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 [xx XXX xxxx]
33 * Add ciphersuites based on DHE_PSK (RFC 4279) and ECDHE_PSK (RFC 5489)
34 to the list of ciphersuites providing Perfect Forward Secrecy as
35 required by SECLEVEL >= 3.
37 *Dmitry Belyavskiy, Nicola Tuveri*
39 * Made the AES constant time code for no-asm configurations
40 optional due to the resulting 95% performance degradation.
41 The AES constant time code can be enabled, for no assembly
42 builds, with: ./config no-asm -DOPENSSL_AES_CONST_TIME
46 * Fixed PEM_write_bio_PKCS8PrivateKey() to make it possible to use empty
51 ### Changes between 3.0.0 and 3.0.1 [14 Dec 2021]
53 * Fixed invalid handling of X509_verify_cert() internal errors in libssl
54 Internally libssl in OpenSSL calls X509_verify_cert() on the client side to
55 verify a certificate supplied by a server. That function may return a
56 negative return value to indicate an internal error (for example out of
57 memory). Such a negative return value is mishandled by OpenSSL and will cause
58 an IO function (such as SSL_connect() or SSL_do_handshake()) to not indicate
59 success and a subsequent call to SSL_get_error() to return the value
60 SSL_ERROR_WANT_RETRY_VERIFY. This return value is only supposed to be
61 returned by OpenSSL if the application has previously called
62 SSL_CTX_set_cert_verify_callback(). Since most applications do not do this
63 the SSL_ERROR_WANT_RETRY_VERIFY return value from SSL_get_error() will be
64 totally unexpected and applications may not behave correctly as a result. The
65 exact behaviour will depend on the application but it could result in
66 crashes, infinite loops or other similar incorrect responses.
68 This issue is made more serious in combination with a separate bug in OpenSSL
69 3.0 that will cause X509_verify_cert() to indicate an internal error when
70 processing a certificate chain. This will occur where a certificate does not
71 include the Subject Alternative Name extension but where a Certificate
72 Authority has enforced name constraints. This issue can occur even with valid
78 * Corrected a few file name and file reference bugs in the build,
79 installation and setup scripts, which lead to installation verification
80 failures. Slightly enhanced the installation verification script.
84 * Fixed EVP_PKEY_eq() to make it possible to use it with strictly private
89 * Fixed PVK encoder to properly query for the passphrase.
93 * Multiple fixes in the OSSL_HTTP API functions.
97 * Allow sign extension in OSSL_PARAM_allocate_from_text() for the
98 OSSL_PARAM_INTEGER data type and return error on negative numbers
99 used with the OSSL_PARAM_UNSIGNED_INTEGER data type. Make
100 OSSL_PARAM_BLD_push_BN{,_pad}() return an error on negative numbers.
104 * Allow copying uninitialized digest contexts with EVP_MD_CTX_copy_ex.
108 * Fixed detection of ARMv7 and ARM64 CPU features on FreeBSD.
112 * Multiple threading fixes.
116 * Added NULL digest implementation to keep compatibility with 1.1.1 version.
120 * Allow fetching an operation from the provider that owns an unexportable key
121 as a fallback if that is still allowed by the property query.
125 ### Changes between 1.1.1 and 3.0.0 [7 sep 2021]
127 * TLS_MAX_VERSION, DTLS_MAX_VERSION and DTLS_MIN_VERSION constants are now
132 * The `OPENSSL_s390xcap` environment variable can be used to set bits in the
133 S390X capability vector to zero. This simplifies testing of different code
134 paths on S390X architecture.
138 * Encrypting more than 2^64 TLS records with AES-GCM is disallowed
139 as per FIPS 140-2 IG A.5 "Key/IV Pair Uniqueness Requirements from
140 SP 800-38D". The communication will fail at this point.
144 * The EC_GROUP_clear_free() function is deprecated as there is nothing
145 confidential in EC_GROUP data.
149 * The byte order mark (BOM) character is ignored if encountered at the
150 beginning of a PEM-formatted file.
154 * Added CMS support for the Russian GOST algorithms.
158 * Due to move of the implementation of cryptographic operations
159 to the providers, validation of various operation parameters can
160 be postponed until the actual operation is executed where previously
161 it happened immediately when an operation parameter was set.
163 For example when setting an unsupported curve with
164 EVP_PKEY_CTX_set_ec_paramgen_curve_nid() this function call will not
165 fail but later keygen operations with the EVP_PKEY_CTX will fail.
167 *OpenSSL team members and many third party contributors*
169 * The EVP_get_cipherbyname() function will return NULL for algorithms such as
170 "AES-128-SIV", "AES-128-CBC-CTS" and "CAMELLIA-128-CBC-CTS" which were
171 previously only accessible via low level interfaces. Use EVP_CIPHER_fetch()
172 instead to retrieve these algorithms from a provider.
176 * On build targets where the multilib postfix is set in the build
177 configuration the libdir directory was changing based on whether
178 the lib directory with the multilib postfix exists on the system
179 or not. This unpredictable behavior was removed and eventual
180 multilib postfix is now always added to the default libdir. Use
181 `--libdir=lib` to override the libdir if adding the postfix is
186 * The triple DES key wrap functionality now conforms to RFC 3217 but is
187 no longer interoperable with OpenSSL 1.1.1.
191 * The ERR_GET_FUNC() function was removed. With the loss of meaningful
192 function codes, this function can only cause problems for calling
197 * Add a configurable flag to output date formats as ISO 8601. Does not
198 change the default date format.
202 * Version of MSVC earlier than 1300 could get link warnings, which could
203 be suppressed if the undocumented -DI_CAN_LIVE_WITH_LNK4049 was set.
204 Support for this flag has been removed.
208 * Rework and make DEBUG macros consistent. Remove unused -DCONF_DEBUG,
209 -DBN_CTX_DEBUG, and REF_PRINT. Add a new tracing category and use it for
210 printing reference counts. Rename -DDEBUG_UNUSED to -DUNUSED_RESULT_DEBUG
211 Fix BN_DEBUG_RAND so it compiles and, when set, force DEBUG_RAND to be set
212 also. Rename engine_debug_ref to be ENGINE_REF_PRINT also for consistency.
216 * The signatures of the functions to get and set options on SSL and
217 SSL_CTX objects changed from "unsigned long" to "uint64_t" type.
218 Some source code changes may be required.
222 * The public definitions of conf_method_st and conf_st have been
223 deprecated. They will be made opaque in a future release.
225 *Rich Salz and Tomáš Mráz*
227 * Client-initiated renegotiation is disabled by default. To allow it, use
228 the -client_renegotiation option, the SSL_OP_ALLOW_CLIENT_RENEGOTIATION
229 flag, or the "ClientRenegotiation" config parameter as appropriate.
233 * Add "abspath" and "includedir" pragma's to config files, to prevent,
234 or modify relative pathname inclusion.
238 * OpenSSL includes a cryptographic module that is intended to be FIPS 140-2
239 validated. Please consult the README-FIPS and
240 README-PROVIDERS files, as well as the migration guide.
242 *OpenSSL team members and many third party contributors*
244 * For the key types DH and DHX the allowed settable parameters are now different.
248 * The openssl commands that read keys, certificates, and CRLs now
249 automatically detect the PEM or DER format of the input files.
251 *David von Oheimb, Richard Levitte, and Tomáš Mráz*
253 * Added enhanced PKCS#12 APIs which accept a library context.
257 * The default manual page suffix ($MANSUFFIX) has been changed to "ossl"
261 * Added support for Kernel TLS (KTLS).
263 *Boris Pismenny, John Baldwin and Andrew Gallatin*
265 * Support for RFC 5746 secure renegotiation is now required by default for
266 SSL or TLS connections to succeed.
270 * The signature of the `copy` functional parameter of the
271 EVP_PKEY_meth_set_copy() function has changed so its `src` argument is
272 now `const EVP_PKEY_CTX *` instead of `EVP_PKEY_CTX *`. Similarly
273 the signature of the `pub_decode` functional parameter of the
274 EVP_PKEY_asn1_set_public() function has changed so its `pub` argument is
275 now `const X509_PUBKEY *` instead of `X509_PUBKEY *`.
279 * The error return values from some control calls (ctrl) have changed.
283 * A public key check is now performed during EVP_PKEY_derive_set_peer().
287 * Many functions in the EVP_ namespace that are getters of values from
288 implementations or contexts were renamed to include get or get0 in their
289 names. Old names are provided as macro aliases for compatibility and
294 * The EVP_PKEY_CTRL_PKCS7_ENCRYPT, EVP_PKEY_CTRL_PKCS7_DECRYPT,
295 EVP_PKEY_CTRL_PKCS7_SIGN, EVP_PKEY_CTRL_CMS_ENCRYPT,
296 EVP_PKEY_CTRL_CMS_DECRYPT, and EVP_PKEY_CTRL_CMS_SIGN control operations
301 * The EVP_PKEY_public_check() and EVP_PKEY_param_check() functions now work for
304 * The output from the command line applications may have minor
309 * The output from numerous "printing" may have minor changes.
313 * Windows thread synchronization uses read/write primitives (SRWLock) when
314 supported by the OS, otherwise CriticalSection continues to be used.
318 * Add filter BIO BIO_f_readbuffer() that allows BIO_tell() and BIO_seek() to
319 work on read only BIO source/sinks that do not support these functions.
320 This allows piping or redirection of a file BIO using stdin to be buffered
321 into memory. This is used internally in OSSL_DECODER_from_bio().
325 * OSSL_STORE_INFO_get_type() may now return an additional value. In 1.1.1
326 this function would return one of the values OSSL_STORE_INFO_NAME,
327 OSSL_STORE_INFO_PKEY, OSSL_STORE_INFO_PARAMS, OSSL_STORE_INFO_CERT or
328 OSSL_STORE_INFO_CRL. Decoded public keys would previously have been reported
329 as type OSSL_STORE_INFO_PKEY in 1.1.1. In 3.0 decoded public keys are now
330 reported as having the new type OSSL_STORE_INFO_PUBKEY. Applications
331 using this function should be amended to handle the changed return value.
335 * Improved adherence to Enhanced Security Services (ESS, RFC 2634 and RFC 5035)
336 for the TSP and CMS Advanced Electronic Signatures (CAdES) implementations.
337 As required by RFC 5035 check both ESSCertID and ESSCertIDv2 if both present.
338 Correct the semantics of checking the validation chain in case ESSCertID{,v2}
339 contains more than one certificate identifier: This means that all
340 certificates referenced there MUST be part of the validation chain.
344 * The implementation of older EVP ciphers related to CAST, IDEA, SEED, RC2, RC4,
345 RC5, DESX and DES have been moved to the legacy provider.
349 * The implementation of the EVP digests MD2, MD4, MDC2, WHIRLPOOL and
350 RIPEMD-160 have been moved to the legacy provider.
354 * The deprecated function EVP_PKEY_get0() now returns NULL being called for a
359 * The deprecated functions EVP_PKEY_get0_RSA(),
360 EVP_PKEY_get0_DSA(), EVP_PKEY_get0_EC_KEY(), EVP_PKEY_get0_DH(),
361 EVP_PKEY_get0_hmac(), EVP_PKEY_get0_poly1305() and EVP_PKEY_get0_siphash() as
362 well as the similarly named "get1" functions behave differently in
367 * A number of functions handling low-level keys or engines were deprecated
368 including EVP_PKEY_set1_engine(), EVP_PKEY_get0_engine(), EVP_PKEY_assign(),
369 EVP_PKEY_get0(), EVP_PKEY_get0_hmac(), EVP_PKEY_get0_poly1305() and
370 EVP_PKEY_get0_siphash().
374 * PKCS#5 PBKDF1 key derivation has been moved from PKCS5_PBE_keyivgen() into
375 the legacy crypto provider as an EVP_KDF. Applications requiring this KDF
376 will need to load the legacy crypto provider. This includes these PBE
377 algorithms which use this KDF:
378 - NID_pbeWithMD2AndDES_CBC
379 - NID_pbeWithMD5AndDES_CBC
380 - NID_pbeWithSHA1AndRC2_CBC
381 - NID_pbeWithMD2AndRC2_CBC
382 - NID_pbeWithMD5AndRC2_CBC
383 - NID_pbeWithSHA1AndDES_CBC
387 * Deprecated obsolete BIO_set_callback(), BIO_get_callback(), and
388 BIO_debug_callback() functions.
392 * Deprecated obsolete EVP_PKEY_CTX_get0_dh_kdf_ukm() and
393 EVP_PKEY_CTX_get0_ecdh_kdf_ukm() functions.
397 * The RAND_METHOD APIs have been deprecated.
401 * The SRP APIs have been deprecated.
405 * Add a compile time option to prevent the caching of provider fetched
406 algorithms. This is enabled by including the no-cached-fetch option
407 at configuration time.
411 * pkcs12 now uses defaults of PBKDF2, AES and SHA-256, with a MAC iteration
412 count of PKCS12_DEFAULT_ITER.
414 *Tomáš Mráz and Sahana Prasad*
416 * The openssl speed command does not use low-level API calls anymore.
420 * Parallel dual-prime 1024-bit modular exponentiation for AVX512_IFMA
423 *Ilya Albrekht, Sergey Kirillov, Andrey Matyukov (Intel Corp)*
425 * Combining the Configure options no-ec and no-dh no longer disables TLSv1.3.
429 * Implemented support for fully "pluggable" TLSv1.3 groups. This means that
430 providers may supply their own group implementations (using either the "key
431 exchange" or the "key encapsulation" methods) which will automatically be
432 detected and used by libssl.
434 *Matt Caswell, Nicola Tuveri*
436 * The undocumented function X509_certificate_type() has been deprecated;
440 * Deprecated the obsolete BN_pseudo_rand() and BN_pseudo_rand_range().
444 * Removed RSA padding mode for SSLv23 (which was only used for
445 SSLv2). This includes the functions RSA_padding_check_SSLv23() and
446 RSA_padding_add_SSLv23() and the `-ssl` option in the deprecated
451 * Deprecated the obsolete X9.31 RSA key generation related functions.
453 * While a callback function set via `SSL_CTX_set_cert_verify_callback()`
454 is not allowed to return a value > 1, this is no more taken as failure.
456 *Viktor Dukhovni and David von Oheimb*
458 * Deprecated the obsolete X9.31 RSA key generation related functions
459 BN_X931_generate_Xpq(), BN_X931_derive_prime_ex(), and
460 BN_X931_generate_prime_ex().
464 * The default key generation method for the regular 2-prime RSA keys was
465 changed to the FIPS 186-4 B.3.6 method.
469 * Deprecated the BN_is_prime_ex() and BN_is_prime_fasttest_ex() functions.
473 * Deprecated EVP_MD_CTX_set_update_fn() and EVP_MD_CTX_update_fn().
477 * Deprecated the type OCSP_REQ_CTX and the functions OCSP_REQ_CTX_*() and
478 replaced with OSSL_HTTP_REQ_CTX and the functions OSSL_HTTP_REQ_CTX_*().
480 *Rich Salz, Richard Levitte, and David von Oheimb*
482 * Deprecated `X509_http_nbio()` and `X509_CRL_http_nbio()`.
486 * Deprecated `OCSP_parse_url()`.
490 * Validation of SM2 keys has been separated from the validation of regular EC
495 * Behavior of the `pkey` app is changed, when using the `-check` or `-pubcheck`
496 switches: a validation failure triggers an early exit, returning a failure
497 exit status to the parent process.
501 * Changed behavior of SSL_CTX_set_ciphersuites() and SSL_set_ciphersuites()
502 to ignore unknown ciphers.
506 * The `-cipher-commands` and `-digest-commands` options
507 of the command line utility `list` have been deprecated.
508 Instead use the `-cipher-algorithms` and `-digest-algorithms` options.
512 * Added convenience functions for generating asymmetric key pairs:
513 The 'quick' one-shot (yet somewhat limited) function L<EVP_PKEY_Q_keygen(3)>
514 and macros for the most common cases: <EVP_RSA_gen(3)> and L<EVP_EC_gen(3)>.
518 * All of the low level EC_KEY functions have been deprecated.
520 *Shane Lontis, Paul Dale, Richard Levitte, and Tomáš Mráz*
522 * Deprecated all the libcrypto and libssl error string loading
527 * The functions SSL_CTX_set_tmp_dh_callback and SSL_set_tmp_dh_callback, as
528 well as the macros SSL_CTX_set_tmp_dh() and SSL_set_tmp_dh() have been
533 * The `-crypt` option to the `passwd` command line tool has been removed.
537 * The -C option to the `x509`, `dhparam`, `dsaparam`, and `ecparam` commands
542 * Add support for AES Key Wrap inverse ciphers to the EVP layer.
546 * Deprecated EVP_PKEY_set1_tls_encodedpoint() and
547 EVP_PKEY_get1_tls_encodedpoint().
551 * The security callback, which can be customised by application code, supports
552 the security operation SSL_SECOP_TMP_DH. One location of the "other" parameter
553 was incorrectly passing a DH object. It now passed an EVP_PKEY in all cases.
557 * Add PKCS7_get_octet_string() and PKCS7_type_is_other() to the public
558 interface. Their functionality remains unchanged.
562 * Added new option for 'openssl list', '-providers', which will display the
563 list of loaded providers, their names, version and status. It optionally
564 displays their gettable parameters.
568 * Removed EVP_PKEY_set_alias_type().
572 * Deprecated `EVP_PKEY_CTX_set_rsa_keygen_pubexp()` and introduced
573 `EVP_PKEY_CTX_set1_rsa_keygen_pubexp()`, which is now preferred.
577 * Changed all "STACK" functions to be macros instead of inline functions. Macro
578 parameters are still checked for type safety at compile time via helper
583 * Remove the RAND_DRBG API
585 *Paul Dale and Matthias St. Pierre*
587 * Allow `SSL_set1_host()` and `SSL_add1_host()` to take IP literal addresses
588 as well as actual hostnames.
592 * The 'MinProtocol' and 'MaxProtocol' configuration commands now silently
593 ignore TLS protocol version bounds when configuring DTLS-based contexts, and
594 conversely, silently ignore DTLS protocol version bounds when configuring
595 TLS-based contexts. The commands can be repeated to set bounds of both
596 types. The same applies with the corresponding "min_protocol" and
597 "max_protocol" command-line switches, in case some application uses both TLS
600 SSL_CTX instances that are created for a fixed protocol version (e.g.
601 `TLSv1_server_method()`) also silently ignore version bounds. Previously
602 attempts to apply bounds to these protocol versions would result in an
603 error. Now only the "version-flexible" SSL_CTX instances are subject to
604 limits in configuration files in command-line options.
608 * Deprecated the `ENGINE` API. Engines should be replaced with providers
613 * Reworked the recorded ERR codes to make better space for system errors.
614 To distinguish them, the macro `ERR_SYSTEM_ERROR()` indicates if the
615 given code is a system error (true) or an OpenSSL error (false).
619 * Reworked the test perl framework to better allow parallel testing.
621 *Nicola Tuveri and David von Oheimb*
623 * Added ciphertext stealing algorithms AES-128-CBC-CTS, AES-192-CBC-CTS and
624 AES-256-CBC-CTS to the providers. CS1, CS2 and CS3 variants are supported.
628 * 'Configure' has been changed to figure out the configuration target if
629 none is given on the command line. Consequently, the 'config' script is
630 now only a mere wrapper. All documentation is changed to only mention
633 *Rich Salz and Richard Levitte*
635 * Added a library context `OSSL_LIB_CTX` that applications as well as
636 other libraries can use to form a separate context within which
637 libcrypto operations are performed.
641 * Added various `_ex` functions to the OpenSSL API that support using
642 a non-default `OSSL_LIB_CTX`.
646 * Handshake now fails if Extended Master Secret extension is dropped
651 * Dropped interactive mode from the `openssl` program.
655 * Deprecated `EVP_PKEY_cmp()` and `EVP_PKEY_cmp_parameters()`.
657 *David von Oheimb and Shane Lontis*
659 * Deprecated `EC_METHOD_get_field_type()`.
663 * Deprecated EC_GFp_simple_method(), EC_GFp_mont_method(),
664 EC_GF2m_simple_method(), EC_GFp_nist_method(), EC_GFp_nistp224_method()
665 EC_GFp_nistp256_method(), and EC_GFp_nistp521_method().
669 * Deprecated EC_GROUP_new(), EC_GROUP_method_of(), and EC_POINT_method_of().
673 * Add CAdES-BES signature verification support, mostly derived
674 from ESSCertIDv2 TS (RFC 5816) contribution by Marek Klein.
676 *Filipe Raimundo da Silva*
678 * Add CAdES-BES signature scheme and attributes support (RFC 5126) to CMS API.
682 * Added the AuthEnvelopedData content type structure (RFC 5083) with AES-GCM
683 parameter (RFC 5084) for the Cryptographic Message Syntax (CMS).
687 * Deprecated EC_POINT_make_affine() and EC_POINTs_make_affine().
691 * Deprecated EC_GROUP_precompute_mult(), EC_GROUP_have_precompute_mult(), and
692 EC_KEY_precompute_mult().
696 * Deprecated EC_POINTs_mul().
700 * Removed FIPS_mode() and FIPS_mode_set().
704 * The SSL option SSL_OP_IGNORE_UNEXPECTED_EOF is introduced.
708 * Deprecated EC_POINT_set_Jprojective_coordinates_GFp() and
709 EC_POINT_get_Jprojective_coordinates_GFp().
713 * Added OSSL_PARAM_BLD to the public interface. This allows OSSL_PARAM
714 arrays to be more easily constructed via a series of utility functions.
715 Create a parameter builder using OSSL_PARAM_BLD_new(), add parameters using
716 the various push functions and finally convert to a passable OSSL_PARAM
717 array using OSSL_PARAM_BLD_to_param().
721 * The security strength of SHA1 and MD5 based signatures in TLS has been
726 * Added EVP_PKEY_set_type_by_keymgmt(), to initialise an EVP_PKEY to
727 contain a provider side internal key.
731 * ASN1_verify(), ASN1_digest() and ASN1_sign() have been deprecated.
735 * Project text documents not yet having a proper file name extension
736 (`HACKING`, `LICENSE`, `NOTES*`, `README*`, `VERSION`) have been renamed to
737 `*.md` as far as reasonable, else `*.txt`, for better use with file managers.
741 * The main project documents (README, NEWS, CHANGES, INSTALL, SUPPORT)
742 have been converted to Markdown with the goal to produce documents
743 which not only look pretty when viewed online in the browser, but
744 remain well readable inside a plain text editor.
746 To achieve this goal, a 'minimalistic' Markdown style has been applied
747 which avoids formatting elements that interfere too much with the
748 reading flow in the text file. For example, it
750 * avoids [ATX headings][] and uses [setext headings][] instead
751 (which works for `<h1>` and `<h2>` headings only).
752 * avoids [inline links][] and uses [reference links][] instead.
753 * avoids [fenced code blocks][] and uses [indented code blocks][] instead.
755 [ATX headings]: https://github.github.com/gfm/#atx-headings
756 [setext headings]: https://github.github.com/gfm/#setext-headings
757 [inline links]: https://github.github.com/gfm/#inline-link
758 [reference links]: https://github.github.com/gfm/#reference-link
759 [fenced code blocks]: https://github.github.com/gfm/#fenced-code-blocks
760 [indented code blocks]: https://github.github.com/gfm/#indented-code-blocks
762 *Matthias St. Pierre*
764 * The test suite is changed to preserve results of each test recipe.
765 A new directory test-runs/ with subdirectories named like the
766 test recipes are created in the build tree for this purpose.
770 * Added an implementation of CMP and CRMF (RFC 4210, RFC 4211 RFC 6712).
771 This adds `crypto/cmp/`, `crpyto/crmf/`, `apps/cmp.c`, and `test/cmp_*`.
772 See L<openssl-cmp(1)> and L<OSSL_CMP_exec_IR_ses(3)> as starting points.
774 *David von Oheimb, Martin Peylo*
776 * Generalized the HTTP client code from `crypto/ocsp/` into `crpyto/http/`.
777 It supports arbitrary request and response content types, GET redirection,
778 TLS, connections via HTTP(S) proxies, connections and exchange via
779 user-defined BIOs (allowing implicit connections), persistent connections,
780 and timeout checks. See L<OSSL_HTTP_transfer(3)> etc. for details.
781 The legacy OCSP-focused (and only partly documented) API
782 is retained for backward compatibility, while most of it is deprecated.
786 * Added `util/check-format.pl`, a tool for checking adherence to the
787 OpenSSL coding style <https://www.openssl.org/policies/codingstyle.html>.
788 The checks performed are incomplete and yield some false positives.
789 Still the tool should be useful for detecting most typical glitches.
793 * `BIO_do_connect()` and `BIO_do_handshake()` have been extended:
794 If domain name resolution yields multiple IP addresses all of them are tried
795 after `connect()` failures.
799 * All of the low level RSA functions have been deprecated.
803 * X509 certificates signed using SHA1 are no longer allowed at security
808 * The command line utilities dhparam, dsa, gendsa and dsaparam have been
809 modified to use PKEY APIs. These commands are now in maintenance mode
810 and no new features will be added to them.
814 * The command line utility rsautl has been deprecated.
818 * The command line utilities genrsa and rsa have been modified to use PKEY
819 APIs. They now write PKCS#8 keys by default. These commands are now in
820 maintenance mode and no new features will be added to them.
824 * All of the low level DH functions have been deprecated.
826 *Paul Dale and Matt Caswell*
828 * All of the low level DSA functions have been deprecated.
832 * Reworked the treatment of EC EVP_PKEYs with the SM2 curve to
833 automatically become EVP_PKEY_SM2 rather than EVP_PKEY_EC.
837 * Deprecated low level ECDH and ECDSA functions.
841 * Deprecated EVP_PKEY_decrypt_old() and EVP_PKEY_encrypt_old().
845 * Enhanced the documentation of EVP_PKEY_get_size(), EVP_PKEY_get_bits()
846 and EVP_PKEY_get_security_bits(). Especially EVP_PKEY_get_size() needed
847 a new formulation to include all the things it can be used for,
848 as well as words of caution.
852 * The SSL_CTX_set_tlsext_ticket_key_cb(3) function has been deprecated.
856 * All of the low level HMAC functions have been deprecated.
858 *Paul Dale and David von Oheimb*
860 * Over two thousand fixes were made to the documentation, including:
861 - Common options (such as -rand/-writerand, TLS version control, etc)
862 were refactored and point to newly-enhanced descriptions in openssl.pod.
863 - Added style conformance for all options (with help from Richard Levitte),
864 documented all reported missing options, added a CI build to check
865 that all options are documented and that no unimplemented options
867 - Documented some internals, such as all use of environment variables.
868 - Addressed all internal broken L<> references.
872 * All of the low level CMAC functions have been deprecated.
876 * The low-level MD2, MD4, MD5, MDC2, RIPEMD160 and Whirlpool digest
877 functions have been deprecated.
879 *Paul Dale and David von Oheimb*
881 * Corrected the documentation of the return values from the `EVP_DigestSign*`
882 set of functions. The documentation mentioned negative values for some
883 errors, but this was never the case, so the mention of negative values
886 Code that followed the documentation and thereby check with something
887 like `EVP_DigestSignInit(...) <= 0` will continue to work undisturbed.
891 * All of the low level cipher functions have been deprecated.
893 *Matt Caswell and Paul Dale*
895 * Removed include/openssl/opensslconf.h.in and replaced it with
896 include/openssl/configuration.h.in, which differs in not including
897 <openssl/macros.h>. A short header include/openssl/opensslconf.h
898 was added to include both.
900 This allows internal hacks where one might need to modify the set
901 of configured macros, for example this if deprecated symbols are
902 still supposed to be available internally:
904 #include <openssl/configuration.h>
906 #undef OPENSSL_NO_DEPRECATED
907 #define OPENSSL_SUPPRESS_DEPRECATED
909 #include <openssl/macros.h>
911 This should not be used by applications that use the exported
912 symbols, as that will lead to linking errors.
916 * Fixed an overflow bug in the x64_64 Montgomery squaring procedure
917 used in exponentiation with 512-bit moduli. No EC algorithms are
918 affected. Analysis suggests that attacks against 2-prime RSA1024,
919 3-prime RSA1536, and DSA1024 as a result of this defect would be very
920 difficult to perform and are not believed likely. Attacks against DH512
921 are considered just feasible. However, for an attack the target would
922 have to re-use the DH512 private key, which is not recommended anyway.
923 Also applications directly using the low-level API BN_mod_exp may be
924 affected if they use BN_FLG_CONSTTIME.
929 * Most memory-debug features have been deprecated, and the functionality
930 replaced with no-ops.
934 * Added documentation for the STACK API.
938 * Introduced a new method type and API, OSSL_ENCODER, to represent
939 generic encoders. These do the same sort of job that PEM writers
940 and d2i functions do, but with support for methods supplied by
941 providers, and the possibility for providers to support other
946 * Introduced a new method type and API, OSSL_DECODER, to represent
947 generic decoders. These do the same sort of job that PEM readers
948 and i2d functions do, but with support for methods supplied by
949 providers, and the possibility for providers to support other
954 * Added a .pragma directive to the syntax of configuration files, to
955 allow varying behavior in a supported and predictable manner.
956 Currently added pragma:
960 This allows dollar signs to be a keyword character unless it's
961 followed by a opening brace or parenthesis. This is useful for
962 platforms where dollar signs are commonly used in names, such as
963 volume names and system directory names on VMS.
967 * Added functionality to create an EVP_PKEY from user data.
971 * Change the interpretation of the '--api' configuration option to
972 mean that this is a desired API compatibility level with no
973 further meaning. The previous interpretation, that this would
974 also mean to remove all deprecated symbols up to and including
975 the given version, no requires that 'no-deprecated' is also used
976 in the configuration.
978 When building applications, the desired API compatibility level
979 can be set with the OPENSSL_API_COMPAT macro like before. For
980 API compatibility version below 3.0, the old style numerical
981 value is valid as before, such as -DOPENSSL_API_COMPAT=0x10100000L.
982 For version 3.0 and on, the value is expected to be the decimal
983 value calculated from the major and minor version like this:
985 MAJOR * 10000 + MINOR * 100
989 -DOPENSSL_API_COMPAT=30000 For 3.0
990 -DOPENSSL_API_COMPAT=30200 For 3.2
992 To hide declarations that are deprecated up to and including the
993 given API compatibility level, -DOPENSSL_NO_DEPRECATED must be
994 given when building the application as well.
998 * Added the X509_LOOKUP_METHOD called X509_LOOKUP_store, to allow
999 access to certificate and CRL stores via URIs and OSSL_STORE
1002 This adds the following functions:
1004 - X509_LOOKUP_store()
1005 - X509_STORE_load_file()
1006 - X509_STORE_load_path()
1007 - X509_STORE_load_store()
1008 - SSL_add_store_cert_subjects_to_stack()
1009 - SSL_CTX_set_default_verify_store()
1010 - SSL_CTX_load_verify_file()
1011 - SSL_CTX_load_verify_dir()
1012 - SSL_CTX_load_verify_store()
1016 * Added a new method to gather entropy on VMS, based on SYS$GET_ENTROPY.
1017 The presence of this system service is determined at run-time.
1021 * Added functionality to create an EVP_PKEY context based on data
1022 for methods from providers. This takes an algorithm name and a
1023 property query string and simply stores them, with the intent
1024 that any operation that uses this context will use those strings
1025 to fetch the needed methods implicitly, thereby making the port
1026 of application written for pre-3.0 OpenSSL easier.
1030 * The undocumented function NCONF_WIN32() has been deprecated; for
1031 conversion details see the HISTORY section of doc/man5/config.pod
1035 * Introduced the new functions EVP_DigestSignInit_ex() and
1036 EVP_DigestVerifyInit_ex(). The macros EVP_DigestSignUpdate() and
1037 EVP_DigestVerifyUpdate() have been converted to functions. See the man
1038 pages for further details.
1042 * Over two thousand fixes were made to the documentation, including:
1043 adding missing command flags, better style conformance, documentation
1046 *Rich Salz, Richard Levitte*
1048 * s390x assembly pack: add hardware-support for P-256, P-384, P-521,
1049 X25519, X448, Ed25519 and Ed448.
1053 * Print all values for a PKCS#12 attribute with 'openssl pkcs12', not just
1058 * Deprecated the public definition of `ERR_STATE` as well as the function
1059 `ERR_get_state()`. This is done in preparation of making `ERR_STATE` an
1064 * Added ERR functionality to give callers access to the stored function
1065 names that have replaced the older function code based functions.
1067 New functions are ERR_peek_error_func(), ERR_peek_last_error_func(),
1068 ERR_peek_error_data(), ERR_peek_last_error_data(), ERR_get_error_all(),
1069 ERR_peek_error_all() and ERR_peek_last_error_all().
1071 Deprecate ERR functions ERR_get_error_line(), ERR_get_error_line_data(),
1072 ERR_peek_error_line_data(), ERR_peek_last_error_line_data() and
1073 ERR_func_error_string().
1077 * Extended testing to be verbose for failing tests only. The make variables
1078 VERBOSE_FAILURE or VF can be used to enable this:
1080 $ make VF=1 test # Unix
1081 $ mms /macro=(VF=1) test ! OpenVMS
1082 $ nmake VF=1 test # Windows
1086 * Added the `-copy_extensions` option to the `x509` command for use with
1087 `-req` and `-x509toreq`. When given with the `copy` or `copyall` argument,
1088 all extensions in the request are copied to the certificate or vice versa.
1090 *David von Oheimb*, *Kirill Stefanenkov <kirill_stefanenkov@rambler.ru>*
1092 * Added the `-copy_extensions` option to the `req` command for use with
1093 `-x509`. When given with the `copy` or `copyall` argument,
1094 all extensions in the certification request are copied to the certificate.
1098 * The `x509`, `req`, and `ca` commands now make sure that X.509v3 certificates
1099 they generate are by default RFC 5280 compliant in the following sense:
1100 There is a subjectKeyIdentifier extension with a hash value of the public key
1101 and for not self-signed certs there is an authorityKeyIdentifier extension
1102 with a keyIdentifier field or issuer information identifying the signing key.
1103 This is done unless some configuration overrides the new default behavior,
1104 such as `subjectKeyIdentifier = none` and `authorityKeyIdentifier = none`.
1108 * Added several checks to `X509_verify_cert()` according to requirements in
1109 RFC 5280 in case `X509_V_FLAG_X509_STRICT` is set
1110 (which may be done by using the CLI option `-x509_strict`):
1111 * The basicConstraints of CA certificates must be marked critical.
1112 * CA certificates must explicitly include the keyUsage extension.
1113 * If a pathlenConstraint is given the key usage keyCertSign must be allowed.
1114 * The issuer name of any certificate must not be empty.
1115 * The subject name of CA certs, certs with keyUsage crlSign,
1116 and certs without subjectAlternativeName must not be empty.
1117 * If a subjectAlternativeName extension is given it must not be empty.
1118 * The signatureAlgorithm field and the cert signature must be consistent.
1119 * Any given authorityKeyIdentifier and any given subjectKeyIdentifier
1120 must not be marked critical.
1121 * The authorityKeyIdentifier must be given for X.509v3 certs
1122 unless they are self-signed.
1123 * The subjectKeyIdentifier must be given for all X.509v3 CA certs.
1127 * Certificate verification using `X509_verify_cert()` meanwhile rejects EC keys
1128 with explicit curve parameters (specifiedCurve) as required by RFC 5480.
1132 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
1133 used even when parsing explicit parameters, when loading a encoded key
1134 or calling `EC_GROUP_new_from_ecpkparameters()`/
1135 `EC_GROUP_new_from_ecparameters()`.
1136 This prevents bypass of security hardening and performance gains,
1137 especially for curves with specialized EC_METHODs.
1138 By default, if a key encoded with explicit parameters is loaded and later
1139 encoded, the output is still encoded with explicit parameters, even if
1140 internally a "named" EC_GROUP is used for computation.
1144 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
1145 this change, EC_GROUP_set_generator would accept order and/or cofactor as
1146 NULL. After this change, only the cofactor parameter can be NULL. It also
1147 does some minimal sanity checks on the passed order.
1152 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
1153 An attack is simple, if the first CMS_recipientInfo is valid but the
1154 second CMS_recipientInfo is chosen ciphertext. If the second
1155 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
1156 encryption key will be replaced by garbage, and the message cannot be
1157 decoded, but if the RSA decryption fails, the correct encryption key is
1158 used and the recipient will not notice the attack.
1159 As a work around for this potential attack the length of the decrypted
1160 key must be equal to the cipher default key length, in case the
1161 certifiate is not given and all recipientInfo are tried out.
1162 The old behaviour can be re-enabled in the CMS code by setting the
1163 CMS_DEBUG_DECRYPT flag.
1167 * Early start up entropy quality from the DEVRANDOM seed source has been
1168 improved for older Linux systems. The RAND subsystem will wait for
1169 /dev/random to be producing output before seeding from /dev/urandom.
1170 The seeded state is stored for future library initialisations using
1171 a system global shared memory segment. The shared memory identifier
1172 can be configured by defining OPENSSL_RAND_SEED_DEVRANDOM_SHM_ID to
1173 the desired value. The default identifier is 114.
1177 * Revised BN_generate_prime_ex to not avoid factors 2..17863 in p-1
1178 when primes for RSA keys are computed.
1179 Since we previously always generated primes == 2 (mod 3) for RSA keys,
1180 the 2-prime and 3-prime RSA modules were easy to distinguish, since
1181 `N = p*q = 1 (mod 3)`, but `N = p*q*r = 2 (mod 3)`. Therefore fingerprinting
1182 2-prime vs. 3-prime RSA keys was possible by computing N mod 3.
1183 This avoids possible fingerprinting of newly generated RSA modules.
1187 * Correct the extended master secret constant on EBCDIC systems. Without this
1188 fix TLS connections between an EBCDIC system and a non-EBCDIC system that
1189 negotiate EMS will fail. Unfortunately this also means that TLS connections
1190 between EBCDIC systems with this fix, and EBCDIC systems without this
1191 fix will fail if they negotiate EMS.
1195 * Changed the library initialisation so that the config file is now loaded
1196 by default. This was already the case for libssl. It now occurs for both
1197 libcrypto and libssl. Use the OPENSSL_INIT_NO_LOAD_CONFIG option to
1198 `OPENSSL_init_crypto()` to suppress automatic loading of a config file.
1202 * Introduced new error raising macros, `ERR_raise()` and `ERR_raise_data()`,
1203 where the former acts as a replacement for `ERR_put_error()`, and the
1204 latter replaces the combination `ERR_put_error()` + `ERR_add_error_data()`.
1205 `ERR_raise_data()` adds more flexibility by taking a format string and
1206 an arbitrary number of arguments following it, to be processed with
1211 * Introduced a new function, `OSSL_PROVIDER_available()`, which can be used
1212 to check if a named provider is loaded and available. When called, it
1213 will also activate all fallback providers if such are still present.
1217 * Enforce a minimum DH modulus size of 512 bits.
1221 * Changed DH parameters to generate the order q subgroup instead of 2q.
1222 Previously generated DH parameters are still accepted by DH_check
1223 but DH_generate_key works around that by clearing bit 0 of the
1224 private key for those. This avoids leaking bit 0 of the private key.
1228 * Significantly reduce secure memory usage by the randomness pools.
1232 * `{CRYPTO,OPENSSL}_mem_debug_{push,pop}` are now no-ops and have been
1237 * A new type, EVP_KEYEXCH, has been introduced to represent key exchange
1238 algorithms. An implementation of a key exchange algorithm can be obtained
1239 by using the function EVP_KEYEXCH_fetch(). An EVP_KEYEXCH algorithm can be
1240 used in a call to EVP_PKEY_derive_init_ex() which works in a similar way to
1241 the older EVP_PKEY_derive_init() function. See the man pages for the new
1242 functions for further details.
1246 * The EVP_PKEY_CTX_set_dh_pad() macro has now been converted to a function.
1250 * Removed the function names from error messages and deprecated the
1255 * Removed NextStep support and the macro OPENSSL_UNISTD
1259 * Removed DES_check_key. Also removed OPENSSL_IMPLEMENT_GLOBAL,
1260 OPENSSL_GLOBAL_REF, OPENSSL_DECLARE_GLOBAL.
1261 Also removed "export var as function" capability; we do not export
1262 variables, only functions.
1266 * RC5_32_set_key has been changed to return an int type, with 0 indicating
1267 an error and 1 indicating success. In previous versions of OpenSSL this
1268 was a void type. If a key was set longer than the maximum possible this
1273 * Support SM2 signing and verification schemes with X509 certificate.
1277 * Use SHA256 as the default digest for TS query in the `ts` app.
1281 * Change PBKDF2 to conform to SP800-132 instead of the older PKCS5 RFC2898.
1285 * Default cipher lists/suites are now available via a function, the
1286 #defines are deprecated.
1290 * Add target VC-WIN32-UWP, VC-WIN64A-UWP, VC-WIN32-ARM-UWP and
1291 VC-WIN64-ARM-UWP in Windows OneCore target for making building libraries
1292 for Windows Store apps easier. Also, the "no-uplink" option has been added.
1296 * Join the directories crypto/x509 and crypto/x509v3
1300 * Added command 'openssl kdf' that uses the EVP_KDF API.
1304 * Added command 'openssl mac' that uses the EVP_MAC API.
1308 * Added OPENSSL_info() to get diverse built-in OpenSSL data, such
1309 as default directories. Also added the command 'openssl info'
1310 for scripting purposes.
1314 * The functions AES_ige_encrypt() and AES_bi_ige_encrypt() have been
1319 * Add prediction resistance to the DRBG reseeding process.
1323 * Limit the number of blocks in a data unit for AES-XTS to 2^20 as
1324 mandated by IEEE Std 1619-2018.
1328 * Added newline escaping functionality to a filename when using openssl dgst.
1329 This output format is to replicate the output format found in the `*sum`
1330 checksum programs. This aims to preserve backward compatibility.
1332 *Matt Eaton, Richard Levitte, and Paul Dale*
1334 * Removed the heartbeat message in DTLS feature, as it has very
1335 little usage and doesn't seem to fulfill a valuable purpose.
1336 The configuration option is now deprecated.
1340 * Changed the output of 'openssl {digestname} < file' to display the
1341 digest name in its output.
1345 * Added a new generic trace API which provides support for enabling
1346 instrumentation through trace output.
1348 *Richard Levitte & Matthias St. Pierre*
1350 * Added build tests for C++. These are generated files that only do one
1351 thing, to include one public OpenSSL head file each. This tests that
1352 the public header files can be usefully included in a C++ application.
1354 This test isn't enabled by default. It can be enabled with the option
1355 'enable-buildtest-c++'.
1359 * Added KB KDF (EVP_KDF_KB) to EVP_KDF.
1363 * Added SSH KDF (EVP_KDF_SSHKDF) and KRB5 KDF (EVP_KDF_KRB5KDF) to EVP_KDF.
1367 * Added Single Step KDF (EVP_KDF_SS), X963 KDF, and X942 KDF to EVP_KDF.
1371 * Added KMAC to EVP_MAC.
1375 * Added property based algorithm implementation selection framework to
1380 * Added SCA hardening for modular field inversion in EC_GROUP through
1381 a new dedicated field_inv() pointer in EC_METHOD.
1382 This also addresses a leakage affecting conversions from projective
1383 to affine coordinates.
1385 *Billy Bob Brumley, Nicola Tuveri*
1387 * Added EVP_KDF, an EVP layer KDF API, to simplify adding KDF and PRF
1388 implementations. This includes an EVP_PKEY to EVP_KDF bridge for
1389 those algorithms that were already supported through the EVP_PKEY API
1390 (scrypt, TLS1 PRF and HKDF). The low-level KDF functions for PBKDF2
1391 and scrypt are now wrappers that call EVP_KDF.
1395 * Build devcrypto engine as a dynamic engine.
1397 *Eneas U de Queiroz*
1399 * Add keyed BLAKE2 to EVP_MAC.
1403 * Fix a bug in the computation of the endpoint-pair shared secret used
1404 by DTLS over SCTP. This breaks interoperability with older versions
1405 of OpenSSL like OpenSSL 1.1.0 and OpenSSL 1.0.2. There is a runtime
1406 switch SSL_MODE_DTLS_SCTP_LABEL_LENGTH_BUG (off by default) enabling
1407 interoperability with such broken implementations. However, enabling
1408 this switch breaks interoperability with correct implementations.
1410 * Fix a use after free bug in d2i_X509_PUBKEY when overwriting a
1411 re-used X509_PUBKEY object if the second PUBKEY is malformed.
1415 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
1419 * Changed the license to the Apache License v2.0.
1423 * Switch to a new version scheme using three numbers MAJOR.MINOR.PATCH.
1425 - Major releases (indicated by incrementing the MAJOR release number)
1426 may introduce incompatible API/ABI changes.
1427 - Minor releases (indicated by incrementing the MINOR release number)
1428 may introduce new features but retain API/ABI compatibility.
1429 - Patch releases (indicated by incrementing the PATCH number)
1430 are intended for bug fixes and other improvements of existing
1431 features only (like improving performance or adding documentation)
1432 and retain API/ABI compatibility.
1436 * Add support for RFC5297 SIV mode (siv128), including AES-SIV.
1440 * Remove the 'dist' target and add a tarball building script. The
1441 'dist' target has fallen out of use, and it shouldn't be
1442 necessary to configure just to create a source distribution.
1446 * Recreate the OS390-Unix config target. It no longer relies on a
1447 special script like it did for OpenSSL pre-1.1.0.
1451 * Instead of having the source directories listed in Configure, add
1452 a 'build.info' keyword SUBDIRS to indicate what sub-directories to
1457 * Add GMAC to EVP_MAC.
1461 * Ported the HMAC, CMAC and SipHash EVP_PKEY_METHODs to EVP_MAC.
1465 * Added EVP_MAC, an EVP layer MAC API, to simplify adding MAC
1466 implementations. This includes a generic EVP_PKEY to EVP_MAC bridge,
1467 to facilitate the continued use of MACs through raw private keys in
1468 functionality such as `EVP_DigestSign*` and `EVP_DigestVerify*`.
1472 * Deprecate ECDH_KDF_X9_62().
1476 * Added EVP_PKEY_ECDH_KDF_X9_63 and ecdh_KDF_X9_63() as replacements for
1477 the EVP_PKEY_ECDH_KDF_X9_62 KDF type and ECDH_KDF_X9_62(). The old names
1478 are retained for backwards compatibility.
1482 * AES-XTS mode now enforces that its two keys are different to mitigate
1483 the attacked described in "Efficient Instantiations of Tweakable
1484 Blockciphers and Refinements to Modes OCB and PMAC" by Phillip Rogaway.
1485 Details of this attack can be obtained from:
1486 <http://web.cs.ucdavis.edu/%7Erogaway/papers/offsets.pdf>
1490 * Rename the object files, i.e. give them other names than in previous
1491 versions. Their names now include the name of the final product, as
1492 well as its type mnemonic (bin, lib, shlib).
1496 * Added new option for 'openssl list', '-objects', which will display the
1497 list of built in objects, i.e. OIDs with names.
1501 * Added the options `-crl_lastupdate` and `-crl_nextupdate` to `openssl ca`,
1502 allowing the `lastUpdate` and `nextUpdate` fields in the generated CRL to
1507 * Added support for Linux Kernel TLS data-path. The Linux Kernel data-path
1508 improves application performance by removing data copies and providing
1509 applications with zero-copy system calls such as sendfile and splice.
1513 * The SSL option SSL_OP_CLEANSE_PLAINTEXT is introduced.
1517 * `PKCS12_parse` now maintains the order of the parsed certificates
1518 when outputting them via `*ca` (rather than reversing it).
1522 * Deprecated pthread fork support methods.
1526 * Added support for FFDHE key exchange in TLS 1.3.
1530 * Added a new concept for OpenSSL plugability: providers. This
1531 functionality is designed to replace the ENGINE API and ENGINE
1532 implementations, and to be much more dynamic, allowing provider
1533 authors to introduce new algorithms among other things, as long as
1534 there's an API that supports the algorithm type.
1536 With this concept comes a new core API for interaction between
1537 libcrypto and provider implementations. Public libcrypto functions
1538 that want to use providers do so through this core API.
1540 The main documentation for this core API is found in
1541 doc/man7/provider.pod, doc/man7/provider-base.pod, and they in turn
1542 refer to other manuals describing the API specific for supported
1543 algorithm types (also called operations).
1550 ### Changes between 1.1.1l and 1.1.1m [xx XXX xxxx]
1552 * Avoid loading of a dynamic engine twice.
1556 * Prioritise DANE TLSA issuer certs over peer certs
1560 * Fixed random API for MacOS prior to 10.12
1562 These MacOS versions don't support the CommonCrypto APIs
1566 ### Changes between 1.1.1k and 1.1.1l [24 Aug 2021]
1568 * Fixed an SM2 Decryption Buffer Overflow.
1570 In order to decrypt SM2 encrypted data an application is expected to
1571 call the API function EVP_PKEY_decrypt(). Typically an application will
1572 call this function twice. The first time, on entry, the "out" parameter
1573 can be NULL and, on exit, the "outlen" parameter is populated with the
1574 buffer size required to hold the decrypted plaintext. The application
1575 can then allocate a sufficiently sized buffer and call EVP_PKEY_decrypt()
1576 again, but this time passing a non-NULL value for the "out" parameter.
1578 A bug in the implementation of the SM2 decryption code means that the
1579 calculation of the buffer size required to hold the plaintext returned
1580 by the first call to EVP_PKEY_decrypt() can be smaller than the actual
1581 size required by the second call. This can lead to a buffer overflow
1582 when EVP_PKEY_decrypt() is called by the application a second time with
1583 a buffer that is too small.
1585 A malicious attacker who is able present SM2 content for decryption to
1586 an application could cause attacker chosen data to overflow the buffer
1587 by up to a maximum of 62 bytes altering the contents of other data held
1588 after the buffer, possibly changing application behaviour or causing
1589 the application to crash. The location of the buffer is application
1590 dependent but is typically heap allocated.
1595 * Fixed various read buffer overruns processing ASN.1 strings
1597 ASN.1 strings are represented internally within OpenSSL as an ASN1_STRING
1598 structure which contains a buffer holding the string data and a field
1599 holding the buffer length. This contrasts with normal C strings which
1600 are repesented as a buffer for the string data which is terminated
1601 with a NUL (0) byte.
1603 Although not a strict requirement, ASN.1 strings that are parsed using
1604 OpenSSL's own "d2i" functions (and other similar parsing functions) as
1605 well as any string whose value has been set with the ASN1_STRING_set()
1606 function will additionally NUL terminate the byte array in the
1607 ASN1_STRING structure.
1609 However, it is possible for applications to directly construct valid
1610 ASN1_STRING structures which do not NUL terminate the byte array by
1611 directly setting the "data" and "length" fields in the ASN1_STRING
1612 array. This can also happen by using the ASN1_STRING_set0() function.
1614 Numerous OpenSSL functions that print ASN.1 data have been found to
1615 assume that the ASN1_STRING byte array will be NUL terminated, even
1616 though this is not guaranteed for strings that have been directly
1617 constructed. Where an application requests an ASN.1 structure to be
1618 printed, and where that ASN.1 structure contains ASN1_STRINGs that have
1619 been directly constructed by the application without NUL terminating
1620 the "data" field, then a read buffer overrun can occur.
1622 The same thing can also occur during name constraints processing
1623 of certificates (for example if a certificate has been directly
1624 constructed by the application instead of loading it via the OpenSSL
1625 parsing functions, and the certificate contains non NUL terminated
1626 ASN1_STRING structures). It can also occur in the X509_get1_email(),
1627 X509_REQ_get1_email() and X509_get1_ocsp() functions.
1629 If a malicious actor can cause an application to directly construct an
1630 ASN1_STRING and then process it through one of the affected OpenSSL
1631 functions then this issue could be hit. This might result in a crash
1632 (causing a Denial of Service attack). It could also result in the
1633 disclosure of private memory contents (such as private keys, or
1634 sensitive plaintext).
1639 ### Changes between 1.1.1j and 1.1.1k [25 Mar 2021]
1641 * Fixed a problem with verifying a certificate chain when using the
1642 X509_V_FLAG_X509_STRICT flag. This flag enables additional security checks of
1643 the certificates present in a certificate chain. It is not set by default.
1645 Starting from OpenSSL version 1.1.1h a check to disallow certificates in
1646 the chain that have explicitly encoded elliptic curve parameters was added
1647 as an additional strict check.
1649 An error in the implementation of this check meant that the result of a
1650 previous check to confirm that certificates in the chain are valid CA
1651 certificates was overwritten. This effectively bypasses the check
1652 that non-CA certificates must not be able to issue other certificates.
1654 If a "purpose" has been configured then there is a subsequent opportunity
1655 for checks that the certificate is a valid CA. All of the named "purpose"
1656 values implemented in libcrypto perform this check. Therefore, where
1657 a purpose is set the certificate chain will still be rejected even when the
1658 strict flag has been used. A purpose is set by default in libssl client and
1659 server certificate verification routines, but it can be overridden or
1660 removed by an application.
1662 In order to be affected, an application must explicitly set the
1663 X509_V_FLAG_X509_STRICT verification flag and either not set a purpose
1664 for the certificate verification or, in the case of TLS client or server
1665 applications, override the default purpose.
1670 * Fixed an issue where an OpenSSL TLS server may crash if sent a maliciously
1671 crafted renegotiation ClientHello message from a client. If a TLSv1.2
1672 renegotiation ClientHello omits the signature_algorithms extension (where it
1673 was present in the initial ClientHello), but includes a
1674 signature_algorithms_cert extension then a NULL pointer dereference will
1675 result, leading to a crash and a denial of service attack.
1677 A server is only vulnerable if it has TLSv1.2 and renegotiation enabled
1678 (which is the default configuration). OpenSSL TLS clients are not impacted by
1682 *Peter Kästle and Samuel Sapalski*
1684 ### Changes between 1.1.1i and 1.1.1j [16 Feb 2021]
1686 * Fixed the X509_issuer_and_serial_hash() function. It attempts to
1687 create a unique hash value based on the issuer and serial number data
1688 contained within an X509 certificate. However it was failing to correctly
1689 handle any errors that may occur while parsing the issuer field (which might
1690 occur if the issuer field is maliciously constructed). This may subsequently
1691 result in a NULL pointer deref and a crash leading to a potential denial of
1697 * Fixed the RSA_padding_check_SSLv23() function and the RSA_SSLV23_PADDING
1698 padding mode to correctly check for rollback attacks. This is considered a
1699 bug in OpenSSL 1.1.1 because it does not support SSLv2. In 1.0.2 this is
1704 Fixed the EVP_CipherUpdate, EVP_EncryptUpdate and EVP_DecryptUpdate
1705 functions. Previously they could overflow the output length argument in some
1706 cases where the input length is close to the maximum permissable length for
1707 an integer on the platform. In such cases the return value from the function
1708 call would be 1 (indicating success), but the output length value would be
1709 negative. This could cause applications to behave incorrectly or crash.
1714 * Fixed SRP_Calc_client_key so that it runs in constant time. The previous
1715 implementation called BN_mod_exp without setting BN_FLG_CONSTTIME. This
1716 could be exploited in a side channel attack to recover the password. Since
1717 the attack is local host only this is outside of the current OpenSSL
1718 threat model and therefore no CVE is assigned.
1720 Thanks to Mohammed Sabt and Daniel De Almeida Braga for reporting this
1725 ### Changes between 1.1.1h and 1.1.1i [8 Dec 2020]
1727 * Fixed NULL pointer deref in the GENERAL_NAME_cmp function
1728 This function could crash if both GENERAL_NAMEs contain an EDIPARTYNAME.
1729 If an attacker can control both items being compared then this could lead
1730 to a possible denial of service attack. OpenSSL itself uses the
1731 GENERAL_NAME_cmp function for two purposes:
1732 1) Comparing CRL distribution point names between an available CRL and a
1733 CRL distribution point embedded in an X509 certificate
1734 2) When verifying that a timestamp response token signer matches the
1735 timestamp authority name (exposed via the API functions
1736 TS_RESP_verify_response and TS_RESP_verify_token)
1741 ### Changes between 1.1.1g and 1.1.1h [22 Sep 2020]
1743 * Certificates with explicit curve parameters are now disallowed in
1744 verification chains if the X509_V_FLAG_X509_STRICT flag is used.
1748 * The 'MinProtocol' and 'MaxProtocol' configuration commands now silently
1749 ignore TLS protocol version bounds when configuring DTLS-based contexts, and
1750 conversely, silently ignore DTLS protocol version bounds when configuring
1751 TLS-based contexts. The commands can be repeated to set bounds of both
1752 types. The same applies with the corresponding "min_protocol" and
1753 "max_protocol" command-line switches, in case some application uses both TLS
1756 SSL_CTX instances that are created for a fixed protocol version (e.g.
1757 TLSv1_server_method()) also silently ignore version bounds. Previously
1758 attempts to apply bounds to these protocol versions would result in an
1759 error. Now only the "version-flexible" SSL_CTX instances are subject to
1760 limits in configuration files in command-line options.
1764 * Handshake now fails if Extended Master Secret extension is dropped
1769 * The Oracle Developer Studio compiler will start reporting deprecated APIs
1771 ### Changes between 1.1.1f and 1.1.1g [21 Apr 2020]
1773 * Fixed segmentation fault in SSL_check_chain()
1774 Server or client applications that call the SSL_check_chain() function
1775 during or after a TLS 1.3 handshake may crash due to a NULL pointer
1776 dereference as a result of incorrect handling of the
1777 "signature_algorithms_cert" TLS extension. The crash occurs if an invalid
1778 or unrecognised signature algorithm is received from the peer. This could
1779 be exploited by a malicious peer in a Denial of Service attack.
1784 * Added AES consttime code for no-asm configurations
1785 an optional constant time support for AES was added
1786 when building openssl for no-asm.
1787 Enable with: ./config no-asm -DOPENSSL_AES_CONST_TIME
1788 Disable with: ./config no-asm -DOPENSSL_NO_AES_CONST_TIME
1789 At this time this feature is by default disabled.
1790 It will be enabled by default in 3.0.
1794 ### Changes between 1.1.1e and 1.1.1f [31 Mar 2020]
1796 * Revert the change of EOF detection while reading in libssl to avoid
1797 regressions in applications depending on the current way of reporting
1798 the EOF. As the existing method is not fully accurate the change to
1799 reporting the EOF via SSL_ERROR_SSL is kept on the current development
1800 branch and will be present in the 3.0 release.
1804 * Revised BN_generate_prime_ex to not avoid factors 3..17863 in p-1
1805 when primes for RSA keys are computed.
1806 Since we previously always generated primes == 2 (mod 3) for RSA keys,
1807 the 2-prime and 3-prime RSA modules were easy to distinguish, since
1808 N = p*q = 1 (mod 3), but N = p*q*r = 2 (mod 3). Therefore fingerprinting
1809 2-prime vs. 3-prime RSA keys was possible by computing N mod 3.
1810 This avoids possible fingerprinting of newly generated RSA modules.
1814 ### Changes between 1.1.1d and 1.1.1e [17 Mar 2020]
1816 * Properly detect EOF while reading in libssl. Previously if we hit an EOF
1817 while reading in libssl then we would report an error back to the
1818 application (SSL_ERROR_SYSCALL) but errno would be 0. We now add
1819 an error to the stack (which means we instead return SSL_ERROR_SSL) and
1820 therefore give a hint as to what went wrong.
1824 * Check that ed25519 and ed448 are allowed by the security level. Previously
1825 signature algorithms not using an MD were not being checked that they were
1826 allowed by the security level.
1830 * Fixed SSL_get_servername() behaviour. The behaviour of SSL_get_servername()
1831 was not quite right. The behaviour was not consistent between resumption
1832 and normal handshakes, and also not quite consistent with historical
1833 behaviour. The behaviour in various scenarios has been clarified and
1834 it has been updated to make it match historical behaviour as closely as
1839 * *[VMS only]* The header files that the VMS compilers include automatically,
1840 `__DECC_INCLUDE_PROLOGUE.H` and `__DECC_INCLUDE_EPILOGUE.H`, use pragmas
1841 that the C++ compiler doesn't understand. This is a shortcoming in the
1842 compiler, but can be worked around with `__cplusplus` guards.
1844 C++ applications that use OpenSSL libraries must be compiled using the
1845 qualifier `/NAMES=(AS_IS,SHORTENED)` to be able to use all the OpenSSL
1846 functions. Otherwise, only functions with symbols of less than 31
1847 characters can be used, as the linker will not be able to successfully
1848 resolve symbols with longer names.
1852 * Added a new method to gather entropy on VMS, based on SYS$GET_ENTROPY.
1853 The presence of this system service is determined at run-time.
1857 * Print all values for a PKCS#12 attribute with 'openssl pkcs12', not just
1862 ### Changes between 1.1.1c and 1.1.1d [10 Sep 2019]
1864 * Fixed a fork protection issue. OpenSSL 1.1.1 introduced a rewritten random
1865 number generator (RNG). This was intended to include protection in the
1866 event of a fork() system call in order to ensure that the parent and child
1867 processes did not share the same RNG state. However this protection was not
1868 being used in the default case.
1870 A partial mitigation for this issue is that the output from a high
1871 precision timer is mixed into the RNG state so the likelihood of a parent
1872 and child process sharing state is significantly reduced.
1874 If an application already calls OPENSSL_init_crypto() explicitly using
1875 OPENSSL_INIT_ATFORK then this problem does not occur at all.
1878 *Matthias St. Pierre*
1880 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
1881 used even when parsing explicit parameters, when loading a encoded key
1882 or calling `EC_GROUP_new_from_ecpkparameters()`/
1883 `EC_GROUP_new_from_ecparameters()`.
1884 This prevents bypass of security hardening and performance gains,
1885 especially for curves with specialized EC_METHODs.
1886 By default, if a key encoded with explicit parameters is loaded and later
1887 encoded, the output is still encoded with explicit parameters, even if
1888 internally a "named" EC_GROUP is used for computation.
1892 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
1893 this change, EC_GROUP_set_generator would accept order and/or cofactor as
1894 NULL. After this change, only the cofactor parameter can be NULL. It also
1895 does some minimal sanity checks on the passed order.
1900 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
1901 An attack is simple, if the first CMS_recipientInfo is valid but the
1902 second CMS_recipientInfo is chosen ciphertext. If the second
1903 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
1904 encryption key will be replaced by garbage, and the message cannot be
1905 decoded, but if the RSA decryption fails, the correct encryption key is
1906 used and the recipient will not notice the attack.
1907 As a work around for this potential attack the length of the decrypted
1908 key must be equal to the cipher default key length, in case the
1909 certifiate is not given and all recipientInfo are tried out.
1910 The old behaviour can be re-enabled in the CMS code by setting the
1911 CMS_DEBUG_DECRYPT flag.
1916 * Early start up entropy quality from the DEVRANDOM seed source has been
1917 improved for older Linux systems. The RAND subsystem will wait for
1918 /dev/random to be producing output before seeding from /dev/urandom.
1919 The seeded state is stored for future library initialisations using
1920 a system global shared memory segment. The shared memory identifier
1921 can be configured by defining OPENSSL_RAND_SEED_DEVRANDOM_SHM_ID to
1922 the desired value. The default identifier is 114.
1926 * Correct the extended master secret constant on EBCDIC systems. Without this
1927 fix TLS connections between an EBCDIC system and a non-EBCDIC system that
1928 negotiate EMS will fail. Unfortunately this also means that TLS connections
1929 between EBCDIC systems with this fix, and EBCDIC systems without this
1930 fix will fail if they negotiate EMS.
1934 * Use Windows installation paths in the mingw builds
1936 Mingw isn't a POSIX environment per se, which means that Windows
1937 paths should be used for installation.
1942 * Changed DH_check to accept parameters with order q and 2q subgroups.
1943 With order 2q subgroups the bit 0 of the private key is not secret
1944 but DH_generate_key works around that by clearing bit 0 of the
1945 private key for those. This avoids leaking bit 0 of the private key.
1949 * Significantly reduce secure memory usage by the randomness pools.
1953 * Revert the DEVRANDOM_WAIT feature for Linux systems
1955 The DEVRANDOM_WAIT feature added a select() call to wait for the
1956 /dev/random device to become readable before reading from the
1957 /dev/urandom device.
1959 It turned out that this change had negative side effects on
1960 performance which were not acceptable. After some discussion it
1961 was decided to revert this feature and leave it up to the OS
1962 resp. the platform maintainer to ensure a proper initialization
1963 during early boot time.
1965 *Matthias St. Pierre*
1967 ### Changes between 1.1.1b and 1.1.1c [28 May 2019]
1969 * Add build tests for C++. These are generated files that only do one
1970 thing, to include one public OpenSSL head file each. This tests that
1971 the public header files can be usefully included in a C++ application.
1973 This test isn't enabled by default. It can be enabled with the option
1974 'enable-buildtest-c++'.
1978 * Enable SHA3 pre-hashing for ECDSA and DSA.
1982 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
1983 This changes the size when using the `genpkey` command when no size is given.
1984 It fixes an omission in earlier changes that changed all RSA, DSA and DH
1985 generation commands to use 2048 bits by default.
1989 * Reorganize the manual pages to consistently have RETURN VALUES,
1990 EXAMPLES, SEE ALSO and HISTORY come in that order, and adjust
1991 util/fix-doc-nits accordingly.
1993 *Paul Yang, Joshua Lock*
1995 * Add the missing accessor EVP_PKEY_get0_engine()
1999 * Have commands like `s_client` and `s_server` output the signature scheme
2000 along with other cipher suite parameters when debugging.
2004 * Make OPENSSL_config() error agnostic again.
2008 * Do the error handling in RSA decryption constant time.
2012 * Prevent over long nonces in ChaCha20-Poly1305.
2014 ChaCha20-Poly1305 is an AEAD cipher, and requires a unique nonce input
2015 for every encryption operation. RFC 7539 specifies that the nonce value
2016 (IV) should be 96 bits (12 bytes). OpenSSL allows a variable nonce length
2017 and front pads the nonce with 0 bytes if it is less than 12
2018 bytes. However it also incorrectly allows a nonce to be set of up to 16
2019 bytes. In this case only the last 12 bytes are significant and any
2020 additional leading bytes are ignored.
2022 It is a requirement of using this cipher that nonce values are
2023 unique. Messages encrypted using a reused nonce value are susceptible to
2024 serious confidentiality and integrity attacks. If an application changes
2025 the default nonce length to be longer than 12 bytes and then makes a
2026 change to the leading bytes of the nonce expecting the new value to be a
2027 new unique nonce then such an application could inadvertently encrypt
2028 messages with a reused nonce.
2030 Additionally the ignored bytes in a long nonce are not covered by the
2031 integrity guarantee of this cipher. Any application that relies on the
2032 integrity of these ignored leading bytes of a long nonce may be further
2033 affected. Any OpenSSL internal use of this cipher, including in SSL/TLS,
2034 is safe because no such use sets such a long nonce value. However user
2035 applications that use this cipher directly and set a non-default nonce
2036 length to be longer than 12 bytes may be vulnerable.
2038 This issue was reported to OpenSSL on 16th of March 2019 by Joran Dirk
2044 * Add DEVRANDOM_WAIT feature for Linux systems
2046 On older Linux systems where the getrandom() system call is not available,
2047 OpenSSL normally uses the /dev/urandom device for seeding its CSPRNG.
2048 Contrary to getrandom(), the /dev/urandom device will not block during
2049 early boot when the kernel CSPRNG has not been seeded yet.
2051 To mitigate this known weakness, use select() to wait for /dev/random to
2052 become readable before reading from /dev/urandom.
2054 * Ensure that SM2 only uses SM3 as digest algorithm
2058 ### Changes between 1.1.1a and 1.1.1b [26 Feb 2019]
2060 * Change the info callback signals for the start and end of a post-handshake
2061 message exchange in TLSv1.3. In 1.1.1/1.1.1a we used SSL_CB_HANDSHAKE_START
2062 and SSL_CB_HANDSHAKE_DONE. Experience has shown that many applications get
2063 confused by this and assume that a TLSv1.2 renegotiation has started. This
2064 can break KeyUpdate handling. Instead we no longer signal the start and end
2065 of a post handshake message exchange (although the messages themselves are
2066 still signalled). This could break some applications that were expecting
2067 the old signals. However without this KeyUpdate is not usable for many
2072 ### Changes between 1.1.1 and 1.1.1a [20 Nov 2018]
2074 * Timing vulnerability in DSA signature generation
2076 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
2077 timing side channel attack. An attacker could use variations in the signing
2078 algorithm to recover the private key.
2080 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
2085 * Timing vulnerability in ECDSA signature generation
2087 The OpenSSL ECDSA signature algorithm has been shown to be vulnerable to a
2088 timing side channel attack. An attacker could use variations in the signing
2089 algorithm to recover the private key.
2091 This issue was reported to OpenSSL on 25th October 2018 by Samuel Weiser.
2096 * Fixed the issue that RAND_add()/RAND_seed() silently discards random input
2097 if its length exceeds 4096 bytes. The limit has been raised to a buffer size
2098 of two gigabytes and the error handling improved.
2100 This issue was reported to OpenSSL by Dr. Falko Strenzke. It has been
2101 categorized as a normal bug, not a security issue, because the DRBG reseeds
2102 automatically and is fully functional even without additional randomness
2103 provided by the application.
2105 ### Changes between 1.1.0i and 1.1.1 [11 Sep 2018]
2107 * Add a new ClientHello callback. Provides a callback interface that gives
2108 the application the ability to adjust the nascent SSL object at the
2109 earliest stage of ClientHello processing, immediately after extensions have
2110 been collected but before they have been processed. In particular, this
2111 callback can adjust the supported TLS versions in response to the contents
2116 * Add SM2 base algorithm support.
2120 * s390x assembly pack: add (improved) hardware-support for the following
2121 cryptographic primitives: sha3, shake, aes-gcm, aes-ccm, aes-ctr, aes-ofb,
2122 aes-cfb/cfb8, aes-ecb.
2126 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
2127 parameter is no longer accepted, as it leads to a corrupt table. NULL
2128 pem_str is reserved for alias entries only.
2132 * Use the new ec_scalar_mul_ladder scaffold to implement a specialized ladder
2133 step for prime curves. The new implementation is based on formulae from
2134 differential addition-and-doubling in homogeneous projective coordinates
2135 from Izu-Takagi "A fast parallel elliptic curve multiplication resistant
2136 against side channel attacks" and Brier-Joye "Weierstrass Elliptic Curves
2137 and Side-Channel Attacks" Eq. (8) for y-coordinate recovery, modified
2138 to work in projective coordinates.
2140 *Billy Bob Brumley, Nicola Tuveri*
2142 * Change generating and checking of primes so that the error rate of not
2143 being prime depends on the intended use based on the size of the input.
2144 For larger primes this will result in more rounds of Miller-Rabin.
2145 The maximal error rate for primes with more than 1080 bits is lowered
2148 *Kurt Roeckx, Annie Yousar*
2150 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
2154 * The 'tsget' script is renamed to 'tsget.pl', to avoid confusion when
2155 moving between systems, and to avoid confusion when a Windows build is
2156 done with mingw vs with MSVC. For POSIX installs, there's still a
2157 symlink or copy named 'tsget' to avoid that confusion as well.
2161 * Revert blinding in ECDSA sign and instead make problematic addition
2162 length-invariant. Switch even to fixed-length Montgomery multiplication.
2166 * Use the new ec_scalar_mul_ladder scaffold to implement a specialized ladder
2167 step for binary curves. The new implementation is based on formulae from
2168 differential addition-and-doubling in mixed Lopez-Dahab projective
2169 coordinates, modified to independently blind the operands.
2171 *Billy Bob Brumley, Sohaib ul Hassan, Nicola Tuveri*
2173 * Add a scaffold to optionally enhance the Montgomery ladder implementation
2174 for `ec_scalar_mul_ladder` (formerly `ec_mul_consttime`) allowing
2175 EC_METHODs to implement their own specialized "ladder step", to take
2176 advantage of more favorable coordinate systems or more efficient
2177 differential addition-and-doubling algorithms.
2179 *Billy Bob Brumley, Sohaib ul Hassan, Nicola Tuveri*
2181 * Modified the random device based seed sources to keep the relevant
2182 file descriptors open rather than reopening them on each access.
2183 This allows such sources to operate in a chroot() jail without
2184 the associated device nodes being available. This behaviour can be
2185 controlled using RAND_keep_random_devices_open().
2189 * Numerous side-channel attack mitigations have been applied. This may have
2190 performance impacts for some algorithms for the benefit of improved
2191 security. Specific changes are noted in this change log by their respective
2196 * AIX shared library support overhaul. Switch to AIX "natural" way of
2197 handling shared libraries, which means collecting shared objects of
2198 different versions and bitnesses in one common archive. This allows to
2199 mitigate conflict between 1.0 and 1.1 side-by-side installations. It
2200 doesn't affect the way 3rd party applications are linked, only how
2201 multi-version installation is managed.
2205 * Make ec_group_do_inverse_ord() more robust and available to other
2206 EC cryptosystems, so that irrespective of BN_FLG_CONSTTIME, SCA
2207 mitigations are applied to the fallback BN_mod_inverse().
2208 When using this function rather than BN_mod_inverse() directly, new
2209 EC cryptosystem implementations are then safer-by-default.
2213 * Add coordinate blinding for EC_POINT and implement projective
2214 coordinate blinding for generic prime curves as a countermeasure to
2215 chosen point SCA attacks.
2217 *Sohaib ul Hassan, Nicola Tuveri, Billy Bob Brumley*
2219 * Add blinding to ECDSA and DSA signatures to protect against side channel
2220 attacks discovered by Keegan Ryan (NCC Group).
2224 * Enforce checking in the `pkeyutl` command to ensure that the input
2225 length does not exceed the maximum supported digest length when performing
2226 a sign, verify or verifyrecover operation.
2230 * SSL_MODE_AUTO_RETRY is enabled by default. Applications that use blocking
2231 I/O in combination with something like select() or poll() will hang. This
2232 can be turned off again using SSL_CTX_clear_mode().
2233 Many applications do not properly handle non-application data records, and
2234 TLS 1.3 sends more of such records. Setting SSL_MODE_AUTO_RETRY works
2235 around the problems in those applications, but can also break some.
2236 It's recommended to read the manpages about SSL_read(), SSL_write(),
2237 SSL_get_error(), SSL_shutdown(), SSL_CTX_set_mode() and
2238 SSL_CTX_set_read_ahead() again.
2242 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
2243 now allow empty (zero character) pass phrases.
2247 * Apply blinding to binary field modular inversion and remove patent
2248 pending (OPENSSL_SUN_GF2M_DIV) BN_GF2m_mod_div implementation.
2252 * Deprecate ec2_mult.c and unify scalar multiplication code paths for
2253 binary and prime elliptic curves.
2257 * Remove ECDSA nonce padding: EC_POINT_mul is now responsible for
2258 constant time fixed point multiplication.
2262 * Revise elliptic curve scalar multiplication with timing attack
2263 defenses: ec_wNAF_mul redirects to a constant time implementation
2264 when computing fixed point and variable point multiplication (which
2265 in OpenSSL are mostly used with secret scalars in keygen, sign,
2266 ECDH derive operations).
2267 *Billy Bob Brumley, Nicola Tuveri, Cesar Pereida García,
2270 * Updated CONTRIBUTING
2274 * Updated DRBG / RAND to request nonce and additional low entropy
2275 randomness from the system.
2277 *Matthias St. Pierre*
2279 * Updated 'openssl rehash' to use OpenSSL consistent default.
2283 * Moved the load of the ssl_conf module to libcrypto, which helps
2284 loading engines that libssl uses before libssl is initialised.
2288 * Added EVP_PKEY_sign() and EVP_PKEY_verify() for EdDSA
2292 * Fixed X509_NAME_ENTRY_set to get multi-valued RDNs right in all cases.
2294 *Ingo Schwarze, Rich Salz*
2296 * Added output of accepting IP address and port for 'openssl s_server'
2300 * Added a new API for TLSv1.3 ciphersuites:
2301 SSL_CTX_set_ciphersuites()
2302 SSL_set_ciphersuites()
2306 * Memory allocation failures consistently add an error to the error
2311 * Don't use OPENSSL_ENGINES and OPENSSL_CONF environment values
2312 in libcrypto when run as setuid/setgid.
2316 * Load any config file by default when libssl is used.
2320 * Added new public header file <openssl/rand_drbg.h> and documentation
2321 for the RAND_DRBG API. See manual page RAND_DRBG(7) for an overview.
2323 *Matthias St. Pierre*
2325 * QNX support removed (cannot find contributors to get their approval
2326 for the license change).
2330 * TLSv1.3 replay protection for early data has been implemented. See the
2331 SSL_read_early_data() man page for further details.
2335 * Separated TLSv1.3 ciphersuite configuration out from TLSv1.2 ciphersuite
2336 configuration. TLSv1.3 ciphersuites are not compatible with TLSv1.2 and
2337 below. Similarly TLSv1.2 ciphersuites are not compatible with TLSv1.3.
2338 In order to avoid issues where legacy TLSv1.2 ciphersuite configuration
2339 would otherwise inadvertently disable all TLSv1.3 ciphersuites the
2340 configuration has been separated out. See the ciphers man page or the
2341 SSL_CTX_set_ciphersuites() man page for more information.
2345 * On POSIX (BSD, Linux, ...) systems the ocsp(1) command running
2346 in responder mode now supports the new "-multi" option, which
2347 spawns the specified number of child processes to handle OCSP
2348 requests. The "-timeout" option now also limits the OCSP
2349 responder's patience to wait to receive the full client request
2350 on a newly accepted connection. Child processes are respawned
2351 as needed, and the CA index file is automatically reloaded
2352 when changed. This makes it possible to run the "ocsp" responder
2353 as a long-running service, making the OpenSSL CA somewhat more
2354 feature-complete. In this mode, most diagnostic messages logged
2355 after entering the event loop are logged via syslog(3) rather than
2360 * Added support for X448 and Ed448. Heavily based on original work by
2365 * Extend OSSL_STORE with capabilities to search and to narrow the set of
2366 objects loaded. This adds the functions OSSL_STORE_expect() and
2367 OSSL_STORE_find() as well as needed tools to construct searches and
2368 get the search data out of them.
2372 * Support for TLSv1.3 added. Note that users upgrading from an earlier
2373 version of OpenSSL should review their configuration settings to ensure
2374 that they are still appropriate for TLSv1.3. For further information see:
2375 <https://wiki.openssl.org/index.php/TLS1.3>
2379 * Grand redesign of the OpenSSL random generator
2381 The default RAND method now utilizes an AES-CTR DRBG according to
2382 NIST standard SP 800-90Ar1. The new random generator is essentially
2383 a port of the default random generator from the OpenSSL FIPS 2.0
2384 object module. It is a hybrid deterministic random bit generator
2385 using an AES-CTR bit stream and which seeds and reseeds itself
2386 automatically using trusted system entropy sources.
2388 Some of its new features are:
2389 - Support for multiple DRBG instances with seed chaining.
2390 - The default RAND method makes use of a DRBG.
2391 - There is a public and private DRBG instance.
2392 - The DRBG instances are fork-safe.
2393 - Keep all global DRBG instances on the secure heap if it is enabled.
2394 - The public and private DRBG instance are per thread for lock free
2397 *Paul Dale, Benjamin Kaduk, Kurt Roeckx, Rich Salz, Matthias St. Pierre*
2399 * Changed Configure so it only says what it does and doesn't dump
2400 so much data. Instead, ./configdata.pm should be used as a script
2401 to display all sorts of configuration data.
2405 * Added processing of "make variables" to Configure.
2409 * Added SHA512/224 and SHA512/256 algorithm support.
2413 * The last traces of Netware support, first removed in 1.1.0, have
2418 * Get rid of Makefile.shared, and in the process, make the processing
2419 of certain files (rc.obj, or the .def/.map/.opt files produced from
2420 the ordinal files) more visible and hopefully easier to trace and
2421 debug (or make silent).
2425 * Make it possible to have environment variable assignments as
2426 arguments to config / Configure.
2430 * Add multi-prime RSA (RFC 8017) support.
2434 * Add SM3 implemented according to GB/T 32905-2016
2435 *Jack Lloyd <jack.lloyd@ribose.com>,*
2436 *Ronald Tse <ronald.tse@ribose.com>,*
2437 *Erick Borsboom <erick.borsboom@ribose.com>*
2439 * Add 'Maximum Fragment Length' TLS extension negotiation and support
2440 as documented in RFC6066.
2441 Based on a patch from Tomasz Moń
2443 *Filipe Raimundo da Silva*
2445 * Add SM4 implemented according to GB/T 32907-2016.
2446 *Jack Lloyd <jack.lloyd@ribose.com>,*
2447 *Ronald Tse <ronald.tse@ribose.com>,*
2448 *Erick Borsboom <erick.borsboom@ribose.com>*
2450 * Reimplement -newreq-nodes and ERR_error_string_n; the
2451 original author does not agree with the license change.
2455 * Add ARIA AEAD TLS support.
2459 * Some macro definitions to support VS6 have been removed. Visual
2460 Studio 6 has not worked since 1.1.0
2464 * Add ERR_clear_last_mark(), to allow callers to clear the last mark
2465 without clearing the errors.
2469 * Add "atfork" functions. If building on a system that without
2470 pthreads, see doc/man3/OPENSSL_fork_prepare.pod for application
2471 requirements. The RAND facility now uses/requires this.
2479 * The UI API becomes a permanent and integral part of libcrypto, i.e.
2480 not possible to disable entirely. However, it's still possible to
2481 disable the console reading UI method, UI_OpenSSL() (use UI_null()
2484 To disable, configure with 'no-ui-console'. 'no-ui' is still
2485 possible to use as an alias. Check at compile time with the
2486 macro OPENSSL_NO_UI_CONSOLE. The macro OPENSSL_NO_UI is still
2487 possible to check and is an alias for OPENSSL_NO_UI_CONSOLE.
2491 * Add a STORE module, which implements a uniform and URI based reader of
2492 stores that can contain keys, certificates, CRLs and numerous other
2493 objects. The main API is loosely based on a few stdio functions,
2494 and includes OSSL_STORE_open, OSSL_STORE_load, OSSL_STORE_eof,
2495 OSSL_STORE_error and OSSL_STORE_close.
2496 The implementation uses backends called "loaders" to implement arbitrary
2497 URI schemes. There is one built in "loader" for the 'file' scheme.
2501 * Add devcrypto engine. This has been implemented against cryptodev-linux,
2502 then adjusted to work on FreeBSD 8.4 as well.
2503 Enable by configuring with 'enable-devcryptoeng'. This is done by default
2504 on BSD implementations, as cryptodev.h is assumed to exist on all of them.
2508 * Module names can prefixed with OSSL_ or OPENSSL_. This affects
2509 util/mkerr.pl, which is adapted to allow those prefixes, leading to
2510 error code calls like this:
2512 OSSL_FOOerr(OSSL_FOO_F_SOMETHING, OSSL_FOO_R_WHATEVER);
2514 With this change, we claim the namespaces OSSL and OPENSSL in a manner
2515 that can be encoded in C. For the foreseeable future, this will only
2518 *Richard Levitte and Tim Hudson*
2520 * Removed BSD cryptodev engine.
2524 * Add a build target 'build_all_generated', to build all generated files
2525 and only that. This can be used to prepare everything that requires
2526 things like perl for a system that lacks perl and then move everything
2527 to that system and do the rest of the build there.
2531 * In the UI interface, make it possible to duplicate the user data. This
2532 can be used by engines that need to retain the data for a longer time
2533 than just the call where this user data is passed.
2537 * Ignore the '-named_curve auto' value for compatibility of applications
2540 *Tomáš Mráz <tmraz@fedoraproject.org>*
2542 * Fragmented SSL/TLS alerts are no longer accepted. An alert message is 2
2543 bytes long. In theory it is permissible in SSLv3 - TLSv1.2 to fragment such
2544 alerts across multiple records (some of which could be empty). In practice
2545 it make no sense to send an empty alert record, or to fragment one. TLSv1.3
2546 prohibits this altogether and other libraries (BoringSSL, NSS) do not
2547 support this at all. Supporting it adds significant complexity to the
2548 record layer, and its removal is unlikely to cause interoperability
2553 * Add the ASN.1 types INT32, UINT32, INT64, UINT64 and variants prefixed
2554 with Z. These are meant to replace LONG and ZLONG and to be size safe.
2555 The use of LONG and ZLONG is discouraged and scheduled for deprecation
2560 * Add the 'z' and 'j' modifiers to BIO_printf() et al formatting string,
2561 'z' is to be used for [s]size_t, and 'j' - with [u]int64_t.
2563 *Richard Levitte, Andy Polyakov*
2565 * Add EC_KEY_get0_engine(), which does for EC_KEY what RSA_get0_engine()
2570 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
2571 platform rather than 'mingw'.
2575 * The functions X509_STORE_add_cert and X509_STORE_add_crl return
2576 success if they are asked to add an object which already exists
2577 in the store. This change cascades to other functions which load
2578 certificates and CRLs.
2582 * x86_64 assembly pack: annotate code with DWARF CFI directives to
2583 facilitate stack unwinding even from assembly subroutines.
2587 * Remove VAX C specific definitions of OPENSSL_EXPORT, OPENSSL_EXTERN.
2588 Also remove OPENSSL_GLOBAL entirely, as it became a no-op.
2592 * Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
2593 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
2594 which is the minimum version we support.
2598 * Certificate time validation (X509_cmp_time) enforces stricter
2599 compliance with RFC 5280. Fractional seconds and timezone offsets
2600 are no longer allowed.
2604 * Add support for ARIA
2608 * s_client will now send the Server Name Indication (SNI) extension by
2609 default unless the new "-noservername" option is used. The server name is
2610 based on the host provided to the "-connect" option unless overridden by
2611 using "-servername".
2615 * Add support for SipHash
2619 * OpenSSL now fails if it receives an unrecognised record type in TLS1.0
2620 or TLS1.1. Previously this only happened in SSLv3 and TLS1.2. This is to
2621 prevent issues where no progress is being made and the peer continually
2622 sends unrecognised record types, using up resources processing them.
2626 * 'openssl passwd' can now produce SHA256 and SHA512 based output,
2627 using the algorithm defined in
2628 <https://www.akkadia.org/drepper/SHA-crypt.txt>
2632 * Heartbeat support has been removed; the ABI is changed for now.
2634 *Richard Levitte, Rich Salz*
2636 * Support for SSL_OP_NO_ENCRYPT_THEN_MAC in SSL_CONF_cmd.
2640 * The RSA "null" method, which was partially supported to avoid patent
2641 issues, has been replaced to always returns NULL.
2648 ### Changes between 1.1.0k and 1.1.0l [10 Sep 2019]
2650 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
2651 used even when parsing explicit parameters, when loading a encoded key
2652 or calling `EC_GROUP_new_from_ecpkparameters()`/
2653 `EC_GROUP_new_from_ecparameters()`.
2654 This prevents bypass of security hardening and performance gains,
2655 especially for curves with specialized EC_METHODs.
2656 By default, if a key encoded with explicit parameters is loaded and later
2657 encoded, the output is still encoded with explicit parameters, even if
2658 internally a "named" EC_GROUP is used for computation.
2662 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
2663 this change, EC_GROUP_set_generator would accept order and/or cofactor as
2664 NULL. After this change, only the cofactor parameter can be NULL. It also
2665 does some minimal sanity checks on the passed order.
2670 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
2671 An attack is simple, if the first CMS_recipientInfo is valid but the
2672 second CMS_recipientInfo is chosen ciphertext. If the second
2673 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
2674 encryption key will be replaced by garbage, and the message cannot be
2675 decoded, but if the RSA decryption fails, the correct encryption key is
2676 used and the recipient will not notice the attack.
2677 As a work around for this potential attack the length of the decrypted
2678 key must be equal to the cipher default key length, in case the
2679 certifiate is not given and all recipientInfo are tried out.
2680 The old behaviour can be re-enabled in the CMS code by setting the
2681 CMS_DEBUG_DECRYPT flag.
2686 * Use Windows installation paths in the mingw builds
2688 Mingw isn't a POSIX environment per se, which means that Windows
2689 paths should be used for installation.
2694 ### Changes between 1.1.0j and 1.1.0k [28 May 2019]
2696 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
2697 This changes the size when using the `genpkey` command when no size is given.
2698 It fixes an omission in earlier changes that changed all RSA, DSA and DH
2699 generation commands to use 2048 bits by default.
2703 * Prevent over long nonces in ChaCha20-Poly1305.
2705 ChaCha20-Poly1305 is an AEAD cipher, and requires a unique nonce input
2706 for every encryption operation. RFC 7539 specifies that the nonce value
2707 (IV) should be 96 bits (12 bytes). OpenSSL allows a variable nonce length
2708 and front pads the nonce with 0 bytes if it is less than 12
2709 bytes. However it also incorrectly allows a nonce to be set of up to 16
2710 bytes. In this case only the last 12 bytes are significant and any
2711 additional leading bytes are ignored.
2713 It is a requirement of using this cipher that nonce values are
2714 unique. Messages encrypted using a reused nonce value are susceptible to
2715 serious confidentiality and integrity attacks. If an application changes
2716 the default nonce length to be longer than 12 bytes and then makes a
2717 change to the leading bytes of the nonce expecting the new value to be a
2718 new unique nonce then such an application could inadvertently encrypt
2719 messages with a reused nonce.
2721 Additionally the ignored bytes in a long nonce are not covered by the
2722 integrity guarantee of this cipher. Any application that relies on the
2723 integrity of these ignored leading bytes of a long nonce may be further
2724 affected. Any OpenSSL internal use of this cipher, including in SSL/TLS,
2725 is safe because no such use sets such a long nonce value. However user
2726 applications that use this cipher directly and set a non-default nonce
2727 length to be longer than 12 bytes may be vulnerable.
2729 This issue was reported to OpenSSL on 16th of March 2019 by Joran Dirk
2735 * Added SCA hardening for modular field inversion in EC_GROUP through
2736 a new dedicated field_inv() pointer in EC_METHOD.
2737 This also addresses a leakage affecting conversions from projective
2738 to affine coordinates.
2740 *Billy Bob Brumley, Nicola Tuveri*
2742 * Fix a use after free bug in d2i_X509_PUBKEY when overwriting a
2743 re-used X509_PUBKEY object if the second PUBKEY is malformed.
2747 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
2751 * Remove the 'dist' target and add a tarball building script. The
2752 'dist' target has fallen out of use, and it shouldn't be
2753 necessary to configure just to create a source distribution.
2757 ### Changes between 1.1.0i and 1.1.0j [20 Nov 2018]
2759 * Timing vulnerability in DSA signature generation
2761 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
2762 timing side channel attack. An attacker could use variations in the signing
2763 algorithm to recover the private key.
2765 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
2770 * Timing vulnerability in ECDSA signature generation
2772 The OpenSSL ECDSA signature algorithm has been shown to be vulnerable to a
2773 timing side channel attack. An attacker could use variations in the signing
2774 algorithm to recover the private key.
2776 This issue was reported to OpenSSL on 25th October 2018 by Samuel Weiser.
2781 * Add coordinate blinding for EC_POINT and implement projective
2782 coordinate blinding for generic prime curves as a countermeasure to
2783 chosen point SCA attacks.
2785 *Sohaib ul Hassan, Nicola Tuveri, Billy Bob Brumley*
2787 ### Changes between 1.1.0h and 1.1.0i [14 Aug 2018]
2789 * Client DoS due to large DH parameter
2791 During key agreement in a TLS handshake using a DH(E) based ciphersuite a
2792 malicious server can send a very large prime value to the client. This will
2793 cause the client to spend an unreasonably long period of time generating a
2794 key for this prime resulting in a hang until the client has finished. This
2795 could be exploited in a Denial Of Service attack.
2797 This issue was reported to OpenSSL on 5th June 2018 by Guido Vranken
2802 * Cache timing vulnerability in RSA Key Generation
2804 The OpenSSL RSA Key generation algorithm has been shown to be vulnerable to
2805 a cache timing side channel attack. An attacker with sufficient access to
2806 mount cache timing attacks during the RSA key generation process could
2807 recover the private key.
2809 This issue was reported to OpenSSL on 4th April 2018 by Alejandro Cabrera
2810 Aldaya, Billy Brumley, Cesar Pereida Garcia and Luis Manuel Alvarez Tapia.
2815 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
2816 parameter is no longer accepted, as it leads to a corrupt table. NULL
2817 pem_str is reserved for alias entries only.
2821 * Revert blinding in ECDSA sign and instead make problematic addition
2822 length-invariant. Switch even to fixed-length Montgomery multiplication.
2826 * Change generating and checking of primes so that the error rate of not
2827 being prime depends on the intended use based on the size of the input.
2828 For larger primes this will result in more rounds of Miller-Rabin.
2829 The maximal error rate for primes with more than 1080 bits is lowered
2832 *Kurt Roeckx, Annie Yousar*
2834 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
2838 * Add blinding to ECDSA and DSA signatures to protect against side channel
2839 attacks discovered by Keegan Ryan (NCC Group).
2843 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
2844 now allow empty (zero character) pass phrases.
2848 * Certificate time validation (X509_cmp_time) enforces stricter
2849 compliance with RFC 5280. Fractional seconds and timezone offsets
2850 are no longer allowed.
2854 * Fixed a text canonicalisation bug in CMS
2856 Where a CMS detached signature is used with text content the text goes
2857 through a canonicalisation process first prior to signing or verifying a
2858 signature. This process strips trailing space at the end of lines, converts
2859 line terminators to CRLF and removes additional trailing line terminators
2860 at the end of a file. A bug in the canonicalisation process meant that
2861 some characters, such as form-feed, were incorrectly treated as whitespace
2862 and removed. This is contrary to the specification (RFC5485). This fix
2863 could mean that detached text data signed with an earlier version of
2864 OpenSSL 1.1.0 may fail to verify using the fixed version, or text data
2865 signed with a fixed OpenSSL may fail to verify with an earlier version of
2866 OpenSSL 1.1.0. A workaround is to only verify the canonicalised text data
2867 and use the "-binary" flag (for the "cms" command line application) or set
2868 the SMIME_BINARY/PKCS7_BINARY/CMS_BINARY flags (if using CMS_verify()).
2872 ### Changes between 1.1.0g and 1.1.0h [27 Mar 2018]
2874 * Constructed ASN.1 types with a recursive definition could exceed the stack
2876 Constructed ASN.1 types with a recursive definition (such as can be found
2877 in PKCS7) could eventually exceed the stack given malicious input with
2878 excessive recursion. This could result in a Denial Of Service attack. There
2879 are no such structures used within SSL/TLS that come from untrusted sources
2880 so this is considered safe.
2882 This issue was reported to OpenSSL on 4th January 2018 by the OSS-fuzz
2888 * Incorrect CRYPTO_memcmp on HP-UX PA-RISC
2890 Because of an implementation bug the PA-RISC CRYPTO_memcmp function is
2891 effectively reduced to only comparing the least significant bit of each
2892 byte. This allows an attacker to forge messages that would be considered as
2893 authenticated in an amount of tries lower than that guaranteed by the
2894 security claims of the scheme. The module can only be compiled by the
2895 HP-UX assembler, so that only HP-UX PA-RISC targets are affected.
2897 This issue was reported to OpenSSL on 2nd March 2018 by Peter Waltenberg
2903 * Add a build target 'build_all_generated', to build all generated files
2904 and only that. This can be used to prepare everything that requires
2905 things like perl for a system that lacks perl and then move everything
2906 to that system and do the rest of the build there.
2910 * Backport SSL_OP_NO_RENGOTIATION
2912 OpenSSL 1.0.2 and below had the ability to disable renegotiation using the
2913 (undocumented) SSL3_FLAGS_NO_RENEGOTIATE_CIPHERS flag. Due to the opacity
2914 changes this is no longer possible in 1.1.0. Therefore the new
2915 SSL_OP_NO_RENEGOTIATION option from 1.1.1-dev has been backported to
2916 1.1.0 to provide equivalent functionality.
2918 Note that if an application built against 1.1.0h headers (or above) is run
2919 using an older version of 1.1.0 (prior to 1.1.0h) then the option will be
2920 accepted but nothing will happen, i.e. renegotiation will not be prevented.
2924 * Removed the OS390-Unix config target. It relied on a script that doesn't
2929 * rsaz_1024_mul_avx2 overflow bug on x86_64
2931 There is an overflow bug in the AVX2 Montgomery multiplication procedure
2932 used in exponentiation with 1024-bit moduli. No EC algorithms are affected.
2933 Analysis suggests that attacks against RSA and DSA as a result of this
2934 defect would be very difficult to perform and are not believed likely.
2935 Attacks against DH1024 are considered just feasible, because most of the
2936 work necessary to deduce information about a private key may be performed
2937 offline. The amount of resources required for such an attack would be
2938 significant. However, for an attack on TLS to be meaningful, the server
2939 would have to share the DH1024 private key among multiple clients, which is
2940 no longer an option since CVE-2016-0701.
2942 This only affects processors that support the AVX2 but not ADX extensions
2943 like Intel Haswell (4th generation).
2945 This issue was reported to OpenSSL by David Benjamin (Google). The issue
2946 was originally found via the OSS-Fuzz project.
2951 ### Changes between 1.1.0f and 1.1.0g [2 Nov 2017]
2953 * bn_sqrx8x_internal carry bug on x86_64
2955 There is a carry propagating bug in the x86_64 Montgomery squaring
2956 procedure. No EC algorithms are affected. Analysis suggests that attacks
2957 against RSA and DSA as a result of this defect would be very difficult to
2958 perform and are not believed likely. Attacks against DH are considered just
2959 feasible (although very difficult) because most of the work necessary to
2960 deduce information about a private key may be performed offline. The amount
2961 of resources required for such an attack would be very significant and
2962 likely only accessible to a limited number of attackers. An attacker would
2963 additionally need online access to an unpatched system using the target
2964 private key in a scenario with persistent DH parameters and a private
2965 key that is shared between multiple clients.
2967 This only affects processors that support the BMI1, BMI2 and ADX extensions
2968 like Intel Broadwell (5th generation) and later or AMD Ryzen.
2970 This issue was reported to OpenSSL by the OSS-Fuzz project.
2975 * Malformed X.509 IPAddressFamily could cause OOB read
2977 If an X.509 certificate has a malformed IPAddressFamily extension,
2978 OpenSSL could do a one-byte buffer overread. The most likely result
2979 would be an erroneous display of the certificate in text format.
2981 This issue was reported to OpenSSL by the OSS-Fuzz project.
2986 ### Changes between 1.1.0e and 1.1.0f [25 May 2017]
2988 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
2989 platform rather than 'mingw'.
2993 * Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
2994 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
2995 which is the minimum version we support.
2999 ### Changes between 1.1.0d and 1.1.0e [16 Feb 2017]
3001 * Encrypt-Then-Mac renegotiation crash
3003 During a renegotiation handshake if the Encrypt-Then-Mac extension is
3004 negotiated where it was not in the original handshake (or vice-versa) then
3005 this can cause OpenSSL to crash (dependant on ciphersuite). Both clients
3006 and servers are affected.
3008 This issue was reported to OpenSSL by Joe Orton (Red Hat).
3013 ### Changes between 1.1.0c and 1.1.0d [26 Jan 2017]
3015 * Truncated packet could crash via OOB read
3017 If one side of an SSL/TLS path is running on a 32-bit host and a specific
3018 cipher is being used, then a truncated packet can cause that host to
3019 perform an out-of-bounds read, usually resulting in a crash.
3021 This issue was reported to OpenSSL by Robert Święcki of Google.
3026 * Bad (EC)DHE parameters cause a client crash
3028 If a malicious server supplies bad parameters for a DHE or ECDHE key
3029 exchange then this can result in the client attempting to dereference a
3030 NULL pointer leading to a client crash. This could be exploited in a Denial
3033 This issue was reported to OpenSSL by Guido Vranken.
3038 * BN_mod_exp may produce incorrect results on x86_64
3040 There is a carry propagating bug in the x86_64 Montgomery squaring
3041 procedure. No EC algorithms are affected. Analysis suggests that attacks
3042 against RSA and DSA as a result of this defect would be very difficult to
3043 perform and are not believed likely. Attacks against DH are considered just
3044 feasible (although very difficult) because most of the work necessary to
3045 deduce information about a private key may be performed offline. The amount
3046 of resources required for such an attack would be very significant and
3047 likely only accessible to a limited number of attackers. An attacker would
3048 additionally need online access to an unpatched system using the target
3049 private key in a scenario with persistent DH parameters and a private
3050 key that is shared between multiple clients. For example this can occur by
3051 default in OpenSSL DHE based SSL/TLS ciphersuites. Note: This issue is very
3052 similar to CVE-2015-3193 but must be treated as a separate problem.
3054 This issue was reported to OpenSSL by the OSS-Fuzz project.
3059 ### Changes between 1.1.0b and 1.1.0c [10 Nov 2016]
3061 * ChaCha20/Poly1305 heap-buffer-overflow
3063 TLS connections using `*-CHACHA20-POLY1305` ciphersuites are susceptible to
3064 a DoS attack by corrupting larger payloads. This can result in an OpenSSL
3065 crash. This issue is not considered to be exploitable beyond a DoS.
3067 This issue was reported to OpenSSL by Robert Święcki (Google Security Team)
3072 * CMS Null dereference
3074 Applications parsing invalid CMS structures can crash with a NULL pointer
3075 dereference. This is caused by a bug in the handling of the ASN.1 CHOICE
3076 type in OpenSSL 1.1.0 which can result in a NULL value being passed to the
3077 structure callback if an attempt is made to free certain invalid encodings.
3078 Only CHOICE structures using a callback which do not handle NULL value are
3081 This issue was reported to OpenSSL by Tyler Nighswander of ForAllSecure.
3086 * Montgomery multiplication may produce incorrect results
3088 There is a carry propagating bug in the Broadwell-specific Montgomery
3089 multiplication procedure that handles input lengths divisible by, but
3090 longer than 256 bits. Analysis suggests that attacks against RSA, DSA
3091 and DH private keys are impossible. This is because the subroutine in
3092 question is not used in operations with the private key itself and an input
3093 of the attacker's direct choice. Otherwise the bug can manifest itself as
3094 transient authentication and key negotiation failures or reproducible
3095 erroneous outcome of public-key operations with specially crafted input.
3096 Among EC algorithms only Brainpool P-512 curves are affected and one
3097 presumably can attack ECDH key negotiation. Impact was not analyzed in
3098 detail, because pre-requisites for attack are considered unlikely. Namely
3099 multiple clients have to choose the curve in question and the server has to
3100 share the private key among them, neither of which is default behaviour.
3101 Even then only clients that chose the curve will be affected.
3103 This issue was publicly reported as transient failures and was not
3104 initially recognized as a security issue. Thanks to Richard Morgan for
3105 providing reproducible case.
3110 * Removed automatic addition of RPATH in shared libraries and executables,
3111 as this was a remainder from OpenSSL 1.0.x and isn't needed any more.
3115 ### Changes between 1.1.0a and 1.1.0b [26 Sep 2016]
3117 * Fix Use After Free for large message sizes
3119 The patch applied to address CVE-2016-6307 resulted in an issue where if a
3120 message larger than approx 16k is received then the underlying buffer to
3121 store the incoming message is reallocated and moved. Unfortunately a
3122 dangling pointer to the old location is left which results in an attempt to
3123 write to the previously freed location. This is likely to result in a
3124 crash, however it could potentially lead to execution of arbitrary code.
3126 This issue only affects OpenSSL 1.1.0a.
3128 This issue was reported to OpenSSL by Robert Święcki.
3133 ### Changes between 1.1.0 and 1.1.0a [22 Sep 2016]
3135 * OCSP Status Request extension unbounded memory growth
3137 A malicious client can send an excessively large OCSP Status Request
3138 extension. If that client continually requests renegotiation, sending a
3139 large OCSP Status Request extension each time, then there will be unbounded
3140 memory growth on the server. This will eventually lead to a Denial Of
3141 Service attack through memory exhaustion. Servers with a default
3142 configuration are vulnerable even if they do not support OCSP. Builds using
3143 the "no-ocsp" build time option are not affected.
3145 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
3150 * SSL_peek() hang on empty record
3152 OpenSSL 1.1.0 SSL/TLS will hang during a call to SSL_peek() if the peer
3153 sends an empty record. This could be exploited by a malicious peer in a
3154 Denial Of Service attack.
3156 This issue was reported to OpenSSL by Alex Gaynor.
3161 * Excessive allocation of memory in tls_get_message_header() and
3162 dtls1_preprocess_fragment()
3164 A (D)TLS message includes 3 bytes for its length in the header for the
3165 message. This would allow for messages up to 16Mb in length. Messages of
3166 this length are excessive and OpenSSL includes a check to ensure that a
3167 peer is sending reasonably sized messages in order to avoid too much memory
3168 being consumed to service a connection. A flaw in the logic of version
3169 1.1.0 means that memory for the message is allocated too early, prior to
3170 the excessive message length check. Due to way memory is allocated in
3171 OpenSSL this could mean an attacker could force up to 21Mb to be allocated
3172 to service a connection. This could lead to a Denial of Service through
3173 memory exhaustion. However, the excessive message length check still takes
3174 place, and this would cause the connection to immediately fail. Assuming
3175 that the application calls SSL_free() on the failed connection in a timely
3176 manner then the 21Mb of allocated memory will then be immediately freed
3177 again. Therefore the excessive memory allocation will be transitory in
3178 nature. This then means that there is only a security impact if:
3180 1) The application does not call SSL_free() in a timely manner in the event
3181 that the connection fails
3183 2) The application is working in a constrained environment where there is
3184 very little free memory
3186 3) The attacker initiates multiple connection attempts such that there are
3187 multiple connections in a state where memory has been allocated for the
3188 connection; SSL_free() has not yet been called; and there is insufficient
3189 memory to service the multiple requests.
3191 Except in the instance of (1) above any Denial Of Service is likely to be
3192 transitory because as soon as the connection fails the memory is
3193 subsequently freed again in the SSL_free() call. However there is an
3194 increased risk during this period of application crashes due to the lack of
3195 memory - which would then mean a more serious Denial of Service.
3197 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
3198 (CVE-2016-6307 and CVE-2016-6308)
3202 * solaris-x86-cc, i.e. 32-bit configuration with vendor compiler,
3203 had to be removed. Primary reason is that vendor assembler can't
3204 assemble our modules with -KPIC flag. As result it, assembly
3205 support, was not even available as option. But its lack means
3206 lack of side-channel resistant code, which is incompatible with
3207 security by todays standards. Fortunately gcc is readily available
3208 prepackaged option, which we firmly point at...
3212 ### Changes between 1.0.2h and 1.1.0 [25 Aug 2016]
3214 * Windows command-line tool supports UTF-8 opt-in option for arguments
3215 and console input. Setting OPENSSL_WIN32_UTF8 environment variable
3216 (to any value) allows Windows user to access PKCS#12 file generated
3217 with Windows CryptoAPI and protected with non-ASCII password, as well
3218 as files generated under UTF-8 locale on Linux also protected with
3223 * To mitigate the SWEET32 attack ([CVE-2016-2183]), 3DES cipher suites
3224 have been disabled by default and removed from DEFAULT, just like RC4.
3225 See the RC4 item below to re-enable both.
3229 * The method for finding the storage location for the Windows RAND seed file
3230 has changed. First we check %RANDFILE%. If that is not set then we check
3231 the directories %HOME%, %USERPROFILE% and %SYSTEMROOT% in that order. If
3232 all else fails we fall back to C:\.
3236 * The EVP_EncryptUpdate() function has had its return type changed from void
3237 to int. A return of 0 indicates and error while a return of 1 indicates
3242 * The flags RSA_FLAG_NO_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME and
3243 DH_FLAG_NO_EXP_CONSTTIME which previously provided the ability to switch
3244 off the constant time implementation for RSA, DSA and DH have been made
3245 no-ops and deprecated.
3249 * Windows RAND implementation was simplified to only get entropy by
3250 calling CryptGenRandom(). Various other RAND-related tickets
3253 *Joseph Wylie Yandle, Rich Salz*
3255 * The stack and lhash API's were renamed to start with `OPENSSL_SK_`
3256 and `OPENSSL_LH_`, respectively. The old names are available
3257 with API compatibility. They new names are now completely documented.
3261 * Unify TYPE_up_ref(obj) methods signature.
3262 SSL_CTX_up_ref(), SSL_up_ref(), X509_up_ref(), EVP_PKEY_up_ref(),
3263 X509_CRL_up_ref(), X509_OBJECT_up_ref_count() methods are now returning an
3264 int (instead of void) like all others TYPE_up_ref() methods.
3265 So now these methods also check the return value of CRYPTO_atomic_add(),
3266 and the validity of object reference counter.
3268 *fdasilvayy@gmail.com*
3270 * With Windows Visual Studio builds, the .pdb files are installed
3271 alongside the installed libraries and executables. For a static
3272 library installation, ossl_static.pdb is the associate compiler
3273 generated .pdb file to be used when linking programs.
3277 * Remove openssl.spec. Packaging files belong with the packagers.
3281 * Automatic Darwin/OSX configuration has had a refresh, it will now
3282 recognise x86_64 architectures automatically. You can still decide
3283 to build for a different bitness with the environment variable
3284 KERNEL_BITS (can be 32 or 64), for example:
3286 KERNEL_BITS=32 ./config
3290 * Change default algorithms in pkcs8 utility to use PKCS#5 v2.0,
3291 256 bit AES and HMAC with SHA256.
3295 * Remove support for MIPS o32 ABI on IRIX (and IRIX only).
3299 * Triple-DES ciphers have been moved from HIGH to MEDIUM.
3303 * To enable users to have their own config files and build file templates,
3304 Configure looks in the directory indicated by the environment variable
3305 OPENSSL_LOCAL_CONFIG_DIR as well as the in-source Configurations/
3306 directory. On VMS, OPENSSL_LOCAL_CONFIG_DIR is expected to be a logical
3307 name and is used as is.
3311 * The following datatypes were made opaque: X509_OBJECT, X509_STORE_CTX,
3312 X509_STORE, X509_LOOKUP, and X509_LOOKUP_METHOD. The unused type
3313 X509_CERT_FILE_CTX was removed.
3317 * "shared" builds are now the default. To create only static libraries use
3318 the "no-shared" Configure option.
3322 * Remove the no-aes, no-hmac, no-rsa, no-sha and no-md5 Configure options.
3323 All of these option have not worked for some while and are fundamental
3328 * Make various cleanup routines no-ops and mark them as deprecated. Most
3329 global cleanup functions are no longer required because they are handled
3330 via auto-deinit (see OPENSSL_init_crypto and OPENSSL_init_ssl man pages).
3331 Explicitly de-initing can cause problems (e.g. where a library that uses
3332 OpenSSL de-inits, but an application is still using it). The affected
3333 functions are CONF_modules_free(), ENGINE_cleanup(), OBJ_cleanup(),
3334 EVP_cleanup(), BIO_sock_cleanup(), CRYPTO_cleanup_all_ex_data(),
3335 RAND_cleanup(), SSL_COMP_free_compression_methods(), ERR_free_strings() and
3336 COMP_zlib_cleanup().
3340 * --strict-warnings no longer enables runtime debugging options
3341 such as REF_DEBUG. Instead, debug options are automatically
3342 enabled with '--debug' builds.
3344 *Andy Polyakov, Emilia Käsper*
3346 * Made DH and DH_METHOD opaque. The structures for managing DH objects
3347 have been moved out of the public header files. New functions for managing
3348 these have been added.
3352 * Made RSA and RSA_METHOD opaque. The structures for managing RSA
3353 objects have been moved out of the public header files. New
3354 functions for managing these have been added.
3358 * Made DSA and DSA_METHOD opaque. The structures for managing DSA objects
3359 have been moved out of the public header files. New functions for managing
3360 these have been added.
3364 * Made BIO and BIO_METHOD opaque. The structures for managing BIOs have been
3365 moved out of the public header files. New functions for managing these
3370 * Removed no-rijndael as a config option. Rijndael is an old name for AES.
3374 * Removed the mk1mf build scripts.
3378 * Headers are now wrapped, if necessary, with OPENSSL_NO_xxx, so
3379 it is always safe to #include a header now.
3383 * Removed the aged BC-32 config and all its supporting scripts
3387 * Removed support for Ultrix, Netware, and OS/2.
3391 * Add support for HKDF.
3393 *Alessandro Ghedini*
3395 * Add support for blake2b and blake2s
3399 * Added support for "pipelining". Ciphers that have the
3400 EVP_CIPH_FLAG_PIPELINE flag set have a capability to process multiple
3401 encryptions/decryptions simultaneously. There are currently no built-in
3402 ciphers with this property but the expectation is that engines will be able
3403 to offer it to significantly improve throughput. Support has been extended
3404 into libssl so that multiple records for a single connection can be
3405 processed in one go (for >=TLS 1.1).
3409 * Added the AFALG engine. This is an async capable engine which is able to
3410 offload work to the Linux kernel. In this initial version it only supports
3411 AES128-CBC. The kernel must be version 4.1.0 or greater.
3415 * OpenSSL now uses a new threading API. It is no longer necessary to
3416 set locking callbacks to use OpenSSL in a multi-threaded environment. There
3417 are two supported threading models: pthreads and windows threads. It is
3418 also possible to configure OpenSSL at compile time for "no-threads". The
3419 old threading API should no longer be used. The functions have been
3420 replaced with "no-op" compatibility macros.
3422 *Alessandro Ghedini, Matt Caswell*
3424 * Modify behavior of ALPN to invoke callback after SNI/servername
3425 callback, such that updates to the SSL_CTX affect ALPN.
3429 * Add SSL_CIPHER queries for authentication and key-exchange.
3433 * Changes to the DEFAULT cipherlist:
3434 - Prefer (EC)DHE handshakes over plain RSA.
3435 - Prefer AEAD ciphers over legacy ciphers.
3436 - Prefer ECDSA over RSA when both certificates are available.
3437 - Prefer TLSv1.2 ciphers/PRF.
3438 - Remove DSS, SEED, IDEA, CAMELLIA, and AES-CCM from the
3443 * Change the ECC default curve list to be this, in order: x25519,
3444 secp256r1, secp521r1, secp384r1.
3448 * RC4 based libssl ciphersuites are now classed as "weak" ciphers and are
3449 disabled by default. They can be re-enabled using the
3450 enable-weak-ssl-ciphers option to Configure.
3454 * If the server has ALPN configured, but supports no protocols that the
3455 client advertises, send a fatal "no_application_protocol" alert.
3456 This behaviour is SHALL in RFC 7301, though it isn't universally
3457 implemented by other servers.
3461 * Add X25519 support.
3462 Add ASN.1 and EVP_PKEY methods for X25519. This includes support
3463 for public and private key encoding using the format documented in
3464 draft-ietf-curdle-pkix-02. The corresponding EVP_PKEY method supports
3465 key generation and key derivation.
3467 TLS support complies with draft-ietf-tls-rfc4492bis-08 and uses
3472 * Deprecate SRP_VBASE_get_by_user.
3473 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
3474 In order to fix an unavoidable memory leak ([CVE-2016-0798]),
3475 SRP_VBASE_get_by_user was changed to ignore the "fake user" SRP
3476 seed, even if the seed is configured.
3478 Users should use SRP_VBASE_get1_by_user instead. Note that in
3479 SRP_VBASE_get1_by_user, caller must free the returned value. Note
3480 also that even though configuring the SRP seed attempts to hide
3481 invalid usernames by continuing the handshake with fake
3482 credentials, this behaviour is not constant time and no strong
3483 guarantees are made that the handshake is indistinguishable from
3484 that of a valid user.
3488 * Configuration change; it's now possible to build dynamic engines
3489 without having to build shared libraries and vice versa. This
3490 only applies to the engines in `engines/`, those in `crypto/engine/`
3491 will always be built into libcrypto (i.e. "static").
3493 Building dynamic engines is enabled by default; to disable, use
3494 the configuration option "disable-dynamic-engine".
3496 The only requirements for building dynamic engines are the
3497 presence of the DSO module and building with position independent
3498 code, so they will also automatically be disabled if configuring
3499 with "disable-dso" or "disable-pic".
3501 The macros OPENSSL_NO_STATIC_ENGINE and OPENSSL_NO_DYNAMIC_ENGINE
3502 are also taken away from openssl/opensslconf.h, as they are
3507 * Configuration change; if there is a known flag to compile
3508 position independent code, it will always be applied on the
3509 libcrypto and libssl object files, and never on the application
3510 object files. This means other libraries that use routines from
3511 libcrypto / libssl can be made into shared libraries regardless
3512 of how OpenSSL was configured.
3514 If this isn't desirable, the configuration options "disable-pic"
3515 or "no-pic" can be used to disable the use of PIC. This will
3516 also disable building shared libraries and dynamic engines.
3520 * Removed JPAKE code. It was experimental and has no wide use.
3524 * The INSTALL_PREFIX Makefile variable has been renamed to
3525 DESTDIR. That makes for less confusion on what this variable
3526 is for. Also, the configuration option --install_prefix is
3531 * Heartbeat for TLS has been removed and is disabled by default
3532 for DTLS; configure with enable-heartbeats. Code that uses the
3533 old #define's might need to be updated.
3535 *Emilia Käsper, Rich Salz*
3537 * Rename REF_CHECK to REF_DEBUG.
3541 * New "unified" build system
3543 The "unified" build system is aimed to be a common system for all
3544 platforms we support. With it comes new support for VMS.
3546 This system builds supports building in a different directory tree
3547 than the source tree. It produces one Makefile (for unix family
3548 or lookalikes), or one descrip.mms (for VMS).
3550 The source of information to make the Makefile / descrip.mms is
3551 small files called 'build.info', holding the necessary
3552 information for each directory with source to compile, and a
3553 template in Configurations, like unix-Makefile.tmpl or
3556 With this change, the library names were also renamed on Windows
3557 and on VMS. They now have names that are closer to the standard
3558 on Unix, and include the major version number, and in certain
3559 cases, the architecture they are built for. See "Notes on shared
3560 libraries" in INSTALL.
3562 We rely heavily on the perl module Text::Template.
3566 * Added support for auto-initialisation and de-initialisation of the library.
3567 OpenSSL no longer requires explicit init or deinit routines to be called,
3568 except in certain circumstances. See the OPENSSL_init_crypto() and
3569 OPENSSL_init_ssl() man pages for further information.
3573 * The arguments to the DTLSv1_listen function have changed. Specifically the
3574 "peer" argument is now expected to be a BIO_ADDR object.
3576 * Rewrite of BIO networking library. The BIO library lacked consistent
3577 support of IPv6, and adding it required some more extensive
3578 modifications. This introduces the BIO_ADDR and BIO_ADDRINFO types,
3579 which hold all types of addresses and chains of address information.
3580 It also introduces a new API, with functions like BIO_socket,
3581 BIO_connect, BIO_listen, BIO_lookup and a rewrite of BIO_accept.
3582 The source/sink BIOs BIO_s_connect, BIO_s_accept and BIO_s_datagram
3583 have been adapted accordingly.
3587 * RSA_padding_check_PKCS1_type_1 now accepts inputs with and without
3592 * CRIME protection: disable compression by default, even if OpenSSL is
3593 compiled with zlib enabled. Applications can still enable compression
3594 by calling SSL_CTX_clear_options(ctx, SSL_OP_NO_COMPRESSION), or by
3595 using the SSL_CONF library to configure compression.
3599 * The signature of the session callback configured with
3600 SSL_CTX_sess_set_get_cb was changed. The read-only input buffer
3601 was explicitly marked as `const unsigned char*` instead of
3606 * Always DPURIFY. Remove the use of uninitialized memory in the
3607 RNG, and other conditional uses of DPURIFY. This makes -DPURIFY a no-op.
3611 * Removed many obsolete configuration items, including
3612 DES_PTR, DES_RISC1, DES_RISC2, DES_INT
3613 MD2_CHAR, MD2_INT, MD2_LONG
3615 IDEA_SHORT, IDEA_LONG
3616 RC2_SHORT, RC2_LONG, RC4_LONG, RC4_CHUNK, RC4_INDEX
3618 *Rich Salz, with advice from Andy Polyakov*
3620 * Many BN internals have been moved to an internal header file.
3622 *Rich Salz with help from Andy Polyakov*
3624 * Configuration and writing out the results from it has changed.
3625 Files such as Makefile include/openssl/opensslconf.h and are now
3626 produced through general templates, such as Makefile.in and
3627 crypto/opensslconf.h.in and some help from the perl module
3630 Also, the center of configuration information is no longer
3631 Makefile. Instead, Configure produces a perl module in
3632 configdata.pm which holds most of the config data (in the hash
3633 table %config), the target data that comes from the target
3634 configuration in one of the `Configurations/*.conf` files (in
3639 * To clarify their intended purposes, the Configure options
3640 --prefix and --openssldir change their semantics, and become more
3641 straightforward and less interdependent.
3643 --prefix shall be used exclusively to give the location INSTALLTOP
3644 where programs, scripts, libraries, include files and manuals are
3645 going to be installed. The default is now /usr/local.
3647 --openssldir shall be used exclusively to give the default
3648 location OPENSSLDIR where certificates, private keys, CRLs are
3649 managed. This is also where the default openssl.cnf gets
3651 If the directory given with this option is a relative path, the
3652 values of both the --prefix value and the --openssldir value will
3653 be combined to become OPENSSLDIR.
3654 The default for --openssldir is INSTALLTOP/ssl.
3656 Anyone who uses --openssldir to specify where OpenSSL is to be
3657 installed MUST change to use --prefix instead.
3661 * The GOST engine was out of date and therefore it has been removed. An up
3662 to date GOST engine is now being maintained in an external repository.
3663 See: <https://wiki.openssl.org/index.php/Binaries>. Libssl still retains
3664 support for GOST ciphersuites (these are only activated if a GOST engine
3669 * EGD is no longer supported by default; use enable-egd when
3672 *Ben Kaduk and Rich Salz*
3674 * The distribution now has Makefile.in files, which are used to
3675 create Makefile's when Configure is run. *Configure must be run
3676 before trying to build now.*
3680 * The return value for SSL_CIPHER_description() for error conditions
3685 * Support for RFC6698/RFC7671 DANE TLSA peer authentication.
3687 Obtaining and performing DNSSEC validation of TLSA records is
3688 the application's responsibility. The application provides
3689 the TLSA records of its choice to OpenSSL, and these are then
3690 used to authenticate the peer.
3692 The TLSA records need not even come from DNS. They can, for
3693 example, be used to implement local end-entity certificate or
3694 trust-anchor "pinning", where the "pin" data takes the form
3695 of TLSA records, which can augment or replace verification
3696 based on the usual WebPKI public certification authorities.
3700 * Revert default OPENSSL_NO_DEPRECATED setting. Instead OpenSSL
3701 continues to support deprecated interfaces in default builds.
3702 However, applications are strongly advised to compile their
3703 source files with -DOPENSSL_API_COMPAT=0x10100000L, which hides
3704 the declarations of all interfaces deprecated in 0.9.8, 1.0.0
3705 or the 1.1.0 releases.
3707 In environments in which all applications have been ported to
3708 not use any deprecated interfaces OpenSSL's Configure script
3709 should be used with the --api=1.1.0 option to entirely remove
3710 support for the deprecated features from the library and
3711 unconditionally disable them in the installed headers.
3712 Essentially the same effect can be achieved with the "no-deprecated"
3713 argument to Configure, except that this will always restrict
3714 the build to just the latest API, rather than a fixed API
3717 As applications are ported to future revisions of the API,
3718 they should update their compile-time OPENSSL_API_COMPAT define
3719 accordingly, but in most cases should be able to continue to
3720 compile with later releases.
3722 The OPENSSL_API_COMPAT versions for 1.0.0, and 0.9.8 are
3723 0x10000000L and 0x00908000L, respectively. However those
3724 versions did not support the OPENSSL_API_COMPAT feature, and
3725 so applications are not typically tested for explicit support
3726 of just the undeprecated features of either release.
3730 * Add support for setting the minimum and maximum supported protocol.
3731 It can bet set via the SSL_set_min_proto_version() and
3732 SSL_set_max_proto_version(), or via the SSL_CONF's MinProtocol and
3733 MaxProtocol. It's recommended to use the new APIs to disable
3734 protocols instead of disabling individual protocols using
3735 SSL_set_options() or SSL_CONF's Protocol. This change also
3736 removes support for disabling TLS 1.2 in the OpenSSL TLS
3737 client at compile time by defining OPENSSL_NO_TLS1_2_CLIENT.
3741 * Support for ChaCha20 and Poly1305 added to libcrypto and libssl.
3745 * New EC_KEY_METHOD, this replaces the older ECDSA_METHOD and ECDH_METHOD
3746 and integrates ECDSA and ECDH functionality into EC. Implementations can
3747 now redirect key generation and no longer need to convert to or from
3750 Note: the ecdsa.h and ecdh.h headers are now no longer needed and just
3751 include the ec.h header file instead.
3755 * Remove support for all 40 and 56 bit ciphers. This includes all the export
3756 ciphers who are no longer supported and drops support the ephemeral RSA key
3757 exchange. The LOW ciphers currently doesn't have any ciphers in it.
3761 * Made EVP_MD_CTX, EVP_MD, EVP_CIPHER_CTX, EVP_CIPHER and HMAC_CTX
3762 opaque. For HMAC_CTX, the following constructors and destructors
3765 HMAC_CTX *HMAC_CTX_new(void);
3766 void HMAC_CTX_free(HMAC_CTX *ctx);
3768 For EVP_MD and EVP_CIPHER, complete APIs to create, fill and
3769 destroy such methods has been added. See EVP_MD_meth_new(3) and
3770 EVP_CIPHER_meth_new(3) for documentation.
3773 1) `EVP_MD_CTX_cleanup()`, `EVP_CIPHER_CTX_cleanup()` and
3774 `HMAC_CTX_cleanup()` were removed. `HMAC_CTX_reset()` and
3775 `EVP_MD_CTX_reset()` should be called instead to reinitialise
3776 an already created structure.
3777 2) For consistency with the majority of our object creators and
3778 destructors, `EVP_MD_CTX_(create|destroy)` were renamed to
3779 `EVP_MD_CTX_(new|free)`. The old names are retained as macros
3780 for deprecated builds.
3784 * Added ASYNC support. Libcrypto now includes the async sub-library to enable
3785 cryptographic operations to be performed asynchronously as long as an
3786 asynchronous capable engine is used. See the ASYNC_start_job() man page for
3787 further details. Libssl has also had this capability integrated with the
3788 introduction of the new mode SSL_MODE_ASYNC and associated error
3789 SSL_ERROR_WANT_ASYNC. See the SSL_CTX_set_mode() and SSL_get_error() man
3790 pages. This work was developed in partnership with Intel Corp.
3794 * SSL_{CTX_}set_ecdh_auto() has been removed and ECDH is support is
3795 always enabled now. If you want to disable the support you should
3796 exclude it using the list of supported ciphers. This also means that the
3797 "-no_ecdhe" option has been removed from s_server.
3801 * SSL_{CTX}_set_tmp_ecdh() which can set 1 EC curve now internally calls
3802 SSL_{CTX_}set1_curves() which can set a list.
3806 * Remove support for SSL_{CTX_}set_tmp_ecdh_callback(). You should set the
3807 curve you want to support using SSL_{CTX_}set1_curves().
3811 * State machine rewrite. The state machine code has been significantly
3812 refactored in order to remove much duplication of code and solve issues
3813 with the old code (see [ssl/statem/README.md](ssl/statem/README.md) for
3814 further details). This change does have some associated API changes.
3815 Notably the SSL_state() function has been removed and replaced by
3816 SSL_get_state which now returns an "OSSL_HANDSHAKE_STATE" instead of an int.
3817 SSL_set_state() has been removed altogether. The previous handshake states
3818 defined in ssl.h and ssl3.h have also been removed.
3822 * All instances of the string "ssleay" in the public API were replaced
3823 with OpenSSL (case-matching; e.g., OPENSSL_VERSION for #define's)
3824 Some error codes related to internal RSA_eay API's were renamed.
3828 * The demo files in crypto/threads were moved to demo/threads.
3832 * Removed obsolete engines: 4758cca, aep, atalla, cswift, nuron, gmp,
3835 *Matt Caswell, Rich Salz*
3837 * New ASN.1 embed macro.
3839 New ASN.1 macro ASN1_EMBED. This is the same as ASN1_SIMPLE except the
3840 structure is not allocated: it is part of the parent. That is instead of
3848 This reduces memory fragmentation and make it impossible to accidentally
3849 set a mandatory field to NULL.
3851 This currently only works for some fields specifically a SEQUENCE, CHOICE,
3852 or ASN1_STRING type which is part of a parent SEQUENCE. Since it is
3853 equivalent to ASN1_SIMPLE it cannot be tagged, OPTIONAL, SET OF or
3858 * Remove EVP_CHECK_DES_KEY, a compile-time option that never compiled.
3862 * Removed DES and RC4 ciphersuites from DEFAULT. Also removed RC2 although
3863 in 1.0.2 EXPORT was already removed and the only RC2 ciphersuite is also
3864 an EXPORT one. COMPLEMENTOFDEFAULT has been updated accordingly to add
3865 DES and RC4 ciphersuites.
3869 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
3870 This changes the decoding behaviour for some invalid messages,
3871 though the change is mostly in the more lenient direction, and
3872 legacy behaviour is preserved as much as possible.
3876 * Fix no-stdio build.
3877 *David Woodhouse <David.Woodhouse@intel.com> and also*
3878 *Ivan Nestlerode <ivan.nestlerode@sonos.com>*
3880 * New testing framework
3881 The testing framework has been largely rewritten and is now using
3882 perl and the perl modules Test::Harness and an extended variant of
3883 Test::More called OpenSSL::Test to do its work. All test scripts in
3884 test/ have been rewritten into test recipes, and all direct calls to
3885 executables in test/Makefile have become individual recipes using the
3886 simplified testing OpenSSL::Test::Simple.
3888 For documentation on our testing modules, do:
3890 perldoc test/testlib/OpenSSL/Test/Simple.pm
3891 perldoc test/testlib/OpenSSL/Test.pm
3895 * Revamped memory debug; only -DCRYPTO_MDEBUG and -DCRYPTO_MDEBUG_ABORT
3896 are used; the latter aborts on memory leaks (usually checked on exit).
3897 Some undocumented "set malloc, etc., hooks" functions were removed
3898 and others were changed. All are now documented.
3902 * In DSA_generate_parameters_ex, if the provided seed is too short,
3905 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
3907 * Rewrite PSK to support ECDHE_PSK, DHE_PSK and RSA_PSK. Add ciphersuites
3908 from RFC4279, RFC4785, RFC5487, RFC5489.
3910 Thanks to Christian J. Dietrich and Giuseppe D'Angelo for the
3911 original RSA_PSK patch.
3915 * Dropped support for the SSL3_FLAGS_DELAY_CLIENT_FINISHED flag. This SSLeay
3916 era flag was never set throughout the codebase (only read). Also removed
3917 SSL3_FLAGS_POP_BUFFER which was only used if
3918 SSL3_FLAGS_DELAY_CLIENT_FINISHED was also set.
3922 * Changed the default name options in the "ca", "crl", "req" and "x509"
3923 to be "oneline" instead of "compat".
3927 * Remove SSL_OP_TLS_BLOCK_PADDING_BUG. This is SSLeay legacy, we're
3928 not aware of clients that still exhibit this bug, and the workaround
3929 hasn't been working properly for a while.
3933 * The return type of BIO_number_read() and BIO_number_written() as well as
3934 the corresponding num_read and num_write members in the BIO structure has
3935 changed from unsigned long to uint64_t. On platforms where an unsigned
3936 long is 32 bits (e.g. Windows) these counters could overflow if >4Gb is
3941 * Given the pervasive nature of TLS extensions it is inadvisable to run
3942 OpenSSL without support for them. It also means that maintaining
3943 the OPENSSL_NO_TLSEXT option within the code is very invasive (and probably
3944 not well tested). Therefore the OPENSSL_NO_TLSEXT option has been removed.
3948 * Removed support for the two export grade static DH ciphersuites
3949 EXP-DH-RSA-DES-CBC-SHA and EXP-DH-DSS-DES-CBC-SHA. These two ciphersuites
3950 were newly added (along with a number of other static DH ciphersuites) to
3951 1.0.2. However the two export ones have *never* worked since they were
3952 introduced. It seems strange in any case to be adding new export
3953 ciphersuites, and given "logjam" it also does not seem correct to fix them.
3957 * Version negotiation has been rewritten. In particular SSLv23_method(),
3958 SSLv23_client_method() and SSLv23_server_method() have been deprecated,
3959 and turned into macros which simply call the new preferred function names
3960 TLS_method(), TLS_client_method() and TLS_server_method(). All new code
3961 should use the new names instead. Also as part of this change the ssl23.h
3962 header file has been removed.
3966 * Support for Kerberos ciphersuites in TLS (RFC2712) has been removed. This
3967 code and the associated standard is no longer considered fit-for-purpose.
3971 * RT2547 was closed. When generating a private key, try to make the
3972 output file readable only by the owner. This behavior change might
3973 be noticeable when interacting with other software.
3975 * Documented all exdata functions. Added CRYPTO_free_ex_index.
3980 * Added HTTP GET support to the ocsp command.
3984 * Changed default digest for the dgst and enc commands from MD5 to
3989 * RAND_pseudo_bytes has been deprecated. Users should use RAND_bytes instead.
3993 * Added support for TLS extended master secret from
3994 draft-ietf-tls-session-hash-03.txt. Thanks for Alfredo Pironti for an
3995 initial patch which was a great help during development.
3999 * All libssl internal structures have been removed from the public header
4000 files, and the OPENSSL_NO_SSL_INTERN option has been removed (since it is
4001 now redundant). Users should not attempt to access internal structures
4002 directly. Instead they should use the provided API functions.
4006 * config has been changed so that by default OPENSSL_NO_DEPRECATED is used.
4007 Access to deprecated functions can be re-enabled by running config with
4008 "enable-deprecated". In addition applications wishing to use deprecated
4009 functions must define OPENSSL_USE_DEPRECATED. Note that this new behaviour
4010 will, by default, disable some transitive includes that previously existed
4011 in the header files (e.g. ec.h will no longer, by default, include bn.h)
4015 * Added support for OCB mode. OpenSSL has been granted a patent license
4016 compatible with the OpenSSL license for use of OCB. Details are available
4017 at <https://www.openssl.org/source/OCB-patent-grant-OpenSSL.pdf>. Support
4018 for OCB can be removed by calling config with no-ocb.
4022 * SSLv2 support has been removed. It still supports receiving a SSLv2
4023 compatible client hello.
4027 * Increased the minimal RSA keysize from 256 to 512 bits [Rich Salz],
4028 done while fixing the error code for the key-too-small case.
4030 *Annie Yousar <a.yousar@informatik.hu-berlin.de>*
4032 * CA.sh has been removed; use CA.pl instead.
4036 * Removed old DES API.
4040 * Remove various unsupported platforms:
4046 Sinix/ReliantUNIX RM400
4051 16-bit platforms such as WIN16
4055 * Clean up OPENSSL_NO_xxx #define's
4056 - Use setbuf() and remove OPENSSL_NO_SETVBUF_IONBF
4057 - Rename OPENSSL_SYSNAME_xxx to OPENSSL_SYS_xxx
4058 - OPENSSL_NO_EC{DH,DSA} merged into OPENSSL_NO_EC
4059 - OPENSSL_NO_RIPEMD160, OPENSSL_NO_RIPEMD merged into OPENSSL_NO_RMD160
4060 - OPENSSL_NO_FP_API merged into OPENSSL_NO_STDIO
4061 - Remove OPENSSL_NO_BIO OPENSSL_NO_BUFFER OPENSSL_NO_CHAIN_VERIFY
4062 OPENSSL_NO_EVP OPENSSL_NO_FIPS_ERR OPENSSL_NO_HASH_COMP
4063 OPENSSL_NO_LHASH OPENSSL_NO_OBJECT OPENSSL_NO_SPEED OPENSSL_NO_STACK
4064 OPENSSL_NO_X509 OPENSSL_NO_X509_VERIFY
4065 - Remove MS_STATIC; it's a relic from platforms <32 bits.
4069 * Cleaned up dead code
4070 Remove all but one '#ifdef undef' which is to be looked at.
4074 * Clean up calling of xxx_free routines.
4075 Just like free(), fix most of the xxx_free routines to accept
4076 NULL. Remove the non-null checks from callers. Save much code.
4080 * Add secure heap for storage of private keys (when possible).
4081 Add BIO_s_secmem(), CBIGNUM, etc.
4082 Contributed by Akamai Technologies under our Corporate CLA.
4086 * Experimental support for a new, fast, unbiased prime candidate generator,
4087 bn_probable_prime_dh_coprime(). Not currently used by any prime generator.
4089 *Felix Laurie von Massenbach <felix@erbridge.co.uk>*
4091 * New output format NSS in the sess_id command line tool. This allows
4092 exporting the session id and the master key in NSS keylog format.
4094 *Martin Kaiser <martin@kaiser.cx>*
4096 * Harmonize version and its documentation. -f flag is used to display
4099 *mancha <mancha1@zoho.com>*
4101 * Fix eckey_priv_encode so it immediately returns an error upon a failure
4102 in i2d_ECPrivateKey. Thanks to Ted Unangst for feedback on this issue.
4104 *mancha <mancha1@zoho.com>*
4106 * Fix some double frees. These are not thought to be exploitable.
4108 *mancha <mancha1@zoho.com>*
4110 * A missing bounds check in the handling of the TLS heartbeat extension
4111 can be used to reveal up to 64k of memory to a connected client or
4114 Thanks for Neel Mehta of Google Security for discovering this bug and to
4115 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
4116 preparing the fix ([CVE-2014-0160])
4118 *Adam Langley, Bodo Moeller*
4120 * Fix for the attack described in the paper "Recovering OpenSSL
4121 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
4122 by Yuval Yarom and Naomi Benger. Details can be obtained from:
4123 <http://eprint.iacr.org/2014/140>
4125 Thanks to Yuval Yarom and Naomi Benger for discovering this
4126 flaw and to Yuval Yarom for supplying a fix ([CVE-2014-0076])
4128 *Yuval Yarom and Naomi Benger*
4130 * Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
4131 this fixes a limitation in previous versions of OpenSSL.
4135 * Experimental encrypt-then-mac support.
4137 Experimental support for encrypt then mac from
4138 draft-gutmann-tls-encrypt-then-mac-02.txt
4140 To enable it set the appropriate extension number (0x42 for the test
4141 server) using e.g. -DTLSEXT_TYPE_encrypt_then_mac=0x42
4143 For non-compliant peers (i.e. just about everything) this should have no
4146 WARNING: EXPERIMENTAL, SUBJECT TO CHANGE.
4150 * Add EVP support for key wrapping algorithms, to avoid problems with
4151 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
4152 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
4153 algorithms and include tests cases.
4157 * Extend CMS code to support RSA-PSS signatures and RSA-OAEP for
4162 * Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
4163 MGF1 digest and OAEP label.
4167 * Make openssl verify return errors.
4169 *Chris Palmer <palmer@google.com> and Ben Laurie*
4171 * New function ASN1_TIME_diff to calculate the difference between two
4172 ASN1_TIME structures or one structure and the current time.
4176 * Update fips_test_suite to support multiple command line options. New
4177 test to induce all self test errors in sequence and check expected
4182 * Add FIPS_{rsa,dsa,ecdsa}_{sign,verify} functions which digest and
4183 sign or verify all in one operation.
4187 * Add fips_algvs: a multicall fips utility incorporating all the algorithm
4188 test programs and fips_test_suite. Includes functionality to parse
4189 the minimal script output of fipsalgest.pl directly.
4193 * Add authorisation parameter to FIPS_module_mode_set().
4197 * Add FIPS selftest for ECDH algorithm using P-224 and B-233 curves.
4201 * Use separate DRBG fields for internal and external flags. New function
4202 FIPS_drbg_health_check() to perform on demand health checking. Add
4203 generation tests to fips_test_suite with reduced health check interval to
4204 demonstrate periodic health checking. Add "nodh" option to
4205 fips_test_suite to skip very slow DH test.
4209 * New function FIPS_get_cipherbynid() to lookup FIPS supported ciphers
4214 * More extensive health check for DRBG checking many more failure modes.
4215 New function FIPS_selftest_drbg_all() to handle every possible DRBG
4216 combination: call this in fips_test_suite.
4220 * Add support for canonical generation of DSA parameter 'g'. See
4223 * Add support for HMAC DRBG from SP800-90. Update DRBG algorithm test and
4224 POST to handle HMAC cases.
4228 * Add functions FIPS_module_version() and FIPS_module_version_text()
4229 to return numerical and string versions of the FIPS module number.
4233 * Rename FIPS_mode_set and FIPS_mode to FIPS_module_mode_set and
4234 FIPS_module_mode. FIPS_mode and FIPS_mode_set will be implemented
4235 outside the validated module in the FIPS capable OpenSSL.
4239 * Minor change to DRBG entropy callback semantics. In some cases
4240 there is no multiple of the block length between min_len and
4241 max_len. Allow the callback to return more than max_len bytes
4242 of entropy but discard any extra: it is the callback's responsibility
4243 to ensure that the extra data discarded does not impact the
4244 requested amount of entropy.
4248 * Add PRNG security strength checks to RSA, DSA and ECDSA using
4249 information in FIPS186-3, SP800-57 and SP800-131A.
4253 * CCM support via EVP. Interface is very similar to GCM case except we
4254 must supply all data in one chunk (i.e. no update, final) and the
4255 message length must be supplied if AAD is used. Add algorithm test
4260 * Initial version of POST overhaul. Add POST callback to allow the status
4261 of POST to be monitored and/or failures induced. Modify fips_test_suite
4262 to use callback. Always run all selftests even if one fails.
4266 * XTS support including algorithm test driver in the fips_gcmtest program.
4267 Note: this does increase the maximum key length from 32 to 64 bytes but
4268 there should be no binary compatibility issues as existing applications
4269 will never use XTS mode.
4273 * Extensive reorganisation of FIPS PRNG behaviour. Remove all dependencies
4274 to OpenSSL RAND code and replace with a tiny FIPS RAND API which also
4275 performs algorithm blocking for unapproved PRNG types. Also do not
4276 set PRNG type in FIPS_mode_set(): leave this to the application.
4277 Add default OpenSSL DRBG handling: sets up FIPS PRNG and seeds with
4278 the standard OpenSSL PRNG: set additional data to a date time vector.
4282 * Rename old X9.31 PRNG functions of the form `FIPS_rand*` to `FIPS_x931*`.
4283 This shouldn't present any incompatibility problems because applications
4284 shouldn't be using these directly and any that are will need to rethink
4285 anyway as the X9.31 PRNG is now deprecated by FIPS 140-2
4289 * Extensive self tests and health checking required by SP800-90 DRBG.
4290 Remove strength parameter from FIPS_drbg_instantiate and always
4291 instantiate at maximum supported strength.
4295 * Add ECDH code to fips module and fips_ecdhvs for primitives only testing.
4299 * New algorithm test program fips_dhvs to handle DH primitives only testing.
4303 * New function DH_compute_key_padded() to compute a DH key and pad with
4304 leading zeroes if needed: this complies with SP800-56A et al.
4308 * Initial implementation of SP800-90 DRBGs for Hash and CTR. Not used by
4309 anything, incomplete, subject to change and largely untested at present.
4313 * Modify fipscanisteronly build option to only build the necessary object
4314 files by filtering FIPS_EX_OBJ through a perl script in crypto/Makefile.
4318 * Add experimental option FIPSSYMS to give all symbols in
4319 fipscanister.o and FIPS or fips prefix. This will avoid
4320 conflicts with future versions of OpenSSL. Add perl script
4321 util/fipsas.pl to preprocess assembly language source files
4322 and rename any affected symbols.
4326 * Add selftest checks and algorithm block of non-fips algorithms in
4327 FIPS mode. Remove DES2 from selftests.
4331 * Add ECDSA code to fips module. Add tiny fips_ecdsa_check to just
4332 return internal method without any ENGINE dependencies. Add new
4333 tiny fips sign and verify functions.
4337 * New build option no-ec2m to disable characteristic 2 code.
4341 * New build option "fipscanisteronly". This only builds fipscanister.o
4342 and (currently) associated fips utilities. Uses the file Makefile.fips
4343 instead of Makefile.org as the prototype.
4347 * Add some FIPS mode restrictions to GCM. Add internal IV generator.
4348 Update fips_gcmtest to use IV generator.
4352 * Initial, experimental EVP support for AES-GCM. AAD can be input by
4353 setting output buffer to NULL. The `*Final` function must be
4354 called although it will not retrieve any additional data. The tag
4355 can be set or retrieved with a ctrl. The IV length is by default 12
4356 bytes (96 bits) but can be set to an alternative value. If the IV
4357 length exceeds the maximum IV length (currently 16 bytes) it cannot be
4362 * New flag in ciphers: EVP_CIPH_FLAG_CUSTOM_CIPHER. This means the
4363 underlying do_cipher function handles all cipher semantics itself
4364 including padding and finalisation. This is useful if (for example)
4365 an ENGINE cipher handles block padding itself. The behaviour of
4366 do_cipher is subtly changed if this flag is set: the return value
4367 is the number of characters written to the output buffer (zero is
4368 no longer an error code) or a negative error code. Also if the
4369 input buffer is NULL and length 0 finalisation should be performed.
4373 * If a candidate issuer certificate is already part of the constructed
4374 path ignore it: new debug notification X509_V_ERR_PATH_LOOP for this case.
4378 * Improve forward-security support: add functions
4380 void SSL_CTX_set_not_resumable_session_callback(
4381 SSL_CTX *ctx, int (*cb)(SSL *ssl, int is_forward_secure))
4382 void SSL_set_not_resumable_session_callback(
4383 SSL *ssl, int (*cb)(SSL *ssl, int is_forward_secure))
4385 for use by SSL/TLS servers; the callback function will be called whenever a
4386 new session is created, and gets to decide whether the session may be
4387 cached to make it resumable (return 0) or not (return 1). (As by the
4388 SSL/TLS protocol specifications, the session_id sent by the server will be
4389 empty to indicate that the session is not resumable; also, the server will
4390 not generate RFC 4507 (RFC 5077) session tickets.)
4392 A simple reasonable callback implementation is to return is_forward_secure.
4393 This parameter will be set to 1 or 0 depending on the ciphersuite selected
4394 by the SSL/TLS server library, indicating whether it can provide forward
4397 *Emilia Käsper <emilia.kasper@esat.kuleuven.be> (Google)*
4399 * New -verify_name option in command line utilities to set verification
4404 * Initial CMAC implementation. WARNING: EXPERIMENTAL, API MAY CHANGE.
4405 Add CMAC pkey methods.
4409 * Experimental renegotiation in s_server -www mode. If the client
4410 browses /reneg connection is renegotiated. If /renegcert it is
4411 renegotiated requesting a certificate.
4415 * Add an "external" session cache for debugging purposes to s_server. This
4416 should help trace issues which normally are only apparent in deployed
4417 multi-process servers.
4421 * Extensive audit of libcrypto with DEBUG_UNUSED. Fix many cases where
4422 return value is ignored. NB. The functions RAND_add(), RAND_seed(),
4423 BIO_set_cipher() and some obscure PEM functions were changed so they
4424 can now return an error. The RAND changes required a change to the
4425 RAND_METHOD structure.
4429 * New macro `__owur` for "OpenSSL Warn Unused Result". This makes use of
4430 a gcc attribute to warn if the result of a function is ignored. This
4431 is enable if DEBUG_UNUSED is set. Add to several functions in evp.h
4432 whose return value is often ignored.
4436 * New -noct, -requestct, -requirect and -ctlogfile options for s_client.
4437 These allow SCTs (signed certificate timestamps) to be requested and
4438 validated when establishing a connection.
4440 *Rob Percival <robpercival@google.com>*
4445 ### Changes between 1.0.2s and 1.0.2t [10 Sep 2019]
4447 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
4448 used even when parsing explicit parameters, when loading a encoded key
4449 or calling `EC_GROUP_new_from_ecpkparameters()`/
4450 `EC_GROUP_new_from_ecparameters()`.
4451 This prevents bypass of security hardening and performance gains,
4452 especially for curves with specialized EC_METHODs.
4453 By default, if a key encoded with explicit parameters is loaded and later
4454 encoded, the output is still encoded with explicit parameters, even if
4455 internally a "named" EC_GROUP is used for computation.
4459 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
4460 this change, EC_GROUP_set_generator would accept order and/or cofactor as
4461 NULL. After this change, only the cofactor parameter can be NULL. It also
4462 does some minimal sanity checks on the passed order.
4467 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
4468 An attack is simple, if the first CMS_recipientInfo is valid but the
4469 second CMS_recipientInfo is chosen ciphertext. If the second
4470 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
4471 encryption key will be replaced by garbage, and the message cannot be
4472 decoded, but if the RSA decryption fails, the correct encryption key is
4473 used and the recipient will not notice the attack.
4474 As a work around for this potential attack the length of the decrypted
4475 key must be equal to the cipher default key length, in case the
4476 certifiate is not given and all recipientInfo are tried out.
4477 The old behaviour can be re-enabled in the CMS code by setting the
4478 CMS_DEBUG_DECRYPT flag.
4483 * Document issue with installation paths in diverse Windows builds
4485 '/usr/local/ssl' is an unsafe prefix for location to install OpenSSL
4486 binaries and run-time config file.
4491 ### Changes between 1.0.2r and 1.0.2s [28 May 2019]
4493 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
4494 This changes the size when using the `genpkey` command when no size is given.
4495 It fixes an omission in earlier changes that changed all RSA, DSA and DH
4496 generation commands to use 2048 bits by default.
4500 * Add FIPS support for Android Arm 64-bit
4502 Support for Android Arm 64-bit was added to the OpenSSL FIPS Object
4503 Module in Version 2.0.10. For some reason, the corresponding target
4504 'android64-aarch64' was missing OpenSSL 1.0.2, whence it could not be
4505 built with FIPS support on Android Arm 64-bit. This omission has been
4508 *Matthias St. Pierre*
4510 ### Changes between 1.0.2q and 1.0.2r [26 Feb 2019]
4512 * 0-byte record padding oracle
4514 If an application encounters a fatal protocol error and then calls
4515 SSL_shutdown() twice (once to send a close_notify, and once to receive one)
4516 then OpenSSL can respond differently to the calling application if a 0 byte
4517 record is received with invalid padding compared to if a 0 byte record is
4518 received with an invalid MAC. If the application then behaves differently
4519 based on that in a way that is detectable to the remote peer, then this
4520 amounts to a padding oracle that could be used to decrypt data.
4522 In order for this to be exploitable "non-stitched" ciphersuites must be in
4523 use. Stitched ciphersuites are optimised implementations of certain
4524 commonly used ciphersuites. Also the application must call SSL_shutdown()
4525 twice even if a protocol error has occurred (applications should not do
4526 this but some do anyway).
4528 This issue was discovered by Juraj Somorovsky, Robert Merget and Nimrod
4529 Aviram, with additional investigation by Steven Collison and Andrew
4530 Hourselt. It was reported to OpenSSL on 10th December 2018.
4535 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
4539 ### Changes between 1.0.2p and 1.0.2q [20 Nov 2018]
4541 * Microarchitecture timing vulnerability in ECC scalar multiplication
4543 OpenSSL ECC scalar multiplication, used in e.g. ECDSA and ECDH, has been
4544 shown to be vulnerable to a microarchitecture timing side channel attack.
4545 An attacker with sufficient access to mount local timing attacks during
4546 ECDSA signature generation could recover the private key.
4548 This issue was reported to OpenSSL on 26th October 2018 by Alejandro
4549 Cabrera Aldaya, Billy Brumley, Sohaib ul Hassan, Cesar Pereida Garcia and
4555 * Timing vulnerability in DSA signature generation
4557 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
4558 timing side channel attack. An attacker could use variations in the signing
4559 algorithm to recover the private key.
4561 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
4566 * Resolve a compatibility issue in EC_GROUP handling with the FIPS Object
4567 Module, accidentally introduced while backporting security fixes from the
4568 development branch and hindering the use of ECC in FIPS mode.
4572 ### Changes between 1.0.2o and 1.0.2p [14 Aug 2018]
4574 * Client DoS due to large DH parameter
4576 During key agreement in a TLS handshake using a DH(E) based ciphersuite a
4577 malicious server can send a very large prime value to the client. This will
4578 cause the client to spend an unreasonably long period of time generating a
4579 key for this prime resulting in a hang until the client has finished. This
4580 could be exploited in a Denial Of Service attack.
4582 This issue was reported to OpenSSL on 5th June 2018 by Guido Vranken
4587 * Cache timing vulnerability in RSA Key Generation
4589 The OpenSSL RSA Key generation algorithm has been shown to be vulnerable to
4590 a cache timing side channel attack. An attacker with sufficient access to
4591 mount cache timing attacks during the RSA key generation process could
4592 recover the private key.
4594 This issue was reported to OpenSSL on 4th April 2018 by Alejandro Cabrera
4595 Aldaya, Billy Brumley, Cesar Pereida Garcia and Luis Manuel Alvarez Tapia.
4600 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
4601 parameter is no longer accepted, as it leads to a corrupt table. NULL
4602 pem_str is reserved for alias entries only.
4606 * Revert blinding in ECDSA sign and instead make problematic addition
4607 length-invariant. Switch even to fixed-length Montgomery multiplication.
4611 * Change generating and checking of primes so that the error rate of not
4612 being prime depends on the intended use based on the size of the input.
4613 For larger primes this will result in more rounds of Miller-Rabin.
4614 The maximal error rate for primes with more than 1080 bits is lowered
4617 *Kurt Roeckx, Annie Yousar*
4619 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
4623 * Add blinding to ECDSA and DSA signatures to protect against side channel
4624 attacks discovered by Keegan Ryan (NCC Group).
4628 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
4629 now allow empty (zero character) pass phrases.
4633 * Certificate time validation (X509_cmp_time) enforces stricter
4634 compliance with RFC 5280. Fractional seconds and timezone offsets
4635 are no longer allowed.
4639 ### Changes between 1.0.2n and 1.0.2o [27 Mar 2018]
4641 * Constructed ASN.1 types with a recursive definition could exceed the stack
4643 Constructed ASN.1 types with a recursive definition (such as can be found
4644 in PKCS7) could eventually exceed the stack given malicious input with
4645 excessive recursion. This could result in a Denial Of Service attack. There
4646 are no such structures used within SSL/TLS that come from untrusted sources
4647 so this is considered safe.
4649 This issue was reported to OpenSSL on 4th January 2018 by the OSS-fuzz
4655 ### Changes between 1.0.2m and 1.0.2n [7 Dec 2017]
4657 * Read/write after SSL object in error state
4659 OpenSSL 1.0.2 (starting from version 1.0.2b) introduced an "error state"
4660 mechanism. The intent was that if a fatal error occurred during a handshake
4661 then OpenSSL would move into the error state and would immediately fail if
4662 you attempted to continue the handshake. This works as designed for the
4663 explicit handshake functions (SSL_do_handshake(), SSL_accept() and
4664 SSL_connect()), however due to a bug it does not work correctly if
4665 SSL_read() or SSL_write() is called directly. In that scenario, if the
4666 handshake fails then a fatal error will be returned in the initial function
4667 call. If SSL_read()/SSL_write() is subsequently called by the application
4668 for the same SSL object then it will succeed and the data is passed without
4669 being decrypted/encrypted directly from the SSL/TLS record layer.
4671 In order to exploit this issue an application bug would have to be present
4672 that resulted in a call to SSL_read()/SSL_write() being issued after having
4673 already received a fatal error.
4675 This issue was reported to OpenSSL by David Benjamin (Google).
4680 * rsaz_1024_mul_avx2 overflow bug on x86_64
4682 There is an overflow bug in the AVX2 Montgomery multiplication procedure
4683 used in exponentiation with 1024-bit moduli. No EC algorithms are affected.
4684 Analysis suggests that attacks against RSA and DSA as a result of this
4685 defect would be very difficult to perform and are not believed likely.
4686 Attacks against DH1024 are considered just feasible, because most of the
4687 work necessary to deduce information about a private key may be performed
4688 offline. The amount of resources required for such an attack would be
4689 significant. However, for an attack on TLS to be meaningful, the server
4690 would have to share the DH1024 private key among multiple clients, which is
4691 no longer an option since CVE-2016-0701.
4693 This only affects processors that support the AVX2 but not ADX extensions
4694 like Intel Haswell (4th generation).
4696 This issue was reported to OpenSSL by David Benjamin (Google). The issue
4697 was originally found via the OSS-Fuzz project.
4702 ### Changes between 1.0.2l and 1.0.2m [2 Nov 2017]
4704 * bn_sqrx8x_internal carry bug on x86_64
4706 There is a carry propagating bug in the x86_64 Montgomery squaring
4707 procedure. No EC algorithms are affected. Analysis suggests that attacks
4708 against RSA and DSA as a result of this defect would be very difficult to
4709 perform and are not believed likely. Attacks against DH are considered just
4710 feasible (although very difficult) because most of the work necessary to
4711 deduce information about a private key may be performed offline. The amount
4712 of resources required for such an attack would be very significant and
4713 likely only accessible to a limited number of attackers. An attacker would
4714 additionally need online access to an unpatched system using the target
4715 private key in a scenario with persistent DH parameters and a private
4716 key that is shared between multiple clients.
4718 This only affects processors that support the BMI1, BMI2 and ADX extensions
4719 like Intel Broadwell (5th generation) and later or AMD Ryzen.
4721 This issue was reported to OpenSSL by the OSS-Fuzz project.
4726 * Malformed X.509 IPAddressFamily could cause OOB read
4728 If an X.509 certificate has a malformed IPAddressFamily extension,
4729 OpenSSL could do a one-byte buffer overread. The most likely result
4730 would be an erroneous display of the certificate in text format.
4732 This issue was reported to OpenSSL by the OSS-Fuzz project.
4736 ### Changes between 1.0.2k and 1.0.2l [25 May 2017]
4738 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
4739 platform rather than 'mingw'.
4743 ### Changes between 1.0.2j and 1.0.2k [26 Jan 2017]
4745 * Truncated packet could crash via OOB read
4747 If one side of an SSL/TLS path is running on a 32-bit host and a specific
4748 cipher is being used, then a truncated packet can cause that host to
4749 perform an out-of-bounds read, usually resulting in a crash.
4751 This issue was reported to OpenSSL by Robert Święcki of Google.
4756 * BN_mod_exp may produce incorrect results on x86_64
4758 There is a carry propagating bug in the x86_64 Montgomery squaring
4759 procedure. No EC algorithms are affected. Analysis suggests that attacks
4760 against RSA and DSA as a result of this defect would be very difficult to
4761 perform and are not believed likely. Attacks against DH are considered just
4762 feasible (although very difficult) because most of the work necessary to
4763 deduce information about a private key may be performed offline. The amount
4764 of resources required for such an attack would be very significant and
4765 likely only accessible to a limited number of attackers. An attacker would
4766 additionally need online access to an unpatched system using the target
4767 private key in a scenario with persistent DH parameters and a private
4768 key that is shared between multiple clients. For example this can occur by
4769 default in OpenSSL DHE based SSL/TLS ciphersuites. Note: This issue is very
4770 similar to CVE-2015-3193 but must be treated as a separate problem.
4772 This issue was reported to OpenSSL by the OSS-Fuzz project.
4777 * Montgomery multiplication may produce incorrect results
4779 There is a carry propagating bug in the Broadwell-specific Montgomery
4780 multiplication procedure that handles input lengths divisible by, but
4781 longer than 256 bits. Analysis suggests that attacks against RSA, DSA
4782 and DH private keys are impossible. This is because the subroutine in
4783 question is not used in operations with the private key itself and an input
4784 of the attacker's direct choice. Otherwise the bug can manifest itself as
4785 transient authentication and key negotiation failures or reproducible
4786 erroneous outcome of public-key operations with specially crafted input.
4787 Among EC algorithms only Brainpool P-512 curves are affected and one
4788 presumably can attack ECDH key negotiation. Impact was not analyzed in
4789 detail, because pre-requisites for attack are considered unlikely. Namely
4790 multiple clients have to choose the curve in question and the server has to
4791 share the private key among them, neither of which is default behaviour.
4792 Even then only clients that chose the curve will be affected.
4794 This issue was publicly reported as transient failures and was not
4795 initially recognized as a security issue. Thanks to Richard Morgan for
4796 providing reproducible case.
4801 * OpenSSL now fails if it receives an unrecognised record type in TLS1.0
4802 or TLS1.1. Previously this only happened in SSLv3 and TLS1.2. This is to
4803 prevent issues where no progress is being made and the peer continually
4804 sends unrecognised record types, using up resources processing them.
4808 ### Changes between 1.0.2i and 1.0.2j [26 Sep 2016]
4810 * Missing CRL sanity check
4812 A bug fix which included a CRL sanity check was added to OpenSSL 1.1.0
4813 but was omitted from OpenSSL 1.0.2i. As a result any attempt to use
4814 CRLs in OpenSSL 1.0.2i will crash with a null pointer exception.
4816 This issue only affects the OpenSSL 1.0.2i
4821 ### Changes between 1.0.2h and 1.0.2i [22 Sep 2016]
4823 * OCSP Status Request extension unbounded memory growth
4825 A malicious client can send an excessively large OCSP Status Request
4826 extension. If that client continually requests renegotiation, sending a
4827 large OCSP Status Request extension each time, then there will be unbounded
4828 memory growth on the server. This will eventually lead to a Denial Of
4829 Service attack through memory exhaustion. Servers with a default
4830 configuration are vulnerable even if they do not support OCSP. Builds using
4831 the "no-ocsp" build time option are not affected.
4833 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4838 * In order to mitigate the SWEET32 attack, the DES ciphers were moved from
4841 This issue was reported to OpenSSL Karthikeyan Bhargavan and Gaetan
4847 * OOB write in MDC2_Update()
4849 An overflow can occur in MDC2_Update() either if called directly or
4850 through the EVP_DigestUpdate() function using MDC2. If an attacker
4851 is able to supply very large amounts of input data after a previous
4852 call to EVP_EncryptUpdate() with a partial block then a length check
4853 can overflow resulting in a heap corruption.
4855 The amount of data needed is comparable to SIZE_MAX which is impractical
4858 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4863 * Malformed SHA512 ticket DoS
4865 If a server uses SHA512 for TLS session ticket HMAC it is vulnerable to a
4866 DoS attack where a malformed ticket will result in an OOB read which will
4869 The use of SHA512 in TLS session tickets is comparatively rare as it requires
4870 a custom server callback and ticket lookup mechanism.
4872 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4877 * OOB write in BN_bn2dec()
4879 The function BN_bn2dec() does not check the return value of BN_div_word().
4880 This can cause an OOB write if an application uses this function with an
4881 overly large BIGNUM. This could be a problem if an overly large certificate
4882 or CRL is printed out from an untrusted source. TLS is not affected because
4883 record limits will reject an oversized certificate before it is parsed.
4885 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4890 * OOB read in TS_OBJ_print_bio()
4892 The function TS_OBJ_print_bio() misuses OBJ_obj2txt(): the return value is
4893 the total length the OID text representation would use and not the amount
4894 of data written. This will result in OOB reads when large OIDs are
4897 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4902 * Pointer arithmetic undefined behaviour
4904 Avoid some undefined pointer arithmetic
4906 A common idiom in the codebase is to check limits in the following manner:
4909 Where "p" points to some malloc'd data of SIZE bytes and
4912 "len" here could be from some externally supplied data (e.g. from a TLS
4915 The rules of C pointer arithmetic are such that "p + len" is only well
4916 defined where len <= SIZE. Therefore the above idiom is actually
4917 undefined behaviour.
4919 For example this could cause problems if some malloc implementation
4920 provides an address for "p" such that "p + len" actually overflows for
4921 values of len that are too big and therefore p + len < limit.
4923 This issue was reported to OpenSSL by Guido Vranken
4928 * Constant time flag not preserved in DSA signing
4930 Operations in the DSA signing algorithm should run in constant time in
4931 order to avoid side channel attacks. A flaw in the OpenSSL DSA
4932 implementation means that a non-constant time codepath is followed for
4933 certain operations. This has been demonstrated through a cache-timing
4934 attack to be sufficient for an attacker to recover the private DSA key.
4936 This issue was reported by César Pereida (Aalto University), Billy Brumley
4937 (Tampere University of Technology), and Yuval Yarom (The University of
4938 Adelaide and NICTA).
4943 * DTLS buffered message DoS
4945 In a DTLS connection where handshake messages are delivered out-of-order
4946 those messages that OpenSSL is not yet ready to process will be buffered
4947 for later use. Under certain circumstances, a flaw in the logic means that
4948 those messages do not get removed from the buffer even though the handshake
4949 has been completed. An attacker could force up to approx. 15 messages to
4950 remain in the buffer when they are no longer required. These messages will
4951 be cleared when the DTLS connection is closed. The default maximum size for
4952 a message is 100k. Therefore the attacker could force an additional 1500k
4953 to be consumed per connection. By opening many simulataneous connections an
4954 attacker could cause a DoS attack through memory exhaustion.
4956 This issue was reported to OpenSSL by Quan Luo.
4961 * DTLS replay protection DoS
4963 A flaw in the DTLS replay attack protection mechanism means that records
4964 that arrive for future epochs update the replay protection "window" before
4965 the MAC for the record has been validated. This could be exploited by an
4966 attacker by sending a record for the next epoch (which does not have to
4967 decrypt or have a valid MAC), with a very large sequence number. This means
4968 that all subsequent legitimate packets are dropped causing a denial of
4969 service for a specific DTLS connection.
4971 This issue was reported to OpenSSL by the OCAP audit team.
4976 * Certificate message OOB reads
4978 In OpenSSL 1.0.2 and earlier some missing message length checks can result
4979 in OOB reads of up to 2 bytes beyond an allocated buffer. There is a
4980 theoretical DoS risk but this has not been observed in practice on common
4983 The messages affected are client certificate, client certificate request
4984 and server certificate. As a result the attack can only be performed
4985 against a client or a server which enables client authentication.
4987 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4992 ### Changes between 1.0.2g and 1.0.2h [3 May 2016]
4994 * Prevent padding oracle in AES-NI CBC MAC check
4996 A MITM attacker can use a padding oracle attack to decrypt traffic
4997 when the connection uses an AES CBC cipher and the server support
5000 This issue was introduced as part of the fix for Lucky 13 padding
5001 attack ([CVE-2013-0169]). The padding check was rewritten to be in
5002 constant time by making sure that always the same bytes are read and
5003 compared against either the MAC or padding bytes. But it no longer
5004 checked that there was enough data to have both the MAC and padding
5007 This issue was reported by Juraj Somorovsky using TLS-Attacker.
5011 * Fix EVP_EncodeUpdate overflow
5013 An overflow can occur in the EVP_EncodeUpdate() function which is used for
5014 Base64 encoding of binary data. If an attacker is able to supply very large
5015 amounts of input data then a length check can overflow resulting in a heap
5018 Internally to OpenSSL the EVP_EncodeUpdate() function is primarily used by
5019 the `PEM_write_bio*` family of functions. These are mainly used within the
5020 OpenSSL command line applications, so any application which processes data
5021 from an untrusted source and outputs it as a PEM file should be considered
5022 vulnerable to this issue. User applications that call these APIs directly
5023 with large amounts of untrusted data may also be vulnerable.
5025 This issue was reported by Guido Vranken.
5030 * Fix EVP_EncryptUpdate overflow
5032 An overflow can occur in the EVP_EncryptUpdate() function. If an attacker
5033 is able to supply very large amounts of input data after a previous call to
5034 EVP_EncryptUpdate() with a partial block then a length check can overflow
5035 resulting in a heap corruption. Following an analysis of all OpenSSL
5036 internal usage of the EVP_EncryptUpdate() function all usage is one of two
5037 forms. The first form is where the EVP_EncryptUpdate() call is known to be
5038 the first called function after an EVP_EncryptInit(), and therefore that
5039 specific call must be safe. The second form is where the length passed to
5040 EVP_EncryptUpdate() can be seen from the code to be some small value and
5041 therefore there is no possibility of an overflow. Since all instances are
5042 one of these two forms, it is believed that there can be no overflows in
5043 internal code due to this problem. It should be noted that
5044 EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths.
5045 Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances
5046 of these calls have also been analysed too and it is believed there are no
5047 instances in internal usage where an overflow could occur.
5049 This issue was reported by Guido Vranken.
5054 * Prevent ASN.1 BIO excessive memory allocation
5056 When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio()
5057 a short invalid encoding can cause allocation of large amounts of memory
5058 potentially consuming excessive resources or exhausting memory.
5060 Any application parsing untrusted data through d2i BIO functions is
5061 affected. The memory based functions such as d2i_X509() are *not* affected.
5062 Since the memory based functions are used by the TLS library, TLS
5063 applications are not affected.
5065 This issue was reported by Brian Carpenter.
5072 ASN1 Strings that are over 1024 bytes can cause an overread in applications
5073 using the X509_NAME_oneline() function on EBCDIC systems. This could result
5074 in arbitrary stack data being returned in the buffer.
5076 This issue was reported by Guido Vranken.
5081 * Modify behavior of ALPN to invoke callback after SNI/servername
5082 callback, such that updates to the SSL_CTX affect ALPN.
5086 * Remove LOW from the DEFAULT cipher list. This removes singles DES from the
5091 * Only remove the SSLv2 methods with the no-ssl2-method option. When the
5092 methods are enabled and ssl2 is disabled the methods return NULL.
5096 ### Changes between 1.0.2f and 1.0.2g [1 Mar 2016]
5098 * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL.
5099 Builds that are not configured with "enable-weak-ssl-ciphers" will not
5100 provide any "EXPORT" or "LOW" strength ciphers.
5104 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2
5105 is by default disabled at build-time. Builds that are not configured with
5106 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used,
5107 users who want to negotiate SSLv2 via the version-flexible SSLv23_method()
5108 will need to explicitly call either of:
5110 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2);
5112 SSL_clear_options(ssl, SSL_OP_NO_SSLv2);
5114 as appropriate. Even if either of those is used, or the application
5115 explicitly uses the version-specific SSLv2_method() or its client and
5116 server variants, SSLv2 ciphers vulnerable to exhaustive search key
5117 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT
5118 ciphers, and SSLv2 56-bit DES are no longer available.
5123 * Fix a double-free in DSA code
5125 A double free bug was discovered when OpenSSL parses malformed DSA private
5126 keys and could lead to a DoS attack or memory corruption for applications
5127 that receive DSA private keys from untrusted sources. This scenario is
5130 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using
5136 * Disable SRP fake user seed to address a server memory leak.
5138 Add a new method SRP_VBASE_get1_by_user that handles the seed properly.
5140 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
5141 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user
5142 was changed to ignore the "fake user" SRP seed, even if the seed
5145 Users should use SRP_VBASE_get1_by_user instead. Note that in
5146 SRP_VBASE_get1_by_user, caller must free the returned value. Note
5147 also that even though configuring the SRP seed attempts to hide
5148 invalid usernames by continuing the handshake with fake
5149 credentials, this behaviour is not constant time and no strong
5150 guarantees are made that the handshake is indistinguishable from
5151 that of a valid user.
5156 * Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption
5158 In the BN_hex2bn function the number of hex digits is calculated using an
5159 int value `i`. Later `bn_expand` is called with a value of `i * 4`. For
5160 large values of `i` this can result in `bn_expand` not allocating any
5161 memory because `i * 4` is negative. This can leave the internal BIGNUM data
5162 field as NULL leading to a subsequent NULL ptr deref. For very large values
5163 of `i`, the calculation `i * 4` could be a positive value smaller than `i`.
5164 In this case memory is allocated to the internal BIGNUM data field, but it
5165 is insufficiently sized leading to heap corruption. A similar issue exists
5166 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn
5167 is ever called by user applications with very large untrusted hex/dec data.
5168 This is anticipated to be a rare occurrence.
5170 All OpenSSL internal usage of these functions use data that is not expected
5171 to be untrusted, e.g. config file data or application command line
5172 arguments. If user developed applications generate config file data based
5173 on untrusted data then it is possible that this could also lead to security
5174 consequences. This is also anticipated to be rare.
5176 This issue was reported to OpenSSL by Guido Vranken.
5181 * Fix memory issues in `BIO_*printf` functions
5183 The internal `fmtstr` function used in processing a "%s" format string in
5184 the `BIO_*printf` functions could overflow while calculating the length of a
5185 string and cause an OOB read when printing very long strings.
5187 Additionally the internal `doapr_outch` function can attempt to write to an
5188 OOB memory location (at an offset from the NULL pointer) in the event of a
5189 memory allocation failure. In 1.0.2 and below this could be caused where
5190 the size of a buffer to be allocated is greater than INT_MAX. E.g. this
5191 could be in processing a very long "%s" format string. Memory leaks can
5194 The first issue may mask the second issue dependent on compiler behaviour.
5195 These problems could enable attacks where large amounts of untrusted data
5196 is passed to the `BIO_*printf` functions. If applications use these functions
5197 in this way then they could be vulnerable. OpenSSL itself uses these
5198 functions when printing out human-readable dumps of ASN.1 data. Therefore
5199 applications that print this data could be vulnerable if the data is from
5200 untrusted sources. OpenSSL command line applications could also be
5201 vulnerable where they print out ASN.1 data, or if untrusted data is passed
5202 as command line arguments.
5204 Libssl is not considered directly vulnerable. Additionally certificates etc
5205 received via remote connections via libssl are also unlikely to be able to
5206 trigger these issues because of message size limits enforced within libssl.
5208 This issue was reported to OpenSSL Guido Vranken.
5213 * Side channel attack on modular exponentiation
5215 A side-channel attack was found which makes use of cache-bank conflicts on
5216 the Intel Sandy-Bridge microarchitecture which could lead to the recovery
5217 of RSA keys. The ability to exploit this issue is limited as it relies on
5218 an attacker who has control of code in a thread running on the same
5219 hyper-threaded core as the victim thread which is performing decryptions.
5221 This issue was reported to OpenSSL by Yuval Yarom, The University of
5222 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and
5223 Nadia Heninger, University of Pennsylvania with more information at
5224 <http://cachebleed.info>.
5229 * Change the `req` command to generate a 2048-bit RSA/DSA key by default,
5230 if no keysize is specified with default_bits. This fixes an
5231 omission in an earlier change that changed all RSA/DSA key generation
5232 commands to use 2048 bits by default.
5236 ### Changes between 1.0.2e and 1.0.2f [28 Jan 2016]
5238 * DH small subgroups
5240 Historically OpenSSL only ever generated DH parameters based on "safe"
5241 primes. More recently (in version 1.0.2) support was provided for
5242 generating X9.42 style parameter files such as those required for RFC 5114
5243 support. The primes used in such files may not be "safe". Where an
5244 application is using DH configured with parameters based on primes that are
5245 not "safe" then an attacker could use this fact to find a peer's private
5246 DH exponent. This attack requires that the attacker complete multiple
5247 handshakes in which the peer uses the same private DH exponent. For example
5248 this could be used to discover a TLS server's private DH exponent if it's
5249 reusing the private DH exponent or it's using a static DH ciphersuite.
5251 OpenSSL provides the option SSL_OP_SINGLE_DH_USE for ephemeral DH (DHE) in
5252 TLS. It is not on by default. If the option is not set then the server
5253 reuses the same private DH exponent for the life of the server process and
5254 would be vulnerable to this attack. It is believed that many popular
5255 applications do set this option and would therefore not be at risk.
5257 The fix for this issue adds an additional check where a "q" parameter is
5258 available (as is the case in X9.42 based parameters). This detects the
5259 only known attack, and is the only possible defense for static DH
5260 ciphersuites. This could have some performance impact.
5262 Additionally the SSL_OP_SINGLE_DH_USE option has been switched on by
5263 default and cannot be disabled. This could have some performance impact.
5265 This issue was reported to OpenSSL by Antonio Sanso (Adobe).
5270 * SSLv2 doesn't block disabled ciphers
5272 A malicious client can negotiate SSLv2 ciphers that have been disabled on
5273 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
5274 been disabled, provided that the SSLv2 protocol was not also disabled via
5277 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
5278 and Sebastian Schinzel.
5283 ### Changes between 1.0.2d and 1.0.2e [3 Dec 2015]
5285 * BN_mod_exp may produce incorrect results on x86_64
5287 There is a carry propagating bug in the x86_64 Montgomery squaring
5288 procedure. No EC algorithms are affected. Analysis suggests that attacks
5289 against RSA and DSA as a result of this defect would be very difficult to
5290 perform and are not believed likely. Attacks against DH are considered just
5291 feasible (although very difficult) because most of the work necessary to
5292 deduce information about a private key may be performed offline. The amount
5293 of resources required for such an attack would be very significant and
5294 likely only accessible to a limited number of attackers. An attacker would
5295 additionally need online access to an unpatched system using the target
5296 private key in a scenario with persistent DH parameters and a private
5297 key that is shared between multiple clients. For example this can occur by
5298 default in OpenSSL DHE based SSL/TLS ciphersuites.
5300 This issue was reported to OpenSSL by Hanno Böck.
5305 * Certificate verify crash with missing PSS parameter
5307 The signature verification routines will crash with a NULL pointer
5308 dereference if presented with an ASN.1 signature using the RSA PSS
5309 algorithm and absent mask generation function parameter. Since these
5310 routines are used to verify certificate signature algorithms this can be
5311 used to crash any certificate verification operation and exploited in a
5312 DoS attack. Any application which performs certificate verification is
5313 vulnerable including OpenSSL clients and servers which enable client
5316 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
5321 * X509_ATTRIBUTE memory leak
5323 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
5324 memory. This structure is used by the PKCS#7 and CMS routines so any
5325 application which reads PKCS#7 or CMS data from untrusted sources is
5326 affected. SSL/TLS is not affected.
5328 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
5334 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
5335 This changes the decoding behaviour for some invalid messages,
5336 though the change is mostly in the more lenient direction, and
5337 legacy behaviour is preserved as much as possible.
5341 * In DSA_generate_parameters_ex, if the provided seed is too short,
5344 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
5346 ### Changes between 1.0.2c and 1.0.2d [9 Jul 2015]
5348 * Alternate chains certificate forgery
5350 During certificate verification, OpenSSL will attempt to find an
5351 alternative certificate chain if the first attempt to build such a chain
5352 fails. An error in the implementation of this logic can mean that an
5353 attacker could cause certain checks on untrusted certificates to be
5354 bypassed, such as the CA flag, enabling them to use a valid leaf
5355 certificate to act as a CA and "issue" an invalid certificate.
5357 This issue was reported to OpenSSL by Adam Langley/David Benjamin
5362 ### Changes between 1.0.2b and 1.0.2c [12 Jun 2015]
5364 * Fix HMAC ABI incompatibility. The previous version introduced an ABI
5365 incompatibility in the handling of HMAC. The previous ABI has now been
5370 ### Changes between 1.0.2a and 1.0.2b [11 Jun 2015]
5372 * Malformed ECParameters causes infinite loop
5374 When processing an ECParameters structure OpenSSL enters an infinite loop
5375 if the curve specified is over a specially malformed binary polynomial
5378 This can be used to perform denial of service against any
5379 system which processes public keys, certificate requests or
5380 certificates. This includes TLS clients and TLS servers with
5381 client authentication enabled.
5383 This issue was reported to OpenSSL by Joseph Barr-Pixton.
5388 * Exploitable out-of-bounds read in X509_cmp_time
5390 X509_cmp_time does not properly check the length of the ASN1_TIME
5391 string and can read a few bytes out of bounds. In addition,
5392 X509_cmp_time accepts an arbitrary number of fractional seconds in the
5395 An attacker can use this to craft malformed certificates and CRLs of
5396 various sizes and potentially cause a segmentation fault, resulting in
5397 a DoS on applications that verify certificates or CRLs. TLS clients
5398 that verify CRLs are affected. TLS clients and servers with client
5399 authentication enabled may be affected if they use custom verification
5402 This issue was reported to OpenSSL by Robert Swiecki (Google), and
5403 independently by Hanno Böck.
5408 * PKCS7 crash with missing EnvelopedContent
5410 The PKCS#7 parsing code does not handle missing inner EncryptedContent
5411 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
5412 with missing content and trigger a NULL pointer dereference on parsing.
5414 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
5415 structures from untrusted sources are affected. OpenSSL clients and
5416 servers are not affected.
5418 This issue was reported to OpenSSL by Michal Zalewski (Google).
5423 * CMS verify infinite loop with unknown hash function
5425 When verifying a signedData message the CMS code can enter an infinite loop
5426 if presented with an unknown hash function OID. This can be used to perform
5427 denial of service against any system which verifies signedData messages using
5429 This issue was reported to OpenSSL by Johannes Bauer.
5434 * Race condition handling NewSessionTicket
5436 If a NewSessionTicket is received by a multi-threaded client when attempting to
5437 reuse a previous ticket then a race condition can occur potentially leading to
5438 a double free of the ticket data.
5443 * Only support 256-bit or stronger elliptic curves with the
5444 'ecdh_auto' setting (server) or by default (client). Of supported
5445 curves, prefer P-256 (both).
5449 ### Changes between 1.0.2 and 1.0.2a [19 Mar 2015]
5451 * ClientHello sigalgs DoS fix
5453 If a client connects to an OpenSSL 1.0.2 server and renegotiates with an
5454 invalid signature algorithms extension a NULL pointer dereference will
5455 occur. This can be exploited in a DoS attack against the server.
5457 This issue was was reported to OpenSSL by David Ramos of Stanford
5461 *Stephen Henson and Matt Caswell*
5463 * Multiblock corrupted pointer fix
5465 OpenSSL 1.0.2 introduced the "multiblock" performance improvement. This
5466 feature only applies on 64 bit x86 architecture platforms that support AES
5467 NI instructions. A defect in the implementation of "multiblock" can cause
5468 OpenSSL's internal write buffer to become incorrectly set to NULL when
5469 using non-blocking IO. Typically, when the user application is using a
5470 socket BIO for writing, this will only result in a failed connection.
5471 However if some other BIO is used then it is likely that a segmentation
5472 fault will be triggered, thus enabling a potential DoS attack.
5474 This issue was reported to OpenSSL by Daniel Danner and Rainer Mueller.
5479 * Segmentation fault in DTLSv1_listen fix
5481 The DTLSv1_listen function is intended to be stateless and processes the
5482 initial ClientHello from many peers. It is common for user code to loop
5483 over the call to DTLSv1_listen until a valid ClientHello is received with
5484 an associated cookie. A defect in the implementation of DTLSv1_listen means
5485 that state is preserved in the SSL object from one invocation to the next
5486 that can lead to a segmentation fault. Errors processing the initial
5487 ClientHello can trigger this scenario. An example of such an error could be
5488 that a DTLS1.0 only client is attempting to connect to a DTLS1.2 only
5491 This issue was reported to OpenSSL by Per Allansson.
5496 * Segmentation fault in ASN1_TYPE_cmp fix
5498 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
5499 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
5500 certificate signature algorithm consistency this can be used to crash any
5501 certificate verification operation and exploited in a DoS attack. Any
5502 application which performs certificate verification is vulnerable including
5503 OpenSSL clients and servers which enable client authentication.
5508 * Segmentation fault for invalid PSS parameters fix
5510 The signature verification routines will crash with a NULL pointer
5511 dereference if presented with an ASN.1 signature using the RSA PSS
5512 algorithm and invalid parameters. Since these routines are used to verify
5513 certificate signature algorithms this can be used to crash any
5514 certificate verification operation and exploited in a DoS attack. Any
5515 application which performs certificate verification is vulnerable including
5516 OpenSSL clients and servers which enable client authentication.
5518 This issue was was reported to OpenSSL by Brian Carpenter.
5523 * ASN.1 structure reuse memory corruption fix
5525 Reusing a structure in ASN.1 parsing may allow an attacker to cause
5526 memory corruption via an invalid write. Such reuse is and has been
5527 strongly discouraged and is believed to be rare.
5529 Applications that parse structures containing CHOICE or ANY DEFINED BY
5530 components may be affected. Certificate parsing (d2i_X509 and related
5531 functions) are however not affected. OpenSSL clients and servers are
5537 * PKCS7 NULL pointer dereferences fix
5539 The PKCS#7 parsing code does not handle missing outer ContentInfo
5540 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
5541 missing content and trigger a NULL pointer dereference on parsing.
5543 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
5544 otherwise parse PKCS#7 structures from untrusted sources are
5545 affected. OpenSSL clients and servers are not affected.
5547 This issue was reported to OpenSSL by Michal Zalewski (Google).
5552 * DoS via reachable assert in SSLv2 servers fix
5554 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
5555 servers that both support SSLv2 and enable export cipher suites by sending
5556 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
5558 This issue was discovered by Sean Burford (Google) and Emilia Käsper
5559 (OpenSSL development team).
5564 * Empty CKE with client auth and DHE fix
5566 If client auth is used then a server can seg fault in the event of a DHE
5567 ciphersuite being selected and a zero length ClientKeyExchange message
5568 being sent by the client. This could be exploited in a DoS attack.
5573 * Handshake with unseeded PRNG fix
5575 Under certain conditions an OpenSSL 1.0.2 client can complete a handshake
5576 with an unseeded PRNG. The conditions are:
5577 - The client is on a platform where the PRNG has not been seeded
5578 automatically, and the user has not seeded manually
5579 - A protocol specific client method version has been used (i.e. not
5580 SSL_client_methodv23)
5581 - A ciphersuite is used that does not require additional random data from
5582 the PRNG beyond the initial ClientHello client random (e.g. PSK-RC4-SHA).
5584 If the handshake succeeds then the client random that has been used will
5585 have been generated from a PRNG with insufficient entropy and therefore the
5586 output may be predictable.
5588 For example using the following command with an unseeded openssl will
5589 succeed on an unpatched platform:
5591 openssl s_client -psk 1a2b3c4d -tls1_2 -cipher PSK-RC4-SHA
5596 * Use After Free following d2i_ECPrivatekey error fix
5598 A malformed EC private key file consumed via the d2i_ECPrivateKey function
5599 could cause a use after free condition. This, in turn, could cause a double
5600 free in several private key parsing functions (such as d2i_PrivateKey
5601 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
5602 for applications that receive EC private keys from untrusted
5603 sources. This scenario is considered rare.
5605 This issue was discovered by the BoringSSL project and fixed in their
5611 * X509_to_X509_REQ NULL pointer deref fix
5613 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
5614 the certificate key is invalid. This function is rarely used in practice.
5616 This issue was discovered by Brian Carpenter.
5621 * Removed the export ciphers from the DEFAULT ciphers
5625 ### Changes between 1.0.1l and 1.0.2 [22 Jan 2015]
5627 * Facilitate "universal" ARM builds targeting range of ARM ISAs, e.g.
5628 ARMv5 through ARMv8, as opposite to "locking" it to single one.
5629 So far those who have to target multiple platforms would compromise
5630 and argue that binary targeting say ARMv5 would still execute on
5631 ARMv8. "Universal" build resolves this compromise by providing
5632 near-optimal performance even on newer platforms.
5636 * Accelerated NIST P-256 elliptic curve implementation for x86_64
5637 (other platforms pending).
5639 *Shay Gueron & Vlad Krasnov (Intel Corp), Andy Polyakov*
5641 * Add support for the SignedCertificateTimestampList certificate and
5642 OCSP response extensions from RFC6962.
5646 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
5647 for corner cases. (Certain input points at infinity could lead to
5648 bogus results, with non-infinity inputs mapped to infinity too.)
5652 * Initial support for PowerISA 2.0.7, first implemented in POWER8.
5653 This covers AES, SHA256/512 and GHASH. "Initial" means that most
5654 common cases are optimized and there still is room for further
5655 improvements. Vector Permutation AES for Altivec is also added.
5659 * Add support for little-endian ppc64 Linux target.
5661 *Marcelo Cerri (IBM)*
5663 * Initial support for AMRv8 ISA crypto extensions. This covers AES,
5664 SHA1, SHA256 and GHASH. "Initial" means that most common cases
5665 are optimized and there still is room for further improvements.
5666 Both 32- and 64-bit modes are supported.
5668 *Andy Polyakov, Ard Biesheuvel (Linaro)*
5670 * Improved ARMv7 NEON support.
5674 * Support for SPARC Architecture 2011 crypto extensions, first
5675 implemented in SPARC T4. This covers AES, DES, Camellia, SHA1,
5676 SHA256/512, MD5, GHASH and modular exponentiation.
5678 *Andy Polyakov, David Miller*
5680 * Accelerated modular exponentiation for Intel processors, a.k.a.
5683 *Shay Gueron & Vlad Krasnov (Intel Corp)*
5685 * Support for new and upcoming Intel processors, including AVX2,
5686 BMI and SHA ISA extensions. This includes additional "stitched"
5687 implementations, AESNI-SHA256 and GCM, and multi-buffer support
5690 This work was sponsored by Intel Corp.
5694 * Support for DTLS 1.2. This adds two sets of DTLS methods: DTLS_*_method()
5695 supports both DTLS 1.2 and 1.0 and should use whatever version the peer
5696 supports and DTLSv1_2_*_method() which supports DTLS 1.2 only.
5700 * Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
5701 this fixes a limitation in previous versions of OpenSSL.
5705 * Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
5706 MGF1 digest and OAEP label.
5710 * Add EVP support for key wrapping algorithms, to avoid problems with
5711 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
5712 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
5713 algorithms and include tests cases.
5717 * Add functions to allocate and set the fields of an ECDSA_METHOD
5720 *Douglas E. Engert, Steve Henson*
5722 * New functions OPENSSL_gmtime_diff and ASN1_TIME_diff to find the
5723 difference in days and seconds between two tm or ASN1_TIME structures.
5727 * Add -rev test option to s_server to just reverse order of characters
5728 received by client and send back to server. Also prints an abbreviated
5729 summary of the connection parameters.
5733 * New option -brief for s_client and s_server to print out a brief summary
5734 of connection parameters.
5738 * Add callbacks for arbitrary TLS extensions.
5740 *Trevor Perrin <trevp@trevp.net> and Ben Laurie*
5742 * New option -crl_download in several openssl utilities to download CRLs
5743 from CRLDP extension in certificates.
5747 * New options -CRL and -CRLform for s_client and s_server for CRLs.
5751 * New function X509_CRL_diff to generate a delta CRL from the difference
5752 of two full CRLs. Add support to "crl" utility.
5756 * New functions to set lookup_crls function and to retrieve
5757 X509_STORE from X509_STORE_CTX.
5761 * Print out deprecated issuer and subject unique ID fields in
5766 * Extend OCSP I/O functions so they can be used for simple general purpose
5767 HTTP as well as OCSP. New wrapper function which can be used to download
5768 CRLs using the OCSP API.
5772 * Delegate command line handling in s_client/s_server to SSL_CONF APIs.
5776 * `SSL_CONF*` functions. These provide a common framework for application
5777 configuration using configuration files or command lines.
5781 * SSL/TLS tracing code. This parses out SSL/TLS records using the
5782 message callback and prints the results. Needs compile time option
5783 "enable-ssl-trace". New options to s_client and s_server to enable
5788 * New ctrl and macro to retrieve supported points extensions.
5789 Print out extension in s_server and s_client.
5793 * New functions to retrieve certificate signature and signature
5798 * Add functions to retrieve and manipulate the raw cipherlist sent by a
5803 * New Suite B modes for TLS code. These use and enforce the requirements
5804 of RFC6460: restrict ciphersuites, only permit Suite B algorithms and
5805 only use Suite B curves. The Suite B modes can be set by using the
5806 strings "SUITEB128", "SUITEB192" or "SUITEB128ONLY" for the cipherstring.
5810 * New chain verification flags for Suite B levels of security. Check
5811 algorithms are acceptable when flags are set in X509_verify_cert.
5815 * Make tls1_check_chain return a set of flags indicating checks passed
5816 by a certificate chain. Add additional tests to handle client
5817 certificates: checks for matching certificate type and issuer name
5822 * If an attempt is made to use a signature algorithm not in the peer
5823 preference list abort the handshake. If client has no suitable
5824 signature algorithms in response to a certificate request do not
5825 use the certificate.
5829 * If server EC tmp key is not in client preference list abort handshake.
5833 * Add support for certificate stores in CERT structure. This makes it
5834 possible to have different stores per SSL structure or one store in
5835 the parent SSL_CTX. Include distinct stores for certificate chain
5836 verification and chain building. New ctrl SSL_CTRL_BUILD_CERT_CHAIN
5837 to build and store a certificate chain in CERT structure: returning
5838 an error if the chain cannot be built: this will allow applications
5839 to test if a chain is correctly configured.
5841 Note: if the CERT based stores are not set then the parent SSL_CTX
5842 store is used to retain compatibility with existing behaviour.
5846 * New function ssl_set_client_disabled to set a ciphersuite disabled
5847 mask based on the current session, check mask when sending client
5848 hello and checking the requested ciphersuite.
5852 * New ctrls to retrieve and set certificate types in a certificate
5853 request message. Print out received values in s_client. If certificate
5854 types is not set with custom values set sensible values based on
5855 supported signature algorithms.
5859 * Support for distinct client and server supported signature algorithms.
5863 * Add certificate callback. If set this is called whenever a certificate
5864 is required by client or server. An application can decide which
5865 certificate chain to present based on arbitrary criteria: for example
5866 supported signature algorithms. Add very simple example to s_server.
5867 This fixes many of the problems and restrictions of the existing client
5868 certificate callback: for example you can now clear an existing
5869 certificate and specify the whole chain.
5873 * Add new "valid_flags" field to CERT_PKEY structure which determines what
5874 the certificate can be used for (if anything). Set valid_flags field
5875 in new tls1_check_chain function. Simplify ssl_set_cert_masks which used
5876 to have similar checks in it.
5878 Add new "cert_flags" field to CERT structure and include a "strict mode".
5879 This enforces some TLS certificate requirements (such as only permitting
5880 certificate signature algorithms contained in the supported algorithms
5881 extension) which some implementations ignore: this option should be used
5882 with caution as it could cause interoperability issues.
5886 * Update and tidy signature algorithm extension processing. Work out
5887 shared signature algorithms based on preferences and peer algorithms
5888 and print them out in s_client and s_server. Abort handshake if no
5889 shared signature algorithms.
5893 * Add new functions to allow customised supported signature algorithms
5894 for SSL and SSL_CTX structures. Add options to s_client and s_server
5899 * New function SSL_certs_clear() to delete all references to certificates
5900 from an SSL structure. Before this once a certificate had been added
5901 it couldn't be removed.
5905 * Integrate hostname, email address and IP address checking with certificate
5906 verification. New verify options supporting checking in openssl utility.
5910 * Fixes and wildcard matching support to hostname and email checking
5911 functions. Add manual page.
5913 *Florian Weimer (Red Hat Product Security Team)*
5915 * New functions to check a hostname email or IP address against a
5916 certificate. Add options x509 utility to print results of checks against
5921 * Fix OCSP checking.
5923 *Rob Stradling <rob.stradling@comodo.com> and Ben Laurie*
5925 * Initial experimental support for explicitly trusted non-root CAs.
5926 OpenSSL still tries to build a complete chain to a root but if an
5927 intermediate CA has a trust setting included that is used. The first
5928 setting is used: whether to trust (e.g., -addtrust option to the x509
5933 * Add -trusted_first option which attempts to find certificates in the
5934 trusted store even if an untrusted chain is also supplied.
5938 * MIPS assembly pack updates: support for MIPS32r2 and SmartMIPS ASE,
5939 platform support for Linux and Android.
5943 * Support for linux-x32, ILP32 environment in x86_64 framework.
5947 * Experimental multi-implementation support for FIPS capable OpenSSL.
5948 When in FIPS mode the approved implementations are used as normal,
5949 when not in FIPS mode the internal unapproved versions are used instead.
5950 This means that the FIPS capable OpenSSL isn't forced to use the
5951 (often lower performance) FIPS implementations outside FIPS mode.
5955 * Transparently support X9.42 DH parameters when calling
5956 PEM_read_bio_DHparameters. This means existing applications can handle
5957 the new parameter format automatically.
5961 * Initial experimental support for X9.42 DH parameter format: mainly
5962 to support use of 'q' parameter for RFC5114 parameters.
5966 * Add DH parameters from RFC5114 including test data to dhtest.
5970 * Support for automatic EC temporary key parameter selection. If enabled
5971 the most preferred EC parameters are automatically used instead of
5972 hardcoded fixed parameters. Now a server just has to call:
5973 SSL_CTX_set_ecdh_auto(ctx, 1) and the server will automatically
5974 support ECDH and use the most appropriate parameters.
5978 * Enhance and tidy EC curve and point format TLS extension code. Use
5979 static structures instead of allocation if default values are used.
5980 New ctrls to set curves we wish to support and to retrieve shared curves.
5981 Print out shared curves in s_server. New options to s_server and s_client
5982 to set list of supported curves.
5986 * New ctrls to retrieve supported signature algorithms and
5987 supported curve values as an array of NIDs. Extend openssl utility
5988 to print out received values.
5992 * Add new APIs EC_curve_nist2nid and EC_curve_nid2nist which convert
5993 between NIDs and the more common NIST names such as "P-256". Enhance
5994 ecparam utility and ECC method to recognise the NIST names for curves.
5998 * Enhance SSL/TLS certificate chain handling to support different
5999 chains for each certificate instead of one chain in the parent SSL_CTX.
6003 * Support for fixed DH ciphersuite client authentication: where both
6004 server and client use DH certificates with common parameters.
6008 * Support for fixed DH ciphersuites: those requiring DH server
6013 * New function i2d_re_X509_tbs for re-encoding the TBS portion of
6015 Note: Related 1.0.2-beta specific macros X509_get_cert_info,
6016 X509_CINF_set_modified, X509_CINF_get_issuer, X509_CINF_get_extensions and
6017 X509_CINF_get_signature were reverted post internal team review.
6022 ### Changes between 1.0.1t and 1.0.1u [22 Sep 2016]
6024 * OCSP Status Request extension unbounded memory growth
6026 A malicious client can send an excessively large OCSP Status Request
6027 extension. If that client continually requests renegotiation, sending a
6028 large OCSP Status Request extension each time, then there will be unbounded
6029 memory growth on the server. This will eventually lead to a Denial Of
6030 Service attack through memory exhaustion. Servers with a default
6031 configuration are vulnerable even if they do not support OCSP. Builds using
6032 the "no-ocsp" build time option are not affected.
6034 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6039 * In order to mitigate the SWEET32 attack, the DES ciphers were moved from
6042 This issue was reported to OpenSSL Karthikeyan Bhargavan and Gaetan
6048 * OOB write in MDC2_Update()
6050 An overflow can occur in MDC2_Update() either if called directly or
6051 through the EVP_DigestUpdate() function using MDC2. If an attacker
6052 is able to supply very large amounts of input data after a previous
6053 call to EVP_EncryptUpdate() with a partial block then a length check
6054 can overflow resulting in a heap corruption.
6056 The amount of data needed is comparable to SIZE_MAX which is impractical
6059 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6064 * Malformed SHA512 ticket DoS
6066 If a server uses SHA512 for TLS session ticket HMAC it is vulnerable to a
6067 DoS attack where a malformed ticket will result in an OOB read which will
6070 The use of SHA512 in TLS session tickets is comparatively rare as it requires
6071 a custom server callback and ticket lookup mechanism.
6073 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6078 * OOB write in BN_bn2dec()
6080 The function BN_bn2dec() does not check the return value of BN_div_word().
6081 This can cause an OOB write if an application uses this function with an
6082 overly large BIGNUM. This could be a problem if an overly large certificate
6083 or CRL is printed out from an untrusted source. TLS is not affected because
6084 record limits will reject an oversized certificate before it is parsed.
6086 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6091 * OOB read in TS_OBJ_print_bio()
6093 The function TS_OBJ_print_bio() misuses OBJ_obj2txt(): the return value is
6094 the total length the OID text representation would use and not the amount
6095 of data written. This will result in OOB reads when large OIDs are
6098 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6103 * Pointer arithmetic undefined behaviour
6105 Avoid some undefined pointer arithmetic
6107 A common idiom in the codebase is to check limits in the following manner:
6110 Where "p" points to some malloc'd data of SIZE bytes and
6113 "len" here could be from some externally supplied data (e.g. from a TLS
6116 The rules of C pointer arithmetic are such that "p + len" is only well
6117 defined where len <= SIZE. Therefore the above idiom is actually
6118 undefined behaviour.
6120 For example this could cause problems if some malloc implementation
6121 provides an address for "p" such that "p + len" actually overflows for
6122 values of len that are too big and therefore p + len < limit.
6124 This issue was reported to OpenSSL by Guido Vranken
6129 * Constant time flag not preserved in DSA signing
6131 Operations in the DSA signing algorithm should run in constant time in
6132 order to avoid side channel attacks. A flaw in the OpenSSL DSA
6133 implementation means that a non-constant time codepath is followed for
6134 certain operations. This has been demonstrated through a cache-timing
6135 attack to be sufficient for an attacker to recover the private DSA key.
6137 This issue was reported by César Pereida (Aalto University), Billy Brumley
6138 (Tampere University of Technology), and Yuval Yarom (The University of
6139 Adelaide and NICTA).
6144 * DTLS buffered message DoS
6146 In a DTLS connection where handshake messages are delivered out-of-order
6147 those messages that OpenSSL is not yet ready to process will be buffered
6148 for later use. Under certain circumstances, a flaw in the logic means that
6149 those messages do not get removed from the buffer even though the handshake
6150 has been completed. An attacker could force up to approx. 15 messages to
6151 remain in the buffer when they are no longer required. These messages will
6152 be cleared when the DTLS connection is closed. The default maximum size for
6153 a message is 100k. Therefore the attacker could force an additional 1500k
6154 to be consumed per connection. By opening many simulataneous connections an
6155 attacker could cause a DoS attack through memory exhaustion.
6157 This issue was reported to OpenSSL by Quan Luo.
6162 * DTLS replay protection DoS
6164 A flaw in the DTLS replay attack protection mechanism means that records
6165 that arrive for future epochs update the replay protection "window" before
6166 the MAC for the record has been validated. This could be exploited by an
6167 attacker by sending a record for the next epoch (which does not have to
6168 decrypt or have a valid MAC), with a very large sequence number. This means
6169 that all subsequent legitimate packets are dropped causing a denial of
6170 service for a specific DTLS connection.
6172 This issue was reported to OpenSSL by the OCAP audit team.
6177 * Certificate message OOB reads
6179 In OpenSSL 1.0.2 and earlier some missing message length checks can result
6180 in OOB reads of up to 2 bytes beyond an allocated buffer. There is a
6181 theoretical DoS risk but this has not been observed in practice on common
6184 The messages affected are client certificate, client certificate request
6185 and server certificate. As a result the attack can only be performed
6186 against a client or a server which enables client authentication.
6188 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
6193 ### Changes between 1.0.1s and 1.0.1t [3 May 2016]
6195 * Prevent padding oracle in AES-NI CBC MAC check
6197 A MITM attacker can use a padding oracle attack to decrypt traffic
6198 when the connection uses an AES CBC cipher and the server support
6201 This issue was introduced as part of the fix for Lucky 13 padding
6202 attack ([CVE-2013-0169]). The padding check was rewritten to be in
6203 constant time by making sure that always the same bytes are read and
6204 compared against either the MAC or padding bytes. But it no longer
6205 checked that there was enough data to have both the MAC and padding
6208 This issue was reported by Juraj Somorovsky using TLS-Attacker.
6213 * Fix EVP_EncodeUpdate overflow
6215 An overflow can occur in the EVP_EncodeUpdate() function which is used for
6216 Base64 encoding of binary data. If an attacker is able to supply very large
6217 amounts of input data then a length check can overflow resulting in a heap
6220 Internally to OpenSSL the EVP_EncodeUpdate() function is primarly used by
6221 the `PEM_write_bio*` family of functions. These are mainly used within the
6222 OpenSSL command line applications, so any application which processes data
6223 from an untrusted source and outputs it as a PEM file should be considered
6224 vulnerable to this issue. User applications that call these APIs directly
6225 with large amounts of untrusted data may also be vulnerable.
6227 This issue was reported by Guido Vranken.
6232 * Fix EVP_EncryptUpdate overflow
6234 An overflow can occur in the EVP_EncryptUpdate() function. If an attacker
6235 is able to supply very large amounts of input data after a previous call to
6236 EVP_EncryptUpdate() with a partial block then a length check can overflow
6237 resulting in a heap corruption. Following an analysis of all OpenSSL
6238 internal usage of the EVP_EncryptUpdate() function all usage is one of two
6239 forms. The first form is where the EVP_EncryptUpdate() call is known to be
6240 the first called function after an EVP_EncryptInit(), and therefore that
6241 specific call must be safe. The second form is where the length passed to
6242 EVP_EncryptUpdate() can be seen from the code to be some small value and
6243 therefore there is no possibility of an overflow. Since all instances are
6244 one of these two forms, it is believed that there can be no overflows in
6245 internal code due to this problem. It should be noted that
6246 EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths.
6247 Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances
6248 of these calls have also been analysed too and it is believed there are no
6249 instances in internal usage where an overflow could occur.
6251 This issue was reported by Guido Vranken.
6256 * Prevent ASN.1 BIO excessive memory allocation
6258 When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio()
6259 a short invalid encoding can casuse allocation of large amounts of memory
6260 potentially consuming excessive resources or exhausting memory.
6262 Any application parsing untrusted data through d2i BIO functions is
6263 affected. The memory based functions such as d2i_X509() are *not* affected.
6264 Since the memory based functions are used by the TLS library, TLS
6265 applications are not affected.
6267 This issue was reported by Brian Carpenter.
6274 ASN1 Strings that are over 1024 bytes can cause an overread in applications
6275 using the X509_NAME_oneline() function on EBCDIC systems. This could result
6276 in arbitrary stack data being returned in the buffer.
6278 This issue was reported by Guido Vranken.
6283 * Modify behavior of ALPN to invoke callback after SNI/servername
6284 callback, such that updates to the SSL_CTX affect ALPN.
6288 * Remove LOW from the DEFAULT cipher list. This removes singles DES from the
6293 * Only remove the SSLv2 methods with the no-ssl2-method option. When the
6294 methods are enabled and ssl2 is disabled the methods return NULL.
6298 ### Changes between 1.0.1r and 1.0.1s [1 Mar 2016]
6300 * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL.
6301 Builds that are not configured with "enable-weak-ssl-ciphers" will not
6302 provide any "EXPORT" or "LOW" strength ciphers.
6306 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2
6307 is by default disabled at build-time. Builds that are not configured with
6308 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used,
6309 users who want to negotiate SSLv2 via the version-flexible SSLv23_method()
6310 will need to explicitly call either of:
6312 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2);
6314 SSL_clear_options(ssl, SSL_OP_NO_SSLv2);
6316 as appropriate. Even if either of those is used, or the application
6317 explicitly uses the version-specific SSLv2_method() or its client and
6318 server variants, SSLv2 ciphers vulnerable to exhaustive search key
6319 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT
6320 ciphers, and SSLv2 56-bit DES are no longer available.
6325 * Fix a double-free in DSA code
6327 A double free bug was discovered when OpenSSL parses malformed DSA private
6328 keys and could lead to a DoS attack or memory corruption for applications
6329 that receive DSA private keys from untrusted sources. This scenario is
6332 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using
6338 * Disable SRP fake user seed to address a server memory leak.
6340 Add a new method SRP_VBASE_get1_by_user that handles the seed properly.
6342 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
6343 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user
6344 was changed to ignore the "fake user" SRP seed, even if the seed
6347 Users should use SRP_VBASE_get1_by_user instead. Note that in
6348 SRP_VBASE_get1_by_user, caller must free the returned value. Note
6349 also that even though configuring the SRP seed attempts to hide
6350 invalid usernames by continuing the handshake with fake
6351 credentials, this behaviour is not constant time and no strong
6352 guarantees are made that the handshake is indistinguishable from
6353 that of a valid user.
6358 * Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption
6360 In the BN_hex2bn function the number of hex digits is calculated using an
6361 int value `i`. Later `bn_expand` is called with a value of `i * 4`. For
6362 large values of `i` this can result in `bn_expand` not allocating any
6363 memory because `i * 4` is negative. This can leave the internal BIGNUM data
6364 field as NULL leading to a subsequent NULL ptr deref. For very large values
6365 of `i`, the calculation `i * 4` could be a positive value smaller than `i`.
6366 In this case memory is allocated to the internal BIGNUM data field, but it
6367 is insufficiently sized leading to heap corruption. A similar issue exists
6368 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn
6369 is ever called by user applications with very large untrusted hex/dec data.
6370 This is anticipated to be a rare occurrence.
6372 All OpenSSL internal usage of these functions use data that is not expected
6373 to be untrusted, e.g. config file data or application command line
6374 arguments. If user developed applications generate config file data based
6375 on untrusted data then it is possible that this could also lead to security
6376 consequences. This is also anticipated to be rare.
6378 This issue was reported to OpenSSL by Guido Vranken.
6383 * Fix memory issues in `BIO_*printf` functions
6385 The internal `fmtstr` function used in processing a "%s" format string in
6386 the `BIO_*printf` functions could overflow while calculating the length of a
6387 string and cause an OOB read when printing very long strings.
6389 Additionally the internal `doapr_outch` function can attempt to write to an
6390 OOB memory location (at an offset from the NULL pointer) in the event of a
6391 memory allocation failure. In 1.0.2 and below this could be caused where
6392 the size of a buffer to be allocated is greater than INT_MAX. E.g. this
6393 could be in processing a very long "%s" format string. Memory leaks can
6396 The first issue may mask the second issue dependent on compiler behaviour.
6397 These problems could enable attacks where large amounts of untrusted data
6398 is passed to the `BIO_*printf` functions. If applications use these functions
6399 in this way then they could be vulnerable. OpenSSL itself uses these
6400 functions when printing out human-readable dumps of ASN.1 data. Therefore
6401 applications that print this data could be vulnerable if the data is from
6402 untrusted sources. OpenSSL command line applications could also be
6403 vulnerable where they print out ASN.1 data, or if untrusted data is passed
6404 as command line arguments.
6406 Libssl is not considered directly vulnerable. Additionally certificates etc
6407 received via remote connections via libssl are also unlikely to be able to
6408 trigger these issues because of message size limits enforced within libssl.
6410 This issue was reported to OpenSSL Guido Vranken.
6415 * Side channel attack on modular exponentiation
6417 A side-channel attack was found which makes use of cache-bank conflicts on
6418 the Intel Sandy-Bridge microarchitecture which could lead to the recovery
6419 of RSA keys. The ability to exploit this issue is limited as it relies on
6420 an attacker who has control of code in a thread running on the same
6421 hyper-threaded core as the victim thread which is performing decryptions.
6423 This issue was reported to OpenSSL by Yuval Yarom, The University of
6424 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and
6425 Nadia Heninger, University of Pennsylvania with more information at
6426 <http://cachebleed.info>.
6431 * Change the req command to generate a 2048-bit RSA/DSA key by default,
6432 if no keysize is specified with default_bits. This fixes an
6433 omission in an earlier change that changed all RSA/DSA key generation
6434 commands to use 2048 bits by default.
6438 ### Changes between 1.0.1q and 1.0.1r [28 Jan 2016]
6440 * Protection for DH small subgroup attacks
6442 As a precautionary measure the SSL_OP_SINGLE_DH_USE option has been
6443 switched on by default and cannot be disabled. This could have some
6448 * SSLv2 doesn't block disabled ciphers
6450 A malicious client can negotiate SSLv2 ciphers that have been disabled on
6451 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
6452 been disabled, provided that the SSLv2 protocol was not also disabled via
6455 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
6456 and Sebastian Schinzel.
6461 * Reject DH handshakes with parameters shorter than 1024 bits.
6465 ### Changes between 1.0.1p and 1.0.1q [3 Dec 2015]
6467 * Certificate verify crash with missing PSS parameter
6469 The signature verification routines will crash with a NULL pointer
6470 dereference if presented with an ASN.1 signature using the RSA PSS
6471 algorithm and absent mask generation function parameter. Since these
6472 routines are used to verify certificate signature algorithms this can be
6473 used to crash any certificate verification operation and exploited in a
6474 DoS attack. Any application which performs certificate verification is
6475 vulnerable including OpenSSL clients and servers which enable client
6478 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
6483 * X509_ATTRIBUTE memory leak
6485 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
6486 memory. This structure is used by the PKCS#7 and CMS routines so any
6487 application which reads PKCS#7 or CMS data from untrusted sources is
6488 affected. SSL/TLS is not affected.
6490 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
6496 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
6497 This changes the decoding behaviour for some invalid messages,
6498 though the change is mostly in the more lenient direction, and
6499 legacy behaviour is preserved as much as possible.
6503 * In DSA_generate_parameters_ex, if the provided seed is too short,
6504 use a random seed, as already documented.
6506 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
6508 ### Changes between 1.0.1o and 1.0.1p [9 Jul 2015]
6510 * Alternate chains certificate forgery
6512 During certificate verfification, OpenSSL will attempt to find an
6513 alternative certificate chain if the first attempt to build such a chain
6514 fails. An error in the implementation of this logic can mean that an
6515 attacker could cause certain checks on untrusted certificates to be
6516 bypassed, such as the CA flag, enabling them to use a valid leaf
6517 certificate to act as a CA and "issue" an invalid certificate.
6519 This issue was reported to OpenSSL by Adam Langley/David Benjamin
6525 * Race condition handling PSK identify hint
6527 If PSK identity hints are received by a multi-threaded client then
6528 the values are wrongly updated in the parent SSL_CTX structure. This can
6529 result in a race condition potentially leading to a double free of the
6535 ### Changes between 1.0.1n and 1.0.1o [12 Jun 2015]
6537 * Fix HMAC ABI incompatibility. The previous version introduced an ABI
6538 incompatibility in the handling of HMAC. The previous ABI has now been
6541 ### Changes between 1.0.1m and 1.0.1n [11 Jun 2015]
6543 * Malformed ECParameters causes infinite loop
6545 When processing an ECParameters structure OpenSSL enters an infinite loop
6546 if the curve specified is over a specially malformed binary polynomial
6549 This can be used to perform denial of service against any
6550 system which processes public keys, certificate requests or
6551 certificates. This includes TLS clients and TLS servers with
6552 client authentication enabled.
6554 This issue was reported to OpenSSL by Joseph Barr-Pixton.
6559 * Exploitable out-of-bounds read in X509_cmp_time
6561 X509_cmp_time does not properly check the length of the ASN1_TIME
6562 string and can read a few bytes out of bounds. In addition,
6563 X509_cmp_time accepts an arbitrary number of fractional seconds in the
6566 An attacker can use this to craft malformed certificates and CRLs of
6567 various sizes and potentially cause a segmentation fault, resulting in
6568 a DoS on applications that verify certificates or CRLs. TLS clients
6569 that verify CRLs are affected. TLS clients and servers with client
6570 authentication enabled may be affected if they use custom verification
6573 This issue was reported to OpenSSL by Robert Swiecki (Google), and
6574 independently by Hanno Böck.
6579 * PKCS7 crash with missing EnvelopedContent
6581 The PKCS#7 parsing code does not handle missing inner EncryptedContent
6582 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
6583 with missing content and trigger a NULL pointer dereference on parsing.
6585 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
6586 structures from untrusted sources are affected. OpenSSL clients and
6587 servers are not affected.
6589 This issue was reported to OpenSSL by Michal Zalewski (Google).
6594 * CMS verify infinite loop with unknown hash function
6596 When verifying a signedData message the CMS code can enter an infinite loop
6597 if presented with an unknown hash function OID. This can be used to perform
6598 denial of service against any system which verifies signedData messages using
6600 This issue was reported to OpenSSL by Johannes Bauer.
6605 * Race condition handling NewSessionTicket
6607 If a NewSessionTicket is received by a multi-threaded client when attempting to
6608 reuse a previous ticket then a race condition can occur potentially leading to
6609 a double free of the ticket data.
6614 * Reject DH handshakes with parameters shorter than 768 bits.
6616 *Kurt Roeckx and Emilia Kasper*
6618 * dhparam: generate 2048-bit parameters by default.
6620 *Kurt Roeckx and Emilia Kasper*
6622 ### Changes between 1.0.1l and 1.0.1m [19 Mar 2015]
6624 * Segmentation fault in ASN1_TYPE_cmp fix
6626 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
6627 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
6628 certificate signature algorithm consistency this can be used to crash any
6629 certificate verification operation and exploited in a DoS attack. Any
6630 application which performs certificate verification is vulnerable including
6631 OpenSSL clients and servers which enable client authentication.
6636 * ASN.1 structure reuse memory corruption fix
6638 Reusing a structure in ASN.1 parsing may allow an attacker to cause
6639 memory corruption via an invalid write. Such reuse is and has been
6640 strongly discouraged and is believed to be rare.
6642 Applications that parse structures containing CHOICE or ANY DEFINED BY
6643 components may be affected. Certificate parsing (d2i_X509 and related
6644 functions) are however not affected. OpenSSL clients and servers are
6650 * PKCS7 NULL pointer dereferences fix
6652 The PKCS#7 parsing code does not handle missing outer ContentInfo
6653 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
6654 missing content and trigger a NULL pointer dereference on parsing.
6656 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
6657 otherwise parse PKCS#7 structures from untrusted sources are
6658 affected. OpenSSL clients and servers are not affected.
6660 This issue was reported to OpenSSL by Michal Zalewski (Google).
6665 * DoS via reachable assert in SSLv2 servers fix
6667 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
6668 servers that both support SSLv2 and enable export cipher suites by sending
6669 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
6671 This issue was discovered by Sean Burford (Google) and Emilia Käsper
6672 (OpenSSL development team).
6677 * Use After Free following d2i_ECPrivatekey error fix
6679 A malformed EC private key file consumed via the d2i_ECPrivateKey function
6680 could cause a use after free condition. This, in turn, could cause a double
6681 free in several private key parsing functions (such as d2i_PrivateKey
6682 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
6683 for applications that receive EC private keys from untrusted
6684 sources. This scenario is considered rare.
6686 This issue was discovered by the BoringSSL project and fixed in their
6692 * X509_to_X509_REQ NULL pointer deref fix
6694 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
6695 the certificate key is invalid. This function is rarely used in practice.
6697 This issue was discovered by Brian Carpenter.
6702 * Removed the export ciphers from the DEFAULT ciphers
6706 ### Changes between 1.0.1k and 1.0.1l [15 Jan 2015]
6708 * Build fixes for the Windows and OpenVMS platforms
6710 *Matt Caswell and Richard Levitte*
6712 ### Changes between 1.0.1j and 1.0.1k [8 Jan 2015]
6714 * Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
6715 message can cause a segmentation fault in OpenSSL due to a NULL pointer
6716 dereference. This could lead to a Denial Of Service attack. Thanks to
6717 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
6722 * Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
6723 dtls1_buffer_record function under certain conditions. In particular this
6724 could occur if an attacker sent repeated DTLS records with the same
6725 sequence number but for the next epoch. The memory leak could be exploited
6726 by an attacker in a Denial of Service attack through memory exhaustion.
6727 Thanks to Chris Mueller for reporting this issue.
6732 * Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
6733 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
6734 method would be set to NULL which could later result in a NULL pointer
6735 dereference. Thanks to Frank Schmirler for reporting this issue.
6740 * Abort handshake if server key exchange message is omitted for ephemeral
6743 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
6744 reporting this issue.
6749 * Remove non-export ephemeral RSA code on client and server. This code
6750 violated the TLS standard by allowing the use of temporary RSA keys in
6751 non-export ciphersuites and could be used by a server to effectively
6752 downgrade the RSA key length used to a value smaller than the server
6753 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
6754 INRIA or reporting this issue.
6759 * Fixed issue where DH client certificates are accepted without verification.
6760 An OpenSSL server will accept a DH certificate for client authentication
6761 without the certificate verify message. This effectively allows a client to
6762 authenticate without the use of a private key. This only affects servers
6763 which trust a client certificate authority which issues certificates
6764 containing DH keys: these are extremely rare and hardly ever encountered.
6765 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
6771 * Ensure that the session ID context of an SSL is updated when its
6772 SSL_CTX is updated via SSL_set_SSL_CTX.
6774 The session ID context is typically set from the parent SSL_CTX,
6775 and can vary with the CTX.
6779 * Fix various certificate fingerprint issues.
6781 By using non-DER or invalid encodings outside the signed portion of a
6782 certificate the fingerprint can be changed without breaking the signature.
6783 Although no details of the signed portion of the certificate can be changed
6784 this can cause problems with some applications: e.g. those using the
6785 certificate fingerprint for blacklists.
6787 1. Reject signatures with non zero unused bits.
6789 If the BIT STRING containing the signature has non zero unused bits reject
6790 the signature. All current signature algorithms require zero unused bits.
6792 2. Check certificate algorithm consistency.
6794 Check the AlgorithmIdentifier inside TBS matches the one in the
6795 certificate signature. NB: this will result in signature failure
6796 errors for some broken certificates.
6798 Thanks to Konrad Kraszewski from Google for reporting this issue.
6800 3. Check DSA/ECDSA signatures use DER.
6802 Re-encode DSA/ECDSA signatures and compare with the original received
6803 signature. Return an error if there is a mismatch.
6805 This will reject various cases including garbage after signature
6806 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
6807 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
6808 (negative or with leading zeroes).
6810 Further analysis was conducted and fixes were developed by Stephen Henson
6811 of the OpenSSL core team.
6817 * Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
6818 results on some platforms, including x86_64. This bug occurs at random
6819 with a very low probability, and is not known to be exploitable in any
6820 way, though its exact impact is difficult to determine. Thanks to Pieter
6821 Wuille (Blockstream) who reported this issue and also suggested an initial
6822 fix. Further analysis was conducted by the OpenSSL development team and
6823 Adam Langley of Google. The final fix was developed by Andy Polyakov of
6824 the OpenSSL core team.
6829 * Do not resume sessions on the server if the negotiated protocol
6830 version does not match the session's version. Resuming with a different
6831 version, while not strictly forbidden by the RFC, is of questionable
6832 sanity and breaks all known clients.
6834 *David Benjamin, Emilia Käsper*
6836 * Tighten handling of the ChangeCipherSpec (CCS) message: reject
6837 early CCS messages during renegotiation. (Note that because
6838 renegotiation is encrypted, this early CCS was not exploitable.)
6842 * Tighten client-side session ticket handling during renegotiation:
6843 ensure that the client only accepts a session ticket if the server sends
6844 the extension anew in the ServerHello. Previously, a TLS client would
6845 reuse the old extension state and thus accept a session ticket if one was
6846 announced in the initial ServerHello.
6848 Similarly, ensure that the client requires a session ticket if one
6849 was advertised in the ServerHello. Previously, a TLS client would
6850 ignore a missing NewSessionTicket message.
6854 ### Changes between 1.0.1i and 1.0.1j [15 Oct 2014]
6858 A flaw in the DTLS SRTP extension parsing code allows an attacker, who
6859 sends a carefully crafted handshake message, to cause OpenSSL to fail
6860 to free up to 64k of memory causing a memory leak. This could be
6861 exploited in a Denial Of Service attack. This issue affects OpenSSL
6862 1.0.1 server implementations for both SSL/TLS and DTLS regardless of
6863 whether SRTP is used or configured. Implementations of OpenSSL that
6864 have been compiled with OPENSSL_NO_SRTP defined are not affected.
6866 The fix was developed by the OpenSSL team.
6871 * Session Ticket Memory Leak.
6873 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
6874 integrity of that ticket is first verified. In the event of a session
6875 ticket integrity check failing, OpenSSL will fail to free memory
6876 causing a memory leak. By sending a large number of invalid session
6877 tickets an attacker could exploit this issue in a Denial Of Service
6883 * Build option no-ssl3 is incomplete.
6885 When OpenSSL is configured with "no-ssl3" as a build option, servers
6886 could accept and complete a SSL 3.0 handshake, and clients could be
6887 configured to send them.
6890 *Akamai and the OpenSSL team*
6892 * Add support for TLS_FALLBACK_SCSV.
6893 Client applications doing fallback retries should call
6894 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
6897 *Adam Langley, Bodo Moeller*
6899 * Add additional DigestInfo checks.
6901 Re-encode DigestInto in DER and check against the original when
6902 verifying RSA signature: this will reject any improperly encoded
6903 DigestInfo structures.
6905 Note: this is a precautionary measure and no attacks are currently known.
6909 ### Changes between 1.0.1h and 1.0.1i [6 Aug 2014]
6911 * Fix SRP buffer overrun vulnerability. Invalid parameters passed to the
6912 SRP code can be overrun an internal buffer. Add sanity check that
6913 g, A, B < N to SRP code.
6915 Thanks to Sean Devlin and Watson Ladd of Cryptography Services, NCC
6916 Group for discovering this issue.
6921 * A flaw in the OpenSSL SSL/TLS server code causes the server to negotiate
6922 TLS 1.0 instead of higher protocol versions when the ClientHello message
6923 is badly fragmented. This allows a man-in-the-middle attacker to force a
6924 downgrade to TLS 1.0 even if both the server and the client support a
6925 higher protocol version, by modifying the client's TLS records.
6927 Thanks to David Benjamin and Adam Langley (Google) for discovering and
6928 researching this issue.
6933 * OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
6934 to a denial of service attack. A malicious server can crash the client
6935 with a null pointer dereference (read) by specifying an anonymous (EC)DH
6936 ciphersuite and sending carefully crafted handshake messages.
6938 Thanks to Felix Gröbert (Google) for discovering and researching this
6944 * By sending carefully crafted DTLS packets an attacker could cause openssl
6945 to leak memory. This can be exploited through a Denial of Service attack.
6946 Thanks to Adam Langley for discovering and researching this issue.
6951 * An attacker can force openssl to consume large amounts of memory whilst
6952 processing DTLS handshake messages. This can be exploited through a
6953 Denial of Service attack.
6954 Thanks to Adam Langley for discovering and researching this issue.
6959 * An attacker can force an error condition which causes openssl to crash
6960 whilst processing DTLS packets due to memory being freed twice. This
6961 can be exploited through a Denial of Service attack.
6962 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
6968 * If a multithreaded client connects to a malicious server using a resumed
6969 session and the server sends an ec point format extension it could write
6970 up to 255 bytes to freed memory.
6972 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
6978 * A malicious server can crash an OpenSSL client with a null pointer
6979 dereference (read) by specifying an SRP ciphersuite even though it was not
6980 properly negotiated with the client. This can be exploited through a
6981 Denial of Service attack.
6983 Thanks to Joonas Kuorilehto and Riku Hietamäki (Codenomicon) for
6984 discovering and researching this issue.
6989 * A flaw in OBJ_obj2txt may cause pretty printing functions such as
6990 X509_name_oneline, X509_name_print_ex et al. to leak some information
6991 from the stack. Applications may be affected if they echo pretty printing
6992 output to the attacker.
6994 Thanks to Ivan Fratric (Google) for discovering this issue.
6997 *Emilia Käsper, and Steve Henson*
6999 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
7000 for corner cases. (Certain input points at infinity could lead to
7001 bogus results, with non-infinity inputs mapped to infinity too.)
7005 ### Changes between 1.0.1g and 1.0.1h [5 Jun 2014]
7007 * Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
7008 handshake can force the use of weak keying material in OpenSSL
7009 SSL/TLS clients and servers.
7011 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
7012 researching this issue. ([CVE-2014-0224])
7014 *KIKUCHI Masashi, Steve Henson*
7016 * Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
7017 OpenSSL DTLS client the code can be made to recurse eventually crashing
7020 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
7023 *Imre Rad, Steve Henson*
7025 * Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
7026 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
7027 client or server. This is potentially exploitable to run arbitrary
7028 code on a vulnerable client or server.
7030 Thanks to Jüri Aedla for reporting this issue. ([CVE-2014-0195])
7032 *Jüri Aedla, Steve Henson*
7034 * Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
7035 are subject to a denial of service attack.
7037 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
7038 this issue. ([CVE-2014-3470])
7040 *Felix Gröbert, Ivan Fratric, Steve Henson*
7042 * Harmonize version and its documentation. -f flag is used to display
7045 *mancha <mancha1@zoho.com>*
7047 * Fix eckey_priv_encode so it immediately returns an error upon a failure
7048 in i2d_ECPrivateKey.
7050 *mancha <mancha1@zoho.com>*
7052 * Fix some double frees. These are not thought to be exploitable.
7054 *mancha <mancha1@zoho.com>*
7056 ### Changes between 1.0.1f and 1.0.1g [7 Apr 2014]
7058 * A missing bounds check in the handling of the TLS heartbeat extension
7059 can be used to reveal up to 64k of memory to a connected client or
7062 Thanks for Neel Mehta of Google Security for discovering this bug and to
7063 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
7064 preparing the fix ([CVE-2014-0160])
7066 *Adam Langley, Bodo Moeller*
7068 * Fix for the attack described in the paper "Recovering OpenSSL
7069 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
7070 by Yuval Yarom and Naomi Benger. Details can be obtained from:
7071 <http://eprint.iacr.org/2014/140>
7073 Thanks to Yuval Yarom and Naomi Benger for discovering this
7074 flaw and to Yuval Yarom for supplying a fix ([CVE-2014-0076])
7076 *Yuval Yarom and Naomi Benger*
7078 * TLS pad extension: draft-agl-tls-padding-03
7080 Workaround for the "TLS hang bug" (see FAQ and PR#2771): if the
7081 TLS client Hello record length value would otherwise be > 255 and
7082 less that 512 pad with a dummy extension containing zeroes so it
7083 is at least 512 bytes long.
7085 *Adam Langley, Steve Henson*
7087 ### Changes between 1.0.1e and 1.0.1f [6 Jan 2014]
7089 * Fix for TLS record tampering bug. A carefully crafted invalid
7090 handshake could crash OpenSSL with a NULL pointer exception.
7091 Thanks to Anton Johansson for reporting this issues.
7094 * Keep original DTLS digest and encryption contexts in retransmission
7095 structures so we can use the previous session parameters if they need
7096 to be resent. ([CVE-2013-6450])
7100 * Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
7101 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
7102 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
7103 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
7104 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
7105 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
7107 *Rob Stradling, Adam Langley*
7109 ### Changes between 1.0.1d and 1.0.1e [11 Feb 2013]
7111 * Correct fix for CVE-2013-0169. The original didn't work on AES-NI
7112 supporting platforms or when small records were transferred.
7114 *Andy Polyakov, Steve Henson*
7116 ### Changes between 1.0.1c and 1.0.1d [5 Feb 2013]
7118 * Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
7120 This addresses the flaw in CBC record processing discovered by
7121 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
7122 at: <http://www.isg.rhul.ac.uk/tls/>
7124 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
7125 Security Group at Royal Holloway, University of London
7126 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
7127 Emilia Käsper for the initial patch.
7130 *Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson*
7132 * Fix flaw in AESNI handling of TLS 1.2 and 1.1 records for CBC mode
7133 ciphersuites which can be exploited in a denial of service attack.
7134 Thanks go to and to Adam Langley <agl@chromium.org> for discovering
7135 and detecting this bug and to Wolfgang Ettlinger
7136 <wolfgang.ettlinger@gmail.com> for independently discovering this issue.
7141 * Return an error when checking OCSP signatures when key is NULL.
7142 This fixes a DoS attack. ([CVE-2013-0166])
7146 * Make openssl verify return errors.
7148 *Chris Palmer <palmer@google.com> and Ben Laurie*
7150 * Call OCSP Stapling callback after ciphersuite has been chosen, so
7151 the right response is stapled. Also change SSL_get_certificate()
7152 so it returns the certificate actually sent.
7153 See <http://rt.openssl.org/Ticket/Display.html?id=2836>.
7155 *Rob Stradling <rob.stradling@comodo.com>*
7157 * Fix possible deadlock when decoding public keys.
7161 * Don't use TLS 1.0 record version number in initial client hello
7166 ### Changes between 1.0.1b and 1.0.1c [10 May 2012]
7168 * Sanity check record length before skipping explicit IV in TLS
7169 1.2, 1.1 and DTLS to fix DoS attack.
7171 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
7172 fuzzing as a service testing platform.
7177 * Initialise tkeylen properly when encrypting CMS messages.
7178 Thanks to Solar Designer of Openwall for reporting this issue.
7182 * In FIPS mode don't try to use composite ciphers as they are not
7187 ### Changes between 1.0.1a and 1.0.1b [26 Apr 2012]
7189 * OpenSSL 1.0.0 sets SSL_OP_ALL to 0x80000FFFL and OpenSSL 1.0.1 and
7190 1.0.1a set SSL_OP_NO_TLSv1_1 to 0x00000400L which would unfortunately
7191 mean any application compiled against OpenSSL 1.0.0 headers setting
7192 SSL_OP_ALL would also set SSL_OP_NO_TLSv1_1, unintentionally disabling
7193 TLS 1.1 also. Fix this by changing the value of SSL_OP_NO_TLSv1_1 to
7194 0x10000000L Any application which was previously compiled against
7195 OpenSSL 1.0.1 or 1.0.1a headers and which cares about SSL_OP_NO_TLSv1_1
7196 will need to be recompiled as a result. Letting be results in
7197 inability to disable specifically TLS 1.1 and in client context,
7198 in unlike event, limit maximum offered version to TLS 1.0 [see below].
7202 * In order to ensure interoperability SSL_OP_NO_protocolX does not
7203 disable just protocol X, but all protocols above X *if* there are
7204 protocols *below* X still enabled. In more practical terms it means
7205 that if application wants to disable TLS1.0 in favor of TLS1.1 and
7206 above, it's not sufficient to pass `SSL_OP_NO_TLSv1`, one has to pass
7207 `SSL_OP_NO_TLSv1|SSL_OP_NO_SSLv3|SSL_OP_NO_SSLv2`. This applies to
7212 ### Changes between 1.0.1 and 1.0.1a [19 Apr 2012]
7214 * Check for potentially exploitable overflows in asn1_d2i_read_bio
7215 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
7216 in CRYPTO_realloc_clean.
7218 Thanks to Tavis Ormandy, Google Security Team, for discovering this
7219 issue and to Adam Langley <agl@chromium.org> for fixing it.
7222 *Adam Langley (Google), Tavis Ormandy, Google Security Team*
7224 * Don't allow TLS 1.2 SHA-256 ciphersuites in TLS 1.0, 1.1 connections.
7228 * Workarounds for some broken servers that "hang" if a client hello
7229 record length exceeds 255 bytes.
7231 1. Do not use record version number > TLS 1.0 in initial client
7232 hello: some (but not all) hanging servers will now work.
7233 2. If we set OPENSSL_MAX_TLS1_2_CIPHER_LENGTH this will truncate
7234 the number of ciphers sent in the client hello. This should be
7235 set to an even number, such as 50, for example by passing:
7236 -DOPENSSL_MAX_TLS1_2_CIPHER_LENGTH=50 to config or Configure.
7237 Most broken servers should now work.
7238 3. If all else fails setting OPENSSL_NO_TLS1_2_CLIENT will disable
7239 TLS 1.2 client support entirely.
7243 * Fix SEGV in Vector Permutation AES module observed in OpenSSH.
7247 ### Changes between 1.0.0h and 1.0.1 [14 Mar 2012]
7249 * Add compatibility with old MDC2 signatures which use an ASN1 OCTET
7250 STRING form instead of a DigestInfo.
7254 * The format used for MDC2 RSA signatures is inconsistent between EVP
7255 and the RSA_sign/RSA_verify functions. This was made more apparent when
7256 OpenSSL used RSA_sign/RSA_verify for some RSA signatures in particular
7257 those which went through EVP_PKEY_METHOD in 1.0.0 and later. Detect
7258 the correct format in RSA_verify so both forms transparently work.
7262 * Some servers which support TLS 1.0 can choke if we initially indicate
7263 support for TLS 1.2 and later renegotiate using TLS 1.0 in the RSA
7264 encrypted premaster secret. As a workaround use the maximum permitted
7265 client version in client hello, this should keep such servers happy
7266 and still work with previous versions of OpenSSL.
7270 * Add support for TLS/DTLS heartbeats.
7272 *Robin Seggelmann <seggelmann@fh-muenster.de>*
7274 * Add support for SCTP.
7276 *Robin Seggelmann <seggelmann@fh-muenster.de>*
7278 * Improved PRNG seeding for VOS.
7280 *Paul Green <Paul.Green@stratus.com>*
7282 * Extensive assembler packs updates, most notably:
7284 - x86[_64]: AES-NI, PCLMULQDQ, RDRAND support;
7285 - x86[_64]: SSSE3 support (SHA1, vector-permutation AES);
7286 - x86_64: bit-sliced AES implementation;
7287 - ARM: NEON support, contemporary platforms optimizations;
7288 - s390x: z196 support;
7289 - `*`: GHASH and GF(2^m) multiplication implementations;
7293 * Make TLS-SRP code conformant with RFC 5054 API cleanup
7294 (removal of unnecessary code)
7296 *Peter Sylvester <peter.sylvester@edelweb.fr>*
7298 * Add TLS key material exporter from RFC 5705.
7302 * Add DTLS-SRTP negotiation from RFC 5764.
7306 * Add Next Protocol Negotiation,
7307 <http://tools.ietf.org/html/draft-agl-tls-nextprotoneg-00>. Can be
7308 disabled with a no-npn flag to config or Configure. Code donated
7311 *Adam Langley <agl@google.com> and Ben Laurie*
7313 * Add optional 64-bit optimized implementations of elliptic curves NIST-P224,
7314 NIST-P256, NIST-P521, with constant-time single point multiplication on
7315 typical inputs. Compiler support for the nonstandard type `__uint128_t` is
7316 required to use this (present in gcc 4.4 and later, for 64-bit builds).
7317 Code made available under Apache License version 2.0.
7319 Specify "enable-ec_nistp_64_gcc_128" on the Configure (or config) command
7320 line to include this in your build of OpenSSL, and run "make depend" (or
7321 "make update"). This enables the following EC_METHODs:
7323 EC_GFp_nistp224_method()
7324 EC_GFp_nistp256_method()
7325 EC_GFp_nistp521_method()
7327 EC_GROUP_new_by_curve_name() will automatically use these (while
7328 EC_GROUP_new_curve_GFp() currently prefers the more flexible
7331 *Emilia Käsper, Adam Langley, Bodo Moeller (Google)*
7333 * Use type ossl_ssize_t instead of ssize_t which isn't available on
7334 all platforms. Move ssize_t definition from e_os.h to the public
7335 header file e_os2.h as it now appears in public header file cms.h
7339 * New -sigopt option to the ca, req and x509 utilities. Additional
7340 signature parameters can be passed using this option and in
7345 * Add RSA PSS signing function. This will generate and set the
7346 appropriate AlgorithmIdentifiers for PSS based on those in the
7347 corresponding EVP_MD_CTX structure. No application support yet.
7351 * Support for companion algorithm specific ASN1 signing routines.
7352 New function ASN1_item_sign_ctx() signs a pre-initialised
7353 EVP_MD_CTX structure and sets AlgorithmIdentifiers based on
7354 the appropriate parameters.
7358 * Add new algorithm specific ASN1 verification initialisation function
7359 to EVP_PKEY_ASN1_METHOD: this is not in EVP_PKEY_METHOD since the ASN1
7360 handling will be the same no matter what EVP_PKEY_METHOD is used.
7361 Add a PSS handler to support verification of PSS signatures: checked
7362 against a number of sample certificates.
7366 * Add signature printing for PSS. Add PSS OIDs.
7368 *Steve Henson, Martin Kaiser <lists@kaiser.cx>*
7370 * Add algorithm specific signature printing. An individual ASN1 method
7371 can now print out signatures instead of the standard hex dump.
7373 More complex signatures (e.g. PSS) can print out more meaningful
7374 information. Include DSA version that prints out the signature
7379 * Password based recipient info support for CMS library: implementing
7384 * Split password based encryption into PBES2 and PBKDF2 functions. This
7385 neatly separates the code into cipher and PBE sections and is required
7386 for some algorithms that split PBES2 into separate pieces (such as
7387 password based CMS).
7391 * Session-handling fixes:
7392 - Fix handling of connections that are resuming with a session ID,
7393 but also support Session Tickets.
7394 - Fix a bug that suppressed issuing of a new ticket if the client
7395 presented a ticket with an expired session.
7396 - Try to set the ticket lifetime hint to something reasonable.
7397 - Make tickets shorter by excluding irrelevant information.
7398 - On the client side, don't ignore renewed tickets.
7400 *Adam Langley, Bodo Moeller (Google)*
7402 * Fix PSK session representation.
7406 * Add RC4-MD5 and AESNI-SHA1 "stitched" implementations.
7408 This work was sponsored by Intel.
7412 * Add GCM support to TLS library. Some custom code is needed to split
7413 the IV between the fixed (from PRF) and explicit (from TLS record)
7414 portions. This adds all GCM ciphersuites supported by RFC5288 and
7415 RFC5289. Generalise some `AES*` cipherstrings to include GCM and
7416 add a special AESGCM string for GCM only.
7420 * Expand range of ctrls for AES GCM. Permit setting invocation
7421 field on decrypt and retrieval of invocation field only on encrypt.
7425 * Add HMAC ECC ciphersuites from RFC5289. Include SHA384 PRF support.
7426 As required by RFC5289 these ciphersuites cannot be used if for
7427 versions of TLS earlier than 1.2.
7431 * For FIPS capable OpenSSL interpret a NULL default public key method
7432 as unset and return the appropriate default but do *not* set the default.
7433 This means we can return the appropriate method in applications that
7434 switch between FIPS and non-FIPS modes.
7438 * Redirect HMAC and CMAC operations to FIPS module in FIPS mode. If an
7439 ENGINE is used then we cannot handle that in the FIPS module so we
7440 keep original code iff non-FIPS operations are allowed.
7444 * Add -attime option to openssl utilities.
7446 *Peter Eckersley <pde@eff.org>, Ben Laurie and Steve Henson*
7448 * Redirect DSA and DH operations to FIPS module in FIPS mode.
7452 * Redirect ECDSA and ECDH operations to FIPS module in FIPS mode. Also use
7453 FIPS EC methods unconditionally for now.
7457 * New build option no-ec2m to disable characteristic 2 code.
7461 * Backport libcrypto audit of return value checking from 1.1.0-dev; not
7462 all cases can be covered as some introduce binary incompatibilities.
7466 * Redirect RSA operations to FIPS module including keygen,
7467 encrypt, decrypt, sign and verify. Block use of non FIPS RSA methods.
7471 * Add similar low-level API blocking to ciphers.
7475 * low-level digest APIs are not approved in FIPS mode: any attempt
7476 to use these will cause a fatal error. Applications that *really* want
7477 to use them can use the `private_*` version instead.
7481 * Redirect cipher operations to FIPS module for FIPS builds.
7485 * Redirect digest operations to FIPS module for FIPS builds.
7489 * Update build system to add "fips" flag which will link in fipscanister.o
7490 for static and shared library builds embedding a signature if needed.
7494 * Output TLS supported curves in preference order instead of numerical
7495 order. This is currently hardcoded for the highest order curves first.
7496 This should be configurable so applications can judge speed vs strength.
7500 * Add TLS v1.2 server support for client authentication.
7504 * Add support for FIPS mode in ssl library: disable SSLv3, non-FIPS ciphers
7509 * Functions FIPS_mode_set() and FIPS_mode() which call the underlying
7510 FIPS modules versions.
7514 * Add TLS v1.2 client side support for client authentication. Keep cache
7515 of handshake records longer as we don't know the hash algorithm to use
7516 until after the certificate request message is received.
7520 * Initial TLS v1.2 client support. Add a default signature algorithms
7521 extension including all the algorithms we support. Parse new signature
7522 format in client key exchange. Relax some ECC signing restrictions for
7523 TLS v1.2 as indicated in RFC5246.
7527 * Add server support for TLS v1.2 signature algorithms extension. Switch
7528 to new signature format when needed using client digest preference.
7529 All server ciphersuites should now work correctly in TLS v1.2. No client
7530 support yet and no support for client certificates.
7534 * Initial TLS v1.2 support. Add new SHA256 digest to ssl code, switch
7535 to SHA256 for PRF when using TLS v1.2 and later. Add new SHA256 based
7536 ciphersuites. At present only RSA key exchange ciphersuites work with
7537 TLS v1.2. Add new option for TLS v1.2 replacing the old and obsolete
7538 SSL_OP_PKCS1_CHECK flags with SSL_OP_NO_TLSv1_2. New TLSv1.2 methods
7539 and version checking.
7543 * New option OPENSSL_NO_SSL_INTERN. If an application can be compiled
7544 with this defined it will not be affected by any changes to ssl internal
7545 structures. Add several utility functions to allow openssl application
7546 to work with OPENSSL_NO_SSL_INTERN defined.
7550 * A long standing patch to add support for SRP from EdelWeb (Peter
7551 Sylvester and Christophe Renou) was integrated.
7552 *Christophe Renou <christophe.renou@edelweb.fr>, Peter Sylvester
7553 <peter.sylvester@edelweb.fr>, Tom Wu <tjw@cs.stanford.edu>, and
7556 * Add functions to copy EVP_PKEY_METHOD and retrieve flags and id.
7560 * Permit abbreviated handshakes when renegotiating using the function
7561 SSL_renegotiate_abbreviated().
7563 *Robin Seggelmann <seggelmann@fh-muenster.de>*
7565 * Add call to ENGINE_register_all_complete() to
7566 ENGINE_load_builtin_engines(), so some implementations get used
7567 automatically instead of needing explicit application support.
7571 * Add support for TLS key exporter as described in RFC5705.
7573 *Robin Seggelmann <seggelmann@fh-muenster.de>, Steve Henson*
7575 * Initial TLSv1.1 support. Since TLSv1.1 is very similar to TLS v1.0 only
7576 a few changes are required:
7578 Add SSL_OP_NO_TLSv1_1 flag.
7579 Add TLSv1_1 methods.
7580 Update version checking logic to handle version 1.1.
7581 Add explicit IV handling (ported from DTLS code).
7582 Add command line options to s_client/s_server.
7589 ### Changes between 1.0.0s and 1.0.0t [3 Dec 2015]
7591 * X509_ATTRIBUTE memory leak
7593 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
7594 memory. This structure is used by the PKCS#7 and CMS routines so any
7595 application which reads PKCS#7 or CMS data from untrusted sources is
7596 affected. SSL/TLS is not affected.
7598 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
7604 * Race condition handling PSK identify hint
7606 If PSK identity hints are received by a multi-threaded client then
7607 the values are wrongly updated in the parent SSL_CTX structure. This can
7608 result in a race condition potentially leading to a double free of the
7614 ### Changes between 1.0.0r and 1.0.0s [11 Jun 2015]
7616 * Malformed ECParameters causes infinite loop
7618 When processing an ECParameters structure OpenSSL enters an infinite loop
7619 if the curve specified is over a specially malformed binary polynomial
7622 This can be used to perform denial of service against any
7623 system which processes public keys, certificate requests or
7624 certificates. This includes TLS clients and TLS servers with
7625 client authentication enabled.
7627 This issue was reported to OpenSSL by Joseph Barr-Pixton.
7632 * Exploitable out-of-bounds read in X509_cmp_time
7634 X509_cmp_time does not properly check the length of the ASN1_TIME
7635 string and can read a few bytes out of bounds. In addition,
7636 X509_cmp_time accepts an arbitrary number of fractional seconds in the
7639 An attacker can use this to craft malformed certificates and CRLs of
7640 various sizes and potentially cause a segmentation fault, resulting in
7641 a DoS on applications that verify certificates or CRLs. TLS clients
7642 that verify CRLs are affected. TLS clients and servers with client
7643 authentication enabled may be affected if they use custom verification
7646 This issue was reported to OpenSSL by Robert Swiecki (Google), and
7647 independently by Hanno Böck.
7652 * PKCS7 crash with missing EnvelopedContent
7654 The PKCS#7 parsing code does not handle missing inner EncryptedContent
7655 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
7656 with missing content and trigger a NULL pointer dereference on parsing.
7658 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
7659 structures from untrusted sources are affected. OpenSSL clients and
7660 servers are not affected.
7662 This issue was reported to OpenSSL by Michal Zalewski (Google).
7667 * CMS verify infinite loop with unknown hash function
7669 When verifying a signedData message the CMS code can enter an infinite loop
7670 if presented with an unknown hash function OID. This can be used to perform
7671 denial of service against any system which verifies signedData messages using
7673 This issue was reported to OpenSSL by Johannes Bauer.
7678 * Race condition handling NewSessionTicket
7680 If a NewSessionTicket is received by a multi-threaded client when attempting to
7681 reuse a previous ticket then a race condition can occur potentially leading to
7682 a double free of the ticket data.
7687 ### Changes between 1.0.0q and 1.0.0r [19 Mar 2015]
7689 * Segmentation fault in ASN1_TYPE_cmp fix
7691 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
7692 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
7693 certificate signature algorithm consistency this can be used to crash any
7694 certificate verification operation and exploited in a DoS attack. Any
7695 application which performs certificate verification is vulnerable including
7696 OpenSSL clients and servers which enable client authentication.
7701 * ASN.1 structure reuse memory corruption fix
7703 Reusing a structure in ASN.1 parsing may allow an attacker to cause
7704 memory corruption via an invalid write. Such reuse is and has been
7705 strongly discouraged and is believed to be rare.
7707 Applications that parse structures containing CHOICE or ANY DEFINED BY
7708 components may be affected. Certificate parsing (d2i_X509 and related
7709 functions) are however not affected. OpenSSL clients and servers are
7715 * PKCS7 NULL pointer dereferences fix
7717 The PKCS#7 parsing code does not handle missing outer ContentInfo
7718 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
7719 missing content and trigger a NULL pointer dereference on parsing.
7721 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
7722 otherwise parse PKCS#7 structures from untrusted sources are
7723 affected. OpenSSL clients and servers are not affected.
7725 This issue was reported to OpenSSL by Michal Zalewski (Google).
7730 * DoS via reachable assert in SSLv2 servers fix
7732 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
7733 servers that both support SSLv2 and enable export cipher suites by sending
7734 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
7736 This issue was discovered by Sean Burford (Google) and Emilia Käsper
7737 (OpenSSL development team).
7742 * Use After Free following d2i_ECPrivatekey error fix
7744 A malformed EC private key file consumed via the d2i_ECPrivateKey function
7745 could cause a use after free condition. This, in turn, could cause a double
7746 free in several private key parsing functions (such as d2i_PrivateKey
7747 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
7748 for applications that receive EC private keys from untrusted
7749 sources. This scenario is considered rare.
7751 This issue was discovered by the BoringSSL project and fixed in their
7757 * X509_to_X509_REQ NULL pointer deref fix
7759 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
7760 the certificate key is invalid. This function is rarely used in practice.
7762 This issue was discovered by Brian Carpenter.
7767 * Removed the export ciphers from the DEFAULT ciphers
7771 ### Changes between 1.0.0p and 1.0.0q [15 Jan 2015]
7773 * Build fixes for the Windows and OpenVMS platforms
7775 *Matt Caswell and Richard Levitte*
7777 ### Changes between 1.0.0o and 1.0.0p [8 Jan 2015]
7779 * Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
7780 message can cause a segmentation fault in OpenSSL due to a NULL pointer
7781 dereference. This could lead to a Denial Of Service attack. Thanks to
7782 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
7787 * Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
7788 dtls1_buffer_record function under certain conditions. In particular this
7789 could occur if an attacker sent repeated DTLS records with the same
7790 sequence number but for the next epoch. The memory leak could be exploited
7791 by an attacker in a Denial of Service attack through memory exhaustion.
7792 Thanks to Chris Mueller for reporting this issue.
7797 * Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
7798 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
7799 method would be set to NULL which could later result in a NULL pointer
7800 dereference. Thanks to Frank Schmirler for reporting this issue.
7805 * Abort handshake if server key exchange message is omitted for ephemeral
7808 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
7809 reporting this issue.
7814 * Remove non-export ephemeral RSA code on client and server. This code
7815 violated the TLS standard by allowing the use of temporary RSA keys in
7816 non-export ciphersuites and could be used by a server to effectively
7817 downgrade the RSA key length used to a value smaller than the server
7818 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
7819 INRIA or reporting this issue.
7824 * Fixed issue where DH client certificates are accepted without verification.
7825 An OpenSSL server will accept a DH certificate for client authentication
7826 without the certificate verify message. This effectively allows a client to
7827 authenticate without the use of a private key. This only affects servers
7828 which trust a client certificate authority which issues certificates
7829 containing DH keys: these are extremely rare and hardly ever encountered.
7830 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
7836 * Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
7837 results on some platforms, including x86_64. This bug occurs at random
7838 with a very low probability, and is not known to be exploitable in any
7839 way, though its exact impact is difficult to determine. Thanks to Pieter
7840 Wuille (Blockstream) who reported this issue and also suggested an initial
7841 fix. Further analysis was conducted by the OpenSSL development team and
7842 Adam Langley of Google. The final fix was developed by Andy Polyakov of
7843 the OpenSSL core team.
7848 * Fix various certificate fingerprint issues.
7850 By using non-DER or invalid encodings outside the signed portion of a
7851 certificate the fingerprint can be changed without breaking the signature.
7852 Although no details of the signed portion of the certificate can be changed
7853 this can cause problems with some applications: e.g. those using the
7854 certificate fingerprint for blacklists.
7856 1. Reject signatures with non zero unused bits.
7858 If the BIT STRING containing the signature has non zero unused bits reject
7859 the signature. All current signature algorithms require zero unused bits.
7861 2. Check certificate algorithm consistency.
7863 Check the AlgorithmIdentifier inside TBS matches the one in the
7864 certificate signature. NB: this will result in signature failure
7865 errors for some broken certificates.
7867 Thanks to Konrad Kraszewski from Google for reporting this issue.
7869 3. Check DSA/ECDSA signatures use DER.
7871 Reencode DSA/ECDSA signatures and compare with the original received
7872 signature. Return an error if there is a mismatch.
7874 This will reject various cases including garbage after signature
7875 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
7876 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
7877 (negative or with leading zeroes).
7879 Further analysis was conducted and fixes were developed by Stephen Henson
7880 of the OpenSSL core team.
7886 ### Changes between 1.0.0n and 1.0.0o [15 Oct 2014]
7888 * Session Ticket Memory Leak.
7890 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
7891 integrity of that ticket is first verified. In the event of a session
7892 ticket integrity check failing, OpenSSL will fail to free memory
7893 causing a memory leak. By sending a large number of invalid session
7894 tickets an attacker could exploit this issue in a Denial Of Service
7900 * Build option no-ssl3 is incomplete.
7902 When OpenSSL is configured with "no-ssl3" as a build option, servers
7903 could accept and complete a SSL 3.0 handshake, and clients could be
7904 configured to send them.
7907 *Akamai and the OpenSSL team*
7909 * Add support for TLS_FALLBACK_SCSV.
7910 Client applications doing fallback retries should call
7911 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
7914 *Adam Langley, Bodo Moeller*
7916 * Add additional DigestInfo checks.
7918 Reencode DigestInto in DER and check against the original when
7919 verifying RSA signature: this will reject any improperly encoded
7920 DigestInfo structures.
7922 Note: this is a precautionary measure and no attacks are currently known.
7926 ### Changes between 1.0.0m and 1.0.0n [6 Aug 2014]
7928 * OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
7929 to a denial of service attack. A malicious server can crash the client
7930 with a null pointer dereference (read) by specifying an anonymous (EC)DH
7931 ciphersuite and sending carefully crafted handshake messages.
7933 Thanks to Felix Gröbert (Google) for discovering and researching this
7939 * By sending carefully crafted DTLS packets an attacker could cause openssl
7940 to leak memory. This can be exploited through a Denial of Service attack.
7941 Thanks to Adam Langley for discovering and researching this issue.
7946 * An attacker can force openssl to consume large amounts of memory whilst
7947 processing DTLS handshake messages. This can be exploited through a
7948 Denial of Service attack.
7949 Thanks to Adam Langley for discovering and researching this issue.
7954 * An attacker can force an error condition which causes openssl to crash
7955 whilst processing DTLS packets due to memory being freed twice. This
7956 can be exploited through a Denial of Service attack.
7957 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
7963 * If a multithreaded client connects to a malicious server using a resumed
7964 session and the server sends an ec point format extension it could write
7965 up to 255 bytes to freed memory.
7967 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
7973 * A flaw in OBJ_obj2txt may cause pretty printing functions such as
7974 X509_name_oneline, X509_name_print_ex et al. to leak some information
7975 from the stack. Applications may be affected if they echo pretty printing
7976 output to the attacker.
7978 Thanks to Ivan Fratric (Google) for discovering this issue.
7981 *Emilia Käsper, and Steve Henson*
7983 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
7984 for corner cases. (Certain input points at infinity could lead to
7985 bogus results, with non-infinity inputs mapped to infinity too.)
7989 ### Changes between 1.0.0l and 1.0.0m [5 Jun 2014]
7991 * Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
7992 handshake can force the use of weak keying material in OpenSSL
7993 SSL/TLS clients and servers.
7995 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
7996 researching this issue. ([CVE-2014-0224])
7998 *KIKUCHI Masashi, Steve Henson*
8000 * Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
8001 OpenSSL DTLS client the code can be made to recurse eventually crashing
8004 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
8007 *Imre Rad, Steve Henson*
8009 * Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
8010 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
8011 client or server. This is potentially exploitable to run arbitrary
8012 code on a vulnerable client or server.
8014 Thanks to Jüri Aedla for reporting this issue. ([CVE-2014-0195])
8016 *Jüri Aedla, Steve Henson*
8018 * Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
8019 are subject to a denial of service attack.
8021 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
8022 this issue. ([CVE-2014-3470])
8024 *Felix Gröbert, Ivan Fratric, Steve Henson*
8026 * Harmonize version and its documentation. -f flag is used to display
8029 *mancha <mancha1@zoho.com>*
8031 * Fix eckey_priv_encode so it immediately returns an error upon a failure
8032 in i2d_ECPrivateKey.
8034 *mancha <mancha1@zoho.com>*
8036 * Fix some double frees. These are not thought to be exploitable.
8038 *mancha <mancha1@zoho.com>*
8040 * Fix for the attack described in the paper "Recovering OpenSSL
8041 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
8042 by Yuval Yarom and Naomi Benger. Details can be obtained from:
8043 <http://eprint.iacr.org/2014/140>
8045 Thanks to Yuval Yarom and Naomi Benger for discovering this
8046 flaw and to Yuval Yarom for supplying a fix ([CVE-2014-0076])
8048 *Yuval Yarom and Naomi Benger*
8050 ### Changes between 1.0.0k and 1.0.0l [6 Jan 2014]
8052 * Keep original DTLS digest and encryption contexts in retransmission
8053 structures so we can use the previous session parameters if they need
8054 to be resent. ([CVE-2013-6450])
8058 * Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
8059 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
8060 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
8061 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
8062 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
8063 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
8065 *Rob Stradling, Adam Langley*
8067 ### Changes between 1.0.0j and 1.0.0k [5 Feb 2013]
8069 * Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
8071 This addresses the flaw in CBC record processing discovered by
8072 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
8073 at: <http://www.isg.rhul.ac.uk/tls/>
8075 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
8076 Security Group at Royal Holloway, University of London
8077 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
8078 Emilia Käsper for the initial patch.
8081 *Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson*
8083 * Return an error when checking OCSP signatures when key is NULL.
8084 This fixes a DoS attack. ([CVE-2013-0166])
8088 * Call OCSP Stapling callback after ciphersuite has been chosen, so
8089 the right response is stapled. Also change SSL_get_certificate()
8090 so it returns the certificate actually sent.
8091 See <http://rt.openssl.org/Ticket/Display.html?id=2836>.
8092 (This is a backport)
8094 *Rob Stradling <rob.stradling@comodo.com>*
8096 * Fix possible deadlock when decoding public keys.
8100 ### Changes between 1.0.0i and 1.0.0j [10 May 2012]
8102 [NB: OpenSSL 1.0.0i and later 1.0.0 patch levels were released after
8105 * Sanity check record length before skipping explicit IV in DTLS
8108 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
8109 fuzzing as a service testing platform.
8114 * Initialise tkeylen properly when encrypting CMS messages.
8115 Thanks to Solar Designer of Openwall for reporting this issue.
8119 ### Changes between 1.0.0h and 1.0.0i [19 Apr 2012]
8121 * Check for potentially exploitable overflows in asn1_d2i_read_bio
8122 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
8123 in CRYPTO_realloc_clean.
8125 Thanks to Tavis Ormandy, Google Security Team, for discovering this
8126 issue and to Adam Langley <agl@chromium.org> for fixing it.
8129 *Adam Langley (Google), Tavis Ormandy, Google Security Team*
8131 ### Changes between 1.0.0g and 1.0.0h [12 Mar 2012]
8133 * Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness
8134 in CMS and PKCS7 code. When RSA decryption fails use a random key for
8135 content decryption and always return the same error. Note: this attack
8136 needs on average 2^20 messages so it only affects automated senders. The
8137 old behaviour can be re-enabled in the CMS code by setting the
8138 CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where
8139 an MMA defence is not necessary.
8140 Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering
8141 this issue. ([CVE-2012-0884])
8145 * Fix CVE-2011-4619: make sure we really are receiving a
8146 client hello before rejecting multiple SGC restarts. Thanks to
8147 Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug.
8151 ### Changes between 1.0.0f and 1.0.0g [18 Jan 2012]
8153 * Fix for DTLS DoS issue introduced by fix for CVE-2011-4109.
8154 Thanks to Antonio Martin, Enterprise Secure Access Research and
8155 Development, Cisco Systems, Inc. for discovering this bug and
8156 preparing a fix. ([CVE-2012-0050])
8160 ### Changes between 1.0.0e and 1.0.0f [4 Jan 2012]
8162 * Nadhem Alfardan and Kenny Paterson have discovered an extension
8163 of the Vaudenay padding oracle attack on CBC mode encryption
8164 which enables an efficient plaintext recovery attack against
8165 the OpenSSL implementation of DTLS. Their attack exploits timing
8166 differences arising during decryption processing. A research
8167 paper describing this attack can be found at:
8168 <http://www.isg.rhul.ac.uk/~kp/dtls.pdf>
8169 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
8170 Security Group at Royal Holloway, University of London
8171 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
8172 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
8173 for preparing the fix. ([CVE-2011-4108])
8175 *Robin Seggelmann, Michael Tuexen*
8177 * Clear bytes used for block padding of SSL 3.0 records.
8180 *Adam Langley (Google)*
8182 * Only allow one SGC handshake restart for SSL/TLS. Thanks to George
8183 Kadianakis <desnacked@gmail.com> for discovering this issue and
8184 Adam Langley for preparing the fix. ([CVE-2011-4619])
8186 *Adam Langley (Google)*
8188 * Check parameters are not NULL in GOST ENGINE. ([CVE-2012-0027])
8190 *Andrey Kulikov <amdeich@gmail.com>*
8192 * Prevent malformed RFC3779 data triggering an assertion failure.
8193 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
8194 and Rob Austein <sra@hactrn.net> for fixing it. ([CVE-2011-4577])
8196 *Rob Austein <sra@hactrn.net>*
8198 * Improved PRNG seeding for VOS.
8200 *Paul Green <Paul.Green@stratus.com>*
8202 * Fix ssl_ciph.c set-up race.
8204 *Adam Langley (Google)*
8206 * Fix spurious failures in ecdsatest.c.
8208 *Emilia Käsper (Google)*
8210 * Fix the BIO_f_buffer() implementation (which was mixing different
8211 interpretations of the `..._len` fields).
8213 *Adam Langley (Google)*
8215 * Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
8216 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
8217 threads won't reuse the same blinding coefficients.
8219 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
8220 lock to call BN_BLINDING_invert_ex, and avoids one use of
8221 BN_BLINDING_update for each BN_BLINDING structure (previously,
8222 the last update always remained unused).
8224 *Emilia Käsper (Google)*
8226 * In ssl3_clear, preserve s3->init_extra along with s3->rbuf.
8228 *Bob Buckholz (Google)*
8230 ### Changes between 1.0.0d and 1.0.0e [6 Sep 2011]
8232 * Fix bug where CRLs with nextUpdate in the past are sometimes accepted
8233 by initialising X509_STORE_CTX properly. ([CVE-2011-3207])
8235 *Kaspar Brand <ossl@velox.ch>*
8237 * Fix SSL memory handling for (EC)DH ciphersuites, in particular
8238 for multi-threaded use of ECDH. ([CVE-2011-3210])
8240 *Adam Langley (Google)*
8242 * Fix x509_name_ex_d2i memory leak on bad inputs.
8246 * Remove hard coded ecdsaWithSHA1 signature tests in ssl code and check
8247 signature public key algorithm by using OID xref utilities instead.
8248 Before this you could only use some ECC ciphersuites with SHA1 only.
8252 * Add protection against ECDSA timing attacks as mentioned in the paper
8253 by Billy Bob Brumley and Nicola Tuveri, see:
8254 <http://eprint.iacr.org/2011/232.pdf>
8256 *Billy Bob Brumley and Nicola Tuveri*
8258 ### Changes between 1.0.0c and 1.0.0d [8 Feb 2011]
8260 * Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
8262 *Neel Mehta, Adam Langley, Bodo Moeller (Google)*
8264 * Fix bug in string printing code: if *any* escaping is enabled we must
8265 escape the escape character (backslash) or the resulting string is
8270 ### Changes between 1.0.0b and 1.0.0c [2 Dec 2010]
8272 * Disable code workaround for ancient and obsolete Netscape browsers
8273 and servers: an attacker can use it in a ciphersuite downgrade attack.
8274 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
8278 * Fixed J-PAKE implementation error, originally discovered by
8279 Sebastien Martini, further info and confirmation from Stefan
8280 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
8284 ### Changes between 1.0.0a and 1.0.0b [16 Nov 2010]
8286 * Fix extension code to avoid race conditions which can result in a buffer
8287 overrun vulnerability: resumed sessions must not be modified as they can
8288 be shared by multiple threads. CVE-2010-3864
8292 * Fix WIN32 build system to correctly link an ENGINE directory into
8297 ### Changes between 1.0.0 and 1.0.0a [01 Jun 2010]
8299 * Check return value of int_rsa_verify in pkey_rsa_verifyrecover
8302 *Steve Henson, Peter-Michael Hager <hager@dortmund.net>*
8304 ### Changes between 0.9.8n and 1.0.0 [29 Mar 2010]
8306 * Add "missing" function EVP_CIPHER_CTX_copy(). This copies a cipher
8307 context. The operation can be customised via the ctrl mechanism in
8308 case ENGINEs want to include additional functionality.
8312 * Tolerate yet another broken PKCS#8 key format: private key value negative.
8316 * Add new -subject_hash_old and -issuer_hash_old options to x509 utility to
8317 output hashes compatible with older versions of OpenSSL.
8319 *Willy Weisz <weisz@vcpc.univie.ac.at>*
8321 * Fix compression algorithm handling: if resuming a session use the
8322 compression algorithm of the resumed session instead of determining
8323 it from client hello again. Don't allow server to change algorithm.
8327 * Add load_crls() function to commands tidying load_certs() too. Add option
8328 to verify utility to allow additional CRLs to be included.
8332 * Update OCSP request code to permit adding custom headers to the request:
8333 some responders need this.
8337 * The function EVP_PKEY_sign() returns <=0 on error: check return code
8340 *Julia Lawall <julia@diku.dk>*
8342 * Update verify callback code in `apps/s_cb.c` and `apps/verify.c`, it
8343 needlessly dereferenced structures, used obsolete functions and
8344 didn't handle all updated verify codes correctly.
8348 * Disable MD2 in the default configuration.
8352 * In BIO_pop() and BIO_push() use the ctrl argument (which was NULL) to
8353 indicate the initial BIO being pushed or popped. This makes it possible
8354 to determine whether the BIO is the one explicitly called or as a result
8355 of the ctrl being passed down the chain. Fix BIO_pop() and SSL BIOs so
8356 it handles reference counts correctly and doesn't zero out the I/O bio
8357 when it is not being explicitly popped. WARNING: applications which
8358 included workarounds for the old buggy behaviour will need to be modified
8359 or they could free up already freed BIOs.
8363 * Extend the uni2asc/asc2uni => OPENSSL_uni2asc/OPENSSL_asc2uni
8364 renaming to all platforms (within the 0.9.8 branch, this was
8365 done conditionally on Netware platforms to avoid a name clash).
8367 *Guenter <lists@gknw.net>*
8369 * Add ECDHE and PSK support to DTLS.
8371 *Michael Tuexen <tuexen@fh-muenster.de>*
8373 * Add CHECKED_STACK_OF macro to safestack.h, otherwise safestack can't
8378 * Add "missing" function EVP_MD_flags() (without this the only way to
8379 retrieve a digest flags is by accessing the structure directly. Update
8380 `EVP_MD_do_all*()` and `EVP_CIPHER_do_all*()` to include the name a digest
8381 or cipher is registered as in the "from" argument. Print out all
8382 registered digests in the dgst usage message instead of manually
8383 attempting to work them out.
8387 * If no SSLv2 ciphers are used don't use an SSLv2 compatible client hello:
8388 this allows the use of compression and extensions. Change default cipher
8389 string to remove SSLv2 ciphersuites. This effectively avoids ancient SSLv2
8390 by default unless an application cipher string requests it.
8394 * Alter match criteria in PKCS12_parse(). It used to try to use local
8395 key ids to find matching certificates and keys but some PKCS#12 files
8396 don't follow the (somewhat unwritten) rules and this strategy fails.
8397 Now just gather all certificates together and the first private key
8398 then look for the first certificate that matches the key.
8402 * Support use of registered digest and cipher names for dgst and cipher
8403 commands instead of having to add each one as a special case. So now
8410 openssl dgst -sha256 foo
8412 and this works for ENGINE based algorithms too.
8416 * Update Gost ENGINE to support parameter files.
8418 *Victor B. Wagner <vitus@cryptocom.ru>*
8420 * Support GeneralizedTime in ca utility.
8422 *Oliver Martin <oliver@volatilevoid.net>, Steve Henson*
8424 * Enhance the hash format used for certificate directory links. The new
8425 form uses the canonical encoding (meaning equivalent names will work
8426 even if they aren't identical) and uses SHA1 instead of MD5. This form
8427 is incompatible with the older format and as a result c_rehash should
8428 be used to rebuild symbolic links.
8432 * Make PKCS#8 the default write format for private keys, replacing the
8433 traditional format. This form is standardised, more secure and doesn't
8434 include an implicit MD5 dependency.
8438 * Add a $gcc_devteam_warn option to Configure. The idea is that any code
8439 committed to OpenSSL should pass this lot as a minimum.
8443 * Add session ticket override functionality for use by EAP-FAST.
8445 *Jouni Malinen <j@w1.fi>*
8447 * Modify HMAC functions to return a value. Since these can be implemented
8448 in an ENGINE errors can occur.
8452 * Type-checked OBJ_bsearch_ex.
8456 * Type-checked OBJ_bsearch. Also some constification necessitated
8457 by type-checking. Still to come: TXT_DB, bsearch(?),
8458 OBJ_bsearch_ex, qsort, CRYPTO_EX_DATA, ASN1_VALUE, ASN1_STRING,
8463 * New function OPENSSL_gmtime_adj() to add a specific number of days and
8464 seconds to a tm structure directly, instead of going through OS
8465 specific date routines. This avoids any issues with OS routines such
8466 as the year 2038 bug. New `*_adj()` functions for ASN1 time structures
8467 and X509_time_adj_ex() to cover the extended range. The existing
8468 X509_time_adj() is still usable and will no longer have any date issues.
8472 * Delta CRL support. New use deltas option which will attempt to locate
8473 and search any appropriate delta CRLs available.
8475 This work was sponsored by Google.
8479 * Support for CRLs partitioned by reason code. Reorganise CRL processing
8480 code and add additional score elements. Validate alternate CRL paths
8481 as part of the CRL checking and indicate a new error "CRL path validation
8482 error" in this case. Applications wanting additional details can use
8483 the verify callback and check the new "parent" field. If this is not
8484 NULL CRL path validation is taking place. Existing applications won't
8485 see this because it requires extended CRL support which is off by
8488 This work was sponsored by Google.
8492 * Support for freshest CRL extension.
8494 This work was sponsored by Google.
8498 * Initial indirect CRL support. Currently only supported in the CRLs
8499 passed directly and not via lookup. Process certificate issuer
8500 CRL entry extension and lookup CRL entries by bother issuer name
8501 and serial number. Check and process CRL issuer entry in IDP extension.
8503 This work was sponsored by Google.
8507 * Add support for distinct certificate and CRL paths. The CRL issuer
8508 certificate is validated separately in this case. Only enabled if
8509 an extended CRL support flag is set: this flag will enable additional
8510 CRL functionality in future.
8512 This work was sponsored by Google.
8516 * Add support for policy mappings extension.
8518 This work was sponsored by Google.
8522 * Fixes to pathlength constraint, self issued certificate handling,
8523 policy processing to align with RFC3280 and PKITS tests.
8525 This work was sponsored by Google.
8529 * Support for name constraints certificate extension. DN, email, DNS
8530 and URI types are currently supported.
8532 This work was sponsored by Google.
8536 * To cater for systems that provide a pointer-based thread ID rather
8537 than numeric, deprecate the current numeric thread ID mechanism and
8538 replace it with a structure and associated callback type. This
8539 mechanism allows a numeric "hash" to be extracted from a thread ID in
8540 either case, and on platforms where pointers are larger than 'long',
8541 mixing is done to help ensure the numeric 'hash' is usable even if it
8542 can't be guaranteed unique. The default mechanism is to use "&errno"
8543 as a pointer-based thread ID to distinguish between threads.
8545 Applications that want to provide their own thread IDs should now use
8546 CRYPTO_THREADID_set_callback() to register a callback that will call
8547 either CRYPTO_THREADID_set_numeric() or CRYPTO_THREADID_set_pointer().
8549 Note that ERR_remove_state() is now deprecated, because it is tied
8550 to the assumption that thread IDs are numeric. ERR_remove_state(0)
8551 to free the current thread's error state should be replaced by
8552 ERR_remove_thread_state(NULL).
8554 (This new approach replaces the functions CRYPTO_set_idptr_callback(),
8555 CRYPTO_get_idptr_callback(), and CRYPTO_thread_idptr() that existed in
8556 OpenSSL 0.9.9-dev between June 2006 and August 2008. Also, if an
8557 application was previously providing a numeric thread callback that
8558 was inappropriate for distinguishing threads, then uniqueness might
8559 have been obtained with &errno that happened immediately in the
8560 intermediate development versions of OpenSSL; this is no longer the
8561 case, the numeric thread callback will now override the automatic use
8564 *Geoff Thorpe, with help from Bodo Moeller*
8566 * Initial support for different CRL issuing certificates. This covers a
8567 simple case where the self issued certificates in the chain exist and
8568 the real CRL issuer is higher in the existing chain.
8570 This work was sponsored by Google.
8574 * Removed effectively defunct crypto/store from the build.
8578 * Revamp of STACK to provide stronger type-checking. Still to come:
8579 TXT_DB, bsearch(?), OBJ_bsearch, qsort, CRYPTO_EX_DATA, ASN1_VALUE,
8580 ASN1_STRING, CONF_VALUE.
8584 * Add a new SSL_MODE_RELEASE_BUFFERS mode flag to release unused buffer
8585 RAM on SSL connections. This option can save about 34k per idle SSL.
8589 * Revamp of LHASH to provide stronger type-checking. Still to come:
8590 STACK, TXT_DB, bsearch, qsort.
8594 * Initial support for Cryptographic Message Syntax (aka CMS) based
8595 on RFC3850, RFC3851 and RFC3852. New cms directory and cms utility,
8596 support for data, signedData, compressedData, digestedData and
8597 encryptedData, envelopedData types included. Scripts to check against
8598 RFC4134 examples draft and interop and consistency checks of many
8599 content types and variants.
8603 * Add options to enc utility to support use of zlib compression BIO.
8607 * Extend mk1mf to support importing of options and assembly language
8608 files from Configure script, currently only included in VC-WIN32.
8609 The assembly language rules can now optionally generate the source
8610 files from the associated perl scripts.
8614 * Implement remaining functionality needed to support GOST ciphersuites.
8615 Interop testing has been performed using CryptoPro implementations.
8617 *Victor B. Wagner <vitus@cryptocom.ru>*
8619 * s390x assembler pack.
8623 * ARMv4 assembler pack. ARMv4 refers to v4 and later ISA, not CPU
8628 * Implement Opaque PRF Input TLS extension as specified in
8629 draft-rescorla-tls-opaque-prf-input-00.txt. Since this is not an
8630 official specification yet and no extension type assignment by
8631 IANA exists, this extension (for now) will have to be explicitly
8632 enabled when building OpenSSL by providing the extension number
8633 to use. For example, specify an option
8635 -DTLSEXT_TYPE_opaque_prf_input=0x9527
8637 to the "config" or "Configure" script to enable the extension,
8638 assuming extension number 0x9527 (which is a completely arbitrary
8639 and unofficial assignment based on the MD5 hash of the Internet
8640 Draft). Note that by doing so, you potentially lose
8641 interoperability with other TLS implementations since these might
8642 be using the same extension number for other purposes.
8644 SSL_set_tlsext_opaque_prf_input(ssl, src, len) is used to set the
8645 opaque PRF input value to use in the handshake. This will create
8646 an internal copy of the length-'len' string at 'src', and will
8647 return non-zero for success.
8649 To get more control and flexibility, provide a callback function
8652 SSL_CTX_set_tlsext_opaque_prf_input_callback(ctx, cb)
8653 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg(ctx, arg)
8657 int (*cb)(SSL *, void *peerinput, size_t len, void *arg);
8660 Callback function 'cb' will be called in handshakes, and is
8661 expected to use SSL_set_tlsext_opaque_prf_input() as appropriate.
8662 Argument 'arg' is for application purposes (the value as given to
8663 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg() will directly
8664 be provided to the callback function). The callback function
8665 has to return non-zero to report success: usually 1 to use opaque
8666 PRF input just if possible, or 2 to enforce use of the opaque PRF
8667 input. In the latter case, the library will abort the handshake
8668 if opaque PRF input is not successfully negotiated.
8670 Arguments 'peerinput' and 'len' given to the callback function
8671 will always be NULL and 0 in the case of a client. A server will
8672 see the client's opaque PRF input through these variables if
8673 available (NULL and 0 otherwise). Note that if the server
8674 provides an opaque PRF input, the length must be the same as the
8675 length of the client's opaque PRF input.
8677 Note that the callback function will only be called when creating
8678 a new session (session resumption can resume whatever was
8679 previously negotiated), and will not be called in SSL 2.0
8680 handshakes; thus, SSL_CTX_set_options(ctx, SSL_OP_NO_SSLv2) or
8681 SSL_set_options(ssl, SSL_OP_NO_SSLv2) is especially recommended
8682 for applications that need to enforce opaque PRF input.
8686 * Update ssl code to support digests other than SHA1+MD5 for handshake
8689 *Victor B. Wagner <vitus@cryptocom.ru>*
8691 * Add RFC4507 support to OpenSSL. This includes the corrections in
8692 RFC4507bis. The encrypted ticket format is an encrypted encoded
8693 SSL_SESSION structure, that way new session features are automatically
8696 If a client application caches session in an SSL_SESSION structure
8697 support is transparent because tickets are now stored in the encoded
8700 The SSL_CTX structure automatically generates keys for ticket
8701 protection in servers so again support should be possible
8702 with no application modification.
8704 If a client or server wishes to disable RFC4507 support then the option
8705 SSL_OP_NO_TICKET can be set.
8707 Add a TLS extension debugging callback to allow the contents of any client
8708 or server extensions to be examined.
8710 This work was sponsored by Google.
8714 * Final changes to avoid use of pointer pointer casts in OpenSSL.
8715 OpenSSL should now compile cleanly on gcc 4.2
8717 *Peter Hartley <pdh@utter.chaos.org.uk>, Steve Henson*
8719 * Update SSL library to use new EVP_PKEY MAC API. Include generic MAC
8720 support including streaming MAC support: this is required for GOST
8721 ciphersuite support.
8723 *Victor B. Wagner <vitus@cryptocom.ru>, Steve Henson*
8725 * Add option -stream to use PKCS#7 streaming in smime utility. New
8726 function i2d_PKCS7_bio_stream() and PEM_write_PKCS7_bio_stream()
8727 to output in BER and PEM format.
8731 * Experimental support for use of HMAC via EVP_PKEY interface. This
8732 allows HMAC to be handled via the `EVP_DigestSign*()` interface. The
8733 EVP_PKEY "key" in this case is the HMAC key, potentially allowing
8734 ENGINE support for HMAC keys which are unextractable. New -mac and
8735 -macopt options to dgst utility.
8739 * New option -sigopt to dgst utility. Update dgst to use
8740 `EVP_Digest{Sign,Verify}*`. These two changes make it possible to use
8741 alternative signing parameters such as X9.31 or PSS in the dgst
8746 * Change ssl_cipher_apply_rule(), the internal function that does
8747 the work each time a ciphersuite string requests enabling
8748 ("foo+bar"), moving ("+foo+bar"), disabling ("-foo+bar", or
8749 removing ("!foo+bar") a class of ciphersuites: Now it maintains
8750 the order of disabled ciphersuites such that those ciphersuites
8751 that most recently went from enabled to disabled not only stay
8752 in order with respect to each other, but also have higher priority
8753 than other disabled ciphersuites the next time ciphersuites are
8756 This means that you can now say, e.g., "PSK:-PSK:HIGH" to enable
8757 the same ciphersuites as with "HIGH" alone, but in a specific
8758 order where the PSK ciphersuites come first (since they are the
8759 most recently disabled ciphersuites when "HIGH" is parsed).
8761 Also, change ssl_create_cipher_list() (using this new
8762 functionality) such that between otherwise identical
8763 ciphersuites, ephemeral ECDH is preferred over ephemeral DH in
8768 * Change ssl_create_cipher_list() so that it automatically
8769 arranges the ciphersuites in reasonable order before starting
8770 to process the rule string. Thus, the definition for "DEFAULT"
8771 (SSL_DEFAULT_CIPHER_LIST) now is just "ALL:!aNULL:!eNULL", but
8772 remains equivalent to `"AES:ALL:!aNULL:!eNULL:+aECDH:+kRSA:+RC4:@STRENGTH"`.
8773 This makes it much easier to arrive at a reasonable default order
8774 in applications for which anonymous ciphers are OK (meaning
8775 that you can't actually use DEFAULT).
8777 *Bodo Moeller; suggested by Victor Duchovni*
8779 * Split the SSL/TLS algorithm mask (as used for ciphersuite string
8780 processing) into multiple integers instead of setting
8781 "SSL_MKEY_MASK" bits, "SSL_AUTH_MASK" bits, "SSL_ENC_MASK",
8782 "SSL_MAC_MASK", and "SSL_SSL_MASK" bits all in a single integer.
8783 (These masks as well as the individual bit definitions are hidden
8784 away into the non-exported interface ssl/ssl_locl.h, so this
8785 change to the definition of the SSL_CIPHER structure shouldn't
8786 affect applications.) This give us more bits for each of these
8787 categories, so there is no longer a need to coagulate AES128 and
8788 AES256 into a single algorithm bit, and to coagulate Camellia128
8789 and Camellia256 into a single algorithm bit, which has led to all
8792 Thus, among other things, the kludge introduced in 0.9.7m and
8793 0.9.8e for masking out AES256 independently of AES128 or masking
8794 out Camellia256 independently of AES256 is not needed here in 0.9.9.
8796 With the change, we also introduce new ciphersuite aliases that
8797 so far were missing: "AES128", "AES256", "CAMELLIA128", and
8802 * Add support for dsa-with-SHA224 and dsa-with-SHA256.
8803 Use the leftmost N bytes of the signature input if the input is
8804 larger than the prime q (with N being the size in bytes of q).
8808 * Very *very* experimental PKCS#7 streaming encoder support. Nothing uses
8809 it yet and it is largely untested.
8813 * Add support for the ecdsa-with-SHA224/256/384/512 signature types.
8817 * Initial incomplete changes to avoid need for function casts in OpenSSL
8818 some compilers (gcc 4.2 and later) reject their use. Safestack is
8819 reimplemented. Update ASN1 to avoid use of legacy functions.
8823 * Win32/64 targets are linked with Winsock2.
8827 * Add an X509_CRL_METHOD structure to allow CRL processing to be redirected
8828 to external functions. This can be used to increase CRL handling
8829 efficiency especially when CRLs are very large by (for example) storing
8830 the CRL revoked certificates in a database.
8834 * Overhaul of by_dir code. Add support for dynamic loading of CRLs so
8835 new CRLs added to a directory can be used. New command line option
8836 -verify_return_error to s_client and s_server. This causes real errors
8837 to be returned by the verify callback instead of carrying on no matter
8838 what. This reflects the way a "real world" verify callback would behave.
8842 * GOST engine, supporting several GOST algorithms and public key formats.
8843 Kindly donated by Cryptocom.
8847 * Partial support for Issuing Distribution Point CRL extension. CRLs
8848 partitioned by DP are handled but no indirect CRL or reason partitioning
8849 (yet). Complete overhaul of CRL handling: now the most suitable CRL is
8850 selected via a scoring technique which handles IDP and AKID in CRLs.
8854 * New X509_STORE_CTX callbacks lookup_crls() and lookup_certs() which
8855 will ultimately be used for all verify operations: this will remove the
8856 X509_STORE dependency on certificate verification and allow alternative
8857 lookup methods. X509_STORE based implementations of these two callbacks.
8861 * Allow multiple CRLs to exist in an X509_STORE with matching issuer names.
8862 Modify get_crl() to find a valid (unexpired) CRL if possible.
8866 * New function X509_CRL_match() to check if two CRLs are identical. Normally
8867 this would be called X509_CRL_cmp() but that name is already used by
8868 a function that just compares CRL issuer names. Cache several CRL
8869 extensions in X509_CRL structure and cache CRLDP in X509.
8873 * Store a "canonical" representation of X509_NAME structure (ASN1 Name)
8874 this maps equivalent X509_NAME structures into a consistent structure.
8875 Name comparison can then be performed rapidly using memcmp().
8879 * Non-blocking OCSP request processing. Add -timeout option to ocsp
8884 * Allow digests to supply their own micalg string for S/MIME type using
8885 the ctrl EVP_MD_CTRL_MICALG.
8889 * During PKCS7 signing pass the PKCS7 SignerInfo structure to the
8890 EVP_PKEY_METHOD before and after signing via the EVP_PKEY_CTRL_PKCS7_SIGN
8891 ctrl. It can then customise the structure before and/or after signing
8896 * New function OBJ_add_sigid() to allow application defined signature OIDs
8897 to be added to OpenSSLs internal tables. New function OBJ_sigid_free()
8898 to free up any added signature OIDs.
8902 * New functions EVP_CIPHER_do_all(), EVP_CIPHER_do_all_sorted(),
8903 EVP_MD_do_all() and EVP_MD_do_all_sorted() to enumerate internal
8904 digest and cipher tables. New options added to openssl utility:
8905 list-message-digest-algorithms and list-cipher-algorithms.
8909 * Change the array representation of binary polynomials: the list
8910 of degrees of non-zero coefficients is now terminated with -1.
8911 Previously it was terminated with 0, which was also part of the
8912 value; thus, the array representation was not applicable to
8913 polynomials where t^0 has coefficient zero. This change makes
8914 the array representation useful in a more general context.
8918 * Various modifications and fixes to SSL/TLS cipher string
8919 handling. For ECC, the code now distinguishes between fixed ECDH
8920 with RSA certificates on the one hand and with ECDSA certificates
8921 on the other hand, since these are separate ciphersuites. The
8922 unused code for Fortezza ciphersuites has been removed.
8924 For consistency with EDH, ephemeral ECDH is now called "EECDH"
8925 (not "ECDHE"). For consistency with the code for DH
8926 certificates, use of ECDH certificates is now considered ECDH
8927 authentication, not RSA or ECDSA authentication (the latter is
8928 merely the CA's signing algorithm and not actively used in the
8931 The temporary ciphersuite alias "ECCdraft" is no longer
8932 available, and ECC ciphersuites are no longer excluded from "ALL"
8933 and "DEFAULT". The following aliases now exist for RFC 4492
8934 ciphersuites, most of these by analogy with the DH case:
8936 kECDHr - ECDH cert, signed with RSA
8937 kECDHe - ECDH cert, signed with ECDSA
8938 kECDH - ECDH cert (signed with either RSA or ECDSA)
8939 kEECDH - ephemeral ECDH
8940 ECDH - ECDH cert or ephemeral ECDH
8946 AECDH - anonymous ECDH
8947 EECDH - non-anonymous ephemeral ECDH (equivalent to "kEECDH:-AECDH")
8951 * Add additional S/MIME capabilities for AES and GOST ciphers if supported.
8952 Use correct micalg parameters depending on digest(s) in signed message.
8956 * Add engine support for EVP_PKEY_ASN1_METHOD. Add functions to process
8957 an ENGINE asn1 method. Support ENGINE lookups in the ASN1 code.
8961 * Initial engine support for EVP_PKEY_METHOD. New functions to permit
8962 an engine to register a method. Add ENGINE lookups for methods and
8963 functional reference processing.
8967 * New functions `EVP_Digest{Sign,Verify)*`. These are enhanced versions of
8968 `EVP_{Sign,Verify}*` which allow an application to customise the signature
8973 * New -resign option to smime utility. This adds one or more signers
8974 to an existing PKCS#7 signedData structure. Also -md option to use an
8975 alternative message digest algorithm for signing.
8979 * Tidy up PKCS#7 routines and add new functions to make it easier to
8980 create PKCS7 structures containing multiple signers. Update smime
8981 application to support multiple signers.
8985 * New -macalg option to pkcs12 utility to allow setting of an alternative
8990 * Initial support for PKCS#5 v2.0 PRFs other than default SHA1 HMAC.
8991 Reorganize PBE internals to lookup from a static table using NIDs,
8992 add support for HMAC PBE OID translation. Add a EVP_CIPHER ctrl:
8993 EVP_CTRL_PBE_PRF_NID this allows a cipher to specify an alternative
8994 PRF which will be automatically used with PBES2.
8998 * Replace the algorithm specific calls to generate keys in "req" with the
9003 * Update PKCS#7 enveloped data routines to use new API. This is now
9004 supported by any public key method supporting the encrypt operation. A
9005 ctrl is added to allow the public key algorithm to examine or modify
9006 the PKCS#7 RecipientInfo structure if it needs to: for RSA this is
9011 * Add a ctrl to asn1 method to allow a public key algorithm to express
9012 a default digest type to use. In most cases this will be SHA1 but some
9013 algorithms (such as GOST) need to specify an alternative digest. The
9014 return value indicates how strong the preference is 1 means optional and
9015 2 is mandatory (that is it is the only supported type). Modify
9016 ASN1_item_sign() to accept a NULL digest argument to indicate it should
9017 use the default md. Update openssl utilities to use the default digest
9018 type for signing if it is not explicitly indicated.
9022 * Use OID cross reference table in ASN1_sign() and ASN1_verify(). New
9023 EVP_MD flag EVP_MD_FLAG_PKEY_METHOD_SIGNATURE. This uses the relevant
9024 signing method from the key type. This effectively removes the link
9025 between digests and public key types.
9029 * Add an OID cross reference table and utility functions. Its purpose is to
9030 translate between signature OIDs such as SHA1WithrsaEncryption and SHA1,
9031 rsaEncryption. This will allow some of the algorithm specific hackery
9032 needed to use the correct OID to be removed.
9036 * Remove algorithm specific dependencies when setting PKCS7_SIGNER_INFO
9037 structures for PKCS7_sign(). They are now set up by the relevant public
9042 * Add provisional EC pkey method with support for ECDSA and ECDH.
9046 * Add support for key derivation (agreement) in the API, DH method and
9051 * Add DSA pkey method and DH pkey methods, extend DH ASN1 method to support
9052 public and private key formats. As a side effect these add additional
9053 command line functionality not previously available: DSA signatures can be
9054 generated and verified using pkeyutl and DH key support and generation in
9061 *Oliver Tappe <zooey@hirschkaefer.de>*
9063 * New make target "install_html_docs" installs HTML renditions of the
9066 *Oliver Tappe <zooey@hirschkaefer.de>*
9068 * New utility "genpkey" this is analogous to "genrsa" etc except it can
9069 generate keys for any algorithm. Extend and update EVP_PKEY_METHOD to
9070 support key and parameter generation and add initial key generation
9071 functionality for RSA.
9075 * Add functions for main EVP_PKEY_method operations. The undocumented
9076 functions `EVP_PKEY_{encrypt,decrypt}` have been renamed to
9077 `EVP_PKEY_{encrypt,decrypt}_old`.
9081 * Initial definitions for EVP_PKEY_METHOD. This will be a high level public
9082 key API, doesn't do much yet.
9086 * New function EVP_PKEY_asn1_get0_info() to retrieve information about
9087 public key algorithms. New option to openssl utility:
9088 "list-public-key-algorithms" to print out info.
9092 * Implement the Supported Elliptic Curves Extension for
9093 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
9097 * Don't free up OIDs in OBJ_cleanup() if they are in use by EVP_MD or
9098 EVP_CIPHER structures to avoid later problems in EVP_cleanup().
9102 * New utilities pkey and pkeyparam. These are similar to algorithm specific
9103 utilities such as rsa, dsa, dsaparam etc except they process any key
9108 * Transfer public key printing routines to EVP_PKEY_ASN1_METHOD. New
9109 functions EVP_PKEY_print_public(), EVP_PKEY_print_private(),
9110 EVP_PKEY_print_param() to print public key data from an EVP_PKEY
9115 * Initial support for pluggable public key ASN1.
9116 De-spaghettify the public key ASN1 handling. Move public and private
9117 key ASN1 handling to a new EVP_PKEY_ASN1_METHOD structure. Relocate
9118 algorithm specific handling to a single module within the relevant
9119 algorithm directory. Add functions to allow (near) opaque processing
9120 of public and private key structures.
9124 * Implement the Supported Point Formats Extension for
9125 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
9129 * Add initial support for RFC 4279 PSK TLS ciphersuites. Add members
9130 for the psk identity [hint] and the psk callback functions to the
9131 SSL_SESSION, SSL and SSL_CTX structure.
9134 PSK-RC4-SHA, PSK-3DES-EDE-CBC-SHA, PSK-AES128-CBC-SHA,
9138 SSL_CTX_use_psk_identity_hint
9139 SSL_get_psk_identity_hint
9140 SSL_get_psk_identity
9141 SSL_use_psk_identity_hint
9143 *Mika Kousa and Pasi Eronen of Nokia Corporation*
9145 * Add RFC 3161 compliant time stamp request creation, response generation
9146 and response verification functionality.
9148 *Zoltán Glózik <zglozik@opentsa.org>, The OpenTSA Project*
9150 * Add initial support for TLS extensions, specifically for the server_name
9151 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
9152 have new members for a host name. The SSL data structure has an
9153 additional member `SSL_CTX *initial_ctx` so that new sessions can be
9154 stored in that context to allow for session resumption, even after the
9155 SSL has been switched to a new SSL_CTX in reaction to a client's
9156 server_name extension.
9158 New functions (subject to change):
9160 SSL_get_servername()
9161 SSL_get_servername_type()
9164 New CTRL codes and macros (subject to change):
9166 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
9167 - SSL_CTX_set_tlsext_servername_callback()
9168 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
9169 - SSL_CTX_set_tlsext_servername_arg()
9170 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
9172 openssl s_client has a new '-servername ...' option.
9174 openssl s_server has new options '-servername_host ...', '-cert2 ...',
9175 '-key2 ...', '-servername_fatal' (subject to change). This allows
9176 testing the HostName extension for a specific single host name ('-cert'
9177 and '-key' remain fallbacks for handshakes without HostName
9178 negotiation). If the unrecognized_name alert has to be sent, this by
9179 default is a warning; it becomes fatal with the '-servername_fatal'
9182 *Peter Sylvester, Remy Allais, Christophe Renou*
9184 * Whirlpool hash implementation is added.
9188 * BIGNUM code on 64-bit SPARCv9 targets is switched from bn(64,64) to
9189 bn(64,32). Because of instruction set limitations it doesn't have
9190 any negative impact on performance. This was done mostly in order
9191 to make it possible to share assembler modules, such as bn_mul_mont
9192 implementations, between 32- and 64-bit builds without hassle.
9196 * Move code previously exiled into file crypto/ec/ec2_smpt.c
9197 to ec2_smpl.c, and no longer require the OPENSSL_EC_BIN_PT_COMP
9202 * New candidate for BIGNUM assembler implementation, bn_mul_mont,
9203 dedicated Montgomery multiplication procedure, is introduced.
9204 BN_MONT_CTX is modified to allow bn_mul_mont to reach for higher
9205 "64-bit" performance on certain 32-bit targets.
9209 * New option SSL_OP_NO_COMP to disable use of compression selectively
9210 in SSL structures. New SSL ctrl to set maximum send fragment size.
9211 Save memory by setting the I/O buffer sizes dynamically instead of
9212 using the maximum available value.
9216 * New option -V for 'openssl ciphers'. This prints the ciphersuite code
9217 in addition to the text details.
9221 * Very, very preliminary EXPERIMENTAL support for printing of general
9222 ASN1 structures. This currently produces rather ugly output and doesn't
9223 handle several customised structures at all.
9227 * Integrated support for PVK file format and some related formats such
9228 as MS PUBLICKEYBLOB and PRIVATEKEYBLOB. Command line switches to support
9229 these in the 'rsa' and 'dsa' utilities.
9233 * Support for PKCS#1 RSAPublicKey format on rsa utility command line.
9237 * Remove the ancient ASN1_METHOD code. This was only ever used in one
9238 place for the (very old) "NETSCAPE" format certificates which are now
9239 handled using new ASN1 code equivalents.
9243 * Let the TLSv1_method() etc. functions return a 'const' SSL_METHOD
9244 pointer and make the SSL_METHOD parameter in SSL_CTX_new,
9245 SSL_CTX_set_ssl_version and SSL_set_ssl_method 'const'.
9249 * Modify CRL distribution points extension code to print out previously
9250 unsupported fields. Enhance extension setting code to allow setting of
9255 * Add print and set support for Issuing Distribution Point CRL extension.
9259 * Change 'Configure' script to enable Camellia by default.
9266 ### Changes between 0.9.8m and 0.9.8n [24 Mar 2010]
9268 * When rejecting SSL/TLS records due to an incorrect version number, never
9269 update s->server with a new major version number. As of
9270 - OpenSSL 0.9.8m if 'short' is a 16-bit type,
9271 - OpenSSL 0.9.8f if 'short' is longer than 16 bits,
9272 the previous behavior could result in a read attempt at NULL when
9273 receiving specific incorrect SSL/TLS records once record payload
9274 protection is active. ([CVE-2010-0740])
9276 *Bodo Moeller, Adam Langley <agl@chromium.org>*
9278 * Fix for CVE-2010-0433 where some kerberos enabled versions of OpenSSL
9279 could be crashed if the relevant tables were not present (e.g. chrooted).
9281 *Tomas Hoger <thoger@redhat.com>*
9283 ### Changes between 0.9.8l and 0.9.8m [25 Feb 2010]
9285 * Always check bn_wexpand() return values for failure. ([CVE-2009-3245])
9287 *Martin Olsson, Neel Mehta*
9289 * Fix X509_STORE locking: Every 'objs' access requires a lock (to
9290 accommodate for stack sorting, always a write lock!).
9294 * On some versions of WIN32 Heap32Next is very slow. This can cause
9295 excessive delays in the RAND_poll(): over a minute. As a workaround
9296 include a time check in the inner Heap32Next loop too.
9300 * The code that handled flushing of data in SSL/TLS originally used the
9301 BIO_CTRL_INFO ctrl to see if any data was pending first. This caused
9302 the problem outlined in PR#1949. The fix suggested there however can
9303 trigger problems with buggy BIO_CTRL_WPENDING (e.g. some versions
9304 of Apache). So instead simplify the code to flush unconditionally.
9305 This should be fine since flushing with no data to flush is a no op.
9309 * Handle TLS versions 2.0 and later properly and correctly use the
9310 highest version of TLS/SSL supported. Although TLS >= 2.0 is some way
9311 off ancient servers have a habit of sticking around for a while...
9315 * Modify compression code so it frees up structures without using the
9316 ex_data callbacks. This works around a problem where some applications
9317 call CRYPTO_cleanup_all_ex_data() before application exit (e.g. when
9318 restarting) then use compression (e.g. SSL with compression) later.
9319 This results in significant per-connection memory leaks and
9320 has caused some security issues including CVE-2008-1678 and
9325 * Constify crypto/cast (i.e., <openssl/cast.h>): a CAST_KEY doesn't
9326 change when encrypting or decrypting.
9330 * Add option SSL_OP_LEGACY_SERVER_CONNECT which will allow clients to
9331 connect and renegotiate with servers which do not support RI.
9332 Until RI is more widely deployed this option is enabled by default.
9336 * Add "missing" ssl ctrls to clear options and mode.
9340 * If client attempts to renegotiate and doesn't support RI respond with
9341 a no_renegotiation alert as required by RFC5746. Some renegotiating
9342 TLS clients will continue a connection gracefully when they receive
9343 the alert. Unfortunately OpenSSL mishandled this alert and would hang
9344 waiting for a server hello which it will never receive. Now we treat a
9345 received no_renegotiation alert as a fatal error. This is because
9346 applications requesting a renegotiation might well expect it to succeed
9347 and would have no code in place to handle the server denying it so the
9348 only safe thing to do is to terminate the connection.
9352 * Add ctrl macro SSL_get_secure_renegotiation_support() which returns 1 if
9353 peer supports secure renegotiation and 0 otherwise. Print out peer
9354 renegotiation support in s_client/s_server.
9358 * Replace the highly broken and deprecated SPKAC certification method with
9359 the updated NID creation version. This should correctly handle UTF8.
9363 * Implement RFC5746. Re-enable renegotiation but require the extension
9364 as needed. Unfortunately, SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION
9365 turns out to be a bad idea. It has been replaced by
9366 SSL_OP_ALLOW_UNSAFE_LEGACY_RENEGOTIATION which can be set with
9367 SSL_CTX_set_options(). This is really not recommended unless you
9368 know what you are doing.
9370 *Eric Rescorla <ekr@networkresonance.com>, Ben Laurie, Steve Henson*
9372 * Fixes to stateless session resumption handling. Use initial_ctx when
9373 issuing and attempting to decrypt tickets in case it has changed during
9374 servername handling. Use a non-zero length session ID when attempting
9375 stateless session resumption: this makes it possible to determine if
9376 a resumption has occurred immediately after receiving server hello
9377 (several places in OpenSSL subtly assume this) instead of later in
9382 * The functions ENGINE_ctrl(), OPENSSL_isservice(),
9383 CMS_get1_RecipientRequest() and RAND_bytes() can return <=0 on error
9384 fixes for a few places where the return code is not checked
9387 *Julia Lawall <julia@diku.dk>*
9389 * Add --strict-warnings option to Configure script to include devteam
9390 warnings in other configurations.
9394 * Add support for --libdir option and LIBDIR variable in makefiles. This
9395 makes it possible to install openssl libraries in locations which
9396 have names other than "lib", for example "/usr/lib64" which some
9399 *Steve Henson, based on patch from Jeremy Utley*
9401 * Don't allow the use of leading 0x80 in OIDs. This is a violation of
9402 X690 8.9.12 and can produce some misleading textual output of OIDs.
9404 *Steve Henson, reported by Dan Kaminsky*
9406 * Delete MD2 from algorithm tables. This follows the recommendation in
9407 several standards that it is not used in new applications due to
9408 several cryptographic weaknesses. For binary compatibility reasons
9409 the MD2 API is still compiled in by default.
9413 * Add compression id to {d2i,i2d}_SSL_SESSION so it is correctly saved
9418 * Rename uni2asc and asc2uni functions to OPENSSL_uni2asc and
9419 OPENSSL_asc2uni conditionally on Netware platforms to avoid a name
9422 *Guenter <lists@gknw.net>*
9424 * Fix the server certificate chain building code to use X509_verify_cert(),
9425 it used to have an ad-hoc builder which was unable to cope with anything
9426 other than a simple chain.
9428 *David Woodhouse <dwmw2@infradead.org>, Steve Henson*
9430 * Don't check self signed certificate signatures in X509_verify_cert()
9431 by default (a flag can override this): it just wastes time without
9432 adding any security. As a useful side effect self signed root CAs
9433 with non-FIPS digests are now usable in FIPS mode.
9437 * In dtls1_process_out_of_seq_message() the check if the current message
9438 is already buffered was missing. For every new message was memory
9439 allocated, allowing an attacker to perform an denial of service attack
9440 with sending out of seq handshake messages until there is no memory
9441 left. Additionally every future message was buffered, even if the
9442 sequence number made no sense and would be part of another handshake.
9443 So only messages with sequence numbers less than 10 in advance will be
9444 buffered. ([CVE-2009-1378])
9446 *Robin Seggelmann, discovered by Daniel Mentz*
9448 * Records are buffered if they arrive with a future epoch to be
9449 processed after finishing the corresponding handshake. There is
9450 currently no limitation to this buffer allowing an attacker to perform
9451 a DOS attack with sending records with future epochs until there is no
9452 memory left. This patch adds the pqueue_size() function to determine
9453 the size of a buffer and limits the record buffer to 100 entries.
9456 *Robin Seggelmann, discovered by Daniel Mentz*
9458 * Keep a copy of frag->msg_header.frag_len so it can be used after the
9459 parent structure is freed. ([CVE-2009-1379])
9463 * Handle non-blocking I/O properly in SSL_shutdown() call.
9465 *Darryl Miles <darryl-mailinglists@netbauds.net>*
9467 * Add `2.5.4.*` OIDs
9469 *Ilya O. <vrghost@gmail.com>*
9471 ### Changes between 0.9.8k and 0.9.8l [5 Nov 2009]
9473 * Disable renegotiation completely - this fixes a severe security
9474 problem ([CVE-2009-3555]) at the cost of breaking all
9475 renegotiation. Renegotiation can be re-enabled by setting
9476 SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION in s3->flags at
9477 run-time. This is really not recommended unless you know what
9482 ### Changes between 0.9.8j and 0.9.8k [25 Mar 2009]
9484 * Don't set val to NULL when freeing up structures, it is freed up by
9485 underlying code. If `sizeof(void *) > sizeof(long)` this can result in
9486 zeroing past the valid field. ([CVE-2009-0789])
9488 *Paolo Ganci <Paolo.Ganci@AdNovum.CH>*
9490 * Fix bug where return value of CMS_SignerInfo_verify_content() was not
9491 checked correctly. This would allow some invalid signed attributes to
9492 appear to verify correctly. ([CVE-2009-0591])
9494 *Ivan Nestlerode <inestlerode@us.ibm.com>*
9496 * Reject UniversalString and BMPString types with invalid lengths. This
9497 prevents a crash in ASN1_STRING_print_ex() which assumes the strings have
9498 a legal length. ([CVE-2009-0590])
9502 * Set S/MIME signing as the default purpose rather than setting it
9503 unconditionally. This allows applications to override it at the store
9508 * Permit restricted recursion of ASN1 strings. This is needed in practice
9509 to handle some structures.
9513 * Improve efficiency of mem_gets: don't search whole buffer each time
9516 *Jeremy Shapiro <jnshapir@us.ibm.com>*
9518 * New -hex option for openssl rand.
9522 * Print out UTF8String and NumericString when parsing ASN1.
9526 * Support NumericString type for name components.
9530 * Allow CC in the environment to override the automatically chosen
9531 compiler. Note that nothing is done to ensure flags work with the
9536 ### Changes between 0.9.8i and 0.9.8j [07 Jan 2009]
9538 * Properly check EVP_VerifyFinal() and similar return values
9541 *Ben Laurie, Bodo Moeller, Google Security Team*
9543 * Enable TLS extensions by default.
9547 * Allow the CHIL engine to be loaded, whether the application is
9548 multithreaded or not. (This does not release the developer from the
9549 obligation to set up the dynamic locking callbacks.)
9551 *Sander Temme <sander@temme.net>*
9553 * Use correct exit code if there is an error in dgst command.
9555 *Steve Henson; problem pointed out by Roland Dirlewanger*
9557 * Tweak Configure so that you need to say "experimental-jpake" to enable
9558 JPAKE, and need to use -DOPENSSL_EXPERIMENTAL_JPAKE in applications.
9562 * Add experimental JPAKE support, including demo authentication in
9563 s_client and s_server.
9567 * Set the comparison function in v3_addr_canonize().
9569 *Rob Austein <sra@hactrn.net>*
9571 * Add support for XMPP STARTTLS in s_client.
9573 *Philip Paeps <philip@freebsd.org>*
9575 * Change the server-side SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG behavior
9576 to ensure that even with this option, only ciphersuites in the
9577 server's preference list will be accepted. (Note that the option
9578 applies only when resuming a session, so the earlier behavior was
9579 just about the algorithm choice for symmetric cryptography.)
9583 ### Changes between 0.9.8h and 0.9.8i [15 Sep 2008]
9585 * Fix NULL pointer dereference if a DTLS server received
9586 ChangeCipherSpec as first record ([CVE-2009-1386]).
9590 * Fix a state transition in s3_srvr.c and d1_srvr.c
9591 (was using SSL3_ST_CW_CLNT_HELLO_B, should be `..._ST_SW_SRVR_...`).
9595 * The fix in 0.9.8c that supposedly got rid of unsafe
9596 double-checked locking was incomplete for RSA blinding,
9597 addressing just one layer of what turns out to have been
9598 doubly unsafe triple-checked locking.
9600 So now fix this for real by retiring the MONT_HELPER macro
9601 in crypto/rsa/rsa_eay.c.
9603 *Bodo Moeller; problem pointed out by Marius Schilder*
9605 * Various precautionary measures:
9607 - Avoid size_t integer overflow in HASH_UPDATE (md32_common.h).
9609 - Avoid a buffer overflow in d2i_SSL_SESSION() (ssl_asn1.c).
9610 (NB: This would require knowledge of the secret session ticket key
9611 to exploit, in which case you'd be SOL either way.)
9613 - Change bn_nist.c so that it will properly handle input BIGNUMs
9614 outside the expected range.
9616 - Enforce the 'num' check in BN_div() (bn_div.c) for non-BN_DEBUG
9619 *Neel Mehta, Bodo Moeller*
9621 * Allow engines to be "soft loaded" - i.e. optionally don't die if
9622 the load fails. Useful for distros.
9624 *Ben Laurie and the FreeBSD team*
9626 * Add support for Local Machine Keyset attribute in PKCS#12 files.
9630 * Fix BN_GF2m_mod_arr() top-bit cleanup code.
9634 * Expand ENGINE to support engine supplied SSL client certificate functions.
9636 This work was sponsored by Logica.
9640 * Add CryptoAPI ENGINE to support use of RSA and DSA keys held in Windows
9641 keystores. Support for SSL/TLS client authentication too.
9642 Not compiled unless enable-capieng specified to Configure.
9644 This work was sponsored by Logica.
9648 * Fix bug in X509_ATTRIBUTE creation: don't set attribute using
9649 ASN1_TYPE_set1 if MBSTRING flag set. This bug would crash certain
9650 attribute creation routines such as certificate requests and PKCS#12
9655 ### Changes between 0.9.8g and 0.9.8h [28 May 2008]
9657 * Fix flaw if 'Server Key exchange message' is omitted from a TLS
9658 handshake which could lead to a client crash as found using the
9659 Codenomicon TLS test suite ([CVE-2008-1672])
9661 *Steve Henson, Mark Cox*
9663 * Fix double free in TLS server name extensions which could lead to
9664 a remote crash found by Codenomicon TLS test suite ([CVE-2008-0891])
9668 * Clear error queue in SSL_CTX_use_certificate_chain_file()
9670 Clear the error queue to ensure that error entries left from
9671 older function calls do not interfere with the correct operation.
9673 *Lutz Jaenicke, Erik de Castro Lopo*
9675 * Remove root CA certificates of commercial CAs:
9677 The OpenSSL project does not recommend any specific CA and does not
9678 have any policy with respect to including or excluding any CA.
9679 Therefore it does not make any sense to ship an arbitrary selection
9680 of root CA certificates with the OpenSSL software.
9684 * RSA OAEP patches to fix two separate invalid memory reads.
9685 The first one involves inputs when 'lzero' is greater than
9686 'SHA_DIGEST_LENGTH' (it would read about SHA_DIGEST_LENGTH bytes
9687 before the beginning of from). The second one involves inputs where
9688 the 'db' section contains nothing but zeroes (there is a one-byte
9689 invalid read after the end of 'db').
9691 *Ivan Nestlerode <inestlerode@us.ibm.com>*
9693 * Partial backport from 0.9.9-dev:
9695 Introduce bn_mul_mont (dedicated Montgomery multiplication
9696 procedure) as a candidate for BIGNUM assembler implementation.
9697 While 0.9.9-dev uses assembler for various architectures, only
9698 x86_64 is available by default here in the 0.9.8 branch, and
9699 32-bit x86 is available through a compile-time setting.
9701 To try the 32-bit x86 assembler implementation, use Configure
9702 option "enable-montasm" (which exists only for this backport).
9704 As "enable-montasm" for 32-bit x86 disclaims code stability
9705 anyway, in this constellation we activate additional code
9706 backported from 0.9.9-dev for further performance improvements,
9707 namely BN_from_montgomery_word. (To enable this otherwise,
9708 e.g. x86_64, try `-DMONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD`.)
9710 *Andy Polyakov (backport partially by Bodo Moeller)*
9712 * Add TLS session ticket callback. This allows an application to set
9713 TLS ticket cipher and HMAC keys rather than relying on hardcoded fixed
9714 values. This is useful for key rollover for example where several key
9715 sets may exist with different names.
9719 * Reverse ENGINE-internal logic for caching default ENGINE handles.
9720 This was broken until now in 0.9.8 releases, such that the only way
9721 a registered ENGINE could be used (assuming it initialises
9722 successfully on the host) was to explicitly set it as the default
9723 for the relevant algorithms. This is in contradiction with 0.9.7
9724 behaviour and the documentation. With this fix, when an ENGINE is
9725 registered into a given algorithm's table of implementations, the
9726 'uptodate' flag is reset so that auto-discovery will be used next
9727 time a new context for that algorithm attempts to select an
9730 *Ian Lister (tweaked by Geoff Thorpe)*
9732 * Backport of CMS code to OpenSSL 0.9.8. This differs from the 0.9.9
9733 implementation in the following ways:
9735 Lack of EVP_PKEY_ASN1_METHOD means algorithm parameters have to be
9738 Lack of BER streaming support means one pass streaming processing is
9739 only supported if data is detached: setting the streaming flag is
9740 ignored for embedded content.
9742 CMS support is disabled by default and must be explicitly enabled
9743 with the enable-cms configuration option.
9747 * Update the GMP engine glue to do direct copies between BIGNUM and
9748 mpz_t when openssl and GMP use the same limb size. Otherwise the
9749 existing "conversion via a text string export" trick is still used.
9751 *Paul Sheer <paulsheer@gmail.com>*
9753 * Zlib compression BIO. This is a filter BIO which compressed and
9754 uncompresses any data passed through it.
9758 * Add AES_wrap_key() and AES_unwrap_key() functions to implement
9759 RFC3394 compatible AES key wrapping.
9763 * Add utility functions to handle ASN1 structures. ASN1_STRING_set0():
9764 sets string data without copying. X509_ALGOR_set0() and
9765 X509_ALGOR_get0(): set and retrieve X509_ALGOR (AlgorithmIdentifier)
9766 data. Attribute function X509at_get0_data_by_OBJ(): retrieves data
9767 from an X509_ATTRIBUTE structure optionally checking it occurs only
9768 once. ASN1_TYPE_set1(): set and ASN1_TYPE structure copying supplied
9773 * Fix BN flag handling in RSA_eay_mod_exp() and BN_MONT_CTX_set()
9774 to get the expected BN_FLG_CONSTTIME behavior.
9776 *Bodo Moeller (Google)*
9780 - fixed wrong usage of ioctlsocket() when build for LIBC BSD sockets
9781 - fixed do_tests.pl to run the test suite with CLIB builds too (CLIB_OPT)
9782 - added some more tests to do_tests.pl
9783 - fixed RunningProcess usage so that it works with newer LIBC NDKs too
9784 - removed usage of BN_LLONG for CLIB builds to avoid runtime dependency
9785 - added new Configure targets netware-clib-bsdsock, netware-clib-gcc,
9786 netware-clib-bsdsock-gcc, netware-libc-bsdsock-gcc
9787 - various changes to netware.pl to enable gcc-cross builds on Win32
9789 - changed crypto/bio/b_sock.c to work with macro functions (CLIB BSD)
9790 - various changes to fix missing prototype warnings
9791 - fixed x86nasm.pl to create correct asm files for NASM COFF output
9792 - added AES, WHIRLPOOL and CPUID assembler code to build files
9793 - added missing AES assembler make rules to mk1mf.pl
9794 - fixed order of includes in `apps/ocsp.c` so that `e_os.h` settings apply
9796 *Guenter Knauf <eflash@gmx.net>*
9798 * Implement certificate status request TLS extension defined in RFC3546.
9799 A client can set the appropriate parameters and receive the encoded
9800 OCSP response via a callback. A server can query the supplied parameters
9801 and set the encoded OCSP response in the callback. Add simplified examples
9802 to s_client and s_server.
9806 ### Changes between 0.9.8f and 0.9.8g [19 Oct 2007]
9809 + Binary incompatibility of ssl_ctx_st structure
9810 + DTLS interoperation with non-compliant servers
9811 + Don't call get_session_cb() without proposed session
9812 + Fix ia64 assembler code
9814 *Andy Polyakov, Steve Henson*
9816 ### Changes between 0.9.8e and 0.9.8f [11 Oct 2007]
9818 * DTLS Handshake overhaul. There were longstanding issues with
9819 OpenSSL DTLS implementation, which were making it impossible for
9820 RFC 4347 compliant client to communicate with OpenSSL server.
9821 Unfortunately just fixing these incompatibilities would "cut off"
9822 pre-0.9.8f clients. To allow for hassle free upgrade post-0.9.8e
9823 server keeps tolerating non RFC compliant syntax. The opposite is
9824 not true, 0.9.8f client can not communicate with earlier server.
9825 This update even addresses CVE-2007-4995.
9829 * Changes to avoid need for function casts in OpenSSL: some compilers
9830 (gcc 4.2 and later) reject their use.
9831 *Kurt Roeckx <kurt@roeckx.be>, Peter Hartley <pdh@utter.chaos.org.uk>,
9834 * Add RFC4507 support to OpenSSL. This includes the corrections in
9835 RFC4507bis. The encrypted ticket format is an encrypted encoded
9836 SSL_SESSION structure, that way new session features are automatically
9839 If a client application caches session in an SSL_SESSION structure
9840 support is transparent because tickets are now stored in the encoded
9843 The SSL_CTX structure automatically generates keys for ticket
9844 protection in servers so again support should be possible
9845 with no application modification.
9847 If a client or server wishes to disable RFC4507 support then the option
9848 SSL_OP_NO_TICKET can be set.
9850 Add a TLS extension debugging callback to allow the contents of any client
9851 or server extensions to be examined.
9853 This work was sponsored by Google.
9857 * Add initial support for TLS extensions, specifically for the server_name
9858 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
9859 have new members for a host name. The SSL data structure has an
9860 additional member `SSL_CTX *initial_ctx` so that new sessions can be
9861 stored in that context to allow for session resumption, even after the
9862 SSL has been switched to a new SSL_CTX in reaction to a client's
9863 server_name extension.
9865 New functions (subject to change):
9867 SSL_get_servername()
9868 SSL_get_servername_type()
9871 New CTRL codes and macros (subject to change):
9873 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
9874 - SSL_CTX_set_tlsext_servername_callback()
9875 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
9876 - SSL_CTX_set_tlsext_servername_arg()
9877 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
9879 openssl s_client has a new '-servername ...' option.
9881 openssl s_server has new options '-servername_host ...', '-cert2 ...',
9882 '-key2 ...', '-servername_fatal' (subject to change). This allows
9883 testing the HostName extension for a specific single host name ('-cert'
9884 and '-key' remain fallbacks for handshakes without HostName
9885 negotiation). If the unrecognized_name alert has to be sent, this by
9886 default is a warning; it becomes fatal with the '-servername_fatal'
9889 *Peter Sylvester, Remy Allais, Christophe Renou, Steve Henson*
9891 * Add AES and SSE2 assembly language support to VC++ build.
9895 * Mitigate attack on final subtraction in Montgomery reduction.
9899 * Fix crypto/ec/ec_mult.c to work properly with scalars of value 0
9900 (which previously caused an internal error).
9904 * Squeeze another 10% out of IGE mode when in != out.
9908 * AES IGE mode speedup.
9910 *Dean Gaudet (Google)*
9912 * Add the Korean symmetric 128-bit cipher SEED (see
9913 <http://www.kisa.or.kr/kisa/seed/jsp/seed_eng.jsp>) and
9914 add SEED ciphersuites from RFC 4162:
9916 TLS_RSA_WITH_SEED_CBC_SHA = "SEED-SHA"
9917 TLS_DHE_DSS_WITH_SEED_CBC_SHA = "DHE-DSS-SEED-SHA"
9918 TLS_DHE_RSA_WITH_SEED_CBC_SHA = "DHE-RSA-SEED-SHA"
9919 TLS_DH_anon_WITH_SEED_CBC_SHA = "ADH-SEED-SHA"
9921 To minimize changes between patchlevels in the OpenSSL 0.9.8
9922 series, SEED remains excluded from compilation unless OpenSSL
9923 is configured with 'enable-seed'.
9925 *KISA, Bodo Moeller*
9927 * Mitigate branch prediction attacks, which can be practical if a
9928 single processor is shared, allowing a spy process to extract
9929 information. For detailed background information, see
9930 <http://eprint.iacr.org/2007/039> (O. Aciicmez, S. Gueron,
9931 J.-P. Seifert, "New Branch Prediction Vulnerabilities in OpenSSL
9932 and Necessary Software Countermeasures"). The core of the change
9933 are new versions BN_div_no_branch() and
9934 BN_mod_inverse_no_branch() of BN_div() and BN_mod_inverse(),
9935 respectively, which are slower, but avoid the security-relevant
9936 conditional branches. These are automatically called by BN_div()
9937 and BN_mod_inverse() if the flag BN_FLG_CONSTTIME is set for one
9938 of the input BIGNUMs. Also, BN_is_bit_set() has been changed to
9939 remove a conditional branch.
9941 BN_FLG_CONSTTIME is the new name for the previous
9942 BN_FLG_EXP_CONSTTIME flag, since it now affects more than just
9943 modular exponentiation. (Since OpenSSL 0.9.7h, setting this flag
9944 in the exponent causes BN_mod_exp_mont() to use the alternative
9945 implementation in BN_mod_exp_mont_consttime().) The old name
9946 remains as a deprecated alias.
9948 Similarly, RSA_FLAG_NO_EXP_CONSTTIME is replaced by a more general
9949 RSA_FLAG_NO_CONSTTIME flag since the RSA implementation now uses
9950 constant-time implementations for more than just exponentiation.
9951 Here too the old name is kept as a deprecated alias.
9953 BN_BLINDING_new() will now use BN_dup() for the modulus so that
9954 the BN_BLINDING structure gets an independent copy of the
9955 modulus. This means that the previous `BIGNUM *m` argument to
9956 BN_BLINDING_new() and to BN_BLINDING_create_param() now
9957 essentially becomes `const BIGNUM *m`, although we can't actually
9958 change this in the header file before 0.9.9. It allows
9959 RSA_setup_blinding() to use BN_with_flags() on the modulus to
9960 enable BN_FLG_CONSTTIME.
9962 *Matthew D Wood (Intel Corp)*
9964 * In the SSL/TLS server implementation, be strict about session ID
9965 context matching (which matters if an application uses a single
9966 external cache for different purposes). Previously,
9967 out-of-context reuse was forbidden only if SSL_VERIFY_PEER was
9968 set. This did ensure strict client verification, but meant that,
9969 with applications using a single external cache for quite
9970 different requirements, clients could circumvent ciphersuite
9971 restrictions for a given session ID context by starting a session
9972 in a different context.
9976 * Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
9977 a ciphersuite string such as "DEFAULT:RSA" cannot enable
9978 authentication-only ciphersuites.
9982 * Update the SSL_get_shared_ciphers() fix CVE-2006-3738 which was
9983 not complete and could lead to a possible single byte overflow
9984 ([CVE-2007-5135]) [Ben Laurie]
9986 ### Changes between 0.9.8d and 0.9.8e [23 Feb 2007]
9988 * Since AES128 and AES256 (and similarly Camellia128 and
9989 Camellia256) share a single mask bit in the logic of
9990 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
9991 kludge to work properly if AES128 is available and AES256 isn't
9992 (or if Camellia128 is available and Camellia256 isn't).
9996 * Fix the BIT STRING encoding generated by crypto/ec/ec_asn1.c
9997 (within i2d_ECPrivateKey, i2d_ECPKParameters, i2d_ECParameters):
9998 When a point or a seed is encoded in a BIT STRING, we need to
9999 prevent the removal of trailing zero bits to get the proper DER
10000 encoding. (By default, crypto/asn1/a_bitstr.c assumes the case
10001 of a NamedBitList, for which trailing 0 bits need to be removed.)
10005 * Have SSL/TLS server implementation tolerate "mismatched" record
10006 protocol version while receiving ClientHello even if the
10007 ClientHello is fragmented. (The server can't insist on the
10008 particular protocol version it has chosen before the ServerHello
10009 message has informed the client about his choice.)
10013 * Add RFC 3779 support.
10015 *Rob Austein for ARIN, Ben Laurie*
10017 * Load error codes if they are not already present instead of using a
10018 static variable. This allows them to be cleanly unloaded and reloaded.
10019 Improve header file function name parsing.
10023 * extend SMTP and IMAP protocol emulation in s_client to use EHLO
10024 or CAPABILITY handshake as required by RFCs.
10026 *Goetz Babin-Ebell*
10028 ### Changes between 0.9.8c and 0.9.8d [28 Sep 2006]
10030 * Introduce limits to prevent malicious keys being able to
10031 cause a denial of service. ([CVE-2006-2940])
10033 *Steve Henson, Bodo Moeller*
10035 * Fix ASN.1 parsing of certain invalid structures that can result
10036 in a denial of service. ([CVE-2006-2937]) [Steve Henson]
10038 * Fix buffer overflow in SSL_get_shared_ciphers() function.
10039 ([CVE-2006-3738]) [Tavis Ormandy and Will Drewry, Google Security Team]
10041 * Fix SSL client code which could crash if connecting to a
10042 malicious SSLv2 server. ([CVE-2006-4343])
10044 *Tavis Ormandy and Will Drewry, Google Security Team*
10046 * Since 0.9.8b, ciphersuite strings naming explicit ciphersuites
10047 match only those. Before that, "AES256-SHA" would be interpreted
10048 as a pattern and match "AES128-SHA" too (since AES128-SHA got
10049 the same strength classification in 0.9.7h) as we currently only
10050 have a single AES bit in the ciphersuite description bitmap.
10051 That change, however, also applied to ciphersuite strings such as
10052 "RC4-MD5" that intentionally matched multiple ciphersuites --
10053 namely, SSL 2.0 ciphersuites in addition to the more common ones
10054 from SSL 3.0/TLS 1.0.
10056 So we change the selection algorithm again: Naming an explicit
10057 ciphersuite selects this one ciphersuite, and any other similar
10058 ciphersuite (same bitmap) from *other* protocol versions.
10059 Thus, "RC4-MD5" again will properly select both the SSL 2.0
10060 ciphersuite and the SSL 3.0/TLS 1.0 ciphersuite.
10062 Since SSL 2.0 does not have any ciphersuites for which the
10063 128/256 bit distinction would be relevant, this works for now.
10064 The proper fix will be to use different bits for AES128 and
10065 AES256, which would have avoided the problems from the beginning;
10066 however, bits are scarce, so we can only do this in a new release
10067 (not just a patchlevel) when we can change the SSL_CIPHER
10068 definition to split the single 'unsigned long mask' bitmap into
10069 multiple values to extend the available space.
10073 ### Changes between 0.9.8b and 0.9.8c [05 Sep 2006]
10075 * Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
10076 ([CVE-2006-4339]) [Ben Laurie and Google Security Team]
10078 * Add AES IGE and biIGE modes.
10082 * Change the Unix randomness entropy gathering to use poll() when
10083 possible instead of select(), since the latter has some
10084 undesirable limitations.
10086 *Darryl Miles via Richard Levitte and Bodo Moeller*
10088 * Disable "ECCdraft" ciphersuites more thoroughly. Now special
10089 treatment in ssl/ssl_ciph.s makes sure that these ciphersuites
10090 cannot be implicitly activated as part of, e.g., the "AES" alias.
10091 However, please upgrade to OpenSSL 0.9.9[-dev] for
10092 non-experimental use of the ECC ciphersuites to get TLS extension
10093 support, which is required for curve and point format negotiation
10094 to avoid potential handshake problems.
10098 * Disable rogue ciphersuites:
10100 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
10101 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
10102 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
10104 The latter two were purportedly from
10105 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
10108 Also deactivate the remaining ciphersuites from
10109 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
10110 unofficial, and the ID has long expired.
10114 * Fix RSA blinding Heisenbug (problems sometimes occurred on
10115 dual-core machines) and other potential thread-safety issues.
10119 * Add the symmetric cipher Camellia (128-bit, 192-bit, 256-bit key
10120 versions), which is now available for royalty-free use
10121 (see <http://info.isl.ntt.co.jp/crypt/eng/info/chiteki.html>).
10122 Also, add Camellia TLS ciphersuites from RFC 4132.
10124 To minimize changes between patchlevels in the OpenSSL 0.9.8
10125 series, Camellia remains excluded from compilation unless OpenSSL
10126 is configured with 'enable-camellia'.
10130 * Disable the padding bug check when compression is in use. The padding
10131 bug check assumes the first packet is of even length, this is not
10132 necessarily true if compression is enabled and can result in false
10133 positives causing handshake failure. The actual bug test is ancient
10134 code so it is hoped that implementations will either have fixed it by
10135 now or any which still have the bug do not support compression.
10139 ### Changes between 0.9.8a and 0.9.8b [04 May 2006]
10141 * When applying a cipher rule check to see if string match is an explicit
10142 cipher suite and only match that one cipher suite if it is.
10146 * Link in manifests for VC++ if needed.
10148 *Austin Ziegler <halostatue@gmail.com>*
10150 * Update support for ECC-based TLS ciphersuites according to
10151 draft-ietf-tls-ecc-12.txt with proposed changes (but without
10152 TLS extensions, which are supported starting with the 0.9.9
10153 branch, not in the OpenSSL 0.9.8 branch).
10157 * New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support
10158 opaque EVP_CIPHER_CTX handling.
10162 * Fixes and enhancements to zlib compression code. We now only use
10163 "zlib1.dll" and use the default `__cdecl` calling convention on Win32
10164 to conform with the standards mentioned here:
10165 <http://www.zlib.net/DLL_FAQ.txt>
10166 Static zlib linking now works on Windows and the new --with-zlib-include
10167 --with-zlib-lib options to Configure can be used to supply the location
10168 of the headers and library. Gracefully handle case where zlib library
10173 * Several fixes and enhancements to the OID generation code. The old code
10174 sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't
10175 handle numbers larger than ULONG_MAX, truncated printing and had a
10176 non standard OBJ_obj2txt() behaviour.
10180 * Add support for building of engines under engine/ as shared libraries
10181 under VC++ build system.
10185 * Corrected the numerous bugs in the Win32 path splitter in DSO.
10186 Hopefully, we will not see any false combination of paths any more.
10190 ### Changes between 0.9.8 and 0.9.8a [11 Oct 2005]
10192 * Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
10193 (part of SSL_OP_ALL). This option used to disable the
10194 countermeasure against man-in-the-middle protocol-version
10195 rollback in the SSL 2.0 server implementation, which is a bad
10196 idea. ([CVE-2005-2969])
10198 *Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
10199 for Information Security, National Institute of Advanced Industrial
10200 Science and Technology [AIST], Japan)*
10202 * Add two function to clear and return the verify parameter flags.
10206 * Keep cipherlists sorted in the source instead of sorting them at
10207 runtime, thus removing the need for a lock.
10211 * Avoid some small subgroup attacks in Diffie-Hellman.
10213 *Nick Mathewson and Ben Laurie*
10215 * Add functions for well-known primes.
10219 * Extended Windows CE support.
10221 *Satoshi Nakamura and Andy Polyakov*
10223 * Initialize SSL_METHOD structures at compile time instead of during
10224 runtime, thus removing the need for a lock.
10228 * Make PKCS7_decrypt() work even if no certificate is supplied by
10229 attempting to decrypt each encrypted key in turn. Add support to
10234 ### Changes between 0.9.7h and 0.9.8 [05 Jul 2005]
10236 [NB: OpenSSL 0.9.7i and later 0.9.7 patch levels were released after
10239 * Add libcrypto.pc and libssl.pc for those who feel they need them.
10243 * Change CA.sh and CA.pl so they don't bundle the CSR and the private
10244 key into the same file any more.
10248 * Add initial support for Win64, both IA64 and AMD64/x64 flavors.
10252 * Add -utf8 command line and config file option to 'ca'.
10254 *Stefan <stf@udoma.org*
10256 * Removed the macro des_crypt(), as it seems to conflict with some
10257 libraries. Use DES_crypt().
10261 * Correct naming of the 'chil' and '4758cca' ENGINEs. This
10262 involves renaming the source and generated shared-libs for
10263 both. The engines will accept the corrected or legacy ids
10264 ('ncipher' and '4758_cca' respectively) when binding. NB,
10265 this only applies when building 'shared'.
10267 *Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe*
10269 * Add attribute functions to EVP_PKEY structure. Modify
10270 PKCS12_create() to recognize a CSP name attribute and
10271 use it. Make -CSP option work again in pkcs12 utility.
10275 * Add new functionality to the bn blinding code:
10276 - automatic re-creation of the BN_BLINDING parameters after
10277 a fixed number of uses (currently 32)
10278 - add new function for parameter creation
10279 - introduce flags to control the update behaviour of the
10280 BN_BLINDING parameters
10281 - hide BN_BLINDING structure
10282 Add a second BN_BLINDING slot to the RSA structure to improve
10283 performance when a single RSA object is shared among several
10288 * Add support for DTLS.
10290 *Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie*
10292 * Add support for DER encoded private keys (SSL_FILETYPE_ASN1)
10293 to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file()
10297 * Remove buggy and incomplete DH cert support from
10298 ssl/ssl_rsa.c and ssl/s3_both.c
10302 * Use SHA-1 instead of MD5 as the default digest algorithm for
10303 the `apps/openssl` commands.
10307 * Compile clean with "-Wall -Wmissing-prototypes
10308 -Wstrict-prototypes -Wmissing-declarations -Werror". Currently
10309 DEBUG_SAFESTACK must also be set.
10313 * Change ./Configure so that certain algorithms can be disabled by default.
10314 The new counterpiece to "no-xxx" is "enable-xxx".
10316 The patented RC5 and MDC2 algorithms will now be disabled unless
10317 "enable-rc5" and "enable-mdc2", respectively, are specified.
10319 (IDEA remains enabled despite being patented. This is because IDEA
10320 is frequently required for interoperability, and there is no license
10321 fee for non-commercial use. As before, "no-idea" can be used to
10322 avoid this algorithm.)
10326 * Add processing of proxy certificates (see RFC 3820). This work was
10327 sponsored by KTH (The Royal Institute of Technology in Stockholm) and
10328 EGEE (Enabling Grids for E-science in Europe).
10332 * RC4 performance overhaul on modern architectures/implementations, such
10333 as Intel P4, IA-64 and AMD64.
10337 * New utility extract-section.pl. This can be used specify an alternative
10338 section number in a pod file instead of having to treat each file as
10339 a separate case in Makefile. This can be done by adding two lines to the
10342 =for comment openssl_section:XXX
10344 The blank line is mandatory.
10348 * New arguments -certform, -keyform and -pass for s_client and s_server
10349 to allow alternative format key and certificate files and passphrase
10354 * New structure X509_VERIFY_PARAM which combines current verify parameters,
10355 update associated structures and add various utility functions.
10357 Add new policy related verify parameters, include policy checking in
10358 standard verify code. Enhance 'smime' application with extra parameters
10359 to support policy checking and print out.
10363 * Add a new engine to support VIA PadLock ACE extensions in the VIA C3
10364 Nehemiah processors. These extensions support AES encryption in hardware
10365 as well as RNG (though RNG support is currently disabled).
10367 *Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov*
10369 * Deprecate `BN_[get|set]_params()` functions (they were ignored internally).
10373 * New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented.
10375 *Andy Polyakov and a number of other people*
10377 * Improved PowerPC platform support. Most notably BIGNUM assembler
10378 implementation contributed by IBM.
10380 *Suresh Chari, Peter Waltenberg, Andy Polyakov*
10382 * The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public
10383 exponent rather than 'unsigned long'. There is a corresponding change to
10384 the new 'rsa_keygen' element of the RSA_METHOD structure.
10386 *Jelte Jansen, Geoff Thorpe*
10388 * Functionality for creating the initial serial number file is now
10389 moved from CA.pl to the 'ca' utility with a new option -create_serial.
10391 (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial
10392 number file to 1, which is bound to cause problems. To avoid
10393 the problems while respecting compatibility between different 0.9.7
10394 patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in
10395 CA.pl for serial number initialization. With the new release 0.9.8,
10396 we can fix the problem directly in the 'ca' utility.)
10400 * Reduced header interdependencies by declaring more opaque objects in
10401 ossl_typ.h. As a consequence, including some headers (eg. engine.h) will
10402 give fewer recursive includes, which could break lazy source code - so
10403 this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always,
10404 developers should define this symbol when building and using openssl to
10405 ensure they track the recommended behaviour, interfaces, [etc], but
10406 backwards-compatible behaviour prevails when this isn't defined.
10410 * New function X509_POLICY_NODE_print() which prints out policy nodes.
10414 * Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality.
10415 This will generate a random key of the appropriate length based on the
10416 cipher context. The EVP_CIPHER can provide its own random key generation
10417 routine to support keys of a specific form. This is used in the des and
10418 3des routines to generate a key of the correct parity. Update S/MIME
10419 code to use new functions and hence generate correct parity DES keys.
10420 Add EVP_CHECK_DES_KEY #define to return an error if the key is not
10421 valid (weak or incorrect parity).
10425 * Add a local set of CRLs that can be used by X509_verify_cert() as well
10426 as looking them up. This is useful when the verified structure may contain
10427 CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs
10428 present unless the new PKCS7_NO_CRL flag is asserted.
10432 * Extend ASN1 oid configuration module. It now additionally accepts the
10435 shortName = some long name, 1.2.3.4
10439 * Reimplemented the BN_CTX implementation. There is now no more static
10440 limitation on the number of variables it can handle nor the depth of the
10441 "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack
10442 information can now expand as required, and rather than having a single
10443 static array of bignums, BN_CTX now uses a linked-list of such arrays
10444 allowing it to expand on demand whilst maintaining the usefulness of
10445 BN_CTX's "bundling".
10449 * Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD
10450 to allow all RSA operations to function using a single BN_CTX.
10454 * Preliminary support for certificate policy evaluation and checking. This
10455 is initially intended to pass the tests outlined in "Conformance Testing
10456 of Relying Party Client Certificate Path Processing Logic" v1.07.
10460 * bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and
10461 remained unused and not that useful. A variety of other little bignum
10462 tweaks and fixes have also been made continuing on from the audit (see
10467 * Constify all or almost all d2i, c2i, s2i and r2i functions, along with
10468 associated ASN1, EVP and SSL functions and old ASN1 macros.
10472 * BN_zero() only needs to set 'top' and 'neg' to zero for correct results,
10473 and this should never fail. So the return value from the use of
10474 BN_set_word() (which can fail due to needless expansion) is now deprecated;
10475 if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro.
10479 * BN_CTX_get() should return zero-valued bignums, providing the same
10480 initialised value as BN_new().
10482 *Geoff Thorpe, suggested by Ulf Möller*
10484 * Support for inhibitAnyPolicy certificate extension.
10488 * An audit of the BIGNUM code is underway, for which debugging code is
10489 enabled when BN_DEBUG is defined. This makes stricter enforcements on what
10490 is considered valid when processing BIGNUMs, and causes execution to
10491 assert() when a problem is discovered. If BN_DEBUG_RAND is defined,
10492 further steps are taken to deliberately pollute unused data in BIGNUM
10493 structures to try and expose faulty code further on. For now, openssl will
10494 (in its default mode of operation) continue to tolerate the inconsistent
10495 forms that it has tolerated in the past, but authors and packagers should
10496 consider trying openssl and their own applications when compiled with
10497 these debugging symbols defined. It will help highlight potential bugs in
10498 their own code, and will improve the test coverage for OpenSSL itself. At
10499 some point, these tighter rules will become openssl's default to improve
10500 maintainability, though the assert()s and other overheads will remain only
10501 in debugging configurations. See bn.h for more details.
10503 *Geoff Thorpe, Nils Larsch, Ulf Möller*
10505 * BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure
10506 that can only be obtained through BN_CTX_new() (which implicitly
10507 initialises it). The presence of this function only made it possible
10508 to overwrite an existing structure (and cause memory leaks).
10512 * Because of the callback-based approach for implementing LHASH as a
10513 template type, lh_insert() adds opaque objects to hash-tables and
10514 lh_doall() or lh_doall_arg() are typically used with a destructor callback
10515 to clean up those corresponding objects before destroying the hash table
10516 (and losing the object pointers). So some over-zealous constifications in
10517 LHASH have been relaxed so that lh_insert() does not take (nor store) the
10518 objects as "const" and the `lh_doall[_arg]` callback wrappers are not
10519 prototyped to have "const" restrictions on the object pointers they are
10520 given (and so aren't required to cast them away any more).
10524 * The tmdiff.h API was so ugly and minimal that our own timing utility
10525 (speed) prefers to use its own implementation. The two implementations
10526 haven't been consolidated as yet (volunteers?) but the tmdiff API has had
10527 its object type properly exposed (MS_TM) instead of casting to/from
10528 `char *`. This may still change yet if someone realises MS_TM and
10530 aren't necessarily the greatest nomenclatures - but this is what was used
10531 internally to the implementation so I've used that for now.
10535 * Ensure that deprecated functions do not get compiled when
10536 OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of
10537 the self-tests were still using deprecated key-generation functions so
10538 these have been updated also.
10542 * Reorganise PKCS#7 code to separate the digest location functionality
10543 into PKCS7_find_digest(), digest addition into PKCS7_bio_add_digest().
10544 New function PKCS7_set_digest() to set the digest type for PKCS#7
10545 digestedData type. Add additional code to correctly generate the
10546 digestedData type and add support for this type in PKCS7 initialization
10551 * New function PKCS7_set0_type_other() this initializes a PKCS7
10552 structure of type "other".
10556 * Fix prime generation loop in crypto/bn/bn_prime.pl by making
10557 sure the loop does correctly stop and breaking ("division by zero")
10558 modulus operations are not performed. The (pre-generated) prime
10559 table crypto/bn/bn_prime.h was already correct, but it could not be
10560 re-generated on some platforms because of the "division by zero"
10561 situation in the script.
10563 *Ralf S. Engelschall*
10565 * Update support for ECC-based TLS ciphersuites according to
10566 draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with
10567 SHA-1 now is only used for "small" curves (where the
10568 representation of a field element takes up to 24 bytes); for
10569 larger curves, the field element resulting from ECDH is directly
10570 used as premaster secret.
10572 *Douglas Stebila (Sun Microsystems Laboratories)*
10574 * Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2
10575 curve secp160r1 to the tests.
10577 *Douglas Stebila (Sun Microsystems Laboratories)*
10579 * Add the possibility to load symbols globally with DSO.
10581 *Götz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte*
10583 * Add the functions ERR_set_mark() and ERR_pop_to_mark() for better
10584 control of the error stack.
10588 * Add support for STORE in ENGINE.
10592 * Add the STORE type. The intention is to provide a common interface
10593 to certificate and key stores, be they simple file-based stores, or
10594 HSM-type store, or LDAP stores, or...
10595 NOTE: The code is currently UNTESTED and isn't really used anywhere.
10599 * Add a generic structure called OPENSSL_ITEM. This can be used to
10600 pass a list of arguments to any function as well as provide a way
10601 for a function to pass data back to the caller.
10605 * Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup()
10606 works like BUF_strdup() but can be used to duplicate a portion of
10607 a string. The copy gets NUL-terminated. BUF_memdup() duplicates
10612 * Add the function sk_find_ex() which works like sk_find(), but will
10613 return an index to an element even if an exact match couldn't be
10614 found. The index is guaranteed to point at the element where the
10615 searched-for key would be inserted to preserve sorting order.
10619 * Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but
10620 takes an extra flags argument for optional functionality. Currently,
10621 the following flags are defined:
10623 OBJ_BSEARCH_VALUE_ON_NOMATCH
10624 This one gets OBJ_bsearch_ex() to return a pointer to the first
10625 element where the comparing function returns a negative or zero
10628 OBJ_BSEARCH_FIRST_VALUE_ON_MATCH
10629 This one gets OBJ_bsearch_ex() to return a pointer to the first
10630 element where the comparing function returns zero. This is useful
10631 if there are more than one element where the comparing function
10636 * Make it possible to create self-signed certificates with 'openssl ca'
10637 in such a way that the self-signed certificate becomes part of the
10638 CA database and uses the same mechanisms for serial number generation
10639 as all other certificate signing. The new flag '-selfsign' enables
10640 this functionality. Adapt CA.sh and CA.pl.in.
10644 * Add functionality to check the public key of a certificate request
10645 against a given private. This is useful to check that a certificate
10646 request can be signed by that key (self-signing).
10650 * Make it possible to have multiple active certificates with the same
10651 subject in the CA index file. This is done only if the keyword
10652 'unique_subject' is set to 'no' in the main CA section (default
10653 if 'CA_default') of the configuration file. The value is saved
10654 with the database itself in a separate index attribute file,
10655 named like the index file with '.attr' appended to the name.
10659 * Generate multi-valued AVAs using '+' notation in config files for
10664 * Support for nameConstraints certificate extension.
10668 * Support for policyConstraints certificate extension.
10672 * Support for policyMappings certificate extension.
10676 * Make sure the default DSA_METHOD implementation only uses its
10677 dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL,
10678 and change its own handlers to be NULL so as to remove unnecessary
10679 indirection. This lets alternative implementations fallback to the
10680 default implementation more easily.
10684 * Support for directoryName in GeneralName related extensions
10689 * Make it possible to link applications using Makefile.shared.
10690 Make that possible even when linking against static libraries!
10694 * Support for single pass processing for S/MIME signing. This now
10695 means that S/MIME signing can be done from a pipe, in addition
10696 cleartext signing (multipart/signed type) is effectively streaming
10697 and the signed data does not need to be all held in memory.
10699 This is done with a new flag PKCS7_STREAM. When this flag is set
10700 PKCS7_sign() only initializes the PKCS7 structure and the actual signing
10701 is done after the data is output (and digests calculated) in
10702 SMIME_write_PKCS7().
10706 * Add full support for -rpath/-R, both in shared libraries and
10707 applications, at least on the platforms where it's known how
10712 * In crypto/ec/ec_mult.c, implement fast point multiplication with
10713 precomputation, based on wNAF splitting: EC_GROUP_precompute_mult()
10714 will now compute a table of multiples of the generator that
10715 makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul()
10716 faster (notably in the case of a single point multiplication,
10717 scalar * generator).
10719 *Nils Larsch, Bodo Moeller*
10721 * IPv6 support for certificate extensions. The various extensions
10722 which use the IP:a.b.c.d can now take IPv6 addresses using the
10723 formats of RFC1884 2.2 . IPv6 addresses are now also displayed
10728 * Added an ENGINE that implements RSA by performing private key
10729 exponentiations with the GMP library. The conversions to and from
10730 GMP's mpz_t format aren't optimised nor are any montgomery forms
10731 cached, and on x86 it appears OpenSSL's own performance has caught up.
10732 However there are likely to be other architectures where GMP could
10733 provide a boost. This ENGINE is not built in by default, but it can be
10734 specified at Configure time and should be accompanied by the necessary
10735 linker additions, eg;
10736 ./config -DOPENSSL_USE_GMP -lgmp
10740 * "openssl engine" will not display ENGINE/DSO load failure errors when
10741 testing availability of engines with "-t" - the old behaviour is
10742 produced by increasing the feature's verbosity with "-tt".
10746 * ECDSA routines: under certain error conditions uninitialized BN objects
10747 could be freed. Solution: make sure initialization is performed early
10748 enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de>
10753 * Key-generation can now be implemented in RSA_METHOD, DSA_METHOD
10754 and DH_METHOD (eg. by ENGINE implementations) to override the normal
10755 software implementations. For DSA and DH, parameter generation can
10756 also be overridden by providing the appropriate method callbacks.
10760 * Change the "progress" mechanism used in key-generation and
10761 primality testing to functions that take a new BN_GENCB pointer in
10762 place of callback/argument pairs. The new API functions have `_ex`
10763 postfixes and the older functions are reimplemented as wrappers for
10764 the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide
10765 declarations of the old functions to help (graceful) attempts to
10766 migrate to the new functions. Also, the new key-generation API
10767 functions operate on a caller-supplied key-structure and return
10768 success/failure rather than returning a key or NULL - this is to
10769 help make "keygen" another member function of RSA_METHOD etc.
10771 Example for using the new callback interface:
10773 int (*my_callback)(int a, int b, BN_GENCB *cb) = ...;
10774 void *my_arg = ...;
10777 BN_GENCB_set(&my_cb, my_callback, my_arg);
10779 return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb);
10780 /* For the meaning of a, b in calls to my_callback(), see the
10781 * documentation of the function that calls the callback.
10782 * cb will point to my_cb; my_arg can be retrieved as cb->arg.
10783 * my_callback should return 1 if it wants BN_is_prime_ex()
10784 * to continue, or 0 to stop.
10789 * Change the ZLIB compression method to be stateful, and make it
10790 available to TLS with the number defined in
10791 draft-ietf-tls-compression-04.txt.
10795 * Add the ASN.1 structures and functions for CertificatePair, which
10796 is defined as follows (according to X.509_4thEditionDraftV6.pdf):
10798 CertificatePair ::= SEQUENCE {
10799 forward [0] Certificate OPTIONAL,
10800 reverse [1] Certificate OPTIONAL,
10801 -- at least one of the pair shall be present -- }
10803 Also implement the PEM functions to read and write certificate
10804 pairs, and defined the PEM tag as "CERTIFICATE PAIR".
10806 This needed to be defined, mostly for the sake of the LDAP
10807 attribute crossCertificatePair, but may prove useful elsewhere as
10812 * Make it possible to inhibit symlinking of shared libraries in
10813 Makefile.shared, for Cygwin's sake.
10817 * Extend the BIGNUM API by creating a function
10818 void BN_set_negative(BIGNUM *a, int neg);
10819 and a macro that behave like
10820 int BN_is_negative(const BIGNUM *a);
10822 to avoid the need to access 'a->neg' directly in applications.
10826 * Implement fast modular reduction for pseudo-Mersenne primes
10827 used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c).
10828 EC_GROUP_new_curve_GFp() will now automatically use this
10831 *Nils Larsch <nla@trustcenter.de>*
10833 * Add new lock type (CRYPTO_LOCK_BN).
10837 * Change the ENGINE framework to automatically load engines
10838 dynamically from specific directories unless they could be
10839 found to already be built in or loaded. Move all the
10840 current engines except for the cryptodev one to a new
10841 directory engines/.
10842 The engines in engines/ are built as shared libraries if
10843 the "shared" options was given to ./Configure or ./config.
10844 Otherwise, they are inserted in libcrypto.a.
10845 /usr/local/ssl/engines is the default directory for dynamic
10846 engines, but that can be overridden at configure time through
10847 the usual use of --prefix and/or --openssldir, and at run
10848 time with the environment variable OPENSSL_ENGINES.
10850 *Geoff Thorpe and Richard Levitte*
10852 * Add Makefile.shared, a helper makefile to build shared
10853 libraries. Adapt Makefile.org.
10857 * Add version info to Win32 DLLs.
10859 *Peter 'Luna' Runestig" <peter@runestig.com>*
10861 * Add new 'medium level' PKCS#12 API. Certificates and keys
10862 can be added using this API to created arbitrary PKCS#12
10863 files while avoiding the low-level API.
10865 New options to PKCS12_create(), key or cert can be NULL and
10866 will then be omitted from the output file. The encryption
10867 algorithm NIDs can be set to -1 for no encryption, the mac
10868 iteration count can be set to 0 to omit the mac.
10870 Enhance pkcs12 utility by making the -nokeys and -nocerts
10871 options work when creating a PKCS#12 file. New option -nomac
10872 to omit the mac, NONE can be set for an encryption algorithm.
10873 New code is modified to use the enhanced PKCS12_create()
10874 instead of the low-level API.
10878 * Extend ASN1 encoder to support indefinite length constructed
10879 encoding. This can output sequences tags and octet strings in
10880 this form. Modify pk7_asn1.c to support indefinite length
10881 encoding. This is experimental and needs additional code to
10882 be useful, such as an ASN1 bio and some enhanced streaming
10885 Extend template encode functionality so that tagging is passed
10886 down to the template encoder.
10890 * Let 'openssl req' fail if an argument to '-newkey' is not
10891 recognized instead of using RSA as a default.
10895 * Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt.
10896 As these are not official, they are not included in "ALL";
10897 the "ECCdraft" ciphersuite group alias can be used to select them.
10899 *Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)*
10901 * Add ECDH engine support.
10903 *Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)*
10905 * Add ECDH in new directory crypto/ecdh/.
10907 *Douglas Stebila (Sun Microsystems Laboratories)*
10909 * Let BN_rand_range() abort with an error after 100 iterations
10910 without success (which indicates a broken PRNG).
10914 * Change BN_mod_sqrt() so that it verifies that the input value
10915 is really the square of the return value. (Previously,
10916 BN_mod_sqrt would show GIGO behaviour.)
10920 * Add named elliptic curves over binary fields from X9.62, SECG,
10921 and WAP/WTLS; add OIDs that were still missing.
10923 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10925 * Extend the EC library for elliptic curves over binary fields
10926 (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/).
10929 EC_GF2m_simple_method
10933 EC_GROUP_new_curve_GF2m
10934 EC_GROUP_set_curve_GF2m
10935 EC_GROUP_get_curve_GF2m
10936 EC_POINT_set_affine_coordinates_GF2m
10937 EC_POINT_get_affine_coordinates_GF2m
10938 EC_POINT_set_compressed_coordinates_GF2m
10940 Point compression for binary fields is disabled by default for
10941 patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to
10944 As binary polynomials are represented as BIGNUMs, various members
10945 of the EC_GROUP and EC_POINT data structures can be shared
10946 between the implementations for prime fields and binary fields;
10947 the above `..._GF2m functions` (except for EX_GROUP_new_curve_GF2m)
10948 are essentially identical to their `..._GFp` counterparts.
10949 (For simplicity, the `..._GFp` prefix has been dropped from
10950 various internal method names.)
10952 An internal 'field_div' method (similar to 'field_mul' and
10953 'field_sqr') has been added; this is used only for binary fields.
10955 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10957 * Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult()
10958 through methods ('mul', 'precompute_mult').
10960 The generic implementations (now internally called 'ec_wNAF_mul'
10961 and 'ec_wNAF_precomputed_mult') remain the default if these
10962 methods are undefined.
10964 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10966 * New function EC_GROUP_get_degree, which is defined through
10967 EC_METHOD. For curves over prime fields, this returns the bit
10968 length of the modulus.
10970 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10972 * New functions EC_GROUP_dup, EC_POINT_dup.
10973 (These simply call ..._new and ..._copy).
10975 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10977 * Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c.
10978 Polynomials are represented as BIGNUMs (where the sign bit is not
10979 used) in the following functions [macros]:
10982 BN_GF2m_sub [= BN_GF2m_add]
10983 BN_GF2m_mod [wrapper for BN_GF2m_mod_arr]
10984 BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr]
10985 BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr]
10987 BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr]
10988 BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr]
10989 BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr]
10990 BN_GF2m_cmp [= BN_ucmp]
10992 (Note that only the 'mod' functions are actually for fields GF(2^m).
10993 BN_GF2m_add() is misnomer, but this is for the sake of consistency.)
10995 For some functions, an the irreducible polynomial defining a
10996 field can be given as an 'unsigned int[]' with strictly
10997 decreasing elements giving the indices of those bits that are set;
10998 i.e., p[] represents the polynomial
10999 f(t) = t^p[0] + t^p[1] + ... + t^p[k]
11001 p[0] > p[1] > ... > p[k] = 0.
11002 This applies to the following functions:
11005 BN_GF2m_mod_mul_arr
11006 BN_GF2m_mod_sqr_arr
11007 BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv]
11008 BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div]
11009 BN_GF2m_mod_exp_arr
11010 BN_GF2m_mod_sqrt_arr
11011 BN_GF2m_mod_solve_quad_arr
11015 Conversion can be performed by the following functions:
11020 bntest.c has additional tests for binary polynomial arithmetic.
11022 Two implementations for BN_GF2m_mod_div() are available.
11023 The default algorithm simply uses BN_GF2m_mod_inv() and
11024 BN_GF2m_mod_mul(). The alternative algorithm is compiled in only
11025 if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the
11026 copyright notice in crypto/bn/bn_gf2m.c before enabling it).
11028 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
11030 * Add new error code 'ERR_R_DISABLED' that can be used when some
11031 functionality is disabled at compile-time.
11033 *Douglas Stebila <douglas.stebila@sun.com>*
11035 * Change default behaviour of 'openssl asn1parse' so that more
11036 information is visible when viewing, e.g., a certificate:
11038 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump'
11039 mode the content of non-printable OCTET STRINGs is output in a
11040 style similar to INTEGERs, but with '[HEX DUMP]' prepended to
11041 avoid the appearance of a printable string.
11043 *Nils Larsch <nla@trustcenter.de>*
11045 * Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access
11047 EC_GROUP_set_asn1_flag()
11048 EC_GROUP_get_asn1_flag()
11049 EC_GROUP_set_point_conversion_form()
11050 EC_GROUP_get_point_conversion_form()
11051 These control ASN1 encoding details:
11052 - Curves (i.e., groups) are encoded explicitly unless asn1_flag
11053 has been set to OPENSSL_EC_NAMED_CURVE.
11054 - Points are encoded in uncompressed form by default; options for
11055 asn1_for are as for point2oct, namely
11056 POINT_CONVERSION_COMPRESSED
11057 POINT_CONVERSION_UNCOMPRESSED
11058 POINT_CONVERSION_HYBRID
11060 Also add 'seed' and 'seed_len' members to EC_GROUP with access
11062 EC_GROUP_set_seed()
11063 EC_GROUP_get0_seed()
11064 EC_GROUP_get_seed_len()
11065 This is used only for ASN1 purposes (so far).
11067 *Nils Larsch <nla@trustcenter.de>*
11069 * Add 'field_type' member to EC_METHOD, which holds the NID
11070 of the appropriate field type OID. The new function
11071 EC_METHOD_get_field_type() returns this value.
11073 *Nils Larsch <nla@trustcenter.de>*
11076 EC_POINT_point2bn()
11077 EC_POINT_bn2point()
11078 EC_POINT_point2hex()
11079 EC_POINT_hex2point()
11080 providing useful interfaces to EC_POINT_point2oct() and
11081 EC_POINT_oct2point().
11083 *Nils Larsch <nla@trustcenter.de>*
11085 * Change internals of the EC library so that the functions
11086 EC_GROUP_set_generator()
11087 EC_GROUP_get_generator()
11088 EC_GROUP_get_order()
11089 EC_GROUP_get_cofactor()
11090 are implemented directly in crypto/ec/ec_lib.c and not dispatched
11091 to methods, which would lead to unnecessary code duplication when
11092 adding different types of curves.
11094 *Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller*
11096 * Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM
11097 arithmetic, and such that modified wNAFs are generated
11098 (which avoid length expansion in many cases).
11102 * Add a function EC_GROUP_check_discriminant() (defined via
11103 EC_METHOD) that verifies that the curve discriminant is non-zero.
11105 Add a function EC_GROUP_check() that makes some sanity tests
11106 on a EC_GROUP, its generator and order. This includes
11107 EC_GROUP_check_discriminant().
11109 *Nils Larsch <nla@trustcenter.de>*
11111 * Add ECDSA in new directory crypto/ecdsa/.
11113 Add applications 'openssl ecparam' and 'openssl ecdsa'
11114 (these are based on 'openssl dsaparam' and 'openssl dsa').
11116 ECDSA support is also included in various other files across the
11117 library. Most notably,
11118 - 'openssl req' now has a '-newkey ecdsa:file' option;
11119 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA;
11120 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and
11121 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make
11122 them suitable for ECDSA where domain parameters must be
11123 extracted before the specific public key;
11124 - ECDSA engine support has been added.
11126 *Nils Larsch <nla@trustcenter.de>*
11128 * Include some named elliptic curves, and add OIDs from X9.62,
11129 SECG, and WAP/WTLS. Each curve can be obtained from the new
11131 EC_GROUP_new_by_curve_name(),
11132 and the list of available named curves can be obtained with
11133 EC_get_builtin_curves().
11134 Also add a 'curve_name' member to EC_GROUP objects, which can be
11136 EC_GROUP_set_curve_name()
11137 EC_GROUP_get_curve_name()
11139 *Nils Larsch <larsch@trustcenter.de, Bodo Moeller*
11141 * Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
11142 was actually never needed) and in BN_mul(). The removal in BN_mul()
11143 required a small change in bn_mul_part_recursive() and the addition
11144 of the functions bn_cmp_part_words(), bn_sub_part_words() and
11145 bn_add_part_words(), which do the same thing as bn_cmp_words(),
11146 bn_sub_words() and bn_add_words() except they take arrays with
11151 ### Changes between 0.9.7l and 0.9.7m [23 Feb 2007]
11153 * Cleanse PEM buffers before freeing them since they may contain
11156 *Benjamin Bennett <ben@psc.edu>*
11158 * Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
11159 a ciphersuite string such as "DEFAULT:RSA" cannot enable
11160 authentication-only ciphersuites.
11164 * Since AES128 and AES256 share a single mask bit in the logic of
11165 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
11166 kludge to work properly if AES128 is available and AES256 isn't.
11170 * Expand security boundary to match 1.1.1 module.
11174 * Remove redundant features: hash file source, editing of test vectors
11175 modify fipsld to use external fips_premain.c signature.
11179 * New perl script mkfipsscr.pl to create shell scripts or batch files to
11180 run algorithm test programs.
11184 * Make algorithm test programs more tolerant of whitespace.
11188 * Have SSL/TLS server implementation tolerate "mismatched" record
11189 protocol version while receiving ClientHello even if the
11190 ClientHello is fragmented. (The server can't insist on the
11191 particular protocol version it has chosen before the ServerHello
11192 message has informed the client about his choice.)
11196 * Load error codes if they are not already present instead of using a
11197 static variable. This allows them to be cleanly unloaded and reloaded.
11201 ### Changes between 0.9.7k and 0.9.7l [28 Sep 2006]
11203 * Introduce limits to prevent malicious keys being able to
11204 cause a denial of service. ([CVE-2006-2940])
11206 *Steve Henson, Bodo Moeller*
11208 * Fix ASN.1 parsing of certain invalid structures that can result
11209 in a denial of service. ([CVE-2006-2937]) [Steve Henson]
11211 * Fix buffer overflow in SSL_get_shared_ciphers() function.
11212 ([CVE-2006-3738]) [Tavis Ormandy and Will Drewry, Google Security Team]
11214 * Fix SSL client code which could crash if connecting to a
11215 malicious SSLv2 server. ([CVE-2006-4343])
11217 *Tavis Ormandy and Will Drewry, Google Security Team*
11219 * Change ciphersuite string processing so that an explicit
11220 ciphersuite selects this one ciphersuite (so that "AES256-SHA"
11221 will no longer include "AES128-SHA"), and any other similar
11222 ciphersuite (same bitmap) from *other* protocol versions (so that
11223 "RC4-MD5" will still include both the SSL 2.0 ciphersuite and the
11224 SSL 3.0/TLS 1.0 ciphersuite). This is a backport combining
11225 changes from 0.9.8b and 0.9.8d.
11229 ### Changes between 0.9.7j and 0.9.7k [05 Sep 2006]
11231 * Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
11232 ([CVE-2006-4339]) [Ben Laurie and Google Security Team]
11234 * Change the Unix randomness entropy gathering to use poll() when
11235 possible instead of select(), since the latter has some
11236 undesirable limitations.
11238 *Darryl Miles via Richard Levitte and Bodo Moeller*
11240 * Disable rogue ciphersuites:
11242 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
11243 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
11244 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
11246 The latter two were purportedly from
11247 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
11250 Also deactivate the remaining ciphersuites from
11251 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
11252 unofficial, and the ID has long expired.
11256 * Fix RSA blinding Heisenbug (problems sometimes occurred on
11257 dual-core machines) and other potential thread-safety issues.
11261 ### Changes between 0.9.7i and 0.9.7j [04 May 2006]
11263 * Adapt fipsld and the build system to link against the validated FIPS
11264 module in FIPS mode.
11268 * Fixes for VC++ 2005 build under Windows.
11272 * Add new Windows build target VC-32-GMAKE for VC++. This uses GNU make
11273 from a Windows bash shell such as MSYS. It is autodetected from the
11274 "config" script when run from a VC++ environment. Modify standard VC++
11275 build to use fipscanister.o from the GNU make build.
11279 ### Changes between 0.9.7h and 0.9.7i [14 Oct 2005]
11281 * Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS.
11282 The value now differs depending on if you build for FIPS or not.
11283 BEWARE! A program linked with a shared FIPSed libcrypto can't be
11284 safely run with a non-FIPSed libcrypto, as it may crash because of
11285 the difference induced by this change.
11289 ### Changes between 0.9.7g and 0.9.7h [11 Oct 2005]
11291 * Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
11292 (part of SSL_OP_ALL). This option used to disable the
11293 countermeasure against man-in-the-middle protocol-version
11294 rollback in the SSL 2.0 server implementation, which is a bad
11295 idea. ([CVE-2005-2969])
11297 *Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
11298 for Information Security, National Institute of Advanced Industrial
11299 Science and Technology [AIST, Japan)]*
11301 * Minimal support for X9.31 signatures and PSS padding modes. This is
11302 mainly for FIPS compliance and not fully integrated at this stage.
11306 * For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform
11307 the exponentiation using a fixed-length exponent. (Otherwise,
11308 the information leaked through timing could expose the secret key
11309 after many signatures; cf. Bleichenbacher's attack on DSA with
11314 * Make a new fixed-window mod_exp implementation the default for
11315 RSA, DSA, and DH private-key operations so that the sequence of
11316 squares and multiplies and the memory access pattern are
11317 independent of the particular secret key. This will mitigate
11318 cache-timing and potential related attacks.
11320 BN_mod_exp_mont_consttime() is the new exponentiation implementation,
11321 and this is automatically used by BN_mod_exp_mont() if the new flag
11322 BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH
11323 will use this BN flag for private exponents unless the flag
11324 RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or
11325 DH_FLAG_NO_EXP_CONSTTIME, respectively, is set.
11327 *Matthew D Wood (Intel Corp), with some changes by Bodo Moeller*
11329 * Change the client implementation for SSLv23_method() and
11330 SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0
11331 Client Hello message format if the SSL_OP_NO_SSLv2 option is set.
11332 (Previously, the SSL 2.0 backwards compatible Client Hello
11333 message format would be used even with SSL_OP_NO_SSLv2.)
11337 * Add support for smime-type MIME parameter in S/MIME messages which some
11342 * New function BN_MONT_CTX_set_locked() to set montgomery parameters in
11343 a threadsafe manner. Modify rsa code to use new function and add calls
11344 to dsa and dh code (which had race conditions before).
11348 * Include the fixed error library code in the C error file definitions
11349 instead of fixing them up at runtime. This keeps the error code
11350 structures constant.
11354 ### Changes between 0.9.7f and 0.9.7g [11 Apr 2005]
11356 [NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after
11359 * Fixes for newer kerberos headers. NB: the casts are needed because
11360 the 'length' field is signed on one version and unsigned on another
11361 with no (?) obvious way to tell the difference, without these VC++
11362 complains. Also the "definition" of FAR (blank) is no longer included
11363 nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up
11364 some needed definitions.
11368 * Undo Cygwin change.
11372 * Added support for proxy certificates according to RFC 3820.
11373 Because they may be a security thread to unaware applications,
11374 they must be explicitly allowed in run-time. See
11375 docs/HOWTO/proxy_certificates.txt for further information.
11379 ### Changes between 0.9.7e and 0.9.7f [22 Mar 2005]
11381 * Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating
11382 server and client random values. Previously
11383 (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in
11384 less random data when sizeof(time_t) > 4 (some 64 bit platforms).
11386 This change has negligible security impact because:
11388 1. Server and client random values still have 24 bytes of pseudo random
11391 2. Server and client random values are sent in the clear in the initial
11394 3. The master secret is derived using the premaster secret (48 bytes in
11395 size for static RSA ciphersuites) as well as client server and random
11398 The OpenSSL team would like to thank the UK NISCC for bringing this issue
11401 *Stephen Henson, reported by UK NISCC*
11403 * Use Windows randomness collection on Cygwin.
11407 * Fix hang in EGD/PRNGD query when communication socket is closed
11408 prematurely by EGD/PRNGD.
11410 *Darren Tucker <dtucker@zip.com.au> via Lutz Jänicke, resolves #1014*
11412 * Prompt for pass phrases when appropriate for PKCS12 input format.
11416 * Back-port of selected performance improvements from development
11417 branch, as well as improved support for PowerPC platforms.
11421 * Add lots of checks for memory allocation failure, error codes to indicate
11422 failure and freeing up memory if a failure occurs.
11424 *Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson*
11426 * Add new -passin argument to dgst.
11430 * Perform some character comparisons of different types in X509_NAME_cmp:
11431 this is needed for some certificates that re-encode DNs into UTF8Strings
11432 (in violation of RFC3280) and can't or won't issue name rollover
11437 * Make an explicit check during certificate validation to see that
11438 the CA setting in each certificate on the chain is correct. As a
11439 side effect always do the following basic checks on extensions,
11440 not just when there's an associated purpose to the check:
11442 - if there is an unhandled critical extension (unless the user
11443 has chosen to ignore this fault)
11444 - if the path length has been exceeded (if one is set at all)
11445 - that certain extensions fit the associated purpose (if one has
11450 ### Changes between 0.9.7d and 0.9.7e [25 Oct 2004]
11452 * Avoid a race condition when CRLs are checked in a multi threaded
11453 environment. This would happen due to the reordering of the revoked
11454 entries during signature checking and serial number lookup. Now the
11455 encoding is cached and the serial number sort performed under a lock.
11456 Add new STACK function sk_is_sorted().
11460 * Add Delta CRL to the extension code.
11464 * Various fixes to s3_pkt.c so alerts are sent properly.
11466 *David Holmes <d.holmes@f5.com>*
11468 * Reduce the chances of duplicate issuer name and serial numbers (in
11469 violation of RFC3280) using the OpenSSL certificate creation utilities.
11470 This is done by creating a random 64 bit value for the initial serial
11471 number when a serial number file is created or when a self signed
11472 certificate is created using 'openssl req -x509'. The initial serial
11473 number file is created using 'openssl x509 -next_serial' in CA.pl
11474 rather than being initialized to 1.
11478 ### Changes between 0.9.7c and 0.9.7d [17 Mar 2004]
11480 * Fix null-pointer assignment in do_change_cipher_spec() revealed
11481 by using the Codenomicon TLS Test Tool ([CVE-2004-0079])
11483 *Joe Orton, Steve Henson*
11485 * Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites
11488 *Joe Orton, Steve Henson*
11490 * Make it possible to have multiple active certificates with the same
11491 subject in the CA index file. This is done only if the keyword
11492 'unique_subject' is set to 'no' in the main CA section (default
11493 if 'CA_default') of the configuration file. The value is saved
11494 with the database itself in a separate index attribute file,
11495 named like the index file with '.attr' appended to the name.
11499 * X509 verify fixes. Disable broken certificate workarounds when
11500 X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if
11501 keyUsage extension present. Don't accept CRLs with unhandled critical
11502 extensions: since verify currently doesn't process CRL extensions this
11503 rejects a CRL with *any* critical extensions. Add new verify error codes
11508 * When creating an OCSP nonce use an OCTET STRING inside the extnValue.
11509 A clarification of RFC2560 will require the use of OCTET STRINGs and
11510 some implementations cannot handle the current raw format. Since OpenSSL
11511 copies and compares OCSP nonces as opaque blobs without any attempt at
11512 parsing them this should not create any compatibility issues.
11516 * New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when
11517 calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without
11518 this HMAC (and other) operations are several times slower than OpenSSL
11523 * Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex().
11525 *Peter Sylvester <Peter.Sylvester@EdelWeb.fr>*
11527 * Use the correct content when signing type "other".
11531 ### Changes between 0.9.7b and 0.9.7c [30 Sep 2003]
11533 * Fix various bugs revealed by running the NISCC test suite:
11535 Stop out of bounds reads in the ASN1 code when presented with
11536 invalid tags (CVE-2003-0543 and CVE-2003-0544).
11538 Free up ASN1_TYPE correctly if ANY type is invalid ([CVE-2003-0545]).
11540 If verify callback ignores invalid public key errors don't try to check
11541 certificate signature with the NULL public key.
11545 * New -ignore_err option in ocsp application to stop the server
11546 exiting on the first error in a request.
11550 * In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
11551 if the server requested one: as stated in TLS 1.0 and SSL 3.0
11556 * In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
11557 extra data after the compression methods not only for TLS 1.0
11558 but also for SSL 3.0 (as required by the specification).
11560 *Bodo Moeller; problem pointed out by Matthias Loepfe*
11562 * Change X509_certificate_type() to mark the key as exported/exportable
11563 when it's 512 *bits* long, not 512 bytes.
11567 * Change AES_cbc_encrypt() so it outputs exact multiple of
11568 blocks during encryption.
11572 * Various fixes to base64 BIO and non blocking I/O. On write
11573 flushes were not handled properly if the BIO retried. On read
11574 data was not being buffered properly and had various logic bugs.
11575 This also affects blocking I/O when the data being decoded is a
11580 * Various S/MIME bugfixes and compatibility changes:
11581 output correct application/pkcs7 MIME type if
11582 PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures.
11583 Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening
11584 of files as .eml work). Correctly handle very long lines in MIME
11589 ### Changes between 0.9.7a and 0.9.7b [10 Apr 2003]
11591 * Countermeasure against the Klima-Pokorny-Rosa extension of
11592 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
11593 a protocol version number mismatch like a decryption error
11594 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
11598 * Turn on RSA blinding by default in the default implementation
11599 to avoid a timing attack. Applications that don't want it can call
11600 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
11601 They would be ill-advised to do so in most cases.
11603 *Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller*
11605 * Change RSA blinding code so that it works when the PRNG is not
11606 seeded (in this case, the secret RSA exponent is abused as
11607 an unpredictable seed -- if it is not unpredictable, there
11608 is no point in blinding anyway). Make RSA blinding thread-safe
11609 by remembering the creator's thread ID in rsa->blinding and
11610 having all other threads use local one-time blinding factors
11611 (this requires more computation than sharing rsa->blinding, but
11612 avoids excessive locking; and if an RSA object is not shared
11613 between threads, blinding will still be very fast).
11617 * Fixed a typo bug that would cause ENGINE_set_default() to set an
11618 ENGINE as defaults for all supported algorithms irrespective of
11619 the 'flags' parameter. 'flags' is now honoured, so applications
11620 should make sure they are passing it correctly.
11624 * Target "mingw" now allows native Windows code to be generated in
11625 the Cygwin environment as well as with the MinGW compiler.
11629 ### Changes between 0.9.7 and 0.9.7a [19 Feb 2003]
11631 * In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
11632 via timing by performing a MAC computation even if incorrect
11633 block cipher padding has been found. This is a countermeasure
11634 against active attacks where the attacker has to distinguish
11635 between bad padding and a MAC verification error. ([CVE-2003-0078])
11637 *Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
11638 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
11639 Martin Vuagnoux (EPFL, Ilion)*
11641 * Make the no-err option work as intended. The intention with no-err
11642 is not to have the whole error stack handling routines removed from
11643 libcrypto, it's only intended to remove all the function name and
11644 reason texts, thereby removing some of the footprint that may not
11645 be interesting if those errors aren't displayed anyway.
11647 NOTE: it's still possible for any application or module to have its
11648 own set of error texts inserted. The routines are there, just not
11649 used by default when no-err is given.
11653 * Add support for FreeBSD on IA64.
11655 *dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454*
11657 * Adjust DES_cbc_cksum() so it returns the same value as the MIT
11658 Kerberos function mit_des_cbc_cksum(). Before this change,
11659 the value returned by DES_cbc_cksum() was like the one from
11660 mit_des_cbc_cksum(), except the bytes were swapped.
11662 *Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte*
11664 * Allow an application to disable the automatic SSL chain building.
11665 Before this a rather primitive chain build was always performed in
11666 ssl3_output_cert_chain(): an application had no way to send the
11667 correct chain if the automatic operation produced an incorrect result.
11669 Now the chain builder is disabled if either:
11671 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert().
11673 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set.
11675 The reasoning behind this is that an application would not want the
11676 auto chain building to take place if extra chain certificates are
11677 present and it might also want a means of sending no additional
11678 certificates (for example the chain has two certificates and the
11683 * Add the possibility to build without the ENGINE framework.
11685 *Steven Reddie <smr@essemer.com.au> via Richard Levitte*
11687 * Under Win32 gmtime() can return NULL: check return value in
11688 OPENSSL_gmtime(). Add error code for case where gmtime() fails.
11692 * DSA routines: under certain error conditions uninitialized BN objects
11693 could be freed. Solution: make sure initialization is performed early
11694 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>,
11695 Nils Larsch <nla@trustcenter.de> via PR#459)
11699 * Another fix for SSLv2 session ID handling: the session ID was incorrectly
11700 checked on reconnect on the client side, therefore session resumption
11701 could still fail with a "ssl session id is different" error. This
11702 behaviour is masked when SSL_OP_ALL is used due to
11703 SSL_OP_MICROSOFT_SESS_ID_BUG being set.
11704 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
11705 followup to PR #377.
11709 * IA-32 assembler support enhancements: unified ELF targets, support
11710 for SCO/Caldera platforms, fix for Cygwin shared build.
11714 * Add support for FreeBSD on sparc64. As a consequence, support for
11715 FreeBSD on non-x86 processors is separate from x86 processors on
11716 the config script, much like the NetBSD support.
11718 *Richard Levitte & Kris Kennaway <kris@obsecurity.org>*
11720 ### Changes between 0.9.6h and 0.9.7 [31 Dec 2002]
11722 [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after
11725 * Fix session ID handling in SSLv2 client code: the SERVER FINISHED
11726 code (06) was taken as the first octet of the session ID and the last
11727 octet was ignored consequently. As a result SSLv2 client side session
11728 caching could not have worked due to the session ID mismatch between
11730 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
11735 * Change the declaration of needed Kerberos libraries to use EX_LIBS
11736 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is
11741 * The hw_ncipher.c engine requires dynamic locks. Unfortunately, it
11742 seems that in spite of existing for more than a year, many application
11743 author have done nothing to provide the necessary callbacks, which
11744 means that this particular engine will not work properly anywhere.
11745 This is a very unfortunate situation which forces us, in the name
11746 of usability, to give the hw_ncipher.c a static lock, which is part
11748 NOTE: This is for the 0.9.7 series ONLY. This hack will never
11749 appear in 0.9.8 or later. We EXPECT application authors to have
11750 dealt properly with this when 0.9.8 is released (unless we actually
11751 make such changes in the libcrypto locking code that changes will
11752 have to be made anyway).
11756 * In asn1_d2i_read_bio() repeatedly call BIO_read() until all content
11757 octets have been read, EOF or an error occurs. Without this change
11758 some truncated ASN1 structures will not produce an error.
11762 * Disable Heimdal support, since it hasn't been fully implemented.
11763 Still give the possibility to force the use of Heimdal, but with
11764 warnings and a request that patches get sent to openssl-dev.
11768 * Add the VC-CE target, introduce the WINCE sysname, and add
11769 INSTALL.WCE and appropriate conditionals to make it build.
11771 *Steven Reddie <smr@essemer.com.au> via Richard Levitte*
11773 * Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and
11774 cygssl-x.y.z.dll, where x, y and z are the major, minor and
11775 edit numbers of the version.
11777 *Corinna Vinschen <vinschen@redhat.com> and Richard Levitte*
11779 * Introduce safe string copy and catenation functions
11780 (BUF_strlcpy() and BUF_strlcat()).
11782 *Ben Laurie (CHATS) and Richard Levitte*
11784 * Avoid using fixed-size buffers for one-line DNs.
11786 *Ben Laurie (CHATS)*
11788 * Add BUF_MEM_grow_clean() to avoid information leakage when
11789 resizing buffers containing secrets, and use where appropriate.
11791 *Ben Laurie (CHATS)*
11793 * Avoid using fixed size buffers for configuration file location.
11795 *Ben Laurie (CHATS)*
11797 * Avoid filename truncation for various CA files.
11799 *Ben Laurie (CHATS)*
11801 * Use sizeof in preference to magic numbers.
11803 *Ben Laurie (CHATS)*
11805 * Avoid filename truncation in cert requests.
11807 *Ben Laurie (CHATS)*
11809 * Add assertions to check for (supposedly impossible) buffer
11812 *Ben Laurie (CHATS)*
11814 * Don't cache truncated DNS entries in the local cache (this could
11815 potentially lead to a spoofing attack).
11817 *Ben Laurie (CHATS)*
11819 * Fix various buffers to be large enough for hex/decimal
11820 representations in a platform independent manner.
11822 *Ben Laurie (CHATS)*
11824 * Add CRYPTO_realloc_clean() to avoid information leakage when
11825 resizing buffers containing secrets, and use where appropriate.
11827 *Ben Laurie (CHATS)*
11829 * Add BIO_indent() to avoid much slightly worrying code to do
11832 *Ben Laurie (CHATS)*
11834 * Convert sprintf()/BIO_puts() to BIO_printf().
11836 *Ben Laurie (CHATS)*
11838 * buffer_gets() could terminate with the buffer only half
11841 *Ben Laurie (CHATS)*
11843 * Add assertions to prevent user-supplied crypto functions from
11844 overflowing internal buffers by having large block sizes, etc.
11846 *Ben Laurie (CHATS)*
11848 * New OPENSSL_assert() macro (similar to assert(), but enabled
11851 *Ben Laurie (CHATS)*
11853 * Eliminate unused copy of key in RC4.
11855 *Ben Laurie (CHATS)*
11857 * Eliminate unused and incorrectly sized buffers for IV in pem.h.
11859 *Ben Laurie (CHATS)*
11861 * Fix off-by-one error in EGD path.
11863 *Ben Laurie (CHATS)*
11865 * If RANDFILE path is too long, ignore instead of truncating.
11867 *Ben Laurie (CHATS)*
11869 * Eliminate unused and incorrectly sized X.509 structure
11872 *Ben Laurie (CHATS)*
11874 * Eliminate unused and dangerous function knumber().
11876 *Ben Laurie (CHATS)*
11878 * Eliminate unused and dangerous structure, KSSL_ERR.
11880 *Ben Laurie (CHATS)*
11882 * Protect against overlong session ID context length in an encoded
11883 session object. Since these are local, this does not appear to be
11886 *Ben Laurie (CHATS)*
11888 * Change from security patch (see 0.9.6e below) that did not affect
11889 the 0.9.6 release series:
11891 Remote buffer overflow in SSL3 protocol - an attacker could
11892 supply an oversized master key in Kerberos-enabled versions.
11895 *Ben Laurie (CHATS)*
11897 * Change the SSL kerb5 codes to match RFC 2712.
11901 * Make -nameopt work fully for req and add -reqopt switch.
11903 *Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson*
11905 * The "block size" for block ciphers in CFB and OFB mode should be 1.
11907 *Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>*
11909 * Make sure tests can be performed even if the corresponding algorithms
11910 have been removed entirely. This was also the last step to make
11911 OpenSSL compilable with DJGPP under all reasonable conditions.
11913 *Richard Levitte, Doug Kaufman <dkaufman@rahul.net>*
11915 * Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT
11916 to allow version independent disabling of normally unselected ciphers,
11917 which may be activated as a side-effect of selecting a single cipher.
11919 (E.g., cipher list string "RSA" enables ciphersuites that are left
11920 out of "ALL" because they do not provide symmetric encryption.
11921 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.)
11923 *Lutz Jaenicke, Bodo Moeller*
11925 * Add appropriate support for separate platform-dependent build
11926 directories. The recommended way to make a platform-dependent
11927 build directory is the following (tested on Linux), maybe with
11930 # Place yourself outside of the OpenSSL source tree. In
11931 # this example, the environment variable OPENSSL_SOURCE
11932 # is assumed to contain the absolute OpenSSL source directory.
11933 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`"
11934 cd objtree/"`uname -s`-`uname -r`-`uname -m`"
11935 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do
11936 mkdir -p `dirname $F`
11937 ln -s $OPENSSL_SOURCE/$F $F
11940 To be absolutely sure not to disturb the source tree, a "make clean"
11941 is a good thing. If it isn't successful, don't worry about it,
11942 it probably means the source directory is very clean.
11946 * Make sure any ENGINE control commands make local copies of string
11947 pointers passed to them whenever necessary. Otherwise it is possible
11948 the caller may have overwritten (or deallocated) the original string
11949 data when a later ENGINE operation tries to use the stored values.
11951 *Götz Babin-Ebell <babinebell@trustcenter.de>*
11953 * Improve diagnostics in file reading and command-line digests.
11955 *Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>*
11957 * Add AES modes CFB and OFB to the object database. Correct an
11958 error in AES-CFB decryption.
11962 * Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this
11963 allows existing EVP_CIPHER_CTX structures to be reused after
11964 calling `EVP_*Final()`. This behaviour is used by encryption
11965 BIOs and some applications. This has the side effect that
11966 applications must explicitly clean up cipher contexts with
11967 EVP_CIPHER_CTX_cleanup() or they will leak memory.
11971 * Check the values of dna and dnb in bn_mul_recursive before calling
11972 bn_mul_comba (a non zero value means the a or b arrays do not contain
11973 n2 elements) and fallback to bn_mul_normal if either is not zero.
11977 * Fix escaping of non-ASCII characters when using the -subj option
11978 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>)
11982 * Make object definitions compliant to LDAP (RFC2256): SN is the short
11983 form for "surname", serialNumber has no short form.
11984 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798;
11985 therefore remove "mail" short name for "internet 7".
11986 The OID for unique identifiers in X509 certificates is
11987 x500UniqueIdentifier, not uniqueIdentifier.
11988 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>)
11992 * Add an "init" command to the ENGINE config module and auto initialize
11993 ENGINEs. Without any "init" command the ENGINE will be initialized
11994 after all ctrl commands have been executed on it. If init=1 the
11995 ENGINE is initialized at that point (ctrls before that point are run
11996 on the uninitialized ENGINE and after on the initialized one). If
11997 init=0 then the ENGINE will not be initialized at all.
12001 * Fix the 'app_verify_callback' interface so that the user-defined
12002 argument is actually passed to the callback: In the
12003 SSL_CTX_set_cert_verify_callback() prototype, the callback
12004 declaration has been changed from
12007 int (*cb)(X509_STORE_CTX *,void *);
12008 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
12009 i=s->ctx->app_verify_callback(&ctx)
12010 has been changed into
12011 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
12013 To update applications using SSL_CTX_set_cert_verify_callback(),
12014 a dummy argument can be added to their callback functions.
12016 *D. K. Smetters <smetters@parc.xerox.com>*
12018 * Added the '4758cca' ENGINE to support IBM 4758 cards.
12020 *Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe*
12022 * Add and OPENSSL_LOAD_CONF define which will cause
12023 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
12024 This allows older applications to transparently support certain
12025 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
12026 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
12027 load the config file and OPENSSL_add_all_algorithms_conf() which will
12028 always load it have also been added.
12032 * Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
12033 Adjust NIDs and EVP layer.
12035 *Stephen Sprunk <stephen@sprunk.org> and Richard Levitte*
12037 * Config modules support in openssl utility.
12039 Most commands now load modules from the config file,
12040 though in a few (such as version) this isn't done
12041 because it couldn't be used for anything.
12043 In the case of ca and req the config file used is
12044 the same as the utility itself: that is the -config
12045 command line option can be used to specify an
12050 * Move default behaviour from OPENSSL_config(). If appname is NULL
12051 use "openssl_conf" if filename is NULL use default openssl config file.
12055 * Add an argument to OPENSSL_config() to allow the use of an alternative
12056 config section name. Add a new flag to tolerate a missing config file
12057 and move code to CONF_modules_load_file().
12061 * Support for crypto accelerator cards from Accelerated Encryption
12062 Processing, www.aep.ie. (Use engine 'aep')
12063 The support was copied from 0.9.6c [engine] and adapted/corrected
12064 to work with the new engine framework.
12066 *AEP Inc. and Richard Levitte*
12068 * Support for SureWare crypto accelerator cards from Baltimore
12069 Technologies. (Use engine 'sureware')
12070 The support was copied from 0.9.6c [engine] and adapted
12071 to work with the new engine framework.
12075 * Have the CHIL engine fork-safe (as defined by nCipher) and actually
12076 make the newer ENGINE framework commands for the CHIL engine work.
12078 *Toomas Kiisk <vix@cyber.ee> and Richard Levitte*
12080 * Make it possible to produce shared libraries on ReliantUNIX.
12082 *Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte*
12084 * Add the configuration target debug-linux-ppro.
12085 Make 'openssl rsa' use the general key loading routines
12086 implemented in `apps.c`, and make those routines able to
12087 handle the key format FORMAT_NETSCAPE and the variant
12090 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
12092 * Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
12094 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
12096 * Add -keyform to rsautl, and document -engine.
12098 *Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>*
12100 * Change BIO_new_file (crypto/bio/bss_file.c) to use new
12101 BIO_R_NO_SUCH_FILE error code rather than the generic
12102 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
12106 * Add new functions
12107 ERR_peek_last_error
12108 ERR_peek_last_error_line
12109 ERR_peek_last_error_line_data.
12110 These are similar to
12112 ERR_peek_error_line
12113 ERR_peek_error_line_data,
12114 but report on the latest error recorded rather than the first one
12115 still in the error queue.
12117 *Ben Laurie, Bodo Moeller*
12119 * default_algorithms option in ENGINE config module. This allows things
12121 default_algorithms = ALL
12122 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
12126 * Preliminary ENGINE config module.
12130 * New experimental application configuration code.
12134 * Change the AES code to follow the same name structure as all other
12135 symmetric ciphers, and behave the same way. Move everything to
12136 the directory crypto/aes, thereby obsoleting crypto/rijndael.
12138 *Stephen Sprunk <stephen@sprunk.org> and Richard Levitte*
12140 * SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
12142 *Ben Laurie and Theo de Raadt*
12144 * Add option to output public keys in req command.
12146 *Massimiliano Pala madwolf@openca.org*
12148 * Use wNAFs in EC_POINTs_mul() for improved efficiency
12149 (up to about 10% better than before for P-192 and P-224).
12153 * New functions/macros
12155 SSL_CTX_set_msg_callback(ctx, cb)
12156 SSL_CTX_set_msg_callback_arg(ctx, arg)
12157 SSL_set_msg_callback(ssl, cb)
12158 SSL_set_msg_callback_arg(ssl, arg)
12160 to request calling a callback function
12162 void cb(int write_p, int version, int content_type,
12163 const void *buf, size_t len, SSL *ssl, void *arg)
12165 whenever a protocol message has been completely received
12166 (write_p == 0) or sent (write_p == 1). Here 'version' is the
12167 protocol version according to which the SSL library interprets
12168 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
12169 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
12170 the content type as defined in the SSL 3.0/TLS 1.0 protocol
12171 specification (change_cipher_spec(20), alert(21), handshake(22)).
12172 'buf' and 'len' point to the actual message, 'ssl' to the
12173 SSL object, and 'arg' is the application-defined value set by
12174 SSL[_CTX]_set_msg_callback_arg().
12176 'openssl s_client' and 'openssl s_server' have new '-msg' options
12177 to enable a callback that displays all protocol messages.
12181 * Change the shared library support so shared libraries are built as
12182 soon as the corresponding static library is finished, and thereby get
12183 openssl and the test programs linked against the shared library.
12184 This still only happens when the keyword "shard" has been given to
12185 the configuration scripts.
12187 NOTE: shared library support is still an experimental thing, and
12188 backward binary compatibility is still not guaranteed.
12190 *"Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte*
12192 * Add support for Subject Information Access extension.
12194 *Peter Sylvester <Peter.Sylvester@EdelWeb.fr>*
12196 * Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
12197 additional bytes when new memory had to be allocated, not just
12198 when reusing an existing buffer.
12202 * New command line and configuration option 'utf8' for the req command.
12203 This allows field values to be specified as UTF8 strings.
12207 * Add -multi and -mr options to "openssl speed" - giving multiple parallel
12208 runs for the former and machine-readable output for the latter.
12212 * Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
12213 of the e-mail address in the DN (i.e., it will go into a certificate
12214 extension only). The new configuration file option 'email_in_dn = no'
12215 has the same effect.
12217 *Massimiliano Pala madwolf@openca.org*
12219 * Change all functions with names starting with `des_` to be starting
12220 with `DES_` instead. Add wrappers that are compatible with libdes,
12221 but are named `_ossl_old_des_*`. Finally, add macros that map the
12222 `des_*` symbols to the corresponding `_ossl_old_des_*` if libdes
12223 compatibility is desired. If OpenSSL 0.9.6c compatibility is
12224 desired, the `des_*` symbols will be mapped to `DES_*`, with one
12227 Since we provide two compatibility mappings, the user needs to
12228 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
12229 compatibility is desired. The default (i.e., when that macro
12230 isn't defined) is OpenSSL 0.9.6c compatibility.
12232 There are also macros that enable and disable the support of old
12233 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
12234 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
12235 are defined, the default will apply: to support the old des routines.
12237 In either case, one must include openssl/des.h to get the correct
12238 definitions. Do not try to just include openssl/des_old.h, that
12241 NOTE: This is a major break of an old API into a new one. Software
12242 authors are encouraged to switch to the `DES_` style functions. Some
12243 time in the future, des_old.h and the libdes compatibility functions
12244 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
12245 default), and then completely removed.
12249 * Test for certificates which contain unsupported critical extensions.
12250 If such a certificate is found during a verify operation it is
12251 rejected by default: this behaviour can be overridden by either
12252 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
12253 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
12254 X509_supported_extension() has also been added which returns 1 if a
12255 particular extension is supported.
12259 * Modify the behaviour of EVP cipher functions in similar way to digests
12260 to retain compatibility with existing code.
12264 * Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
12265 compatibility with existing code. In particular the 'ctx' parameter does
12266 not have to be to be initialized before the call to EVP_DigestInit() and
12267 it is tidied up after a call to EVP_DigestFinal(). New function
12268 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
12269 EVP_MD_CTX_copy() changed to not require the destination to be
12270 initialized valid and new function EVP_MD_CTX_copy_ex() added which
12271 requires the destination to be valid.
12273 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
12274 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
12278 * Change ssl3_get_message (ssl/s3_both.c) and the functions using it
12279 so that complete 'Handshake' protocol structures are kept in memory
12280 instead of overwriting 'msg_type' and 'length' with 'body' data.
12284 * Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
12286 *Massimo Santin via Richard Levitte*
12288 * Major restructuring to the underlying ENGINE code. This includes
12289 reduction of linker bloat, separation of pure "ENGINE" manipulation
12290 (initialisation, etc) from functionality dealing with implementations
12291 of specific crypto interfaces. This change also introduces integrated
12292 support for symmetric ciphers and digest implementations - so ENGINEs
12293 can now accelerate these by providing EVP_CIPHER and EVP_MD
12294 implementations of their own. This is detailed in
12295 [crypto/engine/README.md](crypto/engine/README.md)
12296 as it couldn't be adequately described here. However, there are a few
12297 API changes worth noting - some RSA, DSA, DH, and RAND functions that
12298 were changed in the original introduction of ENGINE code have now
12299 reverted back - the hooking from this code to ENGINE is now a good
12300 deal more passive and at run-time, operations deal directly with
12301 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
12302 dereferencing through an ENGINE pointer any more. Also, the ENGINE
12303 functions dealing with `BN_MOD_EXP[_CRT]` handlers have been removed -
12304 they were not being used by the framework as there is no concept of a
12305 BIGNUM_METHOD and they could not be generalised to the new
12306 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
12307 ENGINE_cpy() has been removed as it cannot be consistently defined in
12312 * Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
12316 * Change mkdef.pl to sort symbols that get the same entry number,
12317 and make sure the automatically generated functions `ERR_load_*`
12318 become part of libeay.num as well.
12322 * New function SSL_renegotiate_pending(). This returns true once
12323 renegotiation has been requested (either SSL_renegotiate() call
12324 or HelloRequest/ClientHello received from the peer) and becomes
12325 false once a handshake has been completed.
12326 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
12327 sends a HelloRequest, but does not ensure that a handshake takes
12328 place. SSL_renegotiate_pending() is useful for checking if the
12329 client has followed the request.)
12333 * New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
12334 By default, clients may request session resumption even during
12335 renegotiation (if session ID contexts permit); with this option,
12336 session resumption is possible only in the first handshake.
12338 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes
12339 more bits available for options that should not be part of
12340 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION).
12344 * Add some demos for certificate and certificate request creation.
12348 * Make maximum certificate chain size accepted from the peer application
12349 settable (`SSL*_get/set_max_cert_list()`), as proposed by
12350 "Douglas E. Engert" <deengert@anl.gov>.
12354 * Add support for shared libraries for Unixware-7
12355 (Boyd Lynn Gerber <gerberb@zenez.com>).
12359 * Add a "destroy" handler to ENGINEs that allows structural cleanup to
12360 be done prior to destruction. Use this to unload error strings from
12361 ENGINEs that load their own error strings. NB: This adds two new API
12362 functions to "get" and "set" this destroy handler in an ENGINE.
12366 * Alter all existing ENGINE implementations (except "openssl" and
12367 "openbsd") to dynamically instantiate their own error strings. This
12368 makes them more flexible to be built both as statically-linked ENGINEs
12369 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
12370 Also, add stub code to each that makes building them as self-contained
12371 shared-libraries easier (see [README-Engine.md](README-Engine.md)).
12375 * Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
12376 implementations into applications that are completely implemented in
12377 self-contained shared-libraries. The "dynamic" ENGINE exposes control
12378 commands that can be used to configure what shared-library to load and
12379 to control aspects of the way it is handled. Also, made an update to
12380 the [README-Engine.md](README-Engine.md) file
12381 that brings its information up-to-date and
12382 provides some information and instructions on the "dynamic" ENGINE
12383 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
12387 * Make it possible to unload ranges of ERR strings with a new
12388 "ERR_unload_strings" function.
12392 * Add a copy() function to EVP_MD.
12396 * Make EVP_MD routines take a context pointer instead of just the
12397 md_data void pointer.
12401 * Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
12402 that the digest can only process a single chunk of data
12403 (typically because it is provided by a piece of
12404 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
12405 is only going to provide a single chunk of data, and hence the
12406 framework needn't accumulate the data for oneshot drivers.
12410 * As with "ERR", make it possible to replace the underlying "ex_data"
12411 functions. This change also alters the storage and management of global
12412 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
12413 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
12414 index counters. The API functions that use this state have been changed
12415 to take a "class_index" rather than pointers to the class's local STACK
12416 and counter, and there is now an API function to dynamically create new
12417 classes. This centralisation allows us to (a) plug a lot of the
12418 thread-safety problems that existed, and (b) makes it possible to clean
12419 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
12420 such data would previously have always leaked in application code and
12421 workarounds were in place to make the memory debugging turn a blind eye
12422 to it. Application code that doesn't use this new function will still
12423 leak as before, but their memory debugging output will announce it now
12424 rather than letting it slide.
12426 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
12427 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
12428 has a return value to indicate success or failure.
12432 * Make it possible to replace the underlying "ERR" functions such that the
12433 global state (2 LHASH tables and 2 locks) is only used by the "default"
12434 implementation. This change also adds two functions to "get" and "set"
12435 the implementation prior to it being automatically set the first time
12436 any other ERR function takes place. Ie. an application can call "get",
12437 pass the return value to a module it has just loaded, and that module
12438 can call its own "set" function using that value. This means the
12439 module's "ERR" operations will use (and modify) the error state in the
12440 application and not in its own statically linked copy of OpenSSL code.
12444 * Give DH, DSA, and RSA types their own `*_up_ref()` function to increment
12445 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
12446 the operation, and provides a more encapsulated way for external code
12447 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
12448 to use these functions rather than manually incrementing the counts.
12450 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
12454 * Add EVP test program.
12458 * Add symmetric cipher support to ENGINE. Expect the API to change!
12462 * New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
12463 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
12464 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
12465 These allow a CRL to be built without having to access X509_CRL fields
12466 directly. Modify 'ca' application to use new functions.
12470 * Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
12471 bug workarounds. Rollback attack detection is a security feature.
12472 The problem will only arise on OpenSSL servers when TLSv1 is not
12473 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
12474 Software authors not wanting to support TLSv1 will have special reasons
12475 for their choice and can explicitly enable this option.
12477 *Bodo Moeller, Lutz Jaenicke*
12479 * Rationalise EVP so it can be extended: don't include a union of
12480 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX
12481 (similar to those existing for EVP_CIPHER_CTX).
12486 EVP_MD_CTX_init(&md); /* new function call */
12487 EVP_DigestInit(&md, EVP_sha1());
12488 EVP_DigestUpdate(&md, in, len);
12489 EVP_DigestFinal(&md, out, NULL);
12490 EVP_MD_CTX_cleanup(&md); /* new function call */
12494 * Make DES key schedule conform to the usual scheme, as well as
12495 correcting its structure. This means that calls to DES functions
12496 now have to pass a pointer to a des_key_schedule instead of a
12497 plain des_key_schedule (which was actually always a pointer
12500 des_key_schedule ks;
12502 des_set_key_checked(..., &ks);
12503 des_ncbc_encrypt(..., &ks, ...);
12505 (Note that a later change renames 'des_...' into 'DES_...'.)
12509 * Initial reduction of linker bloat: the use of some functions, such as
12510 PEM causes large amounts of unused functions to be linked in due to
12511 poor organisation. For example pem_all.c contains every PEM function
12512 which has a knock on effect of linking in large amounts of (unused)
12513 ASN1 code. Grouping together similar functions and splitting unrelated
12514 functions prevents this.
12518 * Cleanup of EVP macros.
12522 * Change historical references to `{NID,SN,LN}_des_ede` and ede3 to add the
12523 correct `_ecb suffix`.
12527 * Add initial OCSP responder support to ocsp application. The
12528 revocation information is handled using the text based index
12529 use by the ca application. The responder can either handle
12530 requests generated internally, supplied in files (for example
12531 via a CGI script) or using an internal minimal server.
12535 * Add configuration choices to get zlib compression for TLS.
12539 * Changes to Kerberos SSL for RFC 2712 compliance:
12540 1. Implemented real KerberosWrapper, instead of just using
12541 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
12542 2. Implemented optional authenticator field of KerberosWrapper.
12544 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
12545 and authenticator structs; see crypto/krb5/.
12547 Generalized Kerberos calls to support multiple Kerberos libraries.
12548 *Vern Staats <staatsvr@asc.hpc.mil>, Jeffrey Altman <jaltman@columbia.edu>
12549 via Richard Levitte*
12551 * Cause 'openssl speed' to use fully hard-coded DSA keys as it
12552 already does with RSA. testdsa.h now has 'priv_key/pub_key'
12553 values for each of the key sizes rather than having just
12554 parameters (and 'speed' generating keys each time).
12558 * Speed up EVP routines.
12561 pe 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
12562 s-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
12563 s-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
12564 s-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
12566 s-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
12567 s-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
12568 s-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
12571 s-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
12573 s-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
12577 * Added the OS2-EMX target.
12579 *"Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte*
12581 * Rewrite commands to use `NCONF` routines instead of the old `CONF`.
12582 New functions to support `NCONF` routines in extension code.
12583 New function `CONF_set_nconf()`
12584 to allow functions which take an `NCONF` to also handle the old `LHASH`
12585 structure: this means that the old `CONF` compatible routines can be
12586 retained (in particular w.rt. extensions) without having to duplicate the
12587 code. New function `X509V3_add_ext_nconf_sk()` to add extensions to a stack.
12591 * Enhance the general user interface with mechanisms for inner control
12592 and with possibilities to have yes/no kind of prompts.
12596 * Change all calls to low-level digest routines in the library and
12597 applications to use EVP. Add missing calls to HMAC_cleanup() and
12598 don't assume HMAC_CTX can be copied using memcpy().
12600 *Verdon Walker <VWalker@novell.com>, Steve Henson*
12602 * Add the possibility to control engines through control names but with
12603 arbitrary arguments instead of just a string.
12604 Change the key loaders to take a UI_METHOD instead of a callback
12605 function pointer. NOTE: this breaks binary compatibility with earlier
12606 versions of OpenSSL [engine].
12607 Adapt the nCipher code for these new conditions and add a card insertion
12612 * Enhance the general user interface with mechanisms to better support
12613 dialog box interfaces, application-defined prompts, the possibility
12614 to use defaults (for example default passwords from somewhere else)
12615 and interrupts/cancellations.
12619 * Tidy up PKCS#12 attribute handling. Add support for the CSP name
12620 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
12624 * Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
12625 tidy up some unnecessarily weird code in 'sk_new()').
12627 *Geoff, reported by Diego Tartara <dtartara@novamens.com>*
12629 * Change the key loading routines for ENGINEs to use the same kind
12630 callback (pem_password_cb) as all other routines that need this
12635 * Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
12636 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
12637 than this minimum value is recommended.
12641 * New random seeder for OpenVMS, using the system process statistics
12642 that are easily reachable.
12646 * Windows apparently can't transparently handle global
12647 variables defined in DLLs. Initialisations such as:
12649 const ASN1_ITEM *it = &ASN1_INTEGER_it;
12651 won't compile. This is used by the any applications that need to
12652 declare their own ASN1 modules. This was fixed by adding the option
12653 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
12654 needed for static libraries under Win32.
12658 * New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
12659 setting of purpose and trust fields. New X509_STORE trust and
12660 purpose functions and tidy up setting in other SSL functions.
12664 * Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
12665 structure. These are inherited by X509_STORE_CTX when it is
12666 initialised. This allows various defaults to be set in the
12667 X509_STORE structure (such as flags for CRL checking and custom
12668 purpose or trust settings) for functions which only use X509_STORE_CTX
12669 internally such as S/MIME.
12671 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
12672 trust settings if they are not set in X509_STORE. This allows X509_STORE
12673 purposes and trust (in S/MIME for example) to override any set by default.
12675 Add command line options for CRL checking to smime, s_client and s_server
12680 * Initial CRL based revocation checking. If the CRL checking flag(s)
12681 are set then the CRL is looked up in the X509_STORE structure and
12682 its validity and signature checked, then if the certificate is found
12683 in the CRL the verify fails with a revoked error.
12685 Various new CRL related callbacks added to X509_STORE_CTX structure.
12687 Command line options added to 'verify' application to support this.
12689 This needs some additional work, such as being able to handle multiple
12690 CRLs with different times, extension based lookup (rather than just
12691 by subject name) and ultimately more complete V2 CRL extension
12696 * Add a general user interface API (crypto/ui/). This is designed
12697 to replace things like des_read_password and friends (backward
12698 compatibility functions using this new API are provided).
12699 The purpose is to remove prompting functions from the DES code
12700 section as well as provide for prompting through dialog boxes in
12701 a window system and the like.
12705 * Add "ex_data" support to ENGINE so implementations can add state at a
12706 per-structure level rather than having to store it globally.
12710 * Make it possible for ENGINE structures to be copied when retrieved by
12711 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
12712 This causes the "original" ENGINE structure to act like a template,
12713 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
12714 operational state can be localised to each ENGINE structure, despite the
12715 fact they all share the same "methods". New ENGINE structures returned in
12716 this case have no functional references and the return value is the single
12717 structural reference. This matches the single structural reference returned
12718 by ENGINE_by_id() normally, when it is incremented on the pre-existing
12723 * Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
12724 needs to match any other type at all we need to manually clear the
12729 * Changes to the "openssl engine" utility to include;
12730 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
12731 about an ENGINE's available control commands.
12732 - executing control commands from command line arguments using the
12733 '-pre' and '-post' switches. '-post' is only used if '-t' is
12734 specified and the ENGINE is successfully initialised. The syntax for
12735 the individual commands are colon-separated, for example;
12736 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
12740 * New dynamic control command support for ENGINEs. ENGINEs can now
12741 declare their own commands (numbers), names (strings), descriptions,
12742 and input types for run-time discovery by calling applications. A
12743 subset of these commands are implicitly classed as "executable"
12744 depending on their input type, and only these can be invoked through
12745 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
12746 can be based on user input, config files, etc). The distinction is
12747 that "executable" commands cannot return anything other than a boolean
12748 result and can only support numeric or string input, whereas some
12749 discoverable commands may only be for direct use through
12750 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
12751 pointers, or other custom uses. The "executable" commands are to
12752 support parameterisations of ENGINE behaviour that can be
12753 unambiguously defined by ENGINEs and used consistently across any
12754 OpenSSL-based application. Commands have been added to all the
12755 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
12756 control over shared-library paths without source code alterations.
12760 * Changed all ENGINE implementations to dynamically allocate their
12761 ENGINEs rather than declaring them statically. Apart from this being
12762 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
12763 this also allows the implementations to compile without using the
12764 internal engine_int.h header.
12768 * Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
12769 'const' value. Any code that should be able to modify a RAND_METHOD
12770 should already have non-const pointers to it (ie. they should only
12771 modify their own ones).
12775 * Made a variety of little tweaks to the ENGINE code.
12776 - "atalla" and "ubsec" string definitions were moved from header files
12777 to C code. "nuron" string definitions were placed in variables
12778 rather than hard-coded - allowing parameterisation of these values
12779 later on via ctrl() commands.
12780 - Removed unused "#if 0"'d code.
12781 - Fixed engine list iteration code so it uses ENGINE_free() to release
12782 structural references.
12783 - Constified the RAND_METHOD element of ENGINE structures.
12784 - Constified various get/set functions as appropriate and added
12785 missing functions (including a catch-all ENGINE_cpy that duplicates
12786 all ENGINE values onto a new ENGINE except reference counts/state).
12787 - Removed NULL parameter checks in get/set functions. Setting a method
12788 or function to NULL is a way of cancelling out a previously set
12789 value. Passing a NULL ENGINE parameter is just plain stupid anyway
12790 and doesn't justify the extra error symbols and code.
12791 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
12792 flags from engine_int.h to engine.h.
12793 - Changed prototypes for ENGINE handler functions (init(), finish(),
12794 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
12798 * Implement binary inversion algorithm for BN_mod_inverse in addition
12799 to the algorithm using long division. The binary algorithm can be
12800 used only if the modulus is odd. On 32-bit systems, it is faster
12801 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
12802 roughly 5-15% for 256-bit moduli), so we use it only for moduli
12803 up to 450 bits. In 64-bit environments, the binary algorithm
12804 appears to be advantageous for much longer moduli; here we use it
12805 for moduli up to 2048 bits.
12809 * Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
12810 could not support the combine flag in choice fields.
12814 * Add a 'copy_extensions' option to the 'ca' utility. This copies
12815 extensions from a certificate request to the certificate.
12819 * Allow multiple 'certopt' and 'nameopt' options to be separated
12820 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
12821 file: this allows the display of the certificate about to be
12822 signed to be customised, to allow certain fields to be included
12823 or excluded and extension details. The old system didn't display
12824 multicharacter strings properly, omitted fields not in the policy
12825 and couldn't display additional details such as extensions.
12829 * Function EC_POINTs_mul for multiple scalar multiplication
12830 of an arbitrary number of elliptic curve points
12831 \sum scalars[i]*points[i],
12832 optionally including the generator defined for the EC_GROUP:
12833 scalar*generator + \sum scalars[i]*points[i].
12835 EC_POINT_mul is a simple wrapper function for the typical case
12836 that the point list has just one item (besides the optional
12841 * First EC_METHODs for curves over GF(p):
12843 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
12844 operations and provides various method functions that can also
12845 operate with faster implementations of modular arithmetic.
12847 EC_GFp_mont_method() reuses most functions that are part of
12848 EC_GFp_simple_method, but uses Montgomery arithmetic.
12850 *Bodo Moeller; point addition and point doubling
12851 implementation directly derived from source code provided by
12852 Lenka Fibikova <fibikova@exp-math.uni-essen.de>*
12854 * Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
12855 crypto/ec/ec_lib.c):
12857 Curves are EC_GROUP objects (with an optional group generator)
12858 based on EC_METHODs that are built into the library.
12860 Points are EC_POINT objects based on EC_GROUP objects.
12862 Most of the framework would be able to handle curves over arbitrary
12863 finite fields, but as there are no obvious types for fields other
12864 than GF(p), some functions are limited to that for now.
12868 * Add the -HTTP option to s_server. It is similar to -WWW, but requires
12869 that the file contains a complete HTTP response.
12873 * Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
12874 change the def and num file printf format specifier from "%-40sXXX"
12875 to "%-39s XXX". The latter will always guarantee a space after the
12876 field while the former will cause them to run together if the field
12877 is 40 of more characters long.
12881 * Constify the cipher and digest 'method' functions and structures
12882 and modify related functions to take constant EVP_MD and EVP_CIPHER
12887 * Hide BN_CTX structure details in bn_lcl.h instead of publishing them
12888 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
12892 * Modify `EVP_Digest*()` routines so they now return values. Although the
12893 internal software routines can never fail additional hardware versions
12898 * Clean up crypto/err/err.h and change some error codes to avoid conflicts:
12900 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
12901 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
12904 ERR_R_NESTED_ASN1_ERROR
12906 ERR_R_MISSING_ASN1_EOS
12907 were 4 .. 9, conflicting with
12908 ERR_LIB_RSA (= ERR_R_RSA_LIB)
12910 ERR_LIB_PEM (= ERR_R_PEM_LIB).
12911 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
12913 Add new error code 'ERR_R_INTERNAL_ERROR'.
12917 * Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
12922 * New option '-subj arg' for 'openssl req' and 'openssl ca'. This
12923 sets the subject name for a new request or supersedes the
12924 subject name in a given request. Formats that can be parsed are
12925 'CN=Some Name, OU=myOU, C=IT'
12927 'CN=Some Name/OU=myOU/C=IT'.
12929 Add options '-batch' and '-verbose' to 'openssl req'.
12931 *Massimiliano Pala <madwolf@hackmasters.net>*
12933 * Introduce the possibility to access global variables through
12934 functions on platform were that's the best way to handle exporting
12935 global variables in shared libraries. To enable this functionality,
12936 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
12937 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
12938 is normally done by Configure or something similar).
12940 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
12941 in the source file (foo.c) like this:
12943 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
12944 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
12946 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
12947 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
12949 OPENSSL_DECLARE_GLOBAL(int,foo);
12950 #define foo OPENSSL_GLOBAL_REF(foo)
12951 OPENSSL_DECLARE_GLOBAL(double,bar);
12952 #define bar OPENSSL_GLOBAL_REF(bar)
12954 The #defines are very important, and therefore so is including the
12955 header file everywhere where the defined globals are used.
12957 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
12958 of ASN.1 items, but that structure is a bit different.
12960 The largest change is in util/mkdef.pl which has been enhanced with
12961 better and easier to understand logic to choose which symbols should
12962 go into the Windows .def files as well as a number of fixes and code
12963 cleanup (among others, algorithm keywords are now sorted
12964 lexicographically to avoid constant rewrites).
12968 * In BN_div() keep a copy of the sign of 'num' before writing the
12969 result to 'rm' because if rm==num the value will be overwritten
12970 and produce the wrong result if 'num' is negative: this caused
12971 problems with BN_mod() and BN_nnmod().
12975 * Function OCSP_request_verify(). This checks the signature on an
12976 OCSP request and verifies the signer certificate. The signer
12977 certificate is just checked for a generic purpose and OCSP request
12982 * Add OCSP_check_validity() function to check the validity of OCSP
12983 responses. OCSP responses are prepared in real time and may only
12984 be a few seconds old. Simply checking that the current time lies
12985 between thisUpdate and nextUpdate max reject otherwise valid responses
12986 caused by either OCSP responder or client clock inaccuracy. Instead
12987 we allow thisUpdate and nextUpdate to fall within a certain period of
12988 the current time. The age of the response can also optionally be
12989 checked. Two new options -validity_period and -status_age added to
12994 * If signature or public key algorithm is unrecognized print out its
12995 OID rather that just UNKNOWN.
12999 * Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
13000 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
13001 ID to be generated from the issuer certificate alone which can then be
13002 passed to OCSP_id_issuer_cmp().
13006 * New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
13007 ASN1 modules to export functions returning ASN1_ITEM pointers
13008 instead of the ASN1_ITEM structures themselves. This adds several
13009 new macros which allow the underlying ASN1 function/structure to
13010 be accessed transparently. As a result code should not use ASN1_ITEM
13011 references directly (such as &X509_it) but instead use the relevant
13012 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
13013 use of the new ASN1 code on platforms where exporting structures
13014 is problematical (for example in shared libraries) but exporting
13015 functions returning pointers to structures is not.
13019 * Add support for overriding the generation of SSL/TLS session IDs.
13020 These callbacks can be registered either in an SSL_CTX or per SSL.
13021 The purpose of this is to allow applications to control, if they wish,
13022 the arbitrary values chosen for use as session IDs, particularly as it
13023 can be useful for session caching in multiple-server environments. A
13024 command-line switch for testing this (and any client code that wishes
13025 to use such a feature) has been added to "s_server".
13027 *Geoff Thorpe, Lutz Jaenicke*
13029 * Modify mkdef.pl to recognise and parse preprocessor conditionals
13030 of the form `#if defined(...) || defined(...) || ...` and
13031 `#if !defined(...) && !defined(...) && ...`. This also avoids
13032 the growing number of special cases it was previously handling.
13036 * Make all configuration macros available for application by making
13037 sure they are available in opensslconf.h, by giving them names starting
13038 with `OPENSSL_` to avoid conflicts with other packages and by making
13039 sure e_os2.h will cover all platform-specific cases together with
13041 Additionally, it is now possible to define configuration/platform-
13042 specific names (called "system identities"). In the C code, these
13043 are prefixed with `OPENSSL_SYSNAME_`. e_os2.h will create another
13044 macro with the name beginning with `OPENSSL_SYS_`, which is determined
13045 from `OPENSSL_SYSNAME_*` or compiler-specific macros depending on
13050 * New option -set_serial to 'req' and 'x509' this allows the serial
13051 number to use to be specified on the command line. Previously self
13052 signed certificates were hard coded with serial number 0 and the
13053 CA options of 'x509' had to use a serial number in a file which was
13058 * New options to 'ca' utility to support V2 CRL entry extensions.
13059 Currently CRL reason, invalidity date and hold instruction are
13060 supported. Add new CRL extensions to V3 code and some new objects.
13064 * New function EVP_CIPHER_CTX_set_padding() this is used to
13065 disable standard block padding (aka PKCS#5 padding) in the EVP
13066 API, which was previously mandatory. This means that the data is
13067 not padded in any way and so the total length much be a multiple
13068 of the block size, otherwise an error occurs.
13072 * Initial (incomplete) OCSP SSL support.
13076 * New function OCSP_parse_url(). This splits up a URL into its host,
13077 port and path components: primarily to parse OCSP URLs. New -url
13078 option to ocsp utility.
13082 * New nonce behavior. The return value of OCSP_check_nonce() now
13083 reflects the various checks performed. Applications can decide
13084 whether to tolerate certain situations such as an absent nonce
13085 in a response when one was present in a request: the ocsp application
13086 just prints out a warning. New function OCSP_add1_basic_nonce()
13087 this is to allow responders to include a nonce in a response even if
13088 the request is nonce-less.
13092 * Disable stdin buffering in `load_cert()` (`apps/apps.c`) so that no certs are
13093 skipped when using openssl x509 multiple times on a single input file,
13094 e.g. `(openssl x509 -out cert1; openssl x509 -out cert2) <certs`.
13098 * Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
13099 set string type: to handle setting ASN1_TIME structures. Fix ca
13100 utility to correctly initialize revocation date of CRLs.
13104 * New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
13105 the clients preferred ciphersuites and rather use its own preferences.
13106 Should help to work around M$ SGC (Server Gated Cryptography) bug in
13107 Internet Explorer by ensuring unchanged hash method during stepup.
13108 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
13112 * Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
13113 to aes and add a new 'exist' option to print out symbols that don't
13118 * Additional options to ocsp utility to allow flags to be set and
13119 additional certificates supplied.
13123 * Add the option -VAfile to 'openssl ocsp', so the user can give the
13124 OCSP client a number of certificate to only verify the response
13129 * Update Rijndael code to version 3.0 and change EVP AES ciphers to
13130 handle the new API. Currently only ECB, CBC modes supported. Add new
13133 Add TLS AES ciphersuites as described in RFC3268, "Advanced
13134 Encryption Standard (AES) Ciphersuites for Transport Layer
13135 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were
13136 not enabled by default and were not part of the "ALL" ciphersuite
13137 alias because they were not yet official; they could be
13138 explicitly requested by specifying the "AESdraft" ciphersuite
13139 group alias. In the final release of OpenSSL 0.9.7, the group
13140 alias is called "AES" and is part of "ALL".)
13142 *Ben Laurie, Steve Henson, Bodo Moeller*
13144 * New function OCSP_copy_nonce() to copy nonce value (if present) from
13145 request to response.
13149 * Functions for OCSP responders. OCSP_request_onereq_count(),
13150 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
13151 extract information from a certificate request. OCSP_response_create()
13152 creates a response and optionally adds a basic response structure.
13153 OCSP_basic_add1_status() adds a complete single response to a basic
13154 response and returns the OCSP_SINGLERESP structure just added (to allow
13155 extensions to be included for example). OCSP_basic_add1_cert() adds a
13156 certificate to a basic response and OCSP_basic_sign() signs a basic
13157 response with various flags. New helper functions ASN1_TIME_check()
13158 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
13159 (converts ASN1_TIME to GeneralizedTime).
13163 * Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
13164 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
13165 structure from a certificate. X509_pubkey_digest() digests the public_key
13166 contents: this is used in various key identifiers.
13170 * Make sk_sort() tolerate a NULL argument.
13172 *Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>*
13174 * New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
13175 passed by the function are trusted implicitly. If any of them signed the
13176 response then it is assumed to be valid and is not verified.
13180 * In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
13181 to data. This was previously part of the PKCS7 ASN1 code. This
13182 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
13183 *Steve Henson, reported by Kenneth R. Robinette
13184 <support@securenetterm.com>*
13186 * Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
13187 routines: without these tracing memory leaks is very painful.
13188 Fix leaks in PKCS12 and PKCS7 routines.
13192 * Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
13193 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
13194 effectively meant GeneralizedTime would never be used. Now it
13195 is initialised to -1 but X509_time_adj() now has to check the value
13196 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
13197 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
13198 *Steve Henson, reported by Kenneth R. Robinette
13199 <support@securenetterm.com>*
13201 * Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
13202 result in a zero length in the ASN1_INTEGER structure which was
13203 not consistent with the structure when d2i_ASN1_INTEGER() was used
13204 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
13205 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
13206 where it did not print out a minus for negative ASN1_INTEGER.
13210 * Add summary printout to ocsp utility. The various functions which
13211 convert status values to strings have been renamed to:
13212 OCSP_response_status_str(), OCSP_cert_status_str() and
13213 OCSP_crl_reason_str() and are no longer static. New options
13214 to verify nonce values and to disable verification. OCSP response
13215 printout format cleaned up.
13219 * Add additional OCSP certificate checks. These are those specified
13220 in RFC2560. This consists of two separate checks: the CA of the
13221 certificate being checked must either be the OCSP signer certificate
13222 or the issuer of the OCSP signer certificate. In the latter case the
13223 OCSP signer certificate must contain the OCSP signing extended key
13224 usage. This check is performed by attempting to match the OCSP
13225 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
13226 in the OCSP_CERTID structures of the response.
13230 * Initial OCSP certificate verification added to OCSP_basic_verify()
13231 and related routines. This uses the standard OpenSSL certificate
13232 verify routines to perform initial checks (just CA validity) and
13233 to obtain the certificate chain. Then additional checks will be
13234 performed on the chain. Currently the root CA is checked to see
13235 if it is explicitly trusted for OCSP signing. This is used to set
13236 a root CA as a global signing root: that is any certificate that
13237 chains to that CA is an acceptable OCSP signing certificate.
13241 * New '-extfile ...' option to 'openssl ca' for reading X.509v3
13242 extensions from a separate configuration file.
13243 As when reading extensions from the main configuration file,
13244 the '-extensions ...' option may be used for specifying the
13247 *Massimiliano Pala <madwolf@comune.modena.it>*
13249 * New OCSP utility. Allows OCSP requests to be generated or
13250 read. The request can be sent to a responder and the output
13251 parsed, outputted or printed in text form. Not complete yet:
13252 still needs to check the OCSP response validity.
13256 * New subcommands for 'openssl ca':
13257 `openssl ca -status <serial>` prints the status of the cert with
13258 the given serial number (according to the index file).
13259 `openssl ca -updatedb` updates the expiry status of certificates
13262 *Massimiliano Pala <madwolf@comune.modena.it>*
13264 * New '-newreq-nodes' command option to CA.pl. This is like
13265 '-newreq', but calls 'openssl req' with the '-nodes' option
13266 so that the resulting key is not encrypted.
13268 *Damien Miller <djm@mindrot.org>*
13270 * New configuration for the GNU Hurd.
13272 *Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte*
13274 * Initial code to implement OCSP basic response verify. This
13275 is currently incomplete. Currently just finds the signer's
13276 certificate and verifies the signature on the response.
13280 * New SSLeay_version code SSLEAY_DIR to determine the compiled-in
13281 value of OPENSSLDIR. This is available via the new '-d' option
13282 to 'openssl version', and is also included in 'openssl version -a'.
13286 * Allowing defining memory allocation callbacks that will be given
13287 file name and line number information in additional arguments
13288 (a `const char*` and an int). The basic functionality remains, as
13289 well as the original possibility to just replace malloc(),
13290 realloc() and free() by functions that do not know about these
13291 additional arguments. To register and find out the current
13292 settings for extended allocation functions, the following
13293 functions are provided:
13295 CRYPTO_set_mem_ex_functions
13296 CRYPTO_set_locked_mem_ex_functions
13297 CRYPTO_get_mem_ex_functions
13298 CRYPTO_get_locked_mem_ex_functions
13300 These work the same way as CRYPTO_set_mem_functions and friends.
13301 `CRYPTO_get_[locked_]mem_functions` now writes 0 where such an
13302 extended allocation function is enabled.
13303 Similarly, `CRYPTO_get_[locked_]mem_ex_functions` writes 0 where
13304 a conventional allocation function is enabled.
13306 *Richard Levitte, Bodo Moeller*
13308 * Finish off removing the remaining LHASH function pointer casts.
13309 There should no longer be any prototype-casting required when using
13310 the LHASH abstraction, and any casts that remain are "bugs". See
13311 the callback types and macros at the head of lhash.h for details
13312 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
13316 * Add automatic query of EGD sockets in RAND_poll() for the unix variant.
13317 If /dev/[u]random devices are not available or do not return enough
13318 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
13320 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
13321 /etc/entropy will be queried once each in this sequence, querying stops
13322 when enough entropy was collected without querying more sockets.
13326 * Change the Unix RAND_poll() variant to be able to poll several
13327 random devices, as specified by DEVRANDOM, until a sufficient amount
13328 of data has been collected. We spend at most 10 ms on each file
13329 (select timeout) and read in non-blocking mode. DEVRANDOM now
13330 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
13331 (previously it was just the string "/dev/urandom"), so on typical
13332 platforms the 10 ms delay will never occur.
13333 Also separate out the Unix variant to its own file, rand_unix.c.
13334 For VMS, there's a currently-empty rand_vms.c.
13338 * Move OCSP client related routines to ocsp_cl.c. These
13339 provide utility functions which an application needing
13340 to issue a request to an OCSP responder and analyse the
13341 response will typically need: as opposed to those which an
13342 OCSP responder itself would need which will be added later.
13344 OCSP_request_sign() signs an OCSP request with an API similar
13345 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
13346 response. OCSP_response_get1_basic() extracts basic response
13347 from response. OCSP_resp_find_status(): finds and extracts status
13348 information from an OCSP_CERTID structure (which will be created
13349 when the request structure is built). These are built from lower
13350 level functions which work on OCSP_SINGLERESP structures but
13351 won't normally be used unless the application wishes to examine
13352 extensions in the OCSP response for example.
13354 Replace nonce routines with a pair of functions.
13355 OCSP_request_add1_nonce() adds a nonce value and optionally
13356 generates a random value. OCSP_check_nonce() checks the
13357 validity of the nonce in an OCSP response.
13361 * Change function OCSP_request_add() to OCSP_request_add0_id().
13362 This doesn't copy the supplied OCSP_CERTID and avoids the
13363 need to free up the newly created id. Change return type
13364 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
13365 This can then be used to add extensions to the request.
13366 Deleted OCSP_request_new(), since most of its functionality
13367 is now in OCSP_REQUEST_new() (and the case insensitive name
13368 clash) apart from the ability to set the request name which
13369 will be added elsewhere.
13373 * Update OCSP API. Remove obsolete extensions argument from
13374 various functions. Extensions are now handled using the new
13375 OCSP extension code. New simple OCSP HTTP function which
13376 can be used to send requests and parse the response.
13380 * Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
13381 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
13382 uses the special reorder version of SET OF to sort the attributes
13383 and reorder them to match the encoded order. This resolves a long
13384 standing problem: a verify on a PKCS7 structure just after signing
13385 it used to fail because the attribute order did not match the
13386 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
13387 it uses the received order. This is necessary to tolerate some broken
13388 software that does not order SET OF. This is handled by encoding
13389 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
13390 to produce the required SET OF.
13394 * Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
13395 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
13396 files to get correct declarations of the ASN.1 item variables.
13400 * Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
13401 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
13402 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
13403 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
13404 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
13405 ASN1_ITEM and no wrapper functions.
13409 * New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
13410 replace the old function pointer based I/O routines. Change most of
13411 the `*_d2i_bio()` and `*_d2i_fp()` functions to use these.
13415 * Enhance mkdef.pl to be more accepting about spacing in C preprocessor
13416 lines, recognize more "algorithms" that can be deselected, and make
13417 it complain about algorithm deselection that isn't recognised.
13421 * New ASN1 functions to handle dup, sign, verify, digest, pack and
13422 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
13423 to use new functions. Add NO_ASN1_OLD which can be set to remove
13424 some old style ASN1 functions: this can be used to determine if old
13425 code will still work when these eventually go away.
13429 * New extension functions for OCSP structures, these follow the
13430 same conventions as certificates and CRLs.
13434 * New function X509V3_add1_i2d(). This automatically encodes and
13435 adds an extension. Its behaviour can be customised with various
13436 flags to append, replace or delete. Various wrappers added for
13437 certificates and CRLs.
13441 * Fix to avoid calling the underlying ASN1 print routine when
13442 an extension cannot be parsed. Correct a typo in the
13443 OCSP_SERVICELOC extension. Tidy up print OCSP format.
13447 * Make mkdef.pl parse some of the ASN1 macros and add appropriate
13448 entries for variables.
13452 * Add functionality to `apps/openssl.c` for detecting locking
13453 problems: As the program is single-threaded, all we have
13454 to do is register a locking callback using an array for
13455 storing which locks are currently held by the program.
13459 * Use a lock around the call to CRYPTO_get_ex_new_index() in
13460 SSL_get_ex_data_X509_STORE_idx(), which is used in
13461 ssl_verify_cert_chain() and thus can be called at any time
13462 during TLS/SSL handshakes so that thread-safety is essential.
13463 Unfortunately, the ex_data design is not at all suited
13464 for multi-threaded use, so it probably should be abolished.
13468 * Added Broadcom "ubsec" ENGINE to OpenSSL.
13470 *Broadcom, tweaked and integrated by Geoff Thorpe*
13472 * Move common extension printing code to new function
13473 X509V3_print_extensions(). Reorganise OCSP print routines and
13474 implement some needed OCSP ASN1 functions. Add OCSP extensions.
13478 * New function X509_signature_print() to remove duplication in some
13483 * Add a special meaning when SET OF and SEQUENCE OF flags are both
13484 set (this was treated exactly the same as SET OF previously). This
13485 is used to reorder the STACK representing the structure to match the
13486 encoding. This will be used to get round a problem where a PKCS7
13487 structure which was signed could not be verified because the STACK
13488 order did not reflect the encoded order.
13492 * Reimplement the OCSP ASN1 module using the new code.
13496 * Update the X509V3 code to permit the use of an ASN1_ITEM structure
13497 for its ASN1 operations. The old style function pointers still exist
13498 for now but they will eventually go away.
13502 * Merge in replacement ASN1 code from the ASN1 branch. This almost
13503 completely replaces the old ASN1 functionality with a table driven
13504 encoder and decoder which interprets an ASN1_ITEM structure describing
13505 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
13506 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
13507 has also been converted to the new form.
13511 * Change BN_mod_exp_recp so that negative moduli are tolerated
13512 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
13513 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
13514 for negative moduli.
13518 * Fix BN_uadd and BN_usub: Always return non-negative results instead
13519 of not touching the result's sign bit.
13523 * BN_div bugfix: If the result is 0, the sign (res->neg) must not be
13528 * Changed the LHASH code to use prototypes for callbacks, and created
13529 macros to declare and implement thin (optionally static) functions
13530 that provide type-safety and avoid function pointer casting for the
13531 type-specific callbacks.
13535 * Added Kerberos Cipher Suites to be used with TLS, as written in
13537 *Veers Staats <staatsvr@asc.hpc.mil>,
13538 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte*
13540 * Reformat the FAQ so the different questions and answers can be divided
13541 in sections depending on the subject.
13545 * Have the zlib compression code load ZLIB.DLL dynamically under
13550 * New function BN_mod_sqrt for computing square roots modulo a prime
13551 (using the probabilistic Tonelli-Shanks algorithm unless
13552 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
13553 be handled deterministically).
13555 *Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller*
13557 * Make BN_mod_inverse faster by explicitly handling small quotients
13558 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
13559 512 bits], about 30% for larger ones [1024 or 2048 bits].)
13563 * New function BN_kronecker.
13567 * Fix BN_gcd so that it works on negative inputs; the result is
13568 positive unless both parameters are zero.
13569 Previously something reasonably close to an infinite loop was
13570 possible because numbers could be growing instead of shrinking
13571 in the implementation of Euclid's algorithm.
13575 * Fix BN_is_word() and BN_is_one() macros to take into account the
13576 sign of the number in question.
13578 Fix BN_is_word(a,w) to work correctly for w == 0.
13580 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
13581 because its test if the absolute value of 'a' equals 'w'.
13582 Note that BN_abs_is_word does *not* handle w == 0 reliably;
13583 it exists mostly for use in the implementations of BN_is_zero(),
13584 BN_is_one(), and BN_is_word().
13588 * New function BN_swap.
13592 * Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
13593 the exponentiation functions are more likely to produce reasonable
13594 results on negative inputs.
13598 * Change BN_mod_mul so that the result is always non-negative.
13599 Previously, it could be negative if one of the factors was negative;
13600 I don't think anyone really wanted that behaviour.
13604 * Move `BN_mod_...` functions into new file `crypto/bn/bn_mod.c`
13605 (except for exponentiation, which stays in `crypto/bn/bn_exp.c`,
13606 and `BN_mod_mul_reciprocal`, which stays in `crypto/bn/bn_recp.c`)
13607 and add new functions:
13616 BN_mod_lshift1_quick
13618 BN_mod_lshift_quick
13620 These functions always generate non-negative results.
13622 `BN_nnmod` otherwise is `like BN_mod` (if `BN_mod` computes a remainder `r`
13623 such that `|m| < r < 0`, `BN_nnmod` will output `rem + |m|` instead).
13625 `BN_mod_XXX_quick(r, a, [b,] m)` generates the same result as
13626 `BN_mod_XXX(r, a, [b,] m, ctx)`, but requires that `a` [and `b`]
13627 be reduced modulo `m`.
13629 *Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller*
13632 The following entry accidentally appeared in the CHANGES file
13633 distributed with OpenSSL 0.9.7. The modifications described in
13634 it do *not* apply to OpenSSL 0.9.7.
13636 * Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
13637 was actually never needed) and in BN_mul(). The removal in BN_mul()
13638 required a small change in bn_mul_part_recursive() and the addition
13639 of the functions bn_cmp_part_words(), bn_sub_part_words() and
13640 bn_add_part_words(), which do the same thing as bn_cmp_words(),
13641 bn_sub_words() and bn_add_words() except they take arrays with
13647 * In 'openssl passwd', verify passwords read from the terminal
13648 unless the '-salt' option is used (which usually means that
13649 verification would just waste user's time since the resulting
13650 hash is going to be compared with some given password hash)
13651 or the new '-noverify' option is used.
13653 This is an incompatible change, but it does not affect
13654 non-interactive use of 'openssl passwd' (passwords on the command
13655 line, '-stdin' option, '-in ...' option) and thus should not
13656 cause any problems.
13660 * Remove all references to RSAref, since there's no more need for it.
13664 * Make DSO load along a path given through an environment variable
13665 (SHLIB_PATH) with shl_load().
13669 * Constify the ENGINE code as a result of BIGNUM constification.
13670 Also constify the RSA code and most things related to it. In a
13671 few places, most notable in the depth of the ASN.1 code, ugly
13672 casts back to non-const were required (to be solved at a later
13677 * Make it so the openssl application has all engines loaded by default.
13681 * Constify the BIGNUM routines a little more.
13685 * Add the following functions:
13687 ENGINE_load_cswift()
13689 ENGINE_load_atalla()
13690 ENGINE_load_nuron()
13691 ENGINE_load_builtin_engines()
13693 That way, an application can itself choose if external engines that
13694 are built-in in OpenSSL shall ever be used or not. The benefit is
13695 that applications won't have to be linked with libdl or other dso
13696 libraries unless it's really needed.
13698 Changed 'openssl engine' to load all engines on demand.
13699 Changed the engine header files to avoid the duplication of some
13700 declarations (they differed!).
13704 * 'openssl engine' can now list capabilities.
13708 * Better error reporting in 'openssl engine'.
13712 * Never call load_dh_param(NULL) in s_server.
13716 * Add engine application. It can currently list engines by name and
13717 identity, and test if they are actually available.
13721 * Improve RPM specification file by forcing symbolic linking and making
13722 sure the installed documentation is also owned by root.root.
13724 *Damien Miller <djm@mindrot.org>*
13726 * Give the OpenSSL applications more possibilities to make use of
13727 keys (public as well as private) handled by engines.
13731 * Add OCSP code that comes from CertCo.
13735 * Add VMS support for the Rijndael code.
13739 * Added untested support for Nuron crypto accelerator.
13743 * Add support for external cryptographic devices. This code was
13744 previously distributed separately as the "engine" branch.
13746 *Geoff Thorpe, Richard Levitte*
13748 * Rework the filename-translation in the DSO code. It is now possible to
13749 have far greater control over how a "name" is turned into a filename
13750 depending on the operating environment and any oddities about the
13751 different shared library filenames on each system.
13755 * Support threads on FreeBSD-elf in Configure.
13759 * Fix for SHA1 assembly problem with MASM: it produces
13760 warnings about corrupt line number information when assembling
13761 with debugging information. This is caused by the overlapping
13764 *Bernd Matthes <mainbug@celocom.de>, Steve Henson*
13767 NCONF_get_number() has no error checking at all. As a replacement,
13768 NCONF_get_number_e() is defined (`_e` for "error checking") and is
13769 promoted strongly. The old NCONF_get_number is kept around for
13770 binary backward compatibility.
13771 Make it possible for methods to load from something other than a BIO,
13772 by providing a function pointer that is given a name instead of a BIO.
13773 For example, this could be used to load configuration data from an
13778 * Fix for non blocking accept BIOs. Added new I/O special reason
13779 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
13780 with non blocking I/O was not possible because no retry code was
13781 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
13786 * Added the beginnings of Rijndael support.
13790 * Fix for bug in DirectoryString mask setting. Add support for
13791 X509_NAME_print_ex() in 'req' and X509_print_ex() function
13792 to allow certificate printing to more controllable, additional
13793 'certopt' option to 'x509' to allow new printing options to be
13798 * Clean old EAY MD5 hack from e_os.h.
13802 ### Changes between 0.9.6l and 0.9.6m [17 Mar 2004]
13804 * Fix null-pointer assignment in do_change_cipher_spec() revealed
13805 by using the Codenomicon TLS Test Tool ([CVE-2004-0079])
13807 *Joe Orton, Steve Henson*
13809 ### Changes between 0.9.6k and 0.9.6l [04 Nov 2003]
13811 * Fix additional bug revealed by the NISCC test suite:
13813 Stop bug triggering large recursion when presented with
13814 certain ASN.1 tags ([CVE-2003-0851])
13818 ### Changes between 0.9.6j and 0.9.6k [30 Sep 2003]
13820 * Fix various bugs revealed by running the NISCC test suite:
13822 Stop out of bounds reads in the ASN1 code when presented with
13823 invalid tags (CVE-2003-0543 and CVE-2003-0544).
13825 If verify callback ignores invalid public key errors don't try to check
13826 certificate signature with the NULL public key.
13830 * In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
13831 if the server requested one: as stated in TLS 1.0 and SSL 3.0
13836 * In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
13837 extra data after the compression methods not only for TLS 1.0
13838 but also for SSL 3.0 (as required by the specification).
13840 *Bodo Moeller; problem pointed out by Matthias Loepfe*
13842 * Change X509_certificate_type() to mark the key as exported/exportable
13843 when it's 512 *bits* long, not 512 bytes.
13847 ### Changes between 0.9.6i and 0.9.6j [10 Apr 2003]
13849 * Countermeasure against the Klima-Pokorny-Rosa extension of
13850 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
13851 a protocol version number mismatch like a decryption error
13852 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
13856 * Turn on RSA blinding by default in the default implementation
13857 to avoid a timing attack. Applications that don't want it can call
13858 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
13859 They would be ill-advised to do so in most cases.
13861 *Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller*
13863 * Change RSA blinding code so that it works when the PRNG is not
13864 seeded (in this case, the secret RSA exponent is abused as
13865 an unpredictable seed -- if it is not unpredictable, there
13866 is no point in blinding anyway). Make RSA blinding thread-safe
13867 by remembering the creator's thread ID in rsa->blinding and
13868 having all other threads use local one-time blinding factors
13869 (this requires more computation than sharing rsa->blinding, but
13870 avoids excessive locking; and if an RSA object is not shared
13871 between threads, blinding will still be very fast).
13875 ### Changes between 0.9.6h and 0.9.6i [19 Feb 2003]
13877 * In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
13878 via timing by performing a MAC computation even if incorrect
13879 block cipher padding has been found. This is a countermeasure
13880 against active attacks where the attacker has to distinguish
13881 between bad padding and a MAC verification error. ([CVE-2003-0078])
13883 *Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
13884 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
13885 Martin Vuagnoux (EPFL, Ilion)*
13887 ### Changes between 0.9.6g and 0.9.6h [5 Dec 2002]
13889 * New function OPENSSL_cleanse(), which is used to cleanse a section of
13890 memory from its contents. This is done with a counter that will
13891 place alternating values in each byte. This can be used to solve
13892 two issues: 1) the removal of calls to memset() by highly optimizing
13893 compilers, and 2) cleansing with other values than 0, since those can
13894 be read through on certain media, for example a swap space on disk.
13898 * Bugfix: client side session caching did not work with external caching,
13899 because the session->cipher setting was not restored when reloading
13900 from the external cache. This problem was masked, when
13901 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set.
13902 (Found by Steve Haslam <steve@araqnid.ddts.net>.)
13906 * Fix client_certificate (ssl/s2_clnt.c): The permissible total
13907 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33.
13909 *Zeev Lieber <zeev-l@yahoo.com>*
13911 * Undo an undocumented change introduced in 0.9.6e which caused
13912 repeated calls to OpenSSL_add_all_ciphers() and
13913 OpenSSL_add_all_digests() to be ignored, even after calling
13918 * Change the default configuration reader to deal with last line not
13919 being properly terminated.
13923 * Change X509_NAME_cmp() so it applies the special rules on handling
13924 DN values that are of type PrintableString, as well as RDNs of type
13925 emailAddress where the value has the type ia5String.
13927 *stefank@valicert.com via Richard Levitte*
13929 * Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half
13930 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently
13931 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be
13932 the bitwise-OR of the two for use by the majority of applications
13933 wanting this behaviour, and update the docs. The documented
13934 behaviour and actual behaviour were inconsistent and had been
13935 changing anyway, so this is more a bug-fix than a behavioural
13938 *Geoff Thorpe, diagnosed by Nadav Har'El*
13940 * Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c
13941 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes).
13945 * Fix initialization code race conditions in
13946 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(),
13947 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(),
13948 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(),
13949 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(),
13950 ssl2_get_cipher_by_char(),
13951 ssl3_get_cipher_by_char().
13953 *Patrick McCormick <patrick@tellme.com>, Bodo Moeller*
13955 * Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after
13956 the cached sessions are flushed, as the remove_cb() might use ex_data
13957 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com>
13958 (see [openssl.org #212]).
13960 *Geoff Thorpe, Lutz Jaenicke*
13962 * Fix typo in OBJ_txt2obj which incorrectly passed the content
13963 length, instead of the encoding length to d2i_ASN1_OBJECT.
13967 ### Changes between 0.9.6f and 0.9.6g [9 Aug 2002]
13969 * [In 0.9.6g-engine release:]
13970 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use `_stdcall`).
13972 *Lynn Gazis <lgazis@rainbow.com>*
13974 ### Changes between 0.9.6e and 0.9.6f [8 Aug 2002]
13976 * Fix ASN1 checks. Check for overflow by comparing with LONG_MAX
13977 and get fix the header length calculation.
13978 *Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>,
13979 Alon Kantor <alonk@checkpoint.com> (and others), Steve Henson*
13981 * Use proper error handling instead of 'assertions' in buffer
13982 overflow checks added in 0.9.6e. This prevents DoS (the
13983 assertions could call abort()).
13985 *Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller*
13987 ### Changes between 0.9.6d and 0.9.6e [30 Jul 2002]
13989 * Add various sanity checks to asn1_get_length() to reject
13990 the ASN1 length bytes if they exceed sizeof(long), will appear
13991 negative or the content length exceeds the length of the
13994 *Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>*
13996 * Fix cipher selection routines: ciphers without encryption had no flags
13997 for the cipher strength set and where therefore not handled correctly
13998 by the selection routines (PR #130).
14002 * Fix EVP_dsa_sha macro.
14007 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS
14008 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure
14009 that was added in OpenSSL 0.9.6d.
14011 As the countermeasure turned out to be incompatible with some
14012 broken SSL implementations, the new option is part of SSL_OP_ALL.
14013 SSL_OP_ALL is usually employed when compatibility with weird SSL
14014 implementations is desired (e.g. '-bugs' option to 's_client' and
14015 's_server'), so the new option is automatically set in many
14020 * Changes in security patch:
14022 Changes marked "(CHATS)" were sponsored by the Defense Advanced
14023 Research Projects Agency (DARPA) and Air Force Research Laboratory,
14024 Air Force Materiel Command, USAF, under agreement number
14027 * Add various sanity checks to asn1_get_length() to reject
14028 the ASN1 length bytes if they exceed sizeof(long), will appear
14029 negative or the content length exceeds the length of the
14030 supplied buffer. ([CVE-2002-0659])
14032 *Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>*
14034 * Assertions for various potential buffer overflows, not known to
14035 happen in practice.
14037 *Ben Laurie (CHATS)*
14039 * Various temporary buffers to hold ASCII versions of integers were
14040 too small for 64 bit platforms. ([CVE-2002-0655])
14041 *Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>*
14043 * Remote buffer overflow in SSL3 protocol - an attacker could
14044 supply an oversized session ID to a client. ([CVE-2002-0656])
14046 *Ben Laurie (CHATS)*
14048 * Remote buffer overflow in SSL2 protocol - an attacker could
14049 supply an oversized client master key. ([CVE-2002-0656])
14051 *Ben Laurie (CHATS)*
14053 ### Changes between 0.9.6c and 0.9.6d [9 May 2002]
14055 * Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
14056 encoded as NULL) with id-dsa-with-sha1.
14058 *Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller*
14060 * Check various `X509_...()` return values in `apps/req.c`.
14062 *Nils Larsch <nla@trustcenter.de>*
14064 * Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
14065 an end-of-file condition would erroneously be flagged, when the CRLF
14066 was just at the end of a processed block. The bug was discovered when
14067 processing data through a buffering memory BIO handing the data to a
14068 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
14069 <ptsekov@syntrex.com> and Nedelcho Stanev.
14073 * Implement a countermeasure against a vulnerability recently found
14074 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
14075 before application data chunks to avoid the use of known IVs
14076 with data potentially chosen by the attacker.
14080 * Fix length checks in ssl3_get_client_hello().
14084 * TLS/SSL library bugfix: use s->s3->in_read_app_data differently
14085 to prevent ssl3_read_internal() from incorrectly assuming that
14086 ssl3_read_bytes() found application data while handshake
14087 processing was enabled when in fact s->s3->in_read_app_data was
14088 merely automatically cleared during the initial handshake.
14090 *Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>*
14092 * Fix object definitions for Private and Enterprise: they were not
14093 recognized in their shortname (=lowercase) representation. Extend
14094 obj_dat.pl to issue an error when using undefined keywords instead
14095 of silently ignoring the problem (Svenning Sorensen
14096 <sss@sss.dnsalias.net>).
14100 * Fix DH_generate_parameters() so that it works for 'non-standard'
14101 generators, i.e. generators other than 2 and 5. (Previously, the
14102 code did not properly initialise the 'add' and 'rem' values to
14103 BN_generate_prime().)
14105 In the new general case, we do not insist that 'generator' is
14106 actually a primitive root: This requirement is rather pointless;
14107 a generator of the order-q subgroup is just as good, if not
14112 * Map new X509 verification errors to alerts. Discovered and submitted by
14113 Tom Wu <tom@arcot.com>.
14117 * Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
14118 returning non-zero before the data has been completely received
14119 when using non-blocking I/O.
14121 *Bodo Moeller; problem pointed out by John Hughes*
14123 * Some of the ciphers missed the strength entry (SSL_LOW etc).
14125 *Ben Laurie, Lutz Jaenicke*
14127 * Fix bug in SSL_clear(): bad sessions were not removed (found by
14128 Yoram Zahavi <YoramZ@gilian.com>).
14132 * Add information about CygWin 1.3 and on, and preserve proper
14133 configuration for the versions before that.
14135 *Corinna Vinschen <vinschen@redhat.com> and Richard Levitte*
14137 * Make removal from session cache (SSL_CTX_remove_session()) more robust:
14138 check whether we deal with a copy of a session and do not delete from
14139 the cache in this case. Problem reported by "Izhar Shoshani Levi"
14140 <izhar@checkpoint.com>.
14144 * Do not store session data into the internal session cache, if it
14145 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
14146 flag is set). Proposed by Aslam <aslam@funk.com>.
14150 * Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
14155 * [In 0.9.6d-engine release:]
14156 Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
14158 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
14160 * Add the configuration target linux-s390x.
14162 *Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte*
14164 * The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
14165 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
14166 variable as an indication that a ClientHello message has been
14167 received. As the flag value will be lost between multiple
14168 invocations of ssl3_accept when using non-blocking I/O, the
14169 function may not be aware that a handshake has actually taken
14170 place, thus preventing a new session from being added to the
14173 To avoid this problem, we now set s->new_session to 2 instead of
14174 using a local variable.
14176 *Lutz Jaenicke, Bodo Moeller*
14178 * Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
14179 if the SSL_R_LENGTH_MISMATCH error is detected.
14181 *Geoff Thorpe, Bodo Moeller*
14183 * New 'shared_ldflag' column in Configure platform table.
14187 * Fix EVP_CIPHER_mode macro.
14189 *"Dan S. Camper" <dan@bti.net>*
14191 * Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
14192 type, we must throw them away by setting rr->length to 0.
14194 *D P Chang <dpc@qualys.com>*
14196 ### Changes between 0.9.6b and 0.9.6c [21 dec 2001]
14198 * Fix BN_rand_range bug pointed out by Dominikus Scherkl
14199 <Dominikus.Scherkl@biodata.com>. (The previous implementation
14200 worked incorrectly for those cases where range = `10..._2` and
14201 `3*range` is two bits longer than range.)
14205 * Only add signing time to PKCS7 structures if it is not already
14210 * Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
14211 OBJ_ld_ce should be OBJ_id_ce.
14212 Also some ip-pda OIDs in crypto/objects/objects.txt were
14213 incorrect (cf. RFC 3039).
14215 *Matt Cooper, Frederic Giudicelli, Bodo Moeller*
14217 * Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
14218 returns early because it has nothing to do.
14220 *Andy Schneider <andy.schneider@bjss.co.uk>*
14222 * [In 0.9.6c-engine release:]
14223 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
14225 *Andy Schneider <andy.schneider@bjss.co.uk>*
14227 * [In 0.9.6c-engine release:]
14228 Add support for Cryptographic Appliance's keyserver technology.
14229 (Use engine 'keyclient')
14231 *Cryptographic Appliances and Geoff Thorpe*
14233 * Add a configuration entry for OS/390 Unix. The C compiler 'c89'
14234 is called via tools/c89.sh because arguments have to be
14235 rearranged (all '-L' options must appear before the first object
14238 *Richard Shapiro <rshapiro@abinitio.com>*
14240 * [In 0.9.6c-engine release:]
14241 Add support for Broadcom crypto accelerator cards, backported
14244 *Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox*
14246 * [In 0.9.6c-engine release:]
14247 Add support for SureWare crypto accelerator cards from
14248 Baltimore Technologies. (Use engine 'sureware')
14250 *Baltimore Technologies and Mark Cox*
14252 * [In 0.9.6c-engine release:]
14253 Add support for crypto accelerator cards from Accelerated
14254 Encryption Processing, www.aep.ie. (Use engine 'aep')
14256 *AEP Inc. and Mark Cox*
14258 * Add a configuration entry for gcc on UnixWare.
14260 *Gary Benson <gbenson@redhat.com>*
14262 * Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
14263 messages are stored in a single piece (fixed-length part and
14264 variable-length part combined) and fix various bugs found on the way.
14268 * Disable caching in BIO_gethostbyname(), directly use gethostbyname()
14269 instead. BIO_gethostbyname() does not know what timeouts are
14270 appropriate, so entries would stay in cache even when they have
14272 *Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>*
14274 * Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
14275 faced with a pathologically small ClientHello fragment that does
14276 not contain client_version: Instead of aborting with an error,
14277 simply choose the highest available protocol version (i.e.,
14278 TLS 1.0 unless it is disabled). In practice, ClientHello
14279 messages are never sent like this, but this change gives us
14280 strictly correct behaviour at least for TLS.
14284 * Fix SSL handshake functions and SSL_clear() such that SSL_clear()
14285 never resets s->method to s->ctx->method when called from within
14286 one of the SSL handshake functions.
14288 *Bodo Moeller; problem pointed out by Niko Baric*
14290 * In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
14291 (sent using the client's version number) if client_version is
14292 smaller than the protocol version in use. Also change
14293 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
14294 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
14295 the client will at least see that alert.
14299 * Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
14304 * Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
14305 client receives HelloRequest while in a handshake.
14307 *Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>*
14309 * Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
14310 should end in 'break', not 'goto end' which circumvents various
14311 cleanups done in state SSL_ST_OK. But session related stuff
14312 must be disabled for SSL_ST_OK in the case that we just sent a
14315 Also avoid some overhead by not calling ssl_init_wbio_buffer()
14316 before just sending a HelloRequest.
14318 *Bodo Moeller, Eric Rescorla <ekr@rtfm.com>*
14320 * Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
14321 reveal whether illegal block cipher padding was found or a MAC
14322 verification error occurred. (Neither SSLerr() codes nor alerts
14323 are directly visible to potential attackers, but the information
14324 may leak via logfiles.)
14326 Similar changes are not required for the SSL 2.0 implementation
14327 because the number of padding bytes is sent in clear for SSL 2.0,
14328 and the extra bytes are just ignored. However ssl/s2_pkt.c
14329 failed to verify that the purported number of padding bytes is in
14334 * Add OpenUNIX-8 support including shared libraries
14335 (Boyd Lynn Gerber <gerberb@zenez.com>).
14339 * Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
14340 'wristwatch attack' using huge encoding parameters (cf.
14341 James H. Manger's CRYPTO 2001 paper). Note that the
14342 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
14343 encoding parameters and hence was not vulnerable.
14347 * BN_sqr() bug fix.
14349 *Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>*
14351 * Rabin-Miller test analyses assume uniformly distributed witnesses,
14352 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
14353 followed by modular reduction.
14355 *Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>*
14357 * Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
14358 equivalent based on BN_pseudo_rand() instead of BN_rand().
14362 * s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
14363 This function was broken, as the check for a new client hello message
14364 to handle SGC did not allow these large messages.
14365 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
14369 * Add alert descriptions for TLSv1 to `SSL_alert_desc_string[_long]()`.
14373 * Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
14374 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
14378 * Rework the configuration and shared library support for Tru64 Unix.
14379 The configuration part makes use of modern compiler features and
14380 still retains old compiler behavior for those that run older versions
14381 of the OS. The shared library support part includes a variant that
14382 uses the RPATH feature, and is available through the special
14383 configuration target "alpha-cc-rpath", which will never be selected
14386 *Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte*
14388 * In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
14389 with the same message size as in ssl3_get_certificate_request().
14390 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
14391 messages might inadvertently be reject as too long.
14393 *Petr Lampa <lampa@fee.vutbr.cz>*
14395 * Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
14399 * Modified SSL library such that the verify_callback that has been set
14400 specifically for an SSL object with SSL_set_verify() is actually being
14401 used. Before the change, a verify_callback set with this function was
14402 ignored and the verify_callback() set in the SSL_CTX at the time of
14403 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
14404 to allow the necessary settings.
14408 * Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
14409 explicitly to NULL, as at least on Solaris 8 this seems not always to be
14410 done automatically (in contradiction to the requirements of the C
14411 standard). This made problems when used from OpenSSH.
14415 * In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
14416 dh->length and always used
14418 BN_rand_range(priv_key, dh->p).
14420 BN_rand_range() is not necessary for Diffie-Hellman, and this
14421 specific range makes Diffie-Hellman unnecessarily inefficient if
14422 dh->length (recommended exponent length) is much smaller than the
14423 length of dh->p. We could use BN_rand_range() if the order of
14424 the subgroup was stored in the DH structure, but we only have
14429 BN_rand(priv_key, l, ...)
14431 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
14438 RSA_eay_public_encrypt
14439 RSA_eay_private_decrypt
14440 RSA_eay_private_encrypt (signing)
14441 RSA_eay_public_decrypt (signature verification)
14443 (default implementations for RSA_public_encrypt,
14444 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
14445 always reject numbers >= n.
14449 * In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
14450 to synchronize access to 'locking_thread'. This is necessary on
14451 systems where access to 'locking_thread' (an 'unsigned long'
14452 variable) is not atomic.
14456 * In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
14457 *before* setting the 'crypto_lock_rand' flag. The previous code had
14458 a race condition if 0 is a valid thread ID.
14460 *Travis Vitek <vitek@roguewave.com>*
14462 * Add support for shared libraries under Irix.
14464 *Albert Chin-A-Young <china@thewrittenword.com>*
14466 * Add configuration option to build on Linux on both big-endian and
14467 little-endian MIPS.
14469 *Ralf Baechle <ralf@uni-koblenz.de>*
14471 * Add the possibility to create shared libraries on HP-UX.
14475 ### Changes between 0.9.6a and 0.9.6b [9 Jul 2001]
14477 * Change ssleay_rand_bytes (crypto/rand/md_rand.c)
14478 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
14479 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
14480 PRNG state recovery was possible based on the output of
14481 one PRNG request appropriately sized to gain knowledge on
14482 'md' followed by enough consecutive 1-byte PRNG requests
14483 to traverse all of 'state'.
14485 1. When updating 'md_local' (the current thread's copy of 'md')
14486 during PRNG output generation, hash all of the previous
14487 'md_local' value, not just the half used for PRNG output.
14489 2. Make the number of bytes from 'state' included into the hash
14490 independent from the number of PRNG bytes requested.
14492 The first measure alone would be sufficient to avoid
14493 Markku-Juhani's attack. (Actually it had never occurred
14494 to me that the half of 'md_local' used for chaining was the
14495 half from which PRNG output bytes were taken -- I had always
14496 assumed that the secret half would be used.) The second
14497 measure makes sure that additional data from 'state' is never
14498 mixed into 'md_local' in small portions; this heuristically
14499 further strengthens the PRNG.
14503 * Fix crypto/bn/asm/mips3.s.
14507 * When only the key is given to "enc", the IV is undefined. Print out
14508 an error message in this case.
14512 * Handle special case when X509_NAME is empty in X509 printing routines.
14516 * In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
14517 positive and less than q.
14521 * Don't change `*pointer` in CRYPTO_add_lock() is add_lock_callback is
14522 used: it isn't thread safe and the add_lock_callback should handle
14525 *Paul Rose <Paul.Rose@bridge.com>*
14527 * Verify that incoming data obeys the block size in
14528 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
14534 *Ulf Möller, Bodo Möller*
14536 * The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
14537 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
14538 when fixing the server behaviour for backwards-compatible 'client
14539 hello' messages. (Note that the attack is impractical against
14540 SSL 3.0 and TLS 1.0 anyway because length and version checking
14541 means that the probability of guessing a valid ciphertext is
14542 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
14545 Before 0.9.5, the countermeasure (hide the error by generating a
14546 random 'decryption result') did not work properly because
14547 ERR_clear_error() was missing, meaning that SSL_get_error() would
14548 detect the supposedly ignored error.
14550 Both problems are now fixed.
14554 * In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
14555 (previously it was 1024).
14559 * Fix for compatibility mode trust settings: ignore trust settings
14560 unless some valid trust or reject settings are present.
14564 * Fix for blowfish EVP: its a variable length cipher.
14568 * Fix various bugs related to DSA S/MIME verification. Handle missing
14569 parameters in DSA public key structures and return an error in the
14570 DSA routines if parameters are absent.
14574 * In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
14575 in the current directory if neither $RANDFILE nor $HOME was set.
14576 RAND_file_name() in 0.9.6a returned NULL in this case. This has
14577 caused some confusion to Windows users who haven't defined $HOME.
14578 Thus RAND_file_name() is changed again: e_os.h can define a
14579 DEFAULT_HOME, which will be used if $HOME is not set.
14580 For Windows, we use "C:"; on other platforms, we still require
14581 environment variables.
14583 * Move 'if (!initialized) RAND_poll()' into regions protected by
14584 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
14585 having multiple threads call RAND_poll() concurrently.
14589 * In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
14590 combination of a flag and a thread ID variable.
14591 Otherwise while one thread is in ssleay_rand_bytes (which sets the
14592 flag), *other* threads can enter ssleay_add_bytes without obeying
14593 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
14594 that they do not hold after the first thread unsets add_do_not_lock).
14598 * Change bctest again: '-x' expressions are not available in all
14599 versions of 'test'.
14603 ### Changes between 0.9.6 and 0.9.6a [5 Apr 2001]
14605 * Fix a couple of memory leaks in PKCS7_dataDecode()
14607 *Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>*
14609 * Change Configure and Makefiles to provide EXE_EXT, which will contain
14610 the default extension for executables, if any. Also, make the perl
14611 scripts that use symlink() to test if it really exists and use "cp"
14612 if it doesn't. All this made OpenSSL compilable and installable in
14617 * Fix for asn1_GetSequence() for indefinite length constructed data.
14618 If SEQUENCE is length is indefinite just set c->slen to the total
14619 amount of data available.
14621 *Steve Henson, reported by shige@FreeBSD.org*
14623 *This change does not apply to 0.9.7.*
14625 * Change bctest to avoid here-documents inside command substitution
14626 (workaround for FreeBSD /bin/sh bug).
14627 For compatibility with Ultrix, avoid shell functions (introduced
14628 in the bctest version that searches along $PATH).
14632 * Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
14633 with des_encrypt() defined on some operating systems, like Solaris
14638 * Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
14639 On the Importance of Eliminating Errors in Cryptographic
14640 Computations, J. Cryptology 14 (2001) 2, 101-119,
14641 <http://theory.stanford.edu/~dabo/papers/faults.ps.gz>).
14645 * MIPS assembler BIGNUM division bug fix.
14649 * Disabled incorrect Alpha assembler code.
14653 * Fix PKCS#7 decode routines so they correctly update the length
14654 after reading an EOC for the EXPLICIT tag.
14658 *This change does not apply to 0.9.7.*
14660 * Fix bug in PKCS#12 key generation routines. This was triggered
14661 if a 3DES key was generated with a 0 initial byte. Include
14662 PKCS12_BROKEN_KEYGEN compilation option to retain the old
14663 (but broken) behaviour.
14667 * Enhance bctest to search for a working bc along $PATH and print
14670 *Tim Rice <tim@multitalents.net> via Richard Levitte*
14672 * Fix memory leaks in err.c: free err_data string if necessary;
14673 don't write to the wrong index in ERR_set_error_data.
14677 * Implement ssl23_peek (analogous to ssl23_read), which previously
14682 * Replace rdtsc with `_emit` statements for VC++ version 5.
14684 *Jeremy Cooper <jeremy@baymoo.org>*
14686 * Make it possible to reuse SSLv2 sessions.
14690 * In copy_email() check for >= 0 as a return value for
14691 X509_NAME_get_index_by_NID() since 0 is a valid index.
14693 *Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>*
14695 * Avoid coredump with unsupported or invalid public keys by checking if
14696 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
14697 PKCS7_verify() fails with non detached data.
14701 * Don't use getenv in library functions when run as setuid/setgid.
14702 New function OPENSSL_issetugid().
14706 * Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
14707 due to incorrect handling of multi-threading:
14709 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
14711 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
14713 3. Count how many times MemCheck_off() has been called so that
14714 nested use can be treated correctly. This also avoids
14715 inband-signalling in the previous code (which relied on the
14716 assumption that thread ID 0 is impossible).
14720 * Add "-rand" option also to s_client and s_server.
14724 * Fix CPU detection on Irix 6.x.
14725 *Kurt Hockenbury <khockenb@stevens-tech.edu> and
14726 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>*
14728 * Fix X509_NAME bug which produced incorrect encoding if X509_NAME
14733 *This change does not apply to 0.9.7.*
14735 * Use the cached encoding of an X509_NAME structure rather than
14736 copying it. This is apparently the reason for the libsafe "errors"
14737 but the code is actually correct.
14741 * Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
14742 Bleichenbacher's DSA attack.
14743 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
14744 to be set and top=0 forces the highest bit to be set; top=-1 is new
14745 and leaves the highest bit random.
14747 *Ulf Moeller, Bodo Moeller*
14749 * In the `NCONF_...`-based implementations for `CONF_...` queries
14750 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
14751 a temporary CONF structure with the data component set to NULL
14752 (which gives segmentation faults in lh_retrieve).
14753 Instead, use NULL for the CONF pointer in CONF_get_string and
14754 CONF_get_number (which may use environment variables) and directly
14755 return NULL from CONF_get_section.
14759 * Fix potential buffer overrun for EBCDIC.
14763 * Tolerate nonRepudiation as being valid for S/MIME signing and certSign
14764 keyUsage if basicConstraints absent for a CA.
14768 * Make SMIME_write_PKCS7() write mail header values with a format that
14769 is more generally accepted (no spaces before the semicolon), since
14770 some programs can't parse those values properly otherwise. Also make
14771 sure BIO's that break lines after each write do not create invalid
14776 * Make the CRL encoding routines work with empty SEQUENCE OF. The
14777 macros previously used would not encode an empty SEQUENCE OF
14778 and break the signature.
14782 *This change does not apply to 0.9.7.*
14784 * Zero the premaster secret after deriving the master secret in
14789 * Add some EVP_add_digest_alias registrations (as found in
14790 OpenSSL_add_all_digests()) to SSL_library_init()
14791 aka OpenSSL_add_ssl_algorithms(). This provides improved
14792 compatibility with peers using X.509 certificates
14793 with unconventional AlgorithmIdentifier OIDs.
14797 * Fix for Irix with NO_ASM.
14799 *"Bruce W. Forsberg" <bruce.forsberg@baesystems.com>*
14801 * ./config script fixes.
14803 *Ulf Moeller, Richard Levitte*
14805 * Fix 'openssl passwd -1'.
14809 * Change PKCS12_key_gen_asc() so it can cope with non null
14810 terminated strings whose length is passed in the passlen
14811 parameter, for example from PEM callbacks. This was done
14812 by adding an extra length parameter to asc2uni().
14814 *Steve Henson, reported by <oddissey@samsung.co.kr>*
14816 * Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
14817 call failed, free the DSA structure.
14821 * Fix to uni2asc() to cope with zero length Unicode strings.
14822 These are present in some PKCS#12 files.
14826 * Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
14827 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
14828 when writing a 32767 byte record.
14830 *Bodo Moeller; problem reported by Eric Day <eday@concentric.net>*
14832 * In `RSA_eay_public_{en,ed}crypt` and RSA_eay_mod_exp (rsa_eay.c),
14833 obtain lock CRYPTO_LOCK_RSA before setting `rsa->_method_mod_{n,p,q}`.
14835 (RSA objects have a reference count access to which is protected
14836 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
14837 so they are meant to be shared between threads.)
14838 *Bodo Moeller, Geoff Thorpe; original patch submitted by
14839 "Reddie, Steven" <Steven.Reddie@ca.com>*
14841 * Fix a deadlock in CRYPTO_mem_leaks().
14845 * Use better test patterns in bntest.
14849 * rand_win.c fix for Borland C.
14853 * BN_rshift bugfix for n == 0.
14857 * Add a 'bctest' script that checks for some known 'bc' bugs
14858 so that 'make test' does not abort just because 'bc' is broken.
14862 * Store verify_result within SSL_SESSION also for client side to
14863 avoid potential security hole. (Re-used sessions on the client side
14864 always resulted in verify_result==X509_V_OK, not using the original
14865 result of the server certificate verification.)
14869 * Fix ssl3_pending: If the record in s->s3->rrec is not of type
14870 SSL3_RT_APPLICATION_DATA, return 0.
14871 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
14876 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
14877 releases, have been re-implemented by renaming the previous
14878 implementations of ssl2_read and ssl3_read to ssl2_read_internal
14879 and ssl3_read_internal, respectively, and adding 'peek' parameters
14880 to them. The new ssl[23]_{read,peek} functions are calls to
14881 ssl[23]_read_internal with the 'peek' flag set appropriately.
14882 A 'peek' parameter has also been added to ssl3_read_bytes, which
14883 does the actual work for ssl3_read_internal.
14887 * Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
14888 the method-specific "init()" handler. Also clean up ex_data after
14889 calling the method-specific "finish()" handler. Previously, this was
14890 happening the other way round.
14894 * Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
14895 The previous value, 12, was not always sufficient for BN_mod_exp().
14899 * Make sure that shared libraries get the internal name engine with
14900 the full version number and not just 0. This should mark the
14901 shared libraries as not backward compatible. Of course, this should
14902 be changed again when we can guarantee backward binary compatibility.
14906 * Fix typo in get_cert_by_subject() in by_dir.c
14908 *Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>*
14910 * Rework the system to generate shared libraries:
14912 - Make note of the expected extension for the shared libraries and
14913 if there is a need for symbolic links from for example libcrypto.so.0
14914 to libcrypto.so.0.9.7. There is extended info in Configure for
14917 - Make as few rebuilds of the shared libraries as possible.
14919 - Still avoid linking the OpenSSL programs with the shared libraries.
14921 - When installing, install the shared libraries separately from the
14926 * Fix SSL_CTX_set_read_ahead macro to actually use its argument.
14928 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
14929 and not in SSL_clear because the latter is also used by the
14930 accept/connect functions; previously, the settings made by
14931 SSL_set_read_ahead would be lost during the handshake.
14933 *Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>*
14935 * Correct util/mkdef.pl to be selective about disabled algorithms.
14936 Previously, it would create entries for disabled algorithms no
14941 * Added several new manual pages for SSL_* function.
14945 ### Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
14947 * In ssl23_get_client_hello, generate an error message when faced
14948 with an initial SSL 3.0/TLS record that is too small to contain the
14949 first two bytes of the ClientHello message, i.e. client_version.
14950 (Note that this is a pathologic case that probably has never happened
14951 in real life.) The previous approach was to use the version number
14952 from the record header as a substitute; but our protocol choice
14953 should not depend on that one because it is not authenticated
14954 by the Finished messages.
14958 * More robust randomness gathering functions for Windows.
14960 *Jeffrey Altman <jaltman@columbia.edu>*
14962 * For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
14963 not set then we don't setup the error code for issuer check errors
14964 to avoid possibly overwriting other errors which the callback does
14965 handle. If an application does set the flag then we assume it knows
14966 what it is doing and can handle the new informational codes
14971 * Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
14972 a general "ANY" type, as such it should be able to decode anything
14973 including tagged types. However it didn't check the class so it would
14974 wrongly interpret tagged types in the same way as their universal
14975 counterpart and unknown types were just rejected. Changed so that the
14976 tagged and unknown types are handled in the same way as a SEQUENCE:
14977 that is the encoding is stored intact. There is also a new type
14978 "V_ASN1_OTHER" which is used when the class is not universal, in this
14979 case we have no idea what the actual type is so we just lump them all
14984 * On VMS, stdout may very well lead to a file that is written to
14985 in a record-oriented fashion. That means that every write() will
14986 write a separate record, which will be read separately by the
14987 programs trying to read from it. This can be very confusing.
14989 The solution is to put a BIO filter in the way that will buffer
14990 text until a linefeed is reached, and then write everything a
14991 line at a time, so every record written will be an actual line,
14992 not chunks of lines and not (usually doesn't happen, but I've
14993 seen it once) several lines in one record. BIO_f_linebuffer() is
14996 Currently, it's a VMS-only method, because that's where it has
14997 been tested well enough.
15001 * Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
15002 it can return incorrect results.
15003 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
15004 but it was in 0.9.6-beta[12].)
15008 * Disable the check for content being present when verifying detached
15009 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
15010 include zero length content when signing messages.
15014 * New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
15015 BIO_ctrl (for BIO pairs).
15019 * Add DSO method for VMS.
15023 * Bug fix: Montgomery multiplication could produce results with the
15028 * Add RPM specification openssl.spec and modify it to build three
15029 packages. The default package contains applications, application
15030 documentation and run-time libraries. The devel package contains
15031 include files, static libraries and function documentation. The
15032 doc package contains the contents of the doc directory. The original
15033 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
15037 * Add a large number of documentation files for many SSL routines.
15039 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>*
15041 * Add a configuration entry for Sony News 4.
15043 *NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>*
15045 * Don't set the two most significant bits to one when generating a
15046 random number < q in the DSA library.
15050 * New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
15051 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
15052 the underlying transport is blocking) if a handshake took place.
15053 (The default behaviour is needed by applications such as s_client
15054 and s_server that use select() to determine when to use SSL_read;
15055 but for applications that know in advance when to expect data, it
15056 just makes things more complicated.)
15060 * Add RAND_egd_bytes(), which gives control over the number of bytes read
15065 * Add a few more EBCDIC conditionals that make `req` and `x509`
15066 work better on such systems.
15068 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
15070 * Add two demo programs for PKCS12_parse() and PKCS12_create().
15071 Update PKCS12_parse() so it copies the friendlyName and the
15072 keyid to the certificates aux info.
15076 * Fix bug in PKCS7_verify() which caused an infinite loop
15077 if there was more than one signature.
15079 *Sven Uszpelkat <su@celocom.de>*
15081 * Major change in util/mkdef.pl to include extra information
15082 about each symbol, as well as presenting variables as well
15083 as functions. This change means that there's n more need
15084 to rebuild the .num files when some algorithms are excluded.
15088 * Allow the verify time to be set by an application,
15089 rather than always using the current time.
15093 * Phase 2 verify code reorganisation. The certificate
15094 verify code now looks up an issuer certificate by a
15095 number of criteria: subject name, authority key id
15096 and key usage. It also verifies self signed certificates
15097 by the same criteria. The main comparison function is
15098 X509_check_issued() which performs these checks.
15100 Lot of changes were necessary in order to support this
15101 without completely rewriting the lookup code.
15103 Authority and subject key identifier are now cached.
15105 The LHASH 'certs' is X509_STORE has now been replaced
15106 by a STACK_OF(X509_OBJECT). This is mainly because an
15107 LHASH can't store or retrieve multiple objects with
15108 the same hash value.
15110 As a result various functions (which were all internal
15111 use only) have changed to handle the new X509_STORE
15112 structure. This will break anything that messed round
15113 with X509_STORE internally.
15115 The functions X509_STORE_add_cert() now checks for an
15116 exact match, rather than just subject name.
15118 The X509_STORE API doesn't directly support the retrieval
15119 of multiple certificates matching a given criteria, however
15120 this can be worked round by performing a lookup first
15121 (which will fill the cache with candidate certificates)
15122 and then examining the cache for matches. This is probably
15123 the best we can do without throwing out X509_LOOKUP
15124 entirely (maybe later...).
15126 The X509_VERIFY_CTX structure has been enhanced considerably.
15128 All certificate lookup operations now go via a get_issuer()
15129 callback. Although this currently uses an X509_STORE it
15130 can be replaced by custom lookups. This is a simple way
15131 to bypass the X509_STORE hackery necessary to make this
15132 work and makes it possible to use more efficient techniques
15133 in future. A very simple version which uses a simple
15134 STACK for its trusted certificate store is also provided
15135 using X509_STORE_CTX_trusted_stack().
15137 The verify_cb() and verify() callbacks now have equivalents
15138 in the X509_STORE_CTX structure.
15140 X509_STORE_CTX also has a 'flags' field which can be used
15141 to customise the verify behaviour.
15145 * Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
15146 excludes S/MIME capabilities.
15150 * When a certificate request is read in keep a copy of the
15151 original encoding of the signed data and use it when outputting
15152 again. Signatures then use the original encoding rather than
15153 a decoded, encoded version which may cause problems if the
15154 request is improperly encoded.
15158 * For consistency with other BIO_puts implementations, call
15159 buffer_write(b, ...) directly in buffer_puts instead of calling
15162 In BIO_puts, increment b->num_write as in BIO_write.
15164 *Peter.Sylvester@EdelWeb.fr*
15166 * Fix BN_mul_word for the case where the word is 0. (We have to use
15167 BN_zero, we may not return a BIGNUM with an array consisting of
15168 words set to zero.)
15172 * Avoid calling abort() from within the library when problems are
15173 detected, except if preprocessor symbols have been defined
15174 (such as REF_CHECK, BN_DEBUG etc.).
15178 * New openssl application 'rsautl'. This utility can be
15179 used for low-level RSA operations. DER public key
15180 BIO/fp routines also added.
15184 * New Configure entry and patches for compiling on QNX 4.
15186 *Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>*
15188 * A demo state-machine implementation was sponsored by
15189 Nuron (<http://www.nuron.com/>) and is now available in
15190 demos/state_machine.
15194 * New options added to the 'dgst' utility for signature
15195 generation and verification.
15199 * Unrecognized PKCS#7 content types are now handled via a
15200 catch all ASN1_TYPE structure. This allows unsupported
15201 types to be stored as a "blob" and an application can
15202 encode and decode it manually.
15206 * Fix various signed/unsigned issues to make a_strex.c
15207 compile under VC++.
15209 *Oscar Jacobsson <oscar.jacobsson@celocom.com>*
15211 * ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
15212 length if passed a buffer. ASN1_INTEGER_to_BN failed
15213 if passed a NULL BN and its argument was negative.
15215 *Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>*
15217 * Modification to PKCS#7 encoding routines to output definite
15218 length encoding. Since currently the whole structures are in
15219 memory there's not real point in using indefinite length
15220 constructed encoding. However if OpenSSL is compiled with
15221 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
15225 * Added BIO_vprintf() and BIO_vsnprintf().
15229 * Added more prefixes to parse for in the strings written
15230 through a logging bio, to cover all the levels that are available
15231 through syslog. The prefixes are now:
15233 PANIC, EMERG, EMR => LOG_EMERG
15234 ALERT, ALR => LOG_ALERT
15235 CRIT, CRI => LOG_CRIT
15236 ERROR, ERR => LOG_ERR
15237 WARNING, WARN, WAR => LOG_WARNING
15238 NOTICE, NOTE, NOT => LOG_NOTICE
15239 INFO, INF => LOG_INFO
15240 DEBUG, DBG => LOG_DEBUG
15242 and as before, if none of those prefixes are present at the
15243 beginning of the string, LOG_ERR is chosen.
15245 On Win32, the `LOG_*` levels are mapped according to this:
15247 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
15248 LOG_WARNING => EVENTLOG_WARNING_TYPE
15249 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
15253 * Made it possible to reconfigure with just the configuration
15254 argument "reconf" or "reconfigure". The command line arguments
15255 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
15256 and are retrieved from there when reconfiguring.
15262 *Assar Westerlund <assar@sics.se>, Richard Levitte*
15264 * Add the arguments -CAfile and -CApath to the pkcs12 utility.
15268 * The obj_dat.pl script was messing up the sorting of object
15269 names. The reason was that it compared the quoted version
15270 of strings as a result "OCSP" > "OCSP Signing" because
15271 " > SPACE. Changed script to store unquoted versions of
15272 names and add quotes on output. It was also omitting some
15273 names from the lookup table if they were given a default
15274 value (that is if SN is missing it is given the same
15275 value as LN and vice versa), these are now added on the
15276 grounds that if an object has a name we should be able to
15277 look it up. Finally added warning output when duplicate
15278 short or long names are found.
15282 * Changes needed for Tandem NSK.
15284 *Scott Uroff <scott@xypro.com>*
15286 * Fix SSL 2.0 rollback checking: Due to an off-by-one error in
15287 RSA_padding_check_SSLv23(), special padding was never detected
15288 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
15289 version rollback attacks was not effective.
15291 In s23_clnt.c, don't use special rollback-attack detection padding
15292 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
15293 client; similarly, in s23_srvr.c, don't do the rollback check if
15294 SSL 2.0 is the only protocol enabled in the server.
15298 * Make it possible to get hexdumps of unprintable data with 'openssl
15299 asn1parse'. By implication, the functions ASN1_parse_dump() and
15300 BIO_dump_indent() are added.
15304 * New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
15305 these print out strings and name structures based on various
15306 flags including RFC2253 support and proper handling of
15307 multibyte characters. Added options to the 'x509' utility
15308 to allow the various flags to be set.
15312 * Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
15313 Also change the functions X509_cmp_current_time() and
15314 X509_gmtime_adj() work with an ASN1_TIME structure,
15315 this will enable certificates using GeneralizedTime in validity
15316 dates to be checked.
15320 * Make the NEG_PUBKEY_BUG code (which tolerates invalid
15321 negative public key encodings) on by default,
15322 NO_NEG_PUBKEY_BUG can be set to disable it.
15326 * New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
15327 content octets. An i2c_ASN1_OBJECT is unnecessary because
15328 the encoding can be trivially obtained from the structure.
15332 * crypto/err.c locking bugfix: Use write locks (`CRYPTO_w_[un]lock`),
15333 not read locks (`CRYPTO_r_[un]lock`).
15337 * A first attempt at creating official support for shared
15338 libraries through configuration. I've kept it so the
15339 default is static libraries only, and the OpenSSL programs
15340 are always statically linked for now, but there are
15341 preparations for dynamic linking in place.
15342 This has been tested on Linux and Tru64.
15346 * Randomness polling function for Win9x, as described in:
15347 Peter Gutmann, Software Generation of Practically Strong
15352 * Fix so PRNG is seeded in req if using an already existing
15357 * New options to smime application. -inform and -outform
15358 allow alternative formats for the S/MIME message including
15359 PEM and DER. The -content option allows the content to be
15360 specified separately. This should allow things like Netscape
15361 form signing output easier to verify.
15365 * Fix the ASN1 encoding of tags using the 'long form'.
15369 * New ASN1 functions, `i2c_*` and `c2i_*` for INTEGER and BIT
15370 STRING types. These convert content octets to and from the
15371 underlying type. The actual tag and length octets are
15372 already assumed to have been read in and checked. These
15373 are needed because all other string types have virtually
15374 identical handling apart from the tag. By having versions
15375 of the ASN1 functions that just operate on content octets
15376 IMPLICIT tagging can be handled properly. It also allows
15377 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
15378 and ASN1_INTEGER are identical apart from the tag.
15382 * Change the handling of OID objects as follows:
15384 - New object identifiers are inserted in objects.txt, following
15385 the syntax given in [crypto/objects/README.md](crypto/objects/README.md).
15386 - objects.pl is used to process obj_mac.num and create a new
15388 - obj_dat.pl is used to create a new obj_dat.h, using the data in
15391 This is currently kind of a hack, and the perl code in objects.pl
15392 isn't very elegant, but it works as I intended. The simplest way
15393 to check that it worked correctly is to look in obj_dat.h and
15394 check the array nid_objs and make sure the objects haven't moved
15395 around (this is important!). Additions are OK, as well as
15396 consistent name changes.
15400 * Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
15404 * Addition of the command line parameter '-rand file' to 'openssl req'.
15405 The given file adds to whatever has already been seeded into the
15406 random pool through the RANDFILE configuration file option or
15407 environment variable, or the default random state file.
15411 * mkstack.pl now sorts each macro group into lexical order.
15412 Previously the output order depended on the order the files
15413 appeared in the directory, resulting in needless rewriting
15418 * Patches to make OpenSSL compile under Win32 again. Mostly
15419 work arounds for the VC++ problem that it treats func() as
15420 func(void). Also stripped out the parts of mkdef.pl that
15421 added extra typesafe functions: these no longer exist.
15425 * Reorganisation of the stack code. The macros are now all
15426 collected in safestack.h . Each macro is defined in terms of
15427 a "stack macro" of the form `SKM_<name>(type, a, b)`. The
15428 DEBUG_SAFESTACK is now handled in terms of function casts,
15429 this has the advantage of retaining type safety without the
15430 use of additional functions. If DEBUG_SAFESTACK is not defined
15431 then the non typesafe macros are used instead. Also modified the
15432 mkstack.pl script to handle the new form. Needs testing to see
15433 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
15434 the default if no major problems. Similar behaviour for ASN1_SET_OF
15435 and PKCS12_STACK_OF.
15439 * When some versions of IIS use the 'NET' form of private key the
15440 key derivation algorithm is different. Normally MD5(password) is
15441 used as a 128 bit RC4 key. In the modified case
15442 MD5(MD5(password) + "SGCKEYSALT") is used instead. Added some
15443 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
15444 as the old Netscape_RSA functions except they have an additional
15445 'sgckey' parameter which uses the modified algorithm. Also added
15446 an -sgckey command line option to the rsa utility. Thanks to
15447 Adrian Peck <bertie@ncipher.com> for posting details of the modified
15448 algorithm to openssl-dev.
15452 * The evp_local.h macros were using 'c.##kname' which resulted in
15453 invalid expansion on some systems (SCO 5.0.5 for example).
15454 Corrected to 'c.kname'.
15456 *Phillip Porch <root@theporch.com>*
15458 * New X509_get1_email() and X509_REQ_get1_email() functions that return
15459 a STACK of email addresses from a certificate or request, these look
15460 in the subject name and the subject alternative name extensions and
15461 omit any duplicate addresses.
15465 * Re-implement BN_mod_exp2_mont using independent (and larger) windows.
15466 This makes DSA verification about 2 % faster.
15470 * Increase maximum window size in `BN_mod_exp_...` to 6 bits instead of 5
15471 (meaning that now 2^5 values will be precomputed, which is only 4 KB
15472 plus overhead for 1024 bit moduli).
15473 This makes exponentiations about 0.5 % faster for 1024 bit
15474 exponents (as measured by "openssl speed rsa2048").
15478 * Rename memory handling macros to avoid conflicts with other
15480 Malloc => OPENSSL_malloc
15481 Malloc_locked => OPENSSL_malloc_locked
15482 Realloc => OPENSSL_realloc
15483 Free => OPENSSL_free
15487 * New function BN_mod_exp_mont_word for small bases (roughly 15%
15488 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
15492 * CygWin32 support.
15494 *John Jarvie <jjarvie@newsguy.com>*
15496 * The type-safe stack code has been rejigged. It is now only compiled
15497 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
15498 by default all type-specific stack functions are "#define"d back to
15499 standard stack functions. This results in more streamlined output
15500 but retains the type-safety checking possibilities of the original
15505 * The STACK code has been cleaned up, and certain type declarations
15506 that didn't make a lot of sense have been brought in line. This has
15507 also involved a cleanup of sorts in safestack.h to more correctly
15508 map type-safe stack functions onto their plain stack counterparts.
15509 This work has also resulted in a variety of "const"ifications of
15510 lots of the code, especially `_cmp` operations which should normally
15511 be prototyped with "const" parameters anyway.
15515 * When generating bytes for the first time in md_rand.c, 'stir the pool'
15516 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
15517 (The PRNG state consists of two parts, the large pool 'state' and 'md',
15518 where all of 'md' is used each time the PRNG is used, but 'state'
15519 is used only indexed by a cyclic counter. As entropy may not be
15520 well distributed from the beginning, 'md' is important as a
15521 chaining variable. However, the output function chains only half
15522 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
15523 all of 'md', and seeding with STATE_SIZE dummy bytes will result
15524 in all of 'state' being rewritten, with the new values depending
15525 on virtually all of 'md'. This overcomes the 80 bit limitation.)
15529 * In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
15530 the handshake is continued after ssl_verify_cert_chain();
15531 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
15532 can lead to 'unexplainable' connection aborts later.
15534 *Bodo Moeller; problem tracked down by Lutz Jaenicke*
15536 * Major EVP API cipher revision.
15537 Add hooks for extra EVP features. This allows various cipher
15538 parameters to be set in the EVP interface. Support added for variable
15539 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
15540 setting of RC2 and RC5 parameters.
15542 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
15545 Remove lots of duplicated code from the EVP library. For example *every*
15546 cipher init() function handles the 'iv' in the same way according to the
15547 cipher mode. They also all do nothing if the 'key' parameter is NULL and
15548 for CFB and OFB modes they zero ctx->num.
15550 New functionality allows removal of S/MIME code RC2 hack.
15552 Most of the routines have the same form and so can be declared in terms
15555 By shifting this to the top level EVP_CipherInit() it can be removed from
15556 all individual ciphers. If the cipher wants to handle IVs or keys
15557 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
15560 Change lots of functions like EVP_EncryptUpdate() to now return a
15561 value: although software versions of the algorithms cannot fail
15562 any installed hardware versions can.
15566 * Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
15567 this option is set, tolerate broken clients that send the negotiated
15568 protocol version number instead of the requested protocol version
15573 * Call dh_tmp_cb (set by `..._TMP_DH_CB`) with correct 'is_export' flag;
15574 i.e. non-zero for export ciphersuites, zero otherwise.
15575 Previous versions had this flag inverted, inconsistent with
15576 rsa_tmp_cb (..._TMP_RSA_CB).
15578 *Bodo Moeller; problem reported by Amit Chopra*
15580 * Add missing DSA library text string. Work around for some IIS
15581 key files with invalid SEQUENCE encoding.
15585 * Add a document (doc/standards.txt) that list all kinds of standards
15586 and so on that are implemented in OpenSSL.
15590 * Enhance c_rehash script. Old version would mishandle certificates
15591 with the same subject name hash and wouldn't handle CRLs at all.
15592 Added -fingerprint option to crl utility, to support new c_rehash
15597 * Eliminate non-ANSI declarations in crypto.h and stack.h.
15601 * Fix for SSL server purpose checking. Server checking was
15602 rejecting certificates which had extended key usage present
15603 but no ssl client purpose.
15605 *Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>*
15607 * Make PKCS#12 code work with no password. The PKCS#12 spec
15608 is a little unclear about how a blank password is handled.
15609 Since the password in encoded as a BMPString with terminating
15610 double NULL a zero length password would end up as just the
15611 double NULL. However no password at all is different and is
15612 handled differently in the PKCS#12 key generation code. NS
15613 treats a blank password as zero length. MSIE treats it as no
15614 password on export: but it will try both on import. We now do
15615 the same: PKCS12_parse() tries zero length and no password if
15616 the password is set to "" or NULL (NULL is now a valid password:
15617 it wasn't before) as does the pkcs12 application.
15621 * Bugfixes in `apps/x509.c`: Avoid a memory leak; and don't use
15622 perror when PEM_read_bio_X509_REQ fails, the error message must
15623 be obtained from the error queue.
15627 * Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
15628 it in ERR_remove_state if appropriate, and change ERR_get_state
15629 accordingly to avoid race conditions (this is necessary because
15630 thread_hash is no longer constant once set).
15634 * Bugfix for linux-elf makefile.one.
15638 * RSA_get_default_method() will now cause a default
15639 RSA_METHOD to be chosen if one doesn't exist already.
15640 Previously this was only set during a call to RSA_new()
15641 or RSA_new_method(NULL) meaning it was possible for
15642 RSA_get_default_method() to return NULL.
15646 * Added native name translation to the existing DSO code
15647 that will convert (if the flag to do so is set) filenames
15648 that are sufficiently small and have no path information
15649 into a canonical native form. Eg. "blah" converted to
15650 "libblah.so" or "blah.dll" etc.
15654 * New function ERR_error_string_n(e, buf, len) which is like
15655 ERR_error_string(e, buf), but writes at most 'len' bytes
15656 including the 0 terminator. For ERR_error_string_n, 'buf'
15659 *Damien Miller <djm@mindrot.org>, Bodo Moeller*
15661 * CONF library reworked to become more general. A new CONF
15662 configuration file reader "class" is implemented as well as a
15663 new functions (`NCONF_*`, for "New CONF") to handle it. The now
15664 old `CONF_*` functions are still there, but are reimplemented to
15665 work in terms of the new functions. Also, a set of functions
15666 to handle the internal storage of the configuration data is
15667 provided to make it easier to write new configuration file
15668 reader "classes" (I can definitely see something reading a
15669 configuration file in XML format, for example), called `_CONF_*`,
15670 or "the configuration storage API"...
15672 The new configuration file reading functions are:
15674 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
15675 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
15677 NCONF_default, NCONF_WIN32
15679 NCONF_dump_fp, NCONF_dump_bio
15681 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
15682 NCONF_new creates a new CONF object. This works in the same way
15683 as other interfaces in OpenSSL, like the BIO interface.
15684 `NCONF_dump_*` dump the internal storage of the configuration file,
15685 which is useful for debugging. All other functions take the same
15686 arguments as the old `CONF_*` functions with the exception of the
15687 first that must be a `CONF *` instead of a `LHASH *`.
15689 To make it easier to use the new classes with the old `CONF_*` functions,
15690 the function CONF_set_default_method is provided.
15694 * Add '-tls1' option to 'openssl ciphers', which was already
15695 mentioned in the documentation but had not been implemented.
15696 (This option is not yet really useful because even the additional
15697 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
15701 * Initial DSO code added into libcrypto for letting OpenSSL (and
15702 OpenSSL-based applications) load shared libraries and bind to
15703 them in a portable way.
15705 *Geoff Thorpe, with contributions from Richard Levitte*
15707 ### Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
15709 * Make sure _lrotl and _lrotr are only used with MSVC.
15711 * Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
15712 (the default implementation of RAND_status).
15714 * Rename openssl x509 option '-crlext', which was added in 0.9.5,
15715 to '-clrext' (= clear extensions), as intended and documented.
15716 *Bodo Moeller; inconsistency pointed out by Michael Attili
15717 <attili@amaxo.com>*
15719 * Fix for HMAC. It wasn't zeroing the rest of the block if the key length
15720 was larger than the MD block size.
15722 *Steve Henson, pointed out by Yost William <YostW@tce.com>*
15724 * Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
15725 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
15726 using the passed key: if the passed key was a private key the result
15727 of X509_print(), for example, would be to print out all the private key
15732 * des_quad_cksum() byte order bug fix.
15733 *Ulf Möller, using the problem description in krb4-0.9.7, where
15734 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>*
15736 * Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
15739 *Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>*
15741 * For easily testing in shell scripts whether some command
15742 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
15743 returns with exit code 0 iff no command of the given name is available.
15744 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
15745 the output goes to stdout and nothing is printed to stderr.
15746 Additional arguments are always ignored.
15748 Since for each cipher there is a command of the same name,
15749 the 'no-cipher' compilation switches can be tested this way.
15751 ('openssl no-XXX' is not able to detect pseudo-commands such
15752 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
15756 * Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
15760 * For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
15761 is set; it will be thrown away anyway because each handshake creates
15763 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
15764 to parameters -- in previous versions (since OpenSSL 0.9.3) the
15765 'default key' from SSL_CTX_set_tmp_dh would always be lost, meaning
15766 you effectively got SSL_OP_SINGLE_DH_USE when using this macro.
15770 * New s_client option -ign_eof: EOF at stdin is ignored, and
15771 'Q' and 'R' lose their special meanings (quit/renegotiate).
15772 This is part of what -quiet does; unlike -quiet, -ign_eof
15773 does not suppress any output.
15777 * Add compatibility options to the purpose and trust code. The
15778 purpose X509_PURPOSE_ANY is "any purpose" which automatically
15779 accepts a certificate or CA, this was the previous behaviour,
15780 with all the associated security issues.
15782 X509_TRUST_COMPAT is the old trust behaviour: only and
15783 automatically trust self signed roots in certificate store. A
15784 new trust setting X509_TRUST_DEFAULT is used to specify that
15785 a purpose has no associated trust setting and it should instead
15786 use the value in the default purpose.
15790 * Fix the PKCS#8 DSA private key code so it decodes keys again
15791 and fix a memory leak.
15795 * In util/mkerr.pl (which implements 'make errors'), preserve
15796 reason strings from the previous version of the .c file, as
15797 the default to have only downcase letters (and digits) in
15798 automatically generated reasons codes is not always appropriate.
15802 * In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
15803 using strerror. Previously, ERR_reason_error_string() returned
15804 library names as reason strings for SYSerr; but SYSerr is a special
15805 case where small numbers are errno values, not library numbers.
15809 * Add '-dsaparam' option to 'openssl dhparam' application. This
15810 converts DSA parameters into DH parameters. (When creating parameters,
15811 DSA_generate_parameters is used.)
15815 * Include 'length' (recommended exponent length) in C code generated
15816 by 'openssl dhparam -C'.
15820 * The second argument to set_label in perlasm was already being used
15821 so couldn't be used as a "file scope" flag. Moved to third argument
15826 * In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
15827 instead of RAND_bytes for encryption IVs and salts.
15831 * Include RAND_status() into RAND_METHOD instead of implementing
15832 it only for md_rand.c Otherwise replacing the PRNG by calling
15833 RAND_set_rand_method would be impossible.
15837 * Don't let DSA_generate_key() enter an infinite loop if the random
15838 number generation fails.
15842 * New 'rand' application for creating pseudo-random output.
15846 * Added configuration support for Linux/IA64
15848 *Rolf Haberrecker <rolf@suse.de>*
15850 * Assembler module support for Mingw32.
15854 * Shared library support for HPUX (in shlib/).
15856 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous*
15858 * Shared library support for Solaris gcc.
15860 *Lutz Behnke <behnke@trustcenter.de>*
15862 ### Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
15864 * PKCS7_encrypt() was adding text MIME headers twice because they
15865 were added manually and by SMIME_crlf_copy().
15869 * In bntest.c don't call BN_rand with zero bits argument.
15871 *Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>*
15873 * BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
15874 case was implemented. This caused BN_div_recp() to fail occasionally.
15878 * Add an optional second argument to the set_label() in the perl
15879 assembly language builder. If this argument exists and is set
15880 to 1 it signals that the assembler should use a symbol whose
15881 scope is the entire file, not just the current function. This
15882 is needed with MASM which uses the format label:: for this scope.
15884 *Steve Henson, pointed out by Peter Runestig <peter@runestig.com>*
15886 * Change the ASN1 types so they are typedefs by default. Before
15887 almost all types were #define'd to ASN1_STRING which was causing
15888 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
15893 * Change names of new functions to the new get1/get0 naming
15894 convention: After 'get1', the caller owns a reference count
15895 and has to call `..._free`; 'get0' returns a pointer to some
15896 data structure without incrementing reference counters.
15897 (Some of the existing 'get' functions increment a reference
15898 counter, some don't.)
15899 Similarly, 'set1' and 'add1' functions increase reference
15900 counters or duplicate objects.
15904 * Allow for the possibility of temp RSA key generation failure:
15905 the code used to assume it always worked and crashed on failure.
15909 * Fix potential buffer overrun problem in BIO_printf().
15910 *Ulf Möller, using public domain code by Patrick Powell; problem
15911 pointed out by David Sacerdote <das33@cornell.edu>*
15913 * Support EGD <http://www.lothar.com/tech/crypto/>. New functions
15914 RAND_egd() and RAND_status(). In the command line application,
15915 the EGD socket can be specified like a seed file using RANDFILE
15920 * Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
15921 Some CAs (e.g. Verisign) distribute certificates in this form.
15925 * Remove the SSL_ALLOW_ADH compile option and set the default cipher
15926 list to exclude them. This means that no special compilation option
15927 is needed to use anonymous DH: it just needs to be included in the
15932 * Change the EVP_MD_CTX_type macro so its meaning consistent with
15933 EVP_MD_type. The old functionality is available in a new macro called
15934 EVP_MD_md(). Change code that uses it and update docs.
15938 * `..._ctrl` functions now have corresponding `..._callback_ctrl` functions
15939 where the `void *` argument is replaced by a function pointer argument.
15940 Previously `void *` was abused to point to functions, which works on
15941 many platforms, but is not correct. As these functions are usually
15942 called by macros defined in OpenSSL header files, most source code
15943 should work without changes.
15947 * `<openssl/opensslconf.h>` (which is created by Configure) now contains
15948 sections with information on -D... compiler switches used for
15949 compiling the library so that applications can see them. To enable
15950 one of these sections, a pre-processor symbol `OPENSSL_..._DEFINES`
15951 must be defined. E.g.,
15952 #define OPENSSL_ALGORITHM_DEFINES
15953 #include <openssl/opensslconf.h>
15954 defines all pertinent `NO_<algo>` symbols, such as NO_IDEA, NO_RSA, etc.
15956 *Richard Levitte, Ulf and Bodo Möller*
15958 * Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
15963 * Change the 'other' type in certificate aux info to a STACK_OF
15964 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
15965 the required ASN1 format: arbitrary types determined by an OID.
15969 * Add some PEM_write_X509_REQ_NEW() functions and a command line
15970 argument to 'req'. This is not because the function is newer or
15971 better than others it just uses the work 'NEW' in the certificate
15972 request header lines. Some software needs this.
15976 * Reorganise password command line arguments: now passwords can be
15977 obtained from various sources. Delete the PEM_cb function and make
15978 it the default behaviour: i.e. if the callback is NULL and the
15979 usrdata argument is not NULL interpret it as a null terminated pass
15980 phrase. If usrdata and the callback are NULL then the pass phrase
15981 is prompted for as usual.
15985 * Add support for the Compaq Atalla crypto accelerator. If it is installed,
15986 the support is automatically enabled. The resulting binaries will
15987 autodetect the card and use it if present.
15989 *Ben Laurie and Compaq Inc.*
15991 * Work around for Netscape hang bug. This sends certificate request
15992 and server done in one record. Since this is perfectly legal in the
15993 SSL/TLS protocol it isn't a "bug" option and is on by default. See
15994 the bugs/SSLv3 entry for more info.
15998 * HP-UX tune-up: new unified configs, HP C compiler bug workaround.
16002 * Add -rand argument to smime and pkcs12 applications and read/write
16007 * New 'passwd' tool for crypt(3) and apr1 password hashes.
16011 * Add command line password options to the remaining applications.
16015 * Bug fix for BN_div_recp() for numerators with an even number of
16020 * More tests in bntest.c, and changed test_bn output.
16024 * ./config recognizes MacOS X now.
16028 * Bug fix for BN_div() when the first words of num and divisor are
16029 equal (it gave wrong results if `(rem=(n1-q*d0)&BN_MASK2) < d0)`.
16033 * Add support for various broken PKCS#8 formats, and command line
16034 options to produce them.
16038 * New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
16039 get temporary BIGNUMs from a BN_CTX.
16043 * Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
16048 * Change the `SSLeay_add_all_*()` functions to `OpenSSL_add_all_*()` and
16049 include a #define from the old name to the new. The original intent
16050 was that statically linked binaries could for example just call
16051 SSLeay_add_all_ciphers() to just add ciphers to the table and not
16052 link with digests. This never worked because SSLeay_add_all_digests()
16053 and SSLeay_add_all_ciphers() were in the same source file so calling
16054 one would link with the other. They are now in separate source files.
16058 * Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
16062 * Use a less unusual form of the Miller-Rabin primality test (it used
16063 a binary algorithm for exponentiation integrated into the Miller-Rabin
16064 loop, our standard modexp algorithms are faster).
16068 * Support for the EBCDIC character set completed.
16070 *Martin Kraemer <Martin.Kraemer@Mch.SNI.De>*
16072 * Source code cleanups: use const where appropriate, eliminate casts,
16073 use `void *` instead of `char *` in lhash.
16077 * Bugfix: ssl3_send_server_key_exchange was not restartable
16078 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
16079 this the server could overwrite ephemeral keys that the client
16084 * Turn DSA_is_prime into a macro that calls BN_is_prime,
16085 using 50 iterations of the Rabin-Miller test.
16087 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
16088 iterations of the Rabin-Miller test as required by the appendix
16089 to FIPS PUB 186[-1]) instead of DSA_is_prime.
16090 As BN_is_prime_fasttest includes trial division, DSA parameter
16091 generation becomes much faster.
16093 This implies a change for the callback functions in DSA_is_prime
16094 and DSA_generate_parameters: The callback function is called once
16095 for each positive witness in the Rabin-Miller test, not just
16096 occasionally in the inner loop; and the parameters to the
16097 callback function now provide an iteration count for the outer
16098 loop rather than for the current invocation of the inner loop.
16099 DSA_generate_parameters additionally can call the callback
16100 function with an 'iteration count' of -1, meaning that a
16101 candidate has passed the trial division test (when q is generated
16102 from an application-provided seed, trial division is skipped).
16106 * New function BN_is_prime_fasttest that optionally does trial
16107 division before starting the Rabin-Miller test and has
16108 an additional BN_CTX * argument (whereas BN_is_prime always
16109 has to allocate at least one BN_CTX).
16110 'callback(1, -1, cb_arg)' is called when a number has passed the
16111 trial division stage.
16115 * Fix for bug in CRL encoding. The validity dates weren't being handled
16120 * New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
16124 * New function BN_pseudo_rand().
16128 * Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
16129 bignum version of BN_from_montgomery() with the working code from
16130 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
16135 * Avoid a race condition in s2_clnt.c (function get_server_hello) that
16136 made it impossible to use the same SSL_SESSION data structure in
16137 SSL2 clients in multiple threads.
16141 * The return value of RAND_load_file() no longer counts bytes obtained
16142 by stat(). RAND_load_file(..., -1) is new and uses the complete file
16143 to seed the PRNG (previously an explicit byte count was required).
16145 *Ulf Möller, Bodo Möller*
16147 * Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
16148 used `char *` instead of `void *` and had casts all over the place.
16152 * Make BN_generate_prime() return NULL on error if ret!=NULL.
16156 * Retain source code compatibility for BN_prime_checks macro:
16157 BN_is_prime(..., BN_prime_checks, ...) now uses
16158 BN_prime_checks_for_size to determine the appropriate number of
16159 Rabin-Miller iterations.
16163 * Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
16164 DH_CHECK_P_NOT_SAFE_PRIME.
16165 (Check if this is true? OpenPGP calls them "strong".)
16169 * Merge the functionality of "dh" and "gendh" programs into a new program
16170 "dhparam". The old programs are retained for now but will handle DH keys
16171 (instead of parameters) in future.
16175 * Make the ciphers, s_server and s_client programs check the return values
16176 when a new cipher list is set.
16180 * Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
16181 ciphers. Before when the 56bit ciphers were enabled the sorting was
16184 The syntax for the cipher sorting has been extended to support sorting by
16185 cipher-strength (using the strength_bits hard coded in the tables).
16186 The new command is `@STRENGTH` (see also `doc/apps/ciphers.pod`).
16188 Fix a bug in the cipher-command parser: when supplying a cipher command
16189 string with an "undefined" symbol (neither command nor alphanumeric
16190 *A-Za-z0-9*, ssl_set_cipher_list used to hang in an endless loop. Now
16191 an error is flagged.
16193 Due to the strength-sorting extension, the code of the
16194 ssl_create_cipher_list() function was completely rearranged. I hope that
16195 the readability was also increased :-)
16197 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>*
16199 * Minor change to 'x509' utility. The -CAcreateserial option now uses 1
16200 for the first serial number and places 2 in the serial number file. This
16201 avoids problems when the root CA is created with serial number zero and
16202 the first user certificate has the same issuer name and serial number
16207 * Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
16208 the new code. Add documentation for this stuff.
16212 * Changes to X509_ATTRIBUTE utilities. These have been renamed from
16213 `X509_*()` to `X509at_*()` on the grounds that they don't handle X509
16214 structures and behave in an analogous way to the X509v3 functions:
16215 they shouldn't be called directly but wrapper functions should be used
16218 So we also now have some wrapper functions that call the X509at functions
16219 when passed certificate requests. (TO DO: similar things can be done with
16220 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
16221 things. Some of these need some d2i or i2d and print functionality
16222 because they handle more complex structures.)
16226 * Add missing #ifndefs that caused missing symbols when building libssl
16227 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
16228 NO_RSA in `ssl/s2*.c`.
16230 *Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller*
16232 * Precautions against using the PRNG uninitialized: RAND_bytes() now
16233 has a return value which indicates the quality of the random data
16234 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
16235 error queue. New function RAND_pseudo_bytes() generates output that is
16236 guaranteed to be unique but not unpredictable. RAND_add is like
16237 RAND_seed, but takes an extra argument for an entropy estimate
16238 (RAND_seed always assumes full entropy).
16242 * Do more iterations of Rabin-Miller probable prime test (specifically,
16243 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
16244 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
16245 in crypto/bn/bn_prime.c for the complete table). This guarantees a
16246 false-positive rate of at most 2^-80 for random input.
16250 * Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
16254 * New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
16255 in the 0.9.5 release), this returns the chain
16256 from an X509_CTX structure with a dup of the stack and all
16257 the X509 reference counts upped: so the stack will exist
16258 after X509_CTX_cleanup() has been called. Modify pkcs12.c
16261 Also make SSL_SESSION_print() print out the verify return
16266 * Add manpage for the pkcs12 command. Also change the default
16267 behaviour so MAC iteration counts are used unless the new
16268 -nomaciter option is used. This improves file security and
16269 only older versions of MSIE (4.0 for example) need it.
16273 * Honor the no-xxx Configure options when creating .DEF files.
16277 * Add PKCS#10 attributes to field table: challengePassword,
16278 unstructuredName and unstructuredAddress. These are taken from
16279 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
16280 international characters are used.
16282 More changes to X509_ATTRIBUTE code: allow the setting of types
16283 based on strings. Remove the 'loc' parameter when adding
16284 attributes because these will be a SET OF encoding which is sorted
16289 * Initial changes to the 'req' utility to allow request generation
16290 automation. This will allow an application to just generate a template
16291 file containing all the field values and have req construct the
16294 Initial support for X509_ATTRIBUTE handling. Stacks of these are
16295 used all over the place including certificate requests and PKCS#7
16296 structures. They are currently handled manually where necessary with
16297 some primitive wrappers for PKCS#7. The new functions behave in a
16298 manner analogous to the X509 extension functions: they allow
16299 attributes to be looked up by NID and added.
16301 Later something similar to the X509V3 code would be desirable to
16302 automatically handle the encoding, decoding and printing of the
16303 more complex types. The string types like challengePassword can
16304 be handled by the string table functions.
16306 Also modified the multi byte string table handling. Now there is
16307 a 'global mask' which masks out certain types. The table itself
16308 can use the flag STABLE_NO_MASK to ignore the mask setting: this
16309 is useful when for example there is only one permissible type
16310 (as in countryName) and using the mask might result in no valid
16315 * Clean up 'Finished' handling, and add functions SSL_get_finished and
16316 SSL_get_peer_finished to allow applications to obtain the latest
16317 Finished messages sent to the peer or expected from the peer,
16318 respectively. (SSL_get_peer_finished is usually the Finished message
16319 actually received from the peer, otherwise the protocol will be aborted.)
16321 As the Finished message are message digests of the complete handshake
16322 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
16323 be used for external authentication procedures when the authentication
16324 provided by SSL/TLS is not desired or is not enough.
16328 * Enhanced support for Alpha Linux is added. Now ./config checks if
16329 the host supports BWX extension and if Compaq C is present on the
16330 $PATH. Just exploiting of the BWX extension results in 20-30%
16331 performance kick for some algorithms, e.g. DES and RC4 to mention
16332 a couple. Compaq C in turn generates ~20% faster code for MD5 and
16337 * Add support for MS "fast SGC". This is arguably a violation of the
16338 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
16339 weak crypto and after checking the certificate is SGC a second one
16340 with strong crypto. MS SGC stops the first handshake after receiving
16341 the server certificate message and sends a second client hello. Since
16342 a server will typically do all the time consuming operations before
16343 expecting any further messages from the client (server key exchange
16344 is the most expensive) there is little difference between the two.
16346 To get OpenSSL to support MS SGC we have to permit a second client
16347 hello message after we have sent server done. In addition we have to
16348 reset the MAC if we do get this second client hello.
16352 * Add a function 'd2i_AutoPrivateKey()' this will automatically decide
16353 if a DER encoded private key is RSA or DSA traditional format. Changed
16354 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
16355 format DER encoded private key. Newer code should use PKCS#8 format which
16356 has the key type encoded in the ASN1 structure. Added DER private key
16357 support to pkcs8 application.
16361 * SSL 3/TLS 1 servers now don't request certificates when an anonymous
16362 ciphersuites has been selected (as required by the SSL 3/TLS 1
16363 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
16364 is set, we interpret this as a request to violate the specification
16365 (the worst that can happen is a handshake failure, and 'correct'
16366 behaviour would result in a handshake failure anyway).
16370 * In SSL_CTX_add_session, take into account that there might be multiple
16371 SSL_SESSION structures with the same session ID (e.g. when two threads
16372 concurrently obtain them from an external cache).
16373 The internal cache can handle only one SSL_SESSION with a given ID,
16374 so if there's a conflict, we now throw out the old one to achieve
16379 * Add OIDs for idea and blowfish in CBC mode. This will allow both
16380 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
16381 some routines that use cipher OIDs: some ciphers do not have OIDs
16382 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
16387 * Simplify the trust setting structure and code. Now we just have
16388 two sequences of OIDs for trusted and rejected settings. These will
16389 typically have values the same as the extended key usage extension
16390 and any application specific purposes.
16392 The trust checking code now has a default behaviour: it will just
16393 check for an object with the same NID as the passed id. Functions can
16394 be provided to override either the default behaviour or the behaviour
16395 for a given id. SSL client, server and email already have functions
16396 in place for compatibility: they check the NID and also return "trusted"
16397 if the certificate is self signed.
16401 * Add d2i,i2d bio/fp functions for PrivateKey: these convert the
16402 traditional format into an EVP_PKEY structure.
16406 * Add a password callback function PEM_cb() which either prompts for
16407 a password if usr_data is NULL or otherwise assumes it is a null
16408 terminated password. Allow passwords to be passed on command line
16409 environment or config files in a few more utilities.
16413 * Add a bunch of DER and PEM functions to handle PKCS#8 format private
16414 keys. Add some short names for PKCS#8 PBE algorithms and allow them
16415 to be specified on the command line for the pkcs8 and pkcs12 utilities.
16416 Update documentation.
16420 * Support for ASN1 "NULL" type. This could be handled before by using
16421 ASN1_TYPE but there wasn't any function that would try to read a NULL
16422 and produce an error if it couldn't. For compatibility we also have
16423 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
16424 don't allocate anything because they don't need to.
16428 * Initial support for MacOS is now provided. Examine INSTALL.MacOS
16431 *Andy Polyakov, Roy Woods <roy@centicsystems.ca>*
16433 * Rebuild of the memory allocation routines used by OpenSSL code and
16434 possibly others as well. The purpose is to make an interface that
16435 provide hooks so anyone can build a separate set of allocation and
16436 deallocation routines to be used by OpenSSL, for example memory
16437 pool implementations, or something else, which was previously hard
16438 since Malloc(), Realloc() and Free() were defined as macros having
16439 the values malloc, realloc and free, respectively (except for Win32
16440 compilations). The same is provided for memory debugging code.
16441 OpenSSL already comes with functionality to find memory leaks, but
16442 this gives people a chance to debug other memory problems.
16444 With these changes, a new set of functions and macros have appeared:
16446 CRYPTO_set_mem_debug_functions() [F]
16447 CRYPTO_get_mem_debug_functions() [F]
16448 CRYPTO_dbg_set_options() [F]
16449 CRYPTO_dbg_get_options() [F]
16450 CRYPTO_malloc_debug_init() [M]
16452 The memory debug functions are NULL by default, unless the library
16453 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
16454 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
16455 gives the standard debugging functions that come with OpenSSL) or
16456 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
16457 provided by the library user) must be used. When the standard
16458 debugging functions are used, CRYPTO_dbg_set_options can be used to
16459 request additional information:
16460 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
16461 the CRYPTO_MDEBUG_xxx macro when compiling the library.
16463 Also, things like CRYPTO_set_mem_functions will always give the
16464 expected result (the new set of functions is used for allocation
16465 and deallocation) at all times, regardless of platform and compiler
16468 To finish it up, some functions that were never use in any other
16469 way than through macros have a new API and new semantic:
16471 CRYPTO_dbg_malloc()
16472 CRYPTO_dbg_realloc()
16475 All macros of value have retained their old syntax.
16477 *Richard Levitte and Bodo Moeller*
16479 * Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
16480 ordering of SMIMECapabilities wasn't in "strength order" and there
16481 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
16486 * Some ASN1 types with illegal zero length encoding (INTEGER,
16487 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
16489 *Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson*
16491 * Merge in my S/MIME library for OpenSSL. This provides a simple
16492 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
16493 functionality to handle multipart/signed properly) and a utility
16494 called 'smime' to call all this stuff. This is based on code I
16495 originally wrote for Celo who have kindly allowed it to be
16496 included in OpenSSL.
16500 * Add variants des_set_key_checked and des_set_key_unchecked of
16501 des_set_key (aka des_key_sched). Global variable des_check_key
16502 decides which of these is called by des_set_key; this way
16503 des_check_key behaves as it always did, but applications and
16504 the library itself, which was buggy for des_check_key == 1,
16505 have a cleaner way to pick the version they need.
16509 * New function PKCS12_newpass() which changes the password of a
16514 * Modify X509_TRUST and X509_PURPOSE so it also uses a static and
16515 dynamic mix. In both cases the ids can be used as an index into the
16516 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
16517 functions so they accept a list of the field values and the
16518 application doesn't need to directly manipulate the X509_TRUST
16523 * Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
16528 * Modify the way the V3 extension code looks up extensions. This now
16529 works in a similar way to the object code: we have some "standard"
16530 extensions in a static table which is searched with OBJ_bsearch()
16531 and the application can add dynamic ones if needed. The file
16532 crypto/x509v3/ext_dat.h now has the info: this file needs to be
16533 updated whenever a new extension is added to the core code and kept
16534 in ext_nid order. There is a simple program 'tabtest.c' which checks
16535 this. New extensions are not added too often so this file can readily
16536 be maintained manually.
16538 There are two big advantages in doing things this way. The extensions
16539 can be looked up immediately and no longer need to be "added" using
16540 X509V3_add_standard_extensions(): this function now does nothing.
16541 Side note: I get *lots* of email saying the extension code doesn't
16542 work because people forget to call this function.
16543 Also no dynamic allocation is done unless new extensions are added:
16544 so if we don't add custom extensions there is no need to call
16545 X509V3_EXT_cleanup().
16549 * Modify enc utility's salting as follows: make salting the default. Add a
16550 magic header, so unsalted files fail gracefully instead of just decrypting
16551 to garbage. This is because not salting is a big security hole, so people
16552 should be discouraged from doing it.
16556 * Fixes and enhancements to the 'x509' utility. It allowed a message
16557 digest to be passed on the command line but it only used this
16558 parameter when signing a certificate. Modified so all relevant
16559 operations are affected by the digest parameter including the
16560 -fingerprint and -x509toreq options. Also -x509toreq choked if a
16561 DSA key was used because it didn't fix the digest.
16565 * Initial certificate chain verify code. Currently tests the untrusted
16566 certificates for consistency with the verify purpose (which is set
16567 when the X509_STORE_CTX structure is set up) and checks the pathlength.
16569 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
16570 this is because it will reject chains with invalid extensions whereas
16571 every previous version of OpenSSL and SSLeay made no checks at all.
16573 Trust code: checks the root CA for the relevant trust settings. Trust
16574 settings have an initial value consistent with the verify purpose: e.g.
16575 if the verify purpose is for SSL client use it expects the CA to be
16576 trusted for SSL client use. However the default value can be changed to
16577 permit custom trust settings: one example of this would be to only trust
16578 certificates from a specific "secure" set of CAs.
16580 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
16581 which should be used for version portability: especially since the
16582 verify structure is likely to change more often now.
16584 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
16585 to set them. If not set then assume SSL clients will verify SSL servers
16588 Two new options to the verify program: -untrusted allows a set of
16589 untrusted certificates to be passed in and -purpose which sets the
16590 intended purpose of the certificate. If a purpose is set then the
16591 new chain verify code is used to check extension consistency.
16595 * Support for the authority information access extension.
16599 * Modify RSA and DSA PEM read routines to transparently handle
16600 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
16601 public keys in a format compatible with certificate
16602 SubjectPublicKeyInfo structures. Unfortunately there were already
16603 functions called *_PublicKey_* which used various odd formats so
16604 these are retained for compatibility: however the DSA variants were
16605 never in a public release so they have been deleted. Changed dsa/rsa
16606 utilities to handle the new format: note no releases ever handled public
16607 keys so we should be OK.
16609 The primary motivation for this change is to avoid the same fiasco
16610 that dogs private keys: there are several incompatible private key
16611 formats some of which are standard and some OpenSSL specific and
16612 require various evil hacks to allow partial transparent handling and
16613 even then it doesn't work with DER formats. Given the option anything
16614 other than PKCS#8 should be dumped: but the other formats have to
16615 stay in the name of compatibility.
16617 With public keys and the benefit of hindsight one standard format
16618 is used which works with EVP_PKEY, RSA or DSA structures: though
16619 it clearly returns an error if you try to read the wrong kind of key.
16621 Added a -pubkey option to the 'x509' utility to output the public key.
16622 Also rename the `EVP_PKEY_get_*()` to `EVP_PKEY_rget_*()`
16623 (renamed to `EVP_PKEY_get1_*()` in the OpenSSL 0.9.5 release) and add
16624 `EVP_PKEY_rset_*()` functions (renamed to `EVP_PKEY_set1_*()`)
16625 that do the same as the `EVP_PKEY_assign_*()` except they up the
16626 reference count of the added key (they don't "swallow" the
16631 * Fixes to crypto/x509/by_file.c the code to read in certificates and
16632 CRLs would fail if the file contained no certificates or no CRLs:
16633 added a new function to read in both types and return the number
16634 read: this means that if none are read it will be an error. The
16635 DER versions of the certificate and CRL reader would always fail
16636 because it isn't possible to mix certificates and CRLs in DER format
16637 without choking one or the other routine. Changed this to just read
16638 a certificate: this is the best we can do. Also modified the code
16639 in `apps/verify.c` to take notice of return codes: it was previously
16640 attempting to read in certificates from NULL pointers and ignoring
16641 any errors: this is one reason why the cert and CRL reader seemed
16642 to work. It doesn't check return codes from the default certificate
16643 routines: these may well fail if the certificates aren't installed.
16647 * Code to support otherName option in GeneralName.
16651 * First update to verify code. Change the verify utility
16652 so it warns if it is passed a self signed certificate:
16653 for consistency with the normal behaviour. X509_verify
16654 has been modified to it will now verify a self signed
16655 certificate if *exactly* the same certificate appears
16656 in the store: it was previously impossible to trust a
16657 single self signed certificate. This means that:
16658 openssl verify ss.pem
16659 now gives a warning about a self signed certificate but
16660 openssl verify -CAfile ss.pem ss.pem
16665 * For servers, store verify_result in SSL_SESSION data structure
16666 (and add it to external session representation).
16667 This is needed when client certificate verifications fails,
16668 but an application-provided verification callback (set by
16669 SSL_CTX_set_cert_verify_callback) allows accepting the session
16670 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
16671 but returns 1): When the session is reused, we have to set
16672 ssl->verify_result to the appropriate error code to avoid
16675 *Bodo Moeller, problem pointed out by Lutz Jaenicke*
16677 * Fix a bug in the new PKCS#7 code: it didn't consider the
16678 case in PKCS7_dataInit() where the signed PKCS7 structure
16679 didn't contain any existing data because it was being created.
16681 *Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson*
16683 * Add a salt to the key derivation routines in enc.c. This
16684 forms the first 8 bytes of the encrypted file. Also add a
16685 -S option to allow a salt to be input on the command line.
16689 * New function X509_cmp(). Oddly enough there wasn't a function
16690 to compare two certificates. We do this by working out the SHA1
16691 hash and comparing that. X509_cmp() will be needed by the trust
16696 * SSL_get1_session() is like SSL_get_session(), but increments
16697 the reference count in the SSL_SESSION returned.
16699 *Geoff Thorpe <geoff@eu.c2.net>*
16701 * Fix for 'req': it was adding a null to request attributes.
16702 Also change the X509_LOOKUP and X509_INFO code to handle
16703 certificate auxiliary information.
16707 * Add support for 40 and 64 bit RC2 and RC4 algorithms: document
16712 * Add the possibility to add extra information to the memory leak
16713 detecting output, to form tracebacks, showing from where each
16714 allocation was originated: CRYPTO_push_info("constant string") adds
16715 the string plus current file name and line number to a per-thread
16716 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
16717 is like calling CYRPTO_pop_info() until the stack is empty.
16718 Also updated memory leak detection code to be multi-thread-safe.
16722 * Add options -text and -noout to pkcs7 utility and delete the
16723 encryption options which never did anything. Update docs.
16727 * Add options to some of the utilities to allow the pass phrase
16728 to be included on either the command line (not recommended on
16729 OSes like Unix) or read from the environment. Update the
16730 manpages and fix a few bugs.
16734 * Add a few manpages for some of the openssl commands.
16738 * Fix the -revoke option in ca. It was freeing up memory twice,
16739 leaking and not finding already revoked certificates.
16743 * Extensive changes to support certificate auxiliary information.
16744 This involves the use of X509_CERT_AUX structure and X509_AUX
16745 functions. An X509_AUX function such as PEM_read_X509_AUX()
16746 can still read in a certificate file in the usual way but it
16747 will also read in any additional "auxiliary information". By
16748 doing things this way a fair degree of compatibility can be
16749 retained: existing certificates can have this information added
16750 using the new 'x509' options.
16752 Current auxiliary information includes an "alias" and some trust
16753 settings. The trust settings will ultimately be used in enhanced
16754 certificate chain verification routines: currently a certificate
16755 can only be trusted if it is self signed and then it is trusted
16760 * Fix assembler for Alpha (tested only on DEC OSF not Linux or `*BSD`).
16761 The problem was that one of the replacement routines had not been working
16762 since SSLeay releases. For now the offending routine has been replaced
16763 with non-optimised assembler. Even so, this now gives around 95%
16764 performance improvement for 1024 bit RSA signs.
16768 * Hack to fix PKCS#7 decryption when used with some unorthodox RC2
16769 handling. Most clients have the effective key size in bits equal to
16770 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
16771 A few however don't do this and instead use the size of the decrypted key
16772 to determine the RC2 key length and the AlgorithmIdentifier to determine
16773 the effective key length. In this case the effective key length can still
16774 be 40 bits but the key length can be 168 bits for example. This is fixed
16775 by manually forcing an RC2 key into the EVP_PKEY structure because the
16776 EVP code can't currently handle unusual RC2 key sizes: it always assumes
16777 the key length and effective key length are equal.
16781 * Add a bunch of functions that should simplify the creation of
16782 X509_NAME structures. Now you should be able to do:
16783 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
16784 and have it automatically work out the correct field type and fill in
16785 the structures. The more adventurous can try:
16786 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
16787 and it will (hopefully) work out the correct multibyte encoding.
16791 * Change the 'req' utility to use the new field handling and multibyte
16792 copy routines. Before the DN field creation was handled in an ad hoc
16793 way in req, ca, and x509 which was rather broken and didn't support
16794 BMPStrings or UTF8Strings. Since some software doesn't implement
16795 BMPStrings or UTF8Strings yet, they can be enabled using the config file
16796 using the dirstring_type option. See the new comment in the default
16797 openssl.cnf for more info.
16801 * Make crypto/rand/md_rand.c more robust:
16802 - Assure unique random numbers after fork().
16803 - Make sure that concurrent threads access the global counter and
16804 md serializably so that we never lose entropy in them
16805 or use exactly the same state in multiple threads.
16806 Access to the large state is not always serializable because
16807 the additional locking could be a performance killer, and
16808 md should be large enough anyway.
16812 * New file `apps/app_rand.c` with commonly needed functionality
16813 for handling the random seed file.
16815 Use the random seed file in some applications that previously did not:
16817 dsaparam -genkey (which also ignored its '-rand' option),
16820 x509 (when signing).
16821 Except on systems with /dev/urandom, it is crucial to have a random
16822 seed file at least for key creation, DSA signing, and for DH exchanges;
16823 for RSA signatures we could do without one.
16825 gendh and gendsa (unlike genrsa) used to read only the first byte
16826 of each file listed in the '-rand' option. The function as previously
16827 found in genrsa is now in app_rand.c and is used by all programs
16828 that support '-rand'.
16832 * In RAND_write_file, use mode 0600 for creating files;
16833 don't just chmod when it may be too late.
16837 * Report an error from X509_STORE_load_locations
16838 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
16842 * New function ASN1_mbstring_copy() this copies a string in either
16843 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
16844 into an ASN1_STRING type. A mask of permissible types is passed
16845 and it chooses the "minimal" type to use or an error if not type
16850 * Add function equivalents to the various macros in asn1.h. The old
16851 macros are retained with an `M_` prefix. Code inside the library can
16852 use the `M_` macros. External code (including the openssl utility)
16853 should *NOT* in order to be "shared library friendly".
16857 * Add various functions that can check a certificate's extensions
16858 to see if it usable for various purposes such as SSL client,
16859 server or S/MIME and CAs of these types. This is currently
16860 VERY EXPERIMENTAL but will ultimately be used for certificate chain
16861 verification. Also added a -purpose flag to x509 utility to
16862 print out all the purposes.
16866 * Add a CRYPTO_EX_DATA to X509 certificate structure and associated
16871 * New `X509V3_{X509,CRL,REVOKED}_get_d2i()` functions. These will search
16872 for, obtain and decode and extension and obtain its critical flag.
16873 This allows all the necessary extension code to be handled in a
16874 single function call.
16878 * RC4 tune-up featuring 30-40% performance improvement on most RISC
16879 platforms. See crypto/rc4/rc4_enc.c for further details.
16883 * New -noout option to asn1parse. This causes no output to be produced
16884 its main use is when combined with -strparse and -out to extract data
16885 from a file (which may not be in ASN.1 format).
16889 * Fix for pkcs12 program. It was hashing an invalid certificate pointer
16890 when producing the local key id.
16892 *Richard Levitte <levitte@stacken.kth.se>*
16894 * New option -dhparam in s_server. This allows a DH parameter file to be
16895 stated explicitly. If it is not stated then it tries the first server
16896 certificate file. The previous behaviour hard coded the filename
16901 * Add -pubin and -pubout options to the rsa and dsa commands. These allow
16902 a public key to be input or output. For example:
16903 openssl rsa -in key.pem -pubout -out pubkey.pem
16904 Also added necessary DSA public key functions to handle this.
16908 * Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
16909 in the message. This was handled by allowing
16910 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
16912 *Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>*
16914 * Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
16915 to the end of the strings whereas this didn't. This would cause problems
16916 if strings read with d2i_ASN1_bytes() were later modified.
16918 *Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>*
16920 * Fix for base64 decode bug. When a base64 bio reads only one line of
16921 data and it contains EOF it will end up returning an error. This is
16922 caused by input 46 bytes long. The cause is due to the way base64
16923 BIOs find the start of base64 encoded data. They do this by trying a
16924 trial decode on each line until they find one that works. When they
16925 do a flag is set and it starts again knowing it can pass all the
16926 data directly through the decoder. Unfortunately it doesn't reset
16927 the context it uses. This means that if EOF is reached an attempt
16928 is made to pass two EOFs through the context and this causes the
16929 resulting error. This can also cause other problems as well. As is
16930 usual with these problems it takes *ages* to find and the fix is
16931 trivial: move one line.
16933 *Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer)*
16935 * Ugly workaround to get s_client and s_server working under Windows. The
16936 old code wouldn't work because it needed to select() on sockets and the
16937 tty (for keypresses and to see if data could be written). Win32 only
16938 supports select() on sockets so we select() with a 1s timeout on the
16939 sockets and then see if any characters are waiting to be read, if none
16940 are present then we retry, we also assume we can always write data to
16941 the tty. This isn't nice because the code then blocks until we've
16942 received a complete line of data and it is effectively polling the
16943 keyboard at 1s intervals: however it's quite a bit better than not
16944 working at all :-) A dedicated Windows application might handle this
16945 with an event loop for example.
16949 * Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
16950 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
16951 will be called when RSA_sign() and RSA_verify() are used. This is useful
16952 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
16953 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
16954 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
16955 This necessitated the support of an extra signature type NID_md5_sha1
16956 for SSL signatures and modifications to the SSL library to use it instead
16957 of calling RSA_public_decrypt() and RSA_private_encrypt().
16961 * Add new -verify -CAfile and -CApath options to the crl program, these
16962 will lookup a CRL issuers certificate and verify the signature in a
16963 similar way to the verify program. Tidy up the crl program so it
16964 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
16965 less strict. It will now permit CRL extensions even if it is not
16966 a V2 CRL: this will allow it to tolerate some broken CRLs.
16970 * Initialize all non-automatic variables each time one of the openssl
16971 sub-programs is started (this is necessary as they may be started
16972 multiple times from the "OpenSSL>" prompt).
16974 *Lennart Bang, Bodo Moeller*
16976 * Preliminary compilation option RSA_NULL which disables RSA crypto without
16977 removing all other RSA functionality (this is what NO_RSA does). This
16978 is so (for example) those in the US can disable those operations covered
16979 by the RSA patent while allowing storage and parsing of RSA keys and RSA
16984 * Non-copying interface to BIO pairs.
16985 (still largely untested)
16989 * New function ASN1_tag2str() to convert an ASN1 tag to a descriptive
16990 ASCII string. This was handled independently in various places before.
16994 * New functions UTF8_getc() and UTF8_putc() that parse and generate
16995 UTF8 strings a character at a time.
16999 * Use client_version from client hello to select the protocol
17000 (s23_srvr.c) and for RSA client key exchange verification
17001 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
17005 * Add various utility functions to handle SPKACs, these were previously
17006 handled by poking round in the structure internals. Added new function
17007 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
17008 print, verify and generate SPKACs. Based on an original idea from
17009 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
17013 * RIPEMD160 is operational on all platforms and is back in 'make test'.
17017 * Allow the config file extension section to be overwritten on the
17018 command line. Based on an original idea from Massimiliano Pala
17019 <madwolf@comune.modena.it>. The new option is called -extensions
17020 and can be applied to ca, req and x509. Also -reqexts to override
17021 the request extensions in req and -crlexts to override the crl extensions
17026 * Add new feature to the SPKAC handling in ca. Now you can include
17027 the same field multiple times by preceding it by "XXXX." for example:
17030 this is the same syntax as used in the req config file.
17034 * Allow certificate extensions to be added to certificate requests. These
17035 are specified in a 'req_extensions' option of the req section of the
17036 config file. They can be printed out with the -text option to req but
17037 are otherwise ignored at present.
17041 * Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
17042 data read consists of only the final block it would not decrypted because
17043 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
17044 A misplaced 'break' also meant the decrypted final block might not be
17045 copied until the next read.
17049 * Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
17050 a few extra parameters to the DH structure: these will be useful if
17051 for example we want the value of 'q' or implement X9.42 DH.
17055 * Initial support for DSA_METHOD. This is based on the RSA_METHOD and
17056 provides hooks that allow the default DSA functions or functions on a
17057 "per key" basis to be replaced. This allows hardware acceleration and
17058 hardware key storage to be handled without major modification to the
17059 library. Also added low-level modexp hooks and CRYPTO_EX structure and
17060 associated functions.
17064 * Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
17065 as "read only": it can't be written to and the buffer it points to will
17066 not be freed. Reading from a read only BIO is much more efficient than
17067 a normal memory BIO. This was added because there are several times when
17068 an area of memory needs to be read from a BIO. The previous method was
17069 to create a memory BIO and write the data to it, this results in two
17070 copies of the data and an O(n^2) reading algorithm. There is a new
17071 function BIO_new_mem_buf() which creates a read only memory BIO from
17072 an area of memory. Also modified the PKCS#7 routines to use read only
17077 * Bugfix: ssl23_get_client_hello did not work properly when called in
17078 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
17079 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
17080 but a retry condition occurred while trying to read the rest.
17084 * The PKCS7_ENC_CONTENT_new() function was setting the content type as
17085 NID_pkcs7_encrypted by default: this was wrong since this should almost
17086 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
17087 the encrypted data type: this is a more sensible place to put it and it
17088 allows the PKCS#12 code to be tidied up that duplicated this
17093 * Changed obj_dat.pl script so it takes its input and output files on
17094 the command line. This should avoid shell escape redirection problems
17099 * Initial support for certificate extension requests, these are included
17100 in things like Xenroll certificate requests. Included functions to allow
17101 extensions to be obtained and added.
17105 * -crlf option to s_client and s_server for sending newlines as
17106 CRLF (as required by many protocols).
17110 ### Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
17112 * Install libRSAglue.a when OpenSSL is built with RSAref.
17114 *Ralf S. Engelschall*
17116 * A few more `#ifndef NO_FP_API / #endif` pairs for consistency.
17118 *Andrija Antonijevic <TheAntony2@bigfoot.com>*
17120 * Fix -startdate and -enddate (which was missing) arguments to 'ca'
17125 * New function DSA_dup_DH, which duplicates DSA parameters/keys as
17126 DH parameters/keys (q is lost during that conversion, but the resulting
17127 DH parameters contain its length).
17129 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
17130 much faster than DH_generate_parameters (which creates parameters
17131 where `p = 2*q + 1`), and also the smaller q makes DH computations
17132 much more efficient (160-bit exponentiation instead of 1024-bit
17133 exponentiation); so this provides a convenient way to support DHE
17134 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
17135 utter importance to use
17136 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
17138 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
17139 when such DH parameters are used, because otherwise small subgroup
17140 attacks may become possible!
17144 * Avoid memory leak in i2d_DHparams.
17148 * Allow the -k option to be used more than once in the enc program:
17149 this allows the same encrypted message to be read by multiple recipients.
17153 * New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
17154 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
17155 it will always use the numerical form of the OID, even if it has a short
17160 * Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
17161 method only got called if p,q,dmp1,dmq1,iqmp components were present,
17162 otherwise bn_mod_exp was called. In the case of hardware keys for example
17163 no private key components need be present and it might store extra data
17164 in the RSA structure, which cannot be accessed from bn_mod_exp.
17165 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
17166 private key operations.
17170 * Added support for SPARC Linux.
17174 * pem_password_cb function type incompatibly changed from
17175 typedef int pem_password_cb(char *buf, int size, int rwflag);
17177 ....(char *buf, int size, int rwflag, void *userdata);
17178 so that applications can pass data to their callbacks:
17179 The `PEM[_ASN1]_{read,write}...` functions and macros now take an
17180 additional void * argument, which is just handed through whenever
17181 the password callback is called.
17183 *Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller*
17185 New function SSL_CTX_set_default_passwd_cb_userdata.
17187 Compatibility note: As many C implementations push function arguments
17188 onto the stack in reverse order, the new library version is likely to
17189 interoperate with programs that have been compiled with the old
17190 pem_password_cb definition (PEM_whatever takes some data that
17191 happens to be on the stack as its last argument, and the callback
17192 just ignores this garbage); but there is no guarantee whatsoever that
17195 * The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
17196 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
17197 problems not only on Windows, but also on some Unix platforms.
17198 To avoid problematic command lines, these definitions are now in an
17199 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
17200 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
17204 * MIPS III/IV assembler module is reimplemented.
17208 * More DES library cleanups: remove references to srand/rand and
17209 delete an unused file.
17213 * Add support for the free Netwide assembler (NASM) under Win32,
17214 since not many people have MASM (ml) and it can be hard to obtain.
17215 This is currently experimental but it seems to work OK and pass all
17216 the tests. Check out INSTALL.W32 for info.
17220 * Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
17221 without temporary keys kept an extra copy of the server key,
17222 and connections with temporary keys did not free everything in case
17227 * New function RSA_check_key and new openssl rsa option -check
17228 for verifying the consistency of RSA keys.
17230 *Ulf Moeller, Bodo Moeller*
17232 * Various changes to make Win32 compile work:
17233 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
17234 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
17235 comparison" warnings.
17236 3. Add `sk_<TYPE>_sort` to DEF file generator and do make update.
17240 * Add a debugging option to PKCS#5 v2 key generation function: when
17241 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
17242 derived keys are printed to stderr.
17246 * Copy the flags in ASN1_STRING_dup().
17248 *Roman E. Pavlov <pre@mo.msk.ru>*
17250 * The x509 application mishandled signing requests containing DSA
17251 keys when the signing key was also DSA and the parameters didn't match.
17253 It was supposed to omit the parameters when they matched the signing key:
17254 the verifying software was then supposed to automatically use the CA's
17255 parameters if they were absent from the end user certificate.
17257 Omitting parameters is no longer recommended. The test was also
17258 the wrong way round! This was probably due to unusual behaviour in
17259 EVP_cmp_parameters() which returns 1 if the parameters match.
17260 This meant that parameters were omitted when they *didn't* match and
17261 the certificate was useless. Certificates signed with 'ca' didn't have
17264 *Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>*
17266 * Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
17267 The interface is as follows:
17268 Applications can use
17269 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
17270 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
17271 "off" is now the default.
17272 The library internally uses
17273 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
17274 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
17275 to disable memory-checking temporarily.
17277 Some inconsistent states that previously were possible (and were
17278 even the default) are now avoided.
17280 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
17281 with each memory chunk allocated; this is occasionally more helpful
17282 than just having a counter.
17284 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
17286 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
17291 * Introduce "mode" for SSL structures (with defaults in SSL_CTX),
17292 which largely parallels "options", but is for changing API behaviour,
17293 whereas "options" are about protocol behaviour.
17294 Initial "mode" flags are:
17296 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
17297 a single record has been written.
17298 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
17299 retries use the same buffer location.
17300 (But all of the contents must be
17305 * Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
17308 * Fix problems with no-hmac etc.
17310 *Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>*
17312 * New functions RSA_get_default_method(), RSA_set_method() and
17313 RSA_get_method(). These allows replacement of RSA_METHODs without having
17314 to mess around with the internals of an RSA structure.
17318 * Fix memory leaks in DSA_do_sign and DSA_is_prime.
17319 Also really enable memory leak checks in openssl.c and in some
17322 *Chad C. Mulligan, Bodo Moeller*
17324 * Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
17325 up the length of negative integers. This has now been simplified to just
17326 store the length when it is first determined and use it later, rather
17327 than trying to keep track of where data is copied and updating it to
17329 *Steve Henson, reported by Brien Wheeler <bwheeler@authentica-security.com>*
17331 * Add a new function PKCS7_signatureVerify. This allows the verification
17332 of a PKCS#7 signature but with the signing certificate passed to the
17333 function itself. This contrasts with PKCS7_dataVerify which assumes the
17334 certificate is present in the PKCS#7 structure. This isn't always the
17335 case: certificates can be omitted from a PKCS#7 structure and be
17336 distributed by "out of band" means (such as a certificate database).
17340 * Complete the `PEM_*` macros with DECLARE_PEM versions to replace the
17341 function prototypes in pem.h, also change util/mkdef.pl to add the
17342 necessary function names.
17346 * mk1mf.pl (used by Windows builds) did not properly read the
17347 options set by Configure in the top level Makefile, and Configure
17348 was not even able to write more than one option correctly.
17349 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
17353 * New functions CONF_load_bio() and CONF_load_fp() to allow a config
17354 file to be loaded from a BIO or FILE pointer. The BIO version will
17355 for example allow memory BIOs to contain config info.
17359 * New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
17360 Whoever hopes to achieve shared-library compatibility across versions
17361 must use this, not the compile-time macro.
17362 (Exercise 0.9.4: Which is the minimum library version required by
17364 Note: All this applies only to multi-threaded programs, others don't
17369 * Add missing case to s3_clnt.c state machine -- one of the new SSL tests
17370 through a BIO pair triggered the default case, i.e.
17371 SSLerr(...,SSL_R_UNKNOWN_STATE).
17375 * New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
17376 can use the SSL library even if none of the specific BIOs is
17381 * Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
17382 for the encoded length.
17384 *Jeon KyoungHo <khjeon@sds.samsung.co.kr>*
17386 * Add initial documentation of the X509V3 functions.
17390 * Add a new pair of functions PEM_write_PKCS8PrivateKey() and
17391 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
17392 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
17393 secure PKCS#8 private key format with a high iteration count.
17397 * Fix determination of Perl interpreter: A perl or perl5
17398 *directory* in $PATH was also accepted as the interpreter.
17400 *Ralf S. Engelschall*
17402 * Fix demos/sign/sign.c: well there wasn't anything strictly speaking
17403 wrong with it but it was very old and did things like calling
17404 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
17405 unusual formatting.
17409 * Fix demos/selfsign.c: it used obsolete and deleted functions, changed
17410 to use the new extension code.
17414 * Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
17415 with macros. This should make it easier to change their form, add extra
17416 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
17421 * Add to configuration table a new entry that can specify an alternative
17422 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
17423 according to Mark Crispin <MRC@Panda.COM>.
17427 * DES CBC did not update the IV. Weird.
17431 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
17432 Changing the behaviour of the former might break existing programs --
17433 where IV updating is needed, des_ncbc_encrypt can be used.
17436 * When bntest is run from "make test" it drives bc to check its
17437 calculations, as well as internally checking them. If an internal check
17438 fails, it needs to cause bc to give a non-zero result or make test carries
17439 on without noticing the failure. Fixed.
17443 * DES library cleanups.
17447 * Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
17448 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
17449 ciphers. NOTE: although the key derivation function has been verified
17450 against some published test vectors it has not been extensively tested
17451 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
17456 * Instead of "mkdir -p", which is not fully portable, use new
17457 Perl script "util/mkdir-p.pl".
17461 * Rewrite the way password based encryption (PBE) is handled. It used to
17462 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
17463 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
17464 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
17465 the 'parameter' field of the AlgorithmIdentifier is passed to the
17466 underlying key generation function so it must do its own ASN1 parsing.
17467 This has also changed the EVP_PBE_CipherInit() function which now has a
17468 'parameter' argument instead of literal salt and iteration count values
17469 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
17473 * Support for PKCS#5 v1.5 compatible password based encryption algorithms
17474 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
17475 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
17476 KEY" because this clashed with PKCS#8 unencrypted string. Since this
17477 value was just used as a "magic string" and not used directly its
17478 value doesn't matter.
17482 * Introduce some semblance of const correctness to BN. Shame C doesn't
17487 * "linux-sparc64" configuration (ultrapenguin).
17489 *Ray Miller <ray.miller@oucs.ox.ac.uk>*
17490 "linux-sparc" configuration.
17492 *Christian Forster <fo@hawo.stw.uni-erlangen.de>*
17494 * config now generates no-xxx options for missing ciphers.
17498 * Support the EBCDIC character set (work in progress).
17499 File ebcdic.c not yet included because it has a different license.
17501 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
17503 * Support BS2000/OSD-POSIX.
17505 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
17507 * Make callbacks for key generation use `void *` instead of `char *`.
17511 * Make S/MIME samples compile (not yet tested).
17515 * Additional typesafe stacks.
17519 * New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
17523 ### Changes between 0.9.3 and 0.9.3a [29 May 1999]
17525 * New configuration variant "sco5-gcc".
17527 * Updated some demos.
17529 *Sean O Riordain, Wade Scholine*
17531 * Add missing BIO_free at exit of pkcs12 application.
17535 * Fix memory leak in conf.c.
17539 * Updates for Win32 to assembler version of MD5.
17543 * Set #! path to perl in `apps/der_chop` to where we found it
17544 instead of using a fixed path.
17548 * SHA library changes for irix64-mips4-cc.
17552 * Improvements for VMS support.
17556 ### Changes between 0.9.2b and 0.9.3 [24 May 1999]
17558 * Bignum library bug fix. IRIX 6 passes "make test" now!
17559 This also avoids the problems with SC4.2 and unpatched SC5.
17561 *Andy Polyakov <appro@fy.chalmers.se>*
17563 * New functions sk_num, sk_value and sk_set to replace the previous macros.
17564 These are required because of the typesafe stack would otherwise break
17565 existing code. If old code used a structure member which used to be STACK
17566 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
17567 sk_num or sk_value it would produce an error because the num, data members
17568 are not present in STACK_OF. Now it just produces a warning. sk_set
17569 replaces the old method of assigning a value to sk_value
17570 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
17571 that does this will no longer work (and should use sk_set instead) but
17572 this could be regarded as a "questionable" behaviour anyway.
17576 * Fix most of the other PKCS#7 bugs. The "experimental" code can now
17577 correctly handle encrypted S/MIME data.
17581 * Change type of various DES function arguments from des_cblock
17582 (which means, in function argument declarations, pointer to char)
17583 to des_cblock * (meaning pointer to array with 8 char elements),
17584 which allows the compiler to do more typechecking; it was like
17585 that back in SSLeay, but with lots of ugly casts.
17587 Introduce new type const_des_cblock.
17591 * Reorganise the PKCS#7 library and get rid of some of the more obvious
17592 problems: find RecipientInfo structure that matches recipient certificate
17593 and initialise the ASN1 structures properly based on passed cipher.
17597 * Belatedly make the BN tests actually check the results.
17601 * Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
17602 to and from BNs: it was completely broken. New compilation option
17603 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
17604 key elements as negative integers.
17608 * Reorganize and speed up MD5.
17610 *Andy Polyakov <appro@fy.chalmers.se>*
17614 *Richard Levitte <richard@levitte.org>*
17616 * New option -out to asn1parse to allow the parsed structure to be
17617 output to a file. This is most useful when combined with the -strparse
17618 option to examine the output of things like OCTET STRINGS.
17622 * Make SSL library a little more fool-proof by not requiring any longer
17623 that `SSL_set_{accept,connect}_state` be called before
17624 `SSL_{accept,connect}` may be used (`SSL_set_..._state` is omitted
17625 in many applications because usually everything *appeared* to work as
17626 intended anyway -- now it really works as intended).
17630 * Move openssl.cnf out of lib/.
17634 * Fix various things to let OpenSSL even pass "egcc -pipe -O2 -Wall
17635 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
17636 -Wmissing-declarations -Wnested-externs -Winline" with EGCS 1.1.2+
17638 *Ralf S. Engelschall*
17640 * Various fixes to the EVP and PKCS#7 code. It may now be able to
17641 handle PKCS#7 enveloped data properly.
17643 *Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve*
17645 * Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
17646 copying pointers. The cert_st handling is changed by this in
17647 various ways (and thus what used to be known as ctx->default_cert
17648 is now called ctx->cert, since we don't resort to `s->ctx->[default_]cert`
17649 any longer when s->cert does not give us what we need).
17650 ssl_cert_instantiate becomes obsolete by this change.
17651 As soon as we've got the new code right (possibly it already is?),
17652 we have solved a couple of bugs of the earlier code where s->cert
17653 was used as if it could not have been shared with other SSL structures.
17655 Note that using the SSL API in certain dirty ways now will result
17656 in different behaviour than observed with earlier library versions:
17657 Changing settings for an `SSL_CTX *ctx` after having done s = SSL_new(ctx)
17658 does not influence s as it used to.
17660 In order to clean up things more thoroughly, inside SSL_SESSION
17661 we don't use CERT any longer, but a new structure SESS_CERT
17662 that holds per-session data (if available); currently, this is
17663 the peer's certificate chain and, for clients, the server's certificate
17664 and temporary key. CERT holds only those values that can have
17665 meaningful defaults in an SSL_CTX.
17669 * New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
17670 from the internal representation. Various PKCS#7 fixes: remove some
17671 evil casts and set the enc_dig_alg field properly based on the signing
17676 * Allow PKCS#12 password to be set from the command line or the
17677 environment. Let 'ca' get its config file name from the environment
17678 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
17683 * Allow certificate policies extension to use an IA5STRING for the
17684 organization field. This is contrary to the PKIX definition but
17685 VeriSign uses it and IE5 only recognises this form. Document 'x509'
17690 * Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
17691 without disallowing inline assembler and the like for non-pedantic builds.
17695 * Support Borland C++ builder.
17697 *Janez Jere <jj@void.si>, modified by Ulf Möller*
17703 * SHA-1 cleanups and performance enhancements.
17705 *Andy Polyakov <appro@fy.chalmers.se>*
17707 * Sparc v8plus assembler for the bignum library.
17709 *Andy Polyakov <appro@fy.chalmers.se>*
17711 * Accept any -xxx and +xxx compiler options in Configure.
17715 * Update HPUX configuration.
17719 * Add missing `sk_<type>_unshift()` function to safestack.h
17721 *Ralf S. Engelschall*
17723 * New function SSL_CTX_use_certificate_chain_file that sets the
17724 "extra_cert"s in addition to the certificate. (This makes sense
17725 only for "PEM" format files, as chains as a whole are not
17730 * Support verify_depth from the SSL API.
17731 x509_vfy.c had what can be considered an off-by-one-error:
17732 Its depth (which was not part of the external interface)
17733 was actually counting the number of certificates in a chain;
17734 now it really counts the depth.
17738 * Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
17739 instead of X509err, which often resulted in confusing error
17740 messages since the error codes are not globally unique
17741 (e.g. an alleged error in ssl3_accept when a certificate
17742 didn't match the private key).
17744 * New function SSL_CTX_set_session_id_context that allows to set a default
17745 value (so that you don't need SSL_set_session_id_context for each
17746 connection using the SSL_CTX).
17750 * OAEP decoding bug fix.
17754 * Support INSTALL_PREFIX for package builders, as proposed by
17759 * New Configure options "threads" and "no-threads". For systems
17760 where the proper compiler options are known (currently Solaris
17761 and Linux), "threads" is the default.
17765 * New script util/mklink.pl as a faster substitute for util/mklink.sh.
17769 * Install various scripts to $(OPENSSLDIR)/misc, not to
17770 $(INSTALLTOP)/bin -- they shouldn't clutter directories
17771 such as /usr/local/bin.
17775 * "make linux-shared" to build shared libraries.
17777 *Niels Poppe <niels@netbox.org>*
17779 * New Configure option `no-<cipher>` (rsa, idea, rc5, ...).
17783 * Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
17784 extension adding in x509 utility.
17788 * Remove NOPROTO sections and error code comments.
17792 * Partial rewrite of the DEF file generator to now parse the ANSI
17797 * New Configure options --prefix=DIR and --openssldir=DIR.
17801 * Complete rewrite of the error code script(s). It is all now handled
17802 by one script at the top level which handles error code gathering,
17803 header rewriting and C source file generation. It should be much better
17804 than the old method: it now uses a modified version of Ulf's parser to
17805 read the ANSI prototypes in all header files (thus the old K&R definitions
17806 aren't needed for error creation any more) and do a better job of
17807 translating function codes into names. The old 'ASN1 error code embedded
17808 in a comment' is no longer necessary and it doesn't use .err files which
17809 have now been deleted. Also the error code call doesn't have to appear all
17810 on one line (which resulted in some large lines...).
17814 * Change #include filenames from `<foo.h>` to `<openssl/foo.h>`.
17818 * Change behaviour of ssl2_read when facing length-0 packets: Don't return
17819 0 (which usually indicates a closed connection), but continue reading.
17823 * Fix some race conditions.
17827 * Add support for CRL distribution points extension. Add Certificate
17828 Policies and CRL distribution points documentation.
17832 * Move the autogenerated header file parts to crypto/opensslconf.h.
17836 * Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
17837 8 of keying material. Merlin has also confirmed interop with this fix
17838 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
17840 *Merlin Hughes <merlin@baltimore.ie>*
17842 * Fix lots of warnings.
17844 *Richard Levitte <levitte@stacken.kth.se>*
17846 * In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
17847 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
17849 *Richard Levitte <levitte@stacken.kth.se>*
17851 * Fix problems with sizeof(long) == 8.
17853 *Andy Polyakov <appro@fy.chalmers.se>*
17855 * Change functions to ANSI C.
17859 * Fix typos in error codes.
17861 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller*
17863 * Remove defunct assembler files from Configure.
17867 * SPARC v8 assembler BIGNUM implementation.
17869 *Andy Polyakov <appro@fy.chalmers.se>*
17871 * Support for Certificate Policies extension: both print and set.
17872 Various additions to support the r2i method this uses.
17876 * A lot of constification, and fix a bug in X509_NAME_oneline() that could
17877 return a const string when you are expecting an allocated buffer.
17881 * Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
17882 types DirectoryString and DisplayText.
17886 * Add code to allow r2i extensions to access the configuration database,
17887 add an LHASH database driver and add several ctx helper functions.
17891 * Fix an evil bug in bn_expand2() which caused various BN functions to
17892 fail when they extended the size of a BIGNUM.
17896 * Various utility functions to handle SXNet extension. Modify mkdef.pl to
17897 support typesafe stack.
17901 * Fix typo in SSL_[gs]et_options().
17903 *Nils Frostberg <nils@medcom.se>*
17905 * Delete various functions and files that belonged to the (now obsolete)
17906 old X509V3 handling code.
17910 * New Configure option "rsaref".
17914 * Don't auto-generate pem.h.
17918 * Introduce type-safe ASN.1 SETs.
17922 * Convert various additional casted stacks to type-safe STACK_OF() variants.
17924 *Ben Laurie, Ralf S. Engelschall, Steve Henson*
17926 * Introduce type-safe STACKs. This will almost certainly break lots of code
17927 that links with OpenSSL (well at least cause lots of warnings), but fear
17928 not: the conversion is trivial, and it eliminates loads of evil casts. A
17929 few STACKed things have been converted already. Feel free to convert more.
17930 In the fullness of time, I'll do away with the STACK type altogether.
17934 * Add `openssl ca -revoke <certfile>` facility which revokes a certificate
17935 specified in `<certfile>` by updating the entry in the index.txt file.
17936 This way one no longer has to edit the index.txt file manually for
17937 revoking a certificate. The -revoke option does the gory details now.
17939 *Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall*
17941 * Fix `openssl crl -noout -text` combination where `-noout` killed the
17942 `-text` option at all and this way the `-noout -text` combination was
17943 inconsistent in `openssl crl` with the friends in `openssl x509|rsa|dsa`.
17945 *Ralf S. Engelschall*
17947 * Make sure a corresponding plain text error message exists for the
17948 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
17949 verify callback function determined that a certificate was revoked.
17951 *Ralf S. Engelschall*
17953 * Bugfix: In test/testenc, don't test `openssl <cipher>` for
17954 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
17955 all available ciphers including rc5, which was forgotten until now.
17956 In order to let the testing shell script know which algorithms
17957 are available, a new (up to now undocumented) command
17958 `openssl list-cipher-commands` is used.
17962 * Bugfix: s_client occasionally would sleep in select() when
17963 it should have checked SSL_pending() first.
17967 * New functions DSA_do_sign and DSA_do_verify to provide access to
17968 the raw DSA values prior to ASN.1 encoding.
17972 * Tweaks to Configure
17974 *Niels Poppe <niels@netbox.org>*
17976 * Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
17981 * New variables $(RANLIB) and $(PERL) in the Makefiles.
17985 * New config option to avoid instructions that are illegal on the 80386.
17986 The default code is faster, but requires at least a 486.
17990 * Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
17991 SSL2_SERVER_VERSION (not used at all) macros, which are now the
17992 same as SSL2_VERSION anyway.
17996 * New "-showcerts" option for s_client.
18000 * Still more PKCS#12 integration. Add pkcs12 application to openssl
18001 application. Various cleanups and fixes.
18005 * More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
18006 modify error routines to work internally. Add error codes and PBE init
18007 to library startup routines.
18011 * Further PKCS#12 integration. Added password based encryption, PKCS#8 and
18012 packing functions to asn1 and evp. Changed function names and error
18013 codes along the way.
18017 * PKCS12 integration: and so it begins... First of several patches to
18018 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
18019 objects to objects.h
18023 * Add a new 'indent' option to some X509V3 extension code. Initial ASN1
18024 and display support for Thawte strong extranet extension.
18028 * Add LinuxPPC support.
18030 *Jeff Dubrule <igor@pobox.org>*
18032 * Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
18033 bn_div_words in alpha.s.
18035 *Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie*
18037 * Make sure the RSA OAEP test is skipped under -DRSAref because
18038 OAEP isn't supported when OpenSSL is built with RSAref.
18040 *Ulf Moeller <ulf@fitug.de>*
18042 * Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
18043 so they no longer are missing under -DNOPROTO.
18045 *Soren S. Jorvang <soren@t.dk>*
18047 ### Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
18049 * Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
18050 doesn't work when the session is reused. Coming soon!
18054 * Fix a security hole, that allows sessions to be reused in the wrong
18055 context thus bypassing client cert protection! All software that uses
18056 client certs and session caches in multiple contexts NEEDS PATCHING to
18057 allow session reuse! A fuller solution is in the works.
18059 *Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)*
18061 * Some more source tree cleanups (removed obsolete files
18062 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
18063 permission on "config" script to be executable) and a fix for the INSTALL
18066 *Ulf Moeller <ulf@fitug.de>*
18068 * Remove some legacy and erroneous uses of malloc, free instead of
18071 *Lennart Bang <lob@netstream.se>, with minor changes by Steve*
18073 * Make rsa_oaep_test return non-zero on error.
18075 *Ulf Moeller <ulf@fitug.de>*
18077 * Add support for native Solaris shared libraries. Configure
18078 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
18079 if someone would make that last step automatic.
18081 *Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>*
18083 * ctx_size was not built with the right compiler during "make links". Fixed.
18087 * Change the meaning of 'ALL' in the cipher list. It now means "everything
18088 except NULL ciphers". This means the default cipher list will no longer
18089 enable NULL ciphers. They need to be specifically enabled e.g. with
18090 the string "DEFAULT:eNULL".
18094 * Fix to RSA private encryption routines: if p < q then it would
18095 occasionally produce an invalid result. This will only happen with
18096 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
18100 * Be less restrictive and allow also `perl util/perlpath.pl
18101 /path/to/bin/perl` in addition to `perl util/perlpath.pl /path/to/bin`,
18102 because this way one can also use an interpreter named `perl5` (which is
18103 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
18104 installed as `perl`).
18106 *Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
18108 * Let util/clean-depend.pl work also with older Perl 5.00x versions.
18110 *Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
18112 * Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
18113 advapi32.lib to Win32 build and change the pem test comparison
18114 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
18115 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
18116 and crypto/des/ede_cbcm_enc.c.
18120 * DES quad checksum was broken on big-endian architectures. Fixed.
18124 * Comment out two functions in bio.h that aren't implemented. Fix up the
18125 Win32 test batch file so it (might) work again. The Win32 test batch file
18126 is horrible: I feel ill....
18130 * Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
18131 in e_os.h. Audit of header files to check ANSI and non ANSI
18132 sections: 10 functions were absent from non ANSI section and not exported
18133 from Windows DLLs. Fixed up libeay.num for new functions.
18137 * Make `openssl version` output lines consistent.
18139 *Ralf S. Engelschall*
18141 * Fix Win32 symbol export lists for BIO functions: Added
18142 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
18143 to ms/libeay{16,32}.def.
18145 *Ralf S. Engelschall*
18147 * Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
18148 fine under Unix and passes some trivial tests I've now added. But the
18149 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
18150 added to make sure no one expects that this stuff really works in the
18151 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
18152 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
18155 *Ralf S. Engelschall*
18157 * Fix the generation of two part addresses in perl.
18159 *Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie*
18161 * Add config entry for Linux on MIPS.
18163 *John Tobey <jtobey@channel1.com>*
18165 * Make links whenever Configure is run, unless we are on Windoze.
18169 * Permit extensions to be added to CRLs using crl_section in openssl.cnf.
18170 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
18175 * Add a useful kludge to allow package maintainers to specify compiler and
18176 other platforms details on the command line without having to patch the
18177 Configure script every time: One now can use
18178 `perl Configure <id>:<details>`,
18179 i.e. platform ids are allowed to have details appended
18180 to them (separated by colons). This is treated as there would be a static
18181 pre-configured entry in Configure's %table under key `<id>` with value
18182 `<details>` and `perl Configure <id>` is called. So, when you want to
18183 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
18184 assembler stuff you can use `perl Configure "FreeBSD-elf:pgcc:-O6:::"`
18185 now, which overrides the FreeBSD-elf entry on-the-fly.
18187 *Ralf S. Engelschall*
18189 * Disable new TLS1 ciphersuites by default: they aren't official yet.
18193 * Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
18194 on the `perl Configure ...` command line. This way one can compile
18195 OpenSSL libraries with Position Independent Code (PIC) which is needed
18196 for linking it into DSOs.
18198 *Ralf S. Engelschall*
18200 * Remarkably, export ciphers were totally broken and no-one had noticed!
18205 * Cleaned up the LICENSE document: The official contact for any license
18206 questions now is the OpenSSL core team under openssl-core@openssl.org.
18207 And add a paragraph about the dual-license situation to make sure people
18208 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
18209 to the OpenSSL toolkit.
18211 *Ralf S. Engelschall*
18213 * General source tree makefile cleanups: Made `making xxx in yyy...`
18214 display consistent in the source tree and replaced `/bin/rm` by `rm`.
18215 Additionally cleaned up the `make links` target: Remove unnecessary
18216 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
18217 to speed processing and no longer clutter the display with confusing
18218 stuff. Instead only the actually done links are displayed.
18220 *Ralf S. Engelschall*
18222 * Permit null encryption ciphersuites, used for authentication only. It used
18223 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
18224 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
18229 * Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
18230 signed attributes when verifying signatures (this would break them),
18231 the detached data encoding was wrong and public keys obtained using
18232 X509_get_pubkey() weren't freed.
18236 * Add text documentation for the BUFFER functions. Also added a work around
18237 to a Win95 console bug. This was triggered by the password read stuff: the
18238 last character typed gets carried over to the next fread(). If you were
18239 generating a new cert request using 'req' for example then the last
18240 character of the passphrase would be CR which would then enter the first
18245 * Added the new 'Includes OpenSSL Cryptography Software' button as
18246 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
18247 button and can be used by applications based on OpenSSL to show the
18248 relationship to the OpenSSL project.
18250 *Ralf S. Engelschall*
18252 * Remove confusing variables in function signatures in files
18253 ssl/ssl_lib.c and ssl/ssl.h.
18255 *Lennart Bong <lob@kulthea.stacken.kth.se>*
18257 * Don't install bss_file.c under PREFIX/include/
18259 *Lennart Bong <lob@kulthea.stacken.kth.se>*
18261 * Get the Win32 compile working again. Modify mkdef.pl so it can handle
18262 functions that return function pointers and has support for NT specific
18263 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
18264 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
18265 unsigned to signed types: this was killing the Win32 compile.
18269 * Add new certificate file to stack functions,
18270 SSL_add_dir_cert_subjects_to_stack() and
18271 SSL_add_file_cert_subjects_to_stack(). These largely supplant
18272 SSL_load_client_CA_file(), and can be used to add multiple certs easily
18273 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
18274 This means that Apache-SSL and similar packages don't have to mess around
18275 to add as many CAs as they want to the preferred list.
18279 * Experiment with doxygen documentation. Currently only partially applied to
18281 See <http://www.stack.nl/~dimitri/doxygen/index.html>, and run doxygen with
18282 openssl.doxy as the configuration file.
18286 * Get rid of remaining C++-style comments which strict C compilers hate.
18288 *Ralf S. Engelschall, pointed out by Carlos Amengual*
18290 * Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
18291 compiled in by default: it has problems with large keys.
18295 * Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
18296 DH private keys and/or callback functions which directly correspond to
18297 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
18298 is needed for applications which have to configure certificates on a
18299 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
18301 For the RSA certificate situation is makes no difference, but
18302 for the DSA certificate situation this fixes the "no shared cipher"
18303 problem where the OpenSSL cipher selection procedure failed because the
18304 temporary keys were not overtaken from the context and the API provided
18305 no way to reconfigure them.
18306 The new functions now let applications reconfigure the stuff and they
18307 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
18308 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
18309 non-public-API function ssl_cert_instantiate() is used as a helper
18310 function and also to reduce code redundancy inside ssl_rsa.c.
18312 *Ralf S. Engelschall*
18314 * Move s_server -dcert and -dkey options out of the undocumented feature
18315 area because they are useful for the DSA situation and should be
18316 recognized by the users.
18318 *Ralf S. Engelschall*
18320 * Fix the cipher decision scheme for export ciphers: the export bits are
18321 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
18322 SSL_EXP_MASK. So, the original variable has to be used instead of the
18323 already masked variable.
18325 *Richard Levitte <levitte@stacken.kth.se>*
18327 * Fix `port` variable from `int` to `unsigned int` in crypto/bio/b_sock.c
18329 *Richard Levitte <levitte@stacken.kth.se>*
18331 * Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
18332 from `int` to `unsigned int` because it is a length and initialized by
18333 EVP_DigestFinal() which expects an `unsigned int *`.
18335 *Richard Levitte <levitte@stacken.kth.se>*
18337 * Don't hard-code path to Perl interpreter on shebang line of Configure
18338 script. Instead use the usual Shell->Perl transition trick.
18340 *Ralf S. Engelschall*
18342 * Make `openssl x509 -noout -modulus`' functional also for DSA certificates
18343 (in addition to RSA certificates) to match the behaviour of `openssl dsa
18344 -noout -modulus` as it's already the case for `openssl rsa -noout
18345 -modulus`. For RSA the -modulus is the real "modulus" while for DSA
18346 currently the public key is printed (a decision which was already done by
18347 `openssl dsa -modulus` in the past) which serves a similar purpose.
18348 Additionally the NO_RSA no longer completely removes the whole -modulus
18349 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
18352 *Ralf S. Engelschall*
18354 * Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
18355 BIO. See the source (crypto/evp/bio_ok.c) for more info.
18357 *Arne Ansper <arne@ats.cyber.ee>*
18359 * Dump the old yucky req code that tried (and failed) to allow raw OIDs
18360 to be added. Now both 'req' and 'ca' can use new objects defined in the
18365 * Add cool BIO that does syslog (or event log on NT).
18367 *Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie*
18369 * Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
18370 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
18371 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
18372 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
18376 * Add preliminary config info for new extension code.
18380 * Make RSA_NO_PADDING really use no padding.
18382 *Ulf Moeller <ulf@fitug.de>*
18384 * Generate errors when private/public key check is done.
18388 * Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
18389 for some CRL extensions and new objects added.
18393 * Really fix the ASN1 IMPLICIT bug this time... Partial support for private
18394 key usage extension and fuller support for authority key id.
18398 * Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
18399 padding method for RSA, which is recommended for new applications in PKCS
18400 #1 v2.0 (RFC 2437, October 1998).
18401 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
18402 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
18403 against Bleichbacher's attack on RSA.
18404 *Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
18407 * Updates to the new SSL compression code
18409 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
18411 * Fix so that the version number in the master secret, when passed
18412 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
18413 (because the server will not accept higher), that the version number
18414 is 0x03,0x01, not 0x03,0x00
18416 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
18418 * Run extensive memory leak checks on SSL commands. Fixed *lots* of memory
18419 leaks in `ssl/` relating to new `X509_get_pubkey()` behaviour. Also fixes
18420 in `apps/` and an unrelated leak in `crypto/dsa/dsa_vrf.c`.
18424 * Support for RAW extensions where an arbitrary extension can be
18425 created by including its DER encoding. See `apps/openssl.cnf` for
18430 * Make sure latest Perl versions don't interpret some generated C array
18431 code as Perl array code in the crypto/err/err_genc.pl script.
18433 *Lars Weber <3weber@informatik.uni-hamburg.de>*
18435 * Modify ms/do_ms.bat to not generate assembly language makefiles since
18436 not many people have the assembler. Various Win32 compilation fixes and
18437 update to the INSTALL.W32 file with (hopefully) more accurate Win32
18438 build instructions.
18442 * Modify configure script 'Configure' to automatically create crypto/date.h
18443 file under Win32 and also build pem.h from pem.org. New script
18444 util/mkfiles.pl to create the MINFO file on environments that can't do a
18445 'make files': perl util/mkfiles.pl >MINFO should work.
18449 * Major rework of DES function declarations, in the pursuit of correctness
18450 and purity. As a result, many evil casts evaporated, and some weirdness,
18451 too. You may find this causes warnings in your code. Zapping your evil
18452 casts will probably fix them. Mostly.
18456 * Fix for a typo in asn1.h. Bug fix to object creation script
18457 obj_dat.pl. It considered a zero in an object definition to mean
18458 "end of object": none of the objects in objects.h have any zeros
18459 so it wasn't spotted.
18461 *Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>*
18463 * Add support for Triple DES Cipher Block Chaining with Output Feedback
18464 Masking (CBCM). In the absence of test vectors, the best I have been able
18465 to do is check that the decrypt undoes the encrypt, so far. Send me test
18466 vectors if you have them.
18470 * Correct calculation of key length for export ciphers (too much space was
18471 allocated for null ciphers). This has not been tested!
18475 * Modifications to the mkdef.pl for Win32 DEF file creation. The usage
18476 message is now correct (it understands "crypto" and "ssl" on its
18477 command line). There is also now an "update" option. This will update
18478 the util/ssleay.num and util/libeay.num files with any new functions.
18480 perl util/mkdef.pl crypto ssl update
18481 it will update them.
18485 * Overhauled the Perl interface:
18486 - ported BN stuff to OpenSSL's different BN library
18487 - made the perl/ source tree CVS-aware
18488 - renamed the package from SSLeay to OpenSSL (the files still contain
18489 their history because I've copied them in the repository)
18490 - removed obsolete files (the test scripts will be replaced
18491 by better Test::Harness variants in the future)
18493 *Ralf S. Engelschall*
18495 * First cut for a very conservative source tree cleanup:
18496 1. merge various obsolete readme texts into doc/ssleay.txt
18497 where we collect the old documents and readme texts.
18498 2. remove the first part of files where I'm already sure that we no
18499 longer need them because of three reasons: either they are just temporary
18500 files which were left by Eric or they are preserved original files where
18501 I've verified that the diff is also available in the CVS via "cvs diff
18502 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
18503 the crypto/md/ stuff).
18505 *Ralf S. Engelschall*
18507 * More extension code. Incomplete support for subject and issuer alt
18508 name, issuer and authority key id. Change the i2v function parameters
18509 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
18510 what that's for :-) Fix to ASN1 macro which messed up
18511 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
18515 * Preliminary support for ENUMERATED type. This is largely copied from the
18520 * Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
18522 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
18524 * Make sure `make rehash` target really finds the `openssl` program.
18526 *Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
18528 * Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
18529 like to hear about it if this slows down other processors.
18533 * Add CygWin32 platform information to Configure script.
18535 *Alan Batie <batie@aahz.jf.intel.com>*
18537 * Fixed ms/32all.bat script: `no_asm` -> `no-asm`
18539 *Rainer W. Gerling <gerling@mpg-gv.mpg.de>*
18541 * New program nseq to manipulate netscape certificate sequences
18545 * Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
18550 * Fixes to BN code. Previously the default was to define BN_RECURSION
18551 but the BN code had some problems that would cause failures when
18552 doing certificate verification and some other functions.
18554 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
18556 * Add ASN1 and PEM code to support netscape certificate sequences.
18560 * Add ASN1 and PEM code to support netscape certificate sequences.
18564 * Add several PKIX and private extended key usage OIDs.
18568 * Modify the 'ca' program to handle the new extension code. Modify
18569 openssl.cnf for new extension format, add comments.
18573 * More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
18574 and add a sample to openssl.cnf so req -x509 now adds appropriate
18579 * Continued X509 V3 changes. Add to other makefiles, integrate with the
18580 error code, add initial support to X509_print() and x509 application.
18584 * Takes a deep breath and start adding X509 V3 extension support code. Add
18585 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
18586 stuff is currently isolated and isn't even compiled yet.
18590 * Continuing patches for GeneralizedTime. Fix up certificate and CRL
18591 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
18592 Removed the versions check from X509 routines when loading extensions:
18593 this allows certain broken certificates that don't set the version
18594 properly to be processed.
18598 * Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
18599 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
18600 can still be regenerated with "make depend".
18604 * Spelling mistake in C version of CAST-128.
18606 *Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>*
18608 * Changes to the error generation code. The perl script err-code.pl
18609 now reads in the old error codes and retains the old numbers, only
18610 adding new ones if necessary. It also only changes the .err files if new
18611 codes are added. The makefiles have been modified to only insert errors
18612 when needed (to avoid needlessly modifying header files). This is done
18613 by only inserting errors if the .err file is newer than the auto generated
18614 C file. To rebuild all the error codes from scratch (the old behaviour)
18615 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
18616 or delete all the .err files.
18620 * CAST-128 was incorrectly implemented for short keys. The C version has
18621 been fixed, but is untested. The assembler versions are also fixed, but
18622 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
18623 to regenerate it if needed.
18624 *Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
18625 Hagino <itojun@kame.net>*
18627 * File was opened incorrectly in randfile.c.
18629 *Ulf Möller <ulf@fitug.de>*
18631 * Beginning of support for GeneralizedTime. d2i, i2d, check and print
18632 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
18633 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
18634 al: it's just almost always a UTCTime. Note this patch adds new error
18635 codes so do a "make errors" if there are problems.
18639 * Correct Linux 1 recognition in config.
18641 *Ulf Möller <ulf@fitug.de>*
18643 * Remove pointless MD5 hash when using DSA keys in ca.
18645 *Anonymous <nobody@replay.com>*
18647 * Generate an error if given an empty string as a cert directory. Also
18648 generate an error if handed NULL (previously returned 0 to indicate an
18649 error, but didn't set one).
18651 *Ben Laurie, reported by Anonymous <nobody@replay.com>*
18653 * Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
18657 * Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
18658 parameters. This was causing a warning which killed off the Win32 compile.
18662 * Remove C++ style comments from crypto/bn/bn_local.h.
18664 *Neil Costigan <neil.costigan@celocom.com>*
18666 * The function OBJ_txt2nid was broken. It was supposed to return a nid
18667 based on a text string, looking up short and long names and finally
18668 "dot" format. The "dot" format stuff didn't work. Added new function
18669 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
18670 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
18671 OID is not part of the table.
18675 * Add prototypes to X509 lookup/verify methods, fixing a bug in
18676 X509_LOOKUP_by_alias().
18680 * Sort openssl functions by name.
18684 * Get the `gendsa` command working and add it to the `list` command. Remove
18685 encryption from sample DSA keys (in case anyone is interested the password
18690 * Make *all* `*_free` functions accept a NULL pointer.
18692 *Frans Heymans <fheymans@isaserver.be>*
18694 * If a DH key is generated in s3_srvr.c, don't blow it by trying to use
18697 *Anonymous <nobody@replay.com>*
18699 * s_server should send the CAfile as acceptable CAs, not its own cert.
18701 *Bodo Moeller <3moeller@informatik.uni-hamburg.de>*
18703 * Don't blow it for numeric `-newkey` arguments to `apps/req`.
18705 *Bodo Moeller <3moeller@informatik.uni-hamburg.de>*
18707 * Temp key "for export" tests were wrong in s3_srvr.c.
18709 *Anonymous <nobody@replay.com>*
18711 * Add prototype for temp key callback functions
18712 SSL_CTX_set_tmp_{rsa,dh}_callback().
18716 * Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
18717 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
18721 * X509_name_add_entry() freed the wrong thing after an error.
18723 *Arne Ansper <arne@ats.cyber.ee>*
18725 * rsa_eay.c would attempt to free a NULL context.
18727 *Arne Ansper <arne@ats.cyber.ee>*
18729 * BIO_s_socket() had a broken should_retry() on Windoze.
18731 *Arne Ansper <arne@ats.cyber.ee>*
18733 * BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
18735 *Arne Ansper <arne@ats.cyber.ee>*
18737 * Make sure the already existing X509_STORE->depth variable is initialized
18738 in X509_STORE_new(), but document the fact that this variable is still
18739 unused in the certificate verification process.
18741 *Ralf S. Engelschall*
18743 * Fix the various library and `apps/` files to free up pkeys obtained from
18744 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
18748 * Fix reference counting in X509_PUBKEY_get(). This makes
18749 demos/maurice/example2.c work, amongst others, probably.
18751 *Steve Henson and Ben Laurie*
18753 * First cut of a cleanup for `apps/`. First the `ssleay` program is now named
18754 `openssl` and second, the shortcut symlinks for the `openssl <command>`
18755 are no longer created. This way we have a single and consistent command
18756 line interface `openssl <command>`, similar to `cvs <command>`.
18758 *Ralf S. Engelschall, Paul Sutton and Ben Laurie*
18760 * ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
18761 BIT STRING wrapper always have zero unused bits.
18765 * Add CA.pl, perl version of CA.sh, add extended key usage OID.
18769 * Make the top-level INSTALL documentation easier to understand.
18773 * Makefiles updated to exit if an error occurs in a sub-directory
18774 make (including if user presses ^C) [Paul Sutton]
18776 * Make Montgomery context stuff explicit in RSA data structure.
18780 * Fix build order of pem and err to allow for generated pem.h.
18784 * Fix renumbering bug in X509_NAME_delete_entry().
18788 * Enhanced the err-ins.pl script so it makes the error library number
18789 global and can add a library name. This is needed for external ASN1 and
18790 other error libraries.
18794 * Fixed sk_insert which never worked properly.
18798 * Fix ASN1 macros so they can handle indefinite length constructed
18799 EXPLICIT tags. Some non standard certificates use these: they can now
18804 * Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
18805 into a single doc/ssleay.txt bundle. This way the information is still
18806 preserved but no longer messes up this directory. Now it's new room for
18807 the new set of documentation files.
18809 *Ralf S. Engelschall*
18811 * SETs were incorrectly DER encoded. This was a major pain, because they
18812 shared code with SEQUENCEs, which aren't coded the same. This means that
18813 almost everything to do with SETs or SEQUENCEs has either changed name or
18814 number of arguments.
18816 *Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>*
18818 * Fix test data to work with the above.
18822 * Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
18823 was already fixed by Eric for 0.9.1 it seems.
18825 *Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>*
18827 * Autodetect FreeBSD3.
18831 * Fix various bugs in Configure. This affects the following platforms:
18835 unixware-2.0-pentium
18840 * Eliminate generated files from CVS. Reorder tests to regenerate files
18841 before they are needed.
18845 * Generate Makefile.ssl from Makefile.org (to keep CVS happy).
18849 ### Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
18851 * Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
18852 changed SSLeay to OpenSSL in version strings.
18854 *Ralf S. Engelschall*
18856 * Some fixups to the top-level documents.
18860 * Fixed the nasty bug where rsaref.h was not found under compile-time
18861 because the symlink to include/ was missing.
18863 *Ralf S. Engelschall*
18865 * Incorporated the popular no-RSA/DSA-only patches
18866 which allow to compile a RSA-free SSLeay.
18868 *Andrew Cooke / Interrader Ldt., Ralf S. Engelschall*
18870 * Fixed nasty rehash problem under `make -f Makefile.ssl links`
18871 when "ssleay" is still not found.
18873 *Ralf S. Engelschall*
18875 * Added more platforms to Configure: Cray T3E, HPUX 11,
18877 *Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>*
18879 * Updated the README file.
18881 *Ralf S. Engelschall*
18883 * Added various .cvsignore files in the CVS repository subdirs
18884 to make a "cvs update" really silent.
18886 *Ralf S. Engelschall*
18888 * Recompiled the error-definition header files and added
18889 missing symbols to the Win32 linker tables.
18891 *Ralf S. Engelschall*
18893 * Cleaned up the top-level documents;
18894 o new files: CHANGES and LICENSE
18895 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
18896 o merged COPYRIGHT into LICENSE
18897 o removed obsolete TODO file
18898 o renamed MICROSOFT to INSTALL.W32
18900 *Ralf S. Engelschall*
18902 * Removed dummy files from the 0.9.1b source tree:
18903 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
18904 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
18905 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
18906 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
18907 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
18909 *Ralf S. Engelschall*
18911 * Added various platform portability fixes.
18915 * The Genesis of the OpenSSL rpject:
18916 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
18917 Young and Tim J. Hudson created while they were working for C2Net until
18920 *The OpenSSL Project*
18922 ### Changes between 0.9.0b and 0.9.1b [not released]
18924 * Updated a few CA certificates under certs/
18928 * Changed some BIGNUM api stuff.
18932 * Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
18933 DGUX x86, Linux Alpha, etc.
18937 * New COMP library [crypto/comp/] for SSL Record Layer Compression:
18938 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
18943 * Add -strparse option to asn1pars program which parses nested
18946 *Dr Stephen Henson <shenson@bigfoot.com>*
18948 * Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
18952 * DSA fix for "ca" program.
18956 * Added "-genkey" option to "dsaparam" program.
18960 * Added RIPE MD160 (rmd160) message digest.
18964 * Added -a (all) option to "ssleay version" command.
18968 * Added PLATFORM define which is the id given to Configure.
18972 * Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
18976 * Extended the ASN.1 parser routines.
18980 * Extended BIO routines to support REUSEADDR, seek, tell, etc.
18984 * Added a BN_CTX to the BN library.
18988 * Fixed the weak key values in DES library
18992 * Changed API in EVP library for cipher aliases.
18996 * Added support for RC2/64bit cipher.
19000 * Converted the lhash library to the crypto/mem.c functions.
19004 * Added more recognized ASN.1 object ids.
19008 * Added more RSA padding checks for SSL/TLS.
19012 * Added BIO proxy/filter functionality.
19016 * Added extra_certs to SSL_CTX which can be used
19017 send extra CA certificates to the client in the CA cert chain sending
19018 process. It can be configured with SSL_CTX_add_extra_chain_cert().
19022 * Now Fortezza is denied in the authentication phase because
19023 this is key exchange mechanism is not supported by SSLeay at all.
19027 * Additional PKCS1 checks.
19031 * Support the string "TLSv1" for all TLS v1 ciphers.
19035 * Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
19036 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
19040 * Fixed a few memory leaks.
19044 * Fixed various code and comment typos.
19048 * A minor bug in ssl/s3_clnt.c where there would always be 4 0
19049 bytes sent in the client random.
19051 *Edward Bishop <ebishop@spyglass.com>*
19055 [CVE-2020-1971]: https://www.openssl.org/news/vulnerabilities.html#CVE-2020-1971
19056 [CVE-2020-1967]: https://www.openssl.org/news/vulnerabilities.html#CVE-2020-1967
19057 [CVE-2019-1563]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1563
19058 [CVE-2019-1559]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1559
19059 [CVE-2019-1552]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1552
19060 [CVE-2019-1551]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1551
19061 [CVE-2019-1549]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1549
19062 [CVE-2019-1547]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1547
19063 [CVE-2019-1543]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1543
19064 [CVE-2018-5407]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-5407
19065 [CVE-2018-0739]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0739
19066 [CVE-2018-0737]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0737
19067 [CVE-2018-0735]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0735
19068 [CVE-2018-0734]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0734
19069 [CVE-2018-0733]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0733
19070 [CVE-2018-0732]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0732
19071 [CVE-2017-3738]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3738
19072 [CVE-2017-3737]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3737
19073 [CVE-2017-3736]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3736
19074 [CVE-2017-3735]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3735
19075 [CVE-2017-3733]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3733
19076 [CVE-2017-3732]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3732
19077 [CVE-2017-3731]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3731
19078 [CVE-2017-3730]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3730
19079 [CVE-2016-7055]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7055
19080 [CVE-2016-7054]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7054
19081 [CVE-2016-7053]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7053
19082 [CVE-2016-7052]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7052
19083 [CVE-2016-6309]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6309
19084 [CVE-2016-6308]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6308
19085 [CVE-2016-6307]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6307
19086 [CVE-2016-6306]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6306
19087 [CVE-2016-6305]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6305
19088 [CVE-2016-6304]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6304
19089 [CVE-2016-6303]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6303
19090 [CVE-2016-6302]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6302
19091 [CVE-2016-2183]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2183
19092 [CVE-2016-2182]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2182
19093 [CVE-2016-2181]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2181
19094 [CVE-2016-2180]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2180
19095 [CVE-2016-2179]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2179
19096 [CVE-2016-2178]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2178
19097 [CVE-2016-2177]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2177
19098 [CVE-2016-2176]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2176
19099 [CVE-2016-2109]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2109
19100 [CVE-2016-2107]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2107
19101 [CVE-2016-2106]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2106
19102 [CVE-2016-2105]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2105
19103 [CVE-2016-0800]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0800
19104 [CVE-2016-0799]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0799
19105 [CVE-2016-0798]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0798
19106 [CVE-2016-0797]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0797
19107 [CVE-2016-0705]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0705
19108 [CVE-2016-0702]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0702
19109 [CVE-2016-0701]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0701
19110 [CVE-2015-3197]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3197
19111 [CVE-2015-3196]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3196
19112 [CVE-2015-3195]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3195
19113 [CVE-2015-3194]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3194
19114 [CVE-2015-3193]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3193
19115 [CVE-2015-1793]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1793
19116 [CVE-2015-1792]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1792
19117 [CVE-2015-1791]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1791
19118 [CVE-2015-1790]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1790
19119 [CVE-2015-1789]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1789
19120 [CVE-2015-1788]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1788
19121 [CVE-2015-1787]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1787
19122 [CVE-2015-0293]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0293
19123 [CVE-2015-0291]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0291
19124 [CVE-2015-0290]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0290
19125 [CVE-2015-0289]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0289
19126 [CVE-2015-0288]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0288
19127 [CVE-2015-0287]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0287
19128 [CVE-2015-0286]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0286
19129 [CVE-2015-0285]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0285
19130 [CVE-2015-0209]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0209
19131 [CVE-2015-0208]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0208
19132 [CVE-2015-0207]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0207
19133 [CVE-2015-0206]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0206
19134 [CVE-2015-0205]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0205
19135 [CVE-2015-0204]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0204
19136 [CVE-2014-8275]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-8275
19137 [CVE-2014-5139]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-5139
19138 [CVE-2014-3572]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3572
19139 [CVE-2014-3571]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3571
19140 [CVE-2014-3570]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3570
19141 [CVE-2014-3569]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3569
19142 [CVE-2014-3568]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3568
19143 [CVE-2014-3567]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3567
19144 [CVE-2014-3566]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3566
19145 [CVE-2014-3513]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3513
19146 [CVE-2014-3512]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3512
19147 [CVE-2014-3511]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3511
19148 [CVE-2014-3510]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3510
19149 [CVE-2014-3509]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3509
19150 [CVE-2014-3508]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3508
19151 [CVE-2014-3507]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3507
19152 [CVE-2014-3506]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3506
19153 [CVE-2014-3505]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3505
19154 [CVE-2014-3470]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3470
19155 [CVE-2014-0224]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0224
19156 [CVE-2014-0221]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0221
19157 [CVE-2014-0195]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0195
19158 [CVE-2014-0160]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0160
19159 [CVE-2014-0076]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0076
19160 [CVE-2013-6450]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-6450
19161 [CVE-2013-4353]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-4353
19162 [CVE-2013-0169]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-0169
19163 [CVE-2013-0166]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-0166
19164 [CVE-2012-2686]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2686
19165 [CVE-2012-2333]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2333
19166 [CVE-2012-2110]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2110
19167 [CVE-2012-0884]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0884
19168 [CVE-2012-0050]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0050
19169 [CVE-2012-0027]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0027
19170 [CVE-2011-4619]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4619
19171 [CVE-2011-4577]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4577
19172 [CVE-2011-4576]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4576
19173 [CVE-2011-4109]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4109
19174 [CVE-2011-4108]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4108
19175 [CVE-2011-3210]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-3210
19176 [CVE-2011-3207]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-3207
19177 [CVE-2011-0014]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-0014
19178 [CVE-2010-4252]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-4252
19179 [CVE-2010-4180]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-4180
19180 [CVE-2010-3864]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-3864
19181 [CVE-2010-1633]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-1633
19182 [CVE-2010-0740]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-0740
19183 [CVE-2010-0433]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-0433
19184 [CVE-2009-4355]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-4355
19185 [CVE-2009-3555]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-3555
19186 [CVE-2009-3245]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-3245
19187 [CVE-2009-1386]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1386
19188 [CVE-2009-1379]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1379
19189 [CVE-2009-1378]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1378
19190 [CVE-2009-1377]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1377
19191 [CVE-2009-0789]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0789
19192 [CVE-2009-0591]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0591
19193 [CVE-2009-0590]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0590
19194 [CVE-2008-5077]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-5077
19195 [CVE-2008-1678]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-1678
19196 [CVE-2008-1672]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-1672
19197 [CVE-2008-0891]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-0891
19198 [CVE-2007-5135]: https://www.openssl.org/news/vulnerabilities.html#CVE-2007-5135
19199 [CVE-2007-4995]: https://www.openssl.org/news/vulnerabilities.html#CVE-2007-4995
19200 [CVE-2006-4343]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-4343
19201 [CVE-2006-4339]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-4339
19202 [CVE-2006-3738]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-3738
19203 [CVE-2006-2940]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-2940
19204 [CVE-2006-2937]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-2937
19205 [CVE-2005-2969]: https://www.openssl.org/news/vulnerabilities.html#CVE-2005-2969
19206 [CVE-2004-0112]: https://www.openssl.org/news/vulnerabilities.html#CVE-2004-0112
19207 [CVE-2004-0079]: https://www.openssl.org/news/vulnerabilities.html#CVE-2004-0079
19208 [CVE-2003-0851]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0851
19209 [CVE-2003-0545]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0545
19210 [CVE-2003-0544]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0544
19211 [CVE-2003-0543]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0543
19212 [CVE-2003-0078]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0078
19213 [CVE-2002-0659]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0659
19214 [CVE-2002-0657]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0657
19215 [CVE-2002-0656]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0656
19216 [CVE-2002-0655]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0655