548 algorithm and invalid parameters. Since these routines are used to verify
5549 certificate signature algorithms this can be used to crash any
5550 certificate verification operation and exploited in a DoS attack. Any
5551 application which performs certificate verification is vulnerable including
5552 OpenSSL clients and servers which enable client authentication.
5554 This issue was was reported to OpenSSL by Brian Carpenter.
5559 * ASN.1 structure reuse memory corruption fix
5561 Reusing a structure in ASN.1 parsing may allow an attacker to cause
5562 memory corruption via an invalid write. Such reuse is and has been
5563 strongly discouraged and is believed to be rare.
5565 Applications that parse structures containing CHOICE or ANY DEFINED BY
5566 components may be affected. Certificate parsing (d2i_X509 and related
5567 functions) are however not affected. OpenSSL clients and servers are
5573 * PKCS7 NULL pointer dereferences fix
5575 The PKCS#7 parsing code does not handle missing outer ContentInfo
5576 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
5577 missing content and trigger a NULL pointer dereference on parsing.
5579 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
5580 otherwise parse PKCS#7 structures from untrusted sources are
5581 affected. OpenSSL clients and servers are not affected.
5583 This issue was reported to OpenSSL by Michal Zalewski (Google).
5588 * DoS via reachable assert in SSLv2 servers fix
5590 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
5591 servers that both support SSLv2 and enable export cipher suites by sending
5592 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
5594 This issue was discovered by Sean Burford (Google) and Emilia Käsper
5595 (OpenSSL development team).
5600 * Empty CKE with client auth and DHE fix
5602 If client auth is used then a server can seg fault in the event of a DHE
5603 ciphersuite being selected and a zero length ClientKeyExchange message
5604 being sent by the client. This could be exploited in a DoS attack.
5609 * Handshake with unseeded PRNG fix
5611 Under certain conditions an OpenSSL 1.0.2 client can complete a handshake
5612 with an unseeded PRNG. The conditions are:
5613 - The client is on a platform where the PRNG has not been seeded
5614 automatically, and the user has not seeded manually
5615 - A protocol specific client method version has been used (i.e. not
5616 SSL_client_methodv23)
5617 - A ciphersuite is used that does not require additional random data from
5618 the PRNG beyond the initial ClientHello client random (e.g. PSK-RC4-SHA).
5620 If the handshake succeeds then the client random that has been used will
5621 have been generated from a PRNG with insufficient entropy and therefore the
5622 output may be predictable.
5624 For example using the following command with an unseeded openssl will
5625 succeed on an unpatched platform:
5627 openssl s_client -psk 1a2b3c4d -tls1_2 -cipher PSK-RC4-SHA
5632 * Use After Free following d2i_ECPrivatekey error fix
5634 A malformed EC private key file consumed via the d2i_ECPrivateKey function
5635 could cause a use after free condition. This, in turn, could cause a double
5636 free in several private key parsing functions (such as d2i_PrivateKey
5637 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
5638 for applications that receive EC private keys from untrusted
5639 sources. This scenario is considered rare.
5641 This issue was discovered by the BoringSSL project and fixed in their
5647 * X509_to_X509_REQ NULL pointer deref fix
5649 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
5650 the certificate key is invalid. This function is rarely used in practice.
5652 This issue was discovered by Brian Carpenter.
5657 * Removed the export ciphers from the DEFAULT ciphers
5661 ### Changes between 1.0.1l and 1.0.2 [22 Jan 2015]
5663 * Facilitate "universal" ARM builds targeting range of ARM ISAs, e.g.
5664 ARMv5 through ARMv8, as opposite to "locking" it to single one.
5665 So far those who have to target multiple platforms would compromise
5666 and argue that binary targeting say ARMv5 would still execute on
5667 ARMv8. "Universal" build resolves this compromise by providing
5668 near-optimal performance even on newer platforms.
5672 * Accelerated NIST P-256 elliptic curve implementation for x86_64
5673 (other platforms pending).
5675 *Shay Gueron & Vlad Krasnov (Intel Corp), Andy Polyakov*
5677 * Add support for the SignedCertificateTimestampList certificate and
5678 OCSP response extensions from RFC6962.
5682 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
5683 for corner cases. (Certain input points at infinity could lead to
5684 bogus results, with non-infinity inputs mapped to infinity too.)
5688 * Initial support for PowerISA 2.0.7, first implemented in POWER8.
5689 This covers AES, SHA256/512 and GHASH. "Initial" means that most
5690 common cases are optimized and there still is room for further
5691 improvements. Vector Permutation AES for Altivec is also added.
5695 * Add support for little-endian ppc64 Linux target.
5697 *Marcelo Cerri (IBM)*
5699 * Initial support for AMRv8 ISA crypto extensions. This covers AES,
5700 SHA1, SHA256 and GHASH. "Initial" means that most common cases
5701 are optimized and there still is room for further improvements.
5702 Both 32- and 64-bit modes are supported.
5704 *Andy Polyakov, Ard Biesheuvel (Linaro)*
5706 * Improved ARMv7 NEON support.
5710 * Support for SPARC Architecture 2011 crypto extensions, first
5711 implemented in SPARC T4. This covers AES, DES, Camellia, SHA1,
5712 SHA256/512, MD5, GHASH and modular exponentiation.
5714 *Andy Polyakov, David Miller*
5716 * Accelerated modular exponentiation for Intel processors, a.k.a.
5719 *Shay Gueron & Vlad Krasnov (Intel Corp)*
5721 * Support for new and upcoming Intel processors, including AVX2,
5722 BMI and SHA ISA extensions. This includes additional "stitched"
5723 implementations, AESNI-SHA256 and GCM, and multi-buffer support
5726 This work was sponsored by Intel Corp.
5730 * Support for DTLS 1.2. This adds two sets of DTLS methods: DTLS_*_method()
5731 supports both DTLS 1.2 and 1.0 and should use whatever version the peer
5732 supports and DTLSv1_2_*_method() which supports DTLS 1.2 only.
5736 * Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
5737 this fixes a limitation in previous versions of OpenSSL.
5741 * Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
5742 MGF1 digest and OAEP label.
5746 * Add EVP support for key wrapping algorithms, to avoid problems with
5747 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
5748 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
5749 algorithms and include tests cases.
5753 * Add functions to allocate and set the fields of an ECDSA_METHOD
5756 *Douglas E. Engert, Steve Henson*
5758 * New functions OPENSSL_gmtime_diff and ASN1_TIME_diff to find the
5759 difference in days and seconds between two tm or ASN1_TIME structures.
5763 * Add -rev test option to s_server to just reverse order of characters
5764 received by client and send back to server. Also prints an abbreviated
5765 summary of the connection parameters.
5769 * New option -brief for s_client and s_server to print out a brief summary
5770 of connection parameters.
5774 * Add callbacks for arbitrary TLS extensions.
5776 *Trevor Perrin <trevp@trevp.net> and Ben Laurie*
5778 * New option -crl_download in several openssl utilities to download CRLs
5779 from CRLDP extension in certificates.
5783 * New options -CRL and -CRLform for s_client and s_server for CRLs.
5787 * New function X509_CRL_diff to generate a delta CRL from the difference
5788 of two full CRLs. Add support to "crl" utility.
5792 * New functions to set lookup_crls function and to retrieve
5793 X509_STORE from X509_STORE_CTX.
5797 * Print out deprecated issuer and subject unique ID fields in
5802 * Extend OCSP I/O functions so they can be used for simple general purpose
5803 HTTP as well as OCSP. New wrapper function which can be used to download
5804 CRLs using the OCSP API.
5808 * Delegate command line handling in s_client/s_server to SSL_CONF APIs.
5812 * `SSL_CONF*` functions. These provide a common framework for application
5813 configuration using configuration files or command lines.
5817 * SSL/TLS tracing code. This parses out SSL/TLS records using the
5818 message callback and prints the results. Needs compile time option
5819 "enable-ssl-trace". New options to s_client and s_server to enable
5824 * New ctrl and macro to retrieve supported points extensions.
5825 Print out extension in s_server and s_client.
5829 * New functions to retrieve certificate signature and signature
5834 * Add functions to retrieve and manipulate the raw cipherlist sent by a
5839 * New Suite B modes for TLS code. These use and enforce the requirements
5840 of RFC6460: restrict ciphersuites, only permit Suite B algorithms and
5841 only use Suite B curves. The Suite B modes can be set by using the
5842 strings "SUITEB128", "SUITEB192" or "SUITEB128ONLY" for the cipherstring.
5846 * New chain verification flags for Suite B levels of security. Check
5847 algorithms are acceptable when flags are set in X509_verify_cert.
5851 * Make tls1_check_chain return a set of flags indicating checks passed
5852 by a certificate chain. Add additional tests to handle client
5853 certificates: checks for matching certificate type and issuer name
5858 * If an attempt is made to use a signature algorithm not in the peer
5859 preference list abort the handshake. If client has no suitable
5860 signature algorithms in response to a certificate request do not
5861 use the certificate.
5865 * If server EC tmp key is not in client preference list abort handshake.
5869 * Add support for certificate stores in CERT structure. This makes it
5870 possible to have different stores per SSL structure or one store in
5871 the parent SSL_CTX. Include distinct stores for certificate chain
5872 verification and chain building. New ctrl SSL_CTRL_BUILD_CERT_CHAIN
5873 to build and store a certificate chain in CERT structure: returning
5874 an error if the chain cannot be built: this will allow applications
5875 to test if a chain is correctly configured.
5877 Note: if the CERT based stores are not set then the parent SSL_CTX
5878 store is used to retain compatibility with existing behaviour.
5882 * New function ssl_set_client_disabled to set a ciphersuite disabled
5883 mask based on the current session, check mask when sending client
5884 hello and checking the requested ciphersuite.
5888 * New ctrls to retrieve and set certificate types in a certificate
5889 request message. Print out received values in s_client. If certificate
5890 types is not set with custom values set sensible values based on
5891 supported signature algorithms.
5895 * Support for distinct client and server supported signature algorithms.
5899 * Add certificate callback. If set this is called whenever a certificate
5900 is required by client or server. An application can decide which
5901 certificate chain to present based on arbitrary criteria: for example
5902 supported signature algorithms. Add very simple example to s_server.
5903 This fixes many of the problems and restrictions of the existing client
5904 certificate callback: for example you can now clear an existing
5905 certificate and specify the whole chain.
5909 * Add new "valid_flags" field to CERT_PKEY structure which determines what
5910 the certificate can be used for (if anything). Set valid_flags field
5911 in new tls1_check_chain function. Simplify ssl_set_cert_masks which used
5912 to have similar checks in it.
5914 Add new "cert_flags" field to CERT structure and include a "strict mode".
5915 This enforces some TLS certificate requirements (such as only permitting
5916 certificate signature algorithms contained in the supported algorithms
5917 extension) which some implementations ignore: this option should be used
5918 with caution as it could cause interoperability issues.
5922 * Update and tidy signature algorithm extension processing. Work out
5923 shared signature algorithms based on preferences and peer algorithms
5924 and print them out in s_client and s_server. Abort handshake if no
5925 shared signature algorithms.
5929 * Add new functions to allow customised supported signature algorithms
5930 for SSL and SSL_CTX structures. Add options to s_client and s_server
5935 * New function SSL_certs_clear() to delete all references to certificates
5936 from an SSL structure. Before this once a certificate had been added
5937 it couldn't be removed.
5941 * Integrate hostname, email address and IP address checking with certificate
5942 verification. New verify options supporting checking in openssl utility.
5946 * Fixes and wildcard matching support to hostname and email checking
5947 functions. Add manual page.
5949 *Florian Weimer (Red Hat Product Security Team)*
5951 * New functions to check a hostname email or IP address against a
5952 certificate. Add options x509 utility to print results of checks against
5957 * Fix OCSP checking.
5959 *Rob Stradling <rob.stradling@comodo.com> and Ben Laurie*
5961 * Initial experimental support for explicitly trusted non-root CAs.
5962 OpenSSL still tries to build a complete chain to a root but if an
5963 intermediate CA has a trust setting included that is used. The first
5964 setting is used: whether to trust (e.g., -addtrust option to the x509
5969 * Add -trusted_first option which attempts to find certificates in the
5970 trusted store even if an untrusted chain is also supplied.
5974 * MIPS assembly pack updates: support for MIPS32r2 and SmartMIPS ASE,
5975 platform support for Linux and Android.
5979 * Support for linux-x32, ILP32 environment in x86_64 framework.
5983 * Experimental multi-implementation support for FIPS capable OpenSSL.
5984 When in FIPS mode the approved implementations are used as normal,
5985 when not in FIPS mode the internal unapproved versions are used instead.
5986 This means that the FIPS capable OpenSSL isn't forced to use the
5987 (often lower performance) FIPS implementations outside FIPS mode.
5991 * Transparently support X9.42 DH parameters when calling
5992 PEM_read_bio_DHparameters. This means existing applications can handle
5993 the new parameter format automatically.
5997 * Initial experimental support for X9.42 DH parameter format: mainly
5998 to support use of 'q' parameter for RFC5114 parameters.
6002 * Add DH parameters from RFC5114 including test data to dhtest.
6006 * Support for automatic EC temporary key parameter selection. If enabled
6007 the most preferred EC parameters are automatically used instead of
6008 hardcoded fixed parameters. Now a server just has to call:
6009 SSL_CTX_set_ecdh_auto(ctx, 1) and the server will automatically
6010 support ECDH and use the most appropriate parameters.
6014 * Enhance and tidy EC curve and point format TLS extension code. Use
6015 static structures instead of allocation if default values are used.
6016 New ctrls to set curves we wish to support and to retrieve shared curves.
6017 Print out shared curves in s_server. New options to s_server and s_client
6018 to set list of supported curves.
6022 * New ctrls to retrieve supported signature algorithms and
6023 supported curve values as an array of NIDs. Extend openssl utility
6024 to print out received values.
6028 * Add new APIs EC_curve_nist2nid and EC_curve_nid2nist which convert
6029 between NIDs and the more common NIST names such as "P-256". Enhance
6030 ecparam utility and ECC method to recognise the NIST names for curves.
6034 * Enhance SSL/TLS certificate chain handling to support different
6035 chains for each certificate instead of one chain in the parent SSL_CTX.
6039 * Support for fixed DH ciphersuite client authentication: where both
6040 server and client use DH certificates with common parameters.
6044 * Support for fixed DH ciphersuites: those requiring DH server
6049 * New function i2d_re_X509_tbs for re-encoding the TBS portion of
6051 Note: Related 1.0.2-beta specific macros X509_get_cert_info,
6052 X509_CINF_set_modified, X509_CINF_get_issuer, X509_CINF_get_extensions and
6053 X509_CINF_get_signature were reverted post internal team review.
6058 ### Changes between 1.0.1t and 1.0.1u [22 Sep 2016]
6060 * OCSP Status Request extension unbounded memory growth
6062 A malicious client can send an excessively large OCSP Status Request
6063 extension. If that client continually requests renegotiation, sending a
6064 large OCSP Status Request extension each time, then there will be unbounded
6065 memory growth on the server. This will eventually lead to a Denial Of
6066 Service attack through memory exhaustion. Servers with a default
6067 configuration are vulnerable even if they do not support OCSP. Builds using
6068 the "no-ocsp" build time option are not affected.
6070 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6075 * In order to mitigate the SWEET32 attack, the DES ciphers were moved from
6078 This issue was reported to OpenSSL Karthikeyan Bhargavan and Gaetan
6084 * OOB write in MDC2_Update()
6086 An overflow can occur in MDC2_Update() either if called directly or
6087 through the EVP_DigestUpdate() function using MDC2. If an attacker
6088 is able to supply very large amounts of input data after a previous
6089 call to EVP_EncryptUpdate() with a partial block then a length check
6090 can overflow resulting in a heap corruption.
6092 The amount of data needed is comparable to SIZE_MAX which is impractical
6095 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6100 * Malformed SHA512 ticket DoS
6102 If a server uses SHA512 for TLS session ticket HMAC it is vulnerable to a
6103 DoS attack where a malformed ticket will result in an OOB read which will
6106 The use of SHA512 in TLS session tickets is comparatively rare as it requires
6107 a custom server callback and ticket lookup mechanism.
6109 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6114 * OOB write in BN_bn2dec()
6116 The function BN_bn2dec() does not check the return value of BN_div_word().
6117 This can cause an OOB write if an application uses this function with an
6118 overly large BIGNUM. This could be a problem if an overly large certificate
6119 or CRL is printed out from an untrusted source. TLS is not affected because
6120 record limits will reject an oversized certificate before it is parsed.
6122 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6127 * OOB read in TS_OBJ_print_bio()
6129 The function TS_OBJ_print_bio() misuses OBJ_obj2txt(): the return value is
6130 the total length the OID text representation would use and not the amount
6131 of data written. This will result in OOB reads when large OIDs are
6134 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6139 * Pointer arithmetic undefined behaviour
6141 Avoid some undefined pointer arithmetic
6143 A common idiom in the codebase is to check limits in the following manner:
6146 Where "p" points to some malloc'd data of SIZE bytes and
6149 "len" here could be from some externally supplied data (e.g. from a TLS
6152 The rules of C pointer arithmetic are such that "p + len" is only well
6153 defined where len <= SIZE. Therefore the above idiom is actually
6154 undefined behaviour.
6156 For example this could cause problems if some malloc implementation
6157 provides an address for "p" such that "p + len" actually overflows for
6158 values of len that are too big and therefore p + len < limit.
6160 This issue was reported to OpenSSL by Guido Vranken
6165 * Constant time flag not preserved in DSA signing
6167 Operations in the DSA signing algorithm should run in constant time in
6168 order to avoid side channel attacks. A flaw in the OpenSSL DSA
6169 implementation means that a non-constant time codepath is followed for
6170 certain operations. This has been demonstrated through a cache-timing
6171 attack to be sufficient for an attacker to recover the private DSA key.
6173 This issue was reported by César Pereida (Aalto University), Billy Brumley
6174 (Tampere University of Technology), and Yuval Yarom (The University of
6175 Adelaide and NICTA).
6180 * DTLS buffered message DoS
6182 In a DTLS connection where handshake messages are delivered out-of-order
6183 those messages that OpenSSL is not yet ready to process will be buffered
6184 for later use. Under certain circumstances, a flaw in the logic means that
6185 those messages do not get removed from the buffer even though the handshake
6186 has been completed. An attacker could force up to approx. 15 messages to
6187 remain in the buffer when they are no longer required. These messages will
6188 be cleared when the DTLS connection is closed. The default maximum size for
6189 a message is 100k. Therefore the attacker could force an additional 1500k
6190 to be consumed per connection. By opening many simulataneous connections an
6191 attacker could cause a DoS attack through memory exhaustion.
6193 This issue was reported to OpenSSL by Quan Luo.
6198 * DTLS replay protection DoS
6200 A flaw in the DTLS replay attack protection mechanism means that records
6201 that arrive for future epochs update the replay protection "window" before
6202 the MAC for the record has been validated. This could be exploited by an
6203 attacker by sending a record for the next epoch (which does not have to
6204 decrypt or have a valid MAC), with a very large sequence number. This means
6205 that all subsequent legitimate packets are dropped causing a denial of
6206 service for a specific DTLS connection.
6208 This issue was reported to OpenSSL by the OCAP audit team.
6213 * Certificate message OOB reads
6215 In OpenSSL 1.0.2 and earlier some missing message length checks can result
6216 in OOB reads of up to 2 bytes beyond an allocated buffer. There is a
6217 theoretical DoS risk but this has not been observed in practice on common
6220 The messages affected are client certificate, client certificate request
6221 and server certificate. As a result the attack can only be performed
6222 against a client or a server which enables client authentication.
6224 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6229 ### Changes between 1.0.1s and 1.0.1t [3 May 2016]
6231 * Prevent padding oracle in AES-NI CBC MAC check
6233 A MITM attacker can use a padding oracle attack to decrypt traffic
6234 when the connection uses an AES CBC cipher and the server support
6237 This issue was introduced as part of the fix for Lucky 13 padding
6238 attack ([CVE-2013-0169]). The padding check was rewritten to be in
6239 constant time by making sure that always the same bytes are read and
6240 compared against either the MAC or padding bytes. But it no longer
6241 checked that there was enough data to have both the MAC and padding
6244 This issue was reported by Juraj Somorovsky using TLS-Attacker.
6249 * Fix EVP_EncodeUpdate overflow
6251 An overflow can occur in the EVP_EncodeUpdate() function which is used for
6252 Base64 encoding of binary data. If an attacker is able to supply very large
6253 amounts of input data then a length check can overflow resulting in a heap
6256 Internally to OpenSSL the EVP_EncodeUpdate() function is primarly used by
6257 the `PEM_write_bio*` family of functions. These are mainly used within the
6258 OpenSSL command line applications, so any application which processes data
6259 from an untrusted source and outputs it as a PEM file should be considered
6260 vulnerable to this issue. User applications that call these APIs directly
6261 with large amounts of untrusted data may also be vulnerable.
6263 This issue was reported by Guido Vranken.
6268 * Fix EVP_EncryptUpdate overflow
6270 An overflow can occur in the EVP_EncryptUpdate() function. If an attacker
6271 is able to supply very large amounts of input data after a previous call to
6272 EVP_EncryptUpdate() with a partial block then a length check can overflow
6273 resulting in a heap corruption. Following an analysis of all OpenSSL
6274 internal usage of the EVP_EncryptUpdate() function all usage is one of two
6275 forms. The first form is where the EVP_EncryptUpdate() call is known to be
6276 the first called function after an EVP_EncryptInit(), and therefore that
6277 specific call must be safe. The second form is where the length passed to
6278 EVP_EncryptUpdate() can be seen from the code to be some small value and
6279 therefore there is no possibility of an overflow. Since all instances are
6280 one of these two forms, it is believed that there can be no overflows in
6281 internal code due to this problem. It should be noted that
6282 EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths.
6283 Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances
6284 of these calls have also been analysed too and it is believed there are no
6285 instances in internal usage where an overflow could occur.
6287 This issue was reported by Guido Vranken.
6292 * Prevent ASN.1 BIO excessive memory allocation
6294 When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio()
6295 a short invalid encoding can casuse allocation of large amounts of memory
6296 potentially consuming excessive resources or exhausting memory.
6298 Any application parsing untrusted data through d2i BIO functions is
6299 affected. The memory based functions such as d2i_X509() are *not* affected.
6300 Since the memory based functions are used by the TLS library, TLS
6301 applications are not affected.
6303 This issue was reported by Brian Carpenter.
6310 ASN1 Strings that are over 1024 bytes can cause an overread in applications
6311 using the X509_NAME_oneline() function on EBCDIC systems. This could result
6312 in arbitrary stack data being returned in the buffer.
6314 This issue was reported by Guido Vranken.
6319 * Modify behavior of ALPN to invoke callback after SNI/servername
6320 callback, such that updates to the SSL_CTX affect ALPN.
6324 * Remove LOW from the DEFAULT cipher list. This removes singles DES from the
6329 * Only remove the SSLv2 methods with the no-ssl2-method option. When the
6330 methods are enabled and ssl2 is disabled the methods return NULL.
6334 ### Changes between 1.0.1r and 1.0.1s [1 Mar 2016]
6336 * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL.
6337 Builds that are not configured with "enable-weak-ssl-ciphers" will not
6338 provide any "EXPORT" or "LOW" strength ciphers.
6342 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2
6343 is by default disabled at build-time. Builds that are not configured with
6344 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used,
6345 users who want to negotiate SSLv2 via the version-flexible SSLv23_method()
6346 will need to explicitly call either of:
6348 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2);
6350 SSL_clear_options(ssl, SSL_OP_NO_SSLv2);
6352 as appropriate. Even if either of those is used, or the application
6353 explicitly uses the version-specific SSLv2_method() or its client and
6354 server variants, SSLv2 ciphers vulnerable to exhaustive search key
6355 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT
6356 ciphers, and SSLv2 56-bit DES are no longer available.
6361 * Fix a double-free in DSA code
6363 A double free bug was discovered when OpenSSL parses malformed DSA private
6364 keys and could lead to a DoS attack or memory corruption for applications
6365 that receive DSA private keys from untrusted sources. This scenario is
6368 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using
6374 * Disable SRP fake user seed to address a server memory leak.
6376 Add a new method SRP_VBASE_get1_by_user that handles the seed properly.
6378 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
6379 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user
6380 was changed to ignore the "fake user" SRP seed, even if the seed
6383 Users should use SRP_VBASE_get1_by_user instead. Note that in
6384 SRP_VBASE_get1_by_user, caller must free the returned value. Note
6385 also that even though configuring the SRP seed attempts to hide
6386 invalid usernames by continuing the handshake with fake
6387 credentials, this behaviour is not constant time and no strong
6388 guarantees are made that the handshake is indistinguishable from
6389 that of a valid user.
6394 * Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption
6396 In the BN_hex2bn function the number of hex digits is calculated using an
6397 int value `i`. Later `bn_expand` is called with a value of `i * 4`. For
6398 large values of `i` this can result in `bn_expand` not allocating any
6399 memory because `i * 4` is negative. This can leave the internal BIGNUM data
6400 field as NULL leading to a subsequent NULL ptr deref. For very large values
6401 of `i`, the calculation `i * 4` could be a positive value smaller than `i`.
6402 In this case memory is allocated to the internal BIGNUM data field, but it
6403 is insufficiently sized leading to heap corruption. A similar issue exists
6404 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn
6405 is ever called by user applications with very large untrusted hex/dec data.
6406 This is anticipated to be a rare occurrence.
6408 All OpenSSL internal usage of these functions use data that is not expected
6409 to be untrusted, e.g. config file data or application command line
6410 arguments. If user developed applications generate config file data based
6411 on untrusted data then it is possible that this could also lead to security
6412 consequences. This is also anticipated to be rare.
6414 This issue was reported to OpenSSL by Guido Vranken.
6419 * Fix memory issues in `BIO_*printf` functions
6421 The internal `fmtstr` function used in processing a "%s" format string in
6422 the `BIO_*printf` functions could overflow while calculating the length of a
6423 string and cause an OOB read when printing very long strings.
6425 Additionally the internal `doapr_outch` function can attempt to write to an
6426 OOB memory location (at an offset from the NULL pointer) in the event of a
6427 memory allocation failure. In 1.0.2 and below this could be caused where
6428 the size of a buffer to be allocated is greater than INT_MAX. E.g. this
6429 could be in processing a very long "%s" format string. Memory leaks can
6432 The first issue may mask the second issue dependent on compiler behaviour.
6433 These problems could enable attacks where large amounts of untrusted data
6434 is passed to the `BIO_*printf` functions. If applications use these functions
6435 in this way then they could be vulnerable. OpenSSL itself uses these
6436 functions when printing out human-readable dumps of ASN.1 data. Therefore
6437 applications that print this data could be vulnerable if the data is from
6438 untrusted sources. OpenSSL command line applications could also be
6439 vulnerable where they print out ASN.1 data, or if untrusted data is passed
6440 as command line arguments.
6442 Libssl is not considered directly vulnerable. Additionally certificates etc
6443 received via remote connections via libssl are also unlikely to be able to
6444 trigger these issues because of message size limits enforced within libssl.
6446 This issue was reported to OpenSSL Guido Vranken.
6451 * Side channel attack on modular exponentiation
6453 A side-channel attack was found which makes use of cache-bank conflicts on
6454 the Intel Sandy-Bridge microarchitecture which could lead to the recovery
6455 of RSA keys. The ability to exploit this issue is limited as it relies on
6456 an attacker who has control of code in a thread running on the same
6457 hyper-threaded core as the victim thread which is performing decryptions.
6459 This issue was reported to OpenSSL by Yuval Yarom, The University of
6460 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and
6461 Nadia Heninger, University of Pennsylvania with more information at
6462 <http://cachebleed.info>.
6467 * Change the req command to generate a 2048-bit RSA/DSA key by default,
6468 if no keysize is specified with default_bits. This fixes an
6469 omission in an earlier change that changed all RSA/DSA key generation
6470 commands to use 2048 bits by default.
6474 ### Changes between 1.0.1q and 1.0.1r [28 Jan 2016]
6476 * Protection for DH small subgroup attacks
6478 As a precautionary measure the SSL_OP_SINGLE_DH_USE option has been
6479 switched on by default and cannot be disabled. This could have some
6484 * SSLv2 doesn't block disabled ciphers
6486 A malicious client can negotiate SSLv2 ciphers that have been disabled on
6487 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
6488 been disabled, provided that the SSLv2 protocol was not also disabled via
6491 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
6492 and Sebastian Schinzel.
6497 * Reject DH handshakes with parameters shorter than 1024 bits.
6501 ### Changes between 1.0.1p and 1.0.1q [3 Dec 2015]
6503 * Certificate verify crash with missing PSS parameter
6505 The signature verification routines will crash with a NULL pointer
6506 dereference if presented with an ASN.1 signature using the RSA PSS
6507 algorithm and absent mask generation function parameter. Since these
6508 routines are used to verify certificate signature algorithms this can be
6509 used to crash any certificate verification operation and exploited in a
6510 DoS attack. Any application which performs certificate verification is
6511 vulnerable including OpenSSL clients and servers which enable client
6514 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
6519 * X509_ATTRIBUTE memory leak
6521 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
6522 memory. This structure is used by the PKCS#7 and CMS routines so any
6523 application which reads PKCS#7 or CMS data from untrusted sources is
6524 affected. SSL/TLS is not affected.
6526 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
6532 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
6533 This changes the decoding behaviour for some invalid messages,
6534 though the change is mostly in the more lenient direction, and
6535 legacy behaviour is preserved as much as possible.
6539 * In DSA_generate_parameters_ex, if the provided seed is too short,
6540 use a random seed, as already documented.
6542 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
6544 ### Changes between 1.0.1o and 1.0.1p [9 Jul 2015]
6546 * Alternate chains certificate forgery
6548 During certificate verfification, OpenSSL will attempt to find an
6549 alternative certificate chain if the first attempt to build such a chain
6550 fails. An error in the implementation of this logic can mean that an
6551 attacker could cause certain checks on untrusted certificates to be
6552 bypassed, such as the CA flag, enabling them to use a valid leaf
6553 certificate to act as a CA and "issue" an invalid certificate.
6555 This issue was reported to OpenSSL by Adam Langley/David Benjamin
6561 * Race condition handling PSK identify hint
6563 If PSK identity hints are received by a multi-threaded client then
6564 the values are wrongly updated in the parent SSL_CTX structure. This can
6565 result in a race condition potentially leading to a double free of the
6571 ### Changes between 1.0.1n and 1.0.1o [12 Jun 2015]
6573 * Fix HMAC ABI incompatibility. The previous version introduced an ABI
6574 incompatibility in the handling of HMAC. The previous ABI has now been
6577 ### Changes between 1.0.1m and 1.0.1n [11 Jun 2015]
6579 * Malformed ECParameters causes infinite loop
6581 When processing an ECParameters structure OpenSSL enters an infinite loop
6582 if the curve specified is over a specially malformed binary polynomial
6585 This can be used to perform denial of service against any
6586 system which processes public keys, certificate requests or
6587 certificates. This includes TLS clients and TLS servers with
6588 client authentication enabled.
6590 This issue was reported to OpenSSL by Joseph Barr-Pixton.
6595 * Exploitable out-of-bounds read in X509_cmp_time
6597 X509_cmp_time does not properly check the length of the ASN1_TIME
6598 string and can read a few bytes out of bounds. In addition,
6599 X509_cmp_time accepts an arbitrary number of fractional seconds in the
6602 An attacker can use this to craft malformed certificates and CRLs of
6603 various sizes and potentially cause a segmentation fault, resulting in
6604 a DoS on applications that verify certificates or CRLs. TLS clients
6605 that verify CRLs are affected. TLS clients and servers with client
6606 authentication enabled may be affected if they use custom verification
6609 This issue was reported to OpenSSL by Robert Swiecki (Google), and
6610 independently by Hanno Böck.
6615 * PKCS7 crash with missing EnvelopedContent
6617 The PKCS#7 parsing code does not handle missing inner EncryptedContent
6618 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
6619 with missing content and trigger a NULL pointer dereference on parsing.
6621 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
6622 structures from untrusted sources are affected. OpenSSL clients and
6623 servers are not affected.
6625 This issue was reported to OpenSSL by Michal Zalewski (Google).
6630 * CMS verify infinite loop with unknown hash function
6632 When verifying a signedData message the CMS code can enter an infinite loop
6633 if presented with an unknown hash function OID. This can be used to perform
6634 denial of service against any system which verifies signedData messages using
6636 This issue was reported to OpenSSL by Johannes Bauer.
6641 * Race condition handling NewSessionTicket
6643 If a NewSessionTicket is received by a multi-threaded client when attempting to
6644 reuse a previous ticket then a race condition can occur potentially leading to
6645 a double free of the ticket data.
6650 * Reject DH handshakes with parameters shorter than 768 bits.
6652 *Kurt Roeckx and Emilia Kasper*
6654 * dhparam: generate 2048-bit parameters by default.
6656 *Kurt Roeckx and Emilia Kasper*
6658 ### Changes between 1.0.1l and 1.0.1m [19 Mar 2015]
6660 * Segmentation fault in ASN1_TYPE_cmp fix
6662 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
6663 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
6664 certificate signature algorithm consistency this can be used to crash any
6665 certificate verification operation and exploited in a DoS attack. Any
6666 application which performs certificate verification is vulnerable including
6667 OpenSSL clients and servers which enable client authentication.
6672 * ASN.1 structure reuse memory corruption fix
6674 Reusing a structure in ASN.1 parsing may allow an attacker to cause
6675 memory corruption via an invalid write. Such reuse is and has been
6676 strongly discouraged and is believed to be rare.
6678 Applications that parse structures containing CHOICE or ANY DEFINED BY
6679 components may be affected. Certificate parsing (d2i_X509 and related
6680 functions) are however not affected. OpenSSL clients and servers are
6686 * PKCS7 NULL pointer dereferences fix
6688 The PKCS#7 parsing code does not handle missing outer ContentInfo
6689 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
6690 missing content and trigger a NULL pointer dereference on parsing.
6692 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
6693 otherwise parse PKCS#7 structures from untrusted sources are
6694 affected. OpenSSL clients and servers are not affected.
6696 This issue was reported to OpenSSL by Michal Zalewski (Google).
6701 * DoS via reachable assert in SSLv2 servers fix
6703 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
6704 servers that both support SSLv2 and enable export cipher suites by sending
6705 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
6707 This issue was discovered by Sean Burford (Google) and Emilia Käsper
6708 (OpenSSL development team).
6713 * Use After Free following d2i_ECPrivatekey error fix
6715 A malformed EC private key file consumed via the d2i_ECPrivateKey function
6716 could cause a use after free condition. This, in turn, could cause a double
6717 free in several private key parsing functions (such as d2i_PrivateKey
6718 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
6719 for applications that receive EC private keys from untrusted
6720 sources. This scenario is considered rare.
6722 This issue was discovered by the BoringSSL project and fixed in their
6728 * X509_to_X509_REQ NULL pointer deref fix
6730 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
6731 the certificate key is invalid. This function is rarely used in practice.
6733 This issue was discovered by Brian Carpenter.
6738 * Removed the export ciphers from the DEFAULT ciphers
6742 ### Changes between 1.0.1k and 1.0.1l [15 Jan 2015]
6744 * Build fixes for the Windows and OpenVMS platforms
6746 *Matt Caswell and Richard Levitte*
6748 ### Changes between 1.0.1j and 1.0.1k [8 Jan 2015]
6750 * Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
6751 message can cause a segmentation fault in OpenSSL due to a NULL pointer
6752 dereference. This could lead to a Denial Of Service attack. Thanks to
6753 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
6758 * Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
6759 dtls1_buffer_record function under certain conditions. In particular this
6760 could occur if an attacker sent repeated DTLS records with the same
6761 sequence number but for the next epoch. The memory leak could be exploited
6762 by an attacker in a Denial of Service attack through memory exhaustion.
6763 Thanks to Chris Mueller for reporting this issue.
6768 * Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
6769 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
6770 method would be set to NULL which could later result in a NULL pointer
6771 dereference. Thanks to Frank Schmirler for reporting this issue.
6776 * Abort handshake if server key exchange message is omitted for ephemeral
6779 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
6780 reporting this issue.
6785 * Remove non-export ephemeral RSA code on client and server. This code
6786 violated the TLS standard by allowing the use of temporary RSA keys in
6787 non-export ciphersuites and could be used by a server to effectively
6788 downgrade the RSA key length used to a value smaller than the server
6789 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
6790 INRIA or reporting this issue.
6795 * Fixed issue where DH client certificates are accepted without verification.
6796 An OpenSSL server will accept a DH certificate for client authentication
6797 without the certificate verify message. This effectively allows a client to
6798 authenticate without the use of a private key. This only affects servers
6799 which trust a client certificate authority which issues certificates
6800 containing DH keys: these are extremely rare and hardly ever encountered.
6801 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
6807 * Ensure that the session ID context of an SSL is updated when its
6808 SSL_CTX is updated via SSL_set_SSL_CTX.
6810 The session ID context is typically set from the parent SSL_CTX,
6811 and can vary with the CTX.
6815 * Fix various certificate fingerprint issues.
6817 By using non-DER or invalid encodings outside the signed portion of a
6818 certificate the fingerprint can be changed without breaking the signature.
6819 Although no details of the signed portion of the certificate can be changed
6820 this can cause problems with some applications: e.g. those using the
6821 certificate fingerprint for blacklists.
6823 1. Reject signatures with non zero unused bits.
6825 If the BIT STRING containing the signature has non zero unused bits reject
6826 the signature. All current signature algorithms require zero unused bits.
6828 2. Check certificate algorithm consistency.
6830 Check the AlgorithmIdentifier inside TBS matches the one in the
6831 certificate signature. NB: this will result in signature failure
6832 errors for some broken certificates.
6834 Thanks to Konrad Kraszewski from Google for reporting this issue.
6836 3. Check DSA/ECDSA signatures use DER.
6838 Re-encode DSA/ECDSA signatures and compare with the original received
6839 signature. Return an error if there is a mismatch.
6841 This will reject various cases including garbage after signature
6842 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
6843 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
6844 (negative or with leading zeroes).
6846 Further analysis was conducted and fixes were developed by Stephen Henson
6847 of the OpenSSL core team.
6853 * Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
6854 results on some platforms, including x86_64. This bug occurs at random
6855 with a very low probability, and is not known to be exploitable in any
6856 way, though its exact impact is difficult to determine. Thanks to Pieter
6857 Wuille (Blockstream) who reported this issue and also suggested an initial
6858 fix. Further analysis was conducted by the OpenSSL development team and
6859 Adam Langley of Google. The final fix was developed by Andy Polyakov of
6860 the OpenSSL core team.
6865 * Do not resume sessions on the server if the negotiated protocol
6866 version does not match the session's version. Resuming with a different
6867 version, while not strictly forbidden by the RFC, is of questionable
6868 sanity and breaks all known clients.
6870 *David Benjamin, Emilia Käsper*
6872 * Tighten handling of the ChangeCipherSpec (CCS) message: reject
6873 early CCS messages during renegotiation. (Note that because
6874 renegotiation is encrypted, this early CCS was not exploitable.)
6878 * Tighten client-side session ticket handling during renegotiation:
6879 ensure that the client only accepts a session ticket if the server sends
6880 the extension anew in the ServerHello. Previously, a TLS client would
6881 reuse the old extension state and thus accept a session ticket if one was
6882 announced in the initial ServerHello.
6884 Similarly, ensure that the client requires a session ticket if one
6885 was advertised in the ServerHello. Previously, a TLS client would
6886 ignore a missing NewSessionTicket message.
6890 ### Changes between 1.0.1i and 1.0.1j [15 Oct 2014]
6894 A flaw in the DTLS SRTP extension parsing code allows an attacker, who
6895 sends a carefully crafted handshake message, to cause OpenSSL to fail
6896 to free up to 64k of memory causing a memory leak. This could be
6897 exploited in a Denial Of Service attack. This issue affects OpenSSL
6898 1.0.1 server implementations for both SSL/TLS and DTLS regardless of
6899 whether SRTP is used or configured. Implementations of OpenSSL that
6900 have been compiled with OPENSSL_NO_SRTP defined are not affected.
6902 The fix was developed by the OpenSSL team.
6907 * Session Ticket Memory Leak.
6909 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
6910 integrity of that ticket is first verified. In the event of a session
6911 ticket integrity check failing, OpenSSL will fail to free memory
6912 causing a memory leak. By sending a large number of invalid session
6913 tickets an attacker could exploit this issue in a Denial Of Service
6919 * Build option no-ssl3 is incomplete.
6921 When OpenSSL is configured with "no-ssl3" as a build option, servers
6922 could accept and complete a SSL 3.0 handshake, and clients could be
6923 configured to send them.
6926 *Akamai and the OpenSSL team*
6928 * Add support for TLS_FALLBACK_SCSV.
6929 Client applications doing fallback retries should call
6930 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
6933 *Adam Langley, Bodo Moeller*
6935 * Add additional DigestInfo checks.
6937 Re-encode DigestInto in DER and check against the original when
6938 verifying RSA signature: this will reject any improperly encoded
6939 DigestInfo structures.
6941 Note: this is a precautionary measure and no attacks are currently known.
6945 ### Changes between 1.0.1h and 1.0.1i [6 Aug 2014]
6947 * Fix SRP buffer overrun vulnerability. Invalid parameters passed to the
6948 SRP code can be overrun an internal buffer. Add sanity check that
6949 g, A, B < N to SRP code.
6951 Thanks to Sean Devlin and Watson Ladd of Cryptography Services, NCC
6952 Group for discovering this issue.
6957 * A flaw in the OpenSSL SSL/TLS server code causes the server to negotiate
6958 TLS 1.0 instead of higher protocol versions when the ClientHello message
6959 is badly fragmented. This allows a man-in-the-middle attacker to force a
6960 downgrade to TLS 1.0 even if both the server and the client support a
6961 higher protocol version, by modifying the client's TLS records.
6963 Thanks to David Benjamin and Adam Langley (Google) for discovering and
6964 researching this issue.
6969 * OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
6970 to a denial of service attack. A malicious server can crash the client
6971 with a null pointer dereference (read) by specifying an anonymous (EC)DH
6972 ciphersuite and sending carefully crafted handshake messages.
6974 Thanks to Felix Gröbert (Google) for discovering and researching this
6980 * By sending carefully crafted DTLS packets an attacker could cause openssl
6981 to leak memory. This can be exploited through a Denial of Service attack.
6982 Thanks to Adam Langley for discovering and researching this issue.
6987 * An attacker can force openssl to consume large amounts of memory whilst
6988 processing DTLS handshake messages. This can be exploited through a
6989 Denial of Service attack.
6990 Thanks to Adam Langley for discovering and researching this issue.
6995 * An attacker can force an error condition which causes openssl to crash
6996 whilst processing DTLS packets due to memory being freed twice. This
6997 can be exploited through a Denial of Service attack.
6998 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
7004 * If a multithreaded client connects to a malicious server using a resumed
7005 session and the server sends an ec point format extension it could write
7006 up to 255 bytes to freed memory.
7008 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
7014 * A malicious server can crash an OpenSSL client with a null pointer
7015 dereference (read) by specifying an SRP ciphersuite even though it was not
7016 properly negotiated with the client. This can be exploited through a
7017 Denial of Service attack.
7019 Thanks to Joonas Kuorilehto and Riku Hietamäki (Codenomicon) for
7020 discovering and researching this issue.
7025 * A flaw in OBJ_obj2txt may cause pretty printing functions such as
7026 X509_name_oneline, X509_name_print_ex et al. to leak some information
7027 from the stack. Applications may be affected if they echo pretty printing
7028 output to the attacker.
7030 Thanks to Ivan Fratric (Google) for discovering this issue.
7033 *Emilia Käsper, and Steve Henson*
7035 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
7036 for corner cases. (Certain input points at infinity could lead to
7037 bogus results, with non-infinity inputs mapped to infinity too.)
7041 ### Changes between 1.0.1g and 1.0.1h [5 Jun 2014]
7043 * Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
7044 handshake can force the use of weak keying material in OpenSSL
7045 SSL/TLS clients and servers.
7047 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
7048 researching this issue. ([CVE-2014-0224])
7050 *KIKUCHI Masashi, Steve Henson*
7052 * Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
7053 OpenSSL DTLS client the code can be made to recurse eventually crashing
7056 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
7059 *Imre Rad, Steve Henson*
7061 * Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
7062 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
7063 client or server. This is potentially exploitable to run arbitrary
7064 code on a vulnerable client or server.
7066 Thanks to Jüri Aedla for reporting this issue. ([CVE-2014-0195])
7068 *Jüri Aedla, Steve Henson*
7070 * Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
7071 are subject to a denial of service attack.
7073 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
7074 this issue. ([CVE-2014-3470])
7076 *Felix Gröbert, Ivan Fratric, Steve Henson*
7078 * Harmonize version and its documentation. -f flag is used to display
7081 *mancha <mancha1@zoho.com>*
7083 * Fix eckey_priv_encode so it immediately returns an error upon a failure
7084 in i2d_ECPrivateKey.
7086 *mancha <mancha1@zoho.com>*
7088 * Fix some double frees. These are not thought to be exploitable.
7090 *mancha <mancha1@zoho.com>*
7092 ### Changes between 1.0.1f and 1.0.1g [7 Apr 2014]
7094 * A missing bounds check in the handling of the TLS heartbeat extension
7095 can be used to reveal up to 64k of memory to a connected client or
7098 Thanks for Neel Mehta of Google Security for discovering this bug and to
7099 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
7100 preparing the fix ([CVE-2014-0160])
7102 *Adam Langley, Bodo Moeller*
7104 * Fix for the attack described in the paper "Recovering OpenSSL
7105 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
7106 by Yuval Yarom and Naomi Benger. Details can be obtained from:
7107 <http://eprint.iacr.org/2014/140>
7109 Thanks to Yuval Yarom and Naomi Benger for discovering this
7110 flaw and to Yuval Yarom for supplying a fix ([CVE-2014-0076])
7112 *Yuval Yarom and Naomi Benger*
7114 * TLS pad extension: draft-agl-tls-padding-03
7116 Workaround for the "TLS hang bug" (see FAQ and PR#2771): if the
7117 TLS client Hello record length value would otherwise be > 255 and
7118 less that 512 pad with a dummy extension containing zeroes so it
7119 is at least 512 bytes long.
7121 *Adam Langley, Steve Henson*
7123 ### Changes between 1.0.1e and 1.0.1f [6 Jan 2014]
7125 * Fix for TLS record tampering bug. A carefully crafted invalid
7126 handshake could crash OpenSSL with a NULL pointer exception.
7127 Thanks to Anton Johansson for reporting this issues.
7130 * Keep original DTLS digest and encryption contexts in retransmission
7131 structures so we can use the previous session parameters if they need
7132 to be resent. ([CVE-2013-6450])
7136 * Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
7137 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
7138 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
7139 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
7140 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
7141 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
7143 *Rob Stradling, Adam Langley*
7145 ### Changes between 1.0.1d and 1.0.1e [11 Feb 2013]
7147 * Correct fix for CVE-2013-0169. The original didn't work on AES-NI
7148 supporting platforms or when small records were transferred.
7150 *Andy Polyakov, Steve Henson*
7152 ### Changes between 1.0.1c and 1.0.1d [5 Feb 2013]
7154 * Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
7156 This addresses the flaw in CBC record processing discovered by
7157 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
7158 at: <http://www.isg.rhul.ac.uk/tls/>
7160 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
7161 Security Group at Royal Holloway, University of London
7162 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
7163 Emilia Käsper for the initial patch.
7166 *Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson*
7168 * Fix flaw in AESNI handling of TLS 1.2 and 1.1 records for CBC mode
7169 ciphersuites which can be exploited in a denial of service attack.
7170 Thanks go to and to Adam Langley <agl@chromium.org> for discovering
7171 and detecting this bug and to Wolfgang Ettlinger
7172 <wolfgang.ettlinger@gmail.com> for independently discovering this issue.
7177 * Return an error when checking OCSP signatures when key is NULL.
7178 This fixes a DoS attack. ([CVE-2013-0166])
7182 * Make openssl verify return errors.
7184 *Chris Palmer <palmer@google.com> and Ben Laurie*
7186 * Call OCSP Stapling callback after ciphersuite has been chosen, so
7187 the right response is stapled. Also change SSL_get_certificate()
7188 so it returns the certificate actually sent.
7189 See <http://rt.openssl.org/Ticket/Display.html?id=2836>.
7191 *Rob Stradling <rob.stradling@comodo.com>*
7193 * Fix possible deadlock when decoding public keys.
7197 * Don't use TLS 1.0 record version number in initial client hello
7202 ### Changes between 1.0.1b and 1.0.1c [10 May 2012]
7204 * Sanity check record length before skipping explicit IV in TLS
7205 1.2, 1.1 and DTLS to fix DoS attack.
7207 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
7208 fuzzing as a service testing platform.
7213 * Initialise tkeylen properly when encrypting CMS messages.
7214 Thanks to Solar Designer of Openwall for reporting this issue.
7218 * In FIPS mode don't try to use composite ciphers as they are not
7223 ### Changes between 1.0.1a and 1.0.1b [26 Apr 2012]
7225 * OpenSSL 1.0.0 sets SSL_OP_ALL to 0x80000FFFL and OpenSSL 1.0.1 and
7226 1.0.1a set SSL_OP_NO_TLSv1_1 to 0x00000400L which would unfortunately
7227 mean any application compiled against OpenSSL 1.0.0 headers setting
7228 SSL_OP_ALL would also set SSL_OP_NO_TLSv1_1, unintentionally disabling
7229 TLS 1.1 also. Fix this by changing the value of SSL_OP_NO_TLSv1_1 to
7230 0x10000000L Any application which was previously compiled against
7231 OpenSSL 1.0.1 or 1.0.1a headers and which cares about SSL_OP_NO_TLSv1_1
7232 will need to be recompiled as a result. Letting be results in
7233 inability to disable specifically TLS 1.1 and in client context,
7234 in unlike event, limit maximum offered version to TLS 1.0 [see below].
7238 * In order to ensure interoperability SSL_OP_NO_protocolX does not
7239 disable just protocol X, but all protocols above X *if* there are
7240 protocols *below* X still enabled. In more practical terms it means
7241 that if application wants to disable TLS1.0 in favor of TLS1.1 and
7242 above, it's not sufficient to pass `SSL_OP_NO_TLSv1`, one has to pass
7243 `SSL_OP_NO_TLSv1|SSL_OP_NO_SSLv3|SSL_OP_NO_SSLv2`. This applies to
7248 ### Changes between 1.0.1 and 1.0.1a [19 Apr 2012]
7250 * Check for potentially exploitable overflows in asn1_d2i_read_bio
7251 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
7252 in CRYPTO_realloc_clean.
7254 Thanks to Tavis Ormandy, Google Security Team, for discovering this
7255 issue and to Adam Langley <agl@chromium.org> for fixing it.
7258 *Adam Langley (Google), Tavis Ormandy, Google Security Team*
7260 * Don't allow TLS 1.2 SHA-256 ciphersuites in TLS 1.0, 1.1 connections.
7264 * Workarounds for some broken servers that "hang" if a client hello
7265 record length exceeds 255 bytes.
7267 1. Do not use record version number > TLS 1.0 in initial client
7268 hello: some (but not all) hanging servers will now work.
7269 2. If we set OPENSSL_MAX_TLS1_2_CIPHER_LENGTH this will truncate
7270 the number of ciphers sent in the client hello. This should be
7271 set to an even number, such as 50, for example by passing:
7272 -DOPENSSL_MAX_TLS1_2_CIPHER_LENGTH=50 to config or Configure.
7273 Most broken servers should now work.
7274 3. If all else fails setting OPENSSL_NO_TLS1_2_CLIENT will disable
7275 TLS 1.2 client support entirely.
7279 * Fix SEGV in Vector Permutation AES module observed in OpenSSH.
7283 ### Changes between 1.0.0h and 1.0.1 [14 Mar 2012]
7285 * Add compatibility with old MDC2 signatures which use an ASN1 OCTET
7286 STRING form instead of a DigestInfo.
7290 * The format used for MDC2 RSA signatures is inconsistent between EVP
7291 and the RSA_sign/RSA_verify functions. This was made more apparent when
7292 OpenSSL used RSA_sign/RSA_verify for some RSA signatures in particular
7293 those which went through EVP_PKEY_METHOD in 1.0.0 and later. Detect
7294 the correct format in RSA_verify so both forms transparently work.
7298 * Some servers which support TLS 1.0 can choke if we initially indicate
7299 support for TLS 1.2 and later renegotiate using TLS 1.0 in the RSA
7300 encrypted premaster secret. As a workaround use the maximum permitted
7301 client version in client hello, this should keep such servers happy
7302 and still work with previous versions of OpenSSL.
7306 * Add support for TLS/DTLS heartbeats.
7308 *Robin Seggelmann <seggelmann@fh-muenster.de>*
7310 * Add support for SCTP.
7312 *Robin Seggelmann <seggelmann@fh-muenster.de>*
7314 * Improved PRNG seeding for VOS.
7316 *Paul Green <Paul.Green@stratus.com>*
7318 * Extensive assembler packs updates, most notably:
7320 - x86[_64]: AES-NI, PCLMULQDQ, RDRAND support;
7321 - x86[_64]: SSSE3 support (SHA1, vector-permutation AES);
7322 - x86_64: bit-sliced AES implementation;
7323 - ARM: NEON support, contemporary platforms optimizations;
7324 - s390x: z196 support;
7325 - `*`: GHASH and GF(2^m) multiplication implementations;
7329 * Make TLS-SRP code conformant with RFC 5054 API cleanup
7330 (removal of unnecessary code)
7332 *Peter Sylvester <peter.sylvester@edelweb.fr>*
7334 * Add TLS key material exporter from RFC 5705.
7338 * Add DTLS-SRTP negotiation from RFC 5764.
7342 * Add Next Protocol Negotiation,
7343 <http://tools.ietf.org/html/draft-agl-tls-nextprotoneg-00>. Can be
7344 disabled with a no-npn flag to config or Configure. Code donated
7347 *Adam Langley <agl@google.com> and Ben Laurie*
7349 * Add optional 64-bit optimized implementations of elliptic curves NIST-P224,
7350 NIST-P256, NIST-P521, with constant-time single point multiplication on
7351 typical inputs. Compiler support for the nonstandard type `__uint128_t` is
7352 required to use this (present in gcc 4.4 and later, for 64-bit builds).
7353 Code made available under Apache License version 2.0.
7355 Specify "enable-ec_nistp_64_gcc_128" on the Configure (or config) command
7356 line to include this in your build of OpenSSL, and run "make depend" (or
7357 "make update"). This enables the following EC_METHODs:
7359 EC_GFp_nistp224_method()
7360 EC_GFp_nistp256_method()
7361 EC_GFp_nistp521_method()
7363 EC_GROUP_new_by_curve_name() will automatically use these (while
7364 EC_GROUP_new_curve_GFp() currently prefers the more flexible
7367 *Emilia Käsper, Adam Langley, Bodo Moeller (Google)*
7369 * Use type ossl_ssize_t instead of ssize_t which isn't available on
7370 all platforms. Move ssize_t definition from e_os.h to the public
7371 header file e_os2.h as it now appears in public header file cms.h
7375 * New -sigopt option to the ca, req and x509 utilities. Additional
7376 signature parameters can be passed using this option and in
7381 * Add RSA PSS signing function. This will generate and set the
7382 appropriate AlgorithmIdentifiers for PSS based on those in the
7383 corresponding EVP_MD_CTX structure. No application support yet.
7387 * Support for companion algorithm specific ASN1 signing routines.
7388 New function ASN1_item_sign_ctx() signs a pre-initialised
7389 EVP_MD_CTX structure and sets AlgorithmIdentifiers based on
7390 the appropriate parameters.
7394 * Add new algorithm specific ASN1 verification initialisation function
7395 to EVP_PKEY_ASN1_METHOD: this is not in EVP_PKEY_METHOD since the ASN1
7396 handling will be the same no matter what EVP_PKEY_METHOD is used.
7397 Add a PSS handler to support verification of PSS signatures: checked
7398 against a number of sample certificates.
7402 * Add signature printing for PSS. Add PSS OIDs.
7404 *Steve Henson, Martin Kaiser <lists@kaiser.cx>*
7406 * Add algorithm specific signature printing. An individual ASN1 method
7407 can now print out signatures instead of the standard hex dump.
7409 More complex signatures (e.g. PSS) can print out more meaningful
7410 information. Include DSA version that prints out the signature
7415 * Password based recipient info support for CMS library: implementing
7420 * Split password based encryption into PBES2 and PBKDF2 functions. This
7421 neatly separates the code into cipher and PBE sections and is required
7422 for some algorithms that split PBES2 into separate pieces (such as
7423 password based CMS).
7427 * Session-handling fixes:
7428 - Fix handling of connections that are resuming with a session ID,
7429 but also support Session Tickets.
7430 - Fix a bug that suppressed issuing of a new ticket if the client
7431 presented a ticket with an expired session.
7432 - Try to set the ticket lifetime hint to something reasonable.
7433 - Make tickets shorter by excluding irrelevant information.
7434 - On the client side, don't ignore renewed tickets.
7436 *Adam Langley, Bodo Moeller (Google)*
7438 * Fix PSK session representation.
7442 * Add RC4-MD5 and AESNI-SHA1 "stitched" implementations.
7444 This work was sponsored by Intel.
7448 * Add GCM support to TLS library. Some custom code is needed to split
7449 the IV between the fixed (from PRF) and explicit (from TLS record)
7450 portions. This adds all GCM ciphersuites supported by RFC5288 and
7451 RFC5289. Generalise some `AES*` cipherstrings to include GCM and
7452 add a special AESGCM string for GCM only.
7456 * Expand range of ctrls for AES GCM. Permit setting invocation
7457 field on decrypt and retrieval of invocation field only on encrypt.
7461 * Add HMAC ECC ciphersuites from RFC5289. Include SHA384 PRF support.
7462 As required by RFC5289 these ciphersuites cannot be used if for
7463 versions of TLS earlier than 1.2.
7467 * For FIPS capable OpenSSL interpret a NULL default public key method
7468 as unset and return the appropriate default but do *not* set the default.
7469 This means we can return the appropriate method in applications that
7470 switch between FIPS and non-FIPS modes.
7474 * Redirect HMAC and CMAC operations to FIPS module in FIPS mode. If an
7475 ENGINE is used then we cannot handle that in the FIPS module so we
7476 keep original code iff non-FIPS operations are allowed.
7480 * Add -attime option to openssl utilities.
7482 *Peter Eckersley <pde@eff.org>, Ben Laurie and Steve Henson*
7484 * Redirect DSA and DH operations to FIPS module in FIPS mode.
7488 * Redirect ECDSA and ECDH operations to FIPS module in FIPS mode. Also use
7489 FIPS EC methods unconditionally for now.
7493 * New build option no-ec2m to disable characteristic 2 code.
7497 * Backport libcrypto audit of return value checking from 1.1.0-dev; not
7498 all cases can be covered as some introduce binary incompatibilities.
7502 * Redirect RSA operations to FIPS module including keygen,
7503 encrypt, decrypt, sign and verify. Block use of non FIPS RSA methods.
7507 * Add similar low-level API blocking to ciphers.
7511 * low-level digest APIs are not approved in FIPS mode: any attempt
7512 to use these will cause a fatal error. Applications that *really* want
7513 to use them can use the `private_*` version instead.
7517 * Redirect cipher operations to FIPS module for FIPS builds.
7521 * Redirect digest operations to FIPS module for FIPS builds.
7525 * Update build system to add "fips" flag which will link in fipscanister.o
7526 for static and shared library builds embedding a signature if needed.
7530 * Output TLS supported curves in preference order instead of numerical
7531 order. This is currently hardcoded for the highest order curves first.
7532 This should be configurable so applications can judge speed vs strength.
7536 * Add TLS v1.2 server support for client authentication.
7540 * Add support for FIPS mode in ssl library: disable SSLv3, non-FIPS ciphers
7545 * Functions FIPS_mode_set() and FIPS_mode() which call the underlying
7546 FIPS modules versions.
7550 * Add TLS v1.2 client side support for client authentication. Keep cache
7551 of handshake records longer as we don't know the hash algorithm to use
7552 until after the certificate request message is received.
7556 * Initial TLS v1.2 client support. Add a default signature algorithms
7557 extension including all the algorithms we support. Parse new signature
7558 format in client key exchange. Relax some ECC signing restrictions for
7559 TLS v1.2 as indicated in RFC5246.
7563 * Add server support for TLS v1.2 signature algorithms extension. Switch
7564 to new signature format when needed using client digest preference.
7565 All server ciphersuites should now work correctly in TLS v1.2. No client
7566 support yet and no support for client certificates.
7570 * Initial TLS v1.2 support. Add new SHA256 digest to ssl code, switch
7571 to SHA256 for PRF when using TLS v1.2 and later. Add new SHA256 based
7572 ciphersuites. At present only RSA key exchange ciphersuites work with
7573 TLS v1.2. Add new option for TLS v1.2 replacing the old and obsolete
7574 SSL_OP_PKCS1_CHECK flags with SSL_OP_NO_TLSv1_2. New TLSv1.2 methods
7575 and version checking.
7579 * New option OPENSSL_NO_SSL_INTERN. If an application can be compiled
7580 with this defined it will not be affected by any changes to ssl internal
7581 structures. Add several utility functions to allow openssl application
7582 to work with OPENSSL_NO_SSL_INTERN defined.
7586 * A long standing patch to add support for SRP from EdelWeb (Peter
7587 Sylvester and Christophe Renou) was integrated.
7588 *Christophe Renou <christophe.renou@edelweb.fr>, Peter Sylvester
7589 <peter.sylvester@edelweb.fr>, Tom Wu <tjw@cs.stanford.edu>, and
7592 * Add functions to copy EVP_PKEY_METHOD and retrieve flags and id.
7596 * Permit abbreviated handshakes when renegotiating using the function
7597 SSL_renegotiate_abbreviated().
7599 *Robin Seggelmann <seggelmann@fh-muenster.de>*
7601 * Add call to ENGINE_register_all_complete() to
7602 ENGINE_load_builtin_engines(), so some implementations get used
7603 automatically instead of needing explicit application support.
7607 * Add support for TLS key exporter as described in RFC5705.
7609 *Robin Seggelmann <seggelmann@fh-muenster.de>, Steve Henson*
7611 * Initial TLSv1.1 support. Since TLSv1.1 is very similar to TLS v1.0 only
7612 a few changes are required:
7614 Add SSL_OP_NO_TLSv1_1 flag.
7615 Add TLSv1_1 methods.
7616 Update version checking logic to handle version 1.1.
7617 Add explicit IV handling (ported from DTLS code).
7618 Add command line options to s_client/s_server.
7625 ### Changes between 1.0.0s and 1.0.0t [3 Dec 2015]
7627 * X509_ATTRIBUTE memory leak
7629 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
7630 memory. This structure is used by the PKCS#7 and CMS routines so any
7631 application which reads PKCS#7 or CMS data from untrusted sources is
7632 affected. SSL/TLS is not affected.
7634 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
7640 * Race condition handling PSK identify hint
7642 If PSK identity hints are received by a multi-threaded client then
7643 the values are wrongly updated in the parent SSL_CTX structure. This can
7644 result in a race condition potentially leading to a double free of the
7650 ### Changes between 1.0.0r and 1.0.0s [11 Jun 2015]
7652 * Malformed ECParameters causes infinite loop
7654 When processing an ECParameters structure OpenSSL enters an infinite loop
7655 if the curve specified is over a specially malformed binary polynomial
7658 This can be used to perform denial of service against any
7659 system which processes public keys, certificate requests or
7660 certificates. This includes TLS clients and TLS servers with
7661 client authentication enabled.
7663 This issue was reported to OpenSSL by Joseph Barr-Pixton.
7668 * Exploitable out-of-bounds read in X509_cmp_time
7670 X509_cmp_time does not properly check the length of the ASN1_TIME
7671 string and can read a few bytes out of bounds. In addition,
7672 X509_cmp_time accepts an arbitrary number of fractional seconds in the
7675 An attacker can use this to craft malformed certificates and CRLs of
7676 various sizes and potentially cause a segmentation fault, resulting in
7677 a DoS on applications that verify certificates or CRLs. TLS clients
7678 that verify CRLs are affected. TLS clients and servers with client
7679 authentication enabled may be affected if they use custom verification
7682 This issue was reported to OpenSSL by Robert Swiecki (Google), and
7683 independently by Hanno Böck.
7688 * PKCS7 crash with missing EnvelopedContent
7690 The PKCS#7 parsing code does not handle missing inner EncryptedContent
7691 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
7692 with missing content and trigger a NULL pointer dereference on parsing.
7694 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
7695 structures from untrusted sources are affected. OpenSSL clients and
7696 servers are not affected.
7698 This issue was reported to OpenSSL by Michal Zalewski (Google).
7703 * CMS verify infinite loop with unknown hash function
7705 When verifying a signedData message the CMS code can enter an infinite loop
7706 if presented with an unknown hash function OID. This can be used to perform
7707 denial of service against any system which verifies signedData messages using
7709 This issue was reported to OpenSSL by Johannes Bauer.
7714 * Race condition handling NewSessionTicket
7716 If a NewSessionTicket is received by a multi-threaded client when attempting to
7717 reuse a previous ticket then a race condition can occur potentially leading to
7718 a double free of the ticket data.
7723 ### Changes between 1.0.0q and 1.0.0r [19 Mar 2015]
7725 * Segmentation fault in ASN1_TYPE_cmp fix
7727 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
7728 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
7729 certificate signature algorithm consistency this can be used to crash any
7730 certificate verification operation and exploited in a DoS attack. Any
7731 application which performs certificate verification is vulnerable including
7732 OpenSSL clients and servers which enable client authentication.
7737 * ASN.1 structure reuse memory corruption fix
7739 Reusing a structure in ASN.1 parsing may allow an attacker to cause
7740 memory corruption via an invalid write. Such reuse is and has been
7741 strongly discouraged and is believed to be rare.
7743 Applications that parse structures containing CHOICE or ANY DEFINED BY
7744 components may be affected. Certificate parsing (d2i_X509 and related
7745 functions) are however not affected. OpenSSL clients and servers are
7751 * PKCS7 NULL pointer dereferences fix
7753 The PKCS#7 parsing code does not handle missing outer ContentInfo
7754 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
7755 missing content and trigger a NULL pointer dereference on parsing.
7757 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
7758 otherwise parse PKCS#7 structures from untrusted sources are
7759 affected. OpenSSL clients and servers are not affected.
7761 This issue was reported to OpenSSL by Michal Zalewski (Google).
7766 * DoS via reachable assert in SSLv2 servers fix
7768 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
7769 servers that both support SSLv2 and enable export cipher suites by sending
7770 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
7772 This issue was discovered by Sean Burford (Google) and Emilia Käsper
7773 (OpenSSL development team).
7778 * Use After Free following d2i_ECPrivatekey error fix
7780 A malformed EC private key file consumed via the d2i_ECPrivateKey function
7781 could cause a use after free condition. This, in turn, could cause a double
7782 free in several private key parsing functions (such as d2i_PrivateKey
7783 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
7784 for applications that receive EC private keys from untrusted
7785 sources. This scenario is considered rare.
7787 This issue was discovered by the BoringSSL project and fixed in their
7793 * X509_to_X509_REQ NULL pointer deref fix
7795 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
7796 the certificate key is invalid. This function is rarely used in practice.
7798 This issue was discovered by Brian Carpenter.
7803 * Removed the export ciphers from the DEFAULT ciphers
7807 ### Changes between 1.0.0p and 1.0.0q [15 Jan 2015]
7809 * Build fixes for the Windows and OpenVMS platforms
7811 *Matt Caswell and Richard Levitte*
7813 ### Changes between 1.0.0o and 1.0.0p [8 Jan 2015]
7815 * Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
7816 message can cause a segmentation fault in OpenSSL due to a NULL pointer
7817 dereference. This could lead to a Denial Of Service attack. Thanks to
7818 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
7823 * Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
7824 dtls1_buffer_record function under certain conditions. In particular this
7825 could occur if an attacker sent repeated DTLS records with the same
7826 sequence number but for the next epoch. The memory leak could be exploited
7827 by an attacker in a Denial of Service attack through memory exhaustion.
7828 Thanks to Chris Mueller for reporting this issue.
7833 * Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
7834 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
7835 method would be set to NULL which could later result in a NULL pointer
7836 dereference. Thanks to Frank Schmirler for reporting this issue.
7841 * Abort handshake if server key exchange message is omitted for ephemeral
7844 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
7845 reporting this issue.
7850 * Remove non-export ephemeral RSA code on client and server. This code
7851 violated the TLS standard by allowing the use of temporary RSA keys in
7852 non-export ciphersuites and could be used by a server to effectively
7853 downgrade the RSA key length used to a value smaller than the server
7854 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
7855 INRIA or reporting this issue.
7860 * Fixed issue where DH client certificates are accepted without verification.
7861 An OpenSSL server will accept a DH certificate for client authentication
7862 without the certificate verify message. This effectively allows a client to
7863 authenticate without the use of a private key. This only affects servers
7864 which trust a client certificate authority which issues certificates
7865 containing DH keys: these are extremely rare and hardly ever encountered.
7866 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
7872 * Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
7873 results on some platforms, including x86_64. This bug occurs at random
7874 with a very low probability, and is not known to be exploitable in any
7875 way, though its exact impact is difficult to determine. Thanks to Pieter
7876 Wuille (Blockstream) who reported this issue and also suggested an initial
7877 fix. Further analysis was conducted by the OpenSSL development team and
7878 Adam Langley of Google. The final fix was developed by Andy Polyakov of
7879 the OpenSSL core team.
7884 * Fix various certificate fingerprint issues.
7886 By using non-DER or invalid encodings outside the signed portion of a
7887 certificate the fingerprint can be changed without breaking the signature.
7888 Although no details of the signed portion of the certificate can be changed
7889 this can cause problems with some applications: e.g. those using the
7890 certificate fingerprint for blacklists.
7892 1. Reject signatures with non zero unused bits.
7894 If the BIT STRING containing the signature has non zero unused bits reject
7895 the signature. All current signature algorithms require zero unused bits.
7897 2. Check certificate algorithm consistency.
7899 Check the AlgorithmIdentifier inside TBS matches the one in the
7900 certificate signature. NB: this will result in signature failure
7901 errors for some broken certificates.
7903 Thanks to Konrad Kraszewski from Google for reporting this issue.
7905 3. Check DSA/ECDSA signatures use DER.
7907 Reencode DSA/ECDSA signatures and compare with the original received
7908 signature. Return an error if there is a mismatch.
7910 This will reject various cases including garbage after signature
7911 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
7912 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
7913 (negative or with leading zeroes).
7915 Further analysis was conducted and fixes were developed by Stephen Henson
7916 of the OpenSSL core team.
7922 ### Changes between 1.0.0n and 1.0.0o [15 Oct 2014]
7924 * Session Ticket Memory Leak.
7926 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
7927 integrity of that ticket is first verified. In the event of a session
7928 ticket integrity check failing, OpenSSL will fail to free memory
7929 causing a memory leak. By sending a large number of invalid session
7930 tickets an attacker could exploit this issue in a Denial Of Service
7936 * Build option no-ssl3 is incomplete.
7938 When OpenSSL is configured with "no-ssl3" as a build option, servers
7939 could accept and complete a SSL 3.0 handshake, and clients could be
7940 configured to send them.
7943 *Akamai and the OpenSSL team*
7945 * Add support for TLS_FALLBACK_SCSV.
7946 Client applications doing fallback retries should call
7947 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
7950 *Adam Langley, Bodo Moeller*
7952 * Add additional DigestInfo checks.
7954 Reencode DigestInto in DER and check against the original when
7955 verifying RSA signature: this will reject any improperly encoded
7956 DigestInfo structures.
7958 Note: this is a precautionary measure and no attacks are currently known.
7962 ### Changes between 1.0.0m and 1.0.0n [6 Aug 2014]
7964 * OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
7965 to a denial of service attack. A malicious server can crash the client
7966 with a null pointer dereference (read) by specifying an anonymous (EC)DH
7967 ciphersuite and sending carefully crafted handshake messages.
7969 Thanks to Felix Gröbert (Google) for discovering and researching this
7975 * By sending carefully crafted DTLS packets an attacker could cause openssl
7976 to leak memory. This can be exploited through a Denial of Service attack.
7977 Thanks to Adam Langley for discovering and researching this issue.
7982 * An attacker can force openssl to consume large amounts of memory whilst
7983 processing DTLS handshake messages. This can be exploited through a
7984 Denial of Service attack.
7985 Thanks to Adam Langley for discovering and researching this issue.
7990 * An attacker can force an error condition which causes openssl to crash
7991 whilst processing DTLS packets due to memory being freed twice. This
7992 can be exploited through a Denial of Service attack.
7993 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
7999 * If a multithreaded client connects to a malicious server using a resumed
8000 session and the server sends an ec point format extension it could write
8001 up to 255 bytes to freed memory.
8003 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
8009 * A flaw in OBJ_obj2txt may cause pretty printing functions such as
8010 X509_name_oneline, X509_name_print_ex et al. to leak some information
8011 from the stack. Applications may be affected if they echo pretty printing
8012 output to the attacker.
8014 Thanks to Ivan Fratric (Google) for discovering this issue.
8017 *Emilia Käsper, and Steve Henson*
8019 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
8020 for corner cases. (Certain input points at infinity could lead to
8021 bogus results, with non-infinity inputs mapped to infinity too.)
8025 ### Changes between 1.0.0l and 1.0.0m [5 Jun 2014]
8027 * Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
8028 handshake can force the use of weak keying material in OpenSSL
8029 SSL/TLS clients and servers.
8031 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
8032 researching this issue. ([CVE-2014-0224])
8034 *KIKUCHI Masashi, Steve Henson*
8036 * Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
8037 OpenSSL DTLS client the code can be made to recurse eventually crashing
8040 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
8043 *Imre Rad, Steve Henson*
8045 * Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
8046 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
8047 client or server. This is potentially exploitable to run arbitrary
8048 code on a vulnerable client or server.
8050 Thanks to Jüri Aedla for reporting this issue. ([CVE-2014-0195])
8052 *Jüri Aedla, Steve Henson*
8054 * Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
8055 are subject to a denial of service attack.
8057 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
8058 this issue. ([CVE-2014-3470])
8060 *Felix Gröbert, Ivan Fratric, Steve Henson*
8062 * Harmonize version and its documentation. -f flag is used to display
8065 *mancha <mancha1@zoho.com>*
8067 * Fix eckey_priv_encode so it immediately returns an error upon a failure
8068 in i2d_ECPrivateKey.
8070 *mancha <mancha1@zoho.com>*
8072 * Fix some double frees. These are not thought to be exploitable.
8074 *mancha <mancha1@zoho.com>*
8076 * Fix for the attack described in the paper "Recovering OpenSSL
8077 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
8078 by Yuval Yarom and Naomi Benger. Details can be obtained from:
8079 <http://eprint.iacr.org/2014/140>
8081 Thanks to Yuval Yarom and Naomi Benger for discovering this
8082 flaw and to Yuval Yarom for supplying a fix ([CVE-2014-0076])
8084 *Yuval Yarom and Naomi Benger*
8086 ### Changes between 1.0.0k and 1.0.0l [6 Jan 2014]
8088 * Keep original DTLS digest and encryption contexts in retransmission
8089 structures so we can use the previous session parameters if they need
8090 to be resent. ([CVE-2013-6450])
8094 * Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
8095 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
8096 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
8097 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
8098 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
8099 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
8101 *Rob Stradling, Adam Langley*
8103 ### Changes between 1.0.0j and 1.0.0k [5 Feb 2013]
8105 * Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
8107 This addresses the flaw in CBC record processing discovered by
8108 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
8109 at: <http://www.isg.rhul.ac.uk/tls/>
8111 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
8112 Security Group at Royal Holloway, University of London
8113 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
8114 Emilia Käsper for the initial patch.
8117 *Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson*
8119 * Return an error when checking OCSP signatures when key is NULL.
8120 This fixes a DoS attack. ([CVE-2013-0166])
8124 * Call OCSP Stapling callback after ciphersuite has been chosen, so
8125 the right response is stapled. Also change SSL_get_certificate()
8126 so it returns the certificate actually sent.
8127 See <http://rt.openssl.org/Ticket/Display.html?id=2836>.
8128 (This is a backport)
8130 *Rob Stradling <rob.stradling@comodo.com>*
8132 * Fix possible deadlock when decoding public keys.
8136 ### Changes between 1.0.0i and 1.0.0j [10 May 2012]
8138 [NB: OpenSSL 1.0.0i and later 1.0.0 patch levels were released after
8141 * Sanity check record length before skipping explicit IV in DTLS
8144 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
8145 fuzzing as a service testing platform.
8150 * Initialise tkeylen properly when encrypting CMS messages.
8151 Thanks to Solar Designer of Openwall for reporting this issue.
8155 ### Changes between 1.0.0h and 1.0.0i [19 Apr 2012]
8157 * Check for potentially exploitable overflows in asn1_d2i_read_bio
8158 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
8159 in CRYPTO_realloc_clean.
8161 Thanks to Tavis Ormandy, Google Security Team, for discovering this
8162 issue and to Adam Langley <agl@chromium.org> for fixing it.
8165 *Adam Langley (Google), Tavis Ormandy, Google Security Team*
8167 ### Changes between 1.0.0g and 1.0.0h [12 Mar 2012]
8169 * Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness
8170 in CMS and PKCS7 code. When RSA decryption fails use a random key for
8171 content decryption and always return the same error. Note: this attack
8172 needs on average 2^20 messages so it only affects automated senders. The
8173 old behaviour can be re-enabled in the CMS code by setting the
8174 CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where
8175 an MMA defence is not necessary.
8176 Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering
8177 this issue. ([CVE-2012-0884])
8181 * Fix CVE-2011-4619: make sure we really are receiving a
8182 client hello before rejecting multiple SGC restarts. Thanks to
8183 Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug.
8187 ### Changes between 1.0.0f and 1.0.0g [18 Jan 2012]
8189 * Fix for DTLS DoS issue introduced by fix for CVE-2011-4109.
8190 Thanks to Antonio Martin, Enterprise Secure Access Research and
8191 Development, Cisco Systems, Inc. for discovering this bug and
8192 preparing a fix. ([CVE-2012-0050])
8196 ### Changes between 1.0.0e and 1.0.0f [4 Jan 2012]
8198 * Nadhem Alfardan and Kenny Paterson have discovered an extension
8199 of the Vaudenay padding oracle attack on CBC mode encryption
8200 which enables an efficient plaintext recovery attack against
8201 the OpenSSL implementation of DTLS. Their attack exploits timing
8202 differences arising during decryption processing. A research
8203 paper describing this attack can be found at:
8204 <http://www.isg.rhul.ac.uk/~kp/dtls.pdf>
8205 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
8206 Security Group at Royal Holloway, University of London
8207 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
8208 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
8209 for preparing the fix. ([CVE-2011-4108])
8211 *Robin Seggelmann, Michael Tuexen*
8213 * Clear bytes used for block padding of SSL 3.0 records.
8216 *Adam Langley (Google)*
8218 * Only allow one SGC handshake restart for SSL/TLS. Thanks to George
8219 Kadianakis <desnacked@gmail.com> for discovering this issue and
8220 Adam Langley for preparing the fix. ([CVE-2011-4619])
8222 *Adam Langley (Google)*
8224 * Check parameters are not NULL in GOST ENGINE. ([CVE-2012-0027])
8226 *Andrey Kulikov <amdeich@gmail.com>*
8228 * Prevent malformed RFC3779 data triggering an assertion failure.
8229 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
8230 and Rob Austein <sra@hactrn.net> for fixing it. ([CVE-2011-4577])
8232 *Rob Austein <sra@hactrn.net>*
8234 * Improved PRNG seeding for VOS.
8236 *Paul Green <Paul.Green@stratus.com>*
8238 * Fix ssl_ciph.c set-up race.
8240 *Adam Langley (Google)*
8242 * Fix spurious failures in ecdsatest.c.
8244 *Emilia Käsper (Google)*
8246 * Fix the BIO_f_buffer() implementation (which was mixing different
8247 interpretations of the `..._len` fields).
8249 *Adam Langley (Google)*
8251 * Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
8252 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
8253 threads won't reuse the same blinding coefficients.
8255 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
8256 lock to call BN_BLINDING_invert_ex, and avoids one use of
8257 BN_BLINDING_update for each BN_BLINDING structure (previously,
8258 the last update always remained unused).
8260 *Emilia Käsper (Google)*
8262 * In ssl3_clear, preserve s3->init_extra along with s3->rbuf.
8264 *Bob Buckholz (Google)*
8266 ### Changes between 1.0.0d and 1.0.0e [6 Sep 2011]
8268 * Fix bug where CRLs with nextUpdate in the past are sometimes accepted
8269 by initialising X509_STORE_CTX properly. ([CVE-2011-3207])
8271 *Kaspar Brand <ossl@velox.ch>*
8273 * Fix SSL memory handling for (EC)DH ciphersuites, in particular
8274 for multi-threaded use of ECDH. ([CVE-2011-3210])
8276 *Adam Langley (Google)*
8278 * Fix x509_name_ex_d2i memory leak on bad inputs.
8282 * Remove hard coded ecdsaWithSHA1 signature tests in ssl code and check
8283 signature public key algorithm by using OID xref utilities instead.
8284 Before this you could only use some ECC ciphersuites with SHA1 only.
8288 * Add protection against ECDSA timing attacks as mentioned in the paper
8289 by Billy Bob Brumley and Nicola Tuveri, see:
8290 <http://eprint.iacr.org/2011/232.pdf>
8292 *Billy Bob Brumley and Nicola Tuveri*
8294 ### Changes between 1.0.0c and 1.0.0d [8 Feb 2011]
8296 * Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
8298 *Neel Mehta, Adam Langley, Bodo Moeller (Google)*
8300 * Fix bug in string printing code: if *any* escaping is enabled we must
8301 escape the escape character (backslash) or the resulting string is
8306 ### Changes between 1.0.0b and 1.0.0c [2 Dec 2010]
8308 * Disable code workaround for ancient and obsolete Netscape browsers
8309 and servers: an attacker can use it in a ciphersuite downgrade attack.
8310 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
8314 * Fixed J-PAKE implementation error, originally discovered by
8315 Sebastien Martini, further info and confirmation from Stefan
8316 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
8320 ### Changes between 1.0.0a and 1.0.0b [16 Nov 2010]
8322 * Fix extension code to avoid race conditions which can result in a buffer
8323 overrun vulnerability: resumed sessions must not be modified as they can
8324 be shared by multiple threads. CVE-2010-3864
8328 * Fix WIN32 build system to correctly link an ENGINE directory into
8333 ### Changes between 1.0.0 and 1.0.0a [01 Jun 2010]
8335 * Check return value of int_rsa_verify in pkey_rsa_verifyrecover
8338 *Steve Henson, Peter-Michael Hager <hager@dortmund.net>*
8340 ### Changes between 0.9.8n and 1.0.0 [29 Mar 2010]
8342 * Add "missing" function EVP_CIPHER_CTX_copy(). This copies a cipher
8343 context. The operation can be customised via the ctrl mechanism in
8344 case ENGINEs want to include additional functionality.
8348 * Tolerate yet another broken PKCS#8 key format: private key value negative.
8352 * Add new -subject_hash_old and -issuer_hash_old options to x509 utility to
8353 output hashes compatible with older versions of OpenSSL.
8355 *Willy Weisz <weisz@vcpc.univie.ac.at>*
8357 * Fix compression algorithm handling: if resuming a session use the
8358 compression algorithm of the resumed session instead of determining
8359 it from client hello again. Don't allow server to change algorithm.
8363 * Add load_crls() function to commands tidying load_certs() too. Add option
8364 to verify utility to allow additional CRLs to be included.
8368 * Update OCSP request code to permit adding custom headers to the request:
8369 some responders need this.
8373 * The function EVP_PKEY_sign() returns <=0 on error: check return code
8376 *Julia Lawall <julia@diku.dk>*
8378 * Update verify callback code in `apps/s_cb.c` and `apps/verify.c`, it
8379 needlessly dereferenced structures, used obsolete functions and
8380 didn't handle all updated verify codes correctly.
8384 * Disable MD2 in the default configuration.
8388 * In BIO_pop() and BIO_push() use the ctrl argument (which was NULL) to
8389 indicate the initial BIO being pushed or popped. This makes it possible
8390 to determine whether the BIO is the one explicitly called or as a result
8391 of the ctrl being passed down the chain. Fix BIO_pop() and SSL BIOs so
8392 it handles reference counts correctly and doesn't zero out the I/O bio
8393 when it is not being explicitly popped. WARNING: applications which
8394 included workarounds for the old buggy behaviour will need to be modified
8395 or they could free up already freed BIOs.
8399 * Extend the uni2asc/asc2uni => OPENSSL_uni2asc/OPENSSL_asc2uni
8400 renaming to all platforms (within the 0.9.8 branch, this was
8401 done conditionally on Netware platforms to avoid a name clash).
8403 *Guenter <lists@gknw.net>*
8405 * Add ECDHE and PSK support to DTLS.
8407 *Michael Tuexen <tuexen@fh-muenster.de>*
8409 * Add CHECKED_STACK_OF macro to safestack.h, otherwise safestack can't
8414 * Add "missing" function EVP_MD_flags() (without this the only way to
8415 retrieve a digest flags is by accessing the structure directly. Update
8416 `EVP_MD_do_all*()` and `EVP_CIPHER_do_all*()` to include the name a digest
8417 or cipher is registered as in the "from" argument. Print out all
8418 registered digests in the dgst usage message instead of manually
8419 attempting to work them out.
8423 * If no SSLv2 ciphers are used don't use an SSLv2 compatible client hello:
8424 this allows the use of compression and extensions. Change default cipher
8425 string to remove SSLv2 ciphersuites. This effectively avoids ancient SSLv2
8426 by default unless an application cipher string requests it.
8430 * Alter match criteria in PKCS12_parse(). It used to try to use local
8431 key ids to find matching certificates and keys but some PKCS#12 files
8432 don't follow the (somewhat unwritten) rules and this strategy fails.
8433 Now just gather all certificates together and the first private key
8434 then look for the first certificate that matches the key.
8438 * Support use of registered digest and cipher names for dgst and cipher
8439 commands instead of having to add each one as a special case. So now
8446 openssl dgst -sha256 foo
8448 and this works for ENGINE based algorithms too.
8452 * Update Gost ENGINE to support parameter files.
8454 *Victor B. Wagner <vitus@cryptocom.ru>*
8456 * Support GeneralizedTime in ca utility.
8458 *Oliver Martin <oliver@volatilevoid.net>, Steve Henson*
8460 * Enhance the hash format used for certificate directory links. The new
8461 form uses the canonical encoding (meaning equivalent names will work
8462 even if they aren't identical) and uses SHA1 instead of MD5. This form
8463 is incompatible with the older format and as a result c_rehash should
8464 be used to rebuild symbolic links.
8468 * Make PKCS#8 the default write format for private keys, replacing the
8469 traditional format. This form is standardised, more secure and doesn't
8470 include an implicit MD5 dependency.
8474 * Add a $gcc_devteam_warn option to Configure. The idea is that any code
8475 committed to OpenSSL should pass this lot as a minimum.
8479 * Add session ticket override functionality for use by EAP-FAST.
8481 *Jouni Malinen <j@w1.fi>*
8483 * Modify HMAC functions to return a value. Since these can be implemented
8484 in an ENGINE errors can occur.
8488 * Type-checked OBJ_bsearch_ex.
8492 * Type-checked OBJ_bsearch. Also some constification necessitated
8493 by type-checking. Still to come: TXT_DB, bsearch(?),
8494 OBJ_bsearch_ex, qsort, CRYPTO_EX_DATA, ASN1_VALUE, ASN1_STRING,
8499 * New function OPENSSL_gmtime_adj() to add a specific number of days and
8500 seconds to a tm structure directly, instead of going through OS
8501 specific date routines. This avoids any issues with OS routines such
8502 as the year 2038 bug. New `*_adj()` functions for ASN1 time structures
8503 and X509_time_adj_ex() to cover the extended range. The existing
8504 X509_time_adj() is still usable and will no longer have any date issues.
8508 * Delta CRL support. New use deltas option which will attempt to locate
8509 and search any appropriate delta CRLs available.
8511 This work was sponsored by Google.
8515 * Support for CRLs partitioned by reason code. Reorganise CRL processing
8516 code and add additional score elements. Validate alternate CRL paths
8517 as part of the CRL checking and indicate a new error "CRL path validation
8518 error" in this case. Applications wanting additional details can use
8519 the verify callback and check the new "parent" field. If this is not
8520 NULL CRL path validation is taking place. Existing applications won't
8521 see this because it requires extended CRL support which is off by
8524 This work was sponsored by Google.
8528 * Support for freshest CRL extension.
8530 This work was sponsored by Google.
8534 * Initial indirect CRL support. Currently only supported in the CRLs
8535 passed directly and not via lookup. Process certificate issuer
8536 CRL entry extension and lookup CRL entries by bother issuer name
8537 and serial number. Check and process CRL issuer entry in IDP extension.
8539 This work was sponsored by Google.
8543 * Add support for distinct certificate and CRL paths. The CRL issuer
8544 certificate is validated separately in this case. Only enabled if
8545 an extended CRL support flag is set: this flag will enable additional
8546 CRL functionality in future.
8548 This work was sponsored by Google.
8552 * Add support for policy mappings extension.
8554 This work was sponsored by Google.
8558 * Fixes to pathlength constraint, self issued certificate handling,
8559 policy processing to align with RFC3280 and PKITS tests.
8561 This work was sponsored by Google.
8565 * Support for name constraints certificate extension. DN, email, DNS
8566 and URI types are currently supported.
8568 This work was sponsored by Google.
8572 * To cater for systems that provide a pointer-based thread ID rather
8573 than numeric, deprecate the current numeric thread ID mechanism and
8574 replace it with a structure and associated callback type. This
8575 mechanism allows a numeric "hash" to be extracted from a thread ID in
8576 either case, and on platforms where pointers are larger than 'long',
8577 mixing is done to help ensure the numeric 'hash' is usable even if it
8578 can't be guaranteed unique. The default mechanism is to use "&errno"
8579 as a pointer-based thread ID to distinguish between threads.
8581 Applications that want to provide their own thread IDs should now use
8582 CRYPTO_THREADID_set_callback() to register a callback that will call
8583 either CRYPTO_THREADID_set_numeric() or CRYPTO_THREADID_set_pointer().
8585 Note that ERR_remove_state() is now deprecated, because it is tied
8586 to the assumption that thread IDs are numeric. ERR_remove_state(0)
8587 to free the current thread's error state should be replaced by
8588 ERR_remove_thread_state(NULL).
8590 (This new approach replaces the functions CRYPTO_set_idptr_callback(),
8591 CRYPTO_get_idptr_callback(), and CRYPTO_thread_idptr() that existed in
8592 OpenSSL 0.9.9-dev between June 2006 and August 2008. Also, if an
8593 application was previously providing a numeric thread callback that
8594 was inappropriate for distinguishing threads, then uniqueness might
8595 have been obtained with &errno that happened immediately in the
8596 intermediate development versions of OpenSSL; this is no longer the
8597 case, the numeric thread callback will now override the automatic use
8600 *Geoff Thorpe, with help from Bodo Moeller*
8602 * Initial support for different CRL issuing certificates. This covers a
8603 simple case where the self issued certificates in the chain exist and
8604 the real CRL issuer is higher in the existing chain.
8606 This work was sponsored by Google.
8610 * Removed effectively defunct crypto/store from the build.
8614 * Revamp of STACK to provide stronger type-checking. Still to come:
8615 TXT_DB, bsearch(?), OBJ_bsearch, qsort, CRYPTO_EX_DATA, ASN1_VALUE,
8616 ASN1_STRING, CONF_VALUE.
8620 * Add a new SSL_MODE_RELEASE_BUFFERS mode flag to release unused buffer
8621 RAM on SSL connections. This option can save about 34k per idle SSL.
8625 * Revamp of LHASH to provide stronger type-checking. Still to come:
8626 STACK, TXT_DB, bsearch, qsort.
8630 * Initial support for Cryptographic Message Syntax (aka CMS) based
8631 on RFC3850, RFC3851 and RFC3852. New cms directory and cms utility,
8632 support for data, signedData, compressedData, digestedData and
8633 encryptedData, envelopedData types included. Scripts to check against
8634 RFC4134 examples draft and interop and consistency checks of many
8635 content types and variants.
8639 * Add options to enc utility to support use of zlib compression BIO.
8643 * Extend mk1mf to support importing of options and assembly language
8644 files from Configure script, currently only included in VC-WIN32.
8645 The assembly language rules can now optionally generate the source
8646 files from the associated perl scripts.
8650 * Implement remaining functionality needed to support GOST ciphersuites.
8651 Interop testing has been performed using CryptoPro implementations.
8653 *Victor B. Wagner <vitus@cryptocom.ru>*
8655 * s390x assembler pack.
8659 * ARMv4 assembler pack. ARMv4 refers to v4 and later ISA, not CPU
8664 * Implement Opaque PRF Input TLS extension as specified in
8665 draft-rescorla-tls-opaque-prf-input-00.txt. Since this is not an
8666 official specification yet and no extension type assignment by
8667 IANA exists, this extension (for now) will have to be explicitly
8668 enabled when building OpenSSL by providing the extension number
8669 to use. For example, specify an option
8671 -DTLSEXT_TYPE_opaque_prf_input=0x9527
8673 to the "config" or "Configure" script to enable the extension,
8674 assuming extension number 0x9527 (which is a completely arbitrary
8675 and unofficial assignment based on the MD5 hash of the Internet
8676 Draft). Note that by doing so, you potentially lose
8677 interoperability with other TLS implementations since these might
8678 be using the same extension number for other purposes.
8680 SSL_set_tlsext_opaque_prf_input(ssl, src, len) is used to set the
8681 opaque PRF input value to use in the handshake. This will create
8682 an internal copy of the length-'len' string at 'src', and will
8683 return non-zero for success.
8685 To get more control and flexibility, provide a callback function
8688 SSL_CTX_set_tlsext_opaque_prf_input_callback(ctx, cb)
8689 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg(ctx, arg)
8693 int (*cb)(SSL *, void *peerinput, size_t len, void *arg);
8696 Callback function 'cb' will be called in handshakes, and is
8697 expected to use SSL_set_tlsext_opaque_prf_input() as appropriate.
8698 Argument 'arg' is for application purposes (the value as given to
8699 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg() will directly
8700 be provided to the callback function). The callback function
8701 has to return non-zero to report success: usually 1 to use opaque
8702 PRF input just if possible, or 2 to enforce use of the opaque PRF
8703 input. In the latter case, the library will abort the handshake
8704 if opaque PRF input is not successfully negotiated.
8706 Arguments 'peerinput' and 'len' given to the callback function
8707 will always be NULL and 0 in the case of a client. A server will
8708 see the client's opaque PRF input through these variables if
8709 available (NULL and 0 otherwise). Note that if the server
8710 provides an opaque PRF input, the length must be the same as the
8711 length of the client's opaque PRF input.
8713 Note that the callback function will only be called when creating
8714 a new session (session resumption can resume whatever was
8715 previously negotiated), and will not be called in SSL 2.0
8716 handshakes; thus, SSL_CTX_set_options(ctx, SSL_OP_NO_SSLv2) or
8717 SSL_set_options(ssl, SSL_OP_NO_SSLv2) is especially recommended
8718 for applications that need to enforce opaque PRF input.
8722 * Update ssl code to support digests other than SHA1+MD5 for handshake
8725 *Victor B. Wagner <vitus@cryptocom.ru>*
8727 * Add RFC4507 support to OpenSSL. This includes the corrections in
8728 RFC4507bis. The encrypted ticket format is an encrypted encoded
8729 SSL_SESSION structure, that way new session features are automatically
8732 If a client application caches session in an SSL_SESSION structure
8733 support is transparent because tickets are now stored in the encoded
8736 The SSL_CTX structure automatically generates keys for ticket
8737 protection in servers so again support should be possible
8738 with no application modification.
8740 If a client or server wishes to disable RFC4507 support then the option
8741 SSL_OP_NO_TICKET can be set.
8743 Add a TLS extension debugging callback to allow the contents of any client
8744 or server extensions to be examined.
8746 This work was sponsored by Google.
8750 * Final changes to avoid use of pointer pointer casts in OpenSSL.
8751 OpenSSL should now compile cleanly on gcc 4.2
8753 *Peter Hartley <pdh@utter.chaos.org.uk>, Steve Henson*
8755 * Update SSL library to use new EVP_PKEY MAC API. Include generic MAC
8756 support including streaming MAC support: this is required for GOST
8757 ciphersuite support.
8759 *Victor B. Wagner <vitus@cryptocom.ru>, Steve Henson*
8761 * Add option -stream to use PKCS#7 streaming in smime utility. New
8762 function i2d_PKCS7_bio_stream() and PEM_write_PKCS7_bio_stream()
8763 to output in BER and PEM format.
8767 * Experimental support for use of HMAC via EVP_PKEY interface. This
8768 allows HMAC to be handled via the `EVP_DigestSign*()` interface. The
8769 EVP_PKEY "key" in this case is the HMAC key, potentially allowing
8770 ENGINE support for HMAC keys which are unextractable. New -mac and
8771 -macopt options to dgst utility.
8775 * New option -sigopt to dgst utility. Update dgst to use
8776 `EVP_Digest{Sign,Verify}*`. These two changes make it possible to use
8777 alternative signing parameters such as X9.31 or PSS in the dgst
8782 * Change ssl_cipher_apply_rule(), the internal function that does
8783 the work each time a ciphersuite string requests enabling
8784 ("foo+bar"), moving ("+foo+bar"), disabling ("-foo+bar", or
8785 removing ("!foo+bar") a class of ciphersuites: Now it maintains
8786 the order of disabled ciphersuites such that those ciphersuites
8787 that most recently went from enabled to disabled not only stay
8788 in order with respect to each other, but also have higher priority
8789 than other disabled ciphersuites the next time ciphersuites are
8792 This means that you can now say, e.g., "PSK:-PSK:HIGH" to enable
8793 the same ciphersuites as with "HIGH" alone, but in a specific
8794 order where the PSK ciphersuites come first (since they are the
8795 most recently disabled ciphersuites when "HIGH" is parsed).
8797 Also, change ssl_create_cipher_list() (using this new
8798 functionality) such that between otherwise identical
8799 ciphersuites, ephemeral ECDH is preferred over ephemeral DH in
8804 * Change ssl_create_cipher_list() so that it automatically
8805 arranges the ciphersuites in reasonable order before starting
8806 to process the rule string. Thus, the definition for "DEFAULT"
8807 (SSL_DEFAULT_CIPHER_LIST) now is just "ALL:!aNULL:!eNULL", but
8808 remains equivalent to `"AES:ALL:!aNULL:!eNULL:+aECDH:+kRSA:+RC4:@STRENGTH"`.
8809 This makes it much easier to arrive at a reasonable default order
8810 in applications for which anonymous ciphers are OK (meaning
8811 that you can't actually use DEFAULT).
8813 *Bodo Moeller; suggested by Victor Duchovni*
8815 * Split the SSL/TLS algorithm mask (as used for ciphersuite string
8816 processing) into multiple integers instead of setting
8817 "SSL_MKEY_MASK" bits, "SSL_AUTH_MASK" bits, "SSL_ENC_MASK",
8818 "SSL_MAC_MASK", and "SSL_SSL_MASK" bits all in a single integer.
8819 (These masks as well as the individual bit definitions are hidden
8820 away into the non-exported interface ssl/ssl_locl.h, so this
8821 change to the definition of the SSL_CIPHER structure shouldn't
8822 affect applications.) This give us more bits for each of these
8823 categories, so there is no longer a need to coagulate AES128 and
8824 AES256 into a single algorithm bit, and to coagulate Camellia128
8825 and Camellia256 into a single algorithm bit, which has led to all
8828 Thus, among other things, the kludge introduced in 0.9.7m and
8829 0.9.8e for masking out AES256 independently of AES128 or masking
8830 out Camellia256 independently of AES256 is not needed here in 0.9.9.
8832 With the change, we also introduce new ciphersuite aliases that
8833 so far were missing: "AES128", "AES256", "CAMELLIA128", and
8838 * Add support for dsa-with-SHA224 and dsa-with-SHA256.
8839 Use the leftmost N bytes of the signature input if the input is
8840 larger than the prime q (with N being the size in bytes of q).
8844 * Very *very* experimental PKCS#7 streaming encoder support. Nothing uses
8845 it yet and it is largely untested.
8849 * Add support for the ecdsa-with-SHA224/256/384/512 signature types.
8853 * Initial incomplete changes to avoid need for function casts in OpenSSL
8854 some compilers (gcc 4.2 and later) reject their use. Safestack is
8855 reimplemented. Update ASN1 to avoid use of legacy functions.
8859 * Win32/64 targets are linked with Winsock2.
8863 * Add an X509_CRL_METHOD structure to allow CRL processing to be redirected
8864 to external functions. This can be used to increase CRL handling
8865 efficiency especially when CRLs are very large by (for example) storing
8866 the CRL revoked certificates in a database.
8870 * Overhaul of by_dir code. Add support for dynamic loading of CRLs so
8871 new CRLs added to a directory can be used. New command line option
8872 -verify_return_error to s_client and s_server. This causes real errors
8873 to be returned by the verify callback instead of carrying on no matter
8874 what. This reflects the way a "real world" verify callback would behave.
8878 * GOST engine, supporting several GOST algorithms and public key formats.
8879 Kindly donated by Cryptocom.
8883 * Partial support for Issuing Distribution Point CRL extension. CRLs
8884 partitioned by DP are handled but no indirect CRL or reason partitioning
8885 (yet). Complete overhaul of CRL handling: now the most suitable CRL is
8886 selected via a scoring technique which handles IDP and AKID in CRLs.
8890 * New X509_STORE_CTX callbacks lookup_crls() and lookup_certs() which
8891 will ultimately be used for all verify operations: this will remove the
8892 X509_STORE dependency on certificate verification and allow alternative
8893 lookup methods. X509_STORE based implementations of these two callbacks.
8897 * Allow multiple CRLs to exist in an X509_STORE with matching issuer names.
8898 Modify get_crl() to find a valid (unexpired) CRL if possible.
8902 * New function X509_CRL_match() to check if two CRLs are identical. Normally
8903 this would be called X509_CRL_cmp() but that name is already used by
8904 a function that just compares CRL issuer names. Cache several CRL
8905 extensions in X509_CRL structure and cache CRLDP in X509.
8909 * Store a "canonical" representation of X509_NAME structure (ASN1 Name)
8910 this maps equivalent X509_NAME structures into a consistent structure.
8911 Name comparison can then be performed rapidly using memcmp().
8915 * Non-blocking OCSP request processing. Add -timeout option to ocsp
8920 * Allow digests to supply their own micalg string for S/MIME type using
8921 the ctrl EVP_MD_CTRL_MICALG.
8925 * During PKCS7 signing pass the PKCS7 SignerInfo structure to the
8926 EVP_PKEY_METHOD before and after signing via the EVP_PKEY_CTRL_PKCS7_SIGN
8927 ctrl. It can then customise the structure before and/or after signing
8932 * New function OBJ_add_sigid() to allow application defined signature OIDs
8933 to be added to OpenSSLs internal tables. New function OBJ_sigid_free()
8934 to free up any added signature OIDs.
8938 * New functions EVP_CIPHER_do_all(), EVP_CIPHER_do_all_sorted(),
8939 EVP_MD_do_all() and EVP_MD_do_all_sorted() to enumerate internal
8940 digest and cipher tables. New options added to openssl utility:
8941 list-message-digest-algorithms and list-cipher-algorithms.
8945 * Change the array representation of binary polynomials: the list
8946 of degrees of non-zero coefficients is now terminated with -1.
8947 Previously it was terminated with 0, which was also part of the
8948 value; thus, the array representation was not applicable to
8949 polynomials where t^0 has coefficient zero. This change makes
8950 the array representation useful in a more general context.
8954 * Various modifications and fixes to SSL/TLS cipher string
8955 handling. For ECC, the code now distinguishes between fixed ECDH
8956 with RSA certificates on the one hand and with ECDSA certificates
8957 on the other hand, since these are separate ciphersuites. The
8958 unused code for Fortezza ciphersuites has been removed.
8960 For consistency with EDH, ephemeral ECDH is now called "EECDH"
8961 (not "ECDHE"). For consistency with the code for DH
8962 certificates, use of ECDH certificates is now considered ECDH
8963 authentication, not RSA or ECDSA authentication (the latter is
8964 merely the CA's signing algorithm and not actively used in the
8967 The temporary ciphersuite alias "ECCdraft" is no longer
8968 available, and ECC ciphersuites are no longer excluded from "ALL"
8969 and "DEFAULT". The following aliases now exist for RFC 4492
8970 ciphersuites, most of these by analogy with the DH case:
8972 kECDHr - ECDH cert, signed with RSA
8973 kECDHe - ECDH cert, signed with ECDSA
8974 kECDH - ECDH cert (signed with either RSA or ECDSA)
8975 kEECDH - ephemeral ECDH
8976 ECDH - ECDH cert or ephemeral ECDH
8982 AECDH - anonymous ECDH
8983 EECDH - non-anonymous ephemeral ECDH (equivalent to "kEECDH:-AECDH")
8987 * Add additional S/MIME capabilities for AES and GOST ciphers if supported.
8988 Use correct micalg parameters depending on digest(s) in signed message.
8992 * Add engine support for EVP_PKEY_ASN1_METHOD. Add functions to process
8993 an ENGINE asn1 method. Support ENGINE lookups in the ASN1 code.
8997 * Initial engine support for EVP_PKEY_METHOD. New functions to permit
8998 an engine to register a method. Add ENGINE lookups for methods and
8999 functional reference processing.
9003 * New functions `EVP_Digest{Sign,Verify)*`. These are enhanced versions of
9004 `EVP_{Sign,Verify}*` which allow an application to customise the signature
9009 * New -resign option to smime utility. This adds one or more signers
9010 to an existing PKCS#7 signedData structure. Also -md option to use an
9011 alternative message digest algorithm for signing.
9015 * Tidy up PKCS#7 routines and add new functions to make it easier to
9016 create PKCS7 structures containing multiple signers. Update smime
9017 application to support multiple signers.
9021 * New -macalg option to pkcs12 utility to allow setting of an alternative
9026 * Initial support for PKCS#5 v2.0 PRFs other than default SHA1 HMAC.
9027 Reorganize PBE internals to lookup from a static table using NIDs,
9028 add support for HMAC PBE OID translation. Add a EVP_CIPHER ctrl:
9029 EVP_CTRL_PBE_PRF_NID this allows a cipher to specify an alternative
9030 PRF which will be automatically used with PBES2.
9034 * Replace the algorithm specific calls to generate keys in "req" with the
9039 * Update PKCS#7 enveloped data routines to use new API. This is now
9040 supported by any public key method supporting the encrypt operation. A
9041 ctrl is added to allow the public key algorithm to examine or modify
9042 the PKCS#7 RecipientInfo structure if it needs to: for RSA this is
9047 * Add a ctrl to asn1 method to allow a public key algorithm to express
9048 a default digest type to use. In most cases this will be SHA1 but some
9049 algorithms (such as GOST) need to specify an alternative digest. The
9050 return value indicates how strong the preference is 1 means optional and
9051 2 is mandatory (that is it is the only supported type). Modify
9052 ASN1_item_sign() to accept a NULL digest argument to indicate it should
9053 use the default md. Update openssl utilities to use the default digest
9054 type for signing if it is not explicitly indicated.
9058 * Use OID cross reference table in ASN1_sign() and ASN1_verify(). New
9059 EVP_MD flag EVP_MD_FLAG_PKEY_METHOD_SIGNATURE. This uses the relevant
9060 signing method from the key type. This effectively removes the link
9061 between digests and public key types.
9065 * Add an OID cross reference table and utility functions. Its purpose is to
9066 translate between signature OIDs such as SHA1WithrsaEncryption and SHA1,
9067 rsaEncryption. This will allow some of the algorithm specific hackery
9068 needed to use the correct OID to be removed.
9072 * Remove algorithm specific dependencies when setting PKCS7_SIGNER_INFO
9073 structures for PKCS7_sign(). They are now set up by the relevant public
9078 * Add provisional EC pkey method with support for ECDSA and ECDH.
9082 * Add support for key derivation (agreement) in the API, DH method and
9087 * Add DSA pkey method and DH pkey methods, extend DH ASN1 method to support
9088 public and private key formats. As a side effect these add additional
9089 command line functionality not previously available: DSA signatures can be
9090 generated and verified using pkeyutl and DH key support and generation in
9097 *Oliver Tappe <zooey@hirschkaefer.de>*
9099 * New make target "install_html_docs" installs HTML renditions of the
9102 *Oliver Tappe <zooey@hirschkaefer.de>*
9104 * New utility "genpkey" this is analogous to "genrsa" etc except it can
9105 generate keys for any algorithm. Extend and update EVP_PKEY_METHOD to
9106 support key and parameter generation and add initial key generation
9107 functionality for RSA.
9111 * Add functions for main EVP_PKEY_method operations. The undocumented
9112 functions `EVP_PKEY_{encrypt,decrypt}` have been renamed to
9113 `EVP_PKEY_{encrypt,decrypt}_old`.
9117 * Initial definitions for EVP_PKEY_METHOD. This will be a high level public
9118 key API, doesn't do much yet.
9122 * New function EVP_PKEY_asn1_get0_info() to retrieve information about
9123 public key algorithms. New option to openssl utility:
9124 "list-public-key-algorithms" to print out info.
9128 * Implement the Supported Elliptic Curves Extension for
9129 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
9133 * Don't free up OIDs in OBJ_cleanup() if they are in use by EVP_MD or
9134 EVP_CIPHER structures to avoid later problems in EVP_cleanup().
9138 * New utilities pkey and pkeyparam. These are similar to algorithm specific
9139 utilities such as rsa, dsa, dsaparam etc except they process any key
9144 * Transfer public key printing routines to EVP_PKEY_ASN1_METHOD. New
9145 functions EVP_PKEY_print_public(), EVP_PKEY_print_private(),
9146 EVP_PKEY_print_param() to print public key data from an EVP_PKEY
9151 * Initial support for pluggable public key ASN1.
9152 De-spaghettify the public key ASN1 handling. Move public and private
9153 key ASN1 handling to a new EVP_PKEY_ASN1_METHOD structure. Relocate
9154 algorithm specific handling to a single module within the relevant
9155 algorithm directory. Add functions to allow (near) opaque processing
9156 of public and private key structures.
9160 * Implement the Supported Point Formats Extension for
9161 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
9165 * Add initial support for RFC 4279 PSK TLS ciphersuites. Add members
9166 for the psk identity [hint] and the psk callback functions to the
9167 SSL_SESSION, SSL and SSL_CTX structure.
9170 PSK-RC4-SHA, PSK-3DES-EDE-CBC-SHA, PSK-AES128-CBC-SHA,
9174 SSL_CTX_use_psk_identity_hint
9175 SSL_get_psk_identity_hint
9176 SSL_get_psk_identity
9177 SSL_use_psk_identity_hint
9179 *Mika Kousa and Pasi Eronen of Nokia Corporation*
9181 * Add RFC 3161 compliant time stamp request creation, response generation
9182 and response verification functionality.
9184 *Zoltán Glózik <zglozik@opentsa.org>, The OpenTSA Project*
9186 * Add initial support for TLS extensions, specifically for the server_name
9187 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
9188 have new members for a host name. The SSL data structure has an
9189 additional member `SSL_CTX *initial_ctx` so that new sessions can be
9190 stored in that context to allow for session resumption, even after the
9191 SSL has been switched to a new SSL_CTX in reaction to a client's
9192 server_name extension.
9194 New functions (subject to change):
9196 SSL_get_servername()
9197 SSL_get_servername_type()
9200 New CTRL codes and macros (subject to change):
9202 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
9203 - SSL_CTX_set_tlsext_servername_callback()
9204 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
9205 - SSL_CTX_set_tlsext_servername_arg()
9206 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
9208 openssl s_client has a new '-servername ...' option.
9210 openssl s_server has new options '-servername_host ...', '-cert2 ...',
9211 '-key2 ...', '-servername_fatal' (subject to change). This allows
9212 testing the HostName extension for a specific single host name ('-cert'
9213 and '-key' remain fallbacks for handshakes without HostName
9214 negotiation). If the unrecognized_name alert has to be sent, this by
9215 default is a warning; it becomes fatal with the '-servername_fatal'
9218 *Peter Sylvester, Remy Allais, Christophe Renou*
9220 * Whirlpool hash implementation is added.
9224 * BIGNUM code on 64-bit SPARCv9 targets is switched from bn(64,64) to
9225 bn(64,32). Because of instruction set limitations it doesn't have
9226 any negative impact on performance. This was done mostly in order
9227 to make it possible to share assembler modules, such as bn_mul_mont
9228 implementations, between 32- and 64-bit builds without hassle.
9232 * Move code previously exiled into file crypto/ec/ec2_smpt.c
9233 to ec2_smpl.c, and no longer require the OPENSSL_EC_BIN_PT_COMP
9238 * New candidate for BIGNUM assembler implementation, bn_mul_mont,
9239 dedicated Montgomery multiplication procedure, is introduced.
9240 BN_MONT_CTX is modified to allow bn_mul_mont to reach for higher
9241 "64-bit" performance on certain 32-bit targets.
9245 * New option SSL_OP_NO_COMP to disable use of compression selectively
9246 in SSL structures. New SSL ctrl to set maximum send fragment size.
9247 Save memory by setting the I/O buffer sizes dynamically instead of
9248 using the maximum available value.
9252 * New option -V for 'openssl ciphers'. This prints the ciphersuite code
9253 in addition to the text details.
9257 * Very, very preliminary EXPERIMENTAL support for printing of general
9258 ASN1 structures. This currently produces rather ugly output and doesn't
9259 handle several customised structures at all.
9263 * Integrated support for PVK file format and some related formats such
9264 as MS PUBLICKEYBLOB and PRIVATEKEYBLOB. Command line switches to support
9265 these in the 'rsa' and 'dsa' utilities.
9269 * Support for PKCS#1 RSAPublicKey format on rsa utility command line.
9273 * Remove the ancient ASN1_METHOD code. This was only ever used in one
9274 place for the (very old) "NETSCAPE" format certificates which are now
9275 handled using new ASN1 code equivalents.
9279 * Let the TLSv1_method() etc. functions return a 'const' SSL_METHOD
9280 pointer and make the SSL_METHOD parameter in SSL_CTX_new,
9281 SSL_CTX_set_ssl_version and SSL_set_ssl_method 'const'.
9285 * Modify CRL distribution points extension code to print out previously
9286 unsupported fields. Enhance extension setting code to allow setting of
9291 * Add print and set support for Issuing Distribution Point CRL extension.
9295 * Change 'Configure' script to enable Camellia by default.
9302 ### Changes between 0.9.8m and 0.9.8n [24 Mar 2010]
9304 * When rejecting SSL/TLS records due to an incorrect version number, never
9305 update s->server with a new major version number. As of
9306 - OpenSSL 0.9.8m if 'short' is a 16-bit type,
9307 - OpenSSL 0.9.8f if 'short' is longer than 16 bits,
9308 the previous behavior could result in a read attempt at NULL when
9309 receiving specific incorrect SSL/TLS records once record payload
9310 protection is active. ([CVE-2010-0740])
9312 *Bodo Moeller, Adam Langley <agl@chromium.org>*
9314 * Fix for CVE-2010-0433 where some kerberos enabled versions of OpenSSL
9315 could be crashed if the relevant tables were not present (e.g. chrooted).
9317 *Tomas Hoger <thoger@redhat.com>*
9319 ### Changes between 0.9.8l and 0.9.8m [25 Feb 2010]
9321 * Always check bn_wexpand() return values for failure. ([CVE-2009-3245])
9323 *Martin Olsson, Neel Mehta*
9325 * Fix X509_STORE locking: Every 'objs' access requires a lock (to
9326 accommodate for stack sorting, always a write lock!).
9330 * On some versions of WIN32 Heap32Next is very slow. This can cause
9331 excessive delays in the RAND_poll(): over a minute. As a workaround
9332 include a time check in the inner Heap32Next loop too.
9336 * The code that handled flushing of data in SSL/TLS originally used the
9337 BIO_CTRL_INFO ctrl to see if any data was pending first. This caused
9338 the problem outlined in PR#1949. The fix suggested there however can
9339 trigger problems with buggy BIO_CTRL_WPENDING (e.g. some versions
9340 of Apache). So instead simplify the code to flush unconditionally.
9341 This should be fine since flushing with no data to flush is a no op.
9345 * Handle TLS versions 2.0 and later properly and correctly use the
9346 highest version of TLS/SSL supported. Although TLS >= 2.0 is some way
9347 off ancient servers have a habit of sticking around for a while...
9351 * Modify compression code so it frees up structures without using the
9352 ex_data callbacks. This works around a problem where some applications
9353 call CRYPTO_cleanup_all_ex_data() before application exit (e.g. when
9354 restarting) then use compression (e.g. SSL with compression) later.
9355 This results in significant per-connection memory leaks and
9356 has caused some security issues including CVE-2008-1678 and
9361 * Constify crypto/cast (i.e., <openssl/cast.h>): a CAST_KEY doesn't
9362 change when encrypting or decrypting.
9366 * Add option SSL_OP_LEGACY_SERVER_CONNECT which will allow clients to
9367 connect and renegotiate with servers which do not support RI.
9368 Until RI is more widely deployed this option is enabled by default.
9372 * Add "missing" ssl ctrls to clear options and mode.
9376 * If client attempts to renegotiate and doesn't support RI respond with
9377 a no_renegotiation alert as required by RFC5746. Some renegotiating
9378 TLS clients will continue a connection gracefully when they receive
9379 the alert. Unfortunately OpenSSL mishandled this alert and would hang
9380 waiting for a server hello which it will never receive. Now we treat a
9381 received no_renegotiation alert as a fatal error. This is because
9382 applications requesting a renegotiation might well expect it to succeed
9383 and would have no code in place to handle the server denying it so the
9384 only safe thing to do is to terminate the connection.
9388 * Add ctrl macro SSL_get_secure_renegotiation_support() which returns 1 if
9389 peer supports secure renegotiation and 0 otherwise. Print out peer
9390 renegotiation support in s_client/s_server.
9394 * Replace the highly broken and deprecated SPKAC certification method with
9395 the updated NID creation version. This should correctly handle UTF8.
9399 * Implement RFC5746. Re-enable renegotiation but require the extension
9400 as needed. Unfortunately, SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION
9401 turns out to be a bad idea. It has been replaced by
9402 SSL_OP_ALLOW_UNSAFE_LEGACY_RENEGOTIATION which can be set with
9403 SSL_CTX_set_options(). This is really not recommended unless you
9404 know what you are doing.
9406 *Eric Rescorla <ekr@networkresonance.com>, Ben Laurie, Steve Henson*
9408 * Fixes to stateless session resumption handling. Use initial_ctx when
9409 issuing and attempting to decrypt tickets in case it has changed during
9410 servername handling. Use a non-zero length session ID when attempting
9411 stateless session resumption: this makes it possible to determine if
9412 a resumption has occurred immediately after receiving server hello
9413 (several places in OpenSSL subtly assume this) instead of later in
9418 * The functions ENGINE_ctrl(), OPENSSL_isservice(),
9419 CMS_get1_RecipientRequest() and RAND_bytes() can return <=0 on error
9420 fixes for a few places where the return code is not checked
9423 *Julia Lawall <julia@diku.dk>*
9425 * Add --strict-warnings option to Configure script to include devteam
9426 warnings in other configurations.
9430 * Add support for --libdir option and LIBDIR variable in makefiles. This
9431 makes it possible to install openssl libraries in locations which
9432 have names other than "lib", for example "/usr/lib64" which some
9435 *Steve Henson, based on patch from Jeremy Utley*
9437 * Don't allow the use of leading 0x80 in OIDs. This is a violation of
9438 X690 8.9.12 and can produce some misleading textual output of OIDs.
9440 *Steve Henson, reported by Dan Kaminsky*
9442 * Delete MD2 from algorithm tables. This follows the recommendation in
9443 several standards that it is not used in new applications due to
9444 several cryptographic weaknesses. For binary compatibility reasons
9445 the MD2 API is still compiled in by default.
9449 * Add compression id to {d2i,i2d}_SSL_SESSION so it is correctly saved
9454 * Rename uni2asc and asc2uni functions to OPENSSL_uni2asc and
9455 OPENSSL_asc2uni conditionally on Netware platforms to avoid a name
9458 *Guenter <lists@gknw.net>*
9460 * Fix the server certificate chain building code to use X509_verify_cert(),
9461 it used to have an ad-hoc builder which was unable to cope with anything
9462 other than a simple chain.
9464 *David Woodhouse <dwmw2@infradead.org>, Steve Henson*
9466 * Don't check self signed certificate signatures in X509_verify_cert()
9467 by default (a flag can override this): it just wastes time without
9468 adding any security. As a useful side effect self signed root CAs
9469 with non-FIPS digests are now usable in FIPS mode.
9473 * In dtls1_process_out_of_seq_message() the check if the current message
9474 is already buffered was missing. For every new message was memory
9475 allocated, allowing an attacker to perform an denial of service attack
9476 with sending out of seq handshake messages until there is no memory
9477 left. Additionally every future message was buffered, even if the
9478 sequence number made no sense and would be part of another handshake.
9479 So only messages with sequence numbers less than 10 in advance will be
9480 buffered. ([CVE-2009-1378])
9482 *Robin Seggelmann, discovered by Daniel Mentz*
9484 * Records are buffered if they arrive with a future epoch to be
9485 processed after finishing the corresponding handshake. There is
9486 currently no limitation to this buffer allowing an attacker to perform
9487 a DOS attack with sending records with future epochs until there is no
9488 memory left. This patch adds the pqueue_size() function to determine
9489 the size of a buffer and limits the record buffer to 100 entries.
9492 *Robin Seggelmann, discovered by Daniel Mentz*
9494 * Keep a copy of frag->msg_header.frag_len so it can be used after the
9495 parent structure is freed. ([CVE-2009-1379])
9499 * Handle non-blocking I/O properly in SSL_shutdown() call.
9501 *Darryl Miles <darryl-mailinglists@netbauds.net>*
9503 * Add `2.5.4.*` OIDs
9505 *Ilya O. <vrghost@gmail.com>*
9507 ### Changes between 0.9.8k and 0.9.8l [5 Nov 2009]
9509 * Disable renegotiation completely - this fixes a severe security
9510 problem ([CVE-2009-3555]) at the cost of breaking all
9511 renegotiation. Renegotiation can be re-enabled by setting
9512 SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION in s3->flags at
9513 run-time. This is really not recommended unless you know what
9518 ### Changes between 0.9.8j and 0.9.8k [25 Mar 2009]
9520 * Don't set val to NULL when freeing up structures, it is freed up by
9521 underlying code. If `sizeof(void *) > sizeof(long)` this can result in
9522 zeroing past the valid field. ([CVE-2009-0789])
9524 *Paolo Ganci <Paolo.Ganci@AdNovum.CH>*
9526 * Fix bug where return value of CMS_SignerInfo_verify_content() was not
9527 checked correctly. This would allow some invalid signed attributes to
9528 appear to verify correctly. ([CVE-2009-0591])
9530 *Ivan Nestlerode <inestlerode@us.ibm.com>*
9532 * Reject UniversalString and BMPString types with invalid lengths. This
9533 prevents a crash in ASN1_STRING_print_ex() which assumes the strings have
9534 a legal length. ([CVE-2009-0590])
9538 * Set S/MIME signing as the default purpose rather than setting it
9539 unconditionally. This allows applications to override it at the store
9544 * Permit restricted recursion of ASN1 strings. This is needed in practice
9545 to handle some structures.
9549 * Improve efficiency of mem_gets: don't search whole buffer each time
9552 *Jeremy Shapiro <jnshapir@us.ibm.com>*
9554 * New -hex option for openssl rand.
9558 * Print out UTF8String and NumericString when parsing ASN1.
9562 * Support NumericString type for name components.
9566 * Allow CC in the environment to override the automatically chosen
9567 compiler. Note that nothing is done to ensure flags work with the
9572 ### Changes between 0.9.8i and 0.9.8j [07 Jan 2009]
9574 * Properly check EVP_VerifyFinal() and similar return values
9577 *Ben Laurie, Bodo Moeller, Google Security Team*
9579 * Enable TLS extensions by default.
9583 * Allow the CHIL engine to be loaded, whether the application is
9584 multithreaded or not. (This does not release the developer from the
9585 obligation to set up the dynamic locking callbacks.)
9587 *Sander Temme <sander@temme.net>*
9589 * Use correct exit code if there is an error in dgst command.
9591 *Steve Henson; problem pointed out by Roland Dirlewanger*
9593 * Tweak Configure so that you need to say "experimental-jpake" to enable
9594 JPAKE, and need to use -DOPENSSL_EXPERIMENTAL_JPAKE in applications.
9598 * Add experimental JPAKE support, including demo authentication in
9599 s_client and s_server.
9603 * Set the comparison function in v3_addr_canonize().
9605 *Rob Austein <sra@hactrn.net>*
9607 * Add support for XMPP STARTTLS in s_client.
9609 *Philip Paeps <philip@freebsd.org>*
9611 * Change the server-side SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG behavior
9612 to ensure that even with this option, only ciphersuites in the
9613 server's preference list will be accepted. (Note that the option
9614 applies only when resuming a session, so the earlier behavior was
9615 just about the algorithm choice for symmetric cryptography.)
9619 ### Changes between 0.9.8h and 0.9.8i [15 Sep 2008]
9621 * Fix NULL pointer dereference if a DTLS server received
9622 ChangeCipherSpec as first record ([CVE-2009-1386]).
9626 * Fix a state transition in s3_srvr.c and d1_srvr.c
9627 (was using SSL3_ST_CW_CLNT_HELLO_B, should be `..._ST_SW_SRVR_...`).
9631 * The fix in 0.9.8c that supposedly got rid of unsafe
9632 double-checked locking was incomplete for RSA blinding,
9633 addressing just one layer of what turns out to have been
9634 doubly unsafe triple-checked locking.
9636 So now fix this for real by retiring the MONT_HELPER macro
9637 in crypto/rsa/rsa_eay.c.
9639 *Bodo Moeller; problem pointed out by Marius Schilder*
9641 * Various precautionary measures:
9643 - Avoid size_t integer overflow in HASH_UPDATE (md32_common.h).
9645 - Avoid a buffer overflow in d2i_SSL_SESSION() (ssl_asn1.c).
9646 (NB: This would require knowledge of the secret session ticket key
9647 to exploit, in which case you'd be SOL either way.)
9649 - Change bn_nist.c so that it will properly handle input BIGNUMs
9650 outside the expected range.
9652 - Enforce the 'num' check in BN_div() (bn_div.c) for non-BN_DEBUG
9655 *Neel Mehta, Bodo Moeller*
9657 * Allow engines to be "soft loaded" - i.e. optionally don't die if
9658 the load fails. Useful for distros.
9660 *Ben Laurie and the FreeBSD team*
9662 * Add support for Local Machine Keyset attribute in PKCS#12 files.
9666 * Fix BN_GF2m_mod_arr() top-bit cleanup code.
9670 * Expand ENGINE to support engine supplied SSL client certificate functions.
9672 This work was sponsored by Logica.
9676 * Add CryptoAPI ENGINE to support use of RSA and DSA keys held in Windows
9677 keystores. Support for SSL/TLS client authentication too.
9678 Not compiled unless enable-capieng specified to Configure.
9680 This work was sponsored by Logica.
9684 * Fix bug in X509_ATTRIBUTE creation: don't set attribute using
9685 ASN1_TYPE_set1 if MBSTRING flag set. This bug would crash certain
9686 attribute creation routines such as certificate requests and PKCS#12
9691 ### Changes between 0.9.8g and 0.9.8h [28 May 2008]
9693 * Fix flaw if 'Server Key exchange message' is omitted from a TLS
9694 handshake which could lead to a client crash as found using the
9695 Codenomicon TLS test suite ([CVE-2008-1672])
9697 *Steve Henson, Mark Cox*
9699 * Fix double free in TLS server name extensions which could lead to
9700 a remote crash found by Codenomicon TLS test suite ([CVE-2008-0891])
9704 * Clear error queue in SSL_CTX_use_certificate_chain_file()
9706 Clear the error queue to ensure that error entries left from
9707 older function calls do not interfere with the correct operation.
9709 *Lutz Jaenicke, Erik de Castro Lopo*
9711 * Remove root CA certificates of commercial CAs:
9713 The OpenSSL project does not recommend any specific CA and does not
9714 have any policy with respect to including or excluding any CA.
9715 Therefore it does not make any sense to ship an arbitrary selection
9716 of root CA certificates with the OpenSSL software.
9720 * RSA OAEP patches to fix two separate invalid memory reads.
9721 The first one involves inputs when 'lzero' is greater than
9722 'SHA_DIGEST_LENGTH' (it would read about SHA_DIGEST_LENGTH bytes
9723 before the beginning of from). The second one involves inputs where
9724 the 'db' section contains nothing but zeroes (there is a one-byte
9725 invalid read after the end of 'db').
9727 *Ivan Nestlerode <inestlerode@us.ibm.com>*
9729 * Partial backport from 0.9.9-dev:
9731 Introduce bn_mul_mont (dedicated Montgomery multiplication
9732 procedure) as a candidate for BIGNUM assembler implementation.
9733 While 0.9.9-dev uses assembler for various architectures, only
9734 x86_64 is available by default here in the 0.9.8 branch, and
9735 32-bit x86 is available through a compile-time setting.
9737 To try the 32-bit x86 assembler implementation, use Configure
9738 option "enable-montasm" (which exists only for this backport).
9740 As "enable-montasm" for 32-bit x86 disclaims code stability
9741 anyway, in this constellation we activate additional code
9742 backported from 0.9.9-dev for further performance improvements,
9743 namely BN_from_montgomery_word. (To enable this otherwise,
9744 e.g. x86_64, try `-DMONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD`.)
9746 *Andy Polyakov (backport partially by Bodo Moeller)*
9748 * Add TLS session ticket callback. This allows an application to set
9749 TLS ticket cipher and HMAC keys rather than relying on hardcoded fixed
9750 values. This is useful for key rollover for example where several key
9751 sets may exist with different names.
9755 * Reverse ENGINE-internal logic for caching default ENGINE handles.
9756 This was broken until now in 0.9.8 releases, such that the only way
9757 a registered ENGINE could be used (assuming it initialises
9758 successfully on the host) was to explicitly set it as the default
9759 for the relevant algorithms. This is in contradiction with 0.9.7
9760 behaviour and the documentation. With this fix, when an ENGINE is
9761 registered into a given algorithm's table of implementations, the
9762 'uptodate' flag is reset so that auto-discovery will be used next
9763 time a new context for that algorithm attempts to select an
9766 *Ian Lister (tweaked by Geoff Thorpe)*
9768 * Backport of CMS code to OpenSSL 0.9.8. This differs from the 0.9.9
9769 implementation in the following ways:
9771 Lack of EVP_PKEY_ASN1_METHOD means algorithm parameters have to be
9774 Lack of BER streaming support means one pass streaming processing is
9775 only supported if data is detached: setting the streaming flag is
9776 ignored for embedded content.
9778 CMS support is disabled by default and must be explicitly enabled
9779 with the enable-cms configuration option.
9783 * Update the GMP engine glue to do direct copies between BIGNUM and
9784 mpz_t when openssl and GMP use the same limb size. Otherwise the
9785 existing "conversion via a text string export" trick is still used.
9787 *Paul Sheer <paulsheer@gmail.com>*
9789 * Zlib compression BIO. This is a filter BIO which compressed and
9790 uncompresses any data passed through it.
9794 * Add AES_wrap_key() and AES_unwrap_key() functions to implement
9795 RFC3394 compatible AES key wrapping.
9799 * Add utility functions to handle ASN1 structures. ASN1_STRING_set0():
9800 sets string data without copying. X509_ALGOR_set0() and
9801 X509_ALGOR_get0(): set and retrieve X509_ALGOR (AlgorithmIdentifier)
9802 data. Attribute function X509at_get0_data_by_OBJ(): retrieves data
9803 from an X509_ATTRIBUTE structure optionally checking it occurs only
9804 once. ASN1_TYPE_set1(): set and ASN1_TYPE structure copying supplied
9809 * Fix BN flag handling in RSA_eay_mod_exp() and BN_MONT_CTX_set()
9810 to get the expected BN_FLG_CONSTTIME behavior.
9812 *Bodo Moeller (Google)*
9816 - fixed wrong usage of ioctlsocket() when build for LIBC BSD sockets
9817 - fixed do_tests.pl to run the test suite with CLIB builds too (CLIB_OPT)
9818 - added some more tests to do_tests.pl
9819 - fixed RunningProcess usage so that it works with newer LIBC NDKs too
9820 - removed usage of BN_LLONG for CLIB builds to avoid runtime dependency
9821 - added new Configure targets netware-clib-bsdsock, netware-clib-gcc,
9822 netware-clib-bsdsock-gcc, netware-libc-bsdsock-gcc
9823 - various changes to netware.pl to enable gcc-cross builds on Win32
9825 - changed crypto/bio/b_sock.c to work with macro functions (CLIB BSD)
9826 - various changes to fix missing prototype warnings
9827 - fixed x86nasm.pl to create correct asm files for NASM COFF output
9828 - added AES, WHIRLPOOL and CPUID assembler code to build files
9829 - added missing AES assembler make rules to mk1mf.pl
9830 - fixed order of includes in `apps/ocsp.c` so that `e_os.h` settings apply
9832 *Guenter Knauf <eflash@gmx.net>*
9834 * Implement certificate status request TLS extension defined in RFC3546.
9835 A client can set the appropriate parameters and receive the encoded
9836 OCSP response via a callback. A server can query the supplied parameters
9837 and set the encoded OCSP response in the callback. Add simplified examples
9838 to s_client and s_server.
9842 ### Changes between 0.9.8f and 0.9.8g [19 Oct 2007]
9845 + Binary incompatibility of ssl_ctx_st structure
9846 + DTLS interoperation with non-compliant servers
9847 + Don't call get_session_cb() without proposed session
9848 + Fix ia64 assembler code
9850 *Andy Polyakov, Steve Henson*
9852 ### Changes between 0.9.8e and 0.9.8f [11 Oct 2007]
9854 * DTLS Handshake overhaul. There were longstanding issues with
9855 OpenSSL DTLS implementation, which were making it impossible for
9856 RFC 4347 compliant client to communicate with OpenSSL server.
9857 Unfortunately just fixing these incompatibilities would "cut off"
9858 pre-0.9.8f clients. To allow for hassle free upgrade post-0.9.8e
9859 server keeps tolerating non RFC compliant syntax. The opposite is
9860 not true, 0.9.8f client can not communicate with earlier server.
9861 This update even addresses CVE-2007-4995.
9865 * Changes to avoid need for function casts in OpenSSL: some compilers
9866 (gcc 4.2 and later) reject their use.
9867 *Kurt Roeckx <kurt@roeckx.be>, Peter Hartley <pdh@utter.chaos.org.uk>,
9870 * Add RFC4507 support to OpenSSL. This includes the corrections in
9871 RFC4507bis. The encrypted ticket format is an encrypted encoded
9872 SSL_SESSION structure, that way new session features are automatically
9875 If a client application caches session in an SSL_SESSION structure
9876 support is transparent because tickets are now stored in the encoded
9879 The SSL_CTX structure automatically generates keys for ticket
9880 protection in servers so again support should be possible
9881 with no application modification.
9883 If a client or server wishes to disable RFC4507 support then the option
9884 SSL_OP_NO_TICKET can be set.
9886 Add a TLS extension debugging callback to allow the contents of any client
9887 or server extensions to be examined.
9889 This work was sponsored by Google.
9893 * Add initial support for TLS extensions, specifically for the server_name
9894 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
9895 have new members for a host name. The SSL data structure has an
9896 additional member `SSL_CTX *initial_ctx` so that new sessions can be
9897 stored in that context to allow for session resumption, even after the
9898 SSL has been switched to a new SSL_CTX in reaction to a client's
9899 server_name extension.
9901 New functions (subject to change):
9903 SSL_get_servername()
9904 SSL_get_servername_type()
9907 New CTRL codes and macros (subject to change):
9909 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
9910 - SSL_CTX_set_tlsext_servername_callback()
9911 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
9912 - SSL_CTX_set_tlsext_servername_arg()
9913 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
9915 openssl s_client has a new '-servername ...' option.
9917 openssl s_server has new options '-servername_host ...', '-cert2 ...',
9918 '-key2 ...', '-servername_fatal' (subject to change). This allows
9919 testing the HostName extension for a specific single host name ('-cert'
9920 and '-key' remain fallbacks for handshakes without HostName
9921 negotiation). If the unrecognized_name alert has to be sent, this by
9922 default is a warning; it becomes fatal with the '-servername_fatal'
9925 *Peter Sylvester, Remy Allais, Christophe Renou, Steve Henson*
9927 * Add AES and SSE2 assembly language support to VC++ build.
9931 * Mitigate attack on final subtraction in Montgomery reduction.
9935 * Fix crypto/ec/ec_mult.c to work properly with scalars of value 0
9936 (which previously caused an internal error).
9940 * Squeeze another 10% out of IGE mode when in != out.
9944 * AES IGE mode speedup.
9946 *Dean Gaudet (Google)*
9948 * Add the Korean symmetric 128-bit cipher SEED (see
9949 <http://www.kisa.or.kr/kisa/seed/jsp/seed_eng.jsp>) and
9950 add SEED ciphersuites from RFC 4162:
9952 TLS_RSA_WITH_SEED_CBC_SHA = "SEED-SHA"
9953 TLS_DHE_DSS_WITH_SEED_CBC_SHA = "DHE-DSS-SEED-SHA"
9954 TLS_DHE_RSA_WITH_SEED_CBC_SHA = "DHE-RSA-SEED-SHA"
9955 TLS_DH_anon_WITH_SEED_CBC_SHA = "ADH-SEED-SHA"
9957 To minimize changes between patchlevels in the OpenSSL 0.9.8
9958 series, SEED remains excluded from compilation unless OpenSSL
9959 is configured with 'enable-seed'.
9961 *KISA, Bodo Moeller*
9963 * Mitigate branch prediction attacks, which can be practical if a
9964 single processor is shared, allowing a spy process to extract
9965 information. For detailed background information, see
9966 <http://eprint.iacr.org/2007/039> (O. Aciicmez, S. Gueron,
9967 J.-P. Seifert, "New Branch Prediction Vulnerabilities in OpenSSL
9968 and Necessary Software Countermeasures"). The core of the change
9969 are new versions BN_div_no_branch() and
9970 BN_mod_inverse_no_branch() of BN_div() and BN_mod_inverse(),
9971 respectively, which are slower, but avoid the security-relevant
9972 conditional branches. These are automatically called by BN_div()
9973 and BN_mod_inverse() if the flag BN_FLG_CONSTTIME is set for one
9974 of the input BIGNUMs. Also, BN_is_bit_set() has been changed to
9975 remove a conditional branch.
9977 BN_FLG_CONSTTIME is the new name for the previous
9978 BN_FLG_EXP_CONSTTIME flag, since it now affects more than just
9979 modular exponentiation. (Since OpenSSL 0.9.7h, setting this flag
9980 in the exponent causes BN_mod_exp_mont() to use the alternative
9981 implementation in BN_mod_exp_mont_consttime().) The old name
9982 remains as a deprecated alias.
9984 Similarly, RSA_FLAG_NO_EXP_CONSTTIME is replaced by a more general
9985 RSA_FLAG_NO_CONSTTIME flag since the RSA implementation now uses
9986 constant-time implementations for more than just exponentiation.
9987 Here too the old name is kept as a deprecated alias.
9989 BN_BLINDING_new() will now use BN_dup() for the modulus so that
9990 the BN_BLINDING structure gets an independent copy of the
9991 modulus. This means that the previous `BIGNUM *m` argument to
9992 BN_BLINDING_new() and to BN_BLINDING_create_param() now
9993 essentially becomes `const BIGNUM *m`, although we can't actually
9994 change this in the header file before 0.9.9. It allows
9995 RSA_setup_blinding() to use BN_with_flags() on the modulus to
9996 enable BN_FLG_CONSTTIME.
9998 *Matthew D Wood (Intel Corp)*
10000 * In the SSL/TLS server implementation, be strict about session ID
10001 context matching (which matters if an application uses a single
10002 external cache for different purposes). Previously,
10003 out-of-context reuse was forbidden only if SSL_VERIFY_PEER was
10004 set. This did ensure strict client verification, but meant that,
10005 with applications using a single external cache for quite
10006 different requirements, clients could circumvent ciphersuite
10007 restrictions for a given session ID context by starting a session
10008 in a different context.
10012 * Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
10013 a ciphersuite string such as "DEFAULT:RSA" cannot enable
10014 authentication-only ciphersuites.
10018 * Update the SSL_get_shared_ciphers() fix CVE-2006-3738 which was
10019 not complete and could lead to a possible single byte overflow
10020 ([CVE-2007-5135]) [Ben Laurie]
10022 ### Changes between 0.9.8d and 0.9.8e [23 Feb 2007]
10024 * Since AES128 and AES256 (and similarly Camellia128 and
10025 Camellia256) share a single mask bit in the logic of
10026 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
10027 kludge to work properly if AES128 is available and AES256 isn't
10028 (or if Camellia128 is available and Camellia256 isn't).
10032 * Fix the BIT STRING encoding generated by crypto/ec/ec_asn1.c
10033 (within i2d_ECPrivateKey, i2d_ECPKParameters, i2d_ECParameters):
10034 When a point or a seed is encoded in a BIT STRING, we need to
10035 prevent the removal of trailing zero bits to get the proper DER
10036 encoding. (By default, crypto/asn1/a_bitstr.c assumes the case
10037 of a NamedBitList, for which trailing 0 bits need to be removed.)
10041 * Have SSL/TLS server implementation tolerate "mismatched" record
10042 protocol version while receiving ClientHello even if the
10043 ClientHello is fragmented. (The server can't insist on the
10044 particular protocol version it has chosen before the ServerHello
10045 message has informed the client about his choice.)
10049 * Add RFC 3779 support.
10051 *Rob Austein for ARIN, Ben Laurie*
10053 * Load error codes if they are not already present instead of using a
10054 static variable. This allows them to be cleanly unloaded and reloaded.
10055 Improve header file function name parsing.
10059 * extend SMTP and IMAP protocol emulation in s_client to use EHLO
10060 or CAPABILITY handshake as required by RFCs.
10062 *Goetz Babin-Ebell*
10064 ### Changes between 0.9.8c and 0.9.8d [28 Sep 2006]
10066 * Introduce limits to prevent malicious keys being able to
10067 cause a denial of service. ([CVE-2006-2940])
10069 *Steve Henson, Bodo Moeller*
10071 * Fix ASN.1 parsing of certain invalid structures that can result
10072 in a denial of service. ([CVE-2006-2937]) [Steve Henson]
10074 * Fix buffer overflow in SSL_get_shared_ciphers() function.
10075 ([CVE-2006-3738]) [Tavis Ormandy and Will Drewry, Google Security Team]
10077 * Fix SSL client code which could crash if connecting to a
10078 malicious SSLv2 server. ([CVE-2006-4343])
10080 *Tavis Ormandy and Will Drewry, Google Security Team*
10082 * Since 0.9.8b, ciphersuite strings naming explicit ciphersuites
10083 match only those. Before that, "AES256-SHA" would be interpreted
10084 as a pattern and match "AES128-SHA" too (since AES128-SHA got
10085 the same strength classification in 0.9.7h) as we currently only
10086 have a single AES bit in the ciphersuite description bitmap.
10087 That change, however, also applied to ciphersuite strings such as
10088 "RC4-MD5" that intentionally matched multiple ciphersuites --
10089 namely, SSL 2.0 ciphersuites in addition to the more common ones
10090 from SSL 3.0/TLS 1.0.
10092 So we change the selection algorithm again: Naming an explicit
10093 ciphersuite selects this one ciphersuite, and any other similar
10094 ciphersuite (same bitmap) from *other* protocol versions.
10095 Thus, "RC4-MD5" again will properly select both the SSL 2.0
10096 ciphersuite and the SSL 3.0/TLS 1.0 ciphersuite.
10098 Since SSL 2.0 does not have any ciphersuites for which the
10099 128/256 bit distinction would be relevant, this works for now.
10100 The proper fix will be to use different bits for AES128 and
10101 AES256, which would have avoided the problems from the beginning;
10102 however, bits are scarce, so we can only do this in a new release
10103 (not just a patchlevel) when we can change the SSL_CIPHER
10104 definition to split the single 'unsigned long mask' bitmap into
10105 multiple values to extend the available space.
10109 ### Changes between 0.9.8b and 0.9.8c [05 Sep 2006]
10111 * Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
10112 ([CVE-2006-4339]) [Ben Laurie and Google Security Team]
10114 * Add AES IGE and biIGE modes.
10118 * Change the Unix randomness entropy gathering to use poll() when
10119 possible instead of select(), since the latter has some
10120 undesirable limitations.
10122 *Darryl Miles via Richard Levitte and Bodo Moeller*
10124 * Disable "ECCdraft" ciphersuites more thoroughly. Now special
10125 treatment in ssl/ssl_ciph.s makes sure that these ciphersuites
10126 cannot be implicitly activated as part of, e.g., the "AES" alias.
10127 However, please upgrade to OpenSSL 0.9.9[-dev] for
10128 non-experimental use of the ECC ciphersuites to get TLS extension
10129 support, which is required for curve and point format negotiation
10130 to avoid potential handshake problems.
10134 * Disable rogue ciphersuites:
10136 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
10137 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
10138 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
10140 The latter two were purportedly from
10141 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
10144 Also deactivate the remaining ciphersuites from
10145 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
10146 unofficial, and the ID has long expired.
10150 * Fix RSA blinding Heisenbug (problems sometimes occurred on
10151 dual-core machines) and other potential thread-safety issues.
10155 * Add the symmetric cipher Camellia (128-bit, 192-bit, 256-bit key
10156 versions), which is now available for royalty-free use
10157 (see <http://info.isl.ntt.co.jp/crypt/eng/info/chiteki.html>).
10158 Also, add Camellia TLS ciphersuites from RFC 4132.
10160 To minimize changes between patchlevels in the OpenSSL 0.9.8
10161 series, Camellia remains excluded from compilation unless OpenSSL
10162 is configured with 'enable-camellia'.
10166 * Disable the padding bug check when compression is in use. The padding
10167 bug check assumes the first packet is of even length, this is not
10168 necessarily true if compression is enabled and can result in false
10169 positives causing handshake failure. The actual bug test is ancient
10170 code so it is hoped that implementations will either have fixed it by
10171 now or any which still have the bug do not support compression.
10175 ### Changes between 0.9.8a and 0.9.8b [04 May 2006]
10177 * When applying a cipher rule check to see if string match is an explicit
10178 cipher suite and only match that one cipher suite if it is.
10182 * Link in manifests for VC++ if needed.
10184 *Austin Ziegler <halostatue@gmail.com>*
10186 * Update support for ECC-based TLS ciphersuites according to
10187 draft-ietf-tls-ecc-12.txt with proposed changes (but without
10188 TLS extensions, which are supported starting with the 0.9.9
10189 branch, not in the OpenSSL 0.9.8 branch).
10193 * New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support
10194 opaque EVP_CIPHER_CTX handling.
10198 * Fixes and enhancements to zlib compression code. We now only use
10199 "zlib1.dll" and use the default `__cdecl` calling convention on Win32
10200 to conform with the standards mentioned here:
10201 <http://www.zlib.net/DLL_FAQ.txt>
10202 Static zlib linking now works on Windows and the new --with-zlib-include
10203 --with-zlib-lib options to Configure can be used to supply the location
10204 of the headers and library. Gracefully handle case where zlib library
10209 * Several fixes and enhancements to the OID generation code. The old code
10210 sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't
10211 handle numbers larger than ULONG_MAX, truncated printing and had a
10212 non standard OBJ_obj2txt() behaviour.
10216 * Add support for building of engines under engine/ as shared libraries
10217 under VC++ build system.
10221 * Corrected the numerous bugs in the Win32 path splitter in DSO.
10222 Hopefully, we will not see any false combination of paths any more.
10226 ### Changes between 0.9.8 and 0.9.8a [11 Oct 2005]
10228 * Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
10229 (part of SSL_OP_ALL). This option used to disable the
10230 countermeasure against man-in-the-middle protocol-version
10231 rollback in the SSL 2.0 server implementation, which is a bad
10232 idea. ([CVE-2005-2969])
10234 *Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
10235 for Information Security, National Institute of Advanced Industrial
10236 Science and Technology [AIST], Japan)*
10238 * Add two function to clear and return the verify parameter flags.
10242 * Keep cipherlists sorted in the source instead of sorting them at
10243 runtime, thus removing the need for a lock.
10247 * Avoid some small subgroup attacks in Diffie-Hellman.
10249 *Nick Mathewson and Ben Laurie*
10251 * Add functions for well-known primes.
10255 * Extended Windows CE support.
10257 *Satoshi Nakamura and Andy Polyakov*
10259 * Initialize SSL_METHOD structures at compile time instead of during
10260 runtime, thus removing the need for a lock.
10264 * Make PKCS7_decrypt() work even if no certificate is supplied by
10265 attempting to decrypt each encrypted key in turn. Add support to
10270 ### Changes between 0.9.7h and 0.9.8 [05 Jul 2005]
10272 [NB: OpenSSL 0.9.7i and later 0.9.7 patch levels were released after
10275 * Add libcrypto.pc and libssl.pc for those who feel they need them.
10279 * Change CA.sh and CA.pl so they don't bundle the CSR and the private
10280 key into the same file any more.
10284 * Add initial support for Win64, both IA64 and AMD64/x64 flavors.
10288 * Add -utf8 command line and config file option to 'ca'.
10290 *Stefan <stf@udoma.org*
10292 * Removed the macro des_crypt(), as it seems to conflict with some
10293 libraries. Use DES_crypt().
10297 * Correct naming of the 'chil' and '4758cca' ENGINEs. This
10298 involves renaming the source and generated shared-libs for
10299 both. The engines will accept the corrected or legacy ids
10300 ('ncipher' and '4758_cca' respectively) when binding. NB,
10301 this only applies when building 'shared'.
10303 *Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe*
10305 * Add attribute functions to EVP_PKEY structure. Modify
10306 PKCS12_create() to recognize a CSP name attribute and
10307 use it. Make -CSP option work again in pkcs12 utility.
10311 * Add new functionality to the bn blinding code:
10312 - automatic re-creation of the BN_BLINDING parameters after
10313 a fixed number of uses (currently 32)
10314 - add new function for parameter creation
10315 - introduce flags to control the update behaviour of the
10316 BN_BLINDING parameters
10317 - hide BN_BLINDING structure
10318 Add a second BN_BLINDING slot to the RSA structure to improve
10319 performance when a single RSA object is shared among several
10324 * Add support for DTLS.
10326 *Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie*
10328 * Add support for DER encoded private keys (SSL_FILETYPE_ASN1)
10329 to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file()
10333 * Remove buggy and incomplete DH cert support from
10334 ssl/ssl_rsa.c and ssl/s3_both.c
10338 * Use SHA-1 instead of MD5 as the default digest algorithm for
10339 the `apps/openssl` commands.
10343 * Compile clean with "-Wall -Wmissing-prototypes
10344 -Wstrict-prototypes -Wmissing-declarations -Werror". Currently
10345 DEBUG_SAFESTACK must also be set.
10349 * Change ./Configure so that certain algorithms can be disabled by default.
10350 The new counterpiece to "no-xxx" is "enable-xxx".
10352 The patented RC5 and MDC2 algorithms will now be disabled unless
10353 "enable-rc5" and "enable-mdc2", respectively, are specified.
10355 (IDEA remains enabled despite being patented. This is because IDEA
10356 is frequently required for interoperability, and there is no license
10357 fee for non-commercial use. As before, "no-idea" can be used to
10358 avoid this algorithm.)
10362 * Add processing of proxy certificates (see RFC 3820). This work was
10363 sponsored by KTH (The Royal Institute of Technology in Stockholm) and
10364 EGEE (Enabling Grids for E-science in Europe).
10368 * RC4 performance overhaul on modern architectures/implementations, such
10369 as Intel P4, IA-64 and AMD64.
10373 * New utility extract-section.pl. This can be used specify an alternative
10374 section number in a pod file instead of having to treat each file as
10375 a separate case in Makefile. This can be done by adding two lines to the
10378 =for comment openssl_section:XXX
10380 The blank line is mandatory.
10384 * New arguments -certform, -keyform and -pass for s_client and s_server
10385 to allow alternative format key and certificate files and passphrase
10390 * New structure X509_VERIFY_PARAM which combines current verify parameters,
10391 update associated structures and add various utility functions.
10393 Add new policy related verify parameters, include policy checking in
10394 standard verify code. Enhance 'smime' application with extra parameters
10395 to support policy checking and print out.
10399 * Add a new engine to support VIA PadLock ACE extensions in the VIA C3
10400 Nehemiah processors. These extensions support AES encryption in hardware
10401 as well as RNG (though RNG support is currently disabled).
10403 *Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov*
10405 * Deprecate `BN_[get|set]_params()` functions (they were ignored internally).
10409 * New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented.
10411 *Andy Polyakov and a number of other people*
10413 * Improved PowerPC platform support. Most notably BIGNUM assembler
10414 implementation contributed by IBM.
10416 *Suresh Chari, Peter Waltenberg, Andy Polyakov*
10418 * The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public
10419 exponent rather than 'unsigned long'. There is a corresponding change to
10420 the new 'rsa_keygen' element of the RSA_METHOD structure.
10422 *Jelte Jansen, Geoff Thorpe*
10424 * Functionality for creating the initial serial number file is now
10425 moved from CA.pl to the 'ca' utility with a new option -create_serial.
10427 (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial
10428 number file to 1, which is bound to cause problems. To avoid
10429 the problems while respecting compatibility between different 0.9.7
10430 patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in
10431 CA.pl for serial number initialization. With the new release 0.9.8,
10432 we can fix the problem directly in the 'ca' utility.)
10436 * Reduced header interdependencies by declaring more opaque objects in
10437 ossl_typ.h. As a consequence, including some headers (eg. engine.h) will
10438 give fewer recursive includes, which could break lazy source code - so
10439 this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always,
10440 developers should define this symbol when building and using openssl to
10441 ensure they track the recommended behaviour, interfaces, [etc], but
10442 backwards-compatible behaviour prevails when this isn't defined.
10446 * New function X509_POLICY_NODE_print() which prints out policy nodes.
10450 * Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality.
10451 This will generate a random key of the appropriate length based on the
10452 cipher context. The EVP_CIPHER can provide its own random key generation
10453 routine to support keys of a specific form. This is used in the des and
10454 3des routines to generate a key of the correct parity. Update S/MIME
10455 code to use new functions and hence generate correct parity DES keys.
10456 Add EVP_CHECK_DES_KEY #define to return an error if the key is not
10457 valid (weak or incorrect parity).
10461 * Add a local set of CRLs that can be used by X509_verify_cert() as well
10462 as looking them up. This is useful when the verified structure may contain
10463 CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs
10464 present unless the new PKCS7_NO_CRL flag is asserted.
10468 * Extend ASN1 oid configuration module. It now additionally accepts the
10471 shortName = some long name, 1.2.3.4
10475 * Reimplemented the BN_CTX implementation. There is now no more static
10476 limitation on the number of variables it can handle nor the depth of the
10477 "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack
10478 information can now expand as required, and rather than having a single
10479 static array of bignums, BN_CTX now uses a linked-list of such arrays
10480 allowing it to expand on demand whilst maintaining the usefulness of
10481 BN_CTX's "bundling".
10485 * Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD
10486 to allow all RSA operations to function using a single BN_CTX.
10490 * Preliminary support for certificate policy evaluation and checking. This
10491 is initially intended to pass the tests outlined in "Conformance Testing
10492 of Relying Party Client Certificate Path Processing Logic" v1.07.
10496 * bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and
10497 remained unused and not that useful. A variety of other little bignum
10498 tweaks and fixes have also been made continuing on from the audit (see
10503 * Constify all or almost all d2i, c2i, s2i and r2i functions, along with
10504 associated ASN1, EVP and SSL functions and old ASN1 macros.
10508 * BN_zero() only needs to set 'top' and 'neg' to zero for correct results,
10509 and this should never fail. So the return value from the use of
10510 BN_set_word() (which can fail due to needless expansion) is now deprecated;
10511 if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro.
10515 * BN_CTX_get() should return zero-valued bignums, providing the same
10516 initialised value as BN_new().
10518 *Geoff Thorpe, suggested by Ulf Möller*
10520 * Support for inhibitAnyPolicy certificate extension.
10524 * An audit of the BIGNUM code is underway, for which debugging code is
10525 enabled when BN_DEBUG is defined. This makes stricter enforcements on what
10526 is considered valid when processing BIGNUMs, and causes execution to
10527 assert() when a problem is discovered. If BN_DEBUG_RAND is defined,
10528 further steps are taken to deliberately pollute unused data in BIGNUM
10529 structures to try and expose faulty code further on. For now, openssl will
10530 (in its default mode of operation) continue to tolerate the inconsistent
10531 forms that it has tolerated in the past, but authors and packagers should
10532 consider trying openssl and their own applications when compiled with
10533 these debugging symbols defined. It will help highlight potential bugs in
10534 their own code, and will improve the test coverage for OpenSSL itself. At
10535 some point, these tighter rules will become openssl's default to improve
10536 maintainability, though the assert()s and other overheads will remain only
10537 in debugging configurations. See bn.h for more details.
10539 *Geoff Thorpe, Nils Larsch, Ulf Möller*
10541 * BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure
10542 that can only be obtained through BN_CTX_new() (which implicitly
10543 initialises it). The presence of this function only made it possible
10544 to overwrite an existing structure (and cause memory leaks).
10548 * Because of the callback-based approach for implementing LHASH as a
10549 template type, lh_insert() adds opaque objects to hash-tables and
10550 lh_doall() or lh_doall_arg() are typically used with a destructor callback
10551 to clean up those corresponding objects before destroying the hash table
10552 (and losing the object pointers). So some over-zealous constifications in
10553 LHASH have been relaxed so that lh_insert() does not take (nor store) the
10554 objects as "const" and the `lh_doall[_arg]` callback wrappers are not
10555 prototyped to have "const" restrictions on the object pointers they are
10556 given (and so aren't required to cast them away any more).
10560 * The tmdiff.h API was so ugly and minimal that our own timing utility
10561 (speed) prefers to use its own implementation. The two implementations
10562 haven't been consolidated as yet (volunteers?) but the tmdiff API has had
10563 its object type properly exposed (MS_TM) instead of casting to/from
10564 `char *`. This may still change yet if someone realises MS_TM and
10566 aren't necessarily the greatest nomenclatures - but this is what was used
10567 internally to the implementation so I've used that for now.
10571 * Ensure that deprecated functions do not get compiled when
10572 OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of
10573 the self-tests were still using deprecated key-generation functions so
10574 these have been updated also.
10578 * Reorganise PKCS#7 code to separate the digest location functionality
10579 into PKCS7_find_digest(), digest addition into PKCS7_bio_add_digest().
10580 New function PKCS7_set_digest() to set the digest type for PKCS#7
10581 digestedData type. Add additional code to correctly generate the
10582 digestedData type and add support for this type in PKCS7 initialization
10587 * New function PKCS7_set0_type_other() this initializes a PKCS7
10588 structure of type "other".
10592 * Fix prime generation loop in crypto/bn/bn_prime.pl by making
10593 sure the loop does correctly stop and breaking ("division by zero")
10594 modulus operations are not performed. The (pre-generated) prime
10595 table crypto/bn/bn_prime.h was already correct, but it could not be
10596 re-generated on some platforms because of the "division by zero"
10597 situation in the script.
10599 *Ralf S. Engelschall*
10601 * Update support for ECC-based TLS ciphersuites according to
10602 draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with
10603 SHA-1 now is only used for "small" curves (where the
10604 representation of a field element takes up to 24 bytes); for
10605 larger curves, the field element resulting from ECDH is directly
10606 used as premaster secret.
10608 *Douglas Stebila (Sun Microsystems Laboratories)*
10610 * Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2
10611 curve secp160r1 to the tests.
10613 *Douglas Stebila (Sun Microsystems Laboratories)*
10615 * Add the possibility to load symbols globally with DSO.
10617 *Götz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte*
10619 * Add the functions ERR_set_mark() and ERR_pop_to_mark() for better
10620 control of the error stack.
10624 * Add support for STORE in ENGINE.
10628 * Add the STORE type. The intention is to provide a common interface
10629 to certificate and key stores, be they simple file-based stores, or
10630 HSM-type store, or LDAP stores, or...
10631 NOTE: The code is currently UNTESTED and isn't really used anywhere.
10635 * Add a generic structure called OPENSSL_ITEM. This can be used to
10636 pass a list of arguments to any function as well as provide a way
10637 for a function to pass data back to the caller.
10641 * Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup()
10642 works like BUF_strdup() but can be used to duplicate a portion of
10643 a string. The copy gets NUL-terminated. BUF_memdup() duplicates
10648 * Add the function sk_find_ex() which works like sk_find(), but will
10649 return an index to an element even if an exact match couldn't be
10650 found. The index is guaranteed to point at the element where the
10651 searched-for key would be inserted to preserve sorting order.
10655 * Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but
10656 takes an extra flags argument for optional functionality. Currently,
10657 the following flags are defined:
10659 OBJ_BSEARCH_VALUE_ON_NOMATCH
10660 This one gets OBJ_bsearch_ex() to return a pointer to the first
10661 element where the comparing function returns a negative or zero
10664 OBJ_BSEARCH_FIRST_VALUE_ON_MATCH
10665 This one gets OBJ_bsearch_ex() to return a pointer to the first
10666 element where the comparing function returns zero. This is useful
10667 if there are more than one element where the comparing function
10672 * Make it possible to create self-signed certificates with 'openssl ca'
10673 in such a way that the self-signed certificate becomes part of the
10674 CA database and uses the same mechanisms for serial number generation
10675 as all other certificate signing. The new flag '-selfsign' enables
10676 this functionality. Adapt CA.sh and CA.pl.in.
10680 * Add functionality to check the public key of a certificate request
10681 against a given private. This is useful to check that a certificate
10682 request can be signed by that key (self-signing).
10686 * Make it possible to have multiple active certificates with the same
10687 subject in the CA index file. This is done only if the keyword
10688 'unique_subject' is set to 'no' in the main CA section (default
10689 if 'CA_default') of the configuration file. The value is saved
10690 with the database itself in a separate index attribute file,
10691 named like the index file with '.attr' appended to the name.
10695 * Generate multi-valued AVAs using '+' notation in config files for
10700 * Support for nameConstraints certificate extension.
10704 * Support for policyConstraints certificate extension.
10708 * Support for policyMappings certificate extension.
10712 * Make sure the default DSA_METHOD implementation only uses its
10713 dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL,
10714 and change its own handlers to be NULL so as to remove unnecessary
10715 indirection. This lets alternative implementations fallback to the
10716 default implementation more easily.
10720 * Support for directoryName in GeneralName related extensions
10725 * Make it possible to link applications using Makefile.shared.
10726 Make that possible even when linking against static libraries!
10730 * Support for single pass processing for S/MIME signing. This now
10731 means that S/MIME signing can be done from a pipe, in addition
10732 cleartext signing (multipart/signed type) is effectively streaming
10733 and the signed data does not need to be all held in memory.
10735 This is done with a new flag PKCS7_STREAM. When this flag is set
10736 PKCS7_sign() only initializes the PKCS7 structure and the actual signing
10737 is done after the data is output (and digests calculated) in
10738 SMIME_write_PKCS7().
10742 * Add full support for -rpath/-R, both in shared libraries and
10743 applications, at least on the platforms where it's known how
10748 * In crypto/ec/ec_mult.c, implement fast point multiplication with
10749 precomputation, based on wNAF splitting: EC_GROUP_precompute_mult()
10750 will now compute a table of multiples of the generator that
10751 makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul()
10752 faster (notably in the case of a single point multiplication,
10753 scalar * generator).
10755 *Nils Larsch, Bodo Moeller*
10757 * IPv6 support for certificate extensions. The various extensions
10758 which use the IP:a.b.c.d can now take IPv6 addresses using the
10759 formats of RFC1884 2.2 . IPv6 addresses are now also displayed
10764 * Added an ENGINE that implements RSA by performing private key
10765 exponentiations with the GMP library. The conversions to and from
10766 GMP's mpz_t format aren't optimised nor are any montgomery forms
10767 cached, and on x86 it appears OpenSSL's own performance has caught up.
10768 However there are likely to be other architectures where GMP could
10769 provide a boost. This ENGINE is not built in by default, but it can be
10770 specified at Configure time and should be accompanied by the necessary
10771 linker additions, eg;
10772 ./config -DOPENSSL_USE_GMP -lgmp
10776 * "openssl engine" will not display ENGINE/DSO load failure errors when
10777 testing availability of engines with "-t" - the old behaviour is
10778 produced by increasing the feature's verbosity with "-tt".
10782 * ECDSA routines: under certain error conditions uninitialized BN objects
10783 could be freed. Solution: make sure initialization is performed early
10784 enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de>
10789 * Key-generation can now be implemented in RSA_METHOD, DSA_METHOD
10790 and DH_METHOD (eg. by ENGINE implementations) to override the normal
10791 software implementations. For DSA and DH, parameter generation can
10792 also be overridden by providing the appropriate method callbacks.
10796 * Change the "progress" mechanism used in key-generation and
10797 primality testing to functions that take a new BN_GENCB pointer in
10798 place of callback/argument pairs. The new API functions have `_ex`
10799 postfixes and the older functions are reimplemented as wrappers for
10800 the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide
10801 declarations of the old functions to help (graceful) attempts to
10802 migrate to the new functions. Also, the new key-generation API
10803 functions operate on a caller-supplied key-structure and return
10804 success/failure rather than returning a key or NULL - this is to
10805 help make "keygen" another member function of RSA_METHOD etc.
10807 Example for using the new callback interface:
10809 int (*my_callback)(int a, int b, BN_GENCB *cb) = ...;
10810 void *my_arg = ...;
10813 BN_GENCB_set(&my_cb, my_callback, my_arg);
10815 return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb);
10816 /* For the meaning of a, b in calls to my_callback(), see the
10817 * documentation of the function that calls the callback.
10818 * cb will point to my_cb; my_arg can be retrieved as cb->arg.
10819 * my_callback should return 1 if it wants BN_is_prime_ex()
10820 * to continue, or 0 to stop.
10825 * Change the ZLIB compression method to be stateful, and make it
10826 available to TLS with the number defined in
10827 draft-ietf-tls-compression-04.txt.
10831 * Add the ASN.1 structures and functions for CertificatePair, which
10832 is defined as follows (according to X.509_4thEditionDraftV6.pdf):
10834 CertificatePair ::= SEQUENCE {
10835 forward [0] Certificate OPTIONAL,
10836 reverse [1] Certificate OPTIONAL,
10837 -- at least one of the pair shall be present -- }
10839 Also implement the PEM functions to read and write certificate
10840 pairs, and defined the PEM tag as "CERTIFICATE PAIR".
10842 This needed to be defined, mostly for the sake of the LDAP
10843 attribute crossCertificatePair, but may prove useful elsewhere as
10848 * Make it possible to inhibit symlinking of shared libraries in
10849 Makefile.shared, for Cygwin's sake.
10853 * Extend the BIGNUM API by creating a function
10854 void BN_set_negative(BIGNUM *a, int neg);
10855 and a macro that behave like
10856 int BN_is_negative(const BIGNUM *a);
10858 to avoid the need to access 'a->neg' directly in applications.
10862 * Implement fast modular reduction for pseudo-Mersenne primes
10863 used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c).
10864 EC_GROUP_new_curve_GFp() will now automatically use this
10867 *Nils Larsch <nla@trustcenter.de>*
10869 * Add new lock type (CRYPTO_LOCK_BN).
10873 * Change the ENGINE framework to automatically load engines
10874 dynamically from specific directories unless they could be
10875 found to already be built in or loaded. Move all the
10876 current engines except for the cryptodev one to a new
10877 directory engines/.
10878 The engines in engines/ are built as shared libraries if
10879 the "shared" options was given to ./Configure or ./config.
10880 Otherwise, they are inserted in libcrypto.a.
10881 /usr/local/ssl/engines is the default directory for dynamic
10882 engines, but that can be overridden at configure time through
10883 the usual use of --prefix and/or --openssldir, and at run
10884 time with the environment variable OPENSSL_ENGINES.
10886 *Geoff Thorpe and Richard Levitte*
10888 * Add Makefile.shared, a helper makefile to build shared
10889 libraries. Adapt Makefile.org.
10893 * Add version info to Win32 DLLs.
10895 *Peter 'Luna' Runestig" <peter@runestig.com>*
10897 * Add new 'medium level' PKCS#12 API. Certificates and keys
10898 can be added using this API to created arbitrary PKCS#12
10899 files while avoiding the low-level API.
10901 New options to PKCS12_create(), key or cert can be NULL and
10902 will then be omitted from the output file. The encryption
10903 algorithm NIDs can be set to -1 for no encryption, the mac
10904 iteration count can be set to 0 to omit the mac.
10906 Enhance pkcs12 utility by making the -nokeys and -nocerts
10907 options work when creating a PKCS#12 file. New option -nomac
10908 to omit the mac, NONE can be set for an encryption algorithm.
10909 New code is modified to use the enhanced PKCS12_create()
10910 instead of the low-level API.
10914 * Extend ASN1 encoder to support indefinite length constructed
10915 encoding. This can output sequences tags and octet strings in
10916 this form. Modify pk7_asn1.c to support indefinite length
10917 encoding. This is experimental and needs additional code to
10918 be useful, such as an ASN1 bio and some enhanced streaming
10921 Extend template encode functionality so that tagging is passed
10922 down to the template encoder.
10926 * Let 'openssl req' fail if an argument to '-newkey' is not
10927 recognized instead of using RSA as a default.
10931 * Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt.
10932 As these are not official, they are not included in "ALL";
10933 the "ECCdraft" ciphersuite group alias can be used to select them.
10935 *Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)*
10937 * Add ECDH engine support.
10939 *Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)*
10941 * Add ECDH in new directory crypto/ecdh/.
10943 *Douglas Stebila (Sun Microsystems Laboratories)*
10945 * Let BN_rand_range() abort with an error after 100 iterations
10946 without success (which indicates a broken PRNG).
10950 * Change BN_mod_sqrt() so that it verifies that the input value
10951 is really the square of the return value. (Previously,
10952 BN_mod_sqrt would show GIGO behaviour.)
10956 * Add named elliptic curves over binary fields from X9.62, SECG,
10957 and WAP/WTLS; add OIDs that were still missing.
10959 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10961 * Extend the EC library for elliptic curves over binary fields
10962 (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/).
10965 EC_GF2m_simple_method
10969 EC_GROUP_new_curve_GF2m
10970 EC_GROUP_set_curve_GF2m
10971 EC_GROUP_get_curve_GF2m
10972 EC_POINT_set_affine_coordinates_GF2m
10973 EC_POINT_get_affine_coordinates_GF2m
10974 EC_POINT_set_compressed_coordinates_GF2m
10976 Point compression for binary fields is disabled by default for
10977 patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to
10980 As binary polynomials are represented as BIGNUMs, various members
10981 of the EC_GROUP and EC_POINT data structures can be shared
10982 between the implementations for prime fields and binary fields;
10983 the above `..._GF2m functions` (except for EX_GROUP_new_curve_GF2m)
10984 are essentially identical to their `..._GFp` counterparts.
10985 (For simplicity, the `..._GFp` prefix has been dropped from
10986 various internal method names.)
10988 An internal 'field_div' method (similar to 'field_mul' and
10989 'field_sqr') has been added; this is used only for binary fields.
10991 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10993 * Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult()
10994 through methods ('mul', 'precompute_mult').
10996 The generic implementations (now internally called 'ec_wNAF_mul'
10997 and 'ec_wNAF_precomputed_mult') remain the default if these
10998 methods are undefined.
11000 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
11002 * New function EC_GROUP_get_degree, which is defined through
11003 EC_METHOD. For curves over prime fields, this returns the bit
11004 length of the modulus.
11006 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
11008 * New functions EC_GROUP_dup, EC_POINT_dup.
11009 (These simply call ..._new and ..._copy).
11011 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
11013 * Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c.
11014 Polynomials are represented as BIGNUMs (where the sign bit is not
11015 used) in the following functions [macros]:
11018 BN_GF2m_sub [= BN_GF2m_add]
11019 BN_GF2m_mod [wrapper for BN_GF2m_mod_arr]
11020 BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr]
11021 BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr]
11023 BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr]
11024 BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr]
11025 BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr]
11026 BN_GF2m_cmp [= BN_ucmp]
11028 (Note that only the 'mod' functions are actually for fields GF(2^m).
11029 BN_GF2m_add() is misnomer, but this is for the sake of consistency.)
11031 For some functions, an the irreducible polynomial defining a
11032 field can be given as an 'unsigned int[]' with strictly
11033 decreasing elements giving the indices of those bits that are set;
11034 i.e., p[] represents the polynomial
11035 f(t) = t^p[0] + t^p[1] + ... + t^p[k]
11037 p[0] > p[1] > ... > p[k] = 0.
11038 This applies to the following functions:
11041 BN_GF2m_mod_mul_arr
11042 BN_GF2m_mod_sqr_arr
11043 BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv]
11044 BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div]
11045 BN_GF2m_mod_exp_arr
11046 BN_GF2m_mod_sqrt_arr
11047 BN_GF2m_mod_solve_quad_arr
11051 Conversion can be performed by the following functions:
11056 bntest.c has additional tests for binary polynomial arithmetic.
11058 Two implementations for BN_GF2m_mod_div() are available.
11059 The default algorithm simply uses BN_GF2m_mod_inv() and
11060 BN_GF2m_mod_mul(). The alternative algorithm is compiled in only
11061 if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the
11062 copyright notice in crypto/bn/bn_gf2m.c before enabling it).
11064 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
11066 * Add new error code 'ERR_R_DISABLED' that can be used when some
11067 functionality is disabled at compile-time.
11069 *Douglas Stebila <douglas.stebila@sun.com>*
11071 * Change default behaviour of 'openssl asn1parse' so that more
11072 information is visible when viewing, e.g., a certificate:
11074 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump'
11075 mode the content of non-printable OCTET STRINGs is output in a
11076 style similar to INTEGERs, but with '[HEX DUMP]' prepended to
11077 avoid the appearance of a printable string.
11079 *Nils Larsch <nla@trustcenter.de>*
11081 * Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access
11083 EC_GROUP_set_asn1_flag()
11084 EC_GROUP_get_asn1_flag()
11085 EC_GROUP_set_point_conversion_form()
11086 EC_GROUP_get_point_conversion_form()
11087 These control ASN1 encoding details:
11088 - Curves (i.e., groups) are encoded explicitly unless asn1_flag
11089 has been set to OPENSSL_EC_NAMED_CURVE.
11090 - Points are encoded in uncompressed form by default; options for
11091 asn1_for are as for point2oct, namely
11092 POINT_CONVERSION_COMPRESSED
11093 POINT_CONVERSION_UNCOMPRESSED
11094 POINT_CONVERSION_HYBRID
11096 Also add 'seed' and 'seed_len' members to EC_GROUP with access
11098 EC_GROUP_set_seed()
11099 EC_GROUP_get0_seed()
11100 EC_GROUP_get_seed_len()
11101 This is used only for ASN1 purposes (so far).
11103 *Nils Larsch <nla@trustcenter.de>*
11105 * Add 'field_type' member to EC_METHOD, which holds the NID
11106 of the appropriate field type OID. The new function
11107 EC_METHOD_get_field_type() returns this value.
11109 *Nils Larsch <nla@trustcenter.de>*
11112 EC_POINT_point2bn()
11113 EC_POINT_bn2point()
11114 EC_POINT_point2hex()
11115 EC_POINT_hex2point()
11116 providing useful interfaces to EC_POINT_point2oct() and
11117 EC_POINT_oct2point().
11119 *Nils Larsch <nla@trustcenter.de>*
11121 * Change internals of the EC library so that the functions
11122 EC_GROUP_set_generator()
11123 EC_GROUP_get_generator()
11124 EC_GROUP_get_order()
11125 EC_GROUP_get_cofactor()
11126 are implemented directly in crypto/ec/ec_lib.c and not dispatched
11127 to methods, which would lead to unnecessary code duplication when
11128 adding different types of curves.
11130 *Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller*
11132 * Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM
11133 arithmetic, and such that modified wNAFs are generated
11134 (which avoid length expansion in many cases).
11138 * Add a function EC_GROUP_check_discriminant() (defined via
11139 EC_METHOD) that verifies that the curve discriminant is non-zero.
11141 Add a function EC_GROUP_check() that makes some sanity tests
11142 on a EC_GROUP, its generator and order. This includes
11143 EC_GROUP_check_discriminant().
11145 *Nils Larsch <nla@trustcenter.de>*
11147 * Add ECDSA in new directory crypto/ecdsa/.
11149 Add applications 'openssl ecparam' and 'openssl ecdsa'
11150 (these are based on 'openssl dsaparam' and 'openssl dsa').
11152 ECDSA support is also included in various other files across the
11153 library. Most notably,
11154 - 'openssl req' now has a '-newkey ecdsa:file' option;
11155 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA;
11156 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and
11157 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make
11158 them suitable for ECDSA where domain parameters must be
11159 extracted before the specific public key;
11160 - ECDSA engine support has been added.
11162 *Nils Larsch <nla@trustcenter.de>*
11164 * Include some named elliptic curves, and add OIDs from X9.62,
11165 SECG, and WAP/WTLS. Each curve can be obtained from the new
11167 EC_GROUP_new_by_curve_name(),
11168 and the list of available named curves can be obtained with
11169 EC_get_builtin_curves().
11170 Also add a 'curve_name' member to EC_GROUP objects, which can be
11172 EC_GROUP_set_curve_name()
11173 EC_GROUP_get_curve_name()
11175 *Nils Larsch <larsch@trustcenter.de, Bodo Moeller*
11177 * Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
11178 was actually never needed) and in BN_mul(). The removal in BN_mul()
11179 required a small change in bn_mul_part_recursive() and the addition
11180 of the functions bn_cmp_part_words(), bn_sub_part_words() and
11181 bn_add_part_words(), which do the same thing as bn_cmp_words(),
11182 bn_sub_words() and bn_add_words() except they take arrays with
11187 ### Changes between 0.9.7l and 0.9.7m [23 Feb 2007]
11189 * Cleanse PEM buffers before freeing them since they may contain
11192 *Benjamin Bennett <ben@psc.edu>*
11194 * Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
11195 a ciphersuite string such as "DEFAULT:RSA" cannot enable
11196 authentication-only ciphersuites.
11200 * Since AES128 and AES256 share a single mask bit in the logic of
11201 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
11202 kludge to work properly if AES128 is available and AES256 isn't.
11206 * Expand security boundary to match 1.1.1 module.
11210 * Remove redundant features: hash file source, editing of test vectors
11211 modify fipsld to use external fips_premain.c signature.
11215 * New perl script mkfipsscr.pl to create shell scripts or batch files to
11216 run algorithm test programs.
11220 * Make algorithm test programs more tolerant of whitespace.
11224 * Have SSL/TLS server implementation tolerate "mismatched" record
11225 protocol version while receiving ClientHello even if the
11226 ClientHello is fragmented. (The server can't insist on the
11227 particular protocol version it has chosen before the ServerHello
11228 message has informed the client about his choice.)
11232 * Load error codes if they are not already present instead of using a
11233 static variable. This allows them to be cleanly unloaded and reloaded.
11237 ### Changes between 0.9.7k and 0.9.7l [28 Sep 2006]
11239 * Introduce limits to prevent malicious keys being able to
11240 cause a denial of service. ([CVE-2006-2940])
11242 *Steve Henson, Bodo Moeller*
11244 * Fix ASN.1 parsing of certain invalid structures that can result
11245 in a denial of service. ([CVE-2006-2937]) [Steve Henson]
11247 * Fix buffer overflow in SSL_get_shared_ciphers() function.
11248 ([CVE-2006-3738]) [Tavis Ormandy and Will Drewry, Google Security Team]
11250 * Fix SSL client code which could crash if connecting to a
11251 malicious SSLv2 server. ([CVE-2006-4343])
11253 *Tavis Ormandy and Will Drewry, Google Security Team*
11255 * Change ciphersuite string processing so that an explicit
11256 ciphersuite selects this one ciphersuite (so that "AES256-SHA"
11257 will no longer include "AES128-SHA"), and any other similar
11258 ciphersuite (same bitmap) from *other* protocol versions (so that
11259 "RC4-MD5" will still include both the SSL 2.0 ciphersuite and the
11260 SSL 3.0/TLS 1.0 ciphersuite). This is a backport combining
11261 changes from 0.9.8b and 0.9.8d.
11265 ### Changes between 0.9.7j and 0.9.7k [05 Sep 2006]
11267 * Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
11268 ([CVE-2006-4339]) [Ben Laurie and Google Security Team]
11270 * Change the Unix randomness entropy gathering to use poll() when
11271 possible instead of select(), since the latter has some
11272 undesirable limitations.
11274 *Darryl Miles via Richard Levitte and Bodo Moeller*
11276 * Disable rogue ciphersuites:
11278 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
11279 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
11280 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
11282 The latter two were purportedly from
11283 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
11286 Also deactivate the remaining ciphersuites from
11287 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
11288 unofficial, and the ID has long expired.
11292 * Fix RSA blinding Heisenbug (problems sometimes occurred on
11293 dual-core machines) and other potential thread-safety issues.
11297 ### Changes between 0.9.7i and 0.9.7j [04 May 2006]
11299 * Adapt fipsld and the build system to link against the validated FIPS
11300 module in FIPS mode.
11304 * Fixes for VC++ 2005 build under Windows.
11308 * Add new Windows build target VC-32-GMAKE for VC++. This uses GNU make
11309 from a Windows bash shell such as MSYS. It is autodetected from the
11310 "config" script when run from a VC++ environment. Modify standard VC++
11311 build to use fipscanister.o from the GNU make build.
11315 ### Changes between 0.9.7h and 0.9.7i [14 Oct 2005]
11317 * Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS.
11318 The value now differs depending on if you build for FIPS or not.
11319 BEWARE! A program linked with a shared FIPSed libcrypto can't be
11320 safely run with a non-FIPSed libcrypto, as it may crash because of
11321 the difference induced by this change.
11325 ### Changes between 0.9.7g and 0.9.7h [11 Oct 2005]
11327 * Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
11328 (part of SSL_OP_ALL). This option used to disable the
11329 countermeasure against man-in-the-middle protocol-version
11330 rollback in the SSL 2.0 server implementation, which is a bad
11331 idea. ([CVE-2005-2969])
11333 *Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
11334 for Information Security, National Institute of Advanced Industrial
11335 Science and Technology [AIST, Japan)]*
11337 * Minimal support for X9.31 signatures and PSS padding modes. This is
11338 mainly for FIPS compliance and not fully integrated at this stage.
11342 * For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform
11343 the exponentiation using a fixed-length exponent. (Otherwise,
11344 the information leaked through timing could expose the secret key
11345 after many signatures; cf. Bleichenbacher's attack on DSA with
11350 * Make a new fixed-window mod_exp implementation the default for
11351 RSA, DSA, and DH private-key operations so that the sequence of
11352 squares and multiplies and the memory access pattern are
11353 independent of the particular secret key. This will mitigate
11354 cache-timing and potential related attacks.
11356 BN_mod_exp_mont_consttime() is the new exponentiation implementation,
11357 and this is automatically used by BN_mod_exp_mont() if the new flag
11358 BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH
11359 will use this BN flag for private exponents unless the flag
11360 RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or
11361 DH_FLAG_NO_EXP_CONSTTIME, respectively, is set.
11363 *Matthew D Wood (Intel Corp), with some changes by Bodo Moeller*
11365 * Change the client implementation for SSLv23_method() and
11366 SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0
11367 Client Hello message format if the SSL_OP_NO_SSLv2 option is set.
11368 (Previously, the SSL 2.0 backwards compatible Client Hello
11369 message format would be used even with SSL_OP_NO_SSLv2.)
11373 * Add support for smime-type MIME parameter in S/MIME messages which some
11378 * New function BN_MONT_CTX_set_locked() to set montgomery parameters in
11379 a threadsafe manner. Modify rsa code to use new function and add calls
11380 to dsa and dh code (which had race conditions before).
11384 * Include the fixed error library code in the C error file definitions
11385 instead of fixing them up at runtime. This keeps the error code
11386 structures constant.
11390 ### Changes between 0.9.7f and 0.9.7g [11 Apr 2005]
11392 [NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after
11395 * Fixes for newer kerberos headers. NB: the casts are needed because
11396 the 'length' field is signed on one version and unsigned on another
11397 with no (?) obvious way to tell the difference, without these VC++
11398 complains. Also the "definition" of FAR (blank) is no longer included
11399 nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up
11400 some needed definitions.
11404 * Undo Cygwin change.
11408 * Added support for proxy certificates according to RFC 3820.
11409 Because they may be a security thread to unaware applications,
11410 they must be explicitly allowed in run-time. See
11411 docs/HOWTO/proxy_certificates.txt for further information.
11415 ### Changes between 0.9.7e and 0.9.7f [22 Mar 2005]
11417 * Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating
11418 server and client random values. Previously
11419 (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in
11420 less random data when sizeof(time_t) > 4 (some 64 bit platforms).
11422 This change has negligible security impact because:
11424 1. Server and client random values still have 24 bytes of pseudo random
11427 2. Server and client random values are sent in the clear in the initial
11430 3. The master secret is derived using the premaster secret (48 bytes in
11431 size for static RSA ciphersuites) as well as client server and random
11434 The OpenSSL team would like to thank the UK NISCC for bringing this issue
11437 *Stephen Henson, reported by UK NISCC*
11439 * Use Windows randomness collection on Cygwin.
11443 * Fix hang in EGD/PRNGD query when communication socket is closed
11444 prematurely by EGD/PRNGD.
11446 *Darren Tucker <dtucker@zip.com.au> via Lutz Jänicke, resolves #1014*
11448 * Prompt for pass phrases when appropriate for PKCS12 input format.
11452 * Back-port of selected performance improvements from development
11453 branch, as well as improved support for PowerPC platforms.
11457 * Add lots of checks for memory allocation failure, error codes to indicate
11458 failure and freeing up memory if a failure occurs.
11460 *Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson*
11462 * Add new -passin argument to dgst.
11466 * Perform some character comparisons of different types in X509_NAME_cmp:
11467 this is needed for some certificates that re-encode DNs into UTF8Strings
11468 (in violation of RFC3280) and can't or won't issue name rollover
11473 * Make an explicit check during certificate validation to see that
11474 the CA setting in each certificate on the chain is correct. As a
11475 side effect always do the following basic checks on extensions,
11476 not just when there's an associated purpose to the check:
11478 - if there is an unhandled critical extension (unless the user
11479 has chosen to ignore this fault)
11480 - if the path length has been exceeded (if one is set at all)
11481 - that certain extensions fit the associated purpose (if one has
11486 ### Changes between 0.9.7d and 0.9.7e [25 Oct 2004]
11488 * Avoid a race condition when CRLs are checked in a multi threaded
11489 environment. This would happen due to the reordering of the revoked
11490 entries during signature checking and serial number lookup. Now the
11491 encoding is cached and the serial number sort performed under a lock.
11492 Add new STACK function sk_is_sorted().
11496 * Add Delta CRL to the extension code.
11500 * Various fixes to s3_pkt.c so alerts are sent properly.
11502 *David Holmes <d.holmes@f5.com>*
11504 * Reduce the chances of duplicate issuer name and serial numbers (in
11505 violation of RFC3280) using the OpenSSL certificate creation utilities.
11506 This is done by creating a random 64 bit value for the initial serial
11507 number when a serial number file is created or when a self signed
11508 certificate is created using 'openssl req -x509'. The initial serial
11509 number file is created using 'openssl x509 -next_serial' in CA.pl
11510 rather than being initialized to 1.
11514 ### Changes between 0.9.7c and 0.9.7d [17 Mar 2004]
11516 * Fix null-pointer assignment in do_change_cipher_spec() revealed
11517 by using the Codenomicon TLS Test Tool ([CVE-2004-0079])
11519 *Joe Orton, Steve Henson*
11521 * Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites
11524 *Joe Orton, Steve Henson*
11526 * Make it possible to have multiple active certificates with the same
11527 subject in the CA index file. This is done only if the keyword
11528 'unique_subject' is set to 'no' in the main CA section (default
11529 if 'CA_default') of the configuration file. The value is saved
11530 with the database itself in a separate index attribute file,
11531 named like the index file with '.attr' appended to the name.
11535 * X509 verify fixes. Disable broken certificate workarounds when
11536 X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if
11537 keyUsage extension present. Don't accept CRLs with unhandled critical
11538 extensions: since verify currently doesn't process CRL extensions this
11539 rejects a CRL with *any* critical extensions. Add new verify error codes
11544 * When creating an OCSP nonce use an OCTET STRING inside the extnValue.
11545 A clarification of RFC2560 will require the use of OCTET STRINGs and
11546 some implementations cannot handle the current raw format. Since OpenSSL
11547 copies and compares OCSP nonces as opaque blobs without any attempt at
11548 parsing them this should not create any compatibility issues.
11552 * New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when
11553 calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without
11554 this HMAC (and other) operations are several times slower than OpenSSL
11559 * Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex().
11561 *Peter Sylvester <Peter.Sylvester@EdelWeb.fr>*
11563 * Use the correct content when signing type "other".
11567 ### Changes between 0.9.7b and 0.9.7c [30 Sep 2003]
11569 * Fix various bugs revealed by running the NISCC test suite:
11571 Stop out of bounds reads in the ASN1 code when presented with
11572 invalid tags (CVE-2003-0543 and CVE-2003-0544).
11574 Free up ASN1_TYPE correctly if ANY type is invalid ([CVE-2003-0545]).
11576 If verify callback ignores invalid public key errors don't try to check
11577 certificate signature with the NULL public key.
11581 * New -ignore_err option in ocsp application to stop the server
11582 exiting on the first error in a request.
11586 * In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
11587 if the server requested one: as stated in TLS 1.0 and SSL 3.0
11592 * In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
11593 extra data after the compression methods not only for TLS 1.0
11594 but also for SSL 3.0 (as required by the specification).
11596 *Bodo Moeller; problem pointed out by Matthias Loepfe*
11598 * Change X509_certificate_type() to mark the key as exported/exportable
11599 when it's 512 *bits* long, not 512 bytes.
11603 * Change AES_cbc_encrypt() so it outputs exact multiple of
11604 blocks during encryption.
11608 * Various fixes to base64 BIO and non blocking I/O. On write
11609 flushes were not handled properly if the BIO retried. On read
11610 data was not being buffered properly and had various logic bugs.
11611 This also affects blocking I/O when the data being decoded is a
11616 * Various S/MIME bugfixes and compatibility changes:
11617 output correct application/pkcs7 MIME type if
11618 PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures.
11619 Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening
11620 of files as .eml work). Correctly handle very long lines in MIME
11625 ### Changes between 0.9.7a and 0.9.7b [10 Apr 2003]
11627 * Countermeasure against the Klima-Pokorny-Rosa extension of
11628 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
11629 a protocol version number mismatch like a decryption error
11630 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
11634 * Turn on RSA blinding by default in the default implementation
11635 to avoid a timing attack. Applications that don't want it can call
11636 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
11637 They would be ill-advised to do so in most cases.
11639 *Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller*
11641 * Change RSA blinding code so that it works when the PRNG is not
11642 seeded (in this case, the secret RSA exponent is abused as
11643 an unpredictable seed -- if it is not unpredictable, there
11644 is no point in blinding anyway). Make RSA blinding thread-safe
11645 by remembering the creator's thread ID in rsa->blinding and
11646 having all other threads use local one-time blinding factors
11647 (this requires more computation than sharing rsa->blinding, but
11648 avoids excessive locking; and if an RSA object is not shared
11649 between threads, blinding will still be very fast).
11653 * Fixed a typo bug that would cause ENGINE_set_default() to set an
11654 ENGINE as defaults for all supported algorithms irrespective of
11655 the 'flags' parameter. 'flags' is now honoured, so applications
11656 should make sure they are passing it correctly.
11660 * Target "mingw" now allows native Windows code to be generated in
11661 the Cygwin environment as well as with the MinGW compiler.
11665 ### Changes between 0.9.7 and 0.9.7a [19 Feb 2003]
11667 * In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
11668 via timing by performing a MAC computation even if incorrect
11669 block cipher padding has been found. This is a countermeasure
11670 against active attacks where the attacker has to distinguish
11671 between bad padding and a MAC verification error. ([CVE-2003-0078])
11673 *Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
11674 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
11675 Martin Vuagnoux (EPFL, Ilion)*
11677 * Make the no-err option work as intended. The intention with no-err
11678 is not to have the whole error stack handling routines removed from
11679 libcrypto, it's only intended to remove all the function name and
11680 reason texts, thereby removing some of the footprint that may not
11681 be interesting if those errors aren't displayed anyway.
11683 NOTE: it's still possible for any application or module to have its
11684 own set of error texts inserted. The routines are there, just not
11685 used by default when no-err is given.
11689 * Add support for FreeBSD on IA64.
11691 *dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454*
11693 * Adjust DES_cbc_cksum() so it returns the same value as the MIT
11694 Kerberos function mit_des_cbc_cksum(). Before this change,
11695 the value returned by DES_cbc_cksum() was like the one from
11696 mit_des_cbc_cksum(), except the bytes were swapped.
11698 *Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte*
11700 * Allow an application to disable the automatic SSL chain building.
11701 Before this a rather primitive chain build was always performed in
11702 ssl3_output_cert_chain(): an application had no way to send the
11703 correct chain if the automatic operation produced an incorrect result.
11705 Now the chain builder is disabled if either:
11707 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert().
11709 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set.
11711 The reasoning behind this is that an application would not want the
11712 auto chain building to take place if extra chain certificates are
11713 present and it might also want a means of sending no additional
11714 certificates (for example the chain has two certificates and the
11719 * Add the possibility to build without the ENGINE framework.
11721 *Steven Reddie <smr@essemer.com.au> via Richard Levitte*
11723 * Under Win32 gmtime() can return NULL: check return value in
11724 OPENSSL_gmtime(). Add error code for case where gmtime() fails.
11728 * DSA routines: under certain error conditions uninitialized BN objects
11729 could be freed. Solution: make sure initialization is performed early
11730 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>,
11731 Nils Larsch <nla@trustcenter.de> via PR#459)
11735 * Another fix for SSLv2 session ID handling: the session ID was incorrectly
11736 checked on reconnect on the client side, therefore session resumption
11737 could still fail with a "ssl session id is different" error. This
11738 behaviour is masked when SSL_OP_ALL is used due to
11739 SSL_OP_MICROSOFT_SESS_ID_BUG being set.
11740 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
11741 followup to PR #377.
11745 * IA-32 assembler support enhancements: unified ELF targets, support
11746 for SCO/Caldera platforms, fix for Cygwin shared build.
11750 * Add support for FreeBSD on sparc64. As a consequence, support for
11751 FreeBSD on non-x86 processors is separate from x86 processors on
11752 the config script, much like the NetBSD support.
11754 *Richard Levitte & Kris Kennaway <kris@obsecurity.org>*
11756 ### Changes between 0.9.6h and 0.9.7 [31 Dec 2002]
11758 [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after
11761 * Fix session ID handling in SSLv2 client code: the SERVER FINISHED
11762 code (06) was taken as the first octet of the session ID and the last
11763 octet was ignored consequently. As a result SSLv2 client side session
11764 caching could not have worked due to the session ID mismatch between
11766 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
11771 * Change the declaration of needed Kerberos libraries to use EX_LIBS
11772 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is
11777 * The hw_ncipher.c engine requires dynamic locks. Unfortunately, it
11778 seems that in spite of existing for more than a year, many application
11779 author have done nothing to provide the necessary callbacks, which
11780 means that this particular engine will not work properly anywhere.
11781 This is a very unfortunate situation which forces us, in the name
11782 of usability, to give the hw_ncipher.c a static lock, which is part
11784 NOTE: This is for the 0.9.7 series ONLY. This hack will never
11785 appear in 0.9.8 or later. We EXPECT application authors to have
11786 dealt properly with this when 0.9.8 is released (unless we actually
11787 make such changes in the libcrypto locking code that changes will
11788 have to be made anyway).
11792 * In asn1_d2i_read_bio() repeatedly call BIO_read() until all content
11793 octets have been read, EOF or an error occurs. Without this change
11794 some truncated ASN1 structures will not produce an error.
11798 * Disable Heimdal support, since it hasn't been fully implemented.
11799 Still give the possibility to force the use of Heimdal, but with
11800 warnings and a request that patches get sent to openssl-dev.
11804 * Add the VC-CE target, introduce the WINCE sysname, and add
11805 INSTALL.WCE and appropriate conditionals to make it build.
11807 *Steven Reddie <smr@essemer.com.au> via Richard Levitte*
11809 * Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and
11810 cygssl-x.y.z.dll, where x, y and z are the major, minor and
11811 edit numbers of the version.
11813 *Corinna Vinschen <vinschen@redhat.com> and Richard Levitte*
11815 * Introduce safe string copy and catenation functions
11816 (BUF_strlcpy() and BUF_strlcat()).
11818 *Ben Laurie (CHATS) and Richard Levitte*
11820 * Avoid using fixed-size buffers for one-line DNs.
11822 *Ben Laurie (CHATS)*
11824 * Add BUF_MEM_grow_clean() to avoid information leakage when
11825 resizing buffers containing secrets, and use where appropriate.
11827 *Ben Laurie (CHATS)*
11829 * Avoid using fixed size buffers for configuration file location.
11831 *Ben Laurie (CHATS)*
11833 * Avoid filename truncation for various CA files.
11835 *Ben Laurie (CHATS)*
11837 * Use sizeof in preference to magic numbers.
11839 *Ben Laurie (CHATS)*
11841 * Avoid filename truncation in cert requests.
11843 *Ben Laurie (CHATS)*
11845 * Add assertions to check for (supposedly impossible) buffer
11848 *Ben Laurie (CHATS)*
11850 * Don't cache truncated DNS entries in the local cache (this could
11851 potentially lead to a spoofing attack).
11853 *Ben Laurie (CHATS)*
11855 * Fix various buffers to be large enough for hex/decimal
11856 representations in a platform independent manner.
11858 *Ben Laurie (CHATS)*
11860 * Add CRYPTO_realloc_clean() to avoid information leakage when
11861 resizing buffers containing secrets, and use where appropriate.
11863 *Ben Laurie (CHATS)*
11865 * Add BIO_indent() to avoid much slightly worrying code to do
11868 *Ben Laurie (CHATS)*
11870 * Convert sprintf()/BIO_puts() to BIO_printf().
11872 *Ben Laurie (CHATS)*
11874 * buffer_gets() could terminate with the buffer only half
11877 *Ben Laurie (CHATS)*
11879 * Add assertions to prevent user-supplied crypto functions from
11880 overflowing internal buffers by having large block sizes, etc.
11882 *Ben Laurie (CHATS)*
11884 * New OPENSSL_assert() macro (similar to assert(), but enabled
11887 *Ben Laurie (CHATS)*
11889 * Eliminate unused copy of key in RC4.
11891 *Ben Laurie (CHATS)*
11893 * Eliminate unused and incorrectly sized buffers for IV in pem.h.
11895 *Ben Laurie (CHATS)*
11897 * Fix off-by-one error in EGD path.
11899 *Ben Laurie (CHATS)*
11901 * If RANDFILE path is too long, ignore instead of truncating.
11903 *Ben Laurie (CHATS)*
11905 * Eliminate unused and incorrectly sized X.509 structure
11908 *Ben Laurie (CHATS)*
11910 * Eliminate unused and dangerous function knumber().
11912 *Ben Laurie (CHATS)*
11914 * Eliminate unused and dangerous structure, KSSL_ERR.
11916 *Ben Laurie (CHATS)*
11918 * Protect against overlong session ID context length in an encoded
11919 session object. Since these are local, this does not appear to be
11922 *Ben Laurie (CHATS)*
11924 * Change from security patch (see 0.9.6e below) that did not affect
11925 the 0.9.6 release series:
11927 Remote buffer overflow in SSL3 protocol - an attacker could
11928 supply an oversized master key in Kerberos-enabled versions.
11931 *Ben Laurie (CHATS)*
11933 * Change the SSL kerb5 codes to match RFC 2712.
11937 * Make -nameopt work fully for req and add -reqopt switch.
11939 *Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson*
11941 * The "block size" for block ciphers in CFB and OFB mode should be 1.
11943 *Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>*
11945 * Make sure tests can be performed even if the corresponding algorithms
11946 have been removed entirely. This was also the last step to make
11947 OpenSSL compilable with DJGPP under all reasonable conditions.
11949 *Richard Levitte, Doug Kaufman <dkaufman@rahul.net>*
11951 * Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT
11952 to allow version independent disabling of normally unselected ciphers,
11953 which may be activated as a side-effect of selecting a single cipher.
11955 (E.g., cipher list string "RSA" enables ciphersuites that are left
11956 out of "ALL" because they do not provide symmetric encryption.
11957 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.)
11959 *Lutz Jaenicke, Bodo Moeller*
11961 * Add appropriate support for separate platform-dependent build
11962 directories. The recommended way to make a platform-dependent
11963 build directory is the following (tested on Linux), maybe with
11966 # Place yourself outside of the OpenSSL source tree. In
11967 # this example, the environment variable OPENSSL_SOURCE
11968 # is assumed to contain the absolute OpenSSL source directory.
11969 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`"
11970 cd objtree/"`uname -s`-`uname -r`-`uname -m`"
11971 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do
11972 mkdir -p `dirname $F`
11973 ln -s $OPENSSL_SOURCE/$F $F
11976 To be absolutely sure not to disturb the source tree, a "make clean"
11977 is a good thing. If it isn't successful, don't worry about it,
11978 it probably means the source directory is very clean.
11982 * Make sure any ENGINE control commands make local copies of string
11983 pointers passed to them whenever necessary. Otherwise it is possible
11984 the caller may have overwritten (or deallocated) the original string
11985 data when a later ENGINE operation tries to use the stored values.
11987 *Götz Babin-Ebell <babinebell@trustcenter.de>*
11989 * Improve diagnostics in file reading and command-line digests.
11991 *Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>*
11993 * Add AES modes CFB and OFB to the object database. Correct an
11994 error in AES-CFB decryption.
11998 * Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this
11999 allows existing EVP_CIPHER_CTX structures to be reused after
12000 calling `EVP_*Final()`. This behaviour is used by encryption
12001 BIOs and some applications. This has the side effect that
12002 applications must explicitly clean up cipher contexts with
12003 EVP_CIPHER_CTX_cleanup() or they will leak memory.
12007 * Check the values of dna and dnb in bn_mul_recursive before calling
12008 bn_mul_comba (a non zero value means the a or b arrays do not contain
12009 n2 elements) and fallback to bn_mul_normal if either is not zero.
12013 * Fix escaping of non-ASCII characters when using the -subj option
12014 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>)
12018 * Make object definitions compliant to LDAP (RFC2256): SN is the short
12019 form for "surname", serialNumber has no short form.
12020 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798;
12021 therefore remove "mail" short name for "internet 7".
12022 The OID for unique identifiers in X509 certificates is
12023 x500UniqueIdentifier, not uniqueIdentifier.
12024 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>)
12028 * Add an "init" command to the ENGINE config module and auto initialize
12029 ENGINEs. Without any "init" command the ENGINE will be initialized
12030 after all ctrl commands have been executed on it. If init=1 the
12031 ENGINE is initialized at that point (ctrls before that point are run
12032 on the uninitialized ENGINE and after on the initialized one). If
12033 init=0 then the ENGINE will not be initialized at all.
12037 * Fix the 'app_verify_callback' interface so that the user-defined
12038 argument is actually passed to the callback: In the
12039 SSL_CTX_set_cert_verify_callback() prototype, the callback
12040 declaration has been changed from
12043 int (*cb)(X509_STORE_CTX *,void *);
12044 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
12045 i=s->ctx->app_verify_callback(&ctx)
12046 has been changed into
12047 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
12049 To update applications using SSL_CTX_set_cert_verify_callback(),
12050 a dummy argument can be added to their callback functions.
12052 *D. K. Smetters <smetters@parc.xerox.com>*
12054 * Added the '4758cca' ENGINE to support IBM 4758 cards.
12056 *Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe*
12058 * Add and OPENSSL_LOAD_CONF define which will cause
12059 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
12060 This allows older applications to transparently support certain
12061 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
12062 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
12063 load the config file and OPENSSL_add_all_algorithms_conf() which will
12064 always load it have also been added.
12068 * Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
12069 Adjust NIDs and EVP layer.
12071 *Stephen Sprunk <stephen@sprunk.org> and Richard Levitte*
12073 * Config modules support in openssl utility.
12075 Most commands now load modules from the config file,
12076 though in a few (such as version) this isn't done
12077 because it couldn't be used for anything.
12079 In the case of ca and req the config file used is
12080 the same as the utility itself: that is the -config
12081 command line option can be used to specify an
12086 * Move default behaviour from OPENSSL_config(). If appname is NULL
12087 use "openssl_conf" if filename is NULL use default openssl config file.
12091 * Add an argument to OPENSSL_config() to allow the use of an alternative
12092 config section name. Add a new flag to tolerate a missing config file
12093 and move code to CONF_modules_load_file().
12097 * Support for crypto accelerator cards from Accelerated Encryption
12098 Processing, www.aep.ie. (Use engine 'aep')
12099 The support was copied from 0.9.6c [engine] and adapted/corrected
12100 to work with the new engine framework.
12102 *AEP Inc. and Richard Levitte*
12104 * Support for SureWare crypto accelerator cards from Baltimore
12105 Technologies. (Use engine 'sureware')
12106 The support was copied from 0.9.6c [engine] and adapted
12107 to work with the new engine framework.
12111 * Have the CHIL engine fork-safe (as defined by nCipher) and actually
12112 make the newer ENGINE framework commands for the CHIL engine work.
12114 *Toomas Kiisk <vix@cyber.ee> and Richard Levitte*
12116 * Make it possible to produce shared libraries on ReliantUNIX.
12118 *Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte*
12120 * Add the configuration target debug-linux-ppro.
12121 Make 'openssl rsa' use the general key loading routines
12122 implemented in `apps.c`, and make those routines able to
12123 handle the key format FORMAT_NETSCAPE and the variant
12126 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
12128 * Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
12130 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
12132 * Add -keyform to rsautl, and document -engine.
12134 *Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>*
12136 * Change BIO_new_file (crypto/bio/bss_file.c) to use new
12137 BIO_R_NO_SUCH_FILE error code rather than the generic
12138 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
12142 * Add new functions
12143 ERR_peek_last_error
12144 ERR_peek_last_error_line
12145 ERR_peek_last_error_line_data.
12146 These are similar to
12148 ERR_peek_error_line
12149 ERR_peek_error_line_data,
12150 but report on the latest error recorded rather than the first one
12151 still in the error queue.
12153 *Ben Laurie, Bodo Moeller*
12155 * default_algorithms option in ENGINE config module. This allows things
12157 default_algorithms = ALL
12158 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
12162 * Preliminary ENGINE config module.
12166 * New experimental application configuration code.
12170 * Change the AES code to follow the same name structure as all other
12171 symmetric ciphers, and behave the same way. Move everything to
12172 the directory crypto/aes, thereby obsoleting crypto/rijndael.
12174 *Stephen Sprunk <stephen@sprunk.org> and Richard Levitte*
12176 * SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
12178 *Ben Laurie and Theo de Raadt*
12180 * Add option to output public keys in req command.
12182 *Massimiliano Pala madwolf@openca.org*
12184 * Use wNAFs in EC_POINTs_mul() for improved efficiency
12185 (up to about 10% better than before for P-192 and P-224).
12189 * New functions/macros
12191 SSL_CTX_set_msg_callback(ctx, cb)
12192 SSL_CTX_set_msg_callback_arg(ctx, arg)
12193 SSL_set_msg_callback(ssl, cb)
12194 SSL_set_msg_callback_arg(ssl, arg)
12196 to request calling a callback function
12198 void cb(int write_p, int version, int content_type,
12199 const void *buf, size_t len, SSL *ssl, void *arg)
12201 whenever a protocol message has been completely received
12202 (write_p == 0) or sent (write_p == 1). Here 'version' is the
12203 protocol version according to which the SSL library interprets
12204 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
12205 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
12206 the content type as defined in the SSL 3.0/TLS 1.0 protocol
12207 specification (change_cipher_spec(20), alert(21), handshake(22)).
12208 'buf' and 'len' point to the actual message, 'ssl' to the
12209 SSL object, and 'arg' is the application-defined value set by
12210 SSL[_CTX]_set_msg_callback_arg().
12212 'openssl s_client' and 'openssl s_server' have new '-msg' options
12213 to enable a callback that displays all protocol messages.
12217 * Change the shared library support so shared libraries are built as
12218 soon as the corresponding static library is finished, and thereby get
12219 openssl and the test programs linked against the shared library.
12220 This still only happens when the keyword "shard" has been given to
12221 the configuration scripts.
12223 NOTE: shared library support is still an experimental thing, and
12224 backward binary compatibility is still not guaranteed.
12226 *"Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte*
12228 * Add support for Subject Information Access extension.
12230 *Peter Sylvester <Peter.Sylvester@EdelWeb.fr>*
12232 * Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
12233 additional bytes when new memory had to be allocated, not just
12234 when reusing an existing buffer.
12238 * New command line and configuration option 'utf8' for the req command.
12239 This allows field values to be specified as UTF8 strings.
12243 * Add -multi and -mr options to "openssl speed" - giving multiple parallel
12244 runs for the former and machine-readable output for the latter.
12248 * Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
12249 of the e-mail address in the DN (i.e., it will go into a certificate
12250 extension only). The new configuration file option 'email_in_dn = no'
12251 has the same effect.
12253 *Massimiliano Pala madwolf@openca.org*
12255 * Change all functions with names starting with `des_` to be starting
12256 with `DES_` instead. Add wrappers that are compatible with libdes,
12257 but are named `_ossl_old_des_*`. Finally, add macros that map the
12258 `des_*` symbols to the corresponding `_ossl_old_des_*` if libdes
12259 compatibility is desired. If OpenSSL 0.9.6c compatibility is
12260 desired, the `des_*` symbols will be mapped to `DES_*`, with one
12263 Since we provide two compatibility mappings, the user needs to
12264 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
12265 compatibility is desired. The default (i.e., when that macro
12266 isn't defined) is OpenSSL 0.9.6c compatibility.
12268 There are also macros that enable and disable the support of old
12269 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
12270 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
12271 are defined, the default will apply: to support the old des routines.
12273 In either case, one must include openssl/des.h to get the correct
12274 definitions. Do not try to just include openssl/des_old.h, that
12277 NOTE: This is a major break of an old API into a new one. Software
12278 authors are encouraged to switch to the `DES_` style functions. Some
12279 time in the future, des_old.h and the libdes compatibility functions
12280 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
12281 default), and then completely removed.
12285 * Test for certificates which contain unsupported critical extensions.
12286 If such a certificate is found during a verify operation it is
12287 rejected by default: this behaviour can be overridden by either
12288 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
12289 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
12290 X509_supported_extension() has also been added which returns 1 if a
12291 particular extension is supported.
12295 * Modify the behaviour of EVP cipher functions in similar way to digests
12296 to retain compatibility with existing code.
12300 * Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
12301 compatibility with existing code. In particular the 'ctx' parameter does
12302 not have to be to be initialized before the call to EVP_DigestInit() and
12303 it is tidied up after a call to EVP_DigestFinal(). New function
12304 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
12305 EVP_MD_CTX_copy() changed to not require the destination to be
12306 initialized valid and new function EVP_MD_CTX_copy_ex() added which
12307 requires the destination to be valid.
12309 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
12310 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
12314 * Change ssl3_get_message (ssl/s3_both.c) and the functions using it
12315 so that complete 'Handshake' protocol structures are kept in memory
12316 instead of overwriting 'msg_type' and 'length' with 'body' data.
12320 * Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
12322 *Massimo Santin via Richard Levitte*
12324 * Major restructuring to the underlying ENGINE code. This includes
12325 reduction of linker bloat, separation of pure "ENGINE" manipulation
12326 (initialisation, etc) from functionality dealing with implementations
12327 of specific crypto interfaces. This change also introduces integrated
12328 support for symmetric ciphers and digest implementations - so ENGINEs
12329 can now accelerate these by providing EVP_CIPHER and EVP_MD
12330 implementations of their own. This is detailed in
12331 [crypto/engine/README.md](crypto/engine/README.md)
12332 as it couldn't be adequately described here. However, there are a few
12333 API changes worth noting - some RSA, DSA, DH, and RAND functions that
12334 were changed in the original introduction of ENGINE code have now
12335 reverted back - the hooking from this code to ENGINE is now a good
12336 deal more passive and at run-time, operations deal directly with
12337 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
12338 dereferencing through an ENGINE pointer any more. Also, the ENGINE
12339 functions dealing with `BN_MOD_EXP[_CRT]` handlers have been removed -
12340 they were not being used by the framework as there is no concept of a
12341 BIGNUM_METHOD and they could not be generalised to the new
12342 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
12343 ENGINE_cpy() has been removed as it cannot be consistently defined in
12348 * Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
12352 * Change mkdef.pl to sort symbols that get the same entry number,
12353 and make sure the automatically generated functions `ERR_load_*`
12354 become part of libeay.num as well.
12358 * New function SSL_renegotiate_pending(). This returns true once
12359 renegotiation has been requested (either SSL_renegotiate() call
12360 or HelloRequest/ClientHello received from the peer) and becomes
12361 false once a handshake has been completed.
12362 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
12363 sends a HelloRequest, but does not ensure that a handshake takes
12364 place. SSL_renegotiate_pending() is useful for checking if the
12365 client has followed the request.)
12369 * New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
12370 By default, clients may request session resumption even during
12371 renegotiation (if session ID contexts permit); with this option,
12372 session resumption is possible only in the first handshake.
12374 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes
12375 more bits available for options that should not be part of
12376 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION).
12380 * Add some demos for certificate and certificate request creation.
12384 * Make maximum certificate chain size accepted from the peer application
12385 settable (`SSL*_get/set_max_cert_list()`), as proposed by
12386 "Douglas E. Engert" <deengert@anl.gov>.
12390 * Add support for shared libraries for Unixware-7
12391 (Boyd Lynn Gerber <gerberb@zenez.com>).
12395 * Add a "destroy" handler to ENGINEs that allows structural cleanup to
12396 be done prior to destruction. Use this to unload error strings from
12397 ENGINEs that load their own error strings. NB: This adds two new API
12398 functions to "get" and "set" this destroy handler in an ENGINE.
12402 * Alter all existing ENGINE implementations (except "openssl" and
12403 "openbsd") to dynamically instantiate their own error strings. This
12404 makes them more flexible to be built both as statically-linked ENGINEs
12405 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
12406 Also, add stub code to each that makes building them as self-contained
12407 shared-libraries easier (see [README-Engine.md](README-Engine.md)).
12411 * Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
12412 implementations into applications that are completely implemented in
12413 self-contained shared-libraries. The "dynamic" ENGINE exposes control
12414 commands that can be used to configure what shared-library to load and
12415 to control aspects of the way it is handled. Also, made an update to
12416 the [README-Engine.md](README-Engine.md) file
12417 that brings its information up-to-date and
12418 provides some information and instructions on the "dynamic" ENGINE
12419 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
12423 * Make it possible to unload ranges of ERR strings with a new
12424 "ERR_unload_strings" function.
12428 * Add a copy() function to EVP_MD.
12432 * Make EVP_MD routines take a context pointer instead of just the
12433 md_data void pointer.
12437 * Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
12438 that the digest can only process a single chunk of data
12439 (typically because it is provided by a piece of
12440 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
12441 is only going to provide a single chunk of data, and hence the
12442 framework needn't accumulate the data for oneshot drivers.
12446 * As with "ERR", make it possible to replace the underlying "ex_data"
12447 functions. This change also alters the storage and management of global
12448 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
12449 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
12450 index counters. The API functions that use this state have been changed
12451 to take a "class_index" rather than pointers to the class's local STACK
12452 and counter, and there is now an API function to dynamically create new
12453 classes. This centralisation allows us to (a) plug a lot of the
12454 thread-safety problems that existed, and (b) makes it possible to clean
12455 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
12456 such data would previously have always leaked in application code and
12457 workarounds were in place to make the memory debugging turn a blind eye
12458 to it. Application code that doesn't use this new function will still
12459 leak as before, but their memory debugging output will announce it now
12460 rather than letting it slide.
12462 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
12463 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
12464 has a return value to indicate success or failure.
12468 * Make it possible to replace the underlying "ERR" functions such that the
12469 global state (2 LHASH tables and 2 locks) is only used by the "default"
12470 implementation. This change also adds two functions to "get" and "set"
12471 the implementation prior to it being automatically set the first time
12472 any other ERR function takes place. Ie. an application can call "get",
12473 pass the return value to a module it has just loaded, and that module
12474 can call its own "set" function using that value. This means the
12475 module's "ERR" operations will use (and modify) the error state in the
12476 application and not in its own statically linked copy of OpenSSL code.
12480 * Give DH, DSA, and RSA types their own `*_up_ref()` function to increment
12481 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
12482 the operation, and provides a more encapsulated way for external code
12483 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
12484 to use these functions rather than manually incrementing the counts.
12486 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
12490 * Add EVP test program.
12494 * Add symmetric cipher support to ENGINE. Expect the API to change!
12498 * New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
12499 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
12500 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
12501 These allow a CRL to be built without having to access X509_CRL fields
12502 directly. Modify 'ca' application to use new functions.
12506 * Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
12507 bug workarounds. Rollback attack detection is a security feature.
12508 The problem will only arise on OpenSSL servers when TLSv1 is not
12509 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
12510 Software authors not wanting to support TLSv1 will have special reasons
12511 for their choice and can explicitly enable this option.
12513 *Bodo Moeller, Lutz Jaenicke*
12515 * Rationalise EVP so it can be extended: don't include a union of
12516 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX
12517 (similar to those existing for EVP_CIPHER_CTX).
12522 EVP_MD_CTX_init(&md); /* new function call */
12523 EVP_DigestInit(&md, EVP_sha1());
12524 EVP_DigestUpdate(&md, in, len);
12525 EVP_DigestFinal(&md, out, NULL);
12526 EVP_MD_CTX_cleanup(&md); /* new function call */
12530 * Make DES key schedule conform to the usual scheme, as well as
12531 correcting its structure. This means that calls to DES functions
12532 now have to pass a pointer to a des_key_schedule instead of a
12533 plain des_key_schedule (which was actually always a pointer
12536 des_key_schedule ks;
12538 des_set_key_checked(..., &ks);
12539 des_ncbc_encrypt(..., &ks, ...);
12541 (Note that a later change renames 'des_...' into 'DES_...'.)
12545 * Initial reduction of linker bloat: the use of some functions, such as
12546 PEM causes large amounts of unused functions to be linked in due to
12547 poor organisation. For example pem_all.c contains every PEM function
12548 which has a knock on effect of linking in large amounts of (unused)
12549 ASN1 code. Grouping together similar functions and splitting unrelated
12550 functions prevents this.
12554 * Cleanup of EVP macros.
12558 * Change historical references to `{NID,SN,LN}_des_ede` and ede3 to add the
12559 correct `_ecb suffix`.
12563 * Add initial OCSP responder support to ocsp application. The
12564 revocation information is handled using the text based index
12565 use by the ca application. The responder can either handle
12566 requests generated internally, supplied in files (for example
12567 via a CGI script) or using an internal minimal server.
12571 * Add configuration choices to get zlib compression for TLS.
12575 * Changes to Kerberos SSL for RFC 2712 compliance:
12576 1. Implemented real KerberosWrapper, instead of just using
12577 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
12578 2. Implemented optional authenticator field of KerberosWrapper.
12580 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
12581 and authenticator structs; see crypto/krb5/.
12583 Generalized Kerberos calls to support multiple Kerberos libraries.
12584 *Vern Staats <staatsvr@asc.hpc.mil>, Jeffrey Altman <jaltman@columbia.edu>
12585 via Richard Levitte*
12587 * Cause 'openssl speed' to use fully hard-coded DSA keys as it
12588 already does with RSA. testdsa.h now has 'priv_key/pub_key'
12589 values for each of the key sizes rather than having just
12590 parameters (and 'speed' generating keys each time).
12594 * Speed up EVP routines.
12597 pe 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
12598 s-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
12599 s-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
12600 s-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
12602 s-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
12603 s-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
12604 s-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
12607 s-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
12609 s-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
12613 * Added the OS2-EMX target.
12615 *"Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte*
12617 * Rewrite commands to use `NCONF` routines instead of the old `CONF`.
12618 New functions to support `NCONF` routines in extension code.
12619 New function `CONF_set_nconf()`
12620 to allow functions which take an `NCONF` to also handle the old `LHASH`
12621 structure: this means that the old `CONF` compatible routines can be
12622 retained (in particular w.rt. extensions) without having to duplicate the
12623 code. New function `X509V3_add_ext_nconf_sk()` to add extensions to a stack.
12627 * Enhance the general user interface with mechanisms for inner control
12628 and with possibilities to have yes/no kind of prompts.
12632 * Change all calls to low-level digest routines in the library and
12633 applications to use EVP. Add missing calls to HMAC_cleanup() and
12634 don't assume HMAC_CTX can be copied using memcpy().
12636 *Verdon Walker <VWalker@novell.com>, Steve Henson*
12638 * Add the possibility to control engines through control names but with
12639 arbitrary arguments instead of just a string.
12640 Change the key loaders to take a UI_METHOD instead of a callback
12641 function pointer. NOTE: this breaks binary compatibility with earlier
12642 versions of OpenSSL [engine].
12643 Adapt the nCipher code for these new conditions and add a card insertion
12648 * Enhance the general user interface with mechanisms to better support
12649 dialog box interfaces, application-defined prompts, the possibility
12650 to use defaults (for example default passwords from somewhere else)
12651 and interrupts/cancellations.
12655 * Tidy up PKCS#12 attribute handling. Add support for the CSP name
12656 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
12660 * Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
12661 tidy up some unnecessarily weird code in 'sk_new()').
12663 *Geoff, reported by Diego Tartara <dtartara@novamens.com>*
12665 * Change the key loading routines for ENGINEs to use the same kind
12666 callback (pem_password_cb) as all other routines that need this
12671 * Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
12672 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
12673 than this minimum value is recommended.
12677 * New random seeder for OpenVMS, using the system process statistics
12678 that are easily reachable.
12682 * Windows apparently can't transparently handle global
12683 variables defined in DLLs. Initialisations such as:
12685 const ASN1_ITEM *it = &ASN1_INTEGER_it;
12687 won't compile. This is used by the any applications that need to
12688 declare their own ASN1 modules. This was fixed by adding the option
12689 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
12690 needed for static libraries under Win32.
12694 * New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
12695 setting of purpose and trust fields. New X509_STORE trust and
12696 purpose functions and tidy up setting in other SSL functions.
12700 * Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
12701 structure. These are inherited by X509_STORE_CTX when it is
12702 initialised. This allows various defaults to be set in the
12703 X509_STORE structure (such as flags for CRL checking and custom
12704 purpose or trust settings) for functions which only use X509_STORE_CTX
12705 internally such as S/MIME.
12707 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
12708 trust settings if they are not set in X509_STORE. This allows X509_STORE
12709 purposes and trust (in S/MIME for example) to override any set by default.
12711 Add command line options for CRL checking to smime, s_client and s_server
12716 * Initial CRL based revocation checking. If the CRL checking flag(s)
12717 are set then the CRL is looked up in the X509_STORE structure and
12718 its validity and signature checked, then if the certificate is found
12719 in the CRL the verify fails with a revoked error.
12721 Various new CRL related callbacks added to X509_STORE_CTX structure.
12723 Command line options added to 'verify' application to support this.
12725 This needs some additional work, such as being able to handle multiple
12726 CRLs with different times, extension based lookup (rather than just
12727 by subject name) and ultimately more complete V2 CRL extension
12732 * Add a general user interface API (crypto/ui/). This is designed
12733 to replace things like des_read_password and friends (backward
12734 compatibility functions using this new API are provided).
12735 The purpose is to remove prompting functions from the DES code
12736 section as well as provide for prompting through dialog boxes in
12737 a window system and the like.
12741 * Add "ex_data" support to ENGINE so implementations can add state at a
12742 per-structure level rather than having to store it globally.
12746 * Make it possible for ENGINE structures to be copied when retrieved by
12747 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
12748 This causes the "original" ENGINE structure to act like a template,
12749 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
12750 operational state can be localised to each ENGINE structure, despite the
12751 fact they all share the same "methods". New ENGINE structures returned in
12752 this case have no functional references and the return value is the single
12753 structural reference. This matches the single structural reference returned
12754 by ENGINE_by_id() normally, when it is incremented on the pre-existing
12759 * Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
12760 needs to match any other type at all we need to manually clear the
12765 * Changes to the "openssl engine" utility to include;
12766 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
12767 about an ENGINE's available control commands.
12768 - executing control commands from command line arguments using the
12769 '-pre' and '-post' switches. '-post' is only used if '-t' is
12770 specified and the ENGINE is successfully initialised. The syntax for
12771 the individual commands are colon-separated, for example;
12772 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
12776 * New dynamic control command support for ENGINEs. ENGINEs can now
12777 declare their own commands (numbers), names (strings), descriptions,
12778 and input types for run-time discovery by calling applications. A
12779 subset of these commands are implicitly classed as "executable"
12780 depending on their input type, and only these can be invoked through
12781 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
12782 can be based on user input, config files, etc). The distinction is
12783 that "executable" commands cannot return anything other than a boolean
12784 result and can only support numeric or string input, whereas some
12785 discoverable commands may only be for direct use through
12786 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
12787 pointers, or other custom uses. The "executable" commands are to
12788 support parameterisations of ENGINE behaviour that can be
12789 unambiguously defined by ENGINEs and used consistently across any
12790 OpenSSL-based application. Commands have been added to all the
12791 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
12792 control over shared-library paths without source code alterations.
12796 * Changed all ENGINE implementations to dynamically allocate their
12797 ENGINEs rather than declaring them statically. Apart from this being
12798 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
12799 this also allows the implementations to compile without using the
12800 internal engine_int.h header.
12804 * Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
12805 'const' value. Any code that should be able to modify a RAND_METHOD
12806 should already have non-const pointers to it (ie. they should only
12807 modify their own ones).
12811 * Made a variety of little tweaks to the ENGINE code.
12812 - "atalla" and "ubsec" string definitions were moved from header files
12813 to C code. "nuron" string definitions were placed in variables
12814 rather than hard-coded - allowing parameterisation of these values
12815 later on via ctrl() commands.
12816 - Removed unused "#if 0"'d code.
12817 - Fixed engine list iteration code so it uses ENGINE_free() to release
12818 structural references.
12819 - Constified the RAND_METHOD element of ENGINE structures.
12820 - Constified various get/set functions as appropriate and added
12821 missing functions (including a catch-all ENGINE_cpy that duplicates
12822 all ENGINE values onto a new ENGINE except reference counts/state).
12823 - Removed NULL parameter checks in get/set functions. Setting a method
12824 or function to NULL is a way of cancelling out a previously set
12825 value. Passing a NULL ENGINE parameter is just plain stupid anyway
12826 and doesn't justify the extra error symbols and code.
12827 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
12828 flags from engine_int.h to engine.h.
12829 - Changed prototypes for ENGINE handler functions (init(), finish(),
12830 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
12834 * Implement binary inversion algorithm for BN_mod_inverse in addition
12835 to the algorithm using long division. The binary algorithm can be
12836 used only if the modulus is odd. On 32-bit systems, it is faster
12837 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
12838 roughly 5-15% for 256-bit moduli), so we use it only for moduli
12839 up to 450 bits. In 64-bit environments, the binary algorithm
12840 appears to be advantageous for much longer moduli; here we use it
12841 for moduli up to 2048 bits.
12845 * Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
12846 could not support the combine flag in choice fields.
12850 * Add a 'copy_extensions' option to the 'ca' utility. This copies
12851 extensions from a certificate request to the certificate.
12855 * Allow multiple 'certopt' and 'nameopt' options to be separated
12856 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
12857 file: this allows the display of the certificate about to be
12858 signed to be customised, to allow certain fields to be included
12859 or excluded and extension details. The old system didn't display
12860 multicharacter strings properly, omitted fields not in the policy
12861 and couldn't display additional details such as extensions.
12865 * Function EC_POINTs_mul for multiple scalar multiplication
12866 of an arbitrary number of elliptic curve points
12867 \sum scalars[i]*points[i],
12868 optionally including the generator defined for the EC_GROUP:
12869 scalar*generator + \sum scalars[i]*points[i].
12871 EC_POINT_mul is a simple wrapper function for the typical case
12872 that the point list has just one item (besides the optional
12877 * First EC_METHODs for curves over GF(p):
12879 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
12880 operations and provides various method functions that can also
12881 operate with faster implementations of modular arithmetic.
12883 EC_GFp_mont_method() reuses most functions that are part of
12884 EC_GFp_simple_method, but uses Montgomery arithmetic.
12886 *Bodo Moeller; point addition and point doubling
12887 implementation directly derived from source code provided by
12888 Lenka Fibikova <fibikova@exp-math.uni-essen.de>*
12890 * Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
12891 crypto/ec/ec_lib.c):
12893 Curves are EC_GROUP objects (with an optional group generator)
12894 based on EC_METHODs that are built into the library.
12896 Points are EC_POINT objects based on EC_GROUP objects.
12898 Most of the framework would be able to handle curves over arbitrary
12899 finite fields, but as there are no obvious types for fields other
12900 than GF(p), some functions are limited to that for now.
12904 * Add the -HTTP option to s_server. It is similar to -WWW, but requires
12905 that the file contains a complete HTTP response.
12909 * Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
12910 change the def and num file printf format specifier from "%-40sXXX"
12911 to "%-39s XXX". The latter will always guarantee a space after the
12912 field while the former will cause them to run together if the field
12913 is 40 of more characters long.
12917 * Constify the cipher and digest 'method' functions and structures
12918 and modify related functions to take constant EVP_MD and EVP_CIPHER
12923 * Hide BN_CTX structure details in bn_lcl.h instead of publishing them
12924 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
12928 * Modify `EVP_Digest*()` routines so they now return values. Although the
12929 internal software routines can never fail additional hardware versions
12934 * Clean up crypto/err/err.h and change some error codes to avoid conflicts:
12936 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
12937 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
12940 ERR_R_NESTED_ASN1_ERROR
12942 ERR_R_MISSING_ASN1_EOS
12943 were 4 .. 9, conflicting with
12944 ERR_LIB_RSA (= ERR_R_RSA_LIB)
12946 ERR_LIB_PEM (= ERR_R_PEM_LIB).
12947 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
12949 Add new error code 'ERR_R_INTERNAL_ERROR'.
12953 * Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
12958 * New option '-subj arg' for 'openssl req' and 'openssl ca'. This
12959 sets the subject name for a new request or supersedes the
12960 subject name in a given request. Formats that can be parsed are
12961 'CN=Some Name, OU=myOU, C=IT'
12963 'CN=Some Name/OU=myOU/C=IT'.
12965 Add options '-batch' and '-verbose' to 'openssl req'.
12967 *Massimiliano Pala <madwolf@hackmasters.net>*
12969 * Introduce the possibility to access global variables through
12970 functions on platform were that's the best way to handle exporting
12971 global variables in shared libraries. To enable this functionality,
12972 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
12973 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
12974 is normally done by Configure or something similar).
12976 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
12977 in the source file (foo.c) like this:
12979 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
12980 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
12982 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
12983 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
12985 OPENSSL_DECLARE_GLOBAL(int,foo);
12986 #define foo OPENSSL_GLOBAL_REF(foo)
12987 OPENSSL_DECLARE_GLOBAL(double,bar);
12988 #define bar OPENSSL_GLOBAL_REF(bar)
12990 The #defines are very important, and therefore so is including the
12991 header file everywhere where the defined globals are used.
12993 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
12994 of ASN.1 items, but that structure is a bit different.
12996 The largest change is in util/mkdef.pl which has been enhanced with
12997 better and easier to understand logic to choose which symbols should
12998 go into the Windows .def files as well as a number of fixes and code
12999 cleanup (among others, algorithm keywords are now sorted
13000 lexicographically to avoid constant rewrites).
13004 * In BN_div() keep a copy of the sign of 'num' before writing the
13005 result to 'rm' because if rm==num the value will be overwritten
13006 and produce the wrong result if 'num' is negative: this caused
13007 problems with BN_mod() and BN_nnmod().
13011 * Function OCSP_request_verify(). This checks the signature on an
13012 OCSP request and verifies the signer certificate. The signer
13013 certificate is just checked for a generic purpose and OCSP request
13018 * Add OCSP_check_validity() function to check the validity of OCSP
13019 responses. OCSP responses are prepared in real time and may only
13020 be a few seconds old. Simply checking that the current time lies
13021 between thisUpdate and nextUpdate max reject otherwise valid responses
13022 caused by either OCSP responder or client clock inaccuracy. Instead
13023 we allow thisUpdate and nextUpdate to fall within a certain period of
13024 the current time. The age of the response can also optionally be
13025 checked. Two new options -validity_period and -status_age added to
13030 * If signature or public key algorithm is unrecognized print out its
13031 OID rather that just UNKNOWN.
13035 * Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
13036 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
13037 ID to be generated from the issuer certificate alone which can then be
13038 passed to OCSP_id_issuer_cmp().
13042 * New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
13043 ASN1 modules to export functions returning ASN1_ITEM pointers
13044 instead of the ASN1_ITEM structures themselves. This adds several
13045 new macros which allow the underlying ASN1 function/structure to
13046 be accessed transparently. As a result code should not use ASN1_ITEM
13047 references directly (such as &X509_it) but instead use the relevant
13048 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
13049 use of the new ASN1 code on platforms where exporting structures
13050 is problematical (for example in shared libraries) but exporting
13051 functions returning pointers to structures is not.
13055 * Add support for overriding the generation of SSL/TLS session IDs.
13056 These callbacks can be registered either in an SSL_CTX or per SSL.
13057 The purpose of this is to allow applications to control, if they wish,
13058 the arbitrary values chosen for use as session IDs, particularly as it
13059 can be useful for session caching in multiple-server environments. A
13060 command-line switch for testing this (and any client code that wishes
13061 to use such a feature) has been added to "s_server".
13063 *Geoff Thorpe, Lutz Jaenicke*
13065 * Modify mkdef.pl to recognise and parse preprocessor conditionals
13066 of the form `#if defined(...) || defined(...) || ...` and
13067 `#if !defined(...) && !defined(...) && ...`. This also avoids
13068 the growing number of special cases it was previously handling.
13072 * Make all configuration macros available for application by making
13073 sure they are available in opensslconf.h, by giving them names starting
13074 with `OPENSSL_` to avoid conflicts with other packages and by making
13075 sure e_os2.h will cover all platform-specific cases together with
13077 Additionally, it is now possible to define configuration/platform-
13078 specific names (called "system identities"). In the C code, these
13079 are prefixed with `OPENSSL_SYSNAME_`. e_os2.h will create another
13080 macro with the name beginning with `OPENSSL_SYS_`, which is determined
13081 from `OPENSSL_SYSNAME_*` or compiler-specific macros depending on
13086 * New option -set_serial to 'req' and 'x509' this allows the serial
13087 number to use to be specified on the command line. Previously self
13088 signed certificates were hard coded with serial number 0 and the
13089 CA options of 'x509' had to use a serial number in a file which was
13094 * New options to 'ca' utility to support V2 CRL entry extensions.
13095 Currently CRL reason, invalidity date and hold instruction are
13096 supported. Add new CRL extensions to V3 code and some new objects.
13100 * New function EVP_CIPHER_CTX_set_padding() this is used to
13101 disable standard block padding (aka PKCS#5 padding) in the EVP
13102 API, which was previously mandatory. This means that the data is
13103 not padded in any way and so the total length much be a multiple
13104 of the block size, otherwise an error occurs.
13108 * Initial (incomplete) OCSP SSL support.
13112 * New function OCSP_parse_url(). This splits up a URL into its host,
13113 port and path components: primarily to parse OCSP URLs. New -url
13114 option to ocsp utility.
13118 * New nonce behavior. The return value of OCSP_check_nonce() now
13119 reflects the various checks performed. Applications can decide
13120 whether to tolerate certain situations such as an absent nonce
13121 in a response when one was present in a request: the ocsp application
13122 just prints out a warning. New function OCSP_add1_basic_nonce()
13123 this is to allow responders to include a nonce in a response even if
13124 the request is nonce-less.
13128 * Disable stdin buffering in `load_cert()` (`apps/apps.c`) so that no certs are
13129 skipped when using openssl x509 multiple times on a single input file,
13130 e.g. `(openssl x509 -out cert1; openssl x509 -out cert2) <certs`.
13134 * Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
13135 set string type: to handle setting ASN1_TIME structures. Fix ca
13136 utility to correctly initialize revocation date of CRLs.
13140 * New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
13141 the clients preferred ciphersuites and rather use its own preferences.
13142 Should help to work around M$ SGC (Server Gated Cryptography) bug in
13143 Internet Explorer by ensuring unchanged hash method during stepup.
13144 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
13148 * Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
13149 to aes and add a new 'exist' option to print out symbols that don't
13154 * Additional options to ocsp utility to allow flags to be set and
13155 additional certificates supplied.
13159 * Add the option -VAfile to 'openssl ocsp', so the user can give the
13160 OCSP client a number of certificate to only verify the response
13165 * Update Rijndael code to version 3.0 and change EVP AES ciphers to
13166 handle the new API. Currently only ECB, CBC modes supported. Add new
13169 Add TLS AES ciphersuites as described in RFC3268, "Advanced
13170 Encryption Standard (AES) Ciphersuites for Transport Layer
13171 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were
13172 not enabled by default and were not part of the "ALL" ciphersuite
13173 alias because they were not yet official; they could be
13174 explicitly requested by specifying the "AESdraft" ciphersuite
13175 group alias. In the final release of OpenSSL 0.9.7, the group
13176 alias is called "AES" and is part of "ALL".)
13178 *Ben Laurie, Steve Henson, Bodo Moeller*
13180 * New function OCSP_copy_nonce() to copy nonce value (if present) from
13181 request to response.
13185 * Functions for OCSP responders. OCSP_request_onereq_count(),
13186 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
13187 extract information from a certificate request. OCSP_response_create()
13188 creates a response and optionally adds a basic response structure.
13189 OCSP_basic_add1_status() adds a complete single response to a basic
13190 response and returns the OCSP_SINGLERESP structure just added (to allow
13191 extensions to be included for example). OCSP_basic_add1_cert() adds a
13192 certificate to a basic response and OCSP_basic_sign() signs a basic
13193 response with various flags. New helper functions ASN1_TIME_check()
13194 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
13195 (converts ASN1_TIME to GeneralizedTime).
13199 * Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
13200 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
13201 structure from a certificate. X509_pubkey_digest() digests the public_key
13202 contents: this is used in various key identifiers.
13206 * Make sk_sort() tolerate a NULL argument.
13208 *Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>*
13210 * New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
13211 passed by the function are trusted implicitly. If any of them signed the
13212 response then it is assumed to be valid and is not verified.
13216 * In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
13217 to data. This was previously part of the PKCS7 ASN1 code. This
13218 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
13219 *Steve Henson, reported by Kenneth R. Robinette
13220 <support@securenetterm.com>*
13222 * Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
13223 routines: without these tracing memory leaks is very painful.
13224 Fix leaks in PKCS12 and PKCS7 routines.
13228 * Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
13229 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
13230 effectively meant GeneralizedTime would never be used. Now it
13231 is initialised to -1 but X509_time_adj() now has to check the value
13232 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
13233 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
13234 *Steve Henson, reported by Kenneth R. Robinette
13235 <support@securenetterm.com>*
13237 * Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
13238 result in a zero length in the ASN1_INTEGER structure which was
13239 not consistent with the structure when d2i_ASN1_INTEGER() was used
13240 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
13241 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
13242 where it did not print out a minus for negative ASN1_INTEGER.
13246 * Add summary printout to ocsp utility. The various functions which
13247 convert status values to strings have been renamed to:
13248 OCSP_response_status_str(), OCSP_cert_status_str() and
13249 OCSP_crl_reason_str() and are no longer static. New options
13250 to verify nonce values and to disable verification. OCSP response
13251 printout format cleaned up.
13255 * Add additional OCSP certificate checks. These are those specified
13256 in RFC2560. This consists of two separate checks: the CA of the
13257 certificate being checked must either be the OCSP signer certificate
13258 or the issuer of the OCSP signer certificate. In the latter case the
13259 OCSP signer certificate must contain the OCSP signing extended key
13260 usage. This check is performed by attempting to match the OCSP
13261 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
13262 in the OCSP_CERTID structures of the response.
13266 * Initial OCSP certificate verification added to OCSP_basic_verify()
13267 and related routines. This uses the standard OpenSSL certificate
13268 verify routines to perform initial checks (just CA validity) and
13269 to obtain the certificate chain. Then additional checks will be
13270 performed on the chain. Currently the root CA is checked to see
13271 if it is explicitly trusted for OCSP signing. This is used to set
13272 a root CA as a global signing root: that is any certificate that
13273 chains to that CA is an acceptable OCSP signing certificate.
13277 * New '-extfile ...' option to 'openssl ca' for reading X.509v3
13278 extensions from a separate configuration file.
13279 As when reading extensions from the main configuration file,
13280 the '-extensions ...' option may be used for specifying the
13283 *Massimiliano Pala <madwolf@comune.modena.it>*
13285 * New OCSP utility. Allows OCSP requests to be generated or
13286 read. The request can be sent to a responder and the output
13287 parsed, outputted or printed in text form. Not complete yet:
13288 still needs to check the OCSP response validity.
13292 * New subcommands for 'openssl ca':
13293 `openssl ca -status <serial>` prints the status of the cert with
13294 the given serial number (according to the index file).
13295 `openssl ca -updatedb` updates the expiry status of certificates
13298 *Massimiliano Pala <madwolf@comune.modena.it>*
13300 * New '-newreq-nodes' command option to CA.pl. This is like
13301 '-newreq', but calls 'openssl req' with the '-nodes' option
13302 so that the resulting key is not encrypted.
13304 *Damien Miller <djm@mindrot.org>*
13306 * New configuration for the GNU Hurd.
13308 *Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte*
13310 * Initial code to implement OCSP basic response verify. This
13311 is currently incomplete. Currently just finds the signer's
13312 certificate and verifies the signature on the response.
13316 * New SSLeay_version code SSLEAY_DIR to determine the compiled-in
13317 value of OPENSSLDIR. This is available via the new '-d' option
13318 to 'openssl version', and is also included in 'openssl version -a'.
13322 * Allowing defining memory allocation callbacks that will be given
13323 file name and line number information in additional arguments
13324 (a `const char*` and an int). The basic functionality remains, as
13325 well as the original possibility to just replace malloc(),
13326 realloc() and free() by functions that do not know about these
13327 additional arguments. To register and find out the current
13328 settings for extended allocation functions, the following
13329 functions are provided:
13331 CRYPTO_set_mem_ex_functions
13332 CRYPTO_set_locked_mem_ex_functions
13333 CRYPTO_get_mem_ex_functions
13334 CRYPTO_get_locked_mem_ex_functions
13336 These work the same way as CRYPTO_set_mem_functions and friends.
13337 `CRYPTO_get_[locked_]mem_functions` now writes 0 where such an
13338 extended allocation function is enabled.
13339 Similarly, `CRYPTO_get_[locked_]mem_ex_functions` writes 0 where
13340 a conventional allocation function is enabled.
13342 *Richard Levitte, Bodo Moeller*
13344 * Finish off removing the remaining LHASH function pointer casts.
13345 There should no longer be any prototype-casting required when using
13346 the LHASH abstraction, and any casts that remain are "bugs". See
13347 the callback types and macros at the head of lhash.h for details
13348 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
13352 * Add automatic query of EGD sockets in RAND_poll() for the unix variant.
13353 If /dev/[u]random devices are not available or do not return enough
13354 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
13356 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
13357 /etc/entropy will be queried once each in this sequence, querying stops
13358 when enough entropy was collected without querying more sockets.
13362 * Change the Unix RAND_poll() variant to be able to poll several
13363 random devices, as specified by DEVRANDOM, until a sufficient amount
13364 of data has been collected. We spend at most 10 ms on each file
13365 (select timeout) and read in non-blocking mode. DEVRANDOM now
13366 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
13367 (previously it was just the string "/dev/urandom"), so on typical
13368 platforms the 10 ms delay will never occur.
13369 Also separate out the Unix variant to its own file, rand_unix.c.
13370 For VMS, there's a currently-empty rand_vms.c.
13374 * Move OCSP client related routines to ocsp_cl.c. These
13375 provide utility functions which an application needing
13376 to issue a request to an OCSP responder and analyse the
13377 response will typically need: as opposed to those which an
13378 OCSP responder itself would need which will be added later.
13380 OCSP_request_sign() signs an OCSP request with an API similar
13381 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
13382 response. OCSP_response_get1_basic() extracts basic response
13383 from response. OCSP_resp_find_status(): finds and extracts status
13384 information from an OCSP_CERTID structure (which will be created
13385 when the request structure is built). These are built from lower
13386 level functions which work on OCSP_SINGLERESP structures but
13387 won't normally be used unless the application wishes to examine
13388 extensions in the OCSP response for example.
13390 Replace nonce routines with a pair of functions.
13391 OCSP_request_add1_nonce() adds a nonce value and optionally
13392 generates a random value. OCSP_check_nonce() checks the
13393 validity of the nonce in an OCSP response.
13397 * Change function OCSP_request_add() to OCSP_request_add0_id().
13398 This doesn't copy the supplied OCSP_CERTID and avoids the
13399 need to free up the newly created id. Change return type
13400 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
13401 This can then be used to add extensions to the request.
13402 Deleted OCSP_request_new(), since most of its functionality
13403 is now in OCSP_REQUEST_new() (and the case insensitive name
13404 clash) apart from the ability to set the request name which
13405 will be added elsewhere.
13409 * Update OCSP API. Remove obsolete extensions argument from
13410 various functions. Extensions are now handled using the new
13411 OCSP extension code. New simple OCSP HTTP function which
13412 can be used to send requests and parse the response.
13416 * Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
13417 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
13418 uses the special reorder version of SET OF to sort the attributes
13419 and reorder them to match the encoded order. This resolves a long
13420 standing problem: a verify on a PKCS7 structure just after signing
13421 it used to fail because the attribute order did not match the
13422 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
13423 it uses the received order. This is necessary to tolerate some broken
13424 software that does not order SET OF. This is handled by encoding
13425 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
13426 to produce the required SET OF.
13430 * Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
13431 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
13432 files to get correct declarations of the ASN.1 item variables.
13436 * Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
13437 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
13438 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
13439 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
13440 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
13441 ASN1_ITEM and no wrapper functions.
13445 * New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
13446 replace the old function pointer based I/O routines. Change most of
13447 the `*_d2i_bio()` and `*_d2i_fp()` functions to use these.
13451 * Enhance mkdef.pl to be more accepting about spacing in C preprocessor
13452 lines, recognize more "algorithms" that can be deselected, and make
13453 it complain about algorithm deselection that isn't recognised.
13457 * New ASN1 functions to handle dup, sign, verify, digest, pack and
13458 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
13459 to use new functions. Add NO_ASN1_OLD which can be set to remove
13460 some old style ASN1 functions: this can be used to determine if old
13461 code will still work when these eventually go away.
13465 * New extension functions for OCSP structures, these follow the
13466 same conventions as certificates and CRLs.
13470 * New function X509V3_add1_i2d(). This automatically encodes and
13471 adds an extension. Its behaviour can be customised with various
13472 flags to append, replace or delete. Various wrappers added for
13473 certificates and CRLs.
13477 * Fix to avoid calling the underlying ASN1 print routine when
13478 an extension cannot be parsed. Correct a typo in the
13479 OCSP_SERVICELOC extension. Tidy up print OCSP format.
13483 * Make mkdef.pl parse some of the ASN1 macros and add appropriate
13484 entries for variables.
13488 * Add functionality to `apps/openssl.c` for detecting locking
13489 problems: As the program is single-threaded, all we have
13490 to do is register a locking callback using an array for
13491 storing which locks are currently held by the program.
13495 * Use a lock around the call to CRYPTO_get_ex_new_index() in
13496 SSL_get_ex_data_X509_STORE_idx(), which is used in
13497 ssl_verify_cert_chain() and thus can be called at any time
13498 during TLS/SSL handshakes so that thread-safety is essential.
13499 Unfortunately, the ex_data design is not at all suited
13500 for multi-threaded use, so it probably should be abolished.
13504 * Added Broadcom "ubsec" ENGINE to OpenSSL.
13506 *Broadcom, tweaked and integrated by Geoff Thorpe*
13508 * Move common extension printing code to new function
13509 X509V3_print_extensions(). Reorganise OCSP print routines and
13510 implement some needed OCSP ASN1 functions. Add OCSP extensions.
13514 * New function X509_signature_print() to remove duplication in some
13519 * Add a special meaning when SET OF and SEQUENCE OF flags are both
13520 set (this was treated exactly the same as SET OF previously). This
13521 is used to reorder the STACK representing the structure to match the
13522 encoding. This will be used to get round a problem where a PKCS7
13523 structure which was signed could not be verified because the STACK
13524 order did not reflect the encoded order.
13528 * Reimplement the OCSP ASN1 module using the new code.
13532 * Update the X509V3 code to permit the use of an ASN1_ITEM structure
13533 for its ASN1 operations. The old style function pointers still exist
13534 for now but they will eventually go away.
13538 * Merge in replacement ASN1 code from the ASN1 branch. This almost
13539 completely replaces the old ASN1 functionality with a table driven
13540 encoder and decoder which interprets an ASN1_ITEM structure describing
13541 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
13542 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
13543 has also been converted to the new form.
13547 * Change BN_mod_exp_recp so that negative moduli are tolerated
13548 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
13549 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
13550 for negative moduli.
13554 * Fix BN_uadd and BN_usub: Always return non-negative results instead
13555 of not touching the result's sign bit.
13559 * BN_div bugfix: If the result is 0, the sign (res->neg) must not be
13564 * Changed the LHASH code to use prototypes for callbacks, and created
13565 macros to declare and implement thin (optionally static) functions
13566 that provide type-safety and avoid function pointer casting for the
13567 type-specific callbacks.
13571 * Added Kerberos Cipher Suites to be used with TLS, as written in
13573 *Veers Staats <staatsvr@asc.hpc.mil>,
13574 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte*
13576 * Reformat the FAQ so the different questions and answers can be divided
13577 in sections depending on the subject.
13581 * Have the zlib compression code load ZLIB.DLL dynamically under
13586 * New function BN_mod_sqrt for computing square roots modulo a prime
13587 (using the probabilistic Tonelli-Shanks algorithm unless
13588 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
13589 be handled deterministically).
13591 *Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller*
13593 * Make BN_mod_inverse faster by explicitly handling small quotients
13594 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
13595 512 bits], about 30% for larger ones [1024 or 2048 bits].)
13599 * New function BN_kronecker.
13603 * Fix BN_gcd so that it works on negative inputs; the result is
13604 positive unless both parameters are zero.
13605 Previously something reasonably close to an infinite loop was
13606 possible because numbers could be growing instead of shrinking
13607 in the implementation of Euclid's algorithm.
13611 * Fix BN_is_word() and BN_is_one() macros to take into account the
13612 sign of the number in question.
13614 Fix BN_is_word(a,w) to work correctly for w == 0.
13616 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
13617 because its test if the absolute value of 'a' equals 'w'.
13618 Note that BN_abs_is_word does *not* handle w == 0 reliably;
13619 it exists mostly for use in the implementations of BN_is_zero(),
13620 BN_is_one(), and BN_is_word().
13624 * New function BN_swap.
13628 * Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
13629 the exponentiation functions are more likely to produce reasonable
13630 results on negative inputs.
13634 * Change BN_mod_mul so that the result is always non-negative.
13635 Previously, it could be negative if one of the factors was negative;
13636 I don't think anyone really wanted that behaviour.
13640 * Move `BN_mod_...` functions into new file `crypto/bn/bn_mod.c`
13641 (except for exponentiation, which stays in `crypto/bn/bn_exp.c`,
13642 and `BN_mod_mul_reciprocal`, which stays in `crypto/bn/bn_recp.c`)
13643 and add new functions:
13652 BN_mod_lshift1_quick
13654 BN_mod_lshift_quick
13656 These functions always generate non-negative results.
13658 `BN_nnmod` otherwise is `like BN_mod` (if `BN_mod` computes a remainder `r`
13659 such that `|m| < r < 0`, `BN_nnmod` will output `rem + |m|` instead).
13661 `BN_mod_XXX_quick(r, a, [b,] m)` generates the same result as
13662 `BN_mod_XXX(r, a, [b,] m, ctx)`, but requires that `a` [and `b`]
13663 be reduced modulo `m`.
13665 *Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller*
13668 The following entry accidentally appeared in the CHANGES file
13669 distributed with OpenSSL 0.9.7. The modifications described in
13670 it do *not* apply to OpenSSL 0.9.7.
13672 * Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
13673 was actually never needed) and in BN_mul(). The removal in BN_mul()
13674 required a small change in bn_mul_part_recursive() and the addition
13675 of the functions bn_cmp_part_words(), bn_sub_part_words() and
13676 bn_add_part_words(), which do the same thing as bn_cmp_words(),
13677 bn_sub_words() and bn_add_words() except they take arrays with
13683 * In 'openssl passwd', verify passwords read from the terminal
13684 unless the '-salt' option is used (which usually means that
13685 verification would just waste user's time since the resulting
13686 hash is going to be compared with some given password hash)
13687 or the new '-noverify' option is used.
13689 This is an incompatible change, but it does not affect
13690 non-interactive use of 'openssl passwd' (passwords on the command
13691 line, '-stdin' option, '-in ...' option) and thus should not
13692 cause any problems.
13696 * Remove all references to RSAref, since there's no more need for it.
13700 * Make DSO load along a path given through an environment variable
13701 (SHLIB_PATH) with shl_load().
13705 * Constify the ENGINE code as a result of BIGNUM constification.
13706 Also constify the RSA code and most things related to it. In a
13707 few places, most notable in the depth of the ASN.1 code, ugly
13708 casts back to non-const were required (to be solved at a later
13713 * Make it so the openssl application has all engines loaded by default.
13717 * Constify the BIGNUM routines a little more.
13721 * Add the following functions:
13723 ENGINE_load_cswift()
13725 ENGINE_load_atalla()
13726 ENGINE_load_nuron()
13727 ENGINE_load_builtin_engines()
13729 That way, an application can itself choose if external engines that
13730 are built-in in OpenSSL shall ever be used or not. The benefit is
13731 that applications won't have to be linked with libdl or other dso
13732 libraries unless it's really needed.
13734 Changed 'openssl engine' to load all engines on demand.
13735 Changed the engine header files to avoid the duplication of some
13736 declarations (they differed!).
13740 * 'openssl engine' can now list capabilities.
13744 * Better error reporting in 'openssl engine'.
13748 * Never call load_dh_param(NULL) in s_server.
13752 * Add engine application. It can currently list engines by name and
13753 identity, and test if they are actually available.
13757 * Improve RPM specification file by forcing symbolic linking and making
13758 sure the installed documentation is also owned by root.root.
13760 *Damien Miller <djm@mindrot.org>*
13762 * Give the OpenSSL applications more possibilities to make use of
13763 keys (public as well as private) handled by engines.
13767 * Add OCSP code that comes from CertCo.
13771 * Add VMS support for the Rijndael code.
13775 * Added untested support for Nuron crypto accelerator.
13779 * Add support for external cryptographic devices. This code was
13780 previously distributed separately as the "engine" branch.
13782 *Geoff Thorpe, Richard Levitte*
13784 * Rework the filename-translation in the DSO code. It is now possible to
13785 have far greater control over how a "name" is turned into a filename
13786 depending on the operating environment and any oddities about the
13787 different shared library filenames on each system.
13791 * Support threads on FreeBSD-elf in Configure.
13795 * Fix for SHA1 assembly problem with MASM: it produces
13796 warnings about corrupt line number information when assembling
13797 with debugging information. This is caused by the overlapping
13800 *Bernd Matthes <mainbug@celocom.de>, Steve Henson*
13803 NCONF_get_number() has no error checking at all. As a replacement,
13804 NCONF_get_number_e() is defined (`_e` for "error checking") and is
13805 promoted strongly. The old NCONF_get_number is kept around for
13806 binary backward compatibility.
13807 Make it possible for methods to load from something other than a BIO,
13808 by providing a function pointer that is given a name instead of a BIO.
13809 For example, this could be used to load configuration data from an
13814 * Fix for non blocking accept BIOs. Added new I/O special reason
13815 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
13816 with non blocking I/O was not possible because no retry code was
13817 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
13822 * Added the beginnings of Rijndael support.
13826 * Fix for bug in DirectoryString mask setting. Add support for
13827 X509_NAME_print_ex() in 'req' and X509_print_ex() function
13828 to allow certificate printing to more controllable, additional
13829 'certopt' option to 'x509' to allow new printing options to be
13834 * Clean old EAY MD5 hack from e_os.h.
13838 ### Changes between 0.9.6l and 0.9.6m [17 Mar 2004]
13840 * Fix null-pointer assignment in do_change_cipher_spec() revealed
13841 by using the Codenomicon TLS Test Tool ([CVE-2004-0079])
13843 *Joe Orton, Steve Henson*
13845 ### Changes between 0.9.6k and 0.9.6l [04 Nov 2003]
13847 * Fix additional bug revealed by the NISCC test suite:
13849 Stop bug triggering large recursion when presented with
13850 certain ASN.1 tags ([CVE-2003-0851])
13854 ### Changes between 0.9.6j and 0.9.6k [30 Sep 2003]
13856 * Fix various bugs revealed by running the NISCC test suite:
13858 Stop out of bounds reads in the ASN1 code when presented with
13859 invalid tags (CVE-2003-0543 and CVE-2003-0544).
13861 If verify callback ignores invalid public key errors don't try to check
13862 certificate signature with the NULL public key.
13866 * In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
13867 if the server requested one: as stated in TLS 1.0 and SSL 3.0
13872 * In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
13873 extra data after the compression methods not only for TLS 1.0
13874 but also for SSL 3.0 (as required by the specification).
13876 *Bodo Moeller; problem pointed out by Matthias Loepfe*
13878 * Change X509_certificate_type() to mark the key as exported/exportable
13879 when it's 512 *bits* long, not 512 bytes.
13883 ### Changes between 0.9.6i and 0.9.6j [10 Apr 2003]
13885 * Countermeasure against the Klima-Pokorny-Rosa extension of
13886 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
13887 a protocol version number mismatch like a decryption error
13888 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
13892 * Turn on RSA blinding by default in the default implementation
13893 to avoid a timing attack. Applications that don't want it can call
13894 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
13895 They would be ill-advised to do so in most cases.
13897 *Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller*
13899 * Change RSA blinding code so that it works when the PRNG is not
13900 seeded (in this case, the secret RSA exponent is abused as
13901 an unpredictable seed -- if it is not unpredictable, there
13902 is no point in blinding anyway). Make RSA blinding thread-safe
13903 by remembering the creator's thread ID in rsa->blinding and
13904 having all other threads use local one-time blinding factors
13905 (this requires more computation than sharing rsa->blinding, but
13906 avoids excessive locking; and if an RSA object is not shared
13907 between threads, blinding will still be very fast).
13911 ### Changes between 0.9.6h and 0.9.6i [19 Feb 2003]
13913 * In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
13914 via timing by performing a MAC computation even if incorrect
13915 block cipher padding has been found. This is a countermeasure
13916 against active attacks where the attacker has to distinguish
13917 between bad padding and a MAC verification error. ([CVE-2003-0078])
13919 *Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
13920 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
13921 Martin Vuagnoux (EPFL, Ilion)*
13923 ### Changes between 0.9.6g and 0.9.6h [5 Dec 2002]
13925 * New function OPENSSL_cleanse(), which is used to cleanse a section of
13926 memory from its contents. This is done with a counter that will
13927 place alternating values in each byte. This can be used to solve
13928 two issues: 1) the removal of calls to memset() by highly optimizing
13929 compilers, and 2) cleansing with other values than 0, since those can
13930 be read through on certain media, for example a swap space on disk.
13934 * Bugfix: client side session caching did not work with external caching,
13935 because the session->cipher setting was not restored when reloading
13936 from the external cache. This problem was masked, when
13937 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set.
13938 (Found by Steve Haslam <steve@araqnid.ddts.net>.)
13942 * Fix client_certificate (ssl/s2_clnt.c): The permissible total
13943 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33.
13945 *Zeev Lieber <zeev-l@yahoo.com>*
13947 * Undo an undocumented change introduced in 0.9.6e which caused
13948 repeated calls to OpenSSL_add_all_ciphers() and
13949 OpenSSL_add_all_digests() to be ignored, even after calling
13954 * Change the default configuration reader to deal with last line not
13955 being properly terminated.
13959 * Change X509_NAME_cmp() so it applies the special rules on handling
13960 DN values that are of type PrintableString, as well as RDNs of type
13961 emailAddress where the value has the type ia5String.
13963 *stefank@valicert.com via Richard Levitte*
13965 * Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half
13966 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently
13967 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be
13968 the bitwise-OR of the two for use by the majority of applications
13969 wanting this behaviour, and update the docs. The documented
13970 behaviour and actual behaviour were inconsistent and had been
13971 changing anyway, so this is more a bug-fix than a behavioural
13974 *Geoff Thorpe, diagnosed by Nadav Har'El*
13976 * Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c
13977 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes).
13981 * Fix initialization code race conditions in
13982 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(),
13983 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(),
13984 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(),
13985 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(),
13986 ssl2_get_cipher_by_char(),
13987 ssl3_get_cipher_by_char().
13989 *Patrick McCormick <patrick@tellme.com>, Bodo Moeller*
13991 * Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after
13992 the cached sessions are flushed, as the remove_cb() might use ex_data
13993 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com>
13994 (see [openssl.org #212]).
13996 *Geoff Thorpe, Lutz Jaenicke*
13998 * Fix typo in OBJ_txt2obj which incorrectly passed the content
13999 length, instead of the encoding length to d2i_ASN1_OBJECT.
14003 ### Changes between 0.9.6f and 0.9.6g [9 Aug 2002]
14005 * [In 0.9.6g-engine release:]
14006 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use `_stdcall`).
14008 *Lynn Gazis <lgazis@rainbow.com>*
14010 ### Changes between 0.9.6e and 0.9.6f [8 Aug 2002]
14012 * Fix ASN1 checks. Check for overflow by comparing with LONG_MAX
14013 and get fix the header length calculation.
14014 *Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>,
14015 Alon Kantor <alonk@checkpoint.com> (and others), Steve Henson*
14017 * Use proper error handling instead of 'assertions' in buffer
14018 overflow checks added in 0.9.6e. This prevents DoS (the
14019 assertions could call abort()).
14021 *Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller*
14023 ### Changes between 0.9.6d and 0.9.6e [30 Jul 2002]
14025 * Add various sanity checks to asn1_get_length() to reject
14026 the ASN1 length bytes if they exceed sizeof(long), will appear
14027 negative or the content length exceeds the length of the
14030 *Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>*
14032 * Fix cipher selection routines: ciphers without encryption had no flags
14033 for the cipher strength set and where therefore not handled correctly
14034 by the selection routines (PR #130).
14038 * Fix EVP_dsa_sha macro.
14043 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS
14044 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure
14045 that was added in OpenSSL 0.9.6d.
14047 As the countermeasure turned out to be incompatible with some
14048 broken SSL implementations, the new option is part of SSL_OP_ALL.
14049 SSL_OP_ALL is usually employed when compatibility with weird SSL
14050 implementations is desired (e.g. '-bugs' option to 's_client' and
14051 's_server'), so the new option is automatically set in many
14056 * Changes in security patch:
14058 Changes marked "(CHATS)" were sponsored by the Defense Advanced
14059 Research Projects Agency (DARPA) and Air Force Research Laboratory,
14060 Air Force Materiel Command, USAF, under agreement number
14063 * Add various sanity checks to asn1_get_length() to reject
14064 the ASN1 length bytes if they exceed sizeof(long), will appear
14065 negative or the content length exceeds the length of the
14066 supplied buffer. ([CVE-2002-0659])
14068 *Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>*
14070 * Assertions for various potential buffer overflows, not known to
14071 happen in practice.
14073 *Ben Laurie (CHATS)*
14075 * Various temporary buffers to hold ASCII versions of integers were
14076 too small for 64 bit platforms. ([CVE-2002-0655])
14077 *Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>*
14079 * Remote buffer overflow in SSL3 protocol - an attacker could
14080 supply an oversized session ID to a client. ([CVE-2002-0656])
14082 *Ben Laurie (CHATS)*
14084 * Remote buffer overflow in SSL2 protocol - an attacker could
14085 supply an oversized client master key. ([CVE-2002-0656])
14087 *Ben Laurie (CHATS)*
14089 ### Changes between 0.9.6c and 0.9.6d [9 May 2002]
14091 * Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
14092 encoded as NULL) with id-dsa-with-sha1.
14094 *Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller*
14096 * Check various `X509_...()` return values in `apps/req.c`.
14098 *Nils Larsch <nla@trustcenter.de>*
14100 * Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
14101 an end-of-file condition would erroneously be flagged, when the CRLF
14102 was just at the end of a processed block. The bug was discovered when
14103 processing data through a buffering memory BIO handing the data to a
14104 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
14105 <ptsekov@syntrex.com> and Nedelcho Stanev.
14109 * Implement a countermeasure against a vulnerability recently found
14110 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
14111 before application data chunks to avoid the use of known IVs
14112 with data potentially chosen by the attacker.
14116 * Fix length checks in ssl3_get_client_hello().
14120 * TLS/SSL library bugfix: use s->s3->in_read_app_data differently
14121 to prevent ssl3_read_internal() from incorrectly assuming that
14122 ssl3_read_bytes() found application data while handshake
14123 processing was enabled when in fact s->s3->in_read_app_data was
14124 merely automatically cleared during the initial handshake.
14126 *Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>*
14128 * Fix object definitions for Private and Enterprise: they were not
14129 recognized in their shortname (=lowercase) representation. Extend
14130 obj_dat.pl to issue an error when using undefined keywords instead
14131 of silently ignoring the problem (Svenning Sorensen
14132 <sss@sss.dnsalias.net>).
14136 * Fix DH_generate_parameters() so that it works for 'non-standard'
14137 generators, i.e. generators other than 2 and 5. (Previously, the
14138 code did not properly initialise the 'add' and 'rem' values to
14139 BN_generate_prime().)
14141 In the new general case, we do not insist that 'generator' is
14142 actually a primitive root: This requirement is rather pointless;
14143 a generator of the order-q subgroup is just as good, if not
14148 * Map new X509 verification errors to alerts. Discovered and submitted by
14149 Tom Wu <tom@arcot.com>.
14153 * Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
14154 returning non-zero before the data has been completely received
14155 when using non-blocking I/O.
14157 *Bodo Moeller; problem pointed out by John Hughes*
14159 * Some of the ciphers missed the strength entry (SSL_LOW etc).
14161 *Ben Laurie, Lutz Jaenicke*
14163 * Fix bug in SSL_clear(): bad sessions were not removed (found by
14164 Yoram Zahavi <YoramZ@gilian.com>).
14168 * Add information about CygWin 1.3 and on, and preserve proper
14169 configuration for the versions before that.
14171 *Corinna Vinschen <vinschen@redhat.com> and Richard Levitte*
14173 * Make removal from session cache (SSL_CTX_remove_session()) more robust:
14174 check whether we deal with a copy of a session and do not delete from
14175 the cache in this case. Problem reported by "Izhar Shoshani Levi"
14176 <izhar@checkpoint.com>.
14180 * Do not store session data into the internal session cache, if it
14181 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
14182 flag is set). Proposed by Aslam <aslam@funk.com>.
14186 * Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
14191 * [In 0.9.6d-engine release:]
14192 Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
14194 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
14196 * Add the configuration target linux-s390x.
14198 *Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte*
14200 * The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
14201 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
14202 variable as an indication that a ClientHello message has been
14203 received. As the flag value will be lost between multiple
14204 invocations of ssl3_accept when using non-blocking I/O, the
14205 function may not be aware that a handshake has actually taken
14206 place, thus preventing a new session from being added to the
14209 To avoid this problem, we now set s->new_session to 2 instead of
14210 using a local variable.
14212 *Lutz Jaenicke, Bodo Moeller*
14214 * Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
14215 if the SSL_R_LENGTH_MISMATCH error is detected.
14217 *Geoff Thorpe, Bodo Moeller*
14219 * New 'shared_ldflag' column in Configure platform table.
14223 * Fix EVP_CIPHER_mode macro.
14225 *"Dan S. Camper" <dan@bti.net>*
14227 * Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
14228 type, we must throw them away by setting rr->length to 0.
14230 *D P Chang <dpc@qualys.com>*
14232 ### Changes between 0.9.6b and 0.9.6c [21 dec 2001]
14234 * Fix BN_rand_range bug pointed out by Dominikus Scherkl
14235 <Dominikus.Scherkl@biodata.com>. (The previous implementation
14236 worked incorrectly for those cases where range = `10..._2` and
14237 `3*range` is two bits longer than range.)
14241 * Only add signing time to PKCS7 structures if it is not already
14246 * Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
14247 OBJ_ld_ce should be OBJ_id_ce.
14248 Also some ip-pda OIDs in crypto/objects/objects.txt were
14249 incorrect (cf. RFC 3039).
14251 *Matt Cooper, Frederic Giudicelli, Bodo Moeller*
14253 * Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
14254 returns early because it has nothing to do.
14256 *Andy Schneider <andy.schneider@bjss.co.uk>*
14258 * [In 0.9.6c-engine release:]
14259 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
14261 *Andy Schneider <andy.schneider@bjss.co.uk>*
14263 * [In 0.9.6c-engine release:]
14264 Add support for Cryptographic Appliance's keyserver technology.
14265 (Use engine 'keyclient')
14267 *Cryptographic Appliances and Geoff Thorpe*
14269 * Add a configuration entry for OS/390 Unix. The C compiler 'c89'
14270 is called via tools/c89.sh because arguments have to be
14271 rearranged (all '-L' options must appear before the first object
14274 *Richard Shapiro <rshapiro@abinitio.com>*
14276 * [In 0.9.6c-engine release:]
14277 Add support for Broadcom crypto accelerator cards, backported
14280 *Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox*
14282 * [In 0.9.6c-engine release:]
14283 Add support for SureWare crypto accelerator cards from
14284 Baltimore Technologies. (Use engine 'sureware')
14286 *Baltimore Technologies and Mark Cox*
14288 * [In 0.9.6c-engine release:]
14289 Add support for crypto accelerator cards from Accelerated
14290 Encryption Processing, www.aep.ie. (Use engine 'aep')
14292 *AEP Inc. and Mark Cox*
14294 * Add a configuration entry for gcc on UnixWare.
14296 *Gary Benson <gbenson@redhat.com>*
14298 * Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
14299 messages are stored in a single piece (fixed-length part and
14300 variable-length part combined) and fix various bugs found on the way.
14304 * Disable caching in BIO_gethostbyname(), directly use gethostbyname()
14305 instead. BIO_gethostbyname() does not know what timeouts are
14306 appropriate, so entries would stay in cache even when they have
14308 *Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>*
14310 * Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
14311 faced with a pathologically small ClientHello fragment that does
14312 not contain client_version: Instead of aborting with an error,
14313 simply choose the highest available protocol version (i.e.,
14314 TLS 1.0 unless it is disabled). In practice, ClientHello
14315 messages are never sent like this, but this change gives us
14316 strictly correct behaviour at least for TLS.
14320 * Fix SSL handshake functions and SSL_clear() such that SSL_clear()
14321 never resets s->method to s->ctx->method when called from within
14322 one of the SSL handshake functions.
14324 *Bodo Moeller; problem pointed out by Niko Baric*
14326 * In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
14327 (sent using the client's version number) if client_version is
14328 smaller than the protocol version in use. Also change
14329 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
14330 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
14331 the client will at least see that alert.
14335 * Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
14340 * Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
14341 client receives HelloRequest while in a handshake.
14343 *Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>*
14345 * Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
14346 should end in 'break', not 'goto end' which circumvents various
14347 cleanups done in state SSL_ST_OK. But session related stuff
14348 must be disabled for SSL_ST_OK in the case that we just sent a
14351 Also avoid some overhead by not calling ssl_init_wbio_buffer()
14352 before just sending a HelloRequest.
14354 *Bodo Moeller, Eric Rescorla <ekr@rtfm.com>*
14356 * Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
14357 reveal whether illegal block cipher padding was found or a MAC
14358 verification error occurred. (Neither SSLerr() codes nor alerts
14359 are directly visible to potential attackers, but the information
14360 may leak via logfiles.)
14362 Similar changes are not required for the SSL 2.0 implementation
14363 because the number of padding bytes is sent in clear for SSL 2.0,
14364 and the extra bytes are just ignored. However ssl/s2_pkt.c
14365 failed to verify that the purported number of padding bytes is in
14370 * Add OpenUNIX-8 support including shared libraries
14371 (Boyd Lynn Gerber <gerberb@zenez.com>).
14375 * Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
14376 'wristwatch attack' using huge encoding parameters (cf.
14377 James H. Manger's CRYPTO 2001 paper). Note that the
14378 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
14379 encoding parameters and hence was not vulnerable.
14383 * BN_sqr() bug fix.
14385 *Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>*
14387 * Rabin-Miller test analyses assume uniformly distributed witnesses,
14388 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
14389 followed by modular reduction.
14391 *Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>*
14393 * Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
14394 equivalent based on BN_pseudo_rand() instead of BN_rand().
14398 * s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
14399 This function was broken, as the check for a new client hello message
14400 to handle SGC did not allow these large messages.
14401 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
14405 * Add alert descriptions for TLSv1 to `SSL_alert_desc_string[_long]()`.
14409 * Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
14410 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
14414 * Rework the configuration and shared library support for Tru64 Unix.
14415 The configuration part makes use of modern compiler features and
14416 still retains old compiler behavior for those that run older versions
14417 of the OS. The shared library support part includes a variant that
14418 uses the RPATH feature, and is available through the special
14419 configuration target "alpha-cc-rpath", which will never be selected
14422 *Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte*
14424 * In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
14425 with the same message size as in ssl3_get_certificate_request().
14426 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
14427 messages might inadvertently be reject as too long.
14429 *Petr Lampa <lampa@fee.vutbr.cz>*
14431 * Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
14435 * Modified SSL library such that the verify_callback that has been set
14436 specifically for an SSL object with SSL_set_verify() is actually being
14437 used. Before the change, a verify_callback set with this function was
14438 ignored and the verify_callback() set in the SSL_CTX at the time of
14439 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
14440 to allow the necessary settings.
14444 * Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
14445 explicitly to NULL, as at least on Solaris 8 this seems not always to be
14446 done automatically (in contradiction to the requirements of the C
14447 standard). This made problems when used from OpenSSH.
14451 * In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
14452 dh->length and always used
14454 BN_rand_range(priv_key, dh->p).
14456 BN_rand_range() is not necessary for Diffie-Hellman, and this
14457 specific range makes Diffie-Hellman unnecessarily inefficient if
14458 dh->length (recommended exponent length) is much smaller than the
14459 length of dh->p. We could use BN_rand_range() if the order of
14460 the subgroup was stored in the DH structure, but we only have
14465 BN_rand(priv_key, l, ...)
14467 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
14474 RSA_eay_public_encrypt
14475 RSA_eay_private_decrypt
14476 RSA_eay_private_encrypt (signing)
14477 RSA_eay_public_decrypt (signature verification)
14479 (default implementations for RSA_public_encrypt,
14480 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
14481 always reject numbers >= n.
14485 * In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
14486 to synchronize access to 'locking_thread'. This is necessary on
14487 systems where access to 'locking_thread' (an 'unsigned long'
14488 variable) is not atomic.
14492 * In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
14493 *before* setting the 'crypto_lock_rand' flag. The previous code had
14494 a race condition if 0 is a valid thread ID.
14496 *Travis Vitek <vitek@roguewave.com>*
14498 * Add support for shared libraries under Irix.
14500 *Albert Chin-A-Young <china@thewrittenword.com>*
14502 * Add configuration option to build on Linux on both big-endian and
14503 little-endian MIPS.
14505 *Ralf Baechle <ralf@uni-koblenz.de>*
14507 * Add the possibility to create shared libraries on HP-UX.
14511 ### Changes between 0.9.6a and 0.9.6b [9 Jul 2001]
14513 * Change ssleay_rand_bytes (crypto/rand/md_rand.c)
14514 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
14515 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
14516 PRNG state recovery was possible based on the output of
14517 one PRNG request appropriately sized to gain knowledge on
14518 'md' followed by enough consecutive 1-byte PRNG requests
14519 to traverse all of 'state'.
14521 1. When updating 'md_local' (the current thread's copy of 'md')
14522 during PRNG output generation, hash all of the previous
14523 'md_local' value, not just the half used for PRNG output.
14525 2. Make the number of bytes from 'state' included into the hash
14526 independent from the number of PRNG bytes requested.
14528 The first measure alone would be sufficient to avoid
14529 Markku-Juhani's attack. (Actually it had never occurred
14530 to me that the half of 'md_local' used for chaining was the
14531 half from which PRNG output bytes were taken -- I had always
14532 assumed that the secret half would be used.) The second
14533 measure makes sure that additional data from 'state' is never
14534 mixed into 'md_local' in small portions; this heuristically
14535 further strengthens the PRNG.
14539 * Fix crypto/bn/asm/mips3.s.
14543 * When only the key is given to "enc", the IV is undefined. Print out
14544 an error message in this case.
14548 * Handle special case when X509_NAME is empty in X509 printing routines.
14552 * In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
14553 positive and less than q.
14557 * Don't change `*pointer` in CRYPTO_add_lock() is add_lock_callback is
14558 used: it isn't thread safe and the add_lock_callback should handle
14561 *Paul Rose <Paul.Rose@bridge.com>*
14563 * Verify that incoming data obeys the block size in
14564 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
14570 *Ulf Möller, Bodo Möller*
14572 * The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
14573 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
14574 when fixing the server behaviour for backwards-compatible 'client
14575 hello' messages. (Note that the attack is impractical against
14576 SSL 3.0 and TLS 1.0 anyway because length and version checking
14577 means that the probability of guessing a valid ciphertext is
14578 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
14581 Before 0.9.5, the countermeasure (hide the error by generating a
14582 random 'decryption result') did not work properly because
14583 ERR_clear_error() was missing, meaning that SSL_get_error() would
14584 detect the supposedly ignored error.
14586 Both problems are now fixed.
14590 * In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
14591 (previously it was 1024).
14595 * Fix for compatibility mode trust settings: ignore trust settings
14596 unless some valid trust or reject settings are present.
14600 * Fix for blowfish EVP: its a variable length cipher.
14604 * Fix various bugs related to DSA S/MIME verification. Handle missing
14605 parameters in DSA public key structures and return an error in the
14606 DSA routines if parameters are absent.
14610 * In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
14611 in the current directory if neither $RANDFILE nor $HOME was set.
14612 RAND_file_name() in 0.9.6a returned NULL in this case. This has
14613 caused some confusion to Windows users who haven't defined $HOME.
14614 Thus RAND_file_name() is changed again: e_os.h can define a
14615 DEFAULT_HOME, which will be used if $HOME is not set.
14616 For Windows, we use "C:"; on other platforms, we still require
14617 environment variables.
14619 * Move 'if (!initialized) RAND_poll()' into regions protected by
14620 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
14621 having multiple threads call RAND_poll() concurrently.
14625 * In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
14626 combination of a flag and a thread ID variable.
14627 Otherwise while one thread is in ssleay_rand_bytes (which sets the
14628 flag), *other* threads can enter ssleay_add_bytes without obeying
14629 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
14630 that they do not hold after the first thread unsets add_do_not_lock).
14634 * Change bctest again: '-x' expressions are not available in all
14635 versions of 'test'.
14639 ### Changes between 0.9.6 and 0.9.6a [5 Apr 2001]
14641 * Fix a couple of memory leaks in PKCS7_dataDecode()
14643 *Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>*
14645 * Change Configure and Makefiles to provide EXE_EXT, which will contain
14646 the default extension for executables, if any. Also, make the perl
14647 scripts that use symlink() to test if it really exists and use "cp"
14648 if it doesn't. All this made OpenSSL compilable and installable in
14653 * Fix for asn1_GetSequence() for indefinite length constructed data.
14654 If SEQUENCE is length is indefinite just set c->slen to the total
14655 amount of data available.
14657 *Steve Henson, reported by shige@FreeBSD.org*
14659 *This change does not apply to 0.9.7.*
14661 * Change bctest to avoid here-documents inside command substitution
14662 (workaround for FreeBSD /bin/sh bug).
14663 For compatibility with Ultrix, avoid shell functions (introduced
14664 in the bctest version that searches along $PATH).
14668 * Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
14669 with des_encrypt() defined on some operating systems, like Solaris
14674 * Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
14675 On the Importance of Eliminating Errors in Cryptographic
14676 Computations, J. Cryptology 14 (2001) 2, 101-119,
14677 <http://theory.stanford.edu/~dabo/papers/faults.ps.gz>).
14681 * MIPS assembler BIGNUM division bug fix.
14685 * Disabled incorrect Alpha assembler code.
14689 * Fix PKCS#7 decode routines so they correctly update the length
14690 after reading an EOC for the EXPLICIT tag.
14694 *This change does not apply to 0.9.7.*
14696 * Fix bug in PKCS#12 key generation routines. This was triggered
14697 if a 3DES key was generated with a 0 initial byte. Include
14698 PKCS12_BROKEN_KEYGEN compilation option to retain the old
14699 (but broken) behaviour.
14703 * Enhance bctest to search for a working bc along $PATH and print
14706 *Tim Rice <tim@multitalents.net> via Richard Levitte*
14708 * Fix memory leaks in err.c: free err_data string if necessary;
14709 don't write to the wrong index in ERR_set_error_data.
14713 * Implement ssl23_peek (analogous to ssl23_read), which previously
14718 * Replace rdtsc with `_emit` statements for VC++ version 5.
14720 *Jeremy Cooper <jeremy@baymoo.org>*
14722 * Make it possible to reuse SSLv2 sessions.
14726 * In copy_email() check for >= 0 as a return value for
14727 X509_NAME_get_index_by_NID() since 0 is a valid index.
14729 *Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>*
14731 * Avoid coredump with unsupported or invalid public keys by checking if
14732 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
14733 PKCS7_verify() fails with non detached data.
14737 * Don't use getenv in library functions when run as setuid/setgid.
14738 New function OPENSSL_issetugid().
14742 * Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
14743 due to incorrect handling of multi-threading:
14745 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
14747 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
14749 3. Count how many times MemCheck_off() has been called so that
14750 nested use can be treated correctly. This also avoids
14751 inband-signalling in the previous code (which relied on the
14752 assumption that thread ID 0 is impossible).
14756 * Add "-rand" option also to s_client and s_server.
14760 * Fix CPU detection on Irix 6.x.
14761 *Kurt Hockenbury <khockenb@stevens-tech.edu> and
14762 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>*
14764 * Fix X509_NAME bug which produced incorrect encoding if X509_NAME
14769 *This change does not apply to 0.9.7.*
14771 * Use the cached encoding of an X509_NAME structure rather than
14772 copying it. This is apparently the reason for the libsafe "errors"
14773 but the code is actually correct.
14777 * Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
14778 Bleichenbacher's DSA attack.
14779 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
14780 to be set and top=0 forces the highest bit to be set; top=-1 is new
14781 and leaves the highest bit random.
14783 *Ulf Moeller, Bodo Moeller*
14785 * In the `NCONF_...`-based implementations for `CONF_...` queries
14786 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
14787 a temporary CONF structure with the data component set to NULL
14788 (which gives segmentation faults in lh_retrieve).
14789 Instead, use NULL for the CONF pointer in CONF_get_string and
14790 CONF_get_number (which may use environment variables) and directly
14791 return NULL from CONF_get_section.
14795 * Fix potential buffer overrun for EBCDIC.
14799 * Tolerate nonRepudiation as being valid for S/MIME signing and certSign
14800 keyUsage if basicConstraints absent for a CA.
14804 * Make SMIME_write_PKCS7() write mail header values with a format that
14805 is more generally accepted (no spaces before the semicolon), since
14806 some programs can't parse those values properly otherwise. Also make
14807 sure BIO's that break lines after each write do not create invalid
14812 * Make the CRL encoding routines work with empty SEQUENCE OF. The
14813 macros previously used would not encode an empty SEQUENCE OF
14814 and break the signature.
14818 *This change does not apply to 0.9.7.*
14820 * Zero the premaster secret after deriving the master secret in
14825 * Add some EVP_add_digest_alias registrations (as found in
14826 OpenSSL_add_all_digests()) to SSL_library_init()
14827 aka OpenSSL_add_ssl_algorithms(). This provides improved
14828 compatibility with peers using X.509 certificates
14829 with unconventional AlgorithmIdentifier OIDs.
14833 * Fix for Irix with NO_ASM.
14835 *"Bruce W. Forsberg" <bruce.forsberg@baesystems.com>*
14837 * ./config script fixes.
14839 *Ulf Moeller, Richard Levitte*
14841 * Fix 'openssl passwd -1'.
14845 * Change PKCS12_key_gen_asc() so it can cope with non null
14846 terminated strings whose length is passed in the passlen
14847 parameter, for example from PEM callbacks. This was done
14848 by adding an extra length parameter to asc2uni().
14850 *Steve Henson, reported by <oddissey@samsung.co.kr>*
14852 * Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
14853 call failed, free the DSA structure.
14857 * Fix to uni2asc() to cope with zero length Unicode strings.
14858 These are present in some PKCS#12 files.
14862 * Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
14863 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
14864 when writing a 32767 byte record.
14866 *Bodo Moeller; problem reported by Eric Day <eday@concentric.net>*
14868 * In `RSA_eay_public_{en,ed}crypt` and RSA_eay_mod_exp (rsa_eay.c),
14869 obtain lock CRYPTO_LOCK_RSA before setting `rsa->_method_mod_{n,p,q}`.
14871 (RSA objects have a reference count access to which is protected
14872 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
14873 so they are meant to be shared between threads.)
14874 *Bodo Moeller, Geoff Thorpe; original patch submitted by
14875 "Reddie, Steven" <Steven.Reddie@ca.com>*
14877 * Fix a deadlock in CRYPTO_mem_leaks().
14881 * Use better test patterns in bntest.
14885 * rand_win.c fix for Borland C.
14889 * BN_rshift bugfix for n == 0.
14893 * Add a 'bctest' script that checks for some known 'bc' bugs
14894 so that 'make test' does not abort just because 'bc' is broken.
14898 * Store verify_result within SSL_SESSION also for client side to
14899 avoid potential security hole. (Re-used sessions on the client side
14900 always resulted in verify_result==X509_V_OK, not using the original
14901 result of the server certificate verification.)
14905 * Fix ssl3_pending: If the record in s->s3->rrec is not of type
14906 SSL3_RT_APPLICATION_DATA, return 0.
14907 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
14912 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
14913 releases, have been re-implemented by renaming the previous
14914 implementations of ssl2_read and ssl3_read to ssl2_read_internal
14915 and ssl3_read_internal, respectively, and adding 'peek' parameters
14916 to them. The new ssl[23]_{read,peek} functions are calls to
14917 ssl[23]_read_internal with the 'peek' flag set appropriately.
14918 A 'peek' parameter has also been added to ssl3_read_bytes, which
14919 does the actual work for ssl3_read_internal.
14923 * Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
14924 the method-specific "init()" handler. Also clean up ex_data after
14925 calling the method-specific "finish()" handler. Previously, this was
14926 happening the other way round.
14930 * Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
14931 The previous value, 12, was not always sufficient for BN_mod_exp().
14935 * Make sure that shared libraries get the internal name engine with
14936 the full version number and not just 0. This should mark the
14937 shared libraries as not backward compatible. Of course, this should
14938 be changed again when we can guarantee backward binary compatibility.
14942 * Fix typo in get_cert_by_subject() in by_dir.c
14944 *Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>*
14946 * Rework the system to generate shared libraries:
14948 - Make note of the expected extension for the shared libraries and
14949 if there is a need for symbolic links from for example libcrypto.so.0
14950 to libcrypto.so.0.9.7. There is extended info in Configure for
14953 - Make as few rebuilds of the shared libraries as possible.
14955 - Still avoid linking the OpenSSL programs with the shared libraries.
14957 - When installing, install the shared libraries separately from the
14962 * Fix SSL_CTX_set_read_ahead macro to actually use its argument.
14964 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
14965 and not in SSL_clear because the latter is also used by the
14966 accept/connect functions; previously, the settings made by
14967 SSL_set_read_ahead would be lost during the handshake.
14969 *Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>*
14971 * Correct util/mkdef.pl to be selective about disabled algorithms.
14972 Previously, it would create entries for disabled algorithms no
14977 * Added several new manual pages for SSL_* function.
14981 ### Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
14983 * In ssl23_get_client_hello, generate an error message when faced
14984 with an initial SSL 3.0/TLS record that is too small to contain the
14985 first two bytes of the ClientHello message, i.e. client_version.
14986 (Note that this is a pathologic case that probably has never happened
14987 in real life.) The previous approach was to use the version number
14988 from the record header as a substitute; but our protocol choice
14989 should not depend on that one because it is not authenticated
14990 by the Finished messages.
14994 * More robust randomness gathering functions for Windows.
14996 *Jeffrey Altman <jaltman@columbia.edu>*
14998 * For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
14999 not set then we don't setup the error code for issuer check errors
15000 to avoid possibly overwriting other errors which the callback does
15001 handle. If an application does set the flag then we assume it knows
15002 what it is doing and can handle the new informational codes
15007 * Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
15008 a general "ANY" type, as such it should be able to decode anything
15009 including tagged types. However it didn't check the class so it would
15010 wrongly interpret tagged types in the same way as their universal
15011 counterpart and unknown types were just rejected. Changed so that the
15012 tagged and unknown types are handled in the same way as a SEQUENCE:
15013 that is the encoding is stored intact. There is also a new type
15014 "V_ASN1_OTHER" which is used when the class is not universal, in this
15015 case we have no idea what the actual type is so we just lump them all
15020 * On VMS, stdout may very well lead to a file that is written to
15021 in a record-oriented fashion. That means that every write() will
15022 write a separate record, which will be read separately by the
15023 programs trying to read from it. This can be very confusing.
15025 The solution is to put a BIO filter in the way that will buffer
15026 text until a linefeed is reached, and then write everything a
15027 line at a time, so every record written will be an actual line,
15028 not chunks of lines and not (usually doesn't happen, but I've
15029 seen it once) several lines in one record. BIO_f_linebuffer() is
15032 Currently, it's a VMS-only method, because that's where it has
15033 been tested well enough.
15037 * Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
15038 it can return incorrect results.
15039 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
15040 but it was in 0.9.6-beta[12].)
15044 * Disable the check for content being present when verifying detached
15045 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
15046 include zero length content when signing messages.
15050 * New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
15051 BIO_ctrl (for BIO pairs).
15055 * Add DSO method for VMS.
15059 * Bug fix: Montgomery multiplication could produce results with the
15064 * Add RPM specification openssl.spec and modify it to build three
15065 packages. The default package contains applications, application
15066 documentation and run-time libraries. The devel package contains
15067 include files, static libraries and function documentation. The
15068 doc package contains the contents of the doc directory. The original
15069 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
15073 * Add a large number of documentation files for many SSL routines.
15075 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>*
15077 * Add a configuration entry for Sony News 4.
15079 *NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>*
15081 * Don't set the two most significant bits to one when generating a
15082 random number < q in the DSA library.
15086 * New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
15087 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
15088 the underlying transport is blocking) if a handshake took place.
15089 (The default behaviour is needed by applications such as s_client
15090 and s_server that use select() to determine when to use SSL_read;
15091 but for applications that know in advance when to expect data, it
15092 just makes things more complicated.)
15096 * Add RAND_egd_bytes(), which gives control over the number of bytes read
15101 * Add a few more EBCDIC conditionals that make `req` and `x509`
15102 work better on such systems.
15104 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
15106 * Add two demo programs for PKCS12_parse() and PKCS12_create().
15107 Update PKCS12_parse() so it copies the friendlyName and the
15108 keyid to the certificates aux info.
15112 * Fix bug in PKCS7_verify() which caused an infinite loop
15113 if there was more than one signature.
15115 *Sven Uszpelkat <su@celocom.de>*
15117 * Major change in util/mkdef.pl to include extra information
15118 about each symbol, as well as presenting variables as well
15119 as functions. This change means that there's n more need
15120 to rebuild the .num files when some algorithms are excluded.
15124 * Allow the verify time to be set by an application,
15125 rather than always using the current time.
15129 * Phase 2 verify code reorganisation. The certificate
15130 verify code now looks up an issuer certificate by a
15131 number of criteria: subject name, authority key id
15132 and key usage. It also verifies self signed certificates
15133 by the same criteria. The main comparison function is
15134 X509_check_issued() which performs these checks.
15136 Lot of changes were necessary in order to support this
15137 without completely rewriting the lookup code.
15139 Authority and subject key identifier are now cached.
15141 The LHASH 'certs' is X509_STORE has now been replaced
15142 by a STACK_OF(X509_OBJECT). This is mainly because an
15143 LHASH can't store or retrieve multiple objects with
15144 the same hash value.
15146 As a result various functions (which were all internal
15147 use only) have changed to handle the new X509_STORE
15148 structure. This will break anything that messed round
15149 with X509_STORE internally.
15151 The functions X509_STORE_add_cert() now checks for an
15152 exact match, rather than just subject name.
15154 The X509_STORE API doesn't directly support the retrieval
15155 of multiple certificates matching a given criteria, however
15156 this can be worked round by performing a lookup first
15157 (which will fill the cache with candidate certificates)
15158 and then examining the cache for matches. This is probably
15159 the best we can do without throwing out X509_LOOKUP
15160 entirely (maybe later...).
15162 The X509_VERIFY_CTX structure has been enhanced considerably.
15164 All certificate lookup operations now go via a get_issuer()
15165 callback. Although this currently uses an X509_STORE it
15166 can be replaced by custom lookups. This is a simple way
15167 to bypass the X509_STORE hackery necessary to make this
15168 work and makes it possible to use more efficient techniques
15169 in future. A very simple version which uses a simple
15170 STACK for its trusted certificate store is also provided
15171 using X509_STORE_CTX_trusted_stack().
15173 The verify_cb() and verify() callbacks now have equivalents
15174 in the X509_STORE_CTX structure.
15176 X509_STORE_CTX also has a 'flags' field which can be used
15177 to customise the verify behaviour.
15181 * Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
15182 excludes S/MIME capabilities.
15186 * When a certificate request is read in keep a copy of the
15187 original encoding of the signed data and use it when outputting
15188 again. Signatures then use the original encoding rather than
15189 a decoded, encoded version which may cause problems if the
15190 request is improperly encoded.
15194 * For consistency with other BIO_puts implementations, call
15195 buffer_write(b, ...) directly in buffer_puts instead of calling
15198 In BIO_puts, increment b->num_write as in BIO_write.
15200 *Peter.Sylvester@EdelWeb.fr*
15202 * Fix BN_mul_word for the case where the word is 0. (We have to use
15203 BN_zero, we may not return a BIGNUM with an array consisting of
15204 words set to zero.)
15208 * Avoid calling abort() from within the library when problems are
15209 detected, except if preprocessor symbols have been defined
15210 (such as REF_CHECK, BN_DEBUG etc.).
15214 * New openssl application 'rsautl'. This utility can be
15215 used for low-level RSA operations. DER public key
15216 BIO/fp routines also added.
15220 * New Configure entry and patches for compiling on QNX 4.
15222 *Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>*
15224 * A demo state-machine implementation was sponsored by
15225 Nuron (<http://www.nuron.com/>) and is now available in
15226 demos/state_machine.
15230 * New options added to the 'dgst' utility for signature
15231 generation and verification.
15235 * Unrecognized PKCS#7 content types are now handled via a
15236 catch all ASN1_TYPE structure. This allows unsupported
15237 types to be stored as a "blob" and an application can
15238 encode and decode it manually.
15242 * Fix various signed/unsigned issues to make a_strex.c
15243 compile under VC++.
15245 *Oscar Jacobsson <oscar.jacobsson@celocom.com>*
15247 * ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
15248 length if passed a buffer. ASN1_INTEGER_to_BN failed
15249 if passed a NULL BN and its argument was negative.
15251 *Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>*
15253 * Modification to PKCS#7 encoding routines to output definite
15254 length encoding. Since currently the whole structures are in
15255 memory there's not real point in using indefinite length
15256 constructed encoding. However if OpenSSL is compiled with
15257 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
15261 * Added BIO_vprintf() and BIO_vsnprintf().
15265 * Added more prefixes to parse for in the strings written
15266 through a logging bio, to cover all the levels that are available
15267 through syslog. The prefixes are now:
15269 PANIC, EMERG, EMR => LOG_EMERG
15270 ALERT, ALR => LOG_ALERT
15271 CRIT, CRI => LOG_CRIT
15272 ERROR, ERR => LOG_ERR
15273 WARNING, WARN, WAR => LOG_WARNING
15274 NOTICE, NOTE, NOT => LOG_NOTICE
15275 INFO, INF => LOG_INFO
15276 DEBUG, DBG => LOG_DEBUG
15278 and as before, if none of those prefixes are present at the
15279 beginning of the string, LOG_ERR is chosen.
15281 On Win32, the `LOG_*` levels are mapped according to this:
15283 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
15284 LOG_WARNING => EVENTLOG_WARNING_TYPE
15285 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
15289 * Made it possible to reconfigure with just the configuration
15290 argument "reconf" or "reconfigure". The command line arguments
15291 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
15292 and are retrieved from there when reconfiguring.
15298 *Assar Westerlund <assar@sics.se>, Richard Levitte*
15300 * Add the arguments -CAfile and -CApath to the pkcs12 utility.
15304 * The obj_dat.pl script was messing up the sorting of object
15305 names. The reason was that it compared the quoted version
15306 of strings as a result "OCSP" > "OCSP Signing" because
15307 " > SPACE. Changed script to store unquoted versions of
15308 names and add quotes on output. It was also omitting some
15309 names from the lookup table if they were given a default
15310 value (that is if SN is missing it is given the same
15311 value as LN and vice versa), these are now added on the
15312 grounds that if an object has a name we should be able to
15313 look it up. Finally added warning output when duplicate
15314 short or long names are found.
15318 * Changes needed for Tandem NSK.
15320 *Scott Uroff <scott@xypro.com>*
15322 * Fix SSL 2.0 rollback checking: Due to an off-by-one error in
15323 RSA_padding_check_SSLv23(), special padding was never detected
15324 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
15325 version rollback attacks was not effective.
15327 In s23_clnt.c, don't use special rollback-attack detection padding
15328 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
15329 client; similarly, in s23_srvr.c, don't do the rollback check if
15330 SSL 2.0 is the only protocol enabled in the server.
15334 * Make it possible to get hexdumps of unprintable data with 'openssl
15335 asn1parse'. By implication, the functions ASN1_parse_dump() and
15336 BIO_dump_indent() are added.
15340 * New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
15341 these print out strings and name structures based on various
15342 flags including RFC2253 support and proper handling of
15343 multibyte characters. Added options to the 'x509' utility
15344 to allow the various flags to be set.
15348 * Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
15349 Also change the functions X509_cmp_current_time() and
15350 X509_gmtime_adj() work with an ASN1_TIME structure,
15351 this will enable certificates using GeneralizedTime in validity
15352 dates to be checked.
15356 * Make the NEG_PUBKEY_BUG code (which tolerates invalid
15357 negative public key encodings) on by default,
15358 NO_NEG_PUBKEY_BUG can be set to disable it.
15362 * New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
15363 content octets. An i2c_ASN1_OBJECT is unnecessary because
15364 the encoding can be trivially obtained from the structure.
15368 * crypto/err.c locking bugfix: Use write locks (`CRYPTO_w_[un]lock`),
15369 not read locks (`CRYPTO_r_[un]lock`).
15373 * A first attempt at creating official support for shared
15374 libraries through configuration. I've kept it so the
15375 default is static libraries only, and the OpenSSL programs
15376 are always statically linked for now, but there are
15377 preparations for dynamic linking in place.
15378 This has been tested on Linux and Tru64.
15382 * Randomness polling function for Win9x, as described in:
15383 Peter Gutmann, Software Generation of Practically Strong
15388 * Fix so PRNG is seeded in req if using an already existing
15393 * New options to smime application. -inform and -outform
15394 allow alternative formats for the S/MIME message including
15395 PEM and DER. The -content option allows the content to be
15396 specified separately. This should allow things like Netscape
15397 form signing output easier to verify.
15401 * Fix the ASN1 encoding of tags using the 'long form'.
15405 * New ASN1 functions, `i2c_*` and `c2i_*` for INTEGER and BIT
15406 STRING types. These convert content octets to and from the
15407 underlying type. The actual tag and length octets are
15408 already assumed to have been read in and checked. These
15409 are needed because all other string types have virtually
15410 identical handling apart from the tag. By having versions
15411 of the ASN1 functions that just operate on content octets
15412 IMPLICIT tagging can be handled properly. It also allows
15413 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
15414 and ASN1_INTEGER are identical apart from the tag.
15418 * Change the handling of OID objects as follows:
15420 - New object identifiers are inserted in objects.txt, following
15421 the syntax given in [crypto/objects/README.md](crypto/objects/README.md).
15422 - objects.pl is used to process obj_mac.num and create a new
15424 - obj_dat.pl is used to create a new obj_dat.h, using the data in
15427 This is currently kind of a hack, and the perl code in objects.pl
15428 isn't very elegant, but it works as I intended. The simplest way
15429 to check that it worked correctly is to look in obj_dat.h and
15430 check the array nid_objs and make sure the objects haven't moved
15431 around (this is important!). Additions are OK, as well as
15432 consistent name changes.
15436 * Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
15440 * Addition of the command line parameter '-rand file' to 'openssl req'.
15441 The given file adds to whatever has already been seeded into the
15442 random pool through the RANDFILE configuration file option or
15443 environment variable, or the default random state file.
15447 * mkstack.pl now sorts each macro group into lexical order.
15448 Previously the output order depended on the order the files
15449 appeared in the directory, resulting in needless rewriting
15454 * Patches to make OpenSSL compile under Win32 again. Mostly
15455 work arounds for the VC++ problem that it treats func() as
15456 func(void). Also stripped out the parts of mkdef.pl that
15457 added extra typesafe functions: these no longer exist.
15461 * Reorganisation of the stack code. The macros are now all
15462 collected in safestack.h . Each macro is defined in terms of
15463 a "stack macro" of the form `SKM_<name>(type, a, b)`. The
15464 DEBUG_SAFESTACK is now handled in terms of function casts,
15465 this has the advantage of retaining type safety without the
15466 use of additional functions. If DEBUG_SAFESTACK is not defined
15467 then the non typesafe macros are used instead. Also modified the
15468 mkstack.pl script to handle the new form. Needs testing to see
15469 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
15470 the default if no major problems. Similar behaviour for ASN1_SET_OF
15471 and PKCS12_STACK_OF.
15475 * When some versions of IIS use the 'NET' form of private key the
15476 key derivation algorithm is different. Normally MD5(password) is
15477 used as a 128 bit RC4 key. In the modified case
15478 MD5(MD5(password) + "SGCKEYSALT") is used instead. Added some
15479 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
15480 as the old Netscape_RSA functions except they have an additional
15481 'sgckey' parameter which uses the modified algorithm. Also added
15482 an -sgckey command line option to the rsa utility. Thanks to
15483 Adrian Peck <bertie@ncipher.com> for posting details of the modified
15484 algorithm to openssl-dev.
15488 * The evp_local.h macros were using 'c.##kname' which resulted in
15489 invalid expansion on some systems (SCO 5.0.5 for example).
15490 Corrected to 'c.kname'.
15492 *Phillip Porch <root@theporch.com>*
15494 * New X509_get1_email() and X509_REQ_get1_email() functions that return
15495 a STACK of email addresses from a certificate or request, these look
15496 in the subject name and the subject alternative name extensions and
15497 omit any duplicate addresses.
15501 * Re-implement BN_mod_exp2_mont using independent (and larger) windows.
15502 This makes DSA verification about 2 % faster.
15506 * Increase maximum window size in `BN_mod_exp_...` to 6 bits instead of 5
15507 (meaning that now 2^5 values will be precomputed, which is only 4 KB
15508 plus overhead for 1024 bit moduli).
15509 This makes exponentiations about 0.5 % faster for 1024 bit
15510 exponents (as measured by "openssl speed rsa2048").
15514 * Rename memory handling macros to avoid conflicts with other
15516 Malloc => OPENSSL_malloc
15517 Malloc_locked => OPENSSL_malloc_locked
15518 Realloc => OPENSSL_realloc
15519 Free => OPENSSL_free
15523 * New function BN_mod_exp_mont_word for small bases (roughly 15%
15524 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
15528 * CygWin32 support.
15530 *John Jarvie <jjarvie@newsguy.com>*
15532 * The type-safe stack code has been rejigged. It is now only compiled
15533 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
15534 by default all type-specific stack functions are "#define"d back to
15535 standard stack functions. This results in more streamlined output
15536 but retains the type-safety checking possibilities of the original
15541 * The STACK code has been cleaned up, and certain type declarations
15542 that didn't make a lot of sense have been brought in line. This has
15543 also involved a cleanup of sorts in safestack.h to more correctly
15544 map type-safe stack functions onto their plain stack counterparts.
15545 This work has also resulted in a variety of "const"ifications of
15546 lots of the code, especially `_cmp` operations which should normally
15547 be prototyped with "const" parameters anyway.
15551 * When generating bytes for the first time in md_rand.c, 'stir the pool'
15552 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
15553 (The PRNG state consists of two parts, the large pool 'state' and 'md',
15554 where all of 'md' is used each time the PRNG is used, but 'state'
15555 is used only indexed by a cyclic counter. As entropy may not be
15556 well distributed from the beginning, 'md' is important as a
15557 chaining variable. However, the output function chains only half
15558 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
15559 all of 'md', and seeding with STATE_SIZE dummy bytes will result
15560 in all of 'state' being rewritten, with the new values depending
15561 on virtually all of 'md'. This overcomes the 80 bit limitation.)
15565 * In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
15566 the handshake is continued after ssl_verify_cert_chain();
15567 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
15568 can lead to 'unexplainable' connection aborts later.
15570 *Bodo Moeller; problem tracked down by Lutz Jaenicke*
15572 * Major EVP API cipher revision.
15573 Add hooks for extra EVP features. This allows various cipher
15574 parameters to be set in the EVP interface. Support added for variable
15575 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
15576 setting of RC2 and RC5 parameters.
15578 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
15581 Remove lots of duplicated code from the EVP library. For example *every*
15582 cipher init() function handles the 'iv' in the same way according to the
15583 cipher mode. They also all do nothing if the 'key' parameter is NULL and
15584 for CFB and OFB modes they zero ctx->num.
15586 New functionality allows removal of S/MIME code RC2 hack.
15588 Most of the routines have the same form and so can be declared in terms
15591 By shifting this to the top level EVP_CipherInit() it can be removed from
15592 all individual ciphers. If the cipher wants to handle IVs or keys
15593 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
15596 Change lots of functions like EVP_EncryptUpdate() to now return a
15597 value: although software versions of the algorithms cannot fail
15598 any installed hardware versions can.
15602 * Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
15603 this option is set, tolerate broken clients that send the negotiated
15604 protocol version number instead of the requested protocol version
15609 * Call dh_tmp_cb (set by `..._TMP_DH_CB`) with correct 'is_export' flag;
15610 i.e. non-zero for export ciphersuites, zero otherwise.
15611 Previous versions had this flag inverted, inconsistent with
15612 rsa_tmp_cb (..._TMP_RSA_CB).
15614 *Bodo Moeller; problem reported by Amit Chopra*
15616 * Add missing DSA library text string. Work around for some IIS
15617 key files with invalid SEQUENCE encoding.
15621 * Add a document (doc/standards.txt) that list all kinds of standards
15622 and so on that are implemented in OpenSSL.
15626 * Enhance c_rehash script. Old version would mishandle certificates
15627 with the same subject name hash and wouldn't handle CRLs at all.
15628 Added -fingerprint option to crl utility, to support new c_rehash
15633 * Eliminate non-ANSI declarations in crypto.h and stack.h.
15637 * Fix for SSL server purpose checking. Server checking was
15638 rejecting certificates which had extended key usage present
15639 but no ssl client purpose.
15641 *Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>*
15643 * Make PKCS#12 code work with no password. The PKCS#12 spec
15644 is a little unclear about how a blank password is handled.
15645 Since the password in encoded as a BMPString with terminating
15646 double NULL a zero length password would end up as just the
15647 double NULL. However no password at all is different and is
15648 handled differently in the PKCS#12 key generation code. NS
15649 treats a blank password as zero length. MSIE treats it as no
15650 password on export: but it will try both on import. We now do
15651 the same: PKCS12_parse() tries zero length and no password if
15652 the password is set to "" or NULL (NULL is now a valid password:
15653 it wasn't before) as does the pkcs12 application.
15657 * Bugfixes in `apps/x509.c`: Avoid a memory leak; and don't use
15658 perror when PEM_read_bio_X509_REQ fails, the error message must
15659 be obtained from the error queue.
15663 * Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
15664 it in ERR_remove_state if appropriate, and change ERR_get_state
15665 accordingly to avoid race conditions (this is necessary because
15666 thread_hash is no longer constant once set).
15670 * Bugfix for linux-elf makefile.one.
15674 * RSA_get_default_method() will now cause a default
15675 RSA_METHOD to be chosen if one doesn't exist already.
15676 Previously this was only set during a call to RSA_new()
15677 or RSA_new_method(NULL) meaning it was possible for
15678 RSA_get_default_method() to return NULL.
15682 * Added native name translation to the existing DSO code
15683 that will convert (if the flag to do so is set) filenames
15684 that are sufficiently small and have no path information
15685 into a canonical native form. Eg. "blah" converted to
15686 "libblah.so" or "blah.dll" etc.
15690 * New function ERR_error_string_n(e, buf, len) which is like
15691 ERR_error_string(e, buf), but writes at most 'len' bytes
15692 including the 0 terminator. For ERR_error_string_n, 'buf'
15695 *Damien Miller <djm@mindrot.org>, Bodo Moeller*
15697 * CONF library reworked to become more general. A new CONF
15698 configuration file reader "class" is implemented as well as a
15699 new functions (`NCONF_*`, for "New CONF") to handle it. The now
15700 old `CONF_*` functions are still there, but are reimplemented to
15701 work in terms of the new functions. Also, a set of functions
15702 to handle the internal storage of the configuration data is
15703 provided to make it easier to write new configuration file
15704 reader "classes" (I can definitely see something reading a
15705 configuration file in XML format, for example), called `_CONF_*`,
15706 or "the configuration storage API"...
15708 The new configuration file reading functions are:
15710 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
15711 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
15713 NCONF_default, NCONF_WIN32
15715 NCONF_dump_fp, NCONF_dump_bio
15717 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
15718 NCONF_new creates a new CONF object. This works in the same way
15719 as other interfaces in OpenSSL, like the BIO interface.
15720 `NCONF_dump_*` dump the internal storage of the configuration file,
15721 which is useful for debugging. All other functions take the same
15722 arguments as the old `CONF_*` functions with the exception of the
15723 first that must be a `CONF *` instead of a `LHASH *`.
15725 To make it easier to use the new classes with the old `CONF_*` functions,
15726 the function CONF_set_default_method is provided.
15730 * Add '-tls1' option to 'openssl ciphers', which was already
15731 mentioned in the documentation but had not been implemented.
15732 (This option is not yet really useful because even the additional
15733 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
15737 * Initial DSO code added into libcrypto for letting OpenSSL (and
15738 OpenSSL-based applications) load shared libraries and bind to
15739 them in a portable way.
15741 *Geoff Thorpe, with contributions from Richard Levitte*
15743 ### Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
15745 * Make sure _lrotl and _lrotr are only used with MSVC.
15747 * Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
15748 (the default implementation of RAND_status).
15750 * Rename openssl x509 option '-crlext', which was added in 0.9.5,
15751 to '-clrext' (= clear extensions), as intended and documented.
15752 *Bodo Moeller; inconsistency pointed out by Michael Attili
15753 <attili@amaxo.com>*
15755 * Fix for HMAC. It wasn't zeroing the rest of the block if the key length
15756 was larger than the MD block size.
15758 *Steve Henson, pointed out by Yost William <YostW@tce.com>*
15760 * Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
15761 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
15762 using the passed key: if the passed key was a private key the result
15763 of X509_print(), for example, would be to print out all the private key
15768 * des_quad_cksum() byte order bug fix.
15769 *Ulf Möller, using the problem description in krb4-0.9.7, where
15770 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>*
15772 * Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
15775 *Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>*
15777 * For easily testing in shell scripts whether some command
15778 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
15779 returns with exit code 0 iff no command of the given name is available.
15780 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
15781 the output goes to stdout and nothing is printed to stderr.
15782 Additional arguments are always ignored.
15784 Since for each cipher there is a command of the same name,
15785 the 'no-cipher' compilation switches can be tested this way.
15787 ('openssl no-XXX' is not able to detect pseudo-commands such
15788 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
15792 * Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
15796 * For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
15797 is set; it will be thrown away anyway because each handshake creates
15799 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
15800 to parameters -- in previous versions (since OpenSSL 0.9.3) the
15801 'default key' from SSL_CTX_set_tmp_dh would always be lost, meaning
15802 you effectively got SSL_OP_SINGLE_DH_USE when using this macro.
15806 * New s_client option -ign_eof: EOF at stdin is ignored, and
15807 'Q' and 'R' lose their special meanings (quit/renegotiate).
15808 This is part of what -quiet does; unlike -quiet, -ign_eof
15809 does not suppress any output.
15813 * Add compatibility options to the purpose and trust code. The
15814 purpose X509_PURPOSE_ANY is "any purpose" which automatically
15815 accepts a certificate or CA, this was the previous behaviour,
15816 with all the associated security issues.
15818 X509_TRUST_COMPAT is the old trust behaviour: only and
15819 automatically trust self signed roots in certificate store. A
15820 new trust setting X509_TRUST_DEFAULT is used to specify that
15821 a purpose has no associated trust setting and it should instead
15822 use the value in the default purpose.
15826 * Fix the PKCS#8 DSA private key code so it decodes keys again
15827 and fix a memory leak.
15831 * In util/mkerr.pl (which implements 'make errors'), preserve
15832 reason strings from the previous version of the .c file, as
15833 the default to have only downcase letters (and digits) in
15834 automatically generated reasons codes is not always appropriate.
15838 * In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
15839 using strerror. Previously, ERR_reason_error_string() returned
15840 library names as reason strings for SYSerr; but SYSerr is a special
15841 case where small numbers are errno values, not library numbers.
15845 * Add '-dsaparam' option to 'openssl dhparam' application. This
15846 converts DSA parameters into DH parameters. (When creating parameters,
15847 DSA_generate_parameters is used.)
15851 * Include 'length' (recommended exponent length) in C code generated
15852 by 'openssl dhparam -C'.
15856 * The second argument to set_label in perlasm was already being used
15857 so couldn't be used as a "file scope" flag. Moved to third argument
15862 * In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
15863 instead of RAND_bytes for encryption IVs and salts.
15867 * Include RAND_status() into RAND_METHOD instead of implementing
15868 it only for md_rand.c Otherwise replacing the PRNG by calling
15869 RAND_set_rand_method would be impossible.
15873 * Don't let DSA_generate_key() enter an infinite loop if the random
15874 number generation fails.
15878 * New 'rand' application for creating pseudo-random output.
15882 * Added configuration support for Linux/IA64
15884 *Rolf Haberrecker <rolf@suse.de>*
15886 * Assembler module support for Mingw32.
15890 * Shared library support for HPUX (in shlib/).
15892 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous*
15894 * Shared library support for Solaris gcc.
15896 *Lutz Behnke <behnke@trustcenter.de>*
15898 ### Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
15900 * PKCS7_encrypt() was adding text MIME headers twice because they
15901 were added manually and by SMIME_crlf_copy().
15905 * In bntest.c don't call BN_rand with zero bits argument.
15907 *Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>*
15909 * BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
15910 case was implemented. This caused BN_div_recp() to fail occasionally.
15914 * Add an optional second argument to the set_label() in the perl
15915 assembly language builder. If this argument exists and is set
15916 to 1 it signals that the assembler should use a symbol whose
15917 scope is the entire file, not just the current function. This
15918 is needed with MASM which uses the format label:: for this scope.
15920 *Steve Henson, pointed out by Peter Runestig <peter@runestig.com>*
15922 * Change the ASN1 types so they are typedefs by default. Before
15923 almost all types were #define'd to ASN1_STRING which was causing
15924 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
15929 * Change names of new functions to the new get1/get0 naming
15930 convention: After 'get1', the caller owns a reference count
15931 and has to call `..._free`; 'get0' returns a pointer to some
15932 data structure without incrementing reference counters.
15933 (Some of the existing 'get' functions increment a reference
15934 counter, some don't.)
15935 Similarly, 'set1' and 'add1' functions increase reference
15936 counters or duplicate objects.
15940 * Allow for the possibility of temp RSA key generation failure:
15941 the code used to assume it always worked and crashed on failure.
15945 * Fix potential buffer overrun problem in BIO_printf().
15946 *Ulf Möller, using public domain code by Patrick Powell; problem
15947 pointed out by David Sacerdote <das33@cornell.edu>*
15949 * Support EGD <http://www.lothar.com/tech/crypto/>. New functions
15950 RAND_egd() and RAND_status(). In the command line application,
15951 the EGD socket can be specified like a seed file using RANDFILE
15956 * Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
15957 Some CAs (e.g. Verisign) distribute certificates in this form.
15961 * Remove the SSL_ALLOW_ADH compile option and set the default cipher
15962 list to exclude them. This means that no special compilation option
15963 is needed to use anonymous DH: it just needs to be included in the
15968 * Change the EVP_MD_CTX_type macro so its meaning consistent with
15969 EVP_MD_type. The old functionality is available in a new macro called
15970 EVP_MD_md(). Change code that uses it and update docs.
15974 * `..._ctrl` functions now have corresponding `..._callback_ctrl` functions
15975 where the `void *` argument is replaced by a function pointer argument.
15976 Previously `void *` was abused to point to functions, which works on
15977 many platforms, but is not correct. As these functions are usually
15978 called by macros defined in OpenSSL header files, most source code
15979 should work without changes.
15983 * `<openssl/opensslconf.h>` (which is created by Configure) now contains
15984 sections with information on -D... compiler switches used for
15985 compiling the library so that applications can see them. To enable
15986 one of these sections, a pre-processor symbol `OPENSSL_..._DEFINES`
15987 must be defined. E.g.,
15988 #define OPENSSL_ALGORITHM_DEFINES
15989 #include <openssl/opensslconf.h>
15990 defines all pertinent `NO_<algo>` symbols, such as NO_IDEA, NO_RSA, etc.
15992 *Richard Levitte, Ulf and Bodo Möller*
15994 * Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
15999 * Change the 'other' type in certificate aux info to a STACK_OF
16000 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
16001 the required ASN1 format: arbitrary types determined by an OID.
16005 * Add some PEM_write_X509_REQ_NEW() functions and a command line
16006 argument to 'req'. This is not because the function is newer or
16007 better than others it just uses the work 'NEW' in the certificate
16008 request header lines. Some software needs this.
16012 * Reorganise password command line arguments: now passwords can be
16013 obtained from various sources. Delete the PEM_cb function and make
16014 it the default behaviour: i.e. if the callback is NULL and the
16015 usrdata argument is not NULL interpret it as a null terminated pass
16016 phrase. If usrdata and the callback are NULL then the pass phrase
16017 is prompted for as usual.
16021 * Add support for the Compaq Atalla crypto accelerator. If it is installed,
16022 the support is automatically enabled. The resulting binaries will
16023 autodetect the card and use it if present.
16025 *Ben Laurie and Compaq Inc.*
16027 * Work around for Netscape hang bug. This sends certificate request
16028 and server done in one record. Since this is perfectly legal in the
16029 SSL/TLS protocol it isn't a "bug" option and is on by default. See
16030 the bugs/SSLv3 entry for more info.
16034 * HP-UX tune-up: new unified configs, HP C compiler bug workaround.
16038 * Add -rand argument to smime and pkcs12 applications and read/write
16043 * New 'passwd' tool for crypt(3) and apr1 password hashes.
16047 * Add command line password options to the remaining applications.
16051 * Bug fix for BN_div_recp() for numerators with an even number of
16056 * More tests in bntest.c, and changed test_bn output.
16060 * ./config recognizes MacOS X now.
16064 * Bug fix for BN_div() when the first words of num and divisor are
16065 equal (it gave wrong results if `(rem=(n1-q*d0)&BN_MASK2) < d0)`.
16069 * Add support for various broken PKCS#8 formats, and command line
16070 options to produce them.
16074 * New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
16075 get temporary BIGNUMs from a BN_CTX.
16079 * Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
16084 * Change the `SSLeay_add_all_*()` functions to `OpenSSL_add_all_*()` and
16085 include a #define from the old name to the new. The original intent
16086 was that statically linked binaries could for example just call
16087 SSLeay_add_all_ciphers() to just add ciphers to the table and not
16088 link with digests. This never worked because SSLeay_add_all_digests()
16089 and SSLeay_add_all_ciphers() were in the same source file so calling
16090 one would link with the other. They are now in separate source files.
16094 * Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
16098 * Use a less unusual form of the Miller-Rabin primality test (it used
16099 a binary algorithm for exponentiation integrated into the Miller-Rabin
16100 loop, our standard modexp algorithms are faster).
16104 * Support for the EBCDIC character set completed.
16106 *Martin Kraemer <Martin.Kraemer@Mch.SNI.De>*
16108 * Source code cleanups: use const where appropriate, eliminate casts,
16109 use `void *` instead of `char *` in lhash.
16113 * Bugfix: ssl3_send_server_key_exchange was not restartable
16114 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
16115 this the server could overwrite ephemeral keys that the client
16120 * Turn DSA_is_prime into a macro that calls BN_is_prime,
16121 using 50 iterations of the Rabin-Miller test.
16123 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
16124 iterations of the Rabin-Miller test as required by the appendix
16125 to FIPS PUB 186[-1]) instead of DSA_is_prime.
16126 As BN_is_prime_fasttest includes trial division, DSA parameter
16127 generation becomes much faster.
16129 This implies a change for the callback functions in DSA_is_prime
16130 and DSA_generate_parameters: The callback function is called once
16131 for each positive witness in the Rabin-Miller test, not just
16132 occasionally in the inner loop; and the parameters to the
16133 callback function now provide an iteration count for the outer
16134 loop rather than for the current invocation of the inner loop.
16135 DSA_generate_parameters additionally can call the callback
16136 function with an 'iteration count' of -1, meaning that a
16137 candidate has passed the trial division test (when q is generated
16138 from an application-provided seed, trial division is skipped).
16142 * New function BN_is_prime_fasttest that optionally does trial
16143 division before starting the Rabin-Miller test and has
16144 an additional BN_CTX * argument (whereas BN_is_prime always
16145 has to allocate at least one BN_CTX).
16146 'callback(1, -1, cb_arg)' is called when a number has passed the
16147 trial division stage.
16151 * Fix for bug in CRL encoding. The validity dates weren't being handled
16156 * New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
16160 * New function BN_pseudo_rand().
16164 * Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
16165 bignum version of BN_from_montgomery() with the working code from
16166 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
16171 * Avoid a race condition in s2_clnt.c (function get_server_hello) that
16172 made it impossible to use the same SSL_SESSION data structure in
16173 SSL2 clients in multiple threads.
16177 * The return value of RAND_load_file() no longer counts bytes obtained
16178 by stat(). RAND_load_file(..., -1) is new and uses the complete file
16179 to seed the PRNG (previously an explicit byte count was required).
16181 *Ulf Möller, Bodo Möller*
16183 * Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
16184 used `char *` instead of `void *` and had casts all over the place.
16188 * Make BN_generate_prime() return NULL on error if ret!=NULL.
16192 * Retain source code compatibility for BN_prime_checks macro:
16193 BN_is_prime(..., BN_prime_checks, ...) now uses
16194 BN_prime_checks_for_size to determine the appropriate number of
16195 Rabin-Miller iterations.
16199 * Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
16200 DH_CHECK_P_NOT_SAFE_PRIME.
16201 (Check if this is true? OpenPGP calls them "strong".)
16205 * Merge the functionality of "dh" and "gendh" programs into a new program
16206 "dhparam". The old programs are retained for now but will handle DH keys
16207 (instead of parameters) in future.
16211 * Make the ciphers, s_server and s_client programs check the return values
16212 when a new cipher list is set.
16216 * Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
16217 ciphers. Before when the 56bit ciphers were enabled the sorting was
16220 The syntax for the cipher sorting has been extended to support sorting by
16221 cipher-strength (using the strength_bits hard coded in the tables).
16222 The new command is `@STRENGTH` (see also `doc/apps/ciphers.pod`).
16224 Fix a bug in the cipher-command parser: when supplying a cipher command
16225 string with an "undefined" symbol (neither command nor alphanumeric
16226 *A-Za-z0-9*, ssl_set_cipher_list used to hang in an endless loop. Now
16227 an error is flagged.
16229 Due to the strength-sorting extension, the code of the
16230 ssl_create_cipher_list() function was completely rearranged. I hope that
16231 the readability was also increased :-)
16233 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>*
16235 * Minor change to 'x509' utility. The -CAcreateserial option now uses 1
16236 for the first serial number and places 2 in the serial number file. This
16237 avoids problems when the root CA is created with serial number zero and
16238 the first user certificate has the same issuer name and serial number
16243 * Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
16244 the new code. Add documentation for this stuff.
16248 * Changes to X509_ATTRIBUTE utilities. These have been renamed from
16249 `X509_*()` to `X509at_*()` on the grounds that they don't handle X509
16250 structures and behave in an analogous way to the X509v3 functions:
16251 they shouldn't be called directly but wrapper functions should be used
16254 So we also now have some wrapper functions that call the X509at functions
16255 when passed certificate requests. (TO DO: similar things can be done with
16256 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
16257 things. Some of these need some d2i or i2d and print functionality
16258 because they handle more complex structures.)
16262 * Add missing #ifndefs that caused missing symbols when building libssl
16263 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
16264 NO_RSA in `ssl/s2*.c`.
16266 *Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller*
16268 * Precautions against using the PRNG uninitialized: RAND_bytes() now
16269 has a return value which indicates the quality of the random data
16270 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
16271 error queue. New function RAND_pseudo_bytes() generates output that is
16272 guaranteed to be unique but not unpredictable. RAND_add is like
16273 RAND_seed, but takes an extra argument for an entropy estimate
16274 (RAND_seed always assumes full entropy).
16278 * Do more iterations of Rabin-Miller probable prime test (specifically,
16279 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
16280 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
16281 in crypto/bn/bn_prime.c for the complete table). This guarantees a
16282 false-positive rate of at most 2^-80 for random input.
16286 * Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
16290 * New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
16291 in the 0.9.5 release), this returns the chain
16292 from an X509_CTX structure with a dup of the stack and all
16293 the X509 reference counts upped: so the stack will exist
16294 after X509_CTX_cleanup() has been called. Modify pkcs12.c
16297 Also make SSL_SESSION_print() print out the verify return
16302 * Add manpage for the pkcs12 command. Also change the default
16303 behaviour so MAC iteration counts are used unless the new
16304 -nomaciter option is used. This improves file security and
16305 only older versions of MSIE (4.0 for example) need it.
16309 * Honor the no-xxx Configure options when creating .DEF files.
16313 * Add PKCS#10 attributes to field table: challengePassword,
16314 unstructuredName and unstructuredAddress. These are taken from
16315 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
16316 international characters are used.
16318 More changes to X509_ATTRIBUTE code: allow the setting of types
16319 based on strings. Remove the 'loc' parameter when adding
16320 attributes because these will be a SET OF encoding which is sorted
16325 * Initial changes to the 'req' utility to allow request generation
16326 automation. This will allow an application to just generate a template
16327 file containing all the field values and have req construct the
16330 Initial support for X509_ATTRIBUTE handling. Stacks of these are
16331 used all over the place including certificate requests and PKCS#7
16332 structures. They are currently handled manually where necessary with
16333 some primitive wrappers for PKCS#7. The new functions behave in a
16334 manner analogous to the X509 extension functions: they allow
16335 attributes to be looked up by NID and added.
16337 Later something similar to the X509V3 code would be desirable to
16338 automatically handle the encoding, decoding and printing of the
16339 more complex types. The string types like challengePassword can
16340 be handled by the string table functions.
16342 Also modified the multi byte string table handling. Now there is
16343 a 'global mask' which masks out certain types. The table itself
16344 can use the flag STABLE_NO_MASK to ignore the mask setting: this
16345 is useful when for example there is only one permissible type
16346 (as in countryName) and using the mask might result in no valid
16351 * Clean up 'Finished' handling, and add functions SSL_get_finished and
16352 SSL_get_peer_finished to allow applications to obtain the latest
16353 Finished messages sent to the peer or expected from the peer,
16354 respectively. (SSL_get_peer_finished is usually the Finished message
16355 actually received from the peer, otherwise the protocol will be aborted.)
16357 As the Finished message are message digests of the complete handshake
16358 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
16359 be used for external authentication procedures when the authentication
16360 provided by SSL/TLS is not desired or is not enough.
16364 * Enhanced support for Alpha Linux is added. Now ./config checks if
16365 the host supports BWX extension and if Compaq C is present on the
16366 $PATH. Just exploiting of the BWX extension results in 20-30%
16367 performance kick for some algorithms, e.g. DES and RC4 to mention
16368 a couple. Compaq C in turn generates ~20% faster code for MD5 and
16373 * Add support for MS "fast SGC". This is arguably a violation of the
16374 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
16375 weak crypto and after checking the certificate is SGC a second one
16376 with strong crypto. MS SGC stops the first handshake after receiving
16377 the server certificate message and sends a second client hello. Since
16378 a server will typically do all the time consuming operations before
16379 expecting any further messages from the client (server key exchange
16380 is the most expensive) there is little difference between the two.
16382 To get OpenSSL to support MS SGC we have to permit a second client
16383 hello message after we have sent server done. In addition we have to
16384 reset the MAC if we do get this second client hello.
16388 * Add a function 'd2i_AutoPrivateKey()' this will automatically decide
16389 if a DER encoded private key is RSA or DSA traditional format. Changed
16390 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
16391 format DER encoded private key. Newer code should use PKCS#8 format which
16392 has the key type encoded in the ASN1 structure. Added DER private key
16393 support to pkcs8 application.
16397 * SSL 3/TLS 1 servers now don't request certificates when an anonymous
16398 ciphersuites has been selected (as required by the SSL 3/TLS 1
16399 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
16400 is set, we interpret this as a request to violate the specification
16401 (the worst that can happen is a handshake failure, and 'correct'
16402 behaviour would result in a handshake failure anyway).
16406 * In SSL_CTX_add_session, take into account that there might be multiple
16407 SSL_SESSION structures with the same session ID (e.g. when two threads
16408 concurrently obtain them from an external cache).
16409 The internal cache can handle only one SSL_SESSION with a given ID,
16410 so if there's a conflict, we now throw out the old one to achieve
16415 * Add OIDs for idea and blowfish in CBC mode. This will allow both
16416 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
16417 some routines that use cipher OIDs: some ciphers do not have OIDs
16418 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
16423 * Simplify the trust setting structure and code. Now we just have
16424 two sequences of OIDs for trusted and rejected settings. These will
16425 typically have values the same as the extended key usage extension
16426 and any application specific purposes.
16428 The trust checking code now has a default behaviour: it will just
16429 check for an object with the same NID as the passed id. Functions can
16430 be provided to override either the default behaviour or the behaviour
16431 for a given id. SSL client, server and email already have functions
16432 in place for compatibility: they check the NID and also return "trusted"
16433 if the certificate is self signed.
16437 * Add d2i,i2d bio/fp functions for PrivateKey: these convert the
16438 traditional format into an EVP_PKEY structure.
16442 * Add a password callback function PEM_cb() which either prompts for
16443 a password if usr_data is NULL or otherwise assumes it is a null
16444 terminated password. Allow passwords to be passed on command line
16445 environment or config files in a few more utilities.
16449 * Add a bunch of DER and PEM functions to handle PKCS#8 format private
16450 keys. Add some short names for PKCS#8 PBE algorithms and allow them
16451 to be specified on the command line for the pkcs8 and pkcs12 utilities.
16452 Update documentation.
16456 * Support for ASN1 "NULL" type. This could be handled before by using
16457 ASN1_TYPE but there wasn't any function that would try to read a NULL
16458 and produce an error if it couldn't. For compatibility we also have
16459 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
16460 don't allocate anything because they don't need to.
16464 * Initial support for MacOS is now provided. Examine INSTALL.MacOS
16467 *Andy Polyakov, Roy Woods <roy@centicsystems.ca>*
16469 * Rebuild of the memory allocation routines used by OpenSSL code and
16470 possibly others as well. The purpose is to make an interface that
16471 provide hooks so anyone can build a separate set of allocation and
16472 deallocation routines to be used by OpenSSL, for example memory
16473 pool implementations, or something else, which was previously hard
16474 since Malloc(), Realloc() and Free() were defined as macros having
16475 the values malloc, realloc and free, respectively (except for Win32
16476 compilations). The same is provided for memory debugging code.
16477 OpenSSL already comes with functionality to find memory leaks, but
16478 this gives people a chance to debug other memory problems.
16480 With these changes, a new set of functions and macros have appeared:
16482 CRYPTO_set_mem_debug_functions() [F]
16483 CRYPTO_get_mem_debug_functions() [F]
16484 CRYPTO_dbg_set_options() [F]
16485 CRYPTO_dbg_get_options() [F]
16486 CRYPTO_malloc_debug_init() [M]
16488 The memory debug functions are NULL by default, unless the library
16489 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
16490 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
16491 gives the standard debugging functions that come with OpenSSL) or
16492 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
16493 provided by the library user) must be used. When the standard
16494 debugging functions are used, CRYPTO_dbg_set_options can be used to
16495 request additional information:
16496 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
16497 the CRYPTO_MDEBUG_xxx macro when compiling the library.
16499 Also, things like CRYPTO_set_mem_functions will always give the
16500 expected result (the new set of functions is used for allocation
16501 and deallocation) at all times, regardless of platform and compiler
16504 To finish it up, some functions that were never use in any other
16505 way than through macros have a new API and new semantic:
16507 CRYPTO_dbg_malloc()
16508 CRYPTO_dbg_realloc()
16511 All macros of value have retained their old syntax.
16513 *Richard Levitte and Bodo Moeller*
16515 * Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
16516 ordering of SMIMECapabilities wasn't in "strength order" and there
16517 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
16522 * Some ASN1 types with illegal zero length encoding (INTEGER,
16523 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
16525 *Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson*
16527 * Merge in my S/MIME library for OpenSSL. This provides a simple
16528 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
16529 functionality to handle multipart/signed properly) and a utility
16530 called 'smime' to call all this stuff. This is based on code I
16531 originally wrote for Celo who have kindly allowed it to be
16532 included in OpenSSL.
16536 * Add variants des_set_key_checked and des_set_key_unchecked of
16537 des_set_key (aka des_key_sched). Global variable des_check_key
16538 decides which of these is called by des_set_key; this way
16539 des_check_key behaves as it always did, but applications and
16540 the library itself, which was buggy for des_check_key == 1,
16541 have a cleaner way to pick the version they need.
16545 * New function PKCS12_newpass() which changes the password of a
16550 * Modify X509_TRUST and X509_PURPOSE so it also uses a static and
16551 dynamic mix. In both cases the ids can be used as an index into the
16552 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
16553 functions so they accept a list of the field values and the
16554 application doesn't need to directly manipulate the X509_TRUST
16559 * Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
16564 * Modify the way the V3 extension code looks up extensions. This now
16565 works in a similar way to the object code: we have some "standard"
16566 extensions in a static table which is searched with OBJ_bsearch()
16567 and the application can add dynamic ones if needed. The file
16568 crypto/x509v3/ext_dat.h now has the info: this file needs to be
16569 updated whenever a new extension is added to the core code and kept
16570 in ext_nid order. There is a simple program 'tabtest.c' which checks
16571 this. New extensions are not added too often so this file can readily
16572 be maintained manually.
16574 There are two big advantages in doing things this way. The extensions
16575 can be looked up immediately and no longer need to be "added" using
16576 X509V3_add_standard_extensions(): this function now does nothing.
16577 Side note: I get *lots* of email saying the extension code doesn't
16578 work because people forget to call this function.
16579 Also no dynamic allocation is done unless new extensions are added:
16580 so if we don't add custom extensions there is no need to call
16581 X509V3_EXT_cleanup().
16585 * Modify enc utility's salting as follows: make salting the default. Add a
16586 magic header, so unsalted files fail gracefully instead of just decrypting
16587 to garbage. This is because not salting is a big security hole, so people
16588 should be discouraged from doing it.
16592 * Fixes and enhancements to the 'x509' utility. It allowed a message
16593 digest to be passed on the command line but it only used this
16594 parameter when signing a certificate. Modified so all relevant
16595 operations are affected by the digest parameter including the
16596 -fingerprint and -x509toreq options. Also -x509toreq choked if a
16597 DSA key was used because it didn't fix the digest.
16601 * Initial certificate chain verify code. Currently tests the untrusted
16602 certificates for consistency with the verify purpose (which is set
16603 when the X509_STORE_CTX structure is set up) and checks the pathlength.
16605 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
16606 this is because it will reject chains with invalid extensions whereas
16607 every previous version of OpenSSL and SSLeay made no checks at all.
16609 Trust code: checks the root CA for the relevant trust settings. Trust
16610 settings have an initial value consistent with the verify purpose: e.g.
16611 if the verify purpose is for SSL client use it expects the CA to be
16612 trusted for SSL client use. However the default value can be changed to
16613 permit custom trust settings: one example of this would be to only trust
16614 certificates from a specific "secure" set of CAs.
16616 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
16617 which should be used for version portability: especially since the
16618 verify structure is likely to change more often now.
16620 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
16621 to set them. If not set then assume SSL clients will verify SSL servers
16624 Two new options to the verify program: -untrusted allows a set of
16625 untrusted certificates to be passed in and -purpose which sets the
16626 intended purpose of the certificate. If a purpose is set then the
16627 new chain verify code is used to check extension consistency.
16631 * Support for the authority information access extension.
16635 * Modify RSA and DSA PEM read routines to transparently handle
16636 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
16637 public keys in a format compatible with certificate
16638 SubjectPublicKeyInfo structures. Unfortunately there were already
16639 functions called *_PublicKey_* which used various odd formats so
16640 these are retained for compatibility: however the DSA variants were
16641 never in a public release so they have been deleted. Changed dsa/rsa
16642 utilities to handle the new format: note no releases ever handled public
16643 keys so we should be OK.
16645 The primary motivation for this change is to avoid the same fiasco
16646 that dogs private keys: there are several incompatible private key
16647 formats some of which are standard and some OpenSSL specific and
16648 require various evil hacks to allow partial transparent handling and
16649 even then it doesn't work with DER formats. Given the option anything
16650 other than PKCS#8 should be dumped: but the other formats have to
16651 stay in the name of compatibility.
16653 With public keys and the benefit of hindsight one standard format
16654 is used which works with EVP_PKEY, RSA or DSA structures: though
16655 it clearly returns an error if you try to read the wrong kind of key.
16657 Added a -pubkey option to the 'x509' utility to output the public key.
16658 Also rename the `EVP_PKEY_get_*()` to `EVP_PKEY_rget_*()`
16659 (renamed to `EVP_PKEY_get1_*()` in the OpenSSL 0.9.5 release) and add
16660 `EVP_PKEY_rset_*()` functions (renamed to `EVP_PKEY_set1_*()`)
16661 that do the same as the `EVP_PKEY_assign_*()` except they up the
16662 reference count of the added key (they don't "swallow" the
16667 * Fixes to crypto/x509/by_file.c the code to read in certificates and
16668 CRLs would fail if the file contained no certificates or no CRLs:
16669 added a new function to read in both types and return the number
16670 read: this means that if none are read it will be an error. The
16671 DER versions of the certificate and CRL reader would always fail
16672 because it isn't possible to mix certificates and CRLs in DER format
16673 without choking one or the other routine. Changed this to just read
16674 a certificate: this is the best we can do. Also modified the code
16675 in `apps/verify.c` to take notice of return codes: it was previously
16676 attempting to read in certificates from NULL pointers and ignoring
16677 any errors: this is one reason why the cert and CRL reader seemed
16678 to work. It doesn't check return codes from the default certificate
16679 routines: these may well fail if the certificates aren't installed.
16683 * Code to support otherName option in GeneralName.
16687 * First update to verify code. Change the verify utility
16688 so it warns if it is passed a self signed certificate:
16689 for consistency with the normal behaviour. X509_verify
16690 has been modified to it will now verify a self signed
16691 certificate if *exactly* the same certificate appears
16692 in the store: it was previously impossible to trust a
16693 single self signed certificate. This means that:
16694 openssl verify ss.pem
16695 now gives a warning about a self signed certificate but
16696 openssl verify -CAfile ss.pem ss.pem
16701 * For servers, store verify_result in SSL_SESSION data structure
16702 (and add it to external session representation).
16703 This is needed when client certificate verifications fails,
16704 but an application-provided verification callback (set by
16705 SSL_CTX_set_cert_verify_callback) allows accepting the session
16706 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
16707 but returns 1): When the session is reused, we have to set
16708 ssl->verify_result to the appropriate error code to avoid
16711 *Bodo Moeller, problem pointed out by Lutz Jaenicke*
16713 * Fix a bug in the new PKCS#7 code: it didn't consider the
16714 case in PKCS7_dataInit() where the signed PKCS7 structure
16715 didn't contain any existing data because it was being created.
16717 *Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson*
16719 * Add a salt to the key derivation routines in enc.c. This
16720 forms the first 8 bytes of the encrypted file. Also add a
16721 -S option to allow a salt to be input on the command line.
16725 * New function X509_cmp(). Oddly enough there wasn't a function
16726 to compare two certificates. We do this by working out the SHA1
16727 hash and comparing that. X509_cmp() will be needed by the trust
16732 * SSL_get1_session() is like SSL_get_session(), but increments
16733 the reference count in the SSL_SESSION returned.
16735 *Geoff Thorpe <geoff@eu.c2.net>*
16737 * Fix for 'req': it was adding a null to request attributes.
16738 Also change the X509_LOOKUP and X509_INFO code to handle
16739 certificate auxiliary information.
16743 * Add support for 40 and 64 bit RC2 and RC4 algorithms: document
16748 * Add the possibility to add extra information to the memory leak
16749 detecting output, to form tracebacks, showing from where each
16750 allocation was originated: CRYPTO_push_info("constant string") adds
16751 the string plus current file name and line number to a per-thread
16752 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
16753 is like calling CYRPTO_pop_info() until the stack is empty.
16754 Also updated memory leak detection code to be multi-thread-safe.
16758 * Add options -text and -noout to pkcs7 utility and delete the
16759 encryption options which never did anything. Update docs.
16763 * Add options to some of the utilities to allow the pass phrase
16764 to be included on either the command line (not recommended on
16765 OSes like Unix) or read from the environment. Update the
16766 manpages and fix a few bugs.
16770 * Add a few manpages for some of the openssl commands.
16774 * Fix the -revoke option in ca. It was freeing up memory twice,
16775 leaking and not finding already revoked certificates.
16779 * Extensive changes to support certificate auxiliary information.
16780 This involves the use of X509_CERT_AUX structure and X509_AUX
16781 functions. An X509_AUX function such as PEM_read_X509_AUX()
16782 can still read in a certificate file in the usual way but it
16783 will also read in any additional "auxiliary information". By
16784 doing things this way a fair degree of compatibility can be
16785 retained: existing certificates can have this information added
16786 using the new 'x509' options.
16788 Current auxiliary information includes an "alias" and some trust
16789 settings. The trust settings will ultimately be used in enhanced
16790 certificate chain verification routines: currently a certificate
16791 can only be trusted if it is self signed and then it is trusted
16796 * Fix assembler for Alpha (tested only on DEC OSF not Linux or `*BSD`).
16797 The problem was that one of the replacement routines had not been working
16798 since SSLeay releases. For now the offending routine has been replaced
16799 with non-optimised assembler. Even so, this now gives around 95%
16800 performance improvement for 1024 bit RSA signs.
16804 * Hack to fix PKCS#7 decryption when used with some unorthodox RC2
16805 handling. Most clients have the effective key size in bits equal to
16806 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
16807 A few however don't do this and instead use the size of the decrypted key
16808 to determine the RC2 key length and the AlgorithmIdentifier to determine
16809 the effective key length. In this case the effective key length can still
16810 be 40 bits but the key length can be 168 bits for example. This is fixed
16811 by manually forcing an RC2 key into the EVP_PKEY structure because the
16812 EVP code can't currently handle unusual RC2 key sizes: it always assumes
16813 the key length and effective key length are equal.
16817 * Add a bunch of functions that should simplify the creation of
16818 X509_NAME structures. Now you should be able to do:
16819 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
16820 and have it automatically work out the correct field type and fill in
16821 the structures. The more adventurous can try:
16822 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
16823 and it will (hopefully) work out the correct multibyte encoding.
16827 * Change the 'req' utility to use the new field handling and multibyte
16828 copy routines. Before the DN field creation was handled in an ad hoc
16829 way in req, ca, and x509 which was rather broken and didn't support
16830 BMPStrings or UTF8Strings. Since some software doesn't implement
16831 BMPStrings or UTF8Strings yet, they can be enabled using the config file
16832 using the dirstring_type option. See the new comment in the default
16833 openssl.cnf for more info.
16837 * Make crypto/rand/md_rand.c more robust:
16838 - Assure unique random numbers after fork().
16839 - Make sure that concurrent threads access the global counter and
16840 md serializably so that we never lose entropy in them
16841 or use exactly the same state in multiple threads.
16842 Access to the large state is not always serializable because
16843 the additional locking could be a performance killer, and
16844 md should be large enough anyway.
16848 * New file `apps/app_rand.c` with commonly needed functionality
16849 for handling the random seed file.
16851 Use the random seed file in some applications that previously did not:
16853 dsaparam -genkey (which also ignored its '-rand' option),
16856 x509 (when signing).
16857 Except on systems with /dev/urandom, it is crucial to have a random
16858 seed file at least for key creation, DSA signing, and for DH exchanges;
16859 for RSA signatures we could do without one.
16861 gendh and gendsa (unlike genrsa) used to read only the first byte
16862 of each file listed in the '-rand' option. The function as previously
16863 found in genrsa is now in app_rand.c and is used by all programs
16864 that support '-rand'.
16868 * In RAND_write_file, use mode 0600 for creating files;
16869 don't just chmod when it may be too late.
16873 * Report an error from X509_STORE_load_locations
16874 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
16878 * New function ASN1_mbstring_copy() this copies a string in either
16879 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
16880 into an ASN1_STRING type. A mask of permissible types is passed
16881 and it chooses the "minimal" type to use or an error if not type
16886 * Add function equivalents to the various macros in asn1.h. The old
16887 macros are retained with an `M_` prefix. Code inside the library can
16888 use the `M_` macros. External code (including the openssl utility)
16889 should *NOT* in order to be "shared library friendly".
16893 * Add various functions that can check a certificate's extensions
16894 to see if it usable for various purposes such as SSL client,
16895 server or S/MIME and CAs of these types. This is currently
16896 VERY EXPERIMENTAL but will ultimately be used for certificate chain
16897 verification. Also added a -purpose flag to x509 utility to
16898 print out all the purposes.
16902 * Add a CRYPTO_EX_DATA to X509 certificate structure and associated
16907 * New `X509V3_{X509,CRL,REVOKED}_get_d2i()` functions. These will search
16908 for, obtain and decode and extension and obtain its critical flag.
16909 This allows all the necessary extension code to be handled in a
16910 single function call.
16914 * RC4 tune-up featuring 30-40% performance improvement on most RISC
16915 platforms. See crypto/rc4/rc4_enc.c for further details.
16919 * New -noout option to asn1parse. This causes no output to be produced
16920 its main use is when combined with -strparse and -out to extract data
16921 from a file (which may not be in ASN.1 format).
16925 * Fix for pkcs12 program. It was hashing an invalid certificate pointer
16926 when producing the local key id.
16928 *Richard Levitte <levitte@stacken.kth.se>*
16930 * New option -dhparam in s_server. This allows a DH parameter file to be
16931 stated explicitly. If it is not stated then it tries the first server
16932 certificate file. The previous behaviour hard coded the filename
16937 * Add -pubin and -pubout options to the rsa and dsa commands. These allow
16938 a public key to be input or output. For example:
16939 openssl rsa -in key.pem -pubout -out pubkey.pem
16940 Also added necessary DSA public key functions to handle this.
16944 * Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
16945 in the message. This was handled by allowing
16946 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
16948 *Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>*
16950 * Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
16951 to the end of the strings whereas this didn't. This would cause problems
16952 if strings read with d2i_ASN1_bytes() were later modified.
16954 *Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>*
16956 * Fix for base64 decode bug. When a base64 bio reads only one line of
16957 data and it contains EOF it will end up returning an error. This is
16958 caused by input 46 bytes long. The cause is due to the way base64
16959 BIOs find the start of base64 encoded data. They do this by trying a
16960 trial decode on each line until they find one that works. When they
16961 do a flag is set and it starts again knowing it can pass all the
16962 data directly through the decoder. Unfortunately it doesn't reset
16963 the context it uses. This means that if EOF is reached an attempt
16964 is made to pass two EOFs through the context and this causes the
16965 resulting error. This can also cause other problems as well. As is
16966 usual with these problems it takes *ages* to find and the fix is
16967 trivial: move one line.
16969 *Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer)*
16971 * Ugly workaround to get s_client and s_server working under Windows. The
16972 old code wouldn't work because it needed to select() on sockets and the
16973 tty (for keypresses and to see if data could be written). Win32 only
16974 supports select() on sockets so we select() with a 1s timeout on the
16975 sockets and then see if any characters are waiting to be read, if none
16976 are present then we retry, we also assume we can always write data to
16977 the tty. This isn't nice because the code then blocks until we've
16978 received a complete line of data and it is effectively polling the
16979 keyboard at 1s intervals: however it's quite a bit better than not
16980 working at all :-) A dedicated Windows application might handle this
16981 with an event loop for example.
16985 * Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
16986 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
16987 will be called when RSA_sign() and RSA_verify() are used. This is useful
16988 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
16989 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
16990 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
16991 This necessitated the support of an extra signature type NID_md5_sha1
16992 for SSL signatures and modifications to the SSL library to use it instead
16993 of calling RSA_public_decrypt() and RSA_private_encrypt().
16997 * Add new -verify -CAfile and -CApath options to the crl program, these
16998 will lookup a CRL issuers certificate and verify the signature in a
16999 similar way to the verify program. Tidy up the crl program so it
17000 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
17001 less strict. It will now permit CRL extensions even if it is not
17002 a V2 CRL: this will allow it to tolerate some broken CRLs.
17006 * Initialize all non-automatic variables each time one of the openssl
17007 sub-programs is started (this is necessary as they may be started
17008 multiple times from the "OpenSSL>" prompt).
17010 *Lennart Bang, Bodo Moeller*
17012 * Preliminary compilation option RSA_NULL which disables RSA crypto without
17013 removing all other RSA functionality (this is what NO_RSA does). This
17014 is so (for example) those in the US can disable those operations covered
17015 by the RSA patent while allowing storage and parsing of RSA keys and RSA
17020 * Non-copying interface to BIO pairs.
17021 (still largely untested)
17025 * New function ASN1_tag2str() to convert an ASN1 tag to a descriptive
17026 ASCII string. This was handled independently in various places before.
17030 * New functions UTF8_getc() and UTF8_putc() that parse and generate
17031 UTF8 strings a character at a time.
17035 * Use client_version from client hello to select the protocol
17036 (s23_srvr.c) and for RSA client key exchange verification
17037 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
17041 * Add various utility functions to handle SPKACs, these were previously
17042 handled by poking round in the structure internals. Added new function
17043 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
17044 print, verify and generate SPKACs. Based on an original idea from
17045 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
17049 * RIPEMD160 is operational on all platforms and is back in 'make test'.
17053 * Allow the config file extension section to be overwritten on the
17054 command line. Based on an original idea from Massimiliano Pala
17055 <madwolf@comune.modena.it>. The new option is called -extensions
17056 and can be applied to ca, req and x509. Also -reqexts to override
17057 the request extensions in req and -crlexts to override the crl extensions
17062 * Add new feature to the SPKAC handling in ca. Now you can include
17063 the same field multiple times by preceding it by "XXXX." for example:
17066 this is the same syntax as used in the req config file.
17070 * Allow certificate extensions to be added to certificate requests. These
17071 are specified in a 'req_extensions' option of the req section of the
17072 config file. They can be printed out with the -text option to req but
17073 are otherwise ignored at present.
17077 * Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
17078 data read consists of only the final block it would not decrypted because
17079 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
17080 A misplaced 'break' also meant the decrypted final block might not be
17081 copied until the next read.
17085 * Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
17086 a few extra parameters to the DH structure: these will be useful if
17087 for example we want the value of 'q' or implement X9.42 DH.
17091 * Initial support for DSA_METHOD. This is based on the RSA_METHOD and
17092 provides hooks that allow the default DSA functions or functions on a
17093 "per key" basis to be replaced. This allows hardware acceleration and
17094 hardware key storage to be handled without major modification to the
17095 library. Also added low-level modexp hooks and CRYPTO_EX structure and
17096 associated functions.
17100 * Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
17101 as "read only": it can't be written to and the buffer it points to will
17102 not be freed. Reading from a read only BIO is much more efficient than
17103 a normal memory BIO. This was added because there are several times when
17104 an area of memory needs to be read from a BIO. The previous method was
17105 to create a memory BIO and write the data to it, this results in two
17106 copies of the data and an O(n^2) reading algorithm. There is a new
17107 function BIO_new_mem_buf() which creates a read only memory BIO from
17108 an area of memory. Also modified the PKCS#7 routines to use read only
17113 * Bugfix: ssl23_get_client_hello did not work properly when called in
17114 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
17115 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
17116 but a retry condition occurred while trying to read the rest.
17120 * The PKCS7_ENC_CONTENT_new() function was setting the content type as
17121 NID_pkcs7_encrypted by default: this was wrong since this should almost
17122 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
17123 the encrypted data type: this is a more sensible place to put it and it
17124 allows the PKCS#12 code to be tidied up that duplicated this
17129 * Changed obj_dat.pl script so it takes its input and output files on
17130 the command line. This should avoid shell escape redirection problems
17135 * Initial support for certificate extension requests, these are included
17136 in things like Xenroll certificate requests. Included functions to allow
17137 extensions to be obtained and added.
17141 * -crlf option to s_client and s_server for sending newlines as
17142 CRLF (as required by many protocols).
17146 ### Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
17148 * Install libRSAglue.a when OpenSSL is built with RSAref.
17150 *Ralf S. Engelschall*
17152 * A few more `#ifndef NO_FP_API / #endif` pairs for consistency.
17154 *Andrija Antonijevic <TheAntony2@bigfoot.com>*
17156 * Fix -startdate and -enddate (which was missing) arguments to 'ca'
17161 * New function DSA_dup_DH, which duplicates DSA parameters/keys as
17162 DH parameters/keys (q is lost during that conversion, but the resulting
17163 DH parameters contain its length).
17165 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
17166 much faster than DH_generate_parameters (which creates parameters
17167 where `p = 2*q + 1`), and also the smaller q makes DH computations
17168 much more efficient (160-bit exponentiation instead of 1024-bit
17169 exponentiation); so this provides a convenient way to support DHE
17170 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
17171 utter importance to use
17172 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
17174 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
17175 when such DH parameters are used, because otherwise small subgroup
17176 attacks may become possible!
17180 * Avoid memory leak in i2d_DHparams.
17184 * Allow the -k option to be used more than once in the enc program:
17185 this allows the same encrypted message to be read by multiple recipients.
17189 * New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
17190 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
17191 it will always use the numerical form of the OID, even if it has a short
17196 * Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
17197 method only got called if p,q,dmp1,dmq1,iqmp components were present,
17198 otherwise bn_mod_exp was called. In the case of hardware keys for example
17199 no private key components need be present and it might store extra data
17200 in the RSA structure, which cannot be accessed from bn_mod_exp.
17201 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
17202 private key operations.
17206 * Added support for SPARC Linux.
17210 * pem_password_cb function type incompatibly changed from
17211 typedef int pem_password_cb(char *buf, int size, int rwflag);
17213 ....(char *buf, int size, int rwflag, void *userdata);
17214 so that applications can pass data to their callbacks:
17215 The `PEM[_ASN1]_{read,write}...` functions and macros now take an
17216 additional void * argument, which is just handed through whenever
17217 the password callback is called.
17219 *Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller*
17221 New function SSL_CTX_set_default_passwd_cb_userdata.
17223 Compatibility note: As many C implementations push function arguments
17224 onto the stack in reverse order, the new library version is likely to
17225 interoperate with programs that have been compiled with the old
17226 pem_password_cb definition (PEM_whatever takes some data that
17227 happens to be on the stack as its last argument, and the callback
17228 just ignores this garbage); but there is no guarantee whatsoever that
17231 * The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
17232 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
17233 problems not only on Windows, but also on some Unix platforms.
17234 To avoid problematic command lines, these definitions are now in an
17235 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
17236 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
17240 * MIPS III/IV assembler module is reimplemented.
17244 * More DES library cleanups: remove references to srand/rand and
17245 delete an unused file.
17249 * Add support for the free Netwide assembler (NASM) under Win32,
17250 since not many people have MASM (ml) and it can be hard to obtain.
17251 This is currently experimental but it seems to work OK and pass all
17252 the tests. Check out INSTALL.W32 for info.
17256 * Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
17257 without temporary keys kept an extra copy of the server key,
17258 and connections with temporary keys did not free everything in case
17263 * New function RSA_check_key and new openssl rsa option -check
17264 for verifying the consistency of RSA keys.
17266 *Ulf Moeller, Bodo Moeller*
17268 * Various changes to make Win32 compile work:
17269 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
17270 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
17271 comparison" warnings.
17272 3. Add `sk_<TYPE>_sort` to DEF file generator and do make update.
17276 * Add a debugging option to PKCS#5 v2 key generation function: when
17277 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
17278 derived keys are printed to stderr.
17282 * Copy the flags in ASN1_STRING_dup().
17284 *Roman E. Pavlov <pre@mo.msk.ru>*
17286 * The x509 application mishandled signing requests containing DSA
17287 keys when the signing key was also DSA and the parameters didn't match.
17289 It was supposed to omit the parameters when they matched the signing key:
17290 the verifying software was then supposed to automatically use the CA's
17291 parameters if they were absent from the end user certificate.
17293 Omitting parameters is no longer recommended. The test was also
17294 the wrong way round! This was probably due to unusual behaviour in
17295 EVP_cmp_parameters() which returns 1 if the parameters match.
17296 This meant that parameters were omitted when they *didn't* match and
17297 the certificate was useless. Certificates signed with 'ca' didn't have
17300 *Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>*
17302 * Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
17303 The interface is as follows:
17304 Applications can use
17305 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
17306 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
17307 "off" is now the default.
17308 The library internally uses
17309 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
17310 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
17311 to disable memory-checking temporarily.
17313 Some inconsistent states that previously were possible (and were
17314 even the default) are now avoided.
17316 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
17317 with each memory chunk allocated; this is occasionally more helpful
17318 than just having a counter.
17320 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
17322 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
17327 * Introduce "mode" for SSL structures (with defaults in SSL_CTX),
17328 which largely parallels "options", but is for changing API behaviour,
17329 whereas "options" are about protocol behaviour.
17330 Initial "mode" flags are:
17332 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
17333 a single record has been written.
17334 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
17335 retries use the same buffer location.
17336 (But all of the contents must be
17341 * Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
17344 * Fix problems with no-hmac etc.
17346 *Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>*
17348 * New functions RSA_get_default_method(), RSA_set_method() and
17349 RSA_get_method(). These allows replacement of RSA_METHODs without having
17350 to mess around with the internals of an RSA structure.
17354 * Fix memory leaks in DSA_do_sign and DSA_is_prime.
17355 Also really enable memory leak checks in openssl.c and in some
17358 *Chad C. Mulligan, Bodo Moeller*
17360 * Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
17361 up the length of negative integers. This has now been simplified to just
17362 store the length when it is first determined and use it later, rather
17363 than trying to keep track of where data is copied and updating it to
17365 *Steve Henson, reported by Brien Wheeler <bwheeler@authentica-security.com>*
17367 * Add a new function PKCS7_signatureVerify. This allows the verification
17368 of a PKCS#7 signature but with the signing certificate passed to the
17369 function itself. This contrasts with PKCS7_dataVerify which assumes the
17370 certificate is present in the PKCS#7 structure. This isn't always the
17371 case: certificates can be omitted from a PKCS#7 structure and be
17372 distributed by "out of band" means (such as a certificate database).
17376 * Complete the `PEM_*` macros with DECLARE_PEM versions to replace the
17377 function prototypes in pem.h, also change util/mkdef.pl to add the
17378 necessary function names.
17382 * mk1mf.pl (used by Windows builds) did not properly read the
17383 options set by Configure in the top level Makefile, and Configure
17384 was not even able to write more than one option correctly.
17385 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
17389 * New functions CONF_load_bio() and CONF_load_fp() to allow a config
17390 file to be loaded from a BIO or FILE pointer. The BIO version will
17391 for example allow memory BIOs to contain config info.
17395 * New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
17396 Whoever hopes to achieve shared-library compatibility across versions
17397 must use this, not the compile-time macro.
17398 (Exercise 0.9.4: Which is the minimum library version required by
17400 Note: All this applies only to multi-threaded programs, others don't
17405 * Add missing case to s3_clnt.c state machine -- one of the new SSL tests
17406 through a BIO pair triggered the default case, i.e.
17407 SSLerr(...,SSL_R_UNKNOWN_STATE).
17411 * New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
17412 can use the SSL library even if none of the specific BIOs is
17417 * Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
17418 for the encoded length.
17420 *Jeon KyoungHo <khjeon@sds.samsung.co.kr>*
17422 * Add initial documentation of the X509V3 functions.
17426 * Add a new pair of functions PEM_write_PKCS8PrivateKey() and
17427 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
17428 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
17429 secure PKCS#8 private key format with a high iteration count.
17433 * Fix determination of Perl interpreter: A perl or perl5
17434 *directory* in $PATH was also accepted as the interpreter.
17436 *Ralf S. Engelschall*
17438 * Fix demos/sign/sign.c: well there wasn't anything strictly speaking
17439 wrong with it but it was very old and did things like calling
17440 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
17441 unusual formatting.
17445 * Fix demos/selfsign.c: it used obsolete and deleted functions, changed
17446 to use the new extension code.
17450 * Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
17451 with macros. This should make it easier to change their form, add extra
17452 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
17457 * Add to configuration table a new entry that can specify an alternative
17458 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
17459 according to Mark Crispin <MRC@Panda.COM>.
17463 * DES CBC did not update the IV. Weird.
17467 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
17468 Changing the behaviour of the former might break existing programs --
17469 where IV updating is needed, des_ncbc_encrypt can be used.
17472 * When bntest is run from "make test" it drives bc to check its
17473 calculations, as well as internally checking them. If an internal check
17474 fails, it needs to cause bc to give a non-zero result or make test carries
17475 on without noticing the failure. Fixed.
17479 * DES library cleanups.
17483 * Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
17484 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
17485 ciphers. NOTE: although the key derivation function has been verified
17486 against some published test vectors it has not been extensively tested
17487 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
17492 * Instead of "mkdir -p", which is not fully portable, use new
17493 Perl script "util/mkdir-p.pl".
17497 * Rewrite the way password based encryption (PBE) is handled. It used to
17498 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
17499 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
17500 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
17501 the 'parameter' field of the AlgorithmIdentifier is passed to the
17502 underlying key generation function so it must do its own ASN1 parsing.
17503 This has also changed the EVP_PBE_CipherInit() function which now has a
17504 'parameter' argument instead of literal salt and iteration count values
17505 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
17509 * Support for PKCS#5 v1.5 compatible password based encryption algorithms
17510 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
17511 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
17512 KEY" because this clashed with PKCS#8 unencrypted string. Since this
17513 value was just used as a "magic string" and not used directly its
17514 value doesn't matter.
17518 * Introduce some semblance of const correctness to BN. Shame C doesn't
17523 * "linux-sparc64" configuration (ultrapenguin).
17525 *Ray Miller <ray.miller@oucs.ox.ac.uk>*
17526 "linux-sparc" configuration.
17528 *Christian Forster <fo@hawo.stw.uni-erlangen.de>*
17530 * config now generates no-xxx options for missing ciphers.
17534 * Support the EBCDIC character set (work in progress).
17535 File ebcdic.c not yet included because it has a different license.
17537 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
17539 * Support BS2000/OSD-POSIX.
17541 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
17543 * Make callbacks for key generation use `void *` instead of `char *`.
17547 * Make S/MIME samples compile (not yet tested).
17551 * Additional typesafe stacks.
17555 * New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
17559 ### Changes between 0.9.3 and 0.9.3a [29 May 1999]
17561 * New configuration variant "sco5-gcc".
17563 * Updated some demos.
17565 *Sean O Riordain, Wade Scholine*
17567 * Add missing BIO_free at exit of pkcs12 application.
17571 * Fix memory leak in conf.c.
17575 * Updates for Win32 to assembler version of MD5.
17579 * Set #! path to perl in `apps/der_chop` to where we found it
17580 instead of using a fixed path.
17584 * SHA library changes for irix64-mips4-cc.
17588 * Improvements for VMS support.
17592 ### Changes between 0.9.2b and 0.9.3 [24 May 1999]
17594 * Bignum library bug fix. IRIX 6 passes "make test" now!
17595 This also avoids the problems with SC4.2 and unpatched SC5.
17597 *Andy Polyakov <appro@fy.chalmers.se>*
17599 * New functions sk_num, sk_value and sk_set to replace the previous macros.
17600 These are required because of the typesafe stack would otherwise break
17601 existing code. If old code used a structure member which used to be STACK
17602 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
17603 sk_num or sk_value it would produce an error because the num, data members
17604 are not present in STACK_OF. Now it just produces a warning. sk_set
17605 replaces the old method of assigning a value to sk_value
17606 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
17607 that does this will no longer work (and should use sk_set instead) but
17608 this could be regarded as a "questionable" behaviour anyway.
17612 * Fix most of the other PKCS#7 bugs. The "experimental" code can now
17613 correctly handle encrypted S/MIME data.
17617 * Change type of various DES function arguments from des_cblock
17618 (which means, in function argument declarations, pointer to char)
17619 to des_cblock * (meaning pointer to array with 8 char elements),
17620 which allows the compiler to do more typechecking; it was like
17621 that back in SSLeay, but with lots of ugly casts.
17623 Introduce new type const_des_cblock.
17627 * Reorganise the PKCS#7 library and get rid of some of the more obvious
17628 problems: find RecipientInfo structure that matches recipient certificate
17629 and initialise the ASN1 structures properly based on passed cipher.
17633 * Belatedly make the BN tests actually check the results.
17637 * Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
17638 to and from BNs: it was completely broken. New compilation option
17639 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
17640 key elements as negative integers.
17644 * Reorganize and speed up MD5.
17646 *Andy Polyakov <appro@fy.chalmers.se>*
17650 *Richard Levitte <richard@levitte.org>*
17652 * New option -out to asn1parse to allow the parsed structure to be
17653 output to a file. This is most useful when combined with the -strparse
17654 option to examine the output of things like OCTET STRINGS.
17658 * Make SSL library a little more fool-proof by not requiring any longer
17659 that `SSL_set_{accept,connect}_state` be called before
17660 `SSL_{accept,connect}` may be used (`SSL_set_..._state` is omitted
17661 in many applications because usually everything *appeared* to work as
17662 intended anyway -- now it really works as intended).
17666 * Move openssl.cnf out of lib/.
17670 * Fix various things to let OpenSSL even pass "egcc -pipe -O2 -Wall
17671 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
17672 -Wmissing-declarations -Wnested-externs -Winline" with EGCS 1.1.2+
17674 *Ralf S. Engelschall*
17676 * Various fixes to the EVP and PKCS#7 code. It may now be able to
17677 handle PKCS#7 enveloped data properly.
17679 *Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve*
17681 * Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
17682 copying pointers. The cert_st handling is changed by this in
17683 various ways (and thus what used to be known as ctx->default_cert
17684 is now called ctx->cert, since we don't resort to `s->ctx->[default_]cert`
17685 any longer when s->cert does not give us what we need).
17686 ssl_cert_instantiate becomes obsolete by this change.
17687 As soon as we've got the new code right (possibly it already is?),
17688 we have solved a couple of bugs of the earlier code where s->cert
17689 was used as if it could not have been shared with other SSL structures.
17691 Note that using the SSL API in certain dirty ways now will result
17692 in different behaviour than observed with earlier library versions:
17693 Changing settings for an `SSL_CTX *ctx` after having done s = SSL_new(ctx)
17694 does not influence s as it used to.
17696 In order to clean up things more thoroughly, inside SSL_SESSION
17697 we don't use CERT any longer, but a new structure SESS_CERT
17698 that holds per-session data (if available); currently, this is
17699 the peer's certificate chain and, for clients, the server's certificate
17700 and temporary key. CERT holds only those values that can have
17701 meaningful defaults in an SSL_CTX.
17705 * New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
17706 from the internal representation. Various PKCS#7 fixes: remove some
17707 evil casts and set the enc_dig_alg field properly based on the signing
17712 * Allow PKCS#12 password to be set from the command line or the
17713 environment. Let 'ca' get its config file name from the environment
17714 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
17719 * Allow certificate policies extension to use an IA5STRING for the
17720 organization field. This is contrary to the PKIX definition but
17721 VeriSign uses it and IE5 only recognises this form. Document 'x509'
17726 * Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
17727 without disallowing inline assembler and the like for non-pedantic builds.
17731 * Support Borland C++ builder.
17733 *Janez Jere <jj@void.si>, modified by Ulf Möller*
17739 * SHA-1 cleanups and performance enhancements.
17741 *Andy Polyakov <appro@fy.chalmers.se>*
17743 * Sparc v8plus assembler for the bignum library.
17745 *Andy Polyakov <appro@fy.chalmers.se>*
17747 * Accept any -xxx and +xxx compiler options in Configure.
17751 * Update HPUX configuration.
17755 * Add missing `sk_<type>_unshift()` function to safestack.h
17757 *Ralf S. Engelschall*
17759 * New function SSL_CTX_use_certificate_chain_file that sets the
17760 "extra_cert"s in addition to the certificate. (This makes sense
17761 only for "PEM" format files, as chains as a whole are not
17766 * Support verify_depth from the SSL API.
17767 x509_vfy.c had what can be considered an off-by-one-error:
17768 Its depth (which was not part of the external interface)
17769 was actually counting the number of certificates in a chain;
17770 now it really counts the depth.
17774 * Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
17775 instead of X509err, which often resulted in confusing error
17776 messages since the error codes are not globally unique
17777 (e.g. an alleged error in ssl3_accept when a certificate
17778 didn't match the private key).
17780 * New function SSL_CTX_set_session_id_context that allows to set a default
17781 value (so that you don't need SSL_set_session_id_context for each
17782 connection using the SSL_CTX).
17786 * OAEP decoding bug fix.
17790 * Support INSTALL_PREFIX for package builders, as proposed by
17795 * New Configure options "threads" and "no-threads". For systems
17796 where the proper compiler options are known (currently Solaris
17797 and Linux), "threads" is the default.
17801 * New script util/mklink.pl as a faster substitute for util/mklink.sh.
17805 * Install various scripts to $(OPENSSLDIR)/misc, not to
17806 $(INSTALLTOP)/bin -- they shouldn't clutter directories
17807 such as /usr/local/bin.
17811 * "make linux-shared" to build shared libraries.
17813 *Niels Poppe <niels@netbox.org>*
17815 * New Configure option `no-<cipher>` (rsa, idea, rc5, ...).
17819 * Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
17820 extension adding in x509 utility.
17824 * Remove NOPROTO sections and error code comments.
17828 * Partial rewrite of the DEF file generator to now parse the ANSI
17833 * New Configure options --prefix=DIR and --openssldir=DIR.
17837 * Complete rewrite of the error code script(s). It is all now handled
17838 by one script at the top level which handles error code gathering,
17839 header rewriting and C source file generation. It should be much better
17840 than the old method: it now uses a modified version of Ulf's parser to
17841 read the ANSI prototypes in all header files (thus the old K&R definitions
17842 aren't needed for error creation any more) and do a better job of
17843 translating function codes into names. The old 'ASN1 error code embedded
17844 in a comment' is no longer necessary and it doesn't use .err files which
17845 have now been deleted. Also the error code call doesn't have to appear all
17846 on one line (which resulted in some large lines...).
17850 * Change #include filenames from `<foo.h>` to `<openssl/foo.h>`.
17854 * Change behaviour of ssl2_read when facing length-0 packets: Don't return
17855 0 (which usually indicates a closed connection), but continue reading.
17859 * Fix some race conditions.
17863 * Add support for CRL distribution points extension. Add Certificate
17864 Policies and CRL distribution points documentation.
17868 * Move the autogenerated header file parts to crypto/opensslconf.h.
17872 * Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
17873 8 of keying material. Merlin has also confirmed interop with this fix
17874 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
17876 *Merlin Hughes <merlin@baltimore.ie>*
17878 * Fix lots of warnings.
17880 *Richard Levitte <levitte@stacken.kth.se>*
17882 * In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
17883 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
17885 *Richard Levitte <levitte@stacken.kth.se>*
17887 * Fix problems with sizeof(long) == 8.
17889 *Andy Polyakov <appro@fy.chalmers.se>*
17891 * Change functions to ANSI C.
17895 * Fix typos in error codes.
17897 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller*
17899 * Remove defunct assembler files from Configure.
17903 * SPARC v8 assembler BIGNUM implementation.
17905 *Andy Polyakov <appro@fy.chalmers.se>*
17907 * Support for Certificate Policies extension: both print and set.
17908 Various additions to support the r2i method this uses.
17912 * A lot of constification, and fix a bug in X509_NAME_oneline() that could
17913 return a const string when you are expecting an allocated buffer.
17917 * Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
17918 types DirectoryString and DisplayText.
17922 * Add code to allow r2i extensions to access the configuration database,
17923 add an LHASH database driver and add several ctx helper functions.
17927 * Fix an evil bug in bn_expand2() which caused various BN functions to
17928 fail when they extended the size of a BIGNUM.
17932 * Various utility functions to handle SXNet extension. Modify mkdef.pl to
17933 support typesafe stack.
17937 * Fix typo in SSL_[gs]et_options().
17939 *Nils Frostberg <nils@medcom.se>*
17941 * Delete various functions and files that belonged to the (now obsolete)
17942 old X509V3 handling code.
17946 * New Configure option "rsaref".
17950 * Don't auto-generate pem.h.
17954 * Introduce type-safe ASN.1 SETs.
17958 * Convert various additional casted stacks to type-safe STACK_OF() variants.
17960 *Ben Laurie, Ralf S. Engelschall, Steve Henson*
17962 * Introduce type-safe STACKs. This will almost certainly break lots of code
17963 that links with OpenSSL (well at least cause lots of warnings), but fear
17964 not: the conversion is trivial, and it eliminates loads of evil casts. A
17965 few STACKed things have been converted already. Feel free to convert more.
17966 In the fullness of time, I'll do away with the STACK type altogether.
17970 * Add `openssl ca -revoke <certfile>` facility which revokes a certificate
17971 specified in `<certfile>` by updating the entry in the index.txt file.
17972 This way one no longer has to edit the index.txt file manually for
17973 revoking a certificate. The -revoke option does the gory details now.
17975 *Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall*
17977 * Fix `openssl crl -noout -text` combination where `-noout` killed the
17978 `-text` option at all and this way the `-noout -text` combination was
17979 inconsistent in `openssl crl` with the friends in `openssl x509|rsa|dsa`.
17981 *Ralf S. Engelschall*
17983 * Make sure a corresponding plain text error message exists for the
17984 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
17985 verify callback function determined that a certificate was revoked.
17987 *Ralf S. Engelschall*
17989 * Bugfix: In test/testenc, don't test `openssl <cipher>` for
17990 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
17991 all available ciphers including rc5, which was forgotten until now.
17992 In order to let the testing shell script know which algorithms
17993 are available, a new (up to now undocumented) command
17994 `openssl list-cipher-commands` is used.
17998 * Bugfix: s_client occasionally would sleep in select() when
17999 it should have checked SSL_pending() first.
18003 * New functions DSA_do_sign and DSA_do_verify to provide access to
18004 the raw DSA values prior to ASN.1 encoding.
18008 * Tweaks to Configure
18010 *Niels Poppe <niels@netbox.org>*
18012 * Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
18017 * New variables $(RANLIB) and $(PERL) in the Makefiles.
18021 * New config option to avoid instructions that are illegal on the 80386.
18022 The default code is faster, but requires at least a 486.
18026 * Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
18027 SSL2_SERVER_VERSION (not used at all) macros, which are now the
18028 same as SSL2_VERSION anyway.
18032 * New "-showcerts" option for s_client.
18036 * Still more PKCS#12 integration. Add pkcs12 application to openssl
18037 application. Various cleanups and fixes.
18041 * More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
18042 modify error routines to work internally. Add error codes and PBE init
18043 to library startup routines.
18047 * Further PKCS#12 integration. Added password based encryption, PKCS#8 and
18048 packing functions to asn1 and evp. Changed function names and error
18049 codes along the way.
18053 * PKCS12 integration: and so it begins... First of several patches to
18054 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
18055 objects to objects.h
18059 * Add a new 'indent' option to some X509V3 extension code. Initial ASN1
18060 and display support for Thawte strong extranet extension.
18064 * Add LinuxPPC support.
18066 *Jeff Dubrule <igor@pobox.org>*
18068 * Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
18069 bn_div_words in alpha.s.
18071 *Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie*
18073 * Make sure the RSA OAEP test is skipped under -DRSAref because
18074 OAEP isn't supported when OpenSSL is built with RSAref.
18076 *Ulf Moeller <ulf@fitug.de>*
18078 * Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
18079 so they no longer are missing under -DNOPROTO.
18081 *Soren S. Jorvang <soren@t.dk>*
18083 ### Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
18085 * Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
18086 doesn't work when the session is reused. Coming soon!
18090 * Fix a security hole, that allows sessions to be reused in the wrong
18091 context thus bypassing client cert protection! All software that uses
18092 client certs and session caches in multiple contexts NEEDS PATCHING to
18093 allow session reuse! A fuller solution is in the works.
18095 *Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)*
18097 * Some more source tree cleanups (removed obsolete files
18098 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
18099 permission on "config" script to be executable) and a fix for the INSTALL
18102 *Ulf Moeller <ulf@fitug.de>*
18104 * Remove some legacy and erroneous uses of malloc, free instead of
18107 *Lennart Bang <lob@netstream.se>, with minor changes by Steve*
18109 * Make rsa_oaep_test return non-zero on error.
18111 *Ulf Moeller <ulf@fitug.de>*
18113 * Add support for native Solaris shared libraries. Configure
18114 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
18115 if someone would make that last step automatic.
18117 *Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>*
18119 * ctx_size was not built with the right compiler during "make links". Fixed.
18123 * Change the meaning of 'ALL' in the cipher list. It now means "everything
18124 except NULL ciphers". This means the default cipher list will no longer
18125 enable NULL ciphers. They need to be specifically enabled e.g. with
18126 the string "DEFAULT:eNULL".
18130 * Fix to RSA private encryption routines: if p < q then it would
18131 occasionally produce an invalid result. This will only happen with
18132 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
18136 * Be less restrictive and allow also `perl util/perlpath.pl
18137 /path/to/bin/perl` in addition to `perl util/perlpath.pl /path/to/bin`,
18138 because this way one can also use an interpreter named `perl5` (which is
18139 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
18140 installed as `perl`).
18142 *Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
18144 * Let util/clean-depend.pl work also with older Perl 5.00x versions.
18146 *Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
18148 * Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
18149 advapi32.lib to Win32 build and change the pem test comparison
18150 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
18151 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
18152 and crypto/des/ede_cbcm_enc.c.
18156 * DES quad checksum was broken on big-endian architectures. Fixed.
18160 * Comment out two functions in bio.h that aren't implemented. Fix up the
18161 Win32 test batch file so it (might) work again. The Win32 test batch file
18162 is horrible: I feel ill....
18166 * Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
18167 in e_os.h. Audit of header files to check ANSI and non ANSI
18168 sections: 10 functions were absent from non ANSI section and not exported
18169 from Windows DLLs. Fixed up libeay.num for new functions.
18173 * Make `openssl version` output lines consistent.
18175 *Ralf S. Engelschall*
18177 * Fix Win32 symbol export lists for BIO functions: Added
18178 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
18179 to ms/libeay{16,32}.def.
18181 *Ralf S. Engelschall*
18183 * Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
18184 fine under Unix and passes some trivial tests I've now added. But the
18185 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
18186 added to make sure no one expects that this stuff really works in the
18187 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
18188 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
18191 *Ralf S. Engelschall*
18193 * Fix the generation of two part addresses in perl.
18195 *Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie*
18197 * Add config entry for Linux on MIPS.
18199 *John Tobey <jtobey@channel1.com>*
18201 * Make links whenever Configure is run, unless we are on Windoze.
18205 * Permit extensions to be added to CRLs using crl_section in openssl.cnf.
18206 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
18211 * Add a useful kludge to allow package maintainers to specify compiler and
18212 other platforms details on the command line without having to patch the
18213 Configure script every time: One now can use
18214 `perl Configure <id>:<details>`,
18215 i.e. platform ids are allowed to have details appended
18216 to them (separated by colons). This is treated as there would be a static
18217 pre-configured entry in Configure's %table under key `<id>` with value
18218 `<details>` and `perl Configure <id>` is called. So, when you want to
18219 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
18220 assembler stuff you can use `perl Configure "FreeBSD-elf:pgcc:-O6:::"`
18221 now, which overrides the FreeBSD-elf entry on-the-fly.
18223 *Ralf S. Engelschall*
18225 * Disable new TLS1 ciphersuites by default: they aren't official yet.
18229 * Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
18230 on the `perl Configure ...` command line. This way one can compile
18231 OpenSSL libraries with Position Independent Code (PIC) which is needed
18232 for linking it into DSOs.
18234 *Ralf S. Engelschall*
18236 * Remarkably, export ciphers were totally broken and no-one had noticed!
18241 * Cleaned up the LICENSE document: The official contact for any license
18242 questions now is the OpenSSL core team under openssl-core@openssl.org.
18243 And add a paragraph about the dual-license situation to make sure people
18244 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
18245 to the OpenSSL toolkit.
18247 *Ralf S. Engelschall*
18249 * General source tree makefile cleanups: Made `making xxx in yyy...`
18250 display consistent in the source tree and replaced `/bin/rm` by `rm`.
18251 Additionally cleaned up the `make links` target: Remove unnecessary
18252 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
18253 to speed processing and no longer clutter the display with confusing
18254 stuff. Instead only the actually done links are displayed.
18256 *Ralf S. Engelschall*
18258 * Permit null encryption ciphersuites, used for authentication only. It used
18259 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
18260 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
18265 * Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
18266 signed attributes when verifying signatures (this would break them),
18267 the detached data encoding was wrong and public keys obtained using
18268 X509_get_pubkey() weren't freed.
18272 * Add text documentation for the BUFFER functions. Also added a work around
18273 to a Win95 console bug. This was triggered by the password read stuff: the
18274 last character typed gets carried over to the next fread(). If you were
18275 generating a new cert request using 'req' for example then the last
18276 character of the passphrase would be CR which would then enter the first
18281 * Added the new 'Includes OpenSSL Cryptography Software' button as
18282 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
18283 button and can be used by applications based on OpenSSL to show the
18284 relationship to the OpenSSL project.
18286 *Ralf S. Engelschall*
18288 * Remove confusing variables in function signatures in files
18289 ssl/ssl_lib.c and ssl/ssl.h.
18291 *Lennart Bong <lob@kulthea.stacken.kth.se>*
18293 * Don't install bss_file.c under PREFIX/include/
18295 *Lennart Bong <lob@kulthea.stacken.kth.se>*
18297 * Get the Win32 compile working again. Modify mkdef.pl so it can handle
18298 functions that return function pointers and has support for NT specific
18299 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
18300 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
18301 unsigned to signed types: this was killing the Win32 compile.
18305 * Add new certificate file to stack functions,
18306 SSL_add_dir_cert_subjects_to_stack() and
18307 SSL_add_file_cert_subjects_to_stack(). These largely supplant
18308 SSL_load_client_CA_file(), and can be used to add multiple certs easily
18309 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
18310 This means that Apache-SSL and similar packages don't have to mess around
18311 to add as many CAs as they want to the preferred list.
18315 * Experiment with doxygen documentation. Currently only partially applied to
18317 See <http://www.stack.nl/~dimitri/doxygen/index.html>, and run doxygen with
18318 openssl.doxy as the configuration file.
18322 * Get rid of remaining C++-style comments which strict C compilers hate.
18324 *Ralf S. Engelschall, pointed out by Carlos Amengual*
18326 * Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
18327 compiled in by default: it has problems with large keys.
18331 * Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
18332 DH private keys and/or callback functions which directly correspond to
18333 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
18334 is needed for applications which have to configure certificates on a
18335 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
18337 For the RSA certificate situation is makes no difference, but
18338 for the DSA certificate situation this fixes the "no shared cipher"
18339 problem where the OpenSSL cipher selection procedure failed because the
18340 temporary keys were not overtaken from the context and the API provided
18341 no way to reconfigure them.
18342 The new functions now let applications reconfigure the stuff and they
18343 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
18344 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
18345 non-public-API function ssl_cert_instantiate() is used as a helper
18346 function and also to reduce code redundancy inside ssl_rsa.c.
18348 *Ralf S. Engelschall*
18350 * Move s_server -dcert and -dkey options out of the undocumented feature
18351 area because they are useful for the DSA situation and should be
18352 recognized by the users.
18354 *Ralf S. Engelschall*
18356 * Fix the cipher decision scheme for export ciphers: the export bits are
18357 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
18358 SSL_EXP_MASK. So, the original variable has to be used instead of the
18359 already masked variable.
18361 *Richard Levitte <levitte@stacken.kth.se>*
18363 * Fix `port` variable from `int` to `unsigned int` in crypto/bio/b_sock.c
18365 *Richard Levitte <levitte@stacken.kth.se>*
18367 * Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
18368 from `int` to `unsigned int` because it is a length and initialized by
18369 EVP_DigestFinal() which expects an `unsigned int *`.
18371 *Richard Levitte <levitte@stacken.kth.se>*
18373 * Don't hard-code path to Perl interpreter on shebang line of Configure
18374 script. Instead use the usual Shell->Perl transition trick.
18376 *Ralf S. Engelschall*
18378 * Make `openssl x509 -noout -modulus`' functional also for DSA certificates
18379 (in addition to RSA certificates) to match the behaviour of `openssl dsa
18380 -noout -modulus` as it's already the case for `openssl rsa -noout
18381 -modulus`. For RSA the -modulus is the real "modulus" while for DSA
18382 currently the public key is printed (a decision which was already done by
18383 `openssl dsa -modulus` in the past) which serves a similar purpose.
18384 Additionally the NO_RSA no longer completely removes the whole -modulus
18385 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
18388 *Ralf S. Engelschall*
18390 * Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
18391 BIO. See the source (crypto/evp/bio_ok.c) for more info.
18393 *Arne Ansper <arne@ats.cyber.ee>*
18395 * Dump the old yucky req code that tried (and failed) to allow raw OIDs
18396 to be added. Now both 'req' and 'ca' can use new objects defined in the
18401 * Add cool BIO that does syslog (or event log on NT).
18403 *Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie*
18405 * Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
18406 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
18407 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
18408 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
18412 * Add preliminary config info for new extension code.
18416 * Make RSA_NO_PADDING really use no padding.
18418 *Ulf Moeller <ulf@fitug.de>*
18420 * Generate errors when private/public key check is done.
18424 * Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
18425 for some CRL extensions and new objects added.
18429 * Really fix the ASN1 IMPLICIT bug this time... Partial support for private
18430 key usage extension and fuller support for authority key id.
18434 * Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
18435 padding method for RSA, which is recommended for new applications in PKCS
18436 #1 v2.0 (RFC 2437, October 1998).
18437 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
18438 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
18439 against Bleichbacher's attack on RSA.
18440 *Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
18443 * Updates to the new SSL compression code
18445 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
18447 * Fix so that the version number in the master secret, when passed
18448 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
18449 (because the server will not accept higher), that the version number
18450 is 0x03,0x01, not 0x03,0x00
18452 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
18454 * Run extensive memory leak checks on SSL commands. Fixed *lots* of memory
18455 leaks in `ssl/` relating to new `X509_get_pubkey()` behaviour. Also fixes
18456 in `apps/` and an unrelated leak in `crypto/dsa/dsa_vrf.c`.
18460 * Support for RAW extensions where an arbitrary extension can be
18461 created by including its DER encoding. See `apps/openssl.cnf` for
18466 * Make sure latest Perl versions don't interpret some generated C array
18467 code as Perl array code in the crypto/err/err_genc.pl script.
18469 *Lars Weber <3weber@informatik.uni-hamburg.de>*
18471 * Modify ms/do_ms.bat to not generate assembly language makefiles since
18472 not many people have the assembler. Various Win32 compilation fixes and
18473 update to the INSTALL.W32 file with (hopefully) more accurate Win32
18474 build instructions.
18478 * Modify configure script 'Configure' to automatically create crypto/date.h
18479 file under Win32 and also build pem.h from pem.org. New script
18480 util/mkfiles.pl to create the MINFO file on environments that can't do a
18481 'make files': perl util/mkfiles.pl >MINFO should work.
18485 * Major rework of DES function declarations, in the pursuit of correctness
18486 and purity. As a result, many evil casts evaporated, and some weirdness,
18487 too. You may find this causes warnings in your code. Zapping your evil
18488 casts will probably fix them. Mostly.
18492 * Fix for a typo in asn1.h. Bug fix to object creation script
18493 obj_dat.pl. It considered a zero in an object definition to mean
18494 "end of object": none of the objects in objects.h have any zeros
18495 so it wasn't spotted.
18497 *Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>*
18499 * Add support for Triple DES Cipher Block Chaining with Output Feedback
18500 Masking (CBCM). In the absence of test vectors, the best I have been able
18501 to do is check that the decrypt undoes the encrypt, so far. Send me test
18502 vectors if you have them.
18506 * Correct calculation of key length for export ciphers (too much space was
18507 allocated for null ciphers). This has not been tested!
18511 * Modifications to the mkdef.pl for Win32 DEF file creation. The usage
18512 message is now correct (it understands "crypto" and "ssl" on its
18513 command line). There is also now an "update" option. This will update
18514 the util/ssleay.num and util/libeay.num files with any new functions.
18516 perl util/mkdef.pl crypto ssl update
18517 it will update them.
18521 * Overhauled the Perl interface:
18522 - ported BN stuff to OpenSSL's different BN library
18523 - made the perl/ source tree CVS-aware
18524 - renamed the package from SSLeay to OpenSSL (the files still contain
18525 their history because I've copied them in the repository)
18526 - removed obsolete files (the test scripts will be replaced
18527 by better Test::Harness variants in the future)
18529 *Ralf S. Engelschall*
18531 * First cut for a very conservative source tree cleanup:
18532 1. merge various obsolete readme texts into doc/ssleay.txt
18533 where we collect the old documents and readme texts.
18534 2. remove the first part of files where I'm already sure that we no
18535 longer need them because of three reasons: either they are just temporary
18536 files which were left by Eric or they are preserved original files where
18537 I've verified that the diff is also available in the CVS via "cvs diff
18538 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
18539 the crypto/md/ stuff).
18541 *Ralf S. Engelschall*
18543 * More extension code. Incomplete support for subject and issuer alt
18544 name, issuer and authority key id. Change the i2v function parameters
18545 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
18546 what that's for :-) Fix to ASN1 macro which messed up
18547 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
18551 * Preliminary support for ENUMERATED type. This is largely copied from the
18556 * Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
18558 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
18560 * Make sure `make rehash` target really finds the `openssl` program.
18562 *Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
18564 * Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
18565 like to hear about it if this slows down other processors.
18569 * Add CygWin32 platform information to Configure script.
18571 *Alan Batie <batie@aahz.jf.intel.com>*
18573 * Fixed ms/32all.bat script: `no_asm` -> `no-asm`
18575 *Rainer W. Gerling <gerling@mpg-gv.mpg.de>*
18577 * New program nseq to manipulate netscape certificate sequences
18581 * Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
18586 * Fixes to BN code. Previously the default was to define BN_RECURSION
18587 but the BN code had some problems that would cause failures when
18588 doing certificate verification and some other functions.
18590 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
18592 * Add ASN1 and PEM code to support netscape certificate sequences.
18596 * Add ASN1 and PEM code to support netscape certificate sequences.
18600 * Add several PKIX and private extended key usage OIDs.
18604 * Modify the 'ca' program to handle the new extension code. Modify
18605 openssl.cnf for new extension format, add comments.
18609 * More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
18610 and add a sample to openssl.cnf so req -x509 now adds appropriate
18615 * Continued X509 V3 changes. Add to other makefiles, integrate with the
18616 error code, add initial support to X509_print() and x509 application.
18620 * Takes a deep breath and start adding X509 V3 extension support code. Add
18621 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
18622 stuff is currently isolated and isn't even compiled yet.
18626 * Continuing patches for GeneralizedTime. Fix up certificate and CRL
18627 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
18628 Removed the versions check from X509 routines when loading extensions:
18629 this allows certain broken certificates that don't set the version
18630 properly to be processed.
18634 * Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
18635 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
18636 can still be regenerated with "make depend".
18640 * Spelling mistake in C version of CAST-128.
18642 *Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>*
18644 * Changes to the error generation code. The perl script err-code.pl
18645 now reads in the old error codes and retains the old numbers, only
18646 adding new ones if necessary. It also only changes the .err files if new
18647 codes are added. The makefiles have been modified to only insert errors
18648 when needed (to avoid needlessly modifying header files). This is done
18649 by only inserting errors if the .err file is newer than the auto generated
18650 C file. To rebuild all the error codes from scratch (the old behaviour)
18651 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
18652 or delete all the .err files.
18656 * CAST-128 was incorrectly implemented for short keys. The C version has
18657 been fixed, but is untested. The assembler versions are also fixed, but
18658 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
18659 to regenerate it if needed.
18660 *Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
18661 Hagino <itojun@kame.net>*
18663 * File was opened incorrectly in randfile.c.
18665 *Ulf Möller <ulf@fitug.de>*
18667 * Beginning of support for GeneralizedTime. d2i, i2d, check and print
18668 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
18669 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
18670 al: it's just almost always a UTCTime. Note this patch adds new error
18671 codes so do a "make errors" if there are problems.
18675 * Correct Linux 1 recognition in config.
18677 *Ulf Möller <ulf@fitug.de>*
18679 * Remove pointless MD5 hash when using DSA keys in ca.
18681 *Anonymous <nobody@replay.com>*
18683 * Generate an error if given an empty string as a cert directory. Also
18684 generate an error if handed NULL (previously returned 0 to indicate an
18685 error, but didn't set one).
18687 *Ben Laurie, reported by Anonymous <nobody@replay.com>*
18689 * Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
18693 * Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
18694 parameters. This was causing a warning which killed off the Win32 compile.
18698 * Remove C++ style comments from crypto/bn/bn_local.h.
18700 *Neil Costigan <neil.costigan@celocom.com>*
18702 * The function OBJ_txt2nid was broken. It was supposed to return a nid
18703 based on a text string, looking up short and long names and finally
18704 "dot" format. The "dot" format stuff didn't work. Added new function
18705 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
18706 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
18707 OID is not part of the table.
18711 * Add prototypes to X509 lookup/verify methods, fixing a bug in
18712 X509_LOOKUP_by_alias().
18716 * Sort openssl functions by name.
18720 * Get the `gendsa` command working and add it to the `list` command. Remove
18721 encryption from sample DSA keys (in case anyone is interested the password
18726 * Make *all* `*_free` functions accept a NULL pointer.
18728 *Frans Heymans <fheymans@isaserver.be>*
18730 * If a DH key is generated in s3_srvr.c, don't blow it by trying to use
18733 *Anonymous <nobody@replay.com>*
18735 * s_server should send the CAfile as acceptable CAs, not its own cert.
18737 *Bodo Moeller <3moeller@informatik.uni-hamburg.de>*
18739 * Don't blow it for numeric `-newkey` arguments to `apps/req`.
18741 *Bodo Moeller <3moeller@informatik.uni-hamburg.de>*
18743 * Temp key "for export" tests were wrong in s3_srvr.c.
18745 *Anonymous <nobody@replay.com>*
18747 * Add prototype for temp key callback functions
18748 SSL_CTX_set_tmp_{rsa,dh}_callback().
18752 * Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
18753 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
18757 * X509_name_add_entry() freed the wrong thing after an error.
18759 *Arne Ansper <arne@ats.cyber.ee>*
18761 * rsa_eay.c would attempt to free a NULL context.
18763 *Arne Ansper <arne@ats.cyber.ee>*
18765 * BIO_s_socket() had a broken should_retry() on Windoze.
18767 *Arne Ansper <arne@ats.cyber.ee>*
18769 * BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
18771 *Arne Ansper <arne@ats.cyber.ee>*
18773 * Make sure the already existing X509_STORE->depth variable is initialized
18774 in X509_STORE_new(), but document the fact that this variable is still
18775 unused in the certificate verification process.
18777 *Ralf S. Engelschall*
18779 * Fix the various library and `apps/` files to free up pkeys obtained from
18780 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
18784 * Fix reference counting in X509_PUBKEY_get(). This makes
18785 demos/maurice/example2.c work, amongst others, probably.
18787 *Steve Henson and Ben Laurie*
18789 * First cut of a cleanup for `apps/`. First the `ssleay` program is now named
18790 `openssl` and second, the shortcut symlinks for the `openssl <command>`
18791 are no longer created. This way we have a single and consistent command
18792 line interface `openssl <command>`, similar to `cvs <command>`.
18794 *Ralf S. Engelschall, Paul Sutton and Ben Laurie*
18796 * ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
18797 BIT STRING wrapper always have zero unused bits.
18801 * Add CA.pl, perl version of CA.sh, add extended key usage OID.
18805 * Make the top-level INSTALL documentation easier to understand.
18809 * Makefiles updated to exit if an error occurs in a sub-directory
18810 make (including if user presses ^C) [Paul Sutton]
18812 * Make Montgomery context stuff explicit in RSA data structure.
18816 * Fix build order of pem and err to allow for generated pem.h.
18820 * Fix renumbering bug in X509_NAME_delete_entry().
18824 * Enhanced the err-ins.pl script so it makes the error library number
18825 global and can add a library name. This is needed for external ASN1 and
18826 other error libraries.
18830 * Fixed sk_insert which never worked properly.
18834 * Fix ASN1 macros so they can handle indefinite length constructed
18835 EXPLICIT tags. Some non standard certificates use these: they can now
18840 * Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
18841 into a single doc/ssleay.txt bundle. This way the information is still
18842 preserved but no longer messes up this directory. Now it's new room for
18843 the new set of documentation files.
18845 *Ralf S. Engelschall*
18847 * SETs were incorrectly DER encoded. This was a major pain, because they
18848 shared code with SEQUENCEs, which aren't coded the same. This means that
18849 almost everything to do with SETs or SEQUENCEs has either changed name or
18850 number of arguments.
18852 *Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>*
18854 * Fix test data to work with the above.
18858 * Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
18859 was already fixed by Eric for 0.9.1 it seems.
18861 *Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>*
18863 * Autodetect FreeBSD3.
18867 * Fix various bugs in Configure. This affects the following platforms:
18871 unixware-2.0-pentium
18876 * Eliminate generated files from CVS. Reorder tests to regenerate files
18877 before they are needed.
18881 * Generate Makefile.ssl from Makefile.org (to keep CVS happy).
18885 ### Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
18887 * Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
18888 changed SSLeay to OpenSSL in version strings.
18890 *Ralf S. Engelschall*
18892 * Some fixups to the top-level documents.
18896 * Fixed the nasty bug where rsaref.h was not found under compile-time
18897 because the symlink to include/ was missing.
18899 *Ralf S. Engelschall*
18901 * Incorporated the popular no-RSA/DSA-only patches
18902 which allow to compile a RSA-free SSLeay.
18904 *Andrew Cooke / Interrader Ldt., Ralf S. Engelschall*
18906 * Fixed nasty rehash problem under `make -f Makefile.ssl links`
18907 when "ssleay" is still not found.
18909 *Ralf S. Engelschall*
18911 * Added more platforms to Configure: Cray T3E, HPUX 11,
18913 *Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>*
18915 * Updated the README file.
18917 *Ralf S. Engelschall*
18919 * Added various .cvsignore files in the CVS repository subdirs
18920 to make a "cvs update" really silent.
18922 *Ralf S. Engelschall*
18924 * Recompiled the error-definition header files and added
18925 missing symbols to the Win32 linker tables.
18927 *Ralf S. Engelschall*
18929 * Cleaned up the top-level documents;
18930 o new files: CHANGES and LICENSE
18931 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
18932 o merged COPYRIGHT into LICENSE
18933 o removed obsolete TODO file
18934 o renamed MICROSOFT to INSTALL.W32
18936 *Ralf S. Engelschall*
18938 * Removed dummy files from the 0.9.1b source tree:
18939 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
18940 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
18941 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
18942 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
18943 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
18945 *Ralf S. Engelschall*
18947 * Added various platform portability fixes.
18951 * The Genesis of the OpenSSL rpject:
18952 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
18953 Young and Tim J. Hudson created while they were working for C2Net until
18956 *The OpenSSL Project*
18958 ### Changes between 0.9.0b and 0.9.1b [not released]
18960 * Updated a few CA certificates under certs/
18964 * Changed some BIGNUM api stuff.
18968 * Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
18969 DGUX x86, Linux Alpha, etc.
18973 * New COMP library [crypto/comp/] for SSL Record Layer Compression:
18974 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
18979 * Add -strparse option to asn1pars program which parses nested
18982 *Dr Stephen Henson <shenson@bigfoot.com>*
18984 * Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
18988 * DSA fix for "ca" program.
18992 * Added "-genkey" option to "dsaparam" program.
18996 * Added RIPE MD160 (rmd160) message digest.
19000 * Added -a (all) option to "ssleay version" command.
19004 * Added PLATFORM define which is the id given to Configure.
19008 * Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
19012 * Extended the ASN.1 parser routines.
19016 * Extended BIO routines to support REUSEADDR, seek, tell, etc.
19020 * Added a BN_CTX to the BN library.
19024 * Fixed the weak key values in DES library
19028 * Changed API in EVP library for cipher aliases.
19032 * Added support for RC2/64bit cipher.
19036 * Converted the lhash library to the crypto/mem.c functions.
19040 * Added more recognized ASN.1 object ids.
19044 * Added more RSA padding checks for SSL/TLS.
19048 * Added BIO proxy/filter functionality.
19052 * Added extra_certs to SSL_CTX which can be used
19053 send extra CA certificates to the client in the CA cert chain sending
19054 process. It can be configured with SSL_CTX_add_extra_chain_cert().
19058 * Now Fortezza is denied in the authentication phase because
19059 this is key exchange mechanism is not supported by SSLeay at all.
19063 * Additional PKCS1 checks.
19067 * Support the string "TLSv1" for all TLS v1 ciphers.
19071 * Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
19072 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
19076 * Fixed a few memory leaks.
19080 * Fixed various code and comment typos.
19084 * A minor bug in ssl/s3_clnt.c where there would always be 4 0
19085 bytes sent in the client random.
19087 *Edward Bishop <ebishop@spyglass.com>*
19091 [CVE-2020-1971]: https://www.openssl.org/news/vulnerabilities.html#CVE-2020-1971
19092 [CVE-2020-1967]: https://www.openssl.org/news/vulnerabilities.html#CVE-2020-1967
19093 [CVE-2019-1563]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1563
19094 [CVE-2019-1559]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1559
19095 [CVE-2019-1552]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1552
19096 [CVE-2019-1551]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1551
19097 [CVE-2019-1549]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1549
19098 [CVE-2019-1547]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1547
19099 [CVE-2019-1543]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1543
19100 [CVE-2018-5407]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-5407
19101 [CVE-2018-0739]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0739
19102 [CVE-2018-0737]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0737
19103 [CVE-2018-0735]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0735
19104 [CVE-2018-0734]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0734
19105 [CVE-2018-0733]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0733
19106 [CVE-2018-0732]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0732
19107 [CVE-2017-3738]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3738
19108 [CVE-2017-3737]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3737
19109 [CVE-2017-3736]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3736
19110 [CVE-2017-3735]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3735
19111 [CVE-2017-3733]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3733
19112 [CVE-2017-3732]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3732
19113 [CVE-2017-3731]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3731
19114 [CVE-2017-3730]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3730
19115 [CVE-2016-7055]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7055
19116 [CVE-2016-7054]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7054
19117 [CVE-2016-7053]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7053
19118 [CVE-2016-7052]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7052
19119 [CVE-2016-6309]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6309
19120 [CVE-2016-6308]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6308
19121 [CVE-2016-6307]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6307
19122 [CVE-2016-6306]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6306
19123 [CVE-2016-6305]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6305
19124 [CVE-2016-6304]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6304
19125 [CVE-2016-6303]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6303
19126 [CVE-2016-6302]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6302
19127 [CVE-2016-2183]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2183
19128 [CVE-2016-2182]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2182
19129 [CVE-2016-2181]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2181
19130 [CVE-2016-2180]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2180
19131 [CVE-2016-2179]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2179
19132 [CVE-2016-2178]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2178
19133 [CVE-2016-2177]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2177
19134 [CVE-2016-2176]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2176
19135 [CVE-2016-2109]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2109
19136 [CVE-2016-2107]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2107
19137 [CVE-2016-2106]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2106
19138 [CVE-2016-2105]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2105
19139 [CVE-2016-0800]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0800
19140 [CVE-2016-0799]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0799
19141 [CVE-2016-0798]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0798
19142 [CVE-2016-0797]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0797
19143 [CVE-2016-0705]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0705
19144 [CVE-2016-0702]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0702
19145 [CVE-2016-0701]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0701
19146 [CVE-2015-3197]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3197
19147 [CVE-2015-3196]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3196
19148 [CVE-2015-3195]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3195
19149 [CVE-2015-3194]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3194
19150 [CVE-2015-3193]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3193
19151 [CVE-2015-1793]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1793
19152 [CVE-2015-1792]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1792
19153 [CVE-2015-1791]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1791
19154 [CVE-2015-1790]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1790
19155 [CVE-2015-1789]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1789
19156 [CVE-2015-1788]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1788
19157 [CVE-2015-1787]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1787
19158 [CVE-2015-0293]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0293
19159 [CVE-2015-0291]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0291
19160 [CVE-2015-0290]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0290
19161 [CVE-2015-0289]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0289
19162 [CVE-2015-0288]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0288
19163 [CVE-2015-0287]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0287
19164 [CVE-2015-0286]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0286
19165 [CVE-2015-0285]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0285
19166 [CVE-2015-0209]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0209
19167 [CVE-2015-0208]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0208
19168 [CVE-2015-0207]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0207
19169 [CVE-2015-0206]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0206
19170 [CVE-2015-0205]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0205
19171 [CVE-2015-0204]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0204
19172 [CVE-2014-8275]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-8275
19173 [CVE-2014-5139]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-5139
19174 [CVE-2014-3572]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3572
19175 [CVE-2014-3571]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3571
19176 [CVE-2014-3570]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3570
19177 [CVE-2014-3569]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3569
19178 [CVE-2014-3568]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3568
19179 [CVE-2014-3567]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3567
19180 [CVE-2014-3566]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3566
19181 [CVE-2014-3513]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3513
19182 [CVE-2014-3512]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3512
19183 [CVE-2014-3511]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3511
19184 [CVE-2014-3510]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3510
19185 [CVE-2014-3509]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3509
19186 [CVE-2014-3508]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3508
19187 [CVE-2014-3507]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3507
19188 [CVE-2014-3506]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3506
19189 [CVE-2014-3505]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3505
19190 [CVE-2014-3470]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3470
19191 [CVE-2014-0224]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0224
19192 [CVE-2014-0221]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0221
19193 [CVE-2014-0195]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0195
19194 [CVE-2014-0160]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0160
19195 [CVE-2014-0076]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0076
19196 [CVE-2013-6450]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-6450
19197 [CVE-2013-4353]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-4353
19198 [CVE-2013-0169]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-0169
19199 [CVE-2013-0166]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-0166
19200 [CVE-2012-2686]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2686
19201 [CVE-2012-2333]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2333
19202 [CVE-2012-2110]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2110
19203 [CVE-2012-0884]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0884
19204 [CVE-2012-0050]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0050
19205 [CVE-2012-0027]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0027
19206 [CVE-2011-4619]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4619
19207 [CVE-2011-4577]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4577
19208 [CVE-2011-4576]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4576
19209 [CVE-2011-4109]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4109
19210 [CVE-2011-4108]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4108
19211 [CVE-2011-3210]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-3210
19212 [CVE-2011-3207]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-3207
19213 [CVE-2011-0014]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-0014
19214 [CVE-2010-4252]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-4252
19215 [CVE-2010-4180]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-4180
19216 [CVE-2010-3864]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-3864
19217 [CVE-2010-1633]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-1633
19218 [CVE-2010-0740]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-0740
19219 [CVE-2010-0433]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-0433
19220 [CVE-2009-4355]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-4355
19221 [CVE-2009-3555]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-3555
19222 [CVE-2009-3245]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-3245
19223 [CVE-2009-1386]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1386
19224 [CVE-2009-1379]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1379
19225 [CVE-2009-1378]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1378
19226 [CVE-2009-1377]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1377
19227 [CVE-2009-0789]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0789
19228 [CVE-2009-0591]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0591
19229 [CVE-2009-0590]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0590
19230 [CVE-2008-5077]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-5077
19231 [CVE-2008-1678]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-1678
19232 [CVE-2008-1672]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-1672
19233 [CVE-2008-0891]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-0891
19234 [CVE-2007-5135]: https://www.openssl.org/news/vulnerabilities.html#CVE-2007-5135
19235 [CVE-2007-4995]: https://www.openssl.org/news/vulnerabilities.html#CVE-2007-4995
19236 [CVE-2006-4343]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-4343
19237 [CVE-2006-4339]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-4339
19238 [CVE-2006-3738]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-3738
19239 [CVE-2006-2940]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-2940
19240 [CVE-2006-2937]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-2937
19241 [CVE-2005-2969]: https://www.openssl.org/news/vulnerabilities.html#CVE-2005-2969
19242 [CVE-2004-0112]: https://www.openssl.org/news/vulnerabilities.html#CVE-2004-0112
19243 [CVE-2004-0079]: https://www.openssl.org/news/vulnerabilities.html#CVE-2004-0079
19244 [CVE-2003-0851]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0851
19245 [CVE-2003-0545]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0545
19246 [CVE-2003-0544]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0544
19247 [CVE-2003-0543]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0543
19248 [CVE-2003-0078]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0078
19249 [CVE-2002-0659]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0659
19250 [CVE-2002-0657]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0657
19251 [CVE-2002-0656]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0656
19252 [CVE-2002-0655]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0655