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 * The ERR_GET_FUNC() function was removed. With the loss of meaningful
34 function codes, this function can only cause problems for calling
39 * Add a configurable flag to output date formats as ISO 8601. Does not
40 change the default date format.
44 * Version of MSVC earlier than 1300 could get link warnings, which could
45 be suppressed if the undocumented -DI_CAN_LIVE_WITH_LNK4049 was set.
46 Support for this flag has been removed.
50 * Rework and make DEBUG macros consistent. Remove unused -DCONF_DEBUG,
51 -DBN_CTX_DEBUG, and REF_PRINT. Add a new tracing category and use it for
52 printing reference counts. Rename -DDEBUG_UNUSED to -DUNUSED_RESULT_DEBUG
53 Fix BN_DEBUG_RAND so it compiles and, when set, force DEBUG_RAND to be set
54 also. Rename engine_debug_ref to be ENGINE_REF_PRINT also for consistency.
58 * The signatures of the functions to get and set options on SSL and
59 SSL_CTX objects changed from "unsigned long" to "uint64_t" type.
60 Some source code changes may be required.
64 * The public definitions of conf_method_st and conf_st have been
65 deprecated. They will be made opaque in a future release.
67 *Rich Salz and Tomáš Mráz*
69 * Client-initiated renegotiation is disabled by default. To allow it, use
70 the -client_renegotiation option, the SSL_OP_ALLOW_CLIENT_RENEGOTIATION
71 flag, or the "ClientRenegotiation" config parameter as appropriate.
75 * Add "abspath" and "includedir" pragma's to config files, to prevent,
76 or modify relative pathname inclusion.
80 * OpenSSL includes a cryptographic module that is intended to be FIPS 140-2
81 validated. Please consult the README-FIPS and
82 README-PROVIDERS files, as well as the migration guide.
84 *OpenSSL team members and many third party contributors*
86 * For the key types DH and DHX the allowed settable parameters are now different.
90 * The openssl commands that read keys, certificates, and CRLs now
91 automatically detect the PEM or DER format of the input files.
93 *David von Oheimb, Richard Levitte, and Tomáš Mráz*
95 * Added enhanced PKCS#12 APIs which accept a library context.
99 * The default manual page suffix ($MANSUFFIX) has been changed to "ossl"
103 * Added support for Kernel TLS (KTLS).
105 *Boris Pismenny, John Baldwin and Andrew Gallatin*
107 * Support for RFC 5746 secure renegotiation is now required by default for
108 SSL or TLS connections to succeed.
112 * The signature of the `copy` functional parameter of the
113 EVP_PKEY_meth_set_copy() function has changed so its `src` argument is
114 now `const EVP_PKEY_CTX *` instead of `EVP_PKEY_CTX *`. Similarly
115 the signature of the `pub_decode` functional parameter of the
116 EVP_PKEY_asn1_set_public() function has changed so its `pub` argument is
117 now `const X509_PUBKEY *` instead of `X509_PUBKEY *`.
121 * The error return values from some control calls (ctrl) have changed.
125 * A public key check is now performed during EVP_PKEY_derive_set_peer().
129 * Many functions in the EVP_ namespace that are getters of values from
130 implementations or contexts were renamed to include get or get0 in their
131 names. Old names are provided as macro aliases for compatibility and
136 * The EVP_PKEY_CTRL_PKCS7_ENCRYPT, EVP_PKEY_CTRL_PKCS7_DECRYPT,
137 EVP_PKEY_CTRL_PKCS7_SIGN, EVP_PKEY_CTRL_CMS_ENCRYPT,
138 EVP_PKEY_CTRL_CMS_DECRYPT, and EVP_PKEY_CTRL_CMS_SIGN control operations
143 * The EVP_PKEY_public_check() and EVP_PKEY_param_check() functions now work for
146 * The output from the command line applications may have minor
151 * The output from numerous "printing" may have minor changes.
155 * Windows thread synchronization uses read/write primitives (SRWLock) when
156 supported by the OS, otherwise CriticalSection continues to be used.
160 * Add filter BIO BIO_f_readbuffer() that allows BIO_tell() and BIO_seek() to
161 work on read only BIO source/sinks that do not support these functions.
162 This allows piping or redirection of a file BIO using stdin to be buffered
163 into memory. This is used internally in OSSL_DECODER_from_bio().
167 * OSSL_STORE_INFO_get_type() may now return an additional value. In 1.1.1
168 this function would return one of the values OSSL_STORE_INFO_NAME,
169 OSSL_STORE_INFO_PKEY, OSSL_STORE_INFO_PARAMS, OSSL_STORE_INFO_CERT or
170 OSSL_STORE_INFO_CRL. Decoded public keys would previously have been reported
171 as type OSSL_STORE_INFO_PKEY in 1.1.1. In 3.0 decoded public keys are now
172 reported as having the new type OSSL_STORE_INFO_PUBKEY. Applications
173 using this function should be amended to handle the changed return value.
177 * Improved adherence to Enhanced Security Services (ESS, RFC 2634 and RFC 5035)
178 for the TSP and CMS Advanced Electronic Signatures (CAdES) implementations.
179 As required by RFC 5035 check both ESSCertID and ESSCertIDv2 if both present.
180 Correct the semantics of checking the validation chain in case ESSCertID{,v2}
181 contains more than one certificate identifier: This means that all
182 certificates referenced there MUST be part of the validation chain.
186 * The implementation of older EVP ciphers related to CAST, IDEA, SEED, RC2, RC4,
187 RC5, DESX and DES have been moved to the legacy provider.
191 * The implementation of the EVP digests MD2, MD4, MDC2, WHIRLPOOL and
192 RIPEMD-160 have been moved to the legacy provider.
196 * The deprecated function EVP_PKEY_get0() now returns NULL being called for a
201 * The deprecated functions EVP_PKEY_get0_RSA(),
202 EVP_PKEY_get0_DSA(), EVP_PKEY_get0_EC_KEY(), EVP_PKEY_get0_DH(),
203 EVP_PKEY_get0_hmac(), EVP_PKEY_get0_poly1305() and EVP_PKEY_get0_siphash() as
204 well as the similarly named "get1" functions behave differently in
209 * A number of functions handling low-level keys or engines were deprecated
210 including EVP_PKEY_set1_engine(), EVP_PKEY_get0_engine(), EVP_PKEY_assign(),
211 EVP_PKEY_get0(), EVP_PKEY_get0_hmac(), EVP_PKEY_get0_poly1305() and
212 EVP_PKEY_get0_siphash().
216 * PKCS#5 PBKDF1 key derivation has been moved from PKCS5_PBE_keyivgen() into
217 the legacy crypto provider as an EVP_KDF. Applications requiring this KDF
218 will need to load the legacy crypto provider. This includes these PBE
219 algorithms which use this KDF:
220 - NID_pbeWithMD2AndDES_CBC
221 - NID_pbeWithMD5AndDES_CBC
222 - NID_pbeWithSHA1AndRC2_CBC
223 - NID_pbeWithMD2AndRC2_CBC
224 - NID_pbeWithMD5AndRC2_CBC
225 - NID_pbeWithSHA1AndDES_CBC
229 * Deprecated obsolete BIO_set_callback(), BIO_get_callback(), and
230 BIO_debug_callback() functions.
234 * Deprecated obsolete EVP_PKEY_CTX_get0_dh_kdf_ukm() and
235 EVP_PKEY_CTX_get0_ecdh_kdf_ukm() functions.
239 * The RAND_METHOD APIs have been deprecated.
243 * The SRP APIs have been deprecated.
247 * Add a compile time option to prevent the caching of provider fetched
248 algorithms. This is enabled by including the no-cached-fetch option
249 at configuration time.
253 * pkcs12 now uses defaults of PBKDF2, AES and SHA-256, with a MAC iteration
254 count of PKCS12_DEFAULT_ITER.
256 *Tomáš Mráz and Sahana Prasad*
258 * The openssl speed command does not use low-level API calls anymore.
262 * Parallel dual-prime 1024-bit modular exponentiation for AVX512_IFMA
265 *Ilya Albrekht, Sergey Kirillov, Andrey Matyukov (Intel Corp)*
267 * Combining the Configure options no-ec and no-dh no longer disables TLSv1.3.
271 * Implemented support for fully "pluggable" TLSv1.3 groups. This means that
272 providers may supply their own group implementations (using either the "key
273 exchange" or the "key encapsulation" methods) which will automatically be
274 detected and used by libssl.
276 *Matt Caswell, Nicola Tuveri*
278 * The undocumented function X509_certificate_type() has been deprecated;
282 * Deprecated the obsolete BN_pseudo_rand() and BN_pseudo_rand_range().
286 * Removed RSA padding mode for SSLv23 (which was only used for
287 SSLv2). This includes the functions RSA_padding_check_SSLv23() and
288 RSA_padding_add_SSLv23() and the `-ssl` option in the deprecated
293 * Deprecated the obsolete X9.31 RSA key generation related functions.
297 * The default key generation method for the regular 2-prime RSA keys was
298 changed to the FIPS 186-4 B.3.6 method.
302 * Deprecated the BN_is_prime_ex() and BN_is_prime_fasttest_ex() functions.
306 * Deprecated EVP_MD_CTX_set_update_fn() and EVP_MD_CTX_update_fn().
310 * Deprecated the type OCSP_REQ_CTX and the functions OCSP_REQ_CTX_*() and
311 replaced with OSSL_HTTP_REQ_CTX and the functions OSSL_HTTP_REQ_CTX_*().
313 *Rich Salz, Richard Levitte, and David von Oheimb*
315 * Deprecated `X509_http_nbio()` and `X509_CRL_http_nbio()`.
319 * Deprecated `OCSP_parse_url()`.
323 * Validation of SM2 keys has been separated from the validation of regular EC
328 * Behavior of the `pkey` app is changed, when using the `-check` or `-pubcheck`
329 switches: a validation failure triggers an early exit, returning a failure
330 exit status to the parent process.
334 * Changed behavior of SSL_CTX_set_ciphersuites() and SSL_set_ciphersuites()
335 to ignore unknown ciphers.
339 * The `-cipher-commands` and `-digest-commands` options
340 of the command line utility `list` have been deprecated.
341 Instead use the `-cipher-algorithms` and `-digest-algorithms` options.
345 * Added convenience functions for generating asymmetric key pairs:
346 The 'quick' one-shot (yet somewhat limited) function L<EVP_PKEY_Q_keygen(3)>
347 and macros for the most common cases: <EVP_RSA_gen(3)> and L<EVP_EC_gen(3)>.
351 * All of the low level EC_KEY functions have been deprecated.
353 *Shane Lontis, Paul Dale, Richard Levitte, and Tomáš Mráz*
355 * Deprecated all the libcrypto and libssl error string loading
360 * The functions SSL_CTX_set_tmp_dh_callback and SSL_set_tmp_dh_callback, as
361 well as the macros SSL_CTX_set_tmp_dh() and SSL_set_tmp_dh() have been
366 * The `-crypt` option to the `passwd` command line tool has been removed.
370 * The -C option to the `x509`, `dhparam`, `dsaparam`, and `ecparam` commands
375 * Add support for AES Key Wrap inverse ciphers to the EVP layer.
379 * Deprecated EVP_PKEY_set1_tls_encodedpoint() and
380 EVP_PKEY_get1_tls_encodedpoint().
384 * The security callback, which can be customised by application code, supports
385 the security operation SSL_SECOP_TMP_DH. One location of the "other" parameter
386 was incorrectly passing a DH object. It now passed an EVP_PKEY in all cases.
390 * Add PKCS7_get_octet_string() and PKCS7_type_is_other() to the public
391 interface. Their functionality remains unchanged.
395 * Added new option for 'openssl list', '-providers', which will display the
396 list of loaded providers, their names, version and status. It optionally
397 displays their gettable parameters.
401 * Removed EVP_PKEY_set_alias_type().
405 * Deprecated `EVP_PKEY_CTX_set_rsa_keygen_pubexp()` and introduced
406 `EVP_PKEY_CTX_set1_rsa_keygen_pubexp()`, which is now preferred.
410 * Changed all "STACK" functions to be macros instead of inline functions. Macro
411 parameters are still checked for type safety at compile time via helper
416 * Remove the RAND_DRBG API
418 *Paul Dale and Matthias St. Pierre*
420 * Allow `SSL_set1_host()` and `SSL_add1_host()` to take IP literal addresses
421 as well as actual hostnames.
425 * The 'MinProtocol' and 'MaxProtocol' configuration commands now silently
426 ignore TLS protocol version bounds when configuring DTLS-based contexts, and
427 conversely, silently ignore DTLS protocol version bounds when configuring
428 TLS-based contexts. The commands can be repeated to set bounds of both
429 types. The same applies with the corresponding "min_protocol" and
430 "max_protocol" command-line switches, in case some application uses both TLS
433 SSL_CTX instances that are created for a fixed protocol version (e.g.
434 `TLSv1_server_method()`) also silently ignore version bounds. Previously
435 attempts to apply bounds to these protocol versions would result in an
436 error. Now only the "version-flexible" SSL_CTX instances are subject to
437 limits in configuration files in command-line options.
441 * Deprecated the `ENGINE` API. Engines should be replaced with providers
446 * Reworked the recorded ERR codes to make better space for system errors.
447 To distinguish them, the macro `ERR_SYSTEM_ERROR()` indicates if the
448 given code is a system error (true) or an OpenSSL error (false).
452 * Reworked the test perl framework to better allow parallel testing.
454 *Nicola Tuveri and David von Oheimb*
456 * Added ciphertext stealing algorithms AES-128-CBC-CTS, AES-192-CBC-CTS and
457 AES-256-CBC-CTS to the providers. CS1, CS2 and CS3 variants are supported.
461 * 'Configure' has been changed to figure out the configuration target if
462 none is given on the command line. Consequently, the 'config' script is
463 now only a mere wrapper. All documentation is changed to only mention
466 *Rich Salz and Richard Levitte*
468 * Added a library context `OSSL_LIB_CTX` that applications as well as
469 other libraries can use to form a separate context within which
470 libcrypto operations are performed.
474 * Handshake now fails if Extended Master Secret extension is dropped
479 * Dropped interactive mode from the `openssl` program.
483 * Deprecated `EVP_PKEY_cmp()` and `EVP_PKEY_cmp_parameters()`.
485 *David von Oheimb and Shane Lontis*
487 * Deprecated `EC_METHOD_get_field_type()`.
491 * Deprecated EC_GFp_simple_method(), EC_GFp_mont_method(),
492 EC_GF2m_simple_method(), EC_GFp_nist_method(), EC_GFp_nistp224_method()
493 EC_GFp_nistp256_method(), and EC_GFp_nistp521_method().
497 * Deprecated EC_GROUP_new(), EC_GROUP_method_of(), and EC_POINT_method_of().
501 * Add CAdES-BES signature verification support, mostly derived
502 from ESSCertIDv2 TS (RFC 5816) contribution by Marek Klein.
504 *Filipe Raimundo da Silva*
506 * Add CAdES-BES signature scheme and attributes support (RFC 5126) to CMS API.
510 * Added the AuthEnvelopedData content type structure (RFC 5083) with AES-GCM
511 parameter (RFC 5084) for the Cryptographic Message Syntax (CMS).
515 * Deprecated EC_POINT_make_affine() and EC_POINTs_make_affine().
519 * Deprecated EC_GROUP_precompute_mult(), EC_GROUP_have_precompute_mult(), and
520 EC_KEY_precompute_mult().
524 * Deprecated EC_POINTs_mul().
528 * Removed FIPS_mode() and FIPS_mode_set().
532 * The SSL option SSL_OP_IGNORE_UNEXPECTED_EOF is introduced.
536 * Deprecated EC_POINT_set_Jprojective_coordinates_GFp() and
537 EC_POINT_get_Jprojective_coordinates_GFp().
541 * Added OSSL_PARAM_BLD to the public interface. This allows OSSL_PARAM
542 arrays to be more easily constructed via a series of utility functions.
543 Create a parameter builder using OSSL_PARAM_BLD_new(), add parameters using
544 the various push functions and finally convert to a passable OSSL_PARAM
545 array using OSSL_PARAM_BLD_to_param().
549 * The security strength of SHA1 and MD5 based signatures in TLS has been
554 * Added EVP_PKEY_set_type_by_keymgmt(), to initialise an EVP_PKEY to
555 contain a provider side internal key.
559 * ASN1_verify(), ASN1_digest() and ASN1_sign() have been deprecated.
563 * Project text documents not yet having a proper file name extension
564 (`HACKING`, `LICENSE`, `NOTES*`, `README*`, `VERSION`) have been renamed to
565 `*.md` as far as reasonable, else `*.txt`, for better use with file managers.
569 * The main project documents (README, NEWS, CHANGES, INSTALL, SUPPORT)
570 have been converted to Markdown with the goal to produce documents
571 which not only look pretty when viewed online in the browser, but
572 remain well readable inside a plain text editor.
574 To achieve this goal, a 'minimalistic' Markdown style has been applied
575 which avoids formatting elements that interfere too much with the
576 reading flow in the text file. For example, it
578 * avoids [ATX headings][] and uses [setext headings][] instead
579 (which works for `<h1>` and `<h2>` headings only).
580 * avoids [inline links][] and uses [reference links][] instead.
581 * avoids [fenced code blocks][] and uses [indented code blocks][] instead.
583 [ATX headings]: https://github.github.com/gfm/#atx-headings
584 [setext headings]: https://github.github.com/gfm/#setext-headings
585 [inline links]: https://github.github.com/gfm/#inline-link
586 [reference links]: https://github.github.com/gfm/#reference-link
587 [fenced code blocks]: https://github.github.com/gfm/#fenced-code-blocks
588 [indented code blocks]: https://github.github.com/gfm/#indented-code-blocks
590 *Matthias St. Pierre*
592 * The test suite is changed to preserve results of each test recipe.
593 A new directory test-runs/ with subdirectories named like the
594 test recipes are created in the build tree for this purpose.
598 * Added an implementation of CMP and CRMF (RFC 4210, RFC 4211 RFC 6712).
599 This adds `crypto/cmp/`, `crpyto/crmf/`, `apps/cmp.c`, and `test/cmp_*`.
600 See L<openssl-cmp(1)> and L<OSSL_CMP_exec_IR_ses(3)> as starting points.
602 *David von Oheimb, Martin Peylo*
604 * Generalized the HTTP client code from `crypto/ocsp/` into `crpyto/http/`.
605 It supports arbitrary request and response content types, GET redirection,
606 TLS, connections via HTTP(S) proxies, connections and exchange via
607 user-defined BIOs (allowing implicit connections), persistent connections,
608 and timeout checks. See L<OSSL_HTTP_transfer(3)> etc. for details.
609 The legacy OCSP-focused (and only partly documented) API
610 is retained for backward compatibility, while most of it is deprecated.
614 * Added `util/check-format.pl`, a tool for checking adherence to the
615 OpenSSL coding style <https://www.openssl.org/policies/codingstyle.html>.
616 The checks performed are incomplete and yield some false positives.
617 Still the tool should be useful for detecting most typical glitches.
621 * `BIO_do_connect()` and `BIO_do_handshake()` have been extended:
622 If domain name resolution yields multiple IP addresses all of them are tried
623 after `connect()` failures.
627 * All of the low level RSA functions have been deprecated.
631 * X509 certificates signed using SHA1 are no longer allowed at security
636 * The command line utilities dhparam, dsa, gendsa and dsaparam have been
637 modified to use PKEY APIs. These commands are now in maintenance mode
638 and no new features will be added to them.
642 * The command line utility rsautl has been deprecated.
646 * The command line utilities genrsa and rsa have been modified to use PKEY
647 APIs. They now write PKCS#8 keys by default. These commands are now in
648 maintenance mode and no new features will be added to them.
652 * All of the low level DH functions have been deprecated.
654 *Paul Dale and Matt Caswell*
656 * All of the low level DSA functions have been deprecated.
660 * Reworked the treatment of EC EVP_PKEYs with the SM2 curve to
661 automatically become EVP_PKEY_SM2 rather than EVP_PKEY_EC.
665 * Deprecated low level ECDH and ECDSA functions.
669 * Deprecated EVP_PKEY_decrypt_old() and EVP_PKEY_encrypt_old().
673 * Enhanced the documentation of EVP_PKEY_get_size(), EVP_PKEY_get_bits()
674 and EVP_PKEY_get_security_bits(). Especially EVP_PKEY_get_size() needed
675 a new formulation to include all the things it can be used for,
676 as well as words of caution.
680 * The SSL_CTX_set_tlsext_ticket_key_cb(3) function has been deprecated.
684 * All of the low level HMAC functions have been deprecated.
686 *Paul Dale and David von Oheimb*
688 * Over two thousand fixes were made to the documentation, including:
689 - Common options (such as -rand/-writerand, TLS version control, etc)
690 were refactored and point to newly-enhanced descriptions in openssl.pod.
691 - Added style conformance for all options (with help from Richard Levitte),
692 documented all reported missing options, added a CI build to check
693 that all options are documented and that no unimplemented options
695 - Documented some internals, such as all use of environment variables.
696 - Addressed all internal broken L<> references.
700 * All of the low level CMAC functions have been deprecated.
704 * The low-level MD2, MD4, MD5, MDC2, RIPEMD160 and Whirlpool digest
705 functions have been deprecated.
707 *Paul Dale and David von Oheimb*
709 * Corrected the documentation of the return values from the `EVP_DigestSign*`
710 set of functions. The documentation mentioned negative values for some
711 errors, but this was never the case, so the mention of negative values
714 Code that followed the documentation and thereby check with something
715 like `EVP_DigestSignInit(...) <= 0` will continue to work undisturbed.
719 * All of the low level cipher functions have been deprecated.
721 *Matt Caswell and Paul Dale*
723 * Removed include/openssl/opensslconf.h.in and replaced it with
724 include/openssl/configuration.h.in, which differs in not including
725 <openssl/macros.h>. A short header include/openssl/opensslconf.h
726 was added to include both.
728 This allows internal hacks where one might need to modify the set
729 of configured macros, for example this if deprecated symbols are
730 still supposed to be available internally:
732 #include <openssl/configuration.h>
734 #undef OPENSSL_NO_DEPRECATED
735 #define OPENSSL_SUPPRESS_DEPRECATED
737 #include <openssl/macros.h>
739 This should not be used by applications that use the exported
740 symbols, as that will lead to linking errors.
744 * Fixed an overflow bug in the x64_64 Montgomery squaring procedure
745 used in exponentiation with 512-bit moduli. No EC algorithms are
746 affected. Analysis suggests that attacks against 2-prime RSA1024,
747 3-prime RSA1536, and DSA1024 as a result of this defect would be very
748 difficult to perform and are not believed likely. Attacks against DH512
749 are considered just feasible. However, for an attack the target would
750 have to re-use the DH512 private key, which is not recommended anyway.
751 Also applications directly using the low-level API BN_mod_exp may be
752 affected if they use BN_FLG_CONSTTIME.
757 * Most memory-debug features have been deprecated, and the functionality
758 replaced with no-ops.
762 * Added documentation for the STACK API.
766 * Introduced a new method type and API, OSSL_ENCODER, to
767 represent generic encoders.
771 * Added a .pragma directive to the syntax of configuration files, to
772 allow varying behavior in a supported and predictable manner.
773 Currently added pragma:
777 This allows dollar signs to be a keyword character unless it's
778 followed by a opening brace or parenthesis. This is useful for
779 platforms where dollar signs are commonly used in names, such as
780 volume names and system directory names on VMS.
784 * Added functionality to create an EVP_PKEY from user data.
788 * Change the interpretation of the '--api' configuration option to
789 mean that this is a desired API compatibility level with no
790 further meaning. The previous interpretation, that this would
791 also mean to remove all deprecated symbols up to and including
792 the given version, no requires that 'no-deprecated' is also used
793 in the configuration.
795 When building applications, the desired API compatibility level
796 can be set with the OPENSSL_API_COMPAT macro like before. For
797 API compatibility version below 3.0, the old style numerical
798 value is valid as before, such as -DOPENSSL_API_COMPAT=0x10100000L.
799 For version 3.0 and on, the value is expected to be the decimal
800 value calculated from the major and minor version like this:
802 MAJOR * 10000 + MINOR * 100
806 -DOPENSSL_API_COMPAT=30000 For 3.0
807 -DOPENSSL_API_COMPAT=30200 For 3.2
809 To hide declarations that are deprecated up to and including the
810 given API compatibility level, -DOPENSSL_NO_DEPRECATED must be
811 given when building the application as well.
815 * Added the X509_LOOKUP_METHOD called X509_LOOKUP_store, to allow
816 access to certificate and CRL stores via URIs and OSSL_STORE
819 This adds the following functions:
821 - X509_LOOKUP_store()
822 - X509_STORE_load_file()
823 - X509_STORE_load_path()
824 - X509_STORE_load_store()
825 - SSL_add_store_cert_subjects_to_stack()
826 - SSL_CTX_set_default_verify_store()
827 - SSL_CTX_load_verify_file()
828 - SSL_CTX_load_verify_dir()
829 - SSL_CTX_load_verify_store()
833 * Added a new method to gather entropy on VMS, based on SYS$GET_ENTROPY.
834 The presence of this system service is determined at run-time.
838 * Added functionality to create an EVP_PKEY context based on data
839 for methods from providers. This takes an algorithm name and a
840 property query string and simply stores them, with the intent
841 that any operation that uses this context will use those strings
842 to fetch the needed methods implicitly, thereby making the port
843 of application written for pre-3.0 OpenSSL easier.
847 * The undocumented function NCONF_WIN32() has been deprecated; for
848 conversion details see the HISTORY section of doc/man5/config.pod
852 * Introduced the new functions EVP_DigestSignInit_ex() and
853 EVP_DigestVerifyInit_ex(). The macros EVP_DigestSignUpdate() and
854 EVP_DigestVerifyUpdate() have been converted to functions. See the man
855 pages for further details.
859 * Over two thousand fixes were made to the documentation, including:
860 adding missing command flags, better style conformance, documentation
863 *Rich Salz, Richard Levitte*
865 * s390x assembly pack: add hardware-support for P-256, P-384, P-521,
866 X25519, X448, Ed25519 and Ed448.
870 * Print all values for a PKCS#12 attribute with 'openssl pkcs12', not just
875 * Deprecated the public definition of `ERR_STATE` as well as the function
876 `ERR_get_state()`. This is done in preparation of making `ERR_STATE` an
881 * Added ERR functionality to give callers access to the stored function
882 names that have replaced the older function code based functions.
884 New functions are ERR_peek_error_func(), ERR_peek_last_error_func(),
885 ERR_peek_error_data(), ERR_peek_last_error_data(), ERR_get_error_all(),
886 ERR_peek_error_all() and ERR_peek_last_error_all().
888 Deprecate ERR functions ERR_get_error_line(), ERR_get_error_line_data(),
889 ERR_peek_error_line_data(), ERR_peek_last_error_line_data() and
890 ERR_func_error_string().
894 * Extended testing to be verbose for failing tests only. The make variables
895 VERBOSE_FAILURE or VF can be used to enable this:
897 $ make VF=1 test # Unix
898 $ mms /macro=(VF=1) test ! OpenVMS
899 $ nmake VF=1 test # Windows
903 * Added the `-copy_extensions` option to the `x509` command for use with
904 `-req` and `-x509toreq`. When given with the `copy` or `copyall` argument,
905 all extensions in the request are copied to the certificate or vice versa.
907 *David von Oheimb*, *Kirill Stefanenkov <kirill_stefanenkov@rambler.ru>*
909 * Added the `-copy_extensions` option to the `req` command for use with
910 `-x509`. When given with the `copy` or `copyall` argument,
911 all extensions in the certification request are copied to the certificate.
915 * The `x509`, `req`, and `ca` commands now make sure that X.509v3 certificates
916 they generate are by default RFC 5280 compliant in the following sense:
917 There is a subjectKeyIdentifier extension with a hash value of the public key
918 and for not self-signed certs there is an authorityKeyIdentifier extension
919 with a keyIdentifier field or issuer information identifying the signing key.
920 This is done unless some configuration overrides the new default behavior,
921 such as `subjectKeyIdentifier = none` and `authorityKeyIdentifier = none`.
925 * Added several checks to `X509_verify_cert()` according to requirements in
926 RFC 5280 in case `X509_V_FLAG_X509_STRICT` is set
927 (which may be done by using the CLI option `-x509_strict`):
928 * The basicConstraints of CA certificates must be marked critical.
929 * CA certificates must explicitly include the keyUsage extension.
930 * If a pathlenConstraint is given the key usage keyCertSign must be allowed.
931 * The issuer name of any certificate must not be empty.
932 * The subject name of CA certs, certs with keyUsage crlSign,
933 and certs without subjectAlternativeName must not be empty.
934 * If a subjectAlternativeName extension is given it must not be empty.
935 * The signatureAlgorithm field and the cert signature must be consistent.
936 * Any given authorityKeyIdentifier and any given subjectKeyIdentifier
937 must not be marked critical.
938 * The authorityKeyIdentifier must be given for X.509v3 certs
939 unless they are self-signed.
940 * The subjectKeyIdentifier must be given for all X.509v3 CA certs.
944 * Certificate verification using `X509_verify_cert()` meanwhile rejects EC keys
945 with explicit curve parameters (specifiedCurve) as required by RFC 5480.
949 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
950 used even when parsing explicit parameters, when loading a encoded key
951 or calling `EC_GROUP_new_from_ecpkparameters()`/
952 `EC_GROUP_new_from_ecparameters()`.
953 This prevents bypass of security hardening and performance gains,
954 especially for curves with specialized EC_METHODs.
955 By default, if a key encoded with explicit parameters is loaded and later
956 encoded, the output is still encoded with explicit parameters, even if
957 internally a "named" EC_GROUP is used for computation.
961 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
962 this change, EC_GROUP_set_generator would accept order and/or cofactor as
963 NULL. After this change, only the cofactor parameter can be NULL. It also
964 does some minimal sanity checks on the passed order.
969 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
970 An attack is simple, if the first CMS_recipientInfo is valid but the
971 second CMS_recipientInfo is chosen ciphertext. If the second
972 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
973 encryption key will be replaced by garbage, and the message cannot be
974 decoded, but if the RSA decryption fails, the correct encryption key is
975 used and the recipient will not notice the attack.
976 As a work around for this potential attack the length of the decrypted
977 key must be equal to the cipher default key length, in case the
978 certifiate is not given and all recipientInfo are tried out.
979 The old behaviour can be re-enabled in the CMS code by setting the
980 CMS_DEBUG_DECRYPT flag.
984 * Early start up entropy quality from the DEVRANDOM seed source has been
985 improved for older Linux systems. The RAND subsystem will wait for
986 /dev/random to be producing output before seeding from /dev/urandom.
987 The seeded state is stored for future library initialisations using
988 a system global shared memory segment. The shared memory identifier
989 can be configured by defining OPENSSL_RAND_SEED_DEVRANDOM_SHM_ID to
990 the desired value. The default identifier is 114.
994 * Revised BN_generate_prime_ex to not avoid factors 2..17863 in p-1
995 when primes for RSA keys are computed.
996 Since we previously always generated primes == 2 (mod 3) for RSA keys,
997 the 2-prime and 3-prime RSA modules were easy to distinguish, since
998 `N = p*q = 1 (mod 3)`, but `N = p*q*r = 2 (mod 3)`. Therefore fingerprinting
999 2-prime vs. 3-prime RSA keys was possible by computing N mod 3.
1000 This avoids possible fingerprinting of newly generated RSA modules.
1004 * Correct the extended master secret constant on EBCDIC systems. Without this
1005 fix TLS connections between an EBCDIC system and a non-EBCDIC system that
1006 negotiate EMS will fail. Unfortunately this also means that TLS connections
1007 between EBCDIC systems with this fix, and EBCDIC systems without this
1008 fix will fail if they negotiate EMS.
1012 * Changed the library initialisation so that the config file is now loaded
1013 by default. This was already the case for libssl. It now occurs for both
1014 libcrypto and libssl. Use the OPENSSL_INIT_NO_LOAD_CONFIG option to
1015 `OPENSSL_init_crypto()` to suppress automatic loading of a config file.
1019 * Introduced new error raising macros, `ERR_raise()` and `ERR_raise_data()`,
1020 where the former acts as a replacement for `ERR_put_error()`, and the
1021 latter replaces the combination `ERR_put_error()` + `ERR_add_error_data()`.
1022 `ERR_raise_data()` adds more flexibility by taking a format string and
1023 an arbitrary number of arguments following it, to be processed with
1028 * Introduced a new function, `OSSL_PROVIDER_available()`, which can be used
1029 to check if a named provider is loaded and available. When called, it
1030 will also activate all fallback providers if such are still present.
1034 * Enforce a minimum DH modulus size of 512 bits.
1038 * Changed DH parameters to generate the order q subgroup instead of 2q.
1039 Previously generated DH parameters are still accepted by DH_check
1040 but DH_generate_key works around that by clearing bit 0 of the
1041 private key for those. This avoids leaking bit 0 of the private key.
1045 * Significantly reduce secure memory usage by the randomness pools.
1049 * `{CRYPTO,OPENSSL}_mem_debug_{push,pop}` are now no-ops and have been
1054 * A new type, EVP_KEYEXCH, has been introduced to represent key exchange
1055 algorithms. An implementation of a key exchange algorithm can be obtained
1056 by using the function EVP_KEYEXCH_fetch(). An EVP_KEYEXCH algorithm can be
1057 used in a call to EVP_PKEY_derive_init_ex() which works in a similar way to
1058 the older EVP_PKEY_derive_init() function. See the man pages for the new
1059 functions for further details.
1063 * The EVP_PKEY_CTX_set_dh_pad() macro has now been converted to a function.
1067 * Removed the function names from error messages and deprecated the
1072 * Removed NextStep support and the macro OPENSSL_UNISTD
1076 * Removed DES_check_key. Also removed OPENSSL_IMPLEMENT_GLOBAL,
1077 OPENSSL_GLOBAL_REF, OPENSSL_DECLARE_GLOBAL.
1078 Also removed "export var as function" capability; we do not export
1079 variables, only functions.
1083 * RC5_32_set_key has been changed to return an int type, with 0 indicating
1084 an error and 1 indicating success. In previous versions of OpenSSL this
1085 was a void type. If a key was set longer than the maximum possible this
1090 * Support SM2 signing and verification schemes with X509 certificate.
1094 * Use SHA256 as the default digest for TS query in the `ts` app.
1098 * Change PBKDF2 to conform to SP800-132 instead of the older PKCS5 RFC2898.
1102 * Default cipher lists/suites are now available via a function, the
1103 #defines are deprecated.
1107 * Add target VC-WIN32-UWP, VC-WIN64A-UWP, VC-WIN32-ARM-UWP and
1108 VC-WIN64-ARM-UWP in Windows OneCore target for making building libraries
1109 for Windows Store apps easier. Also, the "no-uplink" option has been added.
1113 * Join the directories crypto/x509 and crypto/x509v3
1117 * Added command 'openssl kdf' that uses the EVP_KDF API.
1121 * Added command 'openssl mac' that uses the EVP_MAC API.
1125 * Added OPENSSL_info() to get diverse built-in OpenSSL data, such
1126 as default directories. Also added the command 'openssl info'
1127 for scripting purposes.
1131 * The functions AES_ige_encrypt() and AES_bi_ige_encrypt() have been
1136 * Add prediction resistance to the DRBG reseeding process.
1140 * Limit the number of blocks in a data unit for AES-XTS to 2^20 as
1141 mandated by IEEE Std 1619-2018.
1145 * Added newline escaping functionality to a filename when using openssl dgst.
1146 This output format is to replicate the output format found in the `*sum`
1147 checksum programs. This aims to preserve backward compatibility.
1149 *Matt Eaton, Richard Levitte, and Paul Dale*
1151 * Removed the heartbeat message in DTLS feature, as it has very
1152 little usage and doesn't seem to fulfill a valuable purpose.
1153 The configuration option is now deprecated.
1157 * Changed the output of 'openssl {digestname} < file' to display the
1158 digest name in its output.
1162 * Added a new generic trace API which provides support for enabling
1163 instrumentation through trace output.
1165 *Richard Levitte & Matthias St. Pierre*
1167 * Added build tests for C++. These are generated files that only do one
1168 thing, to include one public OpenSSL head file each. This tests that
1169 the public header files can be usefully included in a C++ application.
1171 This test isn't enabled by default. It can be enabled with the option
1172 'enable-buildtest-c++'.
1176 * Add Single Step KDF (EVP_KDF_SS) to EVP_KDF.
1180 * Add KMAC to EVP_MAC.
1184 * Added property based algorithm implementation selection framework to
1189 * Added SCA hardening for modular field inversion in EC_GROUP through
1190 a new dedicated field_inv() pointer in EC_METHOD.
1191 This also addresses a leakage affecting conversions from projective
1192 to affine coordinates.
1194 *Billy Bob Brumley, Nicola Tuveri*
1196 * Added EVP_KDF, an EVP layer KDF API, to simplify adding KDF and PRF
1197 implementations. This includes an EVP_PKEY to EVP_KDF bridge for
1198 those algorithms that were already supported through the EVP_PKEY API
1199 (scrypt, TLS1 PRF and HKDF). The low-level KDF functions for PBKDF2
1200 and scrypt are now wrappers that call EVP_KDF.
1204 * Build devcrypto engine as a dynamic engine.
1206 *Eneas U de Queiroz*
1208 * Add keyed BLAKE2 to EVP_MAC.
1212 * Fix a bug in the computation of the endpoint-pair shared secret used
1213 by DTLS over SCTP. This breaks interoperability with older versions
1214 of OpenSSL like OpenSSL 1.1.0 and OpenSSL 1.0.2. There is a runtime
1215 switch SSL_MODE_DTLS_SCTP_LABEL_LENGTH_BUG (off by default) enabling
1216 interoperability with such broken implementations. However, enabling
1217 this switch breaks interoperability with correct implementations.
1219 * Fix a use after free bug in d2i_X509_PUBKEY when overwriting a
1220 re-used X509_PUBKEY object if the second PUBKEY is malformed.
1224 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
1228 * Changed the license to the Apache License v2.0.
1232 * Switch to a new version scheme using three numbers MAJOR.MINOR.PATCH.
1234 - Major releases (indicated by incrementing the MAJOR release number)
1235 may introduce incompatible API/ABI changes.
1236 - Minor releases (indicated by incrementing the MINOR release number)
1237 may introduce new features but retain API/ABI compatibility.
1238 - Patch releases (indicated by incrementing the PATCH number)
1239 are intended for bug fixes and other improvements of existing
1240 features only (like improving performance or adding documentation)
1241 and retain API/ABI compatibility.
1245 * Add support for RFC5297 SIV mode (siv128), including AES-SIV.
1249 * Remove the 'dist' target and add a tarball building script. The
1250 'dist' target has fallen out of use, and it shouldn't be
1251 necessary to configure just to create a source distribution.
1255 * Recreate the OS390-Unix config target. It no longer relies on a
1256 special script like it did for OpenSSL pre-1.1.0.
1260 * Instead of having the source directories listed in Configure, add
1261 a 'build.info' keyword SUBDIRS to indicate what sub-directories to
1266 * Add GMAC to EVP_MAC.
1270 * Ported the HMAC, CMAC and SipHash EVP_PKEY_METHODs to EVP_MAC.
1274 * Added EVP_MAC, an EVP layer MAC API, to simplify adding MAC
1275 implementations. This includes a generic EVP_PKEY to EVP_MAC bridge,
1276 to facilitate the continued use of MACs through raw private keys in
1277 functionality such as `EVP_DigestSign*` and `EVP_DigestVerify*`.
1281 * Deprecate ECDH_KDF_X9_62().
1285 * Added EVP_PKEY_ECDH_KDF_X9_63 and ecdh_KDF_X9_63() as replacements for
1286 the EVP_PKEY_ECDH_KDF_X9_62 KDF type and ECDH_KDF_X9_62(). The old names
1287 are retained for backwards compatibility.
1291 * AES-XTS mode now enforces that its two keys are different to mitigate
1292 the attacked described in "Efficient Instantiations of Tweakable
1293 Blockciphers and Refinements to Modes OCB and PMAC" by Phillip Rogaway.
1294 Details of this attack can be obtained from:
1295 <http://web.cs.ucdavis.edu/%7Erogaway/papers/offsets.pdf>
1299 * Rename the object files, i.e. give them other names than in previous
1300 versions. Their names now include the name of the final product, as
1301 well as its type mnemonic (bin, lib, shlib).
1305 * Added new option for 'openssl list', '-objects', which will display the
1306 list of built in objects, i.e. OIDs with names.
1310 * Added the options `-crl_lastupdate` and `-crl_nextupdate` to `openssl ca`,
1311 allowing the `lastUpdate` and `nextUpdate` fields in the generated CRL to
1316 * Added support for Linux Kernel TLS data-path. The Linux Kernel data-path
1317 improves application performance by removing data copies and providing
1318 applications with zero-copy system calls such as sendfile and splice.
1322 * The SSL option SSL_OP_CLEANSE_PLAINTEXT is introduced.
1326 * `PKCS12_parse` now maintains the order of the parsed certificates
1327 when outputting them via `*ca` (rather than reversing it).
1331 * Deprecated pthread fork support methods.
1335 * Added support for FFDHE key exchange in TLS 1.3.
1342 ### Changes between 1.1.1j and 1.1.1k [xx XXX xxxx]
1344 * Fixed a problem with verifying a certificate chain when using the
1345 X509_V_FLAG_X509_STRICT flag. This flag enables additional security checks of
1346 the certificates present in a certificate chain. It is not set by default.
1348 Starting from OpenSSL version 1.1.1h a check to disallow certificates in
1349 the chain that have explicitly encoded elliptic curve parameters was added
1350 as an additional strict check.
1352 An error in the implementation of this check meant that the result of a
1353 previous check to confirm that certificates in the chain are valid CA
1354 certificates was overwritten. This effectively bypasses the check
1355 that non-CA certificates must not be able to issue other certificates.
1357 If a "purpose" has been configured then there is a subsequent opportunity
1358 for checks that the certificate is a valid CA. All of the named "purpose"
1359 values implemented in libcrypto perform this check. Therefore, where
1360 a purpose is set the certificate chain will still be rejected even when the
1361 strict flag has been used. A purpose is set by default in libssl client and
1362 server certificate verification routines, but it can be overridden or
1363 removed by an application.
1365 In order to be affected, an application must explicitly set the
1366 X509_V_FLAG_X509_STRICT verification flag and either not set a purpose
1367 for the certificate verification or, in the case of TLS client or server
1368 applications, override the default purpose.
1373 * Fixed an issue where an OpenSSL TLS server may crash if sent a maliciously
1374 crafted renegotiation ClientHello message from a client. If a TLSv1.2
1375 renegotiation ClientHello omits the signature_algorithms extension (where it
1376 was present in the initial ClientHello), but includes a
1377 signature_algorithms_cert extension then a NULL pointer dereference will
1378 result, leading to a crash and a denial of service attack.
1380 A server is only vulnerable if it has TLSv1.2 and renegotiation enabled
1381 (which is the default configuration). OpenSSL TLS clients are not impacted by
1385 *Peter Kästle and Samuel Sapalski*
1387 ### Changes between 1.1.1i and 1.1.1j [16 Feb 2021]
1389 * Fixed the X509_issuer_and_serial_hash() function. It attempts to
1390 create a unique hash value based on the issuer and serial number data
1391 contained within an X509 certificate. However it was failing to correctly
1392 handle any errors that may occur while parsing the issuer field (which might
1393 occur if the issuer field is maliciously constructed). This may subsequently
1394 result in a NULL pointer deref and a crash leading to a potential denial of
1400 * Fixed the RSA_padding_check_SSLv23() function and the RSA_SSLV23_PADDING
1401 padding mode to correctly check for rollback attacks. This is considered a
1402 bug in OpenSSL 1.1.1 because it does not support SSLv2. In 1.0.2 this is
1407 Fixed the EVP_CipherUpdate, EVP_EncryptUpdate and EVP_DecryptUpdate
1408 functions. Previously they could overflow the output length argument in some
1409 cases where the input length is close to the maximum permissable length for
1410 an integer on the platform. In such cases the return value from the function
1411 call would be 1 (indicating success), but the output length value would be
1412 negative. This could cause applications to behave incorrectly or crash.
1417 * Fixed SRP_Calc_client_key so that it runs in constant time. The previous
1418 implementation called BN_mod_exp without setting BN_FLG_CONSTTIME. This
1419 could be exploited in a side channel attack to recover the password. Since
1420 the attack is local host only this is outside of the current OpenSSL
1421 threat model and therefore no CVE is assigned.
1423 Thanks to Mohammed Sabt and Daniel De Almeida Braga for reporting this
1428 ### Changes between 1.1.1h and 1.1.1i [8 Dec 2020]
1430 * Fixed NULL pointer deref in the GENERAL_NAME_cmp function
1431 This function could crash if both GENERAL_NAMEs contain an EDIPARTYNAME.
1432 If an attacker can control both items being compared then this could lead
1433 to a possible denial of service attack. OpenSSL itself uses the
1434 GENERAL_NAME_cmp function for two purposes:
1435 1) Comparing CRL distribution point names between an available CRL and a
1436 CRL distribution point embedded in an X509 certificate
1437 2) When verifying that a timestamp response token signer matches the
1438 timestamp authority name (exposed via the API functions
1439 TS_RESP_verify_response and TS_RESP_verify_token)
1444 ### Changes between 1.1.1g and 1.1.1h [22 Sep 2020]
1446 * Certificates with explicit curve parameters are now disallowed in
1447 verification chains if the X509_V_FLAG_X509_STRICT flag is used.
1451 * The 'MinProtocol' and 'MaxProtocol' configuration commands now silently
1452 ignore TLS protocol version bounds when configuring DTLS-based contexts, and
1453 conversely, silently ignore DTLS protocol version bounds when configuring
1454 TLS-based contexts. The commands can be repeated to set bounds of both
1455 types. The same applies with the corresponding "min_protocol" and
1456 "max_protocol" command-line switches, in case some application uses both TLS
1459 SSL_CTX instances that are created for a fixed protocol version (e.g.
1460 TLSv1_server_method()) also silently ignore version bounds. Previously
1461 attempts to apply bounds to these protocol versions would result in an
1462 error. Now only the "version-flexible" SSL_CTX instances are subject to
1463 limits in configuration files in command-line options.
1467 * Handshake now fails if Extended Master Secret extension is dropped
1472 * The Oracle Developer Studio compiler will start reporting deprecated APIs
1474 ### Changes between 1.1.1f and 1.1.1g [21 Apr 2020]
1476 * Fixed segmentation fault in SSL_check_chain()
1477 Server or client applications that call the SSL_check_chain() function
1478 during or after a TLS 1.3 handshake may crash due to a NULL pointer
1479 dereference as a result of incorrect handling of the
1480 "signature_algorithms_cert" TLS extension. The crash occurs if an invalid
1481 or unrecognised signature algorithm is received from the peer. This could
1482 be exploited by a malicious peer in a Denial of Service attack.
1487 * Added AES consttime code for no-asm configurations
1488 an optional constant time support for AES was added
1489 when building openssl for no-asm.
1490 Enable with: ./config no-asm -DOPENSSL_AES_CONST_TIME
1491 Disable with: ./config no-asm -DOPENSSL_NO_AES_CONST_TIME
1492 At this time this feature is by default disabled.
1493 It will be enabled by default in 3.0.
1497 ### Changes between 1.1.1e and 1.1.1f [31 Mar 2020]
1499 * Revert the change of EOF detection while reading in libssl to avoid
1500 regressions in applications depending on the current way of reporting
1501 the EOF. As the existing method is not fully accurate the change to
1502 reporting the EOF via SSL_ERROR_SSL is kept on the current development
1503 branch and will be present in the 3.0 release.
1507 * Revised BN_generate_prime_ex to not avoid factors 3..17863 in p-1
1508 when primes for RSA keys are computed.
1509 Since we previously always generated primes == 2 (mod 3) for RSA keys,
1510 the 2-prime and 3-prime RSA modules were easy to distinguish, since
1511 N = p*q = 1 (mod 3), but N = p*q*r = 2 (mod 3). Therefore fingerprinting
1512 2-prime vs. 3-prime RSA keys was possible by computing N mod 3.
1513 This avoids possible fingerprinting of newly generated RSA modules.
1517 ### Changes between 1.1.1d and 1.1.1e [17 Mar 2020]
1519 * Properly detect EOF while reading in libssl. Previously if we hit an EOF
1520 while reading in libssl then we would report an error back to the
1521 application (SSL_ERROR_SYSCALL) but errno would be 0. We now add
1522 an error to the stack (which means we instead return SSL_ERROR_SSL) and
1523 therefore give a hint as to what went wrong.
1527 * Check that ed25519 and ed448 are allowed by the security level. Previously
1528 signature algorithms not using an MD were not being checked that they were
1529 allowed by the security level.
1533 * Fixed SSL_get_servername() behaviour. The behaviour of SSL_get_servername()
1534 was not quite right. The behaviour was not consistent between resumption
1535 and normal handshakes, and also not quite consistent with historical
1536 behaviour. The behaviour in various scenarios has been clarified and
1537 it has been updated to make it match historical behaviour as closely as
1542 * *[VMS only]* The header files that the VMS compilers include automatically,
1543 `__DECC_INCLUDE_PROLOGUE.H` and `__DECC_INCLUDE_EPILOGUE.H`, use pragmas
1544 that the C++ compiler doesn't understand. This is a shortcoming in the
1545 compiler, but can be worked around with `__cplusplus` guards.
1547 C++ applications that use OpenSSL libraries must be compiled using the
1548 qualifier `/NAMES=(AS_IS,SHORTENED)` to be able to use all the OpenSSL
1549 functions. Otherwise, only functions with symbols of less than 31
1550 characters can be used, as the linker will not be able to successfully
1551 resolve symbols with longer names.
1555 * Added a new method to gather entropy on VMS, based on SYS$GET_ENTROPY.
1556 The presence of this system service is determined at run-time.
1560 * Added newline escaping functionality to a filename when using openssl dgst.
1561 This output format is to replicate the output format found in the `*sum`
1562 checksum programs. This aims to preserve backward compatibility.
1564 *Matt Eaton, Richard Levitte, and Paul Dale*
1566 * Print all values for a PKCS#12 attribute with 'openssl pkcs12', not just
1571 ### Changes between 1.1.1c and 1.1.1d [10 Sep 2019]
1573 * Fixed a fork protection issue. OpenSSL 1.1.1 introduced a rewritten random
1574 number generator (RNG). This was intended to include protection in the
1575 event of a fork() system call in order to ensure that the parent and child
1576 processes did not share the same RNG state. However this protection was not
1577 being used in the default case.
1579 A partial mitigation for this issue is that the output from a high
1580 precision timer is mixed into the RNG state so the likelihood of a parent
1581 and child process sharing state is significantly reduced.
1583 If an application already calls OPENSSL_init_crypto() explicitly using
1584 OPENSSL_INIT_ATFORK then this problem does not occur at all.
1587 *Matthias St. Pierre*
1589 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
1590 used even when parsing explicit parameters, when loading a encoded key
1591 or calling `EC_GROUP_new_from_ecpkparameters()`/
1592 `EC_GROUP_new_from_ecparameters()`.
1593 This prevents bypass of security hardening and performance gains,
1594 especially for curves with specialized EC_METHODs.
1595 By default, if a key encoded with explicit parameters is loaded and later
1596 encoded, the output is still encoded with explicit parameters, even if
1597 internally a "named" EC_GROUP is used for computation.
1601 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
1602 this change, EC_GROUP_set_generator would accept order and/or cofactor as
1603 NULL. After this change, only the cofactor parameter can be NULL. It also
1604 does some minimal sanity checks on the passed order.
1609 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
1610 An attack is simple, if the first CMS_recipientInfo is valid but the
1611 second CMS_recipientInfo is chosen ciphertext. If the second
1612 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
1613 encryption key will be replaced by garbage, and the message cannot be
1614 decoded, but if the RSA decryption fails, the correct encryption key is
1615 used and the recipient will not notice the attack.
1616 As a work around for this potential attack the length of the decrypted
1617 key must be equal to the cipher default key length, in case the
1618 certifiate is not given and all recipientInfo are tried out.
1619 The old behaviour can be re-enabled in the CMS code by setting the
1620 CMS_DEBUG_DECRYPT flag.
1625 * Early start up entropy quality from the DEVRANDOM seed source has been
1626 improved for older Linux systems. The RAND subsystem will wait for
1627 /dev/random to be producing output before seeding from /dev/urandom.
1628 The seeded state is stored for future library initialisations using
1629 a system global shared memory segment. The shared memory identifier
1630 can be configured by defining OPENSSL_RAND_SEED_DEVRANDOM_SHM_ID to
1631 the desired value. The default identifier is 114.
1635 * Correct the extended master secret constant on EBCDIC systems. Without this
1636 fix TLS connections between an EBCDIC system and a non-EBCDIC system that
1637 negotiate EMS will fail. Unfortunately this also means that TLS connections
1638 between EBCDIC systems with this fix, and EBCDIC systems without this
1639 fix will fail if they negotiate EMS.
1643 * Use Windows installation paths in the mingw builds
1645 Mingw isn't a POSIX environment per se, which means that Windows
1646 paths should be used for installation.
1651 * Changed DH_check to accept parameters with order q and 2q subgroups.
1652 With order 2q subgroups the bit 0 of the private key is not secret
1653 but DH_generate_key works around that by clearing bit 0 of the
1654 private key for those. This avoids leaking bit 0 of the private key.
1658 * Significantly reduce secure memory usage by the randomness pools.
1662 * Revert the DEVRANDOM_WAIT feature for Linux systems
1664 The DEVRANDOM_WAIT feature added a select() call to wait for the
1665 /dev/random device to become readable before reading from the
1666 /dev/urandom device.
1668 It turned out that this change had negative side effects on
1669 performance which were not acceptable. After some discussion it
1670 was decided to revert this feature and leave it up to the OS
1671 resp. the platform maintainer to ensure a proper initialization
1672 during early boot time.
1674 *Matthias St. Pierre*
1676 ### Changes between 1.1.1b and 1.1.1c [28 May 2019]
1678 * Add build tests for C++. These are generated files that only do one
1679 thing, to include one public OpenSSL head file each. This tests that
1680 the public header files can be usefully included in a C++ application.
1682 This test isn't enabled by default. It can be enabled with the option
1683 'enable-buildtest-c++'.
1687 * Enable SHA3 pre-hashing for ECDSA and DSA.
1691 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
1692 This changes the size when using the `genpkey` command when no size is given.
1693 It fixes an omission in earlier changes that changed all RSA, DSA and DH
1694 generation commands to use 2048 bits by default.
1698 * Reorganize the manual pages to consistently have RETURN VALUES,
1699 EXAMPLES, SEE ALSO and HISTORY come in that order, and adjust
1700 util/fix-doc-nits accordingly.
1702 *Paul Yang, Joshua Lock*
1704 * Add the missing accessor EVP_PKEY_get0_engine()
1708 * Have commands like `s_client` and `s_server` output the signature scheme
1709 along with other cipher suite parameters when debugging.
1713 * Make OPENSSL_config() error agnostic again.
1717 * Do the error handling in RSA decryption constant time.
1721 * Prevent over long nonces in ChaCha20-Poly1305.
1723 ChaCha20-Poly1305 is an AEAD cipher, and requires a unique nonce input
1724 for every encryption operation. RFC 7539 specifies that the nonce value
1725 (IV) should be 96 bits (12 bytes). OpenSSL allows a variable nonce length
1726 and front pads the nonce with 0 bytes if it is less than 12
1727 bytes. However it also incorrectly allows a nonce to be set of up to 16
1728 bytes. In this case only the last 12 bytes are significant and any
1729 additional leading bytes are ignored.
1731 It is a requirement of using this cipher that nonce values are
1732 unique. Messages encrypted using a reused nonce value are susceptible to
1733 serious confidentiality and integrity attacks. If an application changes
1734 the default nonce length to be longer than 12 bytes and then makes a
1735 change to the leading bytes of the nonce expecting the new value to be a
1736 new unique nonce then such an application could inadvertently encrypt
1737 messages with a reused nonce.
1739 Additionally the ignored bytes in a long nonce are not covered by the
1740 integrity guarantee of this cipher. Any application that relies on the
1741 integrity of these ignored leading bytes of a long nonce may be further
1742 affected. Any OpenSSL internal use of this cipher, including in SSL/TLS,
1743 is safe because no such use sets such a long nonce value. However user
1744 applications that use this cipher directly and set a non-default nonce
1745 length to be longer than 12 bytes may be vulnerable.
1747 This issue was reported to OpenSSL on 16th of March 2019 by Joran Dirk
1753 * Add DEVRANDOM_WAIT feature for Linux systems
1755 On older Linux systems where the getrandom() system call is not available,
1756 OpenSSL normally uses the /dev/urandom device for seeding its CSPRNG.
1757 Contrary to getrandom(), the /dev/urandom device will not block during
1758 early boot when the kernel CSPRNG has not been seeded yet.
1760 To mitigate this known weakness, use select() to wait for /dev/random to
1761 become readable before reading from /dev/urandom.
1763 * Ensure that SM2 only uses SM3 as digest algorithm
1767 ### Changes between 1.1.1a and 1.1.1b [26 Feb 2019]
1769 * Change the info callback signals for the start and end of a post-handshake
1770 message exchange in TLSv1.3. In 1.1.1/1.1.1a we used SSL_CB_HANDSHAKE_START
1771 and SSL_CB_HANDSHAKE_DONE. Experience has shown that many applications get
1772 confused by this and assume that a TLSv1.2 renegotiation has started. This
1773 can break KeyUpdate handling. Instead we no longer signal the start and end
1774 of a post handshake message exchange (although the messages themselves are
1775 still signalled). This could break some applications that were expecting
1776 the old signals. However without this KeyUpdate is not usable for many
1781 ### Changes between 1.1.1 and 1.1.1a [20 Nov 2018]
1783 * Timing vulnerability in DSA signature generation
1785 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
1786 timing side channel attack. An attacker could use variations in the signing
1787 algorithm to recover the private key.
1789 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
1794 * Timing vulnerability in ECDSA signature generation
1796 The OpenSSL ECDSA signature algorithm has been shown to be vulnerable to a
1797 timing side channel attack. An attacker could use variations in the signing
1798 algorithm to recover the private key.
1800 This issue was reported to OpenSSL on 25th October 2018 by Samuel Weiser.
1805 * Fixed the issue that RAND_add()/RAND_seed() silently discards random input
1806 if its length exceeds 4096 bytes. The limit has been raised to a buffer size
1807 of two gigabytes and the error handling improved.
1809 This issue was reported to OpenSSL by Dr. Falko Strenzke. It has been
1810 categorized as a normal bug, not a security issue, because the DRBG reseeds
1811 automatically and is fully functional even without additional randomness
1812 provided by the application.
1814 ### Changes between 1.1.0i and 1.1.1 [11 Sep 2018]
1816 * Add a new ClientHello callback. Provides a callback interface that gives
1817 the application the ability to adjust the nascent SSL object at the
1818 earliest stage of ClientHello processing, immediately after extensions have
1819 been collected but before they have been processed. In particular, this
1820 callback can adjust the supported TLS versions in response to the contents
1825 * Add SM2 base algorithm support.
1829 * s390x assembly pack: add (improved) hardware-support for the following
1830 cryptographic primitives: sha3, shake, aes-gcm, aes-ccm, aes-ctr, aes-ofb,
1831 aes-cfb/cfb8, aes-ecb.
1835 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
1836 parameter is no longer accepted, as it leads to a corrupt table. NULL
1837 pem_str is reserved for alias entries only.
1841 * Use the new ec_scalar_mul_ladder scaffold to implement a specialized ladder
1842 step for prime curves. The new implementation is based on formulae from
1843 differential addition-and-doubling in homogeneous projective coordinates
1844 from Izu-Takagi "A fast parallel elliptic curve multiplication resistant
1845 against side channel attacks" and Brier-Joye "Weierstrass Elliptic Curves
1846 and Side-Channel Attacks" Eq. (8) for y-coordinate recovery, modified
1847 to work in projective coordinates.
1849 *Billy Bob Brumley, Nicola Tuveri*
1851 * Change generating and checking of primes so that the error rate of not
1852 being prime depends on the intended use based on the size of the input.
1853 For larger primes this will result in more rounds of Miller-Rabin.
1854 The maximal error rate for primes with more than 1080 bits is lowered
1857 *Kurt Roeckx, Annie Yousar*
1859 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
1863 * The 'tsget' script is renamed to 'tsget.pl', to avoid confusion when
1864 moving between systems, and to avoid confusion when a Windows build is
1865 done with mingw vs with MSVC. For POSIX installs, there's still a
1866 symlink or copy named 'tsget' to avoid that confusion as well.
1870 * Revert blinding in ECDSA sign and instead make problematic addition
1871 length-invariant. Switch even to fixed-length Montgomery multiplication.
1875 * Use the new ec_scalar_mul_ladder scaffold to implement a specialized ladder
1876 step for binary curves. The new implementation is based on formulae from
1877 differential addition-and-doubling in mixed Lopez-Dahab projective
1878 coordinates, modified to independently blind the operands.
1880 *Billy Bob Brumley, Sohaib ul Hassan, Nicola Tuveri*
1882 * Add a scaffold to optionally enhance the Montgomery ladder implementation
1883 for `ec_scalar_mul_ladder` (formerly `ec_mul_consttime`) allowing
1884 EC_METHODs to implement their own specialized "ladder step", to take
1885 advantage of more favorable coordinate systems or more efficient
1886 differential addition-and-doubling algorithms.
1888 *Billy Bob Brumley, Sohaib ul Hassan, Nicola Tuveri*
1890 * Modified the random device based seed sources to keep the relevant
1891 file descriptors open rather than reopening them on each access.
1892 This allows such sources to operate in a chroot() jail without
1893 the associated device nodes being available. This behaviour can be
1894 controlled using RAND_keep_random_devices_open().
1898 * Numerous side-channel attack mitigations have been applied. This may have
1899 performance impacts for some algorithms for the benefit of improved
1900 security. Specific changes are noted in this change log by their respective
1905 * AIX shared library support overhaul. Switch to AIX "natural" way of
1906 handling shared libraries, which means collecting shared objects of
1907 different versions and bitnesses in one common archive. This allows to
1908 mitigate conflict between 1.0 and 1.1 side-by-side installations. It
1909 doesn't affect the way 3rd party applications are linked, only how
1910 multi-version installation is managed.
1914 * Make ec_group_do_inverse_ord() more robust and available to other
1915 EC cryptosystems, so that irrespective of BN_FLG_CONSTTIME, SCA
1916 mitigations are applied to the fallback BN_mod_inverse().
1917 When using this function rather than BN_mod_inverse() directly, new
1918 EC cryptosystem implementations are then safer-by-default.
1922 * Add coordinate blinding for EC_POINT and implement projective
1923 coordinate blinding for generic prime curves as a countermeasure to
1924 chosen point SCA attacks.
1926 *Sohaib ul Hassan, Nicola Tuveri, Billy Bob Brumley*
1928 * Add blinding to ECDSA and DSA signatures to protect against side channel
1929 attacks discovered by Keegan Ryan (NCC Group).
1933 * Enforce checking in the `pkeyutl` command to ensure that the input
1934 length does not exceed the maximum supported digest length when performing
1935 a sign, verify or verifyrecover operation.
1939 * SSL_MODE_AUTO_RETRY is enabled by default. Applications that use blocking
1940 I/O in combination with something like select() or poll() will hang. This
1941 can be turned off again using SSL_CTX_clear_mode().
1942 Many applications do not properly handle non-application data records, and
1943 TLS 1.3 sends more of such records. Setting SSL_MODE_AUTO_RETRY works
1944 around the problems in those applications, but can also break some.
1945 It's recommended to read the manpages about SSL_read(), SSL_write(),
1946 SSL_get_error(), SSL_shutdown(), SSL_CTX_set_mode() and
1947 SSL_CTX_set_read_ahead() again.
1951 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
1952 now allow empty (zero character) pass phrases.
1956 * Apply blinding to binary field modular inversion and remove patent
1957 pending (OPENSSL_SUN_GF2M_DIV) BN_GF2m_mod_div implementation.
1961 * Deprecate ec2_mult.c and unify scalar multiplication code paths for
1962 binary and prime elliptic curves.
1966 * Remove ECDSA nonce padding: EC_POINT_mul is now responsible for
1967 constant time fixed point multiplication.
1971 * Revise elliptic curve scalar multiplication with timing attack
1972 defenses: ec_wNAF_mul redirects to a constant time implementation
1973 when computing fixed point and variable point multiplication (which
1974 in OpenSSL are mostly used with secret scalars in keygen, sign,
1975 ECDH derive operations).
1976 *Billy Bob Brumley, Nicola Tuveri, Cesar Pereida García,
1979 * Updated CONTRIBUTING
1983 * Updated DRBG / RAND to request nonce and additional low entropy
1984 randomness from the system.
1986 *Matthias St. Pierre*
1988 * Updated 'openssl rehash' to use OpenSSL consistent default.
1992 * Moved the load of the ssl_conf module to libcrypto, which helps
1993 loading engines that libssl uses before libssl is initialised.
1997 * Added EVP_PKEY_sign() and EVP_PKEY_verify() for EdDSA
2001 * Fixed X509_NAME_ENTRY_set to get multi-valued RDNs right in all cases.
2003 *Ingo Schwarze, Rich Salz*
2005 * Added output of accepting IP address and port for 'openssl s_server'
2009 * Added a new API for TLSv1.3 ciphersuites:
2010 SSL_CTX_set_ciphersuites()
2011 SSL_set_ciphersuites()
2015 * Memory allocation failures consistently add an error to the error
2020 * Don't use OPENSSL_ENGINES and OPENSSL_CONF environment values
2021 in libcrypto when run as setuid/setgid.
2025 * Load any config file by default when libssl is used.
2029 * Added new public header file <openssl/rand_drbg.h> and documentation
2030 for the RAND_DRBG API. See manual page RAND_DRBG(7) for an overview.
2032 *Matthias St. Pierre*
2034 * QNX support removed (cannot find contributors to get their approval
2035 for the license change).
2039 * TLSv1.3 replay protection for early data has been implemented. See the
2040 SSL_read_early_data() man page for further details.
2044 * Separated TLSv1.3 ciphersuite configuration out from TLSv1.2 ciphersuite
2045 configuration. TLSv1.3 ciphersuites are not compatible with TLSv1.2 and
2046 below. Similarly TLSv1.2 ciphersuites are not compatible with TLSv1.3.
2047 In order to avoid issues where legacy TLSv1.2 ciphersuite configuration
2048 would otherwise inadvertently disable all TLSv1.3 ciphersuites the
2049 configuration has been separated out. See the ciphers man page or the
2050 SSL_CTX_set_ciphersuites() man page for more information.
2054 * On POSIX (BSD, Linux, ...) systems the ocsp(1) command running
2055 in responder mode now supports the new "-multi" option, which
2056 spawns the specified number of child processes to handle OCSP
2057 requests. The "-timeout" option now also limits the OCSP
2058 responder's patience to wait to receive the full client request
2059 on a newly accepted connection. Child processes are respawned
2060 as needed, and the CA index file is automatically reloaded
2061 when changed. This makes it possible to run the "ocsp" responder
2062 as a long-running service, making the OpenSSL CA somewhat more
2063 feature-complete. In this mode, most diagnostic messages logged
2064 after entering the event loop are logged via syslog(3) rather than
2069 * Added support for X448 and Ed448. Heavily based on original work by
2074 * Extend OSSL_STORE with capabilities to search and to narrow the set of
2075 objects loaded. This adds the functions OSSL_STORE_expect() and
2076 OSSL_STORE_find() as well as needed tools to construct searches and
2077 get the search data out of them.
2081 * Support for TLSv1.3 added. Note that users upgrading from an earlier
2082 version of OpenSSL should review their configuration settings to ensure
2083 that they are still appropriate for TLSv1.3. For further information see:
2084 <https://wiki.openssl.org/index.php/TLS1.3>
2088 * Grand redesign of the OpenSSL random generator
2090 The default RAND method now utilizes an AES-CTR DRBG according to
2091 NIST standard SP 800-90Ar1. The new random generator is essentially
2092 a port of the default random generator from the OpenSSL FIPS 2.0
2093 object module. It is a hybrid deterministic random bit generator
2094 using an AES-CTR bit stream and which seeds and reseeds itself
2095 automatically using trusted system entropy sources.
2097 Some of its new features are:
2098 - Support for multiple DRBG instances with seed chaining.
2099 - The default RAND method makes use of a DRBG.
2100 - There is a public and private DRBG instance.
2101 - The DRBG instances are fork-safe.
2102 - Keep all global DRBG instances on the secure heap if it is enabled.
2103 - The public and private DRBG instance are per thread for lock free
2106 *Paul Dale, Benjamin Kaduk, Kurt Roeckx, Rich Salz, Matthias St. Pierre*
2108 * Changed Configure so it only says what it does and doesn't dump
2109 so much data. Instead, ./configdata.pm should be used as a script
2110 to display all sorts of configuration data.
2114 * Added processing of "make variables" to Configure.
2118 * Added SHA512/224 and SHA512/256 algorithm support.
2122 * The last traces of Netware support, first removed in 1.1.0, have
2127 * Get rid of Makefile.shared, and in the process, make the processing
2128 of certain files (rc.obj, or the .def/.map/.opt files produced from
2129 the ordinal files) more visible and hopefully easier to trace and
2130 debug (or make silent).
2134 * Make it possible to have environment variable assignments as
2135 arguments to config / Configure.
2139 * Add multi-prime RSA (RFC 8017) support.
2143 * Add SM3 implemented according to GB/T 32905-2016
2144 *Jack Lloyd <jack.lloyd@ribose.com>,*
2145 *Ronald Tse <ronald.tse@ribose.com>,*
2146 *Erick Borsboom <erick.borsboom@ribose.com>*
2148 * Add 'Maximum Fragment Length' TLS extension negotiation and support
2149 as documented in RFC6066.
2150 Based on a patch from Tomasz Moń
2152 *Filipe Raimundo da Silva*
2154 * Add SM4 implemented according to GB/T 32907-2016.
2155 *Jack Lloyd <jack.lloyd@ribose.com>,*
2156 *Ronald Tse <ronald.tse@ribose.com>,*
2157 *Erick Borsboom <erick.borsboom@ribose.com>*
2159 * Reimplement -newreq-nodes and ERR_error_string_n; the
2160 original author does not agree with the license change.
2164 * Add ARIA AEAD TLS support.
2168 * Some macro definitions to support VS6 have been removed. Visual
2169 Studio 6 has not worked since 1.1.0
2173 * Add ERR_clear_last_mark(), to allow callers to clear the last mark
2174 without clearing the errors.
2178 * Add "atfork" functions. If building on a system that without
2179 pthreads, see doc/man3/OPENSSL_fork_prepare.pod for application
2180 requirements. The RAND facility now uses/requires this.
2188 * The UI API becomes a permanent and integral part of libcrypto, i.e.
2189 not possible to disable entirely. However, it's still possible to
2190 disable the console reading UI method, UI_OpenSSL() (use UI_null()
2193 To disable, configure with 'no-ui-console'. 'no-ui' is still
2194 possible to use as an alias. Check at compile time with the
2195 macro OPENSSL_NO_UI_CONSOLE. The macro OPENSSL_NO_UI is still
2196 possible to check and is an alias for OPENSSL_NO_UI_CONSOLE.
2200 * Add a STORE module, which implements a uniform and URI based reader of
2201 stores that can contain keys, certificates, CRLs and numerous other
2202 objects. The main API is loosely based on a few stdio functions,
2203 and includes OSSL_STORE_open, OSSL_STORE_load, OSSL_STORE_eof,
2204 OSSL_STORE_error and OSSL_STORE_close.
2205 The implementation uses backends called "loaders" to implement arbitrary
2206 URI schemes. There is one built in "loader" for the 'file' scheme.
2210 * Add devcrypto engine. This has been implemented against cryptodev-linux,
2211 then adjusted to work on FreeBSD 8.4 as well.
2212 Enable by configuring with 'enable-devcryptoeng'. This is done by default
2213 on BSD implementations, as cryptodev.h is assumed to exist on all of them.
2217 * Module names can prefixed with OSSL_ or OPENSSL_. This affects
2218 util/mkerr.pl, which is adapted to allow those prefixes, leading to
2219 error code calls like this:
2221 OSSL_FOOerr(OSSL_FOO_F_SOMETHING, OSSL_FOO_R_WHATEVER);
2223 With this change, we claim the namespaces OSSL and OPENSSL in a manner
2224 that can be encoded in C. For the foreseeable future, this will only
2227 *Richard Levitte and Tim Hudson*
2229 * Removed BSD cryptodev engine.
2233 * Add a build target 'build_all_generated', to build all generated files
2234 and only that. This can be used to prepare everything that requires
2235 things like perl for a system that lacks perl and then move everything
2236 to that system and do the rest of the build there.
2240 * In the UI interface, make it possible to duplicate the user data. This
2241 can be used by engines that need to retain the data for a longer time
2242 than just the call where this user data is passed.
2246 * Ignore the '-named_curve auto' value for compatibility of applications
2249 *Tomáš Mráz <tmraz@fedoraproject.org>*
2251 * Fragmented SSL/TLS alerts are no longer accepted. An alert message is 2
2252 bytes long. In theory it is permissible in SSLv3 - TLSv1.2 to fragment such
2253 alerts across multiple records (some of which could be empty). In practice
2254 it make no sense to send an empty alert record, or to fragment one. TLSv1.3
2255 prohibits this altogether and other libraries (BoringSSL, NSS) do not
2256 support this at all. Supporting it adds significant complexity to the
2257 record layer, and its removal is unlikely to cause interoperability
2262 * Add the ASN.1 types INT32, UINT32, INT64, UINT64 and variants prefixed
2263 with Z. These are meant to replace LONG and ZLONG and to be size safe.
2264 The use of LONG and ZLONG is discouraged and scheduled for deprecation
2269 * Add the 'z' and 'j' modifiers to BIO_printf() et al formatting string,
2270 'z' is to be used for [s]size_t, and 'j' - with [u]int64_t.
2272 *Richard Levitte, Andy Polyakov*
2274 * Add EC_KEY_get0_engine(), which does for EC_KEY what RSA_get0_engine()
2279 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
2280 platform rather than 'mingw'.
2284 * The functions X509_STORE_add_cert and X509_STORE_add_crl return
2285 success if they are asked to add an object which already exists
2286 in the store. This change cascades to other functions which load
2287 certificates and CRLs.
2291 * x86_64 assembly pack: annotate code with DWARF CFI directives to
2292 facilitate stack unwinding even from assembly subroutines.
2296 * Remove VAX C specific definitions of OPENSSL_EXPORT, OPENSSL_EXTERN.
2297 Also remove OPENSSL_GLOBAL entirely, as it became a no-op.
2301 * Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
2302 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
2303 which is the minimum version we support.
2307 * Certificate time validation (X509_cmp_time) enforces stricter
2308 compliance with RFC 5280. Fractional seconds and timezone offsets
2309 are no longer allowed.
2313 * Add support for ARIA
2317 * s_client will now send the Server Name Indication (SNI) extension by
2318 default unless the new "-noservername" option is used. The server name is
2319 based on the host provided to the "-connect" option unless overridden by
2320 using "-servername".
2324 * Add support for SipHash
2328 * OpenSSL now fails if it receives an unrecognised record type in TLS1.0
2329 or TLS1.1. Previously this only happened in SSLv3 and TLS1.2. This is to
2330 prevent issues where no progress is being made and the peer continually
2331 sends unrecognised record types, using up resources processing them.
2335 * 'openssl passwd' can now produce SHA256 and SHA512 based output,
2336 using the algorithm defined in
2337 <https://www.akkadia.org/drepper/SHA-crypt.txt>
2341 * Heartbeat support has been removed; the ABI is changed for now.
2343 *Richard Levitte, Rich Salz*
2345 * Support for SSL_OP_NO_ENCRYPT_THEN_MAC in SSL_CONF_cmd.
2349 * The RSA "null" method, which was partially supported to avoid patent
2350 issues, has been replaced to always returns NULL.
2357 ### Changes between 1.1.0k and 1.1.0l [10 Sep 2019]
2359 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
2360 used even when parsing explicit parameters, when loading a encoded key
2361 or calling `EC_GROUP_new_from_ecpkparameters()`/
2362 `EC_GROUP_new_from_ecparameters()`.
2363 This prevents bypass of security hardening and performance gains,
2364 especially for curves with specialized EC_METHODs.
2365 By default, if a key encoded with explicit parameters is loaded and later
2366 encoded, the output is still encoded with explicit parameters, even if
2367 internally a "named" EC_GROUP is used for computation.
2371 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
2372 this change, EC_GROUP_set_generator would accept order and/or cofactor as
2373 NULL. After this change, only the cofactor parameter can be NULL. It also
2374 does some minimal sanity checks on the passed order.
2379 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
2380 An attack is simple, if the first CMS_recipientInfo is valid but the
2381 second CMS_recipientInfo is chosen ciphertext. If the second
2382 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
2383 encryption key will be replaced by garbage, and the message cannot be
2384 decoded, but if the RSA decryption fails, the correct encryption key is
2385 used and the recipient will not notice the attack.
2386 As a work around for this potential attack the length of the decrypted
2387 key must be equal to the cipher default key length, in case the
2388 certifiate is not given and all recipientInfo are tried out.
2389 The old behaviour can be re-enabled in the CMS code by setting the
2390 CMS_DEBUG_DECRYPT flag.
2395 * Use Windows installation paths in the mingw builds
2397 Mingw isn't a POSIX environment per se, which means that Windows
2398 paths should be used for installation.
2403 ### Changes between 1.1.0j and 1.1.0k [28 May 2019]
2405 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
2406 This changes the size when using the `genpkey` command when no size is given.
2407 It fixes an omission in earlier changes that changed all RSA, DSA and DH
2408 generation commands to use 2048 bits by default.
2412 * Prevent over long nonces in ChaCha20-Poly1305.
2414 ChaCha20-Poly1305 is an AEAD cipher, and requires a unique nonce input
2415 for every encryption operation. RFC 7539 specifies that the nonce value
2416 (IV) should be 96 bits (12 bytes). OpenSSL allows a variable nonce length
2417 and front pads the nonce with 0 bytes if it is less than 12
2418 bytes. However it also incorrectly allows a nonce to be set of up to 16
2419 bytes. In this case only the last 12 bytes are significant and any
2420 additional leading bytes are ignored.
2422 It is a requirement of using this cipher that nonce values are
2423 unique. Messages encrypted using a reused nonce value are susceptible to
2424 serious confidentiality and integrity attacks. If an application changes
2425 the default nonce length to be longer than 12 bytes and then makes a
2426 change to the leading bytes of the nonce expecting the new value to be a
2427 new unique nonce then such an application could inadvertently encrypt
2428 messages with a reused nonce.
2430 Additionally the ignored bytes in a long nonce are not covered by the
2431 integrity guarantee of this cipher. Any application that relies on the
2432 integrity of these ignored leading bytes of a long nonce may be further
2433 affected. Any OpenSSL internal use of this cipher, including in SSL/TLS,
2434 is safe because no such use sets such a long nonce value. However user
2435 applications that use this cipher directly and set a non-default nonce
2436 length to be longer than 12 bytes may be vulnerable.
2438 This issue was reported to OpenSSL on 16th of March 2019 by Joran Dirk
2444 * Added SCA hardening for modular field inversion in EC_GROUP through
2445 a new dedicated field_inv() pointer in EC_METHOD.
2446 This also addresses a leakage affecting conversions from projective
2447 to affine coordinates.
2449 *Billy Bob Brumley, Nicola Tuveri*
2451 * Fix a use after free bug in d2i_X509_PUBKEY when overwriting a
2452 re-used X509_PUBKEY object if the second PUBKEY is malformed.
2456 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
2460 * Remove the 'dist' target and add a tarball building script. The
2461 'dist' target has fallen out of use, and it shouldn't be
2462 necessary to configure just to create a source distribution.
2466 ### Changes between 1.1.0i and 1.1.0j [20 Nov 2018]
2468 * Timing vulnerability in DSA signature generation
2470 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
2471 timing side channel attack. An attacker could use variations in the signing
2472 algorithm to recover the private key.
2474 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
2479 * Timing vulnerability in ECDSA signature generation
2481 The OpenSSL ECDSA signature algorithm has been shown to be vulnerable to a
2482 timing side channel attack. An attacker could use variations in the signing
2483 algorithm to recover the private key.
2485 This issue was reported to OpenSSL on 25th October 2018 by Samuel Weiser.
2490 * Add coordinate blinding for EC_POINT and implement projective
2491 coordinate blinding for generic prime curves as a countermeasure to
2492 chosen point SCA attacks.
2494 *Sohaib ul Hassan, Nicola Tuveri, Billy Bob Brumley*
2496 ### Changes between 1.1.0h and 1.1.0i [14 Aug 2018]
2498 * Client DoS due to large DH parameter
2500 During key agreement in a TLS handshake using a DH(E) based ciphersuite a
2501 malicious server can send a very large prime value to the client. This will
2502 cause the client to spend an unreasonably long period of time generating a
2503 key for this prime resulting in a hang until the client has finished. This
2504 could be exploited in a Denial Of Service attack.
2506 This issue was reported to OpenSSL on 5th June 2018 by Guido Vranken
2511 * Cache timing vulnerability in RSA Key Generation
2513 The OpenSSL RSA Key generation algorithm has been shown to be vulnerable to
2514 a cache timing side channel attack. An attacker with sufficient access to
2515 mount cache timing attacks during the RSA key generation process could
2516 recover the private key.
2518 This issue was reported to OpenSSL on 4th April 2018 by Alejandro Cabrera
2519 Aldaya, Billy Brumley, Cesar Pereida Garcia and Luis Manuel Alvarez Tapia.
2524 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
2525 parameter is no longer accepted, as it leads to a corrupt table. NULL
2526 pem_str is reserved for alias entries only.
2530 * Revert blinding in ECDSA sign and instead make problematic addition
2531 length-invariant. Switch even to fixed-length Montgomery multiplication.
2535 * Change generating and checking of primes so that the error rate of not
2536 being prime depends on the intended use based on the size of the input.
2537 For larger primes this will result in more rounds of Miller-Rabin.
2538 The maximal error rate for primes with more than 1080 bits is lowered
2541 *Kurt Roeckx, Annie Yousar*
2543 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
2547 * Add blinding to ECDSA and DSA signatures to protect against side channel
2548 attacks discovered by Keegan Ryan (NCC Group).
2552 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
2553 now allow empty (zero character) pass phrases.
2557 * Certificate time validation (X509_cmp_time) enforces stricter
2558 compliance with RFC 5280. Fractional seconds and timezone offsets
2559 are no longer allowed.
2563 * Fixed a text canonicalisation bug in CMS
2565 Where a CMS detached signature is used with text content the text goes
2566 through a canonicalisation process first prior to signing or verifying a
2567 signature. This process strips trailing space at the end of lines, converts
2568 line terminators to CRLF and removes additional trailing line terminators
2569 at the end of a file. A bug in the canonicalisation process meant that
2570 some characters, such as form-feed, were incorrectly treated as whitespace
2571 and removed. This is contrary to the specification (RFC5485). This fix
2572 could mean that detached text data signed with an earlier version of
2573 OpenSSL 1.1.0 may fail to verify using the fixed version, or text data
2574 signed with a fixed OpenSSL may fail to verify with an earlier version of
2575 OpenSSL 1.1.0. A workaround is to only verify the canonicalised text data
2576 and use the "-binary" flag (for the "cms" command line application) or set
2577 the SMIME_BINARY/PKCS7_BINARY/CMS_BINARY flags (if using CMS_verify()).
2581 ### Changes between 1.1.0g and 1.1.0h [27 Mar 2018]
2583 * Constructed ASN.1 types with a recursive definition could exceed the stack
2585 Constructed ASN.1 types with a recursive definition (such as can be found
2586 in PKCS7) could eventually exceed the stack given malicious input with
2587 excessive recursion. This could result in a Denial Of Service attack. There
2588 are no such structures used within SSL/TLS that come from untrusted sources
2589 so this is considered safe.
2591 This issue was reported to OpenSSL on 4th January 2018 by the OSS-fuzz
2597 * Incorrect CRYPTO_memcmp on HP-UX PA-RISC
2599 Because of an implementation bug the PA-RISC CRYPTO_memcmp function is
2600 effectively reduced to only comparing the least significant bit of each
2601 byte. This allows an attacker to forge messages that would be considered as
2602 authenticated in an amount of tries lower than that guaranteed by the
2603 security claims of the scheme. The module can only be compiled by the
2604 HP-UX assembler, so that only HP-UX PA-RISC targets are affected.
2606 This issue was reported to OpenSSL on 2nd March 2018 by Peter Waltenberg
2612 * Add a build target 'build_all_generated', to build all generated files
2613 and only that. This can be used to prepare everything that requires
2614 things like perl for a system that lacks perl and then move everything
2615 to that system and do the rest of the build there.
2619 * Backport SSL_OP_NO_RENGOTIATION
2621 OpenSSL 1.0.2 and below had the ability to disable renegotiation using the
2622 (undocumented) SSL3_FLAGS_NO_RENEGOTIATE_CIPHERS flag. Due to the opacity
2623 changes this is no longer possible in 1.1.0. Therefore the new
2624 SSL_OP_NO_RENEGOTIATION option from 1.1.1-dev has been backported to
2625 1.1.0 to provide equivalent functionality.
2627 Note that if an application built against 1.1.0h headers (or above) is run
2628 using an older version of 1.1.0 (prior to 1.1.0h) then the option will be
2629 accepted but nothing will happen, i.e. renegotiation will not be prevented.
2633 * Removed the OS390-Unix config target. It relied on a script that doesn't
2638 * rsaz_1024_mul_avx2 overflow bug on x86_64
2640 There is an overflow bug in the AVX2 Montgomery multiplication procedure
2641 used in exponentiation with 1024-bit moduli. No EC algorithms are affected.
2642 Analysis suggests that attacks against RSA and DSA as a result of this
2643 defect would be very difficult to perform and are not believed likely.
2644 Attacks against DH1024 are considered just feasible, because most of the
2645 work necessary to deduce information about a private key may be performed
2646 offline. The amount of resources required for such an attack would be
2647 significant. However, for an attack on TLS to be meaningful, the server
2648 would have to share the DH1024 private key among multiple clients, which is
2649 no longer an option since CVE-2016-0701.
2651 This only affects processors that support the AVX2 but not ADX extensions
2652 like Intel Haswell (4th generation).
2654 This issue was reported to OpenSSL by David Benjamin (Google). The issue
2655 was originally found via the OSS-Fuzz project.
2660 ### Changes between 1.1.0f and 1.1.0g [2 Nov 2017]
2662 * bn_sqrx8x_internal carry bug on x86_64
2664 There is a carry propagating bug in the x86_64 Montgomery squaring
2665 procedure. No EC algorithms are affected. Analysis suggests that attacks
2666 against RSA and DSA as a result of this defect would be very difficult to
2667 perform and are not believed likely. Attacks against DH are considered just
2668 feasible (although very difficult) because most of the work necessary to
2669 deduce information about a private key may be performed offline. The amount
2670 of resources required for such an attack would be very significant and
2671 likely only accessible to a limited number of attackers. An attacker would
2672 additionally need online access to an unpatched system using the target
2673 private key in a scenario with persistent DH parameters and a private
2674 key that is shared between multiple clients.
2676 This only affects processors that support the BMI1, BMI2 and ADX extensions
2677 like Intel Broadwell (5th generation) and later or AMD Ryzen.
2679 This issue was reported to OpenSSL by the OSS-Fuzz project.
2684 * Malformed X.509 IPAddressFamily could cause OOB read
2686 If an X.509 certificate has a malformed IPAddressFamily extension,
2687 OpenSSL could do a one-byte buffer overread. The most likely result
2688 would be an erroneous display of the certificate in text format.
2690 This issue was reported to OpenSSL by the OSS-Fuzz project.
2695 ### Changes between 1.1.0e and 1.1.0f [25 May 2017]
2697 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
2698 platform rather than 'mingw'.
2702 * Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
2703 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
2704 which is the minimum version we support.
2708 ### Changes between 1.1.0d and 1.1.0e [16 Feb 2017]
2710 * Encrypt-Then-Mac renegotiation crash
2712 During a renegotiation handshake if the Encrypt-Then-Mac extension is
2713 negotiated where it was not in the original handshake (or vice-versa) then
2714 this can cause OpenSSL to crash (dependant on ciphersuite). Both clients
2715 and servers are affected.
2717 This issue was reported to OpenSSL by Joe Orton (Red Hat).
2722 ### Changes between 1.1.0c and 1.1.0d [26 Jan 2017]
2724 * Truncated packet could crash via OOB read
2726 If one side of an SSL/TLS path is running on a 32-bit host and a specific
2727 cipher is being used, then a truncated packet can cause that host to
2728 perform an out-of-bounds read, usually resulting in a crash.
2730 This issue was reported to OpenSSL by Robert Święcki of Google.
2735 * Bad (EC)DHE parameters cause a client crash
2737 If a malicious server supplies bad parameters for a DHE or ECDHE key
2738 exchange then this can result in the client attempting to dereference a
2739 NULL pointer leading to a client crash. This could be exploited in a Denial
2742 This issue was reported to OpenSSL by Guido Vranken.
2747 * BN_mod_exp may produce incorrect results on x86_64
2749 There is a carry propagating bug in the x86_64 Montgomery squaring
2750 procedure. No EC algorithms are affected. Analysis suggests that attacks
2751 against RSA and DSA as a result of this defect would be very difficult to
2752 perform and are not believed likely. Attacks against DH are considered just
2753 feasible (although very difficult) because most of the work necessary to
2754 deduce information about a private key may be performed offline. The amount
2755 of resources required for such an attack would be very significant and
2756 likely only accessible to a limited number of attackers. An attacker would
2757 additionally need online access to an unpatched system using the target
2758 private key in a scenario with persistent DH parameters and a private
2759 key that is shared between multiple clients. For example this can occur by
2760 default in OpenSSL DHE based SSL/TLS ciphersuites. Note: This issue is very
2761 similar to CVE-2015-3193 but must be treated as a separate problem.
2763 This issue was reported to OpenSSL by the OSS-Fuzz project.
2768 ### Changes between 1.1.0b and 1.1.0c [10 Nov 2016]
2770 * ChaCha20/Poly1305 heap-buffer-overflow
2772 TLS connections using `*-CHACHA20-POLY1305` ciphersuites are susceptible to
2773 a DoS attack by corrupting larger payloads. This can result in an OpenSSL
2774 crash. This issue is not considered to be exploitable beyond a DoS.
2776 This issue was reported to OpenSSL by Robert Święcki (Google Security Team)
2781 * CMS Null dereference
2783 Applications parsing invalid CMS structures can crash with a NULL pointer
2784 dereference. This is caused by a bug in the handling of the ASN.1 CHOICE
2785 type in OpenSSL 1.1.0 which can result in a NULL value being passed to the
2786 structure callback if an attempt is made to free certain invalid encodings.
2787 Only CHOICE structures using a callback which do not handle NULL value are
2790 This issue was reported to OpenSSL by Tyler Nighswander of ForAllSecure.
2795 * Montgomery multiplication may produce incorrect results
2797 There is a carry propagating bug in the Broadwell-specific Montgomery
2798 multiplication procedure that handles input lengths divisible by, but
2799 longer than 256 bits. Analysis suggests that attacks against RSA, DSA
2800 and DH private keys are impossible. This is because the subroutine in
2801 question is not used in operations with the private key itself and an input
2802 of the attacker's direct choice. Otherwise the bug can manifest itself as
2803 transient authentication and key negotiation failures or reproducible
2804 erroneous outcome of public-key operations with specially crafted input.
2805 Among EC algorithms only Brainpool P-512 curves are affected and one
2806 presumably can attack ECDH key negotiation. Impact was not analyzed in
2807 detail, because pre-requisites for attack are considered unlikely. Namely
2808 multiple clients have to choose the curve in question and the server has to
2809 share the private key among them, neither of which is default behaviour.
2810 Even then only clients that chose the curve will be affected.
2812 This issue was publicly reported as transient failures and was not
2813 initially recognized as a security issue. Thanks to Richard Morgan for
2814 providing reproducible case.
2819 * Removed automatic addition of RPATH in shared libraries and executables,
2820 as this was a remainder from OpenSSL 1.0.x and isn't needed any more.
2824 ### Changes between 1.1.0a and 1.1.0b [26 Sep 2016]
2826 * Fix Use After Free for large message sizes
2828 The patch applied to address CVE-2016-6307 resulted in an issue where if a
2829 message larger than approx 16k is received then the underlying buffer to
2830 store the incoming message is reallocated and moved. Unfortunately a
2831 dangling pointer to the old location is left which results in an attempt to
2832 write to the previously freed location. This is likely to result in a
2833 crash, however it could potentially lead to execution of arbitrary code.
2835 This issue only affects OpenSSL 1.1.0a.
2837 This issue was reported to OpenSSL by Robert Święcki.
2842 ### Changes between 1.1.0 and 1.1.0a [22 Sep 2016]
2844 * OCSP Status Request extension unbounded memory growth
2846 A malicious client can send an excessively large OCSP Status Request
2847 extension. If that client continually requests renegotiation, sending a
2848 large OCSP Status Request extension each time, then there will be unbounded
2849 memory growth on the server. This will eventually lead to a Denial Of
2850 Service attack through memory exhaustion. Servers with a default
2851 configuration are vulnerable even if they do not support OCSP. Builds using
2852 the "no-ocsp" build time option are not affected.
2854 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
2859 * SSL_peek() hang on empty record
2861 OpenSSL 1.1.0 SSL/TLS will hang during a call to SSL_peek() if the peer
2862 sends an empty record. This could be exploited by a malicious peer in a
2863 Denial Of Service attack.
2865 This issue was reported to OpenSSL by Alex Gaynor.
2870 * Excessive allocation of memory in tls_get_message_header() and
2871 dtls1_preprocess_fragment()
2873 A (D)TLS message includes 3 bytes for its length in the header for the
2874 message. This would allow for messages up to 16Mb in length. Messages of
2875 this length are excessive and OpenSSL includes a check to ensure that a
2876 peer is sending reasonably sized messages in order to avoid too much memory
2877 being consumed to service a connection. A flaw in the logic of version
2878 1.1.0 means that memory for the message is allocated too early, prior to
2879 the excessive message length check. Due to way memory is allocated in
2880 OpenSSL this could mean an attacker could force up to 21Mb to be allocated
2881 to service a connection. This could lead to a Denial of Service through
2882 memory exhaustion. However, the excessive message length check still takes
2883 place, and this would cause the connection to immediately fail. Assuming
2884 that the application calls SSL_free() on the failed connection in a timely
2885 manner then the 21Mb of allocated memory will then be immediately freed
2886 again. Therefore the excessive memory allocation will be transitory in
2887 nature. This then means that there is only a security impact if:
2889 1) The application does not call SSL_free() in a timely manner in the event
2890 that the connection fails
2892 2) The application is working in a constrained environment where there is
2893 very little free memory
2895 3) The attacker initiates multiple connection attempts such that there are
2896 multiple connections in a state where memory has been allocated for the
2897 connection; SSL_free() has not yet been called; and there is insufficient
2898 memory to service the multiple requests.
2900 Except in the instance of (1) above any Denial Of Service is likely to be
2901 transitory because as soon as the connection fails the memory is
2902 subsequently freed again in the SSL_free() call. However there is an
2903 increased risk during this period of application crashes due to the lack of
2904 memory - which would then mean a more serious Denial of Service.
2906 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
2907 (CVE-2016-6307 and CVE-2016-6308)
2911 * solaris-x86-cc, i.e. 32-bit configuration with vendor compiler,
2912 had to be removed. Primary reason is that vendor assembler can't
2913 assemble our modules with -KPIC flag. As result it, assembly
2914 support, was not even available as option. But its lack means
2915 lack of side-channel resistant code, which is incompatible with
2916 security by todays standards. Fortunately gcc is readily available
2917 prepackaged option, which we firmly point at...
2921 ### Changes between 1.0.2h and 1.1.0 [25 Aug 2016]
2923 * Windows command-line tool supports UTF-8 opt-in option for arguments
2924 and console input. Setting OPENSSL_WIN32_UTF8 environment variable
2925 (to any value) allows Windows user to access PKCS#12 file generated
2926 with Windows CryptoAPI and protected with non-ASCII password, as well
2927 as files generated under UTF-8 locale on Linux also protected with
2932 * To mitigate the SWEET32 attack ([CVE-2016-2183]), 3DES cipher suites
2933 have been disabled by default and removed from DEFAULT, just like RC4.
2934 See the RC4 item below to re-enable both.
2938 * The method for finding the storage location for the Windows RAND seed file
2939 has changed. First we check %RANDFILE%. If that is not set then we check
2940 the directories %HOME%, %USERPROFILE% and %SYSTEMROOT% in that order. If
2941 all else fails we fall back to C:\.
2945 * The EVP_EncryptUpdate() function has had its return type changed from void
2946 to int. A return of 0 indicates and error while a return of 1 indicates
2951 * The flags RSA_FLAG_NO_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME and
2952 DH_FLAG_NO_EXP_CONSTTIME which previously provided the ability to switch
2953 off the constant time implementation for RSA, DSA and DH have been made
2954 no-ops and deprecated.
2958 * Windows RAND implementation was simplified to only get entropy by
2959 calling CryptGenRandom(). Various other RAND-related tickets
2962 *Joseph Wylie Yandle, Rich Salz*
2964 * The stack and lhash API's were renamed to start with `OPENSSL_SK_`
2965 and `OPENSSL_LH_`, respectively. The old names are available
2966 with API compatibility. They new names are now completely documented.
2970 * Unify TYPE_up_ref(obj) methods signature.
2971 SSL_CTX_up_ref(), SSL_up_ref(), X509_up_ref(), EVP_PKEY_up_ref(),
2972 X509_CRL_up_ref(), X509_OBJECT_up_ref_count() methods are now returning an
2973 int (instead of void) like all others TYPE_up_ref() methods.
2974 So now these methods also check the return value of CRYPTO_atomic_add(),
2975 and the validity of object reference counter.
2977 *fdasilvayy@gmail.com*
2979 * With Windows Visual Studio builds, the .pdb files are installed
2980 alongside the installed libraries and executables. For a static
2981 library installation, ossl_static.pdb is the associate compiler
2982 generated .pdb file to be used when linking programs.
2986 * Remove openssl.spec. Packaging files belong with the packagers.
2990 * Automatic Darwin/OSX configuration has had a refresh, it will now
2991 recognise x86_64 architectures automatically. You can still decide
2992 to build for a different bitness with the environment variable
2993 KERNEL_BITS (can be 32 or 64), for example:
2995 KERNEL_BITS=32 ./config
2999 * Change default algorithms in pkcs8 utility to use PKCS#5 v2.0,
3000 256 bit AES and HMAC with SHA256.
3004 * Remove support for MIPS o32 ABI on IRIX (and IRIX only).
3008 * Triple-DES ciphers have been moved from HIGH to MEDIUM.
3012 * To enable users to have their own config files and build file templates,
3013 Configure looks in the directory indicated by the environment variable
3014 OPENSSL_LOCAL_CONFIG_DIR as well as the in-source Configurations/
3015 directory. On VMS, OPENSSL_LOCAL_CONFIG_DIR is expected to be a logical
3016 name and is used as is.
3020 * The following datatypes were made opaque: X509_OBJECT, X509_STORE_CTX,
3021 X509_STORE, X509_LOOKUP, and X509_LOOKUP_METHOD. The unused type
3022 X509_CERT_FILE_CTX was removed.
3026 * "shared" builds are now the default. To create only static libraries use
3027 the "no-shared" Configure option.
3031 * Remove the no-aes, no-hmac, no-rsa, no-sha and no-md5 Configure options.
3032 All of these option have not worked for some while and are fundamental
3037 * Make various cleanup routines no-ops and mark them as deprecated. Most
3038 global cleanup functions are no longer required because they are handled
3039 via auto-deinit (see OPENSSL_init_crypto and OPENSSL_init_ssl man pages).
3040 Explicitly de-initing can cause problems (e.g. where a library that uses
3041 OpenSSL de-inits, but an application is still using it). The affected
3042 functions are CONF_modules_free(), ENGINE_cleanup(), OBJ_cleanup(),
3043 EVP_cleanup(), BIO_sock_cleanup(), CRYPTO_cleanup_all_ex_data(),
3044 RAND_cleanup(), SSL_COMP_free_compression_methods(), ERR_free_strings() and
3045 COMP_zlib_cleanup().
3049 * --strict-warnings no longer enables runtime debugging options
3050 such as REF_DEBUG. Instead, debug options are automatically
3051 enabled with '--debug' builds.
3053 *Andy Polyakov, Emilia Käsper*
3055 * Made DH and DH_METHOD opaque. The structures for managing DH objects
3056 have been moved out of the public header files. New functions for managing
3057 these have been added.
3061 * Made RSA and RSA_METHOD opaque. The structures for managing RSA
3062 objects have been moved out of the public header files. New
3063 functions for managing these have been added.
3067 * Made DSA and DSA_METHOD opaque. The structures for managing DSA objects
3068 have been moved out of the public header files. New functions for managing
3069 these have been added.
3073 * Made BIO and BIO_METHOD opaque. The structures for managing BIOs have been
3074 moved out of the public header files. New functions for managing these
3079 * Removed no-rijndael as a config option. Rijndael is an old name for AES.
3083 * Removed the mk1mf build scripts.
3087 * Headers are now wrapped, if necessary, with OPENSSL_NO_xxx, so
3088 it is always safe to #include a header now.
3092 * Removed the aged BC-32 config and all its supporting scripts
3096 * Removed support for Ultrix, Netware, and OS/2.
3100 * Add support for HKDF.
3102 *Alessandro Ghedini*
3104 * Add support for blake2b and blake2s
3108 * Added support for "pipelining". Ciphers that have the
3109 EVP_CIPH_FLAG_PIPELINE flag set have a capability to process multiple
3110 encryptions/decryptions simultaneously. There are currently no built-in
3111 ciphers with this property but the expectation is that engines will be able
3112 to offer it to significantly improve throughput. Support has been extended
3113 into libssl so that multiple records for a single connection can be
3114 processed in one go (for >=TLS 1.1).
3118 * Added the AFALG engine. This is an async capable engine which is able to
3119 offload work to the Linux kernel. In this initial version it only supports
3120 AES128-CBC. The kernel must be version 4.1.0 or greater.
3124 * OpenSSL now uses a new threading API. It is no longer necessary to
3125 set locking callbacks to use OpenSSL in a multi-threaded environment. There
3126 are two supported threading models: pthreads and windows threads. It is
3127 also possible to configure OpenSSL at compile time for "no-threads". The
3128 old threading API should no longer be used. The functions have been
3129 replaced with "no-op" compatibility macros.
3131 *Alessandro Ghedini, Matt Caswell*
3133 * Modify behavior of ALPN to invoke callback after SNI/servername
3134 callback, such that updates to the SSL_CTX affect ALPN.
3138 * Add SSL_CIPHER queries for authentication and key-exchange.
3142 * Changes to the DEFAULT cipherlist:
3143 - Prefer (EC)DHE handshakes over plain RSA.
3144 - Prefer AEAD ciphers over legacy ciphers.
3145 - Prefer ECDSA over RSA when both certificates are available.
3146 - Prefer TLSv1.2 ciphers/PRF.
3147 - Remove DSS, SEED, IDEA, CAMELLIA, and AES-CCM from the
3152 * Change the ECC default curve list to be this, in order: x25519,
3153 secp256r1, secp521r1, secp384r1.
3157 * RC4 based libssl ciphersuites are now classed as "weak" ciphers and are
3158 disabled by default. They can be re-enabled using the
3159 enable-weak-ssl-ciphers option to Configure.
3163 * If the server has ALPN configured, but supports no protocols that the
3164 client advertises, send a fatal "no_application_protocol" alert.
3165 This behaviour is SHALL in RFC 7301, though it isn't universally
3166 implemented by other servers.
3170 * Add X25519 support.
3171 Add ASN.1 and EVP_PKEY methods for X25519. This includes support
3172 for public and private key encoding using the format documented in
3173 draft-ietf-curdle-pkix-02. The corresponding EVP_PKEY method supports
3174 key generation and key derivation.
3176 TLS support complies with draft-ietf-tls-rfc4492bis-08 and uses
3181 * Deprecate SRP_VBASE_get_by_user.
3182 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
3183 In order to fix an unavoidable memory leak ([CVE-2016-0798]),
3184 SRP_VBASE_get_by_user was changed to ignore the "fake user" SRP
3185 seed, even if the seed is configured.
3187 Users should use SRP_VBASE_get1_by_user instead. Note that in
3188 SRP_VBASE_get1_by_user, caller must free the returned value. Note
3189 also that even though configuring the SRP seed attempts to hide
3190 invalid usernames by continuing the handshake with fake
3191 credentials, this behaviour is not constant time and no strong
3192 guarantees are made that the handshake is indistinguishable from
3193 that of a valid user.
3197 * Configuration change; it's now possible to build dynamic engines
3198 without having to build shared libraries and vice versa. This
3199 only applies to the engines in `engines/`, those in `crypto/engine/`
3200 will always be built into libcrypto (i.e. "static").
3202 Building dynamic engines is enabled by default; to disable, use
3203 the configuration option "disable-dynamic-engine".
3205 The only requirements for building dynamic engines are the
3206 presence of the DSO module and building with position independent
3207 code, so they will also automatically be disabled if configuring
3208 with "disable-dso" or "disable-pic".
3210 The macros OPENSSL_NO_STATIC_ENGINE and OPENSSL_NO_DYNAMIC_ENGINE
3211 are also taken away from openssl/opensslconf.h, as they are
3216 * Configuration change; if there is a known flag to compile
3217 position independent code, it will always be applied on the
3218 libcrypto and libssl object files, and never on the application
3219 object files. This means other libraries that use routines from
3220 libcrypto / libssl can be made into shared libraries regardless
3221 of how OpenSSL was configured.
3223 If this isn't desirable, the configuration options "disable-pic"
3224 or "no-pic" can be used to disable the use of PIC. This will
3225 also disable building shared libraries and dynamic engines.
3229 * Removed JPAKE code. It was experimental and has no wide use.
3233 * The INSTALL_PREFIX Makefile variable has been renamed to
3234 DESTDIR. That makes for less confusion on what this variable
3235 is for. Also, the configuration option --install_prefix is
3240 * Heartbeat for TLS has been removed and is disabled by default
3241 for DTLS; configure with enable-heartbeats. Code that uses the
3242 old #define's might need to be updated.
3244 *Emilia Käsper, Rich Salz*
3246 * Rename REF_CHECK to REF_DEBUG.
3250 * New "unified" build system
3252 The "unified" build system is aimed to be a common system for all
3253 platforms we support. With it comes new support for VMS.
3255 This system builds supports building in a different directory tree
3256 than the source tree. It produces one Makefile (for unix family
3257 or lookalikes), or one descrip.mms (for VMS).
3259 The source of information to make the Makefile / descrip.mms is
3260 small files called 'build.info', holding the necessary
3261 information for each directory with source to compile, and a
3262 template in Configurations, like unix-Makefile.tmpl or
3265 With this change, the library names were also renamed on Windows
3266 and on VMS. They now have names that are closer to the standard
3267 on Unix, and include the major version number, and in certain
3268 cases, the architecture they are built for. See "Notes on shared
3269 libraries" in INSTALL.
3271 We rely heavily on the perl module Text::Template.
3275 * Added support for auto-initialisation and de-initialisation of the library.
3276 OpenSSL no longer requires explicit init or deinit routines to be called,
3277 except in certain circumstances. See the OPENSSL_init_crypto() and
3278 OPENSSL_init_ssl() man pages for further information.
3282 * The arguments to the DTLSv1_listen function have changed. Specifically the
3283 "peer" argument is now expected to be a BIO_ADDR object.
3285 * Rewrite of BIO networking library. The BIO library lacked consistent
3286 support of IPv6, and adding it required some more extensive
3287 modifications. This introduces the BIO_ADDR and BIO_ADDRINFO types,
3288 which hold all types of addresses and chains of address information.
3289 It also introduces a new API, with functions like BIO_socket,
3290 BIO_connect, BIO_listen, BIO_lookup and a rewrite of BIO_accept.
3291 The source/sink BIOs BIO_s_connect, BIO_s_accept and BIO_s_datagram
3292 have been adapted accordingly.
3296 * RSA_padding_check_PKCS1_type_1 now accepts inputs with and without
3301 * CRIME protection: disable compression by default, even if OpenSSL is
3302 compiled with zlib enabled. Applications can still enable compression
3303 by calling SSL_CTX_clear_options(ctx, SSL_OP_NO_COMPRESSION), or by
3304 using the SSL_CONF library to configure compression.
3308 * The signature of the session callback configured with
3309 SSL_CTX_sess_set_get_cb was changed. The read-only input buffer
3310 was explicitly marked as `const unsigned char*` instead of
3315 * Always DPURIFY. Remove the use of uninitialized memory in the
3316 RNG, and other conditional uses of DPURIFY. This makes -DPURIFY a no-op.
3320 * Removed many obsolete configuration items, including
3321 DES_PTR, DES_RISC1, DES_RISC2, DES_INT
3322 MD2_CHAR, MD2_INT, MD2_LONG
3324 IDEA_SHORT, IDEA_LONG
3325 RC2_SHORT, RC2_LONG, RC4_LONG, RC4_CHUNK, RC4_INDEX
3327 *Rich Salz, with advice from Andy Polyakov*
3329 * Many BN internals have been moved to an internal header file.
3331 *Rich Salz with help from Andy Polyakov*
3333 * Configuration and writing out the results from it has changed.
3334 Files such as Makefile include/openssl/opensslconf.h and are now
3335 produced through general templates, such as Makefile.in and
3336 crypto/opensslconf.h.in and some help from the perl module
3339 Also, the center of configuration information is no longer
3340 Makefile. Instead, Configure produces a perl module in
3341 configdata.pm which holds most of the config data (in the hash
3342 table %config), the target data that comes from the target
3343 configuration in one of the `Configurations/*.conf` files (in
3348 * To clarify their intended purposes, the Configure options
3349 --prefix and --openssldir change their semantics, and become more
3350 straightforward and less interdependent.
3352 --prefix shall be used exclusively to give the location INSTALLTOP
3353 where programs, scripts, libraries, include files and manuals are
3354 going to be installed. The default is now /usr/local.
3356 --openssldir shall be used exclusively to give the default
3357 location OPENSSLDIR where certificates, private keys, CRLs are
3358 managed. This is also where the default openssl.cnf gets
3360 If the directory given with this option is a relative path, the
3361 values of both the --prefix value and the --openssldir value will
3362 be combined to become OPENSSLDIR.
3363 The default for --openssldir is INSTALLTOP/ssl.
3365 Anyone who uses --openssldir to specify where OpenSSL is to be
3366 installed MUST change to use --prefix instead.
3370 * The GOST engine was out of date and therefore it has been removed. An up
3371 to date GOST engine is now being maintained in an external repository.
3372 See: <https://wiki.openssl.org/index.php/Binaries>. Libssl still retains
3373 support for GOST ciphersuites (these are only activated if a GOST engine
3378 * EGD is no longer supported by default; use enable-egd when
3381 *Ben Kaduk and Rich Salz*
3383 * The distribution now has Makefile.in files, which are used to
3384 create Makefile's when Configure is run. *Configure must be run
3385 before trying to build now.*
3389 * The return value for SSL_CIPHER_description() for error conditions
3394 * Support for RFC6698/RFC7671 DANE TLSA peer authentication.
3396 Obtaining and performing DNSSEC validation of TLSA records is
3397 the application's responsibility. The application provides
3398 the TLSA records of its choice to OpenSSL, and these are then
3399 used to authenticate the peer.
3401 The TLSA records need not even come from DNS. They can, for
3402 example, be used to implement local end-entity certificate or
3403 trust-anchor "pinning", where the "pin" data takes the form
3404 of TLSA records, which can augment or replace verification
3405 based on the usual WebPKI public certification authorities.
3409 * Revert default OPENSSL_NO_DEPRECATED setting. Instead OpenSSL
3410 continues to support deprecated interfaces in default builds.
3411 However, applications are strongly advised to compile their
3412 source files with -DOPENSSL_API_COMPAT=0x10100000L, which hides
3413 the declarations of all interfaces deprecated in 0.9.8, 1.0.0
3414 or the 1.1.0 releases.
3416 In environments in which all applications have been ported to
3417 not use any deprecated interfaces OpenSSL's Configure script
3418 should be used with the --api=1.1.0 option to entirely remove
3419 support for the deprecated features from the library and
3420 unconditionally disable them in the installed headers.
3421 Essentially the same effect can be achieved with the "no-deprecated"
3422 argument to Configure, except that this will always restrict
3423 the build to just the latest API, rather than a fixed API
3426 As applications are ported to future revisions of the API,
3427 they should update their compile-time OPENSSL_API_COMPAT define
3428 accordingly, but in most cases should be able to continue to
3429 compile with later releases.
3431 The OPENSSL_API_COMPAT versions for 1.0.0, and 0.9.8 are
3432 0x10000000L and 0x00908000L, respectively. However those
3433 versions did not support the OPENSSL_API_COMPAT feature, and
3434 so applications are not typically tested for explicit support
3435 of just the undeprecated features of either release.
3439 * Add support for setting the minimum and maximum supported protocol.
3440 It can bet set via the SSL_set_min_proto_version() and
3441 SSL_set_max_proto_version(), or via the SSL_CONF's MinProtocol and
3442 MaxProtocol. It's recommended to use the new APIs to disable
3443 protocols instead of disabling individual protocols using
3444 SSL_set_options() or SSL_CONF's Protocol. This change also
3445 removes support for disabling TLS 1.2 in the OpenSSL TLS
3446 client at compile time by defining OPENSSL_NO_TLS1_2_CLIENT.
3450 * Support for ChaCha20 and Poly1305 added to libcrypto and libssl.
3454 * New EC_KEY_METHOD, this replaces the older ECDSA_METHOD and ECDH_METHOD
3455 and integrates ECDSA and ECDH functionality into EC. Implementations can
3456 now redirect key generation and no longer need to convert to or from
3459 Note: the ecdsa.h and ecdh.h headers are now no longer needed and just
3460 include the ec.h header file instead.
3464 * Remove support for all 40 and 56 bit ciphers. This includes all the export
3465 ciphers who are no longer supported and drops support the ephemeral RSA key
3466 exchange. The LOW ciphers currently doesn't have any ciphers in it.
3470 * Made EVP_MD_CTX, EVP_MD, EVP_CIPHER_CTX, EVP_CIPHER and HMAC_CTX
3471 opaque. For HMAC_CTX, the following constructors and destructors
3474 HMAC_CTX *HMAC_CTX_new(void);
3475 void HMAC_CTX_free(HMAC_CTX *ctx);
3477 For EVP_MD and EVP_CIPHER, complete APIs to create, fill and
3478 destroy such methods has been added. See EVP_MD_meth_new(3) and
3479 EVP_CIPHER_meth_new(3) for documentation.
3482 1) `EVP_MD_CTX_cleanup()`, `EVP_CIPHER_CTX_cleanup()` and
3483 `HMAC_CTX_cleanup()` were removed. `HMAC_CTX_reset()` and
3484 `EVP_MD_CTX_reset()` should be called instead to reinitialise
3485 an already created structure.
3486 2) For consistency with the majority of our object creators and
3487 destructors, `EVP_MD_CTX_(create|destroy)` were renamed to
3488 `EVP_MD_CTX_(new|free)`. The old names are retained as macros
3489 for deprecated builds.
3493 * Added ASYNC support. Libcrypto now includes the async sub-library to enable
3494 cryptographic operations to be performed asynchronously as long as an
3495 asynchronous capable engine is used. See the ASYNC_start_job() man page for
3496 further details. Libssl has also had this capability integrated with the
3497 introduction of the new mode SSL_MODE_ASYNC and associated error
3498 SSL_ERROR_WANT_ASYNC. See the SSL_CTX_set_mode() and SSL_get_error() man
3499 pages. This work was developed in partnership with Intel Corp.
3503 * SSL_{CTX_}set_ecdh_auto() has been removed and ECDH is support is
3504 always enabled now. If you want to disable the support you should
3505 exclude it using the list of supported ciphers. This also means that the
3506 "-no_ecdhe" option has been removed from s_server.
3510 * SSL_{CTX}_set_tmp_ecdh() which can set 1 EC curve now internally calls
3511 SSL_{CTX_}set1_curves() which can set a list.
3515 * Remove support for SSL_{CTX_}set_tmp_ecdh_callback(). You should set the
3516 curve you want to support using SSL_{CTX_}set1_curves().
3520 * State machine rewrite. The state machine code has been significantly
3521 refactored in order to remove much duplication of code and solve issues
3522 with the old code (see [ssl/statem/README.md](ssl/statem/README.md) for
3523 further details). This change does have some associated API changes.
3524 Notably the SSL_state() function has been removed and replaced by
3525 SSL_get_state which now returns an "OSSL_HANDSHAKE_STATE" instead of an int.
3526 SSL_set_state() has been removed altogether. The previous handshake states
3527 defined in ssl.h and ssl3.h have also been removed.
3531 * All instances of the string "ssleay" in the public API were replaced
3532 with OpenSSL (case-matching; e.g., OPENSSL_VERSION for #define's)
3533 Some error codes related to internal RSA_eay API's were renamed.
3537 * The demo files in crypto/threads were moved to demo/threads.
3541 * Removed obsolete engines: 4758cca, aep, atalla, cswift, nuron, gmp,
3544 *Matt Caswell, Rich Salz*
3546 * New ASN.1 embed macro.
3548 New ASN.1 macro ASN1_EMBED. This is the same as ASN1_SIMPLE except the
3549 structure is not allocated: it is part of the parent. That is instead of
3557 This reduces memory fragmentation and make it impossible to accidentally
3558 set a mandatory field to NULL.
3560 This currently only works for some fields specifically a SEQUENCE, CHOICE,
3561 or ASN1_STRING type which is part of a parent SEQUENCE. Since it is
3562 equivalent to ASN1_SIMPLE it cannot be tagged, OPTIONAL, SET OF or
3567 * Remove EVP_CHECK_DES_KEY, a compile-time option that never compiled.
3571 * Removed DES and RC4 ciphersuites from DEFAULT. Also removed RC2 although
3572 in 1.0.2 EXPORT was already removed and the only RC2 ciphersuite is also
3573 an EXPORT one. COMPLEMENTOFDEFAULT has been updated accordingly to add
3574 DES and RC4 ciphersuites.
3578 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
3579 This changes the decoding behaviour for some invalid messages,
3580 though the change is mostly in the more lenient direction, and
3581 legacy behaviour is preserved as much as possible.
3585 * Fix no-stdio build.
3586 *David Woodhouse <David.Woodhouse@intel.com> and also*
3587 *Ivan Nestlerode <ivan.nestlerode@sonos.com>*
3589 * New testing framework
3590 The testing framework has been largely rewritten and is now using
3591 perl and the perl modules Test::Harness and an extended variant of
3592 Test::More called OpenSSL::Test to do its work. All test scripts in
3593 test/ have been rewritten into test recipes, and all direct calls to
3594 executables in test/Makefile have become individual recipes using the
3595 simplified testing OpenSSL::Test::Simple.
3597 For documentation on our testing modules, do:
3599 perldoc test/testlib/OpenSSL/Test/Simple.pm
3600 perldoc test/testlib/OpenSSL/Test.pm
3604 * Revamped memory debug; only -DCRYPTO_MDEBUG and -DCRYPTO_MDEBUG_ABORT
3605 are used; the latter aborts on memory leaks (usually checked on exit).
3606 Some undocumented "set malloc, etc., hooks" functions were removed
3607 and others were changed. All are now documented.
3611 * In DSA_generate_parameters_ex, if the provided seed is too short,
3614 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
3616 * Rewrite PSK to support ECDHE_PSK, DHE_PSK and RSA_PSK. Add ciphersuites
3617 from RFC4279, RFC4785, RFC5487, RFC5489.
3619 Thanks to Christian J. Dietrich and Giuseppe D'Angelo for the
3620 original RSA_PSK patch.
3624 * Dropped support for the SSL3_FLAGS_DELAY_CLIENT_FINISHED flag. This SSLeay
3625 era flag was never set throughout the codebase (only read). Also removed
3626 SSL3_FLAGS_POP_BUFFER which was only used if
3627 SSL3_FLAGS_DELAY_CLIENT_FINISHED was also set.
3631 * Changed the default name options in the "ca", "crl", "req" and "x509"
3632 to be "oneline" instead of "compat".
3636 * Remove SSL_OP_TLS_BLOCK_PADDING_BUG. This is SSLeay legacy, we're
3637 not aware of clients that still exhibit this bug, and the workaround
3638 hasn't been working properly for a while.
3642 * The return type of BIO_number_read() and BIO_number_written() as well as
3643 the corresponding num_read and num_write members in the BIO structure has
3644 changed from unsigned long to uint64_t. On platforms where an unsigned
3645 long is 32 bits (e.g. Windows) these counters could overflow if >4Gb is
3650 * Given the pervasive nature of TLS extensions it is inadvisable to run
3651 OpenSSL without support for them. It also means that maintaining
3652 the OPENSSL_NO_TLSEXT option within the code is very invasive (and probably
3653 not well tested). Therefore the OPENSSL_NO_TLSEXT option has been removed.
3657 * Removed support for the two export grade static DH ciphersuites
3658 EXP-DH-RSA-DES-CBC-SHA and EXP-DH-DSS-DES-CBC-SHA. These two ciphersuites
3659 were newly added (along with a number of other static DH ciphersuites) to
3660 1.0.2. However the two export ones have *never* worked since they were
3661 introduced. It seems strange in any case to be adding new export
3662 ciphersuites, and given "logjam" it also does not seem correct to fix them.
3666 * Version negotiation has been rewritten. In particular SSLv23_method(),
3667 SSLv23_client_method() and SSLv23_server_method() have been deprecated,
3668 and turned into macros which simply call the new preferred function names
3669 TLS_method(), TLS_client_method() and TLS_server_method(). All new code
3670 should use the new names instead. Also as part of this change the ssl23.h
3671 header file has been removed.
3675 * Support for Kerberos ciphersuites in TLS (RFC2712) has been removed. This
3676 code and the associated standard is no longer considered fit-for-purpose.
3680 * RT2547 was closed. When generating a private key, try to make the
3681 output file readable only by the owner. This behavior change might
3682 be noticeable when interacting with other software.
3684 * Documented all exdata functions. Added CRYPTO_free_ex_index.
3689 * Added HTTP GET support to the ocsp command.
3693 * Changed default digest for the dgst and enc commands from MD5 to
3698 * RAND_pseudo_bytes has been deprecated. Users should use RAND_bytes instead.
3702 * Added support for TLS extended master secret from
3703 draft-ietf-tls-session-hash-03.txt. Thanks for Alfredo Pironti for an
3704 initial patch which was a great help during development.
3708 * All libssl internal structures have been removed from the public header
3709 files, and the OPENSSL_NO_SSL_INTERN option has been removed (since it is
3710 now redundant). Users should not attempt to access internal structures
3711 directly. Instead they should use the provided API functions.
3715 * config has been changed so that by default OPENSSL_NO_DEPRECATED is used.
3716 Access to deprecated functions can be re-enabled by running config with
3717 "enable-deprecated". In addition applications wishing to use deprecated
3718 functions must define OPENSSL_USE_DEPRECATED. Note that this new behaviour
3719 will, by default, disable some transitive includes that previously existed
3720 in the header files (e.g. ec.h will no longer, by default, include bn.h)
3724 * Added support for OCB mode. OpenSSL has been granted a patent license
3725 compatible with the OpenSSL license for use of OCB. Details are available
3726 at <https://www.openssl.org/source/OCB-patent-grant-OpenSSL.pdf>. Support
3727 for OCB can be removed by calling config with no-ocb.
3731 * SSLv2 support has been removed. It still supports receiving a SSLv2
3732 compatible client hello.
3736 * Increased the minimal RSA keysize from 256 to 512 bits [Rich Salz],
3737 done while fixing the error code for the key-too-small case.
3739 *Annie Yousar <a.yousar@informatik.hu-berlin.de>*
3741 * CA.sh has been removed; use CA.pl instead.
3745 * Removed old DES API.
3749 * Remove various unsupported platforms:
3755 Sinix/ReliantUNIX RM400
3760 16-bit platforms such as WIN16
3764 * Clean up OPENSSL_NO_xxx #define's
3765 - Use setbuf() and remove OPENSSL_NO_SETVBUF_IONBF
3766 - Rename OPENSSL_SYSNAME_xxx to OPENSSL_SYS_xxx
3767 - OPENSSL_NO_EC{DH,DSA} merged into OPENSSL_NO_EC
3768 - OPENSSL_NO_RIPEMD160, OPENSSL_NO_RIPEMD merged into OPENSSL_NO_RMD160
3769 - OPENSSL_NO_FP_API merged into OPENSSL_NO_STDIO
3770 - Remove OPENSSL_NO_BIO OPENSSL_NO_BUFFER OPENSSL_NO_CHAIN_VERIFY
3771 OPENSSL_NO_EVP OPENSSL_NO_FIPS_ERR OPENSSL_NO_HASH_COMP
3772 OPENSSL_NO_LHASH OPENSSL_NO_OBJECT OPENSSL_NO_SPEED OPENSSL_NO_STACK
3773 OPENSSL_NO_X509 OPENSSL_NO_X509_VERIFY
3774 - Remove MS_STATIC; it's a relic from platforms <32 bits.
3778 * Cleaned up dead code
3779 Remove all but one '#ifdef undef' which is to be looked at.
3783 * Clean up calling of xxx_free routines.
3784 Just like free(), fix most of the xxx_free routines to accept
3785 NULL. Remove the non-null checks from callers. Save much code.
3789 * Add secure heap for storage of private keys (when possible).
3790 Add BIO_s_secmem(), CBIGNUM, etc.
3791 Contributed by Akamai Technologies under our Corporate CLA.
3795 * Experimental support for a new, fast, unbiased prime candidate generator,
3796 bn_probable_prime_dh_coprime(). Not currently used by any prime generator.
3798 *Felix Laurie von Massenbach <felix@erbridge.co.uk>*
3800 * New output format NSS in the sess_id command line tool. This allows
3801 exporting the session id and the master key in NSS keylog format.
3803 *Martin Kaiser <martin@kaiser.cx>*
3805 * Harmonize version and its documentation. -f flag is used to display
3808 *mancha <mancha1@zoho.com>*
3810 * Fix eckey_priv_encode so it immediately returns an error upon a failure
3811 in i2d_ECPrivateKey. Thanks to Ted Unangst for feedback on this issue.
3813 *mancha <mancha1@zoho.com>*
3815 * Fix some double frees. These are not thought to be exploitable.
3817 *mancha <mancha1@zoho.com>*
3819 * A missing bounds check in the handling of the TLS heartbeat extension
3820 can be used to reveal up to 64k of memory to a connected client or
3823 Thanks for Neel Mehta of Google Security for discovering this bug and to
3824 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
3825 preparing the fix ([CVE-2014-0160])
3827 *Adam Langley, Bodo Moeller*
3829 * Fix for the attack described in the paper "Recovering OpenSSL
3830 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
3831 by Yuval Yarom and Naomi Benger. Details can be obtained from:
3832 <http://eprint.iacr.org/2014/140>
3834 Thanks to Yuval Yarom and Naomi Benger for discovering this
3835 flaw and to Yuval Yarom for supplying a fix ([CVE-2014-0076])
3837 *Yuval Yarom and Naomi Benger*
3839 * Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
3840 this fixes a limitation in previous versions of OpenSSL.
3844 * Experimental encrypt-then-mac support.
3846 Experimental support for encrypt then mac from
3847 draft-gutmann-tls-encrypt-then-mac-02.txt
3849 To enable it set the appropriate extension number (0x42 for the test
3850 server) using e.g. -DTLSEXT_TYPE_encrypt_then_mac=0x42
3852 For non-compliant peers (i.e. just about everything) this should have no
3855 WARNING: EXPERIMENTAL, SUBJECT TO CHANGE.
3859 * Add EVP support for key wrapping algorithms, to avoid problems with
3860 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
3861 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
3862 algorithms and include tests cases.
3866 * Extend CMS code to support RSA-PSS signatures and RSA-OAEP for
3871 * Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
3872 MGF1 digest and OAEP label.
3876 * Make openssl verify return errors.
3878 *Chris Palmer <palmer@google.com> and Ben Laurie*
3880 * New function ASN1_TIME_diff to calculate the difference between two
3881 ASN1_TIME structures or one structure and the current time.
3885 * Update fips_test_suite to support multiple command line options. New
3886 test to induce all self test errors in sequence and check expected
3891 * Add FIPS_{rsa,dsa,ecdsa}_{sign,verify} functions which digest and
3892 sign or verify all in one operation.
3896 * Add fips_algvs: a multicall fips utility incorporating all the algorithm
3897 test programs and fips_test_suite. Includes functionality to parse
3898 the minimal script output of fipsalgest.pl directly.
3902 * Add authorisation parameter to FIPS_module_mode_set().
3906 * Add FIPS selftest for ECDH algorithm using P-224 and B-233 curves.
3910 * Use separate DRBG fields for internal and external flags. New function
3911 FIPS_drbg_health_check() to perform on demand health checking. Add
3912 generation tests to fips_test_suite with reduced health check interval to
3913 demonstrate periodic health checking. Add "nodh" option to
3914 fips_test_suite to skip very slow DH test.
3918 * New function FIPS_get_cipherbynid() to lookup FIPS supported ciphers
3923 * More extensive health check for DRBG checking many more failure modes.
3924 New function FIPS_selftest_drbg_all() to handle every possible DRBG
3925 combination: call this in fips_test_suite.
3929 * Add support for canonical generation of DSA parameter 'g'. See
3932 * Add support for HMAC DRBG from SP800-90. Update DRBG algorithm test and
3933 POST to handle HMAC cases.
3937 * Add functions FIPS_module_version() and FIPS_module_version_text()
3938 to return numerical and string versions of the FIPS module number.
3942 * Rename FIPS_mode_set and FIPS_mode to FIPS_module_mode_set and
3943 FIPS_module_mode. FIPS_mode and FIPS_mode_set will be implemented
3944 outside the validated module in the FIPS capable OpenSSL.
3948 * Minor change to DRBG entropy callback semantics. In some cases
3949 there is no multiple of the block length between min_len and
3950 max_len. Allow the callback to return more than max_len bytes
3951 of entropy but discard any extra: it is the callback's responsibility
3952 to ensure that the extra data discarded does not impact the
3953 requested amount of entropy.
3957 * Add PRNG security strength checks to RSA, DSA and ECDSA using
3958 information in FIPS186-3, SP800-57 and SP800-131A.
3962 * CCM support via EVP. Interface is very similar to GCM case except we
3963 must supply all data in one chunk (i.e. no update, final) and the
3964 message length must be supplied if AAD is used. Add algorithm test
3969 * Initial version of POST overhaul. Add POST callback to allow the status
3970 of POST to be monitored and/or failures induced. Modify fips_test_suite
3971 to use callback. Always run all selftests even if one fails.
3975 * XTS support including algorithm test driver in the fips_gcmtest program.
3976 Note: this does increase the maximum key length from 32 to 64 bytes but
3977 there should be no binary compatibility issues as existing applications
3978 will never use XTS mode.
3982 * Extensive reorganisation of FIPS PRNG behaviour. Remove all dependencies
3983 to OpenSSL RAND code and replace with a tiny FIPS RAND API which also
3984 performs algorithm blocking for unapproved PRNG types. Also do not
3985 set PRNG type in FIPS_mode_set(): leave this to the application.
3986 Add default OpenSSL DRBG handling: sets up FIPS PRNG and seeds with
3987 the standard OpenSSL PRNG: set additional data to a date time vector.
3991 * Rename old X9.31 PRNG functions of the form `FIPS_rand*` to `FIPS_x931*`.
3992 This shouldn't present any incompatibility problems because applications
3993 shouldn't be using these directly and any that are will need to rethink
3994 anyway as the X9.31 PRNG is now deprecated by FIPS 140-2
3998 * Extensive self tests and health checking required by SP800-90 DRBG.
3999 Remove strength parameter from FIPS_drbg_instantiate and always
4000 instantiate at maximum supported strength.
4004 * Add ECDH code to fips module and fips_ecdhvs for primitives only testing.
4008 * New algorithm test program fips_dhvs to handle DH primitives only testing.
4012 * New function DH_compute_key_padded() to compute a DH key and pad with
4013 leading zeroes if needed: this complies with SP800-56A et al.
4017 * Initial implementation of SP800-90 DRBGs for Hash and CTR. Not used by
4018 anything, incomplete, subject to change and largely untested at present.
4022 * Modify fipscanisteronly build option to only build the necessary object
4023 files by filtering FIPS_EX_OBJ through a perl script in crypto/Makefile.
4027 * Add experimental option FIPSSYMS to give all symbols in
4028 fipscanister.o and FIPS or fips prefix. This will avoid
4029 conflicts with future versions of OpenSSL. Add perl script
4030 util/fipsas.pl to preprocess assembly language source files
4031 and rename any affected symbols.
4035 * Add selftest checks and algorithm block of non-fips algorithms in
4036 FIPS mode. Remove DES2 from selftests.
4040 * Add ECDSA code to fips module. Add tiny fips_ecdsa_check to just
4041 return internal method without any ENGINE dependencies. Add new
4042 tiny fips sign and verify functions.
4046 * New build option no-ec2m to disable characteristic 2 code.
4050 * New build option "fipscanisteronly". This only builds fipscanister.o
4051 and (currently) associated fips utilities. Uses the file Makefile.fips
4052 instead of Makefile.org as the prototype.
4056 * Add some FIPS mode restrictions to GCM. Add internal IV generator.
4057 Update fips_gcmtest to use IV generator.
4061 * Initial, experimental EVP support for AES-GCM. AAD can be input by
4062 setting output buffer to NULL. The `*Final` function must be
4063 called although it will not retrieve any additional data. The tag
4064 can be set or retrieved with a ctrl. The IV length is by default 12
4065 bytes (96 bits) but can be set to an alternative value. If the IV
4066 length exceeds the maximum IV length (currently 16 bytes) it cannot be
4071 * New flag in ciphers: EVP_CIPH_FLAG_CUSTOM_CIPHER. This means the
4072 underlying do_cipher function handles all cipher semantics itself
4073 including padding and finalisation. This is useful if (for example)
4074 an ENGINE cipher handles block padding itself. The behaviour of
4075 do_cipher is subtly changed if this flag is set: the return value
4076 is the number of characters written to the output buffer (zero is
4077 no longer an error code) or a negative error code. Also if the
4078 input buffer is NULL and length 0 finalisation should be performed.
4082 * If a candidate issuer certificate is already part of the constructed
4083 path ignore it: new debug notification X509_V_ERR_PATH_LOOP for this case.
4087 * Improve forward-security support: add functions
4089 void SSL_CTX_set_not_resumable_session_callback(
4090 SSL_CTX *ctx, int (*cb)(SSL *ssl, int is_forward_secure))
4091 void SSL_set_not_resumable_session_callback(
4092 SSL *ssl, int (*cb)(SSL *ssl, int is_forward_secure))
4094 for use by SSL/TLS servers; the callback function will be called whenever a
4095 new session is created, and gets to decide whether the session may be
4096 cached to make it resumable (return 0) or not (return 1). (As by the
4097 SSL/TLS protocol specifications, the session_id sent by the server will be
4098 empty to indicate that the session is not resumable; also, the server will
4099 not generate RFC 4507 (RFC 5077) session tickets.)
4101 A simple reasonable callback implementation is to return is_forward_secure.
4102 This parameter will be set to 1 or 0 depending on the ciphersuite selected
4103 by the SSL/TLS server library, indicating whether it can provide forward
4106 *Emilia Käsper <emilia.kasper@esat.kuleuven.be> (Google)*
4108 * New -verify_name option in command line utilities to set verification
4113 * Initial CMAC implementation. WARNING: EXPERIMENTAL, API MAY CHANGE.
4114 Add CMAC pkey methods.
4118 * Experimental renegotiation in s_server -www mode. If the client
4119 browses /reneg connection is renegotiated. If /renegcert it is
4120 renegotiated requesting a certificate.
4124 * Add an "external" session cache for debugging purposes to s_server. This
4125 should help trace issues which normally are only apparent in deployed
4126 multi-process servers.
4130 * Extensive audit of libcrypto with DEBUG_UNUSED. Fix many cases where
4131 return value is ignored. NB. The functions RAND_add(), RAND_seed(),
4132 BIO_set_cipher() and some obscure PEM functions were changed so they
4133 can now return an error. The RAND changes required a change to the
4134 RAND_METHOD structure.
4138 * New macro `__owur` for "OpenSSL Warn Unused Result". This makes use of
4139 a gcc attribute to warn if the result of a function is ignored. This
4140 is enable if DEBUG_UNUSED is set. Add to several functions in evp.h
4141 whose return value is often ignored.
4145 * New -noct, -requestct, -requirect and -ctlogfile options for s_client.
4146 These allow SCTs (signed certificate timestamps) to be requested and
4147 validated when establishing a connection.
4149 *Rob Percival <robpercival@google.com>*
4154 ### Changes between 1.0.2s and 1.0.2t [10 Sep 2019]
4156 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
4157 used even when parsing explicit parameters, when loading a encoded key
4158 or calling `EC_GROUP_new_from_ecpkparameters()`/
4159 `EC_GROUP_new_from_ecparameters()`.
4160 This prevents bypass of security hardening and performance gains,
4161 especially for curves with specialized EC_METHODs.
4162 By default, if a key encoded with explicit parameters is loaded and later
4163 encoded, the output is still encoded with explicit parameters, even if
4164 internally a "named" EC_GROUP is used for computation.
4168 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
4169 this change, EC_GROUP_set_generator would accept order and/or cofactor as
4170 NULL. After this change, only the cofactor parameter can be NULL. It also
4171 does some minimal sanity checks on the passed order.
4176 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
4177 An attack is simple, if the first CMS_recipientInfo is valid but the
4178 second CMS_recipientInfo is chosen ciphertext. If the second
4179 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
4180 encryption key will be replaced by garbage, and the message cannot be
4181 decoded, but if the RSA decryption fails, the correct encryption key is
4182 used and the recipient will not notice the attack.
4183 As a work around for this potential attack the length of the decrypted
4184 key must be equal to the cipher default key length, in case the
4185 certifiate is not given and all recipientInfo are tried out.
4186 The old behaviour can be re-enabled in the CMS code by setting the
4187 CMS_DEBUG_DECRYPT flag.
4192 * Document issue with installation paths in diverse Windows builds
4194 '/usr/local/ssl' is an unsafe prefix for location to install OpenSSL
4195 binaries and run-time config file.
4200 ### Changes between 1.0.2r and 1.0.2s [28 May 2019]
4202 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
4203 This changes the size when using the `genpkey` command when no size is given.
4204 It fixes an omission in earlier changes that changed all RSA, DSA and DH
4205 generation commands to use 2048 bits by default.
4209 * Add FIPS support for Android Arm 64-bit
4211 Support for Android Arm 64-bit was added to the OpenSSL FIPS Object
4212 Module in Version 2.0.10. For some reason, the corresponding target
4213 'android64-aarch64' was missing OpenSSL 1.0.2, whence it could not be
4214 built with FIPS support on Android Arm 64-bit. This omission has been
4217 *Matthias St. Pierre*
4219 ### Changes between 1.0.2q and 1.0.2r [26 Feb 2019]
4221 * 0-byte record padding oracle
4223 If an application encounters a fatal protocol error and then calls
4224 SSL_shutdown() twice (once to send a close_notify, and once to receive one)
4225 then OpenSSL can respond differently to the calling application if a 0 byte
4226 record is received with invalid padding compared to if a 0 byte record is
4227 received with an invalid MAC. If the application then behaves differently
4228 based on that in a way that is detectable to the remote peer, then this
4229 amounts to a padding oracle that could be used to decrypt data.
4231 In order for this to be exploitable "non-stitched" ciphersuites must be in
4232 use. Stitched ciphersuites are optimised implementations of certain
4233 commonly used ciphersuites. Also the application must call SSL_shutdown()
4234 twice even if a protocol error has occurred (applications should not do
4235 this but some do anyway).
4237 This issue was discovered by Juraj Somorovsky, Robert Merget and Nimrod
4238 Aviram, with additional investigation by Steven Collison and Andrew
4239 Hourselt. It was reported to OpenSSL on 10th December 2018.
4244 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
4248 ### Changes between 1.0.2p and 1.0.2q [20 Nov 2018]
4250 * Microarchitecture timing vulnerability in ECC scalar multiplication
4252 OpenSSL ECC scalar multiplication, used in e.g. ECDSA and ECDH, has been
4253 shown to be vulnerable to a microarchitecture timing side channel attack.
4254 An attacker with sufficient access to mount local timing attacks during
4255 ECDSA signature generation could recover the private key.
4257 This issue was reported to OpenSSL on 26th October 2018 by Alejandro
4258 Cabrera Aldaya, Billy Brumley, Sohaib ul Hassan, Cesar Pereida Garcia and
4264 * Timing vulnerability in DSA signature generation
4266 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
4267 timing side channel attack. An attacker could use variations in the signing
4268 algorithm to recover the private key.
4270 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
4275 * Resolve a compatibility issue in EC_GROUP handling with the FIPS Object
4276 Module, accidentally introduced while backporting security fixes from the
4277 development branch and hindering the use of ECC in FIPS mode.
4281 ### Changes between 1.0.2o and 1.0.2p [14 Aug 2018]
4283 * Client DoS due to large DH parameter
4285 During key agreement in a TLS handshake using a DH(E) based ciphersuite a
4286 malicious server can send a very large prime value to the client. This will
4287 cause the client to spend an unreasonably long period of time generating a
4288 key for this prime resulting in a hang until the client has finished. This
4289 could be exploited in a Denial Of Service attack.
4291 This issue was reported to OpenSSL on 5th June 2018 by Guido Vranken
4296 * Cache timing vulnerability in RSA Key Generation
4298 The OpenSSL RSA Key generation algorithm has been shown to be vulnerable to
4299 a cache timing side channel attack. An attacker with sufficient access to
4300 mount cache timing attacks during the RSA key generation process could
4301 recover the private key.
4303 This issue was reported to OpenSSL on 4th April 2018 by Alejandro Cabrera
4304 Aldaya, Billy Brumley, Cesar Pereida Garcia and Luis Manuel Alvarez Tapia.
4309 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
4310 parameter is no longer accepted, as it leads to a corrupt table. NULL
4311 pem_str is reserved for alias entries only.
4315 * Revert blinding in ECDSA sign and instead make problematic addition
4316 length-invariant. Switch even to fixed-length Montgomery multiplication.
4320 * Change generating and checking of primes so that the error rate of not
4321 being prime depends on the intended use based on the size of the input.
4322 For larger primes this will result in more rounds of Miller-Rabin.
4323 The maximal error rate for primes with more than 1080 bits is lowered
4326 *Kurt Roeckx, Annie Yousar*
4328 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
4332 * Add blinding to ECDSA and DSA signatures to protect against side channel
4333 attacks discovered by Keegan Ryan (NCC Group).
4337 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
4338 now allow empty (zero character) pass phrases.
4342 * Certificate time validation (X509_cmp_time) enforces stricter
4343 compliance with RFC 5280. Fractional seconds and timezone offsets
4344 are no longer allowed.
4348 ### Changes between 1.0.2n and 1.0.2o [27 Mar 2018]
4350 * Constructed ASN.1 types with a recursive definition could exceed the stack
4352 Constructed ASN.1 types with a recursive definition (such as can be found
4353 in PKCS7) could eventually exceed the stack given malicious input with
4354 excessive recursion. This could result in a Denial Of Service attack. There
4355 are no such structures used within SSL/TLS that come from untrusted sources
4356 so this is considered safe.
4358 This issue was reported to OpenSSL on 4th January 2018 by the OSS-fuzz
4364 ### Changes between 1.0.2m and 1.0.2n [7 Dec 2017]
4366 * Read/write after SSL object in error state
4368 OpenSSL 1.0.2 (starting from version 1.0.2b) introduced an "error state"
4369 mechanism. The intent was that if a fatal error occurred during a handshake
4370 then OpenSSL would move into the error state and would immediately fail if
4371 you attempted to continue the handshake. This works as designed for the
4372 explicit handshake functions (SSL_do_handshake(), SSL_accept() and
4373 SSL_connect()), however due to a bug it does not work correctly if
4374 SSL_read() or SSL_write() is called directly. In that scenario, if the
4375 handshake fails then a fatal error will be returned in the initial function
4376 call. If SSL_read()/SSL_write() is subsequently called by the application
4377 for the same SSL object then it will succeed and the data is passed without
4378 being decrypted/encrypted directly from the SSL/TLS record layer.
4380 In order to exploit this issue an application bug would have to be present
4381 that resulted in a call to SSL_read()/SSL_write() being issued after having
4382 already received a fatal error.
4384 This issue was reported to OpenSSL by David Benjamin (Google).
4389 * rsaz_1024_mul_avx2 overflow bug on x86_64
4391 There is an overflow bug in the AVX2 Montgomery multiplication procedure
4392 used in exponentiation with 1024-bit moduli. No EC algorithms are affected.
4393 Analysis suggests that attacks against RSA and DSA as a result of this
4394 defect would be very difficult to perform and are not believed likely.
4395 Attacks against DH1024 are considered just feasible, because most of the
4396 work necessary to deduce information about a private key may be performed
4397 offline. The amount of resources required for such an attack would be
4398 significant. However, for an attack on TLS to be meaningful, the server
4399 would have to share the DH1024 private key among multiple clients, which is
4400 no longer an option since CVE-2016-0701.
4402 This only affects processors that support the AVX2 but not ADX extensions
4403 like Intel Haswell (4th generation).
4405 This issue was reported to OpenSSL by David Benjamin (Google). The issue
4406 was originally found via the OSS-Fuzz project.
4411 ### Changes between 1.0.2l and 1.0.2m [2 Nov 2017]
4413 * bn_sqrx8x_internal carry bug on x86_64
4415 There is a carry propagating bug in the x86_64 Montgomery squaring
4416 procedure. No EC algorithms are affected. Analysis suggests that attacks
4417 against RSA and DSA as a result of this defect would be very difficult to
4418 perform and are not believed likely. Attacks against DH are considered just
4419 feasible (although very difficult) because most of the work necessary to
4420 deduce information about a private key may be performed offline. The amount
4421 of resources required for such an attack would be very significant and
4422 likely only accessible to a limited number of attackers. An attacker would
4423 additionally need online access to an unpatched system using the target
4424 private key in a scenario with persistent DH parameters and a private
4425 key that is shared between multiple clients.
4427 This only affects processors that support the BMI1, BMI2 and ADX extensions
4428 like Intel Broadwell (5th generation) and later or AMD Ryzen.
4430 This issue was reported to OpenSSL by the OSS-Fuzz project.
4435 * Malformed X.509 IPAddressFamily could cause OOB read
4437 If an X.509 certificate has a malformed IPAddressFamily extension,
4438 OpenSSL could do a one-byte buffer overread. The most likely result
4439 would be an erroneous display of the certificate in text format.
4441 This issue was reported to OpenSSL by the OSS-Fuzz project.
4445 ### Changes between 1.0.2k and 1.0.2l [25 May 2017]
4447 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
4448 platform rather than 'mingw'.
4452 ### Changes between 1.0.2j and 1.0.2k [26 Jan 2017]
4454 * Truncated packet could crash via OOB read
4456 If one side of an SSL/TLS path is running on a 32-bit host and a specific
4457 cipher is being used, then a truncated packet can cause that host to
4458 perform an out-of-bounds read, usually resulting in a crash.
4460 This issue was reported to OpenSSL by Robert Święcki of Google.
4465 * BN_mod_exp may produce incorrect results on x86_64
4467 There is a carry propagating bug in the x86_64 Montgomery squaring
4468 procedure. No EC algorithms are affected. Analysis suggests that attacks
4469 against RSA and DSA as a result of this defect would be very difficult to
4470 perform and are not believed likely. Attacks against DH are considered just
4471 feasible (although very difficult) because most of the work necessary to
4472 deduce information about a private key may be performed offline. The amount
4473 of resources required for such an attack would be very significant and
4474 likely only accessible to a limited number of attackers. An attacker would
4475 additionally need online access to an unpatched system using the target
4476 private key in a scenario with persistent DH parameters and a private
4477 key that is shared between multiple clients. For example this can occur by
4478 default in OpenSSL DHE based SSL/TLS ciphersuites. Note: This issue is very
4479 similar to CVE-2015-3193 but must be treated as a separate problem.
4481 This issue was reported to OpenSSL by the OSS-Fuzz project.
4486 * Montgomery multiplication may produce incorrect results
4488 There is a carry propagating bug in the Broadwell-specific Montgomery
4489 multiplication procedure that handles input lengths divisible by, but
4490 longer than 256 bits. Analysis suggests that attacks against RSA, DSA
4491 and DH private keys are impossible. This is because the subroutine in
4492 question is not used in operations with the private key itself and an input
4493 of the attacker's direct choice. Otherwise the bug can manifest itself as
4494 transient authentication and key negotiation failures or reproducible
4495 erroneous outcome of public-key operations with specially crafted input.
4496 Among EC algorithms only Brainpool P-512 curves are affected and one
4497 presumably can attack ECDH key negotiation. Impact was not analyzed in
4498 detail, because pre-requisites for attack are considered unlikely. Namely
4499 multiple clients have to choose the curve in question and the server has to
4500 share the private key among them, neither of which is default behaviour.
4501 Even then only clients that chose the curve will be affected.
4503 This issue was publicly reported as transient failures and was not
4504 initially recognized as a security issue. Thanks to Richard Morgan for
4505 providing reproducible case.
4510 * OpenSSL now fails if it receives an unrecognised record type in TLS1.0
4511 or TLS1.1. Previously this only happened in SSLv3 and TLS1.2. This is to
4512 prevent issues where no progress is being made and the peer continually
4513 sends unrecognised record types, using up resources processing them.
4517 ### Changes between 1.0.2i and 1.0.2j [26 Sep 2016]
4519 * Missing CRL sanity check
4521 A bug fix which included a CRL sanity check was added to OpenSSL 1.1.0
4522 but was omitted from OpenSSL 1.0.2i. As a result any attempt to use
4523 CRLs in OpenSSL 1.0.2i will crash with a null pointer exception.
4525 This issue only affects the OpenSSL 1.0.2i
4530 ### Changes between 1.0.2h and 1.0.2i [22 Sep 2016]
4532 * OCSP Status Request extension unbounded memory growth
4534 A malicious client can send an excessively large OCSP Status Request
4535 extension. If that client continually requests renegotiation, sending a
4536 large OCSP Status Request extension each time, then there will be unbounded
4537 memory growth on the server. This will eventually lead to a Denial Of
4538 Service attack through memory exhaustion. Servers with a default
4539 configuration are vulnerable even if they do not support OCSP. Builds using
4540 the "no-ocsp" build time option are not affected.
4542 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4547 * In order to mitigate the SWEET32 attack, the DES ciphers were moved from
4550 This issue was reported to OpenSSL Karthikeyan Bhargavan and Gaetan
4556 * OOB write in MDC2_Update()
4558 An overflow can occur in MDC2_Update() either if called directly or
4559 through the EVP_DigestUpdate() function using MDC2. If an attacker
4560 is able to supply very large amounts of input data after a previous
4561 call to EVP_EncryptUpdate() with a partial block then a length check
4562 can overflow resulting in a heap corruption.
4564 The amount of data needed is comparable to SIZE_MAX which is impractical
4567 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4572 * Malformed SHA512 ticket DoS
4574 If a server uses SHA512 for TLS session ticket HMAC it is vulnerable to a
4575 DoS attack where a malformed ticket will result in an OOB read which will
4578 The use of SHA512 in TLS session tickets is comparatively rare as it requires
4579 a custom server callback and ticket lookup mechanism.
4581 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4586 * OOB write in BN_bn2dec()
4588 The function BN_bn2dec() does not check the return value of BN_div_word().
4589 This can cause an OOB write if an application uses this function with an
4590 overly large BIGNUM. This could be a problem if an overly large certificate
4591 or CRL is printed out from an untrusted source. TLS is not affected because
4592 record limits will reject an oversized certificate before it is parsed.
4594 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4599 * OOB read in TS_OBJ_print_bio()
4601 The function TS_OBJ_print_bio() misuses OBJ_obj2txt(): the return value is
4602 the total length the OID text representation would use and not the amount
4603 of data written. This will result in OOB reads when large OIDs are
4606 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4611 * Pointer arithmetic undefined behaviour
4613 Avoid some undefined pointer arithmetic
4615 A common idiom in the codebase is to check limits in the following manner:
4618 Where "p" points to some malloc'd data of SIZE bytes and
4621 "len" here could be from some externally supplied data (e.g. from a TLS
4624 The rules of C pointer arithmetic are such that "p + len" is only well
4625 defined where len <= SIZE. Therefore the above idiom is actually
4626 undefined behaviour.
4628 For example this could cause problems if some malloc implementation
4629 provides an address for "p" such that "p + len" actually overflows for
4630 values of len that are too big and therefore p + len < limit.
4632 This issue was reported to OpenSSL by Guido Vranken
4637 * Constant time flag not preserved in DSA signing
4639 Operations in the DSA signing algorithm should run in constant time in
4640 order to avoid side channel attacks. A flaw in the OpenSSL DSA
4641 implementation means that a non-constant time codepath is followed for
4642 certain operations. This has been demonstrated through a cache-timing
4643 attack to be sufficient for an attacker to recover the private DSA key.
4645 This issue was reported by César Pereida (Aalto University), Billy Brumley
4646 (Tampere University of Technology), and Yuval Yarom (The University of
4647 Adelaide and NICTA).
4652 * DTLS buffered message DoS
4654 In a DTLS connection where handshake messages are delivered out-of-order
4655 those messages that OpenSSL is not yet ready to process will be buffered
4656 for later use. Under certain circumstances, a flaw in the logic means that
4657 those messages do not get removed from the buffer even though the handshake
4658 has been completed. An attacker could force up to approx. 15 messages to
4659 remain in the buffer when they are no longer required. These messages will
4660 be cleared when the DTLS connection is closed. The default maximum size for
4661 a message is 100k. Therefore the attacker could force an additional 1500k
4662 to be consumed per connection. By opening many simulataneous connections an
4663 attacker could cause a DoS attack through memory exhaustion.
4665 This issue was reported to OpenSSL by Quan Luo.
4670 * DTLS replay protection DoS
4672 A flaw in the DTLS replay attack protection mechanism means that records
4673 that arrive for future epochs update the replay protection "window" before
4674 the MAC for the record has been validated. This could be exploited by an
4675 attacker by sending a record for the next epoch (which does not have to
4676 decrypt or have a valid MAC), with a very large sequence number. This means
4677 that all subsequent legitimate packets are dropped causing a denial of
4678 service for a specific DTLS connection.
4680 This issue was reported to OpenSSL by the OCAP audit team.
4685 * Certificate message OOB reads
4687 In OpenSSL 1.0.2 and earlier some missing message length checks can result
4688 in OOB reads of up to 2 bytes beyond an allocated buffer. There is a
4689 theoretical DoS risk but this has not been observed in practice on common
4692 The messages affected are client certificate, client certificate request
4693 and server certificate. As a result the attack can only be performed
4694 against a client or a server which enables client authentication.
4696 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4701 ### Changes between 1.0.2g and 1.0.2h [3 May 2016]
4703 * Prevent padding oracle in AES-NI CBC MAC check
4705 A MITM attacker can use a padding oracle attack to decrypt traffic
4706 when the connection uses an AES CBC cipher and the server support
4709 This issue was introduced as part of the fix for Lucky 13 padding
4710 attack ([CVE-2013-0169]). The padding check was rewritten to be in
4711 constant time by making sure that always the same bytes are read and
4712 compared against either the MAC or padding bytes. But it no longer
4713 checked that there was enough data to have both the MAC and padding
4716 This issue was reported by Juraj Somorovsky using TLS-Attacker.
4720 * Fix EVP_EncodeUpdate overflow
4722 An overflow can occur in the EVP_EncodeUpdate() function which is used for
4723 Base64 encoding of binary data. If an attacker is able to supply very large
4724 amounts of input data then a length check can overflow resulting in a heap
4727 Internally to OpenSSL the EVP_EncodeUpdate() function is primarily used by
4728 the `PEM_write_bio*` family of functions. These are mainly used within the
4729 OpenSSL command line applications, so any application which processes data
4730 from an untrusted source and outputs it as a PEM file should be considered
4731 vulnerable to this issue. User applications that call these APIs directly
4732 with large amounts of untrusted data may also be vulnerable.
4734 This issue was reported by Guido Vranken.
4739 * Fix EVP_EncryptUpdate overflow
4741 An overflow can occur in the EVP_EncryptUpdate() function. If an attacker
4742 is able to supply very large amounts of input data after a previous call to
4743 EVP_EncryptUpdate() with a partial block then a length check can overflow
4744 resulting in a heap corruption. Following an analysis of all OpenSSL
4745 internal usage of the EVP_EncryptUpdate() function all usage is one of two
4746 forms. The first form is where the EVP_EncryptUpdate() call is known to be
4747 the first called function after an EVP_EncryptInit(), and therefore that
4748 specific call must be safe. The second form is where the length passed to
4749 EVP_EncryptUpdate() can be seen from the code to be some small value and
4750 therefore there is no possibility of an overflow. Since all instances are
4751 one of these two forms, it is believed that there can be no overflows in
4752 internal code due to this problem. It should be noted that
4753 EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths.
4754 Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances
4755 of these calls have also been analysed too and it is believed there are no
4756 instances in internal usage where an overflow could occur.
4758 This issue was reported by Guido Vranken.
4763 * Prevent ASN.1 BIO excessive memory allocation
4765 When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio()
4766 a short invalid encoding can cause allocation of large amounts of memory
4767 potentially consuming excessive resources or exhausting memory.
4769 Any application parsing untrusted data through d2i BIO functions is
4770 affected. The memory based functions such as d2i_X509() are *not* affected.
4771 Since the memory based functions are used by the TLS library, TLS
4772 applications are not affected.
4774 This issue was reported by Brian Carpenter.
4781 ASN1 Strings that are over 1024 bytes can cause an overread in applications
4782 using the X509_NAME_oneline() function on EBCDIC systems. This could result
4783 in arbitrary stack data being returned in the buffer.
4785 This issue was reported by Guido Vranken.
4790 * Modify behavior of ALPN to invoke callback after SNI/servername
4791 callback, such that updates to the SSL_CTX affect ALPN.
4795 * Remove LOW from the DEFAULT cipher list. This removes singles DES from the
4800 * Only remove the SSLv2 methods with the no-ssl2-method option. When the
4801 methods are enabled and ssl2 is disabled the methods return NULL.
4805 ### Changes between 1.0.2f and 1.0.2g [1 Mar 2016]
4807 * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL.
4808 Builds that are not configured with "enable-weak-ssl-ciphers" will not
4809 provide any "EXPORT" or "LOW" strength ciphers.
4813 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2
4814 is by default disabled at build-time. Builds that are not configured with
4815 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used,
4816 users who want to negotiate SSLv2 via the version-flexible SSLv23_method()
4817 will need to explicitly call either of:
4819 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2);
4821 SSL_clear_options(ssl, SSL_OP_NO_SSLv2);
4823 as appropriate. Even if either of those is used, or the application
4824 explicitly uses the version-specific SSLv2_method() or its client and
4825 server variants, SSLv2 ciphers vulnerable to exhaustive search key
4826 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT
4827 ciphers, and SSLv2 56-bit DES are no longer available.
4832 * Fix a double-free in DSA code
4834 A double free bug was discovered when OpenSSL parses malformed DSA private
4835 keys and could lead to a DoS attack or memory corruption for applications
4836 that receive DSA private keys from untrusted sources. This scenario is
4839 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using
4845 * Disable SRP fake user seed to address a server memory leak.
4847 Add a new method SRP_VBASE_get1_by_user that handles the seed properly.
4849 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
4850 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user
4851 was changed to ignore the "fake user" SRP seed, even if the seed
4854 Users should use SRP_VBASE_get1_by_user instead. Note that in
4855 SRP_VBASE_get1_by_user, caller must free the returned value. Note
4856 also that even though configuring the SRP seed attempts to hide
4857 invalid usernames by continuing the handshake with fake
4858 credentials, this behaviour is not constant time and no strong
4859 guarantees are made that the handshake is indistinguishable from
4860 that of a valid user.
4865 * Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption
4867 In the BN_hex2bn function the number of hex digits is calculated using an
4868 int value `i`. Later `bn_expand` is called with a value of `i * 4`. For
4869 large values of `i` this can result in `bn_expand` not allocating any
4870 memory because `i * 4` is negative. This can leave the internal BIGNUM data
4871 field as NULL leading to a subsequent NULL ptr deref. For very large values
4872 of `i`, the calculation `i * 4` could be a positive value smaller than `i`.
4873 In this case memory is allocated to the internal BIGNUM data field, but it
4874 is insufficiently sized leading to heap corruption. A similar issue exists
4875 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn
4876 is ever called by user applications with very large untrusted hex/dec data.
4877 This is anticipated to be a rare occurrence.
4879 All OpenSSL internal usage of these functions use data that is not expected
4880 to be untrusted, e.g. config file data or application command line
4881 arguments. If user developed applications generate config file data based
4882 on untrusted data then it is possible that this could also lead to security
4883 consequences. This is also anticipated to be rare.
4885 This issue was reported to OpenSSL by Guido Vranken.
4890 * Fix memory issues in `BIO_*printf` functions
4892 The internal `fmtstr` function used in processing a "%s" format string in
4893 the `BIO_*printf` functions could overflow while calculating the length of a
4894 string and cause an OOB read when printing very long strings.
4896 Additionally the internal `doapr_outch` function can attempt to write to an
4897 OOB memory location (at an offset from the NULL pointer) in the event of a
4898 memory allocation failure. In 1.0.2 and below this could be caused where
4899 the size of a buffer to be allocated is greater than INT_MAX. E.g. this
4900 could be in processing a very long "%s" format string. Memory leaks can
4903 The first issue may mask the second issue dependent on compiler behaviour.
4904 These problems could enable attacks where large amounts of untrusted data
4905 is passed to the `BIO_*printf` functions. If applications use these functions
4906 in this way then they could be vulnerable. OpenSSL itself uses these
4907 functions when printing out human-readable dumps of ASN.1 data. Therefore
4908 applications that print this data could be vulnerable if the data is from
4909 untrusted sources. OpenSSL command line applications could also be
4910 vulnerable where they print out ASN.1 data, or if untrusted data is passed
4911 as command line arguments.
4913 Libssl is not considered directly vulnerable. Additionally certificates etc
4914 received via remote connections via libssl are also unlikely to be able to
4915 trigger these issues because of message size limits enforced within libssl.
4917 This issue was reported to OpenSSL Guido Vranken.
4922 * Side channel attack on modular exponentiation
4924 A side-channel attack was found which makes use of cache-bank conflicts on
4925 the Intel Sandy-Bridge microarchitecture which could lead to the recovery
4926 of RSA keys. The ability to exploit this issue is limited as it relies on
4927 an attacker who has control of code in a thread running on the same
4928 hyper-threaded core as the victim thread which is performing decryptions.
4930 This issue was reported to OpenSSL by Yuval Yarom, The University of
4931 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and
4932 Nadia Heninger, University of Pennsylvania with more information at
4933 <http://cachebleed.info>.
4938 * Change the `req` command to generate a 2048-bit RSA/DSA key by default,
4939 if no keysize is specified with default_bits. This fixes an
4940 omission in an earlier change that changed all RSA/DSA key generation
4941 commands to use 2048 bits by default.
4945 ### Changes between 1.0.2e and 1.0.2f [28 Jan 2016]
4947 * DH small subgroups
4949 Historically OpenSSL only ever generated DH parameters based on "safe"
4950 primes. More recently (in version 1.0.2) support was provided for
4951 generating X9.42 style parameter files such as those required for RFC 5114
4952 support. The primes used in such files may not be "safe". Where an
4953 application is using DH configured with parameters based on primes that are
4954 not "safe" then an attacker could use this fact to find a peer's private
4955 DH exponent. This attack requires that the attacker complete multiple
4956 handshakes in which the peer uses the same private DH exponent. For example
4957 this could be used to discover a TLS server's private DH exponent if it's
4958 reusing the private DH exponent or it's using a static DH ciphersuite.
4960 OpenSSL provides the option SSL_OP_SINGLE_DH_USE for ephemeral DH (DHE) in
4961 TLS. It is not on by default. If the option is not set then the server
4962 reuses the same private DH exponent for the life of the server process and
4963 would be vulnerable to this attack. It is believed that many popular
4964 applications do set this option and would therefore not be at risk.
4966 The fix for this issue adds an additional check where a "q" parameter is
4967 available (as is the case in X9.42 based parameters). This detects the
4968 only known attack, and is the only possible defense for static DH
4969 ciphersuites. This could have some performance impact.
4971 Additionally the SSL_OP_SINGLE_DH_USE option has been switched on by
4972 default and cannot be disabled. This could have some performance impact.
4974 This issue was reported to OpenSSL by Antonio Sanso (Adobe).
4979 * SSLv2 doesn't block disabled ciphers
4981 A malicious client can negotiate SSLv2 ciphers that have been disabled on
4982 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
4983 been disabled, provided that the SSLv2 protocol was not also disabled via
4986 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
4987 and Sebastian Schinzel.
4992 ### Changes between 1.0.2d and 1.0.2e [3 Dec 2015]
4994 * BN_mod_exp may produce incorrect results on x86_64
4996 There is a carry propagating bug in the x86_64 Montgomery squaring
4997 procedure. No EC algorithms are affected. Analysis suggests that attacks
4998 against RSA and DSA as a result of this defect would be very difficult to
4999 perform and are not believed likely. Attacks against DH are considered just
5000 feasible (although very difficult) because most of the work necessary to
5001 deduce information about a private key may be performed offline. The amount
5002 of resources required for such an attack would be very significant and
5003 likely only accessible to a limited number of attackers. An attacker would
5004 additionally need online access to an unpatched system using the target
5005 private key in a scenario with persistent DH parameters and a private
5006 key that is shared between multiple clients. For example this can occur by
5007 default in OpenSSL DHE based SSL/TLS ciphersuites.
5009 This issue was reported to OpenSSL by Hanno Böck.
5014 * Certificate verify crash with missing PSS parameter
5016 The signature verification routines will crash with a NULL pointer
5017 dereference if presented with an ASN.1 signature using the RSA PSS
5018 algorithm and absent mask generation function parameter. Since these
5019 routines are used to verify certificate signature algorithms this can be
5020 used to crash any certificate verification operation and exploited in a
5021 DoS attack. Any application which performs certificate verification is
5022 vulnerable including OpenSSL clients and servers which enable client
5025 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
5030 * X509_ATTRIBUTE memory leak
5032 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
5033 memory. This structure is used by the PKCS#7 and CMS routines so any
5034 application which reads PKCS#7 or CMS data from untrusted sources is
5035 affected. SSL/TLS is not affected.
5037 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
5043 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
5044 This changes the decoding behaviour for some invalid messages,
5045 though the change is mostly in the more lenient direction, and
5046 legacy behaviour is preserved as much as possible.
5050 * In DSA_generate_parameters_ex, if the provided seed is too short,
5053 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
5055 ### Changes between 1.0.2c and 1.0.2d [9 Jul 2015]
5057 * Alternate chains certificate forgery
5059 During certificate verification, OpenSSL will attempt to find an
5060 alternative certificate chain if the first attempt to build such a chain
5061 fails. An error in the implementation of this logic can mean that an
5062 attacker could cause certain checks on untrusted certificates to be
5063 bypassed, such as the CA flag, enabling them to use a valid leaf
5064 certificate to act as a CA and "issue" an invalid certificate.
5066 This issue was reported to OpenSSL by Adam Langley/David Benjamin
5071 ### Changes between 1.0.2b and 1.0.2c [12 Jun 2015]
5073 * Fix HMAC ABI incompatibility. The previous version introduced an ABI
5074 incompatibility in the handling of HMAC. The previous ABI has now been
5079 ### Changes between 1.0.2a and 1.0.2b [11 Jun 2015]
5081 * Malformed ECParameters causes infinite loop
5083 When processing an ECParameters structure OpenSSL enters an infinite loop
5084 if the curve specified is over a specially malformed binary polynomial
5087 This can be used to perform denial of service against any
5088 system which processes public keys, certificate requests or
5089 certificates. This includes TLS clients and TLS servers with
5090 client authentication enabled.
5092 This issue was reported to OpenSSL by Joseph Barr-Pixton.
5097 * Exploitable out-of-bounds read in X509_cmp_time
5099 X509_cmp_time does not properly check the length of the ASN1_TIME
5100 string and can read a few bytes out of bounds. In addition,
5101 X509_cmp_time accepts an arbitrary number of fractional seconds in the
5104 An attacker can use this to craft malformed certificates and CRLs of
5105 various sizes and potentially cause a segmentation fault, resulting in
5106 a DoS on applications that verify certificates or CRLs. TLS clients
5107 that verify CRLs are affected. TLS clients and servers with client
5108 authentication enabled may be affected if they use custom verification
5111 This issue was reported to OpenSSL by Robert Swiecki (Google), and
5112 independently by Hanno Böck.
5117 * PKCS7 crash with missing EnvelopedContent
5119 The PKCS#7 parsing code does not handle missing inner EncryptedContent
5120 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
5121 with missing content and trigger a NULL pointer dereference on parsing.
5123 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
5124 structures from untrusted sources are affected. OpenSSL clients and
5125 servers are not affected.
5127 This issue was reported to OpenSSL by Michal Zalewski (Google).
5132 * CMS verify infinite loop with unknown hash function
5134 When verifying a signedData message the CMS code can enter an infinite loop
5135 if presented with an unknown hash function OID. This can be used to perform
5136 denial of service against any system which verifies signedData messages using
5138 This issue was reported to OpenSSL by Johannes Bauer.
5143 * Race condition handling NewSessionTicket
5145 If a NewSessionTicket is received by a multi-threaded client when attempting to
5146 reuse a previous ticket then a race condition can occur potentially leading to
5147 a double free of the ticket data.
5152 * Only support 256-bit or stronger elliptic curves with the
5153 'ecdh_auto' setting (server) or by default (client). Of supported
5154 curves, prefer P-256 (both).
5158 ### Changes between 1.0.2 and 1.0.2a [19 Mar 2015]
5160 * ClientHello sigalgs DoS fix
5162 If a client connects to an OpenSSL 1.0.2 server and renegotiates with an
5163 invalid signature algorithms extension a NULL pointer dereference will
5164 occur. This can be exploited in a DoS attack against the server.
5166 This issue was was reported to OpenSSL by David Ramos of Stanford
5170 *Stephen Henson and Matt Caswell*
5172 * Multiblock corrupted pointer fix
5174 OpenSSL 1.0.2 introduced the "multiblock" performance improvement. This
5175 feature only applies on 64 bit x86 architecture platforms that support AES
5176 NI instructions. A defect in the implementation of "multiblock" can cause
5177 OpenSSL's internal write buffer to become incorrectly set to NULL when
5178 using non-blocking IO. Typically, when the user application is using a
5179 socket BIO for writing, this will only result in a failed connection.
5180 However if some other BIO is used then it is likely that a segmentation
5181 fault will be triggered, thus enabling a potential DoS attack.
5183 This issue was reported to OpenSSL by Daniel Danner and Rainer Mueller.
5188 * Segmentation fault in DTLSv1_listen fix
5190 The DTLSv1_listen function is intended to be stateless and processes the
5191 initial ClientHello from many peers. It is common for user code to loop
5192 over the call to DTLSv1_listen until a valid ClientHello is received with
5193 an associated cookie. A defect in the implementation of DTLSv1_listen means
5194 that state is preserved in the SSL object from one invocation to the next
5195 that can lead to a segmentation fault. Errors processing the initial
5196 ClientHello can trigger this scenario. An example of such an error could be
5197 that a DTLS1.0 only client is attempting to connect to a DTLS1.2 only
5200 This issue was reported to OpenSSL by Per Allansson.
5205 * Segmentation fault in ASN1_TYPE_cmp fix
5207 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
5208 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
5209 certificate signature algorithm consistency this can be used to crash any
5210 certificate verification operation and exploited in a DoS attack. Any
5211 application which performs certificate verification is vulnerable including
5212 OpenSSL clients and servers which enable client authentication.
5217 * Segmentation fault for invalid PSS parameters fix
5219 The signature verification routines will crash with a NULL pointer
5220 dereference if presented with an ASN.1 signature using the RSA PSS
5221 algorithm and invalid parameters. Since these routines are used to verify
5222 certificate signature algorithms this can be used to crash any
5223 certificate verification operation and exploited in a DoS attack. Any
5224 application which performs certificate verification is vulnerable including
5225 OpenSSL clients and servers which enable client authentication.
5227 This issue was was reported to OpenSSL by Brian Carpenter.
5232 * ASN.1 structure reuse memory corruption fix
5234 Reusing a structure in ASN.1 parsing may allow an attacker to cause
5235 memory corruption via an invalid write. Such reuse is and has been
5236 strongly discouraged and is believed to be rare.
5238 Applications that parse structures containing CHOICE or ANY DEFINED BY
5239 components may be affected. Certificate parsing (d2i_X509 and related
5240 functions) are however not affected. OpenSSL clients and servers are
5246 * PKCS7 NULL pointer dereferences fix
5248 The PKCS#7 parsing code does not handle missing outer ContentInfo
5249 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
5250 missing content and trigger a NULL pointer dereference on parsing.
5252 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
5253 otherwise parse PKCS#7 structures from untrusted sources are
5254 affected. OpenSSL clients and servers are not affected.
5256 This issue was reported to OpenSSL by Michal Zalewski (Google).
5261 * DoS via reachable assert in SSLv2 servers fix
5263 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
5264 servers that both support SSLv2 and enable export cipher suites by sending
5265 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
5267 This issue was discovered by Sean Burford (Google) and Emilia Käsper
5268 (OpenSSL development team).
5273 * Empty CKE with client auth and DHE fix
5275 If client auth is used then a server can seg fault in the event of a DHE
5276 ciphersuite being selected and a zero length ClientKeyExchange message
5277 being sent by the client. This could be exploited in a DoS attack.
5282 * Handshake with unseeded PRNG fix
5284 Under certain conditions an OpenSSL 1.0.2 client can complete a handshake
5285 with an unseeded PRNG. The conditions are:
5286 - The client is on a platform where the PRNG has not been seeded
5287 automatically, and the user has not seeded manually
5288 - A protocol specific client method version has been used (i.e. not
5289 SSL_client_methodv23)
5290 - A ciphersuite is used that does not require additional random data from
5291 the PRNG beyond the initial ClientHello client random (e.g. PSK-RC4-SHA).
5293 If the handshake succeeds then the client random that has been used will
5294 have been generated from a PRNG with insufficient entropy and therefore the
5295 output may be predictable.
5297 For example using the following command with an unseeded openssl will
5298 succeed on an unpatched platform:
5300 openssl s_client -psk 1a2b3c4d -tls1_2 -cipher PSK-RC4-SHA
5305 * Use After Free following d2i_ECPrivatekey error fix
5307 A malformed EC private key file consumed via the d2i_ECPrivateKey function
5308 could cause a use after free condition. This, in turn, could cause a double
5309 free in several private key parsing functions (such as d2i_PrivateKey
5310 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
5311 for applications that receive EC private keys from untrusted
5312 sources. This scenario is considered rare.
5314 This issue was discovered by the BoringSSL project and fixed in their
5320 * X509_to_X509_REQ NULL pointer deref fix
5322 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
5323 the certificate key is invalid. This function is rarely used in practice.
5325 This issue was discovered by Brian Carpenter.
5330 * Removed the export ciphers from the DEFAULT ciphers
5334 ### Changes between 1.0.1l and 1.0.2 [22 Jan 2015]
5336 * Facilitate "universal" ARM builds targeting range of ARM ISAs, e.g.
5337 ARMv5 through ARMv8, as opposite to "locking" it to single one.
5338 So far those who have to target multiple platforms would compromise
5339 and argue that binary targeting say ARMv5 would still execute on
5340 ARMv8. "Universal" build resolves this compromise by providing
5341 near-optimal performance even on newer platforms.
5345 * Accelerated NIST P-256 elliptic curve implementation for x86_64
5346 (other platforms pending).
5348 *Shay Gueron & Vlad Krasnov (Intel Corp), Andy Polyakov*
5350 * Add support for the SignedCertificateTimestampList certificate and
5351 OCSP response extensions from RFC6962.
5355 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
5356 for corner cases. (Certain input points at infinity could lead to
5357 bogus results, with non-infinity inputs mapped to infinity too.)
5361 * Initial support for PowerISA 2.0.7, first implemented in POWER8.
5362 This covers AES, SHA256/512 and GHASH. "Initial" means that most
5363 common cases are optimized and there still is room for further
5364 improvements. Vector Permutation AES for Altivec is also added.
5368 * Add support for little-endian ppc64 Linux target.
5370 *Marcelo Cerri (IBM)*
5372 * Initial support for AMRv8 ISA crypto extensions. This covers AES,
5373 SHA1, SHA256 and GHASH. "Initial" means that most common cases
5374 are optimized and there still is room for further improvements.
5375 Both 32- and 64-bit modes are supported.
5377 *Andy Polyakov, Ard Biesheuvel (Linaro)*
5379 * Improved ARMv7 NEON support.
5383 * Support for SPARC Architecture 2011 crypto extensions, first
5384 implemented in SPARC T4. This covers AES, DES, Camellia, SHA1,
5385 SHA256/512, MD5, GHASH and modular exponentiation.
5387 *Andy Polyakov, David Miller*
5389 * Accelerated modular exponentiation for Intel processors, a.k.a.
5392 *Shay Gueron & Vlad Krasnov (Intel Corp)*
5394 * Support for new and upcoming Intel processors, including AVX2,
5395 BMI and SHA ISA extensions. This includes additional "stitched"
5396 implementations, AESNI-SHA256 and GCM, and multi-buffer support
5399 This work was sponsored by Intel Corp.
5403 * Support for DTLS 1.2. This adds two sets of DTLS methods: DTLS_*_method()
5404 supports both DTLS 1.2 and 1.0 and should use whatever version the peer
5405 supports and DTLSv1_2_*_method() which supports DTLS 1.2 only.
5409 * Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
5410 this fixes a limitation in previous versions of OpenSSL.
5414 * Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
5415 MGF1 digest and OAEP label.
5419 * Add EVP support for key wrapping algorithms, to avoid problems with
5420 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
5421 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
5422 algorithms and include tests cases.
5426 * Add functions to allocate and set the fields of an ECDSA_METHOD
5429 *Douglas E. Engert, Steve Henson*
5431 * New functions OPENSSL_gmtime_diff and ASN1_TIME_diff to find the
5432 difference in days and seconds between two tm or ASN1_TIME structures.
5436 * Add -rev test option to s_server to just reverse order of characters
5437 received by client and send back to server. Also prints an abbreviated
5438 summary of the connection parameters.
5442 * New option -brief for s_client and s_server to print out a brief summary
5443 of connection parameters.
5447 * Add callbacks for arbitrary TLS extensions.
5449 *Trevor Perrin <trevp@trevp.net> and Ben Laurie*
5451 * New option -crl_download in several openssl utilities to download CRLs
5452 from CRLDP extension in certificates.
5456 * New options -CRL and -CRLform for s_client and s_server for CRLs.
5460 * New function X509_CRL_diff to generate a delta CRL from the difference
5461 of two full CRLs. Add support to "crl" utility.
5465 * New functions to set lookup_crls function and to retrieve
5466 X509_STORE from X509_STORE_CTX.
5470 * Print out deprecated issuer and subject unique ID fields in
5475 * Extend OCSP I/O functions so they can be used for simple general purpose
5476 HTTP as well as OCSP. New wrapper function which can be used to download
5477 CRLs using the OCSP API.
5481 * Delegate command line handling in s_client/s_server to SSL_CONF APIs.
5485 * `SSL_CONF*` functions. These provide a common framework for application
5486 configuration using configuration files or command lines.
5490 * SSL/TLS tracing code. This parses out SSL/TLS records using the
5491 message callback and prints the results. Needs compile time option
5492 "enable-ssl-trace". New options to s_client and s_server to enable
5497 * New ctrl and macro to retrieve supported points extensions.
5498 Print out extension in s_server and s_client.
5502 * New functions to retrieve certificate signature and signature
5507 * Add functions to retrieve and manipulate the raw cipherlist sent by a
5512 * New Suite B modes for TLS code. These use and enforce the requirements
5513 of RFC6460: restrict ciphersuites, only permit Suite B algorithms and
5514 only use Suite B curves. The Suite B modes can be set by using the
5515 strings "SUITEB128", "SUITEB192" or "SUITEB128ONLY" for the cipherstring.
5519 * New chain verification flags for Suite B levels of security. Check
5520 algorithms are acceptable when flags are set in X509_verify_cert.
5524 * Make tls1_check_chain return a set of flags indicating checks passed
5525 by a certificate chain. Add additional tests to handle client
5526 certificates: checks for matching certificate type and issuer name
5531 * If an attempt is made to use a signature algorithm not in the peer
5532 preference list abort the handshake. If client has no suitable
5533 signature algorithms in response to a certificate request do not
5534 use the certificate.
5538 * If server EC tmp key is not in client preference list abort handshake.
5542 * Add support for certificate stores in CERT structure. This makes it
5543 possible to have different stores per SSL structure or one store in
5544 the parent SSL_CTX. Include distinct stores for certificate chain
5545 verification and chain building. New ctrl SSL_CTRL_BUILD_CERT_CHAIN
5546 to build and store a certificate chain in CERT structure: returning
5547 an error if the chain cannot be built: this will allow applications
5548 to test if a chain is correctly configured.
5550 Note: if the CERT based stores are not set then the parent SSL_CTX
5551 store is used to retain compatibility with existing behaviour.
5555 * New function ssl_set_client_disabled to set a ciphersuite disabled
5556 mask based on the current session, check mask when sending client
5557 hello and checking the requested ciphersuite.
5561 * New ctrls to retrieve and set certificate types in a certificate
5562 request message. Print out received values in s_client. If certificate
5563 types is not set with custom values set sensible values based on
5564 supported signature algorithms.
5568 * Support for distinct client and server supported signature algorithms.
5572 * Add certificate callback. If set this is called whenever a certificate
5573 is required by client or server. An application can decide which
5574 certificate chain to present based on arbitrary criteria: for example
5575 supported signature algorithms. Add very simple example to s_server.
5576 This fixes many of the problems and restrictions of the existing client
5577 certificate callback: for example you can now clear an existing
5578 certificate and specify the whole chain.
5582 * Add new "valid_flags" field to CERT_PKEY structure which determines what
5583 the certificate can be used for (if anything). Set valid_flags field
5584 in new tls1_check_chain function. Simplify ssl_set_cert_masks which used
5585 to have similar checks in it.
5587 Add new "cert_flags" field to CERT structure and include a "strict mode".
5588 This enforces some TLS certificate requirements (such as only permitting
5589 certificate signature algorithms contained in the supported algorithms
5590 extension) which some implementations ignore: this option should be used
5591 with caution as it could cause interoperability issues.
5595 * Update and tidy signature algorithm extension processing. Work out
5596 shared signature algorithms based on preferences and peer algorithms
5597 and print them out in s_client and s_server. Abort handshake if no
5598 shared signature algorithms.
5602 * Add new functions to allow customised supported signature algorithms
5603 for SSL and SSL_CTX structures. Add options to s_client and s_server
5608 * New function SSL_certs_clear() to delete all references to certificates
5609 from an SSL structure. Before this once a certificate had been added
5610 it couldn't be removed.
5614 * Integrate hostname, email address and IP address checking with certificate
5615 verification. New verify options supporting checking in openssl utility.
5619 * Fixes and wildcard matching support to hostname and email checking
5620 functions. Add manual page.
5622 *Florian Weimer (Red Hat Product Security Team)*
5624 * New functions to check a hostname email or IP address against a
5625 certificate. Add options x509 utility to print results of checks against
5630 * Fix OCSP checking.
5632 *Rob Stradling <rob.stradling@comodo.com> and Ben Laurie*
5634 * Initial experimental support for explicitly trusted non-root CAs.
5635 OpenSSL still tries to build a complete chain to a root but if an
5636 intermediate CA has a trust setting included that is used. The first
5637 setting is used: whether to trust (e.g., -addtrust option to the x509
5642 * Add -trusted_first option which attempts to find certificates in the
5643 trusted store even if an untrusted chain is also supplied.
5647 * MIPS assembly pack updates: support for MIPS32r2 and SmartMIPS ASE,
5648 platform support for Linux and Android.
5652 * Support for linux-x32, ILP32 environment in x86_64 framework.
5656 * Experimental multi-implementation support for FIPS capable OpenSSL.
5657 When in FIPS mode the approved implementations are used as normal,
5658 when not in FIPS mode the internal unapproved versions are used instead.
5659 This means that the FIPS capable OpenSSL isn't forced to use the
5660 (often lower performance) FIPS implementations outside FIPS mode.
5664 * Transparently support X9.42 DH parameters when calling
5665 PEM_read_bio_DHparameters. This means existing applications can handle
5666 the new parameter format automatically.
5670 * Initial experimental support for X9.42 DH parameter format: mainly
5671 to support use of 'q' parameter for RFC5114 parameters.
5675 * Add DH parameters from RFC5114 including test data to dhtest.
5679 * Support for automatic EC temporary key parameter selection. If enabled
5680 the most preferred EC parameters are automatically used instead of
5681 hardcoded fixed parameters. Now a server just has to call:
5682 SSL_CTX_set_ecdh_auto(ctx, 1) and the server will automatically
5683 support ECDH and use the most appropriate parameters.
5687 * Enhance and tidy EC curve and point format TLS extension code. Use
5688 static structures instead of allocation if default values are used.
5689 New ctrls to set curves we wish to support and to retrieve shared curves.
5690 Print out shared curves in s_server. New options to s_server and s_client
5691 to set list of supported curves.
5695 * New ctrls to retrieve supported signature algorithms and
5696 supported curve values as an array of NIDs. Extend openssl utility
5697 to print out received values.
5701 * Add new APIs EC_curve_nist2nid and EC_curve_nid2nist which convert
5702 between NIDs and the more common NIST names such as "P-256". Enhance
5703 ecparam utility and ECC method to recognise the NIST names for curves.
5707 * Enhance SSL/TLS certificate chain handling to support different
5708 chains for each certificate instead of one chain in the parent SSL_CTX.
5712 * Support for fixed DH ciphersuite client authentication: where both
5713 server and client use DH certificates with common parameters.
5717 * Support for fixed DH ciphersuites: those requiring DH server
5722 * New function i2d_re_X509_tbs for re-encoding the TBS portion of
5724 Note: Related 1.0.2-beta specific macros X509_get_cert_info,
5725 X509_CINF_set_modified, X509_CINF_get_issuer, X509_CINF_get_extensions and
5726 X509_CINF_get_signature were reverted post internal team review.
5731 ### Changes between 1.0.1t and 1.0.1u [22 Sep 2016]
5733 * OCSP Status Request extension unbounded memory growth
5735 A malicious client can send an excessively large OCSP Status Request
5736 extension. If that client continually requests renegotiation, sending a
5737 large OCSP Status Request extension each time, then there will be unbounded
5738 memory growth on the server. This will eventually lead to a Denial Of
5739 Service attack through memory exhaustion. Servers with a default
5740 configuration are vulnerable even if they do not support OCSP. Builds using
5741 the "no-ocsp" build time option are not affected.
5743 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5748 * In order to mitigate the SWEET32 attack, the DES ciphers were moved from
5751 This issue was reported to OpenSSL Karthikeyan Bhargavan and Gaetan
5757 * OOB write in MDC2_Update()
5759 An overflow can occur in MDC2_Update() either if called directly or
5760 through the EVP_DigestUpdate() function using MDC2. If an attacker
5761 is able to supply very large amounts of input data after a previous
5762 call to EVP_EncryptUpdate() with a partial block then a length check
5763 can overflow resulting in a heap corruption.
5765 The amount of data needed is comparable to SIZE_MAX which is impractical
5768 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5773 * Malformed SHA512 ticket DoS
5775 If a server uses SHA512 for TLS session ticket HMAC it is vulnerable to a
5776 DoS attack where a malformed ticket will result in an OOB read which will
5779 The use of SHA512 in TLS session tickets is comparatively rare as it requires
5780 a custom server callback and ticket lookup mechanism.
5782 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5787 * OOB write in BN_bn2dec()
5789 The function BN_bn2dec() does not check the return value of BN_div_word().
5790 This can cause an OOB write if an application uses this function with an
5791 overly large BIGNUM. This could be a problem if an overly large certificate
5792 or CRL is printed out from an untrusted source. TLS is not affected because
5793 record limits will reject an oversized certificate before it is parsed.
5795 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5800 * OOB read in TS_OBJ_print_bio()
5802 The function TS_OBJ_print_bio() misuses OBJ_obj2txt(): the return value is
5803 the total length the OID text representation would use and not the amount
5804 of data written. This will result in OOB reads when large OIDs are
5807 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5812 * Pointer arithmetic undefined behaviour
5814 Avoid some undefined pointer arithmetic
5816 A common idiom in the codebase is to check limits in the following manner:
5819 Where "p" points to some malloc'd data of SIZE bytes and
5822 "len" here could be from some externally supplied data (e.g. from a TLS
5825 The rules of C pointer arithmetic are such that "p + len" is only well
5826 defined where len <= SIZE. Therefore the above idiom is actually
5827 undefined behaviour.
5829 For example this could cause problems if some malloc implementation
5830 provides an address for "p" such that "p + len" actually overflows for
5831 values of len that are too big and therefore p + len < limit.
5833 This issue was reported to OpenSSL by Guido Vranken
5838 * Constant time flag not preserved in DSA signing
5840 Operations in the DSA signing algorithm should run in constant time in
5841 order to avoid side channel attacks. A flaw in the OpenSSL DSA
5842 implementation means that a non-constant time codepath is followed for
5843 certain operations. This has been demonstrated through a cache-timing
5844 attack to be sufficient for an attacker to recover the private DSA key.
5846 This issue was reported by César Pereida (Aalto University), Billy Brumley
5847 (Tampere University of Technology), and Yuval Yarom (The University of
5848 Adelaide and NICTA).
5853 * DTLS buffered message DoS
5855 In a DTLS connection where handshake messages are delivered out-of-order
5856 those messages that OpenSSL is not yet ready to process will be buffered
5857 for later use. Under certain circumstances, a flaw in the logic means that
5858 those messages do not get removed from the buffer even though the handshake
5859 has been completed. An attacker could force up to approx. 15 messages to
5860 remain in the buffer when they are no longer required. These messages will
5861 be cleared when the DTLS connection is closed. The default maximum size for
5862 a message is 100k. Therefore the attacker could force an additional 1500k
5863 to be consumed per connection. By opening many simulataneous connections an
5864 attacker could cause a DoS attack through memory exhaustion.
5866 This issue was reported to OpenSSL by Quan Luo.
5871 * DTLS replay protection DoS
5873 A flaw in the DTLS replay attack protection mechanism means that records
5874 that arrive for future epochs update the replay protection "window" before
5875 the MAC for the record has been validated. This could be exploited by an
5876 attacker by sending a record for the next epoch (which does not have to
5877 decrypt or have a valid MAC), with a very large sequence number. This means
5878 that all subsequent legitimate packets are dropped causing a denial of
5879 service for a specific DTLS connection.
5881 This issue was reported to OpenSSL by the OCAP audit team.
5886 * Certificate message OOB reads
5888 In OpenSSL 1.0.2 and earlier some missing message length checks can result
5889 in OOB reads of up to 2 bytes beyond an allocated buffer. There is a
5890 theoretical DoS risk but this has not been observed in practice on common
5893 The messages affected are client certificate, client certificate request
5894 and server certificate. As a result the attack can only be performed
5895 against a client or a server which enables client authentication.
5897 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5902 ### Changes between 1.0.1s and 1.0.1t [3 May 2016]
5904 * Prevent padding oracle in AES-NI CBC MAC check
5906 A MITM attacker can use a padding oracle attack to decrypt traffic
5907 when the connection uses an AES CBC cipher and the server support
5910 This issue was introduced as part of the fix for Lucky 13 padding
5911 attack ([CVE-2013-0169]). The padding check was rewritten to be in
5912 constant time by making sure that always the same bytes are read and
5913 compared against either the MAC or padding bytes. But it no longer
5914 checked that there was enough data to have both the MAC and padding
5917 This issue was reported by Juraj Somorovsky using TLS-Attacker.
5922 * Fix EVP_EncodeUpdate overflow
5924 An overflow can occur in the EVP_EncodeUpdate() function which is used for
5925 Base64 encoding of binary data. If an attacker is able to supply very large
5926 amounts of input data then a length check can overflow resulting in a heap
5929 Internally to OpenSSL the EVP_EncodeUpdate() function is primarly used by
5930 the `PEM_write_bio*` family of functions. These are mainly used within the
5931 OpenSSL command line applications, so any application which processes data
5932 from an untrusted source and outputs it as a PEM file should be considered
5933 vulnerable to this issue. User applications that call these APIs directly
5934 with large amounts of untrusted data may also be vulnerable.
5936 This issue was reported by Guido Vranken.
5941 * Fix EVP_EncryptUpdate overflow
5943 An overflow can occur in the EVP_EncryptUpdate() function. If an attacker
5944 is able to supply very large amounts of input data after a previous call to
5945 EVP_EncryptUpdate() with a partial block then a length check can overflow
5946 resulting in a heap corruption. Following an analysis of all OpenSSL
5947 internal usage of the EVP_EncryptUpdate() function all usage is one of two
5948 forms. The first form is where the EVP_EncryptUpdate() call is known to be
5949 the first called function after an EVP_EncryptInit(), and therefore that
5950 specific call must be safe. The second form is where the length passed to
5951 EVP_EncryptUpdate() can be seen from the code to be some small value and
5952 therefore there is no possibility of an overflow. Since all instances are
5953 one of these two forms, it is believed that there can be no overflows in
5954 internal code due to this problem. It should be noted that
5955 EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths.
5956 Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances
5957 of these calls have also been analysed too and it is believed there are no
5958 instances in internal usage where an overflow could occur.
5960 This issue was reported by Guido Vranken.
5965 * Prevent ASN.1 BIO excessive memory allocation
5967 When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio()
5968 a short invalid encoding can casuse allocation of large amounts of memory
5969 potentially consuming excessive resources or exhausting memory.
5971 Any application parsing untrusted data through d2i BIO functions is
5972 affected. The memory based functions such as d2i_X509() are *not* affected.
5973 Since the memory based functions are used by the TLS library, TLS
5974 applications are not affected.
5976 This issue was reported by Brian Carpenter.
5983 ASN1 Strings that are over 1024 bytes can cause an overread in applications
5984 using the X509_NAME_oneline() function on EBCDIC systems. This could result
5985 in arbitrary stack data being returned in the buffer.
5987 This issue was reported by Guido Vranken.
5992 * Modify behavior of ALPN to invoke callback after SNI/servername
5993 callback, such that updates to the SSL_CTX affect ALPN.
5997 * Remove LOW from the DEFAULT cipher list. This removes singles DES from the
6002 * Only remove the SSLv2 methods with the no-ssl2-method option. When the
6003 methods are enabled and ssl2 is disabled the methods return NULL.
6007 ### Changes between 1.0.1r and 1.0.1s [1 Mar 2016]
6009 * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL.
6010 Builds that are not configured with "enable-weak-ssl-ciphers" will not
6011 provide any "EXPORT" or "LOW" strength ciphers.
6015 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2
6016 is by default disabled at build-time. Builds that are not configured with
6017 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used,
6018 users who want to negotiate SSLv2 via the version-flexible SSLv23_method()
6019 will need to explicitly call either of:
6021 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2);
6023 SSL_clear_options(ssl, SSL_OP_NO_SSLv2);
6025 as appropriate. Even if either of those is used, or the application
6026 explicitly uses the version-specific SSLv2_method() or its client and
6027 server variants, SSLv2 ciphers vulnerable to exhaustive search key
6028 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT
6029 ciphers, and SSLv2 56-bit DES are no longer available.
6034 * Fix a double-free in DSA code
6036 A double free bug was discovered when OpenSSL parses malformed DSA private
6037 keys and could lead to a DoS attack or memory corruption for applications
6038 that receive DSA private keys from untrusted sources. This scenario is
6041 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using
6047 * Disable SRP fake user seed to address a server memory leak.
6049 Add a new method SRP_VBASE_get1_by_user that handles the seed properly.
6051 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
6052 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user
6053 was changed to ignore the "fake user" SRP seed, even if the seed
6056 Users should use SRP_VBASE_get1_by_user instead. Note that in
6057 SRP_VBASE_get1_by_user, caller must free the returned value. Note
6058 also that even though configuring the SRP seed attempts to hide
6059 invalid usernames by continuing the handshake with fake
6060 credentials, this behaviour is not constant time and no strong
6061 guarantees are made that the handshake is indistinguishable from
6062 that of a valid user.
6067 * Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption
6069 In the BN_hex2bn function the number of hex digits is calculated using an
6070 int value `i`. Later `bn_expand` is called with a value of `i * 4`. For
6071 large values of `i` this can result in `bn_expand` not allocating any
6072 memory because `i * 4` is negative. This can leave the internal BIGNUM data
6073 field as NULL leading to a subsequent NULL ptr deref. For very large values
6074 of `i`, the calculation `i * 4` could be a positive value smaller than `i`.
6075 In this case memory is allocated to the internal BIGNUM data field, but it
6076 is insufficiently sized leading to heap corruption. A similar issue exists
6077 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn
6078 is ever called by user applications with very large untrusted hex/dec data.
6079 This is anticipated to be a rare occurrence.
6081 All OpenSSL internal usage of these functions use data that is not expected
6082 to be untrusted, e.g. config file data or application command line
6083 arguments. If user developed applications generate config file data based
6084 on untrusted data then it is possible that this could also lead to security
6085 consequences. This is also anticipated to be rare.
6087 This issue was reported to OpenSSL by Guido Vranken.
6092 * Fix memory issues in `BIO_*printf` functions
6094 The internal `fmtstr` function used in processing a "%s" format string in
6095 the `BIO_*printf` functions could overflow while calculating the length of a
6096 string and cause an OOB read when printing very long strings.
6098 Additionally the internal `doapr_outch` function can attempt to write to an
6099 OOB memory location (at an offset from the NULL pointer) in the event of a
6100 memory allocation failure. In 1.0.2 and below this could be caused where
6101 the size of a buffer to be allocated is greater than INT_MAX. E.g. this
6102 could be in processing a very long "%s" format string. Memory leaks can
6105 The first issue may mask the second issue dependent on compiler behaviour.
6106 These problems could enable attacks where large amounts of untrusted data
6107 is passed to the `BIO_*printf` functions. If applications use these functions
6108 in this way then they could be vulnerable. OpenSSL itself uses these
6109 functions when printing out human-readable dumps of ASN.1 data. Therefore
6110 applications that print this data could be vulnerable if the data is from
6111 untrusted sources. OpenSSL command line applications could also be
6112 vulnerable where they print out ASN.1 data, or if untrusted data is passed
6113 as command line arguments.
6115 Libssl is not considered directly vulnerable. Additionally certificates etc
6116 received via remote connections via libssl are also unlikely to be able to
6117 trigger these issues because of message size limits enforced within libssl.
6119 This issue was reported to OpenSSL Guido Vranken.
6124 * Side channel attack on modular exponentiation
6126 A side-channel attack was found which makes use of cache-bank conflicts on
6127 the Intel Sandy-Bridge microarchitecture which could lead to the recovery
6128 of RSA keys. The ability to exploit this issue is limited as it relies on
6129 an attacker who has control of code in a thread running on the same
6130 hyper-threaded core as the victim thread which is performing decryptions.
6132 This issue was reported to OpenSSL by Yuval Yarom, The University of
6133 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and
6134 Nadia Heninger, University of Pennsylvania with more information at
6135 <http://cachebleed.info>.
6140 * Change the req command to generate a 2048-bit RSA/DSA key by default,
6141 if no keysize is specified with default_bits. This fixes an
6142 omission in an earlier change that changed all RSA/DSA key generation
6143 commands to use 2048 bits by default.
6147 ### Changes between 1.0.1q and 1.0.1r [28 Jan 2016]
6149 * Protection for DH small subgroup attacks
6151 As a precautionary measure the SSL_OP_SINGLE_DH_USE option has been
6152 switched on by default and cannot be disabled. This could have some
6157 * SSLv2 doesn't block disabled ciphers
6159 A malicious client can negotiate SSLv2 ciphers that have been disabled on
6160 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
6161 been disabled, provided that the SSLv2 protocol was not also disabled via
6164 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
6165 and Sebastian Schinzel.
6170 * Reject DH handshakes with parameters shorter than 1024 bits.
6174 ### Changes between 1.0.1p and 1.0.1q [3 Dec 2015]
6176 * Certificate verify crash with missing PSS parameter
6178 The signature verification routines will crash with a NULL pointer
6179 dereference if presented with an ASN.1 signature using the RSA PSS
6180 algorithm and absent mask generation function parameter. Since these
6181 routines are used to verify certificate signature algorithms this can be
6182 used to crash any certificate verification operation and exploited in a
6183 DoS attack. Any application which performs certificate verification is
6184 vulnerable including OpenSSL clients and servers which enable client
6187 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
6192 * X509_ATTRIBUTE memory leak
6194 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
6195 memory. This structure is used by the PKCS#7 and CMS routines so any
6196 application which reads PKCS#7 or CMS data from untrusted sources is
6197 affected. SSL/TLS is not affected.
6199 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
6205 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
6206 This changes the decoding behaviour for some invalid messages,
6207 though the change is mostly in the more lenient direction, and
6208 legacy behaviour is preserved as much as possible.
6212 * In DSA_generate_parameters_ex, if the provided seed is too short,
6213 use a random seed, as already documented.
6215 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
6217 ### Changes between 1.0.1o and 1.0.1p [9 Jul 2015]
6219 * Alternate chains certificate forgery
6221 During certificate verfification, OpenSSL will attempt to find an
6222 alternative certificate chain if the first attempt to build such a chain
6223 fails. An error in the implementation of this logic can mean that an
6224 attacker could cause certain checks on untrusted certificates to be
6225 bypassed, such as the CA flag, enabling them to use a valid leaf
6226 certificate to act as a CA and "issue" an invalid certificate.
6228 This issue was reported to OpenSSL by Adam Langley/David Benjamin
6234 * Race condition handling PSK identify hint
6236 If PSK identity hints are received by a multi-threaded client then
6237 the values are wrongly updated in the parent SSL_CTX structure. This can
6238 result in a race condition potentially leading to a double free of the
6244 ### Changes between 1.0.1n and 1.0.1o [12 Jun 2015]
6246 * Fix HMAC ABI incompatibility. The previous version introduced an ABI
6247 incompatibility in the handling of HMAC. The previous ABI has now been
6250 ### Changes between 1.0.1m and 1.0.1n [11 Jun 2015]
6252 * Malformed ECParameters causes infinite loop
6254 When processing an ECParameters structure OpenSSL enters an infinite loop
6255 if the curve specified is over a specially malformed binary polynomial
6258 This can be used to perform denial of service against any
6259 system which processes public keys, certificate requests or
6260 certificates. This includes TLS clients and TLS servers with
6261 client authentication enabled.
6263 This issue was reported to OpenSSL by Joseph Barr-Pixton.
6268 * Exploitable out-of-bounds read in X509_cmp_time
6270 X509_cmp_time does not properly check the length of the ASN1_TIME
6271 string and can read a few bytes out of bounds. In addition,
6272 X509_cmp_time accepts an arbitrary number of fractional seconds in the
6275 An attacker can use this to craft malformed certificates and CRLs of
6276 various sizes and potentially cause a segmentation fault, resulting in
6277 a DoS on applications that verify certificates or CRLs. TLS clients
6278 that verify CRLs are affected. TLS clients and servers with client
6279 authentication enabled may be affected if they use custom verification
6282 This issue was reported to OpenSSL by Robert Swiecki (Google), and
6283 independently by Hanno Böck.
6288 * PKCS7 crash with missing EnvelopedContent
6290 The PKCS#7 parsing code does not handle missing inner EncryptedContent
6291 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
6292 with missing content and trigger a NULL pointer dereference on parsing.
6294 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
6295 structures from untrusted sources are affected. OpenSSL clients and
6296 servers are not affected.
6298 This issue was reported to OpenSSL by Michal Zalewski (Google).
6303 * CMS verify infinite loop with unknown hash function
6305 When verifying a signedData message the CMS code can enter an infinite loop
6306 if presented with an unknown hash function OID. This can be used to perform
6307 denial of service against any system which verifies signedData messages using
6309 This issue was reported to OpenSSL by Johannes Bauer.
6314 * Race condition handling NewSessionTicket
6316 If a NewSessionTicket is received by a multi-threaded client when attempting to
6317 reuse a previous ticket then a race condition can occur potentially leading to
6318 a double free of the ticket data.
6323 * Reject DH handshakes with parameters shorter than 768 bits.
6325 *Kurt Roeckx and Emilia Kasper*
6327 * dhparam: generate 2048-bit parameters by default.
6329 *Kurt Roeckx and Emilia Kasper*
6331 ### Changes between 1.0.1l and 1.0.1m [19 Mar 2015]
6333 * Segmentation fault in ASN1_TYPE_cmp fix
6335 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
6336 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
6337 certificate signature algorithm consistency this can be used to crash any
6338 certificate verification operation and exploited in a DoS attack. Any
6339 application which performs certificate verification is vulnerable including
6340 OpenSSL clients and servers which enable client authentication.
6345 * ASN.1 structure reuse memory corruption fix
6347 Reusing a structure in ASN.1 parsing may allow an attacker to cause
6348 memory corruption via an invalid write. Such reuse is and has been
6349 strongly discouraged and is believed to be rare.
6351 Applications that parse structures containing CHOICE or ANY DEFINED BY
6352 components may be affected. Certificate parsing (d2i_X509 and related
6353 functions) are however not affected. OpenSSL clients and servers are
6359 * PKCS7 NULL pointer dereferences fix
6361 The PKCS#7 parsing code does not handle missing outer ContentInfo
6362 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
6363 missing content and trigger a NULL pointer dereference on parsing.
6365 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
6366 otherwise parse PKCS#7 structures from untrusted sources are
6367 affected. OpenSSL clients and servers are not affected.
6369 This issue was reported to OpenSSL by Michal Zalewski (Google).
6374 * DoS via reachable assert in SSLv2 servers fix
6376 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
6377 servers that both support SSLv2 and enable export cipher suites by sending
6378 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
6380 This issue was discovered by Sean Burford (Google) and Emilia Käsper
6381 (OpenSSL development team).
6386 * Use After Free following d2i_ECPrivatekey error fix
6388 A malformed EC private key file consumed via the d2i_ECPrivateKey function
6389 could cause a use after free condition. This, in turn, could cause a double
6390 free in several private key parsing functions (such as d2i_PrivateKey
6391 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
6392 for applications that receive EC private keys from untrusted
6393 sources. This scenario is considered rare.
6395 This issue was discovered by the BoringSSL project and fixed in their
6401 * X509_to_X509_REQ NULL pointer deref fix
6403 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
6404 the certificate key is invalid. This function is rarely used in practice.
6406 This issue was discovered by Brian Carpenter.
6411 * Removed the export ciphers from the DEFAULT ciphers
6415 ### Changes between 1.0.1k and 1.0.1l [15 Jan 2015]
6417 * Build fixes for the Windows and OpenVMS platforms
6419 *Matt Caswell and Richard Levitte*
6421 ### Changes between 1.0.1j and 1.0.1k [8 Jan 2015]
6423 * Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
6424 message can cause a segmentation fault in OpenSSL due to a NULL pointer
6425 dereference. This could lead to a Denial Of Service attack. Thanks to
6426 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
6431 * Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
6432 dtls1_buffer_record function under certain conditions. In particular this
6433 could occur if an attacker sent repeated DTLS records with the same
6434 sequence number but for the next epoch. The memory leak could be exploited
6435 by an attacker in a Denial of Service attack through memory exhaustion.
6436 Thanks to Chris Mueller for reporting this issue.
6441 * Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
6442 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
6443 method would be set to NULL which could later result in a NULL pointer
6444 dereference. Thanks to Frank Schmirler for reporting this issue.
6449 * Abort handshake if server key exchange message is omitted for ephemeral
6452 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
6453 reporting this issue.
6458 * Remove non-export ephemeral RSA code on client and server. This code
6459 violated the TLS standard by allowing the use of temporary RSA keys in
6460 non-export ciphersuites and could be used by a server to effectively
6461 downgrade the RSA key length used to a value smaller than the server
6462 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
6463 INRIA or reporting this issue.
6468 * Fixed issue where DH client certificates are accepted without verification.
6469 An OpenSSL server will accept a DH certificate for client authentication
6470 without the certificate verify message. This effectively allows a client to
6471 authenticate without the use of a private key. This only affects servers
6472 which trust a client certificate authority which issues certificates
6473 containing DH keys: these are extremely rare and hardly ever encountered.
6474 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
6480 * Ensure that the session ID context of an SSL is updated when its
6481 SSL_CTX is updated via SSL_set_SSL_CTX.
6483 The session ID context is typically set from the parent SSL_CTX,
6484 and can vary with the CTX.
6488 * Fix various certificate fingerprint issues.
6490 By using non-DER or invalid encodings outside the signed portion of a
6491 certificate the fingerprint can be changed without breaking the signature.
6492 Although no details of the signed portion of the certificate can be changed
6493 this can cause problems with some applications: e.g. those using the
6494 certificate fingerprint for blacklists.
6496 1. Reject signatures with non zero unused bits.
6498 If the BIT STRING containing the signature has non zero unused bits reject
6499 the signature. All current signature algorithms require zero unused bits.
6501 2. Check certificate algorithm consistency.
6503 Check the AlgorithmIdentifier inside TBS matches the one in the
6504 certificate signature. NB: this will result in signature failure
6505 errors for some broken certificates.
6507 Thanks to Konrad Kraszewski from Google for reporting this issue.
6509 3. Check DSA/ECDSA signatures use DER.
6511 Re-encode DSA/ECDSA signatures and compare with the original received
6512 signature. Return an error if there is a mismatch.
6514 This will reject various cases including garbage after signature
6515 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
6516 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
6517 (negative or with leading zeroes).
6519 Further analysis was conducted and fixes were developed by Stephen Henson
6520 of the OpenSSL core team.
6526 * Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
6527 results on some platforms, including x86_64. This bug occurs at random
6528 with a very low probability, and is not known to be exploitable in any
6529 way, though its exact impact is difficult to determine. Thanks to Pieter
6530 Wuille (Blockstream) who reported this issue and also suggested an initial
6531 fix. Further analysis was conducted by the OpenSSL development team and
6532 Adam Langley of Google. The final fix was developed by Andy Polyakov of
6533 the OpenSSL core team.
6538 * Do not resume sessions on the server if the negotiated protocol
6539 version does not match the session's version. Resuming with a different
6540 version, while not strictly forbidden by the RFC, is of questionable
6541 sanity and breaks all known clients.
6543 *David Benjamin, Emilia Käsper*
6545 * Tighten handling of the ChangeCipherSpec (CCS) message: reject
6546 early CCS messages during renegotiation. (Note that because
6547 renegotiation is encrypted, this early CCS was not exploitable.)
6551 * Tighten client-side session ticket handling during renegotiation:
6552 ensure that the client only accepts a session ticket if the server sends
6553 the extension anew in the ServerHello. Previously, a TLS client would
6554 reuse the old extension state and thus accept a session ticket if one was
6555 announced in the initial ServerHello.
6557 Similarly, ensure that the client requires a session ticket if one
6558 was advertised in the ServerHello. Previously, a TLS client would
6559 ignore a missing NewSessionTicket message.
6563 ### Changes between 1.0.1i and 1.0.1j [15 Oct 2014]
6567 A flaw in the DTLS SRTP extension parsing code allows an attacker, who
6568 sends a carefully crafted handshake message, to cause OpenSSL to fail
6569 to free up to 64k of memory causing a memory leak. This could be
6570 exploited in a Denial Of Service attack. This issue affects OpenSSL
6571 1.0.1 server implementations for both SSL/TLS and DTLS regardless of
6572 whether SRTP is used or configured. Implementations of OpenSSL that
6573 have been compiled with OPENSSL_NO_SRTP defined are not affected.
6575 The fix was developed by the OpenSSL team.
6580 * Session Ticket Memory Leak.
6582 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
6583 integrity of that ticket is first verified. In the event of a session
6584 ticket integrity check failing, OpenSSL will fail to free memory
6585 causing a memory leak. By sending a large number of invalid session
6586 tickets an attacker could exploit this issue in a Denial Of Service
6592 * Build option no-ssl3 is incomplete.
6594 When OpenSSL is configured with "no-ssl3" as a build option, servers
6595 could accept and complete a SSL 3.0 handshake, and clients could be
6596 configured to send them.
6599 *Akamai and the OpenSSL team*
6601 * Add support for TLS_FALLBACK_SCSV.
6602 Client applications doing fallback retries should call
6603 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
6606 *Adam Langley, Bodo Moeller*
6608 * Add additional DigestInfo checks.
6610 Re-encode DigestInto in DER and check against the original when
6611 verifying RSA signature: this will reject any improperly encoded
6612 DigestInfo structures.
6614 Note: this is a precautionary measure and no attacks are currently known.
6618 ### Changes between 1.0.1h and 1.0.1i [6 Aug 2014]
6620 * Fix SRP buffer overrun vulnerability. Invalid parameters passed to the
6621 SRP code can be overrun an internal buffer. Add sanity check that
6622 g, A, B < N to SRP code.
6624 Thanks to Sean Devlin and Watson Ladd of Cryptography Services, NCC
6625 Group for discovering this issue.
6630 * A flaw in the OpenSSL SSL/TLS server code causes the server to negotiate
6631 TLS 1.0 instead of higher protocol versions when the ClientHello message
6632 is badly fragmented. This allows a man-in-the-middle attacker to force a
6633 downgrade to TLS 1.0 even if both the server and the client support a
6634 higher protocol version, by modifying the client's TLS records.
6636 Thanks to David Benjamin and Adam Langley (Google) for discovering and
6637 researching this issue.
6642 * OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
6643 to a denial of service attack. A malicious server can crash the client
6644 with a null pointer dereference (read) by specifying an anonymous (EC)DH
6645 ciphersuite and sending carefully crafted handshake messages.
6647 Thanks to Felix Gröbert (Google) for discovering and researching this
6653 * By sending carefully crafted DTLS packets an attacker could cause openssl
6654 to leak memory. This can be exploited through a Denial of Service attack.
6655 Thanks to Adam Langley for discovering and researching this issue.
6660 * An attacker can force openssl to consume large amounts of memory whilst
6661 processing DTLS handshake messages. This can be exploited through a
6662 Denial of Service attack.
6663 Thanks to Adam Langley for discovering and researching this issue.
6668 * An attacker can force an error condition which causes openssl to crash
6669 whilst processing DTLS packets due to memory being freed twice. This
6670 can be exploited through a Denial of Service attack.
6671 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
6677 * If a multithreaded client connects to a malicious server using a resumed
6678 session and the server sends an ec point format extension it could write
6679 up to 255 bytes to freed memory.
6681 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
6687 * A malicious server can crash an OpenSSL client with a null pointer
6688 dereference (read) by specifying an SRP ciphersuite even though it was not
6689 properly negotiated with the client. This can be exploited through a
6690 Denial of Service attack.
6692 Thanks to Joonas Kuorilehto and Riku Hietamäki (Codenomicon) for
6693 discovering and researching this issue.
6698 * A flaw in OBJ_obj2txt may cause pretty printing functions such as
6699 X509_name_oneline, X509_name_print_ex et al. to leak some information
6700 from the stack. Applications may be affected if they echo pretty printing
6701 output to the attacker.
6703 Thanks to Ivan Fratric (Google) for discovering this issue.
6706 *Emilia Käsper, and Steve Henson*
6708 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
6709 for corner cases. (Certain input points at infinity could lead to
6710 bogus results, with non-infinity inputs mapped to infinity too.)
6714 ### Changes between 1.0.1g and 1.0.1h [5 Jun 2014]
6716 * Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
6717 handshake can force the use of weak keying material in OpenSSL
6718 SSL/TLS clients and servers.
6720 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
6721 researching this issue. ([CVE-2014-0224])
6723 *KIKUCHI Masashi, Steve Henson*
6725 * Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
6726 OpenSSL DTLS client the code can be made to recurse eventually crashing
6729 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
6732 *Imre Rad, Steve Henson*
6734 * Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
6735 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
6736 client or server. This is potentially exploitable to run arbitrary
6737 code on a vulnerable client or server.
6739 Thanks to Jüri Aedla for reporting this issue. ([CVE-2014-0195])
6741 *Jüri Aedla, Steve Henson*
6743 * Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
6744 are subject to a denial of service attack.
6746 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
6747 this issue. ([CVE-2014-3470])
6749 *Felix Gröbert, Ivan Fratric, Steve Henson*
6751 * Harmonize version and its documentation. -f flag is used to display
6754 *mancha <mancha1@zoho.com>*
6756 * Fix eckey_priv_encode so it immediately returns an error upon a failure
6757 in i2d_ECPrivateKey.
6759 *mancha <mancha1@zoho.com>*
6761 * Fix some double frees. These are not thought to be exploitable.
6763 *mancha <mancha1@zoho.com>*
6765 ### Changes between 1.0.1f and 1.0.1g [7 Apr 2014]
6767 * A missing bounds check in the handling of the TLS heartbeat extension
6768 can be used to reveal up to 64k of memory to a connected client or
6771 Thanks for Neel Mehta of Google Security for discovering this bug and to
6772 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
6773 preparing the fix ([CVE-2014-0160])
6775 *Adam Langley, Bodo Moeller*
6777 * Fix for the attack described in the paper "Recovering OpenSSL
6778 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
6779 by Yuval Yarom and Naomi Benger. Details can be obtained from:
6780 <http://eprint.iacr.org/2014/140>
6782 Thanks to Yuval Yarom and Naomi Benger for discovering this
6783 flaw and to Yuval Yarom for supplying a fix ([CVE-2014-0076])
6785 *Yuval Yarom and Naomi Benger*
6787 * TLS pad extension: draft-agl-tls-padding-03
6789 Workaround for the "TLS hang bug" (see FAQ and PR#2771): if the
6790 TLS client Hello record length value would otherwise be > 255 and
6791 less that 512 pad with a dummy extension containing zeroes so it
6792 is at least 512 bytes long.
6794 *Adam Langley, Steve Henson*
6796 ### Changes between 1.0.1e and 1.0.1f [6 Jan 2014]
6798 * Fix for TLS record tampering bug. A carefully crafted invalid
6799 handshake could crash OpenSSL with a NULL pointer exception.
6800 Thanks to Anton Johansson for reporting this issues.
6803 * Keep original DTLS digest and encryption contexts in retransmission
6804 structures so we can use the previous session parameters if they need
6805 to be resent. ([CVE-2013-6450])
6809 * Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
6810 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
6811 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
6812 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
6813 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
6814 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
6816 *Rob Stradling, Adam Langley*
6818 ### Changes between 1.0.1d and 1.0.1e [11 Feb 2013]
6820 * Correct fix for CVE-2013-0169. The original didn't work on AES-NI
6821 supporting platforms or when small records were transferred.
6823 *Andy Polyakov, Steve Henson*
6825 ### Changes between 1.0.1c and 1.0.1d [5 Feb 2013]
6827 * Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
6829 This addresses the flaw in CBC record processing discovered by
6830 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
6831 at: <http://www.isg.rhul.ac.uk/tls/>
6833 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
6834 Security Group at Royal Holloway, University of London
6835 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
6836 Emilia Käsper for the initial patch.
6839 *Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson*
6841 * Fix flaw in AESNI handling of TLS 1.2 and 1.1 records for CBC mode
6842 ciphersuites which can be exploited in a denial of service attack.
6843 Thanks go to and to Adam Langley <agl@chromium.org> for discovering
6844 and detecting this bug and to Wolfgang Ettlinger
6845 <wolfgang.ettlinger@gmail.com> for independently discovering this issue.
6850 * Return an error when checking OCSP signatures when key is NULL.
6851 This fixes a DoS attack. ([CVE-2013-0166])
6855 * Make openssl verify return errors.
6857 *Chris Palmer <palmer@google.com> and Ben Laurie*
6859 * Call OCSP Stapling callback after ciphersuite has been chosen, so
6860 the right response is stapled. Also change SSL_get_certificate()
6861 so it returns the certificate actually sent.
6862 See <http://rt.openssl.org/Ticket/Display.html?id=2836>.
6864 *Rob Stradling <rob.stradling@comodo.com>*
6866 * Fix possible deadlock when decoding public keys.
6870 * Don't use TLS 1.0 record version number in initial client hello
6875 ### Changes between 1.0.1b and 1.0.1c [10 May 2012]
6877 * Sanity check record length before skipping explicit IV in TLS
6878 1.2, 1.1 and DTLS to fix DoS attack.
6880 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
6881 fuzzing as a service testing platform.
6886 * Initialise tkeylen properly when encrypting CMS messages.
6887 Thanks to Solar Designer of Openwall for reporting this issue.
6891 * In FIPS mode don't try to use composite ciphers as they are not
6896 ### Changes between 1.0.1a and 1.0.1b [26 Apr 2012]
6898 * OpenSSL 1.0.0 sets SSL_OP_ALL to 0x80000FFFL and OpenSSL 1.0.1 and
6899 1.0.1a set SSL_OP_NO_TLSv1_1 to 0x00000400L which would unfortunately
6900 mean any application compiled against OpenSSL 1.0.0 headers setting
6901 SSL_OP_ALL would also set SSL_OP_NO_TLSv1_1, unintentionally disabling
6902 TLS 1.1 also. Fix this by changing the value of SSL_OP_NO_TLSv1_1 to
6903 0x10000000L Any application which was previously compiled against
6904 OpenSSL 1.0.1 or 1.0.1a headers and which cares about SSL_OP_NO_TLSv1_1
6905 will need to be recompiled as a result. Letting be results in
6906 inability to disable specifically TLS 1.1 and in client context,
6907 in unlike event, limit maximum offered version to TLS 1.0 [see below].
6911 * In order to ensure interoperability SSL_OP_NO_protocolX does not
6912 disable just protocol X, but all protocols above X *if* there are
6913 protocols *below* X still enabled. In more practical terms it means
6914 that if application wants to disable TLS1.0 in favor of TLS1.1 and
6915 above, it's not sufficient to pass `SSL_OP_NO_TLSv1`, one has to pass
6916 `SSL_OP_NO_TLSv1|SSL_OP_NO_SSLv3|SSL_OP_NO_SSLv2`. This applies to
6921 ### Changes between 1.0.1 and 1.0.1a [19 Apr 2012]
6923 * Check for potentially exploitable overflows in asn1_d2i_read_bio
6924 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
6925 in CRYPTO_realloc_clean.
6927 Thanks to Tavis Ormandy, Google Security Team, for discovering this
6928 issue and to Adam Langley <agl@chromium.org> for fixing it.
6931 *Adam Langley (Google), Tavis Ormandy, Google Security Team*
6933 * Don't allow TLS 1.2 SHA-256 ciphersuites in TLS 1.0, 1.1 connections.
6937 * Workarounds for some broken servers that "hang" if a client hello
6938 record length exceeds 255 bytes.
6940 1. Do not use record version number > TLS 1.0 in initial client
6941 hello: some (but not all) hanging servers will now work.
6942 2. If we set OPENSSL_MAX_TLS1_2_CIPHER_LENGTH this will truncate
6943 the number of ciphers sent in the client hello. This should be
6944 set to an even number, such as 50, for example by passing:
6945 -DOPENSSL_MAX_TLS1_2_CIPHER_LENGTH=50 to config or Configure.
6946 Most broken servers should now work.
6947 3. If all else fails setting OPENSSL_NO_TLS1_2_CLIENT will disable
6948 TLS 1.2 client support entirely.
6952 * Fix SEGV in Vector Permutation AES module observed in OpenSSH.
6956 ### Changes between 1.0.0h and 1.0.1 [14 Mar 2012]
6958 * Add compatibility with old MDC2 signatures which use an ASN1 OCTET
6959 STRING form instead of a DigestInfo.
6963 * The format used for MDC2 RSA signatures is inconsistent between EVP
6964 and the RSA_sign/RSA_verify functions. This was made more apparent when
6965 OpenSSL used RSA_sign/RSA_verify for some RSA signatures in particular
6966 those which went through EVP_PKEY_METHOD in 1.0.0 and later. Detect
6967 the correct format in RSA_verify so both forms transparently work.
6971 * Some servers which support TLS 1.0 can choke if we initially indicate
6972 support for TLS 1.2 and later renegotiate using TLS 1.0 in the RSA
6973 encrypted premaster secret. As a workaround use the maximum permitted
6974 client version in client hello, this should keep such servers happy
6975 and still work with previous versions of OpenSSL.
6979 * Add support for TLS/DTLS heartbeats.
6981 *Robin Seggelmann <seggelmann@fh-muenster.de>*
6983 * Add support for SCTP.
6985 *Robin Seggelmann <seggelmann@fh-muenster.de>*
6987 * Improved PRNG seeding for VOS.
6989 *Paul Green <Paul.Green@stratus.com>*
6991 * Extensive assembler packs updates, most notably:
6993 - x86[_64]: AES-NI, PCLMULQDQ, RDRAND support;
6994 - x86[_64]: SSSE3 support (SHA1, vector-permutation AES);
6995 - x86_64: bit-sliced AES implementation;
6996 - ARM: NEON support, contemporary platforms optimizations;
6997 - s390x: z196 support;
6998 - `*`: GHASH and GF(2^m) multiplication implementations;
7002 * Make TLS-SRP code conformant with RFC 5054 API cleanup
7003 (removal of unnecessary code)
7005 *Peter Sylvester <peter.sylvester@edelweb.fr>*
7007 * Add TLS key material exporter from RFC 5705.
7011 * Add DTLS-SRTP negotiation from RFC 5764.
7015 * Add Next Protocol Negotiation,
7016 <http://tools.ietf.org/html/draft-agl-tls-nextprotoneg-00>. Can be
7017 disabled with a no-npn flag to config or Configure. Code donated
7020 *Adam Langley <agl@google.com> and Ben Laurie*
7022 * Add optional 64-bit optimized implementations of elliptic curves NIST-P224,
7023 NIST-P256, NIST-P521, with constant-time single point multiplication on
7024 typical inputs. Compiler support for the nonstandard type `__uint128_t` is
7025 required to use this (present in gcc 4.4 and later, for 64-bit builds).
7026 Code made available under Apache License version 2.0.
7028 Specify "enable-ec_nistp_64_gcc_128" on the Configure (or config) command
7029 line to include this in your build of OpenSSL, and run "make depend" (or
7030 "make update"). This enables the following EC_METHODs:
7032 EC_GFp_nistp224_method()
7033 EC_GFp_nistp256_method()
7034 EC_GFp_nistp521_method()
7036 EC_GROUP_new_by_curve_name() will automatically use these (while
7037 EC_GROUP_new_curve_GFp() currently prefers the more flexible
7040 *Emilia Käsper, Adam Langley, Bodo Moeller (Google)*
7042 * Use type ossl_ssize_t instead of ssize_t which isn't available on
7043 all platforms. Move ssize_t definition from e_os.h to the public
7044 header file e_os2.h as it now appears in public header file cms.h
7048 * New -sigopt option to the ca, req and x509 utilities. Additional
7049 signature parameters can be passed using this option and in
7054 * Add RSA PSS signing function. This will generate and set the
7055 appropriate AlgorithmIdentifiers for PSS based on those in the
7056 corresponding EVP_MD_CTX structure. No application support yet.
7060 * Support for companion algorithm specific ASN1 signing routines.
7061 New function ASN1_item_sign_ctx() signs a pre-initialised
7062 EVP_MD_CTX structure and sets AlgorithmIdentifiers based on
7063 the appropriate parameters.
7067 * Add new algorithm specific ASN1 verification initialisation function
7068 to EVP_PKEY_ASN1_METHOD: this is not in EVP_PKEY_METHOD since the ASN1
7069 handling will be the same no matter what EVP_PKEY_METHOD is used.
7070 Add a PSS handler to support verification of PSS signatures: checked
7071 against a number of sample certificates.
7075 * Add signature printing for PSS. Add PSS OIDs.
7077 *Steve Henson, Martin Kaiser <lists@kaiser.cx>*
7079 * Add algorithm specific signature printing. An individual ASN1 method
7080 can now print out signatures instead of the standard hex dump.
7082 More complex signatures (e.g. PSS) can print out more meaningful
7083 information. Include DSA version that prints out the signature
7088 * Password based recipient info support for CMS library: implementing
7093 * Split password based encryption into PBES2 and PBKDF2 functions. This
7094 neatly separates the code into cipher and PBE sections and is required
7095 for some algorithms that split PBES2 into separate pieces (such as
7096 password based CMS).
7100 * Session-handling fixes:
7101 - Fix handling of connections that are resuming with a session ID,
7102 but also support Session Tickets.
7103 - Fix a bug that suppressed issuing of a new ticket if the client
7104 presented a ticket with an expired session.
7105 - Try to set the ticket lifetime hint to something reasonable.
7106 - Make tickets shorter by excluding irrelevant information.
7107 - On the client side, don't ignore renewed tickets.
7109 *Adam Langley, Bodo Moeller (Google)*
7111 * Fix PSK session representation.
7115 * Add RC4-MD5 and AESNI-SHA1 "stitched" implementations.
7117 This work was sponsored by Intel.
7121 * Add GCM support to TLS library. Some custom code is needed to split
7122 the IV between the fixed (from PRF) and explicit (from TLS record)
7123 portions. This adds all GCM ciphersuites supported by RFC5288 and
7124 RFC5289. Generalise some `AES*` cipherstrings to include GCM and
7125 add a special AESGCM string for GCM only.
7129 * Expand range of ctrls for AES GCM. Permit setting invocation
7130 field on decrypt and retrieval of invocation field only on encrypt.
7134 * Add HMAC ECC ciphersuites from RFC5289. Include SHA384 PRF support.
7135 As required by RFC5289 these ciphersuites cannot be used if for
7136 versions of TLS earlier than 1.2.
7140 * For FIPS capable OpenSSL interpret a NULL default public key method
7141 as unset and return the appropriate default but do *not* set the default.
7142 This means we can return the appropriate method in applications that
7143 switch between FIPS and non-FIPS modes.
7147 * Redirect HMAC and CMAC operations to FIPS module in FIPS mode. If an
7148 ENGINE is used then we cannot handle that in the FIPS module so we
7149 keep original code iff non-FIPS operations are allowed.
7153 * Add -attime option to openssl utilities.
7155 *Peter Eckersley <pde@eff.org>, Ben Laurie and Steve Henson*
7157 * Redirect DSA and DH operations to FIPS module in FIPS mode.
7161 * Redirect ECDSA and ECDH operations to FIPS module in FIPS mode. Also use
7162 FIPS EC methods unconditionally for now.
7166 * New build option no-ec2m to disable characteristic 2 code.
7170 * Backport libcrypto audit of return value checking from 1.1.0-dev; not
7171 all cases can be covered as some introduce binary incompatibilities.
7175 * Redirect RSA operations to FIPS module including keygen,
7176 encrypt, decrypt, sign and verify. Block use of non FIPS RSA methods.
7180 * Add similar low-level API blocking to ciphers.
7184 * low-level digest APIs are not approved in FIPS mode: any attempt
7185 to use these will cause a fatal error. Applications that *really* want
7186 to use them can use the `private_*` version instead.
7190 * Redirect cipher operations to FIPS module for FIPS builds.
7194 * Redirect digest operations to FIPS module for FIPS builds.
7198 * Update build system to add "fips" flag which will link in fipscanister.o
7199 for static and shared library builds embedding a signature if needed.
7203 * Output TLS supported curves in preference order instead of numerical
7204 order. This is currently hardcoded for the highest order curves first.
7205 This should be configurable so applications can judge speed vs strength.
7209 * Add TLS v1.2 server support for client authentication.
7213 * Add support for FIPS mode in ssl library: disable SSLv3, non-FIPS ciphers
7218 * Functions FIPS_mode_set() and FIPS_mode() which call the underlying
7219 FIPS modules versions.
7223 * Add TLS v1.2 client side support for client authentication. Keep cache
7224 of handshake records longer as we don't know the hash algorithm to use
7225 until after the certificate request message is received.
7229 * Initial TLS v1.2 client support. Add a default signature algorithms
7230 extension including all the algorithms we support. Parse new signature
7231 format in client key exchange. Relax some ECC signing restrictions for
7232 TLS v1.2 as indicated in RFC5246.
7236 * Add server support for TLS v1.2 signature algorithms extension. Switch
7237 to new signature format when needed using client digest preference.
7238 All server ciphersuites should now work correctly in TLS v1.2. No client
7239 support yet and no support for client certificates.
7243 * Initial TLS v1.2 support. Add new SHA256 digest to ssl code, switch
7244 to SHA256 for PRF when using TLS v1.2 and later. Add new SHA256 based
7245 ciphersuites. At present only RSA key exchange ciphersuites work with
7246 TLS v1.2. Add new option for TLS v1.2 replacing the old and obsolete
7247 SSL_OP_PKCS1_CHECK flags with SSL_OP_NO_TLSv1_2. New TLSv1.2 methods
7248 and version checking.
7252 * New option OPENSSL_NO_SSL_INTERN. If an application can be compiled
7253 with this defined it will not be affected by any changes to ssl internal
7254 structures. Add several utility functions to allow openssl application
7255 to work with OPENSSL_NO_SSL_INTERN defined.
7259 * A long standing patch to add support for SRP from EdelWeb (Peter
7260 Sylvester and Christophe Renou) was integrated.
7261 *Christophe Renou <christophe.renou@edelweb.fr>, Peter Sylvester
7262 <peter.sylvester@edelweb.fr>, Tom Wu <tjw@cs.stanford.edu>, and
7265 * Add functions to copy EVP_PKEY_METHOD and retrieve flags and id.
7269 * Permit abbreviated handshakes when renegotiating using the function
7270 SSL_renegotiate_abbreviated().
7272 *Robin Seggelmann <seggelmann@fh-muenster.de>*
7274 * Add call to ENGINE_register_all_complete() to
7275 ENGINE_load_builtin_engines(), so some implementations get used
7276 automatically instead of needing explicit application support.
7280 * Add support for TLS key exporter as described in RFC5705.
7282 *Robin Seggelmann <seggelmann@fh-muenster.de>, Steve Henson*
7284 * Initial TLSv1.1 support. Since TLSv1.1 is very similar to TLS v1.0 only
7285 a few changes are required:
7287 Add SSL_OP_NO_TLSv1_1 flag.
7288 Add TLSv1_1 methods.
7289 Update version checking logic to handle version 1.1.
7290 Add explicit IV handling (ported from DTLS code).
7291 Add command line options to s_client/s_server.
7298 ### Changes between 1.0.0s and 1.0.0t [3 Dec 2015]
7300 * X509_ATTRIBUTE memory leak
7302 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
7303 memory. This structure is used by the PKCS#7 and CMS routines so any
7304 application which reads PKCS#7 or CMS data from untrusted sources is
7305 affected. SSL/TLS is not affected.
7307 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
7313 * Race condition handling PSK identify hint
7315 If PSK identity hints are received by a multi-threaded client then
7316 the values are wrongly updated in the parent SSL_CTX structure. This can
7317 result in a race condition potentially leading to a double free of the
7323 ### Changes between 1.0.0r and 1.0.0s [11 Jun 2015]
7325 * Malformed ECParameters causes infinite loop
7327 When processing an ECParameters structure OpenSSL enters an infinite loop
7328 if the curve specified is over a specially malformed binary polynomial
7331 This can be used to perform denial of service against any
7332 system which processes public keys, certificate requests or
7333 certificates. This includes TLS clients and TLS servers with
7334 client authentication enabled.
7336 This issue was reported to OpenSSL by Joseph Barr-Pixton.
7341 * Exploitable out-of-bounds read in X509_cmp_time
7343 X509_cmp_time does not properly check the length of the ASN1_TIME
7344 string and can read a few bytes out of bounds. In addition,
7345 X509_cmp_time accepts an arbitrary number of fractional seconds in the
7348 An attacker can use this to craft malformed certificates and CRLs of
7349 various sizes and potentially cause a segmentation fault, resulting in
7350 a DoS on applications that verify certificates or CRLs. TLS clients
7351 that verify CRLs are affected. TLS clients and servers with client
7352 authentication enabled may be affected if they use custom verification
7355 This issue was reported to OpenSSL by Robert Swiecki (Google), and
7356 independently by Hanno Böck.
7361 * PKCS7 crash with missing EnvelopedContent
7363 The PKCS#7 parsing code does not handle missing inner EncryptedContent
7364 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
7365 with missing content and trigger a NULL pointer dereference on parsing.
7367 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
7368 structures from untrusted sources are affected. OpenSSL clients and
7369 servers are not affected.
7371 This issue was reported to OpenSSL by Michal Zalewski (Google).
7376 * CMS verify infinite loop with unknown hash function
7378 When verifying a signedData message the CMS code can enter an infinite loop
7379 if presented with an unknown hash function OID. This can be used to perform
7380 denial of service against any system which verifies signedData messages using
7382 This issue was reported to OpenSSL by Johannes Bauer.
7387 * Race condition handling NewSessionTicket
7389 If a NewSessionTicket is received by a multi-threaded client when attempting to
7390 reuse a previous ticket then a race condition can occur potentially leading to
7391 a double free of the ticket data.
7396 ### Changes between 1.0.0q and 1.0.0r [19 Mar 2015]
7398 * Segmentation fault in ASN1_TYPE_cmp fix
7400 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
7401 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
7402 certificate signature algorithm consistency this can be used to crash any
7403 certificate verification operation and exploited in a DoS attack. Any
7404 application which performs certificate verification is vulnerable including
7405 OpenSSL clients and servers which enable client authentication.
7410 * ASN.1 structure reuse memory corruption fix
7412 Reusing a structure in ASN.1 parsing may allow an attacker to cause
7413 memory corruption via an invalid write. Such reuse is and has been
7414 strongly discouraged and is believed to be rare.
7416 Applications that parse structures containing CHOICE or ANY DEFINED BY
7417 components may be affected. Certificate parsing (d2i_X509 and related
7418 functions) are however not affected. OpenSSL clients and servers are
7424 * PKCS7 NULL pointer dereferences fix
7426 The PKCS#7 parsing code does not handle missing outer ContentInfo
7427 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
7428 missing content and trigger a NULL pointer dereference on parsing.
7430 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
7431 otherwise parse PKCS#7 structures from untrusted sources are
7432 affected. OpenSSL clients and servers are not affected.
7434 This issue was reported to OpenSSL by Michal Zalewski (Google).
7439 * DoS via reachable assert in SSLv2 servers fix
7441 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
7442 servers that both support SSLv2 and enable export cipher suites by sending
7443 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
7445 This issue was discovered by Sean Burford (Google) and Emilia Käsper
7446 (OpenSSL development team).
7451 * Use After Free following d2i_ECPrivatekey error fix
7453 A malformed EC private key file consumed via the d2i_ECPrivateKey function
7454 could cause a use after free condition. This, in turn, could cause a double
7455 free in several private key parsing functions (such as d2i_PrivateKey
7456 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
7457 for applications that receive EC private keys from untrusted
7458 sources. This scenario is considered rare.
7460 This issue was discovered by the BoringSSL project and fixed in their
7466 * X509_to_X509_REQ NULL pointer deref fix
7468 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
7469 the certificate key is invalid. This function is rarely used in practice.
7471 This issue was discovered by Brian Carpenter.
7476 * Removed the export ciphers from the DEFAULT ciphers
7480 ### Changes between 1.0.0p and 1.0.0q [15 Jan 2015]
7482 * Build fixes for the Windows and OpenVMS platforms
7484 *Matt Caswell and Richard Levitte*
7486 ### Changes between 1.0.0o and 1.0.0p [8 Jan 2015]
7488 * Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
7489 message can cause a segmentation fault in OpenSSL due to a NULL pointer
7490 dereference. This could lead to a Denial Of Service attack. Thanks to
7491 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
7496 * Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
7497 dtls1_buffer_record function under certain conditions. In particular this
7498 could occur if an attacker sent repeated DTLS records with the same
7499 sequence number but for the next epoch. The memory leak could be exploited
7500 by an attacker in a Denial of Service attack through memory exhaustion.
7501 Thanks to Chris Mueller for reporting this issue.
7506 * Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
7507 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
7508 method would be set to NULL which could later result in a NULL pointer
7509 dereference. Thanks to Frank Schmirler for reporting this issue.
7514 * Abort handshake if server key exchange message is omitted for ephemeral
7517 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
7518 reporting this issue.
7523 * Remove non-export ephemeral RSA code on client and server. This code
7524 violated the TLS standard by allowing the use of temporary RSA keys in
7525 non-export ciphersuites and could be used by a server to effectively
7526 downgrade the RSA key length used to a value smaller than the server
7527 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
7528 INRIA or reporting this issue.
7533 * Fixed issue where DH client certificates are accepted without verification.
7534 An OpenSSL server will accept a DH certificate for client authentication
7535 without the certificate verify message. This effectively allows a client to
7536 authenticate without the use of a private key. This only affects servers
7537 which trust a client certificate authority which issues certificates
7538 containing DH keys: these are extremely rare and hardly ever encountered.
7539 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
7545 * Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
7546 results on some platforms, including x86_64. This bug occurs at random
7547 with a very low probability, and is not known to be exploitable in any
7548 way, though its exact impact is difficult to determine. Thanks to Pieter
7549 Wuille (Blockstream) who reported this issue and also suggested an initial
7550 fix. Further analysis was conducted by the OpenSSL development team and
7551 Adam Langley of Google. The final fix was developed by Andy Polyakov of
7552 the OpenSSL core team.
7557 * Fix various certificate fingerprint issues.
7559 By using non-DER or invalid encodings outside the signed portion of a
7560 certificate the fingerprint can be changed without breaking the signature.
7561 Although no details of the signed portion of the certificate can be changed
7562 this can cause problems with some applications: e.g. those using the
7563 certificate fingerprint for blacklists.
7565 1. Reject signatures with non zero unused bits.
7567 If the BIT STRING containing the signature has non zero unused bits reject
7568 the signature. All current signature algorithms require zero unused bits.
7570 2. Check certificate algorithm consistency.
7572 Check the AlgorithmIdentifier inside TBS matches the one in the
7573 certificate signature. NB: this will result in signature failure
7574 errors for some broken certificates.
7576 Thanks to Konrad Kraszewski from Google for reporting this issue.
7578 3. Check DSA/ECDSA signatures use DER.
7580 Reencode DSA/ECDSA signatures and compare with the original received
7581 signature. Return an error if there is a mismatch.
7583 This will reject various cases including garbage after signature
7584 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
7585 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
7586 (negative or with leading zeroes).
7588 Further analysis was conducted and fixes were developed by Stephen Henson
7589 of the OpenSSL core team.
7595 ### Changes between 1.0.0n and 1.0.0o [15 Oct 2014]
7597 * Session Ticket Memory Leak.
7599 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
7600 integrity of that ticket is first verified. In the event of a session
7601 ticket integrity check failing, OpenSSL will fail to free memory
7602 causing a memory leak. By sending a large number of invalid session
7603 tickets an attacker could exploit this issue in a Denial Of Service
7609 * Build option no-ssl3 is incomplete.
7611 When OpenSSL is configured with "no-ssl3" as a build option, servers
7612 could accept and complete a SSL 3.0 handshake, and clients could be
7613 configured to send them.
7616 *Akamai and the OpenSSL team*
7618 * Add support for TLS_FALLBACK_SCSV.
7619 Client applications doing fallback retries should call
7620 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
7623 *Adam Langley, Bodo Moeller*
7625 * Add additional DigestInfo checks.
7627 Reencode DigestInto in DER and check against the original when
7628 verifying RSA signature: this will reject any improperly encoded
7629 DigestInfo structures.
7631 Note: this is a precautionary measure and no attacks are currently known.
7635 ### Changes between 1.0.0m and 1.0.0n [6 Aug 2014]
7637 * OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
7638 to a denial of service attack. A malicious server can crash the client
7639 with a null pointer dereference (read) by specifying an anonymous (EC)DH
7640 ciphersuite and sending carefully crafted handshake messages.
7642 Thanks to Felix Gröbert (Google) for discovering and researching this
7648 * By sending carefully crafted DTLS packets an attacker could cause openssl
7649 to leak memory. This can be exploited through a Denial of Service attack.
7650 Thanks to Adam Langley for discovering and researching this issue.
7655 * An attacker can force openssl to consume large amounts of memory whilst
7656 processing DTLS handshake messages. This can be exploited through a
7657 Denial of Service attack.
7658 Thanks to Adam Langley for discovering and researching this issue.
7663 * An attacker can force an error condition which causes openssl to crash
7664 whilst processing DTLS packets due to memory being freed twice. This
7665 can be exploited through a Denial of Service attack.
7666 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
7672 * If a multithreaded client connects to a malicious server using a resumed
7673 session and the server sends an ec point format extension it could write
7674 up to 255 bytes to freed memory.
7676 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
7682 * A flaw in OBJ_obj2txt may cause pretty printing functions such as
7683 X509_name_oneline, X509_name_print_ex et al. to leak some information
7684 from the stack. Applications may be affected if they echo pretty printing
7685 output to the attacker.
7687 Thanks to Ivan Fratric (Google) for discovering this issue.
7690 *Emilia Käsper, and Steve Henson*
7692 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
7693 for corner cases. (Certain input points at infinity could lead to
7694 bogus results, with non-infinity inputs mapped to infinity too.)
7698 ### Changes between 1.0.0l and 1.0.0m [5 Jun 2014]
7700 * Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
7701 handshake can force the use of weak keying material in OpenSSL
7702 SSL/TLS clients and servers.
7704 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
7705 researching this issue. ([CVE-2014-0224])
7707 *KIKUCHI Masashi, Steve Henson*
7709 * Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
7710 OpenSSL DTLS client the code can be made to recurse eventually crashing
7713 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
7716 *Imre Rad, Steve Henson*
7718 * Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
7719 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
7720 client or server. This is potentially exploitable to run arbitrary
7721 code on a vulnerable client or server.
7723 Thanks to Jüri Aedla for reporting this issue. ([CVE-2014-0195])
7725 *Jüri Aedla, Steve Henson*
7727 * Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
7728 are subject to a denial of service attack.
7730 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
7731 this issue. ([CVE-2014-3470])
7733 *Felix Gröbert, Ivan Fratric, Steve Henson*
7735 * Harmonize version and its documentation. -f flag is used to display
7738 *mancha <mancha1@zoho.com>*
7740 * Fix eckey_priv_encode so it immediately returns an error upon a failure
7741 in i2d_ECPrivateKey.
7743 *mancha <mancha1@zoho.com>*
7745 * Fix some double frees. These are not thought to be exploitable.
7747 *mancha <mancha1@zoho.com>*
7749 * Fix for the attack described in the paper "Recovering OpenSSL
7750 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
7751 by Yuval Yarom and Naomi Benger. Details can be obtained from:
7752 <http://eprint.iacr.org/2014/140>
7754 Thanks to Yuval Yarom and Naomi Benger for discovering this
7755 flaw and to Yuval Yarom for supplying a fix ([CVE-2014-0076])
7757 *Yuval Yarom and Naomi Benger*
7759 ### Changes between 1.0.0k and 1.0.0l [6 Jan 2014]
7761 * Keep original DTLS digest and encryption contexts in retransmission
7762 structures so we can use the previous session parameters if they need
7763 to be resent. ([CVE-2013-6450])
7767 * Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
7768 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
7769 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
7770 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
7771 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
7772 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
7774 *Rob Stradling, Adam Langley*
7776 ### Changes between 1.0.0j and 1.0.0k [5 Feb 2013]
7778 * Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
7780 This addresses the flaw in CBC record processing discovered by
7781 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
7782 at: <http://www.isg.rhul.ac.uk/tls/>
7784 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
7785 Security Group at Royal Holloway, University of London
7786 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
7787 Emilia Käsper for the initial patch.
7790 *Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson*
7792 * Return an error when checking OCSP signatures when key is NULL.
7793 This fixes a DoS attack. ([CVE-2013-0166])
7797 * Call OCSP Stapling callback after ciphersuite has been chosen, so
7798 the right response is stapled. Also change SSL_get_certificate()
7799 so it returns the certificate actually sent.
7800 See <http://rt.openssl.org/Ticket/Display.html?id=2836>.
7801 (This is a backport)
7803 *Rob Stradling <rob.stradling@comodo.com>*
7805 * Fix possible deadlock when decoding public keys.
7809 ### Changes between 1.0.0i and 1.0.0j [10 May 2012]
7811 [NB: OpenSSL 1.0.0i and later 1.0.0 patch levels were released after
7814 * Sanity check record length before skipping explicit IV in DTLS
7817 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
7818 fuzzing as a service testing platform.
7823 * Initialise tkeylen properly when encrypting CMS messages.
7824 Thanks to Solar Designer of Openwall for reporting this issue.
7828 ### Changes between 1.0.0h and 1.0.0i [19 Apr 2012]
7830 * Check for potentially exploitable overflows in asn1_d2i_read_bio
7831 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
7832 in CRYPTO_realloc_clean.
7834 Thanks to Tavis Ormandy, Google Security Team, for discovering this
7835 issue and to Adam Langley <agl@chromium.org> for fixing it.
7838 *Adam Langley (Google), Tavis Ormandy, Google Security Team*
7840 ### Changes between 1.0.0g and 1.0.0h [12 Mar 2012]
7842 * Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness
7843 in CMS and PKCS7 code. When RSA decryption fails use a random key for
7844 content decryption and always return the same error. Note: this attack
7845 needs on average 2^20 messages so it only affects automated senders. The
7846 old behaviour can be re-enabled in the CMS code by setting the
7847 CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where
7848 an MMA defence is not necessary.
7849 Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering
7850 this issue. ([CVE-2012-0884])
7854 * Fix CVE-2011-4619: make sure we really are receiving a
7855 client hello before rejecting multiple SGC restarts. Thanks to
7856 Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug.
7860 ### Changes between 1.0.0f and 1.0.0g [18 Jan 2012]
7862 * Fix for DTLS DoS issue introduced by fix for CVE-2011-4109.
7863 Thanks to Antonio Martin, Enterprise Secure Access Research and
7864 Development, Cisco Systems, Inc. for discovering this bug and
7865 preparing a fix. ([CVE-2012-0050])
7869 ### Changes between 1.0.0e and 1.0.0f [4 Jan 2012]
7871 * Nadhem Alfardan and Kenny Paterson have discovered an extension
7872 of the Vaudenay padding oracle attack on CBC mode encryption
7873 which enables an efficient plaintext recovery attack against
7874 the OpenSSL implementation of DTLS. Their attack exploits timing
7875 differences arising during decryption processing. A research
7876 paper describing this attack can be found at:
7877 <http://www.isg.rhul.ac.uk/~kp/dtls.pdf>
7878 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
7879 Security Group at Royal Holloway, University of London
7880 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
7881 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
7882 for preparing the fix. ([CVE-2011-4108])
7884 *Robin Seggelmann, Michael Tuexen*
7886 * Clear bytes used for block padding of SSL 3.0 records.
7889 *Adam Langley (Google)*
7891 * Only allow one SGC handshake restart for SSL/TLS. Thanks to George
7892 Kadianakis <desnacked@gmail.com> for discovering this issue and
7893 Adam Langley for preparing the fix. ([CVE-2011-4619])
7895 *Adam Langley (Google)*
7897 * Check parameters are not NULL in GOST ENGINE. ([CVE-2012-0027])
7899 *Andrey Kulikov <amdeich@gmail.com>*
7901 * Prevent malformed RFC3779 data triggering an assertion failure.
7902 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
7903 and Rob Austein <sra@hactrn.net> for fixing it. ([CVE-2011-4577])
7905 *Rob Austein <sra@hactrn.net>*
7907 * Improved PRNG seeding for VOS.
7909 *Paul Green <Paul.Green@stratus.com>*
7911 * Fix ssl_ciph.c set-up race.
7913 *Adam Langley (Google)*
7915 * Fix spurious failures in ecdsatest.c.
7917 *Emilia Käsper (Google)*
7919 * Fix the BIO_f_buffer() implementation (which was mixing different
7920 interpretations of the `..._len` fields).
7922 *Adam Langley (Google)*
7924 * Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
7925 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
7926 threads won't reuse the same blinding coefficients.
7928 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
7929 lock to call BN_BLINDING_invert_ex, and avoids one use of
7930 BN_BLINDING_update for each BN_BLINDING structure (previously,
7931 the last update always remained unused).
7933 *Emilia Käsper (Google)*
7935 * In ssl3_clear, preserve s3->init_extra along with s3->rbuf.
7937 *Bob Buckholz (Google)*
7939 ### Changes between 1.0.0d and 1.0.0e [6 Sep 2011]
7941 * Fix bug where CRLs with nextUpdate in the past are sometimes accepted
7942 by initialising X509_STORE_CTX properly. ([CVE-2011-3207])
7944 *Kaspar Brand <ossl@velox.ch>*
7946 * Fix SSL memory handling for (EC)DH ciphersuites, in particular
7947 for multi-threaded use of ECDH. ([CVE-2011-3210])
7949 *Adam Langley (Google)*
7951 * Fix x509_name_ex_d2i memory leak on bad inputs.
7955 * Remove hard coded ecdsaWithSHA1 signature tests in ssl code and check
7956 signature public key algorithm by using OID xref utilities instead.
7957 Before this you could only use some ECC ciphersuites with SHA1 only.
7961 * Add protection against ECDSA timing attacks as mentioned in the paper
7962 by Billy Bob Brumley and Nicola Tuveri, see:
7963 <http://eprint.iacr.org/2011/232.pdf>
7965 *Billy Bob Brumley and Nicola Tuveri*
7967 ### Changes between 1.0.0c and 1.0.0d [8 Feb 2011]
7969 * Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
7971 *Neel Mehta, Adam Langley, Bodo Moeller (Google)*
7973 * Fix bug in string printing code: if *any* escaping is enabled we must
7974 escape the escape character (backslash) or the resulting string is
7979 ### Changes between 1.0.0b and 1.0.0c [2 Dec 2010]
7981 * Disable code workaround for ancient and obsolete Netscape browsers
7982 and servers: an attacker can use it in a ciphersuite downgrade attack.
7983 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
7987 * Fixed J-PAKE implementation error, originally discovered by
7988 Sebastien Martini, further info and confirmation from Stefan
7989 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
7993 ### Changes between 1.0.0a and 1.0.0b [16 Nov 2010]
7995 * Fix extension code to avoid race conditions which can result in a buffer
7996 overrun vulnerability: resumed sessions must not be modified as they can
7997 be shared by multiple threads. CVE-2010-3864
8001 * Fix WIN32 build system to correctly link an ENGINE directory into
8006 ### Changes between 1.0.0 and 1.0.0a [01 Jun 2010]
8008 * Check return value of int_rsa_verify in pkey_rsa_verifyrecover
8011 *Steve Henson, Peter-Michael Hager <hager@dortmund.net>*
8013 ### Changes between 0.9.8n and 1.0.0 [29 Mar 2010]
8015 * Add "missing" function EVP_CIPHER_CTX_copy(). This copies a cipher
8016 context. The operation can be customised via the ctrl mechanism in
8017 case ENGINEs want to include additional functionality.
8021 * Tolerate yet another broken PKCS#8 key format: private key value negative.
8025 * Add new -subject_hash_old and -issuer_hash_old options to x509 utility to
8026 output hashes compatible with older versions of OpenSSL.
8028 *Willy Weisz <weisz@vcpc.univie.ac.at>*
8030 * Fix compression algorithm handling: if resuming a session use the
8031 compression algorithm of the resumed session instead of determining
8032 it from client hello again. Don't allow server to change algorithm.
8036 * Add load_crls() function to commands tidying load_certs() too. Add option
8037 to verify utility to allow additional CRLs to be included.
8041 * Update OCSP request code to permit adding custom headers to the request:
8042 some responders need this.
8046 * The function EVP_PKEY_sign() returns <=0 on error: check return code
8049 *Julia Lawall <julia@diku.dk>*
8051 * Update verify callback code in `apps/s_cb.c` and `apps/verify.c`, it
8052 needlessly dereferenced structures, used obsolete functions and
8053 didn't handle all updated verify codes correctly.
8057 * Disable MD2 in the default configuration.
8061 * In BIO_pop() and BIO_push() use the ctrl argument (which was NULL) to
8062 indicate the initial BIO being pushed or popped. This makes it possible
8063 to determine whether the BIO is the one explicitly called or as a result
8064 of the ctrl being passed down the chain. Fix BIO_pop() and SSL BIOs so
8065 it handles reference counts correctly and doesn't zero out the I/O bio
8066 when it is not being explicitly popped. WARNING: applications which
8067 included workarounds for the old buggy behaviour will need to be modified
8068 or they could free up already freed BIOs.
8072 * Extend the uni2asc/asc2uni => OPENSSL_uni2asc/OPENSSL_asc2uni
8073 renaming to all platforms (within the 0.9.8 branch, this was
8074 done conditionally on Netware platforms to avoid a name clash).
8076 *Guenter <lists@gknw.net>*
8078 * Add ECDHE and PSK support to DTLS.
8080 *Michael Tuexen <tuexen@fh-muenster.de>*
8082 * Add CHECKED_STACK_OF macro to safestack.h, otherwise safestack can't
8087 * Add "missing" function EVP_MD_flags() (without this the only way to
8088 retrieve a digest flags is by accessing the structure directly. Update
8089 `EVP_MD_do_all*()` and `EVP_CIPHER_do_all*()` to include the name a digest
8090 or cipher is registered as in the "from" argument. Print out all
8091 registered digests in the dgst usage message instead of manually
8092 attempting to work them out.
8096 * If no SSLv2 ciphers are used don't use an SSLv2 compatible client hello:
8097 this allows the use of compression and extensions. Change default cipher
8098 string to remove SSLv2 ciphersuites. This effectively avoids ancient SSLv2
8099 by default unless an application cipher string requests it.
8103 * Alter match criteria in PKCS12_parse(). It used to try to use local
8104 key ids to find matching certificates and keys but some PKCS#12 files
8105 don't follow the (somewhat unwritten) rules and this strategy fails.
8106 Now just gather all certificates together and the first private key
8107 then look for the first certificate that matches the key.
8111 * Support use of registered digest and cipher names for dgst and cipher
8112 commands instead of having to add each one as a special case. So now
8119 openssl dgst -sha256 foo
8121 and this works for ENGINE based algorithms too.
8125 * Update Gost ENGINE to support parameter files.
8127 *Victor B. Wagner <vitus@cryptocom.ru>*
8129 * Support GeneralizedTime in ca utility.
8131 *Oliver Martin <oliver@volatilevoid.net>, Steve Henson*
8133 * Enhance the hash format used for certificate directory links. The new
8134 form uses the canonical encoding (meaning equivalent names will work
8135 even if they aren't identical) and uses SHA1 instead of MD5. This form
8136 is incompatible with the older format and as a result c_rehash should
8137 be used to rebuild symbolic links.
8141 * Make PKCS#8 the default write format for private keys, replacing the
8142 traditional format. This form is standardised, more secure and doesn't
8143 include an implicit MD5 dependency.
8147 * Add a $gcc_devteam_warn option to Configure. The idea is that any code
8148 committed to OpenSSL should pass this lot as a minimum.
8152 * Add session ticket override functionality for use by EAP-FAST.
8154 *Jouni Malinen <j@w1.fi>*
8156 * Modify HMAC functions to return a value. Since these can be implemented
8157 in an ENGINE errors can occur.
8161 * Type-checked OBJ_bsearch_ex.
8165 * Type-checked OBJ_bsearch. Also some constification necessitated
8166 by type-checking. Still to come: TXT_DB, bsearch(?),
8167 OBJ_bsearch_ex, qsort, CRYPTO_EX_DATA, ASN1_VALUE, ASN1_STRING,
8172 * New function OPENSSL_gmtime_adj() to add a specific number of days and
8173 seconds to a tm structure directly, instead of going through OS
8174 specific date routines. This avoids any issues with OS routines such
8175 as the year 2038 bug. New `*_adj()` functions for ASN1 time structures
8176 and X509_time_adj_ex() to cover the extended range. The existing
8177 X509_time_adj() is still usable and will no longer have any date issues.
8181 * Delta CRL support. New use deltas option which will attempt to locate
8182 and search any appropriate delta CRLs available.
8184 This work was sponsored by Google.
8188 * Support for CRLs partitioned by reason code. Reorganise CRL processing
8189 code and add additional score elements. Validate alternate CRL paths
8190 as part of the CRL checking and indicate a new error "CRL path validation
8191 error" in this case. Applications wanting additional details can use
8192 the verify callback and check the new "parent" field. If this is not
8193 NULL CRL path validation is taking place. Existing applications won't
8194 see this because it requires extended CRL support which is off by
8197 This work was sponsored by Google.
8201 * Support for freshest CRL extension.
8203 This work was sponsored by Google.
8207 * Initial indirect CRL support. Currently only supported in the CRLs
8208 passed directly and not via lookup. Process certificate issuer
8209 CRL entry extension and lookup CRL entries by bother issuer name
8210 and serial number. Check and process CRL issuer entry in IDP extension.
8212 This work was sponsored by Google.
8216 * Add support for distinct certificate and CRL paths. The CRL issuer
8217 certificate is validated separately in this case. Only enabled if
8218 an extended CRL support flag is set: this flag will enable additional
8219 CRL functionality in future.
8221 This work was sponsored by Google.
8225 * Add support for policy mappings extension.
8227 This work was sponsored by Google.
8231 * Fixes to pathlength constraint, self issued certificate handling,
8232 policy processing to align with RFC3280 and PKITS tests.
8234 This work was sponsored by Google.
8238 * Support for name constraints certificate extension. DN, email, DNS
8239 and URI types are currently supported.
8241 This work was sponsored by Google.
8245 * To cater for systems that provide a pointer-based thread ID rather
8246 than numeric, deprecate the current numeric thread ID mechanism and
8247 replace it with a structure and associated callback type. This
8248 mechanism allows a numeric "hash" to be extracted from a thread ID in
8249 either case, and on platforms where pointers are larger than 'long',
8250 mixing is done to help ensure the numeric 'hash' is usable even if it
8251 can't be guaranteed unique. The default mechanism is to use "&errno"
8252 as a pointer-based thread ID to distinguish between threads.
8254 Applications that want to provide their own thread IDs should now use
8255 CRYPTO_THREADID_set_callback() to register a callback that will call
8256 either CRYPTO_THREADID_set_numeric() or CRYPTO_THREADID_set_pointer().
8258 Note that ERR_remove_state() is now deprecated, because it is tied
8259 to the assumption that thread IDs are numeric. ERR_remove_state(0)
8260 to free the current thread's error state should be replaced by
8261 ERR_remove_thread_state(NULL).
8263 (This new approach replaces the functions CRYPTO_set_idptr_callback(),
8264 CRYPTO_get_idptr_callback(), and CRYPTO_thread_idptr() that existed in
8265 OpenSSL 0.9.9-dev between June 2006 and August 2008. Also, if an
8266 application was previously providing a numeric thread callback that
8267 was inappropriate for distinguishing threads, then uniqueness might
8268 have been obtained with &errno that happened immediately in the
8269 intermediate development versions of OpenSSL; this is no longer the
8270 case, the numeric thread callback will now override the automatic use
8273 *Geoff Thorpe, with help from Bodo Moeller*
8275 * Initial support for different CRL issuing certificates. This covers a
8276 simple case where the self issued certificates in the chain exist and
8277 the real CRL issuer is higher in the existing chain.
8279 This work was sponsored by Google.
8283 * Removed effectively defunct crypto/store from the build.
8287 * Revamp of STACK to provide stronger type-checking. Still to come:
8288 TXT_DB, bsearch(?), OBJ_bsearch, qsort, CRYPTO_EX_DATA, ASN1_VALUE,
8289 ASN1_STRING, CONF_VALUE.
8293 * Add a new SSL_MODE_RELEASE_BUFFERS mode flag to release unused buffer
8294 RAM on SSL connections. This option can save about 34k per idle SSL.
8298 * Revamp of LHASH to provide stronger type-checking. Still to come:
8299 STACK, TXT_DB, bsearch, qsort.
8303 * Initial support for Cryptographic Message Syntax (aka CMS) based
8304 on RFC3850, RFC3851 and RFC3852. New cms directory and cms utility,
8305 support for data, signedData, compressedData, digestedData and
8306 encryptedData, envelopedData types included. Scripts to check against
8307 RFC4134 examples draft and interop and consistency checks of many
8308 content types and variants.
8312 * Add options to enc utility to support use of zlib compression BIO.
8316 * Extend mk1mf to support importing of options and assembly language
8317 files from Configure script, currently only included in VC-WIN32.
8318 The assembly language rules can now optionally generate the source
8319 files from the associated perl scripts.
8323 * Implement remaining functionality needed to support GOST ciphersuites.
8324 Interop testing has been performed using CryptoPro implementations.
8326 *Victor B. Wagner <vitus@cryptocom.ru>*
8328 * s390x assembler pack.
8332 * ARMv4 assembler pack. ARMv4 refers to v4 and later ISA, not CPU
8337 * Implement Opaque PRF Input TLS extension as specified in
8338 draft-rescorla-tls-opaque-prf-input-00.txt. Since this is not an
8339 official specification yet and no extension type assignment by
8340 IANA exists, this extension (for now) will have to be explicitly
8341 enabled when building OpenSSL by providing the extension number
8342 to use. For example, specify an option
8344 -DTLSEXT_TYPE_opaque_prf_input=0x9527
8346 to the "config" or "Configure" script to enable the extension,
8347 assuming extension number 0x9527 (which is a completely arbitrary
8348 and unofficial assignment based on the MD5 hash of the Internet
8349 Draft). Note that by doing so, you potentially lose
8350 interoperability with other TLS implementations since these might
8351 be using the same extension number for other purposes.
8353 SSL_set_tlsext_opaque_prf_input(ssl, src, len) is used to set the
8354 opaque PRF input value to use in the handshake. This will create
8355 an internal copy of the length-'len' string at 'src', and will
8356 return non-zero for success.
8358 To get more control and flexibility, provide a callback function
8361 SSL_CTX_set_tlsext_opaque_prf_input_callback(ctx, cb)
8362 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg(ctx, arg)
8366 int (*cb)(SSL *, void *peerinput, size_t len, void *arg);
8369 Callback function 'cb' will be called in handshakes, and is
8370 expected to use SSL_set_tlsext_opaque_prf_input() as appropriate.
8371 Argument 'arg' is for application purposes (the value as given to
8372 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg() will directly
8373 be provided to the callback function). The callback function
8374 has to return non-zero to report success: usually 1 to use opaque
8375 PRF input just if possible, or 2 to enforce use of the opaque PRF
8376 input. In the latter case, the library will abort the handshake
8377 if opaque PRF input is not successfully negotiated.
8379 Arguments 'peerinput' and 'len' given to the callback function
8380 will always be NULL and 0 in the case of a client. A server will
8381 see the client's opaque PRF input through these variables if
8382 available (NULL and 0 otherwise). Note that if the server
8383 provides an opaque PRF input, the length must be the same as the
8384 length of the client's opaque PRF input.
8386 Note that the callback function will only be called when creating
8387 a new session (session resumption can resume whatever was
8388 previously negotiated), and will not be called in SSL 2.0
8389 handshakes; thus, SSL_CTX_set_options(ctx, SSL_OP_NO_SSLv2) or
8390 SSL_set_options(ssl, SSL_OP_NO_SSLv2) is especially recommended
8391 for applications that need to enforce opaque PRF input.
8395 * Update ssl code to support digests other than SHA1+MD5 for handshake
8398 *Victor B. Wagner <vitus@cryptocom.ru>*
8400 * Add RFC4507 support to OpenSSL. This includes the corrections in
8401 RFC4507bis. The encrypted ticket format is an encrypted encoded
8402 SSL_SESSION structure, that way new session features are automatically
8405 If a client application caches session in an SSL_SESSION structure
8406 support is transparent because tickets are now stored in the encoded
8409 The SSL_CTX structure automatically generates keys for ticket
8410 protection in servers so again support should be possible
8411 with no application modification.
8413 If a client or server wishes to disable RFC4507 support then the option
8414 SSL_OP_NO_TICKET can be set.
8416 Add a TLS extension debugging callback to allow the contents of any client
8417 or server extensions to be examined.
8419 This work was sponsored by Google.
8423 * Final changes to avoid use of pointer pointer casts in OpenSSL.
8424 OpenSSL should now compile cleanly on gcc 4.2
8426 *Peter Hartley <pdh@utter.chaos.org.uk>, Steve Henson*
8428 * Update SSL library to use new EVP_PKEY MAC API. Include generic MAC
8429 support including streaming MAC support: this is required for GOST
8430 ciphersuite support.
8432 *Victor B. Wagner <vitus@cryptocom.ru>, Steve Henson*
8434 * Add option -stream to use PKCS#7 streaming in smime utility. New
8435 function i2d_PKCS7_bio_stream() and PEM_write_PKCS7_bio_stream()
8436 to output in BER and PEM format.
8440 * Experimental support for use of HMAC via EVP_PKEY interface. This
8441 allows HMAC to be handled via the `EVP_DigestSign*()` interface. The
8442 EVP_PKEY "key" in this case is the HMAC key, potentially allowing
8443 ENGINE support for HMAC keys which are unextractable. New -mac and
8444 -macopt options to dgst utility.
8448 * New option -sigopt to dgst utility. Update dgst to use
8449 `EVP_Digest{Sign,Verify}*`. These two changes make it possible to use
8450 alternative signing parameters such as X9.31 or PSS in the dgst
8455 * Change ssl_cipher_apply_rule(), the internal function that does
8456 the work each time a ciphersuite string requests enabling
8457 ("foo+bar"), moving ("+foo+bar"), disabling ("-foo+bar", or
8458 removing ("!foo+bar") a class of ciphersuites: Now it maintains
8459 the order of disabled ciphersuites such that those ciphersuites
8460 that most recently went from enabled to disabled not only stay
8461 in order with respect to each other, but also have higher priority
8462 than other disabled ciphersuites the next time ciphersuites are
8465 This means that you can now say, e.g., "PSK:-PSK:HIGH" to enable
8466 the same ciphersuites as with "HIGH" alone, but in a specific
8467 order where the PSK ciphersuites come first (since they are the
8468 most recently disabled ciphersuites when "HIGH" is parsed).
8470 Also, change ssl_create_cipher_list() (using this new
8471 functionality) such that between otherwise identical
8472 ciphersuites, ephemeral ECDH is preferred over ephemeral DH in
8477 * Change ssl_create_cipher_list() so that it automatically
8478 arranges the ciphersuites in reasonable order before starting
8479 to process the rule string. Thus, the definition for "DEFAULT"
8480 (SSL_DEFAULT_CIPHER_LIST) now is just "ALL:!aNULL:!eNULL", but
8481 remains equivalent to `"AES:ALL:!aNULL:!eNULL:+aECDH:+kRSA:+RC4:@STRENGTH"`.
8482 This makes it much easier to arrive at a reasonable default order
8483 in applications for which anonymous ciphers are OK (meaning
8484 that you can't actually use DEFAULT).
8486 *Bodo Moeller; suggested by Victor Duchovni*
8488 * Split the SSL/TLS algorithm mask (as used for ciphersuite string
8489 processing) into multiple integers instead of setting
8490 "SSL_MKEY_MASK" bits, "SSL_AUTH_MASK" bits, "SSL_ENC_MASK",
8491 "SSL_MAC_MASK", and "SSL_SSL_MASK" bits all in a single integer.
8492 (These masks as well as the individual bit definitions are hidden
8493 away into the non-exported interface ssl/ssl_locl.h, so this
8494 change to the definition of the SSL_CIPHER structure shouldn't
8495 affect applications.) This give us more bits for each of these
8496 categories, so there is no longer a need to coagulate AES128 and
8497 AES256 into a single algorithm bit, and to coagulate Camellia128
8498 and Camellia256 into a single algorithm bit, which has led to all
8501 Thus, among other things, the kludge introduced in 0.9.7m and
8502 0.9.8e for masking out AES256 independently of AES128 or masking
8503 out Camellia256 independently of AES256 is not needed here in 0.9.9.
8505 With the change, we also introduce new ciphersuite aliases that
8506 so far were missing: "AES128", "AES256", "CAMELLIA128", and
8511 * Add support for dsa-with-SHA224 and dsa-with-SHA256.
8512 Use the leftmost N bytes of the signature input if the input is
8513 larger than the prime q (with N being the size in bytes of q).
8517 * Very *very* experimental PKCS#7 streaming encoder support. Nothing uses
8518 it yet and it is largely untested.
8522 * Add support for the ecdsa-with-SHA224/256/384/512 signature types.
8526 * Initial incomplete changes to avoid need for function casts in OpenSSL
8527 some compilers (gcc 4.2 and later) reject their use. Safestack is
8528 reimplemented. Update ASN1 to avoid use of legacy functions.
8532 * Win32/64 targets are linked with Winsock2.
8536 * Add an X509_CRL_METHOD structure to allow CRL processing to be redirected
8537 to external functions. This can be used to increase CRL handling
8538 efficiency especially when CRLs are very large by (for example) storing
8539 the CRL revoked certificates in a database.
8543 * Overhaul of by_dir code. Add support for dynamic loading of CRLs so
8544 new CRLs added to a directory can be used. New command line option
8545 -verify_return_error to s_client and s_server. This causes real errors
8546 to be returned by the verify callback instead of carrying on no matter
8547 what. This reflects the way a "real world" verify callback would behave.
8551 * GOST engine, supporting several GOST algorithms and public key formats.
8552 Kindly donated by Cryptocom.
8556 * Partial support for Issuing Distribution Point CRL extension. CRLs
8557 partitioned by DP are handled but no indirect CRL or reason partitioning
8558 (yet). Complete overhaul of CRL handling: now the most suitable CRL is
8559 selected via a scoring technique which handles IDP and AKID in CRLs.
8563 * New X509_STORE_CTX callbacks lookup_crls() and lookup_certs() which
8564 will ultimately be used for all verify operations: this will remove the
8565 X509_STORE dependency on certificate verification and allow alternative
8566 lookup methods. X509_STORE based implementations of these two callbacks.
8570 * Allow multiple CRLs to exist in an X509_STORE with matching issuer names.
8571 Modify get_crl() to find a valid (unexpired) CRL if possible.
8575 * New function X509_CRL_match() to check if two CRLs are identical. Normally
8576 this would be called X509_CRL_cmp() but that name is already used by
8577 a function that just compares CRL issuer names. Cache several CRL
8578 extensions in X509_CRL structure and cache CRLDP in X509.
8582 * Store a "canonical" representation of X509_NAME structure (ASN1 Name)
8583 this maps equivalent X509_NAME structures into a consistent structure.
8584 Name comparison can then be performed rapidly using memcmp().
8588 * Non-blocking OCSP request processing. Add -timeout option to ocsp
8593 * Allow digests to supply their own micalg string for S/MIME type using
8594 the ctrl EVP_MD_CTRL_MICALG.
8598 * During PKCS7 signing pass the PKCS7 SignerInfo structure to the
8599 EVP_PKEY_METHOD before and after signing via the EVP_PKEY_CTRL_PKCS7_SIGN
8600 ctrl. It can then customise the structure before and/or after signing
8605 * New function OBJ_add_sigid() to allow application defined signature OIDs
8606 to be added to OpenSSLs internal tables. New function OBJ_sigid_free()
8607 to free up any added signature OIDs.
8611 * New functions EVP_CIPHER_do_all(), EVP_CIPHER_do_all_sorted(),
8612 EVP_MD_do_all() and EVP_MD_do_all_sorted() to enumerate internal
8613 digest and cipher tables. New options added to openssl utility:
8614 list-message-digest-algorithms and list-cipher-algorithms.
8618 * Change the array representation of binary polynomials: the list
8619 of degrees of non-zero coefficients is now terminated with -1.
8620 Previously it was terminated with 0, which was also part of the
8621 value; thus, the array representation was not applicable to
8622 polynomials where t^0 has coefficient zero. This change makes
8623 the array representation useful in a more general context.
8627 * Various modifications and fixes to SSL/TLS cipher string
8628 handling. For ECC, the code now distinguishes between fixed ECDH
8629 with RSA certificates on the one hand and with ECDSA certificates
8630 on the other hand, since these are separate ciphersuites. The
8631 unused code for Fortezza ciphersuites has been removed.
8633 For consistency with EDH, ephemeral ECDH is now called "EECDH"
8634 (not "ECDHE"). For consistency with the code for DH
8635 certificates, use of ECDH certificates is now considered ECDH
8636 authentication, not RSA or ECDSA authentication (the latter is
8637 merely the CA's signing algorithm and not actively used in the
8640 The temporary ciphersuite alias "ECCdraft" is no longer
8641 available, and ECC ciphersuites are no longer excluded from "ALL"
8642 and "DEFAULT". The following aliases now exist for RFC 4492
8643 ciphersuites, most of these by analogy with the DH case:
8645 kECDHr - ECDH cert, signed with RSA
8646 kECDHe - ECDH cert, signed with ECDSA
8647 kECDH - ECDH cert (signed with either RSA or ECDSA)
8648 kEECDH - ephemeral ECDH
8649 ECDH - ECDH cert or ephemeral ECDH
8655 AECDH - anonymous ECDH
8656 EECDH - non-anonymous ephemeral ECDH (equivalent to "kEECDH:-AECDH")
8660 * Add additional S/MIME capabilities for AES and GOST ciphers if supported.
8661 Use correct micalg parameters depending on digest(s) in signed message.
8665 * Add engine support for EVP_PKEY_ASN1_METHOD. Add functions to process
8666 an ENGINE asn1 method. Support ENGINE lookups in the ASN1 code.
8670 * Initial engine support for EVP_PKEY_METHOD. New functions to permit
8671 an engine to register a method. Add ENGINE lookups for methods and
8672 functional reference processing.
8676 * New functions `EVP_Digest{Sign,Verify)*`. These are enhanced versions of
8677 `EVP_{Sign,Verify}*` which allow an application to customise the signature
8682 * New -resign option to smime utility. This adds one or more signers
8683 to an existing PKCS#7 signedData structure. Also -md option to use an
8684 alternative message digest algorithm for signing.
8688 * Tidy up PKCS#7 routines and add new functions to make it easier to
8689 create PKCS7 structures containing multiple signers. Update smime
8690 application to support multiple signers.
8694 * New -macalg option to pkcs12 utility to allow setting of an alternative
8699 * Initial support for PKCS#5 v2.0 PRFs other than default SHA1 HMAC.
8700 Reorganize PBE internals to lookup from a static table using NIDs,
8701 add support for HMAC PBE OID translation. Add a EVP_CIPHER ctrl:
8702 EVP_CTRL_PBE_PRF_NID this allows a cipher to specify an alternative
8703 PRF which will be automatically used with PBES2.
8707 * Replace the algorithm specific calls to generate keys in "req" with the
8712 * Update PKCS#7 enveloped data routines to use new API. This is now
8713 supported by any public key method supporting the encrypt operation. A
8714 ctrl is added to allow the public key algorithm to examine or modify
8715 the PKCS#7 RecipientInfo structure if it needs to: for RSA this is
8720 * Add a ctrl to asn1 method to allow a public key algorithm to express
8721 a default digest type to use. In most cases this will be SHA1 but some
8722 algorithms (such as GOST) need to specify an alternative digest. The
8723 return value indicates how strong the preference is 1 means optional and
8724 2 is mandatory (that is it is the only supported type). Modify
8725 ASN1_item_sign() to accept a NULL digest argument to indicate it should
8726 use the default md. Update openssl utilities to use the default digest
8727 type for signing if it is not explicitly indicated.
8731 * Use OID cross reference table in ASN1_sign() and ASN1_verify(). New
8732 EVP_MD flag EVP_MD_FLAG_PKEY_METHOD_SIGNATURE. This uses the relevant
8733 signing method from the key type. This effectively removes the link
8734 between digests and public key types.
8738 * Add an OID cross reference table and utility functions. Its purpose is to
8739 translate between signature OIDs such as SHA1WithrsaEncryption and SHA1,
8740 rsaEncryption. This will allow some of the algorithm specific hackery
8741 needed to use the correct OID to be removed.
8745 * Remove algorithm specific dependencies when setting PKCS7_SIGNER_INFO
8746 structures for PKCS7_sign(). They are now set up by the relevant public
8751 * Add provisional EC pkey method with support for ECDSA and ECDH.
8755 * Add support for key derivation (agreement) in the API, DH method and
8760 * Add DSA pkey method and DH pkey methods, extend DH ASN1 method to support
8761 public and private key formats. As a side effect these add additional
8762 command line functionality not previously available: DSA signatures can be
8763 generated and verified using pkeyutl and DH key support and generation in
8770 *Oliver Tappe <zooey@hirschkaefer.de>*
8772 * New make target "install_html_docs" installs HTML renditions of the
8775 *Oliver Tappe <zooey@hirschkaefer.de>*
8777 * New utility "genpkey" this is analogous to "genrsa" etc except it can
8778 generate keys for any algorithm. Extend and update EVP_PKEY_METHOD to
8779 support key and parameter generation and add initial key generation
8780 functionality for RSA.
8784 * Add functions for main EVP_PKEY_method operations. The undocumented
8785 functions `EVP_PKEY_{encrypt,decrypt}` have been renamed to
8786 `EVP_PKEY_{encrypt,decrypt}_old`.
8790 * Initial definitions for EVP_PKEY_METHOD. This will be a high level public
8791 key API, doesn't do much yet.
8795 * New function EVP_PKEY_asn1_get0_info() to retrieve information about
8796 public key algorithms. New option to openssl utility:
8797 "list-public-key-algorithms" to print out info.
8801 * Implement the Supported Elliptic Curves Extension for
8802 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
8806 * Don't free up OIDs in OBJ_cleanup() if they are in use by EVP_MD or
8807 EVP_CIPHER structures to avoid later problems in EVP_cleanup().
8811 * New utilities pkey and pkeyparam. These are similar to algorithm specific
8812 utilities such as rsa, dsa, dsaparam etc except they process any key
8817 * Transfer public key printing routines to EVP_PKEY_ASN1_METHOD. New
8818 functions EVP_PKEY_print_public(), EVP_PKEY_print_private(),
8819 EVP_PKEY_print_param() to print public key data from an EVP_PKEY
8824 * Initial support for pluggable public key ASN1.
8825 De-spaghettify the public key ASN1 handling. Move public and private
8826 key ASN1 handling to a new EVP_PKEY_ASN1_METHOD structure. Relocate
8827 algorithm specific handling to a single module within the relevant
8828 algorithm directory. Add functions to allow (near) opaque processing
8829 of public and private key structures.
8833 * Implement the Supported Point Formats Extension for
8834 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
8838 * Add initial support for RFC 4279 PSK TLS ciphersuites. Add members
8839 for the psk identity [hint] and the psk callback functions to the
8840 SSL_SESSION, SSL and SSL_CTX structure.
8843 PSK-RC4-SHA, PSK-3DES-EDE-CBC-SHA, PSK-AES128-CBC-SHA,
8847 SSL_CTX_use_psk_identity_hint
8848 SSL_get_psk_identity_hint
8849 SSL_get_psk_identity
8850 SSL_use_psk_identity_hint
8852 *Mika Kousa and Pasi Eronen of Nokia Corporation*
8854 * Add RFC 3161 compliant time stamp request creation, response generation
8855 and response verification functionality.
8857 *Zoltán Glózik <zglozik@opentsa.org>, The OpenTSA Project*
8859 * Add initial support for TLS extensions, specifically for the server_name
8860 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
8861 have new members for a host name. The SSL data structure has an
8862 additional member `SSL_CTX *initial_ctx` so that new sessions can be
8863 stored in that context to allow for session resumption, even after the
8864 SSL has been switched to a new SSL_CTX in reaction to a client's
8865 server_name extension.
8867 New functions (subject to change):
8869 SSL_get_servername()
8870 SSL_get_servername_type()
8873 New CTRL codes and macros (subject to change):
8875 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
8876 - SSL_CTX_set_tlsext_servername_callback()
8877 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
8878 - SSL_CTX_set_tlsext_servername_arg()
8879 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
8881 openssl s_client has a new '-servername ...' option.
8883 openssl s_server has new options '-servername_host ...', '-cert2 ...',
8884 '-key2 ...', '-servername_fatal' (subject to change). This allows
8885 testing the HostName extension for a specific single host name ('-cert'
8886 and '-key' remain fallbacks for handshakes without HostName
8887 negotiation). If the unrecognized_name alert has to be sent, this by
8888 default is a warning; it becomes fatal with the '-servername_fatal'
8891 *Peter Sylvester, Remy Allais, Christophe Renou*
8893 * Whirlpool hash implementation is added.
8897 * BIGNUM code on 64-bit SPARCv9 targets is switched from bn(64,64) to
8898 bn(64,32). Because of instruction set limitations it doesn't have
8899 any negative impact on performance. This was done mostly in order
8900 to make it possible to share assembler modules, such as bn_mul_mont
8901 implementations, between 32- and 64-bit builds without hassle.
8905 * Move code previously exiled into file crypto/ec/ec2_smpt.c
8906 to ec2_smpl.c, and no longer require the OPENSSL_EC_BIN_PT_COMP
8911 * New candidate for BIGNUM assembler implementation, bn_mul_mont,
8912 dedicated Montgomery multiplication procedure, is introduced.
8913 BN_MONT_CTX is modified to allow bn_mul_mont to reach for higher
8914 "64-bit" performance on certain 32-bit targets.
8918 * New option SSL_OP_NO_COMP to disable use of compression selectively
8919 in SSL structures. New SSL ctrl to set maximum send fragment size.
8920 Save memory by setting the I/O buffer sizes dynamically instead of
8921 using the maximum available value.
8925 * New option -V for 'openssl ciphers'. This prints the ciphersuite code
8926 in addition to the text details.
8930 * Very, very preliminary EXPERIMENTAL support for printing of general
8931 ASN1 structures. This currently produces rather ugly output and doesn't
8932 handle several customised structures at all.
8936 * Integrated support for PVK file format and some related formats such
8937 as MS PUBLICKEYBLOB and PRIVATEKEYBLOB. Command line switches to support
8938 these in the 'rsa' and 'dsa' utilities.
8942 * Support for PKCS#1 RSAPublicKey format on rsa utility command line.
8946 * Remove the ancient ASN1_METHOD code. This was only ever used in one
8947 place for the (very old) "NETSCAPE" format certificates which are now
8948 handled using new ASN1 code equivalents.
8952 * Let the TLSv1_method() etc. functions return a 'const' SSL_METHOD
8953 pointer and make the SSL_METHOD parameter in SSL_CTX_new,
8954 SSL_CTX_set_ssl_version and SSL_set_ssl_method 'const'.
8958 * Modify CRL distribution points extension code to print out previously
8959 unsupported fields. Enhance extension setting code to allow setting of
8964 * Add print and set support for Issuing Distribution Point CRL extension.
8968 * Change 'Configure' script to enable Camellia by default.
8975 ### Changes between 0.9.8m and 0.9.8n [24 Mar 2010]
8977 * When rejecting SSL/TLS records due to an incorrect version number, never
8978 update s->server with a new major version number. As of
8979 - OpenSSL 0.9.8m if 'short' is a 16-bit type,
8980 - OpenSSL 0.9.8f if 'short' is longer than 16 bits,
8981 the previous behavior could result in a read attempt at NULL when
8982 receiving specific incorrect SSL/TLS records once record payload
8983 protection is active. ([CVE-2010-0740])
8985 *Bodo Moeller, Adam Langley <agl@chromium.org>*
8987 * Fix for CVE-2010-0433 where some kerberos enabled versions of OpenSSL
8988 could be crashed if the relevant tables were not present (e.g. chrooted).
8990 *Tomas Hoger <thoger@redhat.com>*
8992 ### Changes between 0.9.8l and 0.9.8m [25 Feb 2010]
8994 * Always check bn_wexpand() return values for failure. ([CVE-2009-3245])
8996 *Martin Olsson, Neel Mehta*
8998 * Fix X509_STORE locking: Every 'objs' access requires a lock (to
8999 accommodate for stack sorting, always a write lock!).
9003 * On some versions of WIN32 Heap32Next is very slow. This can cause
9004 excessive delays in the RAND_poll(): over a minute. As a workaround
9005 include a time check in the inner Heap32Next loop too.
9009 * The code that handled flushing of data in SSL/TLS originally used the
9010 BIO_CTRL_INFO ctrl to see if any data was pending first. This caused
9011 the problem outlined in PR#1949. The fix suggested there however can
9012 trigger problems with buggy BIO_CTRL_WPENDING (e.g. some versions
9013 of Apache). So instead simplify the code to flush unconditionally.
9014 This should be fine since flushing with no data to flush is a no op.
9018 * Handle TLS versions 2.0 and later properly and correctly use the
9019 highest version of TLS/SSL supported. Although TLS >= 2.0 is some way
9020 off ancient servers have a habit of sticking around for a while...
9024 * Modify compression code so it frees up structures without using the
9025 ex_data callbacks. This works around a problem where some applications
9026 call CRYPTO_cleanup_all_ex_data() before application exit (e.g. when
9027 restarting) then use compression (e.g. SSL with compression) later.
9028 This results in significant per-connection memory leaks and
9029 has caused some security issues including CVE-2008-1678 and
9034 * Constify crypto/cast (i.e., <openssl/cast.h>): a CAST_KEY doesn't
9035 change when encrypting or decrypting.
9039 * Add option SSL_OP_LEGACY_SERVER_CONNECT which will allow clients to
9040 connect and renegotiate with servers which do not support RI.
9041 Until RI is more widely deployed this option is enabled by default.
9045 * Add "missing" ssl ctrls to clear options and mode.
9049 * If client attempts to renegotiate and doesn't support RI respond with
9050 a no_renegotiation alert as required by RFC5746. Some renegotiating
9051 TLS clients will continue a connection gracefully when they receive
9052 the alert. Unfortunately OpenSSL mishandled this alert and would hang
9053 waiting for a server hello which it will never receive. Now we treat a
9054 received no_renegotiation alert as a fatal error. This is because
9055 applications requesting a renegotiation might well expect it to succeed
9056 and would have no code in place to handle the server denying it so the
9057 only safe thing to do is to terminate the connection.
9061 * Add ctrl macro SSL_get_secure_renegotiation_support() which returns 1 if
9062 peer supports secure renegotiation and 0 otherwise. Print out peer
9063 renegotiation support in s_client/s_server.
9067 * Replace the highly broken and deprecated SPKAC certification method with
9068 the updated NID creation version. This should correctly handle UTF8.
9072 * Implement RFC5746. Re-enable renegotiation but require the extension
9073 as needed. Unfortunately, SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION
9074 turns out to be a bad idea. It has been replaced by
9075 SSL_OP_ALLOW_UNSAFE_LEGACY_RENEGOTIATION which can be set with
9076 SSL_CTX_set_options(). This is really not recommended unless you
9077 know what you are doing.
9079 *Eric Rescorla <ekr@networkresonance.com>, Ben Laurie, Steve Henson*
9081 * Fixes to stateless session resumption handling. Use initial_ctx when
9082 issuing and attempting to decrypt tickets in case it has changed during
9083 servername handling. Use a non-zero length session ID when attempting
9084 stateless session resumption: this makes it possible to determine if
9085 a resumption has occurred immediately after receiving server hello
9086 (several places in OpenSSL subtly assume this) instead of later in
9091 * The functions ENGINE_ctrl(), OPENSSL_isservice(),
9092 CMS_get1_RecipientRequest() and RAND_bytes() can return <=0 on error
9093 fixes for a few places where the return code is not checked
9096 *Julia Lawall <julia@diku.dk>*
9098 * Add --strict-warnings option to Configure script to include devteam
9099 warnings in other configurations.
9103 * Add support for --libdir option and LIBDIR variable in makefiles. This
9104 makes it possible to install openssl libraries in locations which
9105 have names other than "lib", for example "/usr/lib64" which some
9108 *Steve Henson, based on patch from Jeremy Utley*
9110 * Don't allow the use of leading 0x80 in OIDs. This is a violation of
9111 X690 8.9.12 and can produce some misleading textual output of OIDs.
9113 *Steve Henson, reported by Dan Kaminsky*
9115 * Delete MD2 from algorithm tables. This follows the recommendation in
9116 several standards that it is not used in new applications due to
9117 several cryptographic weaknesses. For binary compatibility reasons
9118 the MD2 API is still compiled in by default.
9122 * Add compression id to {d2i,i2d}_SSL_SESSION so it is correctly saved
9127 * Rename uni2asc and asc2uni functions to OPENSSL_uni2asc and
9128 OPENSSL_asc2uni conditionally on Netware platforms to avoid a name
9131 *Guenter <lists@gknw.net>*
9133 * Fix the server certificate chain building code to use X509_verify_cert(),
9134 it used to have an ad-hoc builder which was unable to cope with anything
9135 other than a simple chain.
9137 *David Woodhouse <dwmw2@infradead.org>, Steve Henson*
9139 * Don't check self signed certificate signatures in X509_verify_cert()
9140 by default (a flag can override this): it just wastes time without
9141 adding any security. As a useful side effect self signed root CAs
9142 with non-FIPS digests are now usable in FIPS mode.
9146 * In dtls1_process_out_of_seq_message() the check if the current message
9147 is already buffered was missing. For every new message was memory
9148 allocated, allowing an attacker to perform an denial of service attack
9149 with sending out of seq handshake messages until there is no memory
9150 left. Additionally every future message was buffered, even if the
9151 sequence number made no sense and would be part of another handshake.
9152 So only messages with sequence numbers less than 10 in advance will be
9153 buffered. ([CVE-2009-1378])
9155 *Robin Seggelmann, discovered by Daniel Mentz*
9157 * Records are buffered if they arrive with a future epoch to be
9158 processed after finishing the corresponding handshake. There is
9159 currently no limitation to this buffer allowing an attacker to perform
9160 a DOS attack with sending records with future epochs until there is no
9161 memory left. This patch adds the pqueue_size() function to determine
9162 the size of a buffer and limits the record buffer to 100 entries.
9165 *Robin Seggelmann, discovered by Daniel Mentz*
9167 * Keep a copy of frag->msg_header.frag_len so it can be used after the
9168 parent structure is freed. ([CVE-2009-1379])
9172 * Handle non-blocking I/O properly in SSL_shutdown() call.
9174 *Darryl Miles <darryl-mailinglists@netbauds.net>*
9176 * Add `2.5.4.*` OIDs
9178 *Ilya O. <vrghost@gmail.com>*
9180 ### Changes between 0.9.8k and 0.9.8l [5 Nov 2009]
9182 * Disable renegotiation completely - this fixes a severe security
9183 problem ([CVE-2009-3555]) at the cost of breaking all
9184 renegotiation. Renegotiation can be re-enabled by setting
9185 SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION in s3->flags at
9186 run-time. This is really not recommended unless you know what
9191 ### Changes between 0.9.8j and 0.9.8k [25 Mar 2009]
9193 * Don't set val to NULL when freeing up structures, it is freed up by
9194 underlying code. If `sizeof(void *) > sizeof(long)` this can result in
9195 zeroing past the valid field. ([CVE-2009-0789])
9197 *Paolo Ganci <Paolo.Ganci@AdNovum.CH>*
9199 * Fix bug where return value of CMS_SignerInfo_verify_content() was not
9200 checked correctly. This would allow some invalid signed attributes to
9201 appear to verify correctly. ([CVE-2009-0591])
9203 *Ivan Nestlerode <inestlerode@us.ibm.com>*
9205 * Reject UniversalString and BMPString types with invalid lengths. This
9206 prevents a crash in ASN1_STRING_print_ex() which assumes the strings have
9207 a legal length. ([CVE-2009-0590])
9211 * Set S/MIME signing as the default purpose rather than setting it
9212 unconditionally. This allows applications to override it at the store
9217 * Permit restricted recursion of ASN1 strings. This is needed in practice
9218 to handle some structures.
9222 * Improve efficiency of mem_gets: don't search whole buffer each time
9225 *Jeremy Shapiro <jnshapir@us.ibm.com>*
9227 * New -hex option for openssl rand.
9231 * Print out UTF8String and NumericString when parsing ASN1.
9235 * Support NumericString type for name components.
9239 * Allow CC in the environment to override the automatically chosen
9240 compiler. Note that nothing is done to ensure flags work with the
9245 ### Changes between 0.9.8i and 0.9.8j [07 Jan 2009]
9247 * Properly check EVP_VerifyFinal() and similar return values
9250 *Ben Laurie, Bodo Moeller, Google Security Team*
9252 * Enable TLS extensions by default.
9256 * Allow the CHIL engine to be loaded, whether the application is
9257 multithreaded or not. (This does not release the developer from the
9258 obligation to set up the dynamic locking callbacks.)
9260 *Sander Temme <sander@temme.net>*
9262 * Use correct exit code if there is an error in dgst command.
9264 *Steve Henson; problem pointed out by Roland Dirlewanger*
9266 * Tweak Configure so that you need to say "experimental-jpake" to enable
9267 JPAKE, and need to use -DOPENSSL_EXPERIMENTAL_JPAKE in applications.
9271 * Add experimental JPAKE support, including demo authentication in
9272 s_client and s_server.
9276 * Set the comparison function in v3_addr_canonize().
9278 *Rob Austein <sra@hactrn.net>*
9280 * Add support for XMPP STARTTLS in s_client.
9282 *Philip Paeps <philip@freebsd.org>*
9284 * Change the server-side SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG behavior
9285 to ensure that even with this option, only ciphersuites in the
9286 server's preference list will be accepted. (Note that the option
9287 applies only when resuming a session, so the earlier behavior was
9288 just about the algorithm choice for symmetric cryptography.)
9292 ### Changes between 0.9.8h and 0.9.8i [15 Sep 2008]
9294 * Fix NULL pointer dereference if a DTLS server received
9295 ChangeCipherSpec as first record ([CVE-2009-1386]).
9299 * Fix a state transition in s3_srvr.c and d1_srvr.c
9300 (was using SSL3_ST_CW_CLNT_HELLO_B, should be `..._ST_SW_SRVR_...`).
9304 * The fix in 0.9.8c that supposedly got rid of unsafe
9305 double-checked locking was incomplete for RSA blinding,
9306 addressing just one layer of what turns out to have been
9307 doubly unsafe triple-checked locking.
9309 So now fix this for real by retiring the MONT_HELPER macro
9310 in crypto/rsa/rsa_eay.c.
9312 *Bodo Moeller; problem pointed out by Marius Schilder*
9314 * Various precautionary measures:
9316 - Avoid size_t integer overflow in HASH_UPDATE (md32_common.h).
9318 - Avoid a buffer overflow in d2i_SSL_SESSION() (ssl_asn1.c).
9319 (NB: This would require knowledge of the secret session ticket key
9320 to exploit, in which case you'd be SOL either way.)
9322 - Change bn_nist.c so that it will properly handle input BIGNUMs
9323 outside the expected range.
9325 - Enforce the 'num' check in BN_div() (bn_div.c) for non-BN_DEBUG
9328 *Neel Mehta, Bodo Moeller*
9330 * Allow engines to be "soft loaded" - i.e. optionally don't die if
9331 the load fails. Useful for distros.
9333 *Ben Laurie and the FreeBSD team*
9335 * Add support for Local Machine Keyset attribute in PKCS#12 files.
9339 * Fix BN_GF2m_mod_arr() top-bit cleanup code.
9343 * Expand ENGINE to support engine supplied SSL client certificate functions.
9345 This work was sponsored by Logica.
9349 * Add CryptoAPI ENGINE to support use of RSA and DSA keys held in Windows
9350 keystores. Support for SSL/TLS client authentication too.
9351 Not compiled unless enable-capieng specified to Configure.
9353 This work was sponsored by Logica.
9357 * Fix bug in X509_ATTRIBUTE creation: don't set attribute using
9358 ASN1_TYPE_set1 if MBSTRING flag set. This bug would crash certain
9359 attribute creation routines such as certificate requests and PKCS#12
9364 ### Changes between 0.9.8g and 0.9.8h [28 May 2008]
9366 * Fix flaw if 'Server Key exchange message' is omitted from a TLS
9367 handshake which could lead to a client crash as found using the
9368 Codenomicon TLS test suite ([CVE-2008-1672])
9370 *Steve Henson, Mark Cox*
9372 * Fix double free in TLS server name extensions which could lead to
9373 a remote crash found by Codenomicon TLS test suite ([CVE-2008-0891])
9377 * Clear error queue in SSL_CTX_use_certificate_chain_file()
9379 Clear the error queue to ensure that error entries left from
9380 older function calls do not interfere with the correct operation.
9382 *Lutz Jaenicke, Erik de Castro Lopo*
9384 * Remove root CA certificates of commercial CAs:
9386 The OpenSSL project does not recommend any specific CA and does not
9387 have any policy with respect to including or excluding any CA.
9388 Therefore it does not make any sense to ship an arbitrary selection
9389 of root CA certificates with the OpenSSL software.
9393 * RSA OAEP patches to fix two separate invalid memory reads.
9394 The first one involves inputs when 'lzero' is greater than
9395 'SHA_DIGEST_LENGTH' (it would read about SHA_DIGEST_LENGTH bytes
9396 before the beginning of from). The second one involves inputs where
9397 the 'db' section contains nothing but zeroes (there is a one-byte
9398 invalid read after the end of 'db').
9400 *Ivan Nestlerode <inestlerode@us.ibm.com>*
9402 * Partial backport from 0.9.9-dev:
9404 Introduce bn_mul_mont (dedicated Montgomery multiplication
9405 procedure) as a candidate for BIGNUM assembler implementation.
9406 While 0.9.9-dev uses assembler for various architectures, only
9407 x86_64 is available by default here in the 0.9.8 branch, and
9408 32-bit x86 is available through a compile-time setting.
9410 To try the 32-bit x86 assembler implementation, use Configure
9411 option "enable-montasm" (which exists only for this backport).
9413 As "enable-montasm" for 32-bit x86 disclaims code stability
9414 anyway, in this constellation we activate additional code
9415 backported from 0.9.9-dev for further performance improvements,
9416 namely BN_from_montgomery_word. (To enable this otherwise,
9417 e.g. x86_64, try `-DMONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD`.)
9419 *Andy Polyakov (backport partially by Bodo Moeller)*
9421 * Add TLS session ticket callback. This allows an application to set
9422 TLS ticket cipher and HMAC keys rather than relying on hardcoded fixed
9423 values. This is useful for key rollover for example where several key
9424 sets may exist with different names.
9428 * Reverse ENGINE-internal logic for caching default ENGINE handles.
9429 This was broken until now in 0.9.8 releases, such that the only way
9430 a registered ENGINE could be used (assuming it initialises
9431 successfully on the host) was to explicitly set it as the default
9432 for the relevant algorithms. This is in contradiction with 0.9.7
9433 behaviour and the documentation. With this fix, when an ENGINE is
9434 registered into a given algorithm's table of implementations, the
9435 'uptodate' flag is reset so that auto-discovery will be used next
9436 time a new context for that algorithm attempts to select an
9439 *Ian Lister (tweaked by Geoff Thorpe)*
9441 * Backport of CMS code to OpenSSL 0.9.8. This differs from the 0.9.9
9442 implementation in the following ways:
9444 Lack of EVP_PKEY_ASN1_METHOD means algorithm parameters have to be
9447 Lack of BER streaming support means one pass streaming processing is
9448 only supported if data is detached: setting the streaming flag is
9449 ignored for embedded content.
9451 CMS support is disabled by default and must be explicitly enabled
9452 with the enable-cms configuration option.
9456 * Update the GMP engine glue to do direct copies between BIGNUM and
9457 mpz_t when openssl and GMP use the same limb size. Otherwise the
9458 existing "conversion via a text string export" trick is still used.
9460 *Paul Sheer <paulsheer@gmail.com>*
9462 * Zlib compression BIO. This is a filter BIO which compressed and
9463 uncompresses any data passed through it.
9467 * Add AES_wrap_key() and AES_unwrap_key() functions to implement
9468 RFC3394 compatible AES key wrapping.
9472 * Add utility functions to handle ASN1 structures. ASN1_STRING_set0():
9473 sets string data without copying. X509_ALGOR_set0() and
9474 X509_ALGOR_get0(): set and retrieve X509_ALGOR (AlgorithmIdentifier)
9475 data. Attribute function X509at_get0_data_by_OBJ(): retrieves data
9476 from an X509_ATTRIBUTE structure optionally checking it occurs only
9477 once. ASN1_TYPE_set1(): set and ASN1_TYPE structure copying supplied
9482 * Fix BN flag handling in RSA_eay_mod_exp() and BN_MONT_CTX_set()
9483 to get the expected BN_FLG_CONSTTIME behavior.
9485 *Bodo Moeller (Google)*
9489 - fixed wrong usage of ioctlsocket() when build for LIBC BSD sockets
9490 - fixed do_tests.pl to run the test suite with CLIB builds too (CLIB_OPT)
9491 - added some more tests to do_tests.pl
9492 - fixed RunningProcess usage so that it works with newer LIBC NDKs too
9493 - removed usage of BN_LLONG for CLIB builds to avoid runtime dependency
9494 - added new Configure targets netware-clib-bsdsock, netware-clib-gcc,
9495 netware-clib-bsdsock-gcc, netware-libc-bsdsock-gcc
9496 - various changes to netware.pl to enable gcc-cross builds on Win32
9498 - changed crypto/bio/b_sock.c to work with macro functions (CLIB BSD)
9499 - various changes to fix missing prototype warnings
9500 - fixed x86nasm.pl to create correct asm files for NASM COFF output
9501 - added AES, WHIRLPOOL and CPUID assembler code to build files
9502 - added missing AES assembler make rules to mk1mf.pl
9503 - fixed order of includes in `apps/ocsp.c` so that `e_os.h` settings apply
9505 *Guenter Knauf <eflash@gmx.net>*
9507 * Implement certificate status request TLS extension defined in RFC3546.
9508 A client can set the appropriate parameters and receive the encoded
9509 OCSP response via a callback. A server can query the supplied parameters
9510 and set the encoded OCSP response in the callback. Add simplified examples
9511 to s_client and s_server.
9515 ### Changes between 0.9.8f and 0.9.8g [19 Oct 2007]
9518 + Binary incompatibility of ssl_ctx_st structure
9519 + DTLS interoperation with non-compliant servers
9520 + Don't call get_session_cb() without proposed session
9521 + Fix ia64 assembler code
9523 *Andy Polyakov, Steve Henson*
9525 ### Changes between 0.9.8e and 0.9.8f [11 Oct 2007]
9527 * DTLS Handshake overhaul. There were longstanding issues with
9528 OpenSSL DTLS implementation, which were making it impossible for
9529 RFC 4347 compliant client to communicate with OpenSSL server.
9530 Unfortunately just fixing these incompatibilities would "cut off"
9531 pre-0.9.8f clients. To allow for hassle free upgrade post-0.9.8e
9532 server keeps tolerating non RFC compliant syntax. The opposite is
9533 not true, 0.9.8f client can not communicate with earlier server.
9534 This update even addresses CVE-2007-4995.
9538 * Changes to avoid need for function casts in OpenSSL: some compilers
9539 (gcc 4.2 and later) reject their use.
9540 *Kurt Roeckx <kurt@roeckx.be>, Peter Hartley <pdh@utter.chaos.org.uk>,
9543 * Add RFC4507 support to OpenSSL. This includes the corrections in
9544 RFC4507bis. The encrypted ticket format is an encrypted encoded
9545 SSL_SESSION structure, that way new session features are automatically
9548 If a client application caches session in an SSL_SESSION structure
9549 support is transparent because tickets are now stored in the encoded
9552 The SSL_CTX structure automatically generates keys for ticket
9553 protection in servers so again support should be possible
9554 with no application modification.
9556 If a client or server wishes to disable RFC4507 support then the option
9557 SSL_OP_NO_TICKET can be set.
9559 Add a TLS extension debugging callback to allow the contents of any client
9560 or server extensions to be examined.
9562 This work was sponsored by Google.
9566 * Add initial support for TLS extensions, specifically for the server_name
9567 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
9568 have new members for a host name. The SSL data structure has an
9569 additional member `SSL_CTX *initial_ctx` so that new sessions can be
9570 stored in that context to allow for session resumption, even after the
9571 SSL has been switched to a new SSL_CTX in reaction to a client's
9572 server_name extension.
9574 New functions (subject to change):
9576 SSL_get_servername()
9577 SSL_get_servername_type()
9580 New CTRL codes and macros (subject to change):
9582 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
9583 - SSL_CTX_set_tlsext_servername_callback()
9584 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
9585 - SSL_CTX_set_tlsext_servername_arg()
9586 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
9588 openssl s_client has a new '-servername ...' option.
9590 openssl s_server has new options '-servername_host ...', '-cert2 ...',
9591 '-key2 ...', '-servername_fatal' (subject to change). This allows
9592 testing the HostName extension for a specific single host name ('-cert'
9593 and '-key' remain fallbacks for handshakes without HostName
9594 negotiation). If the unrecognized_name alert has to be sent, this by
9595 default is a warning; it becomes fatal with the '-servername_fatal'
9598 *Peter Sylvester, Remy Allais, Christophe Renou, Steve Henson*
9600 * Add AES and SSE2 assembly language support to VC++ build.
9604 * Mitigate attack on final subtraction in Montgomery reduction.
9608 * Fix crypto/ec/ec_mult.c to work properly with scalars of value 0
9609 (which previously caused an internal error).
9613 * Squeeze another 10% out of IGE mode when in != out.
9617 * AES IGE mode speedup.
9619 *Dean Gaudet (Google)*
9621 * Add the Korean symmetric 128-bit cipher SEED (see
9622 <http://www.kisa.or.kr/kisa/seed/jsp/seed_eng.jsp>) and
9623 add SEED ciphersuites from RFC 4162:
9625 TLS_RSA_WITH_SEED_CBC_SHA = "SEED-SHA"
9626 TLS_DHE_DSS_WITH_SEED_CBC_SHA = "DHE-DSS-SEED-SHA"
9627 TLS_DHE_RSA_WITH_SEED_CBC_SHA = "DHE-RSA-SEED-SHA"
9628 TLS_DH_anon_WITH_SEED_CBC_SHA = "ADH-SEED-SHA"
9630 To minimize changes between patchlevels in the OpenSSL 0.9.8
9631 series, SEED remains excluded from compilation unless OpenSSL
9632 is configured with 'enable-seed'.
9634 *KISA, Bodo Moeller*
9636 * Mitigate branch prediction attacks, which can be practical if a
9637 single processor is shared, allowing a spy process to extract
9638 information. For detailed background information, see
9639 <http://eprint.iacr.org/2007/039> (O. Aciicmez, S. Gueron,
9640 J.-P. Seifert, "New Branch Prediction Vulnerabilities in OpenSSL
9641 and Necessary Software Countermeasures"). The core of the change
9642 are new versions BN_div_no_branch() and
9643 BN_mod_inverse_no_branch() of BN_div() and BN_mod_inverse(),
9644 respectively, which are slower, but avoid the security-relevant
9645 conditional branches. These are automatically called by BN_div()
9646 and BN_mod_inverse() if the flag BN_FLG_CONSTTIME is set for one
9647 of the input BIGNUMs. Also, BN_is_bit_set() has been changed to
9648 remove a conditional branch.
9650 BN_FLG_CONSTTIME is the new name for the previous
9651 BN_FLG_EXP_CONSTTIME flag, since it now affects more than just
9652 modular exponentiation. (Since OpenSSL 0.9.7h, setting this flag
9653 in the exponent causes BN_mod_exp_mont() to use the alternative
9654 implementation in BN_mod_exp_mont_consttime().) The old name
9655 remains as a deprecated alias.
9657 Similarly, RSA_FLAG_NO_EXP_CONSTTIME is replaced by a more general
9658 RSA_FLAG_NO_CONSTTIME flag since the RSA implementation now uses
9659 constant-time implementations for more than just exponentiation.
9660 Here too the old name is kept as a deprecated alias.
9662 BN_BLINDING_new() will now use BN_dup() for the modulus so that
9663 the BN_BLINDING structure gets an independent copy of the
9664 modulus. This means that the previous `BIGNUM *m` argument to
9665 BN_BLINDING_new() and to BN_BLINDING_create_param() now
9666 essentially becomes `const BIGNUM *m`, although we can't actually
9667 change this in the header file before 0.9.9. It allows
9668 RSA_setup_blinding() to use BN_with_flags() on the modulus to
9669 enable BN_FLG_CONSTTIME.
9671 *Matthew D Wood (Intel Corp)*
9673 * In the SSL/TLS server implementation, be strict about session ID
9674 context matching (which matters if an application uses a single
9675 external cache for different purposes). Previously,
9676 out-of-context reuse was forbidden only if SSL_VERIFY_PEER was
9677 set. This did ensure strict client verification, but meant that,
9678 with applications using a single external cache for quite
9679 different requirements, clients could circumvent ciphersuite
9680 restrictions for a given session ID context by starting a session
9681 in a different context.
9685 * Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
9686 a ciphersuite string such as "DEFAULT:RSA" cannot enable
9687 authentication-only ciphersuites.
9691 * Update the SSL_get_shared_ciphers() fix CVE-2006-3738 which was
9692 not complete and could lead to a possible single byte overflow
9693 ([CVE-2007-5135]) [Ben Laurie]
9695 ### Changes between 0.9.8d and 0.9.8e [23 Feb 2007]
9697 * Since AES128 and AES256 (and similarly Camellia128 and
9698 Camellia256) share a single mask bit in the logic of
9699 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
9700 kludge to work properly if AES128 is available and AES256 isn't
9701 (or if Camellia128 is available and Camellia256 isn't).
9705 * Fix the BIT STRING encoding generated by crypto/ec/ec_asn1.c
9706 (within i2d_ECPrivateKey, i2d_ECPKParameters, i2d_ECParameters):
9707 When a point or a seed is encoded in a BIT STRING, we need to
9708 prevent the removal of trailing zero bits to get the proper DER
9709 encoding. (By default, crypto/asn1/a_bitstr.c assumes the case
9710 of a NamedBitList, for which trailing 0 bits need to be removed.)
9714 * Have SSL/TLS server implementation tolerate "mismatched" record
9715 protocol version while receiving ClientHello even if the
9716 ClientHello is fragmented. (The server can't insist on the
9717 particular protocol version it has chosen before the ServerHello
9718 message has informed the client about his choice.)
9722 * Add RFC 3779 support.
9724 *Rob Austein for ARIN, Ben Laurie*
9726 * Load error codes if they are not already present instead of using a
9727 static variable. This allows them to be cleanly unloaded and reloaded.
9728 Improve header file function name parsing.
9732 * extend SMTP and IMAP protocol emulation in s_client to use EHLO
9733 or CAPABILITY handshake as required by RFCs.
9737 ### Changes between 0.9.8c and 0.9.8d [28 Sep 2006]
9739 * Introduce limits to prevent malicious keys being able to
9740 cause a denial of service. ([CVE-2006-2940])
9742 *Steve Henson, Bodo Moeller*
9744 * Fix ASN.1 parsing of certain invalid structures that can result
9745 in a denial of service. ([CVE-2006-2937]) [Steve Henson]
9747 * Fix buffer overflow in SSL_get_shared_ciphers() function.
9748 ([CVE-2006-3738]) [Tavis Ormandy and Will Drewry, Google Security Team]
9750 * Fix SSL client code which could crash if connecting to a
9751 malicious SSLv2 server. ([CVE-2006-4343])
9753 *Tavis Ormandy and Will Drewry, Google Security Team*
9755 * Since 0.9.8b, ciphersuite strings naming explicit ciphersuites
9756 match only those. Before that, "AES256-SHA" would be interpreted
9757 as a pattern and match "AES128-SHA" too (since AES128-SHA got
9758 the same strength classification in 0.9.7h) as we currently only
9759 have a single AES bit in the ciphersuite description bitmap.
9760 That change, however, also applied to ciphersuite strings such as
9761 "RC4-MD5" that intentionally matched multiple ciphersuites --
9762 namely, SSL 2.0 ciphersuites in addition to the more common ones
9763 from SSL 3.0/TLS 1.0.
9765 So we change the selection algorithm again: Naming an explicit
9766 ciphersuite selects this one ciphersuite, and any other similar
9767 ciphersuite (same bitmap) from *other* protocol versions.
9768 Thus, "RC4-MD5" again will properly select both the SSL 2.0
9769 ciphersuite and the SSL 3.0/TLS 1.0 ciphersuite.
9771 Since SSL 2.0 does not have any ciphersuites for which the
9772 128/256 bit distinction would be relevant, this works for now.
9773 The proper fix will be to use different bits for AES128 and
9774 AES256, which would have avoided the problems from the beginning;
9775 however, bits are scarce, so we can only do this in a new release
9776 (not just a patchlevel) when we can change the SSL_CIPHER
9777 definition to split the single 'unsigned long mask' bitmap into
9778 multiple values to extend the available space.
9782 ### Changes between 0.9.8b and 0.9.8c [05 Sep 2006]
9784 * Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
9785 ([CVE-2006-4339]) [Ben Laurie and Google Security Team]
9787 * Add AES IGE and biIGE modes.
9791 * Change the Unix randomness entropy gathering to use poll() when
9792 possible instead of select(), since the latter has some
9793 undesirable limitations.
9795 *Darryl Miles via Richard Levitte and Bodo Moeller*
9797 * Disable "ECCdraft" ciphersuites more thoroughly. Now special
9798 treatment in ssl/ssl_ciph.s makes sure that these ciphersuites
9799 cannot be implicitly activated as part of, e.g., the "AES" alias.
9800 However, please upgrade to OpenSSL 0.9.9[-dev] for
9801 non-experimental use of the ECC ciphersuites to get TLS extension
9802 support, which is required for curve and point format negotiation
9803 to avoid potential handshake problems.
9807 * Disable rogue ciphersuites:
9809 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
9810 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
9811 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
9813 The latter two were purportedly from
9814 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
9817 Also deactivate the remaining ciphersuites from
9818 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
9819 unofficial, and the ID has long expired.
9823 * Fix RSA blinding Heisenbug (problems sometimes occurred on
9824 dual-core machines) and other potential thread-safety issues.
9828 * Add the symmetric cipher Camellia (128-bit, 192-bit, 256-bit key
9829 versions), which is now available for royalty-free use
9830 (see <http://info.isl.ntt.co.jp/crypt/eng/info/chiteki.html>).
9831 Also, add Camellia TLS ciphersuites from RFC 4132.
9833 To minimize changes between patchlevels in the OpenSSL 0.9.8
9834 series, Camellia remains excluded from compilation unless OpenSSL
9835 is configured with 'enable-camellia'.
9839 * Disable the padding bug check when compression is in use. The padding
9840 bug check assumes the first packet is of even length, this is not
9841 necessarily true if compression is enabled and can result in false
9842 positives causing handshake failure. The actual bug test is ancient
9843 code so it is hoped that implementations will either have fixed it by
9844 now or any which still have the bug do not support compression.
9848 ### Changes between 0.9.8a and 0.9.8b [04 May 2006]
9850 * When applying a cipher rule check to see if string match is an explicit
9851 cipher suite and only match that one cipher suite if it is.
9855 * Link in manifests for VC++ if needed.
9857 *Austin Ziegler <halostatue@gmail.com>*
9859 * Update support for ECC-based TLS ciphersuites according to
9860 draft-ietf-tls-ecc-12.txt with proposed changes (but without
9861 TLS extensions, which are supported starting with the 0.9.9
9862 branch, not in the OpenSSL 0.9.8 branch).
9866 * New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support
9867 opaque EVP_CIPHER_CTX handling.
9871 * Fixes and enhancements to zlib compression code. We now only use
9872 "zlib1.dll" and use the default `__cdecl` calling convention on Win32
9873 to conform with the standards mentioned here:
9874 <http://www.zlib.net/DLL_FAQ.txt>
9875 Static zlib linking now works on Windows and the new --with-zlib-include
9876 --with-zlib-lib options to Configure can be used to supply the location
9877 of the headers and library. Gracefully handle case where zlib library
9882 * Several fixes and enhancements to the OID generation code. The old code
9883 sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't
9884 handle numbers larger than ULONG_MAX, truncated printing and had a
9885 non standard OBJ_obj2txt() behaviour.
9889 * Add support for building of engines under engine/ as shared libraries
9890 under VC++ build system.
9894 * Corrected the numerous bugs in the Win32 path splitter in DSO.
9895 Hopefully, we will not see any false combination of paths any more.
9899 ### Changes between 0.9.8 and 0.9.8a [11 Oct 2005]
9901 * Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
9902 (part of SSL_OP_ALL). This option used to disable the
9903 countermeasure against man-in-the-middle protocol-version
9904 rollback in the SSL 2.0 server implementation, which is a bad
9905 idea. ([CVE-2005-2969])
9907 *Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
9908 for Information Security, National Institute of Advanced Industrial
9909 Science and Technology [AIST], Japan)*
9911 * Add two function to clear and return the verify parameter flags.
9915 * Keep cipherlists sorted in the source instead of sorting them at
9916 runtime, thus removing the need for a lock.
9920 * Avoid some small subgroup attacks in Diffie-Hellman.
9922 *Nick Mathewson and Ben Laurie*
9924 * Add functions for well-known primes.
9928 * Extended Windows CE support.
9930 *Satoshi Nakamura and Andy Polyakov*
9932 * Initialize SSL_METHOD structures at compile time instead of during
9933 runtime, thus removing the need for a lock.
9937 * Make PKCS7_decrypt() work even if no certificate is supplied by
9938 attempting to decrypt each encrypted key in turn. Add support to
9943 ### Changes between 0.9.7h and 0.9.8 [05 Jul 2005]
9945 [NB: OpenSSL 0.9.7i and later 0.9.7 patch levels were released after
9948 * Add libcrypto.pc and libssl.pc for those who feel they need them.
9952 * Change CA.sh and CA.pl so they don't bundle the CSR and the private
9953 key into the same file any more.
9957 * Add initial support for Win64, both IA64 and AMD64/x64 flavors.
9961 * Add -utf8 command line and config file option to 'ca'.
9963 *Stefan <stf@udoma.org*
9965 * Removed the macro des_crypt(), as it seems to conflict with some
9966 libraries. Use DES_crypt().
9970 * Correct naming of the 'chil' and '4758cca' ENGINEs. This
9971 involves renaming the source and generated shared-libs for
9972 both. The engines will accept the corrected or legacy ids
9973 ('ncipher' and '4758_cca' respectively) when binding. NB,
9974 this only applies when building 'shared'.
9976 *Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe*
9978 * Add attribute functions to EVP_PKEY structure. Modify
9979 PKCS12_create() to recognize a CSP name attribute and
9980 use it. Make -CSP option work again in pkcs12 utility.
9984 * Add new functionality to the bn blinding code:
9985 - automatic re-creation of the BN_BLINDING parameters after
9986 a fixed number of uses (currently 32)
9987 - add new function for parameter creation
9988 - introduce flags to control the update behaviour of the
9989 BN_BLINDING parameters
9990 - hide BN_BLINDING structure
9991 Add a second BN_BLINDING slot to the RSA structure to improve
9992 performance when a single RSA object is shared among several
9997 * Add support for DTLS.
9999 *Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie*
10001 * Add support for DER encoded private keys (SSL_FILETYPE_ASN1)
10002 to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file()
10006 * Remove buggy and incomplete DH cert support from
10007 ssl/ssl_rsa.c and ssl/s3_both.c
10011 * Use SHA-1 instead of MD5 as the default digest algorithm for
10012 the `apps/openssl` commands.
10016 * Compile clean with "-Wall -Wmissing-prototypes
10017 -Wstrict-prototypes -Wmissing-declarations -Werror". Currently
10018 DEBUG_SAFESTACK must also be set.
10022 * Change ./Configure so that certain algorithms can be disabled by default.
10023 The new counterpiece to "no-xxx" is "enable-xxx".
10025 The patented RC5 and MDC2 algorithms will now be disabled unless
10026 "enable-rc5" and "enable-mdc2", respectively, are specified.
10028 (IDEA remains enabled despite being patented. This is because IDEA
10029 is frequently required for interoperability, and there is no license
10030 fee for non-commercial use. As before, "no-idea" can be used to
10031 avoid this algorithm.)
10035 * Add processing of proxy certificates (see RFC 3820). This work was
10036 sponsored by KTH (The Royal Institute of Technology in Stockholm) and
10037 EGEE (Enabling Grids for E-science in Europe).
10041 * RC4 performance overhaul on modern architectures/implementations, such
10042 as Intel P4, IA-64 and AMD64.
10046 * New utility extract-section.pl. This can be used specify an alternative
10047 section number in a pod file instead of having to treat each file as
10048 a separate case in Makefile. This can be done by adding two lines to the
10051 =for comment openssl_section:XXX
10053 The blank line is mandatory.
10057 * New arguments -certform, -keyform and -pass for s_client and s_server
10058 to allow alternative format key and certificate files and passphrase
10063 * New structure X509_VERIFY_PARAM which combines current verify parameters,
10064 update associated structures and add various utility functions.
10066 Add new policy related verify parameters, include policy checking in
10067 standard verify code. Enhance 'smime' application with extra parameters
10068 to support policy checking and print out.
10072 * Add a new engine to support VIA PadLock ACE extensions in the VIA C3
10073 Nehemiah processors. These extensions support AES encryption in hardware
10074 as well as RNG (though RNG support is currently disabled).
10076 *Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov*
10078 * Deprecate `BN_[get|set]_params()` functions (they were ignored internally).
10082 * New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented.
10084 *Andy Polyakov and a number of other people*
10086 * Improved PowerPC platform support. Most notably BIGNUM assembler
10087 implementation contributed by IBM.
10089 *Suresh Chari, Peter Waltenberg, Andy Polyakov*
10091 * The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public
10092 exponent rather than 'unsigned long'. There is a corresponding change to
10093 the new 'rsa_keygen' element of the RSA_METHOD structure.
10095 *Jelte Jansen, Geoff Thorpe*
10097 * Functionality for creating the initial serial number file is now
10098 moved from CA.pl to the 'ca' utility with a new option -create_serial.
10100 (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial
10101 number file to 1, which is bound to cause problems. To avoid
10102 the problems while respecting compatibility between different 0.9.7
10103 patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in
10104 CA.pl for serial number initialization. With the new release 0.9.8,
10105 we can fix the problem directly in the 'ca' utility.)
10109 * Reduced header interdependencies by declaring more opaque objects in
10110 ossl_typ.h. As a consequence, including some headers (eg. engine.h) will
10111 give fewer recursive includes, which could break lazy source code - so
10112 this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always,
10113 developers should define this symbol when building and using openssl to
10114 ensure they track the recommended behaviour, interfaces, [etc], but
10115 backwards-compatible behaviour prevails when this isn't defined.
10119 * New function X509_POLICY_NODE_print() which prints out policy nodes.
10123 * Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality.
10124 This will generate a random key of the appropriate length based on the
10125 cipher context. The EVP_CIPHER can provide its own random key generation
10126 routine to support keys of a specific form. This is used in the des and
10127 3des routines to generate a key of the correct parity. Update S/MIME
10128 code to use new functions and hence generate correct parity DES keys.
10129 Add EVP_CHECK_DES_KEY #define to return an error if the key is not
10130 valid (weak or incorrect parity).
10134 * Add a local set of CRLs that can be used by X509_verify_cert() as well
10135 as looking them up. This is useful when the verified structure may contain
10136 CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs
10137 present unless the new PKCS7_NO_CRL flag is asserted.
10141 * Extend ASN1 oid configuration module. It now additionally accepts the
10144 shortName = some long name, 1.2.3.4
10148 * Reimplemented the BN_CTX implementation. There is now no more static
10149 limitation on the number of variables it can handle nor the depth of the
10150 "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack
10151 information can now expand as required, and rather than having a single
10152 static array of bignums, BN_CTX now uses a linked-list of such arrays
10153 allowing it to expand on demand whilst maintaining the usefulness of
10154 BN_CTX's "bundling".
10158 * Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD
10159 to allow all RSA operations to function using a single BN_CTX.
10163 * Preliminary support for certificate policy evaluation and checking. This
10164 is initially intended to pass the tests outlined in "Conformance Testing
10165 of Relying Party Client Certificate Path Processing Logic" v1.07.
10169 * bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and
10170 remained unused and not that useful. A variety of other little bignum
10171 tweaks and fixes have also been made continuing on from the audit (see
10176 * Constify all or almost all d2i, c2i, s2i and r2i functions, along with
10177 associated ASN1, EVP and SSL functions and old ASN1 macros.
10181 * BN_zero() only needs to set 'top' and 'neg' to zero for correct results,
10182 and this should never fail. So the return value from the use of
10183 BN_set_word() (which can fail due to needless expansion) is now deprecated;
10184 if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro.
10188 * BN_CTX_get() should return zero-valued bignums, providing the same
10189 initialised value as BN_new().
10191 *Geoff Thorpe, suggested by Ulf Möller*
10193 * Support for inhibitAnyPolicy certificate extension.
10197 * An audit of the BIGNUM code is underway, for which debugging code is
10198 enabled when BN_DEBUG is defined. This makes stricter enforcements on what
10199 is considered valid when processing BIGNUMs, and causes execution to
10200 assert() when a problem is discovered. If BN_DEBUG_RAND is defined,
10201 further steps are taken to deliberately pollute unused data in BIGNUM
10202 structures to try and expose faulty code further on. For now, openssl will
10203 (in its default mode of operation) continue to tolerate the inconsistent
10204 forms that it has tolerated in the past, but authors and packagers should
10205 consider trying openssl and their own applications when compiled with
10206 these debugging symbols defined. It will help highlight potential bugs in
10207 their own code, and will improve the test coverage for OpenSSL itself. At
10208 some point, these tighter rules will become openssl's default to improve
10209 maintainability, though the assert()s and other overheads will remain only
10210 in debugging configurations. See bn.h for more details.
10212 *Geoff Thorpe, Nils Larsch, Ulf Möller*
10214 * BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure
10215 that can only be obtained through BN_CTX_new() (which implicitly
10216 initialises it). The presence of this function only made it possible
10217 to overwrite an existing structure (and cause memory leaks).
10221 * Because of the callback-based approach for implementing LHASH as a
10222 template type, lh_insert() adds opaque objects to hash-tables and
10223 lh_doall() or lh_doall_arg() are typically used with a destructor callback
10224 to clean up those corresponding objects before destroying the hash table
10225 (and losing the object pointers). So some over-zealous constifications in
10226 LHASH have been relaxed so that lh_insert() does not take (nor store) the
10227 objects as "const" and the `lh_doall[_arg]` callback wrappers are not
10228 prototyped to have "const" restrictions on the object pointers they are
10229 given (and so aren't required to cast them away any more).
10233 * The tmdiff.h API was so ugly and minimal that our own timing utility
10234 (speed) prefers to use its own implementation. The two implementations
10235 haven't been consolidated as yet (volunteers?) but the tmdiff API has had
10236 its object type properly exposed (MS_TM) instead of casting to/from
10237 `char *`. This may still change yet if someone realises MS_TM and
10239 aren't necessarily the greatest nomenclatures - but this is what was used
10240 internally to the implementation so I've used that for now.
10244 * Ensure that deprecated functions do not get compiled when
10245 OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of
10246 the self-tests were still using deprecated key-generation functions so
10247 these have been updated also.
10251 * Reorganise PKCS#7 code to separate the digest location functionality
10252 into PKCS7_find_digest(), digest addition into PKCS7_bio_add_digest().
10253 New function PKCS7_set_digest() to set the digest type for PKCS#7
10254 digestedData type. Add additional code to correctly generate the
10255 digestedData type and add support for this type in PKCS7 initialization
10260 * New function PKCS7_set0_type_other() this initializes a PKCS7
10261 structure of type "other".
10265 * Fix prime generation loop in crypto/bn/bn_prime.pl by making
10266 sure the loop does correctly stop and breaking ("division by zero")
10267 modulus operations are not performed. The (pre-generated) prime
10268 table crypto/bn/bn_prime.h was already correct, but it could not be
10269 re-generated on some platforms because of the "division by zero"
10270 situation in the script.
10272 *Ralf S. Engelschall*
10274 * Update support for ECC-based TLS ciphersuites according to
10275 draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with
10276 SHA-1 now is only used for "small" curves (where the
10277 representation of a field element takes up to 24 bytes); for
10278 larger curves, the field element resulting from ECDH is directly
10279 used as premaster secret.
10281 *Douglas Stebila (Sun Microsystems Laboratories)*
10283 * Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2
10284 curve secp160r1 to the tests.
10286 *Douglas Stebila (Sun Microsystems Laboratories)*
10288 * Add the possibility to load symbols globally with DSO.
10290 *Götz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte*
10292 * Add the functions ERR_set_mark() and ERR_pop_to_mark() for better
10293 control of the error stack.
10297 * Add support for STORE in ENGINE.
10301 * Add the STORE type. The intention is to provide a common interface
10302 to certificate and key stores, be they simple file-based stores, or
10303 HSM-type store, or LDAP stores, or...
10304 NOTE: The code is currently UNTESTED and isn't really used anywhere.
10308 * Add a generic structure called OPENSSL_ITEM. This can be used to
10309 pass a list of arguments to any function as well as provide a way
10310 for a function to pass data back to the caller.
10314 * Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup()
10315 works like BUF_strdup() but can be used to duplicate a portion of
10316 a string. The copy gets NUL-terminated. BUF_memdup() duplicates
10321 * Add the function sk_find_ex() which works like sk_find(), but will
10322 return an index to an element even if an exact match couldn't be
10323 found. The index is guaranteed to point at the element where the
10324 searched-for key would be inserted to preserve sorting order.
10328 * Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but
10329 takes an extra flags argument for optional functionality. Currently,
10330 the following flags are defined:
10332 OBJ_BSEARCH_VALUE_ON_NOMATCH
10333 This one gets OBJ_bsearch_ex() to return a pointer to the first
10334 element where the comparing function returns a negative or zero
10337 OBJ_BSEARCH_FIRST_VALUE_ON_MATCH
10338 This one gets OBJ_bsearch_ex() to return a pointer to the first
10339 element where the comparing function returns zero. This is useful
10340 if there are more than one element where the comparing function
10345 * Make it possible to create self-signed certificates with 'openssl ca'
10346 in such a way that the self-signed certificate becomes part of the
10347 CA database and uses the same mechanisms for serial number generation
10348 as all other certificate signing. The new flag '-selfsign' enables
10349 this functionality. Adapt CA.sh and CA.pl.in.
10353 * Add functionality to check the public key of a certificate request
10354 against a given private. This is useful to check that a certificate
10355 request can be signed by that key (self-signing).
10359 * Make it possible to have multiple active certificates with the same
10360 subject in the CA index file. This is done only if the keyword
10361 'unique_subject' is set to 'no' in the main CA section (default
10362 if 'CA_default') of the configuration file. The value is saved
10363 with the database itself in a separate index attribute file,
10364 named like the index file with '.attr' appended to the name.
10368 * Generate multi-valued AVAs using '+' notation in config files for
10373 * Support for nameConstraints certificate extension.
10377 * Support for policyConstraints certificate extension.
10381 * Support for policyMappings certificate extension.
10385 * Make sure the default DSA_METHOD implementation only uses its
10386 dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL,
10387 and change its own handlers to be NULL so as to remove unnecessary
10388 indirection. This lets alternative implementations fallback to the
10389 default implementation more easily.
10393 * Support for directoryName in GeneralName related extensions
10398 * Make it possible to link applications using Makefile.shared.
10399 Make that possible even when linking against static libraries!
10403 * Support for single pass processing for S/MIME signing. This now
10404 means that S/MIME signing can be done from a pipe, in addition
10405 cleartext signing (multipart/signed type) is effectively streaming
10406 and the signed data does not need to be all held in memory.
10408 This is done with a new flag PKCS7_STREAM. When this flag is set
10409 PKCS7_sign() only initializes the PKCS7 structure and the actual signing
10410 is done after the data is output (and digests calculated) in
10411 SMIME_write_PKCS7().
10415 * Add full support for -rpath/-R, both in shared libraries and
10416 applications, at least on the platforms where it's known how
10421 * In crypto/ec/ec_mult.c, implement fast point multiplication with
10422 precomputation, based on wNAF splitting: EC_GROUP_precompute_mult()
10423 will now compute a table of multiples of the generator that
10424 makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul()
10425 faster (notably in the case of a single point multiplication,
10426 scalar * generator).
10428 *Nils Larsch, Bodo Moeller*
10430 * IPv6 support for certificate extensions. The various extensions
10431 which use the IP:a.b.c.d can now take IPv6 addresses using the
10432 formats of RFC1884 2.2 . IPv6 addresses are now also displayed
10437 * Added an ENGINE that implements RSA by performing private key
10438 exponentiations with the GMP library. The conversions to and from
10439 GMP's mpz_t format aren't optimised nor are any montgomery forms
10440 cached, and on x86 it appears OpenSSL's own performance has caught up.
10441 However there are likely to be other architectures where GMP could
10442 provide a boost. This ENGINE is not built in by default, but it can be
10443 specified at Configure time and should be accompanied by the necessary
10444 linker additions, eg;
10445 ./config -DOPENSSL_USE_GMP -lgmp
10449 * "openssl engine" will not display ENGINE/DSO load failure errors when
10450 testing availability of engines with "-t" - the old behaviour is
10451 produced by increasing the feature's verbosity with "-tt".
10455 * ECDSA routines: under certain error conditions uninitialized BN objects
10456 could be freed. Solution: make sure initialization is performed early
10457 enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de>
10462 * Key-generation can now be implemented in RSA_METHOD, DSA_METHOD
10463 and DH_METHOD (eg. by ENGINE implementations) to override the normal
10464 software implementations. For DSA and DH, parameter generation can
10465 also be overridden by providing the appropriate method callbacks.
10469 * Change the "progress" mechanism used in key-generation and
10470 primality testing to functions that take a new BN_GENCB pointer in
10471 place of callback/argument pairs. The new API functions have `_ex`
10472 postfixes and the older functions are reimplemented as wrappers for
10473 the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide
10474 declarations of the old functions to help (graceful) attempts to
10475 migrate to the new functions. Also, the new key-generation API
10476 functions operate on a caller-supplied key-structure and return
10477 success/failure rather than returning a key or NULL - this is to
10478 help make "keygen" another member function of RSA_METHOD etc.
10480 Example for using the new callback interface:
10482 int (*my_callback)(int a, int b, BN_GENCB *cb) = ...;
10483 void *my_arg = ...;
10486 BN_GENCB_set(&my_cb, my_callback, my_arg);
10488 return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb);
10489 /* For the meaning of a, b in calls to my_callback(), see the
10490 * documentation of the function that calls the callback.
10491 * cb will point to my_cb; my_arg can be retrieved as cb->arg.
10492 * my_callback should return 1 if it wants BN_is_prime_ex()
10493 * to continue, or 0 to stop.
10498 * Change the ZLIB compression method to be stateful, and make it
10499 available to TLS with the number defined in
10500 draft-ietf-tls-compression-04.txt.
10504 * Add the ASN.1 structures and functions for CertificatePair, which
10505 is defined as follows (according to X.509_4thEditionDraftV6.pdf):
10507 CertificatePair ::= SEQUENCE {
10508 forward [0] Certificate OPTIONAL,
10509 reverse [1] Certificate OPTIONAL,
10510 -- at least one of the pair shall be present -- }
10512 Also implement the PEM functions to read and write certificate
10513 pairs, and defined the PEM tag as "CERTIFICATE PAIR".
10515 This needed to be defined, mostly for the sake of the LDAP
10516 attribute crossCertificatePair, but may prove useful elsewhere as
10521 * Make it possible to inhibit symlinking of shared libraries in
10522 Makefile.shared, for Cygwin's sake.
10526 * Extend the BIGNUM API by creating a function
10527 void BN_set_negative(BIGNUM *a, int neg);
10528 and a macro that behave like
10529 int BN_is_negative(const BIGNUM *a);
10531 to avoid the need to access 'a->neg' directly in applications.
10535 * Implement fast modular reduction for pseudo-Mersenne primes
10536 used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c).
10537 EC_GROUP_new_curve_GFp() will now automatically use this
10540 *Nils Larsch <nla@trustcenter.de>*
10542 * Add new lock type (CRYPTO_LOCK_BN).
10546 * Change the ENGINE framework to automatically load engines
10547 dynamically from specific directories unless they could be
10548 found to already be built in or loaded. Move all the
10549 current engines except for the cryptodev one to a new
10550 directory engines/.
10551 The engines in engines/ are built as shared libraries if
10552 the "shared" options was given to ./Configure or ./config.
10553 Otherwise, they are inserted in libcrypto.a.
10554 /usr/local/ssl/engines is the default directory for dynamic
10555 engines, but that can be overridden at configure time through
10556 the usual use of --prefix and/or --openssldir, and at run
10557 time with the environment variable OPENSSL_ENGINES.
10559 *Geoff Thorpe and Richard Levitte*
10561 * Add Makefile.shared, a helper makefile to build shared
10562 libraries. Adapt Makefile.org.
10566 * Add version info to Win32 DLLs.
10568 *Peter 'Luna' Runestig" <peter@runestig.com>*
10570 * Add new 'medium level' PKCS#12 API. Certificates and keys
10571 can be added using this API to created arbitrary PKCS#12
10572 files while avoiding the low-level API.
10574 New options to PKCS12_create(), key or cert can be NULL and
10575 will then be omitted from the output file. The encryption
10576 algorithm NIDs can be set to -1 for no encryption, the mac
10577 iteration count can be set to 0 to omit the mac.
10579 Enhance pkcs12 utility by making the -nokeys and -nocerts
10580 options work when creating a PKCS#12 file. New option -nomac
10581 to omit the mac, NONE can be set for an encryption algorithm.
10582 New code is modified to use the enhanced PKCS12_create()
10583 instead of the low-level API.
10587 * Extend ASN1 encoder to support indefinite length constructed
10588 encoding. This can output sequences tags and octet strings in
10589 this form. Modify pk7_asn1.c to support indefinite length
10590 encoding. This is experimental and needs additional code to
10591 be useful, such as an ASN1 bio and some enhanced streaming
10594 Extend template encode functionality so that tagging is passed
10595 down to the template encoder.
10599 * Let 'openssl req' fail if an argument to '-newkey' is not
10600 recognized instead of using RSA as a default.
10604 * Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt.
10605 As these are not official, they are not included in "ALL";
10606 the "ECCdraft" ciphersuite group alias can be used to select them.
10608 *Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)*
10610 * Add ECDH engine support.
10612 *Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)*
10614 * Add ECDH in new directory crypto/ecdh/.
10616 *Douglas Stebila (Sun Microsystems Laboratories)*
10618 * Let BN_rand_range() abort with an error after 100 iterations
10619 without success (which indicates a broken PRNG).
10623 * Change BN_mod_sqrt() so that it verifies that the input value
10624 is really the square of the return value. (Previously,
10625 BN_mod_sqrt would show GIGO behaviour.)
10629 * Add named elliptic curves over binary fields from X9.62, SECG,
10630 and WAP/WTLS; add OIDs that were still missing.
10632 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10634 * Extend the EC library for elliptic curves over binary fields
10635 (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/).
10638 EC_GF2m_simple_method
10642 EC_GROUP_new_curve_GF2m
10643 EC_GROUP_set_curve_GF2m
10644 EC_GROUP_get_curve_GF2m
10645 EC_POINT_set_affine_coordinates_GF2m
10646 EC_POINT_get_affine_coordinates_GF2m
10647 EC_POINT_set_compressed_coordinates_GF2m
10649 Point compression for binary fields is disabled by default for
10650 patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to
10653 As binary polynomials are represented as BIGNUMs, various members
10654 of the EC_GROUP and EC_POINT data structures can be shared
10655 between the implementations for prime fields and binary fields;
10656 the above `..._GF2m functions` (except for EX_GROUP_new_curve_GF2m)
10657 are essentially identical to their `..._GFp` counterparts.
10658 (For simplicity, the `..._GFp` prefix has been dropped from
10659 various internal method names.)
10661 An internal 'field_div' method (similar to 'field_mul' and
10662 'field_sqr') has been added; this is used only for binary fields.
10664 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10666 * Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult()
10667 through methods ('mul', 'precompute_mult').
10669 The generic implementations (now internally called 'ec_wNAF_mul'
10670 and 'ec_wNAF_precomputed_mult') remain the default if these
10671 methods are undefined.
10673 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10675 * New function EC_GROUP_get_degree, which is defined through
10676 EC_METHOD. For curves over prime fields, this returns the bit
10677 length of the modulus.
10679 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10681 * New functions EC_GROUP_dup, EC_POINT_dup.
10682 (These simply call ..._new and ..._copy).
10684 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10686 * Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c.
10687 Polynomials are represented as BIGNUMs (where the sign bit is not
10688 used) in the following functions [macros]:
10691 BN_GF2m_sub [= BN_GF2m_add]
10692 BN_GF2m_mod [wrapper for BN_GF2m_mod_arr]
10693 BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr]
10694 BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr]
10696 BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr]
10697 BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr]
10698 BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr]
10699 BN_GF2m_cmp [= BN_ucmp]
10701 (Note that only the 'mod' functions are actually for fields GF(2^m).
10702 BN_GF2m_add() is misnomer, but this is for the sake of consistency.)
10704 For some functions, an the irreducible polynomial defining a
10705 field can be given as an 'unsigned int[]' with strictly
10706 decreasing elements giving the indices of those bits that are set;
10707 i.e., p[] represents the polynomial
10708 f(t) = t^p[0] + t^p[1] + ... + t^p[k]
10710 p[0] > p[1] > ... > p[k] = 0.
10711 This applies to the following functions:
10714 BN_GF2m_mod_mul_arr
10715 BN_GF2m_mod_sqr_arr
10716 BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv]
10717 BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div]
10718 BN_GF2m_mod_exp_arr
10719 BN_GF2m_mod_sqrt_arr
10720 BN_GF2m_mod_solve_quad_arr
10724 Conversion can be performed by the following functions:
10729 bntest.c has additional tests for binary polynomial arithmetic.
10731 Two implementations for BN_GF2m_mod_div() are available.
10732 The default algorithm simply uses BN_GF2m_mod_inv() and
10733 BN_GF2m_mod_mul(). The alternative algorithm is compiled in only
10734 if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the
10735 copyright notice in crypto/bn/bn_gf2m.c before enabling it).
10737 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10739 * Add new error code 'ERR_R_DISABLED' that can be used when some
10740 functionality is disabled at compile-time.
10742 *Douglas Stebila <douglas.stebila@sun.com>*
10744 * Change default behaviour of 'openssl asn1parse' so that more
10745 information is visible when viewing, e.g., a certificate:
10747 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump'
10748 mode the content of non-printable OCTET STRINGs is output in a
10749 style similar to INTEGERs, but with '[HEX DUMP]' prepended to
10750 avoid the appearance of a printable string.
10752 *Nils Larsch <nla@trustcenter.de>*
10754 * Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access
10756 EC_GROUP_set_asn1_flag()
10757 EC_GROUP_get_asn1_flag()
10758 EC_GROUP_set_point_conversion_form()
10759 EC_GROUP_get_point_conversion_form()
10760 These control ASN1 encoding details:
10761 - Curves (i.e., groups) are encoded explicitly unless asn1_flag
10762 has been set to OPENSSL_EC_NAMED_CURVE.
10763 - Points are encoded in uncompressed form by default; options for
10764 asn1_for are as for point2oct, namely
10765 POINT_CONVERSION_COMPRESSED
10766 POINT_CONVERSION_UNCOMPRESSED
10767 POINT_CONVERSION_HYBRID
10769 Also add 'seed' and 'seed_len' members to EC_GROUP with access
10771 EC_GROUP_set_seed()
10772 EC_GROUP_get0_seed()
10773 EC_GROUP_get_seed_len()
10774 This is used only for ASN1 purposes (so far).
10776 *Nils Larsch <nla@trustcenter.de>*
10778 * Add 'field_type' member to EC_METHOD, which holds the NID
10779 of the appropriate field type OID. The new function
10780 EC_METHOD_get_field_type() returns this value.
10782 *Nils Larsch <nla@trustcenter.de>*
10785 EC_POINT_point2bn()
10786 EC_POINT_bn2point()
10787 EC_POINT_point2hex()
10788 EC_POINT_hex2point()
10789 providing useful interfaces to EC_POINT_point2oct() and
10790 EC_POINT_oct2point().
10792 *Nils Larsch <nla@trustcenter.de>*
10794 * Change internals of the EC library so that the functions
10795 EC_GROUP_set_generator()
10796 EC_GROUP_get_generator()
10797 EC_GROUP_get_order()
10798 EC_GROUP_get_cofactor()
10799 are implemented directly in crypto/ec/ec_lib.c and not dispatched
10800 to methods, which would lead to unnecessary code duplication when
10801 adding different types of curves.
10803 *Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller*
10805 * Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM
10806 arithmetic, and such that modified wNAFs are generated
10807 (which avoid length expansion in many cases).
10811 * Add a function EC_GROUP_check_discriminant() (defined via
10812 EC_METHOD) that verifies that the curve discriminant is non-zero.
10814 Add a function EC_GROUP_check() that makes some sanity tests
10815 on a EC_GROUP, its generator and order. This includes
10816 EC_GROUP_check_discriminant().
10818 *Nils Larsch <nla@trustcenter.de>*
10820 * Add ECDSA in new directory crypto/ecdsa/.
10822 Add applications 'openssl ecparam' and 'openssl ecdsa'
10823 (these are based on 'openssl dsaparam' and 'openssl dsa').
10825 ECDSA support is also included in various other files across the
10826 library. Most notably,
10827 - 'openssl req' now has a '-newkey ecdsa:file' option;
10828 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA;
10829 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and
10830 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make
10831 them suitable for ECDSA where domain parameters must be
10832 extracted before the specific public key;
10833 - ECDSA engine support has been added.
10835 *Nils Larsch <nla@trustcenter.de>*
10837 * Include some named elliptic curves, and add OIDs from X9.62,
10838 SECG, and WAP/WTLS. Each curve can be obtained from the new
10840 EC_GROUP_new_by_curve_name(),
10841 and the list of available named curves can be obtained with
10842 EC_get_builtin_curves().
10843 Also add a 'curve_name' member to EC_GROUP objects, which can be
10845 EC_GROUP_set_curve_name()
10846 EC_GROUP_get_curve_name()
10848 *Nils Larsch <larsch@trustcenter.de, Bodo Moeller*
10850 * Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
10851 was actually never needed) and in BN_mul(). The removal in BN_mul()
10852 required a small change in bn_mul_part_recursive() and the addition
10853 of the functions bn_cmp_part_words(), bn_sub_part_words() and
10854 bn_add_part_words(), which do the same thing as bn_cmp_words(),
10855 bn_sub_words() and bn_add_words() except they take arrays with
10860 ### Changes between 0.9.7l and 0.9.7m [23 Feb 2007]
10862 * Cleanse PEM buffers before freeing them since they may contain
10865 *Benjamin Bennett <ben@psc.edu>*
10867 * Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
10868 a ciphersuite string such as "DEFAULT:RSA" cannot enable
10869 authentication-only ciphersuites.
10873 * Since AES128 and AES256 share a single mask bit in the logic of
10874 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
10875 kludge to work properly if AES128 is available and AES256 isn't.
10879 * Expand security boundary to match 1.1.1 module.
10883 * Remove redundant features: hash file source, editing of test vectors
10884 modify fipsld to use external fips_premain.c signature.
10888 * New perl script mkfipsscr.pl to create shell scripts or batch files to
10889 run algorithm test programs.
10893 * Make algorithm test programs more tolerant of whitespace.
10897 * Have SSL/TLS server implementation tolerate "mismatched" record
10898 protocol version while receiving ClientHello even if the
10899 ClientHello is fragmented. (The server can't insist on the
10900 particular protocol version it has chosen before the ServerHello
10901 message has informed the client about his choice.)
10905 * Load error codes if they are not already present instead of using a
10906 static variable. This allows them to be cleanly unloaded and reloaded.
10910 ### Changes between 0.9.7k and 0.9.7l [28 Sep 2006]
10912 * Introduce limits to prevent malicious keys being able to
10913 cause a denial of service. ([CVE-2006-2940])
10915 *Steve Henson, Bodo Moeller*
10917 * Fix ASN.1 parsing of certain invalid structures that can result
10918 in a denial of service. ([CVE-2006-2937]) [Steve Henson]
10920 * Fix buffer overflow in SSL_get_shared_ciphers() function.
10921 ([CVE-2006-3738]) [Tavis Ormandy and Will Drewry, Google Security Team]
10923 * Fix SSL client code which could crash if connecting to a
10924 malicious SSLv2 server. ([CVE-2006-4343])
10926 *Tavis Ormandy and Will Drewry, Google Security Team*
10928 * Change ciphersuite string processing so that an explicit
10929 ciphersuite selects this one ciphersuite (so that "AES256-SHA"
10930 will no longer include "AES128-SHA"), and any other similar
10931 ciphersuite (same bitmap) from *other* protocol versions (so that
10932 "RC4-MD5" will still include both the SSL 2.0 ciphersuite and the
10933 SSL 3.0/TLS 1.0 ciphersuite). This is a backport combining
10934 changes from 0.9.8b and 0.9.8d.
10938 ### Changes between 0.9.7j and 0.9.7k [05 Sep 2006]
10940 * Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
10941 ([CVE-2006-4339]) [Ben Laurie and Google Security Team]
10943 * Change the Unix randomness entropy gathering to use poll() when
10944 possible instead of select(), since the latter has some
10945 undesirable limitations.
10947 *Darryl Miles via Richard Levitte and Bodo Moeller*
10949 * Disable rogue ciphersuites:
10951 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
10952 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
10953 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
10955 The latter two were purportedly from
10956 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
10959 Also deactivate the remaining ciphersuites from
10960 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
10961 unofficial, and the ID has long expired.
10965 * Fix RSA blinding Heisenbug (problems sometimes occurred on
10966 dual-core machines) and other potential thread-safety issues.
10970 ### Changes between 0.9.7i and 0.9.7j [04 May 2006]
10972 * Adapt fipsld and the build system to link against the validated FIPS
10973 module in FIPS mode.
10977 * Fixes for VC++ 2005 build under Windows.
10981 * Add new Windows build target VC-32-GMAKE for VC++. This uses GNU make
10982 from a Windows bash shell such as MSYS. It is autodetected from the
10983 "config" script when run from a VC++ environment. Modify standard VC++
10984 build to use fipscanister.o from the GNU make build.
10988 ### Changes between 0.9.7h and 0.9.7i [14 Oct 2005]
10990 * Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS.
10991 The value now differs depending on if you build for FIPS or not.
10992 BEWARE! A program linked with a shared FIPSed libcrypto can't be
10993 safely run with a non-FIPSed libcrypto, as it may crash because of
10994 the difference induced by this change.
10998 ### Changes between 0.9.7g and 0.9.7h [11 Oct 2005]
11000 * Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
11001 (part of SSL_OP_ALL). This option used to disable the
11002 countermeasure against man-in-the-middle protocol-version
11003 rollback in the SSL 2.0 server implementation, which is a bad
11004 idea. ([CVE-2005-2969])
11006 *Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
11007 for Information Security, National Institute of Advanced Industrial
11008 Science and Technology [AIST, Japan)]*
11010 * Minimal support for X9.31 signatures and PSS padding modes. This is
11011 mainly for FIPS compliance and not fully integrated at this stage.
11015 * For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform
11016 the exponentiation using a fixed-length exponent. (Otherwise,
11017 the information leaked through timing could expose the secret key
11018 after many signatures; cf. Bleichenbacher's attack on DSA with
11023 * Make a new fixed-window mod_exp implementation the default for
11024 RSA, DSA, and DH private-key operations so that the sequence of
11025 squares and multiplies and the memory access pattern are
11026 independent of the particular secret key. This will mitigate
11027 cache-timing and potential related attacks.
11029 BN_mod_exp_mont_consttime() is the new exponentiation implementation,
11030 and this is automatically used by BN_mod_exp_mont() if the new flag
11031 BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH
11032 will use this BN flag for private exponents unless the flag
11033 RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or
11034 DH_FLAG_NO_EXP_CONSTTIME, respectively, is set.
11036 *Matthew D Wood (Intel Corp), with some changes by Bodo Moeller*
11038 * Change the client implementation for SSLv23_method() and
11039 SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0
11040 Client Hello message format if the SSL_OP_NO_SSLv2 option is set.
11041 (Previously, the SSL 2.0 backwards compatible Client Hello
11042 message format would be used even with SSL_OP_NO_SSLv2.)
11046 * Add support for smime-type MIME parameter in S/MIME messages which some
11051 * New function BN_MONT_CTX_set_locked() to set montgomery parameters in
11052 a threadsafe manner. Modify rsa code to use new function and add calls
11053 to dsa and dh code (which had race conditions before).
11057 * Include the fixed error library code in the C error file definitions
11058 instead of fixing them up at runtime. This keeps the error code
11059 structures constant.
11063 ### Changes between 0.9.7f and 0.9.7g [11 Apr 2005]
11065 [NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after
11068 * Fixes for newer kerberos headers. NB: the casts are needed because
11069 the 'length' field is signed on one version and unsigned on another
11070 with no (?) obvious way to tell the difference, without these VC++
11071 complains. Also the "definition" of FAR (blank) is no longer included
11072 nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up
11073 some needed definitions.
11077 * Undo Cygwin change.
11081 * Added support for proxy certificates according to RFC 3820.
11082 Because they may be a security thread to unaware applications,
11083 they must be explicitly allowed in run-time. See
11084 docs/HOWTO/proxy_certificates.txt for further information.
11088 ### Changes between 0.9.7e and 0.9.7f [22 Mar 2005]
11090 * Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating
11091 server and client random values. Previously
11092 (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in
11093 less random data when sizeof(time_t) > 4 (some 64 bit platforms).
11095 This change has negligible security impact because:
11097 1. Server and client random values still have 24 bytes of pseudo random
11100 2. Server and client random values are sent in the clear in the initial
11103 3. The master secret is derived using the premaster secret (48 bytes in
11104 size for static RSA ciphersuites) as well as client server and random
11107 The OpenSSL team would like to thank the UK NISCC for bringing this issue
11110 *Stephen Henson, reported by UK NISCC*
11112 * Use Windows randomness collection on Cygwin.
11116 * Fix hang in EGD/PRNGD query when communication socket is closed
11117 prematurely by EGD/PRNGD.
11119 *Darren Tucker <dtucker@zip.com.au> via Lutz Jänicke, resolves #1014*
11121 * Prompt for pass phrases when appropriate for PKCS12 input format.
11125 * Back-port of selected performance improvements from development
11126 branch, as well as improved support for PowerPC platforms.
11130 * Add lots of checks for memory allocation failure, error codes to indicate
11131 failure and freeing up memory if a failure occurs.
11133 *Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson*
11135 * Add new -passin argument to dgst.
11139 * Perform some character comparisons of different types in X509_NAME_cmp:
11140 this is needed for some certificates that re-encode DNs into UTF8Strings
11141 (in violation of RFC3280) and can't or won't issue name rollover
11146 * Make an explicit check during certificate validation to see that
11147 the CA setting in each certificate on the chain is correct. As a
11148 side effect always do the following basic checks on extensions,
11149 not just when there's an associated purpose to the check:
11151 - if there is an unhandled critical extension (unless the user
11152 has chosen to ignore this fault)
11153 - if the path length has been exceeded (if one is set at all)
11154 - that certain extensions fit the associated purpose (if one has
11159 ### Changes between 0.9.7d and 0.9.7e [25 Oct 2004]
11161 * Avoid a race condition when CRLs are checked in a multi threaded
11162 environment. This would happen due to the reordering of the revoked
11163 entries during signature checking and serial number lookup. Now the
11164 encoding is cached and the serial number sort performed under a lock.
11165 Add new STACK function sk_is_sorted().
11169 * Add Delta CRL to the extension code.
11173 * Various fixes to s3_pkt.c so alerts are sent properly.
11175 *David Holmes <d.holmes@f5.com>*
11177 * Reduce the chances of duplicate issuer name and serial numbers (in
11178 violation of RFC3280) using the OpenSSL certificate creation utilities.
11179 This is done by creating a random 64 bit value for the initial serial
11180 number when a serial number file is created or when a self signed
11181 certificate is created using 'openssl req -x509'. The initial serial
11182 number file is created using 'openssl x509 -next_serial' in CA.pl
11183 rather than being initialized to 1.
11187 ### Changes between 0.9.7c and 0.9.7d [17 Mar 2004]
11189 * Fix null-pointer assignment in do_change_cipher_spec() revealed
11190 by using the Codenomicon TLS Test Tool ([CVE-2004-0079])
11192 *Joe Orton, Steve Henson*
11194 * Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites
11197 *Joe Orton, Steve Henson*
11199 * Make it possible to have multiple active certificates with the same
11200 subject in the CA index file. This is done only if the keyword
11201 'unique_subject' is set to 'no' in the main CA section (default
11202 if 'CA_default') of the configuration file. The value is saved
11203 with the database itself in a separate index attribute file,
11204 named like the index file with '.attr' appended to the name.
11208 * X509 verify fixes. Disable broken certificate workarounds when
11209 X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if
11210 keyUsage extension present. Don't accept CRLs with unhandled critical
11211 extensions: since verify currently doesn't process CRL extensions this
11212 rejects a CRL with *any* critical extensions. Add new verify error codes
11217 * When creating an OCSP nonce use an OCTET STRING inside the extnValue.
11218 A clarification of RFC2560 will require the use of OCTET STRINGs and
11219 some implementations cannot handle the current raw format. Since OpenSSL
11220 copies and compares OCSP nonces as opaque blobs without any attempt at
11221 parsing them this should not create any compatibility issues.
11225 * New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when
11226 calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without
11227 this HMAC (and other) operations are several times slower than OpenSSL
11232 * Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex().
11234 *Peter Sylvester <Peter.Sylvester@EdelWeb.fr>*
11236 * Use the correct content when signing type "other".
11240 ### Changes between 0.9.7b and 0.9.7c [30 Sep 2003]
11242 * Fix various bugs revealed by running the NISCC test suite:
11244 Stop out of bounds reads in the ASN1 code when presented with
11245 invalid tags (CVE-2003-0543 and CVE-2003-0544).
11247 Free up ASN1_TYPE correctly if ANY type is invalid ([CVE-2003-0545]).
11249 If verify callback ignores invalid public key errors don't try to check
11250 certificate signature with the NULL public key.
11254 * New -ignore_err option in ocsp application to stop the server
11255 exiting on the first error in a request.
11259 * In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
11260 if the server requested one: as stated in TLS 1.0 and SSL 3.0
11265 * In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
11266 extra data after the compression methods not only for TLS 1.0
11267 but also for SSL 3.0 (as required by the specification).
11269 *Bodo Moeller; problem pointed out by Matthias Loepfe*
11271 * Change X509_certificate_type() to mark the key as exported/exportable
11272 when it's 512 *bits* long, not 512 bytes.
11276 * Change AES_cbc_encrypt() so it outputs exact multiple of
11277 blocks during encryption.
11281 * Various fixes to base64 BIO and non blocking I/O. On write
11282 flushes were not handled properly if the BIO retried. On read
11283 data was not being buffered properly and had various logic bugs.
11284 This also affects blocking I/O when the data being decoded is a
11289 * Various S/MIME bugfixes and compatibility changes:
11290 output correct application/pkcs7 MIME type if
11291 PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures.
11292 Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening
11293 of files as .eml work). Correctly handle very long lines in MIME
11298 ### Changes between 0.9.7a and 0.9.7b [10 Apr 2003]
11300 * Countermeasure against the Klima-Pokorny-Rosa extension of
11301 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
11302 a protocol version number mismatch like a decryption error
11303 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
11307 * Turn on RSA blinding by default in the default implementation
11308 to avoid a timing attack. Applications that don't want it can call
11309 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
11310 They would be ill-advised to do so in most cases.
11312 *Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller*
11314 * Change RSA blinding code so that it works when the PRNG is not
11315 seeded (in this case, the secret RSA exponent is abused as
11316 an unpredictable seed -- if it is not unpredictable, there
11317 is no point in blinding anyway). Make RSA blinding thread-safe
11318 by remembering the creator's thread ID in rsa->blinding and
11319 having all other threads use local one-time blinding factors
11320 (this requires more computation than sharing rsa->blinding, but
11321 avoids excessive locking; and if an RSA object is not shared
11322 between threads, blinding will still be very fast).
11326 * Fixed a typo bug that would cause ENGINE_set_default() to set an
11327 ENGINE as defaults for all supported algorithms irrespective of
11328 the 'flags' parameter. 'flags' is now honoured, so applications
11329 should make sure they are passing it correctly.
11333 * Target "mingw" now allows native Windows code to be generated in
11334 the Cygwin environment as well as with the MinGW compiler.
11338 ### Changes between 0.9.7 and 0.9.7a [19 Feb 2003]
11340 * In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
11341 via timing by performing a MAC computation even if incorrect
11342 block cipher padding has been found. This is a countermeasure
11343 against active attacks where the attacker has to distinguish
11344 between bad padding and a MAC verification error. ([CVE-2003-0078])
11346 *Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
11347 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
11348 Martin Vuagnoux (EPFL, Ilion)*
11350 * Make the no-err option work as intended. The intention with no-err
11351 is not to have the whole error stack handling routines removed from
11352 libcrypto, it's only intended to remove all the function name and
11353 reason texts, thereby removing some of the footprint that may not
11354 be interesting if those errors aren't displayed anyway.
11356 NOTE: it's still possible for any application or module to have its
11357 own set of error texts inserted. The routines are there, just not
11358 used by default when no-err is given.
11362 * Add support for FreeBSD on IA64.
11364 *dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454*
11366 * Adjust DES_cbc_cksum() so it returns the same value as the MIT
11367 Kerberos function mit_des_cbc_cksum(). Before this change,
11368 the value returned by DES_cbc_cksum() was like the one from
11369 mit_des_cbc_cksum(), except the bytes were swapped.
11371 *Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte*
11373 * Allow an application to disable the automatic SSL chain building.
11374 Before this a rather primitive chain build was always performed in
11375 ssl3_output_cert_chain(): an application had no way to send the
11376 correct chain if the automatic operation produced an incorrect result.
11378 Now the chain builder is disabled if either:
11380 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert().
11382 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set.
11384 The reasoning behind this is that an application would not want the
11385 auto chain building to take place if extra chain certificates are
11386 present and it might also want a means of sending no additional
11387 certificates (for example the chain has two certificates and the
11392 * Add the possibility to build without the ENGINE framework.
11394 *Steven Reddie <smr@essemer.com.au> via Richard Levitte*
11396 * Under Win32 gmtime() can return NULL: check return value in
11397 OPENSSL_gmtime(). Add error code for case where gmtime() fails.
11401 * DSA routines: under certain error conditions uninitialized BN objects
11402 could be freed. Solution: make sure initialization is performed early
11403 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>,
11404 Nils Larsch <nla@trustcenter.de> via PR#459)
11408 * Another fix for SSLv2 session ID handling: the session ID was incorrectly
11409 checked on reconnect on the client side, therefore session resumption
11410 could still fail with a "ssl session id is different" error. This
11411 behaviour is masked when SSL_OP_ALL is used due to
11412 SSL_OP_MICROSOFT_SESS_ID_BUG being set.
11413 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
11414 followup to PR #377.
11418 * IA-32 assembler support enhancements: unified ELF targets, support
11419 for SCO/Caldera platforms, fix for Cygwin shared build.
11423 * Add support for FreeBSD on sparc64. As a consequence, support for
11424 FreeBSD on non-x86 processors is separate from x86 processors on
11425 the config script, much like the NetBSD support.
11427 *Richard Levitte & Kris Kennaway <kris@obsecurity.org>*
11429 ### Changes between 0.9.6h and 0.9.7 [31 Dec 2002]
11431 [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after
11434 * Fix session ID handling in SSLv2 client code: the SERVER FINISHED
11435 code (06) was taken as the first octet of the session ID and the last
11436 octet was ignored consequently. As a result SSLv2 client side session
11437 caching could not have worked due to the session ID mismatch between
11439 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
11444 * Change the declaration of needed Kerberos libraries to use EX_LIBS
11445 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is
11450 * The hw_ncipher.c engine requires dynamic locks. Unfortunately, it
11451 seems that in spite of existing for more than a year, many application
11452 author have done nothing to provide the necessary callbacks, which
11453 means that this particular engine will not work properly anywhere.
11454 This is a very unfortunate situation which forces us, in the name
11455 of usability, to give the hw_ncipher.c a static lock, which is part
11457 NOTE: This is for the 0.9.7 series ONLY. This hack will never
11458 appear in 0.9.8 or later. We EXPECT application authors to have
11459 dealt properly with this when 0.9.8 is released (unless we actually
11460 make such changes in the libcrypto locking code that changes will
11461 have to be made anyway).
11465 * In asn1_d2i_read_bio() repeatedly call BIO_read() until all content
11466 octets have been read, EOF or an error occurs. Without this change
11467 some truncated ASN1 structures will not produce an error.
11471 * Disable Heimdal support, since it hasn't been fully implemented.
11472 Still give the possibility to force the use of Heimdal, but with
11473 warnings and a request that patches get sent to openssl-dev.
11477 * Add the VC-CE target, introduce the WINCE sysname, and add
11478 INSTALL.WCE and appropriate conditionals to make it build.
11480 *Steven Reddie <smr@essemer.com.au> via Richard Levitte*
11482 * Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and
11483 cygssl-x.y.z.dll, where x, y and z are the major, minor and
11484 edit numbers of the version.
11486 *Corinna Vinschen <vinschen@redhat.com> and Richard Levitte*
11488 * Introduce safe string copy and catenation functions
11489 (BUF_strlcpy() and BUF_strlcat()).
11491 *Ben Laurie (CHATS) and Richard Levitte*
11493 * Avoid using fixed-size buffers for one-line DNs.
11495 *Ben Laurie (CHATS)*
11497 * Add BUF_MEM_grow_clean() to avoid information leakage when
11498 resizing buffers containing secrets, and use where appropriate.
11500 *Ben Laurie (CHATS)*
11502 * Avoid using fixed size buffers for configuration file location.
11504 *Ben Laurie (CHATS)*
11506 * Avoid filename truncation for various CA files.
11508 *Ben Laurie (CHATS)*
11510 * Use sizeof in preference to magic numbers.
11512 *Ben Laurie (CHATS)*
11514 * Avoid filename truncation in cert requests.
11516 *Ben Laurie (CHATS)*
11518 * Add assertions to check for (supposedly impossible) buffer
11521 *Ben Laurie (CHATS)*
11523 * Don't cache truncated DNS entries in the local cache (this could
11524 potentially lead to a spoofing attack).
11526 *Ben Laurie (CHATS)*
11528 * Fix various buffers to be large enough for hex/decimal
11529 representations in a platform independent manner.
11531 *Ben Laurie (CHATS)*
11533 * Add CRYPTO_realloc_clean() to avoid information leakage when
11534 resizing buffers containing secrets, and use where appropriate.
11536 *Ben Laurie (CHATS)*
11538 * Add BIO_indent() to avoid much slightly worrying code to do
11541 *Ben Laurie (CHATS)*
11543 * Convert sprintf()/BIO_puts() to BIO_printf().
11545 *Ben Laurie (CHATS)*
11547 * buffer_gets() could terminate with the buffer only half
11550 *Ben Laurie (CHATS)*
11552 * Add assertions to prevent user-supplied crypto functions from
11553 overflowing internal buffers by having large block sizes, etc.
11555 *Ben Laurie (CHATS)*
11557 * New OPENSSL_assert() macro (similar to assert(), but enabled
11560 *Ben Laurie (CHATS)*
11562 * Eliminate unused copy of key in RC4.
11564 *Ben Laurie (CHATS)*
11566 * Eliminate unused and incorrectly sized buffers for IV in pem.h.
11568 *Ben Laurie (CHATS)*
11570 * Fix off-by-one error in EGD path.
11572 *Ben Laurie (CHATS)*
11574 * If RANDFILE path is too long, ignore instead of truncating.
11576 *Ben Laurie (CHATS)*
11578 * Eliminate unused and incorrectly sized X.509 structure
11581 *Ben Laurie (CHATS)*
11583 * Eliminate unused and dangerous function knumber().
11585 *Ben Laurie (CHATS)*
11587 * Eliminate unused and dangerous structure, KSSL_ERR.
11589 *Ben Laurie (CHATS)*
11591 * Protect against overlong session ID context length in an encoded
11592 session object. Since these are local, this does not appear to be
11595 *Ben Laurie (CHATS)*
11597 * Change from security patch (see 0.9.6e below) that did not affect
11598 the 0.9.6 release series:
11600 Remote buffer overflow in SSL3 protocol - an attacker could
11601 supply an oversized master key in Kerberos-enabled versions.
11604 *Ben Laurie (CHATS)*
11606 * Change the SSL kerb5 codes to match RFC 2712.
11610 * Make -nameopt work fully for req and add -reqopt switch.
11612 *Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson*
11614 * The "block size" for block ciphers in CFB and OFB mode should be 1.
11616 *Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>*
11618 * Make sure tests can be performed even if the corresponding algorithms
11619 have been removed entirely. This was also the last step to make
11620 OpenSSL compilable with DJGPP under all reasonable conditions.
11622 *Richard Levitte, Doug Kaufman <dkaufman@rahul.net>*
11624 * Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT
11625 to allow version independent disabling of normally unselected ciphers,
11626 which may be activated as a side-effect of selecting a single cipher.
11628 (E.g., cipher list string "RSA" enables ciphersuites that are left
11629 out of "ALL" because they do not provide symmetric encryption.
11630 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.)
11632 *Lutz Jaenicke, Bodo Moeller*
11634 * Add appropriate support for separate platform-dependent build
11635 directories. The recommended way to make a platform-dependent
11636 build directory is the following (tested on Linux), maybe with
11639 # Place yourself outside of the OpenSSL source tree. In
11640 # this example, the environment variable OPENSSL_SOURCE
11641 # is assumed to contain the absolute OpenSSL source directory.
11642 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`"
11643 cd objtree/"`uname -s`-`uname -r`-`uname -m`"
11644 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do
11645 mkdir -p `dirname $F`
11646 ln -s $OPENSSL_SOURCE/$F $F
11649 To be absolutely sure not to disturb the source tree, a "make clean"
11650 is a good thing. If it isn't successful, don't worry about it,
11651 it probably means the source directory is very clean.
11655 * Make sure any ENGINE control commands make local copies of string
11656 pointers passed to them whenever necessary. Otherwise it is possible
11657 the caller may have overwritten (or deallocated) the original string
11658 data when a later ENGINE operation tries to use the stored values.
11660 *Götz Babin-Ebell <babinebell@trustcenter.de>*
11662 * Improve diagnostics in file reading and command-line digests.
11664 *Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>*
11666 * Add AES modes CFB and OFB to the object database. Correct an
11667 error in AES-CFB decryption.
11671 * Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this
11672 allows existing EVP_CIPHER_CTX structures to be reused after
11673 calling `EVP_*Final()`. This behaviour is used by encryption
11674 BIOs and some applications. This has the side effect that
11675 applications must explicitly clean up cipher contexts with
11676 EVP_CIPHER_CTX_cleanup() or they will leak memory.
11680 * Check the values of dna and dnb in bn_mul_recursive before calling
11681 bn_mul_comba (a non zero value means the a or b arrays do not contain
11682 n2 elements) and fallback to bn_mul_normal if either is not zero.
11686 * Fix escaping of non-ASCII characters when using the -subj option
11687 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>)
11691 * Make object definitions compliant to LDAP (RFC2256): SN is the short
11692 form for "surname", serialNumber has no short form.
11693 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798;
11694 therefore remove "mail" short name for "internet 7".
11695 The OID for unique identifiers in X509 certificates is
11696 x500UniqueIdentifier, not uniqueIdentifier.
11697 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>)
11701 * Add an "init" command to the ENGINE config module and auto initialize
11702 ENGINEs. Without any "init" command the ENGINE will be initialized
11703 after all ctrl commands have been executed on it. If init=1 the
11704 ENGINE is initialized at that point (ctrls before that point are run
11705 on the uninitialized ENGINE and after on the initialized one). If
11706 init=0 then the ENGINE will not be initialized at all.
11710 * Fix the 'app_verify_callback' interface so that the user-defined
11711 argument is actually passed to the callback: In the
11712 SSL_CTX_set_cert_verify_callback() prototype, the callback
11713 declaration has been changed from
11716 int (*cb)(X509_STORE_CTX *,void *);
11717 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
11718 i=s->ctx->app_verify_callback(&ctx)
11719 has been changed into
11720 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
11722 To update applications using SSL_CTX_set_cert_verify_callback(),
11723 a dummy argument can be added to their callback functions.
11725 *D. K. Smetters <smetters@parc.xerox.com>*
11727 * Added the '4758cca' ENGINE to support IBM 4758 cards.
11729 *Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe*
11731 * Add and OPENSSL_LOAD_CONF define which will cause
11732 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
11733 This allows older applications to transparently support certain
11734 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
11735 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
11736 load the config file and OPENSSL_add_all_algorithms_conf() which will
11737 always load it have also been added.
11741 * Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
11742 Adjust NIDs and EVP layer.
11744 *Stephen Sprunk <stephen@sprunk.org> and Richard Levitte*
11746 * Config modules support in openssl utility.
11748 Most commands now load modules from the config file,
11749 though in a few (such as version) this isn't done
11750 because it couldn't be used for anything.
11752 In the case of ca and req the config file used is
11753 the same as the utility itself: that is the -config
11754 command line option can be used to specify an
11759 * Move default behaviour from OPENSSL_config(). If appname is NULL
11760 use "openssl_conf" if filename is NULL use default openssl config file.
11764 * Add an argument to OPENSSL_config() to allow the use of an alternative
11765 config section name. Add a new flag to tolerate a missing config file
11766 and move code to CONF_modules_load_file().
11770 * Support for crypto accelerator cards from Accelerated Encryption
11771 Processing, www.aep.ie. (Use engine 'aep')
11772 The support was copied from 0.9.6c [engine] and adapted/corrected
11773 to work with the new engine framework.
11775 *AEP Inc. and Richard Levitte*
11777 * Support for SureWare crypto accelerator cards from Baltimore
11778 Technologies. (Use engine 'sureware')
11779 The support was copied from 0.9.6c [engine] and adapted
11780 to work with the new engine framework.
11784 * Have the CHIL engine fork-safe (as defined by nCipher) and actually
11785 make the newer ENGINE framework commands for the CHIL engine work.
11787 *Toomas Kiisk <vix@cyber.ee> and Richard Levitte*
11789 * Make it possible to produce shared libraries on ReliantUNIX.
11791 *Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte*
11793 * Add the configuration target debug-linux-ppro.
11794 Make 'openssl rsa' use the general key loading routines
11795 implemented in `apps.c`, and make those routines able to
11796 handle the key format FORMAT_NETSCAPE and the variant
11799 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
11801 * Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
11803 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
11805 * Add -keyform to rsautl, and document -engine.
11807 *Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>*
11809 * Change BIO_new_file (crypto/bio/bss_file.c) to use new
11810 BIO_R_NO_SUCH_FILE error code rather than the generic
11811 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
11815 * Add new functions
11816 ERR_peek_last_error
11817 ERR_peek_last_error_line
11818 ERR_peek_last_error_line_data.
11819 These are similar to
11821 ERR_peek_error_line
11822 ERR_peek_error_line_data,
11823 but report on the latest error recorded rather than the first one
11824 still in the error queue.
11826 *Ben Laurie, Bodo Moeller*
11828 * default_algorithms option in ENGINE config module. This allows things
11830 default_algorithms = ALL
11831 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
11835 * Preliminary ENGINE config module.
11839 * New experimental application configuration code.
11843 * Change the AES code to follow the same name structure as all other
11844 symmetric ciphers, and behave the same way. Move everything to
11845 the directory crypto/aes, thereby obsoleting crypto/rijndael.
11847 *Stephen Sprunk <stephen@sprunk.org> and Richard Levitte*
11849 * SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
11851 *Ben Laurie and Theo de Raadt*
11853 * Add option to output public keys in req command.
11855 *Massimiliano Pala madwolf@openca.org*
11857 * Use wNAFs in EC_POINTs_mul() for improved efficiency
11858 (up to about 10% better than before for P-192 and P-224).
11862 * New functions/macros
11864 SSL_CTX_set_msg_callback(ctx, cb)
11865 SSL_CTX_set_msg_callback_arg(ctx, arg)
11866 SSL_set_msg_callback(ssl, cb)
11867 SSL_set_msg_callback_arg(ssl, arg)
11869 to request calling a callback function
11871 void cb(int write_p, int version, int content_type,
11872 const void *buf, size_t len, SSL *ssl, void *arg)
11874 whenever a protocol message has been completely received
11875 (write_p == 0) or sent (write_p == 1). Here 'version' is the
11876 protocol version according to which the SSL library interprets
11877 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
11878 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
11879 the content type as defined in the SSL 3.0/TLS 1.0 protocol
11880 specification (change_cipher_spec(20), alert(21), handshake(22)).
11881 'buf' and 'len' point to the actual message, 'ssl' to the
11882 SSL object, and 'arg' is the application-defined value set by
11883 SSL[_CTX]_set_msg_callback_arg().
11885 'openssl s_client' and 'openssl s_server' have new '-msg' options
11886 to enable a callback that displays all protocol messages.
11890 * Change the shared library support so shared libraries are built as
11891 soon as the corresponding static library is finished, and thereby get
11892 openssl and the test programs linked against the shared library.
11893 This still only happens when the keyword "shard" has been given to
11894 the configuration scripts.
11896 NOTE: shared library support is still an experimental thing, and
11897 backward binary compatibility is still not guaranteed.
11899 *"Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte*
11901 * Add support for Subject Information Access extension.
11903 *Peter Sylvester <Peter.Sylvester@EdelWeb.fr>*
11905 * Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
11906 additional bytes when new memory had to be allocated, not just
11907 when reusing an existing buffer.
11911 * New command line and configuration option 'utf8' for the req command.
11912 This allows field values to be specified as UTF8 strings.
11916 * Add -multi and -mr options to "openssl speed" - giving multiple parallel
11917 runs for the former and machine-readable output for the latter.
11921 * Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
11922 of the e-mail address in the DN (i.e., it will go into a certificate
11923 extension only). The new configuration file option 'email_in_dn = no'
11924 has the same effect.
11926 *Massimiliano Pala madwolf@openca.org*
11928 * Change all functions with names starting with `des_` to be starting
11929 with `DES_` instead. Add wrappers that are compatible with libdes,
11930 but are named `_ossl_old_des_*`. Finally, add macros that map the
11931 `des_*` symbols to the corresponding `_ossl_old_des_*` if libdes
11932 compatibility is desired. If OpenSSL 0.9.6c compatibility is
11933 desired, the `des_*` symbols will be mapped to `DES_*`, with one
11936 Since we provide two compatibility mappings, the user needs to
11937 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
11938 compatibility is desired. The default (i.e., when that macro
11939 isn't defined) is OpenSSL 0.9.6c compatibility.
11941 There are also macros that enable and disable the support of old
11942 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
11943 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
11944 are defined, the default will apply: to support the old des routines.
11946 In either case, one must include openssl/des.h to get the correct
11947 definitions. Do not try to just include openssl/des_old.h, that
11950 NOTE: This is a major break of an old API into a new one. Software
11951 authors are encouraged to switch to the `DES_` style functions. Some
11952 time in the future, des_old.h and the libdes compatibility functions
11953 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
11954 default), and then completely removed.
11958 * Test for certificates which contain unsupported critical extensions.
11959 If such a certificate is found during a verify operation it is
11960 rejected by default: this behaviour can be overridden by either
11961 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
11962 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
11963 X509_supported_extension() has also been added which returns 1 if a
11964 particular extension is supported.
11968 * Modify the behaviour of EVP cipher functions in similar way to digests
11969 to retain compatibility with existing code.
11973 * Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
11974 compatibility with existing code. In particular the 'ctx' parameter does
11975 not have to be to be initialized before the call to EVP_DigestInit() and
11976 it is tidied up after a call to EVP_DigestFinal(). New function
11977 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
11978 EVP_MD_CTX_copy() changed to not require the destination to be
11979 initialized valid and new function EVP_MD_CTX_copy_ex() added which
11980 requires the destination to be valid.
11982 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
11983 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
11987 * Change ssl3_get_message (ssl/s3_both.c) and the functions using it
11988 so that complete 'Handshake' protocol structures are kept in memory
11989 instead of overwriting 'msg_type' and 'length' with 'body' data.
11993 * Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
11995 *Massimo Santin via Richard Levitte*
11997 * Major restructuring to the underlying ENGINE code. This includes
11998 reduction of linker bloat, separation of pure "ENGINE" manipulation
11999 (initialisation, etc) from functionality dealing with implementations
12000 of specific crypto interfaces. This change also introduces integrated
12001 support for symmetric ciphers and digest implementations - so ENGINEs
12002 can now accelerate these by providing EVP_CIPHER and EVP_MD
12003 implementations of their own. This is detailed in
12004 [crypto/engine/README.md](crypto/engine/README.md)
12005 as it couldn't be adequately described here. However, there are a few
12006 API changes worth noting - some RSA, DSA, DH, and RAND functions that
12007 were changed in the original introduction of ENGINE code have now
12008 reverted back - the hooking from this code to ENGINE is now a good
12009 deal more passive and at run-time, operations deal directly with
12010 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
12011 dereferencing through an ENGINE pointer any more. Also, the ENGINE
12012 functions dealing with `BN_MOD_EXP[_CRT]` handlers have been removed -
12013 they were not being used by the framework as there is no concept of a
12014 BIGNUM_METHOD and they could not be generalised to the new
12015 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
12016 ENGINE_cpy() has been removed as it cannot be consistently defined in
12021 * Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
12025 * Change mkdef.pl to sort symbols that get the same entry number,
12026 and make sure the automatically generated functions `ERR_load_*`
12027 become part of libeay.num as well.
12031 * New function SSL_renegotiate_pending(). This returns true once
12032 renegotiation has been requested (either SSL_renegotiate() call
12033 or HelloRequest/ClientHello received from the peer) and becomes
12034 false once a handshake has been completed.
12035 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
12036 sends a HelloRequest, but does not ensure that a handshake takes
12037 place. SSL_renegotiate_pending() is useful for checking if the
12038 client has followed the request.)
12042 * New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
12043 By default, clients may request session resumption even during
12044 renegotiation (if session ID contexts permit); with this option,
12045 session resumption is possible only in the first handshake.
12047 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes
12048 more bits available for options that should not be part of
12049 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION).
12053 * Add some demos for certificate and certificate request creation.
12057 * Make maximum certificate chain size accepted from the peer application
12058 settable (`SSL*_get/set_max_cert_list()`), as proposed by
12059 "Douglas E. Engert" <deengert@anl.gov>.
12063 * Add support for shared libraries for Unixware-7
12064 (Boyd Lynn Gerber <gerberb@zenez.com>).
12068 * Add a "destroy" handler to ENGINEs that allows structural cleanup to
12069 be done prior to destruction. Use this to unload error strings from
12070 ENGINEs that load their own error strings. NB: This adds two new API
12071 functions to "get" and "set" this destroy handler in an ENGINE.
12075 * Alter all existing ENGINE implementations (except "openssl" and
12076 "openbsd") to dynamically instantiate their own error strings. This
12077 makes them more flexible to be built both as statically-linked ENGINEs
12078 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
12079 Also, add stub code to each that makes building them as self-contained
12080 shared-libraries easier (see [README-Engine.md](README-Engine.md)).
12084 * Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
12085 implementations into applications that are completely implemented in
12086 self-contained shared-libraries. The "dynamic" ENGINE exposes control
12087 commands that can be used to configure what shared-library to load and
12088 to control aspects of the way it is handled. Also, made an update to
12089 the [README-Engine.md](README-Engine.md) file
12090 that brings its information up-to-date and
12091 provides some information and instructions on the "dynamic" ENGINE
12092 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
12096 * Make it possible to unload ranges of ERR strings with a new
12097 "ERR_unload_strings" function.
12101 * Add a copy() function to EVP_MD.
12105 * Make EVP_MD routines take a context pointer instead of just the
12106 md_data void pointer.
12110 * Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
12111 that the digest can only process a single chunk of data
12112 (typically because it is provided by a piece of
12113 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
12114 is only going to provide a single chunk of data, and hence the
12115 framework needn't accumulate the data for oneshot drivers.
12119 * As with "ERR", make it possible to replace the underlying "ex_data"
12120 functions. This change also alters the storage and management of global
12121 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
12122 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
12123 index counters. The API functions that use this state have been changed
12124 to take a "class_index" rather than pointers to the class's local STACK
12125 and counter, and there is now an API function to dynamically create new
12126 classes. This centralisation allows us to (a) plug a lot of the
12127 thread-safety problems that existed, and (b) makes it possible to clean
12128 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
12129 such data would previously have always leaked in application code and
12130 workarounds were in place to make the memory debugging turn a blind eye
12131 to it. Application code that doesn't use this new function will still
12132 leak as before, but their memory debugging output will announce it now
12133 rather than letting it slide.
12135 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
12136 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
12137 has a return value to indicate success or failure.
12141 * Make it possible to replace the underlying "ERR" functions such that the
12142 global state (2 LHASH tables and 2 locks) is only used by the "default"
12143 implementation. This change also adds two functions to "get" and "set"
12144 the implementation prior to it being automatically set the first time
12145 any other ERR function takes place. Ie. an application can call "get",
12146 pass the return value to a module it has just loaded, and that module
12147 can call its own "set" function using that value. This means the
12148 module's "ERR" operations will use (and modify) the error state in the
12149 application and not in its own statically linked copy of OpenSSL code.
12153 * Give DH, DSA, and RSA types their own `*_up_ref()` function to increment
12154 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
12155 the operation, and provides a more encapsulated way for external code
12156 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
12157 to use these functions rather than manually incrementing the counts.
12159 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
12163 * Add EVP test program.
12167 * Add symmetric cipher support to ENGINE. Expect the API to change!
12171 * New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
12172 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
12173 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
12174 These allow a CRL to be built without having to access X509_CRL fields
12175 directly. Modify 'ca' application to use new functions.
12179 * Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
12180 bug workarounds. Rollback attack detection is a security feature.
12181 The problem will only arise on OpenSSL servers when TLSv1 is not
12182 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
12183 Software authors not wanting to support TLSv1 will have special reasons
12184 for their choice and can explicitly enable this option.
12186 *Bodo Moeller, Lutz Jaenicke*
12188 * Rationalise EVP so it can be extended: don't include a union of
12189 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX
12190 (similar to those existing for EVP_CIPHER_CTX).
12195 EVP_MD_CTX_init(&md); /* new function call */
12196 EVP_DigestInit(&md, EVP_sha1());
12197 EVP_DigestUpdate(&md, in, len);
12198 EVP_DigestFinal(&md, out, NULL);
12199 EVP_MD_CTX_cleanup(&md); /* new function call */
12203 * Make DES key schedule conform to the usual scheme, as well as
12204 correcting its structure. This means that calls to DES functions
12205 now have to pass a pointer to a des_key_schedule instead of a
12206 plain des_key_schedule (which was actually always a pointer
12209 des_key_schedule ks;
12211 des_set_key_checked(..., &ks);
12212 des_ncbc_encrypt(..., &ks, ...);
12214 (Note that a later change renames 'des_...' into 'DES_...'.)
12218 * Initial reduction of linker bloat: the use of some functions, such as
12219 PEM causes large amounts of unused functions to be linked in due to
12220 poor organisation. For example pem_all.c contains every PEM function
12221 which has a knock on effect of linking in large amounts of (unused)
12222 ASN1 code. Grouping together similar functions and splitting unrelated
12223 functions prevents this.
12227 * Cleanup of EVP macros.
12231 * Change historical references to `{NID,SN,LN}_des_ede` and ede3 to add the
12232 correct `_ecb suffix`.
12236 * Add initial OCSP responder support to ocsp application. The
12237 revocation information is handled using the text based index
12238 use by the ca application. The responder can either handle
12239 requests generated internally, supplied in files (for example
12240 via a CGI script) or using an internal minimal server.
12244 * Add configuration choices to get zlib compression for TLS.
12248 * Changes to Kerberos SSL for RFC 2712 compliance:
12249 1. Implemented real KerberosWrapper, instead of just using
12250 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
12251 2. Implemented optional authenticator field of KerberosWrapper.
12253 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
12254 and authenticator structs; see crypto/krb5/.
12256 Generalized Kerberos calls to support multiple Kerberos libraries.
12257 *Vern Staats <staatsvr@asc.hpc.mil>, Jeffrey Altman <jaltman@columbia.edu>
12258 via Richard Levitte*
12260 * Cause 'openssl speed' to use fully hard-coded DSA keys as it
12261 already does with RSA. testdsa.h now has 'priv_key/pub_key'
12262 values for each of the key sizes rather than having just
12263 parameters (and 'speed' generating keys each time).
12267 * Speed up EVP routines.
12270 pe 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
12271 s-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
12272 s-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
12273 s-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
12275 s-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
12276 s-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
12277 s-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
12280 s-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
12282 s-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
12286 * Added the OS2-EMX target.
12288 *"Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte*
12290 * Rewrite commands to use `NCONF` routines instead of the old `CONF`.
12291 New functions to support `NCONF` routines in extension code.
12292 New function `CONF_set_nconf()`
12293 to allow functions which take an `NCONF` to also handle the old `LHASH`
12294 structure: this means that the old `CONF` compatible routines can be
12295 retained (in particular w.rt. extensions) without having to duplicate the
12296 code. New function `X509V3_add_ext_nconf_sk()` to add extensions to a stack.
12300 * Enhance the general user interface with mechanisms for inner control
12301 and with possibilities to have yes/no kind of prompts.
12305 * Change all calls to low-level digest routines in the library and
12306 applications to use EVP. Add missing calls to HMAC_cleanup() and
12307 don't assume HMAC_CTX can be copied using memcpy().
12309 *Verdon Walker <VWalker@novell.com>, Steve Henson*
12311 * Add the possibility to control engines through control names but with
12312 arbitrary arguments instead of just a string.
12313 Change the key loaders to take a UI_METHOD instead of a callback
12314 function pointer. NOTE: this breaks binary compatibility with earlier
12315 versions of OpenSSL [engine].
12316 Adapt the nCipher code for these new conditions and add a card insertion
12321 * Enhance the general user interface with mechanisms to better support
12322 dialog box interfaces, application-defined prompts, the possibility
12323 to use defaults (for example default passwords from somewhere else)
12324 and interrupts/cancellations.
12328 * Tidy up PKCS#12 attribute handling. Add support for the CSP name
12329 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
12333 * Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
12334 tidy up some unnecessarily weird code in 'sk_new()').
12336 *Geoff, reported by Diego Tartara <dtartara@novamens.com>*
12338 * Change the key loading routines for ENGINEs to use the same kind
12339 callback (pem_password_cb) as all other routines that need this
12344 * Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
12345 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
12346 than this minimum value is recommended.
12350 * New random seeder for OpenVMS, using the system process statistics
12351 that are easily reachable.
12355 * Windows apparently can't transparently handle global
12356 variables defined in DLLs. Initialisations such as:
12358 const ASN1_ITEM *it = &ASN1_INTEGER_it;
12360 won't compile. This is used by the any applications that need to
12361 declare their own ASN1 modules. This was fixed by adding the option
12362 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
12363 needed for static libraries under Win32.
12367 * New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
12368 setting of purpose and trust fields. New X509_STORE trust and
12369 purpose functions and tidy up setting in other SSL functions.
12373 * Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
12374 structure. These are inherited by X509_STORE_CTX when it is
12375 initialised. This allows various defaults to be set in the
12376 X509_STORE structure (such as flags for CRL checking and custom
12377 purpose or trust settings) for functions which only use X509_STORE_CTX
12378 internally such as S/MIME.
12380 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
12381 trust settings if they are not set in X509_STORE. This allows X509_STORE
12382 purposes and trust (in S/MIME for example) to override any set by default.
12384 Add command line options for CRL checking to smime, s_client and s_server
12389 * Initial CRL based revocation checking. If the CRL checking flag(s)
12390 are set then the CRL is looked up in the X509_STORE structure and
12391 its validity and signature checked, then if the certificate is found
12392 in the CRL the verify fails with a revoked error.
12394 Various new CRL related callbacks added to X509_STORE_CTX structure.
12396 Command line options added to 'verify' application to support this.
12398 This needs some additional work, such as being able to handle multiple
12399 CRLs with different times, extension based lookup (rather than just
12400 by subject name) and ultimately more complete V2 CRL extension
12405 * Add a general user interface API (crypto/ui/). This is designed
12406 to replace things like des_read_password and friends (backward
12407 compatibility functions using this new API are provided).
12408 The purpose is to remove prompting functions from the DES code
12409 section as well as provide for prompting through dialog boxes in
12410 a window system and the like.
12414 * Add "ex_data" support to ENGINE so implementations can add state at a
12415 per-structure level rather than having to store it globally.
12419 * Make it possible for ENGINE structures to be copied when retrieved by
12420 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
12421 This causes the "original" ENGINE structure to act like a template,
12422 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
12423 operational state can be localised to each ENGINE structure, despite the
12424 fact they all share the same "methods". New ENGINE structures returned in
12425 this case have no functional references and the return value is the single
12426 structural reference. This matches the single structural reference returned
12427 by ENGINE_by_id() normally, when it is incremented on the pre-existing
12432 * Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
12433 needs to match any other type at all we need to manually clear the
12438 * Changes to the "openssl engine" utility to include;
12439 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
12440 about an ENGINE's available control commands.
12441 - executing control commands from command line arguments using the
12442 '-pre' and '-post' switches. '-post' is only used if '-t' is
12443 specified and the ENGINE is successfully initialised. The syntax for
12444 the individual commands are colon-separated, for example;
12445 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
12449 * New dynamic control command support for ENGINEs. ENGINEs can now
12450 declare their own commands (numbers), names (strings), descriptions,
12451 and input types for run-time discovery by calling applications. A
12452 subset of these commands are implicitly classed as "executable"
12453 depending on their input type, and only these can be invoked through
12454 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
12455 can be based on user input, config files, etc). The distinction is
12456 that "executable" commands cannot return anything other than a boolean
12457 result and can only support numeric or string input, whereas some
12458 discoverable commands may only be for direct use through
12459 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
12460 pointers, or other custom uses. The "executable" commands are to
12461 support parameterisations of ENGINE behaviour that can be
12462 unambiguously defined by ENGINEs and used consistently across any
12463 OpenSSL-based application. Commands have been added to all the
12464 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
12465 control over shared-library paths without source code alterations.
12469 * Changed all ENGINE implementations to dynamically allocate their
12470 ENGINEs rather than declaring them statically. Apart from this being
12471 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
12472 this also allows the implementations to compile without using the
12473 internal engine_int.h header.
12477 * Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
12478 'const' value. Any code that should be able to modify a RAND_METHOD
12479 should already have non-const pointers to it (ie. they should only
12480 modify their own ones).
12484 * Made a variety of little tweaks to the ENGINE code.
12485 - "atalla" and "ubsec" string definitions were moved from header files
12486 to C code. "nuron" string definitions were placed in variables
12487 rather than hard-coded - allowing parameterisation of these values
12488 later on via ctrl() commands.
12489 - Removed unused "#if 0"'d code.
12490 - Fixed engine list iteration code so it uses ENGINE_free() to release
12491 structural references.
12492 - Constified the RAND_METHOD element of ENGINE structures.
12493 - Constified various get/set functions as appropriate and added
12494 missing functions (including a catch-all ENGINE_cpy that duplicates
12495 all ENGINE values onto a new ENGINE except reference counts/state).
12496 - Removed NULL parameter checks in get/set functions. Setting a method
12497 or function to NULL is a way of cancelling out a previously set
12498 value. Passing a NULL ENGINE parameter is just plain stupid anyway
12499 and doesn't justify the extra error symbols and code.
12500 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
12501 flags from engine_int.h to engine.h.
12502 - Changed prototypes for ENGINE handler functions (init(), finish(),
12503 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
12507 * Implement binary inversion algorithm for BN_mod_inverse in addition
12508 to the algorithm using long division. The binary algorithm can be
12509 used only if the modulus is odd. On 32-bit systems, it is faster
12510 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
12511 roughly 5-15% for 256-bit moduli), so we use it only for moduli
12512 up to 450 bits. In 64-bit environments, the binary algorithm
12513 appears to be advantageous for much longer moduli; here we use it
12514 for moduli up to 2048 bits.
12518 * Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
12519 could not support the combine flag in choice fields.
12523 * Add a 'copy_extensions' option to the 'ca' utility. This copies
12524 extensions from a certificate request to the certificate.
12528 * Allow multiple 'certopt' and 'nameopt' options to be separated
12529 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
12530 file: this allows the display of the certificate about to be
12531 signed to be customised, to allow certain fields to be included
12532 or excluded and extension details. The old system didn't display
12533 multicharacter strings properly, omitted fields not in the policy
12534 and couldn't display additional details such as extensions.
12538 * Function EC_POINTs_mul for multiple scalar multiplication
12539 of an arbitrary number of elliptic curve points
12540 \sum scalars[i]*points[i],
12541 optionally including the generator defined for the EC_GROUP:
12542 scalar*generator + \sum scalars[i]*points[i].
12544 EC_POINT_mul is a simple wrapper function for the typical case
12545 that the point list has just one item (besides the optional
12550 * First EC_METHODs for curves over GF(p):
12552 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
12553 operations and provides various method functions that can also
12554 operate with faster implementations of modular arithmetic.
12556 EC_GFp_mont_method() reuses most functions that are part of
12557 EC_GFp_simple_method, but uses Montgomery arithmetic.
12559 *Bodo Moeller; point addition and point doubling
12560 implementation directly derived from source code provided by
12561 Lenka Fibikova <fibikova@exp-math.uni-essen.de>*
12563 * Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
12564 crypto/ec/ec_lib.c):
12566 Curves are EC_GROUP objects (with an optional group generator)
12567 based on EC_METHODs that are built into the library.
12569 Points are EC_POINT objects based on EC_GROUP objects.
12571 Most of the framework would be able to handle curves over arbitrary
12572 finite fields, but as there are no obvious types for fields other
12573 than GF(p), some functions are limited to that for now.
12577 * Add the -HTTP option to s_server. It is similar to -WWW, but requires
12578 that the file contains a complete HTTP response.
12582 * Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
12583 change the def and num file printf format specifier from "%-40sXXX"
12584 to "%-39s XXX". The latter will always guarantee a space after the
12585 field while the former will cause them to run together if the field
12586 is 40 of more characters long.
12590 * Constify the cipher and digest 'method' functions and structures
12591 and modify related functions to take constant EVP_MD and EVP_CIPHER
12596 * Hide BN_CTX structure details in bn_lcl.h instead of publishing them
12597 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
12601 * Modify `EVP_Digest*()` routines so they now return values. Although the
12602 internal software routines can never fail additional hardware versions
12607 * Clean up crypto/err/err.h and change some error codes to avoid conflicts:
12609 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
12610 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
12613 ERR_R_NESTED_ASN1_ERROR
12615 ERR_R_MISSING_ASN1_EOS
12616 were 4 .. 9, conflicting with
12617 ERR_LIB_RSA (= ERR_R_RSA_LIB)
12619 ERR_LIB_PEM (= ERR_R_PEM_LIB).
12620 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
12622 Add new error code 'ERR_R_INTERNAL_ERROR'.
12626 * Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
12631 * New option '-subj arg' for 'openssl req' and 'openssl ca'. This
12632 sets the subject name for a new request or supersedes the
12633 subject name in a given request. Formats that can be parsed are
12634 'CN=Some Name, OU=myOU, C=IT'
12636 'CN=Some Name/OU=myOU/C=IT'.
12638 Add options '-batch' and '-verbose' to 'openssl req'.
12640 *Massimiliano Pala <madwolf@hackmasters.net>*
12642 * Introduce the possibility to access global variables through
12643 functions on platform were that's the best way to handle exporting
12644 global variables in shared libraries. To enable this functionality,
12645 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
12646 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
12647 is normally done by Configure or something similar).
12649 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
12650 in the source file (foo.c) like this:
12652 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
12653 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
12655 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
12656 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
12658 OPENSSL_DECLARE_GLOBAL(int,foo);
12659 #define foo OPENSSL_GLOBAL_REF(foo)
12660 OPENSSL_DECLARE_GLOBAL(double,bar);
12661 #define bar OPENSSL_GLOBAL_REF(bar)
12663 The #defines are very important, and therefore so is including the
12664 header file everywhere where the defined globals are used.
12666 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
12667 of ASN.1 items, but that structure is a bit different.
12669 The largest change is in util/mkdef.pl which has been enhanced with
12670 better and easier to understand logic to choose which symbols should
12671 go into the Windows .def files as well as a number of fixes and code
12672 cleanup (among others, algorithm keywords are now sorted
12673 lexicographically to avoid constant rewrites).
12677 * In BN_div() keep a copy of the sign of 'num' before writing the
12678 result to 'rm' because if rm==num the value will be overwritten
12679 and produce the wrong result if 'num' is negative: this caused
12680 problems with BN_mod() and BN_nnmod().
12684 * Function OCSP_request_verify(). This checks the signature on an
12685 OCSP request and verifies the signer certificate. The signer
12686 certificate is just checked for a generic purpose and OCSP request
12691 * Add OCSP_check_validity() function to check the validity of OCSP
12692 responses. OCSP responses are prepared in real time and may only
12693 be a few seconds old. Simply checking that the current time lies
12694 between thisUpdate and nextUpdate max reject otherwise valid responses
12695 caused by either OCSP responder or client clock inaccuracy. Instead
12696 we allow thisUpdate and nextUpdate to fall within a certain period of
12697 the current time. The age of the response can also optionally be
12698 checked. Two new options -validity_period and -status_age added to
12703 * If signature or public key algorithm is unrecognized print out its
12704 OID rather that just UNKNOWN.
12708 * Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
12709 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
12710 ID to be generated from the issuer certificate alone which can then be
12711 passed to OCSP_id_issuer_cmp().
12715 * New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
12716 ASN1 modules to export functions returning ASN1_ITEM pointers
12717 instead of the ASN1_ITEM structures themselves. This adds several
12718 new macros which allow the underlying ASN1 function/structure to
12719 be accessed transparently. As a result code should not use ASN1_ITEM
12720 references directly (such as &X509_it) but instead use the relevant
12721 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
12722 use of the new ASN1 code on platforms where exporting structures
12723 is problematical (for example in shared libraries) but exporting
12724 functions returning pointers to structures is not.
12728 * Add support for overriding the generation of SSL/TLS session IDs.
12729 These callbacks can be registered either in an SSL_CTX or per SSL.
12730 The purpose of this is to allow applications to control, if they wish,
12731 the arbitrary values chosen for use as session IDs, particularly as it
12732 can be useful for session caching in multiple-server environments. A
12733 command-line switch for testing this (and any client code that wishes
12734 to use such a feature) has been added to "s_server".
12736 *Geoff Thorpe, Lutz Jaenicke*
12738 * Modify mkdef.pl to recognise and parse preprocessor conditionals
12739 of the form `#if defined(...) || defined(...) || ...` and
12740 `#if !defined(...) && !defined(...) && ...`. This also avoids
12741 the growing number of special cases it was previously handling.
12745 * Make all configuration macros available for application by making
12746 sure they are available in opensslconf.h, by giving them names starting
12747 with `OPENSSL_` to avoid conflicts with other packages and by making
12748 sure e_os2.h will cover all platform-specific cases together with
12750 Additionally, it is now possible to define configuration/platform-
12751 specific names (called "system identities"). In the C code, these
12752 are prefixed with `OPENSSL_SYSNAME_`. e_os2.h will create another
12753 macro with the name beginning with `OPENSSL_SYS_`, which is determined
12754 from `OPENSSL_SYSNAME_*` or compiler-specific macros depending on
12759 * New option -set_serial to 'req' and 'x509' this allows the serial
12760 number to use to be specified on the command line. Previously self
12761 signed certificates were hard coded with serial number 0 and the
12762 CA options of 'x509' had to use a serial number in a file which was
12767 * New options to 'ca' utility to support V2 CRL entry extensions.
12768 Currently CRL reason, invalidity date and hold instruction are
12769 supported. Add new CRL extensions to V3 code and some new objects.
12773 * New function EVP_CIPHER_CTX_set_padding() this is used to
12774 disable standard block padding (aka PKCS#5 padding) in the EVP
12775 API, which was previously mandatory. This means that the data is
12776 not padded in any way and so the total length much be a multiple
12777 of the block size, otherwise an error occurs.
12781 * Initial (incomplete) OCSP SSL support.
12785 * New function OCSP_parse_url(). This splits up a URL into its host,
12786 port and path components: primarily to parse OCSP URLs. New -url
12787 option to ocsp utility.
12791 * New nonce behavior. The return value of OCSP_check_nonce() now
12792 reflects the various checks performed. Applications can decide
12793 whether to tolerate certain situations such as an absent nonce
12794 in a response when one was present in a request: the ocsp application
12795 just prints out a warning. New function OCSP_add1_basic_nonce()
12796 this is to allow responders to include a nonce in a response even if
12797 the request is nonce-less.
12801 * Disable stdin buffering in `load_cert()` (`apps/apps.c`) so that no certs are
12802 skipped when using openssl x509 multiple times on a single input file,
12803 e.g. `(openssl x509 -out cert1; openssl x509 -out cert2) <certs`.
12807 * Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
12808 set string type: to handle setting ASN1_TIME structures. Fix ca
12809 utility to correctly initialize revocation date of CRLs.
12813 * New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
12814 the clients preferred ciphersuites and rather use its own preferences.
12815 Should help to work around M$ SGC (Server Gated Cryptography) bug in
12816 Internet Explorer by ensuring unchanged hash method during stepup.
12817 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
12821 * Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
12822 to aes and add a new 'exist' option to print out symbols that don't
12827 * Additional options to ocsp utility to allow flags to be set and
12828 additional certificates supplied.
12832 * Add the option -VAfile to 'openssl ocsp', so the user can give the
12833 OCSP client a number of certificate to only verify the response
12838 * Update Rijndael code to version 3.0 and change EVP AES ciphers to
12839 handle the new API. Currently only ECB, CBC modes supported. Add new
12842 Add TLS AES ciphersuites as described in RFC3268, "Advanced
12843 Encryption Standard (AES) Ciphersuites for Transport Layer
12844 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were
12845 not enabled by default and were not part of the "ALL" ciphersuite
12846 alias because they were not yet official; they could be
12847 explicitly requested by specifying the "AESdraft" ciphersuite
12848 group alias. In the final release of OpenSSL 0.9.7, the group
12849 alias is called "AES" and is part of "ALL".)
12851 *Ben Laurie, Steve Henson, Bodo Moeller*
12853 * New function OCSP_copy_nonce() to copy nonce value (if present) from
12854 request to response.
12858 * Functions for OCSP responders. OCSP_request_onereq_count(),
12859 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
12860 extract information from a certificate request. OCSP_response_create()
12861 creates a response and optionally adds a basic response structure.
12862 OCSP_basic_add1_status() adds a complete single response to a basic
12863 response and returns the OCSP_SINGLERESP structure just added (to allow
12864 extensions to be included for example). OCSP_basic_add1_cert() adds a
12865 certificate to a basic response and OCSP_basic_sign() signs a basic
12866 response with various flags. New helper functions ASN1_TIME_check()
12867 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
12868 (converts ASN1_TIME to GeneralizedTime).
12872 * Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
12873 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
12874 structure from a certificate. X509_pubkey_digest() digests the public_key
12875 contents: this is used in various key identifiers.
12879 * Make sk_sort() tolerate a NULL argument.
12881 *Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>*
12883 * New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
12884 passed by the function are trusted implicitly. If any of them signed the
12885 response then it is assumed to be valid and is not verified.
12889 * In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
12890 to data. This was previously part of the PKCS7 ASN1 code. This
12891 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
12892 *Steve Henson, reported by Kenneth R. Robinette
12893 <support@securenetterm.com>*
12895 * Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
12896 routines: without these tracing memory leaks is very painful.
12897 Fix leaks in PKCS12 and PKCS7 routines.
12901 * Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
12902 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
12903 effectively meant GeneralizedTime would never be used. Now it
12904 is initialised to -1 but X509_time_adj() now has to check the value
12905 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
12906 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
12907 *Steve Henson, reported by Kenneth R. Robinette
12908 <support@securenetterm.com>*
12910 * Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
12911 result in a zero length in the ASN1_INTEGER structure which was
12912 not consistent with the structure when d2i_ASN1_INTEGER() was used
12913 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
12914 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
12915 where it did not print out a minus for negative ASN1_INTEGER.
12919 * Add summary printout to ocsp utility. The various functions which
12920 convert status values to strings have been renamed to:
12921 OCSP_response_status_str(), OCSP_cert_status_str() and
12922 OCSP_crl_reason_str() and are no longer static. New options
12923 to verify nonce values and to disable verification. OCSP response
12924 printout format cleaned up.
12928 * Add additional OCSP certificate checks. These are those specified
12929 in RFC2560. This consists of two separate checks: the CA of the
12930 certificate being checked must either be the OCSP signer certificate
12931 or the issuer of the OCSP signer certificate. In the latter case the
12932 OCSP signer certificate must contain the OCSP signing extended key
12933 usage. This check is performed by attempting to match the OCSP
12934 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
12935 in the OCSP_CERTID structures of the response.
12939 * Initial OCSP certificate verification added to OCSP_basic_verify()
12940 and related routines. This uses the standard OpenSSL certificate
12941 verify routines to perform initial checks (just CA validity) and
12942 to obtain the certificate chain. Then additional checks will be
12943 performed on the chain. Currently the root CA is checked to see
12944 if it is explicitly trusted for OCSP signing. This is used to set
12945 a root CA as a global signing root: that is any certificate that
12946 chains to that CA is an acceptable OCSP signing certificate.
12950 * New '-extfile ...' option to 'openssl ca' for reading X.509v3
12951 extensions from a separate configuration file.
12952 As when reading extensions from the main configuration file,
12953 the '-extensions ...' option may be used for specifying the
12956 *Massimiliano Pala <madwolf@comune.modena.it>*
12958 * New OCSP utility. Allows OCSP requests to be generated or
12959 read. The request can be sent to a responder and the output
12960 parsed, outputted or printed in text form. Not complete yet:
12961 still needs to check the OCSP response validity.
12965 * New subcommands for 'openssl ca':
12966 `openssl ca -status <serial>` prints the status of the cert with
12967 the given serial number (according to the index file).
12968 `openssl ca -updatedb` updates the expiry status of certificates
12971 *Massimiliano Pala <madwolf@comune.modena.it>*
12973 * New '-newreq-nodes' command option to CA.pl. This is like
12974 '-newreq', but calls 'openssl req' with the '-nodes' option
12975 so that the resulting key is not encrypted.
12977 *Damien Miller <djm@mindrot.org>*
12979 * New configuration for the GNU Hurd.
12981 *Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte*
12983 * Initial code to implement OCSP basic response verify. This
12984 is currently incomplete. Currently just finds the signer's
12985 certificate and verifies the signature on the response.
12989 * New SSLeay_version code SSLEAY_DIR to determine the compiled-in
12990 value of OPENSSLDIR. This is available via the new '-d' option
12991 to 'openssl version', and is also included in 'openssl version -a'.
12995 * Allowing defining memory allocation callbacks that will be given
12996 file name and line number information in additional arguments
12997 (a `const char*` and an int). The basic functionality remains, as
12998 well as the original possibility to just replace malloc(),
12999 realloc() and free() by functions that do not know about these
13000 additional arguments. To register and find out the current
13001 settings for extended allocation functions, the following
13002 functions are provided:
13004 CRYPTO_set_mem_ex_functions
13005 CRYPTO_set_locked_mem_ex_functions
13006 CRYPTO_get_mem_ex_functions
13007 CRYPTO_get_locked_mem_ex_functions
13009 These work the same way as CRYPTO_set_mem_functions and friends.
13010 `CRYPTO_get_[locked_]mem_functions` now writes 0 where such an
13011 extended allocation function is enabled.
13012 Similarly, `CRYPTO_get_[locked_]mem_ex_functions` writes 0 where
13013 a conventional allocation function is enabled.
13015 *Richard Levitte, Bodo Moeller*
13017 * Finish off removing the remaining LHASH function pointer casts.
13018 There should no longer be any prototype-casting required when using
13019 the LHASH abstraction, and any casts that remain are "bugs". See
13020 the callback types and macros at the head of lhash.h for details
13021 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
13025 * Add automatic query of EGD sockets in RAND_poll() for the unix variant.
13026 If /dev/[u]random devices are not available or do not return enough
13027 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
13029 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
13030 /etc/entropy will be queried once each in this sequence, querying stops
13031 when enough entropy was collected without querying more sockets.
13035 * Change the Unix RAND_poll() variant to be able to poll several
13036 random devices, as specified by DEVRANDOM, until a sufficient amount
13037 of data has been collected. We spend at most 10 ms on each file
13038 (select timeout) and read in non-blocking mode. DEVRANDOM now
13039 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
13040 (previously it was just the string "/dev/urandom"), so on typical
13041 platforms the 10 ms delay will never occur.
13042 Also separate out the Unix variant to its own file, rand_unix.c.
13043 For VMS, there's a currently-empty rand_vms.c.
13047 * Move OCSP client related routines to ocsp_cl.c. These
13048 provide utility functions which an application needing
13049 to issue a request to an OCSP responder and analyse the
13050 response will typically need: as opposed to those which an
13051 OCSP responder itself would need which will be added later.
13053 OCSP_request_sign() signs an OCSP request with an API similar
13054 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
13055 response. OCSP_response_get1_basic() extracts basic response
13056 from response. OCSP_resp_find_status(): finds and extracts status
13057 information from an OCSP_CERTID structure (which will be created
13058 when the request structure is built). These are built from lower
13059 level functions which work on OCSP_SINGLERESP structures but
13060 won't normally be used unless the application wishes to examine
13061 extensions in the OCSP response for example.
13063 Replace nonce routines with a pair of functions.
13064 OCSP_request_add1_nonce() adds a nonce value and optionally
13065 generates a random value. OCSP_check_nonce() checks the
13066 validity of the nonce in an OCSP response.
13070 * Change function OCSP_request_add() to OCSP_request_add0_id().
13071 This doesn't copy the supplied OCSP_CERTID and avoids the
13072 need to free up the newly created id. Change return type
13073 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
13074 This can then be used to add extensions to the request.
13075 Deleted OCSP_request_new(), since most of its functionality
13076 is now in OCSP_REQUEST_new() (and the case insensitive name
13077 clash) apart from the ability to set the request name which
13078 will be added elsewhere.
13082 * Update OCSP API. Remove obsolete extensions argument from
13083 various functions. Extensions are now handled using the new
13084 OCSP extension code. New simple OCSP HTTP function which
13085 can be used to send requests and parse the response.
13089 * Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
13090 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
13091 uses the special reorder version of SET OF to sort the attributes
13092 and reorder them to match the encoded order. This resolves a long
13093 standing problem: a verify on a PKCS7 structure just after signing
13094 it used to fail because the attribute order did not match the
13095 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
13096 it uses the received order. This is necessary to tolerate some broken
13097 software that does not order SET OF. This is handled by encoding
13098 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
13099 to produce the required SET OF.
13103 * Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
13104 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
13105 files to get correct declarations of the ASN.1 item variables.
13109 * Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
13110 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
13111 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
13112 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
13113 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
13114 ASN1_ITEM and no wrapper functions.
13118 * New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
13119 replace the old function pointer based I/O routines. Change most of
13120 the `*_d2i_bio()` and `*_d2i_fp()` functions to use these.
13124 * Enhance mkdef.pl to be more accepting about spacing in C preprocessor
13125 lines, recognize more "algorithms" that can be deselected, and make
13126 it complain about algorithm deselection that isn't recognised.
13130 * New ASN1 functions to handle dup, sign, verify, digest, pack and
13131 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
13132 to use new functions. Add NO_ASN1_OLD which can be set to remove
13133 some old style ASN1 functions: this can be used to determine if old
13134 code will still work when these eventually go away.
13138 * New extension functions for OCSP structures, these follow the
13139 same conventions as certificates and CRLs.
13143 * New function X509V3_add1_i2d(). This automatically encodes and
13144 adds an extension. Its behaviour can be customised with various
13145 flags to append, replace or delete. Various wrappers added for
13146 certificates and CRLs.
13150 * Fix to avoid calling the underlying ASN1 print routine when
13151 an extension cannot be parsed. Correct a typo in the
13152 OCSP_SERVICELOC extension. Tidy up print OCSP format.
13156 * Make mkdef.pl parse some of the ASN1 macros and add appropriate
13157 entries for variables.
13161 * Add functionality to `apps/openssl.c` for detecting locking
13162 problems: As the program is single-threaded, all we have
13163 to do is register a locking callback using an array for
13164 storing which locks are currently held by the program.
13168 * Use a lock around the call to CRYPTO_get_ex_new_index() in
13169 SSL_get_ex_data_X509_STORE_idx(), which is used in
13170 ssl_verify_cert_chain() and thus can be called at any time
13171 during TLS/SSL handshakes so that thread-safety is essential.
13172 Unfortunately, the ex_data design is not at all suited
13173 for multi-threaded use, so it probably should be abolished.
13177 * Added Broadcom "ubsec" ENGINE to OpenSSL.
13179 *Broadcom, tweaked and integrated by Geoff Thorpe*
13181 * Move common extension printing code to new function
13182 X509V3_print_extensions(). Reorganise OCSP print routines and
13183 implement some needed OCSP ASN1 functions. Add OCSP extensions.
13187 * New function X509_signature_print() to remove duplication in some
13192 * Add a special meaning when SET OF and SEQUENCE OF flags are both
13193 set (this was treated exactly the same as SET OF previously). This
13194 is used to reorder the STACK representing the structure to match the
13195 encoding. This will be used to get round a problem where a PKCS7
13196 structure which was signed could not be verified because the STACK
13197 order did not reflect the encoded order.
13201 * Reimplement the OCSP ASN1 module using the new code.
13205 * Update the X509V3 code to permit the use of an ASN1_ITEM structure
13206 for its ASN1 operations. The old style function pointers still exist
13207 for now but they will eventually go away.
13211 * Merge in replacement ASN1 code from the ASN1 branch. This almost
13212 completely replaces the old ASN1 functionality with a table driven
13213 encoder and decoder which interprets an ASN1_ITEM structure describing
13214 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
13215 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
13216 has also been converted to the new form.
13220 * Change BN_mod_exp_recp so that negative moduli are tolerated
13221 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
13222 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
13223 for negative moduli.
13227 * Fix BN_uadd and BN_usub: Always return non-negative results instead
13228 of not touching the result's sign bit.
13232 * BN_div bugfix: If the result is 0, the sign (res->neg) must not be
13237 * Changed the LHASH code to use prototypes for callbacks, and created
13238 macros to declare and implement thin (optionally static) functions
13239 that provide type-safety and avoid function pointer casting for the
13240 type-specific callbacks.
13244 * Added Kerberos Cipher Suites to be used with TLS, as written in
13246 *Veers Staats <staatsvr@asc.hpc.mil>,
13247 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte*
13249 * Reformat the FAQ so the different questions and answers can be divided
13250 in sections depending on the subject.
13254 * Have the zlib compression code load ZLIB.DLL dynamically under
13259 * New function BN_mod_sqrt for computing square roots modulo a prime
13260 (using the probabilistic Tonelli-Shanks algorithm unless
13261 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
13262 be handled deterministically).
13264 *Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller*
13266 * Make BN_mod_inverse faster by explicitly handling small quotients
13267 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
13268 512 bits], about 30% for larger ones [1024 or 2048 bits].)
13272 * New function BN_kronecker.
13276 * Fix BN_gcd so that it works on negative inputs; the result is
13277 positive unless both parameters are zero.
13278 Previously something reasonably close to an infinite loop was
13279 possible because numbers could be growing instead of shrinking
13280 in the implementation of Euclid's algorithm.
13284 * Fix BN_is_word() and BN_is_one() macros to take into account the
13285 sign of the number in question.
13287 Fix BN_is_word(a,w) to work correctly for w == 0.
13289 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
13290 because its test if the absolute value of 'a' equals 'w'.
13291 Note that BN_abs_is_word does *not* handle w == 0 reliably;
13292 it exists mostly for use in the implementations of BN_is_zero(),
13293 BN_is_one(), and BN_is_word().
13297 * New function BN_swap.
13301 * Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
13302 the exponentiation functions are more likely to produce reasonable
13303 results on negative inputs.
13307 * Change BN_mod_mul so that the result is always non-negative.
13308 Previously, it could be negative if one of the factors was negative;
13309 I don't think anyone really wanted that behaviour.
13313 * Move `BN_mod_...` functions into new file `crypto/bn/bn_mod.c`
13314 (except for exponentiation, which stays in `crypto/bn/bn_exp.c`,
13315 and `BN_mod_mul_reciprocal`, which stays in `crypto/bn/bn_recp.c`)
13316 and add new functions:
13325 BN_mod_lshift1_quick
13327 BN_mod_lshift_quick
13329 These functions always generate non-negative results.
13331 `BN_nnmod` otherwise is `like BN_mod` (if `BN_mod` computes a remainder `r`
13332 such that `|m| < r < 0`, `BN_nnmod` will output `rem + |m|` instead).
13334 `BN_mod_XXX_quick(r, a, [b,] m)` generates the same result as
13335 `BN_mod_XXX(r, a, [b,] m, ctx)`, but requires that `a` [and `b`]
13336 be reduced modulo `m`.
13338 *Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller*
13341 The following entry accidentally appeared in the CHANGES file
13342 distributed with OpenSSL 0.9.7. The modifications described in
13343 it do *not* apply to OpenSSL 0.9.7.
13345 * Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
13346 was actually never needed) and in BN_mul(). The removal in BN_mul()
13347 required a small change in bn_mul_part_recursive() and the addition
13348 of the functions bn_cmp_part_words(), bn_sub_part_words() and
13349 bn_add_part_words(), which do the same thing as bn_cmp_words(),
13350 bn_sub_words() and bn_add_words() except they take arrays with
13356 * In 'openssl passwd', verify passwords read from the terminal
13357 unless the '-salt' option is used (which usually means that
13358 verification would just waste user's time since the resulting
13359 hash is going to be compared with some given password hash)
13360 or the new '-noverify' option is used.
13362 This is an incompatible change, but it does not affect
13363 non-interactive use of 'openssl passwd' (passwords on the command
13364 line, '-stdin' option, '-in ...' option) and thus should not
13365 cause any problems.
13369 * Remove all references to RSAref, since there's no more need for it.
13373 * Make DSO load along a path given through an environment variable
13374 (SHLIB_PATH) with shl_load().
13378 * Constify the ENGINE code as a result of BIGNUM constification.
13379 Also constify the RSA code and most things related to it. In a
13380 few places, most notable in the depth of the ASN.1 code, ugly
13381 casts back to non-const were required (to be solved at a later
13386 * Make it so the openssl application has all engines loaded by default.
13390 * Constify the BIGNUM routines a little more.
13394 * Add the following functions:
13396 ENGINE_load_cswift()
13398 ENGINE_load_atalla()
13399 ENGINE_load_nuron()
13400 ENGINE_load_builtin_engines()
13402 That way, an application can itself choose if external engines that
13403 are built-in in OpenSSL shall ever be used or not. The benefit is
13404 that applications won't have to be linked with libdl or other dso
13405 libraries unless it's really needed.
13407 Changed 'openssl engine' to load all engines on demand.
13408 Changed the engine header files to avoid the duplication of some
13409 declarations (they differed!).
13413 * 'openssl engine' can now list capabilities.
13417 * Better error reporting in 'openssl engine'.
13421 * Never call load_dh_param(NULL) in s_server.
13425 * Add engine application. It can currently list engines by name and
13426 identity, and test if they are actually available.
13430 * Improve RPM specification file by forcing symbolic linking and making
13431 sure the installed documentation is also owned by root.root.
13433 *Damien Miller <djm@mindrot.org>*
13435 * Give the OpenSSL applications more possibilities to make use of
13436 keys (public as well as private) handled by engines.
13440 * Add OCSP code that comes from CertCo.
13444 * Add VMS support for the Rijndael code.
13448 * Added untested support for Nuron crypto accelerator.
13452 * Add support for external cryptographic devices. This code was
13453 previously distributed separately as the "engine" branch.
13455 *Geoff Thorpe, Richard Levitte*
13457 * Rework the filename-translation in the DSO code. It is now possible to
13458 have far greater control over how a "name" is turned into a filename
13459 depending on the operating environment and any oddities about the
13460 different shared library filenames on each system.
13464 * Support threads on FreeBSD-elf in Configure.
13468 * Fix for SHA1 assembly problem with MASM: it produces
13469 warnings about corrupt line number information when assembling
13470 with debugging information. This is caused by the overlapping
13473 *Bernd Matthes <mainbug@celocom.de>, Steve Henson*
13476 NCONF_get_number() has no error checking at all. As a replacement,
13477 NCONF_get_number_e() is defined (`_e` for "error checking") and is
13478 promoted strongly. The old NCONF_get_number is kept around for
13479 binary backward compatibility.
13480 Make it possible for methods to load from something other than a BIO,
13481 by providing a function pointer that is given a name instead of a BIO.
13482 For example, this could be used to load configuration data from an
13487 * Fix for non blocking accept BIOs. Added new I/O special reason
13488 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
13489 with non blocking I/O was not possible because no retry code was
13490 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
13495 * Added the beginnings of Rijndael support.
13499 * Fix for bug in DirectoryString mask setting. Add support for
13500 X509_NAME_print_ex() in 'req' and X509_print_ex() function
13501 to allow certificate printing to more controllable, additional
13502 'certopt' option to 'x509' to allow new printing options to be
13507 * Clean old EAY MD5 hack from e_os.h.
13511 ### Changes between 0.9.6l and 0.9.6m [17 Mar 2004]
13513 * Fix null-pointer assignment in do_change_cipher_spec() revealed
13514 by using the Codenomicon TLS Test Tool ([CVE-2004-0079])
13516 *Joe Orton, Steve Henson*
13518 ### Changes between 0.9.6k and 0.9.6l [04 Nov 2003]
13520 * Fix additional bug revealed by the NISCC test suite:
13522 Stop bug triggering large recursion when presented with
13523 certain ASN.1 tags ([CVE-2003-0851])
13527 ### Changes between 0.9.6j and 0.9.6k [30 Sep 2003]
13529 * Fix various bugs revealed by running the NISCC test suite:
13531 Stop out of bounds reads in the ASN1 code when presented with
13532 invalid tags (CVE-2003-0543 and CVE-2003-0544).
13534 If verify callback ignores invalid public key errors don't try to check
13535 certificate signature with the NULL public key.
13539 * In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
13540 if the server requested one: as stated in TLS 1.0 and SSL 3.0
13545 * In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
13546 extra data after the compression methods not only for TLS 1.0
13547 but also for SSL 3.0 (as required by the specification).
13549 *Bodo Moeller; problem pointed out by Matthias Loepfe*
13551 * Change X509_certificate_type() to mark the key as exported/exportable
13552 when it's 512 *bits* long, not 512 bytes.
13556 ### Changes between 0.9.6i and 0.9.6j [10 Apr 2003]
13558 * Countermeasure against the Klima-Pokorny-Rosa extension of
13559 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
13560 a protocol version number mismatch like a decryption error
13561 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
13565 * Turn on RSA blinding by default in the default implementation
13566 to avoid a timing attack. Applications that don't want it can call
13567 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
13568 They would be ill-advised to do so in most cases.
13570 *Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller*
13572 * Change RSA blinding code so that it works when the PRNG is not
13573 seeded (in this case, the secret RSA exponent is abused as
13574 an unpredictable seed -- if it is not unpredictable, there
13575 is no point in blinding anyway). Make RSA blinding thread-safe
13576 by remembering the creator's thread ID in rsa->blinding and
13577 having all other threads use local one-time blinding factors
13578 (this requires more computation than sharing rsa->blinding, but
13579 avoids excessive locking; and if an RSA object is not shared
13580 between threads, blinding will still be very fast).
13584 ### Changes between 0.9.6h and 0.9.6i [19 Feb 2003]
13586 * In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
13587 via timing by performing a MAC computation even if incorrect
13588 block cipher padding has been found. This is a countermeasure
13589 against active attacks where the attacker has to distinguish
13590 between bad padding and a MAC verification error. ([CVE-2003-0078])
13592 *Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
13593 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
13594 Martin Vuagnoux (EPFL, Ilion)*
13596 ### Changes between 0.9.6g and 0.9.6h [5 Dec 2002]
13598 * New function OPENSSL_cleanse(), which is used to cleanse a section of
13599 memory from its contents. This is done with a counter that will
13600 place alternating values in each byte. This can be used to solve
13601 two issues: 1) the removal of calls to memset() by highly optimizing
13602 compilers, and 2) cleansing with other values than 0, since those can
13603 be read through on certain media, for example a swap space on disk.
13607 * Bugfix: client side session caching did not work with external caching,
13608 because the session->cipher setting was not restored when reloading
13609 from the external cache. This problem was masked, when
13610 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set.
13611 (Found by Steve Haslam <steve@araqnid.ddts.net>.)
13615 * Fix client_certificate (ssl/s2_clnt.c): The permissible total
13616 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33.
13618 *Zeev Lieber <zeev-l@yahoo.com>*
13620 * Undo an undocumented change introduced in 0.9.6e which caused
13621 repeated calls to OpenSSL_add_all_ciphers() and
13622 OpenSSL_add_all_digests() to be ignored, even after calling
13627 * Change the default configuration reader to deal with last line not
13628 being properly terminated.
13632 * Change X509_NAME_cmp() so it applies the special rules on handling
13633 DN values that are of type PrintableString, as well as RDNs of type
13634 emailAddress where the value has the type ia5String.
13636 *stefank@valicert.com via Richard Levitte*
13638 * Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half
13639 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently
13640 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be
13641 the bitwise-OR of the two for use by the majority of applications
13642 wanting this behaviour, and update the docs. The documented
13643 behaviour and actual behaviour were inconsistent and had been
13644 changing anyway, so this is more a bug-fix than a behavioural
13647 *Geoff Thorpe, diagnosed by Nadav Har'El*
13649 * Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c
13650 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes).
13654 * Fix initialization code race conditions in
13655 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(),
13656 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(),
13657 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(),
13658 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(),
13659 ssl2_get_cipher_by_char(),
13660 ssl3_get_cipher_by_char().
13662 *Patrick McCormick <patrick@tellme.com>, Bodo Moeller*
13664 * Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after
13665 the cached sessions are flushed, as the remove_cb() might use ex_data
13666 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com>
13667 (see [openssl.org #212]).
13669 *Geoff Thorpe, Lutz Jaenicke*
13671 * Fix typo in OBJ_txt2obj which incorrectly passed the content
13672 length, instead of the encoding length to d2i_ASN1_OBJECT.
13676 ### Changes between 0.9.6f and 0.9.6g [9 Aug 2002]
13678 * [In 0.9.6g-engine release:]
13679 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use `_stdcall`).
13681 *Lynn Gazis <lgazis@rainbow.com>*
13683 ### Changes between 0.9.6e and 0.9.6f [8 Aug 2002]
13685 * Fix ASN1 checks. Check for overflow by comparing with LONG_MAX
13686 and get fix the header length calculation.
13687 *Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>,
13688 Alon Kantor <alonk@checkpoint.com> (and others), Steve Henson*
13690 * Use proper error handling instead of 'assertions' in buffer
13691 overflow checks added in 0.9.6e. This prevents DoS (the
13692 assertions could call abort()).
13694 *Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller*
13696 ### Changes between 0.9.6d and 0.9.6e [30 Jul 2002]
13698 * Add various sanity checks to asn1_get_length() to reject
13699 the ASN1 length bytes if they exceed sizeof(long), will appear
13700 negative or the content length exceeds the length of the
13703 *Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>*
13705 * Fix cipher selection routines: ciphers without encryption had no flags
13706 for the cipher strength set and where therefore not handled correctly
13707 by the selection routines (PR #130).
13711 * Fix EVP_dsa_sha macro.
13716 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS
13717 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure
13718 that was added in OpenSSL 0.9.6d.
13720 As the countermeasure turned out to be incompatible with some
13721 broken SSL implementations, the new option is part of SSL_OP_ALL.
13722 SSL_OP_ALL is usually employed when compatibility with weird SSL
13723 implementations is desired (e.g. '-bugs' option to 's_client' and
13724 's_server'), so the new option is automatically set in many
13729 * Changes in security patch:
13731 Changes marked "(CHATS)" were sponsored by the Defense Advanced
13732 Research Projects Agency (DARPA) and Air Force Research Laboratory,
13733 Air Force Materiel Command, USAF, under agreement number
13736 * Add various sanity checks to asn1_get_length() to reject
13737 the ASN1 length bytes if they exceed sizeof(long), will appear
13738 negative or the content length exceeds the length of the
13739 supplied buffer. ([CVE-2002-0659])
13741 *Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>*
13743 * Assertions for various potential buffer overflows, not known to
13744 happen in practice.
13746 *Ben Laurie (CHATS)*
13748 * Various temporary buffers to hold ASCII versions of integers were
13749 too small for 64 bit platforms. ([CVE-2002-0655])
13750 *Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>*
13752 * Remote buffer overflow in SSL3 protocol - an attacker could
13753 supply an oversized session ID to a client. ([CVE-2002-0656])
13755 *Ben Laurie (CHATS)*
13757 * Remote buffer overflow in SSL2 protocol - an attacker could
13758 supply an oversized client master key. ([CVE-2002-0656])
13760 *Ben Laurie (CHATS)*
13762 ### Changes between 0.9.6c and 0.9.6d [9 May 2002]
13764 * Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
13765 encoded as NULL) with id-dsa-with-sha1.
13767 *Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller*
13769 * Check various `X509_...()` return values in `apps/req.c`.
13771 *Nils Larsch <nla@trustcenter.de>*
13773 * Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
13774 an end-of-file condition would erroneously be flagged, when the CRLF
13775 was just at the end of a processed block. The bug was discovered when
13776 processing data through a buffering memory BIO handing the data to a
13777 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
13778 <ptsekov@syntrex.com> and Nedelcho Stanev.
13782 * Implement a countermeasure against a vulnerability recently found
13783 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
13784 before application data chunks to avoid the use of known IVs
13785 with data potentially chosen by the attacker.
13789 * Fix length checks in ssl3_get_client_hello().
13793 * TLS/SSL library bugfix: use s->s3->in_read_app_data differently
13794 to prevent ssl3_read_internal() from incorrectly assuming that
13795 ssl3_read_bytes() found application data while handshake
13796 processing was enabled when in fact s->s3->in_read_app_data was
13797 merely automatically cleared during the initial handshake.
13799 *Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>*
13801 * Fix object definitions for Private and Enterprise: they were not
13802 recognized in their shortname (=lowercase) representation. Extend
13803 obj_dat.pl to issue an error when using undefined keywords instead
13804 of silently ignoring the problem (Svenning Sorensen
13805 <sss@sss.dnsalias.net>).
13809 * Fix DH_generate_parameters() so that it works for 'non-standard'
13810 generators, i.e. generators other than 2 and 5. (Previously, the
13811 code did not properly initialise the 'add' and 'rem' values to
13812 BN_generate_prime().)
13814 In the new general case, we do not insist that 'generator' is
13815 actually a primitive root: This requirement is rather pointless;
13816 a generator of the order-q subgroup is just as good, if not
13821 * Map new X509 verification errors to alerts. Discovered and submitted by
13822 Tom Wu <tom@arcot.com>.
13826 * Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
13827 returning non-zero before the data has been completely received
13828 when using non-blocking I/O.
13830 *Bodo Moeller; problem pointed out by John Hughes*
13832 * Some of the ciphers missed the strength entry (SSL_LOW etc).
13834 *Ben Laurie, Lutz Jaenicke*
13836 * Fix bug in SSL_clear(): bad sessions were not removed (found by
13837 Yoram Zahavi <YoramZ@gilian.com>).
13841 * Add information about CygWin 1.3 and on, and preserve proper
13842 configuration for the versions before that.
13844 *Corinna Vinschen <vinschen@redhat.com> and Richard Levitte*
13846 * Make removal from session cache (SSL_CTX_remove_session()) more robust:
13847 check whether we deal with a copy of a session and do not delete from
13848 the cache in this case. Problem reported by "Izhar Shoshani Levi"
13849 <izhar@checkpoint.com>.
13853 * Do not store session data into the internal session cache, if it
13854 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
13855 flag is set). Proposed by Aslam <aslam@funk.com>.
13859 * Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
13864 * [In 0.9.6d-engine release:]
13865 Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
13867 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
13869 * Add the configuration target linux-s390x.
13871 *Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte*
13873 * The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
13874 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
13875 variable as an indication that a ClientHello message has been
13876 received. As the flag value will be lost between multiple
13877 invocations of ssl3_accept when using non-blocking I/O, the
13878 function may not be aware that a handshake has actually taken
13879 place, thus preventing a new session from being added to the
13882 To avoid this problem, we now set s->new_session to 2 instead of
13883 using a local variable.
13885 *Lutz Jaenicke, Bodo Moeller*
13887 * Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
13888 if the SSL_R_LENGTH_MISMATCH error is detected.
13890 *Geoff Thorpe, Bodo Moeller*
13892 * New 'shared_ldflag' column in Configure platform table.
13896 * Fix EVP_CIPHER_mode macro.
13898 *"Dan S. Camper" <dan@bti.net>*
13900 * Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
13901 type, we must throw them away by setting rr->length to 0.
13903 *D P Chang <dpc@qualys.com>*
13905 ### Changes between 0.9.6b and 0.9.6c [21 dec 2001]
13907 * Fix BN_rand_range bug pointed out by Dominikus Scherkl
13908 <Dominikus.Scherkl@biodata.com>. (The previous implementation
13909 worked incorrectly for those cases where range = `10..._2` and
13910 `3*range` is two bits longer than range.)
13914 * Only add signing time to PKCS7 structures if it is not already
13919 * Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
13920 OBJ_ld_ce should be OBJ_id_ce.
13921 Also some ip-pda OIDs in crypto/objects/objects.txt were
13922 incorrect (cf. RFC 3039).
13924 *Matt Cooper, Frederic Giudicelli, Bodo Moeller*
13926 * Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
13927 returns early because it has nothing to do.
13929 *Andy Schneider <andy.schneider@bjss.co.uk>*
13931 * [In 0.9.6c-engine release:]
13932 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
13934 *Andy Schneider <andy.schneider@bjss.co.uk>*
13936 * [In 0.9.6c-engine release:]
13937 Add support for Cryptographic Appliance's keyserver technology.
13938 (Use engine 'keyclient')
13940 *Cryptographic Appliances and Geoff Thorpe*
13942 * Add a configuration entry for OS/390 Unix. The C compiler 'c89'
13943 is called via tools/c89.sh because arguments have to be
13944 rearranged (all '-L' options must appear before the first object
13947 *Richard Shapiro <rshapiro@abinitio.com>*
13949 * [In 0.9.6c-engine release:]
13950 Add support for Broadcom crypto accelerator cards, backported
13953 *Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox*
13955 * [In 0.9.6c-engine release:]
13956 Add support for SureWare crypto accelerator cards from
13957 Baltimore Technologies. (Use engine 'sureware')
13959 *Baltimore Technologies and Mark Cox*
13961 * [In 0.9.6c-engine release:]
13962 Add support for crypto accelerator cards from Accelerated
13963 Encryption Processing, www.aep.ie. (Use engine 'aep')
13965 *AEP Inc. and Mark Cox*
13967 * Add a configuration entry for gcc on UnixWare.
13969 *Gary Benson <gbenson@redhat.com>*
13971 * Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
13972 messages are stored in a single piece (fixed-length part and
13973 variable-length part combined) and fix various bugs found on the way.
13977 * Disable caching in BIO_gethostbyname(), directly use gethostbyname()
13978 instead. BIO_gethostbyname() does not know what timeouts are
13979 appropriate, so entries would stay in cache even when they have
13981 *Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>*
13983 * Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
13984 faced with a pathologically small ClientHello fragment that does
13985 not contain client_version: Instead of aborting with an error,
13986 simply choose the highest available protocol version (i.e.,
13987 TLS 1.0 unless it is disabled). In practice, ClientHello
13988 messages are never sent like this, but this change gives us
13989 strictly correct behaviour at least for TLS.
13993 * Fix SSL handshake functions and SSL_clear() such that SSL_clear()
13994 never resets s->method to s->ctx->method when called from within
13995 one of the SSL handshake functions.
13997 *Bodo Moeller; problem pointed out by Niko Baric*
13999 * In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
14000 (sent using the client's version number) if client_version is
14001 smaller than the protocol version in use. Also change
14002 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
14003 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
14004 the client will at least see that alert.
14008 * Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
14013 * Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
14014 client receives HelloRequest while in a handshake.
14016 *Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>*
14018 * Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
14019 should end in 'break', not 'goto end' which circumvents various
14020 cleanups done in state SSL_ST_OK. But session related stuff
14021 must be disabled for SSL_ST_OK in the case that we just sent a
14024 Also avoid some overhead by not calling ssl_init_wbio_buffer()
14025 before just sending a HelloRequest.
14027 *Bodo Moeller, Eric Rescorla <ekr@rtfm.com>*
14029 * Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
14030 reveal whether illegal block cipher padding was found or a MAC
14031 verification error occurred. (Neither SSLerr() codes nor alerts
14032 are directly visible to potential attackers, but the information
14033 may leak via logfiles.)
14035 Similar changes are not required for the SSL 2.0 implementation
14036 because the number of padding bytes is sent in clear for SSL 2.0,
14037 and the extra bytes are just ignored. However ssl/s2_pkt.c
14038 failed to verify that the purported number of padding bytes is in
14043 * Add OpenUNIX-8 support including shared libraries
14044 (Boyd Lynn Gerber <gerberb@zenez.com>).
14048 * Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
14049 'wristwatch attack' using huge encoding parameters (cf.
14050 James H. Manger's CRYPTO 2001 paper). Note that the
14051 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
14052 encoding parameters and hence was not vulnerable.
14056 * BN_sqr() bug fix.
14058 *Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>*
14060 * Rabin-Miller test analyses assume uniformly distributed witnesses,
14061 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
14062 followed by modular reduction.
14064 *Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>*
14066 * Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
14067 equivalent based on BN_pseudo_rand() instead of BN_rand().
14071 * s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
14072 This function was broken, as the check for a new client hello message
14073 to handle SGC did not allow these large messages.
14074 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
14078 * Add alert descriptions for TLSv1 to `SSL_alert_desc_string[_long]()`.
14082 * Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
14083 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
14087 * Rework the configuration and shared library support for Tru64 Unix.
14088 The configuration part makes use of modern compiler features and
14089 still retains old compiler behavior for those that run older versions
14090 of the OS. The shared library support part includes a variant that
14091 uses the RPATH feature, and is available through the special
14092 configuration target "alpha-cc-rpath", which will never be selected
14095 *Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte*
14097 * In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
14098 with the same message size as in ssl3_get_certificate_request().
14099 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
14100 messages might inadvertently be reject as too long.
14102 *Petr Lampa <lampa@fee.vutbr.cz>*
14104 * Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
14108 * Modified SSL library such that the verify_callback that has been set
14109 specifically for an SSL object with SSL_set_verify() is actually being
14110 used. Before the change, a verify_callback set with this function was
14111 ignored and the verify_callback() set in the SSL_CTX at the time of
14112 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
14113 to allow the necessary settings.
14117 * Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
14118 explicitly to NULL, as at least on Solaris 8 this seems not always to be
14119 done automatically (in contradiction to the requirements of the C
14120 standard). This made problems when used from OpenSSH.
14124 * In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
14125 dh->length and always used
14127 BN_rand_range(priv_key, dh->p).
14129 BN_rand_range() is not necessary for Diffie-Hellman, and this
14130 specific range makes Diffie-Hellman unnecessarily inefficient if
14131 dh->length (recommended exponent length) is much smaller than the
14132 length of dh->p. We could use BN_rand_range() if the order of
14133 the subgroup was stored in the DH structure, but we only have
14138 BN_rand(priv_key, l, ...)
14140 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
14147 RSA_eay_public_encrypt
14148 RSA_eay_private_decrypt
14149 RSA_eay_private_encrypt (signing)
14150 RSA_eay_public_decrypt (signature verification)
14152 (default implementations for RSA_public_encrypt,
14153 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
14154 always reject numbers >= n.
14158 * In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
14159 to synchronize access to 'locking_thread'. This is necessary on
14160 systems where access to 'locking_thread' (an 'unsigned long'
14161 variable) is not atomic.
14165 * In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
14166 *before* setting the 'crypto_lock_rand' flag. The previous code had
14167 a race condition if 0 is a valid thread ID.
14169 *Travis Vitek <vitek@roguewave.com>*
14171 * Add support for shared libraries under Irix.
14173 *Albert Chin-A-Young <china@thewrittenword.com>*
14175 * Add configuration option to build on Linux on both big-endian and
14176 little-endian MIPS.
14178 *Ralf Baechle <ralf@uni-koblenz.de>*
14180 * Add the possibility to create shared libraries on HP-UX.
14184 ### Changes between 0.9.6a and 0.9.6b [9 Jul 2001]
14186 * Change ssleay_rand_bytes (crypto/rand/md_rand.c)
14187 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
14188 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
14189 PRNG state recovery was possible based on the output of
14190 one PRNG request appropriately sized to gain knowledge on
14191 'md' followed by enough consecutive 1-byte PRNG requests
14192 to traverse all of 'state'.
14194 1. When updating 'md_local' (the current thread's copy of 'md')
14195 during PRNG output generation, hash all of the previous
14196 'md_local' value, not just the half used for PRNG output.
14198 2. Make the number of bytes from 'state' included into the hash
14199 independent from the number of PRNG bytes requested.
14201 The first measure alone would be sufficient to avoid
14202 Markku-Juhani's attack. (Actually it had never occurred
14203 to me that the half of 'md_local' used for chaining was the
14204 half from which PRNG output bytes were taken -- I had always
14205 assumed that the secret half would be used.) The second
14206 measure makes sure that additional data from 'state' is never
14207 mixed into 'md_local' in small portions; this heuristically
14208 further strengthens the PRNG.
14212 * Fix crypto/bn/asm/mips3.s.
14216 * When only the key is given to "enc", the IV is undefined. Print out
14217 an error message in this case.
14221 * Handle special case when X509_NAME is empty in X509 printing routines.
14225 * In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
14226 positive and less than q.
14230 * Don't change `*pointer` in CRYPTO_add_lock() is add_lock_callback is
14231 used: it isn't thread safe and the add_lock_callback should handle
14234 *Paul Rose <Paul.Rose@bridge.com>*
14236 * Verify that incoming data obeys the block size in
14237 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
14243 *Ulf Möller, Bodo Möller*
14245 * The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
14246 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
14247 when fixing the server behaviour for backwards-compatible 'client
14248 hello' messages. (Note that the attack is impractical against
14249 SSL 3.0 and TLS 1.0 anyway because length and version checking
14250 means that the probability of guessing a valid ciphertext is
14251 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
14254 Before 0.9.5, the countermeasure (hide the error by generating a
14255 random 'decryption result') did not work properly because
14256 ERR_clear_error() was missing, meaning that SSL_get_error() would
14257 detect the supposedly ignored error.
14259 Both problems are now fixed.
14263 * In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
14264 (previously it was 1024).
14268 * Fix for compatibility mode trust settings: ignore trust settings
14269 unless some valid trust or reject settings are present.
14273 * Fix for blowfish EVP: its a variable length cipher.
14277 * Fix various bugs related to DSA S/MIME verification. Handle missing
14278 parameters in DSA public key structures and return an error in the
14279 DSA routines if parameters are absent.
14283 * In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
14284 in the current directory if neither $RANDFILE nor $HOME was set.
14285 RAND_file_name() in 0.9.6a returned NULL in this case. This has
14286 caused some confusion to Windows users who haven't defined $HOME.
14287 Thus RAND_file_name() is changed again: e_os.h can define a
14288 DEFAULT_HOME, which will be used if $HOME is not set.
14289 For Windows, we use "C:"; on other platforms, we still require
14290 environment variables.
14292 * Move 'if (!initialized) RAND_poll()' into regions protected by
14293 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
14294 having multiple threads call RAND_poll() concurrently.
14298 * In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
14299 combination of a flag and a thread ID variable.
14300 Otherwise while one thread is in ssleay_rand_bytes (which sets the
14301 flag), *other* threads can enter ssleay_add_bytes without obeying
14302 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
14303 that they do not hold after the first thread unsets add_do_not_lock).
14307 * Change bctest again: '-x' expressions are not available in all
14308 versions of 'test'.
14312 ### Changes between 0.9.6 and 0.9.6a [5 Apr 2001]
14314 * Fix a couple of memory leaks in PKCS7_dataDecode()
14316 *Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>*
14318 * Change Configure and Makefiles to provide EXE_EXT, which will contain
14319 the default extension for executables, if any. Also, make the perl
14320 scripts that use symlink() to test if it really exists and use "cp"
14321 if it doesn't. All this made OpenSSL compilable and installable in
14326 * Fix for asn1_GetSequence() for indefinite length constructed data.
14327 If SEQUENCE is length is indefinite just set c->slen to the total
14328 amount of data available.
14330 *Steve Henson, reported by shige@FreeBSD.org*
14332 *This change does not apply to 0.9.7.*
14334 * Change bctest to avoid here-documents inside command substitution
14335 (workaround for FreeBSD /bin/sh bug).
14336 For compatibility with Ultrix, avoid shell functions (introduced
14337 in the bctest version that searches along $PATH).
14341 * Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
14342 with des_encrypt() defined on some operating systems, like Solaris
14347 * Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
14348 On the Importance of Eliminating Errors in Cryptographic
14349 Computations, J. Cryptology 14 (2001) 2, 101-119,
14350 <http://theory.stanford.edu/~dabo/papers/faults.ps.gz>).
14354 * MIPS assembler BIGNUM division bug fix.
14358 * Disabled incorrect Alpha assembler code.
14362 * Fix PKCS#7 decode routines so they correctly update the length
14363 after reading an EOC for the EXPLICIT tag.
14367 *This change does not apply to 0.9.7.*
14369 * Fix bug in PKCS#12 key generation routines. This was triggered
14370 if a 3DES key was generated with a 0 initial byte. Include
14371 PKCS12_BROKEN_KEYGEN compilation option to retain the old
14372 (but broken) behaviour.
14376 * Enhance bctest to search for a working bc along $PATH and print
14379 *Tim Rice <tim@multitalents.net> via Richard Levitte*
14381 * Fix memory leaks in err.c: free err_data string if necessary;
14382 don't write to the wrong index in ERR_set_error_data.
14386 * Implement ssl23_peek (analogous to ssl23_read), which previously
14391 * Replace rdtsc with `_emit` statements for VC++ version 5.
14393 *Jeremy Cooper <jeremy@baymoo.org>*
14395 * Make it possible to reuse SSLv2 sessions.
14399 * In copy_email() check for >= 0 as a return value for
14400 X509_NAME_get_index_by_NID() since 0 is a valid index.
14402 *Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>*
14404 * Avoid coredump with unsupported or invalid public keys by checking if
14405 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
14406 PKCS7_verify() fails with non detached data.
14410 * Don't use getenv in library functions when run as setuid/setgid.
14411 New function OPENSSL_issetugid().
14415 * Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
14416 due to incorrect handling of multi-threading:
14418 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
14420 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
14422 3. Count how many times MemCheck_off() has been called so that
14423 nested use can be treated correctly. This also avoids
14424 inband-signalling in the previous code (which relied on the
14425 assumption that thread ID 0 is impossible).
14429 * Add "-rand" option also to s_client and s_server.
14433 * Fix CPU detection on Irix 6.x.
14434 *Kurt Hockenbury <khockenb@stevens-tech.edu> and
14435 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>*
14437 * Fix X509_NAME bug which produced incorrect encoding if X509_NAME
14442 *This change does not apply to 0.9.7.*
14444 * Use the cached encoding of an X509_NAME structure rather than
14445 copying it. This is apparently the reason for the libsafe "errors"
14446 but the code is actually correct.
14450 * Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
14451 Bleichenbacher's DSA attack.
14452 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
14453 to be set and top=0 forces the highest bit to be set; top=-1 is new
14454 and leaves the highest bit random.
14456 *Ulf Moeller, Bodo Moeller*
14458 * In the `NCONF_...`-based implementations for `CONF_...` queries
14459 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
14460 a temporary CONF structure with the data component set to NULL
14461 (which gives segmentation faults in lh_retrieve).
14462 Instead, use NULL for the CONF pointer in CONF_get_string and
14463 CONF_get_number (which may use environment variables) and directly
14464 return NULL from CONF_get_section.
14468 * Fix potential buffer overrun for EBCDIC.
14472 * Tolerate nonRepudiation as being valid for S/MIME signing and certSign
14473 keyUsage if basicConstraints absent for a CA.
14477 * Make SMIME_write_PKCS7() write mail header values with a format that
14478 is more generally accepted (no spaces before the semicolon), since
14479 some programs can't parse those values properly otherwise. Also make
14480 sure BIO's that break lines after each write do not create invalid
14485 * Make the CRL encoding routines work with empty SEQUENCE OF. The
14486 macros previously used would not encode an empty SEQUENCE OF
14487 and break the signature.
14491 *This change does not apply to 0.9.7.*
14493 * Zero the premaster secret after deriving the master secret in
14498 * Add some EVP_add_digest_alias registrations (as found in
14499 OpenSSL_add_all_digests()) to SSL_library_init()
14500 aka OpenSSL_add_ssl_algorithms(). This provides improved
14501 compatibility with peers using X.509 certificates
14502 with unconventional AlgorithmIdentifier OIDs.
14506 * Fix for Irix with NO_ASM.
14508 *"Bruce W. Forsberg" <bruce.forsberg@baesystems.com>*
14510 * ./config script fixes.
14512 *Ulf Moeller, Richard Levitte*
14514 * Fix 'openssl passwd -1'.
14518 * Change PKCS12_key_gen_asc() so it can cope with non null
14519 terminated strings whose length is passed in the passlen
14520 parameter, for example from PEM callbacks. This was done
14521 by adding an extra length parameter to asc2uni().
14523 *Steve Henson, reported by <oddissey@samsung.co.kr>*
14525 * Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
14526 call failed, free the DSA structure.
14530 * Fix to uni2asc() to cope with zero length Unicode strings.
14531 These are present in some PKCS#12 files.
14535 * Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
14536 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
14537 when writing a 32767 byte record.
14539 *Bodo Moeller; problem reported by Eric Day <eday@concentric.net>*
14541 * In `RSA_eay_public_{en,ed}crypt` and RSA_eay_mod_exp (rsa_eay.c),
14542 obtain lock CRYPTO_LOCK_RSA before setting `rsa->_method_mod_{n,p,q}`.
14544 (RSA objects have a reference count access to which is protected
14545 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
14546 so they are meant to be shared between threads.)
14547 *Bodo Moeller, Geoff Thorpe; original patch submitted by
14548 "Reddie, Steven" <Steven.Reddie@ca.com>*
14550 * Fix a deadlock in CRYPTO_mem_leaks().
14554 * Use better test patterns in bntest.
14558 * rand_win.c fix for Borland C.
14562 * BN_rshift bugfix for n == 0.
14566 * Add a 'bctest' script that checks for some known 'bc' bugs
14567 so that 'make test' does not abort just because 'bc' is broken.
14571 * Store verify_result within SSL_SESSION also for client side to
14572 avoid potential security hole. (Re-used sessions on the client side
14573 always resulted in verify_result==X509_V_OK, not using the original
14574 result of the server certificate verification.)
14578 * Fix ssl3_pending: If the record in s->s3->rrec is not of type
14579 SSL3_RT_APPLICATION_DATA, return 0.
14580 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
14585 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
14586 releases, have been re-implemented by renaming the previous
14587 implementations of ssl2_read and ssl3_read to ssl2_read_internal
14588 and ssl3_read_internal, respectively, and adding 'peek' parameters
14589 to them. The new ssl[23]_{read,peek} functions are calls to
14590 ssl[23]_read_internal with the 'peek' flag set appropriately.
14591 A 'peek' parameter has also been added to ssl3_read_bytes, which
14592 does the actual work for ssl3_read_internal.
14596 * Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
14597 the method-specific "init()" handler. Also clean up ex_data after
14598 calling the method-specific "finish()" handler. Previously, this was
14599 happening the other way round.
14603 * Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
14604 The previous value, 12, was not always sufficient for BN_mod_exp().
14608 * Make sure that shared libraries get the internal name engine with
14609 the full version number and not just 0. This should mark the
14610 shared libraries as not backward compatible. Of course, this should
14611 be changed again when we can guarantee backward binary compatibility.
14615 * Fix typo in get_cert_by_subject() in by_dir.c
14617 *Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>*
14619 * Rework the system to generate shared libraries:
14621 - Make note of the expected extension for the shared libraries and
14622 if there is a need for symbolic links from for example libcrypto.so.0
14623 to libcrypto.so.0.9.7. There is extended info in Configure for
14626 - Make as few rebuilds of the shared libraries as possible.
14628 - Still avoid linking the OpenSSL programs with the shared libraries.
14630 - When installing, install the shared libraries separately from the
14635 * Fix SSL_CTX_set_read_ahead macro to actually use its argument.
14637 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
14638 and not in SSL_clear because the latter is also used by the
14639 accept/connect functions; previously, the settings made by
14640 SSL_set_read_ahead would be lost during the handshake.
14642 *Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>*
14644 * Correct util/mkdef.pl to be selective about disabled algorithms.
14645 Previously, it would create entries for disabled algorithms no
14650 * Added several new manual pages for SSL_* function.
14654 ### Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
14656 * In ssl23_get_client_hello, generate an error message when faced
14657 with an initial SSL 3.0/TLS record that is too small to contain the
14658 first two bytes of the ClientHello message, i.e. client_version.
14659 (Note that this is a pathologic case that probably has never happened
14660 in real life.) The previous approach was to use the version number
14661 from the record header as a substitute; but our protocol choice
14662 should not depend on that one because it is not authenticated
14663 by the Finished messages.
14667 * More robust randomness gathering functions for Windows.
14669 *Jeffrey Altman <jaltman@columbia.edu>*
14671 * For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
14672 not set then we don't setup the error code for issuer check errors
14673 to avoid possibly overwriting other errors which the callback does
14674 handle. If an application does set the flag then we assume it knows
14675 what it is doing and can handle the new informational codes
14680 * Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
14681 a general "ANY" type, as such it should be able to decode anything
14682 including tagged types. However it didn't check the class so it would
14683 wrongly interpret tagged types in the same way as their universal
14684 counterpart and unknown types were just rejected. Changed so that the
14685 tagged and unknown types are handled in the same way as a SEQUENCE:
14686 that is the encoding is stored intact. There is also a new type
14687 "V_ASN1_OTHER" which is used when the class is not universal, in this
14688 case we have no idea what the actual type is so we just lump them all
14693 * On VMS, stdout may very well lead to a file that is written to
14694 in a record-oriented fashion. That means that every write() will
14695 write a separate record, which will be read separately by the
14696 programs trying to read from it. This can be very confusing.
14698 The solution is to put a BIO filter in the way that will buffer
14699 text until a linefeed is reached, and then write everything a
14700 line at a time, so every record written will be an actual line,
14701 not chunks of lines and not (usually doesn't happen, but I've
14702 seen it once) several lines in one record. BIO_f_linebuffer() is
14705 Currently, it's a VMS-only method, because that's where it has
14706 been tested well enough.
14710 * Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
14711 it can return incorrect results.
14712 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
14713 but it was in 0.9.6-beta[12].)
14717 * Disable the check for content being present when verifying detached
14718 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
14719 include zero length content when signing messages.
14723 * New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
14724 BIO_ctrl (for BIO pairs).
14728 * Add DSO method for VMS.
14732 * Bug fix: Montgomery multiplication could produce results with the
14737 * Add RPM specification openssl.spec and modify it to build three
14738 packages. The default package contains applications, application
14739 documentation and run-time libraries. The devel package contains
14740 include files, static libraries and function documentation. The
14741 doc package contains the contents of the doc directory. The original
14742 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
14746 * Add a large number of documentation files for many SSL routines.
14748 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>*
14750 * Add a configuration entry for Sony News 4.
14752 *NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>*
14754 * Don't set the two most significant bits to one when generating a
14755 random number < q in the DSA library.
14759 * New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
14760 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
14761 the underlying transport is blocking) if a handshake took place.
14762 (The default behaviour is needed by applications such as s_client
14763 and s_server that use select() to determine when to use SSL_read;
14764 but for applications that know in advance when to expect data, it
14765 just makes things more complicated.)
14769 * Add RAND_egd_bytes(), which gives control over the number of bytes read
14774 * Add a few more EBCDIC conditionals that make `req` and `x509`
14775 work better on such systems.
14777 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
14779 * Add two demo programs for PKCS12_parse() and PKCS12_create().
14780 Update PKCS12_parse() so it copies the friendlyName and the
14781 keyid to the certificates aux info.
14785 * Fix bug in PKCS7_verify() which caused an infinite loop
14786 if there was more than one signature.
14788 *Sven Uszpelkat <su@celocom.de>*
14790 * Major change in util/mkdef.pl to include extra information
14791 about each symbol, as well as presenting variables as well
14792 as functions. This change means that there's n more need
14793 to rebuild the .num files when some algorithms are excluded.
14797 * Allow the verify time to be set by an application,
14798 rather than always using the current time.
14802 * Phase 2 verify code reorganisation. The certificate
14803 verify code now looks up an issuer certificate by a
14804 number of criteria: subject name, authority key id
14805 and key usage. It also verifies self signed certificates
14806 by the same criteria. The main comparison function is
14807 X509_check_issued() which performs these checks.
14809 Lot of changes were necessary in order to support this
14810 without completely rewriting the lookup code.
14812 Authority and subject key identifier are now cached.
14814 The LHASH 'certs' is X509_STORE has now been replaced
14815 by a STACK_OF(X509_OBJECT). This is mainly because an
14816 LHASH can't store or retrieve multiple objects with
14817 the same hash value.
14819 As a result various functions (which were all internal
14820 use only) have changed to handle the new X509_STORE
14821 structure. This will break anything that messed round
14822 with X509_STORE internally.
14824 The functions X509_STORE_add_cert() now checks for an
14825 exact match, rather than just subject name.
14827 The X509_STORE API doesn't directly support the retrieval
14828 of multiple certificates matching a given criteria, however
14829 this can be worked round by performing a lookup first
14830 (which will fill the cache with candidate certificates)
14831 and then examining the cache for matches. This is probably
14832 the best we can do without throwing out X509_LOOKUP
14833 entirely (maybe later...).
14835 The X509_VERIFY_CTX structure has been enhanced considerably.
14837 All certificate lookup operations now go via a get_issuer()
14838 callback. Although this currently uses an X509_STORE it
14839 can be replaced by custom lookups. This is a simple way
14840 to bypass the X509_STORE hackery necessary to make this
14841 work and makes it possible to use more efficient techniques
14842 in future. A very simple version which uses a simple
14843 STACK for its trusted certificate store is also provided
14844 using X509_STORE_CTX_trusted_stack().
14846 The verify_cb() and verify() callbacks now have equivalents
14847 in the X509_STORE_CTX structure.
14849 X509_STORE_CTX also has a 'flags' field which can be used
14850 to customise the verify behaviour.
14854 * Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
14855 excludes S/MIME capabilities.
14859 * When a certificate request is read in keep a copy of the
14860 original encoding of the signed data and use it when outputting
14861 again. Signatures then use the original encoding rather than
14862 a decoded, encoded version which may cause problems if the
14863 request is improperly encoded.
14867 * For consistency with other BIO_puts implementations, call
14868 buffer_write(b, ...) directly in buffer_puts instead of calling
14871 In BIO_puts, increment b->num_write as in BIO_write.
14873 *Peter.Sylvester@EdelWeb.fr*
14875 * Fix BN_mul_word for the case where the word is 0. (We have to use
14876 BN_zero, we may not return a BIGNUM with an array consisting of
14877 words set to zero.)
14881 * Avoid calling abort() from within the library when problems are
14882 detected, except if preprocessor symbols have been defined
14883 (such as REF_CHECK, BN_DEBUG etc.).
14887 * New openssl application 'rsautl'. This utility can be
14888 used for low-level RSA operations. DER public key
14889 BIO/fp routines also added.
14893 * New Configure entry and patches for compiling on QNX 4.
14895 *Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>*
14897 * A demo state-machine implementation was sponsored by
14898 Nuron (<http://www.nuron.com/>) and is now available in
14899 demos/state_machine.
14903 * New options added to the 'dgst' utility for signature
14904 generation and verification.
14908 * Unrecognized PKCS#7 content types are now handled via a
14909 catch all ASN1_TYPE structure. This allows unsupported
14910 types to be stored as a "blob" and an application can
14911 encode and decode it manually.
14915 * Fix various signed/unsigned issues to make a_strex.c
14916 compile under VC++.
14918 *Oscar Jacobsson <oscar.jacobsson@celocom.com>*
14920 * ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
14921 length if passed a buffer. ASN1_INTEGER_to_BN failed
14922 if passed a NULL BN and its argument was negative.
14924 *Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>*
14926 * Modification to PKCS#7 encoding routines to output definite
14927 length encoding. Since currently the whole structures are in
14928 memory there's not real point in using indefinite length
14929 constructed encoding. However if OpenSSL is compiled with
14930 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
14934 * Added BIO_vprintf() and BIO_vsnprintf().
14938 * Added more prefixes to parse for in the strings written
14939 through a logging bio, to cover all the levels that are available
14940 through syslog. The prefixes are now:
14942 PANIC, EMERG, EMR => LOG_EMERG
14943 ALERT, ALR => LOG_ALERT
14944 CRIT, CRI => LOG_CRIT
14945 ERROR, ERR => LOG_ERR
14946 WARNING, WARN, WAR => LOG_WARNING
14947 NOTICE, NOTE, NOT => LOG_NOTICE
14948 INFO, INF => LOG_INFO
14949 DEBUG, DBG => LOG_DEBUG
14951 and as before, if none of those prefixes are present at the
14952 beginning of the string, LOG_ERR is chosen.
14954 On Win32, the `LOG_*` levels are mapped according to this:
14956 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
14957 LOG_WARNING => EVENTLOG_WARNING_TYPE
14958 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
14962 * Made it possible to reconfigure with just the configuration
14963 argument "reconf" or "reconfigure". The command line arguments
14964 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
14965 and are retrieved from there when reconfiguring.
14971 *Assar Westerlund <assar@sics.se>, Richard Levitte*
14973 * Add the arguments -CAfile and -CApath to the pkcs12 utility.
14977 * The obj_dat.pl script was messing up the sorting of object
14978 names. The reason was that it compared the quoted version
14979 of strings as a result "OCSP" > "OCSP Signing" because
14980 " > SPACE. Changed script to store unquoted versions of
14981 names and add quotes on output. It was also omitting some
14982 names from the lookup table if they were given a default
14983 value (that is if SN is missing it is given the same
14984 value as LN and vice versa), these are now added on the
14985 grounds that if an object has a name we should be able to
14986 look it up. Finally added warning output when duplicate
14987 short or long names are found.
14991 * Changes needed for Tandem NSK.
14993 *Scott Uroff <scott@xypro.com>*
14995 * Fix SSL 2.0 rollback checking: Due to an off-by-one error in
14996 RSA_padding_check_SSLv23(), special padding was never detected
14997 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
14998 version rollback attacks was not effective.
15000 In s23_clnt.c, don't use special rollback-attack detection padding
15001 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
15002 client; similarly, in s23_srvr.c, don't do the rollback check if
15003 SSL 2.0 is the only protocol enabled in the server.
15007 * Make it possible to get hexdumps of unprintable data with 'openssl
15008 asn1parse'. By implication, the functions ASN1_parse_dump() and
15009 BIO_dump_indent() are added.
15013 * New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
15014 these print out strings and name structures based on various
15015 flags including RFC2253 support and proper handling of
15016 multibyte characters. Added options to the 'x509' utility
15017 to allow the various flags to be set.
15021 * Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
15022 Also change the functions X509_cmp_current_time() and
15023 X509_gmtime_adj() work with an ASN1_TIME structure,
15024 this will enable certificates using GeneralizedTime in validity
15025 dates to be checked.
15029 * Make the NEG_PUBKEY_BUG code (which tolerates invalid
15030 negative public key encodings) on by default,
15031 NO_NEG_PUBKEY_BUG can be set to disable it.
15035 * New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
15036 content octets. An i2c_ASN1_OBJECT is unnecessary because
15037 the encoding can be trivially obtained from the structure.
15041 * crypto/err.c locking bugfix: Use write locks (`CRYPTO_w_[un]lock`),
15042 not read locks (`CRYPTO_r_[un]lock`).
15046 * A first attempt at creating official support for shared
15047 libraries through configuration. I've kept it so the
15048 default is static libraries only, and the OpenSSL programs
15049 are always statically linked for now, but there are
15050 preparations for dynamic linking in place.
15051 This has been tested on Linux and Tru64.
15055 * Randomness polling function for Win9x, as described in:
15056 Peter Gutmann, Software Generation of Practically Strong
15061 * Fix so PRNG is seeded in req if using an already existing
15066 * New options to smime application. -inform and -outform
15067 allow alternative formats for the S/MIME message including
15068 PEM and DER. The -content option allows the content to be
15069 specified separately. This should allow things like Netscape
15070 form signing output easier to verify.
15074 * Fix the ASN1 encoding of tags using the 'long form'.
15078 * New ASN1 functions, `i2c_*` and `c2i_*` for INTEGER and BIT
15079 STRING types. These convert content octets to and from the
15080 underlying type. The actual tag and length octets are
15081 already assumed to have been read in and checked. These
15082 are needed because all other string types have virtually
15083 identical handling apart from the tag. By having versions
15084 of the ASN1 functions that just operate on content octets
15085 IMPLICIT tagging can be handled properly. It also allows
15086 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
15087 and ASN1_INTEGER are identical apart from the tag.
15091 * Change the handling of OID objects as follows:
15093 - New object identifiers are inserted in objects.txt, following
15094 the syntax given in [crypto/objects/README.md](crypto/objects/README.md).
15095 - objects.pl is used to process obj_mac.num and create a new
15097 - obj_dat.pl is used to create a new obj_dat.h, using the data in
15100 This is currently kind of a hack, and the perl code in objects.pl
15101 isn't very elegant, but it works as I intended. The simplest way
15102 to check that it worked correctly is to look in obj_dat.h and
15103 check the array nid_objs and make sure the objects haven't moved
15104 around (this is important!). Additions are OK, as well as
15105 consistent name changes.
15109 * Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
15113 * Addition of the command line parameter '-rand file' to 'openssl req'.
15114 The given file adds to whatever has already been seeded into the
15115 random pool through the RANDFILE configuration file option or
15116 environment variable, or the default random state file.
15120 * mkstack.pl now sorts each macro group into lexical order.
15121 Previously the output order depended on the order the files
15122 appeared in the directory, resulting in needless rewriting
15127 * Patches to make OpenSSL compile under Win32 again. Mostly
15128 work arounds for the VC++ problem that it treats func() as
15129 func(void). Also stripped out the parts of mkdef.pl that
15130 added extra typesafe functions: these no longer exist.
15134 * Reorganisation of the stack code. The macros are now all
15135 collected in safestack.h . Each macro is defined in terms of
15136 a "stack macro" of the form `SKM_<name>(type, a, b)`. The
15137 DEBUG_SAFESTACK is now handled in terms of function casts,
15138 this has the advantage of retaining type safety without the
15139 use of additional functions. If DEBUG_SAFESTACK is not defined
15140 then the non typesafe macros are used instead. Also modified the
15141 mkstack.pl script to handle the new form. Needs testing to see
15142 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
15143 the default if no major problems. Similar behaviour for ASN1_SET_OF
15144 and PKCS12_STACK_OF.
15148 * When some versions of IIS use the 'NET' form of private key the
15149 key derivation algorithm is different. Normally MD5(password) is
15150 used as a 128 bit RC4 key. In the modified case
15151 MD5(MD5(password) + "SGCKEYSALT") is used instead. Added some
15152 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
15153 as the old Netscape_RSA functions except they have an additional
15154 'sgckey' parameter which uses the modified algorithm. Also added
15155 an -sgckey command line option to the rsa utility. Thanks to
15156 Adrian Peck <bertie@ncipher.com> for posting details of the modified
15157 algorithm to openssl-dev.
15161 * The evp_local.h macros were using 'c.##kname' which resulted in
15162 invalid expansion on some systems (SCO 5.0.5 for example).
15163 Corrected to 'c.kname'.
15165 *Phillip Porch <root@theporch.com>*
15167 * New X509_get1_email() and X509_REQ_get1_email() functions that return
15168 a STACK of email addresses from a certificate or request, these look
15169 in the subject name and the subject alternative name extensions and
15170 omit any duplicate addresses.
15174 * Re-implement BN_mod_exp2_mont using independent (and larger) windows.
15175 This makes DSA verification about 2 % faster.
15179 * Increase maximum window size in `BN_mod_exp_...` to 6 bits instead of 5
15180 (meaning that now 2^5 values will be precomputed, which is only 4 KB
15181 plus overhead for 1024 bit moduli).
15182 This makes exponentiations about 0.5 % faster for 1024 bit
15183 exponents (as measured by "openssl speed rsa2048").
15187 * Rename memory handling macros to avoid conflicts with other
15189 Malloc => OPENSSL_malloc
15190 Malloc_locked => OPENSSL_malloc_locked
15191 Realloc => OPENSSL_realloc
15192 Free => OPENSSL_free
15196 * New function BN_mod_exp_mont_word for small bases (roughly 15%
15197 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
15201 * CygWin32 support.
15203 *John Jarvie <jjarvie@newsguy.com>*
15205 * The type-safe stack code has been rejigged. It is now only compiled
15206 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
15207 by default all type-specific stack functions are "#define"d back to
15208 standard stack functions. This results in more streamlined output
15209 but retains the type-safety checking possibilities of the original
15214 * The STACK code has been cleaned up, and certain type declarations
15215 that didn't make a lot of sense have been brought in line. This has
15216 also involved a cleanup of sorts in safestack.h to more correctly
15217 map type-safe stack functions onto their plain stack counterparts.
15218 This work has also resulted in a variety of "const"ifications of
15219 lots of the code, especially `_cmp` operations which should normally
15220 be prototyped with "const" parameters anyway.
15224 * When generating bytes for the first time in md_rand.c, 'stir the pool'
15225 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
15226 (The PRNG state consists of two parts, the large pool 'state' and 'md',
15227 where all of 'md' is used each time the PRNG is used, but 'state'
15228 is used only indexed by a cyclic counter. As entropy may not be
15229 well distributed from the beginning, 'md' is important as a
15230 chaining variable. However, the output function chains only half
15231 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
15232 all of 'md', and seeding with STATE_SIZE dummy bytes will result
15233 in all of 'state' being rewritten, with the new values depending
15234 on virtually all of 'md'. This overcomes the 80 bit limitation.)
15238 * In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
15239 the handshake is continued after ssl_verify_cert_chain();
15240 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
15241 can lead to 'unexplainable' connection aborts later.
15243 *Bodo Moeller; problem tracked down by Lutz Jaenicke*
15245 * Major EVP API cipher revision.
15246 Add hooks for extra EVP features. This allows various cipher
15247 parameters to be set in the EVP interface. Support added for variable
15248 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
15249 setting of RC2 and RC5 parameters.
15251 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
15254 Remove lots of duplicated code from the EVP library. For example *every*
15255 cipher init() function handles the 'iv' in the same way according to the
15256 cipher mode. They also all do nothing if the 'key' parameter is NULL and
15257 for CFB and OFB modes they zero ctx->num.
15259 New functionality allows removal of S/MIME code RC2 hack.
15261 Most of the routines have the same form and so can be declared in terms
15264 By shifting this to the top level EVP_CipherInit() it can be removed from
15265 all individual ciphers. If the cipher wants to handle IVs or keys
15266 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
15269 Change lots of functions like EVP_EncryptUpdate() to now return a
15270 value: although software versions of the algorithms cannot fail
15271 any installed hardware versions can.
15275 * Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
15276 this option is set, tolerate broken clients that send the negotiated
15277 protocol version number instead of the requested protocol version
15282 * Call dh_tmp_cb (set by `..._TMP_DH_CB`) with correct 'is_export' flag;
15283 i.e. non-zero for export ciphersuites, zero otherwise.
15284 Previous versions had this flag inverted, inconsistent with
15285 rsa_tmp_cb (..._TMP_RSA_CB).
15287 *Bodo Moeller; problem reported by Amit Chopra*
15289 * Add missing DSA library text string. Work around for some IIS
15290 key files with invalid SEQUENCE encoding.
15294 * Add a document (doc/standards.txt) that list all kinds of standards
15295 and so on that are implemented in OpenSSL.
15299 * Enhance c_rehash script. Old version would mishandle certificates
15300 with the same subject name hash and wouldn't handle CRLs at all.
15301 Added -fingerprint option to crl utility, to support new c_rehash
15306 * Eliminate non-ANSI declarations in crypto.h and stack.h.
15310 * Fix for SSL server purpose checking. Server checking was
15311 rejecting certificates which had extended key usage present
15312 but no ssl client purpose.
15314 *Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>*
15316 * Make PKCS#12 code work with no password. The PKCS#12 spec
15317 is a little unclear about how a blank password is handled.
15318 Since the password in encoded as a BMPString with terminating
15319 double NULL a zero length password would end up as just the
15320 double NULL. However no password at all is different and is
15321 handled differently in the PKCS#12 key generation code. NS
15322 treats a blank password as zero length. MSIE treats it as no
15323 password on export: but it will try both on import. We now do
15324 the same: PKCS12_parse() tries zero length and no password if
15325 the password is set to "" or NULL (NULL is now a valid password:
15326 it wasn't before) as does the pkcs12 application.
15330 * Bugfixes in `apps/x509.c`: Avoid a memory leak; and don't use
15331 perror when PEM_read_bio_X509_REQ fails, the error message must
15332 be obtained from the error queue.
15336 * Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
15337 it in ERR_remove_state if appropriate, and change ERR_get_state
15338 accordingly to avoid race conditions (this is necessary because
15339 thread_hash is no longer constant once set).
15343 * Bugfix for linux-elf makefile.one.
15347 * RSA_get_default_method() will now cause a default
15348 RSA_METHOD to be chosen if one doesn't exist already.
15349 Previously this was only set during a call to RSA_new()
15350 or RSA_new_method(NULL) meaning it was possible for
15351 RSA_get_default_method() to return NULL.
15355 * Added native name translation to the existing DSO code
15356 that will convert (if the flag to do so is set) filenames
15357 that are sufficiently small and have no path information
15358 into a canonical native form. Eg. "blah" converted to
15359 "libblah.so" or "blah.dll" etc.
15363 * New function ERR_error_string_n(e, buf, len) which is like
15364 ERR_error_string(e, buf), but writes at most 'len' bytes
15365 including the 0 terminator. For ERR_error_string_n, 'buf'
15368 *Damien Miller <djm@mindrot.org>, Bodo Moeller*
15370 * CONF library reworked to become more general. A new CONF
15371 configuration file reader "class" is implemented as well as a
15372 new functions (`NCONF_*`, for "New CONF") to handle it. The now
15373 old `CONF_*` functions are still there, but are reimplemented to
15374 work in terms of the new functions. Also, a set of functions
15375 to handle the internal storage of the configuration data is
15376 provided to make it easier to write new configuration file
15377 reader "classes" (I can definitely see something reading a
15378 configuration file in XML format, for example), called `_CONF_*`,
15379 or "the configuration storage API"...
15381 The new configuration file reading functions are:
15383 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
15384 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
15386 NCONF_default, NCONF_WIN32
15388 NCONF_dump_fp, NCONF_dump_bio
15390 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
15391 NCONF_new creates a new CONF object. This works in the same way
15392 as other interfaces in OpenSSL, like the BIO interface.
15393 `NCONF_dump_*` dump the internal storage of the configuration file,
15394 which is useful for debugging. All other functions take the same
15395 arguments as the old `CONF_*` functions with the exception of the
15396 first that must be a `CONF *` instead of a `LHASH *`.
15398 To make it easier to use the new classes with the old `CONF_*` functions,
15399 the function CONF_set_default_method is provided.
15403 * Add '-tls1' option to 'openssl ciphers', which was already
15404 mentioned in the documentation but had not been implemented.
15405 (This option is not yet really useful because even the additional
15406 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
15410 * Initial DSO code added into libcrypto for letting OpenSSL (and
15411 OpenSSL-based applications) load shared libraries and bind to
15412 them in a portable way.
15414 *Geoff Thorpe, with contributions from Richard Levitte*
15416 ### Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
15418 * Make sure _lrotl and _lrotr are only used with MSVC.
15420 * Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
15421 (the default implementation of RAND_status).
15423 * Rename openssl x509 option '-crlext', which was added in 0.9.5,
15424 to '-clrext' (= clear extensions), as intended and documented.
15425 *Bodo Moeller; inconsistency pointed out by Michael Attili
15426 <attili@amaxo.com>*
15428 * Fix for HMAC. It wasn't zeroing the rest of the block if the key length
15429 was larger than the MD block size.
15431 *Steve Henson, pointed out by Yost William <YostW@tce.com>*
15433 * Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
15434 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
15435 using the passed key: if the passed key was a private key the result
15436 of X509_print(), for example, would be to print out all the private key
15441 * des_quad_cksum() byte order bug fix.
15442 *Ulf Möller, using the problem description in krb4-0.9.7, where
15443 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>*
15445 * Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
15448 *Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>*
15450 * For easily testing in shell scripts whether some command
15451 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
15452 returns with exit code 0 iff no command of the given name is available.
15453 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
15454 the output goes to stdout and nothing is printed to stderr.
15455 Additional arguments are always ignored.
15457 Since for each cipher there is a command of the same name,
15458 the 'no-cipher' compilation switches can be tested this way.
15460 ('openssl no-XXX' is not able to detect pseudo-commands such
15461 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
15465 * Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
15469 * For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
15470 is set; it will be thrown away anyway because each handshake creates
15472 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
15473 to parameters -- in previous versions (since OpenSSL 0.9.3) the
15474 'default key' from SSL_CTX_set_tmp_dh would always be lost, meaning
15475 you effectively got SSL_OP_SINGLE_DH_USE when using this macro.
15479 * New s_client option -ign_eof: EOF at stdin is ignored, and
15480 'Q' and 'R' lose their special meanings (quit/renegotiate).
15481 This is part of what -quiet does; unlike -quiet, -ign_eof
15482 does not suppress any output.
15486 * Add compatibility options to the purpose and trust code. The
15487 purpose X509_PURPOSE_ANY is "any purpose" which automatically
15488 accepts a certificate or CA, this was the previous behaviour,
15489 with all the associated security issues.
15491 X509_TRUST_COMPAT is the old trust behaviour: only and
15492 automatically trust self signed roots in certificate store. A
15493 new trust setting X509_TRUST_DEFAULT is used to specify that
15494 a purpose has no associated trust setting and it should instead
15495 use the value in the default purpose.
15499 * Fix the PKCS#8 DSA private key code so it decodes keys again
15500 and fix a memory leak.
15504 * In util/mkerr.pl (which implements 'make errors'), preserve
15505 reason strings from the previous version of the .c file, as
15506 the default to have only downcase letters (and digits) in
15507 automatically generated reasons codes is not always appropriate.
15511 * In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
15512 using strerror. Previously, ERR_reason_error_string() returned
15513 library names as reason strings for SYSerr; but SYSerr is a special
15514 case where small numbers are errno values, not library numbers.
15518 * Add '-dsaparam' option to 'openssl dhparam' application. This
15519 converts DSA parameters into DH parameters. (When creating parameters,
15520 DSA_generate_parameters is used.)
15524 * Include 'length' (recommended exponent length) in C code generated
15525 by 'openssl dhparam -C'.
15529 * The second argument to set_label in perlasm was already being used
15530 so couldn't be used as a "file scope" flag. Moved to third argument
15535 * In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
15536 instead of RAND_bytes for encryption IVs and salts.
15540 * Include RAND_status() into RAND_METHOD instead of implementing
15541 it only for md_rand.c Otherwise replacing the PRNG by calling
15542 RAND_set_rand_method would be impossible.
15546 * Don't let DSA_generate_key() enter an infinite loop if the random
15547 number generation fails.
15551 * New 'rand' application for creating pseudo-random output.
15555 * Added configuration support for Linux/IA64
15557 *Rolf Haberrecker <rolf@suse.de>*
15559 * Assembler module support for Mingw32.
15563 * Shared library support for HPUX (in shlib/).
15565 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous*
15567 * Shared library support for Solaris gcc.
15569 *Lutz Behnke <behnke@trustcenter.de>*
15571 ### Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
15573 * PKCS7_encrypt() was adding text MIME headers twice because they
15574 were added manually and by SMIME_crlf_copy().
15578 * In bntest.c don't call BN_rand with zero bits argument.
15580 *Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>*
15582 * BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
15583 case was implemented. This caused BN_div_recp() to fail occasionally.
15587 * Add an optional second argument to the set_label() in the perl
15588 assembly language builder. If this argument exists and is set
15589 to 1 it signals that the assembler should use a symbol whose
15590 scope is the entire file, not just the current function. This
15591 is needed with MASM which uses the format label:: for this scope.
15593 *Steve Henson, pointed out by Peter Runestig <peter@runestig.com>*
15595 * Change the ASN1 types so they are typedefs by default. Before
15596 almost all types were #define'd to ASN1_STRING which was causing
15597 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
15602 * Change names of new functions to the new get1/get0 naming
15603 convention: After 'get1', the caller owns a reference count
15604 and has to call `..._free`; 'get0' returns a pointer to some
15605 data structure without incrementing reference counters.
15606 (Some of the existing 'get' functions increment a reference
15607 counter, some don't.)
15608 Similarly, 'set1' and 'add1' functions increase reference
15609 counters or duplicate objects.
15613 * Allow for the possibility of temp RSA key generation failure:
15614 the code used to assume it always worked and crashed on failure.
15618 * Fix potential buffer overrun problem in BIO_printf().
15619 *Ulf Möller, using public domain code by Patrick Powell; problem
15620 pointed out by David Sacerdote <das33@cornell.edu>*
15622 * Support EGD <http://www.lothar.com/tech/crypto/>. New functions
15623 RAND_egd() and RAND_status(). In the command line application,
15624 the EGD socket can be specified like a seed file using RANDFILE
15629 * Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
15630 Some CAs (e.g. Verisign) distribute certificates in this form.
15634 * Remove the SSL_ALLOW_ADH compile option and set the default cipher
15635 list to exclude them. This means that no special compilation option
15636 is needed to use anonymous DH: it just needs to be included in the
15641 * Change the EVP_MD_CTX_type macro so its meaning consistent with
15642 EVP_MD_type. The old functionality is available in a new macro called
15643 EVP_MD_md(). Change code that uses it and update docs.
15647 * `..._ctrl` functions now have corresponding `..._callback_ctrl` functions
15648 where the `void *` argument is replaced by a function pointer argument.
15649 Previously `void *` was abused to point to functions, which works on
15650 many platforms, but is not correct. As these functions are usually
15651 called by macros defined in OpenSSL header files, most source code
15652 should work without changes.
15656 * `<openssl/opensslconf.h>` (which is created by Configure) now contains
15657 sections with information on -D... compiler switches used for
15658 compiling the library so that applications can see them. To enable
15659 one of these sections, a pre-processor symbol `OPENSSL_..._DEFINES`
15660 must be defined. E.g.,
15661 #define OPENSSL_ALGORITHM_DEFINES
15662 #include <openssl/opensslconf.h>
15663 defines all pertinent `NO_<algo>` symbols, such as NO_IDEA, NO_RSA, etc.
15665 *Richard Levitte, Ulf and Bodo Möller*
15667 * Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
15672 * Change the 'other' type in certificate aux info to a STACK_OF
15673 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
15674 the required ASN1 format: arbitrary types determined by an OID.
15678 * Add some PEM_write_X509_REQ_NEW() functions and a command line
15679 argument to 'req'. This is not because the function is newer or
15680 better than others it just uses the work 'NEW' in the certificate
15681 request header lines. Some software needs this.
15685 * Reorganise password command line arguments: now passwords can be
15686 obtained from various sources. Delete the PEM_cb function and make
15687 it the default behaviour: i.e. if the callback is NULL and the
15688 usrdata argument is not NULL interpret it as a null terminated pass
15689 phrase. If usrdata and the callback are NULL then the pass phrase
15690 is prompted for as usual.
15694 * Add support for the Compaq Atalla crypto accelerator. If it is installed,
15695 the support is automatically enabled. The resulting binaries will
15696 autodetect the card and use it if present.
15698 *Ben Laurie and Compaq Inc.*
15700 * Work around for Netscape hang bug. This sends certificate request
15701 and server done in one record. Since this is perfectly legal in the
15702 SSL/TLS protocol it isn't a "bug" option and is on by default. See
15703 the bugs/SSLv3 entry for more info.
15707 * HP-UX tune-up: new unified configs, HP C compiler bug workaround.
15711 * Add -rand argument to smime and pkcs12 applications and read/write
15716 * New 'passwd' tool for crypt(3) and apr1 password hashes.
15720 * Add command line password options to the remaining applications.
15724 * Bug fix for BN_div_recp() for numerators with an even number of
15729 * More tests in bntest.c, and changed test_bn output.
15733 * ./config recognizes MacOS X now.
15737 * Bug fix for BN_div() when the first words of num and divisor are
15738 equal (it gave wrong results if `(rem=(n1-q*d0)&BN_MASK2) < d0)`.
15742 * Add support for various broken PKCS#8 formats, and command line
15743 options to produce them.
15747 * New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
15748 get temporary BIGNUMs from a BN_CTX.
15752 * Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
15757 * Change the `SSLeay_add_all_*()` functions to `OpenSSL_add_all_*()` and
15758 include a #define from the old name to the new. The original intent
15759 was that statically linked binaries could for example just call
15760 SSLeay_add_all_ciphers() to just add ciphers to the table and not
15761 link with digests. This never worked because SSLeay_add_all_digests()
15762 and SSLeay_add_all_ciphers() were in the same source file so calling
15763 one would link with the other. They are now in separate source files.
15767 * Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
15771 * Use a less unusual form of the Miller-Rabin primality test (it used
15772 a binary algorithm for exponentiation integrated into the Miller-Rabin
15773 loop, our standard modexp algorithms are faster).
15777 * Support for the EBCDIC character set completed.
15779 *Martin Kraemer <Martin.Kraemer@Mch.SNI.De>*
15781 * Source code cleanups: use const where appropriate, eliminate casts,
15782 use `void *` instead of `char *` in lhash.
15786 * Bugfix: ssl3_send_server_key_exchange was not restartable
15787 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
15788 this the server could overwrite ephemeral keys that the client
15793 * Turn DSA_is_prime into a macro that calls BN_is_prime,
15794 using 50 iterations of the Rabin-Miller test.
15796 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
15797 iterations of the Rabin-Miller test as required by the appendix
15798 to FIPS PUB 186[-1]) instead of DSA_is_prime.
15799 As BN_is_prime_fasttest includes trial division, DSA parameter
15800 generation becomes much faster.
15802 This implies a change for the callback functions in DSA_is_prime
15803 and DSA_generate_parameters: The callback function is called once
15804 for each positive witness in the Rabin-Miller test, not just
15805 occasionally in the inner loop; and the parameters to the
15806 callback function now provide an iteration count for the outer
15807 loop rather than for the current invocation of the inner loop.
15808 DSA_generate_parameters additionally can call the callback
15809 function with an 'iteration count' of -1, meaning that a
15810 candidate has passed the trial division test (when q is generated
15811 from an application-provided seed, trial division is skipped).
15815 * New function BN_is_prime_fasttest that optionally does trial
15816 division before starting the Rabin-Miller test and has
15817 an additional BN_CTX * argument (whereas BN_is_prime always
15818 has to allocate at least one BN_CTX).
15819 'callback(1, -1, cb_arg)' is called when a number has passed the
15820 trial division stage.
15824 * Fix for bug in CRL encoding. The validity dates weren't being handled
15829 * New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
15833 * New function BN_pseudo_rand().
15837 * Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
15838 bignum version of BN_from_montgomery() with the working code from
15839 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
15844 * Avoid a race condition in s2_clnt.c (function get_server_hello) that
15845 made it impossible to use the same SSL_SESSION data structure in
15846 SSL2 clients in multiple threads.
15850 * The return value of RAND_load_file() no longer counts bytes obtained
15851 by stat(). RAND_load_file(..., -1) is new and uses the complete file
15852 to seed the PRNG (previously an explicit byte count was required).
15854 *Ulf Möller, Bodo Möller*
15856 * Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
15857 used `char *` instead of `void *` and had casts all over the place.
15861 * Make BN_generate_prime() return NULL on error if ret!=NULL.
15865 * Retain source code compatibility for BN_prime_checks macro:
15866 BN_is_prime(..., BN_prime_checks, ...) now uses
15867 BN_prime_checks_for_size to determine the appropriate number of
15868 Rabin-Miller iterations.
15872 * Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
15873 DH_CHECK_P_NOT_SAFE_PRIME.
15874 (Check if this is true? OpenPGP calls them "strong".)
15878 * Merge the functionality of "dh" and "gendh" programs into a new program
15879 "dhparam". The old programs are retained for now but will handle DH keys
15880 (instead of parameters) in future.
15884 * Make the ciphers, s_server and s_client programs check the return values
15885 when a new cipher list is set.
15889 * Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
15890 ciphers. Before when the 56bit ciphers were enabled the sorting was
15893 The syntax for the cipher sorting has been extended to support sorting by
15894 cipher-strength (using the strength_bits hard coded in the tables).
15895 The new command is `@STRENGTH` (see also `doc/apps/ciphers.pod`).
15897 Fix a bug in the cipher-command parser: when supplying a cipher command
15898 string with an "undefined" symbol (neither command nor alphanumeric
15899 *A-Za-z0-9*, ssl_set_cipher_list used to hang in an endless loop. Now
15900 an error is flagged.
15902 Due to the strength-sorting extension, the code of the
15903 ssl_create_cipher_list() function was completely rearranged. I hope that
15904 the readability was also increased :-)
15906 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>*
15908 * Minor change to 'x509' utility. The -CAcreateserial option now uses 1
15909 for the first serial number and places 2 in the serial number file. This
15910 avoids problems when the root CA is created with serial number zero and
15911 the first user certificate has the same issuer name and serial number
15916 * Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
15917 the new code. Add documentation for this stuff.
15921 * Changes to X509_ATTRIBUTE utilities. These have been renamed from
15922 `X509_*()` to `X509at_*()` on the grounds that they don't handle X509
15923 structures and behave in an analogous way to the X509v3 functions:
15924 they shouldn't be called directly but wrapper functions should be used
15927 So we also now have some wrapper functions that call the X509at functions
15928 when passed certificate requests. (TO DO: similar things can be done with
15929 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
15930 things. Some of these need some d2i or i2d and print functionality
15931 because they handle more complex structures.)
15935 * Add missing #ifndefs that caused missing symbols when building libssl
15936 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
15937 NO_RSA in `ssl/s2*.c`.
15939 *Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller*
15941 * Precautions against using the PRNG uninitialized: RAND_bytes() now
15942 has a return value which indicates the quality of the random data
15943 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
15944 error queue. New function RAND_pseudo_bytes() generates output that is
15945 guaranteed to be unique but not unpredictable. RAND_add is like
15946 RAND_seed, but takes an extra argument for an entropy estimate
15947 (RAND_seed always assumes full entropy).
15951 * Do more iterations of Rabin-Miller probable prime test (specifically,
15952 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
15953 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
15954 in crypto/bn/bn_prime.c for the complete table). This guarantees a
15955 false-positive rate of at most 2^-80 for random input.
15959 * Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
15963 * New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
15964 in the 0.9.5 release), this returns the chain
15965 from an X509_CTX structure with a dup of the stack and all
15966 the X509 reference counts upped: so the stack will exist
15967 after X509_CTX_cleanup() has been called. Modify pkcs12.c
15970 Also make SSL_SESSION_print() print out the verify return
15975 * Add manpage for the pkcs12 command. Also change the default
15976 behaviour so MAC iteration counts are used unless the new
15977 -nomaciter option is used. This improves file security and
15978 only older versions of MSIE (4.0 for example) need it.
15982 * Honor the no-xxx Configure options when creating .DEF files.
15986 * Add PKCS#10 attributes to field table: challengePassword,
15987 unstructuredName and unstructuredAddress. These are taken from
15988 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
15989 international characters are used.
15991 More changes to X509_ATTRIBUTE code: allow the setting of types
15992 based on strings. Remove the 'loc' parameter when adding
15993 attributes because these will be a SET OF encoding which is sorted
15998 * Initial changes to the 'req' utility to allow request generation
15999 automation. This will allow an application to just generate a template
16000 file containing all the field values and have req construct the
16003 Initial support for X509_ATTRIBUTE handling. Stacks of these are
16004 used all over the place including certificate requests and PKCS#7
16005 structures. They are currently handled manually where necessary with
16006 some primitive wrappers for PKCS#7. The new functions behave in a
16007 manner analogous to the X509 extension functions: they allow
16008 attributes to be looked up by NID and added.
16010 Later something similar to the X509V3 code would be desirable to
16011 automatically handle the encoding, decoding and printing of the
16012 more complex types. The string types like challengePassword can
16013 be handled by the string table functions.
16015 Also modified the multi byte string table handling. Now there is
16016 a 'global mask' which masks out certain types. The table itself
16017 can use the flag STABLE_NO_MASK to ignore the mask setting: this
16018 is useful when for example there is only one permissible type
16019 (as in countryName) and using the mask might result in no valid
16024 * Clean up 'Finished' handling, and add functions SSL_get_finished and
16025 SSL_get_peer_finished to allow applications to obtain the latest
16026 Finished messages sent to the peer or expected from the peer,
16027 respectively. (SSL_get_peer_finished is usually the Finished message
16028 actually received from the peer, otherwise the protocol will be aborted.)
16030 As the Finished message are message digests of the complete handshake
16031 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
16032 be used for external authentication procedures when the authentication
16033 provided by SSL/TLS is not desired or is not enough.
16037 * Enhanced support for Alpha Linux is added. Now ./config checks if
16038 the host supports BWX extension and if Compaq C is present on the
16039 $PATH. Just exploiting of the BWX extension results in 20-30%
16040 performance kick for some algorithms, e.g. DES and RC4 to mention
16041 a couple. Compaq C in turn generates ~20% faster code for MD5 and
16046 * Add support for MS "fast SGC". This is arguably a violation of the
16047 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
16048 weak crypto and after checking the certificate is SGC a second one
16049 with strong crypto. MS SGC stops the first handshake after receiving
16050 the server certificate message and sends a second client hello. Since
16051 a server will typically do all the time consuming operations before
16052 expecting any further messages from the client (server key exchange
16053 is the most expensive) there is little difference between the two.
16055 To get OpenSSL to support MS SGC we have to permit a second client
16056 hello message after we have sent server done. In addition we have to
16057 reset the MAC if we do get this second client hello.
16061 * Add a function 'd2i_AutoPrivateKey()' this will automatically decide
16062 if a DER encoded private key is RSA or DSA traditional format. Changed
16063 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
16064 format DER encoded private key. Newer code should use PKCS#8 format which
16065 has the key type encoded in the ASN1 structure. Added DER private key
16066 support to pkcs8 application.
16070 * SSL 3/TLS 1 servers now don't request certificates when an anonymous
16071 ciphersuites has been selected (as required by the SSL 3/TLS 1
16072 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
16073 is set, we interpret this as a request to violate the specification
16074 (the worst that can happen is a handshake failure, and 'correct'
16075 behaviour would result in a handshake failure anyway).
16079 * In SSL_CTX_add_session, take into account that there might be multiple
16080 SSL_SESSION structures with the same session ID (e.g. when two threads
16081 concurrently obtain them from an external cache).
16082 The internal cache can handle only one SSL_SESSION with a given ID,
16083 so if there's a conflict, we now throw out the old one to achieve
16088 * Add OIDs for idea and blowfish in CBC mode. This will allow both
16089 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
16090 some routines that use cipher OIDs: some ciphers do not have OIDs
16091 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
16096 * Simplify the trust setting structure and code. Now we just have
16097 two sequences of OIDs for trusted and rejected settings. These will
16098 typically have values the same as the extended key usage extension
16099 and any application specific purposes.
16101 The trust checking code now has a default behaviour: it will just
16102 check for an object with the same NID as the passed id. Functions can
16103 be provided to override either the default behaviour or the behaviour
16104 for a given id. SSL client, server and email already have functions
16105 in place for compatibility: they check the NID and also return "trusted"
16106 if the certificate is self signed.
16110 * Add d2i,i2d bio/fp functions for PrivateKey: these convert the
16111 traditional format into an EVP_PKEY structure.
16115 * Add a password callback function PEM_cb() which either prompts for
16116 a password if usr_data is NULL or otherwise assumes it is a null
16117 terminated password. Allow passwords to be passed on command line
16118 environment or config files in a few more utilities.
16122 * Add a bunch of DER and PEM functions to handle PKCS#8 format private
16123 keys. Add some short names for PKCS#8 PBE algorithms and allow them
16124 to be specified on the command line for the pkcs8 and pkcs12 utilities.
16125 Update documentation.
16129 * Support for ASN1 "NULL" type. This could be handled before by using
16130 ASN1_TYPE but there wasn't any function that would try to read a NULL
16131 and produce an error if it couldn't. For compatibility we also have
16132 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
16133 don't allocate anything because they don't need to.
16137 * Initial support for MacOS is now provided. Examine INSTALL.MacOS
16140 *Andy Polyakov, Roy Woods <roy@centicsystems.ca>*
16142 * Rebuild of the memory allocation routines used by OpenSSL code and
16143 possibly others as well. The purpose is to make an interface that
16144 provide hooks so anyone can build a separate set of allocation and
16145 deallocation routines to be used by OpenSSL, for example memory
16146 pool implementations, or something else, which was previously hard
16147 since Malloc(), Realloc() and Free() were defined as macros having
16148 the values malloc, realloc and free, respectively (except for Win32
16149 compilations). The same is provided for memory debugging code.
16150 OpenSSL already comes with functionality to find memory leaks, but
16151 this gives people a chance to debug other memory problems.
16153 With these changes, a new set of functions and macros have appeared:
16155 CRYPTO_set_mem_debug_functions() [F]
16156 CRYPTO_get_mem_debug_functions() [F]
16157 CRYPTO_dbg_set_options() [F]
16158 CRYPTO_dbg_get_options() [F]
16159 CRYPTO_malloc_debug_init() [M]
16161 The memory debug functions are NULL by default, unless the library
16162 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
16163 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
16164 gives the standard debugging functions that come with OpenSSL) or
16165 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
16166 provided by the library user) must be used. When the standard
16167 debugging functions are used, CRYPTO_dbg_set_options can be used to
16168 request additional information:
16169 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
16170 the CRYPTO_MDEBUG_xxx macro when compiling the library.
16172 Also, things like CRYPTO_set_mem_functions will always give the
16173 expected result (the new set of functions is used for allocation
16174 and deallocation) at all times, regardless of platform and compiler
16177 To finish it up, some functions that were never use in any other
16178 way than through macros have a new API and new semantic:
16180 CRYPTO_dbg_malloc()
16181 CRYPTO_dbg_realloc()
16184 All macros of value have retained their old syntax.
16186 *Richard Levitte and Bodo Moeller*
16188 * Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
16189 ordering of SMIMECapabilities wasn't in "strength order" and there
16190 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
16195 * Some ASN1 types with illegal zero length encoding (INTEGER,
16196 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
16198 *Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson*
16200 * Merge in my S/MIME library for OpenSSL. This provides a simple
16201 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
16202 functionality to handle multipart/signed properly) and a utility
16203 called 'smime' to call all this stuff. This is based on code I
16204 originally wrote for Celo who have kindly allowed it to be
16205 included in OpenSSL.
16209 * Add variants des_set_key_checked and des_set_key_unchecked of
16210 des_set_key (aka des_key_sched). Global variable des_check_key
16211 decides which of these is called by des_set_key; this way
16212 des_check_key behaves as it always did, but applications and
16213 the library itself, which was buggy for des_check_key == 1,
16214 have a cleaner way to pick the version they need.
16218 * New function PKCS12_newpass() which changes the password of a
16223 * Modify X509_TRUST and X509_PURPOSE so it also uses a static and
16224 dynamic mix. In both cases the ids can be used as an index into the
16225 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
16226 functions so they accept a list of the field values and the
16227 application doesn't need to directly manipulate the X509_TRUST
16232 * Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
16237 * Modify the way the V3 extension code looks up extensions. This now
16238 works in a similar way to the object code: we have some "standard"
16239 extensions in a static table which is searched with OBJ_bsearch()
16240 and the application can add dynamic ones if needed. The file
16241 crypto/x509v3/ext_dat.h now has the info: this file needs to be
16242 updated whenever a new extension is added to the core code and kept
16243 in ext_nid order. There is a simple program 'tabtest.c' which checks
16244 this. New extensions are not added too often so this file can readily
16245 be maintained manually.
16247 There are two big advantages in doing things this way. The extensions
16248 can be looked up immediately and no longer need to be "added" using
16249 X509V3_add_standard_extensions(): this function now does nothing.
16250 Side note: I get *lots* of email saying the extension code doesn't
16251 work because people forget to call this function.
16252 Also no dynamic allocation is done unless new extensions are added:
16253 so if we don't add custom extensions there is no need to call
16254 X509V3_EXT_cleanup().
16258 * Modify enc utility's salting as follows: make salting the default. Add a
16259 magic header, so unsalted files fail gracefully instead of just decrypting
16260 to garbage. This is because not salting is a big security hole, so people
16261 should be discouraged from doing it.
16265 * Fixes and enhancements to the 'x509' utility. It allowed a message
16266 digest to be passed on the command line but it only used this
16267 parameter when signing a certificate. Modified so all relevant
16268 operations are affected by the digest parameter including the
16269 -fingerprint and -x509toreq options. Also -x509toreq choked if a
16270 DSA key was used because it didn't fix the digest.
16274 * Initial certificate chain verify code. Currently tests the untrusted
16275 certificates for consistency with the verify purpose (which is set
16276 when the X509_STORE_CTX structure is set up) and checks the pathlength.
16278 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
16279 this is because it will reject chains with invalid extensions whereas
16280 every previous version of OpenSSL and SSLeay made no checks at all.
16282 Trust code: checks the root CA for the relevant trust settings. Trust
16283 settings have an initial value consistent with the verify purpose: e.g.
16284 if the verify purpose is for SSL client use it expects the CA to be
16285 trusted for SSL client use. However the default value can be changed to
16286 permit custom trust settings: one example of this would be to only trust
16287 certificates from a specific "secure" set of CAs.
16289 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
16290 which should be used for version portability: especially since the
16291 verify structure is likely to change more often now.
16293 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
16294 to set them. If not set then assume SSL clients will verify SSL servers
16297 Two new options to the verify program: -untrusted allows a set of
16298 untrusted certificates to be passed in and -purpose which sets the
16299 intended purpose of the certificate. If a purpose is set then the
16300 new chain verify code is used to check extension consistency.
16304 * Support for the authority information access extension.
16308 * Modify RSA and DSA PEM read routines to transparently handle
16309 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
16310 public keys in a format compatible with certificate
16311 SubjectPublicKeyInfo structures. Unfortunately there were already
16312 functions called *_PublicKey_* which used various odd formats so
16313 these are retained for compatibility: however the DSA variants were
16314 never in a public release so they have been deleted. Changed dsa/rsa
16315 utilities to handle the new format: note no releases ever handled public
16316 keys so we should be OK.
16318 The primary motivation for this change is to avoid the same fiasco
16319 that dogs private keys: there are several incompatible private key
16320 formats some of which are standard and some OpenSSL specific and
16321 require various evil hacks to allow partial transparent handling and
16322 even then it doesn't work with DER formats. Given the option anything
16323 other than PKCS#8 should be dumped: but the other formats have to
16324 stay in the name of compatibility.
16326 With public keys and the benefit of hindsight one standard format
16327 is used which works with EVP_PKEY, RSA or DSA structures: though
16328 it clearly returns an error if you try to read the wrong kind of key.
16330 Added a -pubkey option to the 'x509' utility to output the public key.
16331 Also rename the `EVP_PKEY_get_*()` to `EVP_PKEY_rget_*()`
16332 (renamed to `EVP_PKEY_get1_*()` in the OpenSSL 0.9.5 release) and add
16333 `EVP_PKEY_rset_*()` functions (renamed to `EVP_PKEY_set1_*()`)
16334 that do the same as the `EVP_PKEY_assign_*()` except they up the
16335 reference count of the added key (they don't "swallow" the
16340 * Fixes to crypto/x509/by_file.c the code to read in certificates and
16341 CRLs would fail if the file contained no certificates or no CRLs:
16342 added a new function to read in both types and return the number
16343 read: this means that if none are read it will be an error. The
16344 DER versions of the certificate and CRL reader would always fail
16345 because it isn't possible to mix certificates and CRLs in DER format
16346 without choking one or the other routine. Changed this to just read
16347 a certificate: this is the best we can do. Also modified the code
16348 in `apps/verify.c` to take notice of return codes: it was previously
16349 attempting to read in certificates from NULL pointers and ignoring
16350 any errors: this is one reason why the cert and CRL reader seemed
16351 to work. It doesn't check return codes from the default certificate
16352 routines: these may well fail if the certificates aren't installed.
16356 * Code to support otherName option in GeneralName.
16360 * First update to verify code. Change the verify utility
16361 so it warns if it is passed a self signed certificate:
16362 for consistency with the normal behaviour. X509_verify
16363 has been modified to it will now verify a self signed
16364 certificate if *exactly* the same certificate appears
16365 in the store: it was previously impossible to trust a
16366 single self signed certificate. This means that:
16367 openssl verify ss.pem
16368 now gives a warning about a self signed certificate but
16369 openssl verify -CAfile ss.pem ss.pem
16374 * For servers, store verify_result in SSL_SESSION data structure
16375 (and add it to external session representation).
16376 This is needed when client certificate verifications fails,
16377 but an application-provided verification callback (set by
16378 SSL_CTX_set_cert_verify_callback) allows accepting the session
16379 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
16380 but returns 1): When the session is reused, we have to set
16381 ssl->verify_result to the appropriate error code to avoid
16384 *Bodo Moeller, problem pointed out by Lutz Jaenicke*
16386 * Fix a bug in the new PKCS#7 code: it didn't consider the
16387 case in PKCS7_dataInit() where the signed PKCS7 structure
16388 didn't contain any existing data because it was being created.
16390 *Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson*
16392 * Add a salt to the key derivation routines in enc.c. This
16393 forms the first 8 bytes of the encrypted file. Also add a
16394 -S option to allow a salt to be input on the command line.
16398 * New function X509_cmp(). Oddly enough there wasn't a function
16399 to compare two certificates. We do this by working out the SHA1
16400 hash and comparing that. X509_cmp() will be needed by the trust
16405 * SSL_get1_session() is like SSL_get_session(), but increments
16406 the reference count in the SSL_SESSION returned.
16408 *Geoff Thorpe <geoff@eu.c2.net>*
16410 * Fix for 'req': it was adding a null to request attributes.
16411 Also change the X509_LOOKUP and X509_INFO code to handle
16412 certificate auxiliary information.
16416 * Add support for 40 and 64 bit RC2 and RC4 algorithms: document
16421 * Add the possibility to add extra information to the memory leak
16422 detecting output, to form tracebacks, showing from where each
16423 allocation was originated: CRYPTO_push_info("constant string") adds
16424 the string plus current file name and line number to a per-thread
16425 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
16426 is like calling CYRPTO_pop_info() until the stack is empty.
16427 Also updated memory leak detection code to be multi-thread-safe.
16431 * Add options -text and -noout to pkcs7 utility and delete the
16432 encryption options which never did anything. Update docs.
16436 * Add options to some of the utilities to allow the pass phrase
16437 to be included on either the command line (not recommended on
16438 OSes like Unix) or read from the environment. Update the
16439 manpages and fix a few bugs.
16443 * Add a few manpages for some of the openssl commands.
16447 * Fix the -revoke option in ca. It was freeing up memory twice,
16448 leaking and not finding already revoked certificates.
16452 * Extensive changes to support certificate auxiliary information.
16453 This involves the use of X509_CERT_AUX structure and X509_AUX
16454 functions. An X509_AUX function such as PEM_read_X509_AUX()
16455 can still read in a certificate file in the usual way but it
16456 will also read in any additional "auxiliary information". By
16457 doing things this way a fair degree of compatibility can be
16458 retained: existing certificates can have this information added
16459 using the new 'x509' options.
16461 Current auxiliary information includes an "alias" and some trust
16462 settings. The trust settings will ultimately be used in enhanced
16463 certificate chain verification routines: currently a certificate
16464 can only be trusted if it is self signed and then it is trusted
16469 * Fix assembler for Alpha (tested only on DEC OSF not Linux or `*BSD`).
16470 The problem was that one of the replacement routines had not been working
16471 since SSLeay releases. For now the offending routine has been replaced
16472 with non-optimised assembler. Even so, this now gives around 95%
16473 performance improvement for 1024 bit RSA signs.
16477 * Hack to fix PKCS#7 decryption when used with some unorthodox RC2
16478 handling. Most clients have the effective key size in bits equal to
16479 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
16480 A few however don't do this and instead use the size of the decrypted key
16481 to determine the RC2 key length and the AlgorithmIdentifier to determine
16482 the effective key length. In this case the effective key length can still
16483 be 40 bits but the key length can be 168 bits for example. This is fixed
16484 by manually forcing an RC2 key into the EVP_PKEY structure because the
16485 EVP code can't currently handle unusual RC2 key sizes: it always assumes
16486 the key length and effective key length are equal.
16490 * Add a bunch of functions that should simplify the creation of
16491 X509_NAME structures. Now you should be able to do:
16492 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
16493 and have it automatically work out the correct field type and fill in
16494 the structures. The more adventurous can try:
16495 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
16496 and it will (hopefully) work out the correct multibyte encoding.
16500 * Change the 'req' utility to use the new field handling and multibyte
16501 copy routines. Before the DN field creation was handled in an ad hoc
16502 way in req, ca, and x509 which was rather broken and didn't support
16503 BMPStrings or UTF8Strings. Since some software doesn't implement
16504 BMPStrings or UTF8Strings yet, they can be enabled using the config file
16505 using the dirstring_type option. See the new comment in the default
16506 openssl.cnf for more info.
16510 * Make crypto/rand/md_rand.c more robust:
16511 - Assure unique random numbers after fork().
16512 - Make sure that concurrent threads access the global counter and
16513 md serializably so that we never lose entropy in them
16514 or use exactly the same state in multiple threads.
16515 Access to the large state is not always serializable because
16516 the additional locking could be a performance killer, and
16517 md should be large enough anyway.
16521 * New file `apps/app_rand.c` with commonly needed functionality
16522 for handling the random seed file.
16524 Use the random seed file in some applications that previously did not:
16526 dsaparam -genkey (which also ignored its '-rand' option),
16529 x509 (when signing).
16530 Except on systems with /dev/urandom, it is crucial to have a random
16531 seed file at least for key creation, DSA signing, and for DH exchanges;
16532 for RSA signatures we could do without one.
16534 gendh and gendsa (unlike genrsa) used to read only the first byte
16535 of each file listed in the '-rand' option. The function as previously
16536 found in genrsa is now in app_rand.c and is used by all programs
16537 that support '-rand'.
16541 * In RAND_write_file, use mode 0600 for creating files;
16542 don't just chmod when it may be too late.
16546 * Report an error from X509_STORE_load_locations
16547 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
16551 * New function ASN1_mbstring_copy() this copies a string in either
16552 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
16553 into an ASN1_STRING type. A mask of permissible types is passed
16554 and it chooses the "minimal" type to use or an error if not type
16559 * Add function equivalents to the various macros in asn1.h. The old
16560 macros are retained with an `M_` prefix. Code inside the library can
16561 use the `M_` macros. External code (including the openssl utility)
16562 should *NOT* in order to be "shared library friendly".
16566 * Add various functions that can check a certificate's extensions
16567 to see if it usable for various purposes such as SSL client,
16568 server or S/MIME and CAs of these types. This is currently
16569 VERY EXPERIMENTAL but will ultimately be used for certificate chain
16570 verification. Also added a -purpose flag to x509 utility to
16571 print out all the purposes.
16575 * Add a CRYPTO_EX_DATA to X509 certificate structure and associated
16580 * New `X509V3_{X509,CRL,REVOKED}_get_d2i()` functions. These will search
16581 for, obtain and decode and extension and obtain its critical flag.
16582 This allows all the necessary extension code to be handled in a
16583 single function call.
16587 * RC4 tune-up featuring 30-40% performance improvement on most RISC
16588 platforms. See crypto/rc4/rc4_enc.c for further details.
16592 * New -noout option to asn1parse. This causes no output to be produced
16593 its main use is when combined with -strparse and -out to extract data
16594 from a file (which may not be in ASN.1 format).
16598 * Fix for pkcs12 program. It was hashing an invalid certificate pointer
16599 when producing the local key id.
16601 *Richard Levitte <levitte@stacken.kth.se>*
16603 * New option -dhparam in s_server. This allows a DH parameter file to be
16604 stated explicitly. If it is not stated then it tries the first server
16605 certificate file. The previous behaviour hard coded the filename
16610 * Add -pubin and -pubout options to the rsa and dsa commands. These allow
16611 a public key to be input or output. For example:
16612 openssl rsa -in key.pem -pubout -out pubkey.pem
16613 Also added necessary DSA public key functions to handle this.
16617 * Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
16618 in the message. This was handled by allowing
16619 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
16621 *Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>*
16623 * Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
16624 to the end of the strings whereas this didn't. This would cause problems
16625 if strings read with d2i_ASN1_bytes() were later modified.
16627 *Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>*
16629 * Fix for base64 decode bug. When a base64 bio reads only one line of
16630 data and it contains EOF it will end up returning an error. This is
16631 caused by input 46 bytes long. The cause is due to the way base64
16632 BIOs find the start of base64 encoded data. They do this by trying a
16633 trial decode on each line until they find one that works. When they
16634 do a flag is set and it starts again knowing it can pass all the
16635 data directly through the decoder. Unfortunately it doesn't reset
16636 the context it uses. This means that if EOF is reached an attempt
16637 is made to pass two EOFs through the context and this causes the
16638 resulting error. This can also cause other problems as well. As is
16639 usual with these problems it takes *ages* to find and the fix is
16640 trivial: move one line.
16642 *Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer)*
16644 * Ugly workaround to get s_client and s_server working under Windows. The
16645 old code wouldn't work because it needed to select() on sockets and the
16646 tty (for keypresses and to see if data could be written). Win32 only
16647 supports select() on sockets so we select() with a 1s timeout on the
16648 sockets and then see if any characters are waiting to be read, if none
16649 are present then we retry, we also assume we can always write data to
16650 the tty. This isn't nice because the code then blocks until we've
16651 received a complete line of data and it is effectively polling the
16652 keyboard at 1s intervals: however it's quite a bit better than not
16653 working at all :-) A dedicated Windows application might handle this
16654 with an event loop for example.
16658 * Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
16659 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
16660 will be called when RSA_sign() and RSA_verify() are used. This is useful
16661 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
16662 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
16663 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
16664 This necessitated the support of an extra signature type NID_md5_sha1
16665 for SSL signatures and modifications to the SSL library to use it instead
16666 of calling RSA_public_decrypt() and RSA_private_encrypt().
16670 * Add new -verify -CAfile and -CApath options to the crl program, these
16671 will lookup a CRL issuers certificate and verify the signature in a
16672 similar way to the verify program. Tidy up the crl program so it
16673 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
16674 less strict. It will now permit CRL extensions even if it is not
16675 a V2 CRL: this will allow it to tolerate some broken CRLs.
16679 * Initialize all non-automatic variables each time one of the openssl
16680 sub-programs is started (this is necessary as they may be started
16681 multiple times from the "OpenSSL>" prompt).
16683 *Lennart Bang, Bodo Moeller*
16685 * Preliminary compilation option RSA_NULL which disables RSA crypto without
16686 removing all other RSA functionality (this is what NO_RSA does). This
16687 is so (for example) those in the US can disable those operations covered
16688 by the RSA patent while allowing storage and parsing of RSA keys and RSA
16693 * Non-copying interface to BIO pairs.
16694 (still largely untested)
16698 * New function ASN1_tag2str() to convert an ASN1 tag to a descriptive
16699 ASCII string. This was handled independently in various places before.
16703 * New functions UTF8_getc() and UTF8_putc() that parse and generate
16704 UTF8 strings a character at a time.
16708 * Use client_version from client hello to select the protocol
16709 (s23_srvr.c) and for RSA client key exchange verification
16710 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
16714 * Add various utility functions to handle SPKACs, these were previously
16715 handled by poking round in the structure internals. Added new function
16716 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
16717 print, verify and generate SPKACs. Based on an original idea from
16718 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
16722 * RIPEMD160 is operational on all platforms and is back in 'make test'.
16726 * Allow the config file extension section to be overwritten on the
16727 command line. Based on an original idea from Massimiliano Pala
16728 <madwolf@comune.modena.it>. The new option is called -extensions
16729 and can be applied to ca, req and x509. Also -reqexts to override
16730 the request extensions in req and -crlexts to override the crl extensions
16735 * Add new feature to the SPKAC handling in ca. Now you can include
16736 the same field multiple times by preceding it by "XXXX." for example:
16739 this is the same syntax as used in the req config file.
16743 * Allow certificate extensions to be added to certificate requests. These
16744 are specified in a 'req_extensions' option of the req section of the
16745 config file. They can be printed out with the -text option to req but
16746 are otherwise ignored at present.
16750 * Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
16751 data read consists of only the final block it would not decrypted because
16752 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
16753 A misplaced 'break' also meant the decrypted final block might not be
16754 copied until the next read.
16758 * Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
16759 a few extra parameters to the DH structure: these will be useful if
16760 for example we want the value of 'q' or implement X9.42 DH.
16764 * Initial support for DSA_METHOD. This is based on the RSA_METHOD and
16765 provides hooks that allow the default DSA functions or functions on a
16766 "per key" basis to be replaced. This allows hardware acceleration and
16767 hardware key storage to be handled without major modification to the
16768 library. Also added low-level modexp hooks and CRYPTO_EX structure and
16769 associated functions.
16773 * Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
16774 as "read only": it can't be written to and the buffer it points to will
16775 not be freed. Reading from a read only BIO is much more efficient than
16776 a normal memory BIO. This was added because there are several times when
16777 an area of memory needs to be read from a BIO. The previous method was
16778 to create a memory BIO and write the data to it, this results in two
16779 copies of the data and an O(n^2) reading algorithm. There is a new
16780 function BIO_new_mem_buf() which creates a read only memory BIO from
16781 an area of memory. Also modified the PKCS#7 routines to use read only
16786 * Bugfix: ssl23_get_client_hello did not work properly when called in
16787 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
16788 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
16789 but a retry condition occurred while trying to read the rest.
16793 * The PKCS7_ENC_CONTENT_new() function was setting the content type as
16794 NID_pkcs7_encrypted by default: this was wrong since this should almost
16795 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
16796 the encrypted data type: this is a more sensible place to put it and it
16797 allows the PKCS#12 code to be tidied up that duplicated this
16802 * Changed obj_dat.pl script so it takes its input and output files on
16803 the command line. This should avoid shell escape redirection problems
16808 * Initial support for certificate extension requests, these are included
16809 in things like Xenroll certificate requests. Included functions to allow
16810 extensions to be obtained and added.
16814 * -crlf option to s_client and s_server for sending newlines as
16815 CRLF (as required by many protocols).
16819 ### Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
16821 * Install libRSAglue.a when OpenSSL is built with RSAref.
16823 *Ralf S. Engelschall*
16825 * A few more `#ifndef NO_FP_API / #endif` pairs for consistency.
16827 *Andrija Antonijevic <TheAntony2@bigfoot.com>*
16829 * Fix -startdate and -enddate (which was missing) arguments to 'ca'
16834 * New function DSA_dup_DH, which duplicates DSA parameters/keys as
16835 DH parameters/keys (q is lost during that conversion, but the resulting
16836 DH parameters contain its length).
16838 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
16839 much faster than DH_generate_parameters (which creates parameters
16840 where `p = 2*q + 1`), and also the smaller q makes DH computations
16841 much more efficient (160-bit exponentiation instead of 1024-bit
16842 exponentiation); so this provides a convenient way to support DHE
16843 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
16844 utter importance to use
16845 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
16847 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
16848 when such DH parameters are used, because otherwise small subgroup
16849 attacks may become possible!
16853 * Avoid memory leak in i2d_DHparams.
16857 * Allow the -k option to be used more than once in the enc program:
16858 this allows the same encrypted message to be read by multiple recipients.
16862 * New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
16863 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
16864 it will always use the numerical form of the OID, even if it has a short
16869 * Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
16870 method only got called if p,q,dmp1,dmq1,iqmp components were present,
16871 otherwise bn_mod_exp was called. In the case of hardware keys for example
16872 no private key components need be present and it might store extra data
16873 in the RSA structure, which cannot be accessed from bn_mod_exp.
16874 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
16875 private key operations.
16879 * Added support for SPARC Linux.
16883 * pem_password_cb function type incompatibly changed from
16884 typedef int pem_password_cb(char *buf, int size, int rwflag);
16886 ....(char *buf, int size, int rwflag, void *userdata);
16887 so that applications can pass data to their callbacks:
16888 The `PEM[_ASN1]_{read,write}...` functions and macros now take an
16889 additional void * argument, which is just handed through whenever
16890 the password callback is called.
16892 *Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller*
16894 New function SSL_CTX_set_default_passwd_cb_userdata.
16896 Compatibility note: As many C implementations push function arguments
16897 onto the stack in reverse order, the new library version is likely to
16898 interoperate with programs that have been compiled with the old
16899 pem_password_cb definition (PEM_whatever takes some data that
16900 happens to be on the stack as its last argument, and the callback
16901 just ignores this garbage); but there is no guarantee whatsoever that
16904 * The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
16905 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
16906 problems not only on Windows, but also on some Unix platforms.
16907 To avoid problematic command lines, these definitions are now in an
16908 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
16909 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
16913 * MIPS III/IV assembler module is reimplemented.
16917 * More DES library cleanups: remove references to srand/rand and
16918 delete an unused file.
16922 * Add support for the free Netwide assembler (NASM) under Win32,
16923 since not many people have MASM (ml) and it can be hard to obtain.
16924 This is currently experimental but it seems to work OK and pass all
16925 the tests. Check out INSTALL.W32 for info.
16929 * Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
16930 without temporary keys kept an extra copy of the server key,
16931 and connections with temporary keys did not free everything in case
16936 * New function RSA_check_key and new openssl rsa option -check
16937 for verifying the consistency of RSA keys.
16939 *Ulf Moeller, Bodo Moeller*
16941 * Various changes to make Win32 compile work:
16942 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
16943 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
16944 comparison" warnings.
16945 3. Add `sk_<TYPE>_sort` to DEF file generator and do make update.
16949 * Add a debugging option to PKCS#5 v2 key generation function: when
16950 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
16951 derived keys are printed to stderr.
16955 * Copy the flags in ASN1_STRING_dup().
16957 *Roman E. Pavlov <pre@mo.msk.ru>*
16959 * The x509 application mishandled signing requests containing DSA
16960 keys when the signing key was also DSA and the parameters didn't match.
16962 It was supposed to omit the parameters when they matched the signing key:
16963 the verifying software was then supposed to automatically use the CA's
16964 parameters if they were absent from the end user certificate.
16966 Omitting parameters is no longer recommended. The test was also
16967 the wrong way round! This was probably due to unusual behaviour in
16968 EVP_cmp_parameters() which returns 1 if the parameters match.
16969 This meant that parameters were omitted when they *didn't* match and
16970 the certificate was useless. Certificates signed with 'ca' didn't have
16973 *Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>*
16975 * Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
16976 The interface is as follows:
16977 Applications can use
16978 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
16979 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
16980 "off" is now the default.
16981 The library internally uses
16982 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
16983 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
16984 to disable memory-checking temporarily.
16986 Some inconsistent states that previously were possible (and were
16987 even the default) are now avoided.
16989 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
16990 with each memory chunk allocated; this is occasionally more helpful
16991 than just having a counter.
16993 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
16995 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
17000 * Introduce "mode" for SSL structures (with defaults in SSL_CTX),
17001 which largely parallels "options", but is for changing API behaviour,
17002 whereas "options" are about protocol behaviour.
17003 Initial "mode" flags are:
17005 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
17006 a single record has been written.
17007 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
17008 retries use the same buffer location.
17009 (But all of the contents must be
17014 * Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
17017 * Fix problems with no-hmac etc.
17019 *Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>*
17021 * New functions RSA_get_default_method(), RSA_set_method() and
17022 RSA_get_method(). These allows replacement of RSA_METHODs without having
17023 to mess around with the internals of an RSA structure.
17027 * Fix memory leaks in DSA_do_sign and DSA_is_prime.
17028 Also really enable memory leak checks in openssl.c and in some
17031 *Chad C. Mulligan, Bodo Moeller*
17033 * Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
17034 up the length of negative integers. This has now been simplified to just
17035 store the length when it is first determined and use it later, rather
17036 than trying to keep track of where data is copied and updating it to
17038 *Steve Henson, reported by Brien Wheeler <bwheeler@authentica-security.com>*
17040 * Add a new function PKCS7_signatureVerify. This allows the verification
17041 of a PKCS#7 signature but with the signing certificate passed to the
17042 function itself. This contrasts with PKCS7_dataVerify which assumes the
17043 certificate is present in the PKCS#7 structure. This isn't always the
17044 case: certificates can be omitted from a PKCS#7 structure and be
17045 distributed by "out of band" means (such as a certificate database).
17049 * Complete the `PEM_*` macros with DECLARE_PEM versions to replace the
17050 function prototypes in pem.h, also change util/mkdef.pl to add the
17051 necessary function names.
17055 * mk1mf.pl (used by Windows builds) did not properly read the
17056 options set by Configure in the top level Makefile, and Configure
17057 was not even able to write more than one option correctly.
17058 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
17062 * New functions CONF_load_bio() and CONF_load_fp() to allow a config
17063 file to be loaded from a BIO or FILE pointer. The BIO version will
17064 for example allow memory BIOs to contain config info.
17068 * New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
17069 Whoever hopes to achieve shared-library compatibility across versions
17070 must use this, not the compile-time macro.
17071 (Exercise 0.9.4: Which is the minimum library version required by
17073 Note: All this applies only to multi-threaded programs, others don't
17078 * Add missing case to s3_clnt.c state machine -- one of the new SSL tests
17079 through a BIO pair triggered the default case, i.e.
17080 SSLerr(...,SSL_R_UNKNOWN_STATE).
17084 * New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
17085 can use the SSL library even if none of the specific BIOs is
17090 * Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
17091 for the encoded length.
17093 *Jeon KyoungHo <khjeon@sds.samsung.co.kr>*
17095 * Add initial documentation of the X509V3 functions.
17099 * Add a new pair of functions PEM_write_PKCS8PrivateKey() and
17100 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
17101 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
17102 secure PKCS#8 private key format with a high iteration count.
17106 * Fix determination of Perl interpreter: A perl or perl5
17107 *directory* in $PATH was also accepted as the interpreter.
17109 *Ralf S. Engelschall*
17111 * Fix demos/sign/sign.c: well there wasn't anything strictly speaking
17112 wrong with it but it was very old and did things like calling
17113 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
17114 unusual formatting.
17118 * Fix demos/selfsign.c: it used obsolete and deleted functions, changed
17119 to use the new extension code.
17123 * Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
17124 with macros. This should make it easier to change their form, add extra
17125 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
17130 * Add to configuration table a new entry that can specify an alternative
17131 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
17132 according to Mark Crispin <MRC@Panda.COM>.
17136 * DES CBC did not update the IV. Weird.
17140 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
17141 Changing the behaviour of the former might break existing programs --
17142 where IV updating is needed, des_ncbc_encrypt can be used.
17145 * When bntest is run from "make test" it drives bc to check its
17146 calculations, as well as internally checking them. If an internal check
17147 fails, it needs to cause bc to give a non-zero result or make test carries
17148 on without noticing the failure. Fixed.
17152 * DES library cleanups.
17156 * Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
17157 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
17158 ciphers. NOTE: although the key derivation function has been verified
17159 against some published test vectors it has not been extensively tested
17160 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
17165 * Instead of "mkdir -p", which is not fully portable, use new
17166 Perl script "util/mkdir-p.pl".
17170 * Rewrite the way password based encryption (PBE) is handled. It used to
17171 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
17172 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
17173 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
17174 the 'parameter' field of the AlgorithmIdentifier is passed to the
17175 underlying key generation function so it must do its own ASN1 parsing.
17176 This has also changed the EVP_PBE_CipherInit() function which now has a
17177 'parameter' argument instead of literal salt and iteration count values
17178 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
17182 * Support for PKCS#5 v1.5 compatible password based encryption algorithms
17183 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
17184 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
17185 KEY" because this clashed with PKCS#8 unencrypted string. Since this
17186 value was just used as a "magic string" and not used directly its
17187 value doesn't matter.
17191 * Introduce some semblance of const correctness to BN. Shame C doesn't
17196 * "linux-sparc64" configuration (ultrapenguin).
17198 *Ray Miller <ray.miller@oucs.ox.ac.uk>*
17199 "linux-sparc" configuration.
17201 *Christian Forster <fo@hawo.stw.uni-erlangen.de>*
17203 * config now generates no-xxx options for missing ciphers.
17207 * Support the EBCDIC character set (work in progress).
17208 File ebcdic.c not yet included because it has a different license.
17210 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
17212 * Support BS2000/OSD-POSIX.
17214 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
17216 * Make callbacks for key generation use `void *` instead of `char *`.
17220 * Make S/MIME samples compile (not yet tested).
17224 * Additional typesafe stacks.
17228 * New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
17232 ### Changes between 0.9.3 and 0.9.3a [29 May 1999]
17234 * New configuration variant "sco5-gcc".
17236 * Updated some demos.
17238 *Sean O Riordain, Wade Scholine*
17240 * Add missing BIO_free at exit of pkcs12 application.
17244 * Fix memory leak in conf.c.
17248 * Updates for Win32 to assembler version of MD5.
17252 * Set #! path to perl in `apps/der_chop` to where we found it
17253 instead of using a fixed path.
17257 * SHA library changes for irix64-mips4-cc.
17261 * Improvements for VMS support.
17265 ### Changes between 0.9.2b and 0.9.3 [24 May 1999]
17267 * Bignum library bug fix. IRIX 6 passes "make test" now!
17268 This also avoids the problems with SC4.2 and unpatched SC5.
17270 *Andy Polyakov <appro@fy.chalmers.se>*
17272 * New functions sk_num, sk_value and sk_set to replace the previous macros.
17273 These are required because of the typesafe stack would otherwise break
17274 existing code. If old code used a structure member which used to be STACK
17275 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
17276 sk_num or sk_value it would produce an error because the num, data members
17277 are not present in STACK_OF. Now it just produces a warning. sk_set
17278 replaces the old method of assigning a value to sk_value
17279 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
17280 that does this will no longer work (and should use sk_set instead) but
17281 this could be regarded as a "questionable" behaviour anyway.
17285 * Fix most of the other PKCS#7 bugs. The "experimental" code can now
17286 correctly handle encrypted S/MIME data.
17290 * Change type of various DES function arguments from des_cblock
17291 (which means, in function argument declarations, pointer to char)
17292 to des_cblock * (meaning pointer to array with 8 char elements),
17293 which allows the compiler to do more typechecking; it was like
17294 that back in SSLeay, but with lots of ugly casts.
17296 Introduce new type const_des_cblock.
17300 * Reorganise the PKCS#7 library and get rid of some of the more obvious
17301 problems: find RecipientInfo structure that matches recipient certificate
17302 and initialise the ASN1 structures properly based on passed cipher.
17306 * Belatedly make the BN tests actually check the results.
17310 * Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
17311 to and from BNs: it was completely broken. New compilation option
17312 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
17313 key elements as negative integers.
17317 * Reorganize and speed up MD5.
17319 *Andy Polyakov <appro@fy.chalmers.se>*
17323 *Richard Levitte <richard@levitte.org>*
17325 * New option -out to asn1parse to allow the parsed structure to be
17326 output to a file. This is most useful when combined with the -strparse
17327 option to examine the output of things like OCTET STRINGS.
17331 * Make SSL library a little more fool-proof by not requiring any longer
17332 that `SSL_set_{accept,connect}_state` be called before
17333 `SSL_{accept,connect}` may be used (`SSL_set_..._state` is omitted
17334 in many applications because usually everything *appeared* to work as
17335 intended anyway -- now it really works as intended).
17339 * Move openssl.cnf out of lib/.
17343 * Fix various things to let OpenSSL even pass "egcc -pipe -O2 -Wall
17344 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
17345 -Wmissing-declarations -Wnested-externs -Winline" with EGCS 1.1.2+
17347 *Ralf S. Engelschall*
17349 * Various fixes to the EVP and PKCS#7 code. It may now be able to
17350 handle PKCS#7 enveloped data properly.
17352 *Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve*
17354 * Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
17355 copying pointers. The cert_st handling is changed by this in
17356 various ways (and thus what used to be known as ctx->default_cert
17357 is now called ctx->cert, since we don't resort to `s->ctx->[default_]cert`
17358 any longer when s->cert does not give us what we need).
17359 ssl_cert_instantiate becomes obsolete by this change.
17360 As soon as we've got the new code right (possibly it already is?),
17361 we have solved a couple of bugs of the earlier code where s->cert
17362 was used as if it could not have been shared with other SSL structures.
17364 Note that using the SSL API in certain dirty ways now will result
17365 in different behaviour than observed with earlier library versions:
17366 Changing settings for an `SSL_CTX *ctx` after having done s = SSL_new(ctx)
17367 does not influence s as it used to.
17369 In order to clean up things more thoroughly, inside SSL_SESSION
17370 we don't use CERT any longer, but a new structure SESS_CERT
17371 that holds per-session data (if available); currently, this is
17372 the peer's certificate chain and, for clients, the server's certificate
17373 and temporary key. CERT holds only those values that can have
17374 meaningful defaults in an SSL_CTX.
17378 * New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
17379 from the internal representation. Various PKCS#7 fixes: remove some
17380 evil casts and set the enc_dig_alg field properly based on the signing
17385 * Allow PKCS#12 password to be set from the command line or the
17386 environment. Let 'ca' get its config file name from the environment
17387 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
17392 * Allow certificate policies extension to use an IA5STRING for the
17393 organization field. This is contrary to the PKIX definition but
17394 VeriSign uses it and IE5 only recognises this form. Document 'x509'
17399 * Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
17400 without disallowing inline assembler and the like for non-pedantic builds.
17404 * Support Borland C++ builder.
17406 *Janez Jere <jj@void.si>, modified by Ulf Möller*
17412 * SHA-1 cleanups and performance enhancements.
17414 *Andy Polyakov <appro@fy.chalmers.se>*
17416 * Sparc v8plus assembler for the bignum library.
17418 *Andy Polyakov <appro@fy.chalmers.se>*
17420 * Accept any -xxx and +xxx compiler options in Configure.
17424 * Update HPUX configuration.
17428 * Add missing `sk_<type>_unshift()` function to safestack.h
17430 *Ralf S. Engelschall*
17432 * New function SSL_CTX_use_certificate_chain_file that sets the
17433 "extra_cert"s in addition to the certificate. (This makes sense
17434 only for "PEM" format files, as chains as a whole are not
17439 * Support verify_depth from the SSL API.
17440 x509_vfy.c had what can be considered an off-by-one-error:
17441 Its depth (which was not part of the external interface)
17442 was actually counting the number of certificates in a chain;
17443 now it really counts the depth.
17447 * Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
17448 instead of X509err, which often resulted in confusing error
17449 messages since the error codes are not globally unique
17450 (e.g. an alleged error in ssl3_accept when a certificate
17451 didn't match the private key).
17453 * New function SSL_CTX_set_session_id_context that allows to set a default
17454 value (so that you don't need SSL_set_session_id_context for each
17455 connection using the SSL_CTX).
17459 * OAEP decoding bug fix.
17463 * Support INSTALL_PREFIX for package builders, as proposed by
17468 * New Configure options "threads" and "no-threads". For systems
17469 where the proper compiler options are known (currently Solaris
17470 and Linux), "threads" is the default.
17474 * New script util/mklink.pl as a faster substitute for util/mklink.sh.
17478 * Install various scripts to $(OPENSSLDIR)/misc, not to
17479 $(INSTALLTOP)/bin -- they shouldn't clutter directories
17480 such as /usr/local/bin.
17484 * "make linux-shared" to build shared libraries.
17486 *Niels Poppe <niels@netbox.org>*
17488 * New Configure option `no-<cipher>` (rsa, idea, rc5, ...).
17492 * Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
17493 extension adding in x509 utility.
17497 * Remove NOPROTO sections and error code comments.
17501 * Partial rewrite of the DEF file generator to now parse the ANSI
17506 * New Configure options --prefix=DIR and --openssldir=DIR.
17510 * Complete rewrite of the error code script(s). It is all now handled
17511 by one script at the top level which handles error code gathering,
17512 header rewriting and C source file generation. It should be much better
17513 than the old method: it now uses a modified version of Ulf's parser to
17514 read the ANSI prototypes in all header files (thus the old K&R definitions
17515 aren't needed for error creation any more) and do a better job of
17516 translating function codes into names. The old 'ASN1 error code embedded
17517 in a comment' is no longer necessary and it doesn't use .err files which
17518 have now been deleted. Also the error code call doesn't have to appear all
17519 on one line (which resulted in some large lines...).
17523 * Change #include filenames from `<foo.h>` to `<openssl/foo.h>`.
17527 * Change behaviour of ssl2_read when facing length-0 packets: Don't return
17528 0 (which usually indicates a closed connection), but continue reading.
17532 * Fix some race conditions.
17536 * Add support for CRL distribution points extension. Add Certificate
17537 Policies and CRL distribution points documentation.
17541 * Move the autogenerated header file parts to crypto/opensslconf.h.
17545 * Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
17546 8 of keying material. Merlin has also confirmed interop with this fix
17547 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
17549 *Merlin Hughes <merlin@baltimore.ie>*
17551 * Fix lots of warnings.
17553 *Richard Levitte <levitte@stacken.kth.se>*
17555 * In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
17556 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
17558 *Richard Levitte <levitte@stacken.kth.se>*
17560 * Fix problems with sizeof(long) == 8.
17562 *Andy Polyakov <appro@fy.chalmers.se>*
17564 * Change functions to ANSI C.
17568 * Fix typos in error codes.
17570 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller*
17572 * Remove defunct assembler files from Configure.
17576 * SPARC v8 assembler BIGNUM implementation.
17578 *Andy Polyakov <appro@fy.chalmers.se>*
17580 * Support for Certificate Policies extension: both print and set.
17581 Various additions to support the r2i method this uses.
17585 * A lot of constification, and fix a bug in X509_NAME_oneline() that could
17586 return a const string when you are expecting an allocated buffer.
17590 * Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
17591 types DirectoryString and DisplayText.
17595 * Add code to allow r2i extensions to access the configuration database,
17596 add an LHASH database driver and add several ctx helper functions.
17600 * Fix an evil bug in bn_expand2() which caused various BN functions to
17601 fail when they extended the size of a BIGNUM.
17605 * Various utility functions to handle SXNet extension. Modify mkdef.pl to
17606 support typesafe stack.
17610 * Fix typo in SSL_[gs]et_options().
17612 *Nils Frostberg <nils@medcom.se>*
17614 * Delete various functions and files that belonged to the (now obsolete)
17615 old X509V3 handling code.
17619 * New Configure option "rsaref".
17623 * Don't auto-generate pem.h.
17627 * Introduce type-safe ASN.1 SETs.
17631 * Convert various additional casted stacks to type-safe STACK_OF() variants.
17633 *Ben Laurie, Ralf S. Engelschall, Steve Henson*
17635 * Introduce type-safe STACKs. This will almost certainly break lots of code
17636 that links with OpenSSL (well at least cause lots of warnings), but fear
17637 not: the conversion is trivial, and it eliminates loads of evil casts. A
17638 few STACKed things have been converted already. Feel free to convert more.
17639 In the fullness of time, I'll do away with the STACK type altogether.
17643 * Add `openssl ca -revoke <certfile>` facility which revokes a certificate
17644 specified in `<certfile>` by updating the entry in the index.txt file.
17645 This way one no longer has to edit the index.txt file manually for
17646 revoking a certificate. The -revoke option does the gory details now.
17648 *Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall*
17650 * Fix `openssl crl -noout -text` combination where `-noout` killed the
17651 `-text` option at all and this way the `-noout -text` combination was
17652 inconsistent in `openssl crl` with the friends in `openssl x509|rsa|dsa`.
17654 *Ralf S. Engelschall*
17656 * Make sure a corresponding plain text error message exists for the
17657 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
17658 verify callback function determined that a certificate was revoked.
17660 *Ralf S. Engelschall*
17662 * Bugfix: In test/testenc, don't test `openssl <cipher>` for
17663 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
17664 all available ciphers including rc5, which was forgotten until now.
17665 In order to let the testing shell script know which algorithms
17666 are available, a new (up to now undocumented) command
17667 `openssl list-cipher-commands` is used.
17671 * Bugfix: s_client occasionally would sleep in select() when
17672 it should have checked SSL_pending() first.
17676 * New functions DSA_do_sign and DSA_do_verify to provide access to
17677 the raw DSA values prior to ASN.1 encoding.
17681 * Tweaks to Configure
17683 *Niels Poppe <niels@netbox.org>*
17685 * Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
17690 * New variables $(RANLIB) and $(PERL) in the Makefiles.
17694 * New config option to avoid instructions that are illegal on the 80386.
17695 The default code is faster, but requires at least a 486.
17699 * Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
17700 SSL2_SERVER_VERSION (not used at all) macros, which are now the
17701 same as SSL2_VERSION anyway.
17705 * New "-showcerts" option for s_client.
17709 * Still more PKCS#12 integration. Add pkcs12 application to openssl
17710 application. Various cleanups and fixes.
17714 * More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
17715 modify error routines to work internally. Add error codes and PBE init
17716 to library startup routines.
17720 * Further PKCS#12 integration. Added password based encryption, PKCS#8 and
17721 packing functions to asn1 and evp. Changed function names and error
17722 codes along the way.
17726 * PKCS12 integration: and so it begins... First of several patches to
17727 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
17728 objects to objects.h
17732 * Add a new 'indent' option to some X509V3 extension code. Initial ASN1
17733 and display support for Thawte strong extranet extension.
17737 * Add LinuxPPC support.
17739 *Jeff Dubrule <igor@pobox.org>*
17741 * Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
17742 bn_div_words in alpha.s.
17744 *Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie*
17746 * Make sure the RSA OAEP test is skipped under -DRSAref because
17747 OAEP isn't supported when OpenSSL is built with RSAref.
17749 *Ulf Moeller <ulf@fitug.de>*
17751 * Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
17752 so they no longer are missing under -DNOPROTO.
17754 *Soren S. Jorvang <soren@t.dk>*
17756 ### Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
17758 * Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
17759 doesn't work when the session is reused. Coming soon!
17763 * Fix a security hole, that allows sessions to be reused in the wrong
17764 context thus bypassing client cert protection! All software that uses
17765 client certs and session caches in multiple contexts NEEDS PATCHING to
17766 allow session reuse! A fuller solution is in the works.
17768 *Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)*
17770 * Some more source tree cleanups (removed obsolete files
17771 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
17772 permission on "config" script to be executable) and a fix for the INSTALL
17775 *Ulf Moeller <ulf@fitug.de>*
17777 * Remove some legacy and erroneous uses of malloc, free instead of
17780 *Lennart Bang <lob@netstream.se>, with minor changes by Steve*
17782 * Make rsa_oaep_test return non-zero on error.
17784 *Ulf Moeller <ulf@fitug.de>*
17786 * Add support for native Solaris shared libraries. Configure
17787 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
17788 if someone would make that last step automatic.
17790 *Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>*
17792 * ctx_size was not built with the right compiler during "make links". Fixed.
17796 * Change the meaning of 'ALL' in the cipher list. It now means "everything
17797 except NULL ciphers". This means the default cipher list will no longer
17798 enable NULL ciphers. They need to be specifically enabled e.g. with
17799 the string "DEFAULT:eNULL".
17803 * Fix to RSA private encryption routines: if p < q then it would
17804 occasionally produce an invalid result. This will only happen with
17805 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
17809 * Be less restrictive and allow also `perl util/perlpath.pl
17810 /path/to/bin/perl` in addition to `perl util/perlpath.pl /path/to/bin`,
17811 because this way one can also use an interpreter named `perl5` (which is
17812 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
17813 installed as `perl`).
17815 *Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
17817 * Let util/clean-depend.pl work also with older Perl 5.00x versions.
17819 *Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
17821 * Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
17822 advapi32.lib to Win32 build and change the pem test comparison
17823 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
17824 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
17825 and crypto/des/ede_cbcm_enc.c.
17829 * DES quad checksum was broken on big-endian architectures. Fixed.
17833 * Comment out two functions in bio.h that aren't implemented. Fix up the
17834 Win32 test batch file so it (might) work again. The Win32 test batch file
17835 is horrible: I feel ill....
17839 * Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
17840 in e_os.h. Audit of header files to check ANSI and non ANSI
17841 sections: 10 functions were absent from non ANSI section and not exported
17842 from Windows DLLs. Fixed up libeay.num for new functions.
17846 * Make `openssl version` output lines consistent.
17848 *Ralf S. Engelschall*
17850 * Fix Win32 symbol export lists for BIO functions: Added
17851 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
17852 to ms/libeay{16,32}.def.
17854 *Ralf S. Engelschall*
17856 * Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
17857 fine under Unix and passes some trivial tests I've now added. But the
17858 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
17859 added to make sure no one expects that this stuff really works in the
17860 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
17861 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
17864 *Ralf S. Engelschall*
17866 * Fix the generation of two part addresses in perl.
17868 *Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie*
17870 * Add config entry for Linux on MIPS.
17872 *John Tobey <jtobey@channel1.com>*
17874 * Make links whenever Configure is run, unless we are on Windoze.
17878 * Permit extensions to be added to CRLs using crl_section in openssl.cnf.
17879 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
17884 * Add a useful kludge to allow package maintainers to specify compiler and
17885 other platforms details on the command line without having to patch the
17886 Configure script every time: One now can use
17887 `perl Configure <id>:<details>`,
17888 i.e. platform ids are allowed to have details appended
17889 to them (separated by colons). This is treated as there would be a static
17890 pre-configured entry in Configure's %table under key `<id>` with value
17891 `<details>` and `perl Configure <id>` is called. So, when you want to
17892 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
17893 assembler stuff you can use `perl Configure "FreeBSD-elf:pgcc:-O6:::"`
17894 now, which overrides the FreeBSD-elf entry on-the-fly.
17896 *Ralf S. Engelschall*
17898 * Disable new TLS1 ciphersuites by default: they aren't official yet.
17902 * Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
17903 on the `perl Configure ...` command line. This way one can compile
17904 OpenSSL libraries with Position Independent Code (PIC) which is needed
17905 for linking it into DSOs.
17907 *Ralf S. Engelschall*
17909 * Remarkably, export ciphers were totally broken and no-one had noticed!
17914 * Cleaned up the LICENSE document: The official contact for any license
17915 questions now is the OpenSSL core team under openssl-core@openssl.org.
17916 And add a paragraph about the dual-license situation to make sure people
17917 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
17918 to the OpenSSL toolkit.
17920 *Ralf S. Engelschall*
17922 * General source tree makefile cleanups: Made `making xxx in yyy...`
17923 display consistent in the source tree and replaced `/bin/rm` by `rm`.
17924 Additionally cleaned up the `make links` target: Remove unnecessary
17925 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
17926 to speed processing and no longer clutter the display with confusing
17927 stuff. Instead only the actually done links are displayed.
17929 *Ralf S. Engelschall*
17931 * Permit null encryption ciphersuites, used for authentication only. It used
17932 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
17933 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
17938 * Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
17939 signed attributes when verifying signatures (this would break them),
17940 the detached data encoding was wrong and public keys obtained using
17941 X509_get_pubkey() weren't freed.
17945 * Add text documentation for the BUFFER functions. Also added a work around
17946 to a Win95 console bug. This was triggered by the password read stuff: the
17947 last character typed gets carried over to the next fread(). If you were
17948 generating a new cert request using 'req' for example then the last
17949 character of the passphrase would be CR which would then enter the first
17954 * Added the new 'Includes OpenSSL Cryptography Software' button as
17955 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
17956 button and can be used by applications based on OpenSSL to show the
17957 relationship to the OpenSSL project.
17959 *Ralf S. Engelschall*
17961 * Remove confusing variables in function signatures in files
17962 ssl/ssl_lib.c and ssl/ssl.h.
17964 *Lennart Bong <lob@kulthea.stacken.kth.se>*
17966 * Don't install bss_file.c under PREFIX/include/
17968 *Lennart Bong <lob@kulthea.stacken.kth.se>*
17970 * Get the Win32 compile working again. Modify mkdef.pl so it can handle
17971 functions that return function pointers and has support for NT specific
17972 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
17973 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
17974 unsigned to signed types: this was killing the Win32 compile.
17978 * Add new certificate file to stack functions,
17979 SSL_add_dir_cert_subjects_to_stack() and
17980 SSL_add_file_cert_subjects_to_stack(). These largely supplant
17981 SSL_load_client_CA_file(), and can be used to add multiple certs easily
17982 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
17983 This means that Apache-SSL and similar packages don't have to mess around
17984 to add as many CAs as they want to the preferred list.
17988 * Experiment with doxygen documentation. Currently only partially applied to
17990 See <http://www.stack.nl/~dimitri/doxygen/index.html>, and run doxygen with
17991 openssl.doxy as the configuration file.
17995 * Get rid of remaining C++-style comments which strict C compilers hate.
17997 *Ralf S. Engelschall, pointed out by Carlos Amengual*
17999 * Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
18000 compiled in by default: it has problems with large keys.
18004 * Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
18005 DH private keys and/or callback functions which directly correspond to
18006 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
18007 is needed for applications which have to configure certificates on a
18008 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
18010 For the RSA certificate situation is makes no difference, but
18011 for the DSA certificate situation this fixes the "no shared cipher"
18012 problem where the OpenSSL cipher selection procedure failed because the
18013 temporary keys were not overtaken from the context and the API provided
18014 no way to reconfigure them.
18015 The new functions now let applications reconfigure the stuff and they
18016 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
18017 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
18018 non-public-API function ssl_cert_instantiate() is used as a helper
18019 function and also to reduce code redundancy inside ssl_rsa.c.
18021 *Ralf S. Engelschall*
18023 * Move s_server -dcert and -dkey options out of the undocumented feature
18024 area because they are useful for the DSA situation and should be
18025 recognized by the users.
18027 *Ralf S. Engelschall*
18029 * Fix the cipher decision scheme for export ciphers: the export bits are
18030 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
18031 SSL_EXP_MASK. So, the original variable has to be used instead of the
18032 already masked variable.
18034 *Richard Levitte <levitte@stacken.kth.se>*
18036 * Fix `port` variable from `int` to `unsigned int` in crypto/bio/b_sock.c
18038 *Richard Levitte <levitte@stacken.kth.se>*
18040 * Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
18041 from `int` to `unsigned int` because it is a length and initialized by
18042 EVP_DigestFinal() which expects an `unsigned int *`.
18044 *Richard Levitte <levitte@stacken.kth.se>*
18046 * Don't hard-code path to Perl interpreter on shebang line of Configure
18047 script. Instead use the usual Shell->Perl transition trick.
18049 *Ralf S. Engelschall*
18051 * Make `openssl x509 -noout -modulus`' functional also for DSA certificates
18052 (in addition to RSA certificates) to match the behaviour of `openssl dsa
18053 -noout -modulus` as it's already the case for `openssl rsa -noout
18054 -modulus`. For RSA the -modulus is the real "modulus" while for DSA
18055 currently the public key is printed (a decision which was already done by
18056 `openssl dsa -modulus` in the past) which serves a similar purpose.
18057 Additionally the NO_RSA no longer completely removes the whole -modulus
18058 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
18061 *Ralf S. Engelschall*
18063 * Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
18064 BIO. See the source (crypto/evp/bio_ok.c) for more info.
18066 *Arne Ansper <arne@ats.cyber.ee>*
18068 * Dump the old yucky req code that tried (and failed) to allow raw OIDs
18069 to be added. Now both 'req' and 'ca' can use new objects defined in the
18074 * Add cool BIO that does syslog (or event log on NT).
18076 *Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie*
18078 * Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
18079 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
18080 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
18081 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
18085 * Add preliminary config info for new extension code.
18089 * Make RSA_NO_PADDING really use no padding.
18091 *Ulf Moeller <ulf@fitug.de>*
18093 * Generate errors when private/public key check is done.
18097 * Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
18098 for some CRL extensions and new objects added.
18102 * Really fix the ASN1 IMPLICIT bug this time... Partial support for private
18103 key usage extension and fuller support for authority key id.
18107 * Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
18108 padding method for RSA, which is recommended for new applications in PKCS
18109 #1 v2.0 (RFC 2437, October 1998).
18110 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
18111 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
18112 against Bleichbacher's attack on RSA.
18113 *Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
18116 * Updates to the new SSL compression code
18118 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
18120 * Fix so that the version number in the master secret, when passed
18121 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
18122 (because the server will not accept higher), that the version number
18123 is 0x03,0x01, not 0x03,0x00
18125 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
18127 * Run extensive memory leak checks on SSL commands. Fixed *lots* of memory
18128 leaks in `ssl/` relating to new `X509_get_pubkey()` behaviour. Also fixes
18129 in `apps/` and an unrelated leak in `crypto/dsa/dsa_vrf.c`.
18133 * Support for RAW extensions where an arbitrary extension can be
18134 created by including its DER encoding. See `apps/openssl.cnf` for
18139 * Make sure latest Perl versions don't interpret some generated C array
18140 code as Perl array code in the crypto/err/err_genc.pl script.
18142 *Lars Weber <3weber@informatik.uni-hamburg.de>*
18144 * Modify ms/do_ms.bat to not generate assembly language makefiles since
18145 not many people have the assembler. Various Win32 compilation fixes and
18146 update to the INSTALL.W32 file with (hopefully) more accurate Win32
18147 build instructions.
18151 * Modify configure script 'Configure' to automatically create crypto/date.h
18152 file under Win32 and also build pem.h from pem.org. New script
18153 util/mkfiles.pl to create the MINFO file on environments that can't do a
18154 'make files': perl util/mkfiles.pl >MINFO should work.
18158 * Major rework of DES function declarations, in the pursuit of correctness
18159 and purity. As a result, many evil casts evaporated, and some weirdness,
18160 too. You may find this causes warnings in your code. Zapping your evil
18161 casts will probably fix them. Mostly.
18165 * Fix for a typo in asn1.h. Bug fix to object creation script
18166 obj_dat.pl. It considered a zero in an object definition to mean
18167 "end of object": none of the objects in objects.h have any zeros
18168 so it wasn't spotted.
18170 *Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>*
18172 * Add support for Triple DES Cipher Block Chaining with Output Feedback
18173 Masking (CBCM). In the absence of test vectors, the best I have been able
18174 to do is check that the decrypt undoes the encrypt, so far. Send me test
18175 vectors if you have them.
18179 * Correct calculation of key length for export ciphers (too much space was
18180 allocated for null ciphers). This has not been tested!
18184 * Modifications to the mkdef.pl for Win32 DEF file creation. The usage
18185 message is now correct (it understands "crypto" and "ssl" on its
18186 command line). There is also now an "update" option. This will update
18187 the util/ssleay.num and util/libeay.num files with any new functions.
18189 perl util/mkdef.pl crypto ssl update
18190 it will update them.
18194 * Overhauled the Perl interface:
18195 - ported BN stuff to OpenSSL's different BN library
18196 - made the perl/ source tree CVS-aware
18197 - renamed the package from SSLeay to OpenSSL (the files still contain
18198 their history because I've copied them in the repository)
18199 - removed obsolete files (the test scripts will be replaced
18200 by better Test::Harness variants in the future)
18202 *Ralf S. Engelschall*
18204 * First cut for a very conservative source tree cleanup:
18205 1. merge various obsolete readme texts into doc/ssleay.txt
18206 where we collect the old documents and readme texts.
18207 2. remove the first part of files where I'm already sure that we no
18208 longer need them because of three reasons: either they are just temporary
18209 files which were left by Eric or they are preserved original files where
18210 I've verified that the diff is also available in the CVS via "cvs diff
18211 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
18212 the crypto/md/ stuff).
18214 *Ralf S. Engelschall*
18216 * More extension code. Incomplete support for subject and issuer alt
18217 name, issuer and authority key id. Change the i2v function parameters
18218 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
18219 what that's for :-) Fix to ASN1 macro which messed up
18220 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
18224 * Preliminary support for ENUMERATED type. This is largely copied from the
18229 * Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
18231 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
18233 * Make sure `make rehash` target really finds the `openssl` program.
18235 *Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
18237 * Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
18238 like to hear about it if this slows down other processors.
18242 * Add CygWin32 platform information to Configure script.
18244 *Alan Batie <batie@aahz.jf.intel.com>*
18246 * Fixed ms/32all.bat script: `no_asm` -> `no-asm`
18248 *Rainer W. Gerling <gerling@mpg-gv.mpg.de>*
18250 * New program nseq to manipulate netscape certificate sequences
18254 * Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
18259 * Fixes to BN code. Previously the default was to define BN_RECURSION
18260 but the BN code had some problems that would cause failures when
18261 doing certificate verification and some other functions.
18263 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
18265 * Add ASN1 and PEM code to support netscape certificate sequences.
18269 * Add ASN1 and PEM code to support netscape certificate sequences.
18273 * Add several PKIX and private extended key usage OIDs.
18277 * Modify the 'ca' program to handle the new extension code. Modify
18278 openssl.cnf for new extension format, add comments.
18282 * More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
18283 and add a sample to openssl.cnf so req -x509 now adds appropriate
18288 * Continued X509 V3 changes. Add to other makefiles, integrate with the
18289 error code, add initial support to X509_print() and x509 application.
18293 * Takes a deep breath and start adding X509 V3 extension support code. Add
18294 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
18295 stuff is currently isolated and isn't even compiled yet.
18299 * Continuing patches for GeneralizedTime. Fix up certificate and CRL
18300 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
18301 Removed the versions check from X509 routines when loading extensions:
18302 this allows certain broken certificates that don't set the version
18303 properly to be processed.
18307 * Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
18308 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
18309 can still be regenerated with "make depend".
18313 * Spelling mistake in C version of CAST-128.
18315 *Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>*
18317 * Changes to the error generation code. The perl script err-code.pl
18318 now reads in the old error codes and retains the old numbers, only
18319 adding new ones if necessary. It also only changes the .err files if new
18320 codes are added. The makefiles have been modified to only insert errors
18321 when needed (to avoid needlessly modifying header files). This is done
18322 by only inserting errors if the .err file is newer than the auto generated
18323 C file. To rebuild all the error codes from scratch (the old behaviour)
18324 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
18325 or delete all the .err files.
18329 * CAST-128 was incorrectly implemented for short keys. The C version has
18330 been fixed, but is untested. The assembler versions are also fixed, but
18331 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
18332 to regenerate it if needed.
18333 *Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
18334 Hagino <itojun@kame.net>*
18336 * File was opened incorrectly in randfile.c.
18338 *Ulf Möller <ulf@fitug.de>*
18340 * Beginning of support for GeneralizedTime. d2i, i2d, check and print
18341 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
18342 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
18343 al: it's just almost always a UTCTime. Note this patch adds new error
18344 codes so do a "make errors" if there are problems.
18348 * Correct Linux 1 recognition in config.
18350 *Ulf Möller <ulf@fitug.de>*
18352 * Remove pointless MD5 hash when using DSA keys in ca.
18354 *Anonymous <nobody@replay.com>*
18356 * Generate an error if given an empty string as a cert directory. Also
18357 generate an error if handed NULL (previously returned 0 to indicate an
18358 error, but didn't set one).
18360 *Ben Laurie, reported by Anonymous <nobody@replay.com>*
18362 * Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
18366 * Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
18367 parameters. This was causing a warning which killed off the Win32 compile.
18371 * Remove C++ style comments from crypto/bn/bn_local.h.
18373 *Neil Costigan <neil.costigan@celocom.com>*
18375 * The function OBJ_txt2nid was broken. It was supposed to return a nid
18376 based on a text string, looking up short and long names and finally
18377 "dot" format. The "dot" format stuff didn't work. Added new function
18378 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
18379 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
18380 OID is not part of the table.
18384 * Add prototypes to X509 lookup/verify methods, fixing a bug in
18385 X509_LOOKUP_by_alias().
18389 * Sort openssl functions by name.
18393 * Get the `gendsa` command working and add it to the `list` command. Remove
18394 encryption from sample DSA keys (in case anyone is interested the password
18399 * Make *all* `*_free` functions accept a NULL pointer.
18401 *Frans Heymans <fheymans@isaserver.be>*
18403 * If a DH key is generated in s3_srvr.c, don't blow it by trying to use
18406 *Anonymous <nobody@replay.com>*
18408 * s_server should send the CAfile as acceptable CAs, not its own cert.
18410 *Bodo Moeller <3moeller@informatik.uni-hamburg.de>*
18412 * Don't blow it for numeric `-newkey` arguments to `apps/req`.
18414 *Bodo Moeller <3moeller@informatik.uni-hamburg.de>*
18416 * Temp key "for export" tests were wrong in s3_srvr.c.
18418 *Anonymous <nobody@replay.com>*
18420 * Add prototype for temp key callback functions
18421 SSL_CTX_set_tmp_{rsa,dh}_callback().
18425 * Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
18426 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
18430 * X509_name_add_entry() freed the wrong thing after an error.
18432 *Arne Ansper <arne@ats.cyber.ee>*
18434 * rsa_eay.c would attempt to free a NULL context.
18436 *Arne Ansper <arne@ats.cyber.ee>*
18438 * BIO_s_socket() had a broken should_retry() on Windoze.
18440 *Arne Ansper <arne@ats.cyber.ee>*
18442 * BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
18444 *Arne Ansper <arne@ats.cyber.ee>*
18446 * Make sure the already existing X509_STORE->depth variable is initialized
18447 in X509_STORE_new(), but document the fact that this variable is still
18448 unused in the certificate verification process.
18450 *Ralf S. Engelschall*
18452 * Fix the various library and `apps/` files to free up pkeys obtained from
18453 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
18457 * Fix reference counting in X509_PUBKEY_get(). This makes
18458 demos/maurice/example2.c work, amongst others, probably.
18460 *Steve Henson and Ben Laurie*
18462 * First cut of a cleanup for `apps/`. First the `ssleay` program is now named
18463 `openssl` and second, the shortcut symlinks for the `openssl <command>`
18464 are no longer created. This way we have a single and consistent command
18465 line interface `openssl <command>`, similar to `cvs <command>`.
18467 *Ralf S. Engelschall, Paul Sutton and Ben Laurie*
18469 * ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
18470 BIT STRING wrapper always have zero unused bits.
18474 * Add CA.pl, perl version of CA.sh, add extended key usage OID.
18478 * Make the top-level INSTALL documentation easier to understand.
18482 * Makefiles updated to exit if an error occurs in a sub-directory
18483 make (including if user presses ^C) [Paul Sutton]
18485 * Make Montgomery context stuff explicit in RSA data structure.
18489 * Fix build order of pem and err to allow for generated pem.h.
18493 * Fix renumbering bug in X509_NAME_delete_entry().
18497 * Enhanced the err-ins.pl script so it makes the error library number
18498 global and can add a library name. This is needed for external ASN1 and
18499 other error libraries.
18503 * Fixed sk_insert which never worked properly.
18507 * Fix ASN1 macros so they can handle indefinite length constructed
18508 EXPLICIT tags. Some non standard certificates use these: they can now
18513 * Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
18514 into a single doc/ssleay.txt bundle. This way the information is still
18515 preserved but no longer messes up this directory. Now it's new room for
18516 the new set of documentation files.
18518 *Ralf S. Engelschall*
18520 * SETs were incorrectly DER encoded. This was a major pain, because they
18521 shared code with SEQUENCEs, which aren't coded the same. This means that
18522 almost everything to do with SETs or SEQUENCEs has either changed name or
18523 number of arguments.
18525 *Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>*
18527 * Fix test data to work with the above.
18531 * Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
18532 was already fixed by Eric for 0.9.1 it seems.
18534 *Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>*
18536 * Autodetect FreeBSD3.
18540 * Fix various bugs in Configure. This affects the following platforms:
18544 unixware-2.0-pentium
18549 * Eliminate generated files from CVS. Reorder tests to regenerate files
18550 before they are needed.
18554 * Generate Makefile.ssl from Makefile.org (to keep CVS happy).
18558 ### Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
18560 * Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
18561 changed SSLeay to OpenSSL in version strings.
18563 *Ralf S. Engelschall*
18565 * Some fixups to the top-level documents.
18569 * Fixed the nasty bug where rsaref.h was not found under compile-time
18570 because the symlink to include/ was missing.
18572 *Ralf S. Engelschall*
18574 * Incorporated the popular no-RSA/DSA-only patches
18575 which allow to compile a RSA-free SSLeay.
18577 *Andrew Cooke / Interrader Ldt., Ralf S. Engelschall*
18579 * Fixed nasty rehash problem under `make -f Makefile.ssl links`
18580 when "ssleay" is still not found.
18582 *Ralf S. Engelschall*
18584 * Added more platforms to Configure: Cray T3E, HPUX 11,
18586 *Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>*
18588 * Updated the README file.
18590 *Ralf S. Engelschall*
18592 * Added various .cvsignore files in the CVS repository subdirs
18593 to make a "cvs update" really silent.
18595 *Ralf S. Engelschall*
18597 * Recompiled the error-definition header files and added
18598 missing symbols to the Win32 linker tables.
18600 *Ralf S. Engelschall*
18602 * Cleaned up the top-level documents;
18603 o new files: CHANGES and LICENSE
18604 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
18605 o merged COPYRIGHT into LICENSE
18606 o removed obsolete TODO file
18607 o renamed MICROSOFT to INSTALL.W32
18609 *Ralf S. Engelschall*
18611 * Removed dummy files from the 0.9.1b source tree:
18612 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
18613 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
18614 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
18615 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
18616 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
18618 *Ralf S. Engelschall*
18620 * Added various platform portability fixes.
18624 * The Genesis of the OpenSSL rpject:
18625 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
18626 Young and Tim J. Hudson created while they were working for C2Net until
18629 *The OpenSSL Project*
18631 ### Changes between 0.9.0b and 0.9.1b [not released]
18633 * Updated a few CA certificates under certs/
18637 * Changed some BIGNUM api stuff.
18641 * Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
18642 DGUX x86, Linux Alpha, etc.
18646 * New COMP library [crypto/comp/] for SSL Record Layer Compression:
18647 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
18652 * Add -strparse option to asn1pars program which parses nested
18655 *Dr Stephen Henson <shenson@bigfoot.com>*
18657 * Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
18661 * DSA fix for "ca" program.
18665 * Added "-genkey" option to "dsaparam" program.
18669 * Added RIPE MD160 (rmd160) message digest.
18673 * Added -a (all) option to "ssleay version" command.
18677 * Added PLATFORM define which is the id given to Configure.
18681 * Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
18685 * Extended the ASN.1 parser routines.
18689 * Extended BIO routines to support REUSEADDR, seek, tell, etc.
18693 * Added a BN_CTX to the BN library.
18697 * Fixed the weak key values in DES library
18701 * Changed API in EVP library for cipher aliases.
18705 * Added support for RC2/64bit cipher.
18709 * Converted the lhash library to the crypto/mem.c functions.
18713 * Added more recognized ASN.1 object ids.
18717 * Added more RSA padding checks for SSL/TLS.
18721 * Added BIO proxy/filter functionality.
18725 * Added extra_certs to SSL_CTX which can be used
18726 send extra CA certificates to the client in the CA cert chain sending
18727 process. It can be configured with SSL_CTX_add_extra_chain_cert().
18731 * Now Fortezza is denied in the authentication phase because
18732 this is key exchange mechanism is not supported by SSLeay at all.
18736 * Additional PKCS1 checks.
18740 * Support the string "TLSv1" for all TLS v1 ciphers.
18744 * Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
18745 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
18749 * Fixed a few memory leaks.
18753 * Fixed various code and comment typos.
18757 * A minor bug in ssl/s3_clnt.c where there would always be 4 0
18758 bytes sent in the client random.
18760 *Edward Bishop <ebishop@spyglass.com>*
18764 [CVE-2020-1971]: https://www.openssl.org/news/vulnerabilities.html#CVE-2020-1971
18765 [CVE-2020-1967]: https://www.openssl.org/news/vulnerabilities.html#CVE-2020-1967
18766 [CVE-2019-1563]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1563
18767 [CVE-2019-1559]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1559
18768 [CVE-2019-1552]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1552
18769 [CVE-2019-1551]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1551
18770 [CVE-2019-1549]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1549
18771 [CVE-2019-1547]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1547
18772 [CVE-2019-1543]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1543
18773 [CVE-2018-5407]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-5407
18774 [CVE-2018-0739]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0739
18775 [CVE-2018-0737]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0737
18776 [CVE-2018-0735]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0735
18777 [CVE-2018-0734]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0734
18778 [CVE-2018-0733]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0733
18779 [CVE-2018-0732]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0732
18780 [CVE-2017-3738]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3738
18781 [CVE-2017-3737]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3737
18782 [CVE-2017-3736]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3736
18783 [CVE-2017-3735]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3735
18784 [CVE-2017-3733]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3733
18785 [CVE-2017-3732]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3732
18786 [CVE-2017-3731]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3731
18787 [CVE-2017-3730]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3730
18788 [CVE-2016-7055]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7055
18789 [CVE-2016-7054]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7054
18790 [CVE-2016-7053]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7053
18791 [CVE-2016-7052]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7052
18792 [CVE-2016-6309]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6309
18793 [CVE-2016-6308]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6308
18794 [CVE-2016-6307]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6307
18795 [CVE-2016-6306]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6306
18796 [CVE-2016-6305]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6305
18797 [CVE-2016-6304]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6304
18798 [CVE-2016-6303]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6303
18799 [CVE-2016-6302]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6302
18800 [CVE-2016-2183]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2183
18801 [CVE-2016-2182]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2182
18802 [CVE-2016-2181]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2181
18803 [CVE-2016-2180]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2180
18804 [CVE-2016-2179]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2179
18805 [CVE-2016-2178]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2178
18806 [CVE-2016-2177]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2177
18807 [CVE-2016-2176]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2176
18808 [CVE-2016-2109]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2109
18809 [CVE-2016-2107]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2107
18810 [CVE-2016-2106]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2106
18811 [CVE-2016-2105]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2105
18812 [CVE-2016-0800]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0800
18813 [CVE-2016-0799]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0799
18814 [CVE-2016-0798]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0798
18815 [CVE-2016-0797]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0797
18816 [CVE-2016-0705]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0705
18817 [CVE-2016-0702]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0702
18818 [CVE-2016-0701]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0701
18819 [CVE-2015-3197]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3197
18820 [CVE-2015-3196]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3196
18821 [CVE-2015-3195]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3195
18822 [CVE-2015-3194]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3194
18823 [CVE-2015-3193]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3193
18824 [CVE-2015-1793]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1793
18825 [CVE-2015-1792]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1792
18826 [CVE-2015-1791]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1791
18827 [CVE-2015-1790]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1790
18828 [CVE-2015-1789]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1789
18829 [CVE-2015-1788]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1788
18830 [CVE-2015-1787]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1787
18831 [CVE-2015-0293]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0293
18832 [CVE-2015-0291]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0291
18833 [CVE-2015-0290]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0290
18834 [CVE-2015-0289]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0289
18835 [CVE-2015-0288]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0288
18836 [CVE-2015-0287]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0287
18837 [CVE-2015-0286]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0286
18838 [CVE-2015-0285]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0285
18839 [CVE-2015-0209]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0209
18840 [CVE-2015-0208]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0208
18841 [CVE-2015-0207]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0207
18842 [CVE-2015-0206]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0206
18843 [CVE-2015-0205]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0205
18844 [CVE-2015-0204]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0204
18845 [CVE-2014-8275]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-8275
18846 [CVE-2014-5139]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-5139
18847 [CVE-2014-3572]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3572
18848 [CVE-2014-3571]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3571
18849 [CVE-2014-3570]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3570
18850 [CVE-2014-3569]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3569
18851 [CVE-2014-3568]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3568
18852 [CVE-2014-3567]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3567
18853 [CVE-2014-3566]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3566
18854 [CVE-2014-3513]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3513
18855 [CVE-2014-3512]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3512
18856 [CVE-2014-3511]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3511
18857 [CVE-2014-3510]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3510
18858 [CVE-2014-3509]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3509
18859 [CVE-2014-3508]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3508
18860 [CVE-2014-3507]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3507
18861 [CVE-2014-3506]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3506
18862 [CVE-2014-3505]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3505
18863 [CVE-2014-3470]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3470
18864 [CVE-2014-0224]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0224
18865 [CVE-2014-0221]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0221
18866 [CVE-2014-0195]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0195
18867 [CVE-2014-0160]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0160
18868 [CVE-2014-0076]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0076
18869 [CVE-2013-6450]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-6450
18870 [CVE-2013-4353]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-4353
18871 [CVE-2013-0169]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-0169
18872 [CVE-2013-0166]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-0166
18873 [CVE-2012-2686]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2686
18874 [CVE-2012-2333]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2333
18875 [CVE-2012-2110]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2110
18876 [CVE-2012-0884]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0884
18877 [CVE-2012-0050]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0050
18878 [CVE-2012-0027]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0027
18879 [CVE-2011-4619]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4619
18880 [CVE-2011-4577]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4577
18881 [CVE-2011-4576]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4576
18882 [CVE-2011-4109]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4109
18883 [CVE-2011-4108]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4108
18884 [CVE-2011-3210]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-3210
18885 [CVE-2011-3207]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-3207
18886 [CVE-2011-0014]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-0014
18887 [CVE-2010-4252]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-4252
18888 [CVE-2010-4180]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-4180
18889 [CVE-2010-3864]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-3864
18890 [CVE-2010-1633]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-1633
18891 [CVE-2010-0740]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-0740
18892 [CVE-2010-0433]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-0433
18893 [CVE-2009-4355]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-4355
18894 [CVE-2009-3555]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-3555
18895 [CVE-2009-3245]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-3245
18896 [CVE-2009-1386]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1386
18897 [CVE-2009-1379]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1379
18898 [CVE-2009-1378]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1378
18899 [CVE-2009-1377]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1377
18900 [CVE-2009-0789]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0789
18901 [CVE-2009-0591]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0591
18902 [CVE-2009-0590]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0590
18903 [CVE-2008-5077]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-5077
18904 [CVE-2008-1678]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-1678
18905 [CVE-2008-1672]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-1672
18906 [CVE-2008-0891]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-0891
18907 [CVE-2007-5135]: https://www.openssl.org/news/vulnerabilities.html#CVE-2007-5135
18908 [CVE-2007-4995]: https://www.openssl.org/news/vulnerabilities.html#CVE-2007-4995
18909 [CVE-2006-4343]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-4343
18910 [CVE-2006-4339]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-4339
18911 [CVE-2006-3738]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-3738
18912 [CVE-2006-2940]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-2940
18913 [CVE-2006-2937]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-2937
18914 [CVE-2005-2969]: https://www.openssl.org/news/vulnerabilities.html#CVE-2005-2969
18915 [CVE-2004-0112]: https://www.openssl.org/news/vulnerabilities.html#CVE-2004-0112
18916 [CVE-2004-0079]: https://www.openssl.org/news/vulnerabilities.html#CVE-2004-0079
18917 [CVE-2003-0851]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0851
18918 [CVE-2003-0545]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0545
18919 [CVE-2003-0544]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0544
18920 [CVE-2003-0543]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0543
18921 [CVE-2003-0078]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0078
18922 [CVE-2002-0659]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0659
18923 [CVE-2002-0657]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0657
18924 [CVE-2002-0656]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0656
18925 [CVE-2002-0655]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0655