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 1.1.1 and 3.0 [xx XXX xxxx]
33 * Add a configurable flag to output date formats as ISO 8601. Does not
34 change the default date format.
38 * Version of MSVC earlier than 1300 could get link warnings, which could
39 be suppressed if the undocumented -DI_CAN_LIVE_WITH_LNK4049 was set.
40 Support for this flag has been removed.
44 * Rework and make DEBUG macros consistent. Remove unused -DCONF_DEBUG,
45 -DBN_CTX_DEBUG, and REF_PRINT. Add a new tracing category and use it for
46 printing reference counts. Rename -DDEBUG_UNUSED to -DUNUSED_RESULT_DEBUG
47 Fix BN_DEBUG_RAND so it compiles and, when set, force DEBUG_RAND to be set
48 also. Rename engine_debug_ref to be ENGINE_REF_PRINT also for consistency.
52 * The signatures of the functions to get and set options on SSL and
53 SSL_CTX objects changed from "unsigned long" to "uint64_t" type.
54 Some source code changes may be required.
58 * The public definitions of conf_method_st and conf_st have been
59 deprecated. They will be made opaque in a future release.
61 *Rich Salz and Tomáš Mráz*
63 * Client-initiated renegotiation is disabled by default. To allow it, use
64 the -client_renegotiation option, the SSL_OP_ALLOW_CLIENT_RENEGOTIATION
65 flag, or the "ClientRenegotiation" config parameter as appropriate.
69 * Add "abspath" and "includedir" pragma's to config files, to prevent,
70 or modify relative pathname inclusion.
74 * OpenSSL includes a cryptographic module that is intended to be FIPS 140-2
75 validated. Please consult the README-FIPS and
76 README-PROVIDERS files, as well as the migration guide.
78 *OpenSSL team members and many third party contributors*
80 * For the key types DH and DHX the allowed settable parameters are now different.
84 * The openssl commands that read keys, certificates, and CRLs now
85 automatically detect the PEM or DER format of the input files.
87 *David von Oheimb, Richard Levitte, and Tomáš Mráz*
89 * Added enhanced PKCS#12 APIs which accept a library context.
93 * The default manual page suffix ($MANSUFFIX) has been changed to "ossl"
97 * Added support for Kernel TLS (KTLS).
99 *Boris Pismenny, John Baldwin and Andrew Gallatin*
101 * Support for RFC 5746 secure renegotiation is now required by default for
102 SSL or TLS connections to succeed.
106 * The signature of the `copy` functional parameter of the
107 EVP_PKEY_meth_set_copy() function has changed so its `src` argument is
108 now `const EVP_PKEY_CTX *` instead of `EVP_PKEY_CTX *`. Similarly
109 the signature of the `pub_decode` functional parameter of the
110 EVP_PKEY_asn1_set_public() function has changed so its `pub` argument is
111 now `const X509_PUBKEY *` instead of `X509_PUBKEY *`.
115 * The error return values from some control calls (ctrl) have changed.
119 * A public key check is now performed during EVP_PKEY_derive_set_peer().
123 * Many functions in the EVP_ namespace that are getters of values from
124 implementations or contexts were renamed to include get or get0 in their
125 names. Old names are provided as macro aliases for compatibility and
130 * The EVP_PKEY_CTRL_PKCS7_ENCRYPT, EVP_PKEY_CTRL_PKCS7_DECRYPT,
131 EVP_PKEY_CTRL_PKCS7_SIGN, EVP_PKEY_CTRL_CMS_ENCRYPT,
132 EVP_PKEY_CTRL_CMS_DECRYPT, and EVP_PKEY_CTRL_CMS_SIGN control operations
137 * The EVP_PKEY_public_check() and EVP_PKEY_param_check() functions now work for
140 * The output from the command line applications may have minor
145 * The output from numerous "printing" may have minor changes.
149 * Windows thread synchronization uses read/write primitives (SRWLock) when
150 supported by the OS, otherwise CriticalSection continues to be used.
154 * Add filter BIO BIO_f_readbuffer() that allows BIO_tell() and BIO_seek() to
155 work on read only BIO source/sinks that do not support these functions.
156 This allows piping or redirection of a file BIO using stdin to be buffered
157 into memory. This is used internally in OSSL_DECODER_from_bio().
161 * OSSL_STORE_INFO_get_type() may now return an additional value. In 1.1.1
162 this function would return one of the values OSSL_STORE_INFO_NAME,
163 OSSL_STORE_INFO_PKEY, OSSL_STORE_INFO_PARAMS, OSSL_STORE_INFO_CERT or
164 OSSL_STORE_INFO_CRL. Decoded public keys would previously have been reported
165 as type OSSL_STORE_INFO_PKEY in 1.1.1. In 3.0 decoded public keys are now
166 reported as having the new type OSSL_STORE_INFO_PUBKEY. Applications
167 using this function should be amended to handle the changed return value.
171 * Improved adherence to Enhanced Security Services (ESS, RFC 2634 and RFC 5035)
172 for the TSP and CMS Advanced Electronic Signatures (CAdES) implementations.
173 As required by RFC 5035 check both ESSCertID and ESSCertIDv2 if both present.
174 Correct the semantics of checking the validation chain in case ESSCertID{,v2}
175 contains more than one certificate identifier: This means that all
176 certificates referenced there MUST be part of the validation chain.
180 * The implementation of older EVP ciphers related to CAST, IDEA, SEED, RC2, RC4,
181 RC5, DESX and DES have been moved to the legacy provider.
185 * The implementation of the EVP digests MD2, MD4, MDC2, WHIRLPOOL and
186 RIPEMD-160 have been moved to the legacy provider.
190 * The deprecated function EVP_PKEY_get0() now returns NULL being called for a
195 * The deprecated functions EVP_PKEY_get0_RSA(),
196 EVP_PKEY_get0_DSA(), EVP_PKEY_get0_EC_KEY(), EVP_PKEY_get0_DH(),
197 EVP_PKEY_get0_hmac(), EVP_PKEY_get0_poly1305() and EVP_PKEY_get0_siphash() as
198 well as the similarly named "get1" functions behave differently in
203 * A number of functions handling low-level keys or engines were deprecated
204 including EVP_PKEY_set1_engine(), EVP_PKEY_get0_engine(), EVP_PKEY_assign(),
205 EVP_PKEY_get0(), EVP_PKEY_get0_hmac(), EVP_PKEY_get0_poly1305() and
206 EVP_PKEY_get0_siphash().
210 * PKCS#5 PBKDF1 key derivation has been moved from PKCS5_PBE_keyivgen() into
211 the legacy crypto provider as an EVP_KDF. Applications requiring this KDF
212 will need to load the legacy crypto provider. This includes these PBE
213 algorithms which use this KDF:
214 - NID_pbeWithMD2AndDES_CBC
215 - NID_pbeWithMD5AndDES_CBC
216 - NID_pbeWithSHA1AndRC2_CBC
217 - NID_pbeWithMD2AndRC2_CBC
218 - NID_pbeWithMD5AndRC2_CBC
219 - NID_pbeWithSHA1AndDES_CBC
223 * Deprecated obsolete BIO_set_callback(), BIO_get_callback(), and
224 BIO_debug_callback() functions.
228 * Deprecated obsolete EVP_PKEY_CTX_get0_dh_kdf_ukm() and
229 EVP_PKEY_CTX_get0_ecdh_kdf_ukm() functions.
233 * The RAND_METHOD APIs have been deprecated.
237 * The SRP APIs have been deprecated.
241 * Add a compile time option to prevent the caching of provider fetched
242 algorithms. This is enabled by including the no-cached-fetch option
243 at configuration time.
247 * pkcs12 now uses defaults of PBKDF2, AES and SHA-256, with a MAC iteration
248 count of PKCS12_DEFAULT_ITER.
250 *Tomáš Mráz and Sahana Prasad*
252 * The openssl speed command does not use low-level API calls anymore.
256 * Parallel dual-prime 1024-bit modular exponentiation for AVX512_IFMA
259 *Ilya Albrekht, Sergey Kirillov, Andrey Matyukov (Intel Corp)*
261 * Combining the Configure options no-ec and no-dh no longer disables TLSv1.3.
265 * Implemented support for fully "pluggable" TLSv1.3 groups. This means that
266 providers may supply their own group implementations (using either the "key
267 exchange" or the "key encapsulation" methods) which will automatically be
268 detected and used by libssl.
270 *Matt Caswell, Nicola Tuveri*
272 * The undocumented function X509_certificate_type() has been deprecated;
276 * Deprecated the obsolete BN_pseudo_rand() and BN_pseudo_rand_range().
280 * Removed RSA padding mode for SSLv23 (which was only used for
281 SSLv2). This includes the functions RSA_padding_check_SSLv23() and
282 RSA_padding_add_SSLv23() and the `-ssl` option in the deprecated
287 * Deprecated the obsolete X9.31 RSA key generation related functions.
291 * The default key generation method for the regular 2-prime RSA keys was
292 changed to the FIPS 186-4 B.3.6 method.
296 * Deprecated the BN_is_prime_ex() and BN_is_prime_fasttest_ex() functions.
300 * Deprecated EVP_MD_CTX_set_update_fn() and EVP_MD_CTX_update_fn().
304 * Deprecated the type OCSP_REQ_CTX and the functions OCSP_REQ_CTX_*() and
305 replaced with OSSL_HTTP_REQ_CTX and the functions OSSL_HTTP_REQ_CTX_*().
307 *Rich Salz, Richard Levitte, and David von Oheimb*
309 * Deprecated `X509_http_nbio()` and `X509_CRL_http_nbio()`.
313 * Deprecated `OCSP_parse_url()`.
317 * Validation of SM2 keys has been separated from the validation of regular EC
322 * Behavior of the `pkey` app is changed, when using the `-check` or `-pubcheck`
323 switches: a validation failure triggers an early exit, returning a failure
324 exit status to the parent process.
328 * Changed behavior of SSL_CTX_set_ciphersuites() and SSL_set_ciphersuites()
329 to ignore unknown ciphers.
333 * The `-cipher-commands` and `-digest-commands` options
334 of the command line utility `list` have been deprecated.
335 Instead use the `-cipher-algorithms` and `-digest-algorithms` options.
339 * Added convenience functions for generating asymmetric key pairs:
340 The 'quick' one-shot (yet somewhat limited) function L<EVP_PKEY_Q_keygen(3)>
341 and macros for the most common cases: <EVP_RSA_gen(3)> and L<EVP_EC_gen(3)>.
345 * All of the low level EC_KEY functions have been deprecated.
347 *Shane Lontis, Paul Dale, Richard Levitte, and Tomáš Mráz*
349 * Deprecated all the libcrypto and libssl error string loading
354 * The functions SSL_CTX_set_tmp_dh_callback and SSL_set_tmp_dh_callback, as
355 well as the macros SSL_CTX_set_tmp_dh() and SSL_set_tmp_dh() have been
360 * The `-crypt` option to the `passwd` command line tool has been removed.
364 * The -C option to the `x509`, `dhparam`, `dsaparam`, and `ecparam` commands
369 * Add support for AES Key Wrap inverse ciphers to the EVP layer.
373 * Deprecated EVP_PKEY_set1_tls_encodedpoint() and
374 EVP_PKEY_get1_tls_encodedpoint().
378 * The security callback, which can be customised by application code, supports
379 the security operation SSL_SECOP_TMP_DH. One location of the "other" parameter
380 was incorrectly passing a DH object. It now passed an EVP_PKEY in all cases.
384 * Add PKCS7_get_octet_string() and PKCS7_type_is_other() to the public
385 interface. Their functionality remains unchanged.
389 * Added new option for 'openssl list', '-providers', which will display the
390 list of loaded providers, their names, version and status. It optionally
391 displays their gettable parameters.
395 * Removed EVP_PKEY_set_alias_type().
399 * Deprecated `EVP_PKEY_CTX_set_rsa_keygen_pubexp()` and introduced
400 `EVP_PKEY_CTX_set1_rsa_keygen_pubexp()`, which is now preferred.
404 * Changed all "STACK" functions to be macros instead of inline functions. Macro
405 parameters are still checked for type safety at compile time via helper
410 * Remove the RAND_DRBG API
412 *Paul Dale and Matthias St. Pierre*
414 * Allow `SSL_set1_host()` and `SSL_add1_host()` to take IP literal addresses
415 as well as actual hostnames.
419 * The 'MinProtocol' and 'MaxProtocol' configuration commands now silently
420 ignore TLS protocol version bounds when configuring DTLS-based contexts, and
421 conversely, silently ignore DTLS protocol version bounds when configuring
422 TLS-based contexts. The commands can be repeated to set bounds of both
423 types. The same applies with the corresponding "min_protocol" and
424 "max_protocol" command-line switches, in case some application uses both TLS
427 SSL_CTX instances that are created for a fixed protocol version (e.g.
428 `TLSv1_server_method()`) also silently ignore version bounds. Previously
429 attempts to apply bounds to these protocol versions would result in an
430 error. Now only the "version-flexible" SSL_CTX instances are subject to
431 limits in configuration files in command-line options.
435 * Deprecated the `ENGINE` API. Engines should be replaced with providers
440 * Reworked the recorded ERR codes to make better space for system errors.
441 To distinguish them, the macro `ERR_SYSTEM_ERROR()` indicates if the
442 given code is a system error (true) or an OpenSSL error (false).
446 * Reworked the test perl framework to better allow parallel testing.
448 *Nicola Tuveri and David von Oheimb*
450 * Added ciphertext stealing algorithms AES-128-CBC-CTS, AES-192-CBC-CTS and
451 AES-256-CBC-CTS to the providers. CS1, CS2 and CS3 variants are supported.
455 * 'Configure' has been changed to figure out the configuration target if
456 none is given on the command line. Consequently, the 'config' script is
457 now only a mere wrapper. All documentation is changed to only mention
460 *Rich Salz and Richard Levitte*
462 * Added a library context `OSSL_LIB_CTX` that applications as well as
463 other libraries can use to form a separate context within which
464 libcrypto operations are performed.
468 * Handshake now fails if Extended Master Secret extension is dropped
473 * Dropped interactive mode from the `openssl` program.
477 * Deprecated `EVP_PKEY_cmp()` and `EVP_PKEY_cmp_parameters()`.
479 *David von Oheimb and Shane Lontis*
481 * Deprecated `EC_METHOD_get_field_type()`.
485 * Deprecated EC_GFp_simple_method(), EC_GFp_mont_method(),
486 EC_GF2m_simple_method(), EC_GFp_nist_method(), EC_GFp_nistp224_method()
487 EC_GFp_nistp256_method(), and EC_GFp_nistp521_method().
491 * Deprecated EC_GROUP_new(), EC_GROUP_method_of(), and EC_POINT_method_of().
495 * Add CAdES-BES signature verification support, mostly derived
496 from ESSCertIDv2 TS (RFC 5816) contribution by Marek Klein.
498 *Filipe Raimundo da Silva*
500 * Add CAdES-BES signature scheme and attributes support (RFC 5126) to CMS API.
504 * Added the AuthEnvelopedData content type structure (RFC 5083) with AES-GCM
505 parameter (RFC 5084) for the Cryptographic Message Syntax (CMS).
509 * Deprecated EC_POINT_make_affine() and EC_POINTs_make_affine().
513 * Deprecated EC_GROUP_precompute_mult(), EC_GROUP_have_precompute_mult(), and
514 EC_KEY_precompute_mult().
518 * Deprecated EC_POINTs_mul().
522 * Removed FIPS_mode() and FIPS_mode_set().
526 * The SSL option SSL_OP_IGNORE_UNEXPECTED_EOF is introduced.
530 * Deprecated EC_POINT_set_Jprojective_coordinates_GFp() and
531 EC_POINT_get_Jprojective_coordinates_GFp().
535 * Added OSSL_PARAM_BLD to the public interface. This allows OSSL_PARAM
536 arrays to be more easily constructed via a series of utility functions.
537 Create a parameter builder using OSSL_PARAM_BLD_new(), add parameters using
538 the various push functions and finally convert to a passable OSSL_PARAM
539 array using OSSL_PARAM_BLD_to_param().
543 * The security strength of SHA1 and MD5 based signatures in TLS has been
548 * Added EVP_PKEY_set_type_by_keymgmt(), to initialise an EVP_PKEY to
549 contain a provider side internal key.
553 * ASN1_verify(), ASN1_digest() and ASN1_sign() have been deprecated.
557 * Project text documents not yet having a proper file name extension
558 (`HACKING`, `LICENSE`, `NOTES*`, `README*`, `VERSION`) have been renamed to
559 `*.md` as far as reasonable, else `*.txt`, for better use with file managers.
563 * The main project documents (README, NEWS, CHANGES, INSTALL, SUPPORT)
564 have been converted to Markdown with the goal to produce documents
565 which not only look pretty when viewed online in the browser, but
566 remain well readable inside a plain text editor.
568 To achieve this goal, a 'minimalistic' Markdown style has been applied
569 which avoids formatting elements that interfere too much with the
570 reading flow in the text file. For example, it
572 * avoids [ATX headings][] and uses [setext headings][] instead
573 (which works for `<h1>` and `<h2>` headings only).
574 * avoids [inline links][] and uses [reference links][] instead.
575 * avoids [fenced code blocks][] and uses [indented code blocks][] instead.
577 [ATX headings]: https://github.github.com/gfm/#atx-headings
578 [setext headings]: https://github.github.com/gfm/#setext-headings
579 [inline links]: https://github.github.com/gfm/#inline-link
580 [reference links]: https://github.github.com/gfm/#reference-link
581 [fenced code blocks]: https://github.github.com/gfm/#fenced-code-blocks
582 [indented code blocks]: https://github.github.com/gfm/#indented-code-blocks
584 *Matthias St. Pierre*
586 * The test suite is changed to preserve results of each test recipe.
587 A new directory test-runs/ with subdirectories named like the
588 test recipes are created in the build tree for this purpose.
592 * Added an implementation of CMP and CRMF (RFC 4210, RFC 4211 RFC 6712).
593 This adds `crypto/cmp/`, `crpyto/crmf/`, `apps/cmp.c`, and `test/cmp_*`.
594 See L<openssl-cmp(1)> and L<OSSL_CMP_exec_IR_ses(3)> as starting points.
596 *David von Oheimb, Martin Peylo*
598 * Generalized the HTTP client code from `crypto/ocsp/` into `crpyto/http/`.
599 It supports arbitrary request and response content types, GET redirection,
600 TLS, connections via HTTP(S) proxies, connections and exchange via
601 user-defined BIOs (allowing implicit connections), persistent connections,
602 and timeout checks. See L<OSSL_HTTP_transfer(3)> etc. for details.
603 The legacy OCSP-focused (and only partly documented) API
604 is retained for backward compatibility, while most of it is deprecated.
608 * Added `util/check-format.pl`, a tool for checking adherence to the
609 OpenSSL coding style <https://www.openssl.org/policies/codingstyle.html>.
610 The checks performed are incomplete and yield some false positives.
611 Still the tool should be useful for detecting most typical glitches.
615 * `BIO_do_connect()` and `BIO_do_handshake()` have been extended:
616 If domain name resolution yields multiple IP addresses all of them are tried
617 after `connect()` failures.
621 * All of the low level RSA functions have been deprecated.
625 * X509 certificates signed using SHA1 are no longer allowed at security
630 * The command line utilities dhparam, dsa, gendsa and dsaparam have been
631 modified to use PKEY APIs. These commands are now in maintenance mode
632 and no new features will be added to them.
636 * The command line utility rsautl has been deprecated.
640 * The command line utilities genrsa and rsa have been modified to use PKEY
641 APIs. They now write PKCS#8 keys by default. These commands are now in
642 maintenance mode and no new features will be added to them.
646 * All of the low level DH functions have been deprecated.
648 *Paul Dale and Matt Caswell*
650 * All of the low level DSA functions have been deprecated.
654 * Reworked the treatment of EC EVP_PKEYs with the SM2 curve to
655 automatically become EVP_PKEY_SM2 rather than EVP_PKEY_EC.
659 * Deprecated low level ECDH and ECDSA functions.
663 * Deprecated EVP_PKEY_decrypt_old() and EVP_PKEY_encrypt_old().
667 * Enhanced the documentation of EVP_PKEY_get_size(), EVP_PKEY_get_bits()
668 and EVP_PKEY_get_security_bits(). Especially EVP_PKEY_get_size() needed
669 a new formulation to include all the things it can be used for,
670 as well as words of caution.
674 * The SSL_CTX_set_tlsext_ticket_key_cb(3) function has been deprecated.
678 * All of the low level HMAC functions have been deprecated.
680 *Paul Dale and David von Oheimb*
682 * Over two thousand fixes were made to the documentation, including:
683 - Common options (such as -rand/-writerand, TLS version control, etc)
684 were refactored and point to newly-enhanced descriptions in openssl.pod.
685 - Added style conformance for all options (with help from Richard Levitte),
686 documented all reported missing options, added a CI build to check
687 that all options are documented and that no unimplemented options
689 - Documented some internals, such as all use of environment variables.
690 - Addressed all internal broken L<> references.
694 * All of the low level CMAC functions have been deprecated.
698 * The low-level MD2, MD4, MD5, MDC2, RIPEMD160 and Whirlpool digest
699 functions have been deprecated.
701 *Paul Dale and David von Oheimb*
703 * Corrected the documentation of the return values from the `EVP_DigestSign*`
704 set of functions. The documentation mentioned negative values for some
705 errors, but this was never the case, so the mention of negative values
708 Code that followed the documentation and thereby check with something
709 like `EVP_DigestSignInit(...) <= 0` will continue to work undisturbed.
713 * All of the low level cipher functions have been deprecated.
715 *Matt Caswell and Paul Dale*
717 * Removed include/openssl/opensslconf.h.in and replaced it with
718 include/openssl/configuration.h.in, which differs in not including
719 <openssl/macros.h>. A short header include/openssl/opensslconf.h
720 was added to include both.
722 This allows internal hacks where one might need to modify the set
723 of configured macros, for example this if deprecated symbols are
724 still supposed to be available internally:
726 #include <openssl/configuration.h>
728 #undef OPENSSL_NO_DEPRECATED
729 #define OPENSSL_SUPPRESS_DEPRECATED
731 #include <openssl/macros.h>
733 This should not be used by applications that use the exported
734 symbols, as that will lead to linking errors.
738 * Fixed an overflow bug in the x64_64 Montgomery squaring procedure
739 used in exponentiation with 512-bit moduli. No EC algorithms are
740 affected. Analysis suggests that attacks against 2-prime RSA1024,
741 3-prime RSA1536, and DSA1024 as a result of this defect would be very
742 difficult to perform and are not believed likely. Attacks against DH512
743 are considered just feasible. However, for an attack the target would
744 have to re-use the DH512 private key, which is not recommended anyway.
745 Also applications directly using the low-level API BN_mod_exp may be
746 affected if they use BN_FLG_CONSTTIME.
751 * Most memory-debug features have been deprecated, and the functionality
752 replaced with no-ops.
756 * Added documentation for the STACK API.
760 * Introduced a new method type and API, OSSL_ENCODER, to
761 represent generic encoders.
765 * Added a .pragma directive to the syntax of configuration files, to
766 allow varying behavior in a supported and predictable manner.
767 Currently added pragma:
771 This allows dollar signs to be a keyword character unless it's
772 followed by a opening brace or parenthesis. This is useful for
773 platforms where dollar signs are commonly used in names, such as
774 volume names and system directory names on VMS.
778 * Added functionality to create an EVP_PKEY from user data.
782 * Change the interpretation of the '--api' configuration option to
783 mean that this is a desired API compatibility level with no
784 further meaning. The previous interpretation, that this would
785 also mean to remove all deprecated symbols up to and including
786 the given version, no requires that 'no-deprecated' is also used
787 in the configuration.
789 When building applications, the desired API compatibility level
790 can be set with the OPENSSL_API_COMPAT macro like before. For
791 API compatibility version below 3.0, the old style numerical
792 value is valid as before, such as -DOPENSSL_API_COMPAT=0x10100000L.
793 For version 3.0 and on, the value is expected to be the decimal
794 value calculated from the major and minor version like this:
796 MAJOR * 10000 + MINOR * 100
800 -DOPENSSL_API_COMPAT=30000 For 3.0
801 -DOPENSSL_API_COMPAT=30200 For 3.2
803 To hide declarations that are deprecated up to and including the
804 given API compatibility level, -DOPENSSL_NO_DEPRECATED must be
805 given when building the application as well.
809 * Added the X509_LOOKUP_METHOD called X509_LOOKUP_store, to allow
810 access to certificate and CRL stores via URIs and OSSL_STORE
813 This adds the following functions:
815 - X509_LOOKUP_store()
816 - X509_STORE_load_file()
817 - X509_STORE_load_path()
818 - X509_STORE_load_store()
819 - SSL_add_store_cert_subjects_to_stack()
820 - SSL_CTX_set_default_verify_store()
821 - SSL_CTX_load_verify_file()
822 - SSL_CTX_load_verify_dir()
823 - SSL_CTX_load_verify_store()
827 * Added a new method to gather entropy on VMS, based on SYS$GET_ENTROPY.
828 The presence of this system service is determined at run-time.
832 * Added functionality to create an EVP_PKEY context based on data
833 for methods from providers. This takes an algorithm name and a
834 property query string and simply stores them, with the intent
835 that any operation that uses this context will use those strings
836 to fetch the needed methods implicitly, thereby making the port
837 of application written for pre-3.0 OpenSSL easier.
841 * The undocumented function NCONF_WIN32() has been deprecated; for
842 conversion details see the HISTORY section of doc/man5/config.pod
846 * Introduced the new functions EVP_DigestSignInit_ex() and
847 EVP_DigestVerifyInit_ex(). The macros EVP_DigestSignUpdate() and
848 EVP_DigestVerifyUpdate() have been converted to functions. See the man
849 pages for further details.
853 * Over two thousand fixes were made to the documentation, including:
854 adding missing command flags, better style conformance, documentation
857 *Rich Salz, Richard Levitte*
859 * s390x assembly pack: add hardware-support for P-256, P-384, P-521,
860 X25519, X448, Ed25519 and Ed448.
864 * Print all values for a PKCS#12 attribute with 'openssl pkcs12', not just
869 * Deprecated the public definition of `ERR_STATE` as well as the function
870 `ERR_get_state()`. This is done in preparation of making `ERR_STATE` an
875 * Added ERR functionality to give callers access to the stored function
876 names that have replaced the older function code based functions.
878 New functions are ERR_peek_error_func(), ERR_peek_last_error_func(),
879 ERR_peek_error_data(), ERR_peek_last_error_data(), ERR_get_error_all(),
880 ERR_peek_error_all() and ERR_peek_last_error_all().
882 Deprecate ERR functions ERR_get_error_line(), ERR_get_error_line_data(),
883 ERR_peek_error_line_data(), ERR_peek_last_error_line_data() and
884 ERR_func_error_string().
888 * Extended testing to be verbose for failing tests only. The make variables
889 VERBOSE_FAILURE or VF can be used to enable this:
891 $ make VF=1 test # Unix
892 $ mms /macro=(VF=1) test ! OpenVMS
893 $ nmake VF=1 test # Windows
897 * Added the `-copy_extensions` option to the `x509` command for use with
898 `-req` and `-x509toreq`. When given with the `copy` or `copyall` argument,
899 all extensions in the request are copied to the certificate or vice versa.
901 *David von Oheimb*, *Kirill Stefanenkov <kirill_stefanenkov@rambler.ru>*
903 * Added the `-copy_extensions` option to the `req` command for use with
904 `-x509`. When given with the `copy` or `copyall` argument,
905 all extensions in the certification request are copied to the certificate.
909 * The `x509`, `req`, and `ca` commands now make sure that X.509v3 certificates
910 they generate are by default RFC 5280 compliant in the following sense:
911 There is a subjectKeyIdentifier extension with a hash value of the public key
912 and for not self-signed certs there is an authorityKeyIdentifier extension
913 with a keyIdentifier field or issuer information identifying the signing key.
914 This is done unless some configuration overrides the new default behavior,
915 such as `subjectKeyIdentifier = none` and `authorityKeyIdentifier = none`.
919 * Added several checks to `X509_verify_cert()` according to requirements in
920 RFC 5280 in case `X509_V_FLAG_X509_STRICT` is set
921 (which may be done by using the CLI option `-x509_strict`):
922 * The basicConstraints of CA certificates must be marked critical.
923 * CA certificates must explicitly include the keyUsage extension.
924 * If a pathlenConstraint is given the key usage keyCertSign must be allowed.
925 * The issuer name of any certificate must not be empty.
926 * The subject name of CA certs, certs with keyUsage crlSign,
927 and certs without subjectAlternativeName must not be empty.
928 * If a subjectAlternativeName extension is given it must not be empty.
929 * The signatureAlgorithm field and the cert signature must be consistent.
930 * Any given authorityKeyIdentifier and any given subjectKeyIdentifier
931 must not be marked critical.
932 * The authorityKeyIdentifier must be given for X.509v3 certs
933 unless they are self-signed.
934 * The subjectKeyIdentifier must be given for all X.509v3 CA certs.
938 * Certificate verification using `X509_verify_cert()` meanwhile rejects EC keys
939 with explicit curve parameters (specifiedCurve) as required by RFC 5480.
943 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
944 used even when parsing explicit parameters, when loading a encoded key
945 or calling `EC_GROUP_new_from_ecpkparameters()`/
946 `EC_GROUP_new_from_ecparameters()`.
947 This prevents bypass of security hardening and performance gains,
948 especially for curves with specialized EC_METHODs.
949 By default, if a key encoded with explicit parameters is loaded and later
950 encoded, the output is still encoded with explicit parameters, even if
951 internally a "named" EC_GROUP is used for computation.
955 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
956 this change, EC_GROUP_set_generator would accept order and/or cofactor as
957 NULL. After this change, only the cofactor parameter can be NULL. It also
958 does some minimal sanity checks on the passed order.
963 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
964 An attack is simple, if the first CMS_recipientInfo is valid but the
965 second CMS_recipientInfo is chosen ciphertext. If the second
966 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
967 encryption key will be replaced by garbage, and the message cannot be
968 decoded, but if the RSA decryption fails, the correct encryption key is
969 used and the recipient will not notice the attack.
970 As a work around for this potential attack the length of the decrypted
971 key must be equal to the cipher default key length, in case the
972 certifiate is not given and all recipientInfo are tried out.
973 The old behaviour can be re-enabled in the CMS code by setting the
974 CMS_DEBUG_DECRYPT flag.
978 * Early start up entropy quality from the DEVRANDOM seed source has been
979 improved for older Linux systems. The RAND subsystem will wait for
980 /dev/random to be producing output before seeding from /dev/urandom.
981 The seeded state is stored for future library initialisations using
982 a system global shared memory segment. The shared memory identifier
983 can be configured by defining OPENSSL_RAND_SEED_DEVRANDOM_SHM_ID to
984 the desired value. The default identifier is 114.
988 * Revised BN_generate_prime_ex to not avoid factors 2..17863 in p-1
989 when primes for RSA keys are computed.
990 Since we previously always generated primes == 2 (mod 3) for RSA keys,
991 the 2-prime and 3-prime RSA modules were easy to distinguish, since
992 `N = p*q = 1 (mod 3)`, but `N = p*q*r = 2 (mod 3)`. Therefore fingerprinting
993 2-prime vs. 3-prime RSA keys was possible by computing N mod 3.
994 This avoids possible fingerprinting of newly generated RSA modules.
998 * Correct the extended master secret constant on EBCDIC systems. Without this
999 fix TLS connections between an EBCDIC system and a non-EBCDIC system that
1000 negotiate EMS will fail. Unfortunately this also means that TLS connections
1001 between EBCDIC systems with this fix, and EBCDIC systems without this
1002 fix will fail if they negotiate EMS.
1006 * Changed the library initialisation so that the config file is now loaded
1007 by default. This was already the case for libssl. It now occurs for both
1008 libcrypto and libssl. Use the OPENSSL_INIT_NO_LOAD_CONFIG option to
1009 `OPENSSL_init_crypto()` to suppress automatic loading of a config file.
1013 * Introduced new error raising macros, `ERR_raise()` and `ERR_raise_data()`,
1014 where the former acts as a replacement for `ERR_put_error()`, and the
1015 latter replaces the combination `ERR_put_error()` + `ERR_add_error_data()`.
1016 `ERR_raise_data()` adds more flexibility by taking a format string and
1017 an arbitrary number of arguments following it, to be processed with
1022 * Introduced a new function, `OSSL_PROVIDER_available()`, which can be used
1023 to check if a named provider is loaded and available. When called, it
1024 will also activate all fallback providers if such are still present.
1028 * Enforce a minimum DH modulus size of 512 bits.
1032 * Changed DH parameters to generate the order q subgroup instead of 2q.
1033 Previously generated DH parameters are still accepted by DH_check
1034 but DH_generate_key works around that by clearing bit 0 of the
1035 private key for those. This avoids leaking bit 0 of the private key.
1039 * Significantly reduce secure memory usage by the randomness pools.
1043 * `{CRYPTO,OPENSSL}_mem_debug_{push,pop}` are now no-ops and have been
1048 * A new type, EVP_KEYEXCH, has been introduced to represent key exchange
1049 algorithms. An implementation of a key exchange algorithm can be obtained
1050 by using the function EVP_KEYEXCH_fetch(). An EVP_KEYEXCH algorithm can be
1051 used in a call to EVP_PKEY_derive_init_ex() which works in a similar way to
1052 the older EVP_PKEY_derive_init() function. See the man pages for the new
1053 functions for further details.
1057 * The EVP_PKEY_CTX_set_dh_pad() macro has now been converted to a function.
1061 * Removed the function names from error messages and deprecated the
1064 * Removed NextStep support and the macro OPENSSL_UNISTD
1068 * Removed DES_check_key. Also removed OPENSSL_IMPLEMENT_GLOBAL,
1069 OPENSSL_GLOBAL_REF, OPENSSL_DECLARE_GLOBAL.
1070 Also removed "export var as function" capability; we do not export
1071 variables, only functions.
1075 * RC5_32_set_key has been changed to return an int type, with 0 indicating
1076 an error and 1 indicating success. In previous versions of OpenSSL this
1077 was a void type. If a key was set longer than the maximum possible this
1082 * Support SM2 signing and verification schemes with X509 certificate.
1086 * Use SHA256 as the default digest for TS query in the `ts` app.
1090 * Change PBKDF2 to conform to SP800-132 instead of the older PKCS5 RFC2898.
1094 * Default cipher lists/suites are now available via a function, the
1095 #defines are deprecated.
1099 * Add target VC-WIN32-UWP, VC-WIN64A-UWP, VC-WIN32-ARM-UWP and
1100 VC-WIN64-ARM-UWP in Windows OneCore target for making building libraries
1101 for Windows Store apps easier. Also, the "no-uplink" option has been added.
1105 * Join the directories crypto/x509 and crypto/x509v3
1109 * Added command 'openssl kdf' that uses the EVP_KDF API.
1113 * Added command 'openssl mac' that uses the EVP_MAC API.
1117 * Added OPENSSL_info() to get diverse built-in OpenSSL data, such
1118 as default directories. Also added the command 'openssl info'
1119 for scripting purposes.
1123 * The functions AES_ige_encrypt() and AES_bi_ige_encrypt() have been
1128 * Add prediction resistance to the DRBG reseeding process.
1132 * Limit the number of blocks in a data unit for AES-XTS to 2^20 as
1133 mandated by IEEE Std 1619-2018.
1137 * Added newline escaping functionality to a filename when using openssl dgst.
1138 This output format is to replicate the output format found in the `*sum`
1139 checksum programs. This aims to preserve backward compatibility.
1141 *Matt Eaton, Richard Levitte, and Paul Dale*
1143 * Removed the heartbeat message in DTLS feature, as it has very
1144 little usage and doesn't seem to fulfill a valuable purpose.
1145 The configuration option is now deprecated.
1149 * Changed the output of 'openssl {digestname} < file' to display the
1150 digest name in its output.
1154 * Added a new generic trace API which provides support for enabling
1155 instrumentation through trace output.
1157 *Richard Levitte & Matthias St. Pierre*
1159 * Added build tests for C++. These are generated files that only do one
1160 thing, to include one public OpenSSL head file each. This tests that
1161 the public header files can be usefully included in a C++ application.
1163 This test isn't enabled by default. It can be enabled with the option
1164 'enable-buildtest-c++'.
1168 * Add Single Step KDF (EVP_KDF_SS) to EVP_KDF.
1172 * Add KMAC to EVP_MAC.
1176 * Added property based algorithm implementation selection framework to
1181 * Added SCA hardening for modular field inversion in EC_GROUP through
1182 a new dedicated field_inv() pointer in EC_METHOD.
1183 This also addresses a leakage affecting conversions from projective
1184 to affine coordinates.
1186 *Billy Bob Brumley, Nicola Tuveri*
1188 * Added EVP_KDF, an EVP layer KDF API, to simplify adding KDF and PRF
1189 implementations. This includes an EVP_PKEY to EVP_KDF bridge for
1190 those algorithms that were already supported through the EVP_PKEY API
1191 (scrypt, TLS1 PRF and HKDF). The low-level KDF functions for PBKDF2
1192 and scrypt are now wrappers that call EVP_KDF.
1196 * Build devcrypto engine as a dynamic engine.
1198 *Eneas U de Queiroz*
1200 * Add keyed BLAKE2 to EVP_MAC.
1204 * Fix a bug in the computation of the endpoint-pair shared secret used
1205 by DTLS over SCTP. This breaks interoperability with older versions
1206 of OpenSSL like OpenSSL 1.1.0 and OpenSSL 1.0.2. There is a runtime
1207 switch SSL_MODE_DTLS_SCTP_LABEL_LENGTH_BUG (off by default) enabling
1208 interoperability with such broken implementations. However, enabling
1209 this switch breaks interoperability with correct implementations.
1211 * Fix a use after free bug in d2i_X509_PUBKEY when overwriting a
1212 re-used X509_PUBKEY object if the second PUBKEY is malformed.
1216 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
1220 * Changed the license to the Apache License v2.0.
1224 * Switch to a new version scheme using three numbers MAJOR.MINOR.PATCH.
1226 - Major releases (indicated by incrementing the MAJOR release number)
1227 may introduce incompatible API/ABI changes.
1228 - Minor releases (indicated by incrementing the MINOR release number)
1229 may introduce new features but retain API/ABI compatibility.
1230 - Patch releases (indicated by incrementing the PATCH number)
1231 are intended for bug fixes and other improvements of existing
1232 features only (like improving performance or adding documentation)
1233 and retain API/ABI compatibility.
1237 * Add support for RFC5297 SIV mode (siv128), including AES-SIV.
1241 * Remove the 'dist' target and add a tarball building script. The
1242 'dist' target has fallen out of use, and it shouldn't be
1243 necessary to configure just to create a source distribution.
1247 * Recreate the OS390-Unix config target. It no longer relies on a
1248 special script like it did for OpenSSL pre-1.1.0.
1252 * Instead of having the source directories listed in Configure, add
1253 a 'build.info' keyword SUBDIRS to indicate what sub-directories to
1258 * Add GMAC to EVP_MAC.
1262 * Ported the HMAC, CMAC and SipHash EVP_PKEY_METHODs to EVP_MAC.
1266 * Added EVP_MAC, an EVP layer MAC API, to simplify adding MAC
1267 implementations. This includes a generic EVP_PKEY to EVP_MAC bridge,
1268 to facilitate the continued use of MACs through raw private keys in
1269 functionality such as `EVP_DigestSign*` and `EVP_DigestVerify*`.
1273 * Deprecate ECDH_KDF_X9_62().
1277 * Added EVP_PKEY_ECDH_KDF_X9_63 and ecdh_KDF_X9_63() as replacements for
1278 the EVP_PKEY_ECDH_KDF_X9_62 KDF type and ECDH_KDF_X9_62(). The old names
1279 are retained for backwards compatibility.
1283 * AES-XTS mode now enforces that its two keys are different to mitigate
1284 the attacked described in "Efficient Instantiations of Tweakable
1285 Blockciphers and Refinements to Modes OCB and PMAC" by Phillip Rogaway.
1286 Details of this attack can be obtained from:
1287 <http://web.cs.ucdavis.edu/%7Erogaway/papers/offsets.pdf>
1291 * Rename the object files, i.e. give them other names than in previous
1292 versions. Their names now include the name of the final product, as
1293 well as its type mnemonic (bin, lib, shlib).
1297 * Added new option for 'openssl list', '-objects', which will display the
1298 list of built in objects, i.e. OIDs with names.
1302 * Added the options `-crl_lastupdate` and `-crl_nextupdate` to `openssl ca`,
1303 allowing the `lastUpdate` and `nextUpdate` fields in the generated CRL to
1308 * Added support for Linux Kernel TLS data-path. The Linux Kernel data-path
1309 improves application performance by removing data copies and providing
1310 applications with zero-copy system calls such as sendfile and splice.
1314 * The SSL option SSL_OP_CLEANSE_PLAINTEXT is introduced.
1318 * `PKCS12_parse` now maintains the order of the parsed certificates
1319 when outputting them via `*ca` (rather than reversing it).
1323 * Deprecated pthread fork support methods.
1327 * Added support for FFDHE key exchange in TLS 1.3.
1334 ### Changes between 1.1.1j and 1.1.1k [xx XXX xxxx]
1336 * Fixed a problem with verifying a certificate chain when using the
1337 X509_V_FLAG_X509_STRICT flag. This flag enables additional security checks of
1338 the certificates present in a certificate chain. It is not set by default.
1340 Starting from OpenSSL version 1.1.1h a check to disallow certificates in
1341 the chain that have explicitly encoded elliptic curve parameters was added
1342 as an additional strict check.
1344 An error in the implementation of this check meant that the result of a
1345 previous check to confirm that certificates in the chain are valid CA
1346 certificates was overwritten. This effectively bypasses the check
1347 that non-CA certificates must not be able to issue other certificates.
1349 If a "purpose" has been configured then there is a subsequent opportunity
1350 for checks that the certificate is a valid CA. All of the named "purpose"
1351 values implemented in libcrypto perform this check. Therefore, where
1352 a purpose is set the certificate chain will still be rejected even when the
1353 strict flag has been used. A purpose is set by default in libssl client and
1354 server certificate verification routines, but it can be overridden or
1355 removed by an application.
1357 In order to be affected, an application must explicitly set the
1358 X509_V_FLAG_X509_STRICT verification flag and either not set a purpose
1359 for the certificate verification or, in the case of TLS client or server
1360 applications, override the default purpose.
1365 * Fixed an issue where an OpenSSL TLS server may crash if sent a maliciously
1366 crafted renegotiation ClientHello message from a client. If a TLSv1.2
1367 renegotiation ClientHello omits the signature_algorithms extension (where it
1368 was present in the initial ClientHello), but includes a
1369 signature_algorithms_cert extension then a NULL pointer dereference will
1370 result, leading to a crash and a denial of service attack.
1372 A server is only vulnerable if it has TLSv1.2 and renegotiation enabled
1373 (which is the default configuration). OpenSSL TLS clients are not impacted by
1377 *Peter Kästle and Samuel Sapalski*
1379 ### Changes between 1.1.1i and 1.1.1j [16 Feb 2021]
1381 * Fixed the X509_issuer_and_serial_hash() function. It attempts to
1382 create a unique hash value based on the issuer and serial number data
1383 contained within an X509 certificate. However it was failing to correctly
1384 handle any errors that may occur while parsing the issuer field (which might
1385 occur if the issuer field is maliciously constructed). This may subsequently
1386 result in a NULL pointer deref and a crash leading to a potential denial of
1392 * Fixed the RSA_padding_check_SSLv23() function and the RSA_SSLV23_PADDING
1393 padding mode to correctly check for rollback attacks. This is considered a
1394 bug in OpenSSL 1.1.1 because it does not support SSLv2. In 1.0.2 this is
1399 Fixed the EVP_CipherUpdate, EVP_EncryptUpdate and EVP_DecryptUpdate
1400 functions. Previously they could overflow the output length argument in some
1401 cases where the input length is close to the maximum permissable length for
1402 an integer on the platform. In such cases the return value from the function
1403 call would be 1 (indicating success), but the output length value would be
1404 negative. This could cause applications to behave incorrectly or crash.
1409 * Fixed SRP_Calc_client_key so that it runs in constant time. The previous
1410 implementation called BN_mod_exp without setting BN_FLG_CONSTTIME. This
1411 could be exploited in a side channel attack to recover the password. Since
1412 the attack is local host only this is outside of the current OpenSSL
1413 threat model and therefore no CVE is assigned.
1415 Thanks to Mohammed Sabt and Daniel De Almeida Braga for reporting this
1420 ### Changes between 1.1.1h and 1.1.1i [8 Dec 2020]
1422 * Fixed NULL pointer deref in the GENERAL_NAME_cmp function
1423 This function could crash if both GENERAL_NAMEs contain an EDIPARTYNAME.
1424 If an attacker can control both items being compared then this could lead
1425 to a possible denial of service attack. OpenSSL itself uses the
1426 GENERAL_NAME_cmp function for two purposes:
1427 1) Comparing CRL distribution point names between an available CRL and a
1428 CRL distribution point embedded in an X509 certificate
1429 2) When verifying that a timestamp response token signer matches the
1430 timestamp authority name (exposed via the API functions
1431 TS_RESP_verify_response and TS_RESP_verify_token)
1436 ### Changes between 1.1.1g and 1.1.1h [22 Sep 2020]
1438 * Certificates with explicit curve parameters are now disallowed in
1439 verification chains if the X509_V_FLAG_X509_STRICT flag is used.
1443 * The 'MinProtocol' and 'MaxProtocol' configuration commands now silently
1444 ignore TLS protocol version bounds when configuring DTLS-based contexts, and
1445 conversely, silently ignore DTLS protocol version bounds when configuring
1446 TLS-based contexts. The commands can be repeated to set bounds of both
1447 types. The same applies with the corresponding "min_protocol" and
1448 "max_protocol" command-line switches, in case some application uses both TLS
1451 SSL_CTX instances that are created for a fixed protocol version (e.g.
1452 TLSv1_server_method()) also silently ignore version bounds. Previously
1453 attempts to apply bounds to these protocol versions would result in an
1454 error. Now only the "version-flexible" SSL_CTX instances are subject to
1455 limits in configuration files in command-line options.
1459 * Handshake now fails if Extended Master Secret extension is dropped
1464 * The Oracle Developer Studio compiler will start reporting deprecated APIs
1466 ### Changes between 1.1.1f and 1.1.1g [21 Apr 2020]
1468 * Fixed segmentation fault in SSL_check_chain()
1469 Server or client applications that call the SSL_check_chain() function
1470 during or after a TLS 1.3 handshake may crash due to a NULL pointer
1471 dereference as a result of incorrect handling of the
1472 "signature_algorithms_cert" TLS extension. The crash occurs if an invalid
1473 or unrecognised signature algorithm is received from the peer. This could
1474 be exploited by a malicious peer in a Denial of Service attack.
1479 * Added AES consttime code for no-asm configurations
1480 an optional constant time support for AES was added
1481 when building openssl for no-asm.
1482 Enable with: ./config no-asm -DOPENSSL_AES_CONST_TIME
1483 Disable with: ./config no-asm -DOPENSSL_NO_AES_CONST_TIME
1484 At this time this feature is by default disabled.
1485 It will be enabled by default in 3.0.
1489 ### Changes between 1.1.1e and 1.1.1f [31 Mar 2020]
1491 * Revert the change of EOF detection while reading in libssl to avoid
1492 regressions in applications depending on the current way of reporting
1493 the EOF. As the existing method is not fully accurate the change to
1494 reporting the EOF via SSL_ERROR_SSL is kept on the current development
1495 branch and will be present in the 3.0 release.
1499 * Revised BN_generate_prime_ex to not avoid factors 3..17863 in p-1
1500 when primes for RSA keys are computed.
1501 Since we previously always generated primes == 2 (mod 3) for RSA keys,
1502 the 2-prime and 3-prime RSA modules were easy to distinguish, since
1503 N = p*q = 1 (mod 3), but N = p*q*r = 2 (mod 3). Therefore fingerprinting
1504 2-prime vs. 3-prime RSA keys was possible by computing N mod 3.
1505 This avoids possible fingerprinting of newly generated RSA modules.
1509 ### Changes between 1.1.1d and 1.1.1e [17 Mar 2020]
1511 * Properly detect EOF while reading in libssl. Previously if we hit an EOF
1512 while reading in libssl then we would report an error back to the
1513 application (SSL_ERROR_SYSCALL) but errno would be 0. We now add
1514 an error to the stack (which means we instead return SSL_ERROR_SSL) and
1515 therefore give a hint as to what went wrong.
1519 * Check that ed25519 and ed448 are allowed by the security level. Previously
1520 signature algorithms not using an MD were not being checked that they were
1521 allowed by the security level.
1525 * Fixed SSL_get_servername() behaviour. The behaviour of SSL_get_servername()
1526 was not quite right. The behaviour was not consistent between resumption
1527 and normal handshakes, and also not quite consistent with historical
1528 behaviour. The behaviour in various scenarios has been clarified and
1529 it has been updated to make it match historical behaviour as closely as
1534 * *[VMS only]* The header files that the VMS compilers include automatically,
1535 `__DECC_INCLUDE_PROLOGUE.H` and `__DECC_INCLUDE_EPILOGUE.H`, use pragmas
1536 that the C++ compiler doesn't understand. This is a shortcoming in the
1537 compiler, but can be worked around with `__cplusplus` guards.
1539 C++ applications that use OpenSSL libraries must be compiled using the
1540 qualifier `/NAMES=(AS_IS,SHORTENED)` to be able to use all the OpenSSL
1541 functions. Otherwise, only functions with symbols of less than 31
1542 characters can be used, as the linker will not be able to successfully
1543 resolve symbols with longer names.
1547 * Added a new method to gather entropy on VMS, based on SYS$GET_ENTROPY.
1548 The presence of this system service is determined at run-time.
1552 * Added newline escaping functionality to a filename when using openssl dgst.
1553 This output format is to replicate the output format found in the `*sum`
1554 checksum programs. This aims to preserve backward compatibility.
1556 *Matt Eaton, Richard Levitte, and Paul Dale*
1558 * Print all values for a PKCS#12 attribute with 'openssl pkcs12', not just
1563 ### Changes between 1.1.1c and 1.1.1d [10 Sep 2019]
1565 * Fixed a fork protection issue. OpenSSL 1.1.1 introduced a rewritten random
1566 number generator (RNG). This was intended to include protection in the
1567 event of a fork() system call in order to ensure that the parent and child
1568 processes did not share the same RNG state. However this protection was not
1569 being used in the default case.
1571 A partial mitigation for this issue is that the output from a high
1572 precision timer is mixed into the RNG state so the likelihood of a parent
1573 and child process sharing state is significantly reduced.
1575 If an application already calls OPENSSL_init_crypto() explicitly using
1576 OPENSSL_INIT_ATFORK then this problem does not occur at all.
1579 *Matthias St. Pierre*
1581 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
1582 used even when parsing explicit parameters, when loading a encoded key
1583 or calling `EC_GROUP_new_from_ecpkparameters()`/
1584 `EC_GROUP_new_from_ecparameters()`.
1585 This prevents bypass of security hardening and performance gains,
1586 especially for curves with specialized EC_METHODs.
1587 By default, if a key encoded with explicit parameters is loaded and later
1588 encoded, the output is still encoded with explicit parameters, even if
1589 internally a "named" EC_GROUP is used for computation.
1593 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
1594 this change, EC_GROUP_set_generator would accept order and/or cofactor as
1595 NULL. After this change, only the cofactor parameter can be NULL. It also
1596 does some minimal sanity checks on the passed order.
1601 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
1602 An attack is simple, if the first CMS_recipientInfo is valid but the
1603 second CMS_recipientInfo is chosen ciphertext. If the second
1604 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
1605 encryption key will be replaced by garbage, and the message cannot be
1606 decoded, but if the RSA decryption fails, the correct encryption key is
1607 used and the recipient will not notice the attack.
1608 As a work around for this potential attack the length of the decrypted
1609 key must be equal to the cipher default key length, in case the
1610 certifiate is not given and all recipientInfo are tried out.
1611 The old behaviour can be re-enabled in the CMS code by setting the
1612 CMS_DEBUG_DECRYPT flag.
1617 * Early start up entropy quality from the DEVRANDOM seed source has been
1618 improved for older Linux systems. The RAND subsystem will wait for
1619 /dev/random to be producing output before seeding from /dev/urandom.
1620 The seeded state is stored for future library initialisations using
1621 a system global shared memory segment. The shared memory identifier
1622 can be configured by defining OPENSSL_RAND_SEED_DEVRANDOM_SHM_ID to
1623 the desired value. The default identifier is 114.
1627 * Correct the extended master secret constant on EBCDIC systems. Without this
1628 fix TLS connections between an EBCDIC system and a non-EBCDIC system that
1629 negotiate EMS will fail. Unfortunately this also means that TLS connections
1630 between EBCDIC systems with this fix, and EBCDIC systems without this
1631 fix will fail if they negotiate EMS.
1635 * Use Windows installation paths in the mingw builds
1637 Mingw isn't a POSIX environment per se, which means that Windows
1638 paths should be used for installation.
1643 * Changed DH_check to accept parameters with order q and 2q subgroups.
1644 With order 2q subgroups the bit 0 of the private key is not secret
1645 but DH_generate_key works around that by clearing bit 0 of the
1646 private key for those. This avoids leaking bit 0 of the private key.
1650 * Significantly reduce secure memory usage by the randomness pools.
1654 * Revert the DEVRANDOM_WAIT feature for Linux systems
1656 The DEVRANDOM_WAIT feature added a select() call to wait for the
1657 /dev/random device to become readable before reading from the
1658 /dev/urandom device.
1660 It turned out that this change had negative side effects on
1661 performance which were not acceptable. After some discussion it
1662 was decided to revert this feature and leave it up to the OS
1663 resp. the platform maintainer to ensure a proper initialization
1664 during early boot time.
1666 *Matthias St. Pierre*
1668 ### Changes between 1.1.1b and 1.1.1c [28 May 2019]
1670 * Add build tests for C++. These are generated files that only do one
1671 thing, to include one public OpenSSL head file each. This tests that
1672 the public header files can be usefully included in a C++ application.
1674 This test isn't enabled by default. It can be enabled with the option
1675 'enable-buildtest-c++'.
1679 * Enable SHA3 pre-hashing for ECDSA and DSA.
1683 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
1684 This changes the size when using the `genpkey` command when no size is given.
1685 It fixes an omission in earlier changes that changed all RSA, DSA and DH
1686 generation commands to use 2048 bits by default.
1690 * Reorganize the manual pages to consistently have RETURN VALUES,
1691 EXAMPLES, SEE ALSO and HISTORY come in that order, and adjust
1692 util/fix-doc-nits accordingly.
1694 *Paul Yang, Joshua Lock*
1696 * Add the missing accessor EVP_PKEY_get0_engine()
1700 * Have commands like `s_client` and `s_server` output the signature scheme
1701 along with other cipher suite parameters when debugging.
1705 * Make OPENSSL_config() error agnostic again.
1709 * Do the error handling in RSA decryption constant time.
1713 * Prevent over long nonces in ChaCha20-Poly1305.
1715 ChaCha20-Poly1305 is an AEAD cipher, and requires a unique nonce input
1716 for every encryption operation. RFC 7539 specifies that the nonce value
1717 (IV) should be 96 bits (12 bytes). OpenSSL allows a variable nonce length
1718 and front pads the nonce with 0 bytes if it is less than 12
1719 bytes. However it also incorrectly allows a nonce to be set of up to 16
1720 bytes. In this case only the last 12 bytes are significant and any
1721 additional leading bytes are ignored.
1723 It is a requirement of using this cipher that nonce values are
1724 unique. Messages encrypted using a reused nonce value are susceptible to
1725 serious confidentiality and integrity attacks. If an application changes
1726 the default nonce length to be longer than 12 bytes and then makes a
1727 change to the leading bytes of the nonce expecting the new value to be a
1728 new unique nonce then such an application could inadvertently encrypt
1729 messages with a reused nonce.
1731 Additionally the ignored bytes in a long nonce are not covered by the
1732 integrity guarantee of this cipher. Any application that relies on the
1733 integrity of these ignored leading bytes of a long nonce may be further
1734 affected. Any OpenSSL internal use of this cipher, including in SSL/TLS,
1735 is safe because no such use sets such a long nonce value. However user
1736 applications that use this cipher directly and set a non-default nonce
1737 length to be longer than 12 bytes may be vulnerable.
1739 This issue was reported to OpenSSL on 16th of March 2019 by Joran Dirk
1745 * Add DEVRANDOM_WAIT feature for Linux systems
1747 On older Linux systems where the getrandom() system call is not available,
1748 OpenSSL normally uses the /dev/urandom device for seeding its CSPRNG.
1749 Contrary to getrandom(), the /dev/urandom device will not block during
1750 early boot when the kernel CSPRNG has not been seeded yet.
1752 To mitigate this known weakness, use select() to wait for /dev/random to
1753 become readable before reading from /dev/urandom.
1755 * Ensure that SM2 only uses SM3 as digest algorithm
1759 ### Changes between 1.1.1a and 1.1.1b [26 Feb 2019]
1761 * Change the info callback signals for the start and end of a post-handshake
1762 message exchange in TLSv1.3. In 1.1.1/1.1.1a we used SSL_CB_HANDSHAKE_START
1763 and SSL_CB_HANDSHAKE_DONE. Experience has shown that many applications get
1764 confused by this and assume that a TLSv1.2 renegotiation has started. This
1765 can break KeyUpdate handling. Instead we no longer signal the start and end
1766 of a post handshake message exchange (although the messages themselves are
1767 still signalled). This could break some applications that were expecting
1768 the old signals. However without this KeyUpdate is not usable for many
1773 ### Changes between 1.1.1 and 1.1.1a [20 Nov 2018]
1775 * Timing vulnerability in DSA signature generation
1777 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
1778 timing side channel attack. An attacker could use variations in the signing
1779 algorithm to recover the private key.
1781 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
1786 * Timing vulnerability in ECDSA signature generation
1788 The OpenSSL ECDSA signature algorithm has been shown to be vulnerable to a
1789 timing side channel attack. An attacker could use variations in the signing
1790 algorithm to recover the private key.
1792 This issue was reported to OpenSSL on 25th October 2018 by Samuel Weiser.
1797 * Fixed the issue that RAND_add()/RAND_seed() silently discards random input
1798 if its length exceeds 4096 bytes. The limit has been raised to a buffer size
1799 of two gigabytes and the error handling improved.
1801 This issue was reported to OpenSSL by Dr. Falko Strenzke. It has been
1802 categorized as a normal bug, not a security issue, because the DRBG reseeds
1803 automatically and is fully functional even without additional randomness
1804 provided by the application.
1806 ### Changes between 1.1.0i and 1.1.1 [11 Sep 2018]
1808 * Add a new ClientHello callback. Provides a callback interface that gives
1809 the application the ability to adjust the nascent SSL object at the
1810 earliest stage of ClientHello processing, immediately after extensions have
1811 been collected but before they have been processed. In particular, this
1812 callback can adjust the supported TLS versions in response to the contents
1817 * Add SM2 base algorithm support.
1821 * s390x assembly pack: add (improved) hardware-support for the following
1822 cryptographic primitives: sha3, shake, aes-gcm, aes-ccm, aes-ctr, aes-ofb,
1823 aes-cfb/cfb8, aes-ecb.
1827 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
1828 parameter is no longer accepted, as it leads to a corrupt table. NULL
1829 pem_str is reserved for alias entries only.
1833 * Use the new ec_scalar_mul_ladder scaffold to implement a specialized ladder
1834 step for prime curves. The new implementation is based on formulae from
1835 differential addition-and-doubling in homogeneous projective coordinates
1836 from Izu-Takagi "A fast parallel elliptic curve multiplication resistant
1837 against side channel attacks" and Brier-Joye "Weierstrass Elliptic Curves
1838 and Side-Channel Attacks" Eq. (8) for y-coordinate recovery, modified
1839 to work in projective coordinates.
1841 *Billy Bob Brumley, Nicola Tuveri*
1843 * Change generating and checking of primes so that the error rate of not
1844 being prime depends on the intended use based on the size of the input.
1845 For larger primes this will result in more rounds of Miller-Rabin.
1846 The maximal error rate for primes with more than 1080 bits is lowered
1849 *Kurt Roeckx, Annie Yousar*
1851 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
1855 * The 'tsget' script is renamed to 'tsget.pl', to avoid confusion when
1856 moving between systems, and to avoid confusion when a Windows build is
1857 done with mingw vs with MSVC. For POSIX installs, there's still a
1858 symlink or copy named 'tsget' to avoid that confusion as well.
1862 * Revert blinding in ECDSA sign and instead make problematic addition
1863 length-invariant. Switch even to fixed-length Montgomery multiplication.
1867 * Use the new ec_scalar_mul_ladder scaffold to implement a specialized ladder
1868 step for binary curves. The new implementation is based on formulae from
1869 differential addition-and-doubling in mixed Lopez-Dahab projective
1870 coordinates, modified to independently blind the operands.
1872 *Billy Bob Brumley, Sohaib ul Hassan, Nicola Tuveri*
1874 * Add a scaffold to optionally enhance the Montgomery ladder implementation
1875 for `ec_scalar_mul_ladder` (formerly `ec_mul_consttime`) allowing
1876 EC_METHODs to implement their own specialized "ladder step", to take
1877 advantage of more favorable coordinate systems or more efficient
1878 differential addition-and-doubling algorithms.
1880 *Billy Bob Brumley, Sohaib ul Hassan, Nicola Tuveri*
1882 * Modified the random device based seed sources to keep the relevant
1883 file descriptors open rather than reopening them on each access.
1884 This allows such sources to operate in a chroot() jail without
1885 the associated device nodes being available. This behaviour can be
1886 controlled using RAND_keep_random_devices_open().
1890 * Numerous side-channel attack mitigations have been applied. This may have
1891 performance impacts for some algorithms for the benefit of improved
1892 security. Specific changes are noted in this change log by their respective
1897 * AIX shared library support overhaul. Switch to AIX "natural" way of
1898 handling shared libraries, which means collecting shared objects of
1899 different versions and bitnesses in one common archive. This allows to
1900 mitigate conflict between 1.0 and 1.1 side-by-side installations. It
1901 doesn't affect the way 3rd party applications are linked, only how
1902 multi-version installation is managed.
1906 * Make ec_group_do_inverse_ord() more robust and available to other
1907 EC cryptosystems, so that irrespective of BN_FLG_CONSTTIME, SCA
1908 mitigations are applied to the fallback BN_mod_inverse().
1909 When using this function rather than BN_mod_inverse() directly, new
1910 EC cryptosystem implementations are then safer-by-default.
1914 * Add coordinate blinding for EC_POINT and implement projective
1915 coordinate blinding for generic prime curves as a countermeasure to
1916 chosen point SCA attacks.
1918 *Sohaib ul Hassan, Nicola Tuveri, Billy Bob Brumley*
1920 * Add blinding to ECDSA and DSA signatures to protect against side channel
1921 attacks discovered by Keegan Ryan (NCC Group).
1925 * Enforce checking in the `pkeyutl` command to ensure that the input
1926 length does not exceed the maximum supported digest length when performing
1927 a sign, verify or verifyrecover operation.
1931 * SSL_MODE_AUTO_RETRY is enabled by default. Applications that use blocking
1932 I/O in combination with something like select() or poll() will hang. This
1933 can be turned off again using SSL_CTX_clear_mode().
1934 Many applications do not properly handle non-application data records, and
1935 TLS 1.3 sends more of such records. Setting SSL_MODE_AUTO_RETRY works
1936 around the problems in those applications, but can also break some.
1937 It's recommended to read the manpages about SSL_read(), SSL_write(),
1938 SSL_get_error(), SSL_shutdown(), SSL_CTX_set_mode() and
1939 SSL_CTX_set_read_ahead() again.
1943 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
1944 now allow empty (zero character) pass phrases.
1948 * Apply blinding to binary field modular inversion and remove patent
1949 pending (OPENSSL_SUN_GF2M_DIV) BN_GF2m_mod_div implementation.
1953 * Deprecate ec2_mult.c and unify scalar multiplication code paths for
1954 binary and prime elliptic curves.
1958 * Remove ECDSA nonce padding: EC_POINT_mul is now responsible for
1959 constant time fixed point multiplication.
1963 * Revise elliptic curve scalar multiplication with timing attack
1964 defenses: ec_wNAF_mul redirects to a constant time implementation
1965 when computing fixed point and variable point multiplication (which
1966 in OpenSSL are mostly used with secret scalars in keygen, sign,
1967 ECDH derive operations).
1968 *Billy Bob Brumley, Nicola Tuveri, Cesar Pereida García,
1971 * Updated CONTRIBUTING
1975 * Updated DRBG / RAND to request nonce and additional low entropy
1976 randomness from the system.
1978 *Matthias St. Pierre*
1980 * Updated 'openssl rehash' to use OpenSSL consistent default.
1984 * Moved the load of the ssl_conf module to libcrypto, which helps
1985 loading engines that libssl uses before libssl is initialised.
1989 * Added EVP_PKEY_sign() and EVP_PKEY_verify() for EdDSA
1993 * Fixed X509_NAME_ENTRY_set to get multi-valued RDNs right in all cases.
1995 *Ingo Schwarze, Rich Salz*
1997 * Added output of accepting IP address and port for 'openssl s_server'
2001 * Added a new API for TLSv1.3 ciphersuites:
2002 SSL_CTX_set_ciphersuites()
2003 SSL_set_ciphersuites()
2007 * Memory allocation failures consistently add an error to the error
2012 * Don't use OPENSSL_ENGINES and OPENSSL_CONF environment values
2013 in libcrypto when run as setuid/setgid.
2017 * Load any config file by default when libssl is used.
2021 * Added new public header file <openssl/rand_drbg.h> and documentation
2022 for the RAND_DRBG API. See manual page RAND_DRBG(7) for an overview.
2024 *Matthias St. Pierre*
2026 * QNX support removed (cannot find contributors to get their approval
2027 for the license change).
2031 * TLSv1.3 replay protection for early data has been implemented. See the
2032 SSL_read_early_data() man page for further details.
2036 * Separated TLSv1.3 ciphersuite configuration out from TLSv1.2 ciphersuite
2037 configuration. TLSv1.3 ciphersuites are not compatible with TLSv1.2 and
2038 below. Similarly TLSv1.2 ciphersuites are not compatible with TLSv1.3.
2039 In order to avoid issues where legacy TLSv1.2 ciphersuite configuration
2040 would otherwise inadvertently disable all TLSv1.3 ciphersuites the
2041 configuration has been separated out. See the ciphers man page or the
2042 SSL_CTX_set_ciphersuites() man page for more information.
2046 * On POSIX (BSD, Linux, ...) systems the ocsp(1) command running
2047 in responder mode now supports the new "-multi" option, which
2048 spawns the specified number of child processes to handle OCSP
2049 requests. The "-timeout" option now also limits the OCSP
2050 responder's patience to wait to receive the full client request
2051 on a newly accepted connection. Child processes are respawned
2052 as needed, and the CA index file is automatically reloaded
2053 when changed. This makes it possible to run the "ocsp" responder
2054 as a long-running service, making the OpenSSL CA somewhat more
2055 feature-complete. In this mode, most diagnostic messages logged
2056 after entering the event loop are logged via syslog(3) rather than
2061 * Added support for X448 and Ed448. Heavily based on original work by
2066 * Extend OSSL_STORE with capabilities to search and to narrow the set of
2067 objects loaded. This adds the functions OSSL_STORE_expect() and
2068 OSSL_STORE_find() as well as needed tools to construct searches and
2069 get the search data out of them.
2073 * Support for TLSv1.3 added. Note that users upgrading from an earlier
2074 version of OpenSSL should review their configuration settings to ensure
2075 that they are still appropriate for TLSv1.3. For further information see:
2076 <https://wiki.openssl.org/index.php/TLS1.3>
2080 * Grand redesign of the OpenSSL random generator
2082 The default RAND method now utilizes an AES-CTR DRBG according to
2083 NIST standard SP 800-90Ar1. The new random generator is essentially
2084 a port of the default random generator from the OpenSSL FIPS 2.0
2085 object module. It is a hybrid deterministic random bit generator
2086 using an AES-CTR bit stream and which seeds and reseeds itself
2087 automatically using trusted system entropy sources.
2089 Some of its new features are:
2090 - Support for multiple DRBG instances with seed chaining.
2091 - The default RAND method makes use of a DRBG.
2092 - There is a public and private DRBG instance.
2093 - The DRBG instances are fork-safe.
2094 - Keep all global DRBG instances on the secure heap if it is enabled.
2095 - The public and private DRBG instance are per thread for lock free
2098 *Paul Dale, Benjamin Kaduk, Kurt Roeckx, Rich Salz, Matthias St. Pierre*
2100 * Changed Configure so it only says what it does and doesn't dump
2101 so much data. Instead, ./configdata.pm should be used as a script
2102 to display all sorts of configuration data.
2106 * Added processing of "make variables" to Configure.
2110 * Added SHA512/224 and SHA512/256 algorithm support.
2114 * The last traces of Netware support, first removed in 1.1.0, have
2119 * Get rid of Makefile.shared, and in the process, make the processing
2120 of certain files (rc.obj, or the .def/.map/.opt files produced from
2121 the ordinal files) more visible and hopefully easier to trace and
2122 debug (or make silent).
2126 * Make it possible to have environment variable assignments as
2127 arguments to config / Configure.
2131 * Add multi-prime RSA (RFC 8017) support.
2135 * Add SM3 implemented according to GB/T 32905-2016
2136 *Jack Lloyd <jack.lloyd@ribose.com>,*
2137 *Ronald Tse <ronald.tse@ribose.com>,*
2138 *Erick Borsboom <erick.borsboom@ribose.com>*
2140 * Add 'Maximum Fragment Length' TLS extension negotiation and support
2141 as documented in RFC6066.
2142 Based on a patch from Tomasz Moń
2144 *Filipe Raimundo da Silva*
2146 * Add SM4 implemented according to GB/T 32907-2016.
2147 *Jack Lloyd <jack.lloyd@ribose.com>,*
2148 *Ronald Tse <ronald.tse@ribose.com>,*
2149 *Erick Borsboom <erick.borsboom@ribose.com>*
2151 * Reimplement -newreq-nodes and ERR_error_string_n; the
2152 original author does not agree with the license change.
2156 * Add ARIA AEAD TLS support.
2160 * Some macro definitions to support VS6 have been removed. Visual
2161 Studio 6 has not worked since 1.1.0
2165 * Add ERR_clear_last_mark(), to allow callers to clear the last mark
2166 without clearing the errors.
2170 * Add "atfork" functions. If building on a system that without
2171 pthreads, see doc/man3/OPENSSL_fork_prepare.pod for application
2172 requirements. The RAND facility now uses/requires this.
2180 * The UI API becomes a permanent and integral part of libcrypto, i.e.
2181 not possible to disable entirely. However, it's still possible to
2182 disable the console reading UI method, UI_OpenSSL() (use UI_null()
2185 To disable, configure with 'no-ui-console'. 'no-ui' is still
2186 possible to use as an alias. Check at compile time with the
2187 macro OPENSSL_NO_UI_CONSOLE. The macro OPENSSL_NO_UI is still
2188 possible to check and is an alias for OPENSSL_NO_UI_CONSOLE.
2192 * Add a STORE module, which implements a uniform and URI based reader of
2193 stores that can contain keys, certificates, CRLs and numerous other
2194 objects. The main API is loosely based on a few stdio functions,
2195 and includes OSSL_STORE_open, OSSL_STORE_load, OSSL_STORE_eof,
2196 OSSL_STORE_error and OSSL_STORE_close.
2197 The implementation uses backends called "loaders" to implement arbitrary
2198 URI schemes. There is one built in "loader" for the 'file' scheme.
2202 * Add devcrypto engine. This has been implemented against cryptodev-linux,
2203 then adjusted to work on FreeBSD 8.4 as well.
2204 Enable by configuring with 'enable-devcryptoeng'. This is done by default
2205 on BSD implementations, as cryptodev.h is assumed to exist on all of them.
2209 * Module names can prefixed with OSSL_ or OPENSSL_. This affects
2210 util/mkerr.pl, which is adapted to allow those prefixes, leading to
2211 error code calls like this:
2213 OSSL_FOOerr(OSSL_FOO_F_SOMETHING, OSSL_FOO_R_WHATEVER);
2215 With this change, we claim the namespaces OSSL and OPENSSL in a manner
2216 that can be encoded in C. For the foreseeable future, this will only
2219 *Richard Levitte and Tim Hudson*
2221 * Removed BSD cryptodev engine.
2225 * Add a build target 'build_all_generated', to build all generated files
2226 and only that. This can be used to prepare everything that requires
2227 things like perl for a system that lacks perl and then move everything
2228 to that system and do the rest of the build there.
2232 * In the UI interface, make it possible to duplicate the user data. This
2233 can be used by engines that need to retain the data for a longer time
2234 than just the call where this user data is passed.
2238 * Ignore the '-named_curve auto' value for compatibility of applications
2241 *Tomáš Mráz <tmraz@fedoraproject.org>*
2243 * Fragmented SSL/TLS alerts are no longer accepted. An alert message is 2
2244 bytes long. In theory it is permissible in SSLv3 - TLSv1.2 to fragment such
2245 alerts across multiple records (some of which could be empty). In practice
2246 it make no sense to send an empty alert record, or to fragment one. TLSv1.3
2247 prohibits this altogether and other libraries (BoringSSL, NSS) do not
2248 support this at all. Supporting it adds significant complexity to the
2249 record layer, and its removal is unlikely to cause interoperability
2254 * Add the ASN.1 types INT32, UINT32, INT64, UINT64 and variants prefixed
2255 with Z. These are meant to replace LONG and ZLONG and to be size safe.
2256 The use of LONG and ZLONG is discouraged and scheduled for deprecation
2261 * Add the 'z' and 'j' modifiers to BIO_printf() et al formatting string,
2262 'z' is to be used for [s]size_t, and 'j' - with [u]int64_t.
2264 *Richard Levitte, Andy Polyakov*
2266 * Add EC_KEY_get0_engine(), which does for EC_KEY what RSA_get0_engine()
2271 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
2272 platform rather than 'mingw'.
2276 * The functions X509_STORE_add_cert and X509_STORE_add_crl return
2277 success if they are asked to add an object which already exists
2278 in the store. This change cascades to other functions which load
2279 certificates and CRLs.
2283 * x86_64 assembly pack: annotate code with DWARF CFI directives to
2284 facilitate stack unwinding even from assembly subroutines.
2288 * Remove VAX C specific definitions of OPENSSL_EXPORT, OPENSSL_EXTERN.
2289 Also remove OPENSSL_GLOBAL entirely, as it became a no-op.
2293 * Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
2294 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
2295 which is the minimum version we support.
2299 * Certificate time validation (X509_cmp_time) enforces stricter
2300 compliance with RFC 5280. Fractional seconds and timezone offsets
2301 are no longer allowed.
2305 * Add support for ARIA
2309 * s_client will now send the Server Name Indication (SNI) extension by
2310 default unless the new "-noservername" option is used. The server name is
2311 based on the host provided to the "-connect" option unless overridden by
2312 using "-servername".
2316 * Add support for SipHash
2320 * OpenSSL now fails if it receives an unrecognised record type in TLS1.0
2321 or TLS1.1. Previously this only happened in SSLv3 and TLS1.2. This is to
2322 prevent issues where no progress is being made and the peer continually
2323 sends unrecognised record types, using up resources processing them.
2327 * 'openssl passwd' can now produce SHA256 and SHA512 based output,
2328 using the algorithm defined in
2329 <https://www.akkadia.org/drepper/SHA-crypt.txt>
2333 * Heartbeat support has been removed; the ABI is changed for now.
2335 *Richard Levitte, Rich Salz*
2337 * Support for SSL_OP_NO_ENCRYPT_THEN_MAC in SSL_CONF_cmd.
2341 * The RSA "null" method, which was partially supported to avoid patent
2342 issues, has been replaced to always returns NULL.
2349 ### Changes between 1.1.0k and 1.1.0l [10 Sep 2019]
2351 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
2352 used even when parsing explicit parameters, when loading a encoded key
2353 or calling `EC_GROUP_new_from_ecpkparameters()`/
2354 `EC_GROUP_new_from_ecparameters()`.
2355 This prevents bypass of security hardening and performance gains,
2356 especially for curves with specialized EC_METHODs.
2357 By default, if a key encoded with explicit parameters is loaded and later
2358 encoded, the output is still encoded with explicit parameters, even if
2359 internally a "named" EC_GROUP is used for computation.
2363 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
2364 this change, EC_GROUP_set_generator would accept order and/or cofactor as
2365 NULL. After this change, only the cofactor parameter can be NULL. It also
2366 does some minimal sanity checks on the passed order.
2371 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
2372 An attack is simple, if the first CMS_recipientInfo is valid but the
2373 second CMS_recipientInfo is chosen ciphertext. If the second
2374 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
2375 encryption key will be replaced by garbage, and the message cannot be
2376 decoded, but if the RSA decryption fails, the correct encryption key is
2377 used and the recipient will not notice the attack.
2378 As a work around for this potential attack the length of the decrypted
2379 key must be equal to the cipher default key length, in case the
2380 certifiate is not given and all recipientInfo are tried out.
2381 The old behaviour can be re-enabled in the CMS code by setting the
2382 CMS_DEBUG_DECRYPT flag.
2387 * Use Windows installation paths in the mingw builds
2389 Mingw isn't a POSIX environment per se, which means that Windows
2390 paths should be used for installation.
2395 ### Changes between 1.1.0j and 1.1.0k [28 May 2019]
2397 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
2398 This changes the size when using the `genpkey` command when no size is given.
2399 It fixes an omission in earlier changes that changed all RSA, DSA and DH
2400 generation commands to use 2048 bits by default.
2404 * Prevent over long nonces in ChaCha20-Poly1305.
2406 ChaCha20-Poly1305 is an AEAD cipher, and requires a unique nonce input
2407 for every encryption operation. RFC 7539 specifies that the nonce value
2408 (IV) should be 96 bits (12 bytes). OpenSSL allows a variable nonce length
2409 and front pads the nonce with 0 bytes if it is less than 12
2410 bytes. However it also incorrectly allows a nonce to be set of up to 16
2411 bytes. In this case only the last 12 bytes are significant and any
2412 additional leading bytes are ignored.
2414 It is a requirement of using this cipher that nonce values are
2415 unique. Messages encrypted using a reused nonce value are susceptible to
2416 serious confidentiality and integrity attacks. If an application changes
2417 the default nonce length to be longer than 12 bytes and then makes a
2418 change to the leading bytes of the nonce expecting the new value to be a
2419 new unique nonce then such an application could inadvertently encrypt
2420 messages with a reused nonce.
2422 Additionally the ignored bytes in a long nonce are not covered by the
2423 integrity guarantee of this cipher. Any application that relies on the
2424 integrity of these ignored leading bytes of a long nonce may be further
2425 affected. Any OpenSSL internal use of this cipher, including in SSL/TLS,
2426 is safe because no such use sets such a long nonce value. However user
2427 applications that use this cipher directly and set a non-default nonce
2428 length to be longer than 12 bytes may be vulnerable.
2430 This issue was reported to OpenSSL on 16th of March 2019 by Joran Dirk
2436 * Added SCA hardening for modular field inversion in EC_GROUP through
2437 a new dedicated field_inv() pointer in EC_METHOD.
2438 This also addresses a leakage affecting conversions from projective
2439 to affine coordinates.
2441 *Billy Bob Brumley, Nicola Tuveri*
2443 * Fix a use after free bug in d2i_X509_PUBKEY when overwriting a
2444 re-used X509_PUBKEY object if the second PUBKEY is malformed.
2448 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
2452 * Remove the 'dist' target and add a tarball building script. The
2453 'dist' target has fallen out of use, and it shouldn't be
2454 necessary to configure just to create a source distribution.
2458 ### Changes between 1.1.0i and 1.1.0j [20 Nov 2018]
2460 * Timing vulnerability in DSA signature generation
2462 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
2463 timing side channel attack. An attacker could use variations in the signing
2464 algorithm to recover the private key.
2466 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
2471 * Timing vulnerability in ECDSA signature generation
2473 The OpenSSL ECDSA signature algorithm has been shown to be vulnerable to a
2474 timing side channel attack. An attacker could use variations in the signing
2475 algorithm to recover the private key.
2477 This issue was reported to OpenSSL on 25th October 2018 by Samuel Weiser.
2482 * Add coordinate blinding for EC_POINT and implement projective
2483 coordinate blinding for generic prime curves as a countermeasure to
2484 chosen point SCA attacks.
2486 *Sohaib ul Hassan, Nicola Tuveri, Billy Bob Brumley*
2488 ### Changes between 1.1.0h and 1.1.0i [14 Aug 2018]
2490 * Client DoS due to large DH parameter
2492 During key agreement in a TLS handshake using a DH(E) based ciphersuite a
2493 malicious server can send a very large prime value to the client. This will
2494 cause the client to spend an unreasonably long period of time generating a
2495 key for this prime resulting in a hang until the client has finished. This
2496 could be exploited in a Denial Of Service attack.
2498 This issue was reported to OpenSSL on 5th June 2018 by Guido Vranken
2503 * Cache timing vulnerability in RSA Key Generation
2505 The OpenSSL RSA Key generation algorithm has been shown to be vulnerable to
2506 a cache timing side channel attack. An attacker with sufficient access to
2507 mount cache timing attacks during the RSA key generation process could
2508 recover the private key.
2510 This issue was reported to OpenSSL on 4th April 2018 by Alejandro Cabrera
2511 Aldaya, Billy Brumley, Cesar Pereida Garcia and Luis Manuel Alvarez Tapia.
2516 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
2517 parameter is no longer accepted, as it leads to a corrupt table. NULL
2518 pem_str is reserved for alias entries only.
2522 * Revert blinding in ECDSA sign and instead make problematic addition
2523 length-invariant. Switch even to fixed-length Montgomery multiplication.
2527 * Change generating and checking of primes so that the error rate of not
2528 being prime depends on the intended use based on the size of the input.
2529 For larger primes this will result in more rounds of Miller-Rabin.
2530 The maximal error rate for primes with more than 1080 bits is lowered
2533 *Kurt Roeckx, Annie Yousar*
2535 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
2539 * Add blinding to ECDSA and DSA signatures to protect against side channel
2540 attacks discovered by Keegan Ryan (NCC Group).
2544 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
2545 now allow empty (zero character) pass phrases.
2549 * Certificate time validation (X509_cmp_time) enforces stricter
2550 compliance with RFC 5280. Fractional seconds and timezone offsets
2551 are no longer allowed.
2555 * Fixed a text canonicalisation bug in CMS
2557 Where a CMS detached signature is used with text content the text goes
2558 through a canonicalisation process first prior to signing or verifying a
2559 signature. This process strips trailing space at the end of lines, converts
2560 line terminators to CRLF and removes additional trailing line terminators
2561 at the end of a file. A bug in the canonicalisation process meant that
2562 some characters, such as form-feed, were incorrectly treated as whitespace
2563 and removed. This is contrary to the specification (RFC5485). This fix
2564 could mean that detached text data signed with an earlier version of
2565 OpenSSL 1.1.0 may fail to verify using the fixed version, or text data
2566 signed with a fixed OpenSSL may fail to verify with an earlier version of
2567 OpenSSL 1.1.0. A workaround is to only verify the canonicalised text data
2568 and use the "-binary" flag (for the "cms" command line application) or set
2569 the SMIME_BINARY/PKCS7_BINARY/CMS_BINARY flags (if using CMS_verify()).
2573 ### Changes between 1.1.0g and 1.1.0h [27 Mar 2018]
2575 * Constructed ASN.1 types with a recursive definition could exceed the stack
2577 Constructed ASN.1 types with a recursive definition (such as can be found
2578 in PKCS7) could eventually exceed the stack given malicious input with
2579 excessive recursion. This could result in a Denial Of Service attack. There
2580 are no such structures used within SSL/TLS that come from untrusted sources
2581 so this is considered safe.
2583 This issue was reported to OpenSSL on 4th January 2018 by the OSS-fuzz
2589 * Incorrect CRYPTO_memcmp on HP-UX PA-RISC
2591 Because of an implementation bug the PA-RISC CRYPTO_memcmp function is
2592 effectively reduced to only comparing the least significant bit of each
2593 byte. This allows an attacker to forge messages that would be considered as
2594 authenticated in an amount of tries lower than that guaranteed by the
2595 security claims of the scheme. The module can only be compiled by the
2596 HP-UX assembler, so that only HP-UX PA-RISC targets are affected.
2598 This issue was reported to OpenSSL on 2nd March 2018 by Peter Waltenberg
2604 * Add a build target 'build_all_generated', to build all generated files
2605 and only that. This can be used to prepare everything that requires
2606 things like perl for a system that lacks perl and then move everything
2607 to that system and do the rest of the build there.
2611 * Backport SSL_OP_NO_RENGOTIATION
2613 OpenSSL 1.0.2 and below had the ability to disable renegotiation using the
2614 (undocumented) SSL3_FLAGS_NO_RENEGOTIATE_CIPHERS flag. Due to the opacity
2615 changes this is no longer possible in 1.1.0. Therefore the new
2616 SSL_OP_NO_RENEGOTIATION option from 1.1.1-dev has been backported to
2617 1.1.0 to provide equivalent functionality.
2619 Note that if an application built against 1.1.0h headers (or above) is run
2620 using an older version of 1.1.0 (prior to 1.1.0h) then the option will be
2621 accepted but nothing will happen, i.e. renegotiation will not be prevented.
2625 * Removed the OS390-Unix config target. It relied on a script that doesn't
2630 * rsaz_1024_mul_avx2 overflow bug on x86_64
2632 There is an overflow bug in the AVX2 Montgomery multiplication procedure
2633 used in exponentiation with 1024-bit moduli. No EC algorithms are affected.
2634 Analysis suggests that attacks against RSA and DSA as a result of this
2635 defect would be very difficult to perform and are not believed likely.
2636 Attacks against DH1024 are considered just feasible, because most of the
2637 work necessary to deduce information about a private key may be performed
2638 offline. The amount of resources required for such an attack would be
2639 significant. However, for an attack on TLS to be meaningful, the server
2640 would have to share the DH1024 private key among multiple clients, which is
2641 no longer an option since CVE-2016-0701.
2643 This only affects processors that support the AVX2 but not ADX extensions
2644 like Intel Haswell (4th generation).
2646 This issue was reported to OpenSSL by David Benjamin (Google). The issue
2647 was originally found via the OSS-Fuzz project.
2652 ### Changes between 1.1.0f and 1.1.0g [2 Nov 2017]
2654 * bn_sqrx8x_internal carry bug on x86_64
2656 There is a carry propagating bug in the x86_64 Montgomery squaring
2657 procedure. No EC algorithms are affected. Analysis suggests that attacks
2658 against RSA and DSA as a result of this defect would be very difficult to
2659 perform and are not believed likely. Attacks against DH are considered just
2660 feasible (although very difficult) because most of the work necessary to
2661 deduce information about a private key may be performed offline. The amount
2662 of resources required for such an attack would be very significant and
2663 likely only accessible to a limited number of attackers. An attacker would
2664 additionally need online access to an unpatched system using the target
2665 private key in a scenario with persistent DH parameters and a private
2666 key that is shared between multiple clients.
2668 This only affects processors that support the BMI1, BMI2 and ADX extensions
2669 like Intel Broadwell (5th generation) and later or AMD Ryzen.
2671 This issue was reported to OpenSSL by the OSS-Fuzz project.
2676 * Malformed X.509 IPAddressFamily could cause OOB read
2678 If an X.509 certificate has a malformed IPAddressFamily extension,
2679 OpenSSL could do a one-byte buffer overread. The most likely result
2680 would be an erroneous display of the certificate in text format.
2682 This issue was reported to OpenSSL by the OSS-Fuzz project.
2687 ### Changes between 1.1.0e and 1.1.0f [25 May 2017]
2689 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
2690 platform rather than 'mingw'.
2694 * Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
2695 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
2696 which is the minimum version we support.
2700 ### Changes between 1.1.0d and 1.1.0e [16 Feb 2017]
2702 * Encrypt-Then-Mac renegotiation crash
2704 During a renegotiation handshake if the Encrypt-Then-Mac extension is
2705 negotiated where it was not in the original handshake (or vice-versa) then
2706 this can cause OpenSSL to crash (dependant on ciphersuite). Both clients
2707 and servers are affected.
2709 This issue was reported to OpenSSL by Joe Orton (Red Hat).
2714 ### Changes between 1.1.0c and 1.1.0d [26 Jan 2017]
2716 * Truncated packet could crash via OOB read
2718 If one side of an SSL/TLS path is running on a 32-bit host and a specific
2719 cipher is being used, then a truncated packet can cause that host to
2720 perform an out-of-bounds read, usually resulting in a crash.
2722 This issue was reported to OpenSSL by Robert Święcki of Google.
2727 * Bad (EC)DHE parameters cause a client crash
2729 If a malicious server supplies bad parameters for a DHE or ECDHE key
2730 exchange then this can result in the client attempting to dereference a
2731 NULL pointer leading to a client crash. This could be exploited in a Denial
2734 This issue was reported to OpenSSL by Guido Vranken.
2739 * BN_mod_exp may produce incorrect results on x86_64
2741 There is a carry propagating bug in the x86_64 Montgomery squaring
2742 procedure. No EC algorithms are affected. Analysis suggests that attacks
2743 against RSA and DSA as a result of this defect would be very difficult to
2744 perform and are not believed likely. Attacks against DH are considered just
2745 feasible (although very difficult) because most of the work necessary to
2746 deduce information about a private key may be performed offline. The amount
2747 of resources required for such an attack would be very significant and
2748 likely only accessible to a limited number of attackers. An attacker would
2749 additionally need online access to an unpatched system using the target
2750 private key in a scenario with persistent DH parameters and a private
2751 key that is shared between multiple clients. For example this can occur by
2752 default in OpenSSL DHE based SSL/TLS ciphersuites. Note: This issue is very
2753 similar to CVE-2015-3193 but must be treated as a separate problem.
2755 This issue was reported to OpenSSL by the OSS-Fuzz project.
2760 ### Changes between 1.1.0b and 1.1.0c [10 Nov 2016]
2762 * ChaCha20/Poly1305 heap-buffer-overflow
2764 TLS connections using `*-CHACHA20-POLY1305` ciphersuites are susceptible to
2765 a DoS attack by corrupting larger payloads. This can result in an OpenSSL
2766 crash. This issue is not considered to be exploitable beyond a DoS.
2768 This issue was reported to OpenSSL by Robert Święcki (Google Security Team)
2773 * CMS Null dereference
2775 Applications parsing invalid CMS structures can crash with a NULL pointer
2776 dereference. This is caused by a bug in the handling of the ASN.1 CHOICE
2777 type in OpenSSL 1.1.0 which can result in a NULL value being passed to the
2778 structure callback if an attempt is made to free certain invalid encodings.
2779 Only CHOICE structures using a callback which do not handle NULL value are
2782 This issue was reported to OpenSSL by Tyler Nighswander of ForAllSecure.
2787 * Montgomery multiplication may produce incorrect results
2789 There is a carry propagating bug in the Broadwell-specific Montgomery
2790 multiplication procedure that handles input lengths divisible by, but
2791 longer than 256 bits. Analysis suggests that attacks against RSA, DSA
2792 and DH private keys are impossible. This is because the subroutine in
2793 question is not used in operations with the private key itself and an input
2794 of the attacker's direct choice. Otherwise the bug can manifest itself as
2795 transient authentication and key negotiation failures or reproducible
2796 erroneous outcome of public-key operations with specially crafted input.
2797 Among EC algorithms only Brainpool P-512 curves are affected and one
2798 presumably can attack ECDH key negotiation. Impact was not analyzed in
2799 detail, because pre-requisites for attack are considered unlikely. Namely
2800 multiple clients have to choose the curve in question and the server has to
2801 share the private key among them, neither of which is default behaviour.
2802 Even then only clients that chose the curve will be affected.
2804 This issue was publicly reported as transient failures and was not
2805 initially recognized as a security issue. Thanks to Richard Morgan for
2806 providing reproducible case.
2811 * Removed automatic addition of RPATH in shared libraries and executables,
2812 as this was a remainder from OpenSSL 1.0.x and isn't needed any more.
2816 ### Changes between 1.1.0a and 1.1.0b [26 Sep 2016]
2818 * Fix Use After Free for large message sizes
2820 The patch applied to address CVE-2016-6307 resulted in an issue where if a
2821 message larger than approx 16k is received then the underlying buffer to
2822 store the incoming message is reallocated and moved. Unfortunately a
2823 dangling pointer to the old location is left which results in an attempt to
2824 write to the previously freed location. This is likely to result in a
2825 crash, however it could potentially lead to execution of arbitrary code.
2827 This issue only affects OpenSSL 1.1.0a.
2829 This issue was reported to OpenSSL by Robert Święcki.
2834 ### Changes between 1.1.0 and 1.1.0a [22 Sep 2016]
2836 * OCSP Status Request extension unbounded memory growth
2838 A malicious client can send an excessively large OCSP Status Request
2839 extension. If that client continually requests renegotiation, sending a
2840 large OCSP Status Request extension each time, then there will be unbounded
2841 memory growth on the server. This will eventually lead to a Denial Of
2842 Service attack through memory exhaustion. Servers with a default
2843 configuration are vulnerable even if they do not support OCSP. Builds using
2844 the "no-ocsp" build time option are not affected.
2846 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
2851 * SSL_peek() hang on empty record
2853 OpenSSL 1.1.0 SSL/TLS will hang during a call to SSL_peek() if the peer
2854 sends an empty record. This could be exploited by a malicious peer in a
2855 Denial Of Service attack.
2857 This issue was reported to OpenSSL by Alex Gaynor.
2862 * Excessive allocation of memory in tls_get_message_header() and
2863 dtls1_preprocess_fragment()
2865 A (D)TLS message includes 3 bytes for its length in the header for the
2866 message. This would allow for messages up to 16Mb in length. Messages of
2867 this length are excessive and OpenSSL includes a check to ensure that a
2868 peer is sending reasonably sized messages in order to avoid too much memory
2869 being consumed to service a connection. A flaw in the logic of version
2870 1.1.0 means that memory for the message is allocated too early, prior to
2871 the excessive message length check. Due to way memory is allocated in
2872 OpenSSL this could mean an attacker could force up to 21Mb to be allocated
2873 to service a connection. This could lead to a Denial of Service through
2874 memory exhaustion. However, the excessive message length check still takes
2875 place, and this would cause the connection to immediately fail. Assuming
2876 that the application calls SSL_free() on the failed connection in a timely
2877 manner then the 21Mb of allocated memory will then be immediately freed
2878 again. Therefore the excessive memory allocation will be transitory in
2879 nature. This then means that there is only a security impact if:
2881 1) The application does not call SSL_free() in a timely manner in the event
2882 that the connection fails
2884 2) The application is working in a constrained environment where there is
2885 very little free memory
2887 3) The attacker initiates multiple connection attempts such that there are
2888 multiple connections in a state where memory has been allocated for the
2889 connection; SSL_free() has not yet been called; and there is insufficient
2890 memory to service the multiple requests.
2892 Except in the instance of (1) above any Denial Of Service is likely to be
2893 transitory because as soon as the connection fails the memory is
2894 subsequently freed again in the SSL_free() call. However there is an
2895 increased risk during this period of application crashes due to the lack of
2896 memory - which would then mean a more serious Denial of Service.
2898 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
2899 (CVE-2016-6307 and CVE-2016-6308)
2903 * solaris-x86-cc, i.e. 32-bit configuration with vendor compiler,
2904 had to be removed. Primary reason is that vendor assembler can't
2905 assemble our modules with -KPIC flag. As result it, assembly
2906 support, was not even available as option. But its lack means
2907 lack of side-channel resistant code, which is incompatible with
2908 security by todays standards. Fortunately gcc is readily available
2909 prepackaged option, which we firmly point at...
2913 ### Changes between 1.0.2h and 1.1.0 [25 Aug 2016]
2915 * Windows command-line tool supports UTF-8 opt-in option for arguments
2916 and console input. Setting OPENSSL_WIN32_UTF8 environment variable
2917 (to any value) allows Windows user to access PKCS#12 file generated
2918 with Windows CryptoAPI and protected with non-ASCII password, as well
2919 as files generated under UTF-8 locale on Linux also protected with
2924 * To mitigate the SWEET32 attack ([CVE-2016-2183]), 3DES cipher suites
2925 have been disabled by default and removed from DEFAULT, just like RC4.
2926 See the RC4 item below to re-enable both.
2930 * The method for finding the storage location for the Windows RAND seed file
2931 has changed. First we check %RANDFILE%. If that is not set then we check
2932 the directories %HOME%, %USERPROFILE% and %SYSTEMROOT% in that order. If
2933 all else fails we fall back to C:\.
2937 * The EVP_EncryptUpdate() function has had its return type changed from void
2938 to int. A return of 0 indicates and error while a return of 1 indicates
2943 * The flags RSA_FLAG_NO_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME and
2944 DH_FLAG_NO_EXP_CONSTTIME which previously provided the ability to switch
2945 off the constant time implementation for RSA, DSA and DH have been made
2946 no-ops and deprecated.
2950 * Windows RAND implementation was simplified to only get entropy by
2951 calling CryptGenRandom(). Various other RAND-related tickets
2954 *Joseph Wylie Yandle, Rich Salz*
2956 * The stack and lhash API's were renamed to start with `OPENSSL_SK_`
2957 and `OPENSSL_LH_`, respectively. The old names are available
2958 with API compatibility. They new names are now completely documented.
2962 * Unify TYPE_up_ref(obj) methods signature.
2963 SSL_CTX_up_ref(), SSL_up_ref(), X509_up_ref(), EVP_PKEY_up_ref(),
2964 X509_CRL_up_ref(), X509_OBJECT_up_ref_count() methods are now returning an
2965 int (instead of void) like all others TYPE_up_ref() methods.
2966 So now these methods also check the return value of CRYPTO_atomic_add(),
2967 and the validity of object reference counter.
2969 *fdasilvayy@gmail.com*
2971 * With Windows Visual Studio builds, the .pdb files are installed
2972 alongside the installed libraries and executables. For a static
2973 library installation, ossl_static.pdb is the associate compiler
2974 generated .pdb file to be used when linking programs.
2978 * Remove openssl.spec. Packaging files belong with the packagers.
2982 * Automatic Darwin/OSX configuration has had a refresh, it will now
2983 recognise x86_64 architectures automatically. You can still decide
2984 to build for a different bitness with the environment variable
2985 KERNEL_BITS (can be 32 or 64), for example:
2987 KERNEL_BITS=32 ./config
2991 * Change default algorithms in pkcs8 utility to use PKCS#5 v2.0,
2992 256 bit AES and HMAC with SHA256.
2996 * Remove support for MIPS o32 ABI on IRIX (and IRIX only).
3000 * Triple-DES ciphers have been moved from HIGH to MEDIUM.
3004 * To enable users to have their own config files and build file templates,
3005 Configure looks in the directory indicated by the environment variable
3006 OPENSSL_LOCAL_CONFIG_DIR as well as the in-source Configurations/
3007 directory. On VMS, OPENSSL_LOCAL_CONFIG_DIR is expected to be a logical
3008 name and is used as is.
3012 * The following datatypes were made opaque: X509_OBJECT, X509_STORE_CTX,
3013 X509_STORE, X509_LOOKUP, and X509_LOOKUP_METHOD. The unused type
3014 X509_CERT_FILE_CTX was removed.
3018 * "shared" builds are now the default. To create only static libraries use
3019 the "no-shared" Configure option.
3023 * Remove the no-aes, no-hmac, no-rsa, no-sha and no-md5 Configure options.
3024 All of these option have not worked for some while and are fundamental
3029 * Make various cleanup routines no-ops and mark them as deprecated. Most
3030 global cleanup functions are no longer required because they are handled
3031 via auto-deinit (see OPENSSL_init_crypto and OPENSSL_init_ssl man pages).
3032 Explicitly de-initing can cause problems (e.g. where a library that uses
3033 OpenSSL de-inits, but an application is still using it). The affected
3034 functions are CONF_modules_free(), ENGINE_cleanup(), OBJ_cleanup(),
3035 EVP_cleanup(), BIO_sock_cleanup(), CRYPTO_cleanup_all_ex_data(),
3036 RAND_cleanup(), SSL_COMP_free_compression_methods(), ERR_free_strings() and
3037 COMP_zlib_cleanup().
3041 * --strict-warnings no longer enables runtime debugging options
3042 such as REF_DEBUG. Instead, debug options are automatically
3043 enabled with '--debug' builds.
3045 *Andy Polyakov, Emilia Käsper*
3047 * Made DH and DH_METHOD opaque. The structures for managing DH objects
3048 have been moved out of the public header files. New functions for managing
3049 these have been added.
3053 * Made RSA and RSA_METHOD opaque. The structures for managing RSA
3054 objects have been moved out of the public header files. New
3055 functions for managing these have been added.
3059 * Made DSA and DSA_METHOD opaque. The structures for managing DSA objects
3060 have been moved out of the public header files. New functions for managing
3061 these have been added.
3065 * Made BIO and BIO_METHOD opaque. The structures for managing BIOs have been
3066 moved out of the public header files. New functions for managing these
3071 * Removed no-rijndael as a config option. Rijndael is an old name for AES.
3075 * Removed the mk1mf build scripts.
3079 * Headers are now wrapped, if necessary, with OPENSSL_NO_xxx, so
3080 it is always safe to #include a header now.
3084 * Removed the aged BC-32 config and all its supporting scripts
3088 * Removed support for Ultrix, Netware, and OS/2.
3092 * Add support for HKDF.
3094 *Alessandro Ghedini*
3096 * Add support for blake2b and blake2s
3100 * Added support for "pipelining". Ciphers that have the
3101 EVP_CIPH_FLAG_PIPELINE flag set have a capability to process multiple
3102 encryptions/decryptions simultaneously. There are currently no built-in
3103 ciphers with this property but the expectation is that engines will be able
3104 to offer it to significantly improve throughput. Support has been extended
3105 into libssl so that multiple records for a single connection can be
3106 processed in one go (for >=TLS 1.1).
3110 * Added the AFALG engine. This is an async capable engine which is able to
3111 offload work to the Linux kernel. In this initial version it only supports
3112 AES128-CBC. The kernel must be version 4.1.0 or greater.
3116 * OpenSSL now uses a new threading API. It is no longer necessary to
3117 set locking callbacks to use OpenSSL in a multi-threaded environment. There
3118 are two supported threading models: pthreads and windows threads. It is
3119 also possible to configure OpenSSL at compile time for "no-threads". The
3120 old threading API should no longer be used. The functions have been
3121 replaced with "no-op" compatibility macros.
3123 *Alessandro Ghedini, Matt Caswell*
3125 * Modify behavior of ALPN to invoke callback after SNI/servername
3126 callback, such that updates to the SSL_CTX affect ALPN.
3130 * Add SSL_CIPHER queries for authentication and key-exchange.
3134 * Changes to the DEFAULT cipherlist:
3135 - Prefer (EC)DHE handshakes over plain RSA.
3136 - Prefer AEAD ciphers over legacy ciphers.
3137 - Prefer ECDSA over RSA when both certificates are available.
3138 - Prefer TLSv1.2 ciphers/PRF.
3139 - Remove DSS, SEED, IDEA, CAMELLIA, and AES-CCM from the
3144 * Change the ECC default curve list to be this, in order: x25519,
3145 secp256r1, secp521r1, secp384r1.
3149 * RC4 based libssl ciphersuites are now classed as "weak" ciphers and are
3150 disabled by default. They can be re-enabled using the
3151 enable-weak-ssl-ciphers option to Configure.
3155 * If the server has ALPN configured, but supports no protocols that the
3156 client advertises, send a fatal "no_application_protocol" alert.
3157 This behaviour is SHALL in RFC 7301, though it isn't universally
3158 implemented by other servers.
3162 * Add X25519 support.
3163 Add ASN.1 and EVP_PKEY methods for X25519. This includes support
3164 for public and private key encoding using the format documented in
3165 draft-ietf-curdle-pkix-02. The corresponding EVP_PKEY method supports
3166 key generation and key derivation.
3168 TLS support complies with draft-ietf-tls-rfc4492bis-08 and uses
3173 * Deprecate SRP_VBASE_get_by_user.
3174 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
3175 In order to fix an unavoidable memory leak ([CVE-2016-0798]),
3176 SRP_VBASE_get_by_user was changed to ignore the "fake user" SRP
3177 seed, even if the seed is configured.
3179 Users should use SRP_VBASE_get1_by_user instead. Note that in
3180 SRP_VBASE_get1_by_user, caller must free the returned value. Note
3181 also that even though configuring the SRP seed attempts to hide
3182 invalid usernames by continuing the handshake with fake
3183 credentials, this behaviour is not constant time and no strong
3184 guarantees are made that the handshake is indistinguishable from
3185 that of a valid user.
3189 * Configuration change; it's now possible to build dynamic engines
3190 without having to build shared libraries and vice versa. This
3191 only applies to the engines in `engines/`, those in `crypto/engine/`
3192 will always be built into libcrypto (i.e. "static").
3194 Building dynamic engines is enabled by default; to disable, use
3195 the configuration option "disable-dynamic-engine".
3197 The only requirements for building dynamic engines are the
3198 presence of the DSO module and building with position independent
3199 code, so they will also automatically be disabled if configuring
3200 with "disable-dso" or "disable-pic".
3202 The macros OPENSSL_NO_STATIC_ENGINE and OPENSSL_NO_DYNAMIC_ENGINE
3203 are also taken away from openssl/opensslconf.h, as they are
3208 * Configuration change; if there is a known flag to compile
3209 position independent code, it will always be applied on the
3210 libcrypto and libssl object files, and never on the application
3211 object files. This means other libraries that use routines from
3212 libcrypto / libssl can be made into shared libraries regardless
3213 of how OpenSSL was configured.
3215 If this isn't desirable, the configuration options "disable-pic"
3216 or "no-pic" can be used to disable the use of PIC. This will
3217 also disable building shared libraries and dynamic engines.
3221 * Removed JPAKE code. It was experimental and has no wide use.
3225 * The INSTALL_PREFIX Makefile variable has been renamed to
3226 DESTDIR. That makes for less confusion on what this variable
3227 is for. Also, the configuration option --install_prefix is
3232 * Heartbeat for TLS has been removed and is disabled by default
3233 for DTLS; configure with enable-heartbeats. Code that uses the
3234 old #define's might need to be updated.
3236 *Emilia Käsper, Rich Salz*
3238 * Rename REF_CHECK to REF_DEBUG.
3242 * New "unified" build system
3244 The "unified" build system is aimed to be a common system for all
3245 platforms we support. With it comes new support for VMS.
3247 This system builds supports building in a different directory tree
3248 than the source tree. It produces one Makefile (for unix family
3249 or lookalikes), or one descrip.mms (for VMS).
3251 The source of information to make the Makefile / descrip.mms is
3252 small files called 'build.info', holding the necessary
3253 information for each directory with source to compile, and a
3254 template in Configurations, like unix-Makefile.tmpl or
3257 With this change, the library names were also renamed on Windows
3258 and on VMS. They now have names that are closer to the standard
3259 on Unix, and include the major version number, and in certain
3260 cases, the architecture they are built for. See "Notes on shared
3261 libraries" in INSTALL.
3263 We rely heavily on the perl module Text::Template.
3267 * Added support for auto-initialisation and de-initialisation of the library.
3268 OpenSSL no longer requires explicit init or deinit routines to be called,
3269 except in certain circumstances. See the OPENSSL_init_crypto() and
3270 OPENSSL_init_ssl() man pages for further information.
3274 * The arguments to the DTLSv1_listen function have changed. Specifically the
3275 "peer" argument is now expected to be a BIO_ADDR object.
3277 * Rewrite of BIO networking library. The BIO library lacked consistent
3278 support of IPv6, and adding it required some more extensive
3279 modifications. This introduces the BIO_ADDR and BIO_ADDRINFO types,
3280 which hold all types of addresses and chains of address information.
3281 It also introduces a new API, with functions like BIO_socket,
3282 BIO_connect, BIO_listen, BIO_lookup and a rewrite of BIO_accept.
3283 The source/sink BIOs BIO_s_connect, BIO_s_accept and BIO_s_datagram
3284 have been adapted accordingly.
3288 * RSA_padding_check_PKCS1_type_1 now accepts inputs with and without
3293 * CRIME protection: disable compression by default, even if OpenSSL is
3294 compiled with zlib enabled. Applications can still enable compression
3295 by calling SSL_CTX_clear_options(ctx, SSL_OP_NO_COMPRESSION), or by
3296 using the SSL_CONF library to configure compression.
3300 * The signature of the session callback configured with
3301 SSL_CTX_sess_set_get_cb was changed. The read-only input buffer
3302 was explicitly marked as `const unsigned char*` instead of
3307 * Always DPURIFY. Remove the use of uninitialized memory in the
3308 RNG, and other conditional uses of DPURIFY. This makes -DPURIFY a no-op.
3312 * Removed many obsolete configuration items, including
3313 DES_PTR, DES_RISC1, DES_RISC2, DES_INT
3314 MD2_CHAR, MD2_INT, MD2_LONG
3316 IDEA_SHORT, IDEA_LONG
3317 RC2_SHORT, RC2_LONG, RC4_LONG, RC4_CHUNK, RC4_INDEX
3319 *Rich Salz, with advice from Andy Polyakov*
3321 * Many BN internals have been moved to an internal header file.
3323 *Rich Salz with help from Andy Polyakov*
3325 * Configuration and writing out the results from it has changed.
3326 Files such as Makefile include/openssl/opensslconf.h and are now
3327 produced through general templates, such as Makefile.in and
3328 crypto/opensslconf.h.in and some help from the perl module
3331 Also, the center of configuration information is no longer
3332 Makefile. Instead, Configure produces a perl module in
3333 configdata.pm which holds most of the config data (in the hash
3334 table %config), the target data that comes from the target
3335 configuration in one of the `Configurations/*.conf` files (in
3340 * To clarify their intended purposes, the Configure options
3341 --prefix and --openssldir change their semantics, and become more
3342 straightforward and less interdependent.
3344 --prefix shall be used exclusively to give the location INSTALLTOP
3345 where programs, scripts, libraries, include files and manuals are
3346 going to be installed. The default is now /usr/local.
3348 --openssldir shall be used exclusively to give the default
3349 location OPENSSLDIR where certificates, private keys, CRLs are
3350 managed. This is also where the default openssl.cnf gets
3352 If the directory given with this option is a relative path, the
3353 values of both the --prefix value and the --openssldir value will
3354 be combined to become OPENSSLDIR.
3355 The default for --openssldir is INSTALLTOP/ssl.
3357 Anyone who uses --openssldir to specify where OpenSSL is to be
3358 installed MUST change to use --prefix instead.
3362 * The GOST engine was out of date and therefore it has been removed. An up
3363 to date GOST engine is now being maintained in an external repository.
3364 See: <https://wiki.openssl.org/index.php/Binaries>. Libssl still retains
3365 support for GOST ciphersuites (these are only activated if a GOST engine
3370 * EGD is no longer supported by default; use enable-egd when
3373 *Ben Kaduk and Rich Salz*
3375 * The distribution now has Makefile.in files, which are used to
3376 create Makefile's when Configure is run. *Configure must be run
3377 before trying to build now.*
3381 * The return value for SSL_CIPHER_description() for error conditions
3386 * Support for RFC6698/RFC7671 DANE TLSA peer authentication.
3388 Obtaining and performing DNSSEC validation of TLSA records is
3389 the application's responsibility. The application provides
3390 the TLSA records of its choice to OpenSSL, and these are then
3391 used to authenticate the peer.
3393 The TLSA records need not even come from DNS. They can, for
3394 example, be used to implement local end-entity certificate or
3395 trust-anchor "pinning", where the "pin" data takes the form
3396 of TLSA records, which can augment or replace verification
3397 based on the usual WebPKI public certification authorities.
3401 * Revert default OPENSSL_NO_DEPRECATED setting. Instead OpenSSL
3402 continues to support deprecated interfaces in default builds.
3403 However, applications are strongly advised to compile their
3404 source files with -DOPENSSL_API_COMPAT=0x10100000L, which hides
3405 the declarations of all interfaces deprecated in 0.9.8, 1.0.0
3406 or the 1.1.0 releases.
3408 In environments in which all applications have been ported to
3409 not use any deprecated interfaces OpenSSL's Configure script
3410 should be used with the --api=1.1.0 option to entirely remove
3411 support for the deprecated features from the library and
3412 unconditionally disable them in the installed headers.
3413 Essentially the same effect can be achieved with the "no-deprecated"
3414 argument to Configure, except that this will always restrict
3415 the build to just the latest API, rather than a fixed API
3418 As applications are ported to future revisions of the API,
3419 they should update their compile-time OPENSSL_API_COMPAT define
3420 accordingly, but in most cases should be able to continue to
3421 compile with later releases.
3423 The OPENSSL_API_COMPAT versions for 1.0.0, and 0.9.8 are
3424 0x10000000L and 0x00908000L, respectively. However those
3425 versions did not support the OPENSSL_API_COMPAT feature, and
3426 so applications are not typically tested for explicit support
3427 of just the undeprecated features of either release.
3431 * Add support for setting the minimum and maximum supported protocol.
3432 It can bet set via the SSL_set_min_proto_version() and
3433 SSL_set_max_proto_version(), or via the SSL_CONF's MinProtocol and
3434 MaxProtocol. It's recommended to use the new APIs to disable
3435 protocols instead of disabling individual protocols using
3436 SSL_set_options() or SSL_CONF's Protocol. This change also
3437 removes support for disabling TLS 1.2 in the OpenSSL TLS
3438 client at compile time by defining OPENSSL_NO_TLS1_2_CLIENT.
3442 * Support for ChaCha20 and Poly1305 added to libcrypto and libssl.
3446 * New EC_KEY_METHOD, this replaces the older ECDSA_METHOD and ECDH_METHOD
3447 and integrates ECDSA and ECDH functionality into EC. Implementations can
3448 now redirect key generation and no longer need to convert to or from
3451 Note: the ecdsa.h and ecdh.h headers are now no longer needed and just
3452 include the ec.h header file instead.
3456 * Remove support for all 40 and 56 bit ciphers. This includes all the export
3457 ciphers who are no longer supported and drops support the ephemeral RSA key
3458 exchange. The LOW ciphers currently doesn't have any ciphers in it.
3462 * Made EVP_MD_CTX, EVP_MD, EVP_CIPHER_CTX, EVP_CIPHER and HMAC_CTX
3463 opaque. For HMAC_CTX, the following constructors and destructors
3466 HMAC_CTX *HMAC_CTX_new(void);
3467 void HMAC_CTX_free(HMAC_CTX *ctx);
3469 For EVP_MD and EVP_CIPHER, complete APIs to create, fill and
3470 destroy such methods has been added. See EVP_MD_meth_new(3) and
3471 EVP_CIPHER_meth_new(3) for documentation.
3474 1) `EVP_MD_CTX_cleanup()`, `EVP_CIPHER_CTX_cleanup()` and
3475 `HMAC_CTX_cleanup()` were removed. `HMAC_CTX_reset()` and
3476 `EVP_MD_CTX_reset()` should be called instead to reinitialise
3477 an already created structure.
3478 2) For consistency with the majority of our object creators and
3479 destructors, `EVP_MD_CTX_(create|destroy)` were renamed to
3480 `EVP_MD_CTX_(new|free)`. The old names are retained as macros
3481 for deprecated builds.
3485 * Added ASYNC support. Libcrypto now includes the async sub-library to enable
3486 cryptographic operations to be performed asynchronously as long as an
3487 asynchronous capable engine is used. See the ASYNC_start_job() man page for
3488 further details. Libssl has also had this capability integrated with the
3489 introduction of the new mode SSL_MODE_ASYNC and associated error
3490 SSL_ERROR_WANT_ASYNC. See the SSL_CTX_set_mode() and SSL_get_error() man
3491 pages. This work was developed in partnership with Intel Corp.
3495 * SSL_{CTX_}set_ecdh_auto() has been removed and ECDH is support is
3496 always enabled now. If you want to disable the support you should
3497 exclude it using the list of supported ciphers. This also means that the
3498 "-no_ecdhe" option has been removed from s_server.
3502 * SSL_{CTX}_set_tmp_ecdh() which can set 1 EC curve now internally calls
3503 SSL_{CTX_}set1_curves() which can set a list.
3507 * Remove support for SSL_{CTX_}set_tmp_ecdh_callback(). You should set the
3508 curve you want to support using SSL_{CTX_}set1_curves().
3512 * State machine rewrite. The state machine code has been significantly
3513 refactored in order to remove much duplication of code and solve issues
3514 with the old code (see [ssl/statem/README.md](ssl/statem/README.md) for
3515 further details). This change does have some associated API changes.
3516 Notably the SSL_state() function has been removed and replaced by
3517 SSL_get_state which now returns an "OSSL_HANDSHAKE_STATE" instead of an int.
3518 SSL_set_state() has been removed altogether. The previous handshake states
3519 defined in ssl.h and ssl3.h have also been removed.
3523 * All instances of the string "ssleay" in the public API were replaced
3524 with OpenSSL (case-matching; e.g., OPENSSL_VERSION for #define's)
3525 Some error codes related to internal RSA_eay API's were renamed.
3529 * The demo files in crypto/threads were moved to demo/threads.
3533 * Removed obsolete engines: 4758cca, aep, atalla, cswift, nuron, gmp,
3536 *Matt Caswell, Rich Salz*
3538 * New ASN.1 embed macro.
3540 New ASN.1 macro ASN1_EMBED. This is the same as ASN1_SIMPLE except the
3541 structure is not allocated: it is part of the parent. That is instead of
3549 This reduces memory fragmentation and make it impossible to accidentally
3550 set a mandatory field to NULL.
3552 This currently only works for some fields specifically a SEQUENCE, CHOICE,
3553 or ASN1_STRING type which is part of a parent SEQUENCE. Since it is
3554 equivalent to ASN1_SIMPLE it cannot be tagged, OPTIONAL, SET OF or
3559 * Remove EVP_CHECK_DES_KEY, a compile-time option that never compiled.
3563 * Removed DES and RC4 ciphersuites from DEFAULT. Also removed RC2 although
3564 in 1.0.2 EXPORT was already removed and the only RC2 ciphersuite is also
3565 an EXPORT one. COMPLEMENTOFDEFAULT has been updated accordingly to add
3566 DES and RC4 ciphersuites.
3570 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
3571 This changes the decoding behaviour for some invalid messages,
3572 though the change is mostly in the more lenient direction, and
3573 legacy behaviour is preserved as much as possible.
3577 * Fix no-stdio build.
3578 *David Woodhouse <David.Woodhouse@intel.com> and also*
3579 *Ivan Nestlerode <ivan.nestlerode@sonos.com>*
3581 * New testing framework
3582 The testing framework has been largely rewritten and is now using
3583 perl and the perl modules Test::Harness and an extended variant of
3584 Test::More called OpenSSL::Test to do its work. All test scripts in
3585 test/ have been rewritten into test recipes, and all direct calls to
3586 executables in test/Makefile have become individual recipes using the
3587 simplified testing OpenSSL::Test::Simple.
3589 For documentation on our testing modules, do:
3591 perldoc test/testlib/OpenSSL/Test/Simple.pm
3592 perldoc test/testlib/OpenSSL/Test.pm
3596 * Revamped memory debug; only -DCRYPTO_MDEBUG and -DCRYPTO_MDEBUG_ABORT
3597 are used; the latter aborts on memory leaks (usually checked on exit).
3598 Some undocumented "set malloc, etc., hooks" functions were removed
3599 and others were changed. All are now documented.
3603 * In DSA_generate_parameters_ex, if the provided seed is too short,
3606 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
3608 * Rewrite PSK to support ECDHE_PSK, DHE_PSK and RSA_PSK. Add ciphersuites
3609 from RFC4279, RFC4785, RFC5487, RFC5489.
3611 Thanks to Christian J. Dietrich and Giuseppe D'Angelo for the
3612 original RSA_PSK patch.
3616 * Dropped support for the SSL3_FLAGS_DELAY_CLIENT_FINISHED flag. This SSLeay
3617 era flag was never set throughout the codebase (only read). Also removed
3618 SSL3_FLAGS_POP_BUFFER which was only used if
3619 SSL3_FLAGS_DELAY_CLIENT_FINISHED was also set.
3623 * Changed the default name options in the "ca", "crl", "req" and "x509"
3624 to be "oneline" instead of "compat".
3628 * Remove SSL_OP_TLS_BLOCK_PADDING_BUG. This is SSLeay legacy, we're
3629 not aware of clients that still exhibit this bug, and the workaround
3630 hasn't been working properly for a while.
3634 * The return type of BIO_number_read() and BIO_number_written() as well as
3635 the corresponding num_read and num_write members in the BIO structure has
3636 changed from unsigned long to uint64_t. On platforms where an unsigned
3637 long is 32 bits (e.g. Windows) these counters could overflow if >4Gb is
3642 * Given the pervasive nature of TLS extensions it is inadvisable to run
3643 OpenSSL without support for them. It also means that maintaining
3644 the OPENSSL_NO_TLSEXT option within the code is very invasive (and probably
3645 not well tested). Therefore the OPENSSL_NO_TLSEXT option has been removed.
3649 * Removed support for the two export grade static DH ciphersuites
3650 EXP-DH-RSA-DES-CBC-SHA and EXP-DH-DSS-DES-CBC-SHA. These two ciphersuites
3651 were newly added (along with a number of other static DH ciphersuites) to
3652 1.0.2. However the two export ones have *never* worked since they were
3653 introduced. It seems strange in any case to be adding new export
3654 ciphersuites, and given "logjam" it also does not seem correct to fix them.
3658 * Version negotiation has been rewritten. In particular SSLv23_method(),
3659 SSLv23_client_method() and SSLv23_server_method() have been deprecated,
3660 and turned into macros which simply call the new preferred function names
3661 TLS_method(), TLS_client_method() and TLS_server_method(). All new code
3662 should use the new names instead. Also as part of this change the ssl23.h
3663 header file has been removed.
3667 * Support for Kerberos ciphersuites in TLS (RFC2712) has been removed. This
3668 code and the associated standard is no longer considered fit-for-purpose.
3672 * RT2547 was closed. When generating a private key, try to make the
3673 output file readable only by the owner. This behavior change might
3674 be noticeable when interacting with other software.
3676 * Documented all exdata functions. Added CRYPTO_free_ex_index.
3681 * Added HTTP GET support to the ocsp command.
3685 * Changed default digest for the dgst and enc commands from MD5 to
3690 * RAND_pseudo_bytes has been deprecated. Users should use RAND_bytes instead.
3694 * Added support for TLS extended master secret from
3695 draft-ietf-tls-session-hash-03.txt. Thanks for Alfredo Pironti for an
3696 initial patch which was a great help during development.
3700 * All libssl internal structures have been removed from the public header
3701 files, and the OPENSSL_NO_SSL_INTERN option has been removed (since it is
3702 now redundant). Users should not attempt to access internal structures
3703 directly. Instead they should use the provided API functions.
3707 * config has been changed so that by default OPENSSL_NO_DEPRECATED is used.
3708 Access to deprecated functions can be re-enabled by running config with
3709 "enable-deprecated". In addition applications wishing to use deprecated
3710 functions must define OPENSSL_USE_DEPRECATED. Note that this new behaviour
3711 will, by default, disable some transitive includes that previously existed
3712 in the header files (e.g. ec.h will no longer, by default, include bn.h)
3716 * Added support for OCB mode. OpenSSL has been granted a patent license
3717 compatible with the OpenSSL license for use of OCB. Details are available
3718 at <https://www.openssl.org/source/OCB-patent-grant-OpenSSL.pdf>. Support
3719 for OCB can be removed by calling config with no-ocb.
3723 * SSLv2 support has been removed. It still supports receiving a SSLv2
3724 compatible client hello.
3728 * Increased the minimal RSA keysize from 256 to 512 bits [Rich Salz],
3729 done while fixing the error code for the key-too-small case.
3731 *Annie Yousar <a.yousar@informatik.hu-berlin.de>*
3733 * CA.sh has been removed; use CA.pl instead.
3737 * Removed old DES API.
3741 * Remove various unsupported platforms:
3747 Sinix/ReliantUNIX RM400
3752 16-bit platforms such as WIN16
3756 * Clean up OPENSSL_NO_xxx #define's
3757 - Use setbuf() and remove OPENSSL_NO_SETVBUF_IONBF
3758 - Rename OPENSSL_SYSNAME_xxx to OPENSSL_SYS_xxx
3759 - OPENSSL_NO_EC{DH,DSA} merged into OPENSSL_NO_EC
3760 - OPENSSL_NO_RIPEMD160, OPENSSL_NO_RIPEMD merged into OPENSSL_NO_RMD160
3761 - OPENSSL_NO_FP_API merged into OPENSSL_NO_STDIO
3762 - Remove OPENSSL_NO_BIO OPENSSL_NO_BUFFER OPENSSL_NO_CHAIN_VERIFY
3763 OPENSSL_NO_EVP OPENSSL_NO_FIPS_ERR OPENSSL_NO_HASH_COMP
3764 OPENSSL_NO_LHASH OPENSSL_NO_OBJECT OPENSSL_NO_SPEED OPENSSL_NO_STACK
3765 OPENSSL_NO_X509 OPENSSL_NO_X509_VERIFY
3766 - Remove MS_STATIC; it's a relic from platforms <32 bits.
3770 * Cleaned up dead code
3771 Remove all but one '#ifdef undef' which is to be looked at.
3775 * Clean up calling of xxx_free routines.
3776 Just like free(), fix most of the xxx_free routines to accept
3777 NULL. Remove the non-null checks from callers. Save much code.
3781 * Add secure heap for storage of private keys (when possible).
3782 Add BIO_s_secmem(), CBIGNUM, etc.
3783 Contributed by Akamai Technologies under our Corporate CLA.
3787 * Experimental support for a new, fast, unbiased prime candidate generator,
3788 bn_probable_prime_dh_coprime(). Not currently used by any prime generator.
3790 *Felix Laurie von Massenbach <felix@erbridge.co.uk>*
3792 * New output format NSS in the sess_id command line tool. This allows
3793 exporting the session id and the master key in NSS keylog format.
3795 *Martin Kaiser <martin@kaiser.cx>*
3797 * Harmonize version and its documentation. -f flag is used to display
3800 *mancha <mancha1@zoho.com>*
3802 * Fix eckey_priv_encode so it immediately returns an error upon a failure
3803 in i2d_ECPrivateKey. Thanks to Ted Unangst for feedback on this issue.
3805 *mancha <mancha1@zoho.com>*
3807 * Fix some double frees. These are not thought to be exploitable.
3809 *mancha <mancha1@zoho.com>*
3811 * A missing bounds check in the handling of the TLS heartbeat extension
3812 can be used to reveal up to 64k of memory to a connected client or
3815 Thanks for Neel Mehta of Google Security for discovering this bug and to
3816 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
3817 preparing the fix ([CVE-2014-0160])
3819 *Adam Langley, Bodo Moeller*
3821 * Fix for the attack described in the paper "Recovering OpenSSL
3822 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
3823 by Yuval Yarom and Naomi Benger. Details can be obtained from:
3824 <http://eprint.iacr.org/2014/140>
3826 Thanks to Yuval Yarom and Naomi Benger for discovering this
3827 flaw and to Yuval Yarom for supplying a fix ([CVE-2014-0076])
3829 *Yuval Yarom and Naomi Benger*
3831 * Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
3832 this fixes a limitation in previous versions of OpenSSL.
3836 * Experimental encrypt-then-mac support.
3838 Experimental support for encrypt then mac from
3839 draft-gutmann-tls-encrypt-then-mac-02.txt
3841 To enable it set the appropriate extension number (0x42 for the test
3842 server) using e.g. -DTLSEXT_TYPE_encrypt_then_mac=0x42
3844 For non-compliant peers (i.e. just about everything) this should have no
3847 WARNING: EXPERIMENTAL, SUBJECT TO CHANGE.
3851 * Add EVP support for key wrapping algorithms, to avoid problems with
3852 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
3853 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
3854 algorithms and include tests cases.
3858 * Extend CMS code to support RSA-PSS signatures and RSA-OAEP for
3863 * Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
3864 MGF1 digest and OAEP label.
3868 * Make openssl verify return errors.
3870 *Chris Palmer <palmer@google.com> and Ben Laurie*
3872 * New function ASN1_TIME_diff to calculate the difference between two
3873 ASN1_TIME structures or one structure and the current time.
3877 * Update fips_test_suite to support multiple command line options. New
3878 test to induce all self test errors in sequence and check expected
3883 * Add FIPS_{rsa,dsa,ecdsa}_{sign,verify} functions which digest and
3884 sign or verify all in one operation.
3888 * Add fips_algvs: a multicall fips utility incorporating all the algorithm
3889 test programs and fips_test_suite. Includes functionality to parse
3890 the minimal script output of fipsalgest.pl directly.
3894 * Add authorisation parameter to FIPS_module_mode_set().
3898 * Add FIPS selftest for ECDH algorithm using P-224 and B-233 curves.
3902 * Use separate DRBG fields for internal and external flags. New function
3903 FIPS_drbg_health_check() to perform on demand health checking. Add
3904 generation tests to fips_test_suite with reduced health check interval to
3905 demonstrate periodic health checking. Add "nodh" option to
3906 fips_test_suite to skip very slow DH test.
3910 * New function FIPS_get_cipherbynid() to lookup FIPS supported ciphers
3915 * More extensive health check for DRBG checking many more failure modes.
3916 New function FIPS_selftest_drbg_all() to handle every possible DRBG
3917 combination: call this in fips_test_suite.
3921 * Add support for canonical generation of DSA parameter 'g'. See
3924 * Add support for HMAC DRBG from SP800-90. Update DRBG algorithm test and
3925 POST to handle HMAC cases.
3929 * Add functions FIPS_module_version() and FIPS_module_version_text()
3930 to return numerical and string versions of the FIPS module number.
3934 * Rename FIPS_mode_set and FIPS_mode to FIPS_module_mode_set and
3935 FIPS_module_mode. FIPS_mode and FIPS_mode_set will be implemented
3936 outside the validated module in the FIPS capable OpenSSL.
3940 * Minor change to DRBG entropy callback semantics. In some cases
3941 there is no multiple of the block length between min_len and
3942 max_len. Allow the callback to return more than max_len bytes
3943 of entropy but discard any extra: it is the callback's responsibility
3944 to ensure that the extra data discarded does not impact the
3945 requested amount of entropy.
3949 * Add PRNG security strength checks to RSA, DSA and ECDSA using
3950 information in FIPS186-3, SP800-57 and SP800-131A.
3954 * CCM support via EVP. Interface is very similar to GCM case except we
3955 must supply all data in one chunk (i.e. no update, final) and the
3956 message length must be supplied if AAD is used. Add algorithm test
3961 * Initial version of POST overhaul. Add POST callback to allow the status
3962 of POST to be monitored and/or failures induced. Modify fips_test_suite
3963 to use callback. Always run all selftests even if one fails.
3967 * XTS support including algorithm test driver in the fips_gcmtest program.
3968 Note: this does increase the maximum key length from 32 to 64 bytes but
3969 there should be no binary compatibility issues as existing applications
3970 will never use XTS mode.
3974 * Extensive reorganisation of FIPS PRNG behaviour. Remove all dependencies
3975 to OpenSSL RAND code and replace with a tiny FIPS RAND API which also
3976 performs algorithm blocking for unapproved PRNG types. Also do not
3977 set PRNG type in FIPS_mode_set(): leave this to the application.
3978 Add default OpenSSL DRBG handling: sets up FIPS PRNG and seeds with
3979 the standard OpenSSL PRNG: set additional data to a date time vector.
3983 * Rename old X9.31 PRNG functions of the form `FIPS_rand*` to `FIPS_x931*`.
3984 This shouldn't present any incompatibility problems because applications
3985 shouldn't be using these directly and any that are will need to rethink
3986 anyway as the X9.31 PRNG is now deprecated by FIPS 140-2
3990 * Extensive self tests and health checking required by SP800-90 DRBG.
3991 Remove strength parameter from FIPS_drbg_instantiate and always
3992 instantiate at maximum supported strength.
3996 * Add ECDH code to fips module and fips_ecdhvs for primitives only testing.
4000 * New algorithm test program fips_dhvs to handle DH primitives only testing.
4004 * New function DH_compute_key_padded() to compute a DH key and pad with
4005 leading zeroes if needed: this complies with SP800-56A et al.
4009 * Initial implementation of SP800-90 DRBGs for Hash and CTR. Not used by
4010 anything, incomplete, subject to change and largely untested at present.
4014 * Modify fipscanisteronly build option to only build the necessary object
4015 files by filtering FIPS_EX_OBJ through a perl script in crypto/Makefile.
4019 * Add experimental option FIPSSYMS to give all symbols in
4020 fipscanister.o and FIPS or fips prefix. This will avoid
4021 conflicts with future versions of OpenSSL. Add perl script
4022 util/fipsas.pl to preprocess assembly language source files
4023 and rename any affected symbols.
4027 * Add selftest checks and algorithm block of non-fips algorithms in
4028 FIPS mode. Remove DES2 from selftests.
4032 * Add ECDSA code to fips module. Add tiny fips_ecdsa_check to just
4033 return internal method without any ENGINE dependencies. Add new
4034 tiny fips sign and verify functions.
4038 * New build option no-ec2m to disable characteristic 2 code.
4042 * New build option "fipscanisteronly". This only builds fipscanister.o
4043 and (currently) associated fips utilities. Uses the file Makefile.fips
4044 instead of Makefile.org as the prototype.
4048 * Add some FIPS mode restrictions to GCM. Add internal IV generator.
4049 Update fips_gcmtest to use IV generator.
4053 * Initial, experimental EVP support for AES-GCM. AAD can be input by
4054 setting output buffer to NULL. The `*Final` function must be
4055 called although it will not retrieve any additional data. The tag
4056 can be set or retrieved with a ctrl. The IV length is by default 12
4057 bytes (96 bits) but can be set to an alternative value. If the IV
4058 length exceeds the maximum IV length (currently 16 bytes) it cannot be
4063 * New flag in ciphers: EVP_CIPH_FLAG_CUSTOM_CIPHER. This means the
4064 underlying do_cipher function handles all cipher semantics itself
4065 including padding and finalisation. This is useful if (for example)
4066 an ENGINE cipher handles block padding itself. The behaviour of
4067 do_cipher is subtly changed if this flag is set: the return value
4068 is the number of characters written to the output buffer (zero is
4069 no longer an error code) or a negative error code. Also if the
4070 input buffer is NULL and length 0 finalisation should be performed.
4074 * If a candidate issuer certificate is already part of the constructed
4075 path ignore it: new debug notification X509_V_ERR_PATH_LOOP for this case.
4079 * Improve forward-security support: add functions
4081 void SSL_CTX_set_not_resumable_session_callback(
4082 SSL_CTX *ctx, int (*cb)(SSL *ssl, int is_forward_secure))
4083 void SSL_set_not_resumable_session_callback(
4084 SSL *ssl, int (*cb)(SSL *ssl, int is_forward_secure))
4086 for use by SSL/TLS servers; the callback function will be called whenever a
4087 new session is created, and gets to decide whether the session may be
4088 cached to make it resumable (return 0) or not (return 1). (As by the
4089 SSL/TLS protocol specifications, the session_id sent by the server will be
4090 empty to indicate that the session is not resumable; also, the server will
4091 not generate RFC 4507 (RFC 5077) session tickets.)
4093 A simple reasonable callback implementation is to return is_forward_secure.
4094 This parameter will be set to 1 or 0 depending on the ciphersuite selected
4095 by the SSL/TLS server library, indicating whether it can provide forward
4098 *Emilia Käsper <emilia.kasper@esat.kuleuven.be> (Google)*
4100 * New -verify_name option in command line utilities to set verification
4105 * Initial CMAC implementation. WARNING: EXPERIMENTAL, API MAY CHANGE.
4106 Add CMAC pkey methods.
4110 * Experimental renegotiation in s_server -www mode. If the client
4111 browses /reneg connection is renegotiated. If /renegcert it is
4112 renegotiated requesting a certificate.
4116 * Add an "external" session cache for debugging purposes to s_server. This
4117 should help trace issues which normally are only apparent in deployed
4118 multi-process servers.
4122 * Extensive audit of libcrypto with DEBUG_UNUSED. Fix many cases where
4123 return value is ignored. NB. The functions RAND_add(), RAND_seed(),
4124 BIO_set_cipher() and some obscure PEM functions were changed so they
4125 can now return an error. The RAND changes required a change to the
4126 RAND_METHOD structure.
4130 * New macro `__owur` for "OpenSSL Warn Unused Result". This makes use of
4131 a gcc attribute to warn if the result of a function is ignored. This
4132 is enable if DEBUG_UNUSED is set. Add to several functions in evp.h
4133 whose return value is often ignored.
4137 * New -noct, -requestct, -requirect and -ctlogfile options for s_client.
4138 These allow SCTs (signed certificate timestamps) to be requested and
4139 validated when establishing a connection.
4141 *Rob Percival <robpercival@google.com>*
4146 ### Changes between 1.0.2s and 1.0.2t [10 Sep 2019]
4148 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
4149 used even when parsing explicit parameters, when loading a encoded key
4150 or calling `EC_GROUP_new_from_ecpkparameters()`/
4151 `EC_GROUP_new_from_ecparameters()`.
4152 This prevents bypass of security hardening and performance gains,
4153 especially for curves with specialized EC_METHODs.
4154 By default, if a key encoded with explicit parameters is loaded and later
4155 encoded, the output is still encoded with explicit parameters, even if
4156 internally a "named" EC_GROUP is used for computation.
4160 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
4161 this change, EC_GROUP_set_generator would accept order and/or cofactor as
4162 NULL. After this change, only the cofactor parameter can be NULL. It also
4163 does some minimal sanity checks on the passed order.
4168 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
4169 An attack is simple, if the first CMS_recipientInfo is valid but the
4170 second CMS_recipientInfo is chosen ciphertext. If the second
4171 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
4172 encryption key will be replaced by garbage, and the message cannot be
4173 decoded, but if the RSA decryption fails, the correct encryption key is
4174 used and the recipient will not notice the attack.
4175 As a work around for this potential attack the length of the decrypted
4176 key must be equal to the cipher default key length, in case the
4177 certifiate is not given and all recipientInfo are tried out.
4178 The old behaviour can be re-enabled in the CMS code by setting the
4179 CMS_DEBUG_DECRYPT flag.
4184 * Document issue with installation paths in diverse Windows builds
4186 '/usr/local/ssl' is an unsafe prefix for location to install OpenSSL
4187 binaries and run-time config file.
4192 ### Changes between 1.0.2r and 1.0.2s [28 May 2019]
4194 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
4195 This changes the size when using the `genpkey` command when no size is given.
4196 It fixes an omission in earlier changes that changed all RSA, DSA and DH
4197 generation commands to use 2048 bits by default.
4201 * Add FIPS support for Android Arm 64-bit
4203 Support for Android Arm 64-bit was added to the OpenSSL FIPS Object
4204 Module in Version 2.0.10. For some reason, the corresponding target
4205 'android64-aarch64' was missing OpenSSL 1.0.2, whence it could not be
4206 built with FIPS support on Android Arm 64-bit. This omission has been
4209 *Matthias St. Pierre*
4211 ### Changes between 1.0.2q and 1.0.2r [26 Feb 2019]
4213 * 0-byte record padding oracle
4215 If an application encounters a fatal protocol error and then calls
4216 SSL_shutdown() twice (once to send a close_notify, and once to receive one)
4217 then OpenSSL can respond differently to the calling application if a 0 byte
4218 record is received with invalid padding compared to if a 0 byte record is
4219 received with an invalid MAC. If the application then behaves differently
4220 based on that in a way that is detectable to the remote peer, then this
4221 amounts to a padding oracle that could be used to decrypt data.
4223 In order for this to be exploitable "non-stitched" ciphersuites must be in
4224 use. Stitched ciphersuites are optimised implementations of certain
4225 commonly used ciphersuites. Also the application must call SSL_shutdown()
4226 twice even if a protocol error has occurred (applications should not do
4227 this but some do anyway).
4229 This issue was discovered by Juraj Somorovsky, Robert Merget and Nimrod
4230 Aviram, with additional investigation by Steven Collison and Andrew
4231 Hourselt. It was reported to OpenSSL on 10th December 2018.
4236 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
4240 ### Changes between 1.0.2p and 1.0.2q [20 Nov 2018]
4242 * Microarchitecture timing vulnerability in ECC scalar multiplication
4244 OpenSSL ECC scalar multiplication, used in e.g. ECDSA and ECDH, has been
4245 shown to be vulnerable to a microarchitecture timing side channel attack.
4246 An attacker with sufficient access to mount local timing attacks during
4247 ECDSA signature generation could recover the private key.
4249 This issue was reported to OpenSSL on 26th October 2018 by Alejandro
4250 Cabrera Aldaya, Billy Brumley, Sohaib ul Hassan, Cesar Pereida Garcia and
4256 * Timing vulnerability in DSA signature generation
4258 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
4259 timing side channel attack. An attacker could use variations in the signing
4260 algorithm to recover the private key.
4262 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
4267 * Resolve a compatibility issue in EC_GROUP handling with the FIPS Object
4268 Module, accidentally introduced while backporting security fixes from the
4269 development branch and hindering the use of ECC in FIPS mode.
4273 ### Changes between 1.0.2o and 1.0.2p [14 Aug 2018]
4275 * Client DoS due to large DH parameter
4277 During key agreement in a TLS handshake using a DH(E) based ciphersuite a
4278 malicious server can send a very large prime value to the client. This will
4279 cause the client to spend an unreasonably long period of time generating a
4280 key for this prime resulting in a hang until the client has finished. This
4281 could be exploited in a Denial Of Service attack.
4283 This issue was reported to OpenSSL on 5th June 2018 by Guido Vranken
4288 * Cache timing vulnerability in RSA Key Generation
4290 The OpenSSL RSA Key generation algorithm has been shown to be vulnerable to
4291 a cache timing side channel attack. An attacker with sufficient access to
4292 mount cache timing attacks during the RSA key generation process could
4293 recover the private key.
4295 This issue was reported to OpenSSL on 4th April 2018 by Alejandro Cabrera
4296 Aldaya, Billy Brumley, Cesar Pereida Garcia and Luis Manuel Alvarez Tapia.
4301 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
4302 parameter is no longer accepted, as it leads to a corrupt table. NULL
4303 pem_str is reserved for alias entries only.
4307 * Revert blinding in ECDSA sign and instead make problematic addition
4308 length-invariant. Switch even to fixed-length Montgomery multiplication.
4312 * Change generating and checking of primes so that the error rate of not
4313 being prime depends on the intended use based on the size of the input.
4314 For larger primes this will result in more rounds of Miller-Rabin.
4315 The maximal error rate for primes with more than 1080 bits is lowered
4318 *Kurt Roeckx, Annie Yousar*
4320 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
4324 * Add blinding to ECDSA and DSA signatures to protect against side channel
4325 attacks discovered by Keegan Ryan (NCC Group).
4329 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
4330 now allow empty (zero character) pass phrases.
4334 * Certificate time validation (X509_cmp_time) enforces stricter
4335 compliance with RFC 5280. Fractional seconds and timezone offsets
4336 are no longer allowed.
4340 ### Changes between 1.0.2n and 1.0.2o [27 Mar 2018]
4342 * Constructed ASN.1 types with a recursive definition could exceed the stack
4344 Constructed ASN.1 types with a recursive definition (such as can be found
4345 in PKCS7) could eventually exceed the stack given malicious input with
4346 excessive recursion. This could result in a Denial Of Service attack. There
4347 are no such structures used within SSL/TLS that come from untrusted sources
4348 so this is considered safe.
4350 This issue was reported to OpenSSL on 4th January 2018 by the OSS-fuzz
4356 ### Changes between 1.0.2m and 1.0.2n [7 Dec 2017]
4358 * Read/write after SSL object in error state
4360 OpenSSL 1.0.2 (starting from version 1.0.2b) introduced an "error state"
4361 mechanism. The intent was that if a fatal error occurred during a handshake
4362 then OpenSSL would move into the error state and would immediately fail if
4363 you attempted to continue the handshake. This works as designed for the
4364 explicit handshake functions (SSL_do_handshake(), SSL_accept() and
4365 SSL_connect()), however due to a bug it does not work correctly if
4366 SSL_read() or SSL_write() is called directly. In that scenario, if the
4367 handshake fails then a fatal error will be returned in the initial function
4368 call. If SSL_read()/SSL_write() is subsequently called by the application
4369 for the same SSL object then it will succeed and the data is passed without
4370 being decrypted/encrypted directly from the SSL/TLS record layer.
4372 In order to exploit this issue an application bug would have to be present
4373 that resulted in a call to SSL_read()/SSL_write() being issued after having
4374 already received a fatal error.
4376 This issue was reported to OpenSSL by David Benjamin (Google).
4381 * rsaz_1024_mul_avx2 overflow bug on x86_64
4383 There is an overflow bug in the AVX2 Montgomery multiplication procedure
4384 used in exponentiation with 1024-bit moduli. No EC algorithms are affected.
4385 Analysis suggests that attacks against RSA and DSA as a result of this
4386 defect would be very difficult to perform and are not believed likely.
4387 Attacks against DH1024 are considered just feasible, because most of the
4388 work necessary to deduce information about a private key may be performed
4389 offline. The amount of resources required for such an attack would be
4390 significant. However, for an attack on TLS to be meaningful, the server
4391 would have to share the DH1024 private key among multiple clients, which is
4392 no longer an option since CVE-2016-0701.
4394 This only affects processors that support the AVX2 but not ADX extensions
4395 like Intel Haswell (4th generation).
4397 This issue was reported to OpenSSL by David Benjamin (Google). The issue
4398 was originally found via the OSS-Fuzz project.
4403 ### Changes between 1.0.2l and 1.0.2m [2 Nov 2017]
4405 * bn_sqrx8x_internal carry bug on x86_64
4407 There is a carry propagating bug in the x86_64 Montgomery squaring
4408 procedure. No EC algorithms are affected. Analysis suggests that attacks
4409 against RSA and DSA as a result of this defect would be very difficult to
4410 perform and are not believed likely. Attacks against DH are considered just
4411 feasible (although very difficult) because most of the work necessary to
4412 deduce information about a private key may be performed offline. The amount
4413 of resources required for such an attack would be very significant and
4414 likely only accessible to a limited number of attackers. An attacker would
4415 additionally need online access to an unpatched system using the target
4416 private key in a scenario with persistent DH parameters and a private
4417 key that is shared between multiple clients.
4419 This only affects processors that support the BMI1, BMI2 and ADX extensions
4420 like Intel Broadwell (5th generation) and later or AMD Ryzen.
4422 This issue was reported to OpenSSL by the OSS-Fuzz project.
4427 * Malformed X.509 IPAddressFamily could cause OOB read
4429 If an X.509 certificate has a malformed IPAddressFamily extension,
4430 OpenSSL could do a one-byte buffer overread. The most likely result
4431 would be an erroneous display of the certificate in text format.
4433 This issue was reported to OpenSSL by the OSS-Fuzz project.
4437 ### Changes between 1.0.2k and 1.0.2l [25 May 2017]
4439 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
4440 platform rather than 'mingw'.
4444 ### Changes between 1.0.2j and 1.0.2k [26 Jan 2017]
4446 * Truncated packet could crash via OOB read
4448 If one side of an SSL/TLS path is running on a 32-bit host and a specific
4449 cipher is being used, then a truncated packet can cause that host to
4450 perform an out-of-bounds read, usually resulting in a crash.
4452 This issue was reported to OpenSSL by Robert Święcki of Google.
4457 * BN_mod_exp may produce incorrect results on x86_64
4459 There is a carry propagating bug in the x86_64 Montgomery squaring
4460 procedure. No EC algorithms are affected. Analysis suggests that attacks
4461 against RSA and DSA as a result of this defect would be very difficult to
4462 perform and are not believed likely. Attacks against DH are considered just
4463 feasible (although very difficult) because most of the work necessary to
4464 deduce information about a private key may be performed offline. The amount
4465 of resources required for such an attack would be very significant and
4466 likely only accessible to a limited number of attackers. An attacker would
4467 additionally need online access to an unpatched system using the target
4468 private key in a scenario with persistent DH parameters and a private
4469 key that is shared between multiple clients. For example this can occur by
4470 default in OpenSSL DHE based SSL/TLS ciphersuites. Note: This issue is very
4471 similar to CVE-2015-3193 but must be treated as a separate problem.
4473 This issue was reported to OpenSSL by the OSS-Fuzz project.
4478 * Montgomery multiplication may produce incorrect results
4480 There is a carry propagating bug in the Broadwell-specific Montgomery
4481 multiplication procedure that handles input lengths divisible by, but
4482 longer than 256 bits. Analysis suggests that attacks against RSA, DSA
4483 and DH private keys are impossible. This is because the subroutine in
4484 question is not used in operations with the private key itself and an input
4485 of the attacker's direct choice. Otherwise the bug can manifest itself as
4486 transient authentication and key negotiation failures or reproducible
4487 erroneous outcome of public-key operations with specially crafted input.
4488 Among EC algorithms only Brainpool P-512 curves are affected and one
4489 presumably can attack ECDH key negotiation. Impact was not analyzed in
4490 detail, because pre-requisites for attack are considered unlikely. Namely
4491 multiple clients have to choose the curve in question and the server has to
4492 share the private key among them, neither of which is default behaviour.
4493 Even then only clients that chose the curve will be affected.
4495 This issue was publicly reported as transient failures and was not
4496 initially recognized as a security issue. Thanks to Richard Morgan for
4497 providing reproducible case.
4502 * OpenSSL now fails if it receives an unrecognised record type in TLS1.0
4503 or TLS1.1. Previously this only happened in SSLv3 and TLS1.2. This is to
4504 prevent issues where no progress is being made and the peer continually
4505 sends unrecognised record types, using up resources processing them.
4509 ### Changes between 1.0.2i and 1.0.2j [26 Sep 2016]
4511 * Missing CRL sanity check
4513 A bug fix which included a CRL sanity check was added to OpenSSL 1.1.0
4514 but was omitted from OpenSSL 1.0.2i. As a result any attempt to use
4515 CRLs in OpenSSL 1.0.2i will crash with a null pointer exception.
4517 This issue only affects the OpenSSL 1.0.2i
4522 ### Changes between 1.0.2h and 1.0.2i [22 Sep 2016]
4524 * OCSP Status Request extension unbounded memory growth
4526 A malicious client can send an excessively large OCSP Status Request
4527 extension. If that client continually requests renegotiation, sending a
4528 large OCSP Status Request extension each time, then there will be unbounded
4529 memory growth on the server. This will eventually lead to a Denial Of
4530 Service attack through memory exhaustion. Servers with a default
4531 configuration are vulnerable even if they do not support OCSP. Builds using
4532 the "no-ocsp" build time option are not affected.
4534 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4539 * In order to mitigate the SWEET32 attack, the DES ciphers were moved from
4542 This issue was reported to OpenSSL Karthikeyan Bhargavan and Gaetan
4548 * OOB write in MDC2_Update()
4550 An overflow can occur in MDC2_Update() either if called directly or
4551 through the EVP_DigestUpdate() function using MDC2. If an attacker
4552 is able to supply very large amounts of input data after a previous
4553 call to EVP_EncryptUpdate() with a partial block then a length check
4554 can overflow resulting in a heap corruption.
4556 The amount of data needed is comparable to SIZE_MAX which is impractical
4559 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4564 * Malformed SHA512 ticket DoS
4566 If a server uses SHA512 for TLS session ticket HMAC it is vulnerable to a
4567 DoS attack where a malformed ticket will result in an OOB read which will
4570 The use of SHA512 in TLS session tickets is comparatively rare as it requires
4571 a custom server callback and ticket lookup mechanism.
4573 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4578 * OOB write in BN_bn2dec()
4580 The function BN_bn2dec() does not check the return value of BN_div_word().
4581 This can cause an OOB write if an application uses this function with an
4582 overly large BIGNUM. This could be a problem if an overly large certificate
4583 or CRL is printed out from an untrusted source. TLS is not affected because
4584 record limits will reject an oversized certificate before it is parsed.
4586 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4591 * OOB read in TS_OBJ_print_bio()
4593 The function TS_OBJ_print_bio() misuses OBJ_obj2txt(): the return value is
4594 the total length the OID text representation would use and not the amount
4595 of data written. This will result in OOB reads when large OIDs are
4598 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4603 * Pointer arithmetic undefined behaviour
4605 Avoid some undefined pointer arithmetic
4607 A common idiom in the codebase is to check limits in the following manner:
4610 Where "p" points to some malloc'd data of SIZE bytes and
4613 "len" here could be from some externally supplied data (e.g. from a TLS
4616 The rules of C pointer arithmetic are such that "p + len" is only well
4617 defined where len <= SIZE. Therefore the above idiom is actually
4618 undefined behaviour.
4620 For example this could cause problems if some malloc implementation
4621 provides an address for "p" such that "p + len" actually overflows for
4622 values of len that are too big and therefore p + len < limit.
4624 This issue was reported to OpenSSL by Guido Vranken
4629 * Constant time flag not preserved in DSA signing
4631 Operations in the DSA signing algorithm should run in constant time in
4632 order to avoid side channel attacks. A flaw in the OpenSSL DSA
4633 implementation means that a non-constant time codepath is followed for
4634 certain operations. This has been demonstrated through a cache-timing
4635 attack to be sufficient for an attacker to recover the private DSA key.
4637 This issue was reported by César Pereida (Aalto University), Billy Brumley
4638 (Tampere University of Technology), and Yuval Yarom (The University of
4639 Adelaide and NICTA).
4644 * DTLS buffered message DoS
4646 In a DTLS connection where handshake messages are delivered out-of-order
4647 those messages that OpenSSL is not yet ready to process will be buffered
4648 for later use. Under certain circumstances, a flaw in the logic means that
4649 those messages do not get removed from the buffer even though the handshake
4650 has been completed. An attacker could force up to approx. 15 messages to
4651 remain in the buffer when they are no longer required. These messages will
4652 be cleared when the DTLS connection is closed. The default maximum size for
4653 a message is 100k. Therefore the attacker could force an additional 1500k
4654 to be consumed per connection. By opening many simulataneous connections an
4655 attacker could cause a DoS attack through memory exhaustion.
4657 This issue was reported to OpenSSL by Quan Luo.
4662 * DTLS replay protection DoS
4664 A flaw in the DTLS replay attack protection mechanism means that records
4665 that arrive for future epochs update the replay protection "window" before
4666 the MAC for the record has been validated. This could be exploited by an
4667 attacker by sending a record for the next epoch (which does not have to
4668 decrypt or have a valid MAC), with a very large sequence number. This means
4669 that all subsequent legitimate packets are dropped causing a denial of
4670 service for a specific DTLS connection.
4672 This issue was reported to OpenSSL by the OCAP audit team.
4677 * Certificate message OOB reads
4679 In OpenSSL 1.0.2 and earlier some missing message length checks can result
4680 in OOB reads of up to 2 bytes beyond an allocated buffer. There is a
4681 theoretical DoS risk but this has not been observed in practice on common
4684 The messages affected are client certificate, client certificate request
4685 and server certificate. As a result the attack can only be performed
4686 against a client or a server which enables client authentication.
4688 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4693 ### Changes between 1.0.2g and 1.0.2h [3 May 2016]
4695 * Prevent padding oracle in AES-NI CBC MAC check
4697 A MITM attacker can use a padding oracle attack to decrypt traffic
4698 when the connection uses an AES CBC cipher and the server support
4701 This issue was introduced as part of the fix for Lucky 13 padding
4702 attack ([CVE-2013-0169]). The padding check was rewritten to be in
4703 constant time by making sure that always the same bytes are read and
4704 compared against either the MAC or padding bytes. But it no longer
4705 checked that there was enough data to have both the MAC and padding
4708 This issue was reported by Juraj Somorovsky using TLS-Attacker.
4712 * Fix EVP_EncodeUpdate overflow
4714 An overflow can occur in the EVP_EncodeUpdate() function which is used for
4715 Base64 encoding of binary data. If an attacker is able to supply very large
4716 amounts of input data then a length check can overflow resulting in a heap
4719 Internally to OpenSSL the EVP_EncodeUpdate() function is primarily used by
4720 the `PEM_write_bio*` family of functions. These are mainly used within the
4721 OpenSSL command line applications, so any application which processes data
4722 from an untrusted source and outputs it as a PEM file should be considered
4723 vulnerable to this issue. User applications that call these APIs directly
4724 with large amounts of untrusted data may also be vulnerable.
4726 This issue was reported by Guido Vranken.
4731 * Fix EVP_EncryptUpdate overflow
4733 An overflow can occur in the EVP_EncryptUpdate() function. If an attacker
4734 is able to supply very large amounts of input data after a previous call to
4735 EVP_EncryptUpdate() with a partial block then a length check can overflow
4736 resulting in a heap corruption. Following an analysis of all OpenSSL
4737 internal usage of the EVP_EncryptUpdate() function all usage is one of two
4738 forms. The first form is where the EVP_EncryptUpdate() call is known to be
4739 the first called function after an EVP_EncryptInit(), and therefore that
4740 specific call must be safe. The second form is where the length passed to
4741 EVP_EncryptUpdate() can be seen from the code to be some small value and
4742 therefore there is no possibility of an overflow. Since all instances are
4743 one of these two forms, it is believed that there can be no overflows in
4744 internal code due to this problem. It should be noted that
4745 EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths.
4746 Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances
4747 of these calls have also been analysed too and it is believed there are no
4748 instances in internal usage where an overflow could occur.
4750 This issue was reported by Guido Vranken.
4755 * Prevent ASN.1 BIO excessive memory allocation
4757 When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio()
4758 a short invalid encoding can cause allocation of large amounts of memory
4759 potentially consuming excessive resources or exhausting memory.
4761 Any application parsing untrusted data through d2i BIO functions is
4762 affected. The memory based functions such as d2i_X509() are *not* affected.
4763 Since the memory based functions are used by the TLS library, TLS
4764 applications are not affected.
4766 This issue was reported by Brian Carpenter.
4773 ASN1 Strings that are over 1024 bytes can cause an overread in applications
4774 using the X509_NAME_oneline() function on EBCDIC systems. This could result
4775 in arbitrary stack data being returned in the buffer.
4777 This issue was reported by Guido Vranken.
4782 * Modify behavior of ALPN to invoke callback after SNI/servername
4783 callback, such that updates to the SSL_CTX affect ALPN.
4787 * Remove LOW from the DEFAULT cipher list. This removes singles DES from the
4792 * Only remove the SSLv2 methods with the no-ssl2-method option. When the
4793 methods are enabled and ssl2 is disabled the methods return NULL.
4797 ### Changes between 1.0.2f and 1.0.2g [1 Mar 2016]
4799 * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL.
4800 Builds that are not configured with "enable-weak-ssl-ciphers" will not
4801 provide any "EXPORT" or "LOW" strength ciphers.
4805 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2
4806 is by default disabled at build-time. Builds that are not configured with
4807 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used,
4808 users who want to negotiate SSLv2 via the version-flexible SSLv23_method()
4809 will need to explicitly call either of:
4811 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2);
4813 SSL_clear_options(ssl, SSL_OP_NO_SSLv2);
4815 as appropriate. Even if either of those is used, or the application
4816 explicitly uses the version-specific SSLv2_method() or its client and
4817 server variants, SSLv2 ciphers vulnerable to exhaustive search key
4818 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT
4819 ciphers, and SSLv2 56-bit DES are no longer available.
4824 * Fix a double-free in DSA code
4826 A double free bug was discovered when OpenSSL parses malformed DSA private
4827 keys and could lead to a DoS attack or memory corruption for applications
4828 that receive DSA private keys from untrusted sources. This scenario is
4831 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using
4837 * Disable SRP fake user seed to address a server memory leak.
4839 Add a new method SRP_VBASE_get1_by_user that handles the seed properly.
4841 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
4842 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user
4843 was changed to ignore the "fake user" SRP seed, even if the seed
4846 Users should use SRP_VBASE_get1_by_user instead. Note that in
4847 SRP_VBASE_get1_by_user, caller must free the returned value. Note
4848 also that even though configuring the SRP seed attempts to hide
4849 invalid usernames by continuing the handshake with fake
4850 credentials, this behaviour is not constant time and no strong
4851 guarantees are made that the handshake is indistinguishable from
4852 that of a valid user.
4857 * Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption
4859 In the BN_hex2bn function the number of hex digits is calculated using an
4860 int value `i`. Later `bn_expand` is called with a value of `i * 4`. For
4861 large values of `i` this can result in `bn_expand` not allocating any
4862 memory because `i * 4` is negative. This can leave the internal BIGNUM data
4863 field as NULL leading to a subsequent NULL ptr deref. For very large values
4864 of `i`, the calculation `i * 4` could be a positive value smaller than `i`.
4865 In this case memory is allocated to the internal BIGNUM data field, but it
4866 is insufficiently sized leading to heap corruption. A similar issue exists
4867 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn
4868 is ever called by user applications with very large untrusted hex/dec data.
4869 This is anticipated to be a rare occurrence.
4871 All OpenSSL internal usage of these functions use data that is not expected
4872 to be untrusted, e.g. config file data or application command line
4873 arguments. If user developed applications generate config file data based
4874 on untrusted data then it is possible that this could also lead to security
4875 consequences. This is also anticipated to be rare.
4877 This issue was reported to OpenSSL by Guido Vranken.
4882 * Fix memory issues in `BIO_*printf` functions
4884 The internal `fmtstr` function used in processing a "%s" format string in
4885 the `BIO_*printf` functions could overflow while calculating the length of a
4886 string and cause an OOB read when printing very long strings.
4888 Additionally the internal `doapr_outch` function can attempt to write to an
4889 OOB memory location (at an offset from the NULL pointer) in the event of a
4890 memory allocation failure. In 1.0.2 and below this could be caused where
4891 the size of a buffer to be allocated is greater than INT_MAX. E.g. this
4892 could be in processing a very long "%s" format string. Memory leaks can
4895 The first issue may mask the second issue dependent on compiler behaviour.
4896 These problems could enable attacks where large amounts of untrusted data
4897 is passed to the `BIO_*printf` functions. If applications use these functions
4898 in this way then they could be vulnerable. OpenSSL itself uses these
4899 functions when printing out human-readable dumps of ASN.1 data. Therefore
4900 applications that print this data could be vulnerable if the data is from
4901 untrusted sources. OpenSSL command line applications could also be
4902 vulnerable where they print out ASN.1 data, or if untrusted data is passed
4903 as command line arguments.
4905 Libssl is not considered directly vulnerable. Additionally certificates etc
4906 received via remote connections via libssl are also unlikely to be able to
4907 trigger these issues because of message size limits enforced within libssl.
4909 This issue was reported to OpenSSL Guido Vranken.
4914 * Side channel attack on modular exponentiation
4916 A side-channel attack was found which makes use of cache-bank conflicts on
4917 the Intel Sandy-Bridge microarchitecture which could lead to the recovery
4918 of RSA keys. The ability to exploit this issue is limited as it relies on
4919 an attacker who has control of code in a thread running on the same
4920 hyper-threaded core as the victim thread which is performing decryptions.
4922 This issue was reported to OpenSSL by Yuval Yarom, The University of
4923 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and
4924 Nadia Heninger, University of Pennsylvania with more information at
4925 <http://cachebleed.info>.
4930 * Change the `req` command to generate a 2048-bit RSA/DSA key by default,
4931 if no keysize is specified with default_bits. This fixes an
4932 omission in an earlier change that changed all RSA/DSA key generation
4933 commands to use 2048 bits by default.
4937 ### Changes between 1.0.2e and 1.0.2f [28 Jan 2016]
4939 * DH small subgroups
4941 Historically OpenSSL only ever generated DH parameters based on "safe"
4942 primes. More recently (in version 1.0.2) support was provided for
4943 generating X9.42 style parameter files such as those required for RFC 5114
4944 support. The primes used in such files may not be "safe". Where an
4945 application is using DH configured with parameters based on primes that are
4946 not "safe" then an attacker could use this fact to find a peer's private
4947 DH exponent. This attack requires that the attacker complete multiple
4948 handshakes in which the peer uses the same private DH exponent. For example
4949 this could be used to discover a TLS server's private DH exponent if it's
4950 reusing the private DH exponent or it's using a static DH ciphersuite.
4952 OpenSSL provides the option SSL_OP_SINGLE_DH_USE for ephemeral DH (DHE) in
4953 TLS. It is not on by default. If the option is not set then the server
4954 reuses the same private DH exponent for the life of the server process and
4955 would be vulnerable to this attack. It is believed that many popular
4956 applications do set this option and would therefore not be at risk.
4958 The fix for this issue adds an additional check where a "q" parameter is
4959 available (as is the case in X9.42 based parameters). This detects the
4960 only known attack, and is the only possible defense for static DH
4961 ciphersuites. This could have some performance impact.
4963 Additionally the SSL_OP_SINGLE_DH_USE option has been switched on by
4964 default and cannot be disabled. This could have some performance impact.
4966 This issue was reported to OpenSSL by Antonio Sanso (Adobe).
4971 * SSLv2 doesn't block disabled ciphers
4973 A malicious client can negotiate SSLv2 ciphers that have been disabled on
4974 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
4975 been disabled, provided that the SSLv2 protocol was not also disabled via
4978 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
4979 and Sebastian Schinzel.
4984 ### Changes between 1.0.2d and 1.0.2e [3 Dec 2015]
4986 * BN_mod_exp may produce incorrect results on x86_64
4988 There is a carry propagating bug in the x86_64 Montgomery squaring
4989 procedure. No EC algorithms are affected. Analysis suggests that attacks
4990 against RSA and DSA as a result of this defect would be very difficult to
4991 perform and are not believed likely. Attacks against DH are considered just
4992 feasible (although very difficult) because most of the work necessary to
4993 deduce information about a private key may be performed offline. The amount
4994 of resources required for such an attack would be very significant and
4995 likely only accessible to a limited number of attackers. An attacker would
4996 additionally need online access to an unpatched system using the target
4997 private key in a scenario with persistent DH parameters and a private
4998 key that is shared between multiple clients. For example this can occur by
4999 default in OpenSSL DHE based SSL/TLS ciphersuites.
5001 This issue was reported to OpenSSL by Hanno Böck.
5006 * Certificate verify crash with missing PSS parameter
5008 The signature verification routines will crash with a NULL pointer
5009 dereference if presented with an ASN.1 signature using the RSA PSS
5010 algorithm and absent mask generation function parameter. Since these
5011 routines are used to verify certificate signature algorithms this can be
5012 used to crash any certificate verification operation and exploited in a
5013 DoS attack. Any application which performs certificate verification is
5014 vulnerable including OpenSSL clients and servers which enable client
5017 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
5022 * X509_ATTRIBUTE memory leak
5024 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
5025 memory. This structure is used by the PKCS#7 and CMS routines so any
5026 application which reads PKCS#7 or CMS data from untrusted sources is
5027 affected. SSL/TLS is not affected.
5029 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
5035 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
5036 This changes the decoding behaviour for some invalid messages,
5037 though the change is mostly in the more lenient direction, and
5038 legacy behaviour is preserved as much as possible.
5042 * In DSA_generate_parameters_ex, if the provided seed is too short,
5045 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
5047 ### Changes between 1.0.2c and 1.0.2d [9 Jul 2015]
5049 * Alternate chains certificate forgery
5051 During certificate verification, OpenSSL will attempt to find an
5052 alternative certificate chain if the first attempt to build such a chain
5053 fails. An error in the implementation of this logic can mean that an
5054 attacker could cause certain checks on untrusted certificates to be
5055 bypassed, such as the CA flag, enabling them to use a valid leaf
5056 certificate to act as a CA and "issue" an invalid certificate.
5058 This issue was reported to OpenSSL by Adam Langley/David Benjamin
5063 ### Changes between 1.0.2b and 1.0.2c [12 Jun 2015]
5065 * Fix HMAC ABI incompatibility. The previous version introduced an ABI
5066 incompatibility in the handling of HMAC. The previous ABI has now been
5071 ### Changes between 1.0.2a and 1.0.2b [11 Jun 2015]
5073 * Malformed ECParameters causes infinite loop
5075 When processing an ECParameters structure OpenSSL enters an infinite loop
5076 if the curve specified is over a specially malformed binary polynomial
5079 This can be used to perform denial of service against any
5080 system which processes public keys, certificate requests or
5081 certificates. This includes TLS clients and TLS servers with
5082 client authentication enabled.
5084 This issue was reported to OpenSSL by Joseph Barr-Pixton.
5089 * Exploitable out-of-bounds read in X509_cmp_time
5091 X509_cmp_time does not properly check the length of the ASN1_TIME
5092 string and can read a few bytes out of bounds. In addition,
5093 X509_cmp_time accepts an arbitrary number of fractional seconds in the
5096 An attacker can use this to craft malformed certificates and CRLs of
5097 various sizes and potentially cause a segmentation fault, resulting in
5098 a DoS on applications that verify certificates or CRLs. TLS clients
5099 that verify CRLs are affected. TLS clients and servers with client
5100 authentication enabled may be affected if they use custom verification
5103 This issue was reported to OpenSSL by Robert Swiecki (Google), and
5104 independently by Hanno Böck.
5109 * PKCS7 crash with missing EnvelopedContent
5111 The PKCS#7 parsing code does not handle missing inner EncryptedContent
5112 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
5113 with missing content and trigger a NULL pointer dereference on parsing.
5115 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
5116 structures from untrusted sources are affected. OpenSSL clients and
5117 servers are not affected.
5119 This issue was reported to OpenSSL by Michal Zalewski (Google).
5124 * CMS verify infinite loop with unknown hash function
5126 When verifying a signedData message the CMS code can enter an infinite loop
5127 if presented with an unknown hash function OID. This can be used to perform
5128 denial of service against any system which verifies signedData messages using
5130 This issue was reported to OpenSSL by Johannes Bauer.
5135 * Race condition handling NewSessionTicket
5137 If a NewSessionTicket is received by a multi-threaded client when attempting to
5138 reuse a previous ticket then a race condition can occur potentially leading to
5139 a double free of the ticket data.
5144 * Only support 256-bit or stronger elliptic curves with the
5145 'ecdh_auto' setting (server) or by default (client). Of supported
5146 curves, prefer P-256 (both).
5150 ### Changes between 1.0.2 and 1.0.2a [19 Mar 2015]
5152 * ClientHello sigalgs DoS fix
5154 If a client connects to an OpenSSL 1.0.2 server and renegotiates with an
5155 invalid signature algorithms extension a NULL pointer dereference will
5156 occur. This can be exploited in a DoS attack against the server.
5158 This issue was was reported to OpenSSL by David Ramos of Stanford
5162 *Stephen Henson and Matt Caswell*
5164 * Multiblock corrupted pointer fix
5166 OpenSSL 1.0.2 introduced the "multiblock" performance improvement. This
5167 feature only applies on 64 bit x86 architecture platforms that support AES
5168 NI instructions. A defect in the implementation of "multiblock" can cause
5169 OpenSSL's internal write buffer to become incorrectly set to NULL when
5170 using non-blocking IO. Typically, when the user application is using a
5171 socket BIO for writing, this will only result in a failed connection.
5172 However if some other BIO is used then it is likely that a segmentation
5173 fault will be triggered, thus enabling a potential DoS attack.
5175 This issue was reported to OpenSSL by Daniel Danner and Rainer Mueller.
5180 * Segmentation fault in DTLSv1_listen fix
5182 The DTLSv1_listen function is intended to be stateless and processes the
5183 initial ClientHello from many peers. It is common for user code to loop
5184 over the call to DTLSv1_listen until a valid ClientHello is received with
5185 an associated cookie. A defect in the implementation of DTLSv1_listen means
5186 that state is preserved in the SSL object from one invocation to the next
5187 that can lead to a segmentation fault. Errors processing the initial
5188 ClientHello can trigger this scenario. An example of such an error could be
5189 that a DTLS1.0 only client is attempting to connect to a DTLS1.2 only
5192 This issue was reported to OpenSSL by Per Allansson.
5197 * Segmentation fault in ASN1_TYPE_cmp fix
5199 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
5200 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
5201 certificate signature algorithm consistency this can be used to crash any
5202 certificate verification operation and exploited in a DoS attack. Any
5203 application which performs certificate verification is vulnerable including
5204 OpenSSL clients and servers which enable client authentication.
5209 * Segmentation fault for invalid PSS parameters fix
5211 The signature verification routines will crash with a NULL pointer
5212 dereference if presented with an ASN.1 signature using the RSA PSS
5213 algorithm and invalid parameters. Since these routines are used to verify
5214 certificate signature algorithms this can be used to crash any
5215 certificate verification operation and exploited in a DoS attack. Any
5216 application which performs certificate verification is vulnerable including
5217 OpenSSL clients and servers which enable client authentication.
5219 This issue was was reported to OpenSSL by Brian Carpenter.
5224 * ASN.1 structure reuse memory corruption fix
5226 Reusing a structure in ASN.1 parsing may allow an attacker to cause
5227 memory corruption via an invalid write. Such reuse is and has been
5228 strongly discouraged and is believed to be rare.
5230 Applications that parse structures containing CHOICE or ANY DEFINED BY
5231 components may be affected. Certificate parsing (d2i_X509 and related
5232 functions) are however not affected. OpenSSL clients and servers are
5238 * PKCS7 NULL pointer dereferences fix
5240 The PKCS#7 parsing code does not handle missing outer ContentInfo
5241 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
5242 missing content and trigger a NULL pointer dereference on parsing.
5244 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
5245 otherwise parse PKCS#7 structures from untrusted sources are
5246 affected. OpenSSL clients and servers are not affected.
5248 This issue was reported to OpenSSL by Michal Zalewski (Google).
5253 * DoS via reachable assert in SSLv2 servers fix
5255 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
5256 servers that both support SSLv2 and enable export cipher suites by sending
5257 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
5259 This issue was discovered by Sean Burford (Google) and Emilia Käsper
5260 (OpenSSL development team).
5265 * Empty CKE with client auth and DHE fix
5267 If client auth is used then a server can seg fault in the event of a DHE
5268 ciphersuite being selected and a zero length ClientKeyExchange message
5269 being sent by the client. This could be exploited in a DoS attack.
5274 * Handshake with unseeded PRNG fix
5276 Under certain conditions an OpenSSL 1.0.2 client can complete a handshake
5277 with an unseeded PRNG. The conditions are:
5278 - The client is on a platform where the PRNG has not been seeded
5279 automatically, and the user has not seeded manually
5280 - A protocol specific client method version has been used (i.e. not
5281 SSL_client_methodv23)
5282 - A ciphersuite is used that does not require additional random data from
5283 the PRNG beyond the initial ClientHello client random (e.g. PSK-RC4-SHA).
5285 If the handshake succeeds then the client random that has been used will
5286 have been generated from a PRNG with insufficient entropy and therefore the
5287 output may be predictable.
5289 For example using the following command with an unseeded openssl will
5290 succeed on an unpatched platform:
5292 openssl s_client -psk 1a2b3c4d -tls1_2 -cipher PSK-RC4-SHA
5297 * Use After Free following d2i_ECPrivatekey error fix
5299 A malformed EC private key file consumed via the d2i_ECPrivateKey function
5300 could cause a use after free condition. This, in turn, could cause a double
5301 free in several private key parsing functions (such as d2i_PrivateKey
5302 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
5303 for applications that receive EC private keys from untrusted
5304 sources. This scenario is considered rare.
5306 This issue was discovered by the BoringSSL project and fixed in their
5312 * X509_to_X509_REQ NULL pointer deref fix
5314 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
5315 the certificate key is invalid. This function is rarely used in practice.
5317 This issue was discovered by Brian Carpenter.
5322 * Removed the export ciphers from the DEFAULT ciphers
5326 ### Changes between 1.0.1l and 1.0.2 [22 Jan 2015]
5328 * Facilitate "universal" ARM builds targeting range of ARM ISAs, e.g.
5329 ARMv5 through ARMv8, as opposite to "locking" it to single one.
5330 So far those who have to target multiple platforms would compromise
5331 and argue that binary targeting say ARMv5 would still execute on
5332 ARMv8. "Universal" build resolves this compromise by providing
5333 near-optimal performance even on newer platforms.
5337 * Accelerated NIST P-256 elliptic curve implementation for x86_64
5338 (other platforms pending).
5340 *Shay Gueron & Vlad Krasnov (Intel Corp), Andy Polyakov*
5342 * Add support for the SignedCertificateTimestampList certificate and
5343 OCSP response extensions from RFC6962.
5347 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
5348 for corner cases. (Certain input points at infinity could lead to
5349 bogus results, with non-infinity inputs mapped to infinity too.)
5353 * Initial support for PowerISA 2.0.7, first implemented in POWER8.
5354 This covers AES, SHA256/512 and GHASH. "Initial" means that most
5355 common cases are optimized and there still is room for further
5356 improvements. Vector Permutation AES for Altivec is also added.
5360 * Add support for little-endian ppc64 Linux target.
5362 *Marcelo Cerri (IBM)*
5364 * Initial support for AMRv8 ISA crypto extensions. This covers AES,
5365 SHA1, SHA256 and GHASH. "Initial" means that most common cases
5366 are optimized and there still is room for further improvements.
5367 Both 32- and 64-bit modes are supported.
5369 *Andy Polyakov, Ard Biesheuvel (Linaro)*
5371 * Improved ARMv7 NEON support.
5375 * Support for SPARC Architecture 2011 crypto extensions, first
5376 implemented in SPARC T4. This covers AES, DES, Camellia, SHA1,
5377 SHA256/512, MD5, GHASH and modular exponentiation.
5379 *Andy Polyakov, David Miller*
5381 * Accelerated modular exponentiation for Intel processors, a.k.a.
5384 *Shay Gueron & Vlad Krasnov (Intel Corp)*
5386 * Support for new and upcoming Intel processors, including AVX2,
5387 BMI and SHA ISA extensions. This includes additional "stitched"
5388 implementations, AESNI-SHA256 and GCM, and multi-buffer support
5391 This work was sponsored by Intel Corp.
5395 * Support for DTLS 1.2. This adds two sets of DTLS methods: DTLS_*_method()
5396 supports both DTLS 1.2 and 1.0 and should use whatever version the peer
5397 supports and DTLSv1_2_*_method() which supports DTLS 1.2 only.
5401 * Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
5402 this fixes a limitation in previous versions of OpenSSL.
5406 * Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
5407 MGF1 digest and OAEP label.
5411 * Add EVP support for key wrapping algorithms, to avoid problems with
5412 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
5413 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
5414 algorithms and include tests cases.
5418 * Add functions to allocate and set the fields of an ECDSA_METHOD
5421 *Douglas E. Engert, Steve Henson*
5423 * New functions OPENSSL_gmtime_diff and ASN1_TIME_diff to find the
5424 difference in days and seconds between two tm or ASN1_TIME structures.
5428 * Add -rev test option to s_server to just reverse order of characters
5429 received by client and send back to server. Also prints an abbreviated
5430 summary of the connection parameters.
5434 * New option -brief for s_client and s_server to print out a brief summary
5435 of connection parameters.
5439 * Add callbacks for arbitrary TLS extensions.
5441 *Trevor Perrin <trevp@trevp.net> and Ben Laurie*
5443 * New option -crl_download in several openssl utilities to download CRLs
5444 from CRLDP extension in certificates.
5448 * New options -CRL and -CRLform for s_client and s_server for CRLs.
5452 * New function X509_CRL_diff to generate a delta CRL from the difference
5453 of two full CRLs. Add support to "crl" utility.
5457 * New functions to set lookup_crls function and to retrieve
5458 X509_STORE from X509_STORE_CTX.
5462 * Print out deprecated issuer and subject unique ID fields in
5467 * Extend OCSP I/O functions so they can be used for simple general purpose
5468 HTTP as well as OCSP. New wrapper function which can be used to download
5469 CRLs using the OCSP API.
5473 * Delegate command line handling in s_client/s_server to SSL_CONF APIs.
5477 * `SSL_CONF*` functions. These provide a common framework for application
5478 configuration using configuration files or command lines.
5482 * SSL/TLS tracing code. This parses out SSL/TLS records using the
5483 message callback and prints the results. Needs compile time option
5484 "enable-ssl-trace". New options to s_client and s_server to enable
5489 * New ctrl and macro to retrieve supported points extensions.
5490 Print out extension in s_server and s_client.
5494 * New functions to retrieve certificate signature and signature
5499 * Add functions to retrieve and manipulate the raw cipherlist sent by a
5504 * New Suite B modes for TLS code. These use and enforce the requirements
5505 of RFC6460: restrict ciphersuites, only permit Suite B algorithms and
5506 only use Suite B curves. The Suite B modes can be set by using the
5507 strings "SUITEB128", "SUITEB192" or "SUITEB128ONLY" for the cipherstring.
5511 * New chain verification flags for Suite B levels of security. Check
5512 algorithms are acceptable when flags are set in X509_verify_cert.
5516 * Make tls1_check_chain return a set of flags indicating checks passed
5517 by a certificate chain. Add additional tests to handle client
5518 certificates: checks for matching certificate type and issuer name
5523 * If an attempt is made to use a signature algorithm not in the peer
5524 preference list abort the handshake. If client has no suitable
5525 signature algorithms in response to a certificate request do not
5526 use the certificate.
5530 * If server EC tmp key is not in client preference list abort handshake.
5534 * Add support for certificate stores in CERT structure. This makes it
5535 possible to have different stores per SSL structure or one store in
5536 the parent SSL_CTX. Include distinct stores for certificate chain
5537 verification and chain building. New ctrl SSL_CTRL_BUILD_CERT_CHAIN
5538 to build and store a certificate chain in CERT structure: returning
5539 an error if the chain cannot be built: this will allow applications
5540 to test if a chain is correctly configured.
5542 Note: if the CERT based stores are not set then the parent SSL_CTX
5543 store is used to retain compatibility with existing behaviour.
5547 * New function ssl_set_client_disabled to set a ciphersuite disabled
5548 mask based on the current session, check mask when sending client
5549 hello and checking the requested ciphersuite.
5553 * New ctrls to retrieve and set certificate types in a certificate
5554 request message. Print out received values in s_client. If certificate
5555 types is not set with custom values set sensible values based on
5556 supported signature algorithms.
5560 * Support for distinct client and server supported signature algorithms.
5564 * Add certificate callback. If set this is called whenever a certificate
5565 is required by client or server. An application can decide which
5566 certificate chain to present based on arbitrary criteria: for example
5567 supported signature algorithms. Add very simple example to s_server.
5568 This fixes many of the problems and restrictions of the existing client
5569 certificate callback: for example you can now clear an existing
5570 certificate and specify the whole chain.
5574 * Add new "valid_flags" field to CERT_PKEY structure which determines what
5575 the certificate can be used for (if anything). Set valid_flags field
5576 in new tls1_check_chain function. Simplify ssl_set_cert_masks which used
5577 to have similar checks in it.
5579 Add new "cert_flags" field to CERT structure and include a "strict mode".
5580 This enforces some TLS certificate requirements (such as only permitting
5581 certificate signature algorithms contained in the supported algorithms
5582 extension) which some implementations ignore: this option should be used
5583 with caution as it could cause interoperability issues.
5587 * Update and tidy signature algorithm extension processing. Work out
5588 shared signature algorithms based on preferences and peer algorithms
5589 and print them out in s_client and s_server. Abort handshake if no
5590 shared signature algorithms.
5594 * Add new functions to allow customised supported signature algorithms
5595 for SSL and SSL_CTX structures. Add options to s_client and s_server
5600 * New function SSL_certs_clear() to delete all references to certificates
5601 from an SSL structure. Before this once a certificate had been added
5602 it couldn't be removed.
5606 * Integrate hostname, email address and IP address checking with certificate
5607 verification. New verify options supporting checking in openssl utility.
5611 * Fixes and wildcard matching support to hostname and email checking
5612 functions. Add manual page.
5614 *Florian Weimer (Red Hat Product Security Team)*
5616 * New functions to check a hostname email or IP address against a
5617 certificate. Add options x509 utility to print results of checks against
5622 * Fix OCSP checking.
5624 *Rob Stradling <rob.stradling@comodo.com> and Ben Laurie*
5626 * Initial experimental support for explicitly trusted non-root CAs.
5627 OpenSSL still tries to build a complete chain to a root but if an
5628 intermediate CA has a trust setting included that is used. The first
5629 setting is used: whether to trust (e.g., -addtrust option to the x509
5634 * Add -trusted_first option which attempts to find certificates in the
5635 trusted store even if an untrusted chain is also supplied.
5639 * MIPS assembly pack updates: support for MIPS32r2 and SmartMIPS ASE,
5640 platform support for Linux and Android.
5644 * Support for linux-x32, ILP32 environment in x86_64 framework.
5648 * Experimental multi-implementation support for FIPS capable OpenSSL.
5649 When in FIPS mode the approved implementations are used as normal,
5650 when not in FIPS mode the internal unapproved versions are used instead.
5651 This means that the FIPS capable OpenSSL isn't forced to use the
5652 (often lower performance) FIPS implementations outside FIPS mode.
5656 * Transparently support X9.42 DH parameters when calling
5657 PEM_read_bio_DHparameters. This means existing applications can handle
5658 the new parameter format automatically.
5662 * Initial experimental support for X9.42 DH parameter format: mainly
5663 to support use of 'q' parameter for RFC5114 parameters.
5667 * Add DH parameters from RFC5114 including test data to dhtest.
5671 * Support for automatic EC temporary key parameter selection. If enabled
5672 the most preferred EC parameters are automatically used instead of
5673 hardcoded fixed parameters. Now a server just has to call:
5674 SSL_CTX_set_ecdh_auto(ctx, 1) and the server will automatically
5675 support ECDH and use the most appropriate parameters.
5679 * Enhance and tidy EC curve and point format TLS extension code. Use
5680 static structures instead of allocation if default values are used.
5681 New ctrls to set curves we wish to support and to retrieve shared curves.
5682 Print out shared curves in s_server. New options to s_server and s_client
5683 to set list of supported curves.
5687 * New ctrls to retrieve supported signature algorithms and
5688 supported curve values as an array of NIDs. Extend openssl utility
5689 to print out received values.
5693 * Add new APIs EC_curve_nist2nid and EC_curve_nid2nist which convert
5694 between NIDs and the more common NIST names such as "P-256". Enhance
5695 ecparam utility and ECC method to recognise the NIST names for curves.
5699 * Enhance SSL/TLS certificate chain handling to support different
5700 chains for each certificate instead of one chain in the parent SSL_CTX.
5704 * Support for fixed DH ciphersuite client authentication: where both
5705 server and client use DH certificates with common parameters.
5709 * Support for fixed DH ciphersuites: those requiring DH server
5714 * New function i2d_re_X509_tbs for re-encoding the TBS portion of
5716 Note: Related 1.0.2-beta specific macros X509_get_cert_info,
5717 X509_CINF_set_modified, X509_CINF_get_issuer, X509_CINF_get_extensions and
5718 X509_CINF_get_signature were reverted post internal team review.
5723 ### Changes between 1.0.1t and 1.0.1u [22 Sep 2016]
5725 * OCSP Status Request extension unbounded memory growth
5727 A malicious client can send an excessively large OCSP Status Request
5728 extension. If that client continually requests renegotiation, sending a
5729 large OCSP Status Request extension each time, then there will be unbounded
5730 memory growth on the server. This will eventually lead to a Denial Of
5731 Service attack through memory exhaustion. Servers with a default
5732 configuration are vulnerable even if they do not support OCSP. Builds using
5733 the "no-ocsp" build time option are not affected.
5735 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5740 * In order to mitigate the SWEET32 attack, the DES ciphers were moved from
5743 This issue was reported to OpenSSL Karthikeyan Bhargavan and Gaetan
5749 * OOB write in MDC2_Update()
5751 An overflow can occur in MDC2_Update() either if called directly or
5752 through the EVP_DigestUpdate() function using MDC2. If an attacker
5753 is able to supply very large amounts of input data after a previous
5754 call to EVP_EncryptUpdate() with a partial block then a length check
5755 can overflow resulting in a heap corruption.
5757 The amount of data needed is comparable to SIZE_MAX which is impractical
5760 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5765 * Malformed SHA512 ticket DoS
5767 If a server uses SHA512 for TLS session ticket HMAC it is vulnerable to a
5768 DoS attack where a malformed ticket will result in an OOB read which will
5771 The use of SHA512 in TLS session tickets is comparatively rare as it requires
5772 a custom server callback and ticket lookup mechanism.
5774 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5779 * OOB write in BN_bn2dec()
5781 The function BN_bn2dec() does not check the return value of BN_div_word().
5782 This can cause an OOB write if an application uses this function with an
5783 overly large BIGNUM. This could be a problem if an overly large certificate
5784 or CRL is printed out from an untrusted source. TLS is not affected because
5785 record limits will reject an oversized certificate before it is parsed.
5787 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5792 * OOB read in TS_OBJ_print_bio()
5794 The function TS_OBJ_print_bio() misuses OBJ_obj2txt(): the return value is
5795 the total length the OID text representation would use and not the amount
5796 of data written. This will result in OOB reads when large OIDs are
5799 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5804 * Pointer arithmetic undefined behaviour
5806 Avoid some undefined pointer arithmetic
5808 A common idiom in the codebase is to check limits in the following manner:
5811 Where "p" points to some malloc'd data of SIZE bytes and
5814 "len" here could be from some externally supplied data (e.g. from a TLS
5817 The rules of C pointer arithmetic are such that "p + len" is only well
5818 defined where len <= SIZE. Therefore the above idiom is actually
5819 undefined behaviour.
5821 For example this could cause problems if some malloc implementation
5822 provides an address for "p" such that "p + len" actually overflows for
5823 values of len that are too big and therefore p + len < limit.
5825 This issue was reported to OpenSSL by Guido Vranken
5830 * Constant time flag not preserved in DSA signing
5832 Operations in the DSA signing algorithm should run in constant time in
5833 order to avoid side channel attacks. A flaw in the OpenSSL DSA
5834 implementation means that a non-constant time codepath is followed for
5835 certain operations. This has been demonstrated through a cache-timing
5836 attack to be sufficient for an attacker to recover the private DSA key.
5838 This issue was reported by César Pereida (Aalto University), Billy Brumley
5839 (Tampere University of Technology), and Yuval Yarom (The University of
5840 Adelaide and NICTA).
5845 * DTLS buffered message DoS
5847 In a DTLS connection where handshake messages are delivered out-of-order
5848 those messages that OpenSSL is not yet ready to process will be buffered
5849 for later use. Under certain circumstances, a flaw in the logic means that
5850 those messages do not get removed from the buffer even though the handshake
5851 has been completed. An attacker could force up to approx. 15 messages to
5852 remain in the buffer when they are no longer required. These messages will
5853 be cleared when the DTLS connection is closed. The default maximum size for
5854 a message is 100k. Therefore the attacker could force an additional 1500k
5855 to be consumed per connection. By opening many simulataneous connections an
5856 attacker could cause a DoS attack through memory exhaustion.
5858 This issue was reported to OpenSSL by Quan Luo.
5863 * DTLS replay protection DoS
5865 A flaw in the DTLS replay attack protection mechanism means that records
5866 that arrive for future epochs update the replay protection "window" before
5867 the MAC for the record has been validated. This could be exploited by an
5868 attacker by sending a record for the next epoch (which does not have to
5869 decrypt or have a valid MAC), with a very large sequence number. This means
5870 that all subsequent legitimate packets are dropped causing a denial of
5871 service for a specific DTLS connection.
5873 This issue was reported to OpenSSL by the OCAP audit team.
5878 * Certificate message OOB reads
5880 In OpenSSL 1.0.2 and earlier some missing message length checks can result
5881 in OOB reads of up to 2 bytes beyond an allocated buffer. There is a
5882 theoretical DoS risk but this has not been observed in practice on common
5885 The messages affected are client certificate, client certificate request
5886 and server certificate. As a result the attack can only be performed
5887 against a client or a server which enables client authentication.
5889 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5894 ### Changes between 1.0.1s and 1.0.1t [3 May 2016]
5896 * Prevent padding oracle in AES-NI CBC MAC check
5898 A MITM attacker can use a padding oracle attack to decrypt traffic
5899 when the connection uses an AES CBC cipher and the server support
5902 This issue was introduced as part of the fix for Lucky 13 padding
5903 attack ([CVE-2013-0169]). The padding check was rewritten to be in
5904 constant time by making sure that always the same bytes are read and
5905 compared against either the MAC or padding bytes. But it no longer
5906 checked that there was enough data to have both the MAC and padding
5909 This issue was reported by Juraj Somorovsky using TLS-Attacker.
5914 * Fix EVP_EncodeUpdate overflow
5916 An overflow can occur in the EVP_EncodeUpdate() function which is used for
5917 Base64 encoding of binary data. If an attacker is able to supply very large
5918 amounts of input data then a length check can overflow resulting in a heap
5921 Internally to OpenSSL the EVP_EncodeUpdate() function is primarly used by
5922 the `PEM_write_bio*` family of functions. These are mainly used within the
5923 OpenSSL command line applications, so any application which processes data
5924 from an untrusted source and outputs it as a PEM file should be considered
5925 vulnerable to this issue. User applications that call these APIs directly
5926 with large amounts of untrusted data may also be vulnerable.
5928 This issue was reported by Guido Vranken.
5933 * Fix EVP_EncryptUpdate overflow
5935 An overflow can occur in the EVP_EncryptUpdate() function. If an attacker
5936 is able to supply very large amounts of input data after a previous call to
5937 EVP_EncryptUpdate() with a partial block then a length check can overflow
5938 resulting in a heap corruption. Following an analysis of all OpenSSL
5939 internal usage of the EVP_EncryptUpdate() function all usage is one of two
5940 forms. The first form is where the EVP_EncryptUpdate() call is known to be
5941 the first called function after an EVP_EncryptInit(), and therefore that
5942 specific call must be safe. The second form is where the length passed to
5943 EVP_EncryptUpdate() can be seen from the code to be some small value and
5944 therefore there is no possibility of an overflow. Since all instances are
5945 one of these two forms, it is believed that there can be no overflows in
5946 internal code due to this problem. It should be noted that
5947 EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths.
5948 Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances
5949 of these calls have also been analysed too and it is believed there are no
5950 instances in internal usage where an overflow could occur.
5952 This issue was reported by Guido Vranken.
5957 * Prevent ASN.1 BIO excessive memory allocation
5959 When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio()
5960 a short invalid encoding can casuse allocation of large amounts of memory
5961 potentially consuming excessive resources or exhausting memory.
5963 Any application parsing untrusted data through d2i BIO functions is
5964 affected. The memory based functions such as d2i_X509() are *not* affected.
5965 Since the memory based functions are used by the TLS library, TLS
5966 applications are not affected.
5968 This issue was reported by Brian Carpenter.
5975 ASN1 Strings that are over 1024 bytes can cause an overread in applications
5976 using the X509_NAME_oneline() function on EBCDIC systems. This could result
5977 in arbitrary stack data being returned in the buffer.
5979 This issue was reported by Guido Vranken.
5984 * Modify behavior of ALPN to invoke callback after SNI/servername
5985 callback, such that updates to the SSL_CTX affect ALPN.
5989 * Remove LOW from the DEFAULT cipher list. This removes singles DES from the
5994 * Only remove the SSLv2 methods with the no-ssl2-method option. When the
5995 methods are enabled and ssl2 is disabled the methods return NULL.
5999 ### Changes between 1.0.1r and 1.0.1s [1 Mar 2016]
6001 * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL.
6002 Builds that are not configured with "enable-weak-ssl-ciphers" will not
6003 provide any "EXPORT" or "LOW" strength ciphers.
6007 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2
6008 is by default disabled at build-time. Builds that are not configured with
6009 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used,
6010 users who want to negotiate SSLv2 via the version-flexible SSLv23_method()
6011 will need to explicitly call either of:
6013 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2);
6015 SSL_clear_options(ssl, SSL_OP_NO_SSLv2);
6017 as appropriate. Even if either of those is used, or the application
6018 explicitly uses the version-specific SSLv2_method() or its client and
6019 server variants, SSLv2 ciphers vulnerable to exhaustive search key
6020 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT
6021 ciphers, and SSLv2 56-bit DES are no longer available.
6026 * Fix a double-free in DSA code
6028 A double free bug was discovered when OpenSSL parses malformed DSA private
6029 keys and could lead to a DoS attack or memory corruption for applications
6030 that receive DSA private keys from untrusted sources. This scenario is
6033 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using
6039 * Disable SRP fake user seed to address a server memory leak.
6041 Add a new method SRP_VBASE_get1_by_user that handles the seed properly.
6043 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
6044 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user
6045 was changed to ignore the "fake user" SRP seed, even if the seed
6048 Users should use SRP_VBASE_get1_by_user instead. Note that in
6049 SRP_VBASE_get1_by_user, caller must free the returned value. Note
6050 also that even though configuring the SRP seed attempts to hide
6051 invalid usernames by continuing the handshake with fake
6052 credentials, this behaviour is not constant time and no strong
6053 guarantees are made that the handshake is indistinguishable from
6054 that of a valid user.
6059 * Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption
6061 In the BN_hex2bn function the number of hex digits is calculated using an
6062 int value `i`. Later `bn_expand` is called with a value of `i * 4`. For
6063 large values of `i` this can result in `bn_expand` not allocating any
6064 memory because `i * 4` is negative. This can leave the internal BIGNUM data
6065 field as NULL leading to a subsequent NULL ptr deref. For very large values
6066 of `i`, the calculation `i * 4` could be a positive value smaller than `i`.
6067 In this case memory is allocated to the internal BIGNUM data field, but it
6068 is insufficiently sized leading to heap corruption. A similar issue exists
6069 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn
6070 is ever called by user applications with very large untrusted hex/dec data.
6071 This is anticipated to be a rare occurrence.
6073 All OpenSSL internal usage of these functions use data that is not expected
6074 to be untrusted, e.g. config file data or application command line
6075 arguments. If user developed applications generate config file data based
6076 on untrusted data then it is possible that this could also lead to security
6077 consequences. This is also anticipated to be rare.
6079 This issue was reported to OpenSSL by Guido Vranken.
6084 * Fix memory issues in `BIO_*printf` functions
6086 The internal `fmtstr` function used in processing a "%s" format string in
6087 the `BIO_*printf` functions could overflow while calculating the length of a
6088 string and cause an OOB read when printing very long strings.
6090 Additionally the internal `doapr_outch` function can attempt to write to an
6091 OOB memory location (at an offset from the NULL pointer) in the event of a
6092 memory allocation failure. In 1.0.2 and below this could be caused where
6093 the size of a buffer to be allocated is greater than INT_MAX. E.g. this
6094 could be in processing a very long "%s" format string. Memory leaks can
6097 The first issue may mask the second issue dependent on compiler behaviour.
6098 These problems could enable attacks where large amounts of untrusted data
6099 is passed to the `BIO_*printf` functions. If applications use these functions
6100 in this way then they could be vulnerable. OpenSSL itself uses these
6101 functions when printing out human-readable dumps of ASN.1 data. Therefore
6102 applications that print this data could be vulnerable if the data is from
6103 untrusted sources. OpenSSL command line applications could also be
6104 vulnerable where they print out ASN.1 data, or if untrusted data is passed
6105 as command line arguments.
6107 Libssl is not considered directly vulnerable. Additionally certificates etc
6108 received via remote connections via libssl are also unlikely to be able to
6109 trigger these issues because of message size limits enforced within libssl.
6111 This issue was reported to OpenSSL Guido Vranken.
6116 * Side channel attack on modular exponentiation
6118 A side-channel attack was found which makes use of cache-bank conflicts on
6119 the Intel Sandy-Bridge microarchitecture which could lead to the recovery
6120 of RSA keys. The ability to exploit this issue is limited as it relies on
6121 an attacker who has control of code in a thread running on the same
6122 hyper-threaded core as the victim thread which is performing decryptions.
6124 This issue was reported to OpenSSL by Yuval Yarom, The University of
6125 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and
6126 Nadia Heninger, University of Pennsylvania with more information at
6127 <http://cachebleed.info>.
6132 * Change the req command to generate a 2048-bit RSA/DSA key by default,
6133 if no keysize is specified with default_bits. This fixes an
6134 omission in an earlier change that changed all RSA/DSA key generation
6135 commands to use 2048 bits by default.
6139 ### Changes between 1.0.1q and 1.0.1r [28 Jan 2016]
6141 * Protection for DH small subgroup attacks
6143 As a precautionary measure the SSL_OP_SINGLE_DH_USE option has been
6144 switched on by default and cannot be disabled. This could have some
6149 * SSLv2 doesn't block disabled ciphers
6151 A malicious client can negotiate SSLv2 ciphers that have been disabled on
6152 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
6153 been disabled, provided that the SSLv2 protocol was not also disabled via
6156 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
6157 and Sebastian Schinzel.
6162 * Reject DH handshakes with parameters shorter than 1024 bits.
6166 ### Changes between 1.0.1p and 1.0.1q [3 Dec 2015]
6168 * Certificate verify crash with missing PSS parameter
6170 The signature verification routines will crash with a NULL pointer
6171 dereference if presented with an ASN.1 signature using the RSA PSS
6172 algorithm and absent mask generation function parameter. Since these
6173 routines are used to verify certificate signature algorithms this can be
6174 used to crash any certificate verification operation and exploited in a
6175 DoS attack. Any application which performs certificate verification is
6176 vulnerable including OpenSSL clients and servers which enable client
6179 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
6184 * X509_ATTRIBUTE memory leak
6186 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
6187 memory. This structure is used by the PKCS#7 and CMS routines so any
6188 application which reads PKCS#7 or CMS data from untrusted sources is
6189 affected. SSL/TLS is not affected.
6191 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
6197 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
6198 This changes the decoding behaviour for some invalid messages,
6199 though the change is mostly in the more lenient direction, and
6200 legacy behaviour is preserved as much as possible.
6204 * In DSA_generate_parameters_ex, if the provided seed is too short,
6205 use a random seed, as already documented.
6207 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
6209 ### Changes between 1.0.1o and 1.0.1p [9 Jul 2015]
6211 * Alternate chains certificate forgery
6213 During certificate verfification, OpenSSL will attempt to find an
6214 alternative certificate chain if the first attempt to build such a chain
6215 fails. An error in the implementation of this logic can mean that an
6216 attacker could cause certain checks on untrusted certificates to be
6217 bypassed, such as the CA flag, enabling them to use a valid leaf
6218 certificate to act as a CA and "issue" an invalid certificate.
6220 This issue was reported to OpenSSL by Adam Langley/David Benjamin
6226 * Race condition handling PSK identify hint
6228 If PSK identity hints are received by a multi-threaded client then
6229 the values are wrongly updated in the parent SSL_CTX structure. This can
6230 result in a race condition potentially leading to a double free of the
6236 ### Changes between 1.0.1n and 1.0.1o [12 Jun 2015]
6238 * Fix HMAC ABI incompatibility. The previous version introduced an ABI
6239 incompatibility in the handling of HMAC. The previous ABI has now been
6242 ### Changes between 1.0.1m and 1.0.1n [11 Jun 2015]
6244 * Malformed ECParameters causes infinite loop
6246 When processing an ECParameters structure OpenSSL enters an infinite loop
6247 if the curve specified is over a specially malformed binary polynomial
6250 This can be used to perform denial of service against any
6251 system which processes public keys, certificate requests or
6252 certificates. This includes TLS clients and TLS servers with
6253 client authentication enabled.
6255 This issue was reported to OpenSSL by Joseph Barr-Pixton.
6260 * Exploitable out-of-bounds read in X509_cmp_time
6262 X509_cmp_time does not properly check the length of the ASN1_TIME
6263 string and can read a few bytes out of bounds. In addition,
6264 X509_cmp_time accepts an arbitrary number of fractional seconds in the
6267 An attacker can use this to craft malformed certificates and CRLs of
6268 various sizes and potentially cause a segmentation fault, resulting in
6269 a DoS on applications that verify certificates or CRLs. TLS clients
6270 that verify CRLs are affected. TLS clients and servers with client
6271 authentication enabled may be affected if they use custom verification
6274 This issue was reported to OpenSSL by Robert Swiecki (Google), and
6275 independently by Hanno Böck.
6280 * PKCS7 crash with missing EnvelopedContent
6282 The PKCS#7 parsing code does not handle missing inner EncryptedContent
6283 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
6284 with missing content and trigger a NULL pointer dereference on parsing.
6286 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
6287 structures from untrusted sources are affected. OpenSSL clients and
6288 servers are not affected.
6290 This issue was reported to OpenSSL by Michal Zalewski (Google).
6295 * CMS verify infinite loop with unknown hash function
6297 When verifying a signedData message the CMS code can enter an infinite loop
6298 if presented with an unknown hash function OID. This can be used to perform
6299 denial of service against any system which verifies signedData messages using
6301 This issue was reported to OpenSSL by Johannes Bauer.
6306 * Race condition handling NewSessionTicket
6308 If a NewSessionTicket is received by a multi-threaded client when attempting to
6309 reuse a previous ticket then a race condition can occur potentially leading to
6310 a double free of the ticket data.
6315 * Reject DH handshakes with parameters shorter than 768 bits.
6317 *Kurt Roeckx and Emilia Kasper*
6319 * dhparam: generate 2048-bit parameters by default.
6321 *Kurt Roeckx and Emilia Kasper*
6323 ### Changes between 1.0.1l and 1.0.1m [19 Mar 2015]
6325 * Segmentation fault in ASN1_TYPE_cmp fix
6327 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
6328 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
6329 certificate signature algorithm consistency this can be used to crash any
6330 certificate verification operation and exploited in a DoS attack. Any
6331 application which performs certificate verification is vulnerable including
6332 OpenSSL clients and servers which enable client authentication.
6337 * ASN.1 structure reuse memory corruption fix
6339 Reusing a structure in ASN.1 parsing may allow an attacker to cause
6340 memory corruption via an invalid write. Such reuse is and has been
6341 strongly discouraged and is believed to be rare.
6343 Applications that parse structures containing CHOICE or ANY DEFINED BY
6344 components may be affected. Certificate parsing (d2i_X509 and related
6345 functions) are however not affected. OpenSSL clients and servers are
6351 * PKCS7 NULL pointer dereferences fix
6353 The PKCS#7 parsing code does not handle missing outer ContentInfo
6354 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
6355 missing content and trigger a NULL pointer dereference on parsing.
6357 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
6358 otherwise parse PKCS#7 structures from untrusted sources are
6359 affected. OpenSSL clients and servers are not affected.
6361 This issue was reported to OpenSSL by Michal Zalewski (Google).
6366 * DoS via reachable assert in SSLv2 servers fix
6368 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
6369 servers that both support SSLv2 and enable export cipher suites by sending
6370 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
6372 This issue was discovered by Sean Burford (Google) and Emilia Käsper
6373 (OpenSSL development team).
6378 * Use After Free following d2i_ECPrivatekey error fix
6380 A malformed EC private key file consumed via the d2i_ECPrivateKey function
6381 could cause a use after free condition. This, in turn, could cause a double
6382 free in several private key parsing functions (such as d2i_PrivateKey
6383 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
6384 for applications that receive EC private keys from untrusted
6385 sources. This scenario is considered rare.
6387 This issue was discovered by the BoringSSL project and fixed in their
6393 * X509_to_X509_REQ NULL pointer deref fix
6395 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
6396 the certificate key is invalid. This function is rarely used in practice.
6398 This issue was discovered by Brian Carpenter.
6403 * Removed the export ciphers from the DEFAULT ciphers
6407 ### Changes between 1.0.1k and 1.0.1l [15 Jan 2015]
6409 * Build fixes for the Windows and OpenVMS platforms
6411 *Matt Caswell and Richard Levitte*
6413 ### Changes between 1.0.1j and 1.0.1k [8 Jan 2015]
6415 * Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
6416 message can cause a segmentation fault in OpenSSL due to a NULL pointer
6417 dereference. This could lead to a Denial Of Service attack. Thanks to
6418 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
6423 * Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
6424 dtls1_buffer_record function under certain conditions. In particular this
6425 could occur if an attacker sent repeated DTLS records with the same
6426 sequence number but for the next epoch. The memory leak could be exploited
6427 by an attacker in a Denial of Service attack through memory exhaustion.
6428 Thanks to Chris Mueller for reporting this issue.
6433 * Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
6434 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
6435 method would be set to NULL which could later result in a NULL pointer
6436 dereference. Thanks to Frank Schmirler for reporting this issue.
6441 * Abort handshake if server key exchange message is omitted for ephemeral
6444 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
6445 reporting this issue.
6450 * Remove non-export ephemeral RSA code on client and server. This code
6451 violated the TLS standard by allowing the use of temporary RSA keys in
6452 non-export ciphersuites and could be used by a server to effectively
6453 downgrade the RSA key length used to a value smaller than the server
6454 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
6455 INRIA or reporting this issue.
6460 * Fixed issue where DH client certificates are accepted without verification.
6461 An OpenSSL server will accept a DH certificate for client authentication
6462 without the certificate verify message. This effectively allows a client to
6463 authenticate without the use of a private key. This only affects servers
6464 which trust a client certificate authority which issues certificates
6465 containing DH keys: these are extremely rare and hardly ever encountered.
6466 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
6472 * Ensure that the session ID context of an SSL is updated when its
6473 SSL_CTX is updated via SSL_set_SSL_CTX.
6475 The session ID context is typically set from the parent SSL_CTX,
6476 and can vary with the CTX.
6480 * Fix various certificate fingerprint issues.
6482 By using non-DER or invalid encodings outside the signed portion of a
6483 certificate the fingerprint can be changed without breaking the signature.
6484 Although no details of the signed portion of the certificate can be changed
6485 this can cause problems with some applications: e.g. those using the
6486 certificate fingerprint for blacklists.
6488 1. Reject signatures with non zero unused bits.
6490 If the BIT STRING containing the signature has non zero unused bits reject
6491 the signature. All current signature algorithms require zero unused bits.
6493 2. Check certificate algorithm consistency.
6495 Check the AlgorithmIdentifier inside TBS matches the one in the
6496 certificate signature. NB: this will result in signature failure
6497 errors for some broken certificates.
6499 Thanks to Konrad Kraszewski from Google for reporting this issue.
6501 3. Check DSA/ECDSA signatures use DER.
6503 Re-encode DSA/ECDSA signatures and compare with the original received
6504 signature. Return an error if there is a mismatch.
6506 This will reject various cases including garbage after signature
6507 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
6508 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
6509 (negative or with leading zeroes).
6511 Further analysis was conducted and fixes were developed by Stephen Henson
6512 of the OpenSSL core team.
6518 * Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
6519 results on some platforms, including x86_64. This bug occurs at random
6520 with a very low probability, and is not known to be exploitable in any
6521 way, though its exact impact is difficult to determine. Thanks to Pieter
6522 Wuille (Blockstream) who reported this issue and also suggested an initial
6523 fix. Further analysis was conducted by the OpenSSL development team and
6524 Adam Langley of Google. The final fix was developed by Andy Polyakov of
6525 the OpenSSL core team.
6530 * Do not resume sessions on the server if the negotiated protocol
6531 version does not match the session's version. Resuming with a different
6532 version, while not strictly forbidden by the RFC, is of questionable
6533 sanity and breaks all known clients.
6535 *David Benjamin, Emilia Käsper*
6537 * Tighten handling of the ChangeCipherSpec (CCS) message: reject
6538 early CCS messages during renegotiation. (Note that because
6539 renegotiation is encrypted, this early CCS was not exploitable.)
6543 * Tighten client-side session ticket handling during renegotiation:
6544 ensure that the client only accepts a session ticket if the server sends
6545 the extension anew in the ServerHello. Previously, a TLS client would
6546 reuse the old extension state and thus accept a session ticket if one was
6547 announced in the initial ServerHello.
6549 Similarly, ensure that the client requires a session ticket if one
6550 was advertised in the ServerHello. Previously, a TLS client would
6551 ignore a missing NewSessionTicket message.
6555 ### Changes between 1.0.1i and 1.0.1j [15 Oct 2014]
6559 A flaw in the DTLS SRTP extension parsing code allows an attacker, who
6560 sends a carefully crafted handshake message, to cause OpenSSL to fail
6561 to free up to 64k of memory causing a memory leak. This could be
6562 exploited in a Denial Of Service attack. This issue affects OpenSSL
6563 1.0.1 server implementations for both SSL/TLS and DTLS regardless of
6564 whether SRTP is used or configured. Implementations of OpenSSL that
6565 have been compiled with OPENSSL_NO_SRTP defined are not affected.
6567 The fix was developed by the OpenSSL team.
6572 * Session Ticket Memory Leak.
6574 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
6575 integrity of that ticket is first verified. In the event of a session
6576 ticket integrity check failing, OpenSSL will fail to free memory
6577 causing a memory leak. By sending a large number of invalid session
6578 tickets an attacker could exploit this issue in a Denial Of Service
6584 * Build option no-ssl3 is incomplete.
6586 When OpenSSL is configured with "no-ssl3" as a build option, servers
6587 could accept and complete a SSL 3.0 handshake, and clients could be
6588 configured to send them.
6591 *Akamai and the OpenSSL team*
6593 * Add support for TLS_FALLBACK_SCSV.
6594 Client applications doing fallback retries should call
6595 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
6598 *Adam Langley, Bodo Moeller*
6600 * Add additional DigestInfo checks.
6602 Re-encode DigestInto in DER and check against the original when
6603 verifying RSA signature: this will reject any improperly encoded
6604 DigestInfo structures.
6606 Note: this is a precautionary measure and no attacks are currently known.
6610 ### Changes between 1.0.1h and 1.0.1i [6 Aug 2014]
6612 * Fix SRP buffer overrun vulnerability. Invalid parameters passed to the
6613 SRP code can be overrun an internal buffer. Add sanity check that
6614 g, A, B < N to SRP code.
6616 Thanks to Sean Devlin and Watson Ladd of Cryptography Services, NCC
6617 Group for discovering this issue.
6622 * A flaw in the OpenSSL SSL/TLS server code causes the server to negotiate
6623 TLS 1.0 instead of higher protocol versions when the ClientHello message
6624 is badly fragmented. This allows a man-in-the-middle attacker to force a
6625 downgrade to TLS 1.0 even if both the server and the client support a
6626 higher protocol version, by modifying the client's TLS records.
6628 Thanks to David Benjamin and Adam Langley (Google) for discovering and
6629 researching this issue.
6634 * OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
6635 to a denial of service attack. A malicious server can crash the client
6636 with a null pointer dereference (read) by specifying an anonymous (EC)DH
6637 ciphersuite and sending carefully crafted handshake messages.
6639 Thanks to Felix Gröbert (Google) for discovering and researching this
6645 * By sending carefully crafted DTLS packets an attacker could cause openssl
6646 to leak memory. This can be exploited through a Denial of Service attack.
6647 Thanks to Adam Langley for discovering and researching this issue.
6652 * An attacker can force openssl to consume large amounts of memory whilst
6653 processing DTLS handshake messages. This can be exploited through a
6654 Denial of Service attack.
6655 Thanks to Adam Langley for discovering and researching this issue.
6660 * An attacker can force an error condition which causes openssl to crash
6661 whilst processing DTLS packets due to memory being freed twice. This
6662 can be exploited through a Denial of Service attack.
6663 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
6669 * If a multithreaded client connects to a malicious server using a resumed
6670 session and the server sends an ec point format extension it could write
6671 up to 255 bytes to freed memory.
6673 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
6679 * A malicious server can crash an OpenSSL client with a null pointer
6680 dereference (read) by specifying an SRP ciphersuite even though it was not
6681 properly negotiated with the client. This can be exploited through a
6682 Denial of Service attack.
6684 Thanks to Joonas Kuorilehto and Riku Hietamäki (Codenomicon) for
6685 discovering and researching this issue.
6690 * A flaw in OBJ_obj2txt may cause pretty printing functions such as
6691 X509_name_oneline, X509_name_print_ex et al. to leak some information
6692 from the stack. Applications may be affected if they echo pretty printing
6693 output to the attacker.
6695 Thanks to Ivan Fratric (Google) for discovering this issue.
6698 *Emilia Käsper, and Steve Henson*
6700 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
6701 for corner cases. (Certain input points at infinity could lead to
6702 bogus results, with non-infinity inputs mapped to infinity too.)
6706 ### Changes between 1.0.1g and 1.0.1h [5 Jun 2014]
6708 * Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
6709 handshake can force the use of weak keying material in OpenSSL
6710 SSL/TLS clients and servers.
6712 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
6713 researching this issue. ([CVE-2014-0224])
6715 *KIKUCHI Masashi, Steve Henson*
6717 * Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
6718 OpenSSL DTLS client the code can be made to recurse eventually crashing
6721 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
6724 *Imre Rad, Steve Henson*
6726 * Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
6727 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
6728 client or server. This is potentially exploitable to run arbitrary
6729 code on a vulnerable client or server.
6731 Thanks to Jüri Aedla for reporting this issue. ([CVE-2014-0195])
6733 *Jüri Aedla, Steve Henson*
6735 * Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
6736 are subject to a denial of service attack.
6738 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
6739 this issue. ([CVE-2014-3470])
6741 *Felix Gröbert, Ivan Fratric, Steve Henson*
6743 * Harmonize version and its documentation. -f flag is used to display
6746 *mancha <mancha1@zoho.com>*
6748 * Fix eckey_priv_encode so it immediately returns an error upon a failure
6749 in i2d_ECPrivateKey.
6751 *mancha <mancha1@zoho.com>*
6753 * Fix some double frees. These are not thought to be exploitable.
6755 *mancha <mancha1@zoho.com>*
6757 ### Changes between 1.0.1f and 1.0.1g [7 Apr 2014]
6759 * A missing bounds check in the handling of the TLS heartbeat extension
6760 can be used to reveal up to 64k of memory to a connected client or
6763 Thanks for Neel Mehta of Google Security for discovering this bug and to
6764 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
6765 preparing the fix ([CVE-2014-0160])
6767 *Adam Langley, Bodo Moeller*
6769 * Fix for the attack described in the paper "Recovering OpenSSL
6770 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
6771 by Yuval Yarom and Naomi Benger. Details can be obtained from:
6772 <http://eprint.iacr.org/2014/140>
6774 Thanks to Yuval Yarom and Naomi Benger for discovering this
6775 flaw and to Yuval Yarom for supplying a fix ([CVE-2014-0076])
6777 *Yuval Yarom and Naomi Benger*
6779 * TLS pad extension: draft-agl-tls-padding-03
6781 Workaround for the "TLS hang bug" (see FAQ and PR#2771): if the
6782 TLS client Hello record length value would otherwise be > 255 and
6783 less that 512 pad with a dummy extension containing zeroes so it
6784 is at least 512 bytes long.
6786 *Adam Langley, Steve Henson*
6788 ### Changes between 1.0.1e and 1.0.1f [6 Jan 2014]
6790 * Fix for TLS record tampering bug. A carefully crafted invalid
6791 handshake could crash OpenSSL with a NULL pointer exception.
6792 Thanks to Anton Johansson for reporting this issues.
6795 * Keep original DTLS digest and encryption contexts in retransmission
6796 structures so we can use the previous session parameters if they need
6797 to be resent. ([CVE-2013-6450])
6801 * Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
6802 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
6803 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
6804 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
6805 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
6806 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
6808 *Rob Stradling, Adam Langley*
6810 ### Changes between 1.0.1d and 1.0.1e [11 Feb 2013]
6812 * Correct fix for CVE-2013-0169. The original didn't work on AES-NI
6813 supporting platforms or when small records were transferred.
6815 *Andy Polyakov, Steve Henson*
6817 ### Changes between 1.0.1c and 1.0.1d [5 Feb 2013]
6819 * Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
6821 This addresses the flaw in CBC record processing discovered by
6822 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
6823 at: <http://www.isg.rhul.ac.uk/tls/>
6825 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
6826 Security Group at Royal Holloway, University of London
6827 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
6828 Emilia Käsper for the initial patch.
6831 *Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson*
6833 * Fix flaw in AESNI handling of TLS 1.2 and 1.1 records for CBC mode
6834 ciphersuites which can be exploited in a denial of service attack.
6835 Thanks go to and to Adam Langley <agl@chromium.org> for discovering
6836 and detecting this bug and to Wolfgang Ettlinger
6837 <wolfgang.ettlinger@gmail.com> for independently discovering this issue.
6842 * Return an error when checking OCSP signatures when key is NULL.
6843 This fixes a DoS attack. ([CVE-2013-0166])
6847 * Make openssl verify return errors.
6849 *Chris Palmer <palmer@google.com> and Ben Laurie*
6851 * Call OCSP Stapling callback after ciphersuite has been chosen, so
6852 the right response is stapled. Also change SSL_get_certificate()
6853 so it returns the certificate actually sent.
6854 See <http://rt.openssl.org/Ticket/Display.html?id=2836>.
6856 *Rob Stradling <rob.stradling@comodo.com>*
6858 * Fix possible deadlock when decoding public keys.
6862 * Don't use TLS 1.0 record version number in initial client hello
6867 ### Changes between 1.0.1b and 1.0.1c [10 May 2012]
6869 * Sanity check record length before skipping explicit IV in TLS
6870 1.2, 1.1 and DTLS to fix DoS attack.
6872 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
6873 fuzzing as a service testing platform.
6878 * Initialise tkeylen properly when encrypting CMS messages.
6879 Thanks to Solar Designer of Openwall for reporting this issue.
6883 * In FIPS mode don't try to use composite ciphers as they are not
6888 ### Changes between 1.0.1a and 1.0.1b [26 Apr 2012]
6890 * OpenSSL 1.0.0 sets SSL_OP_ALL to 0x80000FFFL and OpenSSL 1.0.1 and
6891 1.0.1a set SSL_OP_NO_TLSv1_1 to 0x00000400L which would unfortunately
6892 mean any application compiled against OpenSSL 1.0.0 headers setting
6893 SSL_OP_ALL would also set SSL_OP_NO_TLSv1_1, unintentionally disabling
6894 TLS 1.1 also. Fix this by changing the value of SSL_OP_NO_TLSv1_1 to
6895 0x10000000L Any application which was previously compiled against
6896 OpenSSL 1.0.1 or 1.0.1a headers and which cares about SSL_OP_NO_TLSv1_1
6897 will need to be recompiled as a result. Letting be results in
6898 inability to disable specifically TLS 1.1 and in client context,
6899 in unlike event, limit maximum offered version to TLS 1.0 [see below].
6903 * In order to ensure interoperability SSL_OP_NO_protocolX does not
6904 disable just protocol X, but all protocols above X *if* there are
6905 protocols *below* X still enabled. In more practical terms it means
6906 that if application wants to disable TLS1.0 in favor of TLS1.1 and
6907 above, it's not sufficient to pass `SSL_OP_NO_TLSv1`, one has to pass
6908 `SSL_OP_NO_TLSv1|SSL_OP_NO_SSLv3|SSL_OP_NO_SSLv2`. This applies to
6913 ### Changes between 1.0.1 and 1.0.1a [19 Apr 2012]
6915 * Check for potentially exploitable overflows in asn1_d2i_read_bio
6916 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
6917 in CRYPTO_realloc_clean.
6919 Thanks to Tavis Ormandy, Google Security Team, for discovering this
6920 issue and to Adam Langley <agl@chromium.org> for fixing it.
6923 *Adam Langley (Google), Tavis Ormandy, Google Security Team*
6925 * Don't allow TLS 1.2 SHA-256 ciphersuites in TLS 1.0, 1.1 connections.
6929 * Workarounds for some broken servers that "hang" if a client hello
6930 record length exceeds 255 bytes.
6932 1. Do not use record version number > TLS 1.0 in initial client
6933 hello: some (but not all) hanging servers will now work.
6934 2. If we set OPENSSL_MAX_TLS1_2_CIPHER_LENGTH this will truncate
6935 the number of ciphers sent in the client hello. This should be
6936 set to an even number, such as 50, for example by passing:
6937 -DOPENSSL_MAX_TLS1_2_CIPHER_LENGTH=50 to config or Configure.
6938 Most broken servers should now work.
6939 3. If all else fails setting OPENSSL_NO_TLS1_2_CLIENT will disable
6940 TLS 1.2 client support entirely.
6944 * Fix SEGV in Vector Permutation AES module observed in OpenSSH.
6948 ### Changes between 1.0.0h and 1.0.1 [14 Mar 2012]
6950 * Add compatibility with old MDC2 signatures which use an ASN1 OCTET
6951 STRING form instead of a DigestInfo.
6955 * The format used for MDC2 RSA signatures is inconsistent between EVP
6956 and the RSA_sign/RSA_verify functions. This was made more apparent when
6957 OpenSSL used RSA_sign/RSA_verify for some RSA signatures in particular
6958 those which went through EVP_PKEY_METHOD in 1.0.0 and later. Detect
6959 the correct format in RSA_verify so both forms transparently work.
6963 * Some servers which support TLS 1.0 can choke if we initially indicate
6964 support for TLS 1.2 and later renegotiate using TLS 1.0 in the RSA
6965 encrypted premaster secret. As a workaround use the maximum permitted
6966 client version in client hello, this should keep such servers happy
6967 and still work with previous versions of OpenSSL.
6971 * Add support for TLS/DTLS heartbeats.
6973 *Robin Seggelmann <seggelmann@fh-muenster.de>*
6975 * Add support for SCTP.
6977 *Robin Seggelmann <seggelmann@fh-muenster.de>*
6979 * Improved PRNG seeding for VOS.
6981 *Paul Green <Paul.Green@stratus.com>*
6983 * Extensive assembler packs updates, most notably:
6985 - x86[_64]: AES-NI, PCLMULQDQ, RDRAND support;
6986 - x86[_64]: SSSE3 support (SHA1, vector-permutation AES);
6987 - x86_64: bit-sliced AES implementation;
6988 - ARM: NEON support, contemporary platforms optimizations;
6989 - s390x: z196 support;
6990 - `*`: GHASH and GF(2^m) multiplication implementations;
6994 * Make TLS-SRP code conformant with RFC 5054 API cleanup
6995 (removal of unnecessary code)
6997 *Peter Sylvester <peter.sylvester@edelweb.fr>*
6999 * Add TLS key material exporter from RFC 5705.
7003 * Add DTLS-SRTP negotiation from RFC 5764.
7007 * Add Next Protocol Negotiation,
7008 <http://tools.ietf.org/html/draft-agl-tls-nextprotoneg-00>. Can be
7009 disabled with a no-npn flag to config or Configure. Code donated
7012 *Adam Langley <agl@google.com> and Ben Laurie*
7014 * Add optional 64-bit optimized implementations of elliptic curves NIST-P224,
7015 NIST-P256, NIST-P521, with constant-time single point multiplication on
7016 typical inputs. Compiler support for the nonstandard type `__uint128_t` is
7017 required to use this (present in gcc 4.4 and later, for 64-bit builds).
7018 Code made available under Apache License version 2.0.
7020 Specify "enable-ec_nistp_64_gcc_128" on the Configure (or config) command
7021 line to include this in your build of OpenSSL, and run "make depend" (or
7022 "make update"). This enables the following EC_METHODs:
7024 EC_GFp_nistp224_method()
7025 EC_GFp_nistp256_method()
7026 EC_GFp_nistp521_method()
7028 EC_GROUP_new_by_curve_name() will automatically use these (while
7029 EC_GROUP_new_curve_GFp() currently prefers the more flexible
7032 *Emilia Käsper, Adam Langley, Bodo Moeller (Google)*
7034 * Use type ossl_ssize_t instead of ssize_t which isn't available on
7035 all platforms. Move ssize_t definition from e_os.h to the public
7036 header file e_os2.h as it now appears in public header file cms.h
7040 * New -sigopt option to the ca, req and x509 utilities. Additional
7041 signature parameters can be passed using this option and in
7046 * Add RSA PSS signing function. This will generate and set the
7047 appropriate AlgorithmIdentifiers for PSS based on those in the
7048 corresponding EVP_MD_CTX structure. No application support yet.
7052 * Support for companion algorithm specific ASN1 signing routines.
7053 New function ASN1_item_sign_ctx() signs a pre-initialised
7054 EVP_MD_CTX structure and sets AlgorithmIdentifiers based on
7055 the appropriate parameters.
7059 * Add new algorithm specific ASN1 verification initialisation function
7060 to EVP_PKEY_ASN1_METHOD: this is not in EVP_PKEY_METHOD since the ASN1
7061 handling will be the same no matter what EVP_PKEY_METHOD is used.
7062 Add a PSS handler to support verification of PSS signatures: checked
7063 against a number of sample certificates.
7067 * Add signature printing for PSS. Add PSS OIDs.
7069 *Steve Henson, Martin Kaiser <lists@kaiser.cx>*
7071 * Add algorithm specific signature printing. An individual ASN1 method
7072 can now print out signatures instead of the standard hex dump.
7074 More complex signatures (e.g. PSS) can print out more meaningful
7075 information. Include DSA version that prints out the signature
7080 * Password based recipient info support for CMS library: implementing
7085 * Split password based encryption into PBES2 and PBKDF2 functions. This
7086 neatly separates the code into cipher and PBE sections and is required
7087 for some algorithms that split PBES2 into separate pieces (such as
7088 password based CMS).
7092 * Session-handling fixes:
7093 - Fix handling of connections that are resuming with a session ID,
7094 but also support Session Tickets.
7095 - Fix a bug that suppressed issuing of a new ticket if the client
7096 presented a ticket with an expired session.
7097 - Try to set the ticket lifetime hint to something reasonable.
7098 - Make tickets shorter by excluding irrelevant information.
7099 - On the client side, don't ignore renewed tickets.
7101 *Adam Langley, Bodo Moeller (Google)*
7103 * Fix PSK session representation.
7107 * Add RC4-MD5 and AESNI-SHA1 "stitched" implementations.
7109 This work was sponsored by Intel.
7113 * Add GCM support to TLS library. Some custom code is needed to split
7114 the IV between the fixed (from PRF) and explicit (from TLS record)
7115 portions. This adds all GCM ciphersuites supported by RFC5288 and
7116 RFC5289. Generalise some `AES*` cipherstrings to include GCM and
7117 add a special AESGCM string for GCM only.
7121 * Expand range of ctrls for AES GCM. Permit setting invocation
7122 field on decrypt and retrieval of invocation field only on encrypt.
7126 * Add HMAC ECC ciphersuites from RFC5289. Include SHA384 PRF support.
7127 As required by RFC5289 these ciphersuites cannot be used if for
7128 versions of TLS earlier than 1.2.
7132 * For FIPS capable OpenSSL interpret a NULL default public key method
7133 as unset and return the appropriate default but do *not* set the default.
7134 This means we can return the appropriate method in applications that
7135 switch between FIPS and non-FIPS modes.
7139 * Redirect HMAC and CMAC operations to FIPS module in FIPS mode. If an
7140 ENGINE is used then we cannot handle that in the FIPS module so we
7141 keep original code iff non-FIPS operations are allowed.
7145 * Add -attime option to openssl utilities.
7147 *Peter Eckersley <pde@eff.org>, Ben Laurie and Steve Henson*
7149 * Redirect DSA and DH operations to FIPS module in FIPS mode.
7153 * Redirect ECDSA and ECDH operations to FIPS module in FIPS mode. Also use
7154 FIPS EC methods unconditionally for now.
7158 * New build option no-ec2m to disable characteristic 2 code.
7162 * Backport libcrypto audit of return value checking from 1.1.0-dev; not
7163 all cases can be covered as some introduce binary incompatibilities.
7167 * Redirect RSA operations to FIPS module including keygen,
7168 encrypt, decrypt, sign and verify. Block use of non FIPS RSA methods.
7172 * Add similar low-level API blocking to ciphers.
7176 * low-level digest APIs are not approved in FIPS mode: any attempt
7177 to use these will cause a fatal error. Applications that *really* want
7178 to use them can use the `private_*` version instead.
7182 * Redirect cipher operations to FIPS module for FIPS builds.
7186 * Redirect digest operations to FIPS module for FIPS builds.
7190 * Update build system to add "fips" flag which will link in fipscanister.o
7191 for static and shared library builds embedding a signature if needed.
7195 * Output TLS supported curves in preference order instead of numerical
7196 order. This is currently hardcoded for the highest order curves first.
7197 This should be configurable so applications can judge speed vs strength.
7201 * Add TLS v1.2 server support for client authentication.
7205 * Add support for FIPS mode in ssl library: disable SSLv3, non-FIPS ciphers
7210 * Functions FIPS_mode_set() and FIPS_mode() which call the underlying
7211 FIPS modules versions.
7215 * Add TLS v1.2 client side support for client authentication. Keep cache
7216 of handshake records longer as we don't know the hash algorithm to use
7217 until after the certificate request message is received.
7221 * Initial TLS v1.2 client support. Add a default signature algorithms
7222 extension including all the algorithms we support. Parse new signature
7223 format in client key exchange. Relax some ECC signing restrictions for
7224 TLS v1.2 as indicated in RFC5246.
7228 * Add server support for TLS v1.2 signature algorithms extension. Switch
7229 to new signature format when needed using client digest preference.
7230 All server ciphersuites should now work correctly in TLS v1.2. No client
7231 support yet and no support for client certificates.
7235 * Initial TLS v1.2 support. Add new SHA256 digest to ssl code, switch
7236 to SHA256 for PRF when using TLS v1.2 and later. Add new SHA256 based
7237 ciphersuites. At present only RSA key exchange ciphersuites work with
7238 TLS v1.2. Add new option for TLS v1.2 replacing the old and obsolete
7239 SSL_OP_PKCS1_CHECK flags with SSL_OP_NO_TLSv1_2. New TLSv1.2 methods
7240 and version checking.
7244 * New option OPENSSL_NO_SSL_INTERN. If an application can be compiled
7245 with this defined it will not be affected by any changes to ssl internal
7246 structures. Add several utility functions to allow openssl application
7247 to work with OPENSSL_NO_SSL_INTERN defined.
7251 * A long standing patch to add support for SRP from EdelWeb (Peter
7252 Sylvester and Christophe Renou) was integrated.
7253 *Christophe Renou <christophe.renou@edelweb.fr>, Peter Sylvester
7254 <peter.sylvester@edelweb.fr>, Tom Wu <tjw@cs.stanford.edu>, and
7257 * Add functions to copy EVP_PKEY_METHOD and retrieve flags and id.
7261 * Permit abbreviated handshakes when renegotiating using the function
7262 SSL_renegotiate_abbreviated().
7264 *Robin Seggelmann <seggelmann@fh-muenster.de>*
7266 * Add call to ENGINE_register_all_complete() to
7267 ENGINE_load_builtin_engines(), so some implementations get used
7268 automatically instead of needing explicit application support.
7272 * Add support for TLS key exporter as described in RFC5705.
7274 *Robin Seggelmann <seggelmann@fh-muenster.de>, Steve Henson*
7276 * Initial TLSv1.1 support. Since TLSv1.1 is very similar to TLS v1.0 only
7277 a few changes are required:
7279 Add SSL_OP_NO_TLSv1_1 flag.
7280 Add TLSv1_1 methods.
7281 Update version checking logic to handle version 1.1.
7282 Add explicit IV handling (ported from DTLS code).
7283 Add command line options to s_client/s_server.
7290 ### Changes between 1.0.0s and 1.0.0t [3 Dec 2015]
7292 * X509_ATTRIBUTE memory leak
7294 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
7295 memory. This structure is used by the PKCS#7 and CMS routines so any
7296 application which reads PKCS#7 or CMS data from untrusted sources is
7297 affected. SSL/TLS is not affected.
7299 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
7305 * Race condition handling PSK identify hint
7307 If PSK identity hints are received by a multi-threaded client then
7308 the values are wrongly updated in the parent SSL_CTX structure. This can
7309 result in a race condition potentially leading to a double free of the
7315 ### Changes between 1.0.0r and 1.0.0s [11 Jun 2015]
7317 * Malformed ECParameters causes infinite loop
7319 When processing an ECParameters structure OpenSSL enters an infinite loop
7320 if the curve specified is over a specially malformed binary polynomial
7323 This can be used to perform denial of service against any
7324 system which processes public keys, certificate requests or
7325 certificates. This includes TLS clients and TLS servers with
7326 client authentication enabled.
7328 This issue was reported to OpenSSL by Joseph Barr-Pixton.
7333 * Exploitable out-of-bounds read in X509_cmp_time
7335 X509_cmp_time does not properly check the length of the ASN1_TIME
7336 string and can read a few bytes out of bounds. In addition,
7337 X509_cmp_time accepts an arbitrary number of fractional seconds in the
7340 An attacker can use this to craft malformed certificates and CRLs of
7341 various sizes and potentially cause a segmentation fault, resulting in
7342 a DoS on applications that verify certificates or CRLs. TLS clients
7343 that verify CRLs are affected. TLS clients and servers with client
7344 authentication enabled may be affected if they use custom verification
7347 This issue was reported to OpenSSL by Robert Swiecki (Google), and
7348 independently by Hanno Böck.
7353 * PKCS7 crash with missing EnvelopedContent
7355 The PKCS#7 parsing code does not handle missing inner EncryptedContent
7356 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
7357 with missing content and trigger a NULL pointer dereference on parsing.
7359 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
7360 structures from untrusted sources are affected. OpenSSL clients and
7361 servers are not affected.
7363 This issue was reported to OpenSSL by Michal Zalewski (Google).
7368 * CMS verify infinite loop with unknown hash function
7370 When verifying a signedData message the CMS code can enter an infinite loop
7371 if presented with an unknown hash function OID. This can be used to perform
7372 denial of service against any system which verifies signedData messages using
7374 This issue was reported to OpenSSL by Johannes Bauer.
7379 * Race condition handling NewSessionTicket
7381 If a NewSessionTicket is received by a multi-threaded client when attempting to
7382 reuse a previous ticket then a race condition can occur potentially leading to
7383 a double free of the ticket data.
7388 ### Changes between 1.0.0q and 1.0.0r [19 Mar 2015]
7390 * Segmentation fault in ASN1_TYPE_cmp fix
7392 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
7393 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
7394 certificate signature algorithm consistency this can be used to crash any
7395 certificate verification operation and exploited in a DoS attack. Any
7396 application which performs certificate verification is vulnerable including
7397 OpenSSL clients and servers which enable client authentication.
7402 * ASN.1 structure reuse memory corruption fix
7404 Reusing a structure in ASN.1 parsing may allow an attacker to cause
7405 memory corruption via an invalid write. Such reuse is and has been
7406 strongly discouraged and is believed to be rare.
7408 Applications that parse structures containing CHOICE or ANY DEFINED BY
7409 components may be affected. Certificate parsing (d2i_X509 and related
7410 functions) are however not affected. OpenSSL clients and servers are
7416 * PKCS7 NULL pointer dereferences fix
7418 The PKCS#7 parsing code does not handle missing outer ContentInfo
7419 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
7420 missing content and trigger a NULL pointer dereference on parsing.
7422 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
7423 otherwise parse PKCS#7 structures from untrusted sources are
7424 affected. OpenSSL clients and servers are not affected.
7426 This issue was reported to OpenSSL by Michal Zalewski (Google).
7431 * DoS via reachable assert in SSLv2 servers fix
7433 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
7434 servers that both support SSLv2 and enable export cipher suites by sending
7435 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
7437 This issue was discovered by Sean Burford (Google) and Emilia Käsper
7438 (OpenSSL development team).
7443 * Use After Free following d2i_ECPrivatekey error fix
7445 A malformed EC private key file consumed via the d2i_ECPrivateKey function
7446 could cause a use after free condition. This, in turn, could cause a double
7447 free in several private key parsing functions (such as d2i_PrivateKey
7448 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
7449 for applications that receive EC private keys from untrusted
7450 sources. This scenario is considered rare.
7452 This issue was discovered by the BoringSSL project and fixed in their
7458 * X509_to_X509_REQ NULL pointer deref fix
7460 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
7461 the certificate key is invalid. This function is rarely used in practice.
7463 This issue was discovered by Brian Carpenter.
7468 * Removed the export ciphers from the DEFAULT ciphers
7472 ### Changes between 1.0.0p and 1.0.0q [15 Jan 2015]
7474 * Build fixes for the Windows and OpenVMS platforms
7476 *Matt Caswell and Richard Levitte*
7478 ### Changes between 1.0.0o and 1.0.0p [8 Jan 2015]
7480 * Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
7481 message can cause a segmentation fault in OpenSSL due to a NULL pointer
7482 dereference. This could lead to a Denial Of Service attack. Thanks to
7483 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
7488 * Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
7489 dtls1_buffer_record function under certain conditions. In particular this
7490 could occur if an attacker sent repeated DTLS records with the same
7491 sequence number but for the next epoch. The memory leak could be exploited
7492 by an attacker in a Denial of Service attack through memory exhaustion.
7493 Thanks to Chris Mueller for reporting this issue.
7498 * Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
7499 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
7500 method would be set to NULL which could later result in a NULL pointer
7501 dereference. Thanks to Frank Schmirler for reporting this issue.
7506 * Abort handshake if server key exchange message is omitted for ephemeral
7509 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
7510 reporting this issue.
7515 * Remove non-export ephemeral RSA code on client and server. This code
7516 violated the TLS standard by allowing the use of temporary RSA keys in
7517 non-export ciphersuites and could be used by a server to effectively
7518 downgrade the RSA key length used to a value smaller than the server
7519 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
7520 INRIA or reporting this issue.
7525 * Fixed issue where DH client certificates are accepted without verification.
7526 An OpenSSL server will accept a DH certificate for client authentication
7527 without the certificate verify message. This effectively allows a client to
7528 authenticate without the use of a private key. This only affects servers
7529 which trust a client certificate authority which issues certificates
7530 containing DH keys: these are extremely rare and hardly ever encountered.
7531 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
7537 * Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
7538 results on some platforms, including x86_64. This bug occurs at random
7539 with a very low probability, and is not known to be exploitable in any
7540 way, though its exact impact is difficult to determine. Thanks to Pieter
7541 Wuille (Blockstream) who reported this issue and also suggested an initial
7542 fix. Further analysis was conducted by the OpenSSL development team and
7543 Adam Langley of Google. The final fix was developed by Andy Polyakov of
7544 the OpenSSL core team.
7549 * Fix various certificate fingerprint issues.
7551 By using non-DER or invalid encodings outside the signed portion of a
7552 certificate the fingerprint can be changed without breaking the signature.
7553 Although no details of the signed portion of the certificate can be changed
7554 this can cause problems with some applications: e.g. those using the
7555 certificate fingerprint for blacklists.
7557 1. Reject signatures with non zero unused bits.
7559 If the BIT STRING containing the signature has non zero unused bits reject
7560 the signature. All current signature algorithms require zero unused bits.
7562 2. Check certificate algorithm consistency.
7564 Check the AlgorithmIdentifier inside TBS matches the one in the
7565 certificate signature. NB: this will result in signature failure
7566 errors for some broken certificates.
7568 Thanks to Konrad Kraszewski from Google for reporting this issue.
7570 3. Check DSA/ECDSA signatures use DER.
7572 Reencode DSA/ECDSA signatures and compare with the original received
7573 signature. Return an error if there is a mismatch.
7575 This will reject various cases including garbage after signature
7576 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
7577 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
7578 (negative or with leading zeroes).
7580 Further analysis was conducted and fixes were developed by Stephen Henson
7581 of the OpenSSL core team.
7587 ### Changes between 1.0.0n and 1.0.0o [15 Oct 2014]
7589 * Session Ticket Memory Leak.
7591 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
7592 integrity of that ticket is first verified. In the event of a session
7593 ticket integrity check failing, OpenSSL will fail to free memory
7594 causing a memory leak. By sending a large number of invalid session
7595 tickets an attacker could exploit this issue in a Denial Of Service
7601 * Build option no-ssl3 is incomplete.
7603 When OpenSSL is configured with "no-ssl3" as a build option, servers
7604 could accept and complete a SSL 3.0 handshake, and clients could be
7605 configured to send them.
7608 *Akamai and the OpenSSL team*
7610 * Add support for TLS_FALLBACK_SCSV.
7611 Client applications doing fallback retries should call
7612 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
7615 *Adam Langley, Bodo Moeller*
7617 * Add additional DigestInfo checks.
7619 Reencode DigestInto in DER and check against the original when
7620 verifying RSA signature: this will reject any improperly encoded
7621 DigestInfo structures.
7623 Note: this is a precautionary measure and no attacks are currently known.
7627 ### Changes between 1.0.0m and 1.0.0n [6 Aug 2014]
7629 * OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
7630 to a denial of service attack. A malicious server can crash the client
7631 with a null pointer dereference (read) by specifying an anonymous (EC)DH
7632 ciphersuite and sending carefully crafted handshake messages.
7634 Thanks to Felix Gröbert (Google) for discovering and researching this
7640 * By sending carefully crafted DTLS packets an attacker could cause openssl
7641 to leak memory. This can be exploited through a Denial of Service attack.
7642 Thanks to Adam Langley for discovering and researching this issue.
7647 * An attacker can force openssl to consume large amounts of memory whilst
7648 processing DTLS handshake messages. This can be exploited through a
7649 Denial of Service attack.
7650 Thanks to Adam Langley for discovering and researching this issue.
7655 * An attacker can force an error condition which causes openssl to crash
7656 whilst processing DTLS packets due to memory being freed twice. This
7657 can be exploited through a Denial of Service attack.
7658 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
7664 * If a multithreaded client connects to a malicious server using a resumed
7665 session and the server sends an ec point format extension it could write
7666 up to 255 bytes to freed memory.
7668 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
7674 * A flaw in OBJ_obj2txt may cause pretty printing functions such as
7675 X509_name_oneline, X509_name_print_ex et al. to leak some information
7676 from the stack. Applications may be affected if they echo pretty printing
7677 output to the attacker.
7679 Thanks to Ivan Fratric (Google) for discovering this issue.
7682 *Emilia Käsper, and Steve Henson*
7684 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
7685 for corner cases. (Certain input points at infinity could lead to
7686 bogus results, with non-infinity inputs mapped to infinity too.)
7690 ### Changes between 1.0.0l and 1.0.0m [5 Jun 2014]
7692 * Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
7693 handshake can force the use of weak keying material in OpenSSL
7694 SSL/TLS clients and servers.
7696 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
7697 researching this issue. ([CVE-2014-0224])
7699 *KIKUCHI Masashi, Steve Henson*
7701 * Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
7702 OpenSSL DTLS client the code can be made to recurse eventually crashing
7705 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
7708 *Imre Rad, Steve Henson*
7710 * Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
7711 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
7712 client or server. This is potentially exploitable to run arbitrary
7713 code on a vulnerable client or server.
7715 Thanks to Jüri Aedla for reporting this issue. ([CVE-2014-0195])
7717 *Jüri Aedla, Steve Henson*
7719 * Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
7720 are subject to a denial of service attack.
7722 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
7723 this issue. ([CVE-2014-3470])
7725 *Felix Gröbert, Ivan Fratric, Steve Henson*
7727 * Harmonize version and its documentation. -f flag is used to display
7730 *mancha <mancha1@zoho.com>*
7732 * Fix eckey_priv_encode so it immediately returns an error upon a failure
7733 in i2d_ECPrivateKey.
7735 *mancha <mancha1@zoho.com>*
7737 * Fix some double frees. These are not thought to be exploitable.
7739 *mancha <mancha1@zoho.com>*
7741 * Fix for the attack described in the paper "Recovering OpenSSL
7742 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
7743 by Yuval Yarom and Naomi Benger. Details can be obtained from:
7744 <http://eprint.iacr.org/2014/140>
7746 Thanks to Yuval Yarom and Naomi Benger for discovering this
7747 flaw and to Yuval Yarom for supplying a fix ([CVE-2014-0076])
7749 *Yuval Yarom and Naomi Benger*
7751 ### Changes between 1.0.0k and 1.0.0l [6 Jan 2014]
7753 * Keep original DTLS digest and encryption contexts in retransmission
7754 structures so we can use the previous session parameters if they need
7755 to be resent. ([CVE-2013-6450])
7759 * Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
7760 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
7761 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
7762 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
7763 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
7764 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
7766 *Rob Stradling, Adam Langley*
7768 ### Changes between 1.0.0j and 1.0.0k [5 Feb 2013]
7770 * Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
7772 This addresses the flaw in CBC record processing discovered by
7773 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
7774 at: <http://www.isg.rhul.ac.uk/tls/>
7776 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
7777 Security Group at Royal Holloway, University of London
7778 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
7779 Emilia Käsper for the initial patch.
7782 *Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson*
7784 * Return an error when checking OCSP signatures when key is NULL.
7785 This fixes a DoS attack. ([CVE-2013-0166])
7789 * Call OCSP Stapling callback after ciphersuite has been chosen, so
7790 the right response is stapled. Also change SSL_get_certificate()
7791 so it returns the certificate actually sent.
7792 See <http://rt.openssl.org/Ticket/Display.html?id=2836>.
7793 (This is a backport)
7795 *Rob Stradling <rob.stradling@comodo.com>*
7797 * Fix possible deadlock when decoding public keys.
7801 ### Changes between 1.0.0i and 1.0.0j [10 May 2012]
7803 [NB: OpenSSL 1.0.0i and later 1.0.0 patch levels were released after
7806 * Sanity check record length before skipping explicit IV in DTLS
7809 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
7810 fuzzing as a service testing platform.
7815 * Initialise tkeylen properly when encrypting CMS messages.
7816 Thanks to Solar Designer of Openwall for reporting this issue.
7820 ### Changes between 1.0.0h and 1.0.0i [19 Apr 2012]
7822 * Check for potentially exploitable overflows in asn1_d2i_read_bio
7823 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
7824 in CRYPTO_realloc_clean.
7826 Thanks to Tavis Ormandy, Google Security Team, for discovering this
7827 issue and to Adam Langley <agl@chromium.org> for fixing it.
7830 *Adam Langley (Google), Tavis Ormandy, Google Security Team*
7832 ### Changes between 1.0.0g and 1.0.0h [12 Mar 2012]
7834 * Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness
7835 in CMS and PKCS7 code. When RSA decryption fails use a random key for
7836 content decryption and always return the same error. Note: this attack
7837 needs on average 2^20 messages so it only affects automated senders. The
7838 old behaviour can be re-enabled in the CMS code by setting the
7839 CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where
7840 an MMA defence is not necessary.
7841 Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering
7842 this issue. ([CVE-2012-0884])
7846 * Fix CVE-2011-4619: make sure we really are receiving a
7847 client hello before rejecting multiple SGC restarts. Thanks to
7848 Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug.
7852 ### Changes between 1.0.0f and 1.0.0g [18 Jan 2012]
7854 * Fix for DTLS DoS issue introduced by fix for CVE-2011-4109.
7855 Thanks to Antonio Martin, Enterprise Secure Access Research and
7856 Development, Cisco Systems, Inc. for discovering this bug and
7857 preparing a fix. ([CVE-2012-0050])
7861 ### Changes between 1.0.0e and 1.0.0f [4 Jan 2012]
7863 * Nadhem Alfardan and Kenny Paterson have discovered an extension
7864 of the Vaudenay padding oracle attack on CBC mode encryption
7865 which enables an efficient plaintext recovery attack against
7866 the OpenSSL implementation of DTLS. Their attack exploits timing
7867 differences arising during decryption processing. A research
7868 paper describing this attack can be found at:
7869 <http://www.isg.rhul.ac.uk/~kp/dtls.pdf>
7870 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
7871 Security Group at Royal Holloway, University of London
7872 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
7873 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
7874 for preparing the fix. ([CVE-2011-4108])
7876 *Robin Seggelmann, Michael Tuexen*
7878 * Clear bytes used for block padding of SSL 3.0 records.
7881 *Adam Langley (Google)*
7883 * Only allow one SGC handshake restart for SSL/TLS. Thanks to George
7884 Kadianakis <desnacked@gmail.com> for discovering this issue and
7885 Adam Langley for preparing the fix. ([CVE-2011-4619])
7887 *Adam Langley (Google)*
7889 * Check parameters are not NULL in GOST ENGINE. ([CVE-2012-0027])
7891 *Andrey Kulikov <amdeich@gmail.com>*
7893 * Prevent malformed RFC3779 data triggering an assertion failure.
7894 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
7895 and Rob Austein <sra@hactrn.net> for fixing it. ([CVE-2011-4577])
7897 *Rob Austein <sra@hactrn.net>*
7899 * Improved PRNG seeding for VOS.
7901 *Paul Green <Paul.Green@stratus.com>*
7903 * Fix ssl_ciph.c set-up race.
7905 *Adam Langley (Google)*
7907 * Fix spurious failures in ecdsatest.c.
7909 *Emilia Käsper (Google)*
7911 * Fix the BIO_f_buffer() implementation (which was mixing different
7912 interpretations of the `..._len` fields).
7914 *Adam Langley (Google)*
7916 * Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
7917 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
7918 threads won't reuse the same blinding coefficients.
7920 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
7921 lock to call BN_BLINDING_invert_ex, and avoids one use of
7922 BN_BLINDING_update for each BN_BLINDING structure (previously,
7923 the last update always remained unused).
7925 *Emilia Käsper (Google)*
7927 * In ssl3_clear, preserve s3->init_extra along with s3->rbuf.
7929 *Bob Buckholz (Google)*
7931 ### Changes between 1.0.0d and 1.0.0e [6 Sep 2011]
7933 * Fix bug where CRLs with nextUpdate in the past are sometimes accepted
7934 by initialising X509_STORE_CTX properly. ([CVE-2011-3207])
7936 *Kaspar Brand <ossl@velox.ch>*
7938 * Fix SSL memory handling for (EC)DH ciphersuites, in particular
7939 for multi-threaded use of ECDH. ([CVE-2011-3210])
7941 *Adam Langley (Google)*
7943 * Fix x509_name_ex_d2i memory leak on bad inputs.
7947 * Remove hard coded ecdsaWithSHA1 signature tests in ssl code and check
7948 signature public key algorithm by using OID xref utilities instead.
7949 Before this you could only use some ECC ciphersuites with SHA1 only.
7953 * Add protection against ECDSA timing attacks as mentioned in the paper
7954 by Billy Bob Brumley and Nicola Tuveri, see:
7955 <http://eprint.iacr.org/2011/232.pdf>
7957 *Billy Bob Brumley and Nicola Tuveri*
7959 ### Changes between 1.0.0c and 1.0.0d [8 Feb 2011]
7961 * Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
7963 *Neel Mehta, Adam Langley, Bodo Moeller (Google)*
7965 * Fix bug in string printing code: if *any* escaping is enabled we must
7966 escape the escape character (backslash) or the resulting string is
7971 ### Changes between 1.0.0b and 1.0.0c [2 Dec 2010]
7973 * Disable code workaround for ancient and obsolete Netscape browsers
7974 and servers: an attacker can use it in a ciphersuite downgrade attack.
7975 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
7979 * Fixed J-PAKE implementation error, originally discovered by
7980 Sebastien Martini, further info and confirmation from Stefan
7981 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
7985 ### Changes between 1.0.0a and 1.0.0b [16 Nov 2010]
7987 * Fix extension code to avoid race conditions which can result in a buffer
7988 overrun vulnerability: resumed sessions must not be modified as they can
7989 be shared by multiple threads. CVE-2010-3864
7993 * Fix WIN32 build system to correctly link an ENGINE directory into
7998 ### Changes between 1.0.0 and 1.0.0a [01 Jun 2010]
8000 * Check return value of int_rsa_verify in pkey_rsa_verifyrecover
8003 *Steve Henson, Peter-Michael Hager <hager@dortmund.net>*
8005 ### Changes between 0.9.8n and 1.0.0 [29 Mar 2010]
8007 * Add "missing" function EVP_CIPHER_CTX_copy(). This copies a cipher
8008 context. The operation can be customised via the ctrl mechanism in
8009 case ENGINEs want to include additional functionality.
8013 * Tolerate yet another broken PKCS#8 key format: private key value negative.
8017 * Add new -subject_hash_old and -issuer_hash_old options to x509 utility to
8018 output hashes compatible with older versions of OpenSSL.
8020 *Willy Weisz <weisz@vcpc.univie.ac.at>*
8022 * Fix compression algorithm handling: if resuming a session use the
8023 compression algorithm of the resumed session instead of determining
8024 it from client hello again. Don't allow server to change algorithm.
8028 * Add load_crls() function to commands tidying load_certs() too. Add option
8029 to verify utility to allow additional CRLs to be included.
8033 * Update OCSP request code to permit adding custom headers to the request:
8034 some responders need this.
8038 * The function EVP_PKEY_sign() returns <=0 on error: check return code
8041 *Julia Lawall <julia@diku.dk>*
8043 * Update verify callback code in `apps/s_cb.c` and `apps/verify.c`, it
8044 needlessly dereferenced structures, used obsolete functions and
8045 didn't handle all updated verify codes correctly.
8049 * Disable MD2 in the default configuration.
8053 * In BIO_pop() and BIO_push() use the ctrl argument (which was NULL) to
8054 indicate the initial BIO being pushed or popped. This makes it possible
8055 to determine whether the BIO is the one explicitly called or as a result
8056 of the ctrl being passed down the chain. Fix BIO_pop() and SSL BIOs so
8057 it handles reference counts correctly and doesn't zero out the I/O bio
8058 when it is not being explicitly popped. WARNING: applications which
8059 included workarounds for the old buggy behaviour will need to be modified
8060 or they could free up already freed BIOs.
8064 * Extend the uni2asc/asc2uni => OPENSSL_uni2asc/OPENSSL_asc2uni
8065 renaming to all platforms (within the 0.9.8 branch, this was
8066 done conditionally on Netware platforms to avoid a name clash).
8068 *Guenter <lists@gknw.net>*
8070 * Add ECDHE and PSK support to DTLS.
8072 *Michael Tuexen <tuexen@fh-muenster.de>*
8074 * Add CHECKED_STACK_OF macro to safestack.h, otherwise safestack can't
8079 * Add "missing" function EVP_MD_flags() (without this the only way to
8080 retrieve a digest flags is by accessing the structure directly. Update
8081 `EVP_MD_do_all*()` and `EVP_CIPHER_do_all*()` to include the name a digest
8082 or cipher is registered as in the "from" argument. Print out all
8083 registered digests in the dgst usage message instead of manually
8084 attempting to work them out.
8088 * If no SSLv2 ciphers are used don't use an SSLv2 compatible client hello:
8089 this allows the use of compression and extensions. Change default cipher
8090 string to remove SSLv2 ciphersuites. This effectively avoids ancient SSLv2
8091 by default unless an application cipher string requests it.
8095 * Alter match criteria in PKCS12_parse(). It used to try to use local
8096 key ids to find matching certificates and keys but some PKCS#12 files
8097 don't follow the (somewhat unwritten) rules and this strategy fails.
8098 Now just gather all certificates together and the first private key
8099 then look for the first certificate that matches the key.
8103 * Support use of registered digest and cipher names for dgst and cipher
8104 commands instead of having to add each one as a special case. So now
8111 openssl dgst -sha256 foo
8113 and this works for ENGINE based algorithms too.
8117 * Update Gost ENGINE to support parameter files.
8119 *Victor B. Wagner <vitus@cryptocom.ru>*
8121 * Support GeneralizedTime in ca utility.
8123 *Oliver Martin <oliver@volatilevoid.net>, Steve Henson*
8125 * Enhance the hash format used for certificate directory links. The new
8126 form uses the canonical encoding (meaning equivalent names will work
8127 even if they aren't identical) and uses SHA1 instead of MD5. This form
8128 is incompatible with the older format and as a result c_rehash should
8129 be used to rebuild symbolic links.
8133 * Make PKCS#8 the default write format for private keys, replacing the
8134 traditional format. This form is standardised, more secure and doesn't
8135 include an implicit MD5 dependency.
8139 * Add a $gcc_devteam_warn option to Configure. The idea is that any code
8140 committed to OpenSSL should pass this lot as a minimum.
8144 * Add session ticket override functionality for use by EAP-FAST.
8146 *Jouni Malinen <j@w1.fi>*
8148 * Modify HMAC functions to return a value. Since these can be implemented
8149 in an ENGINE errors can occur.
8153 * Type-checked OBJ_bsearch_ex.
8157 * Type-checked OBJ_bsearch. Also some constification necessitated
8158 by type-checking. Still to come: TXT_DB, bsearch(?),
8159 OBJ_bsearch_ex, qsort, CRYPTO_EX_DATA, ASN1_VALUE, ASN1_STRING,
8164 * New function OPENSSL_gmtime_adj() to add a specific number of days and
8165 seconds to a tm structure directly, instead of going through OS
8166 specific date routines. This avoids any issues with OS routines such
8167 as the year 2038 bug. New `*_adj()` functions for ASN1 time structures
8168 and X509_time_adj_ex() to cover the extended range. The existing
8169 X509_time_adj() is still usable and will no longer have any date issues.
8173 * Delta CRL support. New use deltas option which will attempt to locate
8174 and search any appropriate delta CRLs available.
8176 This work was sponsored by Google.
8180 * Support for CRLs partitioned by reason code. Reorganise CRL processing
8181 code and add additional score elements. Validate alternate CRL paths
8182 as part of the CRL checking and indicate a new error "CRL path validation
8183 error" in this case. Applications wanting additional details can use
8184 the verify callback and check the new "parent" field. If this is not
8185 NULL CRL path validation is taking place. Existing applications won't
8186 see this because it requires extended CRL support which is off by
8189 This work was sponsored by Google.
8193 * Support for freshest CRL extension.
8195 This work was sponsored by Google.
8199 * Initial indirect CRL support. Currently only supported in the CRLs
8200 passed directly and not via lookup. Process certificate issuer
8201 CRL entry extension and lookup CRL entries by bother issuer name
8202 and serial number. Check and process CRL issuer entry in IDP extension.
8204 This work was sponsored by Google.
8208 * Add support for distinct certificate and CRL paths. The CRL issuer
8209 certificate is validated separately in this case. Only enabled if
8210 an extended CRL support flag is set: this flag will enable additional
8211 CRL functionality in future.
8213 This work was sponsored by Google.
8217 * Add support for policy mappings extension.
8219 This work was sponsored by Google.
8223 * Fixes to pathlength constraint, self issued certificate handling,
8224 policy processing to align with RFC3280 and PKITS tests.
8226 This work was sponsored by Google.
8230 * Support for name constraints certificate extension. DN, email, DNS
8231 and URI types are currently supported.
8233 This work was sponsored by Google.
8237 * To cater for systems that provide a pointer-based thread ID rather
8238 than numeric, deprecate the current numeric thread ID mechanism and
8239 replace it with a structure and associated callback type. This
8240 mechanism allows a numeric "hash" to be extracted from a thread ID in
8241 either case, and on platforms where pointers are larger than 'long',
8242 mixing is done to help ensure the numeric 'hash' is usable even if it
8243 can't be guaranteed unique. The default mechanism is to use "&errno"
8244 as a pointer-based thread ID to distinguish between threads.
8246 Applications that want to provide their own thread IDs should now use
8247 CRYPTO_THREADID_set_callback() to register a callback that will call
8248 either CRYPTO_THREADID_set_numeric() or CRYPTO_THREADID_set_pointer().
8250 Note that ERR_remove_state() is now deprecated, because it is tied
8251 to the assumption that thread IDs are numeric. ERR_remove_state(0)
8252 to free the current thread's error state should be replaced by
8253 ERR_remove_thread_state(NULL).
8255 (This new approach replaces the functions CRYPTO_set_idptr_callback(),
8256 CRYPTO_get_idptr_callback(), and CRYPTO_thread_idptr() that existed in
8257 OpenSSL 0.9.9-dev between June 2006 and August 2008. Also, if an
8258 application was previously providing a numeric thread callback that
8259 was inappropriate for distinguishing threads, then uniqueness might
8260 have been obtained with &errno that happened immediately in the
8261 intermediate development versions of OpenSSL; this is no longer the
8262 case, the numeric thread callback will now override the automatic use
8265 *Geoff Thorpe, with help from Bodo Moeller*
8267 * Initial support for different CRL issuing certificates. This covers a
8268 simple case where the self issued certificates in the chain exist and
8269 the real CRL issuer is higher in the existing chain.
8271 This work was sponsored by Google.
8275 * Removed effectively defunct crypto/store from the build.
8279 * Revamp of STACK to provide stronger type-checking. Still to come:
8280 TXT_DB, bsearch(?), OBJ_bsearch, qsort, CRYPTO_EX_DATA, ASN1_VALUE,
8281 ASN1_STRING, CONF_VALUE.
8285 * Add a new SSL_MODE_RELEASE_BUFFERS mode flag to release unused buffer
8286 RAM on SSL connections. This option can save about 34k per idle SSL.
8290 * Revamp of LHASH to provide stronger type-checking. Still to come:
8291 STACK, TXT_DB, bsearch, qsort.
8295 * Initial support for Cryptographic Message Syntax (aka CMS) based
8296 on RFC3850, RFC3851 and RFC3852. New cms directory and cms utility,
8297 support for data, signedData, compressedData, digestedData and
8298 encryptedData, envelopedData types included. Scripts to check against
8299 RFC4134 examples draft and interop and consistency checks of many
8300 content types and variants.
8304 * Add options to enc utility to support use of zlib compression BIO.
8308 * Extend mk1mf to support importing of options and assembly language
8309 files from Configure script, currently only included in VC-WIN32.
8310 The assembly language rules can now optionally generate the source
8311 files from the associated perl scripts.
8315 * Implement remaining functionality needed to support GOST ciphersuites.
8316 Interop testing has been performed using CryptoPro implementations.
8318 *Victor B. Wagner <vitus@cryptocom.ru>*
8320 * s390x assembler pack.
8324 * ARMv4 assembler pack. ARMv4 refers to v4 and later ISA, not CPU
8329 * Implement Opaque PRF Input TLS extension as specified in
8330 draft-rescorla-tls-opaque-prf-input-00.txt. Since this is not an
8331 official specification yet and no extension type assignment by
8332 IANA exists, this extension (for now) will have to be explicitly
8333 enabled when building OpenSSL by providing the extension number
8334 to use. For example, specify an option
8336 -DTLSEXT_TYPE_opaque_prf_input=0x9527
8338 to the "config" or "Configure" script to enable the extension,
8339 assuming extension number 0x9527 (which is a completely arbitrary
8340 and unofficial assignment based on the MD5 hash of the Internet
8341 Draft). Note that by doing so, you potentially lose
8342 interoperability with other TLS implementations since these might
8343 be using the same extension number for other purposes.
8345 SSL_set_tlsext_opaque_prf_input(ssl, src, len) is used to set the
8346 opaque PRF input value to use in the handshake. This will create
8347 an internal copy of the length-'len' string at 'src', and will
8348 return non-zero for success.
8350 To get more control and flexibility, provide a callback function
8353 SSL_CTX_set_tlsext_opaque_prf_input_callback(ctx, cb)
8354 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg(ctx, arg)
8358 int (*cb)(SSL *, void *peerinput, size_t len, void *arg);
8361 Callback function 'cb' will be called in handshakes, and is
8362 expected to use SSL_set_tlsext_opaque_prf_input() as appropriate.
8363 Argument 'arg' is for application purposes (the value as given to
8364 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg() will directly
8365 be provided to the callback function). The callback function
8366 has to return non-zero to report success: usually 1 to use opaque
8367 PRF input just if possible, or 2 to enforce use of the opaque PRF
8368 input. In the latter case, the library will abort the handshake
8369 if opaque PRF input is not successfully negotiated.
8371 Arguments 'peerinput' and 'len' given to the callback function
8372 will always be NULL and 0 in the case of a client. A server will
8373 see the client's opaque PRF input through these variables if
8374 available (NULL and 0 otherwise). Note that if the server
8375 provides an opaque PRF input, the length must be the same as the
8376 length of the client's opaque PRF input.
8378 Note that the callback function will only be called when creating
8379 a new session (session resumption can resume whatever was
8380 previously negotiated), and will not be called in SSL 2.0
8381 handshakes; thus, SSL_CTX_set_options(ctx, SSL_OP_NO_SSLv2) or
8382 SSL_set_options(ssl, SSL_OP_NO_SSLv2) is especially recommended
8383 for applications that need to enforce opaque PRF input.
8387 * Update ssl code to support digests other than SHA1+MD5 for handshake
8390 *Victor B. Wagner <vitus@cryptocom.ru>*
8392 * Add RFC4507 support to OpenSSL. This includes the corrections in
8393 RFC4507bis. The encrypted ticket format is an encrypted encoded
8394 SSL_SESSION structure, that way new session features are automatically
8397 If a client application caches session in an SSL_SESSION structure
8398 support is transparent because tickets are now stored in the encoded
8401 The SSL_CTX structure automatically generates keys for ticket
8402 protection in servers so again support should be possible
8403 with no application modification.
8405 If a client or server wishes to disable RFC4507 support then the option
8406 SSL_OP_NO_TICKET can be set.
8408 Add a TLS extension debugging callback to allow the contents of any client
8409 or server extensions to be examined.
8411 This work was sponsored by Google.
8415 * Final changes to avoid use of pointer pointer casts in OpenSSL.
8416 OpenSSL should now compile cleanly on gcc 4.2
8418 *Peter Hartley <pdh@utter.chaos.org.uk>, Steve Henson*
8420 * Update SSL library to use new EVP_PKEY MAC API. Include generic MAC
8421 support including streaming MAC support: this is required for GOST
8422 ciphersuite support.
8424 *Victor B. Wagner <vitus@cryptocom.ru>, Steve Henson*
8426 * Add option -stream to use PKCS#7 streaming in smime utility. New
8427 function i2d_PKCS7_bio_stream() and PEM_write_PKCS7_bio_stream()
8428 to output in BER and PEM format.
8432 * Experimental support for use of HMAC via EVP_PKEY interface. This
8433 allows HMAC to be handled via the `EVP_DigestSign*()` interface. The
8434 EVP_PKEY "key" in this case is the HMAC key, potentially allowing
8435 ENGINE support for HMAC keys which are unextractable. New -mac and
8436 -macopt options to dgst utility.
8440 * New option -sigopt to dgst utility. Update dgst to use
8441 `EVP_Digest{Sign,Verify}*`. These two changes make it possible to use
8442 alternative signing parameters such as X9.31 or PSS in the dgst
8447 * Change ssl_cipher_apply_rule(), the internal function that does
8448 the work each time a ciphersuite string requests enabling
8449 ("foo+bar"), moving ("+foo+bar"), disabling ("-foo+bar", or
8450 removing ("!foo+bar") a class of ciphersuites: Now it maintains
8451 the order of disabled ciphersuites such that those ciphersuites
8452 that most recently went from enabled to disabled not only stay
8453 in order with respect to each other, but also have higher priority
8454 than other disabled ciphersuites the next time ciphersuites are
8457 This means that you can now say, e.g., "PSK:-PSK:HIGH" to enable
8458 the same ciphersuites as with "HIGH" alone, but in a specific
8459 order where the PSK ciphersuites come first (since they are the
8460 most recently disabled ciphersuites when "HIGH" is parsed).
8462 Also, change ssl_create_cipher_list() (using this new
8463 functionality) such that between otherwise identical
8464 ciphersuites, ephemeral ECDH is preferred over ephemeral DH in
8469 * Change ssl_create_cipher_list() so that it automatically
8470 arranges the ciphersuites in reasonable order before starting
8471 to process the rule string. Thus, the definition for "DEFAULT"
8472 (SSL_DEFAULT_CIPHER_LIST) now is just "ALL:!aNULL:!eNULL", but
8473 remains equivalent to `"AES:ALL:!aNULL:!eNULL:+aECDH:+kRSA:+RC4:@STRENGTH"`.
8474 This makes it much easier to arrive at a reasonable default order
8475 in applications for which anonymous ciphers are OK (meaning
8476 that you can't actually use DEFAULT).
8478 *Bodo Moeller; suggested by Victor Duchovni*
8480 * Split the SSL/TLS algorithm mask (as used for ciphersuite string
8481 processing) into multiple integers instead of setting
8482 "SSL_MKEY_MASK" bits, "SSL_AUTH_MASK" bits, "SSL_ENC_MASK",
8483 "SSL_MAC_MASK", and "SSL_SSL_MASK" bits all in a single integer.
8484 (These masks as well as the individual bit definitions are hidden
8485 away into the non-exported interface ssl/ssl_locl.h, so this
8486 change to the definition of the SSL_CIPHER structure shouldn't
8487 affect applications.) This give us more bits for each of these
8488 categories, so there is no longer a need to coagulate AES128 and
8489 AES256 into a single algorithm bit, and to coagulate Camellia128
8490 and Camellia256 into a single algorithm bit, which has led to all
8493 Thus, among other things, the kludge introduced in 0.9.7m and
8494 0.9.8e for masking out AES256 independently of AES128 or masking
8495 out Camellia256 independently of AES256 is not needed here in 0.9.9.
8497 With the change, we also introduce new ciphersuite aliases that
8498 so far were missing: "AES128", "AES256", "CAMELLIA128", and
8503 * Add support for dsa-with-SHA224 and dsa-with-SHA256.
8504 Use the leftmost N bytes of the signature input if the input is
8505 larger than the prime q (with N being the size in bytes of q).
8509 * Very *very* experimental PKCS#7 streaming encoder support. Nothing uses
8510 it yet and it is largely untested.
8514 * Add support for the ecdsa-with-SHA224/256/384/512 signature types.
8518 * Initial incomplete changes to avoid need for function casts in OpenSSL
8519 some compilers (gcc 4.2 and later) reject their use. Safestack is
8520 reimplemented. Update ASN1 to avoid use of legacy functions.
8524 * Win32/64 targets are linked with Winsock2.
8528 * Add an X509_CRL_METHOD structure to allow CRL processing to be redirected
8529 to external functions. This can be used to increase CRL handling
8530 efficiency especially when CRLs are very large by (for example) storing
8531 the CRL revoked certificates in a database.
8535 * Overhaul of by_dir code. Add support for dynamic loading of CRLs so
8536 new CRLs added to a directory can be used. New command line option
8537 -verify_return_error to s_client and s_server. This causes real errors
8538 to be returned by the verify callback instead of carrying on no matter
8539 what. This reflects the way a "real world" verify callback would behave.
8543 * GOST engine, supporting several GOST algorithms and public key formats.
8544 Kindly donated by Cryptocom.
8548 * Partial support for Issuing Distribution Point CRL extension. CRLs
8549 partitioned by DP are handled but no indirect CRL or reason partitioning
8550 (yet). Complete overhaul of CRL handling: now the most suitable CRL is
8551 selected via a scoring technique which handles IDP and AKID in CRLs.
8555 * New X509_STORE_CTX callbacks lookup_crls() and lookup_certs() which
8556 will ultimately be used for all verify operations: this will remove the
8557 X509_STORE dependency on certificate verification and allow alternative
8558 lookup methods. X509_STORE based implementations of these two callbacks.
8562 * Allow multiple CRLs to exist in an X509_STORE with matching issuer names.
8563 Modify get_crl() to find a valid (unexpired) CRL if possible.
8567 * New function X509_CRL_match() to check if two CRLs are identical. Normally
8568 this would be called X509_CRL_cmp() but that name is already used by
8569 a function that just compares CRL issuer names. Cache several CRL
8570 extensions in X509_CRL structure and cache CRLDP in X509.
8574 * Store a "canonical" representation of X509_NAME structure (ASN1 Name)
8575 this maps equivalent X509_NAME structures into a consistent structure.
8576 Name comparison can then be performed rapidly using memcmp().
8580 * Non-blocking OCSP request processing. Add -timeout option to ocsp
8585 * Allow digests to supply their own micalg string for S/MIME type using
8586 the ctrl EVP_MD_CTRL_MICALG.
8590 * During PKCS7 signing pass the PKCS7 SignerInfo structure to the
8591 EVP_PKEY_METHOD before and after signing via the EVP_PKEY_CTRL_PKCS7_SIGN
8592 ctrl. It can then customise the structure before and/or after signing
8597 * New function OBJ_add_sigid() to allow application defined signature OIDs
8598 to be added to OpenSSLs internal tables. New function OBJ_sigid_free()
8599 to free up any added signature OIDs.
8603 * New functions EVP_CIPHER_do_all(), EVP_CIPHER_do_all_sorted(),
8604 EVP_MD_do_all() and EVP_MD_do_all_sorted() to enumerate internal
8605 digest and cipher tables. New options added to openssl utility:
8606 list-message-digest-algorithms and list-cipher-algorithms.
8610 * Change the array representation of binary polynomials: the list
8611 of degrees of non-zero coefficients is now terminated with -1.
8612 Previously it was terminated with 0, which was also part of the
8613 value; thus, the array representation was not applicable to
8614 polynomials where t^0 has coefficient zero. This change makes
8615 the array representation useful in a more general context.
8619 * Various modifications and fixes to SSL/TLS cipher string
8620 handling. For ECC, the code now distinguishes between fixed ECDH
8621 with RSA certificates on the one hand and with ECDSA certificates
8622 on the other hand, since these are separate ciphersuites. The
8623 unused code for Fortezza ciphersuites has been removed.
8625 For consistency with EDH, ephemeral ECDH is now called "EECDH"
8626 (not "ECDHE"). For consistency with the code for DH
8627 certificates, use of ECDH certificates is now considered ECDH
8628 authentication, not RSA or ECDSA authentication (the latter is
8629 merely the CA's signing algorithm and not actively used in the
8632 The temporary ciphersuite alias "ECCdraft" is no longer
8633 available, and ECC ciphersuites are no longer excluded from "ALL"
8634 and "DEFAULT". The following aliases now exist for RFC 4492
8635 ciphersuites, most of these by analogy with the DH case:
8637 kECDHr - ECDH cert, signed with RSA
8638 kECDHe - ECDH cert, signed with ECDSA
8639 kECDH - ECDH cert (signed with either RSA or ECDSA)
8640 kEECDH - ephemeral ECDH
8641 ECDH - ECDH cert or ephemeral ECDH
8647 AECDH - anonymous ECDH
8648 EECDH - non-anonymous ephemeral ECDH (equivalent to "kEECDH:-AECDH")
8652 * Add additional S/MIME capabilities for AES and GOST ciphers if supported.
8653 Use correct micalg parameters depending on digest(s) in signed message.
8657 * Add engine support for EVP_PKEY_ASN1_METHOD. Add functions to process
8658 an ENGINE asn1 method. Support ENGINE lookups in the ASN1 code.
8662 * Initial engine support for EVP_PKEY_METHOD. New functions to permit
8663 an engine to register a method. Add ENGINE lookups for methods and
8664 functional reference processing.
8668 * New functions `EVP_Digest{Sign,Verify)*`. These are enhanced versions of
8669 `EVP_{Sign,Verify}*` which allow an application to customise the signature
8674 * New -resign option to smime utility. This adds one or more signers
8675 to an existing PKCS#7 signedData structure. Also -md option to use an
8676 alternative message digest algorithm for signing.
8680 * Tidy up PKCS#7 routines and add new functions to make it easier to
8681 create PKCS7 structures containing multiple signers. Update smime
8682 application to support multiple signers.
8686 * New -macalg option to pkcs12 utility to allow setting of an alternative
8691 * Initial support for PKCS#5 v2.0 PRFs other than default SHA1 HMAC.
8692 Reorganize PBE internals to lookup from a static table using NIDs,
8693 add support for HMAC PBE OID translation. Add a EVP_CIPHER ctrl:
8694 EVP_CTRL_PBE_PRF_NID this allows a cipher to specify an alternative
8695 PRF which will be automatically used with PBES2.
8699 * Replace the algorithm specific calls to generate keys in "req" with the
8704 * Update PKCS#7 enveloped data routines to use new API. This is now
8705 supported by any public key method supporting the encrypt operation. A
8706 ctrl is added to allow the public key algorithm to examine or modify
8707 the PKCS#7 RecipientInfo structure if it needs to: for RSA this is
8712 * Add a ctrl to asn1 method to allow a public key algorithm to express
8713 a default digest type to use. In most cases this will be SHA1 but some
8714 algorithms (such as GOST) need to specify an alternative digest. The
8715 return value indicates how strong the preference is 1 means optional and
8716 2 is mandatory (that is it is the only supported type). Modify
8717 ASN1_item_sign() to accept a NULL digest argument to indicate it should
8718 use the default md. Update openssl utilities to use the default digest
8719 type for signing if it is not explicitly indicated.
8723 * Use OID cross reference table in ASN1_sign() and ASN1_verify(). New
8724 EVP_MD flag EVP_MD_FLAG_PKEY_METHOD_SIGNATURE. This uses the relevant
8725 signing method from the key type. This effectively removes the link
8726 between digests and public key types.
8730 * Add an OID cross reference table and utility functions. Its purpose is to
8731 translate between signature OIDs such as SHA1WithrsaEncryption and SHA1,
8732 rsaEncryption. This will allow some of the algorithm specific hackery
8733 needed to use the correct OID to be removed.
8737 * Remove algorithm specific dependencies when setting PKCS7_SIGNER_INFO
8738 structures for PKCS7_sign(). They are now set up by the relevant public
8743 * Add provisional EC pkey method with support for ECDSA and ECDH.
8747 * Add support for key derivation (agreement) in the API, DH method and
8752 * Add DSA pkey method and DH pkey methods, extend DH ASN1 method to support
8753 public and private key formats. As a side effect these add additional
8754 command line functionality not previously available: DSA signatures can be
8755 generated and verified using pkeyutl and DH key support and generation in
8762 *Oliver Tappe <zooey@hirschkaefer.de>*
8764 * New make target "install_html_docs" installs HTML renditions of the
8767 *Oliver Tappe <zooey@hirschkaefer.de>*
8769 * New utility "genpkey" this is analogous to "genrsa" etc except it can
8770 generate keys for any algorithm. Extend and update EVP_PKEY_METHOD to
8771 support key and parameter generation and add initial key generation
8772 functionality for RSA.
8776 * Add functions for main EVP_PKEY_method operations. The undocumented
8777 functions `EVP_PKEY_{encrypt,decrypt}` have been renamed to
8778 `EVP_PKEY_{encrypt,decrypt}_old`.
8782 * Initial definitions for EVP_PKEY_METHOD. This will be a high level public
8783 key API, doesn't do much yet.
8787 * New function EVP_PKEY_asn1_get0_info() to retrieve information about
8788 public key algorithms. New option to openssl utility:
8789 "list-public-key-algorithms" to print out info.
8793 * Implement the Supported Elliptic Curves Extension for
8794 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
8798 * Don't free up OIDs in OBJ_cleanup() if they are in use by EVP_MD or
8799 EVP_CIPHER structures to avoid later problems in EVP_cleanup().
8803 * New utilities pkey and pkeyparam. These are similar to algorithm specific
8804 utilities such as rsa, dsa, dsaparam etc except they process any key
8809 * Transfer public key printing routines to EVP_PKEY_ASN1_METHOD. New
8810 functions EVP_PKEY_print_public(), EVP_PKEY_print_private(),
8811 EVP_PKEY_print_param() to print public key data from an EVP_PKEY
8816 * Initial support for pluggable public key ASN1.
8817 De-spaghettify the public key ASN1 handling. Move public and private
8818 key ASN1 handling to a new EVP_PKEY_ASN1_METHOD structure. Relocate
8819 algorithm specific handling to a single module within the relevant
8820 algorithm directory. Add functions to allow (near) opaque processing
8821 of public and private key structures.
8825 * Implement the Supported Point Formats Extension for
8826 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
8830 * Add initial support for RFC 4279 PSK TLS ciphersuites. Add members
8831 for the psk identity [hint] and the psk callback functions to the
8832 SSL_SESSION, SSL and SSL_CTX structure.
8835 PSK-RC4-SHA, PSK-3DES-EDE-CBC-SHA, PSK-AES128-CBC-SHA,
8839 SSL_CTX_use_psk_identity_hint
8840 SSL_get_psk_identity_hint
8841 SSL_get_psk_identity
8842 SSL_use_psk_identity_hint
8844 *Mika Kousa and Pasi Eronen of Nokia Corporation*
8846 * Add RFC 3161 compliant time stamp request creation, response generation
8847 and response verification functionality.
8849 *Zoltán Glózik <zglozik@opentsa.org>, The OpenTSA Project*
8851 * Add initial support for TLS extensions, specifically for the server_name
8852 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
8853 have new members for a host name. The SSL data structure has an
8854 additional member `SSL_CTX *initial_ctx` so that new sessions can be
8855 stored in that context to allow for session resumption, even after the
8856 SSL has been switched to a new SSL_CTX in reaction to a client's
8857 server_name extension.
8859 New functions (subject to change):
8861 SSL_get_servername()
8862 SSL_get_servername_type()
8865 New CTRL codes and macros (subject to change):
8867 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
8868 - SSL_CTX_set_tlsext_servername_callback()
8869 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
8870 - SSL_CTX_set_tlsext_servername_arg()
8871 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
8873 openssl s_client has a new '-servername ...' option.
8875 openssl s_server has new options '-servername_host ...', '-cert2 ...',
8876 '-key2 ...', '-servername_fatal' (subject to change). This allows
8877 testing the HostName extension for a specific single host name ('-cert'
8878 and '-key' remain fallbacks for handshakes without HostName
8879 negotiation). If the unrecognized_name alert has to be sent, this by
8880 default is a warning; it becomes fatal with the '-servername_fatal'
8883 *Peter Sylvester, Remy Allais, Christophe Renou*
8885 * Whirlpool hash implementation is added.
8889 * BIGNUM code on 64-bit SPARCv9 targets is switched from bn(64,64) to
8890 bn(64,32). Because of instruction set limitations it doesn't have
8891 any negative impact on performance. This was done mostly in order
8892 to make it possible to share assembler modules, such as bn_mul_mont
8893 implementations, between 32- and 64-bit builds without hassle.
8897 * Move code previously exiled into file crypto/ec/ec2_smpt.c
8898 to ec2_smpl.c, and no longer require the OPENSSL_EC_BIN_PT_COMP
8903 * New candidate for BIGNUM assembler implementation, bn_mul_mont,
8904 dedicated Montgomery multiplication procedure, is introduced.
8905 BN_MONT_CTX is modified to allow bn_mul_mont to reach for higher
8906 "64-bit" performance on certain 32-bit targets.
8910 * New option SSL_OP_NO_COMP to disable use of compression selectively
8911 in SSL structures. New SSL ctrl to set maximum send fragment size.
8912 Save memory by setting the I/O buffer sizes dynamically instead of
8913 using the maximum available value.
8917 * New option -V for 'openssl ciphers'. This prints the ciphersuite code
8918 in addition to the text details.
8922 * Very, very preliminary EXPERIMENTAL support for printing of general
8923 ASN1 structures. This currently produces rather ugly output and doesn't
8924 handle several customised structures at all.
8928 * Integrated support for PVK file format and some related formats such
8929 as MS PUBLICKEYBLOB and PRIVATEKEYBLOB. Command line switches to support
8930 these in the 'rsa' and 'dsa' utilities.
8934 * Support for PKCS#1 RSAPublicKey format on rsa utility command line.
8938 * Remove the ancient ASN1_METHOD code. This was only ever used in one
8939 place for the (very old) "NETSCAPE" format certificates which are now
8940 handled using new ASN1 code equivalents.
8944 * Let the TLSv1_method() etc. functions return a 'const' SSL_METHOD
8945 pointer and make the SSL_METHOD parameter in SSL_CTX_new,
8946 SSL_CTX_set_ssl_version and SSL_set_ssl_method 'const'.
8950 * Modify CRL distribution points extension code to print out previously
8951 unsupported fields. Enhance extension setting code to allow setting of
8956 * Add print and set support for Issuing Distribution Point CRL extension.
8960 * Change 'Configure' script to enable Camellia by default.
8967 ### Changes between 0.9.8m and 0.9.8n [24 Mar 2010]
8969 * When rejecting SSL/TLS records due to an incorrect version number, never
8970 update s->server with a new major version number. As of
8971 - OpenSSL 0.9.8m if 'short' is a 16-bit type,
8972 - OpenSSL 0.9.8f if 'short' is longer than 16 bits,
8973 the previous behavior could result in a read attempt at NULL when
8974 receiving specific incorrect SSL/TLS records once record payload
8975 protection is active. ([CVE-2010-0740])
8977 *Bodo Moeller, Adam Langley <agl@chromium.org>*
8979 * Fix for CVE-2010-0433 where some kerberos enabled versions of OpenSSL
8980 could be crashed if the relevant tables were not present (e.g. chrooted).
8982 *Tomas Hoger <thoger@redhat.com>*
8984 ### Changes between 0.9.8l and 0.9.8m [25 Feb 2010]
8986 * Always check bn_wexpand() return values for failure. ([CVE-2009-3245])
8988 *Martin Olsson, Neel Mehta*
8990 * Fix X509_STORE locking: Every 'objs' access requires a lock (to
8991 accommodate for stack sorting, always a write lock!).
8995 * On some versions of WIN32 Heap32Next is very slow. This can cause
8996 excessive delays in the RAND_poll(): over a minute. As a workaround
8997 include a time check in the inner Heap32Next loop too.
9001 * The code that handled flushing of data in SSL/TLS originally used the
9002 BIO_CTRL_INFO ctrl to see if any data was pending first. This caused
9003 the problem outlined in PR#1949. The fix suggested there however can
9004 trigger problems with buggy BIO_CTRL_WPENDING (e.g. some versions
9005 of Apache). So instead simplify the code to flush unconditionally.
9006 This should be fine since flushing with no data to flush is a no op.
9010 * Handle TLS versions 2.0 and later properly and correctly use the
9011 highest version of TLS/SSL supported. Although TLS >= 2.0 is some way
9012 off ancient servers have a habit of sticking around for a while...
9016 * Modify compression code so it frees up structures without using the
9017 ex_data callbacks. This works around a problem where some applications
9018 call CRYPTO_cleanup_all_ex_data() before application exit (e.g. when
9019 restarting) then use compression (e.g. SSL with compression) later.
9020 This results in significant per-connection memory leaks and
9021 has caused some security issues including CVE-2008-1678 and
9026 * Constify crypto/cast (i.e., <openssl/cast.h>): a CAST_KEY doesn't
9027 change when encrypting or decrypting.
9031 * Add option SSL_OP_LEGACY_SERVER_CONNECT which will allow clients to
9032 connect and renegotiate with servers which do not support RI.
9033 Until RI is more widely deployed this option is enabled by default.
9037 * Add "missing" ssl ctrls to clear options and mode.
9041 * If client attempts to renegotiate and doesn't support RI respond with
9042 a no_renegotiation alert as required by RFC5746. Some renegotiating
9043 TLS clients will continue a connection gracefully when they receive
9044 the alert. Unfortunately OpenSSL mishandled this alert and would hang
9045 waiting for a server hello which it will never receive. Now we treat a
9046 received no_renegotiation alert as a fatal error. This is because
9047 applications requesting a renegotiation might well expect it to succeed
9048 and would have no code in place to handle the server denying it so the
9049 only safe thing to do is to terminate the connection.
9053 * Add ctrl macro SSL_get_secure_renegotiation_support() which returns 1 if
9054 peer supports secure renegotiation and 0 otherwise. Print out peer
9055 renegotiation support in s_client/s_server.
9059 * Replace the highly broken and deprecated SPKAC certification method with
9060 the updated NID creation version. This should correctly handle UTF8.
9064 * Implement RFC5746. Re-enable renegotiation but require the extension
9065 as needed. Unfortunately, SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION
9066 turns out to be a bad idea. It has been replaced by
9067 SSL_OP_ALLOW_UNSAFE_LEGACY_RENEGOTIATION which can be set with
9068 SSL_CTX_set_options(). This is really not recommended unless you
9069 know what you are doing.
9071 *Eric Rescorla <ekr@networkresonance.com>, Ben Laurie, Steve Henson*
9073 * Fixes to stateless session resumption handling. Use initial_ctx when
9074 issuing and attempting to decrypt tickets in case it has changed during
9075 servername handling. Use a non-zero length session ID when attempting
9076 stateless session resumption: this makes it possible to determine if
9077 a resumption has occurred immediately after receiving server hello
9078 (several places in OpenSSL subtly assume this) instead of later in
9083 * The functions ENGINE_ctrl(), OPENSSL_isservice(),
9084 CMS_get1_RecipientRequest() and RAND_bytes() can return <=0 on error
9085 fixes for a few places where the return code is not checked
9088 *Julia Lawall <julia@diku.dk>*
9090 * Add --strict-warnings option to Configure script to include devteam
9091 warnings in other configurations.
9095 * Add support for --libdir option and LIBDIR variable in makefiles. This
9096 makes it possible to install openssl libraries in locations which
9097 have names other than "lib", for example "/usr/lib64" which some
9100 *Steve Henson, based on patch from Jeremy Utley*
9102 * Don't allow the use of leading 0x80 in OIDs. This is a violation of
9103 X690 8.9.12 and can produce some misleading textual output of OIDs.
9105 *Steve Henson, reported by Dan Kaminsky*
9107 * Delete MD2 from algorithm tables. This follows the recommendation in
9108 several standards that it is not used in new applications due to
9109 several cryptographic weaknesses. For binary compatibility reasons
9110 the MD2 API is still compiled in by default.
9114 * Add compression id to {d2i,i2d}_SSL_SESSION so it is correctly saved
9119 * Rename uni2asc and asc2uni functions to OPENSSL_uni2asc and
9120 OPENSSL_asc2uni conditionally on Netware platforms to avoid a name
9123 *Guenter <lists@gknw.net>*
9125 * Fix the server certificate chain building code to use X509_verify_cert(),
9126 it used to have an ad-hoc builder which was unable to cope with anything
9127 other than a simple chain.
9129 *David Woodhouse <dwmw2@infradead.org>, Steve Henson*
9131 * Don't check self signed certificate signatures in X509_verify_cert()
9132 by default (a flag can override this): it just wastes time without
9133 adding any security. As a useful side effect self signed root CAs
9134 with non-FIPS digests are now usable in FIPS mode.
9138 * In dtls1_process_out_of_seq_message() the check if the current message
9139 is already buffered was missing. For every new message was memory
9140 allocated, allowing an attacker to perform an denial of service attack
9141 with sending out of seq handshake messages until there is no memory
9142 left. Additionally every future message was buffered, even if the
9143 sequence number made no sense and would be part of another handshake.
9144 So only messages with sequence numbers less than 10 in advance will be
9145 buffered. ([CVE-2009-1378])
9147 *Robin Seggelmann, discovered by Daniel Mentz*
9149 * Records are buffered if they arrive with a future epoch to be
9150 processed after finishing the corresponding handshake. There is
9151 currently no limitation to this buffer allowing an attacker to perform
9152 a DOS attack with sending records with future epochs until there is no
9153 memory left. This patch adds the pqueue_size() function to determine
9154 the size of a buffer and limits the record buffer to 100 entries.
9157 *Robin Seggelmann, discovered by Daniel Mentz*
9159 * Keep a copy of frag->msg_header.frag_len so it can be used after the
9160 parent structure is freed. ([CVE-2009-1379])
9164 * Handle non-blocking I/O properly in SSL_shutdown() call.
9166 *Darryl Miles <darryl-mailinglists@netbauds.net>*
9168 * Add `2.5.4.*` OIDs
9170 *Ilya O. <vrghost@gmail.com>*
9172 ### Changes between 0.9.8k and 0.9.8l [5 Nov 2009]
9174 * Disable renegotiation completely - this fixes a severe security
9175 problem ([CVE-2009-3555]) at the cost of breaking all
9176 renegotiation. Renegotiation can be re-enabled by setting
9177 SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION in s3->flags at
9178 run-time. This is really not recommended unless you know what
9183 ### Changes between 0.9.8j and 0.9.8k [25 Mar 2009]
9185 * Don't set val to NULL when freeing up structures, it is freed up by
9186 underlying code. If `sizeof(void *) > sizeof(long)` this can result in
9187 zeroing past the valid field. ([CVE-2009-0789])
9189 *Paolo Ganci <Paolo.Ganci@AdNovum.CH>*
9191 * Fix bug where return value of CMS_SignerInfo_verify_content() was not
9192 checked correctly. This would allow some invalid signed attributes to
9193 appear to verify correctly. ([CVE-2009-0591])
9195 *Ivan Nestlerode <inestlerode@us.ibm.com>*
9197 * Reject UniversalString and BMPString types with invalid lengths. This
9198 prevents a crash in ASN1_STRING_print_ex() which assumes the strings have
9199 a legal length. ([CVE-2009-0590])
9203 * Set S/MIME signing as the default purpose rather than setting it
9204 unconditionally. This allows applications to override it at the store
9209 * Permit restricted recursion of ASN1 strings. This is needed in practice
9210 to handle some structures.
9214 * Improve efficiency of mem_gets: don't search whole buffer each time
9217 *Jeremy Shapiro <jnshapir@us.ibm.com>*
9219 * New -hex option for openssl rand.
9223 * Print out UTF8String and NumericString when parsing ASN1.
9227 * Support NumericString type for name components.
9231 * Allow CC in the environment to override the automatically chosen
9232 compiler. Note that nothing is done to ensure flags work with the
9237 ### Changes between 0.9.8i and 0.9.8j [07 Jan 2009]
9239 * Properly check EVP_VerifyFinal() and similar return values
9242 *Ben Laurie, Bodo Moeller, Google Security Team*
9244 * Enable TLS extensions by default.
9248 * Allow the CHIL engine to be loaded, whether the application is
9249 multithreaded or not. (This does not release the developer from the
9250 obligation to set up the dynamic locking callbacks.)
9252 *Sander Temme <sander@temme.net>*
9254 * Use correct exit code if there is an error in dgst command.
9256 *Steve Henson; problem pointed out by Roland Dirlewanger*
9258 * Tweak Configure so that you need to say "experimental-jpake" to enable
9259 JPAKE, and need to use -DOPENSSL_EXPERIMENTAL_JPAKE in applications.
9263 * Add experimental JPAKE support, including demo authentication in
9264 s_client and s_server.
9268 * Set the comparison function in v3_addr_canonize().
9270 *Rob Austein <sra@hactrn.net>*
9272 * Add support for XMPP STARTTLS in s_client.
9274 *Philip Paeps <philip@freebsd.org>*
9276 * Change the server-side SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG behavior
9277 to ensure that even with this option, only ciphersuites in the
9278 server's preference list will be accepted. (Note that the option
9279 applies only when resuming a session, so the earlier behavior was
9280 just about the algorithm choice for symmetric cryptography.)
9284 ### Changes between 0.9.8h and 0.9.8i [15 Sep 2008]
9286 * Fix NULL pointer dereference if a DTLS server received
9287 ChangeCipherSpec as first record ([CVE-2009-1386]).
9291 * Fix a state transition in s3_srvr.c and d1_srvr.c
9292 (was using SSL3_ST_CW_CLNT_HELLO_B, should be `..._ST_SW_SRVR_...`).
9296 * The fix in 0.9.8c that supposedly got rid of unsafe
9297 double-checked locking was incomplete for RSA blinding,
9298 addressing just one layer of what turns out to have been
9299 doubly unsafe triple-checked locking.
9301 So now fix this for real by retiring the MONT_HELPER macro
9302 in crypto/rsa/rsa_eay.c.
9304 *Bodo Moeller; problem pointed out by Marius Schilder*
9306 * Various precautionary measures:
9308 - Avoid size_t integer overflow in HASH_UPDATE (md32_common.h).
9310 - Avoid a buffer overflow in d2i_SSL_SESSION() (ssl_asn1.c).
9311 (NB: This would require knowledge of the secret session ticket key
9312 to exploit, in which case you'd be SOL either way.)
9314 - Change bn_nist.c so that it will properly handle input BIGNUMs
9315 outside the expected range.
9317 - Enforce the 'num' check in BN_div() (bn_div.c) for non-BN_DEBUG
9320 *Neel Mehta, Bodo Moeller*
9322 * Allow engines to be "soft loaded" - i.e. optionally don't die if
9323 the load fails. Useful for distros.
9325 *Ben Laurie and the FreeBSD team*
9327 * Add support for Local Machine Keyset attribute in PKCS#12 files.
9331 * Fix BN_GF2m_mod_arr() top-bit cleanup code.
9335 * Expand ENGINE to support engine supplied SSL client certificate functions.
9337 This work was sponsored by Logica.
9341 * Add CryptoAPI ENGINE to support use of RSA and DSA keys held in Windows
9342 keystores. Support for SSL/TLS client authentication too.
9343 Not compiled unless enable-capieng specified to Configure.
9345 This work was sponsored by Logica.
9349 * Fix bug in X509_ATTRIBUTE creation: don't set attribute using
9350 ASN1_TYPE_set1 if MBSTRING flag set. This bug would crash certain
9351 attribute creation routines such as certificate requests and PKCS#12
9356 ### Changes between 0.9.8g and 0.9.8h [28 May 2008]
9358 * Fix flaw if 'Server Key exchange message' is omitted from a TLS
9359 handshake which could lead to a client crash as found using the
9360 Codenomicon TLS test suite ([CVE-2008-1672])
9362 *Steve Henson, Mark Cox*
9364 * Fix double free in TLS server name extensions which could lead to
9365 a remote crash found by Codenomicon TLS test suite ([CVE-2008-0891])
9369 * Clear error queue in SSL_CTX_use_certificate_chain_file()
9371 Clear the error queue to ensure that error entries left from
9372 older function calls do not interfere with the correct operation.
9374 *Lutz Jaenicke, Erik de Castro Lopo*
9376 * Remove root CA certificates of commercial CAs:
9378 The OpenSSL project does not recommend any specific CA and does not
9379 have any policy with respect to including or excluding any CA.
9380 Therefore it does not make any sense to ship an arbitrary selection
9381 of root CA certificates with the OpenSSL software.
9385 * RSA OAEP patches to fix two separate invalid memory reads.
9386 The first one involves inputs when 'lzero' is greater than
9387 'SHA_DIGEST_LENGTH' (it would read about SHA_DIGEST_LENGTH bytes
9388 before the beginning of from). The second one involves inputs where
9389 the 'db' section contains nothing but zeroes (there is a one-byte
9390 invalid read after the end of 'db').
9392 *Ivan Nestlerode <inestlerode@us.ibm.com>*
9394 * Partial backport from 0.9.9-dev:
9396 Introduce bn_mul_mont (dedicated Montgomery multiplication
9397 procedure) as a candidate for BIGNUM assembler implementation.
9398 While 0.9.9-dev uses assembler for various architectures, only
9399 x86_64 is available by default here in the 0.9.8 branch, and
9400 32-bit x86 is available through a compile-time setting.
9402 To try the 32-bit x86 assembler implementation, use Configure
9403 option "enable-montasm" (which exists only for this backport).
9405 As "enable-montasm" for 32-bit x86 disclaims code stability
9406 anyway, in this constellation we activate additional code
9407 backported from 0.9.9-dev for further performance improvements,
9408 namely BN_from_montgomery_word. (To enable this otherwise,
9409 e.g. x86_64, try `-DMONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD`.)
9411 *Andy Polyakov (backport partially by Bodo Moeller)*
9413 * Add TLS session ticket callback. This allows an application to set
9414 TLS ticket cipher and HMAC keys rather than relying on hardcoded fixed
9415 values. This is useful for key rollover for example where several key
9416 sets may exist with different names.
9420 * Reverse ENGINE-internal logic for caching default ENGINE handles.
9421 This was broken until now in 0.9.8 releases, such that the only way
9422 a registered ENGINE could be used (assuming it initialises
9423 successfully on the host) was to explicitly set it as the default
9424 for the relevant algorithms. This is in contradiction with 0.9.7
9425 behaviour and the documentation. With this fix, when an ENGINE is
9426 registered into a given algorithm's table of implementations, the
9427 'uptodate' flag is reset so that auto-discovery will be used next
9428 time a new context for that algorithm attempts to select an
9431 *Ian Lister (tweaked by Geoff Thorpe)*
9433 * Backport of CMS code to OpenSSL 0.9.8. This differs from the 0.9.9
9434 implementation in the following ways:
9436 Lack of EVP_PKEY_ASN1_METHOD means algorithm parameters have to be
9439 Lack of BER streaming support means one pass streaming processing is
9440 only supported if data is detached: setting the streaming flag is
9441 ignored for embedded content.
9443 CMS support is disabled by default and must be explicitly enabled
9444 with the enable-cms configuration option.
9448 * Update the GMP engine glue to do direct copies between BIGNUM and
9449 mpz_t when openssl and GMP use the same limb size. Otherwise the
9450 existing "conversion via a text string export" trick is still used.
9452 *Paul Sheer <paulsheer@gmail.com>*
9454 * Zlib compression BIO. This is a filter BIO which compressed and
9455 uncompresses any data passed through it.
9459 * Add AES_wrap_key() and AES_unwrap_key() functions to implement
9460 RFC3394 compatible AES key wrapping.
9464 * Add utility functions to handle ASN1 structures. ASN1_STRING_set0():
9465 sets string data without copying. X509_ALGOR_set0() and
9466 X509_ALGOR_get0(): set and retrieve X509_ALGOR (AlgorithmIdentifier)
9467 data. Attribute function X509at_get0_data_by_OBJ(): retrieves data
9468 from an X509_ATTRIBUTE structure optionally checking it occurs only
9469 once. ASN1_TYPE_set1(): set and ASN1_TYPE structure copying supplied
9474 * Fix BN flag handling in RSA_eay_mod_exp() and BN_MONT_CTX_set()
9475 to get the expected BN_FLG_CONSTTIME behavior.
9477 *Bodo Moeller (Google)*
9481 - fixed wrong usage of ioctlsocket() when build for LIBC BSD sockets
9482 - fixed do_tests.pl to run the test suite with CLIB builds too (CLIB_OPT)
9483 - added some more tests to do_tests.pl
9484 - fixed RunningProcess usage so that it works with newer LIBC NDKs too
9485 - removed usage of BN_LLONG for CLIB builds to avoid runtime dependency
9486 - added new Configure targets netware-clib-bsdsock, netware-clib-gcc,
9487 netware-clib-bsdsock-gcc, netware-libc-bsdsock-gcc
9488 - various changes to netware.pl to enable gcc-cross builds on Win32
9490 - changed crypto/bio/b_sock.c to work with macro functions (CLIB BSD)
9491 - various changes to fix missing prototype warnings
9492 - fixed x86nasm.pl to create correct asm files for NASM COFF output
9493 - added AES, WHIRLPOOL and CPUID assembler code to build files
9494 - added missing AES assembler make rules to mk1mf.pl
9495 - fixed order of includes in `apps/ocsp.c` so that `e_os.h` settings apply
9497 *Guenter Knauf <eflash@gmx.net>*
9499 * Implement certificate status request TLS extension defined in RFC3546.
9500 A client can set the appropriate parameters and receive the encoded
9501 OCSP response via a callback. A server can query the supplied parameters
9502 and set the encoded OCSP response in the callback. Add simplified examples
9503 to s_client and s_server.
9507 ### Changes between 0.9.8f and 0.9.8g [19 Oct 2007]
9510 + Binary incompatibility of ssl_ctx_st structure
9511 + DTLS interoperation with non-compliant servers
9512 + Don't call get_session_cb() without proposed session
9513 + Fix ia64 assembler code
9515 *Andy Polyakov, Steve Henson*
9517 ### Changes between 0.9.8e and 0.9.8f [11 Oct 2007]
9519 * DTLS Handshake overhaul. There were longstanding issues with
9520 OpenSSL DTLS implementation, which were making it impossible for
9521 RFC 4347 compliant client to communicate with OpenSSL server.
9522 Unfortunately just fixing these incompatibilities would "cut off"
9523 pre-0.9.8f clients. To allow for hassle free upgrade post-0.9.8e
9524 server keeps tolerating non RFC compliant syntax. The opposite is
9525 not true, 0.9.8f client can not communicate with earlier server.
9526 This update even addresses CVE-2007-4995.
9530 * Changes to avoid need for function casts in OpenSSL: some compilers
9531 (gcc 4.2 and later) reject their use.
9532 *Kurt Roeckx <kurt@roeckx.be>, Peter Hartley <pdh@utter.chaos.org.uk>,
9535 * Add RFC4507 support to OpenSSL. This includes the corrections in
9536 RFC4507bis. The encrypted ticket format is an encrypted encoded
9537 SSL_SESSION structure, that way new session features are automatically
9540 If a client application caches session in an SSL_SESSION structure
9541 support is transparent because tickets are now stored in the encoded
9544 The SSL_CTX structure automatically generates keys for ticket
9545 protection in servers so again support should be possible
9546 with no application modification.
9548 If a client or server wishes to disable RFC4507 support then the option
9549 SSL_OP_NO_TICKET can be set.
9551 Add a TLS extension debugging callback to allow the contents of any client
9552 or server extensions to be examined.
9554 This work was sponsored by Google.
9558 * Add initial support for TLS extensions, specifically for the server_name
9559 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
9560 have new members for a host name. The SSL data structure has an
9561 additional member `SSL_CTX *initial_ctx` so that new sessions can be
9562 stored in that context to allow for session resumption, even after the
9563 SSL has been switched to a new SSL_CTX in reaction to a client's
9564 server_name extension.
9566 New functions (subject to change):
9568 SSL_get_servername()
9569 SSL_get_servername_type()
9572 New CTRL codes and macros (subject to change):
9574 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
9575 - SSL_CTX_set_tlsext_servername_callback()
9576 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
9577 - SSL_CTX_set_tlsext_servername_arg()
9578 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
9580 openssl s_client has a new '-servername ...' option.
9582 openssl s_server has new options '-servername_host ...', '-cert2 ...',
9583 '-key2 ...', '-servername_fatal' (subject to change). This allows
9584 testing the HostName extension for a specific single host name ('-cert'
9585 and '-key' remain fallbacks for handshakes without HostName
9586 negotiation). If the unrecognized_name alert has to be sent, this by
9587 default is a warning; it becomes fatal with the '-servername_fatal'
9590 *Peter Sylvester, Remy Allais, Christophe Renou, Steve Henson*
9592 * Add AES and SSE2 assembly language support to VC++ build.
9596 * Mitigate attack on final subtraction in Montgomery reduction.
9600 * Fix crypto/ec/ec_mult.c to work properly with scalars of value 0
9601 (which previously caused an internal error).
9605 * Squeeze another 10% out of IGE mode when in != out.
9609 * AES IGE mode speedup.
9611 *Dean Gaudet (Google)*
9613 * Add the Korean symmetric 128-bit cipher SEED (see
9614 <http://www.kisa.or.kr/kisa/seed/jsp/seed_eng.jsp>) and
9615 add SEED ciphersuites from RFC 4162:
9617 TLS_RSA_WITH_SEED_CBC_SHA = "SEED-SHA"
9618 TLS_DHE_DSS_WITH_SEED_CBC_SHA = "DHE-DSS-SEED-SHA"
9619 TLS_DHE_RSA_WITH_SEED_CBC_SHA = "DHE-RSA-SEED-SHA"
9620 TLS_DH_anon_WITH_SEED_CBC_SHA = "ADH-SEED-SHA"
9622 To minimize changes between patchlevels in the OpenSSL 0.9.8
9623 series, SEED remains excluded from compilation unless OpenSSL
9624 is configured with 'enable-seed'.
9626 *KISA, Bodo Moeller*
9628 * Mitigate branch prediction attacks, which can be practical if a
9629 single processor is shared, allowing a spy process to extract
9630 information. For detailed background information, see
9631 <http://eprint.iacr.org/2007/039> (O. Aciicmez, S. Gueron,
9632 J.-P. Seifert, "New Branch Prediction Vulnerabilities in OpenSSL
9633 and Necessary Software Countermeasures"). The core of the change
9634 are new versions BN_div_no_branch() and
9635 BN_mod_inverse_no_branch() of BN_div() and BN_mod_inverse(),
9636 respectively, which are slower, but avoid the security-relevant
9637 conditional branches. These are automatically called by BN_div()
9638 and BN_mod_inverse() if the flag BN_FLG_CONSTTIME is set for one
9639 of the input BIGNUMs. Also, BN_is_bit_set() has been changed to
9640 remove a conditional branch.
9642 BN_FLG_CONSTTIME is the new name for the previous
9643 BN_FLG_EXP_CONSTTIME flag, since it now affects more than just
9644 modular exponentiation. (Since OpenSSL 0.9.7h, setting this flag
9645 in the exponent causes BN_mod_exp_mont() to use the alternative
9646 implementation in BN_mod_exp_mont_consttime().) The old name
9647 remains as a deprecated alias.
9649 Similarly, RSA_FLAG_NO_EXP_CONSTTIME is replaced by a more general
9650 RSA_FLAG_NO_CONSTTIME flag since the RSA implementation now uses
9651 constant-time implementations for more than just exponentiation.
9652 Here too the old name is kept as a deprecated alias.
9654 BN_BLINDING_new() will now use BN_dup() for the modulus so that
9655 the BN_BLINDING structure gets an independent copy of the
9656 modulus. This means that the previous `BIGNUM *m` argument to
9657 BN_BLINDING_new() and to BN_BLINDING_create_param() now
9658 essentially becomes `const BIGNUM *m`, although we can't actually
9659 change this in the header file before 0.9.9. It allows
9660 RSA_setup_blinding() to use BN_with_flags() on the modulus to
9661 enable BN_FLG_CONSTTIME.
9663 *Matthew D Wood (Intel Corp)*
9665 * In the SSL/TLS server implementation, be strict about session ID
9666 context matching (which matters if an application uses a single
9667 external cache for different purposes). Previously,
9668 out-of-context reuse was forbidden only if SSL_VERIFY_PEER was
9669 set. This did ensure strict client verification, but meant that,
9670 with applications using a single external cache for quite
9671 different requirements, clients could circumvent ciphersuite
9672 restrictions for a given session ID context by starting a session
9673 in a different context.
9677 * Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
9678 a ciphersuite string such as "DEFAULT:RSA" cannot enable
9679 authentication-only ciphersuites.
9683 * Update the SSL_get_shared_ciphers() fix CVE-2006-3738 which was
9684 not complete and could lead to a possible single byte overflow
9685 ([CVE-2007-5135]) [Ben Laurie]
9687 ### Changes between 0.9.8d and 0.9.8e [23 Feb 2007]
9689 * Since AES128 and AES256 (and similarly Camellia128 and
9690 Camellia256) share a single mask bit in the logic of
9691 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
9692 kludge to work properly if AES128 is available and AES256 isn't
9693 (or if Camellia128 is available and Camellia256 isn't).
9697 * Fix the BIT STRING encoding generated by crypto/ec/ec_asn1.c
9698 (within i2d_ECPrivateKey, i2d_ECPKParameters, i2d_ECParameters):
9699 When a point or a seed is encoded in a BIT STRING, we need to
9700 prevent the removal of trailing zero bits to get the proper DER
9701 encoding. (By default, crypto/asn1/a_bitstr.c assumes the case
9702 of a NamedBitList, for which trailing 0 bits need to be removed.)
9706 * Have SSL/TLS server implementation tolerate "mismatched" record
9707 protocol version while receiving ClientHello even if the
9708 ClientHello is fragmented. (The server can't insist on the
9709 particular protocol version it has chosen before the ServerHello
9710 message has informed the client about his choice.)
9714 * Add RFC 3779 support.
9716 *Rob Austein for ARIN, Ben Laurie*
9718 * Load error codes if they are not already present instead of using a
9719 static variable. This allows them to be cleanly unloaded and reloaded.
9720 Improve header file function name parsing.
9724 * extend SMTP and IMAP protocol emulation in s_client to use EHLO
9725 or CAPABILITY handshake as required by RFCs.
9729 ### Changes between 0.9.8c and 0.9.8d [28 Sep 2006]
9731 * Introduce limits to prevent malicious keys being able to
9732 cause a denial of service. ([CVE-2006-2940])
9734 *Steve Henson, Bodo Moeller*
9736 * Fix ASN.1 parsing of certain invalid structures that can result
9737 in a denial of service. ([CVE-2006-2937]) [Steve Henson]
9739 * Fix buffer overflow in SSL_get_shared_ciphers() function.
9740 ([CVE-2006-3738]) [Tavis Ormandy and Will Drewry, Google Security Team]
9742 * Fix SSL client code which could crash if connecting to a
9743 malicious SSLv2 server. ([CVE-2006-4343])
9745 *Tavis Ormandy and Will Drewry, Google Security Team*
9747 * Since 0.9.8b, ciphersuite strings naming explicit ciphersuites
9748 match only those. Before that, "AES256-SHA" would be interpreted
9749 as a pattern and match "AES128-SHA" too (since AES128-SHA got
9750 the same strength classification in 0.9.7h) as we currently only
9751 have a single AES bit in the ciphersuite description bitmap.
9752 That change, however, also applied to ciphersuite strings such as
9753 "RC4-MD5" that intentionally matched multiple ciphersuites --
9754 namely, SSL 2.0 ciphersuites in addition to the more common ones
9755 from SSL 3.0/TLS 1.0.
9757 So we change the selection algorithm again: Naming an explicit
9758 ciphersuite selects this one ciphersuite, and any other similar
9759 ciphersuite (same bitmap) from *other* protocol versions.
9760 Thus, "RC4-MD5" again will properly select both the SSL 2.0
9761 ciphersuite and the SSL 3.0/TLS 1.0 ciphersuite.
9763 Since SSL 2.0 does not have any ciphersuites for which the
9764 128/256 bit distinction would be relevant, this works for now.
9765 The proper fix will be to use different bits for AES128 and
9766 AES256, which would have avoided the problems from the beginning;
9767 however, bits are scarce, so we can only do this in a new release
9768 (not just a patchlevel) when we can change the SSL_CIPHER
9769 definition to split the single 'unsigned long mask' bitmap into
9770 multiple values to extend the available space.
9774 ### Changes between 0.9.8b and 0.9.8c [05 Sep 2006]
9776 * Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
9777 ([CVE-2006-4339]) [Ben Laurie and Google Security Team]
9779 * Add AES IGE and biIGE modes.
9783 * Change the Unix randomness entropy gathering to use poll() when
9784 possible instead of select(), since the latter has some
9785 undesirable limitations.
9787 *Darryl Miles via Richard Levitte and Bodo Moeller*
9789 * Disable "ECCdraft" ciphersuites more thoroughly. Now special
9790 treatment in ssl/ssl_ciph.s makes sure that these ciphersuites
9791 cannot be implicitly activated as part of, e.g., the "AES" alias.
9792 However, please upgrade to OpenSSL 0.9.9[-dev] for
9793 non-experimental use of the ECC ciphersuites to get TLS extension
9794 support, which is required for curve and point format negotiation
9795 to avoid potential handshake problems.
9799 * Disable rogue ciphersuites:
9801 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
9802 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
9803 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
9805 The latter two were purportedly from
9806 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
9809 Also deactivate the remaining ciphersuites from
9810 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
9811 unofficial, and the ID has long expired.
9815 * Fix RSA blinding Heisenbug (problems sometimes occurred on
9816 dual-core machines) and other potential thread-safety issues.
9820 * Add the symmetric cipher Camellia (128-bit, 192-bit, 256-bit key
9821 versions), which is now available for royalty-free use
9822 (see <http://info.isl.ntt.co.jp/crypt/eng/info/chiteki.html>).
9823 Also, add Camellia TLS ciphersuites from RFC 4132.
9825 To minimize changes between patchlevels in the OpenSSL 0.9.8
9826 series, Camellia remains excluded from compilation unless OpenSSL
9827 is configured with 'enable-camellia'.
9831 * Disable the padding bug check when compression is in use. The padding
9832 bug check assumes the first packet is of even length, this is not
9833 necessarily true if compression is enabled and can result in false
9834 positives causing handshake failure. The actual bug test is ancient
9835 code so it is hoped that implementations will either have fixed it by
9836 now or any which still have the bug do not support compression.
9840 ### Changes between 0.9.8a and 0.9.8b [04 May 2006]
9842 * When applying a cipher rule check to see if string match is an explicit
9843 cipher suite and only match that one cipher suite if it is.
9847 * Link in manifests for VC++ if needed.
9849 *Austin Ziegler <halostatue@gmail.com>*
9851 * Update support for ECC-based TLS ciphersuites according to
9852 draft-ietf-tls-ecc-12.txt with proposed changes (but without
9853 TLS extensions, which are supported starting with the 0.9.9
9854 branch, not in the OpenSSL 0.9.8 branch).
9858 * New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support
9859 opaque EVP_CIPHER_CTX handling.
9863 * Fixes and enhancements to zlib compression code. We now only use
9864 "zlib1.dll" and use the default `__cdecl` calling convention on Win32
9865 to conform with the standards mentioned here:
9866 <http://www.zlib.net/DLL_FAQ.txt>
9867 Static zlib linking now works on Windows and the new --with-zlib-include
9868 --with-zlib-lib options to Configure can be used to supply the location
9869 of the headers and library. Gracefully handle case where zlib library
9874 * Several fixes and enhancements to the OID generation code. The old code
9875 sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't
9876 handle numbers larger than ULONG_MAX, truncated printing and had a
9877 non standard OBJ_obj2txt() behaviour.
9881 * Add support for building of engines under engine/ as shared libraries
9882 under VC++ build system.
9886 * Corrected the numerous bugs in the Win32 path splitter in DSO.
9887 Hopefully, we will not see any false combination of paths any more.
9891 ### Changes between 0.9.8 and 0.9.8a [11 Oct 2005]
9893 * Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
9894 (part of SSL_OP_ALL). This option used to disable the
9895 countermeasure against man-in-the-middle protocol-version
9896 rollback in the SSL 2.0 server implementation, which is a bad
9897 idea. ([CVE-2005-2969])
9899 *Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
9900 for Information Security, National Institute of Advanced Industrial
9901 Science and Technology [AIST], Japan)*
9903 * Add two function to clear and return the verify parameter flags.
9907 * Keep cipherlists sorted in the source instead of sorting them at
9908 runtime, thus removing the need for a lock.
9912 * Avoid some small subgroup attacks in Diffie-Hellman.
9914 *Nick Mathewson and Ben Laurie*
9916 * Add functions for well-known primes.
9920 * Extended Windows CE support.
9922 *Satoshi Nakamura and Andy Polyakov*
9924 * Initialize SSL_METHOD structures at compile time instead of during
9925 runtime, thus removing the need for a lock.
9929 * Make PKCS7_decrypt() work even if no certificate is supplied by
9930 attempting to decrypt each encrypted key in turn. Add support to
9935 ### Changes between 0.9.7h and 0.9.8 [05 Jul 2005]
9937 [NB: OpenSSL 0.9.7i and later 0.9.7 patch levels were released after
9940 * Add libcrypto.pc and libssl.pc for those who feel they need them.
9944 * Change CA.sh and CA.pl so they don't bundle the CSR and the private
9945 key into the same file any more.
9949 * Add initial support for Win64, both IA64 and AMD64/x64 flavors.
9953 * Add -utf8 command line and config file option to 'ca'.
9955 *Stefan <stf@udoma.org*
9957 * Removed the macro des_crypt(), as it seems to conflict with some
9958 libraries. Use DES_crypt().
9962 * Correct naming of the 'chil' and '4758cca' ENGINEs. This
9963 involves renaming the source and generated shared-libs for
9964 both. The engines will accept the corrected or legacy ids
9965 ('ncipher' and '4758_cca' respectively) when binding. NB,
9966 this only applies when building 'shared'.
9968 *Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe*
9970 * Add attribute functions to EVP_PKEY structure. Modify
9971 PKCS12_create() to recognize a CSP name attribute and
9972 use it. Make -CSP option work again in pkcs12 utility.
9976 * Add new functionality to the bn blinding code:
9977 - automatic re-creation of the BN_BLINDING parameters after
9978 a fixed number of uses (currently 32)
9979 - add new function for parameter creation
9980 - introduce flags to control the update behaviour of the
9981 BN_BLINDING parameters
9982 - hide BN_BLINDING structure
9983 Add a second BN_BLINDING slot to the RSA structure to improve
9984 performance when a single RSA object is shared among several
9989 * Add support for DTLS.
9991 *Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie*
9993 * Add support for DER encoded private keys (SSL_FILETYPE_ASN1)
9994 to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file()
9998 * Remove buggy and incomplete DH cert support from
9999 ssl/ssl_rsa.c and ssl/s3_both.c
10003 * Use SHA-1 instead of MD5 as the default digest algorithm for
10004 the `apps/openssl` commands.
10008 * Compile clean with "-Wall -Wmissing-prototypes
10009 -Wstrict-prototypes -Wmissing-declarations -Werror". Currently
10010 DEBUG_SAFESTACK must also be set.
10014 * Change ./Configure so that certain algorithms can be disabled by default.
10015 The new counterpiece to "no-xxx" is "enable-xxx".
10017 The patented RC5 and MDC2 algorithms will now be disabled unless
10018 "enable-rc5" and "enable-mdc2", respectively, are specified.
10020 (IDEA remains enabled despite being patented. This is because IDEA
10021 is frequently required for interoperability, and there is no license
10022 fee for non-commercial use. As before, "no-idea" can be used to
10023 avoid this algorithm.)
10027 * Add processing of proxy certificates (see RFC 3820). This work was
10028 sponsored by KTH (The Royal Institute of Technology in Stockholm) and
10029 EGEE (Enabling Grids for E-science in Europe).
10033 * RC4 performance overhaul on modern architectures/implementations, such
10034 as Intel P4, IA-64 and AMD64.
10038 * New utility extract-section.pl. This can be used specify an alternative
10039 section number in a pod file instead of having to treat each file as
10040 a separate case in Makefile. This can be done by adding two lines to the
10043 =for comment openssl_section:XXX
10045 The blank line is mandatory.
10049 * New arguments -certform, -keyform and -pass for s_client and s_server
10050 to allow alternative format key and certificate files and passphrase
10055 * New structure X509_VERIFY_PARAM which combines current verify parameters,
10056 update associated structures and add various utility functions.
10058 Add new policy related verify parameters, include policy checking in
10059 standard verify code. Enhance 'smime' application with extra parameters
10060 to support policy checking and print out.
10064 * Add a new engine to support VIA PadLock ACE extensions in the VIA C3
10065 Nehemiah processors. These extensions support AES encryption in hardware
10066 as well as RNG (though RNG support is currently disabled).
10068 *Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov*
10070 * Deprecate `BN_[get|set]_params()` functions (they were ignored internally).
10074 * New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented.
10076 *Andy Polyakov and a number of other people*
10078 * Improved PowerPC platform support. Most notably BIGNUM assembler
10079 implementation contributed by IBM.
10081 *Suresh Chari, Peter Waltenberg, Andy Polyakov*
10083 * The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public
10084 exponent rather than 'unsigned long'. There is a corresponding change to
10085 the new 'rsa_keygen' element of the RSA_METHOD structure.
10087 *Jelte Jansen, Geoff Thorpe*
10089 * Functionality for creating the initial serial number file is now
10090 moved from CA.pl to the 'ca' utility with a new option -create_serial.
10092 (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial
10093 number file to 1, which is bound to cause problems. To avoid
10094 the problems while respecting compatibility between different 0.9.7
10095 patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in
10096 CA.pl for serial number initialization. With the new release 0.9.8,
10097 we can fix the problem directly in the 'ca' utility.)
10101 * Reduced header interdependencies by declaring more opaque objects in
10102 ossl_typ.h. As a consequence, including some headers (eg. engine.h) will
10103 give fewer recursive includes, which could break lazy source code - so
10104 this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always,
10105 developers should define this symbol when building and using openssl to
10106 ensure they track the recommended behaviour, interfaces, [etc], but
10107 backwards-compatible behaviour prevails when this isn't defined.
10111 * New function X509_POLICY_NODE_print() which prints out policy nodes.
10115 * Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality.
10116 This will generate a random key of the appropriate length based on the
10117 cipher context. The EVP_CIPHER can provide its own random key generation
10118 routine to support keys of a specific form. This is used in the des and
10119 3des routines to generate a key of the correct parity. Update S/MIME
10120 code to use new functions and hence generate correct parity DES keys.
10121 Add EVP_CHECK_DES_KEY #define to return an error if the key is not
10122 valid (weak or incorrect parity).
10126 * Add a local set of CRLs that can be used by X509_verify_cert() as well
10127 as looking them up. This is useful when the verified structure may contain
10128 CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs
10129 present unless the new PKCS7_NO_CRL flag is asserted.
10133 * Extend ASN1 oid configuration module. It now additionally accepts the
10136 shortName = some long name, 1.2.3.4
10140 * Reimplemented the BN_CTX implementation. There is now no more static
10141 limitation on the number of variables it can handle nor the depth of the
10142 "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack
10143 information can now expand as required, and rather than having a single
10144 static array of bignums, BN_CTX now uses a linked-list of such arrays
10145 allowing it to expand on demand whilst maintaining the usefulness of
10146 BN_CTX's "bundling".
10150 * Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD
10151 to allow all RSA operations to function using a single BN_CTX.
10155 * Preliminary support for certificate policy evaluation and checking. This
10156 is initially intended to pass the tests outlined in "Conformance Testing
10157 of Relying Party Client Certificate Path Processing Logic" v1.07.
10161 * bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and
10162 remained unused and not that useful. A variety of other little bignum
10163 tweaks and fixes have also been made continuing on from the audit (see
10168 * Constify all or almost all d2i, c2i, s2i and r2i functions, along with
10169 associated ASN1, EVP and SSL functions and old ASN1 macros.
10173 * BN_zero() only needs to set 'top' and 'neg' to zero for correct results,
10174 and this should never fail. So the return value from the use of
10175 BN_set_word() (which can fail due to needless expansion) is now deprecated;
10176 if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro.
10180 * BN_CTX_get() should return zero-valued bignums, providing the same
10181 initialised value as BN_new().
10183 *Geoff Thorpe, suggested by Ulf Möller*
10185 * Support for inhibitAnyPolicy certificate extension.
10189 * An audit of the BIGNUM code is underway, for which debugging code is
10190 enabled when BN_DEBUG is defined. This makes stricter enforcements on what
10191 is considered valid when processing BIGNUMs, and causes execution to
10192 assert() when a problem is discovered. If BN_DEBUG_RAND is defined,
10193 further steps are taken to deliberately pollute unused data in BIGNUM
10194 structures to try and expose faulty code further on. For now, openssl will
10195 (in its default mode of operation) continue to tolerate the inconsistent
10196 forms that it has tolerated in the past, but authors and packagers should
10197 consider trying openssl and their own applications when compiled with
10198 these debugging symbols defined. It will help highlight potential bugs in
10199 their own code, and will improve the test coverage for OpenSSL itself. At
10200 some point, these tighter rules will become openssl's default to improve
10201 maintainability, though the assert()s and other overheads will remain only
10202 in debugging configurations. See bn.h for more details.
10204 *Geoff Thorpe, Nils Larsch, Ulf Möller*
10206 * BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure
10207 that can only be obtained through BN_CTX_new() (which implicitly
10208 initialises it). The presence of this function only made it possible
10209 to overwrite an existing structure (and cause memory leaks).
10213 * Because of the callback-based approach for implementing LHASH as a
10214 template type, lh_insert() adds opaque objects to hash-tables and
10215 lh_doall() or lh_doall_arg() are typically used with a destructor callback
10216 to clean up those corresponding objects before destroying the hash table
10217 (and losing the object pointers). So some over-zealous constifications in
10218 LHASH have been relaxed so that lh_insert() does not take (nor store) the
10219 objects as "const" and the `lh_doall[_arg]` callback wrappers are not
10220 prototyped to have "const" restrictions on the object pointers they are
10221 given (and so aren't required to cast them away any more).
10225 * The tmdiff.h API was so ugly and minimal that our own timing utility
10226 (speed) prefers to use its own implementation. The two implementations
10227 haven't been consolidated as yet (volunteers?) but the tmdiff API has had
10228 its object type properly exposed (MS_TM) instead of casting to/from
10229 `char *`. This may still change yet if someone realises MS_TM and
10231 aren't necessarily the greatest nomenclatures - but this is what was used
10232 internally to the implementation so I've used that for now.
10236 * Ensure that deprecated functions do not get compiled when
10237 OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of
10238 the self-tests were still using deprecated key-generation functions so
10239 these have been updated also.
10243 * Reorganise PKCS#7 code to separate the digest location functionality
10244 into PKCS7_find_digest(), digest addition into PKCS7_bio_add_digest().
10245 New function PKCS7_set_digest() to set the digest type for PKCS#7
10246 digestedData type. Add additional code to correctly generate the
10247 digestedData type and add support for this type in PKCS7 initialization
10252 * New function PKCS7_set0_type_other() this initializes a PKCS7
10253 structure of type "other".
10257 * Fix prime generation loop in crypto/bn/bn_prime.pl by making
10258 sure the loop does correctly stop and breaking ("division by zero")
10259 modulus operations are not performed. The (pre-generated) prime
10260 table crypto/bn/bn_prime.h was already correct, but it could not be
10261 re-generated on some platforms because of the "division by zero"
10262 situation in the script.
10264 *Ralf S. Engelschall*
10266 * Update support for ECC-based TLS ciphersuites according to
10267 draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with
10268 SHA-1 now is only used for "small" curves (where the
10269 representation of a field element takes up to 24 bytes); for
10270 larger curves, the field element resulting from ECDH is directly
10271 used as premaster secret.
10273 *Douglas Stebila (Sun Microsystems Laboratories)*
10275 * Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2
10276 curve secp160r1 to the tests.
10278 *Douglas Stebila (Sun Microsystems Laboratories)*
10280 * Add the possibility to load symbols globally with DSO.
10282 *Götz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte*
10284 * Add the functions ERR_set_mark() and ERR_pop_to_mark() for better
10285 control of the error stack.
10289 * Add support for STORE in ENGINE.
10293 * Add the STORE type. The intention is to provide a common interface
10294 to certificate and key stores, be they simple file-based stores, or
10295 HSM-type store, or LDAP stores, or...
10296 NOTE: The code is currently UNTESTED and isn't really used anywhere.
10300 * Add a generic structure called OPENSSL_ITEM. This can be used to
10301 pass a list of arguments to any function as well as provide a way
10302 for a function to pass data back to the caller.
10306 * Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup()
10307 works like BUF_strdup() but can be used to duplicate a portion of
10308 a string. The copy gets NUL-terminated. BUF_memdup() duplicates
10313 * Add the function sk_find_ex() which works like sk_find(), but will
10314 return an index to an element even if an exact match couldn't be
10315 found. The index is guaranteed to point at the element where the
10316 searched-for key would be inserted to preserve sorting order.
10320 * Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but
10321 takes an extra flags argument for optional functionality. Currently,
10322 the following flags are defined:
10324 OBJ_BSEARCH_VALUE_ON_NOMATCH
10325 This one gets OBJ_bsearch_ex() to return a pointer to the first
10326 element where the comparing function returns a negative or zero
10329 OBJ_BSEARCH_FIRST_VALUE_ON_MATCH
10330 This one gets OBJ_bsearch_ex() to return a pointer to the first
10331 element where the comparing function returns zero. This is useful
10332 if there are more than one element where the comparing function
10337 * Make it possible to create self-signed certificates with 'openssl ca'
10338 in such a way that the self-signed certificate becomes part of the
10339 CA database and uses the same mechanisms for serial number generation
10340 as all other certificate signing. The new flag '-selfsign' enables
10341 this functionality. Adapt CA.sh and CA.pl.in.
10345 * Add functionality to check the public key of a certificate request
10346 against a given private. This is useful to check that a certificate
10347 request can be signed by that key (self-signing).
10351 * Make it possible to have multiple active certificates with the same
10352 subject in the CA index file. This is done only if the keyword
10353 'unique_subject' is set to 'no' in the main CA section (default
10354 if 'CA_default') of the configuration file. The value is saved
10355 with the database itself in a separate index attribute file,
10356 named like the index file with '.attr' appended to the name.
10360 * Generate multi-valued AVAs using '+' notation in config files for
10365 * Support for nameConstraints certificate extension.
10369 * Support for policyConstraints certificate extension.
10373 * Support for policyMappings certificate extension.
10377 * Make sure the default DSA_METHOD implementation only uses its
10378 dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL,
10379 and change its own handlers to be NULL so as to remove unnecessary
10380 indirection. This lets alternative implementations fallback to the
10381 default implementation more easily.
10385 * Support for directoryName in GeneralName related extensions
10390 * Make it possible to link applications using Makefile.shared.
10391 Make that possible even when linking against static libraries!
10395 * Support for single pass processing for S/MIME signing. This now
10396 means that S/MIME signing can be done from a pipe, in addition
10397 cleartext signing (multipart/signed type) is effectively streaming
10398 and the signed data does not need to be all held in memory.
10400 This is done with a new flag PKCS7_STREAM. When this flag is set
10401 PKCS7_sign() only initializes the PKCS7 structure and the actual signing
10402 is done after the data is output (and digests calculated) in
10403 SMIME_write_PKCS7().
10407 * Add full support for -rpath/-R, both in shared libraries and
10408 applications, at least on the platforms where it's known how
10413 * In crypto/ec/ec_mult.c, implement fast point multiplication with
10414 precomputation, based on wNAF splitting: EC_GROUP_precompute_mult()
10415 will now compute a table of multiples of the generator that
10416 makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul()
10417 faster (notably in the case of a single point multiplication,
10418 scalar * generator).
10420 *Nils Larsch, Bodo Moeller*
10422 * IPv6 support for certificate extensions. The various extensions
10423 which use the IP:a.b.c.d can now take IPv6 addresses using the
10424 formats of RFC1884 2.2 . IPv6 addresses are now also displayed
10429 * Added an ENGINE that implements RSA by performing private key
10430 exponentiations with the GMP library. The conversions to and from
10431 GMP's mpz_t format aren't optimised nor are any montgomery forms
10432 cached, and on x86 it appears OpenSSL's own performance has caught up.
10433 However there are likely to be other architectures where GMP could
10434 provide a boost. This ENGINE is not built in by default, but it can be
10435 specified at Configure time and should be accompanied by the necessary
10436 linker additions, eg;
10437 ./config -DOPENSSL_USE_GMP -lgmp
10441 * "openssl engine" will not display ENGINE/DSO load failure errors when
10442 testing availability of engines with "-t" - the old behaviour is
10443 produced by increasing the feature's verbosity with "-tt".
10447 * ECDSA routines: under certain error conditions uninitialized BN objects
10448 could be freed. Solution: make sure initialization is performed early
10449 enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de>
10454 * Key-generation can now be implemented in RSA_METHOD, DSA_METHOD
10455 and DH_METHOD (eg. by ENGINE implementations) to override the normal
10456 software implementations. For DSA and DH, parameter generation can
10457 also be overridden by providing the appropriate method callbacks.
10461 * Change the "progress" mechanism used in key-generation and
10462 primality testing to functions that take a new BN_GENCB pointer in
10463 place of callback/argument pairs. The new API functions have `_ex`
10464 postfixes and the older functions are reimplemented as wrappers for
10465 the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide
10466 declarations of the old functions to help (graceful) attempts to
10467 migrate to the new functions. Also, the new key-generation API
10468 functions operate on a caller-supplied key-structure and return
10469 success/failure rather than returning a key or NULL - this is to
10470 help make "keygen" another member function of RSA_METHOD etc.
10472 Example for using the new callback interface:
10474 int (*my_callback)(int a, int b, BN_GENCB *cb) = ...;
10475 void *my_arg = ...;
10478 BN_GENCB_set(&my_cb, my_callback, my_arg);
10480 return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb);
10481 /* For the meaning of a, b in calls to my_callback(), see the
10482 * documentation of the function that calls the callback.
10483 * cb will point to my_cb; my_arg can be retrieved as cb->arg.
10484 * my_callback should return 1 if it wants BN_is_prime_ex()
10485 * to continue, or 0 to stop.
10490 * Change the ZLIB compression method to be stateful, and make it
10491 available to TLS with the number defined in
10492 draft-ietf-tls-compression-04.txt.
10496 * Add the ASN.1 structures and functions for CertificatePair, which
10497 is defined as follows (according to X.509_4thEditionDraftV6.pdf):
10499 CertificatePair ::= SEQUENCE {
10500 forward [0] Certificate OPTIONAL,
10501 reverse [1] Certificate OPTIONAL,
10502 -- at least one of the pair shall be present -- }
10504 Also implement the PEM functions to read and write certificate
10505 pairs, and defined the PEM tag as "CERTIFICATE PAIR".
10507 This needed to be defined, mostly for the sake of the LDAP
10508 attribute crossCertificatePair, but may prove useful elsewhere as
10513 * Make it possible to inhibit symlinking of shared libraries in
10514 Makefile.shared, for Cygwin's sake.
10518 * Extend the BIGNUM API by creating a function
10519 void BN_set_negative(BIGNUM *a, int neg);
10520 and a macro that behave like
10521 int BN_is_negative(const BIGNUM *a);
10523 to avoid the need to access 'a->neg' directly in applications.
10527 * Implement fast modular reduction for pseudo-Mersenne primes
10528 used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c).
10529 EC_GROUP_new_curve_GFp() will now automatically use this
10532 *Nils Larsch <nla@trustcenter.de>*
10534 * Add new lock type (CRYPTO_LOCK_BN).
10538 * Change the ENGINE framework to automatically load engines
10539 dynamically from specific directories unless they could be
10540 found to already be built in or loaded. Move all the
10541 current engines except for the cryptodev one to a new
10542 directory engines/.
10543 The engines in engines/ are built as shared libraries if
10544 the "shared" options was given to ./Configure or ./config.
10545 Otherwise, they are inserted in libcrypto.a.
10546 /usr/local/ssl/engines is the default directory for dynamic
10547 engines, but that can be overridden at configure time through
10548 the usual use of --prefix and/or --openssldir, and at run
10549 time with the environment variable OPENSSL_ENGINES.
10551 *Geoff Thorpe and Richard Levitte*
10553 * Add Makefile.shared, a helper makefile to build shared
10554 libraries. Adapt Makefile.org.
10558 * Add version info to Win32 DLLs.
10560 *Peter 'Luna' Runestig" <peter@runestig.com>*
10562 * Add new 'medium level' PKCS#12 API. Certificates and keys
10563 can be added using this API to created arbitrary PKCS#12
10564 files while avoiding the low-level API.
10566 New options to PKCS12_create(), key or cert can be NULL and
10567 will then be omitted from the output file. The encryption
10568 algorithm NIDs can be set to -1 for no encryption, the mac
10569 iteration count can be set to 0 to omit the mac.
10571 Enhance pkcs12 utility by making the -nokeys and -nocerts
10572 options work when creating a PKCS#12 file. New option -nomac
10573 to omit the mac, NONE can be set for an encryption algorithm.
10574 New code is modified to use the enhanced PKCS12_create()
10575 instead of the low-level API.
10579 * Extend ASN1 encoder to support indefinite length constructed
10580 encoding. This can output sequences tags and octet strings in
10581 this form. Modify pk7_asn1.c to support indefinite length
10582 encoding. This is experimental and needs additional code to
10583 be useful, such as an ASN1 bio and some enhanced streaming
10586 Extend template encode functionality so that tagging is passed
10587 down to the template encoder.
10591 * Let 'openssl req' fail if an argument to '-newkey' is not
10592 recognized instead of using RSA as a default.
10596 * Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt.
10597 As these are not official, they are not included in "ALL";
10598 the "ECCdraft" ciphersuite group alias can be used to select them.
10600 *Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)*
10602 * Add ECDH engine support.
10604 *Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)*
10606 * Add ECDH in new directory crypto/ecdh/.
10608 *Douglas Stebila (Sun Microsystems Laboratories)*
10610 * Let BN_rand_range() abort with an error after 100 iterations
10611 without success (which indicates a broken PRNG).
10615 * Change BN_mod_sqrt() so that it verifies that the input value
10616 is really the square of the return value. (Previously,
10617 BN_mod_sqrt would show GIGO behaviour.)
10621 * Add named elliptic curves over binary fields from X9.62, SECG,
10622 and WAP/WTLS; add OIDs that were still missing.
10624 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10626 * Extend the EC library for elliptic curves over binary fields
10627 (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/).
10630 EC_GF2m_simple_method
10634 EC_GROUP_new_curve_GF2m
10635 EC_GROUP_set_curve_GF2m
10636 EC_GROUP_get_curve_GF2m
10637 EC_POINT_set_affine_coordinates_GF2m
10638 EC_POINT_get_affine_coordinates_GF2m
10639 EC_POINT_set_compressed_coordinates_GF2m
10641 Point compression for binary fields is disabled by default for
10642 patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to
10645 As binary polynomials are represented as BIGNUMs, various members
10646 of the EC_GROUP and EC_POINT data structures can be shared
10647 between the implementations for prime fields and binary fields;
10648 the above `..._GF2m functions` (except for EX_GROUP_new_curve_GF2m)
10649 are essentially identical to their `..._GFp` counterparts.
10650 (For simplicity, the `..._GFp` prefix has been dropped from
10651 various internal method names.)
10653 An internal 'field_div' method (similar to 'field_mul' and
10654 'field_sqr') has been added; this is used only for binary fields.
10656 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10658 * Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult()
10659 through methods ('mul', 'precompute_mult').
10661 The generic implementations (now internally called 'ec_wNAF_mul'
10662 and 'ec_wNAF_precomputed_mult') remain the default if these
10663 methods are undefined.
10665 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10667 * New function EC_GROUP_get_degree, which is defined through
10668 EC_METHOD. For curves over prime fields, this returns the bit
10669 length of the modulus.
10671 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10673 * New functions EC_GROUP_dup, EC_POINT_dup.
10674 (These simply call ..._new and ..._copy).
10676 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10678 * Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c.
10679 Polynomials are represented as BIGNUMs (where the sign bit is not
10680 used) in the following functions [macros]:
10683 BN_GF2m_sub [= BN_GF2m_add]
10684 BN_GF2m_mod [wrapper for BN_GF2m_mod_arr]
10685 BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr]
10686 BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr]
10688 BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr]
10689 BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr]
10690 BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr]
10691 BN_GF2m_cmp [= BN_ucmp]
10693 (Note that only the 'mod' functions are actually for fields GF(2^m).
10694 BN_GF2m_add() is misnomer, but this is for the sake of consistency.)
10696 For some functions, an the irreducible polynomial defining a
10697 field can be given as an 'unsigned int[]' with strictly
10698 decreasing elements giving the indices of those bits that are set;
10699 i.e., p[] represents the polynomial
10700 f(t) = t^p[0] + t^p[1] + ... + t^p[k]
10702 p[0] > p[1] > ... > p[k] = 0.
10703 This applies to the following functions:
10706 BN_GF2m_mod_mul_arr
10707 BN_GF2m_mod_sqr_arr
10708 BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv]
10709 BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div]
10710 BN_GF2m_mod_exp_arr
10711 BN_GF2m_mod_sqrt_arr
10712 BN_GF2m_mod_solve_quad_arr
10716 Conversion can be performed by the following functions:
10721 bntest.c has additional tests for binary polynomial arithmetic.
10723 Two implementations for BN_GF2m_mod_div() are available.
10724 The default algorithm simply uses BN_GF2m_mod_inv() and
10725 BN_GF2m_mod_mul(). The alternative algorithm is compiled in only
10726 if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the
10727 copyright notice in crypto/bn/bn_gf2m.c before enabling it).
10729 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10731 * Add new error code 'ERR_R_DISABLED' that can be used when some
10732 functionality is disabled at compile-time.
10734 *Douglas Stebila <douglas.stebila@sun.com>*
10736 * Change default behaviour of 'openssl asn1parse' so that more
10737 information is visible when viewing, e.g., a certificate:
10739 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump'
10740 mode the content of non-printable OCTET STRINGs is output in a
10741 style similar to INTEGERs, but with '[HEX DUMP]' prepended to
10742 avoid the appearance of a printable string.
10744 *Nils Larsch <nla@trustcenter.de>*
10746 * Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access
10748 EC_GROUP_set_asn1_flag()
10749 EC_GROUP_get_asn1_flag()
10750 EC_GROUP_set_point_conversion_form()
10751 EC_GROUP_get_point_conversion_form()
10752 These control ASN1 encoding details:
10753 - Curves (i.e., groups) are encoded explicitly unless asn1_flag
10754 has been set to OPENSSL_EC_NAMED_CURVE.
10755 - Points are encoded in uncompressed form by default; options for
10756 asn1_for are as for point2oct, namely
10757 POINT_CONVERSION_COMPRESSED
10758 POINT_CONVERSION_UNCOMPRESSED
10759 POINT_CONVERSION_HYBRID
10761 Also add 'seed' and 'seed_len' members to EC_GROUP with access
10763 EC_GROUP_set_seed()
10764 EC_GROUP_get0_seed()
10765 EC_GROUP_get_seed_len()
10766 This is used only for ASN1 purposes (so far).
10768 *Nils Larsch <nla@trustcenter.de>*
10770 * Add 'field_type' member to EC_METHOD, which holds the NID
10771 of the appropriate field type OID. The new function
10772 EC_METHOD_get_field_type() returns this value.
10774 *Nils Larsch <nla@trustcenter.de>*
10777 EC_POINT_point2bn()
10778 EC_POINT_bn2point()
10779 EC_POINT_point2hex()
10780 EC_POINT_hex2point()
10781 providing useful interfaces to EC_POINT_point2oct() and
10782 EC_POINT_oct2point().
10784 *Nils Larsch <nla@trustcenter.de>*
10786 * Change internals of the EC library so that the functions
10787 EC_GROUP_set_generator()
10788 EC_GROUP_get_generator()
10789 EC_GROUP_get_order()
10790 EC_GROUP_get_cofactor()
10791 are implemented directly in crypto/ec/ec_lib.c and not dispatched
10792 to methods, which would lead to unnecessary code duplication when
10793 adding different types of curves.
10795 *Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller*
10797 * Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM
10798 arithmetic, and such that modified wNAFs are generated
10799 (which avoid length expansion in many cases).
10803 * Add a function EC_GROUP_check_discriminant() (defined via
10804 EC_METHOD) that verifies that the curve discriminant is non-zero.
10806 Add a function EC_GROUP_check() that makes some sanity tests
10807 on a EC_GROUP, its generator and order. This includes
10808 EC_GROUP_check_discriminant().
10810 *Nils Larsch <nla@trustcenter.de>*
10812 * Add ECDSA in new directory crypto/ecdsa/.
10814 Add applications 'openssl ecparam' and 'openssl ecdsa'
10815 (these are based on 'openssl dsaparam' and 'openssl dsa').
10817 ECDSA support is also included in various other files across the
10818 library. Most notably,
10819 - 'openssl req' now has a '-newkey ecdsa:file' option;
10820 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA;
10821 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and
10822 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make
10823 them suitable for ECDSA where domain parameters must be
10824 extracted before the specific public key;
10825 - ECDSA engine support has been added.
10827 *Nils Larsch <nla@trustcenter.de>*
10829 * Include some named elliptic curves, and add OIDs from X9.62,
10830 SECG, and WAP/WTLS. Each curve can be obtained from the new
10832 EC_GROUP_new_by_curve_name(),
10833 and the list of available named curves can be obtained with
10834 EC_get_builtin_curves().
10835 Also add a 'curve_name' member to EC_GROUP objects, which can be
10837 EC_GROUP_set_curve_name()
10838 EC_GROUP_get_curve_name()
10840 *Nils Larsch <larsch@trustcenter.de, Bodo Moeller*
10842 * Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
10843 was actually never needed) and in BN_mul(). The removal in BN_mul()
10844 required a small change in bn_mul_part_recursive() and the addition
10845 of the functions bn_cmp_part_words(), bn_sub_part_words() and
10846 bn_add_part_words(), which do the same thing as bn_cmp_words(),
10847 bn_sub_words() and bn_add_words() except they take arrays with
10852 ### Changes between 0.9.7l and 0.9.7m [23 Feb 2007]
10854 * Cleanse PEM buffers before freeing them since they may contain
10857 *Benjamin Bennett <ben@psc.edu>*
10859 * Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
10860 a ciphersuite string such as "DEFAULT:RSA" cannot enable
10861 authentication-only ciphersuites.
10865 * Since AES128 and AES256 share a single mask bit in the logic of
10866 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
10867 kludge to work properly if AES128 is available and AES256 isn't.
10871 * Expand security boundary to match 1.1.1 module.
10875 * Remove redundant features: hash file source, editing of test vectors
10876 modify fipsld to use external fips_premain.c signature.
10880 * New perl script mkfipsscr.pl to create shell scripts or batch files to
10881 run algorithm test programs.
10885 * Make algorithm test programs more tolerant of whitespace.
10889 * Have SSL/TLS server implementation tolerate "mismatched" record
10890 protocol version while receiving ClientHello even if the
10891 ClientHello is fragmented. (The server can't insist on the
10892 particular protocol version it has chosen before the ServerHello
10893 message has informed the client about his choice.)
10897 * Load error codes if they are not already present instead of using a
10898 static variable. This allows them to be cleanly unloaded and reloaded.
10902 ### Changes between 0.9.7k and 0.9.7l [28 Sep 2006]
10904 * Introduce limits to prevent malicious keys being able to
10905 cause a denial of service. ([CVE-2006-2940])
10907 *Steve Henson, Bodo Moeller*
10909 * Fix ASN.1 parsing of certain invalid structures that can result
10910 in a denial of service. ([CVE-2006-2937]) [Steve Henson]
10912 * Fix buffer overflow in SSL_get_shared_ciphers() function.
10913 ([CVE-2006-3738]) [Tavis Ormandy and Will Drewry, Google Security Team]
10915 * Fix SSL client code which could crash if connecting to a
10916 malicious SSLv2 server. ([CVE-2006-4343])
10918 *Tavis Ormandy and Will Drewry, Google Security Team*
10920 * Change ciphersuite string processing so that an explicit
10921 ciphersuite selects this one ciphersuite (so that "AES256-SHA"
10922 will no longer include "AES128-SHA"), and any other similar
10923 ciphersuite (same bitmap) from *other* protocol versions (so that
10924 "RC4-MD5" will still include both the SSL 2.0 ciphersuite and the
10925 SSL 3.0/TLS 1.0 ciphersuite). This is a backport combining
10926 changes from 0.9.8b and 0.9.8d.
10930 ### Changes between 0.9.7j and 0.9.7k [05 Sep 2006]
10932 * Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
10933 ([CVE-2006-4339]) [Ben Laurie and Google Security Team]
10935 * Change the Unix randomness entropy gathering to use poll() when
10936 possible instead of select(), since the latter has some
10937 undesirable limitations.
10939 *Darryl Miles via Richard Levitte and Bodo Moeller*
10941 * Disable rogue ciphersuites:
10943 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
10944 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
10945 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
10947 The latter two were purportedly from
10948 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
10951 Also deactivate the remaining ciphersuites from
10952 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
10953 unofficial, and the ID has long expired.
10957 * Fix RSA blinding Heisenbug (problems sometimes occurred on
10958 dual-core machines) and other potential thread-safety issues.
10962 ### Changes between 0.9.7i and 0.9.7j [04 May 2006]
10964 * Adapt fipsld and the build system to link against the validated FIPS
10965 module in FIPS mode.
10969 * Fixes for VC++ 2005 build under Windows.
10973 * Add new Windows build target VC-32-GMAKE for VC++. This uses GNU make
10974 from a Windows bash shell such as MSYS. It is autodetected from the
10975 "config" script when run from a VC++ environment. Modify standard VC++
10976 build to use fipscanister.o from the GNU make build.
10980 ### Changes between 0.9.7h and 0.9.7i [14 Oct 2005]
10982 * Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS.
10983 The value now differs depending on if you build for FIPS or not.
10984 BEWARE! A program linked with a shared FIPSed libcrypto can't be
10985 safely run with a non-FIPSed libcrypto, as it may crash because of
10986 the difference induced by this change.
10990 ### Changes between 0.9.7g and 0.9.7h [11 Oct 2005]
10992 * Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
10993 (part of SSL_OP_ALL). This option used to disable the
10994 countermeasure against man-in-the-middle protocol-version
10995 rollback in the SSL 2.0 server implementation, which is a bad
10996 idea. ([CVE-2005-2969])
10998 *Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
10999 for Information Security, National Institute of Advanced Industrial
11000 Science and Technology [AIST, Japan)]*
11002 * Minimal support for X9.31 signatures and PSS padding modes. This is
11003 mainly for FIPS compliance and not fully integrated at this stage.
11007 * For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform
11008 the exponentiation using a fixed-length exponent. (Otherwise,
11009 the information leaked through timing could expose the secret key
11010 after many signatures; cf. Bleichenbacher's attack on DSA with
11015 * Make a new fixed-window mod_exp implementation the default for
11016 RSA, DSA, and DH private-key operations so that the sequence of
11017 squares and multiplies and the memory access pattern are
11018 independent of the particular secret key. This will mitigate
11019 cache-timing and potential related attacks.
11021 BN_mod_exp_mont_consttime() is the new exponentiation implementation,
11022 and this is automatically used by BN_mod_exp_mont() if the new flag
11023 BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH
11024 will use this BN flag for private exponents unless the flag
11025 RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or
11026 DH_FLAG_NO_EXP_CONSTTIME, respectively, is set.
11028 *Matthew D Wood (Intel Corp), with some changes by Bodo Moeller*
11030 * Change the client implementation for SSLv23_method() and
11031 SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0
11032 Client Hello message format if the SSL_OP_NO_SSLv2 option is set.
11033 (Previously, the SSL 2.0 backwards compatible Client Hello
11034 message format would be used even with SSL_OP_NO_SSLv2.)
11038 * Add support for smime-type MIME parameter in S/MIME messages which some
11043 * New function BN_MONT_CTX_set_locked() to set montgomery parameters in
11044 a threadsafe manner. Modify rsa code to use new function and add calls
11045 to dsa and dh code (which had race conditions before).
11049 * Include the fixed error library code in the C error file definitions
11050 instead of fixing them up at runtime. This keeps the error code
11051 structures constant.
11055 ### Changes between 0.9.7f and 0.9.7g [11 Apr 2005]
11057 [NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after
11060 * Fixes for newer kerberos headers. NB: the casts are needed because
11061 the 'length' field is signed on one version and unsigned on another
11062 with no (?) obvious way to tell the difference, without these VC++
11063 complains. Also the "definition" of FAR (blank) is no longer included
11064 nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up
11065 some needed definitions.
11069 * Undo Cygwin change.
11073 * Added support for proxy certificates according to RFC 3820.
11074 Because they may be a security thread to unaware applications,
11075 they must be explicitly allowed in run-time. See
11076 docs/HOWTO/proxy_certificates.txt for further information.
11080 ### Changes between 0.9.7e and 0.9.7f [22 Mar 2005]
11082 * Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating
11083 server and client random values. Previously
11084 (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in
11085 less random data when sizeof(time_t) > 4 (some 64 bit platforms).
11087 This change has negligible security impact because:
11089 1. Server and client random values still have 24 bytes of pseudo random
11092 2. Server and client random values are sent in the clear in the initial
11095 3. The master secret is derived using the premaster secret (48 bytes in
11096 size for static RSA ciphersuites) as well as client server and random
11099 The OpenSSL team would like to thank the UK NISCC for bringing this issue
11102 *Stephen Henson, reported by UK NISCC*
11104 * Use Windows randomness collection on Cygwin.
11108 * Fix hang in EGD/PRNGD query when communication socket is closed
11109 prematurely by EGD/PRNGD.
11111 *Darren Tucker <dtucker@zip.com.au> via Lutz Jänicke, resolves #1014*
11113 * Prompt for pass phrases when appropriate for PKCS12 input format.
11117 * Back-port of selected performance improvements from development
11118 branch, as well as improved support for PowerPC platforms.
11122 * Add lots of checks for memory allocation failure, error codes to indicate
11123 failure and freeing up memory if a failure occurs.
11125 *Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson*
11127 * Add new -passin argument to dgst.
11131 * Perform some character comparisons of different types in X509_NAME_cmp:
11132 this is needed for some certificates that re-encode DNs into UTF8Strings
11133 (in violation of RFC3280) and can't or won't issue name rollover
11138 * Make an explicit check during certificate validation to see that
11139 the CA setting in each certificate on the chain is correct. As a
11140 side effect always do the following basic checks on extensions,
11141 not just when there's an associated purpose to the check:
11143 - if there is an unhandled critical extension (unless the user
11144 has chosen to ignore this fault)
11145 - if the path length has been exceeded (if one is set at all)
11146 - that certain extensions fit the associated purpose (if one has
11151 ### Changes between 0.9.7d and 0.9.7e [25 Oct 2004]
11153 * Avoid a race condition when CRLs are checked in a multi threaded
11154 environment. This would happen due to the reordering of the revoked
11155 entries during signature checking and serial number lookup. Now the
11156 encoding is cached and the serial number sort performed under a lock.
11157 Add new STACK function sk_is_sorted().
11161 * Add Delta CRL to the extension code.
11165 * Various fixes to s3_pkt.c so alerts are sent properly.
11167 *David Holmes <d.holmes@f5.com>*
11169 * Reduce the chances of duplicate issuer name and serial numbers (in
11170 violation of RFC3280) using the OpenSSL certificate creation utilities.
11171 This is done by creating a random 64 bit value for the initial serial
11172 number when a serial number file is created or when a self signed
11173 certificate is created using 'openssl req -x509'. The initial serial
11174 number file is created using 'openssl x509 -next_serial' in CA.pl
11175 rather than being initialized to 1.
11179 ### Changes between 0.9.7c and 0.9.7d [17 Mar 2004]
11181 * Fix null-pointer assignment in do_change_cipher_spec() revealed
11182 by using the Codenomicon TLS Test Tool ([CVE-2004-0079])
11184 *Joe Orton, Steve Henson*
11186 * Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites
11189 *Joe Orton, Steve Henson*
11191 * Make it possible to have multiple active certificates with the same
11192 subject in the CA index file. This is done only if the keyword
11193 'unique_subject' is set to 'no' in the main CA section (default
11194 if 'CA_default') of the configuration file. The value is saved
11195 with the database itself in a separate index attribute file,
11196 named like the index file with '.attr' appended to the name.
11200 * X509 verify fixes. Disable broken certificate workarounds when
11201 X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if
11202 keyUsage extension present. Don't accept CRLs with unhandled critical
11203 extensions: since verify currently doesn't process CRL extensions this
11204 rejects a CRL with *any* critical extensions. Add new verify error codes
11209 * When creating an OCSP nonce use an OCTET STRING inside the extnValue.
11210 A clarification of RFC2560 will require the use of OCTET STRINGs and
11211 some implementations cannot handle the current raw format. Since OpenSSL
11212 copies and compares OCSP nonces as opaque blobs without any attempt at
11213 parsing them this should not create any compatibility issues.
11217 * New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when
11218 calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without
11219 this HMAC (and other) operations are several times slower than OpenSSL
11224 * Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex().
11226 *Peter Sylvester <Peter.Sylvester@EdelWeb.fr>*
11228 * Use the correct content when signing type "other".
11232 ### Changes between 0.9.7b and 0.9.7c [30 Sep 2003]
11234 * Fix various bugs revealed by running the NISCC test suite:
11236 Stop out of bounds reads in the ASN1 code when presented with
11237 invalid tags (CVE-2003-0543 and CVE-2003-0544).
11239 Free up ASN1_TYPE correctly if ANY type is invalid ([CVE-2003-0545]).
11241 If verify callback ignores invalid public key errors don't try to check
11242 certificate signature with the NULL public key.
11246 * New -ignore_err option in ocsp application to stop the server
11247 exiting on the first error in a request.
11251 * In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
11252 if the server requested one: as stated in TLS 1.0 and SSL 3.0
11257 * In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
11258 extra data after the compression methods not only for TLS 1.0
11259 but also for SSL 3.0 (as required by the specification).
11261 *Bodo Moeller; problem pointed out by Matthias Loepfe*
11263 * Change X509_certificate_type() to mark the key as exported/exportable
11264 when it's 512 *bits* long, not 512 bytes.
11268 * Change AES_cbc_encrypt() so it outputs exact multiple of
11269 blocks during encryption.
11273 * Various fixes to base64 BIO and non blocking I/O. On write
11274 flushes were not handled properly if the BIO retried. On read
11275 data was not being buffered properly and had various logic bugs.
11276 This also affects blocking I/O when the data being decoded is a
11281 * Various S/MIME bugfixes and compatibility changes:
11282 output correct application/pkcs7 MIME type if
11283 PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures.
11284 Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening
11285 of files as .eml work). Correctly handle very long lines in MIME
11290 ### Changes between 0.9.7a and 0.9.7b [10 Apr 2003]
11292 * Countermeasure against the Klima-Pokorny-Rosa extension of
11293 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
11294 a protocol version number mismatch like a decryption error
11295 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
11299 * Turn on RSA blinding by default in the default implementation
11300 to avoid a timing attack. Applications that don't want it can call
11301 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
11302 They would be ill-advised to do so in most cases.
11304 *Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller*
11306 * Change RSA blinding code so that it works when the PRNG is not
11307 seeded (in this case, the secret RSA exponent is abused as
11308 an unpredictable seed -- if it is not unpredictable, there
11309 is no point in blinding anyway). Make RSA blinding thread-safe
11310 by remembering the creator's thread ID in rsa->blinding and
11311 having all other threads use local one-time blinding factors
11312 (this requires more computation than sharing rsa->blinding, but
11313 avoids excessive locking; and if an RSA object is not shared
11314 between threads, blinding will still be very fast).
11318 * Fixed a typo bug that would cause ENGINE_set_default() to set an
11319 ENGINE as defaults for all supported algorithms irrespective of
11320 the 'flags' parameter. 'flags' is now honoured, so applications
11321 should make sure they are passing it correctly.
11325 * Target "mingw" now allows native Windows code to be generated in
11326 the Cygwin environment as well as with the MinGW compiler.
11330 ### Changes between 0.9.7 and 0.9.7a [19 Feb 2003]
11332 * In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
11333 via timing by performing a MAC computation even if incorrect
11334 block cipher padding has been found. This is a countermeasure
11335 against active attacks where the attacker has to distinguish
11336 between bad padding and a MAC verification error. ([CVE-2003-0078])
11338 *Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
11339 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
11340 Martin Vuagnoux (EPFL, Ilion)*
11342 * Make the no-err option work as intended. The intention with no-err
11343 is not to have the whole error stack handling routines removed from
11344 libcrypto, it's only intended to remove all the function name and
11345 reason texts, thereby removing some of the footprint that may not
11346 be interesting if those errors aren't displayed anyway.
11348 NOTE: it's still possible for any application or module to have its
11349 own set of error texts inserted. The routines are there, just not
11350 used by default when no-err is given.
11354 * Add support for FreeBSD on IA64.
11356 *dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454*
11358 * Adjust DES_cbc_cksum() so it returns the same value as the MIT
11359 Kerberos function mit_des_cbc_cksum(). Before this change,
11360 the value returned by DES_cbc_cksum() was like the one from
11361 mit_des_cbc_cksum(), except the bytes were swapped.
11363 *Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte*
11365 * Allow an application to disable the automatic SSL chain building.
11366 Before this a rather primitive chain build was always performed in
11367 ssl3_output_cert_chain(): an application had no way to send the
11368 correct chain if the automatic operation produced an incorrect result.
11370 Now the chain builder is disabled if either:
11372 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert().
11374 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set.
11376 The reasoning behind this is that an application would not want the
11377 auto chain building to take place if extra chain certificates are
11378 present and it might also want a means of sending no additional
11379 certificates (for example the chain has two certificates and the
11384 * Add the possibility to build without the ENGINE framework.
11386 *Steven Reddie <smr@essemer.com.au> via Richard Levitte*
11388 * Under Win32 gmtime() can return NULL: check return value in
11389 OPENSSL_gmtime(). Add error code for case where gmtime() fails.
11393 * DSA routines: under certain error conditions uninitialized BN objects
11394 could be freed. Solution: make sure initialization is performed early
11395 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>,
11396 Nils Larsch <nla@trustcenter.de> via PR#459)
11400 * Another fix for SSLv2 session ID handling: the session ID was incorrectly
11401 checked on reconnect on the client side, therefore session resumption
11402 could still fail with a "ssl session id is different" error. This
11403 behaviour is masked when SSL_OP_ALL is used due to
11404 SSL_OP_MICROSOFT_SESS_ID_BUG being set.
11405 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
11406 followup to PR #377.
11410 * IA-32 assembler support enhancements: unified ELF targets, support
11411 for SCO/Caldera platforms, fix for Cygwin shared build.
11415 * Add support for FreeBSD on sparc64. As a consequence, support for
11416 FreeBSD on non-x86 processors is separate from x86 processors on
11417 the config script, much like the NetBSD support.
11419 *Richard Levitte & Kris Kennaway <kris@obsecurity.org>*
11421 ### Changes between 0.9.6h and 0.9.7 [31 Dec 2002]
11423 [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after
11426 * Fix session ID handling in SSLv2 client code: the SERVER FINISHED
11427 code (06) was taken as the first octet of the session ID and the last
11428 octet was ignored consequently. As a result SSLv2 client side session
11429 caching could not have worked due to the session ID mismatch between
11431 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
11436 * Change the declaration of needed Kerberos libraries to use EX_LIBS
11437 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is
11442 * The hw_ncipher.c engine requires dynamic locks. Unfortunately, it
11443 seems that in spite of existing for more than a year, many application
11444 author have done nothing to provide the necessary callbacks, which
11445 means that this particular engine will not work properly anywhere.
11446 This is a very unfortunate situation which forces us, in the name
11447 of usability, to give the hw_ncipher.c a static lock, which is part
11449 NOTE: This is for the 0.9.7 series ONLY. This hack will never
11450 appear in 0.9.8 or later. We EXPECT application authors to have
11451 dealt properly with this when 0.9.8 is released (unless we actually
11452 make such changes in the libcrypto locking code that changes will
11453 have to be made anyway).
11457 * In asn1_d2i_read_bio() repeatedly call BIO_read() until all content
11458 octets have been read, EOF or an error occurs. Without this change
11459 some truncated ASN1 structures will not produce an error.
11463 * Disable Heimdal support, since it hasn't been fully implemented.
11464 Still give the possibility to force the use of Heimdal, but with
11465 warnings and a request that patches get sent to openssl-dev.
11469 * Add the VC-CE target, introduce the WINCE sysname, and add
11470 INSTALL.WCE and appropriate conditionals to make it build.
11472 *Steven Reddie <smr@essemer.com.au> via Richard Levitte*
11474 * Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and
11475 cygssl-x.y.z.dll, where x, y and z are the major, minor and
11476 edit numbers of the version.
11478 *Corinna Vinschen <vinschen@redhat.com> and Richard Levitte*
11480 * Introduce safe string copy and catenation functions
11481 (BUF_strlcpy() and BUF_strlcat()).
11483 *Ben Laurie (CHATS) and Richard Levitte*
11485 * Avoid using fixed-size buffers for one-line DNs.
11487 *Ben Laurie (CHATS)*
11489 * Add BUF_MEM_grow_clean() to avoid information leakage when
11490 resizing buffers containing secrets, and use where appropriate.
11492 *Ben Laurie (CHATS)*
11494 * Avoid using fixed size buffers for configuration file location.
11496 *Ben Laurie (CHATS)*
11498 * Avoid filename truncation for various CA files.
11500 *Ben Laurie (CHATS)*
11502 * Use sizeof in preference to magic numbers.
11504 *Ben Laurie (CHATS)*
11506 * Avoid filename truncation in cert requests.
11508 *Ben Laurie (CHATS)*
11510 * Add assertions to check for (supposedly impossible) buffer
11513 *Ben Laurie (CHATS)*
11515 * Don't cache truncated DNS entries in the local cache (this could
11516 potentially lead to a spoofing attack).
11518 *Ben Laurie (CHATS)*
11520 * Fix various buffers to be large enough for hex/decimal
11521 representations in a platform independent manner.
11523 *Ben Laurie (CHATS)*
11525 * Add CRYPTO_realloc_clean() to avoid information leakage when
11526 resizing buffers containing secrets, and use where appropriate.
11528 *Ben Laurie (CHATS)*
11530 * Add BIO_indent() to avoid much slightly worrying code to do
11533 *Ben Laurie (CHATS)*
11535 * Convert sprintf()/BIO_puts() to BIO_printf().
11537 *Ben Laurie (CHATS)*
11539 * buffer_gets() could terminate with the buffer only half
11542 *Ben Laurie (CHATS)*
11544 * Add assertions to prevent user-supplied crypto functions from
11545 overflowing internal buffers by having large block sizes, etc.
11547 *Ben Laurie (CHATS)*
11549 * New OPENSSL_assert() macro (similar to assert(), but enabled
11552 *Ben Laurie (CHATS)*
11554 * Eliminate unused copy of key in RC4.
11556 *Ben Laurie (CHATS)*
11558 * Eliminate unused and incorrectly sized buffers for IV in pem.h.
11560 *Ben Laurie (CHATS)*
11562 * Fix off-by-one error in EGD path.
11564 *Ben Laurie (CHATS)*
11566 * If RANDFILE path is too long, ignore instead of truncating.
11568 *Ben Laurie (CHATS)*
11570 * Eliminate unused and incorrectly sized X.509 structure
11573 *Ben Laurie (CHATS)*
11575 * Eliminate unused and dangerous function knumber().
11577 *Ben Laurie (CHATS)*
11579 * Eliminate unused and dangerous structure, KSSL_ERR.
11581 *Ben Laurie (CHATS)*
11583 * Protect against overlong session ID context length in an encoded
11584 session object. Since these are local, this does not appear to be
11587 *Ben Laurie (CHATS)*
11589 * Change from security patch (see 0.9.6e below) that did not affect
11590 the 0.9.6 release series:
11592 Remote buffer overflow in SSL3 protocol - an attacker could
11593 supply an oversized master key in Kerberos-enabled versions.
11596 *Ben Laurie (CHATS)*
11598 * Change the SSL kerb5 codes to match RFC 2712.
11602 * Make -nameopt work fully for req and add -reqopt switch.
11604 *Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson*
11606 * The "block size" for block ciphers in CFB and OFB mode should be 1.
11608 *Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>*
11610 * Make sure tests can be performed even if the corresponding algorithms
11611 have been removed entirely. This was also the last step to make
11612 OpenSSL compilable with DJGPP under all reasonable conditions.
11614 *Richard Levitte, Doug Kaufman <dkaufman@rahul.net>*
11616 * Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT
11617 to allow version independent disabling of normally unselected ciphers,
11618 which may be activated as a side-effect of selecting a single cipher.
11620 (E.g., cipher list string "RSA" enables ciphersuites that are left
11621 out of "ALL" because they do not provide symmetric encryption.
11622 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.)
11624 *Lutz Jaenicke, Bodo Moeller*
11626 * Add appropriate support for separate platform-dependent build
11627 directories. The recommended way to make a platform-dependent
11628 build directory is the following (tested on Linux), maybe with
11631 # Place yourself outside of the OpenSSL source tree. In
11632 # this example, the environment variable OPENSSL_SOURCE
11633 # is assumed to contain the absolute OpenSSL source directory.
11634 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`"
11635 cd objtree/"`uname -s`-`uname -r`-`uname -m`"
11636 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do
11637 mkdir -p `dirname $F`
11638 ln -s $OPENSSL_SOURCE/$F $F
11641 To be absolutely sure not to disturb the source tree, a "make clean"
11642 is a good thing. If it isn't successful, don't worry about it,
11643 it probably means the source directory is very clean.
11647 * Make sure any ENGINE control commands make local copies of string
11648 pointers passed to them whenever necessary. Otherwise it is possible
11649 the caller may have overwritten (or deallocated) the original string
11650 data when a later ENGINE operation tries to use the stored values.
11652 *Götz Babin-Ebell <babinebell@trustcenter.de>*
11654 * Improve diagnostics in file reading and command-line digests.
11656 *Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>*
11658 * Add AES modes CFB and OFB to the object database. Correct an
11659 error in AES-CFB decryption.
11663 * Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this
11664 allows existing EVP_CIPHER_CTX structures to be reused after
11665 calling `EVP_*Final()`. This behaviour is used by encryption
11666 BIOs and some applications. This has the side effect that
11667 applications must explicitly clean up cipher contexts with
11668 EVP_CIPHER_CTX_cleanup() or they will leak memory.
11672 * Check the values of dna and dnb in bn_mul_recursive before calling
11673 bn_mul_comba (a non zero value means the a or b arrays do not contain
11674 n2 elements) and fallback to bn_mul_normal if either is not zero.
11678 * Fix escaping of non-ASCII characters when using the -subj option
11679 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>)
11683 * Make object definitions compliant to LDAP (RFC2256): SN is the short
11684 form for "surname", serialNumber has no short form.
11685 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798;
11686 therefore remove "mail" short name for "internet 7".
11687 The OID for unique identifiers in X509 certificates is
11688 x500UniqueIdentifier, not uniqueIdentifier.
11689 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>)
11693 * Add an "init" command to the ENGINE config module and auto initialize
11694 ENGINEs. Without any "init" command the ENGINE will be initialized
11695 after all ctrl commands have been executed on it. If init=1 the
11696 ENGINE is initialized at that point (ctrls before that point are run
11697 on the uninitialized ENGINE and after on the initialized one). If
11698 init=0 then the ENGINE will not be initialized at all.
11702 * Fix the 'app_verify_callback' interface so that the user-defined
11703 argument is actually passed to the callback: In the
11704 SSL_CTX_set_cert_verify_callback() prototype, the callback
11705 declaration has been changed from
11708 int (*cb)(X509_STORE_CTX *,void *);
11709 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
11710 i=s->ctx->app_verify_callback(&ctx)
11711 has been changed into
11712 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
11714 To update applications using SSL_CTX_set_cert_verify_callback(),
11715 a dummy argument can be added to their callback functions.
11717 *D. K. Smetters <smetters@parc.xerox.com>*
11719 * Added the '4758cca' ENGINE to support IBM 4758 cards.
11721 *Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe*
11723 * Add and OPENSSL_LOAD_CONF define which will cause
11724 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
11725 This allows older applications to transparently support certain
11726 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
11727 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
11728 load the config file and OPENSSL_add_all_algorithms_conf() which will
11729 always load it have also been added.
11733 * Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
11734 Adjust NIDs and EVP layer.
11736 *Stephen Sprunk <stephen@sprunk.org> and Richard Levitte*
11738 * Config modules support in openssl utility.
11740 Most commands now load modules from the config file,
11741 though in a few (such as version) this isn't done
11742 because it couldn't be used for anything.
11744 In the case of ca and req the config file used is
11745 the same as the utility itself: that is the -config
11746 command line option can be used to specify an
11751 * Move default behaviour from OPENSSL_config(). If appname is NULL
11752 use "openssl_conf" if filename is NULL use default openssl config file.
11756 * Add an argument to OPENSSL_config() to allow the use of an alternative
11757 config section name. Add a new flag to tolerate a missing config file
11758 and move code to CONF_modules_load_file().
11762 * Support for crypto accelerator cards from Accelerated Encryption
11763 Processing, www.aep.ie. (Use engine 'aep')
11764 The support was copied from 0.9.6c [engine] and adapted/corrected
11765 to work with the new engine framework.
11767 *AEP Inc. and Richard Levitte*
11769 * Support for SureWare crypto accelerator cards from Baltimore
11770 Technologies. (Use engine 'sureware')
11771 The support was copied from 0.9.6c [engine] and adapted
11772 to work with the new engine framework.
11776 * Have the CHIL engine fork-safe (as defined by nCipher) and actually
11777 make the newer ENGINE framework commands for the CHIL engine work.
11779 *Toomas Kiisk <vix@cyber.ee> and Richard Levitte*
11781 * Make it possible to produce shared libraries on ReliantUNIX.
11783 *Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte*
11785 * Add the configuration target debug-linux-ppro.
11786 Make 'openssl rsa' use the general key loading routines
11787 implemented in `apps.c`, and make those routines able to
11788 handle the key format FORMAT_NETSCAPE and the variant
11791 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
11793 * Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
11795 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
11797 * Add -keyform to rsautl, and document -engine.
11799 *Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>*
11801 * Change BIO_new_file (crypto/bio/bss_file.c) to use new
11802 BIO_R_NO_SUCH_FILE error code rather than the generic
11803 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
11807 * Add new functions
11808 ERR_peek_last_error
11809 ERR_peek_last_error_line
11810 ERR_peek_last_error_line_data.
11811 These are similar to
11813 ERR_peek_error_line
11814 ERR_peek_error_line_data,
11815 but report on the latest error recorded rather than the first one
11816 still in the error queue.
11818 *Ben Laurie, Bodo Moeller*
11820 * default_algorithms option in ENGINE config module. This allows things
11822 default_algorithms = ALL
11823 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
11827 * Preliminary ENGINE config module.
11831 * New experimental application configuration code.
11835 * Change the AES code to follow the same name structure as all other
11836 symmetric ciphers, and behave the same way. Move everything to
11837 the directory crypto/aes, thereby obsoleting crypto/rijndael.
11839 *Stephen Sprunk <stephen@sprunk.org> and Richard Levitte*
11841 * SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
11843 *Ben Laurie and Theo de Raadt*
11845 * Add option to output public keys in req command.
11847 *Massimiliano Pala madwolf@openca.org*
11849 * Use wNAFs in EC_POINTs_mul() for improved efficiency
11850 (up to about 10% better than before for P-192 and P-224).
11854 * New functions/macros
11856 SSL_CTX_set_msg_callback(ctx, cb)
11857 SSL_CTX_set_msg_callback_arg(ctx, arg)
11858 SSL_set_msg_callback(ssl, cb)
11859 SSL_set_msg_callback_arg(ssl, arg)
11861 to request calling a callback function
11863 void cb(int write_p, int version, int content_type,
11864 const void *buf, size_t len, SSL *ssl, void *arg)
11866 whenever a protocol message has been completely received
11867 (write_p == 0) or sent (write_p == 1). Here 'version' is the
11868 protocol version according to which the SSL library interprets
11869 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
11870 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
11871 the content type as defined in the SSL 3.0/TLS 1.0 protocol
11872 specification (change_cipher_spec(20), alert(21), handshake(22)).
11873 'buf' and 'len' point to the actual message, 'ssl' to the
11874 SSL object, and 'arg' is the application-defined value set by
11875 SSL[_CTX]_set_msg_callback_arg().
11877 'openssl s_client' and 'openssl s_server' have new '-msg' options
11878 to enable a callback that displays all protocol messages.
11882 * Change the shared library support so shared libraries are built as
11883 soon as the corresponding static library is finished, and thereby get
11884 openssl and the test programs linked against the shared library.
11885 This still only happens when the keyword "shard" has been given to
11886 the configuration scripts.
11888 NOTE: shared library support is still an experimental thing, and
11889 backward binary compatibility is still not guaranteed.
11891 *"Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte*
11893 * Add support for Subject Information Access extension.
11895 *Peter Sylvester <Peter.Sylvester@EdelWeb.fr>*
11897 * Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
11898 additional bytes when new memory had to be allocated, not just
11899 when reusing an existing buffer.
11903 * New command line and configuration option 'utf8' for the req command.
11904 This allows field values to be specified as UTF8 strings.
11908 * Add -multi and -mr options to "openssl speed" - giving multiple parallel
11909 runs for the former and machine-readable output for the latter.
11913 * Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
11914 of the e-mail address in the DN (i.e., it will go into a certificate
11915 extension only). The new configuration file option 'email_in_dn = no'
11916 has the same effect.
11918 *Massimiliano Pala madwolf@openca.org*
11920 * Change all functions with names starting with `des_` to be starting
11921 with `DES_` instead. Add wrappers that are compatible with libdes,
11922 but are named `_ossl_old_des_*`. Finally, add macros that map the
11923 `des_*` symbols to the corresponding `_ossl_old_des_*` if libdes
11924 compatibility is desired. If OpenSSL 0.9.6c compatibility is
11925 desired, the `des_*` symbols will be mapped to `DES_*`, with one
11928 Since we provide two compatibility mappings, the user needs to
11929 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
11930 compatibility is desired. The default (i.e., when that macro
11931 isn't defined) is OpenSSL 0.9.6c compatibility.
11933 There are also macros that enable and disable the support of old
11934 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
11935 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
11936 are defined, the default will apply: to support the old des routines.
11938 In either case, one must include openssl/des.h to get the correct
11939 definitions. Do not try to just include openssl/des_old.h, that
11942 NOTE: This is a major break of an old API into a new one. Software
11943 authors are encouraged to switch to the `DES_` style functions. Some
11944 time in the future, des_old.h and the libdes compatibility functions
11945 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
11946 default), and then completely removed.
11950 * Test for certificates which contain unsupported critical extensions.
11951 If such a certificate is found during a verify operation it is
11952 rejected by default: this behaviour can be overridden by either
11953 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
11954 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
11955 X509_supported_extension() has also been added which returns 1 if a
11956 particular extension is supported.
11960 * Modify the behaviour of EVP cipher functions in similar way to digests
11961 to retain compatibility with existing code.
11965 * Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
11966 compatibility with existing code. In particular the 'ctx' parameter does
11967 not have to be to be initialized before the call to EVP_DigestInit() and
11968 it is tidied up after a call to EVP_DigestFinal(). New function
11969 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
11970 EVP_MD_CTX_copy() changed to not require the destination to be
11971 initialized valid and new function EVP_MD_CTX_copy_ex() added which
11972 requires the destination to be valid.
11974 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
11975 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
11979 * Change ssl3_get_message (ssl/s3_both.c) and the functions using it
11980 so that complete 'Handshake' protocol structures are kept in memory
11981 instead of overwriting 'msg_type' and 'length' with 'body' data.
11985 * Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
11987 *Massimo Santin via Richard Levitte*
11989 * Major restructuring to the underlying ENGINE code. This includes
11990 reduction of linker bloat, separation of pure "ENGINE" manipulation
11991 (initialisation, etc) from functionality dealing with implementations
11992 of specific crypto interfaces. This change also introduces integrated
11993 support for symmetric ciphers and digest implementations - so ENGINEs
11994 can now accelerate these by providing EVP_CIPHER and EVP_MD
11995 implementations of their own. This is detailed in
11996 [crypto/engine/README.md](crypto/engine/README.md)
11997 as it couldn't be adequately described here. However, there are a few
11998 API changes worth noting - some RSA, DSA, DH, and RAND functions that
11999 were changed in the original introduction of ENGINE code have now
12000 reverted back - the hooking from this code to ENGINE is now a good
12001 deal more passive and at run-time, operations deal directly with
12002 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
12003 dereferencing through an ENGINE pointer any more. Also, the ENGINE
12004 functions dealing with `BN_MOD_EXP[_CRT]` handlers have been removed -
12005 they were not being used by the framework as there is no concept of a
12006 BIGNUM_METHOD and they could not be generalised to the new
12007 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
12008 ENGINE_cpy() has been removed as it cannot be consistently defined in
12013 * Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
12017 * Change mkdef.pl to sort symbols that get the same entry number,
12018 and make sure the automatically generated functions `ERR_load_*`
12019 become part of libeay.num as well.
12023 * New function SSL_renegotiate_pending(). This returns true once
12024 renegotiation has been requested (either SSL_renegotiate() call
12025 or HelloRequest/ClientHello received from the peer) and becomes
12026 false once a handshake has been completed.
12027 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
12028 sends a HelloRequest, but does not ensure that a handshake takes
12029 place. SSL_renegotiate_pending() is useful for checking if the
12030 client has followed the request.)
12034 * New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
12035 By default, clients may request session resumption even during
12036 renegotiation (if session ID contexts permit); with this option,
12037 session resumption is possible only in the first handshake.
12039 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes
12040 more bits available for options that should not be part of
12041 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION).
12045 * Add some demos for certificate and certificate request creation.
12049 * Make maximum certificate chain size accepted from the peer application
12050 settable (`SSL*_get/set_max_cert_list()`), as proposed by
12051 "Douglas E. Engert" <deengert@anl.gov>.
12055 * Add support for shared libraries for Unixware-7
12056 (Boyd Lynn Gerber <gerberb@zenez.com>).
12060 * Add a "destroy" handler to ENGINEs that allows structural cleanup to
12061 be done prior to destruction. Use this to unload error strings from
12062 ENGINEs that load their own error strings. NB: This adds two new API
12063 functions to "get" and "set" this destroy handler in an ENGINE.
12067 * Alter all existing ENGINE implementations (except "openssl" and
12068 "openbsd") to dynamically instantiate their own error strings. This
12069 makes them more flexible to be built both as statically-linked ENGINEs
12070 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
12071 Also, add stub code to each that makes building them as self-contained
12072 shared-libraries easier (see [README-Engine.md](README-Engine.md)).
12076 * Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
12077 implementations into applications that are completely implemented in
12078 self-contained shared-libraries. The "dynamic" ENGINE exposes control
12079 commands that can be used to configure what shared-library to load and
12080 to control aspects of the way it is handled. Also, made an update to
12081 the [README-Engine.md](README-Engine.md) file
12082 that brings its information up-to-date and
12083 provides some information and instructions on the "dynamic" ENGINE
12084 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
12088 * Make it possible to unload ranges of ERR strings with a new
12089 "ERR_unload_strings" function.
12093 * Add a copy() function to EVP_MD.
12097 * Make EVP_MD routines take a context pointer instead of just the
12098 md_data void pointer.
12102 * Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
12103 that the digest can only process a single chunk of data
12104 (typically because it is provided by a piece of
12105 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
12106 is only going to provide a single chunk of data, and hence the
12107 framework needn't accumulate the data for oneshot drivers.
12111 * As with "ERR", make it possible to replace the underlying "ex_data"
12112 functions. This change also alters the storage and management of global
12113 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
12114 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
12115 index counters. The API functions that use this state have been changed
12116 to take a "class_index" rather than pointers to the class's local STACK
12117 and counter, and there is now an API function to dynamically create new
12118 classes. This centralisation allows us to (a) plug a lot of the
12119 thread-safety problems that existed, and (b) makes it possible to clean
12120 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
12121 such data would previously have always leaked in application code and
12122 workarounds were in place to make the memory debugging turn a blind eye
12123 to it. Application code that doesn't use this new function will still
12124 leak as before, but their memory debugging output will announce it now
12125 rather than letting it slide.
12127 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
12128 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
12129 has a return value to indicate success or failure.
12133 * Make it possible to replace the underlying "ERR" functions such that the
12134 global state (2 LHASH tables and 2 locks) is only used by the "default"
12135 implementation. This change also adds two functions to "get" and "set"
12136 the implementation prior to it being automatically set the first time
12137 any other ERR function takes place. Ie. an application can call "get",
12138 pass the return value to a module it has just loaded, and that module
12139 can call its own "set" function using that value. This means the
12140 module's "ERR" operations will use (and modify) the error state in the
12141 application and not in its own statically linked copy of OpenSSL code.
12145 * Give DH, DSA, and RSA types their own `*_up_ref()` function to increment
12146 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
12147 the operation, and provides a more encapsulated way for external code
12148 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
12149 to use these functions rather than manually incrementing the counts.
12151 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
12155 * Add EVP test program.
12159 * Add symmetric cipher support to ENGINE. Expect the API to change!
12163 * New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
12164 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
12165 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
12166 These allow a CRL to be built without having to access X509_CRL fields
12167 directly. Modify 'ca' application to use new functions.
12171 * Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
12172 bug workarounds. Rollback attack detection is a security feature.
12173 The problem will only arise on OpenSSL servers when TLSv1 is not
12174 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
12175 Software authors not wanting to support TLSv1 will have special reasons
12176 for their choice and can explicitly enable this option.
12178 *Bodo Moeller, Lutz Jaenicke*
12180 * Rationalise EVP so it can be extended: don't include a union of
12181 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX
12182 (similar to those existing for EVP_CIPHER_CTX).
12187 EVP_MD_CTX_init(&md); /* new function call */
12188 EVP_DigestInit(&md, EVP_sha1());
12189 EVP_DigestUpdate(&md, in, len);
12190 EVP_DigestFinal(&md, out, NULL);
12191 EVP_MD_CTX_cleanup(&md); /* new function call */
12195 * Make DES key schedule conform to the usual scheme, as well as
12196 correcting its structure. This means that calls to DES functions
12197 now have to pass a pointer to a des_key_schedule instead of a
12198 plain des_key_schedule (which was actually always a pointer
12201 des_key_schedule ks;
12203 des_set_key_checked(..., &ks);
12204 des_ncbc_encrypt(..., &ks, ...);
12206 (Note that a later change renames 'des_...' into 'DES_...'.)
12210 * Initial reduction of linker bloat: the use of some functions, such as
12211 PEM causes large amounts of unused functions to be linked in due to
12212 poor organisation. For example pem_all.c contains every PEM function
12213 which has a knock on effect of linking in large amounts of (unused)
12214 ASN1 code. Grouping together similar functions and splitting unrelated
12215 functions prevents this.
12219 * Cleanup of EVP macros.
12223 * Change historical references to `{NID,SN,LN}_des_ede` and ede3 to add the
12224 correct `_ecb suffix`.
12228 * Add initial OCSP responder support to ocsp application. The
12229 revocation information is handled using the text based index
12230 use by the ca application. The responder can either handle
12231 requests generated internally, supplied in files (for example
12232 via a CGI script) or using an internal minimal server.
12236 * Add configuration choices to get zlib compression for TLS.
12240 * Changes to Kerberos SSL for RFC 2712 compliance:
12241 1. Implemented real KerberosWrapper, instead of just using
12242 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
12243 2. Implemented optional authenticator field of KerberosWrapper.
12245 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
12246 and authenticator structs; see crypto/krb5/.
12248 Generalized Kerberos calls to support multiple Kerberos libraries.
12249 *Vern Staats <staatsvr@asc.hpc.mil>, Jeffrey Altman <jaltman@columbia.edu>
12250 via Richard Levitte*
12252 * Cause 'openssl speed' to use fully hard-coded DSA keys as it
12253 already does with RSA. testdsa.h now has 'priv_key/pub_key'
12254 values for each of the key sizes rather than having just
12255 parameters (and 'speed' generating keys each time).
12259 * Speed up EVP routines.
12262 pe 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
12263 s-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
12264 s-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
12265 s-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
12267 s-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
12268 s-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
12269 s-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
12272 s-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
12274 s-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
12278 * Added the OS2-EMX target.
12280 *"Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte*
12282 * Rewrite commands to use `NCONF` routines instead of the old `CONF`.
12283 New functions to support `NCONF` routines in extension code.
12284 New function `CONF_set_nconf()`
12285 to allow functions which take an `NCONF` to also handle the old `LHASH`
12286 structure: this means that the old `CONF` compatible routines can be
12287 retained (in particular w.rt. extensions) without having to duplicate the
12288 code. New function `X509V3_add_ext_nconf_sk()` to add extensions to a stack.
12292 * Enhance the general user interface with mechanisms for inner control
12293 and with possibilities to have yes/no kind of prompts.
12297 * Change all calls to low-level digest routines in the library and
12298 applications to use EVP. Add missing calls to HMAC_cleanup() and
12299 don't assume HMAC_CTX can be copied using memcpy().
12301 *Verdon Walker <VWalker@novell.com>, Steve Henson*
12303 * Add the possibility to control engines through control names but with
12304 arbitrary arguments instead of just a string.
12305 Change the key loaders to take a UI_METHOD instead of a callback
12306 function pointer. NOTE: this breaks binary compatibility with earlier
12307 versions of OpenSSL [engine].
12308 Adapt the nCipher code for these new conditions and add a card insertion
12313 * Enhance the general user interface with mechanisms to better support
12314 dialog box interfaces, application-defined prompts, the possibility
12315 to use defaults (for example default passwords from somewhere else)
12316 and interrupts/cancellations.
12320 * Tidy up PKCS#12 attribute handling. Add support for the CSP name
12321 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
12325 * Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
12326 tidy up some unnecessarily weird code in 'sk_new()').
12328 *Geoff, reported by Diego Tartara <dtartara@novamens.com>*
12330 * Change the key loading routines for ENGINEs to use the same kind
12331 callback (pem_password_cb) as all other routines that need this
12336 * Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
12337 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
12338 than this minimum value is recommended.
12342 * New random seeder for OpenVMS, using the system process statistics
12343 that are easily reachable.
12347 * Windows apparently can't transparently handle global
12348 variables defined in DLLs. Initialisations such as:
12350 const ASN1_ITEM *it = &ASN1_INTEGER_it;
12352 won't compile. This is used by the any applications that need to
12353 declare their own ASN1 modules. This was fixed by adding the option
12354 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
12355 needed for static libraries under Win32.
12359 * New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
12360 setting of purpose and trust fields. New X509_STORE trust and
12361 purpose functions and tidy up setting in other SSL functions.
12365 * Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
12366 structure. These are inherited by X509_STORE_CTX when it is
12367 initialised. This allows various defaults to be set in the
12368 X509_STORE structure (such as flags for CRL checking and custom
12369 purpose or trust settings) for functions which only use X509_STORE_CTX
12370 internally such as S/MIME.
12372 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
12373 trust settings if they are not set in X509_STORE. This allows X509_STORE
12374 purposes and trust (in S/MIME for example) to override any set by default.
12376 Add command line options for CRL checking to smime, s_client and s_server
12381 * Initial CRL based revocation checking. If the CRL checking flag(s)
12382 are set then the CRL is looked up in the X509_STORE structure and
12383 its validity and signature checked, then if the certificate is found
12384 in the CRL the verify fails with a revoked error.
12386 Various new CRL related callbacks added to X509_STORE_CTX structure.
12388 Command line options added to 'verify' application to support this.
12390 This needs some additional work, such as being able to handle multiple
12391 CRLs with different times, extension based lookup (rather than just
12392 by subject name) and ultimately more complete V2 CRL extension
12397 * Add a general user interface API (crypto/ui/). This is designed
12398 to replace things like des_read_password and friends (backward
12399 compatibility functions using this new API are provided).
12400 The purpose is to remove prompting functions from the DES code
12401 section as well as provide for prompting through dialog boxes in
12402 a window system and the like.
12406 * Add "ex_data" support to ENGINE so implementations can add state at a
12407 per-structure level rather than having to store it globally.
12411 * Make it possible for ENGINE structures to be copied when retrieved by
12412 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
12413 This causes the "original" ENGINE structure to act like a template,
12414 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
12415 operational state can be localised to each ENGINE structure, despite the
12416 fact they all share the same "methods". New ENGINE structures returned in
12417 this case have no functional references and the return value is the single
12418 structural reference. This matches the single structural reference returned
12419 by ENGINE_by_id() normally, when it is incremented on the pre-existing
12424 * Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
12425 needs to match any other type at all we need to manually clear the
12430 * Changes to the "openssl engine" utility to include;
12431 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
12432 about an ENGINE's available control commands.
12433 - executing control commands from command line arguments using the
12434 '-pre' and '-post' switches. '-post' is only used if '-t' is
12435 specified and the ENGINE is successfully initialised. The syntax for
12436 the individual commands are colon-separated, for example;
12437 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
12441 * New dynamic control command support for ENGINEs. ENGINEs can now
12442 declare their own commands (numbers), names (strings), descriptions,
12443 and input types for run-time discovery by calling applications. A
12444 subset of these commands are implicitly classed as "executable"
12445 depending on their input type, and only these can be invoked through
12446 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
12447 can be based on user input, config files, etc). The distinction is
12448 that "executable" commands cannot return anything other than a boolean
12449 result and can only support numeric or string input, whereas some
12450 discoverable commands may only be for direct use through
12451 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
12452 pointers, or other custom uses. The "executable" commands are to
12453 support parameterisations of ENGINE behaviour that can be
12454 unambiguously defined by ENGINEs and used consistently across any
12455 OpenSSL-based application. Commands have been added to all the
12456 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
12457 control over shared-library paths without source code alterations.
12461 * Changed all ENGINE implementations to dynamically allocate their
12462 ENGINEs rather than declaring them statically. Apart from this being
12463 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
12464 this also allows the implementations to compile without using the
12465 internal engine_int.h header.
12469 * Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
12470 'const' value. Any code that should be able to modify a RAND_METHOD
12471 should already have non-const pointers to it (ie. they should only
12472 modify their own ones).
12476 * Made a variety of little tweaks to the ENGINE code.
12477 - "atalla" and "ubsec" string definitions were moved from header files
12478 to C code. "nuron" string definitions were placed in variables
12479 rather than hard-coded - allowing parameterisation of these values
12480 later on via ctrl() commands.
12481 - Removed unused "#if 0"'d code.
12482 - Fixed engine list iteration code so it uses ENGINE_free() to release
12483 structural references.
12484 - Constified the RAND_METHOD element of ENGINE structures.
12485 - Constified various get/set functions as appropriate and added
12486 missing functions (including a catch-all ENGINE_cpy that duplicates
12487 all ENGINE values onto a new ENGINE except reference counts/state).
12488 - Removed NULL parameter checks in get/set functions. Setting a method
12489 or function to NULL is a way of cancelling out a previously set
12490 value. Passing a NULL ENGINE parameter is just plain stupid anyway
12491 and doesn't justify the extra error symbols and code.
12492 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
12493 flags from engine_int.h to engine.h.
12494 - Changed prototypes for ENGINE handler functions (init(), finish(),
12495 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
12499 * Implement binary inversion algorithm for BN_mod_inverse in addition
12500 to the algorithm using long division. The binary algorithm can be
12501 used only if the modulus is odd. On 32-bit systems, it is faster
12502 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
12503 roughly 5-15% for 256-bit moduli), so we use it only for moduli
12504 up to 450 bits. In 64-bit environments, the binary algorithm
12505 appears to be advantageous for much longer moduli; here we use it
12506 for moduli up to 2048 bits.
12510 * Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
12511 could not support the combine flag in choice fields.
12515 * Add a 'copy_extensions' option to the 'ca' utility. This copies
12516 extensions from a certificate request to the certificate.
12520 * Allow multiple 'certopt' and 'nameopt' options to be separated
12521 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
12522 file: this allows the display of the certificate about to be
12523 signed to be customised, to allow certain fields to be included
12524 or excluded and extension details. The old system didn't display
12525 multicharacter strings properly, omitted fields not in the policy
12526 and couldn't display additional details such as extensions.
12530 * Function EC_POINTs_mul for multiple scalar multiplication
12531 of an arbitrary number of elliptic curve points
12532 \sum scalars[i]*points[i],
12533 optionally including the generator defined for the EC_GROUP:
12534 scalar*generator + \sum scalars[i]*points[i].
12536 EC_POINT_mul is a simple wrapper function for the typical case
12537 that the point list has just one item (besides the optional
12542 * First EC_METHODs for curves over GF(p):
12544 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
12545 operations and provides various method functions that can also
12546 operate with faster implementations of modular arithmetic.
12548 EC_GFp_mont_method() reuses most functions that are part of
12549 EC_GFp_simple_method, but uses Montgomery arithmetic.
12551 *Bodo Moeller; point addition and point doubling
12552 implementation directly derived from source code provided by
12553 Lenka Fibikova <fibikova@exp-math.uni-essen.de>*
12555 * Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
12556 crypto/ec/ec_lib.c):
12558 Curves are EC_GROUP objects (with an optional group generator)
12559 based on EC_METHODs that are built into the library.
12561 Points are EC_POINT objects based on EC_GROUP objects.
12563 Most of the framework would be able to handle curves over arbitrary
12564 finite fields, but as there are no obvious types for fields other
12565 than GF(p), some functions are limited to that for now.
12569 * Add the -HTTP option to s_server. It is similar to -WWW, but requires
12570 that the file contains a complete HTTP response.
12574 * Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
12575 change the def and num file printf format specifier from "%-40sXXX"
12576 to "%-39s XXX". The latter will always guarantee a space after the
12577 field while the former will cause them to run together if the field
12578 is 40 of more characters long.
12582 * Constify the cipher and digest 'method' functions and structures
12583 and modify related functions to take constant EVP_MD and EVP_CIPHER
12588 * Hide BN_CTX structure details in bn_lcl.h instead of publishing them
12589 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
12593 * Modify `EVP_Digest*()` routines so they now return values. Although the
12594 internal software routines can never fail additional hardware versions
12599 * Clean up crypto/err/err.h and change some error codes to avoid conflicts:
12601 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
12602 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
12605 ERR_R_NESTED_ASN1_ERROR
12607 ERR_R_MISSING_ASN1_EOS
12608 were 4 .. 9, conflicting with
12609 ERR_LIB_RSA (= ERR_R_RSA_LIB)
12611 ERR_LIB_PEM (= ERR_R_PEM_LIB).
12612 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
12614 Add new error code 'ERR_R_INTERNAL_ERROR'.
12618 * Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
12623 * New option '-subj arg' for 'openssl req' and 'openssl ca'. This
12624 sets the subject name for a new request or supersedes the
12625 subject name in a given request. Formats that can be parsed are
12626 'CN=Some Name, OU=myOU, C=IT'
12628 'CN=Some Name/OU=myOU/C=IT'.
12630 Add options '-batch' and '-verbose' to 'openssl req'.
12632 *Massimiliano Pala <madwolf@hackmasters.net>*
12634 * Introduce the possibility to access global variables through
12635 functions on platform were that's the best way to handle exporting
12636 global variables in shared libraries. To enable this functionality,
12637 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
12638 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
12639 is normally done by Configure or something similar).
12641 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
12642 in the source file (foo.c) like this:
12644 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
12645 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
12647 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
12648 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
12650 OPENSSL_DECLARE_GLOBAL(int,foo);
12651 #define foo OPENSSL_GLOBAL_REF(foo)
12652 OPENSSL_DECLARE_GLOBAL(double,bar);
12653 #define bar OPENSSL_GLOBAL_REF(bar)
12655 The #defines are very important, and therefore so is including the
12656 header file everywhere where the defined globals are used.
12658 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
12659 of ASN.1 items, but that structure is a bit different.
12661 The largest change is in util/mkdef.pl which has been enhanced with
12662 better and easier to understand logic to choose which symbols should
12663 go into the Windows .def files as well as a number of fixes and code
12664 cleanup (among others, algorithm keywords are now sorted
12665 lexicographically to avoid constant rewrites).
12669 * In BN_div() keep a copy of the sign of 'num' before writing the
12670 result to 'rm' because if rm==num the value will be overwritten
12671 and produce the wrong result if 'num' is negative: this caused
12672 problems with BN_mod() and BN_nnmod().
12676 * Function OCSP_request_verify(). This checks the signature on an
12677 OCSP request and verifies the signer certificate. The signer
12678 certificate is just checked for a generic purpose and OCSP request
12683 * Add OCSP_check_validity() function to check the validity of OCSP
12684 responses. OCSP responses are prepared in real time and may only
12685 be a few seconds old. Simply checking that the current time lies
12686 between thisUpdate and nextUpdate max reject otherwise valid responses
12687 caused by either OCSP responder or client clock inaccuracy. Instead
12688 we allow thisUpdate and nextUpdate to fall within a certain period of
12689 the current time. The age of the response can also optionally be
12690 checked. Two new options -validity_period and -status_age added to
12695 * If signature or public key algorithm is unrecognized print out its
12696 OID rather that just UNKNOWN.
12700 * Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
12701 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
12702 ID to be generated from the issuer certificate alone which can then be
12703 passed to OCSP_id_issuer_cmp().
12707 * New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
12708 ASN1 modules to export functions returning ASN1_ITEM pointers
12709 instead of the ASN1_ITEM structures themselves. This adds several
12710 new macros which allow the underlying ASN1 function/structure to
12711 be accessed transparently. As a result code should not use ASN1_ITEM
12712 references directly (such as &X509_it) but instead use the relevant
12713 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
12714 use of the new ASN1 code on platforms where exporting structures
12715 is problematical (for example in shared libraries) but exporting
12716 functions returning pointers to structures is not.
12720 * Add support for overriding the generation of SSL/TLS session IDs.
12721 These callbacks can be registered either in an SSL_CTX or per SSL.
12722 The purpose of this is to allow applications to control, if they wish,
12723 the arbitrary values chosen for use as session IDs, particularly as it
12724 can be useful for session caching in multiple-server environments. A
12725 command-line switch for testing this (and any client code that wishes
12726 to use such a feature) has been added to "s_server".
12728 *Geoff Thorpe, Lutz Jaenicke*
12730 * Modify mkdef.pl to recognise and parse preprocessor conditionals
12731 of the form `#if defined(...) || defined(...) || ...` and
12732 `#if !defined(...) && !defined(...) && ...`. This also avoids
12733 the growing number of special cases it was previously handling.
12737 * Make all configuration macros available for application by making
12738 sure they are available in opensslconf.h, by giving them names starting
12739 with `OPENSSL_` to avoid conflicts with other packages and by making
12740 sure e_os2.h will cover all platform-specific cases together with
12742 Additionally, it is now possible to define configuration/platform-
12743 specific names (called "system identities"). In the C code, these
12744 are prefixed with `OPENSSL_SYSNAME_`. e_os2.h will create another
12745 macro with the name beginning with `OPENSSL_SYS_`, which is determined
12746 from `OPENSSL_SYSNAME_*` or compiler-specific macros depending on
12751 * New option -set_serial to 'req' and 'x509' this allows the serial
12752 number to use to be specified on the command line. Previously self
12753 signed certificates were hard coded with serial number 0 and the
12754 CA options of 'x509' had to use a serial number in a file which was
12759 * New options to 'ca' utility to support V2 CRL entry extensions.
12760 Currently CRL reason, invalidity date and hold instruction are
12761 supported. Add new CRL extensions to V3 code and some new objects.
12765 * New function EVP_CIPHER_CTX_set_padding() this is used to
12766 disable standard block padding (aka PKCS#5 padding) in the EVP
12767 API, which was previously mandatory. This means that the data is
12768 not padded in any way and so the total length much be a multiple
12769 of the block size, otherwise an error occurs.
12773 * Initial (incomplete) OCSP SSL support.
12777 * New function OCSP_parse_url(). This splits up a URL into its host,
12778 port and path components: primarily to parse OCSP URLs. New -url
12779 option to ocsp utility.
12783 * New nonce behavior. The return value of OCSP_check_nonce() now
12784 reflects the various checks performed. Applications can decide
12785 whether to tolerate certain situations such as an absent nonce
12786 in a response when one was present in a request: the ocsp application
12787 just prints out a warning. New function OCSP_add1_basic_nonce()
12788 this is to allow responders to include a nonce in a response even if
12789 the request is nonce-less.
12793 * Disable stdin buffering in `load_cert()` (`apps/apps.c`) so that no certs are
12794 skipped when using openssl x509 multiple times on a single input file,
12795 e.g. `(openssl x509 -out cert1; openssl x509 -out cert2) <certs`.
12799 * Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
12800 set string type: to handle setting ASN1_TIME structures. Fix ca
12801 utility to correctly initialize revocation date of CRLs.
12805 * New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
12806 the clients preferred ciphersuites and rather use its own preferences.
12807 Should help to work around M$ SGC (Server Gated Cryptography) bug in
12808 Internet Explorer by ensuring unchanged hash method during stepup.
12809 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
12813 * Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
12814 to aes and add a new 'exist' option to print out symbols that don't
12819 * Additional options to ocsp utility to allow flags to be set and
12820 additional certificates supplied.
12824 * Add the option -VAfile to 'openssl ocsp', so the user can give the
12825 OCSP client a number of certificate to only verify the response
12830 * Update Rijndael code to version 3.0 and change EVP AES ciphers to
12831 handle the new API. Currently only ECB, CBC modes supported. Add new
12834 Add TLS AES ciphersuites as described in RFC3268, "Advanced
12835 Encryption Standard (AES) Ciphersuites for Transport Layer
12836 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were
12837 not enabled by default and were not part of the "ALL" ciphersuite
12838 alias because they were not yet official; they could be
12839 explicitly requested by specifying the "AESdraft" ciphersuite
12840 group alias. In the final release of OpenSSL 0.9.7, the group
12841 alias is called "AES" and is part of "ALL".)
12843 *Ben Laurie, Steve Henson, Bodo Moeller*
12845 * New function OCSP_copy_nonce() to copy nonce value (if present) from
12846 request to response.
12850 * Functions for OCSP responders. OCSP_request_onereq_count(),
12851 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
12852 extract information from a certificate request. OCSP_response_create()
12853 creates a response and optionally adds a basic response structure.
12854 OCSP_basic_add1_status() adds a complete single response to a basic
12855 response and returns the OCSP_SINGLERESP structure just added (to allow
12856 extensions to be included for example). OCSP_basic_add1_cert() adds a
12857 certificate to a basic response and OCSP_basic_sign() signs a basic
12858 response with various flags. New helper functions ASN1_TIME_check()
12859 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
12860 (converts ASN1_TIME to GeneralizedTime).
12864 * Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
12865 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
12866 structure from a certificate. X509_pubkey_digest() digests the public_key
12867 contents: this is used in various key identifiers.
12871 * Make sk_sort() tolerate a NULL argument.
12873 *Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>*
12875 * New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
12876 passed by the function are trusted implicitly. If any of them signed the
12877 response then it is assumed to be valid and is not verified.
12881 * In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
12882 to data. This was previously part of the PKCS7 ASN1 code. This
12883 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
12884 *Steve Henson, reported by Kenneth R. Robinette
12885 <support@securenetterm.com>*
12887 * Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
12888 routines: without these tracing memory leaks is very painful.
12889 Fix leaks in PKCS12 and PKCS7 routines.
12893 * Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
12894 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
12895 effectively meant GeneralizedTime would never be used. Now it
12896 is initialised to -1 but X509_time_adj() now has to check the value
12897 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
12898 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
12899 *Steve Henson, reported by Kenneth R. Robinette
12900 <support@securenetterm.com>*
12902 * Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
12903 result in a zero length in the ASN1_INTEGER structure which was
12904 not consistent with the structure when d2i_ASN1_INTEGER() was used
12905 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
12906 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
12907 where it did not print out a minus for negative ASN1_INTEGER.
12911 * Add summary printout to ocsp utility. The various functions which
12912 convert status values to strings have been renamed to:
12913 OCSP_response_status_str(), OCSP_cert_status_str() and
12914 OCSP_crl_reason_str() and are no longer static. New options
12915 to verify nonce values and to disable verification. OCSP response
12916 printout format cleaned up.
12920 * Add additional OCSP certificate checks. These are those specified
12921 in RFC2560. This consists of two separate checks: the CA of the
12922 certificate being checked must either be the OCSP signer certificate
12923 or the issuer of the OCSP signer certificate. In the latter case the
12924 OCSP signer certificate must contain the OCSP signing extended key
12925 usage. This check is performed by attempting to match the OCSP
12926 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
12927 in the OCSP_CERTID structures of the response.
12931 * Initial OCSP certificate verification added to OCSP_basic_verify()
12932 and related routines. This uses the standard OpenSSL certificate
12933 verify routines to perform initial checks (just CA validity) and
12934 to obtain the certificate chain. Then additional checks will be
12935 performed on the chain. Currently the root CA is checked to see
12936 if it is explicitly trusted for OCSP signing. This is used to set
12937 a root CA as a global signing root: that is any certificate that
12938 chains to that CA is an acceptable OCSP signing certificate.
12942 * New '-extfile ...' option to 'openssl ca' for reading X.509v3
12943 extensions from a separate configuration file.
12944 As when reading extensions from the main configuration file,
12945 the '-extensions ...' option may be used for specifying the
12948 *Massimiliano Pala <madwolf@comune.modena.it>*
12950 * New OCSP utility. Allows OCSP requests to be generated or
12951 read. The request can be sent to a responder and the output
12952 parsed, outputted or printed in text form. Not complete yet:
12953 still needs to check the OCSP response validity.
12957 * New subcommands for 'openssl ca':
12958 `openssl ca -status <serial>` prints the status of the cert with
12959 the given serial number (according to the index file).
12960 `openssl ca -updatedb` updates the expiry status of certificates
12963 *Massimiliano Pala <madwolf@comune.modena.it>*
12965 * New '-newreq-nodes' command option to CA.pl. This is like
12966 '-newreq', but calls 'openssl req' with the '-nodes' option
12967 so that the resulting key is not encrypted.
12969 *Damien Miller <djm@mindrot.org>*
12971 * New configuration for the GNU Hurd.
12973 *Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte*
12975 * Initial code to implement OCSP basic response verify. This
12976 is currently incomplete. Currently just finds the signer's
12977 certificate and verifies the signature on the response.
12981 * New SSLeay_version code SSLEAY_DIR to determine the compiled-in
12982 value of OPENSSLDIR. This is available via the new '-d' option
12983 to 'openssl version', and is also included in 'openssl version -a'.
12987 * Allowing defining memory allocation callbacks that will be given
12988 file name and line number information in additional arguments
12989 (a `const char*` and an int). The basic functionality remains, as
12990 well as the original possibility to just replace malloc(),
12991 realloc() and free() by functions that do not know about these
12992 additional arguments. To register and find out the current
12993 settings for extended allocation functions, the following
12994 functions are provided:
12996 CRYPTO_set_mem_ex_functions
12997 CRYPTO_set_locked_mem_ex_functions
12998 CRYPTO_get_mem_ex_functions
12999 CRYPTO_get_locked_mem_ex_functions
13001 These work the same way as CRYPTO_set_mem_functions and friends.
13002 `CRYPTO_get_[locked_]mem_functions` now writes 0 where such an
13003 extended allocation function is enabled.
13004 Similarly, `CRYPTO_get_[locked_]mem_ex_functions` writes 0 where
13005 a conventional allocation function is enabled.
13007 *Richard Levitte, Bodo Moeller*
13009 * Finish off removing the remaining LHASH function pointer casts.
13010 There should no longer be any prototype-casting required when using
13011 the LHASH abstraction, and any casts that remain are "bugs". See
13012 the callback types and macros at the head of lhash.h for details
13013 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
13017 * Add automatic query of EGD sockets in RAND_poll() for the unix variant.
13018 If /dev/[u]random devices are not available or do not return enough
13019 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
13021 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
13022 /etc/entropy will be queried once each in this sequence, querying stops
13023 when enough entropy was collected without querying more sockets.
13027 * Change the Unix RAND_poll() variant to be able to poll several
13028 random devices, as specified by DEVRANDOM, until a sufficient amount
13029 of data has been collected. We spend at most 10 ms on each file
13030 (select timeout) and read in non-blocking mode. DEVRANDOM now
13031 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
13032 (previously it was just the string "/dev/urandom"), so on typical
13033 platforms the 10 ms delay will never occur.
13034 Also separate out the Unix variant to its own file, rand_unix.c.
13035 For VMS, there's a currently-empty rand_vms.c.
13039 * Move OCSP client related routines to ocsp_cl.c. These
13040 provide utility functions which an application needing
13041 to issue a request to an OCSP responder and analyse the
13042 response will typically need: as opposed to those which an
13043 OCSP responder itself would need which will be added later.
13045 OCSP_request_sign() signs an OCSP request with an API similar
13046 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
13047 response. OCSP_response_get1_basic() extracts basic response
13048 from response. OCSP_resp_find_status(): finds and extracts status
13049 information from an OCSP_CERTID structure (which will be created
13050 when the request structure is built). These are built from lower
13051 level functions which work on OCSP_SINGLERESP structures but
13052 won't normally be used unless the application wishes to examine
13053 extensions in the OCSP response for example.
13055 Replace nonce routines with a pair of functions.
13056 OCSP_request_add1_nonce() adds a nonce value and optionally
13057 generates a random value. OCSP_check_nonce() checks the
13058 validity of the nonce in an OCSP response.
13062 * Change function OCSP_request_add() to OCSP_request_add0_id().
13063 This doesn't copy the supplied OCSP_CERTID and avoids the
13064 need to free up the newly created id. Change return type
13065 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
13066 This can then be used to add extensions to the request.
13067 Deleted OCSP_request_new(), since most of its functionality
13068 is now in OCSP_REQUEST_new() (and the case insensitive name
13069 clash) apart from the ability to set the request name which
13070 will be added elsewhere.
13074 * Update OCSP API. Remove obsolete extensions argument from
13075 various functions. Extensions are now handled using the new
13076 OCSP extension code. New simple OCSP HTTP function which
13077 can be used to send requests and parse the response.
13081 * Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
13082 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
13083 uses the special reorder version of SET OF to sort the attributes
13084 and reorder them to match the encoded order. This resolves a long
13085 standing problem: a verify on a PKCS7 structure just after signing
13086 it used to fail because the attribute order did not match the
13087 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
13088 it uses the received order. This is necessary to tolerate some broken
13089 software that does not order SET OF. This is handled by encoding
13090 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
13091 to produce the required SET OF.
13095 * Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
13096 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
13097 files to get correct declarations of the ASN.1 item variables.
13101 * Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
13102 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
13103 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
13104 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
13105 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
13106 ASN1_ITEM and no wrapper functions.
13110 * New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
13111 replace the old function pointer based I/O routines. Change most of
13112 the `*_d2i_bio()` and `*_d2i_fp()` functions to use these.
13116 * Enhance mkdef.pl to be more accepting about spacing in C preprocessor
13117 lines, recognize more "algorithms" that can be deselected, and make
13118 it complain about algorithm deselection that isn't recognised.
13122 * New ASN1 functions to handle dup, sign, verify, digest, pack and
13123 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
13124 to use new functions. Add NO_ASN1_OLD which can be set to remove
13125 some old style ASN1 functions: this can be used to determine if old
13126 code will still work when these eventually go away.
13130 * New extension functions for OCSP structures, these follow the
13131 same conventions as certificates and CRLs.
13135 * New function X509V3_add1_i2d(). This automatically encodes and
13136 adds an extension. Its behaviour can be customised with various
13137 flags to append, replace or delete. Various wrappers added for
13138 certificates and CRLs.
13142 * Fix to avoid calling the underlying ASN1 print routine when
13143 an extension cannot be parsed. Correct a typo in the
13144 OCSP_SERVICELOC extension. Tidy up print OCSP format.
13148 * Make mkdef.pl parse some of the ASN1 macros and add appropriate
13149 entries for variables.
13153 * Add functionality to `apps/openssl.c` for detecting locking
13154 problems: As the program is single-threaded, all we have
13155 to do is register a locking callback using an array for
13156 storing which locks are currently held by the program.
13160 * Use a lock around the call to CRYPTO_get_ex_new_index() in
13161 SSL_get_ex_data_X509_STORE_idx(), which is used in
13162 ssl_verify_cert_chain() and thus can be called at any time
13163 during TLS/SSL handshakes so that thread-safety is essential.
13164 Unfortunately, the ex_data design is not at all suited
13165 for multi-threaded use, so it probably should be abolished.
13169 * Added Broadcom "ubsec" ENGINE to OpenSSL.
13171 *Broadcom, tweaked and integrated by Geoff Thorpe*
13173 * Move common extension printing code to new function
13174 X509V3_print_extensions(). Reorganise OCSP print routines and
13175 implement some needed OCSP ASN1 functions. Add OCSP extensions.
13179 * New function X509_signature_print() to remove duplication in some
13184 * Add a special meaning when SET OF and SEQUENCE OF flags are both
13185 set (this was treated exactly the same as SET OF previously). This
13186 is used to reorder the STACK representing the structure to match the
13187 encoding. This will be used to get round a problem where a PKCS7
13188 structure which was signed could not be verified because the STACK
13189 order did not reflect the encoded order.
13193 * Reimplement the OCSP ASN1 module using the new code.
13197 * Update the X509V3 code to permit the use of an ASN1_ITEM structure
13198 for its ASN1 operations. The old style function pointers still exist
13199 for now but they will eventually go away.
13203 * Merge in replacement ASN1 code from the ASN1 branch. This almost
13204 completely replaces the old ASN1 functionality with a table driven
13205 encoder and decoder which interprets an ASN1_ITEM structure describing
13206 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
13207 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
13208 has also been converted to the new form.
13212 * Change BN_mod_exp_recp so that negative moduli are tolerated
13213 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
13214 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
13215 for negative moduli.
13219 * Fix BN_uadd and BN_usub: Always return non-negative results instead
13220 of not touching the result's sign bit.
13224 * BN_div bugfix: If the result is 0, the sign (res->neg) must not be
13229 * Changed the LHASH code to use prototypes for callbacks, and created
13230 macros to declare and implement thin (optionally static) functions
13231 that provide type-safety and avoid function pointer casting for the
13232 type-specific callbacks.
13236 * Added Kerberos Cipher Suites to be used with TLS, as written in
13238 *Veers Staats <staatsvr@asc.hpc.mil>,
13239 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte*
13241 * Reformat the FAQ so the different questions and answers can be divided
13242 in sections depending on the subject.
13246 * Have the zlib compression code load ZLIB.DLL dynamically under
13251 * New function BN_mod_sqrt for computing square roots modulo a prime
13252 (using the probabilistic Tonelli-Shanks algorithm unless
13253 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
13254 be handled deterministically).
13256 *Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller*
13258 * Make BN_mod_inverse faster by explicitly handling small quotients
13259 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
13260 512 bits], about 30% for larger ones [1024 or 2048 bits].)
13264 * New function BN_kronecker.
13268 * Fix BN_gcd so that it works on negative inputs; the result is
13269 positive unless both parameters are zero.
13270 Previously something reasonably close to an infinite loop was
13271 possible because numbers could be growing instead of shrinking
13272 in the implementation of Euclid's algorithm.
13276 * Fix BN_is_word() and BN_is_one() macros to take into account the
13277 sign of the number in question.
13279 Fix BN_is_word(a,w) to work correctly for w == 0.
13281 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
13282 because its test if the absolute value of 'a' equals 'w'.
13283 Note that BN_abs_is_word does *not* handle w == 0 reliably;
13284 it exists mostly for use in the implementations of BN_is_zero(),
13285 BN_is_one(), and BN_is_word().
13289 * New function BN_swap.
13293 * Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
13294 the exponentiation functions are more likely to produce reasonable
13295 results on negative inputs.
13299 * Change BN_mod_mul so that the result is always non-negative.
13300 Previously, it could be negative if one of the factors was negative;
13301 I don't think anyone really wanted that behaviour.
13305 * Move `BN_mod_...` functions into new file `crypto/bn/bn_mod.c`
13306 (except for exponentiation, which stays in `crypto/bn/bn_exp.c`,
13307 and `BN_mod_mul_reciprocal`, which stays in `crypto/bn/bn_recp.c`)
13308 and add new functions:
13317 BN_mod_lshift1_quick
13319 BN_mod_lshift_quick
13321 These functions always generate non-negative results.
13323 `BN_nnmod` otherwise is `like BN_mod` (if `BN_mod` computes a remainder `r`
13324 such that `|m| < r < 0`, `BN_nnmod` will output `rem + |m|` instead).
13326 `BN_mod_XXX_quick(r, a, [b,] m)` generates the same result as
13327 `BN_mod_XXX(r, a, [b,] m, ctx)`, but requires that `a` [and `b`]
13328 be reduced modulo `m`.
13330 *Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller*
13333 The following entry accidentally appeared in the CHANGES file
13334 distributed with OpenSSL 0.9.7. The modifications described in
13335 it do *not* apply to OpenSSL 0.9.7.
13337 * Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
13338 was actually never needed) and in BN_mul(). The removal in BN_mul()
13339 required a small change in bn_mul_part_recursive() and the addition
13340 of the functions bn_cmp_part_words(), bn_sub_part_words() and
13341 bn_add_part_words(), which do the same thing as bn_cmp_words(),
13342 bn_sub_words() and bn_add_words() except they take arrays with
13348 * In 'openssl passwd', verify passwords read from the terminal
13349 unless the '-salt' option is used (which usually means that
13350 verification would just waste user's time since the resulting
13351 hash is going to be compared with some given password hash)
13352 or the new '-noverify' option is used.
13354 This is an incompatible change, but it does not affect
13355 non-interactive use of 'openssl passwd' (passwords on the command
13356 line, '-stdin' option, '-in ...' option) and thus should not
13357 cause any problems.
13361 * Remove all references to RSAref, since there's no more need for it.
13365 * Make DSO load along a path given through an environment variable
13366 (SHLIB_PATH) with shl_load().
13370 * Constify the ENGINE code as a result of BIGNUM constification.
13371 Also constify the RSA code and most things related to it. In a
13372 few places, most notable in the depth of the ASN.1 code, ugly
13373 casts back to non-const were required (to be solved at a later
13378 * Make it so the openssl application has all engines loaded by default.
13382 * Constify the BIGNUM routines a little more.
13386 * Add the following functions:
13388 ENGINE_load_cswift()
13390 ENGINE_load_atalla()
13391 ENGINE_load_nuron()
13392 ENGINE_load_builtin_engines()
13394 That way, an application can itself choose if external engines that
13395 are built-in in OpenSSL shall ever be used or not. The benefit is
13396 that applications won't have to be linked with libdl or other dso
13397 libraries unless it's really needed.
13399 Changed 'openssl engine' to load all engines on demand.
13400 Changed the engine header files to avoid the duplication of some
13401 declarations (they differed!).
13405 * 'openssl engine' can now list capabilities.
13409 * Better error reporting in 'openssl engine'.
13413 * Never call load_dh_param(NULL) in s_server.
13417 * Add engine application. It can currently list engines by name and
13418 identity, and test if they are actually available.
13422 * Improve RPM specification file by forcing symbolic linking and making
13423 sure the installed documentation is also owned by root.root.
13425 *Damien Miller <djm@mindrot.org>*
13427 * Give the OpenSSL applications more possibilities to make use of
13428 keys (public as well as private) handled by engines.
13432 * Add OCSP code that comes from CertCo.
13436 * Add VMS support for the Rijndael code.
13440 * Added untested support for Nuron crypto accelerator.
13444 * Add support for external cryptographic devices. This code was
13445 previously distributed separately as the "engine" branch.
13447 *Geoff Thorpe, Richard Levitte*
13449 * Rework the filename-translation in the DSO code. It is now possible to
13450 have far greater control over how a "name" is turned into a filename
13451 depending on the operating environment and any oddities about the
13452 different shared library filenames on each system.
13456 * Support threads on FreeBSD-elf in Configure.
13460 * Fix for SHA1 assembly problem with MASM: it produces
13461 warnings about corrupt line number information when assembling
13462 with debugging information. This is caused by the overlapping
13465 *Bernd Matthes <mainbug@celocom.de>, Steve Henson*
13468 NCONF_get_number() has no error checking at all. As a replacement,
13469 NCONF_get_number_e() is defined (`_e` for "error checking") and is
13470 promoted strongly. The old NCONF_get_number is kept around for
13471 binary backward compatibility.
13472 Make it possible for methods to load from something other than a BIO,
13473 by providing a function pointer that is given a name instead of a BIO.
13474 For example, this could be used to load configuration data from an
13479 * Fix for non blocking accept BIOs. Added new I/O special reason
13480 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
13481 with non blocking I/O was not possible because no retry code was
13482 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
13487 * Added the beginnings of Rijndael support.
13491 * Fix for bug in DirectoryString mask setting. Add support for
13492 X509_NAME_print_ex() in 'req' and X509_print_ex() function
13493 to allow certificate printing to more controllable, additional
13494 'certopt' option to 'x509' to allow new printing options to be
13499 * Clean old EAY MD5 hack from e_os.h.
13503 ### Changes between 0.9.6l and 0.9.6m [17 Mar 2004]
13505 * Fix null-pointer assignment in do_change_cipher_spec() revealed
13506 by using the Codenomicon TLS Test Tool ([CVE-2004-0079])
13508 *Joe Orton, Steve Henson*
13510 ### Changes between 0.9.6k and 0.9.6l [04 Nov 2003]
13512 * Fix additional bug revealed by the NISCC test suite:
13514 Stop bug triggering large recursion when presented with
13515 certain ASN.1 tags ([CVE-2003-0851])
13519 ### Changes between 0.9.6j and 0.9.6k [30 Sep 2003]
13521 * Fix various bugs revealed by running the NISCC test suite:
13523 Stop out of bounds reads in the ASN1 code when presented with
13524 invalid tags (CVE-2003-0543 and CVE-2003-0544).
13526 If verify callback ignores invalid public key errors don't try to check
13527 certificate signature with the NULL public key.
13531 * In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
13532 if the server requested one: as stated in TLS 1.0 and SSL 3.0
13537 * In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
13538 extra data after the compression methods not only for TLS 1.0
13539 but also for SSL 3.0 (as required by the specification).
13541 *Bodo Moeller; problem pointed out by Matthias Loepfe*
13543 * Change X509_certificate_type() to mark the key as exported/exportable
13544 when it's 512 *bits* long, not 512 bytes.
13548 ### Changes between 0.9.6i and 0.9.6j [10 Apr 2003]
13550 * Countermeasure against the Klima-Pokorny-Rosa extension of
13551 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
13552 a protocol version number mismatch like a decryption error
13553 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
13557 * Turn on RSA blinding by default in the default implementation
13558 to avoid a timing attack. Applications that don't want it can call
13559 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
13560 They would be ill-advised to do so in most cases.
13562 *Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller*
13564 * Change RSA blinding code so that it works when the PRNG is not
13565 seeded (in this case, the secret RSA exponent is abused as
13566 an unpredictable seed -- if it is not unpredictable, there
13567 is no point in blinding anyway). Make RSA blinding thread-safe
13568 by remembering the creator's thread ID in rsa->blinding and
13569 having all other threads use local one-time blinding factors
13570 (this requires more computation than sharing rsa->blinding, but
13571 avoids excessive locking; and if an RSA object is not shared
13572 between threads, blinding will still be very fast).
13576 ### Changes between 0.9.6h and 0.9.6i [19 Feb 2003]
13578 * In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
13579 via timing by performing a MAC computation even if incorrect
13580 block cipher padding has been found. This is a countermeasure
13581 against active attacks where the attacker has to distinguish
13582 between bad padding and a MAC verification error. ([CVE-2003-0078])
13584 *Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
13585 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
13586 Martin Vuagnoux (EPFL, Ilion)*
13588 ### Changes between 0.9.6g and 0.9.6h [5 Dec 2002]
13590 * New function OPENSSL_cleanse(), which is used to cleanse a section of
13591 memory from its contents. This is done with a counter that will
13592 place alternating values in each byte. This can be used to solve
13593 two issues: 1) the removal of calls to memset() by highly optimizing
13594 compilers, and 2) cleansing with other values than 0, since those can
13595 be read through on certain media, for example a swap space on disk.
13599 * Bugfix: client side session caching did not work with external caching,
13600 because the session->cipher setting was not restored when reloading
13601 from the external cache. This problem was masked, when
13602 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set.
13603 (Found by Steve Haslam <steve@araqnid.ddts.net>.)
13607 * Fix client_certificate (ssl/s2_clnt.c): The permissible total
13608 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33.
13610 *Zeev Lieber <zeev-l@yahoo.com>*
13612 * Undo an undocumented change introduced in 0.9.6e which caused
13613 repeated calls to OpenSSL_add_all_ciphers() and
13614 OpenSSL_add_all_digests() to be ignored, even after calling
13619 * Change the default configuration reader to deal with last line not
13620 being properly terminated.
13624 * Change X509_NAME_cmp() so it applies the special rules on handling
13625 DN values that are of type PrintableString, as well as RDNs of type
13626 emailAddress where the value has the type ia5String.
13628 *stefank@valicert.com via Richard Levitte*
13630 * Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half
13631 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently
13632 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be
13633 the bitwise-OR of the two for use by the majority of applications
13634 wanting this behaviour, and update the docs. The documented
13635 behaviour and actual behaviour were inconsistent and had been
13636 changing anyway, so this is more a bug-fix than a behavioural
13639 *Geoff Thorpe, diagnosed by Nadav Har'El*
13641 * Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c
13642 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes).
13646 * Fix initialization code race conditions in
13647 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(),
13648 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(),
13649 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(),
13650 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(),
13651 ssl2_get_cipher_by_char(),
13652 ssl3_get_cipher_by_char().
13654 *Patrick McCormick <patrick@tellme.com>, Bodo Moeller*
13656 * Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after
13657 the cached sessions are flushed, as the remove_cb() might use ex_data
13658 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com>
13659 (see [openssl.org #212]).
13661 *Geoff Thorpe, Lutz Jaenicke*
13663 * Fix typo in OBJ_txt2obj which incorrectly passed the content
13664 length, instead of the encoding length to d2i_ASN1_OBJECT.
13668 ### Changes between 0.9.6f and 0.9.6g [9 Aug 2002]
13670 * [In 0.9.6g-engine release:]
13671 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use `_stdcall`).
13673 *Lynn Gazis <lgazis@rainbow.com>*
13675 ### Changes between 0.9.6e and 0.9.6f [8 Aug 2002]
13677 * Fix ASN1 checks. Check for overflow by comparing with LONG_MAX
13678 and get fix the header length calculation.
13679 *Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>,
13680 Alon Kantor <alonk@checkpoint.com> (and others), Steve Henson*
13682 * Use proper error handling instead of 'assertions' in buffer
13683 overflow checks added in 0.9.6e. This prevents DoS (the
13684 assertions could call abort()).
13686 *Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller*
13688 ### Changes between 0.9.6d and 0.9.6e [30 Jul 2002]
13690 * Add various sanity checks to asn1_get_length() to reject
13691 the ASN1 length bytes if they exceed sizeof(long), will appear
13692 negative or the content length exceeds the length of the
13695 *Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>*
13697 * Fix cipher selection routines: ciphers without encryption had no flags
13698 for the cipher strength set and where therefore not handled correctly
13699 by the selection routines (PR #130).
13703 * Fix EVP_dsa_sha macro.
13708 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS
13709 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure
13710 that was added in OpenSSL 0.9.6d.
13712 As the countermeasure turned out to be incompatible with some
13713 broken SSL implementations, the new option is part of SSL_OP_ALL.
13714 SSL_OP_ALL is usually employed when compatibility with weird SSL
13715 implementations is desired (e.g. '-bugs' option to 's_client' and
13716 's_server'), so the new option is automatically set in many
13721 * Changes in security patch:
13723 Changes marked "(CHATS)" were sponsored by the Defense Advanced
13724 Research Projects Agency (DARPA) and Air Force Research Laboratory,
13725 Air Force Materiel Command, USAF, under agreement number
13728 * Add various sanity checks to asn1_get_length() to reject
13729 the ASN1 length bytes if they exceed sizeof(long), will appear
13730 negative or the content length exceeds the length of the
13731 supplied buffer. ([CVE-2002-0659])
13733 *Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>*
13735 * Assertions for various potential buffer overflows, not known to
13736 happen in practice.
13738 *Ben Laurie (CHATS)*
13740 * Various temporary buffers to hold ASCII versions of integers were
13741 too small for 64 bit platforms. ([CVE-2002-0655])
13742 *Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>*
13744 * Remote buffer overflow in SSL3 protocol - an attacker could
13745 supply an oversized session ID to a client. ([CVE-2002-0656])
13747 *Ben Laurie (CHATS)*
13749 * Remote buffer overflow in SSL2 protocol - an attacker could
13750 supply an oversized client master key. ([CVE-2002-0656])
13752 *Ben Laurie (CHATS)*
13754 ### Changes between 0.9.6c and 0.9.6d [9 May 2002]
13756 * Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
13757 encoded as NULL) with id-dsa-with-sha1.
13759 *Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller*
13761 * Check various `X509_...()` return values in `apps/req.c`.
13763 *Nils Larsch <nla@trustcenter.de>*
13765 * Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
13766 an end-of-file condition would erroneously be flagged, when the CRLF
13767 was just at the end of a processed block. The bug was discovered when
13768 processing data through a buffering memory BIO handing the data to a
13769 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
13770 <ptsekov@syntrex.com> and Nedelcho Stanev.
13774 * Implement a countermeasure against a vulnerability recently found
13775 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
13776 before application data chunks to avoid the use of known IVs
13777 with data potentially chosen by the attacker.
13781 * Fix length checks in ssl3_get_client_hello().
13785 * TLS/SSL library bugfix: use s->s3->in_read_app_data differently
13786 to prevent ssl3_read_internal() from incorrectly assuming that
13787 ssl3_read_bytes() found application data while handshake
13788 processing was enabled when in fact s->s3->in_read_app_data was
13789 merely automatically cleared during the initial handshake.
13791 *Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>*
13793 * Fix object definitions for Private and Enterprise: they were not
13794 recognized in their shortname (=lowercase) representation. Extend
13795 obj_dat.pl to issue an error when using undefined keywords instead
13796 of silently ignoring the problem (Svenning Sorensen
13797 <sss@sss.dnsalias.net>).
13801 * Fix DH_generate_parameters() so that it works for 'non-standard'
13802 generators, i.e. generators other than 2 and 5. (Previously, the
13803 code did not properly initialise the 'add' and 'rem' values to
13804 BN_generate_prime().)
13806 In the new general case, we do not insist that 'generator' is
13807 actually a primitive root: This requirement is rather pointless;
13808 a generator of the order-q subgroup is just as good, if not
13813 * Map new X509 verification errors to alerts. Discovered and submitted by
13814 Tom Wu <tom@arcot.com>.
13818 * Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
13819 returning non-zero before the data has been completely received
13820 when using non-blocking I/O.
13822 *Bodo Moeller; problem pointed out by John Hughes*
13824 * Some of the ciphers missed the strength entry (SSL_LOW etc).
13826 *Ben Laurie, Lutz Jaenicke*
13828 * Fix bug in SSL_clear(): bad sessions were not removed (found by
13829 Yoram Zahavi <YoramZ@gilian.com>).
13833 * Add information about CygWin 1.3 and on, and preserve proper
13834 configuration for the versions before that.
13836 *Corinna Vinschen <vinschen@redhat.com> and Richard Levitte*
13838 * Make removal from session cache (SSL_CTX_remove_session()) more robust:
13839 check whether we deal with a copy of a session and do not delete from
13840 the cache in this case. Problem reported by "Izhar Shoshani Levi"
13841 <izhar@checkpoint.com>.
13845 * Do not store session data into the internal session cache, if it
13846 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
13847 flag is set). Proposed by Aslam <aslam@funk.com>.
13851 * Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
13856 * [In 0.9.6d-engine release:]
13857 Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
13859 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
13861 * Add the configuration target linux-s390x.
13863 *Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte*
13865 * The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
13866 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
13867 variable as an indication that a ClientHello message has been
13868 received. As the flag value will be lost between multiple
13869 invocations of ssl3_accept when using non-blocking I/O, the
13870 function may not be aware that a handshake has actually taken
13871 place, thus preventing a new session from being added to the
13874 To avoid this problem, we now set s->new_session to 2 instead of
13875 using a local variable.
13877 *Lutz Jaenicke, Bodo Moeller*
13879 * Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
13880 if the SSL_R_LENGTH_MISMATCH error is detected.
13882 *Geoff Thorpe, Bodo Moeller*
13884 * New 'shared_ldflag' column in Configure platform table.
13888 * Fix EVP_CIPHER_mode macro.
13890 *"Dan S. Camper" <dan@bti.net>*
13892 * Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
13893 type, we must throw them away by setting rr->length to 0.
13895 *D P Chang <dpc@qualys.com>*
13897 ### Changes between 0.9.6b and 0.9.6c [21 dec 2001]
13899 * Fix BN_rand_range bug pointed out by Dominikus Scherkl
13900 <Dominikus.Scherkl@biodata.com>. (The previous implementation
13901 worked incorrectly for those cases where range = `10..._2` and
13902 `3*range` is two bits longer than range.)
13906 * Only add signing time to PKCS7 structures if it is not already
13911 * Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
13912 OBJ_ld_ce should be OBJ_id_ce.
13913 Also some ip-pda OIDs in crypto/objects/objects.txt were
13914 incorrect (cf. RFC 3039).
13916 *Matt Cooper, Frederic Giudicelli, Bodo Moeller*
13918 * Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
13919 returns early because it has nothing to do.
13921 *Andy Schneider <andy.schneider@bjss.co.uk>*
13923 * [In 0.9.6c-engine release:]
13924 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
13926 *Andy Schneider <andy.schneider@bjss.co.uk>*
13928 * [In 0.9.6c-engine release:]
13929 Add support for Cryptographic Appliance's keyserver technology.
13930 (Use engine 'keyclient')
13932 *Cryptographic Appliances and Geoff Thorpe*
13934 * Add a configuration entry for OS/390 Unix. The C compiler 'c89'
13935 is called via tools/c89.sh because arguments have to be
13936 rearranged (all '-L' options must appear before the first object
13939 *Richard Shapiro <rshapiro@abinitio.com>*
13941 * [In 0.9.6c-engine release:]
13942 Add support for Broadcom crypto accelerator cards, backported
13945 *Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox*
13947 * [In 0.9.6c-engine release:]
13948 Add support for SureWare crypto accelerator cards from
13949 Baltimore Technologies. (Use engine 'sureware')
13951 *Baltimore Technologies and Mark Cox*
13953 * [In 0.9.6c-engine release:]
13954 Add support for crypto accelerator cards from Accelerated
13955 Encryption Processing, www.aep.ie. (Use engine 'aep')
13957 *AEP Inc. and Mark Cox*
13959 * Add a configuration entry for gcc on UnixWare.
13961 *Gary Benson <gbenson@redhat.com>*
13963 * Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
13964 messages are stored in a single piece (fixed-length part and
13965 variable-length part combined) and fix various bugs found on the way.
13969 * Disable caching in BIO_gethostbyname(), directly use gethostbyname()
13970 instead. BIO_gethostbyname() does not know what timeouts are
13971 appropriate, so entries would stay in cache even when they have
13973 *Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>*
13975 * Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
13976 faced with a pathologically small ClientHello fragment that does
13977 not contain client_version: Instead of aborting with an error,
13978 simply choose the highest available protocol version (i.e.,
13979 TLS 1.0 unless it is disabled). In practice, ClientHello
13980 messages are never sent like this, but this change gives us
13981 strictly correct behaviour at least for TLS.
13985 * Fix SSL handshake functions and SSL_clear() such that SSL_clear()
13986 never resets s->method to s->ctx->method when called from within
13987 one of the SSL handshake functions.
13989 *Bodo Moeller; problem pointed out by Niko Baric*
13991 * In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
13992 (sent using the client's version number) if client_version is
13993 smaller than the protocol version in use. Also change
13994 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
13995 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
13996 the client will at least see that alert.
14000 * Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
14005 * Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
14006 client receives HelloRequest while in a handshake.
14008 *Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>*
14010 * Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
14011 should end in 'break', not 'goto end' which circumvents various
14012 cleanups done in state SSL_ST_OK. But session related stuff
14013 must be disabled for SSL_ST_OK in the case that we just sent a
14016 Also avoid some overhead by not calling ssl_init_wbio_buffer()
14017 before just sending a HelloRequest.
14019 *Bodo Moeller, Eric Rescorla <ekr@rtfm.com>*
14021 * Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
14022 reveal whether illegal block cipher padding was found or a MAC
14023 verification error occurred. (Neither SSLerr() codes nor alerts
14024 are directly visible to potential attackers, but the information
14025 may leak via logfiles.)
14027 Similar changes are not required for the SSL 2.0 implementation
14028 because the number of padding bytes is sent in clear for SSL 2.0,
14029 and the extra bytes are just ignored. However ssl/s2_pkt.c
14030 failed to verify that the purported number of padding bytes is in
14035 * Add OpenUNIX-8 support including shared libraries
14036 (Boyd Lynn Gerber <gerberb@zenez.com>).
14040 * Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
14041 'wristwatch attack' using huge encoding parameters (cf.
14042 James H. Manger's CRYPTO 2001 paper). Note that the
14043 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
14044 encoding parameters and hence was not vulnerable.
14048 * BN_sqr() bug fix.
14050 *Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>*
14052 * Rabin-Miller test analyses assume uniformly distributed witnesses,
14053 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
14054 followed by modular reduction.
14056 *Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>*
14058 * Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
14059 equivalent based on BN_pseudo_rand() instead of BN_rand().
14063 * s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
14064 This function was broken, as the check for a new client hello message
14065 to handle SGC did not allow these large messages.
14066 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
14070 * Add alert descriptions for TLSv1 to `SSL_alert_desc_string[_long]()`.
14074 * Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
14075 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
14079 * Rework the configuration and shared library support for Tru64 Unix.
14080 The configuration part makes use of modern compiler features and
14081 still retains old compiler behavior for those that run older versions
14082 of the OS. The shared library support part includes a variant that
14083 uses the RPATH feature, and is available through the special
14084 configuration target "alpha-cc-rpath", which will never be selected
14087 *Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte*
14089 * In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
14090 with the same message size as in ssl3_get_certificate_request().
14091 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
14092 messages might inadvertently be reject as too long.
14094 *Petr Lampa <lampa@fee.vutbr.cz>*
14096 * Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
14100 * Modified SSL library such that the verify_callback that has been set
14101 specifically for an SSL object with SSL_set_verify() is actually being
14102 used. Before the change, a verify_callback set with this function was
14103 ignored and the verify_callback() set in the SSL_CTX at the time of
14104 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
14105 to allow the necessary settings.
14109 * Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
14110 explicitly to NULL, as at least on Solaris 8 this seems not always to be
14111 done automatically (in contradiction to the requirements of the C
14112 standard). This made problems when used from OpenSSH.
14116 * In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
14117 dh->length and always used
14119 BN_rand_range(priv_key, dh->p).
14121 BN_rand_range() is not necessary for Diffie-Hellman, and this
14122 specific range makes Diffie-Hellman unnecessarily inefficient if
14123 dh->length (recommended exponent length) is much smaller than the
14124 length of dh->p. We could use BN_rand_range() if the order of
14125 the subgroup was stored in the DH structure, but we only have
14130 BN_rand(priv_key, l, ...)
14132 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
14139 RSA_eay_public_encrypt
14140 RSA_eay_private_decrypt
14141 RSA_eay_private_encrypt (signing)
14142 RSA_eay_public_decrypt (signature verification)
14144 (default implementations for RSA_public_encrypt,
14145 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
14146 always reject numbers >= n.
14150 * In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
14151 to synchronize access to 'locking_thread'. This is necessary on
14152 systems where access to 'locking_thread' (an 'unsigned long'
14153 variable) is not atomic.
14157 * In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
14158 *before* setting the 'crypto_lock_rand' flag. The previous code had
14159 a race condition if 0 is a valid thread ID.
14161 *Travis Vitek <vitek@roguewave.com>*
14163 * Add support for shared libraries under Irix.
14165 *Albert Chin-A-Young <china@thewrittenword.com>*
14167 * Add configuration option to build on Linux on both big-endian and
14168 little-endian MIPS.
14170 *Ralf Baechle <ralf@uni-koblenz.de>*
14172 * Add the possibility to create shared libraries on HP-UX.
14176 ### Changes between 0.9.6a and 0.9.6b [9 Jul 2001]
14178 * Change ssleay_rand_bytes (crypto/rand/md_rand.c)
14179 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
14180 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
14181 PRNG state recovery was possible based on the output of
14182 one PRNG request appropriately sized to gain knowledge on
14183 'md' followed by enough consecutive 1-byte PRNG requests
14184 to traverse all of 'state'.
14186 1. When updating 'md_local' (the current thread's copy of 'md')
14187 during PRNG output generation, hash all of the previous
14188 'md_local' value, not just the half used for PRNG output.
14190 2. Make the number of bytes from 'state' included into the hash
14191 independent from the number of PRNG bytes requested.
14193 The first measure alone would be sufficient to avoid
14194 Markku-Juhani's attack. (Actually it had never occurred
14195 to me that the half of 'md_local' used for chaining was the
14196 half from which PRNG output bytes were taken -- I had always
14197 assumed that the secret half would be used.) The second
14198 measure makes sure that additional data from 'state' is never
14199 mixed into 'md_local' in small portions; this heuristically
14200 further strengthens the PRNG.
14204 * Fix crypto/bn/asm/mips3.s.
14208 * When only the key is given to "enc", the IV is undefined. Print out
14209 an error message in this case.
14213 * Handle special case when X509_NAME is empty in X509 printing routines.
14217 * In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
14218 positive and less than q.
14222 * Don't change `*pointer` in CRYPTO_add_lock() is add_lock_callback is
14223 used: it isn't thread safe and the add_lock_callback should handle
14226 *Paul Rose <Paul.Rose@bridge.com>*
14228 * Verify that incoming data obeys the block size in
14229 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
14235 *Ulf Möller, Bodo Möller*
14237 * The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
14238 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
14239 when fixing the server behaviour for backwards-compatible 'client
14240 hello' messages. (Note that the attack is impractical against
14241 SSL 3.0 and TLS 1.0 anyway because length and version checking
14242 means that the probability of guessing a valid ciphertext is
14243 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
14246 Before 0.9.5, the countermeasure (hide the error by generating a
14247 random 'decryption result') did not work properly because
14248 ERR_clear_error() was missing, meaning that SSL_get_error() would
14249 detect the supposedly ignored error.
14251 Both problems are now fixed.
14255 * In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
14256 (previously it was 1024).
14260 * Fix for compatibility mode trust settings: ignore trust settings
14261 unless some valid trust or reject settings are present.
14265 * Fix for blowfish EVP: its a variable length cipher.
14269 * Fix various bugs related to DSA S/MIME verification. Handle missing
14270 parameters in DSA public key structures and return an error in the
14271 DSA routines if parameters are absent.
14275 * In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
14276 in the current directory if neither $RANDFILE nor $HOME was set.
14277 RAND_file_name() in 0.9.6a returned NULL in this case. This has
14278 caused some confusion to Windows users who haven't defined $HOME.
14279 Thus RAND_file_name() is changed again: e_os.h can define a
14280 DEFAULT_HOME, which will be used if $HOME is not set.
14281 For Windows, we use "C:"; on other platforms, we still require
14282 environment variables.
14284 * Move 'if (!initialized) RAND_poll()' into regions protected by
14285 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
14286 having multiple threads call RAND_poll() concurrently.
14290 * In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
14291 combination of a flag and a thread ID variable.
14292 Otherwise while one thread is in ssleay_rand_bytes (which sets the
14293 flag), *other* threads can enter ssleay_add_bytes without obeying
14294 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
14295 that they do not hold after the first thread unsets add_do_not_lock).
14299 * Change bctest again: '-x' expressions are not available in all
14300 versions of 'test'.
14304 ### Changes between 0.9.6 and 0.9.6a [5 Apr 2001]
14306 * Fix a couple of memory leaks in PKCS7_dataDecode()
14308 *Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>*
14310 * Change Configure and Makefiles to provide EXE_EXT, which will contain
14311 the default extension for executables, if any. Also, make the perl
14312 scripts that use symlink() to test if it really exists and use "cp"
14313 if it doesn't. All this made OpenSSL compilable and installable in
14318 * Fix for asn1_GetSequence() for indefinite length constructed data.
14319 If SEQUENCE is length is indefinite just set c->slen to the total
14320 amount of data available.
14322 *Steve Henson, reported by shige@FreeBSD.org*
14324 *This change does not apply to 0.9.7.*
14326 * Change bctest to avoid here-documents inside command substitution
14327 (workaround for FreeBSD /bin/sh bug).
14328 For compatibility with Ultrix, avoid shell functions (introduced
14329 in the bctest version that searches along $PATH).
14333 * Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
14334 with des_encrypt() defined on some operating systems, like Solaris
14339 * Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
14340 On the Importance of Eliminating Errors in Cryptographic
14341 Computations, J. Cryptology 14 (2001) 2, 101-119,
14342 <http://theory.stanford.edu/~dabo/papers/faults.ps.gz>).
14346 * MIPS assembler BIGNUM division bug fix.
14350 * Disabled incorrect Alpha assembler code.
14354 * Fix PKCS#7 decode routines so they correctly update the length
14355 after reading an EOC for the EXPLICIT tag.
14359 *This change does not apply to 0.9.7.*
14361 * Fix bug in PKCS#12 key generation routines. This was triggered
14362 if a 3DES key was generated with a 0 initial byte. Include
14363 PKCS12_BROKEN_KEYGEN compilation option to retain the old
14364 (but broken) behaviour.
14368 * Enhance bctest to search for a working bc along $PATH and print
14371 *Tim Rice <tim@multitalents.net> via Richard Levitte*
14373 * Fix memory leaks in err.c: free err_data string if necessary;
14374 don't write to the wrong index in ERR_set_error_data.
14378 * Implement ssl23_peek (analogous to ssl23_read), which previously
14383 * Replace rdtsc with `_emit` statements for VC++ version 5.
14385 *Jeremy Cooper <jeremy@baymoo.org>*
14387 * Make it possible to reuse SSLv2 sessions.
14391 * In copy_email() check for >= 0 as a return value for
14392 X509_NAME_get_index_by_NID() since 0 is a valid index.
14394 *Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>*
14396 * Avoid coredump with unsupported or invalid public keys by checking if
14397 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
14398 PKCS7_verify() fails with non detached data.
14402 * Don't use getenv in library functions when run as setuid/setgid.
14403 New function OPENSSL_issetugid().
14407 * Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
14408 due to incorrect handling of multi-threading:
14410 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
14412 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
14414 3. Count how many times MemCheck_off() has been called so that
14415 nested use can be treated correctly. This also avoids
14416 inband-signalling in the previous code (which relied on the
14417 assumption that thread ID 0 is impossible).
14421 * Add "-rand" option also to s_client and s_server.
14425 * Fix CPU detection on Irix 6.x.
14426 *Kurt Hockenbury <khockenb@stevens-tech.edu> and
14427 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>*
14429 * Fix X509_NAME bug which produced incorrect encoding if X509_NAME
14434 *This change does not apply to 0.9.7.*
14436 * Use the cached encoding of an X509_NAME structure rather than
14437 copying it. This is apparently the reason for the libsafe "errors"
14438 but the code is actually correct.
14442 * Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
14443 Bleichenbacher's DSA attack.
14444 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
14445 to be set and top=0 forces the highest bit to be set; top=-1 is new
14446 and leaves the highest bit random.
14448 *Ulf Moeller, Bodo Moeller*
14450 * In the `NCONF_...`-based implementations for `CONF_...` queries
14451 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
14452 a temporary CONF structure with the data component set to NULL
14453 (which gives segmentation faults in lh_retrieve).
14454 Instead, use NULL for the CONF pointer in CONF_get_string and
14455 CONF_get_number (which may use environment variables) and directly
14456 return NULL from CONF_get_section.
14460 * Fix potential buffer overrun for EBCDIC.
14464 * Tolerate nonRepudiation as being valid for S/MIME signing and certSign
14465 keyUsage if basicConstraints absent for a CA.
14469 * Make SMIME_write_PKCS7() write mail header values with a format that
14470 is more generally accepted (no spaces before the semicolon), since
14471 some programs can't parse those values properly otherwise. Also make
14472 sure BIO's that break lines after each write do not create invalid
14477 * Make the CRL encoding routines work with empty SEQUENCE OF. The
14478 macros previously used would not encode an empty SEQUENCE OF
14479 and break the signature.
14483 *This change does not apply to 0.9.7.*
14485 * Zero the premaster secret after deriving the master secret in
14490 * Add some EVP_add_digest_alias registrations (as found in
14491 OpenSSL_add_all_digests()) to SSL_library_init()
14492 aka OpenSSL_add_ssl_algorithms(). This provides improved
14493 compatibility with peers using X.509 certificates
14494 with unconventional AlgorithmIdentifier OIDs.
14498 * Fix for Irix with NO_ASM.
14500 *"Bruce W. Forsberg" <bruce.forsberg@baesystems.com>*
14502 * ./config script fixes.
14504 *Ulf Moeller, Richard Levitte*
14506 * Fix 'openssl passwd -1'.
14510 * Change PKCS12_key_gen_asc() so it can cope with non null
14511 terminated strings whose length is passed in the passlen
14512 parameter, for example from PEM callbacks. This was done
14513 by adding an extra length parameter to asc2uni().
14515 *Steve Henson, reported by <oddissey@samsung.co.kr>*
14517 * Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
14518 call failed, free the DSA structure.
14522 * Fix to uni2asc() to cope with zero length Unicode strings.
14523 These are present in some PKCS#12 files.
14527 * Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
14528 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
14529 when writing a 32767 byte record.
14531 *Bodo Moeller; problem reported by Eric Day <eday@concentric.net>*
14533 * In `RSA_eay_public_{en,ed}crypt` and RSA_eay_mod_exp (rsa_eay.c),
14534 obtain lock CRYPTO_LOCK_RSA before setting `rsa->_method_mod_{n,p,q}`.
14536 (RSA objects have a reference count access to which is protected
14537 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
14538 so they are meant to be shared between threads.)
14539 *Bodo Moeller, Geoff Thorpe; original patch submitted by
14540 "Reddie, Steven" <Steven.Reddie@ca.com>*
14542 * Fix a deadlock in CRYPTO_mem_leaks().
14546 * Use better test patterns in bntest.
14550 * rand_win.c fix for Borland C.
14554 * BN_rshift bugfix for n == 0.
14558 * Add a 'bctest' script that checks for some known 'bc' bugs
14559 so that 'make test' does not abort just because 'bc' is broken.
14563 * Store verify_result within SSL_SESSION also for client side to
14564 avoid potential security hole. (Re-used sessions on the client side
14565 always resulted in verify_result==X509_V_OK, not using the original
14566 result of the server certificate verification.)
14570 * Fix ssl3_pending: If the record in s->s3->rrec is not of type
14571 SSL3_RT_APPLICATION_DATA, return 0.
14572 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
14577 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
14578 releases, have been re-implemented by renaming the previous
14579 implementations of ssl2_read and ssl3_read to ssl2_read_internal
14580 and ssl3_read_internal, respectively, and adding 'peek' parameters
14581 to them. The new ssl[23]_{read,peek} functions are calls to
14582 ssl[23]_read_internal with the 'peek' flag set appropriately.
14583 A 'peek' parameter has also been added to ssl3_read_bytes, which
14584 does the actual work for ssl3_read_internal.
14588 * Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
14589 the method-specific "init()" handler. Also clean up ex_data after
14590 calling the method-specific "finish()" handler. Previously, this was
14591 happening the other way round.
14595 * Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
14596 The previous value, 12, was not always sufficient for BN_mod_exp().
14600 * Make sure that shared libraries get the internal name engine with
14601 the full version number and not just 0. This should mark the
14602 shared libraries as not backward compatible. Of course, this should
14603 be changed again when we can guarantee backward binary compatibility.
14607 * Fix typo in get_cert_by_subject() in by_dir.c
14609 *Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>*
14611 * Rework the system to generate shared libraries:
14613 - Make note of the expected extension for the shared libraries and
14614 if there is a need for symbolic links from for example libcrypto.so.0
14615 to libcrypto.so.0.9.7. There is extended info in Configure for
14618 - Make as few rebuilds of the shared libraries as possible.
14620 - Still avoid linking the OpenSSL programs with the shared libraries.
14622 - When installing, install the shared libraries separately from the
14627 * Fix SSL_CTX_set_read_ahead macro to actually use its argument.
14629 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
14630 and not in SSL_clear because the latter is also used by the
14631 accept/connect functions; previously, the settings made by
14632 SSL_set_read_ahead would be lost during the handshake.
14634 *Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>*
14636 * Correct util/mkdef.pl to be selective about disabled algorithms.
14637 Previously, it would create entries for disabled algorithms no
14642 * Added several new manual pages for SSL_* function.
14646 ### Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
14648 * In ssl23_get_client_hello, generate an error message when faced
14649 with an initial SSL 3.0/TLS record that is too small to contain the
14650 first two bytes of the ClientHello message, i.e. client_version.
14651 (Note that this is a pathologic case that probably has never happened
14652 in real life.) The previous approach was to use the version number
14653 from the record header as a substitute; but our protocol choice
14654 should not depend on that one because it is not authenticated
14655 by the Finished messages.
14659 * More robust randomness gathering functions for Windows.
14661 *Jeffrey Altman <jaltman@columbia.edu>*
14663 * For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
14664 not set then we don't setup the error code for issuer check errors
14665 to avoid possibly overwriting other errors which the callback does
14666 handle. If an application does set the flag then we assume it knows
14667 what it is doing and can handle the new informational codes
14672 * Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
14673 a general "ANY" type, as such it should be able to decode anything
14674 including tagged types. However it didn't check the class so it would
14675 wrongly interpret tagged types in the same way as their universal
14676 counterpart and unknown types were just rejected. Changed so that the
14677 tagged and unknown types are handled in the same way as a SEQUENCE:
14678 that is the encoding is stored intact. There is also a new type
14679 "V_ASN1_OTHER" which is used when the class is not universal, in this
14680 case we have no idea what the actual type is so we just lump them all
14685 * On VMS, stdout may very well lead to a file that is written to
14686 in a record-oriented fashion. That means that every write() will
14687 write a separate record, which will be read separately by the
14688 programs trying to read from it. This can be very confusing.
14690 The solution is to put a BIO filter in the way that will buffer
14691 text until a linefeed is reached, and then write everything a
14692 line at a time, so every record written will be an actual line,
14693 not chunks of lines and not (usually doesn't happen, but I've
14694 seen it once) several lines in one record. BIO_f_linebuffer() is
14697 Currently, it's a VMS-only method, because that's where it has
14698 been tested well enough.
14702 * Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
14703 it can return incorrect results.
14704 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
14705 but it was in 0.9.6-beta[12].)
14709 * Disable the check for content being present when verifying detached
14710 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
14711 include zero length content when signing messages.
14715 * New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
14716 BIO_ctrl (for BIO pairs).
14720 * Add DSO method for VMS.
14724 * Bug fix: Montgomery multiplication could produce results with the
14729 * Add RPM specification openssl.spec and modify it to build three
14730 packages. The default package contains applications, application
14731 documentation and run-time libraries. The devel package contains
14732 include files, static libraries and function documentation. The
14733 doc package contains the contents of the doc directory. The original
14734 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
14738 * Add a large number of documentation files for many SSL routines.
14740 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>*
14742 * Add a configuration entry for Sony News 4.
14744 *NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>*
14746 * Don't set the two most significant bits to one when generating a
14747 random number < q in the DSA library.
14751 * New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
14752 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
14753 the underlying transport is blocking) if a handshake took place.
14754 (The default behaviour is needed by applications such as s_client
14755 and s_server that use select() to determine when to use SSL_read;
14756 but for applications that know in advance when to expect data, it
14757 just makes things more complicated.)
14761 * Add RAND_egd_bytes(), which gives control over the number of bytes read
14766 * Add a few more EBCDIC conditionals that make `req` and `x509`
14767 work better on such systems.
14769 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
14771 * Add two demo programs for PKCS12_parse() and PKCS12_create().
14772 Update PKCS12_parse() so it copies the friendlyName and the
14773 keyid to the certificates aux info.
14777 * Fix bug in PKCS7_verify() which caused an infinite loop
14778 if there was more than one signature.
14780 *Sven Uszpelkat <su@celocom.de>*
14782 * Major change in util/mkdef.pl to include extra information
14783 about each symbol, as well as presenting variables as well
14784 as functions. This change means that there's n more need
14785 to rebuild the .num files when some algorithms are excluded.
14789 * Allow the verify time to be set by an application,
14790 rather than always using the current time.
14794 * Phase 2 verify code reorganisation. The certificate
14795 verify code now looks up an issuer certificate by a
14796 number of criteria: subject name, authority key id
14797 and key usage. It also verifies self signed certificates
14798 by the same criteria. The main comparison function is
14799 X509_check_issued() which performs these checks.
14801 Lot of changes were necessary in order to support this
14802 without completely rewriting the lookup code.
14804 Authority and subject key identifier are now cached.
14806 The LHASH 'certs' is X509_STORE has now been replaced
14807 by a STACK_OF(X509_OBJECT). This is mainly because an
14808 LHASH can't store or retrieve multiple objects with
14809 the same hash value.
14811 As a result various functions (which were all internal
14812 use only) have changed to handle the new X509_STORE
14813 structure. This will break anything that messed round
14814 with X509_STORE internally.
14816 The functions X509_STORE_add_cert() now checks for an
14817 exact match, rather than just subject name.
14819 The X509_STORE API doesn't directly support the retrieval
14820 of multiple certificates matching a given criteria, however
14821 this can be worked round by performing a lookup first
14822 (which will fill the cache with candidate certificates)
14823 and then examining the cache for matches. This is probably
14824 the best we can do without throwing out X509_LOOKUP
14825 entirely (maybe later...).
14827 The X509_VERIFY_CTX structure has been enhanced considerably.
14829 All certificate lookup operations now go via a get_issuer()
14830 callback. Although this currently uses an X509_STORE it
14831 can be replaced by custom lookups. This is a simple way
14832 to bypass the X509_STORE hackery necessary to make this
14833 work and makes it possible to use more efficient techniques
14834 in future. A very simple version which uses a simple
14835 STACK for its trusted certificate store is also provided
14836 using X509_STORE_CTX_trusted_stack().
14838 The verify_cb() and verify() callbacks now have equivalents
14839 in the X509_STORE_CTX structure.
14841 X509_STORE_CTX also has a 'flags' field which can be used
14842 to customise the verify behaviour.
14846 * Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
14847 excludes S/MIME capabilities.
14851 * When a certificate request is read in keep a copy of the
14852 original encoding of the signed data and use it when outputting
14853 again. Signatures then use the original encoding rather than
14854 a decoded, encoded version which may cause problems if the
14855 request is improperly encoded.
14859 * For consistency with other BIO_puts implementations, call
14860 buffer_write(b, ...) directly in buffer_puts instead of calling
14863 In BIO_puts, increment b->num_write as in BIO_write.
14865 *Peter.Sylvester@EdelWeb.fr*
14867 * Fix BN_mul_word for the case where the word is 0. (We have to use
14868 BN_zero, we may not return a BIGNUM with an array consisting of
14869 words set to zero.)
14873 * Avoid calling abort() from within the library when problems are
14874 detected, except if preprocessor symbols have been defined
14875 (such as REF_CHECK, BN_DEBUG etc.).
14879 * New openssl application 'rsautl'. This utility can be
14880 used for low-level RSA operations. DER public key
14881 BIO/fp routines also added.
14885 * New Configure entry and patches for compiling on QNX 4.
14887 *Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>*
14889 * A demo state-machine implementation was sponsored by
14890 Nuron (<http://www.nuron.com/>) and is now available in
14891 demos/state_machine.
14895 * New options added to the 'dgst' utility for signature
14896 generation and verification.
14900 * Unrecognized PKCS#7 content types are now handled via a
14901 catch all ASN1_TYPE structure. This allows unsupported
14902 types to be stored as a "blob" and an application can
14903 encode and decode it manually.
14907 * Fix various signed/unsigned issues to make a_strex.c
14908 compile under VC++.
14910 *Oscar Jacobsson <oscar.jacobsson@celocom.com>*
14912 * ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
14913 length if passed a buffer. ASN1_INTEGER_to_BN failed
14914 if passed a NULL BN and its argument was negative.
14916 *Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>*
14918 * Modification to PKCS#7 encoding routines to output definite
14919 length encoding. Since currently the whole structures are in
14920 memory there's not real point in using indefinite length
14921 constructed encoding. However if OpenSSL is compiled with
14922 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
14926 * Added BIO_vprintf() and BIO_vsnprintf().
14930 * Added more prefixes to parse for in the strings written
14931 through a logging bio, to cover all the levels that are available
14932 through syslog. The prefixes are now:
14934 PANIC, EMERG, EMR => LOG_EMERG
14935 ALERT, ALR => LOG_ALERT
14936 CRIT, CRI => LOG_CRIT
14937 ERROR, ERR => LOG_ERR
14938 WARNING, WARN, WAR => LOG_WARNING
14939 NOTICE, NOTE, NOT => LOG_NOTICE
14940 INFO, INF => LOG_INFO
14941 DEBUG, DBG => LOG_DEBUG
14943 and as before, if none of those prefixes are present at the
14944 beginning of the string, LOG_ERR is chosen.
14946 On Win32, the `LOG_*` levels are mapped according to this:
14948 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
14949 LOG_WARNING => EVENTLOG_WARNING_TYPE
14950 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
14954 * Made it possible to reconfigure with just the configuration
14955 argument "reconf" or "reconfigure". The command line arguments
14956 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
14957 and are retrieved from there when reconfiguring.
14963 *Assar Westerlund <assar@sics.se>, Richard Levitte*
14965 * Add the arguments -CAfile and -CApath to the pkcs12 utility.
14969 * The obj_dat.pl script was messing up the sorting of object
14970 names. The reason was that it compared the quoted version
14971 of strings as a result "OCSP" > "OCSP Signing" because
14972 " > SPACE. Changed script to store unquoted versions of
14973 names and add quotes on output. It was also omitting some
14974 names from the lookup table if they were given a default
14975 value (that is if SN is missing it is given the same
14976 value as LN and vice versa), these are now added on the
14977 grounds that if an object has a name we should be able to
14978 look it up. Finally added warning output when duplicate
14979 short or long names are found.
14983 * Changes needed for Tandem NSK.
14985 *Scott Uroff <scott@xypro.com>*
14987 * Fix SSL 2.0 rollback checking: Due to an off-by-one error in
14988 RSA_padding_check_SSLv23(), special padding was never detected
14989 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
14990 version rollback attacks was not effective.
14992 In s23_clnt.c, don't use special rollback-attack detection padding
14993 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
14994 client; similarly, in s23_srvr.c, don't do the rollback check if
14995 SSL 2.0 is the only protocol enabled in the server.
14999 * Make it possible to get hexdumps of unprintable data with 'openssl
15000 asn1parse'. By implication, the functions ASN1_parse_dump() and
15001 BIO_dump_indent() are added.
15005 * New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
15006 these print out strings and name structures based on various
15007 flags including RFC2253 support and proper handling of
15008 multibyte characters. Added options to the 'x509' utility
15009 to allow the various flags to be set.
15013 * Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
15014 Also change the functions X509_cmp_current_time() and
15015 X509_gmtime_adj() work with an ASN1_TIME structure,
15016 this will enable certificates using GeneralizedTime in validity
15017 dates to be checked.
15021 * Make the NEG_PUBKEY_BUG code (which tolerates invalid
15022 negative public key encodings) on by default,
15023 NO_NEG_PUBKEY_BUG can be set to disable it.
15027 * New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
15028 content octets. An i2c_ASN1_OBJECT is unnecessary because
15029 the encoding can be trivially obtained from the structure.
15033 * crypto/err.c locking bugfix: Use write locks (`CRYPTO_w_[un]lock`),
15034 not read locks (`CRYPTO_r_[un]lock`).
15038 * A first attempt at creating official support for shared
15039 libraries through configuration. I've kept it so the
15040 default is static libraries only, and the OpenSSL programs
15041 are always statically linked for now, but there are
15042 preparations for dynamic linking in place.
15043 This has been tested on Linux and Tru64.
15047 * Randomness polling function for Win9x, as described in:
15048 Peter Gutmann, Software Generation of Practically Strong
15053 * Fix so PRNG is seeded in req if using an already existing
15058 * New options to smime application. -inform and -outform
15059 allow alternative formats for the S/MIME message including
15060 PEM and DER. The -content option allows the content to be
15061 specified separately. This should allow things like Netscape
15062 form signing output easier to verify.
15066 * Fix the ASN1 encoding of tags using the 'long form'.
15070 * New ASN1 functions, `i2c_*` and `c2i_*` for INTEGER and BIT
15071 STRING types. These convert content octets to and from the
15072 underlying type. The actual tag and length octets are
15073 already assumed to have been read in and checked. These
15074 are needed because all other string types have virtually
15075 identical handling apart from the tag. By having versions
15076 of the ASN1 functions that just operate on content octets
15077 IMPLICIT tagging can be handled properly. It also allows
15078 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
15079 and ASN1_INTEGER are identical apart from the tag.
15083 * Change the handling of OID objects as follows:
15085 - New object identifiers are inserted in objects.txt, following
15086 the syntax given in [crypto/objects/README.md](crypto/objects/README.md).
15087 - objects.pl is used to process obj_mac.num and create a new
15089 - obj_dat.pl is used to create a new obj_dat.h, using the data in
15092 This is currently kind of a hack, and the perl code in objects.pl
15093 isn't very elegant, but it works as I intended. The simplest way
15094 to check that it worked correctly is to look in obj_dat.h and
15095 check the array nid_objs and make sure the objects haven't moved
15096 around (this is important!). Additions are OK, as well as
15097 consistent name changes.
15101 * Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
15105 * Addition of the command line parameter '-rand file' to 'openssl req'.
15106 The given file adds to whatever has already been seeded into the
15107 random pool through the RANDFILE configuration file option or
15108 environment variable, or the default random state file.
15112 * mkstack.pl now sorts each macro group into lexical order.
15113 Previously the output order depended on the order the files
15114 appeared in the directory, resulting in needless rewriting
15119 * Patches to make OpenSSL compile under Win32 again. Mostly
15120 work arounds for the VC++ problem that it treats func() as
15121 func(void). Also stripped out the parts of mkdef.pl that
15122 added extra typesafe functions: these no longer exist.
15126 * Reorganisation of the stack code. The macros are now all
15127 collected in safestack.h . Each macro is defined in terms of
15128 a "stack macro" of the form `SKM_<name>(type, a, b)`. The
15129 DEBUG_SAFESTACK is now handled in terms of function casts,
15130 this has the advantage of retaining type safety without the
15131 use of additional functions. If DEBUG_SAFESTACK is not defined
15132 then the non typesafe macros are used instead. Also modified the
15133 mkstack.pl script to handle the new form. Needs testing to see
15134 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
15135 the default if no major problems. Similar behaviour for ASN1_SET_OF
15136 and PKCS12_STACK_OF.
15140 * When some versions of IIS use the 'NET' form of private key the
15141 key derivation algorithm is different. Normally MD5(password) is
15142 used as a 128 bit RC4 key. In the modified case
15143 MD5(MD5(password) + "SGCKEYSALT") is used instead. Added some
15144 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
15145 as the old Netscape_RSA functions except they have an additional
15146 'sgckey' parameter which uses the modified algorithm. Also added
15147 an -sgckey command line option to the rsa utility. Thanks to
15148 Adrian Peck <bertie@ncipher.com> for posting details of the modified
15149 algorithm to openssl-dev.
15153 * The evp_local.h macros were using 'c.##kname' which resulted in
15154 invalid expansion on some systems (SCO 5.0.5 for example).
15155 Corrected to 'c.kname'.
15157 *Phillip Porch <root@theporch.com>*
15159 * New X509_get1_email() and X509_REQ_get1_email() functions that return
15160 a STACK of email addresses from a certificate or request, these look
15161 in the subject name and the subject alternative name extensions and
15162 omit any duplicate addresses.
15166 * Re-implement BN_mod_exp2_mont using independent (and larger) windows.
15167 This makes DSA verification about 2 % faster.
15171 * Increase maximum window size in `BN_mod_exp_...` to 6 bits instead of 5
15172 (meaning that now 2^5 values will be precomputed, which is only 4 KB
15173 plus overhead for 1024 bit moduli).
15174 This makes exponentiations about 0.5 % faster for 1024 bit
15175 exponents (as measured by "openssl speed rsa2048").
15179 * Rename memory handling macros to avoid conflicts with other
15181 Malloc => OPENSSL_malloc
15182 Malloc_locked => OPENSSL_malloc_locked
15183 Realloc => OPENSSL_realloc
15184 Free => OPENSSL_free
15188 * New function BN_mod_exp_mont_word for small bases (roughly 15%
15189 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
15193 * CygWin32 support.
15195 *John Jarvie <jjarvie@newsguy.com>*
15197 * The type-safe stack code has been rejigged. It is now only compiled
15198 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
15199 by default all type-specific stack functions are "#define"d back to
15200 standard stack functions. This results in more streamlined output
15201 but retains the type-safety checking possibilities of the original
15206 * The STACK code has been cleaned up, and certain type declarations
15207 that didn't make a lot of sense have been brought in line. This has
15208 also involved a cleanup of sorts in safestack.h to more correctly
15209 map type-safe stack functions onto their plain stack counterparts.
15210 This work has also resulted in a variety of "const"ifications of
15211 lots of the code, especially `_cmp` operations which should normally
15212 be prototyped with "const" parameters anyway.
15216 * When generating bytes for the first time in md_rand.c, 'stir the pool'
15217 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
15218 (The PRNG state consists of two parts, the large pool 'state' and 'md',
15219 where all of 'md' is used each time the PRNG is used, but 'state'
15220 is used only indexed by a cyclic counter. As entropy may not be
15221 well distributed from the beginning, 'md' is important as a
15222 chaining variable. However, the output function chains only half
15223 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
15224 all of 'md', and seeding with STATE_SIZE dummy bytes will result
15225 in all of 'state' being rewritten, with the new values depending
15226 on virtually all of 'md'. This overcomes the 80 bit limitation.)
15230 * In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
15231 the handshake is continued after ssl_verify_cert_chain();
15232 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
15233 can lead to 'unexplainable' connection aborts later.
15235 *Bodo Moeller; problem tracked down by Lutz Jaenicke*
15237 * Major EVP API cipher revision.
15238 Add hooks for extra EVP features. This allows various cipher
15239 parameters to be set in the EVP interface. Support added for variable
15240 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
15241 setting of RC2 and RC5 parameters.
15243 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
15246 Remove lots of duplicated code from the EVP library. For example *every*
15247 cipher init() function handles the 'iv' in the same way according to the
15248 cipher mode. They also all do nothing if the 'key' parameter is NULL and
15249 for CFB and OFB modes they zero ctx->num.
15251 New functionality allows removal of S/MIME code RC2 hack.
15253 Most of the routines have the same form and so can be declared in terms
15256 By shifting this to the top level EVP_CipherInit() it can be removed from
15257 all individual ciphers. If the cipher wants to handle IVs or keys
15258 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
15261 Change lots of functions like EVP_EncryptUpdate() to now return a
15262 value: although software versions of the algorithms cannot fail
15263 any installed hardware versions can.
15267 * Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
15268 this option is set, tolerate broken clients that send the negotiated
15269 protocol version number instead of the requested protocol version
15274 * Call dh_tmp_cb (set by `..._TMP_DH_CB`) with correct 'is_export' flag;
15275 i.e. non-zero for export ciphersuites, zero otherwise.
15276 Previous versions had this flag inverted, inconsistent with
15277 rsa_tmp_cb (..._TMP_RSA_CB).
15279 *Bodo Moeller; problem reported by Amit Chopra*
15281 * Add missing DSA library text string. Work around for some IIS
15282 key files with invalid SEQUENCE encoding.
15286 * Add a document (doc/standards.txt) that list all kinds of standards
15287 and so on that are implemented in OpenSSL.
15291 * Enhance c_rehash script. Old version would mishandle certificates
15292 with the same subject name hash and wouldn't handle CRLs at all.
15293 Added -fingerprint option to crl utility, to support new c_rehash
15298 * Eliminate non-ANSI declarations in crypto.h and stack.h.
15302 * Fix for SSL server purpose checking. Server checking was
15303 rejecting certificates which had extended key usage present
15304 but no ssl client purpose.
15306 *Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>*
15308 * Make PKCS#12 code work with no password. The PKCS#12 spec
15309 is a little unclear about how a blank password is handled.
15310 Since the password in encoded as a BMPString with terminating
15311 double NULL a zero length password would end up as just the
15312 double NULL. However no password at all is different and is
15313 handled differently in the PKCS#12 key generation code. NS
15314 treats a blank password as zero length. MSIE treats it as no
15315 password on export: but it will try both on import. We now do
15316 the same: PKCS12_parse() tries zero length and no password if
15317 the password is set to "" or NULL (NULL is now a valid password:
15318 it wasn't before) as does the pkcs12 application.
15322 * Bugfixes in `apps/x509.c`: Avoid a memory leak; and don't use
15323 perror when PEM_read_bio_X509_REQ fails, the error message must
15324 be obtained from the error queue.
15328 * Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
15329 it in ERR_remove_state if appropriate, and change ERR_get_state
15330 accordingly to avoid race conditions (this is necessary because
15331 thread_hash is no longer constant once set).
15335 * Bugfix for linux-elf makefile.one.
15339 * RSA_get_default_method() will now cause a default
15340 RSA_METHOD to be chosen if one doesn't exist already.
15341 Previously this was only set during a call to RSA_new()
15342 or RSA_new_method(NULL) meaning it was possible for
15343 RSA_get_default_method() to return NULL.
15347 * Added native name translation to the existing DSO code
15348 that will convert (if the flag to do so is set) filenames
15349 that are sufficiently small and have no path information
15350 into a canonical native form. Eg. "blah" converted to
15351 "libblah.so" or "blah.dll" etc.
15355 * New function ERR_error_string_n(e, buf, len) which is like
15356 ERR_error_string(e, buf), but writes at most 'len' bytes
15357 including the 0 terminator. For ERR_error_string_n, 'buf'
15360 *Damien Miller <djm@mindrot.org>, Bodo Moeller*
15362 * CONF library reworked to become more general. A new CONF
15363 configuration file reader "class" is implemented as well as a
15364 new functions (`NCONF_*`, for "New CONF") to handle it. The now
15365 old `CONF_*` functions are still there, but are reimplemented to
15366 work in terms of the new functions. Also, a set of functions
15367 to handle the internal storage of the configuration data is
15368 provided to make it easier to write new configuration file
15369 reader "classes" (I can definitely see something reading a
15370 configuration file in XML format, for example), called `_CONF_*`,
15371 or "the configuration storage API"...
15373 The new configuration file reading functions are:
15375 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
15376 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
15378 NCONF_default, NCONF_WIN32
15380 NCONF_dump_fp, NCONF_dump_bio
15382 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
15383 NCONF_new creates a new CONF object. This works in the same way
15384 as other interfaces in OpenSSL, like the BIO interface.
15385 `NCONF_dump_*` dump the internal storage of the configuration file,
15386 which is useful for debugging. All other functions take the same
15387 arguments as the old `CONF_*` functions with the exception of the
15388 first that must be a `CONF *` instead of a `LHASH *`.
15390 To make it easier to use the new classes with the old `CONF_*` functions,
15391 the function CONF_set_default_method is provided.
15395 * Add '-tls1' option to 'openssl ciphers', which was already
15396 mentioned in the documentation but had not been implemented.
15397 (This option is not yet really useful because even the additional
15398 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
15402 * Initial DSO code added into libcrypto for letting OpenSSL (and
15403 OpenSSL-based applications) load shared libraries and bind to
15404 them in a portable way.
15406 *Geoff Thorpe, with contributions from Richard Levitte*
15408 ### Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
15410 * Make sure _lrotl and _lrotr are only used with MSVC.
15412 * Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
15413 (the default implementation of RAND_status).
15415 * Rename openssl x509 option '-crlext', which was added in 0.9.5,
15416 to '-clrext' (= clear extensions), as intended and documented.
15417 *Bodo Moeller; inconsistency pointed out by Michael Attili
15418 <attili@amaxo.com>*
15420 * Fix for HMAC. It wasn't zeroing the rest of the block if the key length
15421 was larger than the MD block size.
15423 *Steve Henson, pointed out by Yost William <YostW@tce.com>*
15425 * Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
15426 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
15427 using the passed key: if the passed key was a private key the result
15428 of X509_print(), for example, would be to print out all the private key
15433 * des_quad_cksum() byte order bug fix.
15434 *Ulf Möller, using the problem description in krb4-0.9.7, where
15435 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>*
15437 * Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
15440 *Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>*
15442 * For easily testing in shell scripts whether some command
15443 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
15444 returns with exit code 0 iff no command of the given name is available.
15445 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
15446 the output goes to stdout and nothing is printed to stderr.
15447 Additional arguments are always ignored.
15449 Since for each cipher there is a command of the same name,
15450 the 'no-cipher' compilation switches can be tested this way.
15452 ('openssl no-XXX' is not able to detect pseudo-commands such
15453 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
15457 * Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
15461 * For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
15462 is set; it will be thrown away anyway because each handshake creates
15464 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
15465 to parameters -- in previous versions (since OpenSSL 0.9.3) the
15466 'default key' from SSL_CTX_set_tmp_dh would always be lost, meaning
15467 you effectively got SSL_OP_SINGLE_DH_USE when using this macro.
15471 * New s_client option -ign_eof: EOF at stdin is ignored, and
15472 'Q' and 'R' lose their special meanings (quit/renegotiate).
15473 This is part of what -quiet does; unlike -quiet, -ign_eof
15474 does not suppress any output.
15478 * Add compatibility options to the purpose and trust code. The
15479 purpose X509_PURPOSE_ANY is "any purpose" which automatically
15480 accepts a certificate or CA, this was the previous behaviour,
15481 with all the associated security issues.
15483 X509_TRUST_COMPAT is the old trust behaviour: only and
15484 automatically trust self signed roots in certificate store. A
15485 new trust setting X509_TRUST_DEFAULT is used to specify that
15486 a purpose has no associated trust setting and it should instead
15487 use the value in the default purpose.
15491 * Fix the PKCS#8 DSA private key code so it decodes keys again
15492 and fix a memory leak.
15496 * In util/mkerr.pl (which implements 'make errors'), preserve
15497 reason strings from the previous version of the .c file, as
15498 the default to have only downcase letters (and digits) in
15499 automatically generated reasons codes is not always appropriate.
15503 * In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
15504 using strerror. Previously, ERR_reason_error_string() returned
15505 library names as reason strings for SYSerr; but SYSerr is a special
15506 case where small numbers are errno values, not library numbers.
15510 * Add '-dsaparam' option to 'openssl dhparam' application. This
15511 converts DSA parameters into DH parameters. (When creating parameters,
15512 DSA_generate_parameters is used.)
15516 * Include 'length' (recommended exponent length) in C code generated
15517 by 'openssl dhparam -C'.
15521 * The second argument to set_label in perlasm was already being used
15522 so couldn't be used as a "file scope" flag. Moved to third argument
15527 * In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
15528 instead of RAND_bytes for encryption IVs and salts.
15532 * Include RAND_status() into RAND_METHOD instead of implementing
15533 it only for md_rand.c Otherwise replacing the PRNG by calling
15534 RAND_set_rand_method would be impossible.
15538 * Don't let DSA_generate_key() enter an infinite loop if the random
15539 number generation fails.
15543 * New 'rand' application for creating pseudo-random output.
15547 * Added configuration support for Linux/IA64
15549 *Rolf Haberrecker <rolf@suse.de>*
15551 * Assembler module support for Mingw32.
15555 * Shared library support for HPUX (in shlib/).
15557 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous*
15559 * Shared library support for Solaris gcc.
15561 *Lutz Behnke <behnke@trustcenter.de>*
15563 ### Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
15565 * PKCS7_encrypt() was adding text MIME headers twice because they
15566 were added manually and by SMIME_crlf_copy().
15570 * In bntest.c don't call BN_rand with zero bits argument.
15572 *Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>*
15574 * BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
15575 case was implemented. This caused BN_div_recp() to fail occasionally.
15579 * Add an optional second argument to the set_label() in the perl
15580 assembly language builder. If this argument exists and is set
15581 to 1 it signals that the assembler should use a symbol whose
15582 scope is the entire file, not just the current function. This
15583 is needed with MASM which uses the format label:: for this scope.
15585 *Steve Henson, pointed out by Peter Runestig <peter@runestig.com>*
15587 * Change the ASN1 types so they are typedefs by default. Before
15588 almost all types were #define'd to ASN1_STRING which was causing
15589 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
15594 * Change names of new functions to the new get1/get0 naming
15595 convention: After 'get1', the caller owns a reference count
15596 and has to call `..._free`; 'get0' returns a pointer to some
15597 data structure without incrementing reference counters.
15598 (Some of the existing 'get' functions increment a reference
15599 counter, some don't.)
15600 Similarly, 'set1' and 'add1' functions increase reference
15601 counters or duplicate objects.
15605 * Allow for the possibility of temp RSA key generation failure:
15606 the code used to assume it always worked and crashed on failure.
15610 * Fix potential buffer overrun problem in BIO_printf().
15611 *Ulf Möller, using public domain code by Patrick Powell; problem
15612 pointed out by David Sacerdote <das33@cornell.edu>*
15614 * Support EGD <http://www.lothar.com/tech/crypto/>. New functions
15615 RAND_egd() and RAND_status(). In the command line application,
15616 the EGD socket can be specified like a seed file using RANDFILE
15621 * Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
15622 Some CAs (e.g. Verisign) distribute certificates in this form.
15626 * Remove the SSL_ALLOW_ADH compile option and set the default cipher
15627 list to exclude them. This means that no special compilation option
15628 is needed to use anonymous DH: it just needs to be included in the
15633 * Change the EVP_MD_CTX_type macro so its meaning consistent with
15634 EVP_MD_type. The old functionality is available in a new macro called
15635 EVP_MD_md(). Change code that uses it and update docs.
15639 * `..._ctrl` functions now have corresponding `..._callback_ctrl` functions
15640 where the `void *` argument is replaced by a function pointer argument.
15641 Previously `void *` was abused to point to functions, which works on
15642 many platforms, but is not correct. As these functions are usually
15643 called by macros defined in OpenSSL header files, most source code
15644 should work without changes.
15648 * `<openssl/opensslconf.h>` (which is created by Configure) now contains
15649 sections with information on -D... compiler switches used for
15650 compiling the library so that applications can see them. To enable
15651 one of these sections, a pre-processor symbol `OPENSSL_..._DEFINES`
15652 must be defined. E.g.,
15653 #define OPENSSL_ALGORITHM_DEFINES
15654 #include <openssl/opensslconf.h>
15655 defines all pertinent `NO_<algo>` symbols, such as NO_IDEA, NO_RSA, etc.
15657 *Richard Levitte, Ulf and Bodo Möller*
15659 * Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
15664 * Change the 'other' type in certificate aux info to a STACK_OF
15665 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
15666 the required ASN1 format: arbitrary types determined by an OID.
15670 * Add some PEM_write_X509_REQ_NEW() functions and a command line
15671 argument to 'req'. This is not because the function is newer or
15672 better than others it just uses the work 'NEW' in the certificate
15673 request header lines. Some software needs this.
15677 * Reorganise password command line arguments: now passwords can be
15678 obtained from various sources. Delete the PEM_cb function and make
15679 it the default behaviour: i.e. if the callback is NULL and the
15680 usrdata argument is not NULL interpret it as a null terminated pass
15681 phrase. If usrdata and the callback are NULL then the pass phrase
15682 is prompted for as usual.
15686 * Add support for the Compaq Atalla crypto accelerator. If it is installed,
15687 the support is automatically enabled. The resulting binaries will
15688 autodetect the card and use it if present.
15690 *Ben Laurie and Compaq Inc.*
15692 * Work around for Netscape hang bug. This sends certificate request
15693 and server done in one record. Since this is perfectly legal in the
15694 SSL/TLS protocol it isn't a "bug" option and is on by default. See
15695 the bugs/SSLv3 entry for more info.
15699 * HP-UX tune-up: new unified configs, HP C compiler bug workaround.
15703 * Add -rand argument to smime and pkcs12 applications and read/write
15708 * New 'passwd' tool for crypt(3) and apr1 password hashes.
15712 * Add command line password options to the remaining applications.
15716 * Bug fix for BN_div_recp() for numerators with an even number of
15721 * More tests in bntest.c, and changed test_bn output.
15725 * ./config recognizes MacOS X now.
15729 * Bug fix for BN_div() when the first words of num and divisor are
15730 equal (it gave wrong results if `(rem=(n1-q*d0)&BN_MASK2) < d0)`.
15734 * Add support for various broken PKCS#8 formats, and command line
15735 options to produce them.
15739 * New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
15740 get temporary BIGNUMs from a BN_CTX.
15744 * Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
15749 * Change the `SSLeay_add_all_*()` functions to `OpenSSL_add_all_*()` and
15750 include a #define from the old name to the new. The original intent
15751 was that statically linked binaries could for example just call
15752 SSLeay_add_all_ciphers() to just add ciphers to the table and not
15753 link with digests. This never worked because SSLeay_add_all_digests()
15754 and SSLeay_add_all_ciphers() were in the same source file so calling
15755 one would link with the other. They are now in separate source files.
15759 * Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
15763 * Use a less unusual form of the Miller-Rabin primality test (it used
15764 a binary algorithm for exponentiation integrated into the Miller-Rabin
15765 loop, our standard modexp algorithms are faster).
15769 * Support for the EBCDIC character set completed.
15771 *Martin Kraemer <Martin.Kraemer@Mch.SNI.De>*
15773 * Source code cleanups: use const where appropriate, eliminate casts,
15774 use `void *` instead of `char *` in lhash.
15778 * Bugfix: ssl3_send_server_key_exchange was not restartable
15779 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
15780 this the server could overwrite ephemeral keys that the client
15785 * Turn DSA_is_prime into a macro that calls BN_is_prime,
15786 using 50 iterations of the Rabin-Miller test.
15788 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
15789 iterations of the Rabin-Miller test as required by the appendix
15790 to FIPS PUB 186[-1]) instead of DSA_is_prime.
15791 As BN_is_prime_fasttest includes trial division, DSA parameter
15792 generation becomes much faster.
15794 This implies a change for the callback functions in DSA_is_prime
15795 and DSA_generate_parameters: The callback function is called once
15796 for each positive witness in the Rabin-Miller test, not just
15797 occasionally in the inner loop; and the parameters to the
15798 callback function now provide an iteration count for the outer
15799 loop rather than for the current invocation of the inner loop.
15800 DSA_generate_parameters additionally can call the callback
15801 function with an 'iteration count' of -1, meaning that a
15802 candidate has passed the trial division test (when q is generated
15803 from an application-provided seed, trial division is skipped).
15807 * New function BN_is_prime_fasttest that optionally does trial
15808 division before starting the Rabin-Miller test and has
15809 an additional BN_CTX * argument (whereas BN_is_prime always
15810 has to allocate at least one BN_CTX).
15811 'callback(1, -1, cb_arg)' is called when a number has passed the
15812 trial division stage.
15816 * Fix for bug in CRL encoding. The validity dates weren't being handled
15821 * New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
15825 * New function BN_pseudo_rand().
15829 * Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
15830 bignum version of BN_from_montgomery() with the working code from
15831 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
15836 * Avoid a race condition in s2_clnt.c (function get_server_hello) that
15837 made it impossible to use the same SSL_SESSION data structure in
15838 SSL2 clients in multiple threads.
15842 * The return value of RAND_load_file() no longer counts bytes obtained
15843 by stat(). RAND_load_file(..., -1) is new and uses the complete file
15844 to seed the PRNG (previously an explicit byte count was required).
15846 *Ulf Möller, Bodo Möller*
15848 * Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
15849 used `char *` instead of `void *` and had casts all over the place.
15853 * Make BN_generate_prime() return NULL on error if ret!=NULL.
15857 * Retain source code compatibility for BN_prime_checks macro:
15858 BN_is_prime(..., BN_prime_checks, ...) now uses
15859 BN_prime_checks_for_size to determine the appropriate number of
15860 Rabin-Miller iterations.
15864 * Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
15865 DH_CHECK_P_NOT_SAFE_PRIME.
15866 (Check if this is true? OpenPGP calls them "strong".)
15870 * Merge the functionality of "dh" and "gendh" programs into a new program
15871 "dhparam". The old programs are retained for now but will handle DH keys
15872 (instead of parameters) in future.
15876 * Make the ciphers, s_server and s_client programs check the return values
15877 when a new cipher list is set.
15881 * Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
15882 ciphers. Before when the 56bit ciphers were enabled the sorting was
15885 The syntax for the cipher sorting has been extended to support sorting by
15886 cipher-strength (using the strength_bits hard coded in the tables).
15887 The new command is `@STRENGTH` (see also `doc/apps/ciphers.pod`).
15889 Fix a bug in the cipher-command parser: when supplying a cipher command
15890 string with an "undefined" symbol (neither command nor alphanumeric
15891 *A-Za-z0-9*, ssl_set_cipher_list used to hang in an endless loop. Now
15892 an error is flagged.
15894 Due to the strength-sorting extension, the code of the
15895 ssl_create_cipher_list() function was completely rearranged. I hope that
15896 the readability was also increased :-)
15898 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>*
15900 * Minor change to 'x509' utility. The -CAcreateserial option now uses 1
15901 for the first serial number and places 2 in the serial number file. This
15902 avoids problems when the root CA is created with serial number zero and
15903 the first user certificate has the same issuer name and serial number
15908 * Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
15909 the new code. Add documentation for this stuff.
15913 * Changes to X509_ATTRIBUTE utilities. These have been renamed from
15914 `X509_*()` to `X509at_*()` on the grounds that they don't handle X509
15915 structures and behave in an analogous way to the X509v3 functions:
15916 they shouldn't be called directly but wrapper functions should be used
15919 So we also now have some wrapper functions that call the X509at functions
15920 when passed certificate requests. (TO DO: similar things can be done with
15921 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
15922 things. Some of these need some d2i or i2d and print functionality
15923 because they handle more complex structures.)
15927 * Add missing #ifndefs that caused missing symbols when building libssl
15928 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
15929 NO_RSA in `ssl/s2*.c`.
15931 *Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller*
15933 * Precautions against using the PRNG uninitialized: RAND_bytes() now
15934 has a return value which indicates the quality of the random data
15935 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
15936 error queue. New function RAND_pseudo_bytes() generates output that is
15937 guaranteed to be unique but not unpredictable. RAND_add is like
15938 RAND_seed, but takes an extra argument for an entropy estimate
15939 (RAND_seed always assumes full entropy).
15943 * Do more iterations of Rabin-Miller probable prime test (specifically,
15944 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
15945 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
15946 in crypto/bn/bn_prime.c for the complete table). This guarantees a
15947 false-positive rate of at most 2^-80 for random input.
15951 * Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
15955 * New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
15956 in the 0.9.5 release), this returns the chain
15957 from an X509_CTX structure with a dup of the stack and all
15958 the X509 reference counts upped: so the stack will exist
15959 after X509_CTX_cleanup() has been called. Modify pkcs12.c
15962 Also make SSL_SESSION_print() print out the verify return
15967 * Add manpage for the pkcs12 command. Also change the default
15968 behaviour so MAC iteration counts are used unless the new
15969 -nomaciter option is used. This improves file security and
15970 only older versions of MSIE (4.0 for example) need it.
15974 * Honor the no-xxx Configure options when creating .DEF files.
15978 * Add PKCS#10 attributes to field table: challengePassword,
15979 unstructuredName and unstructuredAddress. These are taken from
15980 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
15981 international characters are used.
15983 More changes to X509_ATTRIBUTE code: allow the setting of types
15984 based on strings. Remove the 'loc' parameter when adding
15985 attributes because these will be a SET OF encoding which is sorted
15990 * Initial changes to the 'req' utility to allow request generation
15991 automation. This will allow an application to just generate a template
15992 file containing all the field values and have req construct the
15995 Initial support for X509_ATTRIBUTE handling. Stacks of these are
15996 used all over the place including certificate requests and PKCS#7
15997 structures. They are currently handled manually where necessary with
15998 some primitive wrappers for PKCS#7. The new functions behave in a
15999 manner analogous to the X509 extension functions: they allow
16000 attributes to be looked up by NID and added.
16002 Later something similar to the X509V3 code would be desirable to
16003 automatically handle the encoding, decoding and printing of the
16004 more complex types. The string types like challengePassword can
16005 be handled by the string table functions.
16007 Also modified the multi byte string table handling. Now there is
16008 a 'global mask' which masks out certain types. The table itself
16009 can use the flag STABLE_NO_MASK to ignore the mask setting: this
16010 is useful when for example there is only one permissible type
16011 (as in countryName) and using the mask might result in no valid
16016 * Clean up 'Finished' handling, and add functions SSL_get_finished and
16017 SSL_get_peer_finished to allow applications to obtain the latest
16018 Finished messages sent to the peer or expected from the peer,
16019 respectively. (SSL_get_peer_finished is usually the Finished message
16020 actually received from the peer, otherwise the protocol will be aborted.)
16022 As the Finished message are message digests of the complete handshake
16023 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
16024 be used for external authentication procedures when the authentication
16025 provided by SSL/TLS is not desired or is not enough.
16029 * Enhanced support for Alpha Linux is added. Now ./config checks if
16030 the host supports BWX extension and if Compaq C is present on the
16031 $PATH. Just exploiting of the BWX extension results in 20-30%
16032 performance kick for some algorithms, e.g. DES and RC4 to mention
16033 a couple. Compaq C in turn generates ~20% faster code for MD5 and
16038 * Add support for MS "fast SGC". This is arguably a violation of the
16039 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
16040 weak crypto and after checking the certificate is SGC a second one
16041 with strong crypto. MS SGC stops the first handshake after receiving
16042 the server certificate message and sends a second client hello. Since
16043 a server will typically do all the time consuming operations before
16044 expecting any further messages from the client (server key exchange
16045 is the most expensive) there is little difference between the two.
16047 To get OpenSSL to support MS SGC we have to permit a second client
16048 hello message after we have sent server done. In addition we have to
16049 reset the MAC if we do get this second client hello.
16053 * Add a function 'd2i_AutoPrivateKey()' this will automatically decide
16054 if a DER encoded private key is RSA or DSA traditional format. Changed
16055 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
16056 format DER encoded private key. Newer code should use PKCS#8 format which
16057 has the key type encoded in the ASN1 structure. Added DER private key
16058 support to pkcs8 application.
16062 * SSL 3/TLS 1 servers now don't request certificates when an anonymous
16063 ciphersuites has been selected (as required by the SSL 3/TLS 1
16064 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
16065 is set, we interpret this as a request to violate the specification
16066 (the worst that can happen is a handshake failure, and 'correct'
16067 behaviour would result in a handshake failure anyway).
16071 * In SSL_CTX_add_session, take into account that there might be multiple
16072 SSL_SESSION structures with the same session ID (e.g. when two threads
16073 concurrently obtain them from an external cache).
16074 The internal cache can handle only one SSL_SESSION with a given ID,
16075 so if there's a conflict, we now throw out the old one to achieve
16080 * Add OIDs for idea and blowfish in CBC mode. This will allow both
16081 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
16082 some routines that use cipher OIDs: some ciphers do not have OIDs
16083 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
16088 * Simplify the trust setting structure and code. Now we just have
16089 two sequences of OIDs for trusted and rejected settings. These will
16090 typically have values the same as the extended key usage extension
16091 and any application specific purposes.
16093 The trust checking code now has a default behaviour: it will just
16094 check for an object with the same NID as the passed id. Functions can
16095 be provided to override either the default behaviour or the behaviour
16096 for a given id. SSL client, server and email already have functions
16097 in place for compatibility: they check the NID and also return "trusted"
16098 if the certificate is self signed.
16102 * Add d2i,i2d bio/fp functions for PrivateKey: these convert the
16103 traditional format into an EVP_PKEY structure.
16107 * Add a password callback function PEM_cb() which either prompts for
16108 a password if usr_data is NULL or otherwise assumes it is a null
16109 terminated password. Allow passwords to be passed on command line
16110 environment or config files in a few more utilities.
16114 * Add a bunch of DER and PEM functions to handle PKCS#8 format private
16115 keys. Add some short names for PKCS#8 PBE algorithms and allow them
16116 to be specified on the command line for the pkcs8 and pkcs12 utilities.
16117 Update documentation.
16121 * Support for ASN1 "NULL" type. This could be handled before by using
16122 ASN1_TYPE but there wasn't any function that would try to read a NULL
16123 and produce an error if it couldn't. For compatibility we also have
16124 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
16125 don't allocate anything because they don't need to.
16129 * Initial support for MacOS is now provided. Examine INSTALL.MacOS
16132 *Andy Polyakov, Roy Woods <roy@centicsystems.ca>*
16134 * Rebuild of the memory allocation routines used by OpenSSL code and
16135 possibly others as well. The purpose is to make an interface that
16136 provide hooks so anyone can build a separate set of allocation and
16137 deallocation routines to be used by OpenSSL, for example memory
16138 pool implementations, or something else, which was previously hard
16139 since Malloc(), Realloc() and Free() were defined as macros having
16140 the values malloc, realloc and free, respectively (except for Win32
16141 compilations). The same is provided for memory debugging code.
16142 OpenSSL already comes with functionality to find memory leaks, but
16143 this gives people a chance to debug other memory problems.
16145 With these changes, a new set of functions and macros have appeared:
16147 CRYPTO_set_mem_debug_functions() [F]
16148 CRYPTO_get_mem_debug_functions() [F]
16149 CRYPTO_dbg_set_options() [F]
16150 CRYPTO_dbg_get_options() [F]
16151 CRYPTO_malloc_debug_init() [M]
16153 The memory debug functions are NULL by default, unless the library
16154 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
16155 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
16156 gives the standard debugging functions that come with OpenSSL) or
16157 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
16158 provided by the library user) must be used. When the standard
16159 debugging functions are used, CRYPTO_dbg_set_options can be used to
16160 request additional information:
16161 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
16162 the CRYPTO_MDEBUG_xxx macro when compiling the library.
16164 Also, things like CRYPTO_set_mem_functions will always give the
16165 expected result (the new set of functions is used for allocation
16166 and deallocation) at all times, regardless of platform and compiler
16169 To finish it up, some functions that were never use in any other
16170 way than through macros have a new API and new semantic:
16172 CRYPTO_dbg_malloc()
16173 CRYPTO_dbg_realloc()
16176 All macros of value have retained their old syntax.
16178 *Richard Levitte and Bodo Moeller*
16180 * Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
16181 ordering of SMIMECapabilities wasn't in "strength order" and there
16182 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
16187 * Some ASN1 types with illegal zero length encoding (INTEGER,
16188 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
16190 *Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson*
16192 * Merge in my S/MIME library for OpenSSL. This provides a simple
16193 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
16194 functionality to handle multipart/signed properly) and a utility
16195 called 'smime' to call all this stuff. This is based on code I
16196 originally wrote for Celo who have kindly allowed it to be
16197 included in OpenSSL.
16201 * Add variants des_set_key_checked and des_set_key_unchecked of
16202 des_set_key (aka des_key_sched). Global variable des_check_key
16203 decides which of these is called by des_set_key; this way
16204 des_check_key behaves as it always did, but applications and
16205 the library itself, which was buggy for des_check_key == 1,
16206 have a cleaner way to pick the version they need.
16210 * New function PKCS12_newpass() which changes the password of a
16215 * Modify X509_TRUST and X509_PURPOSE so it also uses a static and
16216 dynamic mix. In both cases the ids can be used as an index into the
16217 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
16218 functions so they accept a list of the field values and the
16219 application doesn't need to directly manipulate the X509_TRUST
16224 * Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
16229 * Modify the way the V3 extension code looks up extensions. This now
16230 works in a similar way to the object code: we have some "standard"
16231 extensions in a static table which is searched with OBJ_bsearch()
16232 and the application can add dynamic ones if needed. The file
16233 crypto/x509v3/ext_dat.h now has the info: this file needs to be
16234 updated whenever a new extension is added to the core code and kept
16235 in ext_nid order. There is a simple program 'tabtest.c' which checks
16236 this. New extensions are not added too often so this file can readily
16237 be maintained manually.
16239 There are two big advantages in doing things this way. The extensions
16240 can be looked up immediately and no longer need to be "added" using
16241 X509V3_add_standard_extensions(): this function now does nothing.
16242 Side note: I get *lots* of email saying the extension code doesn't
16243 work because people forget to call this function.
16244 Also no dynamic allocation is done unless new extensions are added:
16245 so if we don't add custom extensions there is no need to call
16246 X509V3_EXT_cleanup().
16250 * Modify enc utility's salting as follows: make salting the default. Add a
16251 magic header, so unsalted files fail gracefully instead of just decrypting
16252 to garbage. This is because not salting is a big security hole, so people
16253 should be discouraged from doing it.
16257 * Fixes and enhancements to the 'x509' utility. It allowed a message
16258 digest to be passed on the command line but it only used this
16259 parameter when signing a certificate. Modified so all relevant
16260 operations are affected by the digest parameter including the
16261 -fingerprint and -x509toreq options. Also -x509toreq choked if a
16262 DSA key was used because it didn't fix the digest.
16266 * Initial certificate chain verify code. Currently tests the untrusted
16267 certificates for consistency with the verify purpose (which is set
16268 when the X509_STORE_CTX structure is set up) and checks the pathlength.
16270 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
16271 this is because it will reject chains with invalid extensions whereas
16272 every previous version of OpenSSL and SSLeay made no checks at all.
16274 Trust code: checks the root CA for the relevant trust settings. Trust
16275 settings have an initial value consistent with the verify purpose: e.g.
16276 if the verify purpose is for SSL client use it expects the CA to be
16277 trusted for SSL client use. However the default value can be changed to
16278 permit custom trust settings: one example of this would be to only trust
16279 certificates from a specific "secure" set of CAs.
16281 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
16282 which should be used for version portability: especially since the
16283 verify structure is likely to change more often now.
16285 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
16286 to set them. If not set then assume SSL clients will verify SSL servers
16289 Two new options to the verify program: -untrusted allows a set of
16290 untrusted certificates to be passed in and -purpose which sets the
16291 intended purpose of the certificate. If a purpose is set then the
16292 new chain verify code is used to check extension consistency.
16296 * Support for the authority information access extension.
16300 * Modify RSA and DSA PEM read routines to transparently handle
16301 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
16302 public keys in a format compatible with certificate
16303 SubjectPublicKeyInfo structures. Unfortunately there were already
16304 functions called *_PublicKey_* which used various odd formats so
16305 these are retained for compatibility: however the DSA variants were
16306 never in a public release so they have been deleted. Changed dsa/rsa
16307 utilities to handle the new format: note no releases ever handled public
16308 keys so we should be OK.
16310 The primary motivation for this change is to avoid the same fiasco
16311 that dogs private keys: there are several incompatible private key
16312 formats some of which are standard and some OpenSSL specific and
16313 require various evil hacks to allow partial transparent handling and
16314 even then it doesn't work with DER formats. Given the option anything
16315 other than PKCS#8 should be dumped: but the other formats have to
16316 stay in the name of compatibility.
16318 With public keys and the benefit of hindsight one standard format
16319 is used which works with EVP_PKEY, RSA or DSA structures: though
16320 it clearly returns an error if you try to read the wrong kind of key.
16322 Added a -pubkey option to the 'x509' utility to output the public key.
16323 Also rename the `EVP_PKEY_get_*()` to `EVP_PKEY_rget_*()`
16324 (renamed to `EVP_PKEY_get1_*()` in the OpenSSL 0.9.5 release) and add
16325 `EVP_PKEY_rset_*()` functions (renamed to `EVP_PKEY_set1_*()`)
16326 that do the same as the `EVP_PKEY_assign_*()` except they up the
16327 reference count of the added key (they don't "swallow" the
16332 * Fixes to crypto/x509/by_file.c the code to read in certificates and
16333 CRLs would fail if the file contained no certificates or no CRLs:
16334 added a new function to read in both types and return the number
16335 read: this means that if none are read it will be an error. The
16336 DER versions of the certificate and CRL reader would always fail
16337 because it isn't possible to mix certificates and CRLs in DER format
16338 without choking one or the other routine. Changed this to just read
16339 a certificate: this is the best we can do. Also modified the code
16340 in `apps/verify.c` to take notice of return codes: it was previously
16341 attempting to read in certificates from NULL pointers and ignoring
16342 any errors: this is one reason why the cert and CRL reader seemed
16343 to work. It doesn't check return codes from the default certificate
16344 routines: these may well fail if the certificates aren't installed.
16348 * Code to support otherName option in GeneralName.
16352 * First update to verify code. Change the verify utility
16353 so it warns if it is passed a self signed certificate:
16354 for consistency with the normal behaviour. X509_verify
16355 has been modified to it will now verify a self signed
16356 certificate if *exactly* the same certificate appears
16357 in the store: it was previously impossible to trust a
16358 single self signed certificate. This means that:
16359 openssl verify ss.pem
16360 now gives a warning about a self signed certificate but
16361 openssl verify -CAfile ss.pem ss.pem
16366 * For servers, store verify_result in SSL_SESSION data structure
16367 (and add it to external session representation).
16368 This is needed when client certificate verifications fails,
16369 but an application-provided verification callback (set by
16370 SSL_CTX_set_cert_verify_callback) allows accepting the session
16371 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
16372 but returns 1): When the session is reused, we have to set
16373 ssl->verify_result to the appropriate error code to avoid
16376 *Bodo Moeller, problem pointed out by Lutz Jaenicke*
16378 * Fix a bug in the new PKCS#7 code: it didn't consider the
16379 case in PKCS7_dataInit() where the signed PKCS7 structure
16380 didn't contain any existing data because it was being created.
16382 *Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson*
16384 * Add a salt to the key derivation routines in enc.c. This
16385 forms the first 8 bytes of the encrypted file. Also add a
16386 -S option to allow a salt to be input on the command line.
16390 * New function X509_cmp(). Oddly enough there wasn't a function
16391 to compare two certificates. We do this by working out the SHA1
16392 hash and comparing that. X509_cmp() will be needed by the trust
16397 * SSL_get1_session() is like SSL_get_session(), but increments
16398 the reference count in the SSL_SESSION returned.
16400 *Geoff Thorpe <geoff@eu.c2.net>*
16402 * Fix for 'req': it was adding a null to request attributes.
16403 Also change the X509_LOOKUP and X509_INFO code to handle
16404 certificate auxiliary information.
16408 * Add support for 40 and 64 bit RC2 and RC4 algorithms: document
16413 * Add the possibility to add extra information to the memory leak
16414 detecting output, to form tracebacks, showing from where each
16415 allocation was originated: CRYPTO_push_info("constant string") adds
16416 the string plus current file name and line number to a per-thread
16417 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
16418 is like calling CYRPTO_pop_info() until the stack is empty.
16419 Also updated memory leak detection code to be multi-thread-safe.
16423 * Add options -text and -noout to pkcs7 utility and delete the
16424 encryption options which never did anything. Update docs.
16428 * Add options to some of the utilities to allow the pass phrase
16429 to be included on either the command line (not recommended on
16430 OSes like Unix) or read from the environment. Update the
16431 manpages and fix a few bugs.
16435 * Add a few manpages for some of the openssl commands.
16439 * Fix the -revoke option in ca. It was freeing up memory twice,
16440 leaking and not finding already revoked certificates.
16444 * Extensive changes to support certificate auxiliary information.
16445 This involves the use of X509_CERT_AUX structure and X509_AUX
16446 functions. An X509_AUX function such as PEM_read_X509_AUX()
16447 can still read in a certificate file in the usual way but it
16448 will also read in any additional "auxiliary information". By
16449 doing things this way a fair degree of compatibility can be
16450 retained: existing certificates can have this information added
16451 using the new 'x509' options.
16453 Current auxiliary information includes an "alias" and some trust
16454 settings. The trust settings will ultimately be used in enhanced
16455 certificate chain verification routines: currently a certificate
16456 can only be trusted if it is self signed and then it is trusted
16461 * Fix assembler for Alpha (tested only on DEC OSF not Linux or `*BSD`).
16462 The problem was that one of the replacement routines had not been working
16463 since SSLeay releases. For now the offending routine has been replaced
16464 with non-optimised assembler. Even so, this now gives around 95%
16465 performance improvement for 1024 bit RSA signs.
16469 * Hack to fix PKCS#7 decryption when used with some unorthodox RC2
16470 handling. Most clients have the effective key size in bits equal to
16471 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
16472 A few however don't do this and instead use the size of the decrypted key
16473 to determine the RC2 key length and the AlgorithmIdentifier to determine
16474 the effective key length. In this case the effective key length can still
16475 be 40 bits but the key length can be 168 bits for example. This is fixed
16476 by manually forcing an RC2 key into the EVP_PKEY structure because the
16477 EVP code can't currently handle unusual RC2 key sizes: it always assumes
16478 the key length and effective key length are equal.
16482 * Add a bunch of functions that should simplify the creation of
16483 X509_NAME structures. Now you should be able to do:
16484 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
16485 and have it automatically work out the correct field type and fill in
16486 the structures. The more adventurous can try:
16487 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
16488 and it will (hopefully) work out the correct multibyte encoding.
16492 * Change the 'req' utility to use the new field handling and multibyte
16493 copy routines. Before the DN field creation was handled in an ad hoc
16494 way in req, ca, and x509 which was rather broken and didn't support
16495 BMPStrings or UTF8Strings. Since some software doesn't implement
16496 BMPStrings or UTF8Strings yet, they can be enabled using the config file
16497 using the dirstring_type option. See the new comment in the default
16498 openssl.cnf for more info.
16502 * Make crypto/rand/md_rand.c more robust:
16503 - Assure unique random numbers after fork().
16504 - Make sure that concurrent threads access the global counter and
16505 md serializably so that we never lose entropy in them
16506 or use exactly the same state in multiple threads.
16507 Access to the large state is not always serializable because
16508 the additional locking could be a performance killer, and
16509 md should be large enough anyway.
16513 * New file `apps/app_rand.c` with commonly needed functionality
16514 for handling the random seed file.
16516 Use the random seed file in some applications that previously did not:
16518 dsaparam -genkey (which also ignored its '-rand' option),
16521 x509 (when signing).
16522 Except on systems with /dev/urandom, it is crucial to have a random
16523 seed file at least for key creation, DSA signing, and for DH exchanges;
16524 for RSA signatures we could do without one.
16526 gendh and gendsa (unlike genrsa) used to read only the first byte
16527 of each file listed in the '-rand' option. The function as previously
16528 found in genrsa is now in app_rand.c and is used by all programs
16529 that support '-rand'.
16533 * In RAND_write_file, use mode 0600 for creating files;
16534 don't just chmod when it may be too late.
16538 * Report an error from X509_STORE_load_locations
16539 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
16543 * New function ASN1_mbstring_copy() this copies a string in either
16544 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
16545 into an ASN1_STRING type. A mask of permissible types is passed
16546 and it chooses the "minimal" type to use or an error if not type
16551 * Add function equivalents to the various macros in asn1.h. The old
16552 macros are retained with an `M_` prefix. Code inside the library can
16553 use the `M_` macros. External code (including the openssl utility)
16554 should *NOT* in order to be "shared library friendly".
16558 * Add various functions that can check a certificate's extensions
16559 to see if it usable for various purposes such as SSL client,
16560 server or S/MIME and CAs of these types. This is currently
16561 VERY EXPERIMENTAL but will ultimately be used for certificate chain
16562 verification. Also added a -purpose flag to x509 utility to
16563 print out all the purposes.
16567 * Add a CRYPTO_EX_DATA to X509 certificate structure and associated
16572 * New `X509V3_{X509,CRL,REVOKED}_get_d2i()` functions. These will search
16573 for, obtain and decode and extension and obtain its critical flag.
16574 This allows all the necessary extension code to be handled in a
16575 single function call.
16579 * RC4 tune-up featuring 30-40% performance improvement on most RISC
16580 platforms. See crypto/rc4/rc4_enc.c for further details.
16584 * New -noout option to asn1parse. This causes no output to be produced
16585 its main use is when combined with -strparse and -out to extract data
16586 from a file (which may not be in ASN.1 format).
16590 * Fix for pkcs12 program. It was hashing an invalid certificate pointer
16591 when producing the local key id.
16593 *Richard Levitte <levitte@stacken.kth.se>*
16595 * New option -dhparam in s_server. This allows a DH parameter file to be
16596 stated explicitly. If it is not stated then it tries the first server
16597 certificate file. The previous behaviour hard coded the filename
16602 * Add -pubin and -pubout options to the rsa and dsa commands. These allow
16603 a public key to be input or output. For example:
16604 openssl rsa -in key.pem -pubout -out pubkey.pem
16605 Also added necessary DSA public key functions to handle this.
16609 * Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
16610 in the message. This was handled by allowing
16611 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
16613 *Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>*
16615 * Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
16616 to the end of the strings whereas this didn't. This would cause problems
16617 if strings read with d2i_ASN1_bytes() were later modified.
16619 *Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>*
16621 * Fix for base64 decode bug. When a base64 bio reads only one line of
16622 data and it contains EOF it will end up returning an error. This is
16623 caused by input 46 bytes long. The cause is due to the way base64
16624 BIOs find the start of base64 encoded data. They do this by trying a
16625 trial decode on each line until they find one that works. When they
16626 do a flag is set and it starts again knowing it can pass all the
16627 data directly through the decoder. Unfortunately it doesn't reset
16628 the context it uses. This means that if EOF is reached an attempt
16629 is made to pass two EOFs through the context and this causes the
16630 resulting error. This can also cause other problems as well. As is
16631 usual with these problems it takes *ages* to find and the fix is
16632 trivial: move one line.
16634 *Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer)*
16636 * Ugly workaround to get s_client and s_server working under Windows. The
16637 old code wouldn't work because it needed to select() on sockets and the
16638 tty (for keypresses and to see if data could be written). Win32 only
16639 supports select() on sockets so we select() with a 1s timeout on the
16640 sockets and then see if any characters are waiting to be read, if none
16641 are present then we retry, we also assume we can always write data to
16642 the tty. This isn't nice because the code then blocks until we've
16643 received a complete line of data and it is effectively polling the
16644 keyboard at 1s intervals: however it's quite a bit better than not
16645 working at all :-) A dedicated Windows application might handle this
16646 with an event loop for example.
16650 * Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
16651 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
16652 will be called when RSA_sign() and RSA_verify() are used. This is useful
16653 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
16654 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
16655 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
16656 This necessitated the support of an extra signature type NID_md5_sha1
16657 for SSL signatures and modifications to the SSL library to use it instead
16658 of calling RSA_public_decrypt() and RSA_private_encrypt().
16662 * Add new -verify -CAfile and -CApath options to the crl program, these
16663 will lookup a CRL issuers certificate and verify the signature in a
16664 similar way to the verify program. Tidy up the crl program so it
16665 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
16666 less strict. It will now permit CRL extensions even if it is not
16667 a V2 CRL: this will allow it to tolerate some broken CRLs.
16671 * Initialize all non-automatic variables each time one of the openssl
16672 sub-programs is started (this is necessary as they may be started
16673 multiple times from the "OpenSSL>" prompt).
16675 *Lennart Bang, Bodo Moeller*
16677 * Preliminary compilation option RSA_NULL which disables RSA crypto without
16678 removing all other RSA functionality (this is what NO_RSA does). This
16679 is so (for example) those in the US can disable those operations covered
16680 by the RSA patent while allowing storage and parsing of RSA keys and RSA
16685 * Non-copying interface to BIO pairs.
16686 (still largely untested)
16690 * New function ASN1_tag2str() to convert an ASN1 tag to a descriptive
16691 ASCII string. This was handled independently in various places before.
16695 * New functions UTF8_getc() and UTF8_putc() that parse and generate
16696 UTF8 strings a character at a time.
16700 * Use client_version from client hello to select the protocol
16701 (s23_srvr.c) and for RSA client key exchange verification
16702 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
16706 * Add various utility functions to handle SPKACs, these were previously
16707 handled by poking round in the structure internals. Added new function
16708 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
16709 print, verify and generate SPKACs. Based on an original idea from
16710 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
16714 * RIPEMD160 is operational on all platforms and is back in 'make test'.
16718 * Allow the config file extension section to be overwritten on the
16719 command line. Based on an original idea from Massimiliano Pala
16720 <madwolf@comune.modena.it>. The new option is called -extensions
16721 and can be applied to ca, req and x509. Also -reqexts to override
16722 the request extensions in req and -crlexts to override the crl extensions
16727 * Add new feature to the SPKAC handling in ca. Now you can include
16728 the same field multiple times by preceding it by "XXXX." for example:
16731 this is the same syntax as used in the req config file.
16735 * Allow certificate extensions to be added to certificate requests. These
16736 are specified in a 'req_extensions' option of the req section of the
16737 config file. They can be printed out with the -text option to req but
16738 are otherwise ignored at present.
16742 * Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
16743 data read consists of only the final block it would not decrypted because
16744 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
16745 A misplaced 'break' also meant the decrypted final block might not be
16746 copied until the next read.
16750 * Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
16751 a few extra parameters to the DH structure: these will be useful if
16752 for example we want the value of 'q' or implement X9.42 DH.
16756 * Initial support for DSA_METHOD. This is based on the RSA_METHOD and
16757 provides hooks that allow the default DSA functions or functions on a
16758 "per key" basis to be replaced. This allows hardware acceleration and
16759 hardware key storage to be handled without major modification to the
16760 library. Also added low-level modexp hooks and CRYPTO_EX structure and
16761 associated functions.
16765 * Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
16766 as "read only": it can't be written to and the buffer it points to will
16767 not be freed. Reading from a read only BIO is much more efficient than
16768 a normal memory BIO. This was added because there are several times when
16769 an area of memory needs to be read from a BIO. The previous method was
16770 to create a memory BIO and write the data to it, this results in two
16771 copies of the data and an O(n^2) reading algorithm. There is a new
16772 function BIO_new_mem_buf() which creates a read only memory BIO from
16773 an area of memory. Also modified the PKCS#7 routines to use read only
16778 * Bugfix: ssl23_get_client_hello did not work properly when called in
16779 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
16780 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
16781 but a retry condition occurred while trying to read the rest.
16785 * The PKCS7_ENC_CONTENT_new() function was setting the content type as
16786 NID_pkcs7_encrypted by default: this was wrong since this should almost
16787 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
16788 the encrypted data type: this is a more sensible place to put it and it
16789 allows the PKCS#12 code to be tidied up that duplicated this
16794 * Changed obj_dat.pl script so it takes its input and output files on
16795 the command line. This should avoid shell escape redirection problems
16800 * Initial support for certificate extension requests, these are included
16801 in things like Xenroll certificate requests. Included functions to allow
16802 extensions to be obtained and added.
16806 * -crlf option to s_client and s_server for sending newlines as
16807 CRLF (as required by many protocols).
16811 ### Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
16813 * Install libRSAglue.a when OpenSSL is built with RSAref.
16815 *Ralf S. Engelschall*
16817 * A few more `#ifndef NO_FP_API / #endif` pairs for consistency.
16819 *Andrija Antonijevic <TheAntony2@bigfoot.com>*
16821 * Fix -startdate and -enddate (which was missing) arguments to 'ca'
16826 * New function DSA_dup_DH, which duplicates DSA parameters/keys as
16827 DH parameters/keys (q is lost during that conversion, but the resulting
16828 DH parameters contain its length).
16830 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
16831 much faster than DH_generate_parameters (which creates parameters
16832 where `p = 2*q + 1`), and also the smaller q makes DH computations
16833 much more efficient (160-bit exponentiation instead of 1024-bit
16834 exponentiation); so this provides a convenient way to support DHE
16835 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
16836 utter importance to use
16837 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
16839 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
16840 when such DH parameters are used, because otherwise small subgroup
16841 attacks may become possible!
16845 * Avoid memory leak in i2d_DHparams.
16849 * Allow the -k option to be used more than once in the enc program:
16850 this allows the same encrypted message to be read by multiple recipients.
16854 * New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
16855 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
16856 it will always use the numerical form of the OID, even if it has a short
16861 * Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
16862 method only got called if p,q,dmp1,dmq1,iqmp components were present,
16863 otherwise bn_mod_exp was called. In the case of hardware keys for example
16864 no private key components need be present and it might store extra data
16865 in the RSA structure, which cannot be accessed from bn_mod_exp.
16866 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
16867 private key operations.
16871 * Added support for SPARC Linux.
16875 * pem_password_cb function type incompatibly changed from
16876 typedef int pem_password_cb(char *buf, int size, int rwflag);
16878 ....(char *buf, int size, int rwflag, void *userdata);
16879 so that applications can pass data to their callbacks:
16880 The `PEM[_ASN1]_{read,write}...` functions and macros now take an
16881 additional void * argument, which is just handed through whenever
16882 the password callback is called.
16884 *Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller*
16886 New function SSL_CTX_set_default_passwd_cb_userdata.
16888 Compatibility note: As many C implementations push function arguments
16889 onto the stack in reverse order, the new library version is likely to
16890 interoperate with programs that have been compiled with the old
16891 pem_password_cb definition (PEM_whatever takes some data that
16892 happens to be on the stack as its last argument, and the callback
16893 just ignores this garbage); but there is no guarantee whatsoever that
16896 * The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
16897 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
16898 problems not only on Windows, but also on some Unix platforms.
16899 To avoid problematic command lines, these definitions are now in an
16900 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
16901 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
16905 * MIPS III/IV assembler module is reimplemented.
16909 * More DES library cleanups: remove references to srand/rand and
16910 delete an unused file.
16914 * Add support for the free Netwide assembler (NASM) under Win32,
16915 since not many people have MASM (ml) and it can be hard to obtain.
16916 This is currently experimental but it seems to work OK and pass all
16917 the tests. Check out INSTALL.W32 for info.
16921 * Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
16922 without temporary keys kept an extra copy of the server key,
16923 and connections with temporary keys did not free everything in case
16928 * New function RSA_check_key and new openssl rsa option -check
16929 for verifying the consistency of RSA keys.
16931 *Ulf Moeller, Bodo Moeller*
16933 * Various changes to make Win32 compile work:
16934 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
16935 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
16936 comparison" warnings.
16937 3. Add `sk_<TYPE>_sort` to DEF file generator and do make update.
16941 * Add a debugging option to PKCS#5 v2 key generation function: when
16942 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
16943 derived keys are printed to stderr.
16947 * Copy the flags in ASN1_STRING_dup().
16949 *Roman E. Pavlov <pre@mo.msk.ru>*
16951 * The x509 application mishandled signing requests containing DSA
16952 keys when the signing key was also DSA and the parameters didn't match.
16954 It was supposed to omit the parameters when they matched the signing key:
16955 the verifying software was then supposed to automatically use the CA's
16956 parameters if they were absent from the end user certificate.
16958 Omitting parameters is no longer recommended. The test was also
16959 the wrong way round! This was probably due to unusual behaviour in
16960 EVP_cmp_parameters() which returns 1 if the parameters match.
16961 This meant that parameters were omitted when they *didn't* match and
16962 the certificate was useless. Certificates signed with 'ca' didn't have
16965 *Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>*
16967 * Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
16968 The interface is as follows:
16969 Applications can use
16970 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
16971 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
16972 "off" is now the default.
16973 The library internally uses
16974 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
16975 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
16976 to disable memory-checking temporarily.
16978 Some inconsistent states that previously were possible (and were
16979 even the default) are now avoided.
16981 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
16982 with each memory chunk allocated; this is occasionally more helpful
16983 than just having a counter.
16985 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
16987 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
16992 * Introduce "mode" for SSL structures (with defaults in SSL_CTX),
16993 which largely parallels "options", but is for changing API behaviour,
16994 whereas "options" are about protocol behaviour.
16995 Initial "mode" flags are:
16997 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
16998 a single record has been written.
16999 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
17000 retries use the same buffer location.
17001 (But all of the contents must be
17006 * Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
17009 * Fix problems with no-hmac etc.
17011 *Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>*
17013 * New functions RSA_get_default_method(), RSA_set_method() and
17014 RSA_get_method(). These allows replacement of RSA_METHODs without having
17015 to mess around with the internals of an RSA structure.
17019 * Fix memory leaks in DSA_do_sign and DSA_is_prime.
17020 Also really enable memory leak checks in openssl.c and in some
17023 *Chad C. Mulligan, Bodo Moeller*
17025 * Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
17026 up the length of negative integers. This has now been simplified to just
17027 store the length when it is first determined and use it later, rather
17028 than trying to keep track of where data is copied and updating it to
17030 *Steve Henson, reported by Brien Wheeler <bwheeler@authentica-security.com>*
17032 * Add a new function PKCS7_signatureVerify. This allows the verification
17033 of a PKCS#7 signature but with the signing certificate passed to the
17034 function itself. This contrasts with PKCS7_dataVerify which assumes the
17035 certificate is present in the PKCS#7 structure. This isn't always the
17036 case: certificates can be omitted from a PKCS#7 structure and be
17037 distributed by "out of band" means (such as a certificate database).
17041 * Complete the `PEM_*` macros with DECLARE_PEM versions to replace the
17042 function prototypes in pem.h, also change util/mkdef.pl to add the
17043 necessary function names.
17047 * mk1mf.pl (used by Windows builds) did not properly read the
17048 options set by Configure in the top level Makefile, and Configure
17049 was not even able to write more than one option correctly.
17050 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
17054 * New functions CONF_load_bio() and CONF_load_fp() to allow a config
17055 file to be loaded from a BIO or FILE pointer. The BIO version will
17056 for example allow memory BIOs to contain config info.
17060 * New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
17061 Whoever hopes to achieve shared-library compatibility across versions
17062 must use this, not the compile-time macro.
17063 (Exercise 0.9.4: Which is the minimum library version required by
17065 Note: All this applies only to multi-threaded programs, others don't
17070 * Add missing case to s3_clnt.c state machine -- one of the new SSL tests
17071 through a BIO pair triggered the default case, i.e.
17072 SSLerr(...,SSL_R_UNKNOWN_STATE).
17076 * New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
17077 can use the SSL library even if none of the specific BIOs is
17082 * Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
17083 for the encoded length.
17085 *Jeon KyoungHo <khjeon@sds.samsung.co.kr>*
17087 * Add initial documentation of the X509V3 functions.
17091 * Add a new pair of functions PEM_write_PKCS8PrivateKey() and
17092 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
17093 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
17094 secure PKCS#8 private key format with a high iteration count.
17098 * Fix determination of Perl interpreter: A perl or perl5
17099 *directory* in $PATH was also accepted as the interpreter.
17101 *Ralf S. Engelschall*
17103 * Fix demos/sign/sign.c: well there wasn't anything strictly speaking
17104 wrong with it but it was very old and did things like calling
17105 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
17106 unusual formatting.
17110 * Fix demos/selfsign.c: it used obsolete and deleted functions, changed
17111 to use the new extension code.
17115 * Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
17116 with macros. This should make it easier to change their form, add extra
17117 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
17122 * Add to configuration table a new entry that can specify an alternative
17123 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
17124 according to Mark Crispin <MRC@Panda.COM>.
17128 * DES CBC did not update the IV. Weird.
17132 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
17133 Changing the behaviour of the former might break existing programs --
17134 where IV updating is needed, des_ncbc_encrypt can be used.
17137 * When bntest is run from "make test" it drives bc to check its
17138 calculations, as well as internally checking them. If an internal check
17139 fails, it needs to cause bc to give a non-zero result or make test carries
17140 on without noticing the failure. Fixed.
17144 * DES library cleanups.
17148 * Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
17149 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
17150 ciphers. NOTE: although the key derivation function has been verified
17151 against some published test vectors it has not been extensively tested
17152 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
17157 * Instead of "mkdir -p", which is not fully portable, use new
17158 Perl script "util/mkdir-p.pl".
17162 * Rewrite the way password based encryption (PBE) is handled. It used to
17163 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
17164 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
17165 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
17166 the 'parameter' field of the AlgorithmIdentifier is passed to the
17167 underlying key generation function so it must do its own ASN1 parsing.
17168 This has also changed the EVP_PBE_CipherInit() function which now has a
17169 'parameter' argument instead of literal salt and iteration count values
17170 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
17174 * Support for PKCS#5 v1.5 compatible password based encryption algorithms
17175 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
17176 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
17177 KEY" because this clashed with PKCS#8 unencrypted string. Since this
17178 value was just used as a "magic string" and not used directly its
17179 value doesn't matter.
17183 * Introduce some semblance of const correctness to BN. Shame C doesn't
17188 * "linux-sparc64" configuration (ultrapenguin).
17190 *Ray Miller <ray.miller@oucs.ox.ac.uk>*
17191 "linux-sparc" configuration.
17193 *Christian Forster <fo@hawo.stw.uni-erlangen.de>*
17195 * config now generates no-xxx options for missing ciphers.
17199 * Support the EBCDIC character set (work in progress).
17200 File ebcdic.c not yet included because it has a different license.
17202 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
17204 * Support BS2000/OSD-POSIX.
17206 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
17208 * Make callbacks for key generation use `void *` instead of `char *`.
17212 * Make S/MIME samples compile (not yet tested).
17216 * Additional typesafe stacks.
17220 * New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
17224 ### Changes between 0.9.3 and 0.9.3a [29 May 1999]
17226 * New configuration variant "sco5-gcc".
17228 * Updated some demos.
17230 *Sean O Riordain, Wade Scholine*
17232 * Add missing BIO_free at exit of pkcs12 application.
17236 * Fix memory leak in conf.c.
17240 * Updates for Win32 to assembler version of MD5.
17244 * Set #! path to perl in `apps/der_chop` to where we found it
17245 instead of using a fixed path.
17249 * SHA library changes for irix64-mips4-cc.
17253 * Improvements for VMS support.
17257 ### Changes between 0.9.2b and 0.9.3 [24 May 1999]
17259 * Bignum library bug fix. IRIX 6 passes "make test" now!
17260 This also avoids the problems with SC4.2 and unpatched SC5.
17262 *Andy Polyakov <appro@fy.chalmers.se>*
17264 * New functions sk_num, sk_value and sk_set to replace the previous macros.
17265 These are required because of the typesafe stack would otherwise break
17266 existing code. If old code used a structure member which used to be STACK
17267 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
17268 sk_num or sk_value it would produce an error because the num, data members
17269 are not present in STACK_OF. Now it just produces a warning. sk_set
17270 replaces the old method of assigning a value to sk_value
17271 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
17272 that does this will no longer work (and should use sk_set instead) but
17273 this could be regarded as a "questionable" behaviour anyway.
17277 * Fix most of the other PKCS#7 bugs. The "experimental" code can now
17278 correctly handle encrypted S/MIME data.
17282 * Change type of various DES function arguments from des_cblock
17283 (which means, in function argument declarations, pointer to char)
17284 to des_cblock * (meaning pointer to array with 8 char elements),
17285 which allows the compiler to do more typechecking; it was like
17286 that back in SSLeay, but with lots of ugly casts.
17288 Introduce new type const_des_cblock.
17292 * Reorganise the PKCS#7 library and get rid of some of the more obvious
17293 problems: find RecipientInfo structure that matches recipient certificate
17294 and initialise the ASN1 structures properly based on passed cipher.
17298 * Belatedly make the BN tests actually check the results.
17302 * Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
17303 to and from BNs: it was completely broken. New compilation option
17304 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
17305 key elements as negative integers.
17309 * Reorganize and speed up MD5.
17311 *Andy Polyakov <appro@fy.chalmers.se>*
17315 *Richard Levitte <richard@levitte.org>*
17317 * New option -out to asn1parse to allow the parsed structure to be
17318 output to a file. This is most useful when combined with the -strparse
17319 option to examine the output of things like OCTET STRINGS.
17323 * Make SSL library a little more fool-proof by not requiring any longer
17324 that `SSL_set_{accept,connect}_state` be called before
17325 `SSL_{accept,connect}` may be used (`SSL_set_..._state` is omitted
17326 in many applications because usually everything *appeared* to work as
17327 intended anyway -- now it really works as intended).
17331 * Move openssl.cnf out of lib/.
17335 * Fix various things to let OpenSSL even pass "egcc -pipe -O2 -Wall
17336 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
17337 -Wmissing-declarations -Wnested-externs -Winline" with EGCS 1.1.2+
17339 *Ralf S. Engelschall*
17341 * Various fixes to the EVP and PKCS#7 code. It may now be able to
17342 handle PKCS#7 enveloped data properly.
17344 *Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve*
17346 * Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
17347 copying pointers. The cert_st handling is changed by this in
17348 various ways (and thus what used to be known as ctx->default_cert
17349 is now called ctx->cert, since we don't resort to `s->ctx->[default_]cert`
17350 any longer when s->cert does not give us what we need).
17351 ssl_cert_instantiate becomes obsolete by this change.
17352 As soon as we've got the new code right (possibly it already is?),
17353 we have solved a couple of bugs of the earlier code where s->cert
17354 was used as if it could not have been shared with other SSL structures.
17356 Note that using the SSL API in certain dirty ways now will result
17357 in different behaviour than observed with earlier library versions:
17358 Changing settings for an `SSL_CTX *ctx` after having done s = SSL_new(ctx)
17359 does not influence s as it used to.
17361 In order to clean up things more thoroughly, inside SSL_SESSION
17362 we don't use CERT any longer, but a new structure SESS_CERT
17363 that holds per-session data (if available); currently, this is
17364 the peer's certificate chain and, for clients, the server's certificate
17365 and temporary key. CERT holds only those values that can have
17366 meaningful defaults in an SSL_CTX.
17370 * New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
17371 from the internal representation. Various PKCS#7 fixes: remove some
17372 evil casts and set the enc_dig_alg field properly based on the signing
17377 * Allow PKCS#12 password to be set from the command line or the
17378 environment. Let 'ca' get its config file name from the environment
17379 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
17384 * Allow certificate policies extension to use an IA5STRING for the
17385 organization field. This is contrary to the PKIX definition but
17386 VeriSign uses it and IE5 only recognises this form. Document 'x509'
17391 * Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
17392 without disallowing inline assembler and the like for non-pedantic builds.
17396 * Support Borland C++ builder.
17398 *Janez Jere <jj@void.si>, modified by Ulf Möller*
17404 * SHA-1 cleanups and performance enhancements.
17406 *Andy Polyakov <appro@fy.chalmers.se>*
17408 * Sparc v8plus assembler for the bignum library.
17410 *Andy Polyakov <appro@fy.chalmers.se>*
17412 * Accept any -xxx and +xxx compiler options in Configure.
17416 * Update HPUX configuration.
17420 * Add missing `sk_<type>_unshift()` function to safestack.h
17422 *Ralf S. Engelschall*
17424 * New function SSL_CTX_use_certificate_chain_file that sets the
17425 "extra_cert"s in addition to the certificate. (This makes sense
17426 only for "PEM" format files, as chains as a whole are not
17431 * Support verify_depth from the SSL API.
17432 x509_vfy.c had what can be considered an off-by-one-error:
17433 Its depth (which was not part of the external interface)
17434 was actually counting the number of certificates in a chain;
17435 now it really counts the depth.
17439 * Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
17440 instead of X509err, which often resulted in confusing error
17441 messages since the error codes are not globally unique
17442 (e.g. an alleged error in ssl3_accept when a certificate
17443 didn't match the private key).
17445 * New function SSL_CTX_set_session_id_context that allows to set a default
17446 value (so that you don't need SSL_set_session_id_context for each
17447 connection using the SSL_CTX).
17451 * OAEP decoding bug fix.
17455 * Support INSTALL_PREFIX for package builders, as proposed by
17460 * New Configure options "threads" and "no-threads". For systems
17461 where the proper compiler options are known (currently Solaris
17462 and Linux), "threads" is the default.
17466 * New script util/mklink.pl as a faster substitute for util/mklink.sh.
17470 * Install various scripts to $(OPENSSLDIR)/misc, not to
17471 $(INSTALLTOP)/bin -- they shouldn't clutter directories
17472 such as /usr/local/bin.
17476 * "make linux-shared" to build shared libraries.
17478 *Niels Poppe <niels@netbox.org>*
17480 * New Configure option `no-<cipher>` (rsa, idea, rc5, ...).
17484 * Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
17485 extension adding in x509 utility.
17489 * Remove NOPROTO sections and error code comments.
17493 * Partial rewrite of the DEF file generator to now parse the ANSI
17498 * New Configure options --prefix=DIR and --openssldir=DIR.
17502 * Complete rewrite of the error code script(s). It is all now handled
17503 by one script at the top level which handles error code gathering,
17504 header rewriting and C source file generation. It should be much better
17505 than the old method: it now uses a modified version of Ulf's parser to
17506 read the ANSI prototypes in all header files (thus the old K&R definitions
17507 aren't needed for error creation any more) and do a better job of
17508 translating function codes into names. The old 'ASN1 error code embedded
17509 in a comment' is no longer necessary and it doesn't use .err files which
17510 have now been deleted. Also the error code call doesn't have to appear all
17511 on one line (which resulted in some large lines...).
17515 * Change #include filenames from `<foo.h>` to `<openssl/foo.h>`.
17519 * Change behaviour of ssl2_read when facing length-0 packets: Don't return
17520 0 (which usually indicates a closed connection), but continue reading.
17524 * Fix some race conditions.
17528 * Add support for CRL distribution points extension. Add Certificate
17529 Policies and CRL distribution points documentation.
17533 * Move the autogenerated header file parts to crypto/opensslconf.h.
17537 * Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
17538 8 of keying material. Merlin has also confirmed interop with this fix
17539 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
17541 *Merlin Hughes <merlin@baltimore.ie>*
17543 * Fix lots of warnings.
17545 *Richard Levitte <levitte@stacken.kth.se>*
17547 * In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
17548 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
17550 *Richard Levitte <levitte@stacken.kth.se>*
17552 * Fix problems with sizeof(long) == 8.
17554 *Andy Polyakov <appro@fy.chalmers.se>*
17556 * Change functions to ANSI C.
17560 * Fix typos in error codes.
17562 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller*
17564 * Remove defunct assembler files from Configure.
17568 * SPARC v8 assembler BIGNUM implementation.
17570 *Andy Polyakov <appro@fy.chalmers.se>*
17572 * Support for Certificate Policies extension: both print and set.
17573 Various additions to support the r2i method this uses.
17577 * A lot of constification, and fix a bug in X509_NAME_oneline() that could
17578 return a const string when you are expecting an allocated buffer.
17582 * Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
17583 types DirectoryString and DisplayText.
17587 * Add code to allow r2i extensions to access the configuration database,
17588 add an LHASH database driver and add several ctx helper functions.
17592 * Fix an evil bug in bn_expand2() which caused various BN functions to
17593 fail when they extended the size of a BIGNUM.
17597 * Various utility functions to handle SXNet extension. Modify mkdef.pl to
17598 support typesafe stack.
17602 * Fix typo in SSL_[gs]et_options().
17604 *Nils Frostberg <nils@medcom.se>*
17606 * Delete various functions and files that belonged to the (now obsolete)
17607 old X509V3 handling code.
17611 * New Configure option "rsaref".
17615 * Don't auto-generate pem.h.
17619 * Introduce type-safe ASN.1 SETs.
17623 * Convert various additional casted stacks to type-safe STACK_OF() variants.
17625 *Ben Laurie, Ralf S. Engelschall, Steve Henson*
17627 * Introduce type-safe STACKs. This will almost certainly break lots of code
17628 that links with OpenSSL (well at least cause lots of warnings), but fear
17629 not: the conversion is trivial, and it eliminates loads of evil casts. A
17630 few STACKed things have been converted already. Feel free to convert more.
17631 In the fullness of time, I'll do away with the STACK type altogether.
17635 * Add `openssl ca -revoke <certfile>` facility which revokes a certificate
17636 specified in `<certfile>` by updating the entry in the index.txt file.
17637 This way one no longer has to edit the index.txt file manually for
17638 revoking a certificate. The -revoke option does the gory details now.
17640 *Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall*
17642 * Fix `openssl crl -noout -text` combination where `-noout` killed the
17643 `-text` option at all and this way the `-noout -text` combination was
17644 inconsistent in `openssl crl` with the friends in `openssl x509|rsa|dsa`.
17646 *Ralf S. Engelschall*
17648 * Make sure a corresponding plain text error message exists for the
17649 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
17650 verify callback function determined that a certificate was revoked.
17652 *Ralf S. Engelschall*
17654 * Bugfix: In test/testenc, don't test `openssl <cipher>` for
17655 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
17656 all available ciphers including rc5, which was forgotten until now.
17657 In order to let the testing shell script know which algorithms
17658 are available, a new (up to now undocumented) command
17659 `openssl list-cipher-commands` is used.
17663 * Bugfix: s_client occasionally would sleep in select() when
17664 it should have checked SSL_pending() first.
17668 * New functions DSA_do_sign and DSA_do_verify to provide access to
17669 the raw DSA values prior to ASN.1 encoding.
17673 * Tweaks to Configure
17675 *Niels Poppe <niels@netbox.org>*
17677 * Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
17682 * New variables $(RANLIB) and $(PERL) in the Makefiles.
17686 * New config option to avoid instructions that are illegal on the 80386.
17687 The default code is faster, but requires at least a 486.
17691 * Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
17692 SSL2_SERVER_VERSION (not used at all) macros, which are now the
17693 same as SSL2_VERSION anyway.
17697 * New "-showcerts" option for s_client.
17701 * Still more PKCS#12 integration. Add pkcs12 application to openssl
17702 application. Various cleanups and fixes.
17706 * More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
17707 modify error routines to work internally. Add error codes and PBE init
17708 to library startup routines.
17712 * Further PKCS#12 integration. Added password based encryption, PKCS#8 and
17713 packing functions to asn1 and evp. Changed function names and error
17714 codes along the way.
17718 * PKCS12 integration: and so it begins... First of several patches to
17719 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
17720 objects to objects.h
17724 * Add a new 'indent' option to some X509V3 extension code. Initial ASN1
17725 and display support for Thawte strong extranet extension.
17729 * Add LinuxPPC support.
17731 *Jeff Dubrule <igor@pobox.org>*
17733 * Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
17734 bn_div_words in alpha.s.
17736 *Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie*
17738 * Make sure the RSA OAEP test is skipped under -DRSAref because
17739 OAEP isn't supported when OpenSSL is built with RSAref.
17741 *Ulf Moeller <ulf@fitug.de>*
17743 * Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
17744 so they no longer are missing under -DNOPROTO.
17746 *Soren S. Jorvang <soren@t.dk>*
17748 ### Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
17750 * Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
17751 doesn't work when the session is reused. Coming soon!
17755 * Fix a security hole, that allows sessions to be reused in the wrong
17756 context thus bypassing client cert protection! All software that uses
17757 client certs and session caches in multiple contexts NEEDS PATCHING to
17758 allow session reuse! A fuller solution is in the works.
17760 *Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)*
17762 * Some more source tree cleanups (removed obsolete files
17763 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
17764 permission on "config" script to be executable) and a fix for the INSTALL
17767 *Ulf Moeller <ulf@fitug.de>*
17769 * Remove some legacy and erroneous uses of malloc, free instead of
17772 *Lennart Bang <lob@netstream.se>, with minor changes by Steve*
17774 * Make rsa_oaep_test return non-zero on error.
17776 *Ulf Moeller <ulf@fitug.de>*
17778 * Add support for native Solaris shared libraries. Configure
17779 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
17780 if someone would make that last step automatic.
17782 *Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>*
17784 * ctx_size was not built with the right compiler during "make links". Fixed.
17788 * Change the meaning of 'ALL' in the cipher list. It now means "everything
17789 except NULL ciphers". This means the default cipher list will no longer
17790 enable NULL ciphers. They need to be specifically enabled e.g. with
17791 the string "DEFAULT:eNULL".
17795 * Fix to RSA private encryption routines: if p < q then it would
17796 occasionally produce an invalid result. This will only happen with
17797 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
17801 * Be less restrictive and allow also `perl util/perlpath.pl
17802 /path/to/bin/perl` in addition to `perl util/perlpath.pl /path/to/bin`,
17803 because this way one can also use an interpreter named `perl5` (which is
17804 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
17805 installed as `perl`).
17807 *Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
17809 * Let util/clean-depend.pl work also with older Perl 5.00x versions.
17811 *Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
17813 * Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
17814 advapi32.lib to Win32 build and change the pem test comparison
17815 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
17816 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
17817 and crypto/des/ede_cbcm_enc.c.
17821 * DES quad checksum was broken on big-endian architectures. Fixed.
17825 * Comment out two functions in bio.h that aren't implemented. Fix up the
17826 Win32 test batch file so it (might) work again. The Win32 test batch file
17827 is horrible: I feel ill....
17831 * Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
17832 in e_os.h. Audit of header files to check ANSI and non ANSI
17833 sections: 10 functions were absent from non ANSI section and not exported
17834 from Windows DLLs. Fixed up libeay.num for new functions.
17838 * Make `openssl version` output lines consistent.
17840 *Ralf S. Engelschall*
17842 * Fix Win32 symbol export lists for BIO functions: Added
17843 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
17844 to ms/libeay{16,32}.def.
17846 *Ralf S. Engelschall*
17848 * Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
17849 fine under Unix and passes some trivial tests I've now added. But the
17850 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
17851 added to make sure no one expects that this stuff really works in the
17852 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
17853 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
17856 *Ralf S. Engelschall*
17858 * Fix the generation of two part addresses in perl.
17860 *Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie*
17862 * Add config entry for Linux on MIPS.
17864 *John Tobey <jtobey@channel1.com>*
17866 * Make links whenever Configure is run, unless we are on Windoze.
17870 * Permit extensions to be added to CRLs using crl_section in openssl.cnf.
17871 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
17876 * Add a useful kludge to allow package maintainers to specify compiler and
17877 other platforms details on the command line without having to patch the
17878 Configure script every time: One now can use
17879 `perl Configure <id>:<details>`,
17880 i.e. platform ids are allowed to have details appended
17881 to them (separated by colons). This is treated as there would be a static
17882 pre-configured entry in Configure's %table under key `<id>` with value
17883 `<details>` and `perl Configure <id>` is called. So, when you want to
17884 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
17885 assembler stuff you can use `perl Configure "FreeBSD-elf:pgcc:-O6:::"`
17886 now, which overrides the FreeBSD-elf entry on-the-fly.
17888 *Ralf S. Engelschall*
17890 * Disable new TLS1 ciphersuites by default: they aren't official yet.
17894 * Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
17895 on the `perl Configure ...` command line. This way one can compile
17896 OpenSSL libraries with Position Independent Code (PIC) which is needed
17897 for linking it into DSOs.
17899 *Ralf S. Engelschall*
17901 * Remarkably, export ciphers were totally broken and no-one had noticed!
17906 * Cleaned up the LICENSE document: The official contact for any license
17907 questions now is the OpenSSL core team under openssl-core@openssl.org.
17908 And add a paragraph about the dual-license situation to make sure people
17909 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
17910 to the OpenSSL toolkit.
17912 *Ralf S. Engelschall*
17914 * General source tree makefile cleanups: Made `making xxx in yyy...`
17915 display consistent in the source tree and replaced `/bin/rm` by `rm`.
17916 Additionally cleaned up the `make links` target: Remove unnecessary
17917 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
17918 to speed processing and no longer clutter the display with confusing
17919 stuff. Instead only the actually done links are displayed.
17921 *Ralf S. Engelschall*
17923 * Permit null encryption ciphersuites, used for authentication only. It used
17924 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
17925 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
17930 * Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
17931 signed attributes when verifying signatures (this would break them),
17932 the detached data encoding was wrong and public keys obtained using
17933 X509_get_pubkey() weren't freed.
17937 * Add text documentation for the BUFFER functions. Also added a work around
17938 to a Win95 console bug. This was triggered by the password read stuff: the
17939 last character typed gets carried over to the next fread(). If you were
17940 generating a new cert request using 'req' for example then the last
17941 character of the passphrase would be CR which would then enter the first
17946 * Added the new 'Includes OpenSSL Cryptography Software' button as
17947 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
17948 button and can be used by applications based on OpenSSL to show the
17949 relationship to the OpenSSL project.
17951 *Ralf S. Engelschall*
17953 * Remove confusing variables in function signatures in files
17954 ssl/ssl_lib.c and ssl/ssl.h.
17956 *Lennart Bong <lob@kulthea.stacken.kth.se>*
17958 * Don't install bss_file.c under PREFIX/include/
17960 *Lennart Bong <lob@kulthea.stacken.kth.se>*
17962 * Get the Win32 compile working again. Modify mkdef.pl so it can handle
17963 functions that return function pointers and has support for NT specific
17964 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
17965 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
17966 unsigned to signed types: this was killing the Win32 compile.
17970 * Add new certificate file to stack functions,
17971 SSL_add_dir_cert_subjects_to_stack() and
17972 SSL_add_file_cert_subjects_to_stack(). These largely supplant
17973 SSL_load_client_CA_file(), and can be used to add multiple certs easily
17974 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
17975 This means that Apache-SSL and similar packages don't have to mess around
17976 to add as many CAs as they want to the preferred list.
17980 * Experiment with doxygen documentation. Currently only partially applied to
17982 See <http://www.stack.nl/~dimitri/doxygen/index.html>, and run doxygen with
17983 openssl.doxy as the configuration file.
17987 * Get rid of remaining C++-style comments which strict C compilers hate.
17989 *Ralf S. Engelschall, pointed out by Carlos Amengual*
17991 * Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
17992 compiled in by default: it has problems with large keys.
17996 * Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
17997 DH private keys and/or callback functions which directly correspond to
17998 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
17999 is needed for applications which have to configure certificates on a
18000 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
18002 For the RSA certificate situation is makes no difference, but
18003 for the DSA certificate situation this fixes the "no shared cipher"
18004 problem where the OpenSSL cipher selection procedure failed because the
18005 temporary keys were not overtaken from the context and the API provided
18006 no way to reconfigure them.
18007 The new functions now let applications reconfigure the stuff and they
18008 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
18009 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
18010 non-public-API function ssl_cert_instantiate() is used as a helper
18011 function and also to reduce code redundancy inside ssl_rsa.c.
18013 *Ralf S. Engelschall*
18015 * Move s_server -dcert and -dkey options out of the undocumented feature
18016 area because they are useful for the DSA situation and should be
18017 recognized by the users.
18019 *Ralf S. Engelschall*
18021 * Fix the cipher decision scheme for export ciphers: the export bits are
18022 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
18023 SSL_EXP_MASK. So, the original variable has to be used instead of the
18024 already masked variable.
18026 *Richard Levitte <levitte@stacken.kth.se>*
18028 * Fix `port` variable from `int` to `unsigned int` in crypto/bio/b_sock.c
18030 *Richard Levitte <levitte@stacken.kth.se>*
18032 * Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
18033 from `int` to `unsigned int` because it is a length and initialized by
18034 EVP_DigestFinal() which expects an `unsigned int *`.
18036 *Richard Levitte <levitte@stacken.kth.se>*
18038 * Don't hard-code path to Perl interpreter on shebang line of Configure
18039 script. Instead use the usual Shell->Perl transition trick.
18041 *Ralf S. Engelschall*
18043 * Make `openssl x509 -noout -modulus`' functional also for DSA certificates
18044 (in addition to RSA certificates) to match the behaviour of `openssl dsa
18045 -noout -modulus` as it's already the case for `openssl rsa -noout
18046 -modulus`. For RSA the -modulus is the real "modulus" while for DSA
18047 currently the public key is printed (a decision which was already done by
18048 `openssl dsa -modulus` in the past) which serves a similar purpose.
18049 Additionally the NO_RSA no longer completely removes the whole -modulus
18050 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
18053 *Ralf S. Engelschall*
18055 * Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
18056 BIO. See the source (crypto/evp/bio_ok.c) for more info.
18058 *Arne Ansper <arne@ats.cyber.ee>*
18060 * Dump the old yucky req code that tried (and failed) to allow raw OIDs
18061 to be added. Now both 'req' and 'ca' can use new objects defined in the
18066 * Add cool BIO that does syslog (or event log on NT).
18068 *Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie*
18070 * Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
18071 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
18072 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
18073 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
18077 * Add preliminary config info for new extension code.
18081 * Make RSA_NO_PADDING really use no padding.
18083 *Ulf Moeller <ulf@fitug.de>*
18085 * Generate errors when private/public key check is done.
18089 * Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
18090 for some CRL extensions and new objects added.
18094 * Really fix the ASN1 IMPLICIT bug this time... Partial support for private
18095 key usage extension and fuller support for authority key id.
18099 * Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
18100 padding method for RSA, which is recommended for new applications in PKCS
18101 #1 v2.0 (RFC 2437, October 1998).
18102 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
18103 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
18104 against Bleichbacher's attack on RSA.
18105 *Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
18108 * Updates to the new SSL compression code
18110 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
18112 * Fix so that the version number in the master secret, when passed
18113 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
18114 (because the server will not accept higher), that the version number
18115 is 0x03,0x01, not 0x03,0x00
18117 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
18119 * Run extensive memory leak checks on SSL commands. Fixed *lots* of memory
18120 leaks in `ssl/` relating to new `X509_get_pubkey()` behaviour. Also fixes
18121 in `apps/` and an unrelated leak in `crypto/dsa/dsa_vrf.c`.
18125 * Support for RAW extensions where an arbitrary extension can be
18126 created by including its DER encoding. See `apps/openssl.cnf` for
18131 * Make sure latest Perl versions don't interpret some generated C array
18132 code as Perl array code in the crypto/err/err_genc.pl script.
18134 *Lars Weber <3weber@informatik.uni-hamburg.de>*
18136 * Modify ms/do_ms.bat to not generate assembly language makefiles since
18137 not many people have the assembler. Various Win32 compilation fixes and
18138 update to the INSTALL.W32 file with (hopefully) more accurate Win32
18139 build instructions.
18143 * Modify configure script 'Configure' to automatically create crypto/date.h
18144 file under Win32 and also build pem.h from pem.org. New script
18145 util/mkfiles.pl to create the MINFO file on environments that can't do a
18146 'make files': perl util/mkfiles.pl >MINFO should work.
18150 * Major rework of DES function declarations, in the pursuit of correctness
18151 and purity. As a result, many evil casts evaporated, and some weirdness,
18152 too. You may find this causes warnings in your code. Zapping your evil
18153 casts will probably fix them. Mostly.
18157 * Fix for a typo in asn1.h. Bug fix to object creation script
18158 obj_dat.pl. It considered a zero in an object definition to mean
18159 "end of object": none of the objects in objects.h have any zeros
18160 so it wasn't spotted.
18162 *Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>*
18164 * Add support for Triple DES Cipher Block Chaining with Output Feedback
18165 Masking (CBCM). In the absence of test vectors, the best I have been able
18166 to do is check that the decrypt undoes the encrypt, so far. Send me test
18167 vectors if you have them.
18171 * Correct calculation of key length for export ciphers (too much space was
18172 allocated for null ciphers). This has not been tested!
18176 * Modifications to the mkdef.pl for Win32 DEF file creation. The usage
18177 message is now correct (it understands "crypto" and "ssl" on its
18178 command line). There is also now an "update" option. This will update
18179 the util/ssleay.num and util/libeay.num files with any new functions.
18181 perl util/mkdef.pl crypto ssl update
18182 it will update them.
18186 * Overhauled the Perl interface:
18187 - ported BN stuff to OpenSSL's different BN library
18188 - made the perl/ source tree CVS-aware
18189 - renamed the package from SSLeay to OpenSSL (the files still contain
18190 their history because I've copied them in the repository)
18191 - removed obsolete files (the test scripts will be replaced
18192 by better Test::Harness variants in the future)
18194 *Ralf S. Engelschall*
18196 * First cut for a very conservative source tree cleanup:
18197 1. merge various obsolete readme texts into doc/ssleay.txt
18198 where we collect the old documents and readme texts.
18199 2. remove the first part of files where I'm already sure that we no
18200 longer need them because of three reasons: either they are just temporary
18201 files which were left by Eric or they are preserved original files where
18202 I've verified that the diff is also available in the CVS via "cvs diff
18203 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
18204 the crypto/md/ stuff).
18206 *Ralf S. Engelschall*
18208 * More extension code. Incomplete support for subject and issuer alt
18209 name, issuer and authority key id. Change the i2v function parameters
18210 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
18211 what that's for :-) Fix to ASN1 macro which messed up
18212 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
18216 * Preliminary support for ENUMERATED type. This is largely copied from the
18221 * Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
18223 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
18225 * Make sure `make rehash` target really finds the `openssl` program.
18227 *Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
18229 * Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
18230 like to hear about it if this slows down other processors.
18234 * Add CygWin32 platform information to Configure script.
18236 *Alan Batie <batie@aahz.jf.intel.com>*
18238 * Fixed ms/32all.bat script: `no_asm` -> `no-asm`
18240 *Rainer W. Gerling <gerling@mpg-gv.mpg.de>*
18242 * New program nseq to manipulate netscape certificate sequences
18246 * Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
18251 * Fixes to BN code. Previously the default was to define BN_RECURSION
18252 but the BN code had some problems that would cause failures when
18253 doing certificate verification and some other functions.
18255 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
18257 * Add ASN1 and PEM code to support netscape certificate sequences.
18261 * Add ASN1 and PEM code to support netscape certificate sequences.
18265 * Add several PKIX and private extended key usage OIDs.
18269 * Modify the 'ca' program to handle the new extension code. Modify
18270 openssl.cnf for new extension format, add comments.
18274 * More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
18275 and add a sample to openssl.cnf so req -x509 now adds appropriate
18280 * Continued X509 V3 changes. Add to other makefiles, integrate with the
18281 error code, add initial support to X509_print() and x509 application.
18285 * Takes a deep breath and start adding X509 V3 extension support code. Add
18286 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
18287 stuff is currently isolated and isn't even compiled yet.
18291 * Continuing patches for GeneralizedTime. Fix up certificate and CRL
18292 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
18293 Removed the versions check from X509 routines when loading extensions:
18294 this allows certain broken certificates that don't set the version
18295 properly to be processed.
18299 * Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
18300 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
18301 can still be regenerated with "make depend".
18305 * Spelling mistake in C version of CAST-128.
18307 *Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>*
18309 * Changes to the error generation code. The perl script err-code.pl
18310 now reads in the old error codes and retains the old numbers, only
18311 adding new ones if necessary. It also only changes the .err files if new
18312 codes are added. The makefiles have been modified to only insert errors
18313 when needed (to avoid needlessly modifying header files). This is done
18314 by only inserting errors if the .err file is newer than the auto generated
18315 C file. To rebuild all the error codes from scratch (the old behaviour)
18316 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
18317 or delete all the .err files.
18321 * CAST-128 was incorrectly implemented for short keys. The C version has
18322 been fixed, but is untested. The assembler versions are also fixed, but
18323 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
18324 to regenerate it if needed.
18325 *Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
18326 Hagino <itojun@kame.net>*
18328 * File was opened incorrectly in randfile.c.
18330 *Ulf Möller <ulf@fitug.de>*
18332 * Beginning of support for GeneralizedTime. d2i, i2d, check and print
18333 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
18334 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
18335 al: it's just almost always a UTCTime. Note this patch adds new error
18336 codes so do a "make errors" if there are problems.
18340 * Correct Linux 1 recognition in config.
18342 *Ulf Möller <ulf@fitug.de>*
18344 * Remove pointless MD5 hash when using DSA keys in ca.
18346 *Anonymous <nobody@replay.com>*
18348 * Generate an error if given an empty string as a cert directory. Also
18349 generate an error if handed NULL (previously returned 0 to indicate an
18350 error, but didn't set one).
18352 *Ben Laurie, reported by Anonymous <nobody@replay.com>*
18354 * Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
18358 * Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
18359 parameters. This was causing a warning which killed off the Win32 compile.
18363 * Remove C++ style comments from crypto/bn/bn_local.h.
18365 *Neil Costigan <neil.costigan@celocom.com>*
18367 * The function OBJ_txt2nid was broken. It was supposed to return a nid
18368 based on a text string, looking up short and long names and finally
18369 "dot" format. The "dot" format stuff didn't work. Added new function
18370 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
18371 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
18372 OID is not part of the table.
18376 * Add prototypes to X509 lookup/verify methods, fixing a bug in
18377 X509_LOOKUP_by_alias().
18381 * Sort openssl functions by name.
18385 * Get the `gendsa` command working and add it to the `list` command. Remove
18386 encryption from sample DSA keys (in case anyone is interested the password
18391 * Make *all* `*_free` functions accept a NULL pointer.
18393 *Frans Heymans <fheymans@isaserver.be>*
18395 * If a DH key is generated in s3_srvr.c, don't blow it by trying to use
18398 *Anonymous <nobody@replay.com>*
18400 * s_server should send the CAfile as acceptable CAs, not its own cert.
18402 *Bodo Moeller <3moeller@informatik.uni-hamburg.de>*
18404 * Don't blow it for numeric `-newkey` arguments to `apps/req`.
18406 *Bodo Moeller <3moeller@informatik.uni-hamburg.de>*
18408 * Temp key "for export" tests were wrong in s3_srvr.c.
18410 *Anonymous <nobody@replay.com>*
18412 * Add prototype for temp key callback functions
18413 SSL_CTX_set_tmp_{rsa,dh}_callback().
18417 * Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
18418 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
18422 * X509_name_add_entry() freed the wrong thing after an error.
18424 *Arne Ansper <arne@ats.cyber.ee>*
18426 * rsa_eay.c would attempt to free a NULL context.
18428 *Arne Ansper <arne@ats.cyber.ee>*
18430 * BIO_s_socket() had a broken should_retry() on Windoze.
18432 *Arne Ansper <arne@ats.cyber.ee>*
18434 * BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
18436 *Arne Ansper <arne@ats.cyber.ee>*
18438 * Make sure the already existing X509_STORE->depth variable is initialized
18439 in X509_STORE_new(), but document the fact that this variable is still
18440 unused in the certificate verification process.
18442 *Ralf S. Engelschall*
18444 * Fix the various library and `apps/` files to free up pkeys obtained from
18445 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
18449 * Fix reference counting in X509_PUBKEY_get(). This makes
18450 demos/maurice/example2.c work, amongst others, probably.
18452 *Steve Henson and Ben Laurie*
18454 * First cut of a cleanup for `apps/`. First the `ssleay` program is now named
18455 `openssl` and second, the shortcut symlinks for the `openssl <command>`
18456 are no longer created. This way we have a single and consistent command
18457 line interface `openssl <command>`, similar to `cvs <command>`.
18459 *Ralf S. Engelschall, Paul Sutton and Ben Laurie*
18461 * ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
18462 BIT STRING wrapper always have zero unused bits.
18466 * Add CA.pl, perl version of CA.sh, add extended key usage OID.
18470 * Make the top-level INSTALL documentation easier to understand.
18474 * Makefiles updated to exit if an error occurs in a sub-directory
18475 make (including if user presses ^C) [Paul Sutton]
18477 * Make Montgomery context stuff explicit in RSA data structure.
18481 * Fix build order of pem and err to allow for generated pem.h.
18485 * Fix renumbering bug in X509_NAME_delete_entry().
18489 * Enhanced the err-ins.pl script so it makes the error library number
18490 global and can add a library name. This is needed for external ASN1 and
18491 other error libraries.
18495 * Fixed sk_insert which never worked properly.
18499 * Fix ASN1 macros so they can handle indefinite length constructed
18500 EXPLICIT tags. Some non standard certificates use these: they can now
18505 * Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
18506 into a single doc/ssleay.txt bundle. This way the information is still
18507 preserved but no longer messes up this directory. Now it's new room for
18508 the new set of documentation files.
18510 *Ralf S. Engelschall*
18512 * SETs were incorrectly DER encoded. This was a major pain, because they
18513 shared code with SEQUENCEs, which aren't coded the same. This means that
18514 almost everything to do with SETs or SEQUENCEs has either changed name or
18515 number of arguments.
18517 *Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>*
18519 * Fix test data to work with the above.
18523 * Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
18524 was already fixed by Eric for 0.9.1 it seems.
18526 *Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>*
18528 * Autodetect FreeBSD3.
18532 * Fix various bugs in Configure. This affects the following platforms:
18536 unixware-2.0-pentium
18541 * Eliminate generated files from CVS. Reorder tests to regenerate files
18542 before they are needed.
18546 * Generate Makefile.ssl from Makefile.org (to keep CVS happy).
18550 ### Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
18552 * Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
18553 changed SSLeay to OpenSSL in version strings.
18555 *Ralf S. Engelschall*
18557 * Some fixups to the top-level documents.
18561 * Fixed the nasty bug where rsaref.h was not found under compile-time
18562 because the symlink to include/ was missing.
18564 *Ralf S. Engelschall*
18566 * Incorporated the popular no-RSA/DSA-only patches
18567 which allow to compile a RSA-free SSLeay.
18569 *Andrew Cooke / Interrader Ldt., Ralf S. Engelschall*
18571 * Fixed nasty rehash problem under `make -f Makefile.ssl links`
18572 when "ssleay" is still not found.
18574 *Ralf S. Engelschall*
18576 * Added more platforms to Configure: Cray T3E, HPUX 11,
18578 *Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>*
18580 * Updated the README file.
18582 *Ralf S. Engelschall*
18584 * Added various .cvsignore files in the CVS repository subdirs
18585 to make a "cvs update" really silent.
18587 *Ralf S. Engelschall*
18589 * Recompiled the error-definition header files and added
18590 missing symbols to the Win32 linker tables.
18592 *Ralf S. Engelschall*
18594 * Cleaned up the top-level documents;
18595 o new files: CHANGES and LICENSE
18596 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
18597 o merged COPYRIGHT into LICENSE
18598 o removed obsolete TODO file
18599 o renamed MICROSOFT to INSTALL.W32
18601 *Ralf S. Engelschall*
18603 * Removed dummy files from the 0.9.1b source tree:
18604 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
18605 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
18606 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
18607 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
18608 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
18610 *Ralf S. Engelschall*
18612 * Added various platform portability fixes.
18616 * The Genesis of the OpenSSL rpject:
18617 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
18618 Young and Tim J. Hudson created while they were working for C2Net until
18621 *The OpenSSL Project*
18623 ### Changes between 0.9.0b and 0.9.1b [not released]
18625 * Updated a few CA certificates under certs/
18629 * Changed some BIGNUM api stuff.
18633 * Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
18634 DGUX x86, Linux Alpha, etc.
18638 * New COMP library [crypto/comp/] for SSL Record Layer Compression:
18639 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
18644 * Add -strparse option to asn1pars program which parses nested
18647 *Dr Stephen Henson <shenson@bigfoot.com>*
18649 * Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
18653 * DSA fix for "ca" program.
18657 * Added "-genkey" option to "dsaparam" program.
18661 * Added RIPE MD160 (rmd160) message digest.
18665 * Added -a (all) option to "ssleay version" command.
18669 * Added PLATFORM define which is the id given to Configure.
18673 * Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
18677 * Extended the ASN.1 parser routines.
18681 * Extended BIO routines to support REUSEADDR, seek, tell, etc.
18685 * Added a BN_CTX to the BN library.
18689 * Fixed the weak key values in DES library
18693 * Changed API in EVP library for cipher aliases.
18697 * Added support for RC2/64bit cipher.
18701 * Converted the lhash library to the crypto/mem.c functions.
18705 * Added more recognized ASN.1 object ids.
18709 * Added more RSA padding checks for SSL/TLS.
18713 * Added BIO proxy/filter functionality.
18717 * Added extra_certs to SSL_CTX which can be used
18718 send extra CA certificates to the client in the CA cert chain sending
18719 process. It can be configured with SSL_CTX_add_extra_chain_cert().
18723 * Now Fortezza is denied in the authentication phase because
18724 this is key exchange mechanism is not supported by SSLeay at all.
18728 * Additional PKCS1 checks.
18732 * Support the string "TLSv1" for all TLS v1 ciphers.
18736 * Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
18737 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
18741 * Fixed a few memory leaks.
18745 * Fixed various code and comment typos.
18749 * A minor bug in ssl/s3_clnt.c where there would always be 4 0
18750 bytes sent in the client random.
18752 *Edward Bishop <ebishop@spyglass.com>*
18756 [CVE-2020-1971]: https://www.openssl.org/news/vulnerabilities.html#CVE-2020-1971
18757 [CVE-2020-1967]: https://www.openssl.org/news/vulnerabilities.html#CVE-2020-1967
18758 [CVE-2019-1563]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1563
18759 [CVE-2019-1559]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1559
18760 [CVE-2019-1552]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1552
18761 [CVE-2019-1551]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1551
18762 [CVE-2019-1549]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1549
18763 [CVE-2019-1547]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1547
18764 [CVE-2019-1543]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1543
18765 [CVE-2018-5407]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-5407
18766 [CVE-2018-0739]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0739
18767 [CVE-2018-0737]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0737
18768 [CVE-2018-0735]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0735
18769 [CVE-2018-0734]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0734
18770 [CVE-2018-0733]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0733
18771 [CVE-2018-0732]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0732
18772 [CVE-2017-3738]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3738
18773 [CVE-2017-3737]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3737
18774 [CVE-2017-3736]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3736
18775 [CVE-2017-3735]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3735
18776 [CVE-2017-3733]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3733
18777 [CVE-2017-3732]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3732
18778 [CVE-2017-3731]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3731
18779 [CVE-2017-3730]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3730
18780 [CVE-2016-7055]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7055
18781 [CVE-2016-7054]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7054
18782 [CVE-2016-7053]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7053
18783 [CVE-2016-7052]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7052
18784 [CVE-2016-6309]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6309
18785 [CVE-2016-6308]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6308
18786 [CVE-2016-6307]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6307
18787 [CVE-2016-6306]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6306
18788 [CVE-2016-6305]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6305
18789 [CVE-2016-6304]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6304
18790 [CVE-2016-6303]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6303
18791 [CVE-2016-6302]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6302
18792 [CVE-2016-2183]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2183
18793 [CVE-2016-2182]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2182
18794 [CVE-2016-2181]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2181
18795 [CVE-2016-2180]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2180
18796 [CVE-2016-2179]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2179
18797 [CVE-2016-2178]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2178
18798 [CVE-2016-2177]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2177
18799 [CVE-2016-2176]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2176
18800 [CVE-2016-2109]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2109
18801 [CVE-2016-2107]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2107
18802 [CVE-2016-2106]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2106
18803 [CVE-2016-2105]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2105
18804 [CVE-2016-0800]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0800
18805 [CVE-2016-0799]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0799
18806 [CVE-2016-0798]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0798
18807 [CVE-2016-0797]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0797
18808 [CVE-2016-0705]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0705
18809 [CVE-2016-0702]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0702
18810 [CVE-2016-0701]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0701
18811 [CVE-2015-3197]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3197
18812 [CVE-2015-3196]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3196
18813 [CVE-2015-3195]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3195
18814 [CVE-2015-3194]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3194
18815 [CVE-2015-3193]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3193
18816 [CVE-2015-1793]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1793
18817 [CVE-2015-1792]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1792
18818 [CVE-2015-1791]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1791
18819 [CVE-2015-1790]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1790
18820 [CVE-2015-1789]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1789
18821 [CVE-2015-1788]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1788
18822 [CVE-2015-1787]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1787
18823 [CVE-2015-0293]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0293
18824 [CVE-2015-0291]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0291
18825 [CVE-2015-0290]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0290
18826 [CVE-2015-0289]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0289
18827 [CVE-2015-0288]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0288
18828 [CVE-2015-0287]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0287
18829 [CVE-2015-0286]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0286
18830 [CVE-2015-0285]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0285
18831 [CVE-2015-0209]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0209
18832 [CVE-2015-0208]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0208
18833 [CVE-2015-0207]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0207
18834 [CVE-2015-0206]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0206
18835 [CVE-2015-0205]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0205
18836 [CVE-2015-0204]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0204
18837 [CVE-2014-8275]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-8275
18838 [CVE-2014-5139]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-5139
18839 [CVE-2014-3572]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3572
18840 [CVE-2014-3571]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3571
18841 [CVE-2014-3570]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3570
18842 [CVE-2014-3569]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3569
18843 [CVE-2014-3568]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3568
18844 [CVE-2014-3567]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3567
18845 [CVE-2014-3566]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3566
18846 [CVE-2014-3513]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3513
18847 [CVE-2014-3512]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3512
18848 [CVE-2014-3511]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3511
18849 [CVE-2014-3510]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3510
18850 [CVE-2014-3509]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3509
18851 [CVE-2014-3508]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3508
18852 [CVE-2014-3507]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3507
18853 [CVE-2014-3506]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3506
18854 [CVE-2014-3505]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3505
18855 [CVE-2014-3470]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3470
18856 [CVE-2014-0224]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0224
18857 [CVE-2014-0221]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0221
18858 [CVE-2014-0195]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0195
18859 [CVE-2014-0160]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0160
18860 [CVE-2014-0076]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0076
18861 [CVE-2013-6450]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-6450
18862 [CVE-2013-4353]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-4353
18863 [CVE-2013-0169]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-0169
18864 [CVE-2013-0166]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-0166
18865 [CVE-2012-2686]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2686
18866 [CVE-2012-2333]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2333
18867 [CVE-2012-2110]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2110
18868 [CVE-2012-0884]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0884
18869 [CVE-2012-0050]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0050
18870 [CVE-2012-0027]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0027
18871 [CVE-2011-4619]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4619
18872 [CVE-2011-4577]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4577
18873 [CVE-2011-4576]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4576
18874 [CVE-2011-4109]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4109
18875 [CVE-2011-4108]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4108
18876 [CVE-2011-3210]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-3210
18877 [CVE-2011-3207]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-3207
18878 [CVE-2011-0014]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-0014
18879 [CVE-2010-4252]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-4252
18880 [CVE-2010-4180]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-4180
18881 [CVE-2010-3864]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-3864
18882 [CVE-2010-1633]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-1633
18883 [CVE-2010-0740]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-0740
18884 [CVE-2010-0433]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-0433
18885 [CVE-2009-4355]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-4355
18886 [CVE-2009-3555]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-3555
18887 [CVE-2009-3245]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-3245
18888 [CVE-2009-1386]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1386
18889 [CVE-2009-1379]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1379
18890 [CVE-2009-1378]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1378
18891 [CVE-2009-1377]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1377
18892 [CVE-2009-0789]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0789
18893 [CVE-2009-0591]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0591
18894 [CVE-2009-0590]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0590
18895 [CVE-2008-5077]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-5077
18896 [CVE-2008-1678]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-1678
18897 [CVE-2008-1672]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-1672
18898 [CVE-2008-0891]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-0891
18899 [CVE-2007-5135]: https://www.openssl.org/news/vulnerabilities.html#CVE-2007-5135
18900 [CVE-2007-4995]: https://www.openssl.org/news/vulnerabilities.html#CVE-2007-4995
18901 [CVE-2006-4343]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-4343
18902 [CVE-2006-4339]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-4339
18903 [CVE-2006-3738]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-3738
18904 [CVE-2006-2940]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-2940
18905 [CVE-2006-2937]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-2937
18906 [CVE-2005-2969]: https://www.openssl.org/news/vulnerabilities.html#CVE-2005-2969
18907 [CVE-2004-0112]: https://www.openssl.org/news/vulnerabilities.html#CVE-2004-0112
18908 [CVE-2004-0079]: https://www.openssl.org/news/vulnerabilities.html#CVE-2004-0079
18909 [CVE-2003-0851]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0851
18910 [CVE-2003-0545]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0545
18911 [CVE-2003-0544]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0544
18912 [CVE-2003-0543]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0543
18913 [CVE-2003-0078]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0078
18914 [CVE-2002-0659]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0659
18915 [CVE-2002-0657]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0657
18916 [CVE-2002-0656]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0656
18917 [CVE-2002-0655]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0655