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 * Version of MSVC earlier than 1300 could get link warnings, which could
34 be suppressed if the undocumented -DI_CAN_LIVE_WITH_LNK4049 was set.
35 Support for this flag has been removed.
39 * Rework and make DEBUG macros consistent. Remove unused -DCONF_DEBUG,
40 -DBN_CTX_DEBUG, and REF_PRINT. Add a new tracing category and use it for
41 printing reference counts. Rename -DDEBUG_UNUSED to -DUNUSED_RESULT_DEBUG
42 Fix BN_DEBUG_RAND so it compiles and, when set, force DEBUG_RAND to be set
43 also. Rename engine_debug_ref to be ENGINE_REF_PRINT also for consistency.
47 * The signatures of the functions to get and set options on SSL and
48 SSL_CTX objects changed from "unsigned long" to "uint64_t" type.
49 Some source code changes may be required.
53 * The public definitions of conf_method_st and conf_st have been
54 deprecated. They will be made opaque in a future release.
56 *Rich Salz and Tomáš Mráz*
58 * Client-initiated renegotiation is disabled by default. To allow it, use
59 the -client_renegotiation option, the SSL_OP_ALLOW_CLIENT_RENEGOTIATION
60 flag, or the "ClientRenegotiation" config parameter as appropriate.
64 * Add "abspath" and "includedir" pragma's to config files, to prevent,
65 or modify relative pathname inclusion.
69 * OpenSSL includes a cryptographic module that is intended to be FIPS 140-2
70 validated. Please consult the README-FIPS and
71 README-PROVIDERS files, as well as the migration guide.
73 *OpenSSL team members and many third party contributors*
75 * For the key types DH and DHX the allowed settable parameters are now different.
79 * The openssl commands that read keys, certificates, and CRLs now
80 automatically detect the PEM or DER format of the input files.
82 *David von Oheimb, Richard Levitte, and Tomáš Mráz*
84 * Added enhanced PKCS#12 APIs which accept a library context.
88 * The default manual page suffix ($MANSUFFIX) has been changed to "ossl"
92 * Added support for Kernel TLS (KTLS).
94 *Boris Pismenny, John Baldwin and Andrew Gallatin*
96 * Support for RFC 5746 secure renegotiation is now required by default for
97 SSL or TLS connections to succeed.
101 * The signature of the `copy` functional parameter of the
102 EVP_PKEY_meth_set_copy() function has changed so its `src` argument is
103 now `const EVP_PKEY_CTX *` instead of `EVP_PKEY_CTX *`. Similarly
104 the signature of the `pub_decode` functional parameter of the
105 EVP_PKEY_asn1_set_public() function has changed so its `pub` argument is
106 now `const X509_PUBKEY *` instead of `X509_PUBKEY *`.
110 * The error return values from some control calls (ctrl) have changed.
114 * A public key check is now performed during EVP_PKEY_derive_set_peer().
118 * Many functions in the EVP_ namespace that are getters of values from
119 implementations or contexts were renamed to include get or get0 in their
120 names. Old names are provided as macro aliases for compatibility and
125 * The EVP_PKEY_CTRL_PKCS7_ENCRYPT, EVP_PKEY_CTRL_PKCS7_DECRYPT,
126 EVP_PKEY_CTRL_PKCS7_SIGN, EVP_PKEY_CTRL_CMS_ENCRYPT,
127 EVP_PKEY_CTRL_CMS_DECRYPT, and EVP_PKEY_CTRL_CMS_SIGN control operations
132 * The EVP_PKEY_public_check() and EVP_PKEY_param_check() functions now work for
135 * The output from the command line applications may have minor
140 * The output from numerous "printing" may have minor changes.
144 * Windows thread synchronization uses read/write primitives (SRWLock) when
145 supported by the OS, otherwise CriticalSection continues to be used.
149 * Add filter BIO BIO_f_readbuffer() that allows BIO_tell() and BIO_seek() to
150 work on read only BIO source/sinks that do not support these functions.
151 This allows piping or redirection of a file BIO using stdin to be buffered
152 into memory. This is used internally in OSSL_DECODER_from_bio().
156 * OSSL_STORE_INFO_get_type() may now return an additional value. In 1.1.1
157 this function would return one of the values OSSL_STORE_INFO_NAME,
158 OSSL_STORE_INFO_PKEY, OSSL_STORE_INFO_PARAMS, OSSL_STORE_INFO_CERT or
159 OSSL_STORE_INFO_CRL. Decoded public keys would previously have been reported
160 as type OSSL_STORE_INFO_PKEY in 1.1.1. In 3.0 decoded public keys are now
161 reported as having the new type OSSL_STORE_INFO_PUBKEY. Applications
162 using this function should be amended to handle the changed return value.
166 * Improved adherence to Enhanced Security Services (ESS, RFC 2634 and RFC 5035)
167 for the TSP and CMS Advanced Electronic Signatures (CAdES) implementations.
168 As required by RFC 5035 check both ESSCertID and ESSCertIDv2 if both present.
169 Correct the semantics of checking the validation chain in case ESSCertID{,v2}
170 contains more than one certificate identifier: This means that all
171 certificates referenced there MUST be part of the validation chain.
175 * The implementation of older EVP ciphers related to CAST, IDEA, SEED, RC2, RC4,
176 RC5, DESX and DES have been moved to the legacy provider.
180 * The implementation of the EVP digests MD2, MD4, MDC2, WHIRLPOOL and
181 RIPEMD-160 have been moved to the legacy provider.
185 * The deprecated function EVP_PKEY_get0() now returns NULL being called for a
190 * The deprecated functions EVP_PKEY_get0_RSA(),
191 EVP_PKEY_get0_DSA(), EVP_PKEY_get0_EC_KEY(), EVP_PKEY_get0_DH(),
192 EVP_PKEY_get0_hmac(), EVP_PKEY_get0_poly1305() and EVP_PKEY_get0_siphash() as
193 well as the similarly named "get1" functions behave differently in
198 * A number of functions handling low-level keys or engines were deprecated
199 including EVP_PKEY_set1_engine(), EVP_PKEY_get0_engine(), EVP_PKEY_assign(),
200 EVP_PKEY_get0(), EVP_PKEY_get0_hmac(), EVP_PKEY_get0_poly1305() and
201 EVP_PKEY_get0_siphash().
205 * PKCS#5 PBKDF1 key derivation has been moved from PKCS5_PBE_keyivgen() into
206 the legacy crypto provider as an EVP_KDF. Applications requiring this KDF
207 will need to load the legacy crypto provider. This includes these PBE
208 algorithms which use this KDF:
209 - NID_pbeWithMD2AndDES_CBC
210 - NID_pbeWithMD5AndDES_CBC
211 - NID_pbeWithSHA1AndRC2_CBC
212 - NID_pbeWithMD2AndRC2_CBC
213 - NID_pbeWithMD5AndRC2_CBC
214 - NID_pbeWithSHA1AndDES_CBC
218 * Deprecated obsolete BIO_set_callback(), BIO_get_callback(), and
219 BIO_debug_callback() functions.
223 * Deprecated obsolete EVP_PKEY_CTX_get0_dh_kdf_ukm() and
224 EVP_PKEY_CTX_get0_ecdh_kdf_ukm() functions.
228 * The RAND_METHOD APIs have been deprecated.
232 * The SRP APIs have been deprecated.
236 * Add a compile time option to prevent the caching of provider fetched
237 algorithms. This is enabled by including the no-cached-fetch option
238 at configuration time.
242 * pkcs12 now uses defaults of PBKDF2, AES and SHA-256, with a MAC iteration
243 count of PKCS12_DEFAULT_ITER.
245 *Tomáš Mráz and Sahana Prasad*
247 * The openssl speed command does not use low-level API calls anymore.
251 * Parallel dual-prime 1024-bit modular exponentiation for AVX512_IFMA
254 *Ilya Albrekht, Sergey Kirillov, Andrey Matyukov (Intel Corp)*
256 * Combining the Configure options no-ec and no-dh no longer disables TLSv1.3.
260 * Implemented support for fully "pluggable" TLSv1.3 groups. This means that
261 providers may supply their own group implementations (using either the "key
262 exchange" or the "key encapsulation" methods) which will automatically be
263 detected and used by libssl.
265 *Matt Caswell, Nicola Tuveri*
267 * The undocumented function X509_certificate_type() has been deprecated;
271 * Deprecated the obsolete BN_pseudo_rand() and BN_pseudo_rand_range().
275 * Removed RSA padding mode for SSLv23 (which was only used for
276 SSLv2). This includes the functions RSA_padding_check_SSLv23() and
277 RSA_padding_add_SSLv23() and the `-ssl` option in the deprecated
282 * Deprecated the obsolete X9.31 RSA key generation related functions.
286 * The default key generation method for the regular 2-prime RSA keys was
287 changed to the FIPS 186-4 B.3.6 method.
291 * Deprecated the BN_is_prime_ex() and BN_is_prime_fasttest_ex() functions.
295 * Deprecated EVP_MD_CTX_set_update_fn() and EVP_MD_CTX_update_fn().
299 * Deprecated the type OCSP_REQ_CTX and the functions OCSP_REQ_CTX_*() and
300 replaced with OSSL_HTTP_REQ_CTX and the functions OSSL_HTTP_REQ_CTX_*().
302 *Rich Salz, Richard Levitte, and David von Oheimb*
304 * Deprecated `X509_http_nbio()` and `X509_CRL_http_nbio()`.
308 * Deprecated `OCSP_parse_url()`.
312 * Validation of SM2 keys has been separated from the validation of regular EC
317 * Behavior of the `pkey` app is changed, when using the `-check` or `-pubcheck`
318 switches: a validation failure triggers an early exit, returning a failure
319 exit status to the parent process.
323 * Changed behavior of SSL_CTX_set_ciphersuites() and SSL_set_ciphersuites()
324 to ignore unknown ciphers.
328 * The `-cipher-commands` and `-digest-commands` options
329 of the command line utility `list` have been deprecated.
330 Instead use the `-cipher-algorithms` and `-digest-algorithms` options.
334 * Added convenience functions for generating asymmetric key pairs:
335 The 'quick' one-shot (yet somewhat limited) function L<EVP_PKEY_Q_keygen(3)>
336 and macros for the most common cases: <EVP_RSA_gen(3)> and L<EVP_EC_gen(3)>.
340 * All of the low level EC_KEY functions have been deprecated.
342 *Shane Lontis, Paul Dale, Richard Levitte, and Tomáš Mráz*
344 * Deprecated all the libcrypto and libssl error string loading
349 * The functions SSL_CTX_set_tmp_dh_callback and SSL_set_tmp_dh_callback, as
350 well as the macros SSL_CTX_set_tmp_dh() and SSL_set_tmp_dh() have been
355 * The `-crypt` option to the `passwd` command line tool has been removed.
359 * The -C option to the `x509`, `dhparam`, `dsaparam`, and `ecparam` commands
364 * Add support for AES Key Wrap inverse ciphers to the EVP layer.
368 * Deprecated EVP_PKEY_set1_tls_encodedpoint() and
369 EVP_PKEY_get1_tls_encodedpoint().
373 * The security callback, which can be customised by application code, supports
374 the security operation SSL_SECOP_TMP_DH. One location of the "other" parameter
375 was incorrectly passing a DH object. It now passed an EVP_PKEY in all cases.
379 * Add PKCS7_get_octet_string() and PKCS7_type_is_other() to the public
380 interface. Their functionality remains unchanged.
384 * Added new option for 'openssl list', '-providers', which will display the
385 list of loaded providers, their names, version and status. It optionally
386 displays their gettable parameters.
390 * Removed EVP_PKEY_set_alias_type().
394 * Deprecated `EVP_PKEY_CTX_set_rsa_keygen_pubexp()` and introduced
395 `EVP_PKEY_CTX_set1_rsa_keygen_pubexp()`, which is now preferred.
399 * Changed all "STACK" functions to be macros instead of inline functions. Macro
400 parameters are still checked for type safety at compile time via helper
405 * Remove the RAND_DRBG API
407 *Paul Dale and Matthias St. Pierre*
409 * Allow `SSL_set1_host()` and `SSL_add1_host()` to take IP literal addresses
410 as well as actual hostnames.
414 * The 'MinProtocol' and 'MaxProtocol' configuration commands now silently
415 ignore TLS protocol version bounds when configuring DTLS-based contexts, and
416 conversely, silently ignore DTLS protocol version bounds when configuring
417 TLS-based contexts. The commands can be repeated to set bounds of both
418 types. The same applies with the corresponding "min_protocol" and
419 "max_protocol" command-line switches, in case some application uses both TLS
422 SSL_CTX instances that are created for a fixed protocol version (e.g.
423 `TLSv1_server_method()`) also silently ignore version bounds. Previously
424 attempts to apply bounds to these protocol versions would result in an
425 error. Now only the "version-flexible" SSL_CTX instances are subject to
426 limits in configuration files in command-line options.
430 * Deprecated the `ENGINE` API. Engines should be replaced with providers
435 * Reworked the recorded ERR codes to make better space for system errors.
436 To distinguish them, the macro `ERR_SYSTEM_ERROR()` indicates if the
437 given code is a system error (true) or an OpenSSL error (false).
441 * Reworked the test perl framework to better allow parallel testing.
443 *Nicola Tuveri and David von Oheimb*
445 * Added ciphertext stealing algorithms AES-128-CBC-CTS, AES-192-CBC-CTS and
446 AES-256-CBC-CTS to the providers. CS1, CS2 and CS3 variants are supported.
450 * 'Configure' has been changed to figure out the configuration target if
451 none is given on the command line. Consequently, the 'config' script is
452 now only a mere wrapper. All documentation is changed to only mention
455 *Rich Salz and Richard Levitte*
457 * Added a library context `OSSL_LIB_CTX` that applications as well as
458 other libraries can use to form a separate context within which
459 libcrypto operations are performed.
463 * Handshake now fails if Extended Master Secret extension is dropped
468 * Dropped interactive mode from the `openssl` program.
472 * Deprecated `EVP_PKEY_cmp()` and `EVP_PKEY_cmp_parameters()`.
474 *David von Oheimb and Shane Lontis*
476 * Deprecated `EC_METHOD_get_field_type()`.
480 * Deprecated EC_GFp_simple_method(), EC_GFp_mont_method(),
481 EC_GF2m_simple_method(), EC_GFp_nist_method(), EC_GFp_nistp224_method()
482 EC_GFp_nistp256_method(), and EC_GFp_nistp521_method().
486 * Deprecated EC_GROUP_new(), EC_GROUP_method_of(), and EC_POINT_method_of().
490 * Add CAdES-BES signature verification support, mostly derived
491 from ESSCertIDv2 TS (RFC 5816) contribution by Marek Klein.
493 *Filipe Raimundo da Silva*
495 * Add CAdES-BES signature scheme and attributes support (RFC 5126) to CMS API.
499 * Added the AuthEnvelopedData content type structure (RFC 5083) with AES-GCM
500 parameter (RFC 5084) for the Cryptographic Message Syntax (CMS).
504 * Deprecated EC_POINT_make_affine() and EC_POINTs_make_affine().
508 * Deprecated EC_GROUP_precompute_mult(), EC_GROUP_have_precompute_mult(), and
509 EC_KEY_precompute_mult().
513 * Deprecated EC_POINTs_mul().
517 * Removed FIPS_mode() and FIPS_mode_set().
521 * The SSL option SSL_OP_IGNORE_UNEXPECTED_EOF is introduced.
525 * Deprecated EC_POINT_set_Jprojective_coordinates_GFp() and
526 EC_POINT_get_Jprojective_coordinates_GFp().
530 * Added OSSL_PARAM_BLD to the public interface. This allows OSSL_PARAM
531 arrays to be more easily constructed via a series of utility functions.
532 Create a parameter builder using OSSL_PARAM_BLD_new(), add parameters using
533 the various push functions and finally convert to a passable OSSL_PARAM
534 array using OSSL_PARAM_BLD_to_param().
538 * The security strength of SHA1 and MD5 based signatures in TLS has been
543 * Added EVP_PKEY_set_type_by_keymgmt(), to initialise an EVP_PKEY to
544 contain a provider side internal key.
548 * ASN1_verify(), ASN1_digest() and ASN1_sign() have been deprecated.
552 * Project text documents not yet having a proper file name extension
553 (`HACKING`, `LICENSE`, `NOTES*`, `README*`, `VERSION`) have been renamed to
554 `*.md` as far as reasonable, else `*.txt`, for better use with file managers.
558 * The main project documents (README, NEWS, CHANGES, INSTALL, SUPPORT)
559 have been converted to Markdown with the goal to produce documents
560 which not only look pretty when viewed online in the browser, but
561 remain well readable inside a plain text editor.
563 To achieve this goal, a 'minimalistic' Markdown style has been applied
564 which avoids formatting elements that interfere too much with the
565 reading flow in the text file. For example, it
567 * avoids [ATX headings][] and uses [setext headings][] instead
568 (which works for `<h1>` and `<h2>` headings only).
569 * avoids [inline links][] and uses [reference links][] instead.
570 * avoids [fenced code blocks][] and uses [indented code blocks][] instead.
572 [ATX headings]: https://github.github.com/gfm/#atx-headings
573 [setext headings]: https://github.github.com/gfm/#setext-headings
574 [inline links]: https://github.github.com/gfm/#inline-link
575 [reference links]: https://github.github.com/gfm/#reference-link
576 [fenced code blocks]: https://github.github.com/gfm/#fenced-code-blocks
577 [indented code blocks]: https://github.github.com/gfm/#indented-code-blocks
579 *Matthias St. Pierre*
581 * The test suite is changed to preserve results of each test recipe.
582 A new directory test-runs/ with subdirectories named like the
583 test recipes are created in the build tree for this purpose.
587 * Added an implementation of CMP and CRMF (RFC 4210, RFC 4211 RFC 6712).
588 This adds `crypto/cmp/`, `crpyto/crmf/`, `apps/cmp.c`, and `test/cmp_*`.
589 See L<openssl-cmp(1)> and L<OSSL_CMP_exec_IR_ses(3)> as starting points.
591 *David von Oheimb, Martin Peylo*
593 * Generalized the HTTP client code from `crypto/ocsp/` into `crpyto/http/`.
594 It supports arbitrary request and response content types, GET redirection,
595 TLS, connections via HTTP(S) proxies, connections and exchange via
596 user-defined BIOs (allowing implicit connections), persistent connections,
597 and timeout checks. See L<OSSL_HTTP_transfer(3)> etc. for details.
598 The legacy OCSP-focused (and only partly documented) API
599 is retained for backward compatibility, while most of it is deprecated.
603 * Added `util/check-format.pl`, a tool for checking adherence to the
604 OpenSSL coding style <https://www.openssl.org/policies/codingstyle.html>.
605 The checks performed are incomplete and yield some false positives.
606 Still the tool should be useful for detecting most typical glitches.
610 * `BIO_do_connect()` and `BIO_do_handshake()` have been extended:
611 If domain name resolution yields multiple IP addresses all of them are tried
612 after `connect()` failures.
616 * All of the low level RSA functions have been deprecated.
620 * X509 certificates signed using SHA1 are no longer allowed at security
625 * The command line utilities dhparam, dsa, gendsa and dsaparam have been
626 modified to use PKEY APIs. These commands are now in maintenance mode
627 and no new features will be added to them.
631 * The command line utility rsautl has been deprecated.
635 * The command line utilities genrsa and rsa have been modified to use PKEY
636 APIs. They now write PKCS#8 keys by default. These commands are now in
637 maintenance mode and no new features will be added to them.
641 * All of the low level DH functions have been deprecated.
643 *Paul Dale and Matt Caswell*
645 * All of the low level DSA functions have been deprecated.
649 * Reworked the treatment of EC EVP_PKEYs with the SM2 curve to
650 automatically become EVP_PKEY_SM2 rather than EVP_PKEY_EC.
654 * Deprecated low level ECDH and ECDSA functions.
658 * Deprecated EVP_PKEY_decrypt_old() and EVP_PKEY_encrypt_old().
662 * Enhanced the documentation of EVP_PKEY_get_size(), EVP_PKEY_get_bits()
663 and EVP_PKEY_get_security_bits(). Especially EVP_PKEY_get_size() needed
664 a new formulation to include all the things it can be used for,
665 as well as words of caution.
669 * The SSL_CTX_set_tlsext_ticket_key_cb(3) function has been deprecated.
673 * All of the low level HMAC functions have been deprecated.
675 *Paul Dale and David von Oheimb*
677 * Over two thousand fixes were made to the documentation, including:
678 - Common options (such as -rand/-writerand, TLS version control, etc)
679 were refactored and point to newly-enhanced descriptions in openssl.pod.
680 - Added style conformance for all options (with help from Richard Levitte),
681 documented all reported missing options, added a CI build to check
682 that all options are documented and that no unimplemented options
684 - Documented some internals, such as all use of environment variables.
685 - Addressed all internal broken L<> references.
689 * All of the low level CMAC functions have been deprecated.
693 * The low-level MD2, MD4, MD5, MDC2, RIPEMD160 and Whirlpool digest
694 functions have been deprecated.
696 *Paul Dale and David von Oheimb*
698 * Corrected the documentation of the return values from the `EVP_DigestSign*`
699 set of functions. The documentation mentioned negative values for some
700 errors, but this was never the case, so the mention of negative values
703 Code that followed the documentation and thereby check with something
704 like `EVP_DigestSignInit(...) <= 0` will continue to work undisturbed.
708 * All of the low level cipher functions have been deprecated.
710 *Matt Caswell and Paul Dale*
712 * Removed include/openssl/opensslconf.h.in and replaced it with
713 include/openssl/configuration.h.in, which differs in not including
714 <openssl/macros.h>. A short header include/openssl/opensslconf.h
715 was added to include both.
717 This allows internal hacks where one might need to modify the set
718 of configured macros, for example this if deprecated symbols are
719 still supposed to be available internally:
721 #include <openssl/configuration.h>
723 #undef OPENSSL_NO_DEPRECATED
724 #define OPENSSL_SUPPRESS_DEPRECATED
726 #include <openssl/macros.h>
728 This should not be used by applications that use the exported
729 symbols, as that will lead to linking errors.
733 * Fixed an overflow bug in the x64_64 Montgomery squaring procedure
734 used in exponentiation with 512-bit moduli. No EC algorithms are
735 affected. Analysis suggests that attacks against 2-prime RSA1024,
736 3-prime RSA1536, and DSA1024 as a result of this defect would be very
737 difficult to perform and are not believed likely. Attacks against DH512
738 are considered just feasible. However, for an attack the target would
739 have to re-use the DH512 private key, which is not recommended anyway.
740 Also applications directly using the low-level API BN_mod_exp may be
741 affected if they use BN_FLG_CONSTTIME.
746 * Most memory-debug features have been deprecated, and the functionality
747 replaced with no-ops.
751 * Added documentation for the STACK API.
755 * Introduced a new method type and API, OSSL_ENCODER, to
756 represent generic encoders.
760 * Added a .pragma directive to the syntax of configuration files, to
761 allow varying behavior in a supported and predictable manner.
762 Currently added pragma:
766 This allows dollar signs to be a keyword character unless it's
767 followed by a opening brace or parenthesis. This is useful for
768 platforms where dollar signs are commonly used in names, such as
769 volume names and system directory names on VMS.
773 * Added functionality to create an EVP_PKEY from user data.
777 * Change the interpretation of the '--api' configuration option to
778 mean that this is a desired API compatibility level with no
779 further meaning. The previous interpretation, that this would
780 also mean to remove all deprecated symbols up to and including
781 the given version, no requires that 'no-deprecated' is also used
782 in the configuration.
784 When building applications, the desired API compatibility level
785 can be set with the OPENSSL_API_COMPAT macro like before. For
786 API compatibility version below 3.0, the old style numerical
787 value is valid as before, such as -DOPENSSL_API_COMPAT=0x10100000L.
788 For version 3.0 and on, the value is expected to be the decimal
789 value calculated from the major and minor version like this:
791 MAJOR * 10000 + MINOR * 100
795 -DOPENSSL_API_COMPAT=30000 For 3.0
796 -DOPENSSL_API_COMPAT=30200 For 3.2
798 To hide declarations that are deprecated up to and including the
799 given API compatibility level, -DOPENSSL_NO_DEPRECATED must be
800 given when building the application as well.
804 * Added the X509_LOOKUP_METHOD called X509_LOOKUP_store, to allow
805 access to certificate and CRL stores via URIs and OSSL_STORE
808 This adds the following functions:
810 - X509_LOOKUP_store()
811 - X509_STORE_load_file()
812 - X509_STORE_load_path()
813 - X509_STORE_load_store()
814 - SSL_add_store_cert_subjects_to_stack()
815 - SSL_CTX_set_default_verify_store()
816 - SSL_CTX_load_verify_file()
817 - SSL_CTX_load_verify_dir()
818 - SSL_CTX_load_verify_store()
822 * Added a new method to gather entropy on VMS, based on SYS$GET_ENTROPY.
823 The presence of this system service is determined at run-time.
827 * Added functionality to create an EVP_PKEY context based on data
828 for methods from providers. This takes an algorithm name and a
829 property query string and simply stores them, with the intent
830 that any operation that uses this context will use those strings
831 to fetch the needed methods implicitly, thereby making the port
832 of application written for pre-3.0 OpenSSL easier.
836 * The undocumented function NCONF_WIN32() has been deprecated; for
837 conversion details see the HISTORY section of doc/man5/config.pod
841 * Introduced the new functions EVP_DigestSignInit_ex() and
842 EVP_DigestVerifyInit_ex(). The macros EVP_DigestSignUpdate() and
843 EVP_DigestVerifyUpdate() have been converted to functions. See the man
844 pages for further details.
848 * Over two thousand fixes were made to the documentation, including:
849 adding missing command flags, better style conformance, documentation
852 *Rich Salz, Richard Levitte*
854 * s390x assembly pack: add hardware-support for P-256, P-384, P-521,
855 X25519, X448, Ed25519 and Ed448.
859 * Print all values for a PKCS#12 attribute with 'openssl pkcs12', not just
864 * Deprecated the public definition of `ERR_STATE` as well as the function
865 `ERR_get_state()`. This is done in preparation of making `ERR_STATE` an
870 * Added ERR functionality to give callers access to the stored function
871 names that have replaced the older function code based functions.
873 New functions are ERR_peek_error_func(), ERR_peek_last_error_func(),
874 ERR_peek_error_data(), ERR_peek_last_error_data(), ERR_get_error_all(),
875 ERR_peek_error_all() and ERR_peek_last_error_all().
877 Deprecate ERR functions ERR_get_error_line(), ERR_get_error_line_data(),
878 ERR_peek_error_line_data(), ERR_peek_last_error_line_data() and
879 ERR_func_error_string().
883 * Extended testing to be verbose for failing tests only. The make variables
884 VERBOSE_FAILURE or VF can be used to enable this:
886 $ make VF=1 test # Unix
887 $ mms /macro=(VF=1) test ! OpenVMS
888 $ nmake VF=1 test # Windows
892 * Added the `-copy_extensions` option to the `x509` command for use with
893 `-req` and `-x509toreq`. When given with the `copy` or `copyall` argument,
894 all extensions in the request are copied to the certificate or vice versa.
896 *David von Oheimb*, *Kirill Stefanenkov <kirill_stefanenkov@rambler.ru>*
898 * Added the `-copy_extensions` option to the `req` command for use with
899 `-x509`. When given with the `copy` or `copyall` argument,
900 all extensions in the certification request are copied to the certificate.
904 * The `x509`, `req`, and `ca` commands now make sure that X.509v3 certificates
905 they generate are by default RFC 5280 compliant in the following sense:
906 There is a subjectKeyIdentifier extension with a hash value of the public key
907 and for not self-signed certs there is an authorityKeyIdentifier extension
908 with a keyIdentifier field or issuer information identifying the signing key.
909 This is done unless some configuration overrides the new default behavior,
910 such as `subjectKeyIdentifier = none` and `authorityKeyIdentifier = none`.
914 * Added several checks to `X509_verify_cert()` according to requirements in
915 RFC 5280 in case `X509_V_FLAG_X509_STRICT` is set
916 (which may be done by using the CLI option `-x509_strict`):
917 * The basicConstraints of CA certificates must be marked critical.
918 * CA certificates must explicitly include the keyUsage extension.
919 * If a pathlenConstraint is given the key usage keyCertSign must be allowed.
920 * The issuer name of any certificate must not be empty.
921 * The subject name of CA certs, certs with keyUsage crlSign,
922 and certs without subjectAlternativeName must not be empty.
923 * If a subjectAlternativeName extension is given it must not be empty.
924 * The signatureAlgorithm field and the cert signature must be consistent.
925 * Any given authorityKeyIdentifier and any given subjectKeyIdentifier
926 must not be marked critical.
927 * The authorityKeyIdentifier must be given for X.509v3 certs
928 unless they are self-signed.
929 * The subjectKeyIdentifier must be given for all X.509v3 CA certs.
933 * Certificate verification using `X509_verify_cert()` meanwhile rejects EC keys
934 with explicit curve parameters (specifiedCurve) as required by RFC 5480.
938 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
939 used even when parsing explicit parameters, when loading a encoded key
940 or calling `EC_GROUP_new_from_ecpkparameters()`/
941 `EC_GROUP_new_from_ecparameters()`.
942 This prevents bypass of security hardening and performance gains,
943 especially for curves with specialized EC_METHODs.
944 By default, if a key encoded with explicit parameters is loaded and later
945 encoded, the output is still encoded with explicit parameters, even if
946 internally a "named" EC_GROUP is used for computation.
950 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
951 this change, EC_GROUP_set_generator would accept order and/or cofactor as
952 NULL. After this change, only the cofactor parameter can be NULL. It also
953 does some minimal sanity checks on the passed order.
958 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
959 An attack is simple, if the first CMS_recipientInfo is valid but the
960 second CMS_recipientInfo is chosen ciphertext. If the second
961 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
962 encryption key will be replaced by garbage, and the message cannot be
963 decoded, but if the RSA decryption fails, the correct encryption key is
964 used and the recipient will not notice the attack.
965 As a work around for this potential attack the length of the decrypted
966 key must be equal to the cipher default key length, in case the
967 certifiate is not given and all recipientInfo are tried out.
968 The old behaviour can be re-enabled in the CMS code by setting the
969 CMS_DEBUG_DECRYPT flag.
973 * Early start up entropy quality from the DEVRANDOM seed source has been
974 improved for older Linux systems. The RAND subsystem will wait for
975 /dev/random to be producing output before seeding from /dev/urandom.
976 The seeded state is stored for future library initialisations using
977 a system global shared memory segment. The shared memory identifier
978 can be configured by defining OPENSSL_RAND_SEED_DEVRANDOM_SHM_ID to
979 the desired value. The default identifier is 114.
983 * Revised BN_generate_prime_ex to not avoid factors 2..17863 in p-1
984 when primes for RSA keys are computed.
985 Since we previously always generated primes == 2 (mod 3) for RSA keys,
986 the 2-prime and 3-prime RSA modules were easy to distinguish, since
987 `N = p*q = 1 (mod 3)`, but `N = p*q*r = 2 (mod 3)`. Therefore fingerprinting
988 2-prime vs. 3-prime RSA keys was possible by computing N mod 3.
989 This avoids possible fingerprinting of newly generated RSA modules.
993 * Correct the extended master secret constant on EBCDIC systems. Without this
994 fix TLS connections between an EBCDIC system and a non-EBCDIC system that
995 negotiate EMS will fail. Unfortunately this also means that TLS connections
996 between EBCDIC systems with this fix, and EBCDIC systems without this
997 fix will fail if they negotiate EMS.
1001 * Changed the library initialisation so that the config file is now loaded
1002 by default. This was already the case for libssl. It now occurs for both
1003 libcrypto and libssl. Use the OPENSSL_INIT_NO_LOAD_CONFIG option to
1004 `OPENSSL_init_crypto()` to suppress automatic loading of a config file.
1008 * Introduced new error raising macros, `ERR_raise()` and `ERR_raise_data()`,
1009 where the former acts as a replacement for `ERR_put_error()`, and the
1010 latter replaces the combination `ERR_put_error()` + `ERR_add_error_data()`.
1011 `ERR_raise_data()` adds more flexibility by taking a format string and
1012 an arbitrary number of arguments following it, to be processed with
1017 * Introduced a new function, `OSSL_PROVIDER_available()`, which can be used
1018 to check if a named provider is loaded and available. When called, it
1019 will also activate all fallback providers if such are still present.
1023 * Enforce a minimum DH modulus size of 512 bits.
1027 * Changed DH parameters to generate the order q subgroup instead of 2q.
1028 Previously generated DH parameters are still accepted by DH_check
1029 but DH_generate_key works around that by clearing bit 0 of the
1030 private key for those. This avoids leaking bit 0 of the private key.
1034 * Significantly reduce secure memory usage by the randomness pools.
1038 * `{CRYPTO,OPENSSL}_mem_debug_{push,pop}` are now no-ops and have been
1043 * A new type, EVP_KEYEXCH, has been introduced to represent key exchange
1044 algorithms. An implementation of a key exchange algorithm can be obtained
1045 by using the function EVP_KEYEXCH_fetch(). An EVP_KEYEXCH algorithm can be
1046 used in a call to EVP_PKEY_derive_init_ex() which works in a similar way to
1047 the older EVP_PKEY_derive_init() function. See the man pages for the new
1048 functions for further details.
1052 * The EVP_PKEY_CTX_set_dh_pad() macro has now been converted to a function.
1056 * Removed the function names from error messages and deprecated the
1059 * Removed NextStep support and the macro OPENSSL_UNISTD
1063 * Removed DES_check_key. Also removed OPENSSL_IMPLEMENT_GLOBAL,
1064 OPENSSL_GLOBAL_REF, OPENSSL_DECLARE_GLOBAL.
1065 Also removed "export var as function" capability; we do not export
1066 variables, only functions.
1070 * RC5_32_set_key has been changed to return an int type, with 0 indicating
1071 an error and 1 indicating success. In previous versions of OpenSSL this
1072 was a void type. If a key was set longer than the maximum possible this
1077 * Support SM2 signing and verification schemes with X509 certificate.
1081 * Use SHA256 as the default digest for TS query in the `ts` app.
1085 * Change PBKDF2 to conform to SP800-132 instead of the older PKCS5 RFC2898.
1089 * Default cipher lists/suites are now available via a function, the
1090 #defines are deprecated.
1094 * Add target VC-WIN32-UWP, VC-WIN64A-UWP, VC-WIN32-ARM-UWP and
1095 VC-WIN64-ARM-UWP in Windows OneCore target for making building libraries
1096 for Windows Store apps easier. Also, the "no-uplink" option has been added.
1100 * Join the directories crypto/x509 and crypto/x509v3
1104 * Added command 'openssl kdf' that uses the EVP_KDF API.
1108 * Added command 'openssl mac' that uses the EVP_MAC API.
1112 * Added OPENSSL_info() to get diverse built-in OpenSSL data, such
1113 as default directories. Also added the command 'openssl info'
1114 for scripting purposes.
1118 * The functions AES_ige_encrypt() and AES_bi_ige_encrypt() have been
1123 * Add prediction resistance to the DRBG reseeding process.
1127 * Limit the number of blocks in a data unit for AES-XTS to 2^20 as
1128 mandated by IEEE Std 1619-2018.
1132 * Added newline escaping functionality to a filename when using openssl dgst.
1133 This output format is to replicate the output format found in the `*sum`
1134 checksum programs. This aims to preserve backward compatibility.
1136 *Matt Eaton, Richard Levitte, and Paul Dale*
1138 * Removed the heartbeat message in DTLS feature, as it has very
1139 little usage and doesn't seem to fulfill a valuable purpose.
1140 The configuration option is now deprecated.
1144 * Changed the output of 'openssl {digestname} < file' to display the
1145 digest name in its output.
1149 * Added a new generic trace API which provides support for enabling
1150 instrumentation through trace output.
1152 *Richard Levitte & Matthias St. Pierre*
1154 * Added build tests for C++. These are generated files that only do one
1155 thing, to include one public OpenSSL head file each. This tests that
1156 the public header files can be usefully included in a C++ application.
1158 This test isn't enabled by default. It can be enabled with the option
1159 'enable-buildtest-c++'.
1163 * Add Single Step KDF (EVP_KDF_SS) to EVP_KDF.
1167 * Add KMAC to EVP_MAC.
1171 * Added property based algorithm implementation selection framework to
1176 * Added SCA hardening for modular field inversion in EC_GROUP through
1177 a new dedicated field_inv() pointer in EC_METHOD.
1178 This also addresses a leakage affecting conversions from projective
1179 to affine coordinates.
1181 *Billy Bob Brumley, Nicola Tuveri*
1183 * Added EVP_KDF, an EVP layer KDF API, to simplify adding KDF and PRF
1184 implementations. This includes an EVP_PKEY to EVP_KDF bridge for
1185 those algorithms that were already supported through the EVP_PKEY API
1186 (scrypt, TLS1 PRF and HKDF). The low-level KDF functions for PBKDF2
1187 and scrypt are now wrappers that call EVP_KDF.
1191 * Build devcrypto engine as a dynamic engine.
1193 *Eneas U de Queiroz*
1195 * Add keyed BLAKE2 to EVP_MAC.
1199 * Fix a bug in the computation of the endpoint-pair shared secret used
1200 by DTLS over SCTP. This breaks interoperability with older versions
1201 of OpenSSL like OpenSSL 1.1.0 and OpenSSL 1.0.2. There is a runtime
1202 switch SSL_MODE_DTLS_SCTP_LABEL_LENGTH_BUG (off by default) enabling
1203 interoperability with such broken implementations. However, enabling
1204 this switch breaks interoperability with correct implementations.
1206 * Fix a use after free bug in d2i_X509_PUBKEY when overwriting a
1207 re-used X509_PUBKEY object if the second PUBKEY is malformed.
1211 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
1215 * Changed the license to the Apache License v2.0.
1219 * Switch to a new version scheme using three numbers MAJOR.MINOR.PATCH.
1221 - Major releases (indicated by incrementing the MAJOR release number)
1222 may introduce incompatible API/ABI changes.
1223 - Minor releases (indicated by incrementing the MINOR release number)
1224 may introduce new features but retain API/ABI compatibility.
1225 - Patch releases (indicated by incrementing the PATCH number)
1226 are intended for bug fixes and other improvements of existing
1227 features only (like improving performance or adding documentation)
1228 and retain API/ABI compatibility.
1232 * Add support for RFC5297 SIV mode (siv128), including AES-SIV.
1236 * Remove the 'dist' target and add a tarball building script. The
1237 'dist' target has fallen out of use, and it shouldn't be
1238 necessary to configure just to create a source distribution.
1242 * Recreate the OS390-Unix config target. It no longer relies on a
1243 special script like it did for OpenSSL pre-1.1.0.
1247 * Instead of having the source directories listed in Configure, add
1248 a 'build.info' keyword SUBDIRS to indicate what sub-directories to
1253 * Add GMAC to EVP_MAC.
1257 * Ported the HMAC, CMAC and SipHash EVP_PKEY_METHODs to EVP_MAC.
1261 * Added EVP_MAC, an EVP layer MAC API, to simplify adding MAC
1262 implementations. This includes a generic EVP_PKEY to EVP_MAC bridge,
1263 to facilitate the continued use of MACs through raw private keys in
1264 functionality such as `EVP_DigestSign*` and `EVP_DigestVerify*`.
1268 * Deprecate ECDH_KDF_X9_62().
1272 * Added EVP_PKEY_ECDH_KDF_X9_63 and ecdh_KDF_X9_63() as replacements for
1273 the EVP_PKEY_ECDH_KDF_X9_62 KDF type and ECDH_KDF_X9_62(). The old names
1274 are retained for backwards compatibility.
1278 * AES-XTS mode now enforces that its two keys are different to mitigate
1279 the attacked described in "Efficient Instantiations of Tweakable
1280 Blockciphers and Refinements to Modes OCB and PMAC" by Phillip Rogaway.
1281 Details of this attack can be obtained from:
1282 <http://web.cs.ucdavis.edu/%7Erogaway/papers/offsets.pdf>
1286 * Rename the object files, i.e. give them other names than in previous
1287 versions. Their names now include the name of the final product, as
1288 well as its type mnemonic (bin, lib, shlib).
1292 * Added new option for 'openssl list', '-objects', which will display the
1293 list of built in objects, i.e. OIDs with names.
1297 * Added the options `-crl_lastupdate` and `-crl_nextupdate` to `openssl ca`,
1298 allowing the `lastUpdate` and `nextUpdate` fields in the generated CRL to
1303 * Added support for Linux Kernel TLS data-path. The Linux Kernel data-path
1304 improves application performance by removing data copies and providing
1305 applications with zero-copy system calls such as sendfile and splice.
1309 * The SSL option SSL_OP_CLEANSE_PLAINTEXT is introduced.
1313 * `PKCS12_parse` now maintains the order of the parsed certificates
1314 when outputting them via `*ca` (rather than reversing it).
1318 * Deprecated pthread fork support methods.
1322 * Added support for FFDHE key exchange in TLS 1.3.
1329 ### Changes between 1.1.1j and 1.1.1k [xx XXX xxxx]
1331 * Fixed a problem with verifying a certificate chain when using the
1332 X509_V_FLAG_X509_STRICT flag. This flag enables additional security checks of
1333 the certificates present in a certificate chain. It is not set by default.
1335 Starting from OpenSSL version 1.1.1h a check to disallow certificates in
1336 the chain that have explicitly encoded elliptic curve parameters was added
1337 as an additional strict check.
1339 An error in the implementation of this check meant that the result of a
1340 previous check to confirm that certificates in the chain are valid CA
1341 certificates was overwritten. This effectively bypasses the check
1342 that non-CA certificates must not be able to issue other certificates.
1344 If a "purpose" has been configured then there is a subsequent opportunity
1345 for checks that the certificate is a valid CA. All of the named "purpose"
1346 values implemented in libcrypto perform this check. Therefore, where
1347 a purpose is set the certificate chain will still be rejected even when the
1348 strict flag has been used. A purpose is set by default in libssl client and
1349 server certificate verification routines, but it can be overridden or
1350 removed by an application.
1352 In order to be affected, an application must explicitly set the
1353 X509_V_FLAG_X509_STRICT verification flag and either not set a purpose
1354 for the certificate verification or, in the case of TLS client or server
1355 applications, override the default purpose.
1360 * Fixed an issue where an OpenSSL TLS server may crash if sent a maliciously
1361 crafted renegotiation ClientHello message from a client. If a TLSv1.2
1362 renegotiation ClientHello omits the signature_algorithms extension (where it
1363 was present in the initial ClientHello), but includes a
1364 signature_algorithms_cert extension then a NULL pointer dereference will
1365 result, leading to a crash and a denial of service attack.
1367 A server is only vulnerable if it has TLSv1.2 and renegotiation enabled
1368 (which is the default configuration). OpenSSL TLS clients are not impacted by
1372 *Peter Kästle and Samuel Sapalski*
1374 ### Changes between 1.1.1i and 1.1.1j [16 Feb 2021]
1376 * Fixed the X509_issuer_and_serial_hash() function. It attempts to
1377 create a unique hash value based on the issuer and serial number data
1378 contained within an X509 certificate. However it was failing to correctly
1379 handle any errors that may occur while parsing the issuer field (which might
1380 occur if the issuer field is maliciously constructed). This may subsequently
1381 result in a NULL pointer deref and a crash leading to a potential denial of
1387 * Fixed the RSA_padding_check_SSLv23() function and the RSA_SSLV23_PADDING
1388 padding mode to correctly check for rollback attacks. This is considered a
1389 bug in OpenSSL 1.1.1 because it does not support SSLv2. In 1.0.2 this is
1394 Fixed the EVP_CipherUpdate, EVP_EncryptUpdate and EVP_DecryptUpdate
1395 functions. Previously they could overflow the output length argument in some
1396 cases where the input length is close to the maximum permissable length for
1397 an integer on the platform. In such cases the return value from the function
1398 call would be 1 (indicating success), but the output length value would be
1399 negative. This could cause applications to behave incorrectly or crash.
1404 * Fixed SRP_Calc_client_key so that it runs in constant time. The previous
1405 implementation called BN_mod_exp without setting BN_FLG_CONSTTIME. This
1406 could be exploited in a side channel attack to recover the password. Since
1407 the attack is local host only this is outside of the current OpenSSL
1408 threat model and therefore no CVE is assigned.
1410 Thanks to Mohammed Sabt and Daniel De Almeida Braga for reporting this
1415 ### Changes between 1.1.1h and 1.1.1i [8 Dec 2020]
1417 * Fixed NULL pointer deref in the GENERAL_NAME_cmp function
1418 This function could crash if both GENERAL_NAMEs contain an EDIPARTYNAME.
1419 If an attacker can control both items being compared then this could lead
1420 to a possible denial of service attack. OpenSSL itself uses the
1421 GENERAL_NAME_cmp function for two purposes:
1422 1) Comparing CRL distribution point names between an available CRL and a
1423 CRL distribution point embedded in an X509 certificate
1424 2) When verifying that a timestamp response token signer matches the
1425 timestamp authority name (exposed via the API functions
1426 TS_RESP_verify_response and TS_RESP_verify_token)
1431 ### Changes between 1.1.1g and 1.1.1h [22 Sep 2020]
1433 * Certificates with explicit curve parameters are now disallowed in
1434 verification chains if the X509_V_FLAG_X509_STRICT flag is used.
1438 * The 'MinProtocol' and 'MaxProtocol' configuration commands now silently
1439 ignore TLS protocol version bounds when configuring DTLS-based contexts, and
1440 conversely, silently ignore DTLS protocol version bounds when configuring
1441 TLS-based contexts. The commands can be repeated to set bounds of both
1442 types. The same applies with the corresponding "min_protocol" and
1443 "max_protocol" command-line switches, in case some application uses both TLS
1446 SSL_CTX instances that are created for a fixed protocol version (e.g.
1447 TLSv1_server_method()) also silently ignore version bounds. Previously
1448 attempts to apply bounds to these protocol versions would result in an
1449 error. Now only the "version-flexible" SSL_CTX instances are subject to
1450 limits in configuration files in command-line options.
1454 * Handshake now fails if Extended Master Secret extension is dropped
1459 * The Oracle Developer Studio compiler will start reporting deprecated APIs
1461 ### Changes between 1.1.1f and 1.1.1g [21 Apr 2020]
1463 * Fixed segmentation fault in SSL_check_chain()
1464 Server or client applications that call the SSL_check_chain() function
1465 during or after a TLS 1.3 handshake may crash due to a NULL pointer
1466 dereference as a result of incorrect handling of the
1467 "signature_algorithms_cert" TLS extension. The crash occurs if an invalid
1468 or unrecognised signature algorithm is received from the peer. This could
1469 be exploited by a malicious peer in a Denial of Service attack.
1474 * Added AES consttime code for no-asm configurations
1475 an optional constant time support for AES was added
1476 when building openssl for no-asm.
1477 Enable with: ./config no-asm -DOPENSSL_AES_CONST_TIME
1478 Disable with: ./config no-asm -DOPENSSL_NO_AES_CONST_TIME
1479 At this time this feature is by default disabled.
1480 It will be enabled by default in 3.0.
1484 ### Changes between 1.1.1e and 1.1.1f [31 Mar 2020]
1486 * Revert the change of EOF detection while reading in libssl to avoid
1487 regressions in applications depending on the current way of reporting
1488 the EOF. As the existing method is not fully accurate the change to
1489 reporting the EOF via SSL_ERROR_SSL is kept on the current development
1490 branch and will be present in the 3.0 release.
1494 * Revised BN_generate_prime_ex to not avoid factors 3..17863 in p-1
1495 when primes for RSA keys are computed.
1496 Since we previously always generated primes == 2 (mod 3) for RSA keys,
1497 the 2-prime and 3-prime RSA modules were easy to distinguish, since
1498 N = p*q = 1 (mod 3), but N = p*q*r = 2 (mod 3). Therefore fingerprinting
1499 2-prime vs. 3-prime RSA keys was possible by computing N mod 3.
1500 This avoids possible fingerprinting of newly generated RSA modules.
1504 ### Changes between 1.1.1d and 1.1.1e [17 Mar 2020]
1506 * Properly detect EOF while reading in libssl. Previously if we hit an EOF
1507 while reading in libssl then we would report an error back to the
1508 application (SSL_ERROR_SYSCALL) but errno would be 0. We now add
1509 an error to the stack (which means we instead return SSL_ERROR_SSL) and
1510 therefore give a hint as to what went wrong.
1514 * Check that ed25519 and ed448 are allowed by the security level. Previously
1515 signature algorithms not using an MD were not being checked that they were
1516 allowed by the security level.
1520 * Fixed SSL_get_servername() behaviour. The behaviour of SSL_get_servername()
1521 was not quite right. The behaviour was not consistent between resumption
1522 and normal handshakes, and also not quite consistent with historical
1523 behaviour. The behaviour in various scenarios has been clarified and
1524 it has been updated to make it match historical behaviour as closely as
1529 * *[VMS only]* The header files that the VMS compilers include automatically,
1530 `__DECC_INCLUDE_PROLOGUE.H` and `__DECC_INCLUDE_EPILOGUE.H`, use pragmas
1531 that the C++ compiler doesn't understand. This is a shortcoming in the
1532 compiler, but can be worked around with `__cplusplus` guards.
1534 C++ applications that use OpenSSL libraries must be compiled using the
1535 qualifier `/NAMES=(AS_IS,SHORTENED)` to be able to use all the OpenSSL
1536 functions. Otherwise, only functions with symbols of less than 31
1537 characters can be used, as the linker will not be able to successfully
1538 resolve symbols with longer names.
1542 * Added a new method to gather entropy on VMS, based on SYS$GET_ENTROPY.
1543 The presence of this system service is determined at run-time.
1547 * Added newline escaping functionality to a filename when using openssl dgst.
1548 This output format is to replicate the output format found in the `*sum`
1549 checksum programs. This aims to preserve backward compatibility.
1551 *Matt Eaton, Richard Levitte, and Paul Dale*
1553 * Print all values for a PKCS#12 attribute with 'openssl pkcs12', not just
1558 ### Changes between 1.1.1c and 1.1.1d [10 Sep 2019]
1560 * Fixed a fork protection issue. OpenSSL 1.1.1 introduced a rewritten random
1561 number generator (RNG). This was intended to include protection in the
1562 event of a fork() system call in order to ensure that the parent and child
1563 processes did not share the same RNG state. However this protection was not
1564 being used in the default case.
1566 A partial mitigation for this issue is that the output from a high
1567 precision timer is mixed into the RNG state so the likelihood of a parent
1568 and child process sharing state is significantly reduced.
1570 If an application already calls OPENSSL_init_crypto() explicitly using
1571 OPENSSL_INIT_ATFORK then this problem does not occur at all.
1574 *Matthias St. Pierre*
1576 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
1577 used even when parsing explicit parameters, when loading a encoded key
1578 or calling `EC_GROUP_new_from_ecpkparameters()`/
1579 `EC_GROUP_new_from_ecparameters()`.
1580 This prevents bypass of security hardening and performance gains,
1581 especially for curves with specialized EC_METHODs.
1582 By default, if a key encoded with explicit parameters is loaded and later
1583 encoded, the output is still encoded with explicit parameters, even if
1584 internally a "named" EC_GROUP is used for computation.
1588 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
1589 this change, EC_GROUP_set_generator would accept order and/or cofactor as
1590 NULL. After this change, only the cofactor parameter can be NULL. It also
1591 does some minimal sanity checks on the passed order.
1596 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
1597 An attack is simple, if the first CMS_recipientInfo is valid but the
1598 second CMS_recipientInfo is chosen ciphertext. If the second
1599 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
1600 encryption key will be replaced by garbage, and the message cannot be
1601 decoded, but if the RSA decryption fails, the correct encryption key is
1602 used and the recipient will not notice the attack.
1603 As a work around for this potential attack the length of the decrypted
1604 key must be equal to the cipher default key length, in case the
1605 certifiate is not given and all recipientInfo are tried out.
1606 The old behaviour can be re-enabled in the CMS code by setting the
1607 CMS_DEBUG_DECRYPT flag.
1612 * Early start up entropy quality from the DEVRANDOM seed source has been
1613 improved for older Linux systems. The RAND subsystem will wait for
1614 /dev/random to be producing output before seeding from /dev/urandom.
1615 The seeded state is stored for future library initialisations using
1616 a system global shared memory segment. The shared memory identifier
1617 can be configured by defining OPENSSL_RAND_SEED_DEVRANDOM_SHM_ID to
1618 the desired value. The default identifier is 114.
1622 * Correct the extended master secret constant on EBCDIC systems. Without this
1623 fix TLS connections between an EBCDIC system and a non-EBCDIC system that
1624 negotiate EMS will fail. Unfortunately this also means that TLS connections
1625 between EBCDIC systems with this fix, and EBCDIC systems without this
1626 fix will fail if they negotiate EMS.
1630 * Use Windows installation paths in the mingw builds
1632 Mingw isn't a POSIX environment per se, which means that Windows
1633 paths should be used for installation.
1638 * Changed DH_check to accept parameters with order q and 2q subgroups.
1639 With order 2q subgroups the bit 0 of the private key is not secret
1640 but DH_generate_key works around that by clearing bit 0 of the
1641 private key for those. This avoids leaking bit 0 of the private key.
1645 * Significantly reduce secure memory usage by the randomness pools.
1649 * Revert the DEVRANDOM_WAIT feature for Linux systems
1651 The DEVRANDOM_WAIT feature added a select() call to wait for the
1652 /dev/random device to become readable before reading from the
1653 /dev/urandom device.
1655 It turned out that this change had negative side effects on
1656 performance which were not acceptable. After some discussion it
1657 was decided to revert this feature and leave it up to the OS
1658 resp. the platform maintainer to ensure a proper initialization
1659 during early boot time.
1661 *Matthias St. Pierre*
1663 ### Changes between 1.1.1b and 1.1.1c [28 May 2019]
1665 * Add build tests for C++. These are generated files that only do one
1666 thing, to include one public OpenSSL head file each. This tests that
1667 the public header files can be usefully included in a C++ application.
1669 This test isn't enabled by default. It can be enabled with the option
1670 'enable-buildtest-c++'.
1674 * Enable SHA3 pre-hashing for ECDSA and DSA.
1678 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
1679 This changes the size when using the `genpkey` command when no size is given.
1680 It fixes an omission in earlier changes that changed all RSA, DSA and DH
1681 generation commands to use 2048 bits by default.
1685 * Reorganize the manual pages to consistently have RETURN VALUES,
1686 EXAMPLES, SEE ALSO and HISTORY come in that order, and adjust
1687 util/fix-doc-nits accordingly.
1689 *Paul Yang, Joshua Lock*
1691 * Add the missing accessor EVP_PKEY_get0_engine()
1695 * Have commands like `s_client` and `s_server` output the signature scheme
1696 along with other cipher suite parameters when debugging.
1700 * Make OPENSSL_config() error agnostic again.
1704 * Do the error handling in RSA decryption constant time.
1708 * Prevent over long nonces in ChaCha20-Poly1305.
1710 ChaCha20-Poly1305 is an AEAD cipher, and requires a unique nonce input
1711 for every encryption operation. RFC 7539 specifies that the nonce value
1712 (IV) should be 96 bits (12 bytes). OpenSSL allows a variable nonce length
1713 and front pads the nonce with 0 bytes if it is less than 12
1714 bytes. However it also incorrectly allows a nonce to be set of up to 16
1715 bytes. In this case only the last 12 bytes are significant and any
1716 additional leading bytes are ignored.
1718 It is a requirement of using this cipher that nonce values are
1719 unique. Messages encrypted using a reused nonce value are susceptible to
1720 serious confidentiality and integrity attacks. If an application changes
1721 the default nonce length to be longer than 12 bytes and then makes a
1722 change to the leading bytes of the nonce expecting the new value to be a
1723 new unique nonce then such an application could inadvertently encrypt
1724 messages with a reused nonce.
1726 Additionally the ignored bytes in a long nonce are not covered by the
1727 integrity guarantee of this cipher. Any application that relies on the
1728 integrity of these ignored leading bytes of a long nonce may be further
1729 affected. Any OpenSSL internal use of this cipher, including in SSL/TLS,
1730 is safe because no such use sets such a long nonce value. However user
1731 applications that use this cipher directly and set a non-default nonce
1732 length to be longer than 12 bytes may be vulnerable.
1734 This issue was reported to OpenSSL on 16th of March 2019 by Joran Dirk
1740 * Add DEVRANDOM_WAIT feature for Linux systems
1742 On older Linux systems where the getrandom() system call is not available,
1743 OpenSSL normally uses the /dev/urandom device for seeding its CSPRNG.
1744 Contrary to getrandom(), the /dev/urandom device will not block during
1745 early boot when the kernel CSPRNG has not been seeded yet.
1747 To mitigate this known weakness, use select() to wait for /dev/random to
1748 become readable before reading from /dev/urandom.
1750 * Ensure that SM2 only uses SM3 as digest algorithm
1754 ### Changes between 1.1.1a and 1.1.1b [26 Feb 2019]
1756 * Change the info callback signals for the start and end of a post-handshake
1757 message exchange in TLSv1.3. In 1.1.1/1.1.1a we used SSL_CB_HANDSHAKE_START
1758 and SSL_CB_HANDSHAKE_DONE. Experience has shown that many applications get
1759 confused by this and assume that a TLSv1.2 renegotiation has started. This
1760 can break KeyUpdate handling. Instead we no longer signal the start and end
1761 of a post handshake message exchange (although the messages themselves are
1762 still signalled). This could break some applications that were expecting
1763 the old signals. However without this KeyUpdate is not usable for many
1768 ### Changes between 1.1.1 and 1.1.1a [20 Nov 2018]
1770 * Timing vulnerability in DSA signature generation
1772 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
1773 timing side channel attack. An attacker could use variations in the signing
1774 algorithm to recover the private key.
1776 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
1781 * Timing vulnerability in ECDSA signature generation
1783 The OpenSSL ECDSA signature algorithm has been shown to be vulnerable to a
1784 timing side channel attack. An attacker could use variations in the signing
1785 algorithm to recover the private key.
1787 This issue was reported to OpenSSL on 25th October 2018 by Samuel Weiser.
1792 * Fixed the issue that RAND_add()/RAND_seed() silently discards random input
1793 if its length exceeds 4096 bytes. The limit has been raised to a buffer size
1794 of two gigabytes and the error handling improved.
1796 This issue was reported to OpenSSL by Dr. Falko Strenzke. It has been
1797 categorized as a normal bug, not a security issue, because the DRBG reseeds
1798 automatically and is fully functional even without additional randomness
1799 provided by the application.
1801 ### Changes between 1.1.0i and 1.1.1 [11 Sep 2018]
1803 * Add a new ClientHello callback. Provides a callback interface that gives
1804 the application the ability to adjust the nascent SSL object at the
1805 earliest stage of ClientHello processing, immediately after extensions have
1806 been collected but before they have been processed. In particular, this
1807 callback can adjust the supported TLS versions in response to the contents
1812 * Add SM2 base algorithm support.
1816 * s390x assembly pack: add (improved) hardware-support for the following
1817 cryptographic primitives: sha3, shake, aes-gcm, aes-ccm, aes-ctr, aes-ofb,
1818 aes-cfb/cfb8, aes-ecb.
1822 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
1823 parameter is no longer accepted, as it leads to a corrupt table. NULL
1824 pem_str is reserved for alias entries only.
1828 * Use the new ec_scalar_mul_ladder scaffold to implement a specialized ladder
1829 step for prime curves. The new implementation is based on formulae from
1830 differential addition-and-doubling in homogeneous projective coordinates
1831 from Izu-Takagi "A fast parallel elliptic curve multiplication resistant
1832 against side channel attacks" and Brier-Joye "Weierstrass Elliptic Curves
1833 and Side-Channel Attacks" Eq. (8) for y-coordinate recovery, modified
1834 to work in projective coordinates.
1836 *Billy Bob Brumley, Nicola Tuveri*
1838 * Change generating and checking of primes so that the error rate of not
1839 being prime depends on the intended use based on the size of the input.
1840 For larger primes this will result in more rounds of Miller-Rabin.
1841 The maximal error rate for primes with more than 1080 bits is lowered
1844 *Kurt Roeckx, Annie Yousar*
1846 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
1850 * The 'tsget' script is renamed to 'tsget.pl', to avoid confusion when
1851 moving between systems, and to avoid confusion when a Windows build is
1852 done with mingw vs with MSVC. For POSIX installs, there's still a
1853 symlink or copy named 'tsget' to avoid that confusion as well.
1857 * Revert blinding in ECDSA sign and instead make problematic addition
1858 length-invariant. Switch even to fixed-length Montgomery multiplication.
1862 * Use the new ec_scalar_mul_ladder scaffold to implement a specialized ladder
1863 step for binary curves. The new implementation is based on formulae from
1864 differential addition-and-doubling in mixed Lopez-Dahab projective
1865 coordinates, modified to independently blind the operands.
1867 *Billy Bob Brumley, Sohaib ul Hassan, Nicola Tuveri*
1869 * Add a scaffold to optionally enhance the Montgomery ladder implementation
1870 for `ec_scalar_mul_ladder` (formerly `ec_mul_consttime`) allowing
1871 EC_METHODs to implement their own specialized "ladder step", to take
1872 advantage of more favorable coordinate systems or more efficient
1873 differential addition-and-doubling algorithms.
1875 *Billy Bob Brumley, Sohaib ul Hassan, Nicola Tuveri*
1877 * Modified the random device based seed sources to keep the relevant
1878 file descriptors open rather than reopening them on each access.
1879 This allows such sources to operate in a chroot() jail without
1880 the associated device nodes being available. This behaviour can be
1881 controlled using RAND_keep_random_devices_open().
1885 * Numerous side-channel attack mitigations have been applied. This may have
1886 performance impacts for some algorithms for the benefit of improved
1887 security. Specific changes are noted in this change log by their respective
1892 * AIX shared library support overhaul. Switch to AIX "natural" way of
1893 handling shared libraries, which means collecting shared objects of
1894 different versions and bitnesses in one common archive. This allows to
1895 mitigate conflict between 1.0 and 1.1 side-by-side installations. It
1896 doesn't affect the way 3rd party applications are linked, only how
1897 multi-version installation is managed.
1901 * Make ec_group_do_inverse_ord() more robust and available to other
1902 EC cryptosystems, so that irrespective of BN_FLG_CONSTTIME, SCA
1903 mitigations are applied to the fallback BN_mod_inverse().
1904 When using this function rather than BN_mod_inverse() directly, new
1905 EC cryptosystem implementations are then safer-by-default.
1909 * Add coordinate blinding for EC_POINT and implement projective
1910 coordinate blinding for generic prime curves as a countermeasure to
1911 chosen point SCA attacks.
1913 *Sohaib ul Hassan, Nicola Tuveri, Billy Bob Brumley*
1915 * Add blinding to ECDSA and DSA signatures to protect against side channel
1916 attacks discovered by Keegan Ryan (NCC Group).
1920 * Enforce checking in the `pkeyutl` command to ensure that the input
1921 length does not exceed the maximum supported digest length when performing
1922 a sign, verify or verifyrecover operation.
1926 * SSL_MODE_AUTO_RETRY is enabled by default. Applications that use blocking
1927 I/O in combination with something like select() or poll() will hang. This
1928 can be turned off again using SSL_CTX_clear_mode().
1929 Many applications do not properly handle non-application data records, and
1930 TLS 1.3 sends more of such records. Setting SSL_MODE_AUTO_RETRY works
1931 around the problems in those applications, but can also break some.
1932 It's recommended to read the manpages about SSL_read(), SSL_write(),
1933 SSL_get_error(), SSL_shutdown(), SSL_CTX_set_mode() and
1934 SSL_CTX_set_read_ahead() again.
1938 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
1939 now allow empty (zero character) pass phrases.
1943 * Apply blinding to binary field modular inversion and remove patent
1944 pending (OPENSSL_SUN_GF2M_DIV) BN_GF2m_mod_div implementation.
1948 * Deprecate ec2_mult.c and unify scalar multiplication code paths for
1949 binary and prime elliptic curves.
1953 * Remove ECDSA nonce padding: EC_POINT_mul is now responsible for
1954 constant time fixed point multiplication.
1958 * Revise elliptic curve scalar multiplication with timing attack
1959 defenses: ec_wNAF_mul redirects to a constant time implementation
1960 when computing fixed point and variable point multiplication (which
1961 in OpenSSL are mostly used with secret scalars in keygen, sign,
1962 ECDH derive operations).
1963 *Billy Bob Brumley, Nicola Tuveri, Cesar Pereida García,
1966 * Updated CONTRIBUTING
1970 * Updated DRBG / RAND to request nonce and additional low entropy
1971 randomness from the system.
1973 *Matthias St. Pierre*
1975 * Updated 'openssl rehash' to use OpenSSL consistent default.
1979 * Moved the load of the ssl_conf module to libcrypto, which helps
1980 loading engines that libssl uses before libssl is initialised.
1984 * Added EVP_PKEY_sign() and EVP_PKEY_verify() for EdDSA
1988 * Fixed X509_NAME_ENTRY_set to get multi-valued RDNs right in all cases.
1990 *Ingo Schwarze, Rich Salz*
1992 * Added output of accepting IP address and port for 'openssl s_server'
1996 * Added a new API for TLSv1.3 ciphersuites:
1997 SSL_CTX_set_ciphersuites()
1998 SSL_set_ciphersuites()
2002 * Memory allocation failures consistently add an error to the error
2007 * Don't use OPENSSL_ENGINES and OPENSSL_CONF environment values
2008 in libcrypto when run as setuid/setgid.
2012 * Load any config file by default when libssl is used.
2016 * Added new public header file <openssl/rand_drbg.h> and documentation
2017 for the RAND_DRBG API. See manual page RAND_DRBG(7) for an overview.
2019 *Matthias St. Pierre*
2021 * QNX support removed (cannot find contributors to get their approval
2022 for the license change).
2026 * TLSv1.3 replay protection for early data has been implemented. See the
2027 SSL_read_early_data() man page for further details.
2031 * Separated TLSv1.3 ciphersuite configuration out from TLSv1.2 ciphersuite
2032 configuration. TLSv1.3 ciphersuites are not compatible with TLSv1.2 and
2033 below. Similarly TLSv1.2 ciphersuites are not compatible with TLSv1.3.
2034 In order to avoid issues where legacy TLSv1.2 ciphersuite configuration
2035 would otherwise inadvertently disable all TLSv1.3 ciphersuites the
2036 configuration has been separated out. See the ciphers man page or the
2037 SSL_CTX_set_ciphersuites() man page for more information.
2041 * On POSIX (BSD, Linux, ...) systems the ocsp(1) command running
2042 in responder mode now supports the new "-multi" option, which
2043 spawns the specified number of child processes to handle OCSP
2044 requests. The "-timeout" option now also limits the OCSP
2045 responder's patience to wait to receive the full client request
2046 on a newly accepted connection. Child processes are respawned
2047 as needed, and the CA index file is automatically reloaded
2048 when changed. This makes it possible to run the "ocsp" responder
2049 as a long-running service, making the OpenSSL CA somewhat more
2050 feature-complete. In this mode, most diagnostic messages logged
2051 after entering the event loop are logged via syslog(3) rather than
2056 * Added support for X448 and Ed448. Heavily based on original work by
2061 * Extend OSSL_STORE with capabilities to search and to narrow the set of
2062 objects loaded. This adds the functions OSSL_STORE_expect() and
2063 OSSL_STORE_find() as well as needed tools to construct searches and
2064 get the search data out of them.
2068 * Support for TLSv1.3 added. Note that users upgrading from an earlier
2069 version of OpenSSL should review their configuration settings to ensure
2070 that they are still appropriate for TLSv1.3. For further information see:
2071 <https://wiki.openssl.org/index.php/TLS1.3>
2075 * Grand redesign of the OpenSSL random generator
2077 The default RAND method now utilizes an AES-CTR DRBG according to
2078 NIST standard SP 800-90Ar1. The new random generator is essentially
2079 a port of the default random generator from the OpenSSL FIPS 2.0
2080 object module. It is a hybrid deterministic random bit generator
2081 using an AES-CTR bit stream and which seeds and reseeds itself
2082 automatically using trusted system entropy sources.
2084 Some of its new features are:
2085 - Support for multiple DRBG instances with seed chaining.
2086 - The default RAND method makes use of a DRBG.
2087 - There is a public and private DRBG instance.
2088 - The DRBG instances are fork-safe.
2089 - Keep all global DRBG instances on the secure heap if it is enabled.
2090 - The public and private DRBG instance are per thread for lock free
2093 *Paul Dale, Benjamin Kaduk, Kurt Roeckx, Rich Salz, Matthias St. Pierre*
2095 * Changed Configure so it only says what it does and doesn't dump
2096 so much data. Instead, ./configdata.pm should be used as a script
2097 to display all sorts of configuration data.
2101 * Added processing of "make variables" to Configure.
2105 * Added SHA512/224 and SHA512/256 algorithm support.
2109 * The last traces of Netware support, first removed in 1.1.0, have
2114 * Get rid of Makefile.shared, and in the process, make the processing
2115 of certain files (rc.obj, or the .def/.map/.opt files produced from
2116 the ordinal files) more visible and hopefully easier to trace and
2117 debug (or make silent).
2121 * Make it possible to have environment variable assignments as
2122 arguments to config / Configure.
2126 * Add multi-prime RSA (RFC 8017) support.
2130 * Add SM3 implemented according to GB/T 32905-2016
2131 *Jack Lloyd <jack.lloyd@ribose.com>,*
2132 *Ronald Tse <ronald.tse@ribose.com>,*
2133 *Erick Borsboom <erick.borsboom@ribose.com>*
2135 * Add 'Maximum Fragment Length' TLS extension negotiation and support
2136 as documented in RFC6066.
2137 Based on a patch from Tomasz Moń
2139 *Filipe Raimundo da Silva*
2141 * Add SM4 implemented according to GB/T 32907-2016.
2142 *Jack Lloyd <jack.lloyd@ribose.com>,*
2143 *Ronald Tse <ronald.tse@ribose.com>,*
2144 *Erick Borsboom <erick.borsboom@ribose.com>*
2146 * Reimplement -newreq-nodes and ERR_error_string_n; the
2147 original author does not agree with the license change.
2151 * Add ARIA AEAD TLS support.
2155 * Some macro definitions to support VS6 have been removed. Visual
2156 Studio 6 has not worked since 1.1.0
2160 * Add ERR_clear_last_mark(), to allow callers to clear the last mark
2161 without clearing the errors.
2165 * Add "atfork" functions. If building on a system that without
2166 pthreads, see doc/man3/OPENSSL_fork_prepare.pod for application
2167 requirements. The RAND facility now uses/requires this.
2175 * The UI API becomes a permanent and integral part of libcrypto, i.e.
2176 not possible to disable entirely. However, it's still possible to
2177 disable the console reading UI method, UI_OpenSSL() (use UI_null()
2180 To disable, configure with 'no-ui-console'. 'no-ui' is still
2181 possible to use as an alias. Check at compile time with the
2182 macro OPENSSL_NO_UI_CONSOLE. The macro OPENSSL_NO_UI is still
2183 possible to check and is an alias for OPENSSL_NO_UI_CONSOLE.
2187 * Add a STORE module, which implements a uniform and URI based reader of
2188 stores that can contain keys, certificates, CRLs and numerous other
2189 objects. The main API is loosely based on a few stdio functions,
2190 and includes OSSL_STORE_open, OSSL_STORE_load, OSSL_STORE_eof,
2191 OSSL_STORE_error and OSSL_STORE_close.
2192 The implementation uses backends called "loaders" to implement arbitrary
2193 URI schemes. There is one built in "loader" for the 'file' scheme.
2197 * Add devcrypto engine. This has been implemented against cryptodev-linux,
2198 then adjusted to work on FreeBSD 8.4 as well.
2199 Enable by configuring with 'enable-devcryptoeng'. This is done by default
2200 on BSD implementations, as cryptodev.h is assumed to exist on all of them.
2204 * Module names can prefixed with OSSL_ or OPENSSL_. This affects
2205 util/mkerr.pl, which is adapted to allow those prefixes, leading to
2206 error code calls like this:
2208 OSSL_FOOerr(OSSL_FOO_F_SOMETHING, OSSL_FOO_R_WHATEVER);
2210 With this change, we claim the namespaces OSSL and OPENSSL in a manner
2211 that can be encoded in C. For the foreseeable future, this will only
2214 *Richard Levitte and Tim Hudson*
2216 * Removed BSD cryptodev engine.
2220 * Add a build target 'build_all_generated', to build all generated files
2221 and only that. This can be used to prepare everything that requires
2222 things like perl for a system that lacks perl and then move everything
2223 to that system and do the rest of the build there.
2227 * In the UI interface, make it possible to duplicate the user data. This
2228 can be used by engines that need to retain the data for a longer time
2229 than just the call where this user data is passed.
2233 * Ignore the '-named_curve auto' value for compatibility of applications
2236 *Tomáš Mráz <tmraz@fedoraproject.org>*
2238 * Fragmented SSL/TLS alerts are no longer accepted. An alert message is 2
2239 bytes long. In theory it is permissible in SSLv3 - TLSv1.2 to fragment such
2240 alerts across multiple records (some of which could be empty). In practice
2241 it make no sense to send an empty alert record, or to fragment one. TLSv1.3
2242 prohibits this altogether and other libraries (BoringSSL, NSS) do not
2243 support this at all. Supporting it adds significant complexity to the
2244 record layer, and its removal is unlikely to cause interoperability
2249 * Add the ASN.1 types INT32, UINT32, INT64, UINT64 and variants prefixed
2250 with Z. These are meant to replace LONG and ZLONG and to be size safe.
2251 The use of LONG and ZLONG is discouraged and scheduled for deprecation
2256 * Add the 'z' and 'j' modifiers to BIO_printf() et al formatting string,
2257 'z' is to be used for [s]size_t, and 'j' - with [u]int64_t.
2259 *Richard Levitte, Andy Polyakov*
2261 * Add EC_KEY_get0_engine(), which does for EC_KEY what RSA_get0_engine()
2266 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
2267 platform rather than 'mingw'.
2271 * The functions X509_STORE_add_cert and X509_STORE_add_crl return
2272 success if they are asked to add an object which already exists
2273 in the store. This change cascades to other functions which load
2274 certificates and CRLs.
2278 * x86_64 assembly pack: annotate code with DWARF CFI directives to
2279 facilitate stack unwinding even from assembly subroutines.
2283 * Remove VAX C specific definitions of OPENSSL_EXPORT, OPENSSL_EXTERN.
2284 Also remove OPENSSL_GLOBAL entirely, as it became a no-op.
2288 * Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
2289 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
2290 which is the minimum version we support.
2294 * Certificate time validation (X509_cmp_time) enforces stricter
2295 compliance with RFC 5280. Fractional seconds and timezone offsets
2296 are no longer allowed.
2300 * Add support for ARIA
2304 * s_client will now send the Server Name Indication (SNI) extension by
2305 default unless the new "-noservername" option is used. The server name is
2306 based on the host provided to the "-connect" option unless overridden by
2307 using "-servername".
2311 * Add support for SipHash
2315 * OpenSSL now fails if it receives an unrecognised record type in TLS1.0
2316 or TLS1.1. Previously this only happened in SSLv3 and TLS1.2. This is to
2317 prevent issues where no progress is being made and the peer continually
2318 sends unrecognised record types, using up resources processing them.
2322 * 'openssl passwd' can now produce SHA256 and SHA512 based output,
2323 using the algorithm defined in
2324 <https://www.akkadia.org/drepper/SHA-crypt.txt>
2328 * Heartbeat support has been removed; the ABI is changed for now.
2330 *Richard Levitte, Rich Salz*
2332 * Support for SSL_OP_NO_ENCRYPT_THEN_MAC in SSL_CONF_cmd.
2336 * The RSA "null" method, which was partially supported to avoid patent
2337 issues, has been replaced to always returns NULL.
2344 ### Changes between 1.1.0k and 1.1.0l [10 Sep 2019]
2346 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
2347 used even when parsing explicit parameters, when loading a encoded key
2348 or calling `EC_GROUP_new_from_ecpkparameters()`/
2349 `EC_GROUP_new_from_ecparameters()`.
2350 This prevents bypass of security hardening and performance gains,
2351 especially for curves with specialized EC_METHODs.
2352 By default, if a key encoded with explicit parameters is loaded and later
2353 encoded, the output is still encoded with explicit parameters, even if
2354 internally a "named" EC_GROUP is used for computation.
2358 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
2359 this change, EC_GROUP_set_generator would accept order and/or cofactor as
2360 NULL. After this change, only the cofactor parameter can be NULL. It also
2361 does some minimal sanity checks on the passed order.
2366 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
2367 An attack is simple, if the first CMS_recipientInfo is valid but the
2368 second CMS_recipientInfo is chosen ciphertext. If the second
2369 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
2370 encryption key will be replaced by garbage, and the message cannot be
2371 decoded, but if the RSA decryption fails, the correct encryption key is
2372 used and the recipient will not notice the attack.
2373 As a work around for this potential attack the length of the decrypted
2374 key must be equal to the cipher default key length, in case the
2375 certifiate is not given and all recipientInfo are tried out.
2376 The old behaviour can be re-enabled in the CMS code by setting the
2377 CMS_DEBUG_DECRYPT flag.
2382 * Use Windows installation paths in the mingw builds
2384 Mingw isn't a POSIX environment per se, which means that Windows
2385 paths should be used for installation.
2390 ### Changes between 1.1.0j and 1.1.0k [28 May 2019]
2392 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
2393 This changes the size when using the `genpkey` command when no size is given.
2394 It fixes an omission in earlier changes that changed all RSA, DSA and DH
2395 generation commands to use 2048 bits by default.
2399 * Prevent over long nonces in ChaCha20-Poly1305.
2401 ChaCha20-Poly1305 is an AEAD cipher, and requires a unique nonce input
2402 for every encryption operation. RFC 7539 specifies that the nonce value
2403 (IV) should be 96 bits (12 bytes). OpenSSL allows a variable nonce length
2404 and front pads the nonce with 0 bytes if it is less than 12
2405 bytes. However it also incorrectly allows a nonce to be set of up to 16
2406 bytes. In this case only the last 12 bytes are significant and any
2407 additional leading bytes are ignored.
2409 It is a requirement of using this cipher that nonce values are
2410 unique. Messages encrypted using a reused nonce value are susceptible to
2411 serious confidentiality and integrity attacks. If an application changes
2412 the default nonce length to be longer than 12 bytes and then makes a
2413 change to the leading bytes of the nonce expecting the new value to be a
2414 new unique nonce then such an application could inadvertently encrypt
2415 messages with a reused nonce.
2417 Additionally the ignored bytes in a long nonce are not covered by the
2418 integrity guarantee of this cipher. Any application that relies on the
2419 integrity of these ignored leading bytes of a long nonce may be further
2420 affected. Any OpenSSL internal use of this cipher, including in SSL/TLS,
2421 is safe because no such use sets such a long nonce value. However user
2422 applications that use this cipher directly and set a non-default nonce
2423 length to be longer than 12 bytes may be vulnerable.
2425 This issue was reported to OpenSSL on 16th of March 2019 by Joran Dirk
2431 * Added SCA hardening for modular field inversion in EC_GROUP through
2432 a new dedicated field_inv() pointer in EC_METHOD.
2433 This also addresses a leakage affecting conversions from projective
2434 to affine coordinates.
2436 *Billy Bob Brumley, Nicola Tuveri*
2438 * Fix a use after free bug in d2i_X509_PUBKEY when overwriting a
2439 re-used X509_PUBKEY object if the second PUBKEY is malformed.
2443 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
2447 * Remove the 'dist' target and add a tarball building script. The
2448 'dist' target has fallen out of use, and it shouldn't be
2449 necessary to configure just to create a source distribution.
2453 ### Changes between 1.1.0i and 1.1.0j [20 Nov 2018]
2455 * Timing vulnerability in DSA signature generation
2457 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
2458 timing side channel attack. An attacker could use variations in the signing
2459 algorithm to recover the private key.
2461 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
2466 * Timing vulnerability in ECDSA signature generation
2468 The OpenSSL ECDSA signature algorithm has been shown to be vulnerable to a
2469 timing side channel attack. An attacker could use variations in the signing
2470 algorithm to recover the private key.
2472 This issue was reported to OpenSSL on 25th October 2018 by Samuel Weiser.
2477 * Add coordinate blinding for EC_POINT and implement projective
2478 coordinate blinding for generic prime curves as a countermeasure to
2479 chosen point SCA attacks.
2481 *Sohaib ul Hassan, Nicola Tuveri, Billy Bob Brumley*
2483 ### Changes between 1.1.0h and 1.1.0i [14 Aug 2018]
2485 * Client DoS due to large DH parameter
2487 During key agreement in a TLS handshake using a DH(E) based ciphersuite a
2488 malicious server can send a very large prime value to the client. This will
2489 cause the client to spend an unreasonably long period of time generating a
2490 key for this prime resulting in a hang until the client has finished. This
2491 could be exploited in a Denial Of Service attack.
2493 This issue was reported to OpenSSL on 5th June 2018 by Guido Vranken
2498 * Cache timing vulnerability in RSA Key Generation
2500 The OpenSSL RSA Key generation algorithm has been shown to be vulnerable to
2501 a cache timing side channel attack. An attacker with sufficient access to
2502 mount cache timing attacks during the RSA key generation process could
2503 recover the private key.
2505 This issue was reported to OpenSSL on 4th April 2018 by Alejandro Cabrera
2506 Aldaya, Billy Brumley, Cesar Pereida Garcia and Luis Manuel Alvarez Tapia.
2511 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
2512 parameter is no longer accepted, as it leads to a corrupt table. NULL
2513 pem_str is reserved for alias entries only.
2517 * Revert blinding in ECDSA sign and instead make problematic addition
2518 length-invariant. Switch even to fixed-length Montgomery multiplication.
2522 * Change generating and checking of primes so that the error rate of not
2523 being prime depends on the intended use based on the size of the input.
2524 For larger primes this will result in more rounds of Miller-Rabin.
2525 The maximal error rate for primes with more than 1080 bits is lowered
2528 *Kurt Roeckx, Annie Yousar*
2530 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
2534 * Add blinding to ECDSA and DSA signatures to protect against side channel
2535 attacks discovered by Keegan Ryan (NCC Group).
2539 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
2540 now allow empty (zero character) pass phrases.
2544 * Certificate time validation (X509_cmp_time) enforces stricter
2545 compliance with RFC 5280. Fractional seconds and timezone offsets
2546 are no longer allowed.
2550 * Fixed a text canonicalisation bug in CMS
2552 Where a CMS detached signature is used with text content the text goes
2553 through a canonicalisation process first prior to signing or verifying a
2554 signature. This process strips trailing space at the end of lines, converts
2555 line terminators to CRLF and removes additional trailing line terminators
2556 at the end of a file. A bug in the canonicalisation process meant that
2557 some characters, such as form-feed, were incorrectly treated as whitespace
2558 and removed. This is contrary to the specification (RFC5485). This fix
2559 could mean that detached text data signed with an earlier version of
2560 OpenSSL 1.1.0 may fail to verify using the fixed version, or text data
2561 signed with a fixed OpenSSL may fail to verify with an earlier version of
2562 OpenSSL 1.1.0. A workaround is to only verify the canonicalised text data
2563 and use the "-binary" flag (for the "cms" command line application) or set
2564 the SMIME_BINARY/PKCS7_BINARY/CMS_BINARY flags (if using CMS_verify()).
2568 ### Changes between 1.1.0g and 1.1.0h [27 Mar 2018]
2570 * Constructed ASN.1 types with a recursive definition could exceed the stack
2572 Constructed ASN.1 types with a recursive definition (such as can be found
2573 in PKCS7) could eventually exceed the stack given malicious input with
2574 excessive recursion. This could result in a Denial Of Service attack. There
2575 are no such structures used within SSL/TLS that come from untrusted sources
2576 so this is considered safe.
2578 This issue was reported to OpenSSL on 4th January 2018 by the OSS-fuzz
2584 * Incorrect CRYPTO_memcmp on HP-UX PA-RISC
2586 Because of an implementation bug the PA-RISC CRYPTO_memcmp function is
2587 effectively reduced to only comparing the least significant bit of each
2588 byte. This allows an attacker to forge messages that would be considered as
2589 authenticated in an amount of tries lower than that guaranteed by the
2590 security claims of the scheme. The module can only be compiled by the
2591 HP-UX assembler, so that only HP-UX PA-RISC targets are affected.
2593 This issue was reported to OpenSSL on 2nd March 2018 by Peter Waltenberg
2599 * Add a build target 'build_all_generated', to build all generated files
2600 and only that. This can be used to prepare everything that requires
2601 things like perl for a system that lacks perl and then move everything
2602 to that system and do the rest of the build there.
2606 * Backport SSL_OP_NO_RENGOTIATION
2608 OpenSSL 1.0.2 and below had the ability to disable renegotiation using the
2609 (undocumented) SSL3_FLAGS_NO_RENEGOTIATE_CIPHERS flag. Due to the opacity
2610 changes this is no longer possible in 1.1.0. Therefore the new
2611 SSL_OP_NO_RENEGOTIATION option from 1.1.1-dev has been backported to
2612 1.1.0 to provide equivalent functionality.
2614 Note that if an application built against 1.1.0h headers (or above) is run
2615 using an older version of 1.1.0 (prior to 1.1.0h) then the option will be
2616 accepted but nothing will happen, i.e. renegotiation will not be prevented.
2620 * Removed the OS390-Unix config target. It relied on a script that doesn't
2625 * rsaz_1024_mul_avx2 overflow bug on x86_64
2627 There is an overflow bug in the AVX2 Montgomery multiplication procedure
2628 used in exponentiation with 1024-bit moduli. No EC algorithms are affected.
2629 Analysis suggests that attacks against RSA and DSA as a result of this
2630 defect would be very difficult to perform and are not believed likely.
2631 Attacks against DH1024 are considered just feasible, because most of the
2632 work necessary to deduce information about a private key may be performed
2633 offline. The amount of resources required for such an attack would be
2634 significant. However, for an attack on TLS to be meaningful, the server
2635 would have to share the DH1024 private key among multiple clients, which is
2636 no longer an option since CVE-2016-0701.
2638 This only affects processors that support the AVX2 but not ADX extensions
2639 like Intel Haswell (4th generation).
2641 This issue was reported to OpenSSL by David Benjamin (Google). The issue
2642 was originally found via the OSS-Fuzz project.
2647 ### Changes between 1.1.0f and 1.1.0g [2 Nov 2017]
2649 * bn_sqrx8x_internal carry bug on x86_64
2651 There is a carry propagating bug in the x86_64 Montgomery squaring
2652 procedure. No EC algorithms are affected. Analysis suggests that attacks
2653 against RSA and DSA as a result of this defect would be very difficult to
2654 perform and are not believed likely. Attacks against DH are considered just
2655 feasible (although very difficult) because most of the work necessary to
2656 deduce information about a private key may be performed offline. The amount
2657 of resources required for such an attack would be very significant and
2658 likely only accessible to a limited number of attackers. An attacker would
2659 additionally need online access to an unpatched system using the target
2660 private key in a scenario with persistent DH parameters and a private
2661 key that is shared between multiple clients.
2663 This only affects processors that support the BMI1, BMI2 and ADX extensions
2664 like Intel Broadwell (5th generation) and later or AMD Ryzen.
2666 This issue was reported to OpenSSL by the OSS-Fuzz project.
2671 * Malformed X.509 IPAddressFamily could cause OOB read
2673 If an X.509 certificate has a malformed IPAddressFamily extension,
2674 OpenSSL could do a one-byte buffer overread. The most likely result
2675 would be an erroneous display of the certificate in text format.
2677 This issue was reported to OpenSSL by the OSS-Fuzz project.
2682 ### Changes between 1.1.0e and 1.1.0f [25 May 2017]
2684 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
2685 platform rather than 'mingw'.
2689 * Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
2690 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
2691 which is the minimum version we support.
2695 ### Changes between 1.1.0d and 1.1.0e [16 Feb 2017]
2697 * Encrypt-Then-Mac renegotiation crash
2699 During a renegotiation handshake if the Encrypt-Then-Mac extension is
2700 negotiated where it was not in the original handshake (or vice-versa) then
2701 this can cause OpenSSL to crash (dependant on ciphersuite). Both clients
2702 and servers are affected.
2704 This issue was reported to OpenSSL by Joe Orton (Red Hat).
2709 ### Changes between 1.1.0c and 1.1.0d [26 Jan 2017]
2711 * Truncated packet could crash via OOB read
2713 If one side of an SSL/TLS path is running on a 32-bit host and a specific
2714 cipher is being used, then a truncated packet can cause that host to
2715 perform an out-of-bounds read, usually resulting in a crash.
2717 This issue was reported to OpenSSL by Robert Święcki of Google.
2722 * Bad (EC)DHE parameters cause a client crash
2724 If a malicious server supplies bad parameters for a DHE or ECDHE key
2725 exchange then this can result in the client attempting to dereference a
2726 NULL pointer leading to a client crash. This could be exploited in a Denial
2729 This issue was reported to OpenSSL by Guido Vranken.
2734 * BN_mod_exp may produce incorrect results on x86_64
2736 There is a carry propagating bug in the x86_64 Montgomery squaring
2737 procedure. No EC algorithms are affected. Analysis suggests that attacks
2738 against RSA and DSA as a result of this defect would be very difficult to
2739 perform and are not believed likely. Attacks against DH are considered just
2740 feasible (although very difficult) because most of the work necessary to
2741 deduce information about a private key may be performed offline. The amount
2742 of resources required for such an attack would be very significant and
2743 likely only accessible to a limited number of attackers. An attacker would
2744 additionally need online access to an unpatched system using the target
2745 private key in a scenario with persistent DH parameters and a private
2746 key that is shared between multiple clients. For example this can occur by
2747 default in OpenSSL DHE based SSL/TLS ciphersuites. Note: This issue is very
2748 similar to CVE-2015-3193 but must be treated as a separate problem.
2750 This issue was reported to OpenSSL by the OSS-Fuzz project.
2755 ### Changes between 1.1.0b and 1.1.0c [10 Nov 2016]
2757 * ChaCha20/Poly1305 heap-buffer-overflow
2759 TLS connections using `*-CHACHA20-POLY1305` ciphersuites are susceptible to
2760 a DoS attack by corrupting larger payloads. This can result in an OpenSSL
2761 crash. This issue is not considered to be exploitable beyond a DoS.
2763 This issue was reported to OpenSSL by Robert Święcki (Google Security Team)
2768 * CMS Null dereference
2770 Applications parsing invalid CMS structures can crash with a NULL pointer
2771 dereference. This is caused by a bug in the handling of the ASN.1 CHOICE
2772 type in OpenSSL 1.1.0 which can result in a NULL value being passed to the
2773 structure callback if an attempt is made to free certain invalid encodings.
2774 Only CHOICE structures using a callback which do not handle NULL value are
2777 This issue was reported to OpenSSL by Tyler Nighswander of ForAllSecure.
2782 * Montgomery multiplication may produce incorrect results
2784 There is a carry propagating bug in the Broadwell-specific Montgomery
2785 multiplication procedure that handles input lengths divisible by, but
2786 longer than 256 bits. Analysis suggests that attacks against RSA, DSA
2787 and DH private keys are impossible. This is because the subroutine in
2788 question is not used in operations with the private key itself and an input
2789 of the attacker's direct choice. Otherwise the bug can manifest itself as
2790 transient authentication and key negotiation failures or reproducible
2791 erroneous outcome of public-key operations with specially crafted input.
2792 Among EC algorithms only Brainpool P-512 curves are affected and one
2793 presumably can attack ECDH key negotiation. Impact was not analyzed in
2794 detail, because pre-requisites for attack are considered unlikely. Namely
2795 multiple clients have to choose the curve in question and the server has to
2796 share the private key among them, neither of which is default behaviour.
2797 Even then only clients that chose the curve will be affected.
2799 This issue was publicly reported as transient failures and was not
2800 initially recognized as a security issue. Thanks to Richard Morgan for
2801 providing reproducible case.
2806 * Removed automatic addition of RPATH in shared libraries and executables,
2807 as this was a remainder from OpenSSL 1.0.x and isn't needed any more.
2811 ### Changes between 1.1.0a and 1.1.0b [26 Sep 2016]
2813 * Fix Use After Free for large message sizes
2815 The patch applied to address CVE-2016-6307 resulted in an issue where if a
2816 message larger than approx 16k is received then the underlying buffer to
2817 store the incoming message is reallocated and moved. Unfortunately a
2818 dangling pointer to the old location is left which results in an attempt to
2819 write to the previously freed location. This is likely to result in a
2820 crash, however it could potentially lead to execution of arbitrary code.
2822 This issue only affects OpenSSL 1.1.0a.
2824 This issue was reported to OpenSSL by Robert Święcki.
2829 ### Changes between 1.1.0 and 1.1.0a [22 Sep 2016]
2831 * OCSP Status Request extension unbounded memory growth
2833 A malicious client can send an excessively large OCSP Status Request
2834 extension. If that client continually requests renegotiation, sending a
2835 large OCSP Status Request extension each time, then there will be unbounded
2836 memory growth on the server. This will eventually lead to a Denial Of
2837 Service attack through memory exhaustion. Servers with a default
2838 configuration are vulnerable even if they do not support OCSP. Builds using
2839 the "no-ocsp" build time option are not affected.
2841 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
2846 * SSL_peek() hang on empty record
2848 OpenSSL 1.1.0 SSL/TLS will hang during a call to SSL_peek() if the peer
2849 sends an empty record. This could be exploited by a malicious peer in a
2850 Denial Of Service attack.
2852 This issue was reported to OpenSSL by Alex Gaynor.
2857 * Excessive allocation of memory in tls_get_message_header() and
2858 dtls1_preprocess_fragment()
2860 A (D)TLS message includes 3 bytes for its length in the header for the
2861 message. This would allow for messages up to 16Mb in length. Messages of
2862 this length are excessive and OpenSSL includes a check to ensure that a
2863 peer is sending reasonably sized messages in order to avoid too much memory
2864 being consumed to service a connection. A flaw in the logic of version
2865 1.1.0 means that memory for the message is allocated too early, prior to
2866 the excessive message length check. Due to way memory is allocated in
2867 OpenSSL this could mean an attacker could force up to 21Mb to be allocated
2868 to service a connection. This could lead to a Denial of Service through
2869 memory exhaustion. However, the excessive message length check still takes
2870 place, and this would cause the connection to immediately fail. Assuming
2871 that the application calls SSL_free() on the failed connection in a timely
2872 manner then the 21Mb of allocated memory will then be immediately freed
2873 again. Therefore the excessive memory allocation will be transitory in
2874 nature. This then means that there is only a security impact if:
2876 1) The application does not call SSL_free() in a timely manner in the event
2877 that the connection fails
2879 2) The application is working in a constrained environment where there is
2880 very little free memory
2882 3) The attacker initiates multiple connection attempts such that there are
2883 multiple connections in a state where memory has been allocated for the
2884 connection; SSL_free() has not yet been called; and there is insufficient
2885 memory to service the multiple requests.
2887 Except in the instance of (1) above any Denial Of Service is likely to be
2888 transitory because as soon as the connection fails the memory is
2889 subsequently freed again in the SSL_free() call. However there is an
2890 increased risk during this period of application crashes due to the lack of
2891 memory - which would then mean a more serious Denial of Service.
2893 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
2894 (CVE-2016-6307 and CVE-2016-6308)
2898 * solaris-x86-cc, i.e. 32-bit configuration with vendor compiler,
2899 had to be removed. Primary reason is that vendor assembler can't
2900 assemble our modules with -KPIC flag. As result it, assembly
2901 support, was not even available as option. But its lack means
2902 lack of side-channel resistant code, which is incompatible with
2903 security by todays standards. Fortunately gcc is readily available
2904 prepackaged option, which we firmly point at...
2908 ### Changes between 1.0.2h and 1.1.0 [25 Aug 2016]
2910 * Windows command-line tool supports UTF-8 opt-in option for arguments
2911 and console input. Setting OPENSSL_WIN32_UTF8 environment variable
2912 (to any value) allows Windows user to access PKCS#12 file generated
2913 with Windows CryptoAPI and protected with non-ASCII password, as well
2914 as files generated under UTF-8 locale on Linux also protected with
2919 * To mitigate the SWEET32 attack ([CVE-2016-2183]), 3DES cipher suites
2920 have been disabled by default and removed from DEFAULT, just like RC4.
2921 See the RC4 item below to re-enable both.
2925 * The method for finding the storage location for the Windows RAND seed file
2926 has changed. First we check %RANDFILE%. If that is not set then we check
2927 the directories %HOME%, %USERPROFILE% and %SYSTEMROOT% in that order. If
2928 all else fails we fall back to C:\.
2932 * The EVP_EncryptUpdate() function has had its return type changed from void
2933 to int. A return of 0 indicates and error while a return of 1 indicates
2938 * The flags RSA_FLAG_NO_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME and
2939 DH_FLAG_NO_EXP_CONSTTIME which previously provided the ability to switch
2940 off the constant time implementation for RSA, DSA and DH have been made
2941 no-ops and deprecated.
2945 * Windows RAND implementation was simplified to only get entropy by
2946 calling CryptGenRandom(). Various other RAND-related tickets
2949 *Joseph Wylie Yandle, Rich Salz*
2951 * The stack and lhash API's were renamed to start with `OPENSSL_SK_`
2952 and `OPENSSL_LH_`, respectively. The old names are available
2953 with API compatibility. They new names are now completely documented.
2957 * Unify TYPE_up_ref(obj) methods signature.
2958 SSL_CTX_up_ref(), SSL_up_ref(), X509_up_ref(), EVP_PKEY_up_ref(),
2959 X509_CRL_up_ref(), X509_OBJECT_up_ref_count() methods are now returning an
2960 int (instead of void) like all others TYPE_up_ref() methods.
2961 So now these methods also check the return value of CRYPTO_atomic_add(),
2962 and the validity of object reference counter.
2964 *fdasilvayy@gmail.com*
2966 * With Windows Visual Studio builds, the .pdb files are installed
2967 alongside the installed libraries and executables. For a static
2968 library installation, ossl_static.pdb is the associate compiler
2969 generated .pdb file to be used when linking programs.
2973 * Remove openssl.spec. Packaging files belong with the packagers.
2977 * Automatic Darwin/OSX configuration has had a refresh, it will now
2978 recognise x86_64 architectures automatically. You can still decide
2979 to build for a different bitness with the environment variable
2980 KERNEL_BITS (can be 32 or 64), for example:
2982 KERNEL_BITS=32 ./config
2986 * Change default algorithms in pkcs8 utility to use PKCS#5 v2.0,
2987 256 bit AES and HMAC with SHA256.
2991 * Remove support for MIPS o32 ABI on IRIX (and IRIX only).
2995 * Triple-DES ciphers have been moved from HIGH to MEDIUM.
2999 * To enable users to have their own config files and build file templates,
3000 Configure looks in the directory indicated by the environment variable
3001 OPENSSL_LOCAL_CONFIG_DIR as well as the in-source Configurations/
3002 directory. On VMS, OPENSSL_LOCAL_CONFIG_DIR is expected to be a logical
3003 name and is used as is.
3007 * The following datatypes were made opaque: X509_OBJECT, X509_STORE_CTX,
3008 X509_STORE, X509_LOOKUP, and X509_LOOKUP_METHOD. The unused type
3009 X509_CERT_FILE_CTX was removed.
3013 * "shared" builds are now the default. To create only static libraries use
3014 the "no-shared" Configure option.
3018 * Remove the no-aes, no-hmac, no-rsa, no-sha and no-md5 Configure options.
3019 All of these option have not worked for some while and are fundamental
3024 * Make various cleanup routines no-ops and mark them as deprecated. Most
3025 global cleanup functions are no longer required because they are handled
3026 via auto-deinit (see OPENSSL_init_crypto and OPENSSL_init_ssl man pages).
3027 Explicitly de-initing can cause problems (e.g. where a library that uses
3028 OpenSSL de-inits, but an application is still using it). The affected
3029 functions are CONF_modules_free(), ENGINE_cleanup(), OBJ_cleanup(),
3030 EVP_cleanup(), BIO_sock_cleanup(), CRYPTO_cleanup_all_ex_data(),
3031 RAND_cleanup(), SSL_COMP_free_compression_methods(), ERR_free_strings() and
3032 COMP_zlib_cleanup().
3036 * --strict-warnings no longer enables runtime debugging options
3037 such as REF_DEBUG. Instead, debug options are automatically
3038 enabled with '--debug' builds.
3040 *Andy Polyakov, Emilia Käsper*
3042 * Made DH and DH_METHOD opaque. The structures for managing DH objects
3043 have been moved out of the public header files. New functions for managing
3044 these have been added.
3048 * Made RSA and RSA_METHOD opaque. The structures for managing RSA
3049 objects have been moved out of the public header files. New
3050 functions for managing these have been added.
3054 * Made DSA and DSA_METHOD opaque. The structures for managing DSA objects
3055 have been moved out of the public header files. New functions for managing
3056 these have been added.
3060 * Made BIO and BIO_METHOD opaque. The structures for managing BIOs have been
3061 moved out of the public header files. New functions for managing these
3066 * Removed no-rijndael as a config option. Rijndael is an old name for AES.
3070 * Removed the mk1mf build scripts.
3074 * Headers are now wrapped, if necessary, with OPENSSL_NO_xxx, so
3075 it is always safe to #include a header now.
3079 * Removed the aged BC-32 config and all its supporting scripts
3083 * Removed support for Ultrix, Netware, and OS/2.
3087 * Add support for HKDF.
3089 *Alessandro Ghedini*
3091 * Add support for blake2b and blake2s
3095 * Added support for "pipelining". Ciphers that have the
3096 EVP_CIPH_FLAG_PIPELINE flag set have a capability to process multiple
3097 encryptions/decryptions simultaneously. There are currently no built-in
3098 ciphers with this property but the expectation is that engines will be able
3099 to offer it to significantly improve throughput. Support has been extended
3100 into libssl so that multiple records for a single connection can be
3101 processed in one go (for >=TLS 1.1).
3105 * Added the AFALG engine. This is an async capable engine which is able to
3106 offload work to the Linux kernel. In this initial version it only supports
3107 AES128-CBC. The kernel must be version 4.1.0 or greater.
3111 * OpenSSL now uses a new threading API. It is no longer necessary to
3112 set locking callbacks to use OpenSSL in a multi-threaded environment. There
3113 are two supported threading models: pthreads and windows threads. It is
3114 also possible to configure OpenSSL at compile time for "no-threads". The
3115 old threading API should no longer be used. The functions have been
3116 replaced with "no-op" compatibility macros.
3118 *Alessandro Ghedini, Matt Caswell*
3120 * Modify behavior of ALPN to invoke callback after SNI/servername
3121 callback, such that updates to the SSL_CTX affect ALPN.
3125 * Add SSL_CIPHER queries for authentication and key-exchange.
3129 * Changes to the DEFAULT cipherlist:
3130 - Prefer (EC)DHE handshakes over plain RSA.
3131 - Prefer AEAD ciphers over legacy ciphers.
3132 - Prefer ECDSA over RSA when both certificates are available.
3133 - Prefer TLSv1.2 ciphers/PRF.
3134 - Remove DSS, SEED, IDEA, CAMELLIA, and AES-CCM from the
3139 * Change the ECC default curve list to be this, in order: x25519,
3140 secp256r1, secp521r1, secp384r1.
3144 * RC4 based libssl ciphersuites are now classed as "weak" ciphers and are
3145 disabled by default. They can be re-enabled using the
3146 enable-weak-ssl-ciphers option to Configure.
3150 * If the server has ALPN configured, but supports no protocols that the
3151 client advertises, send a fatal "no_application_protocol" alert.
3152 This behaviour is SHALL in RFC 7301, though it isn't universally
3153 implemented by other servers.
3157 * Add X25519 support.
3158 Add ASN.1 and EVP_PKEY methods for X25519. This includes support
3159 for public and private key encoding using the format documented in
3160 draft-ietf-curdle-pkix-02. The corresponding EVP_PKEY method supports
3161 key generation and key derivation.
3163 TLS support complies with draft-ietf-tls-rfc4492bis-08 and uses
3168 * Deprecate SRP_VBASE_get_by_user.
3169 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
3170 In order to fix an unavoidable memory leak ([CVE-2016-0798]),
3171 SRP_VBASE_get_by_user was changed to ignore the "fake user" SRP
3172 seed, even if the seed is configured.
3174 Users should use SRP_VBASE_get1_by_user instead. Note that in
3175 SRP_VBASE_get1_by_user, caller must free the returned value. Note
3176 also that even though configuring the SRP seed attempts to hide
3177 invalid usernames by continuing the handshake with fake
3178 credentials, this behaviour is not constant time and no strong
3179 guarantees are made that the handshake is indistinguishable from
3180 that of a valid user.
3184 * Configuration change; it's now possible to build dynamic engines
3185 without having to build shared libraries and vice versa. This
3186 only applies to the engines in `engines/`, those in `crypto/engine/`
3187 will always be built into libcrypto (i.e. "static").
3189 Building dynamic engines is enabled by default; to disable, use
3190 the configuration option "disable-dynamic-engine".
3192 The only requirements for building dynamic engines are the
3193 presence of the DSO module and building with position independent
3194 code, so they will also automatically be disabled if configuring
3195 with "disable-dso" or "disable-pic".
3197 The macros OPENSSL_NO_STATIC_ENGINE and OPENSSL_NO_DYNAMIC_ENGINE
3198 are also taken away from openssl/opensslconf.h, as they are
3203 * Configuration change; if there is a known flag to compile
3204 position independent code, it will always be applied on the
3205 libcrypto and libssl object files, and never on the application
3206 object files. This means other libraries that use routines from
3207 libcrypto / libssl can be made into shared libraries regardless
3208 of how OpenSSL was configured.
3210 If this isn't desirable, the configuration options "disable-pic"
3211 or "no-pic" can be used to disable the use of PIC. This will
3212 also disable building shared libraries and dynamic engines.
3216 * Removed JPAKE code. It was experimental and has no wide use.
3220 * The INSTALL_PREFIX Makefile variable has been renamed to
3221 DESTDIR. That makes for less confusion on what this variable
3222 is for. Also, the configuration option --install_prefix is
3227 * Heartbeat for TLS has been removed and is disabled by default
3228 for DTLS; configure with enable-heartbeats. Code that uses the
3229 old #define's might need to be updated.
3231 *Emilia Käsper, Rich Salz*
3233 * Rename REF_CHECK to REF_DEBUG.
3237 * New "unified" build system
3239 The "unified" build system is aimed to be a common system for all
3240 platforms we support. With it comes new support for VMS.
3242 This system builds supports building in a different directory tree
3243 than the source tree. It produces one Makefile (for unix family
3244 or lookalikes), or one descrip.mms (for VMS).
3246 The source of information to make the Makefile / descrip.mms is
3247 small files called 'build.info', holding the necessary
3248 information for each directory with source to compile, and a
3249 template in Configurations, like unix-Makefile.tmpl or
3252 With this change, the library names were also renamed on Windows
3253 and on VMS. They now have names that are closer to the standard
3254 on Unix, and include the major version number, and in certain
3255 cases, the architecture they are built for. See "Notes on shared
3256 libraries" in INSTALL.
3258 We rely heavily on the perl module Text::Template.
3262 * Added support for auto-initialisation and de-initialisation of the library.
3263 OpenSSL no longer requires explicit init or deinit routines to be called,
3264 except in certain circumstances. See the OPENSSL_init_crypto() and
3265 OPENSSL_init_ssl() man pages for further information.
3269 * The arguments to the DTLSv1_listen function have changed. Specifically the
3270 "peer" argument is now expected to be a BIO_ADDR object.
3272 * Rewrite of BIO networking library. The BIO library lacked consistent
3273 support of IPv6, and adding it required some more extensive
3274 modifications. This introduces the BIO_ADDR and BIO_ADDRINFO types,
3275 which hold all types of addresses and chains of address information.
3276 It also introduces a new API, with functions like BIO_socket,
3277 BIO_connect, BIO_listen, BIO_lookup and a rewrite of BIO_accept.
3278 The source/sink BIOs BIO_s_connect, BIO_s_accept and BIO_s_datagram
3279 have been adapted accordingly.
3283 * RSA_padding_check_PKCS1_type_1 now accepts inputs with and without
3288 * CRIME protection: disable compression by default, even if OpenSSL is
3289 compiled with zlib enabled. Applications can still enable compression
3290 by calling SSL_CTX_clear_options(ctx, SSL_OP_NO_COMPRESSION), or by
3291 using the SSL_CONF library to configure compression.
3295 * The signature of the session callback configured with
3296 SSL_CTX_sess_set_get_cb was changed. The read-only input buffer
3297 was explicitly marked as `const unsigned char*` instead of
3302 * Always DPURIFY. Remove the use of uninitialized memory in the
3303 RNG, and other conditional uses of DPURIFY. This makes -DPURIFY a no-op.
3307 * Removed many obsolete configuration items, including
3308 DES_PTR, DES_RISC1, DES_RISC2, DES_INT
3309 MD2_CHAR, MD2_INT, MD2_LONG
3311 IDEA_SHORT, IDEA_LONG
3312 RC2_SHORT, RC2_LONG, RC4_LONG, RC4_CHUNK, RC4_INDEX
3314 *Rich Salz, with advice from Andy Polyakov*
3316 * Many BN internals have been moved to an internal header file.
3318 *Rich Salz with help from Andy Polyakov*
3320 * Configuration and writing out the results from it has changed.
3321 Files such as Makefile include/openssl/opensslconf.h and are now
3322 produced through general templates, such as Makefile.in and
3323 crypto/opensslconf.h.in and some help from the perl module
3326 Also, the center of configuration information is no longer
3327 Makefile. Instead, Configure produces a perl module in
3328 configdata.pm which holds most of the config data (in the hash
3329 table %config), the target data that comes from the target
3330 configuration in one of the `Configurations/*.conf` files (in
3335 * To clarify their intended purposes, the Configure options
3336 --prefix and --openssldir change their semantics, and become more
3337 straightforward and less interdependent.
3339 --prefix shall be used exclusively to give the location INSTALLTOP
3340 where programs, scripts, libraries, include files and manuals are
3341 going to be installed. The default is now /usr/local.
3343 --openssldir shall be used exclusively to give the default
3344 location OPENSSLDIR where certificates, private keys, CRLs are
3345 managed. This is also where the default openssl.cnf gets
3347 If the directory given with this option is a relative path, the
3348 values of both the --prefix value and the --openssldir value will
3349 be combined to become OPENSSLDIR.
3350 The default for --openssldir is INSTALLTOP/ssl.
3352 Anyone who uses --openssldir to specify where OpenSSL is to be
3353 installed MUST change to use --prefix instead.
3357 * The GOST engine was out of date and therefore it has been removed. An up
3358 to date GOST engine is now being maintained in an external repository.
3359 See: <https://wiki.openssl.org/index.php/Binaries>. Libssl still retains
3360 support for GOST ciphersuites (these are only activated if a GOST engine
3365 * EGD is no longer supported by default; use enable-egd when
3368 *Ben Kaduk and Rich Salz*
3370 * The distribution now has Makefile.in files, which are used to
3371 create Makefile's when Configure is run. *Configure must be run
3372 before trying to build now.*
3376 * The return value for SSL_CIPHER_description() for error conditions
3381 * Support for RFC6698/RFC7671 DANE TLSA peer authentication.
3383 Obtaining and performing DNSSEC validation of TLSA records is
3384 the application's responsibility. The application provides
3385 the TLSA records of its choice to OpenSSL, and these are then
3386 used to authenticate the peer.
3388 The TLSA records need not even come from DNS. They can, for
3389 example, be used to implement local end-entity certificate or
3390 trust-anchor "pinning", where the "pin" data takes the form
3391 of TLSA records, which can augment or replace verification
3392 based on the usual WebPKI public certification authorities.
3396 * Revert default OPENSSL_NO_DEPRECATED setting. Instead OpenSSL
3397 continues to support deprecated interfaces in default builds.
3398 However, applications are strongly advised to compile their
3399 source files with -DOPENSSL_API_COMPAT=0x10100000L, which hides
3400 the declarations of all interfaces deprecated in 0.9.8, 1.0.0
3401 or the 1.1.0 releases.
3403 In environments in which all applications have been ported to
3404 not use any deprecated interfaces OpenSSL's Configure script
3405 should be used with the --api=1.1.0 option to entirely remove
3406 support for the deprecated features from the library and
3407 unconditionally disable them in the installed headers.
3408 Essentially the same effect can be achieved with the "no-deprecated"
3409 argument to Configure, except that this will always restrict
3410 the build to just the latest API, rather than a fixed API
3413 As applications are ported to future revisions of the API,
3414 they should update their compile-time OPENSSL_API_COMPAT define
3415 accordingly, but in most cases should be able to continue to
3416 compile with later releases.
3418 The OPENSSL_API_COMPAT versions for 1.0.0, and 0.9.8 are
3419 0x10000000L and 0x00908000L, respectively. However those
3420 versions did not support the OPENSSL_API_COMPAT feature, and
3421 so applications are not typically tested for explicit support
3422 of just the undeprecated features of either release.
3426 * Add support for setting the minimum and maximum supported protocol.
3427 It can bet set via the SSL_set_min_proto_version() and
3428 SSL_set_max_proto_version(), or via the SSL_CONF's MinProtocol and
3429 MaxProtocol. It's recommended to use the new APIs to disable
3430 protocols instead of disabling individual protocols using
3431 SSL_set_options() or SSL_CONF's Protocol. This change also
3432 removes support for disabling TLS 1.2 in the OpenSSL TLS
3433 client at compile time by defining OPENSSL_NO_TLS1_2_CLIENT.
3437 * Support for ChaCha20 and Poly1305 added to libcrypto and libssl.
3441 * New EC_KEY_METHOD, this replaces the older ECDSA_METHOD and ECDH_METHOD
3442 and integrates ECDSA and ECDH functionality into EC. Implementations can
3443 now redirect key generation and no longer need to convert to or from
3446 Note: the ecdsa.h and ecdh.h headers are now no longer needed and just
3447 include the ec.h header file instead.
3451 * Remove support for all 40 and 56 bit ciphers. This includes all the export
3452 ciphers who are no longer supported and drops support the ephemeral RSA key
3453 exchange. The LOW ciphers currently doesn't have any ciphers in it.
3457 * Made EVP_MD_CTX, EVP_MD, EVP_CIPHER_CTX, EVP_CIPHER and HMAC_CTX
3458 opaque. For HMAC_CTX, the following constructors and destructors
3461 HMAC_CTX *HMAC_CTX_new(void);
3462 void HMAC_CTX_free(HMAC_CTX *ctx);
3464 For EVP_MD and EVP_CIPHER, complete APIs to create, fill and
3465 destroy such methods has been added. See EVP_MD_meth_new(3) and
3466 EVP_CIPHER_meth_new(3) for documentation.
3469 1) `EVP_MD_CTX_cleanup()`, `EVP_CIPHER_CTX_cleanup()` and
3470 `HMAC_CTX_cleanup()` were removed. `HMAC_CTX_reset()` and
3471 `EVP_MD_CTX_reset()` should be called instead to reinitialise
3472 an already created structure.
3473 2) For consistency with the majority of our object creators and
3474 destructors, `EVP_MD_CTX_(create|destroy)` were renamed to
3475 `EVP_MD_CTX_(new|free)`. The old names are retained as macros
3476 for deprecated builds.
3480 * Added ASYNC support. Libcrypto now includes the async sub-library to enable
3481 cryptographic operations to be performed asynchronously as long as an
3482 asynchronous capable engine is used. See the ASYNC_start_job() man page for
3483 further details. Libssl has also had this capability integrated with the
3484 introduction of the new mode SSL_MODE_ASYNC and associated error
3485 SSL_ERROR_WANT_ASYNC. See the SSL_CTX_set_mode() and SSL_get_error() man
3486 pages. This work was developed in partnership with Intel Corp.
3490 * SSL_{CTX_}set_ecdh_auto() has been removed and ECDH is support is
3491 always enabled now. If you want to disable the support you should
3492 exclude it using the list of supported ciphers. This also means that the
3493 "-no_ecdhe" option has been removed from s_server.
3497 * SSL_{CTX}_set_tmp_ecdh() which can set 1 EC curve now internally calls
3498 SSL_{CTX_}set1_curves() which can set a list.
3502 * Remove support for SSL_{CTX_}set_tmp_ecdh_callback(). You should set the
3503 curve you want to support using SSL_{CTX_}set1_curves().
3507 * State machine rewrite. The state machine code has been significantly
3508 refactored in order to remove much duplication of code and solve issues
3509 with the old code (see [ssl/statem/README.md](ssl/statem/README.md) for
3510 further details). This change does have some associated API changes.
3511 Notably the SSL_state() function has been removed and replaced by
3512 SSL_get_state which now returns an "OSSL_HANDSHAKE_STATE" instead of an int.
3513 SSL_set_state() has been removed altogether. The previous handshake states
3514 defined in ssl.h and ssl3.h have also been removed.
3518 * All instances of the string "ssleay" in the public API were replaced
3519 with OpenSSL (case-matching; e.g., OPENSSL_VERSION for #define's)
3520 Some error codes related to internal RSA_eay API's were renamed.
3524 * The demo files in crypto/threads were moved to demo/threads.
3528 * Removed obsolete engines: 4758cca, aep, atalla, cswift, nuron, gmp,
3531 *Matt Caswell, Rich Salz*
3533 * New ASN.1 embed macro.
3535 New ASN.1 macro ASN1_EMBED. This is the same as ASN1_SIMPLE except the
3536 structure is not allocated: it is part of the parent. That is instead of
3544 This reduces memory fragmentation and make it impossible to accidentally
3545 set a mandatory field to NULL.
3547 This currently only works for some fields specifically a SEQUENCE, CHOICE,
3548 or ASN1_STRING type which is part of a parent SEQUENCE. Since it is
3549 equivalent to ASN1_SIMPLE it cannot be tagged, OPTIONAL, SET OF or
3554 * Remove EVP_CHECK_DES_KEY, a compile-time option that never compiled.
3558 * Removed DES and RC4 ciphersuites from DEFAULT. Also removed RC2 although
3559 in 1.0.2 EXPORT was already removed and the only RC2 ciphersuite is also
3560 an EXPORT one. COMPLEMENTOFDEFAULT has been updated accordingly to add
3561 DES and RC4 ciphersuites.
3565 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
3566 This changes the decoding behaviour for some invalid messages,
3567 though the change is mostly in the more lenient direction, and
3568 legacy behaviour is preserved as much as possible.
3572 * Fix no-stdio build.
3573 *David Woodhouse <David.Woodhouse@intel.com> and also*
3574 *Ivan Nestlerode <ivan.nestlerode@sonos.com>*
3576 * New testing framework
3577 The testing framework has been largely rewritten and is now using
3578 perl and the perl modules Test::Harness and an extended variant of
3579 Test::More called OpenSSL::Test to do its work. All test scripts in
3580 test/ have been rewritten into test recipes, and all direct calls to
3581 executables in test/Makefile have become individual recipes using the
3582 simplified testing OpenSSL::Test::Simple.
3584 For documentation on our testing modules, do:
3586 perldoc test/testlib/OpenSSL/Test/Simple.pm
3587 perldoc test/testlib/OpenSSL/Test.pm
3591 * Revamped memory debug; only -DCRYPTO_MDEBUG and -DCRYPTO_MDEBUG_ABORT
3592 are used; the latter aborts on memory leaks (usually checked on exit).
3593 Some undocumented "set malloc, etc., hooks" functions were removed
3594 and others were changed. All are now documented.
3598 * In DSA_generate_parameters_ex, if the provided seed is too short,
3601 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
3603 * Rewrite PSK to support ECDHE_PSK, DHE_PSK and RSA_PSK. Add ciphersuites
3604 from RFC4279, RFC4785, RFC5487, RFC5489.
3606 Thanks to Christian J. Dietrich and Giuseppe D'Angelo for the
3607 original RSA_PSK patch.
3611 * Dropped support for the SSL3_FLAGS_DELAY_CLIENT_FINISHED flag. This SSLeay
3612 era flag was never set throughout the codebase (only read). Also removed
3613 SSL3_FLAGS_POP_BUFFER which was only used if
3614 SSL3_FLAGS_DELAY_CLIENT_FINISHED was also set.
3618 * Changed the default name options in the "ca", "crl", "req" and "x509"
3619 to be "oneline" instead of "compat".
3623 * Remove SSL_OP_TLS_BLOCK_PADDING_BUG. This is SSLeay legacy, we're
3624 not aware of clients that still exhibit this bug, and the workaround
3625 hasn't been working properly for a while.
3629 * The return type of BIO_number_read() and BIO_number_written() as well as
3630 the corresponding num_read and num_write members in the BIO structure has
3631 changed from unsigned long to uint64_t. On platforms where an unsigned
3632 long is 32 bits (e.g. Windows) these counters could overflow if >4Gb is
3637 * Given the pervasive nature of TLS extensions it is inadvisable to run
3638 OpenSSL without support for them. It also means that maintaining
3639 the OPENSSL_NO_TLSEXT option within the code is very invasive (and probably
3640 not well tested). Therefore the OPENSSL_NO_TLSEXT option has been removed.
3644 * Removed support for the two export grade static DH ciphersuites
3645 EXP-DH-RSA-DES-CBC-SHA and EXP-DH-DSS-DES-CBC-SHA. These two ciphersuites
3646 were newly added (along with a number of other static DH ciphersuites) to
3647 1.0.2. However the two export ones have *never* worked since they were
3648 introduced. It seems strange in any case to be adding new export
3649 ciphersuites, and given "logjam" it also does not seem correct to fix them.
3653 * Version negotiation has been rewritten. In particular SSLv23_method(),
3654 SSLv23_client_method() and SSLv23_server_method() have been deprecated,
3655 and turned into macros which simply call the new preferred function names
3656 TLS_method(), TLS_client_method() and TLS_server_method(). All new code
3657 should use the new names instead. Also as part of this change the ssl23.h
3658 header file has been removed.
3662 * Support for Kerberos ciphersuites in TLS (RFC2712) has been removed. This
3663 code and the associated standard is no longer considered fit-for-purpose.
3667 * RT2547 was closed. When generating a private key, try to make the
3668 output file readable only by the owner. This behavior change might
3669 be noticeable when interacting with other software.
3671 * Documented all exdata functions. Added CRYPTO_free_ex_index.
3676 * Added HTTP GET support to the ocsp command.
3680 * Changed default digest for the dgst and enc commands from MD5 to
3685 * RAND_pseudo_bytes has been deprecated. Users should use RAND_bytes instead.
3689 * Added support for TLS extended master secret from
3690 draft-ietf-tls-session-hash-03.txt. Thanks for Alfredo Pironti for an
3691 initial patch which was a great help during development.
3695 * All libssl internal structures have been removed from the public header
3696 files, and the OPENSSL_NO_SSL_INTERN option has been removed (since it is
3697 now redundant). Users should not attempt to access internal structures
3698 directly. Instead they should use the provided API functions.
3702 * config has been changed so that by default OPENSSL_NO_DEPRECATED is used.
3703 Access to deprecated functions can be re-enabled by running config with
3704 "enable-deprecated". In addition applications wishing to use deprecated
3705 functions must define OPENSSL_USE_DEPRECATED. Note that this new behaviour
3706 will, by default, disable some transitive includes that previously existed
3707 in the header files (e.g. ec.h will no longer, by default, include bn.h)
3711 * Added support for OCB mode. OpenSSL has been granted a patent license
3712 compatible with the OpenSSL license for use of OCB. Details are available
3713 at <https://www.openssl.org/source/OCB-patent-grant-OpenSSL.pdf>. Support
3714 for OCB can be removed by calling config with no-ocb.
3718 * SSLv2 support has been removed. It still supports receiving a SSLv2
3719 compatible client hello.
3723 * Increased the minimal RSA keysize from 256 to 512 bits [Rich Salz],
3724 done while fixing the error code for the key-too-small case.
3726 *Annie Yousar <a.yousar@informatik.hu-berlin.de>*
3728 * CA.sh has been removed; use CA.pl instead.
3732 * Removed old DES API.
3736 * Remove various unsupported platforms:
3742 Sinix/ReliantUNIX RM400
3747 16-bit platforms such as WIN16
3751 * Clean up OPENSSL_NO_xxx #define's
3752 - Use setbuf() and remove OPENSSL_NO_SETVBUF_IONBF
3753 - Rename OPENSSL_SYSNAME_xxx to OPENSSL_SYS_xxx
3754 - OPENSSL_NO_EC{DH,DSA} merged into OPENSSL_NO_EC
3755 - OPENSSL_NO_RIPEMD160, OPENSSL_NO_RIPEMD merged into OPENSSL_NO_RMD160
3756 - OPENSSL_NO_FP_API merged into OPENSSL_NO_STDIO
3757 - Remove OPENSSL_NO_BIO OPENSSL_NO_BUFFER OPENSSL_NO_CHAIN_VERIFY
3758 OPENSSL_NO_EVP OPENSSL_NO_FIPS_ERR OPENSSL_NO_HASH_COMP
3759 OPENSSL_NO_LHASH OPENSSL_NO_OBJECT OPENSSL_NO_SPEED OPENSSL_NO_STACK
3760 OPENSSL_NO_X509 OPENSSL_NO_X509_VERIFY
3761 - Remove MS_STATIC; it's a relic from platforms <32 bits.
3765 * Cleaned up dead code
3766 Remove all but one '#ifdef undef' which is to be looked at.
3770 * Clean up calling of xxx_free routines.
3771 Just like free(), fix most of the xxx_free routines to accept
3772 NULL. Remove the non-null checks from callers. Save much code.
3776 * Add secure heap for storage of private keys (when possible).
3777 Add BIO_s_secmem(), CBIGNUM, etc.
3778 Contributed by Akamai Technologies under our Corporate CLA.
3782 * Experimental support for a new, fast, unbiased prime candidate generator,
3783 bn_probable_prime_dh_coprime(). Not currently used by any prime generator.
3785 *Felix Laurie von Massenbach <felix@erbridge.co.uk>*
3787 * New output format NSS in the sess_id command line tool. This allows
3788 exporting the session id and the master key in NSS keylog format.
3790 *Martin Kaiser <martin@kaiser.cx>*
3792 * Harmonize version and its documentation. -f flag is used to display
3795 *mancha <mancha1@zoho.com>*
3797 * Fix eckey_priv_encode so it immediately returns an error upon a failure
3798 in i2d_ECPrivateKey. Thanks to Ted Unangst for feedback on this issue.
3800 *mancha <mancha1@zoho.com>*
3802 * Fix some double frees. These are not thought to be exploitable.
3804 *mancha <mancha1@zoho.com>*
3806 * A missing bounds check in the handling of the TLS heartbeat extension
3807 can be used to reveal up to 64k of memory to a connected client or
3810 Thanks for Neel Mehta of Google Security for discovering this bug and to
3811 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
3812 preparing the fix ([CVE-2014-0160])
3814 *Adam Langley, Bodo Moeller*
3816 * Fix for the attack described in the paper "Recovering OpenSSL
3817 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
3818 by Yuval Yarom and Naomi Benger. Details can be obtained from:
3819 <http://eprint.iacr.org/2014/140>
3821 Thanks to Yuval Yarom and Naomi Benger for discovering this
3822 flaw and to Yuval Yarom for supplying a fix ([CVE-2014-0076])
3824 *Yuval Yarom and Naomi Benger*
3826 * Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
3827 this fixes a limitation in previous versions of OpenSSL.
3831 * Experimental encrypt-then-mac support.
3833 Experimental support for encrypt then mac from
3834 draft-gutmann-tls-encrypt-then-mac-02.txt
3836 To enable it set the appropriate extension number (0x42 for the test
3837 server) using e.g. -DTLSEXT_TYPE_encrypt_then_mac=0x42
3839 For non-compliant peers (i.e. just about everything) this should have no
3842 WARNING: EXPERIMENTAL, SUBJECT TO CHANGE.
3846 * Add EVP support for key wrapping algorithms, to avoid problems with
3847 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
3848 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
3849 algorithms and include tests cases.
3853 * Extend CMS code to support RSA-PSS signatures and RSA-OAEP for
3858 * Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
3859 MGF1 digest and OAEP label.
3863 * Make openssl verify return errors.
3865 *Chris Palmer <palmer@google.com> and Ben Laurie*
3867 * New function ASN1_TIME_diff to calculate the difference between two
3868 ASN1_TIME structures or one structure and the current time.
3872 * Update fips_test_suite to support multiple command line options. New
3873 test to induce all self test errors in sequence and check expected
3878 * Add FIPS_{rsa,dsa,ecdsa}_{sign,verify} functions which digest and
3879 sign or verify all in one operation.
3883 * Add fips_algvs: a multicall fips utility incorporating all the algorithm
3884 test programs and fips_test_suite. Includes functionality to parse
3885 the minimal script output of fipsalgest.pl directly.
3889 * Add authorisation parameter to FIPS_module_mode_set().
3893 * Add FIPS selftest for ECDH algorithm using P-224 and B-233 curves.
3897 * Use separate DRBG fields for internal and external flags. New function
3898 FIPS_drbg_health_check() to perform on demand health checking. Add
3899 generation tests to fips_test_suite with reduced health check interval to
3900 demonstrate periodic health checking. Add "nodh" option to
3901 fips_test_suite to skip very slow DH test.
3905 * New function FIPS_get_cipherbynid() to lookup FIPS supported ciphers
3910 * More extensive health check for DRBG checking many more failure modes.
3911 New function FIPS_selftest_drbg_all() to handle every possible DRBG
3912 combination: call this in fips_test_suite.
3916 * Add support for canonical generation of DSA parameter 'g'. See
3919 * Add support for HMAC DRBG from SP800-90. Update DRBG algorithm test and
3920 POST to handle HMAC cases.
3924 * Add functions FIPS_module_version() and FIPS_module_version_text()
3925 to return numerical and string versions of the FIPS module number.
3929 * Rename FIPS_mode_set and FIPS_mode to FIPS_module_mode_set and
3930 FIPS_module_mode. FIPS_mode and FIPS_mode_set will be implemented
3931 outside the validated module in the FIPS capable OpenSSL.
3935 * Minor change to DRBG entropy callback semantics. In some cases
3936 there is no multiple of the block length between min_len and
3937 max_len. Allow the callback to return more than max_len bytes
3938 of entropy but discard any extra: it is the callback's responsibility
3939 to ensure that the extra data discarded does not impact the
3940 requested amount of entropy.
3944 * Add PRNG security strength checks to RSA, DSA and ECDSA using
3945 information in FIPS186-3, SP800-57 and SP800-131A.
3949 * CCM support via EVP. Interface is very similar to GCM case except we
3950 must supply all data in one chunk (i.e. no update, final) and the
3951 message length must be supplied if AAD is used. Add algorithm test
3956 * Initial version of POST overhaul. Add POST callback to allow the status
3957 of POST to be monitored and/or failures induced. Modify fips_test_suite
3958 to use callback. Always run all selftests even if one fails.
3962 * XTS support including algorithm test driver in the fips_gcmtest program.
3963 Note: this does increase the maximum key length from 32 to 64 bytes but
3964 there should be no binary compatibility issues as existing applications
3965 will never use XTS mode.
3969 * Extensive reorganisation of FIPS PRNG behaviour. Remove all dependencies
3970 to OpenSSL RAND code and replace with a tiny FIPS RAND API which also
3971 performs algorithm blocking for unapproved PRNG types. Also do not
3972 set PRNG type in FIPS_mode_set(): leave this to the application.
3973 Add default OpenSSL DRBG handling: sets up FIPS PRNG and seeds with
3974 the standard OpenSSL PRNG: set additional data to a date time vector.
3978 * Rename old X9.31 PRNG functions of the form `FIPS_rand*` to `FIPS_x931*`.
3979 This shouldn't present any incompatibility problems because applications
3980 shouldn't be using these directly and any that are will need to rethink
3981 anyway as the X9.31 PRNG is now deprecated by FIPS 140-2
3985 * Extensive self tests and health checking required by SP800-90 DRBG.
3986 Remove strength parameter from FIPS_drbg_instantiate and always
3987 instantiate at maximum supported strength.
3991 * Add ECDH code to fips module and fips_ecdhvs for primitives only testing.
3995 * New algorithm test program fips_dhvs to handle DH primitives only testing.
3999 * New function DH_compute_key_padded() to compute a DH key and pad with
4000 leading zeroes if needed: this complies with SP800-56A et al.
4004 * Initial implementation of SP800-90 DRBGs for Hash and CTR. Not used by
4005 anything, incomplete, subject to change and largely untested at present.
4009 * Modify fipscanisteronly build option to only build the necessary object
4010 files by filtering FIPS_EX_OBJ through a perl script in crypto/Makefile.
4014 * Add experimental option FIPSSYMS to give all symbols in
4015 fipscanister.o and FIPS or fips prefix. This will avoid
4016 conflicts with future versions of OpenSSL. Add perl script
4017 util/fipsas.pl to preprocess assembly language source files
4018 and rename any affected symbols.
4022 * Add selftest checks and algorithm block of non-fips algorithms in
4023 FIPS mode. Remove DES2 from selftests.
4027 * Add ECDSA code to fips module. Add tiny fips_ecdsa_check to just
4028 return internal method without any ENGINE dependencies. Add new
4029 tiny fips sign and verify functions.
4033 * New build option no-ec2m to disable characteristic 2 code.
4037 * New build option "fipscanisteronly". This only builds fipscanister.o
4038 and (currently) associated fips utilities. Uses the file Makefile.fips
4039 instead of Makefile.org as the prototype.
4043 * Add some FIPS mode restrictions to GCM. Add internal IV generator.
4044 Update fips_gcmtest to use IV generator.
4048 * Initial, experimental EVP support for AES-GCM. AAD can be input by
4049 setting output buffer to NULL. The `*Final` function must be
4050 called although it will not retrieve any additional data. The tag
4051 can be set or retrieved with a ctrl. The IV length is by default 12
4052 bytes (96 bits) but can be set to an alternative value. If the IV
4053 length exceeds the maximum IV length (currently 16 bytes) it cannot be
4058 * New flag in ciphers: EVP_CIPH_FLAG_CUSTOM_CIPHER. This means the
4059 underlying do_cipher function handles all cipher semantics itself
4060 including padding and finalisation. This is useful if (for example)
4061 an ENGINE cipher handles block padding itself. The behaviour of
4062 do_cipher is subtly changed if this flag is set: the return value
4063 is the number of characters written to the output buffer (zero is
4064 no longer an error code) or a negative error code. Also if the
4065 input buffer is NULL and length 0 finalisation should be performed.
4069 * If a candidate issuer certificate is already part of the constructed
4070 path ignore it: new debug notification X509_V_ERR_PATH_LOOP for this case.
4074 * Improve forward-security support: add functions
4076 void SSL_CTX_set_not_resumable_session_callback(
4077 SSL_CTX *ctx, int (*cb)(SSL *ssl, int is_forward_secure))
4078 void SSL_set_not_resumable_session_callback(
4079 SSL *ssl, int (*cb)(SSL *ssl, int is_forward_secure))
4081 for use by SSL/TLS servers; the callback function will be called whenever a
4082 new session is created, and gets to decide whether the session may be
4083 cached to make it resumable (return 0) or not (return 1). (As by the
4084 SSL/TLS protocol specifications, the session_id sent by the server will be
4085 empty to indicate that the session is not resumable; also, the server will
4086 not generate RFC 4507 (RFC 5077) session tickets.)
4088 A simple reasonable callback implementation is to return is_forward_secure.
4089 This parameter will be set to 1 or 0 depending on the ciphersuite selected
4090 by the SSL/TLS server library, indicating whether it can provide forward
4093 *Emilia Käsper <emilia.kasper@esat.kuleuven.be> (Google)*
4095 * New -verify_name option in command line utilities to set verification
4100 * Initial CMAC implementation. WARNING: EXPERIMENTAL, API MAY CHANGE.
4101 Add CMAC pkey methods.
4105 * Experimental renegotiation in s_server -www mode. If the client
4106 browses /reneg connection is renegotiated. If /renegcert it is
4107 renegotiated requesting a certificate.
4111 * Add an "external" session cache for debugging purposes to s_server. This
4112 should help trace issues which normally are only apparent in deployed
4113 multi-process servers.
4117 * Extensive audit of libcrypto with DEBUG_UNUSED. Fix many cases where
4118 return value is ignored. NB. The functions RAND_add(), RAND_seed(),
4119 BIO_set_cipher() and some obscure PEM functions were changed so they
4120 can now return an error. The RAND changes required a change to the
4121 RAND_METHOD structure.
4125 * New macro `__owur` for "OpenSSL Warn Unused Result". This makes use of
4126 a gcc attribute to warn if the result of a function is ignored. This
4127 is enable if DEBUG_UNUSED is set. Add to several functions in evp.h
4128 whose return value is often ignored.
4132 * New -noct, -requestct, -requirect and -ctlogfile options for s_client.
4133 These allow SCTs (signed certificate timestamps) to be requested and
4134 validated when establishing a connection.
4136 *Rob Percival <robpercival@google.com>*
4141 ### Changes between 1.0.2s and 1.0.2t [10 Sep 2019]
4143 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
4144 used even when parsing explicit parameters, when loading a encoded key
4145 or calling `EC_GROUP_new_from_ecpkparameters()`/
4146 `EC_GROUP_new_from_ecparameters()`.
4147 This prevents bypass of security hardening and performance gains,
4148 especially for curves with specialized EC_METHODs.
4149 By default, if a key encoded with explicit parameters is loaded and later
4150 encoded, the output is still encoded with explicit parameters, even if
4151 internally a "named" EC_GROUP is used for computation.
4155 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
4156 this change, EC_GROUP_set_generator would accept order and/or cofactor as
4157 NULL. After this change, only the cofactor parameter can be NULL. It also
4158 does some minimal sanity checks on the passed order.
4163 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
4164 An attack is simple, if the first CMS_recipientInfo is valid but the
4165 second CMS_recipientInfo is chosen ciphertext. If the second
4166 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
4167 encryption key will be replaced by garbage, and the message cannot be
4168 decoded, but if the RSA decryption fails, the correct encryption key is
4169 used and the recipient will not notice the attack.
4170 As a work around for this potential attack the length of the decrypted
4171 key must be equal to the cipher default key length, in case the
4172 certifiate is not given and all recipientInfo are tried out.
4173 The old behaviour can be re-enabled in the CMS code by setting the
4174 CMS_DEBUG_DECRYPT flag.
4179 * Document issue with installation paths in diverse Windows builds
4181 '/usr/local/ssl' is an unsafe prefix for location to install OpenSSL
4182 binaries and run-time config file.
4187 ### Changes between 1.0.2r and 1.0.2s [28 May 2019]
4189 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
4190 This changes the size when using the `genpkey` command when no size is given.
4191 It fixes an omission in earlier changes that changed all RSA, DSA and DH
4192 generation commands to use 2048 bits by default.
4196 * Add FIPS support for Android Arm 64-bit
4198 Support for Android Arm 64-bit was added to the OpenSSL FIPS Object
4199 Module in Version 2.0.10. For some reason, the corresponding target
4200 'android64-aarch64' was missing OpenSSL 1.0.2, whence it could not be
4201 built with FIPS support on Android Arm 64-bit. This omission has been
4204 *Matthias St. Pierre*
4206 ### Changes between 1.0.2q and 1.0.2r [26 Feb 2019]
4208 * 0-byte record padding oracle
4210 If an application encounters a fatal protocol error and then calls
4211 SSL_shutdown() twice (once to send a close_notify, and once to receive one)
4212 then OpenSSL can respond differently to the calling application if a 0 byte
4213 record is received with invalid padding compared to if a 0 byte record is
4214 received with an invalid MAC. If the application then behaves differently
4215 based on that in a way that is detectable to the remote peer, then this
4216 amounts to a padding oracle that could be used to decrypt data.
4218 In order for this to be exploitable "non-stitched" ciphersuites must be in
4219 use. Stitched ciphersuites are optimised implementations of certain
4220 commonly used ciphersuites. Also the application must call SSL_shutdown()
4221 twice even if a protocol error has occurred (applications should not do
4222 this but some do anyway).
4224 This issue was discovered by Juraj Somorovsky, Robert Merget and Nimrod
4225 Aviram, with additional investigation by Steven Collison and Andrew
4226 Hourselt. It was reported to OpenSSL on 10th December 2018.
4231 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
4235 ### Changes between 1.0.2p and 1.0.2q [20 Nov 2018]
4237 * Microarchitecture timing vulnerability in ECC scalar multiplication
4239 OpenSSL ECC scalar multiplication, used in e.g. ECDSA and ECDH, has been
4240 shown to be vulnerable to a microarchitecture timing side channel attack.
4241 An attacker with sufficient access to mount local timing attacks during
4242 ECDSA signature generation could recover the private key.
4244 This issue was reported to OpenSSL on 26th October 2018 by Alejandro
4245 Cabrera Aldaya, Billy Brumley, Sohaib ul Hassan, Cesar Pereida Garcia and
4251 * Timing vulnerability in DSA signature generation
4253 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
4254 timing side channel attack. An attacker could use variations in the signing
4255 algorithm to recover the private key.
4257 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
4262 * Resolve a compatibility issue in EC_GROUP handling with the FIPS Object
4263 Module, accidentally introduced while backporting security fixes from the
4264 development branch and hindering the use of ECC in FIPS mode.
4268 ### Changes between 1.0.2o and 1.0.2p [14 Aug 2018]
4270 * Client DoS due to large DH parameter
4272 During key agreement in a TLS handshake using a DH(E) based ciphersuite a
4273 malicious server can send a very large prime value to the client. This will
4274 cause the client to spend an unreasonably long period of time generating a
4275 key for this prime resulting in a hang until the client has finished. This
4276 could be exploited in a Denial Of Service attack.
4278 This issue was reported to OpenSSL on 5th June 2018 by Guido Vranken
4283 * Cache timing vulnerability in RSA Key Generation
4285 The OpenSSL RSA Key generation algorithm has been shown to be vulnerable to
4286 a cache timing side channel attack. An attacker with sufficient access to
4287 mount cache timing attacks during the RSA key generation process could
4288 recover the private key.
4290 This issue was reported to OpenSSL on 4th April 2018 by Alejandro Cabrera
4291 Aldaya, Billy Brumley, Cesar Pereida Garcia and Luis Manuel Alvarez Tapia.
4296 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
4297 parameter is no longer accepted, as it leads to a corrupt table. NULL
4298 pem_str is reserved for alias entries only.
4302 * Revert blinding in ECDSA sign and instead make problematic addition
4303 length-invariant. Switch even to fixed-length Montgomery multiplication.
4307 * Change generating and checking of primes so that the error rate of not
4308 being prime depends on the intended use based on the size of the input.
4309 For larger primes this will result in more rounds of Miller-Rabin.
4310 The maximal error rate for primes with more than 1080 bits is lowered
4313 *Kurt Roeckx, Annie Yousar*
4315 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
4319 * Add blinding to ECDSA and DSA signatures to protect against side channel
4320 attacks discovered by Keegan Ryan (NCC Group).
4324 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
4325 now allow empty (zero character) pass phrases.
4329 * Certificate time validation (X509_cmp_time) enforces stricter
4330 compliance with RFC 5280. Fractional seconds and timezone offsets
4331 are no longer allowed.
4335 ### Changes between 1.0.2n and 1.0.2o [27 Mar 2018]
4337 * Constructed ASN.1 types with a recursive definition could exceed the stack
4339 Constructed ASN.1 types with a recursive definition (such as can be found
4340 in PKCS7) could eventually exceed the stack given malicious input with
4341 excessive recursion. This could result in a Denial Of Service attack. There
4342 are no such structures used within SSL/TLS that come from untrusted sources
4343 so this is considered safe.
4345 This issue was reported to OpenSSL on 4th January 2018 by the OSS-fuzz
4351 ### Changes between 1.0.2m and 1.0.2n [7 Dec 2017]
4353 * Read/write after SSL object in error state
4355 OpenSSL 1.0.2 (starting from version 1.0.2b) introduced an "error state"
4356 mechanism. The intent was that if a fatal error occurred during a handshake
4357 then OpenSSL would move into the error state and would immediately fail if
4358 you attempted to continue the handshake. This works as designed for the
4359 explicit handshake functions (SSL_do_handshake(), SSL_accept() and
4360 SSL_connect()), however due to a bug it does not work correctly if
4361 SSL_read() or SSL_write() is called directly. In that scenario, if the
4362 handshake fails then a fatal error will be returned in the initial function
4363 call. If SSL_read()/SSL_write() is subsequently called by the application
4364 for the same SSL object then it will succeed and the data is passed without
4365 being decrypted/encrypted directly from the SSL/TLS record layer.
4367 In order to exploit this issue an application bug would have to be present
4368 that resulted in a call to SSL_read()/SSL_write() being issued after having
4369 already received a fatal error.
4371 This issue was reported to OpenSSL by David Benjamin (Google).
4376 * rsaz_1024_mul_avx2 overflow bug on x86_64
4378 There is an overflow bug in the AVX2 Montgomery multiplication procedure
4379 used in exponentiation with 1024-bit moduli. No EC algorithms are affected.
4380 Analysis suggests that attacks against RSA and DSA as a result of this
4381 defect would be very difficult to perform and are not believed likely.
4382 Attacks against DH1024 are considered just feasible, because most of the
4383 work necessary to deduce information about a private key may be performed
4384 offline. The amount of resources required for such an attack would be
4385 significant. However, for an attack on TLS to be meaningful, the server
4386 would have to share the DH1024 private key among multiple clients, which is
4387 no longer an option since CVE-2016-0701.
4389 This only affects processors that support the AVX2 but not ADX extensions
4390 like Intel Haswell (4th generation).
4392 This issue was reported to OpenSSL by David Benjamin (Google). The issue
4393 was originally found via the OSS-Fuzz project.
4398 ### Changes between 1.0.2l and 1.0.2m [2 Nov 2017]
4400 * bn_sqrx8x_internal carry bug on x86_64
4402 There is a carry propagating bug in the x86_64 Montgomery squaring
4403 procedure. No EC algorithms are affected. Analysis suggests that attacks
4404 against RSA and DSA as a result of this defect would be very difficult to
4405 perform and are not believed likely. Attacks against DH are considered just
4406 feasible (although very difficult) because most of the work necessary to
4407 deduce information about a private key may be performed offline. The amount
4408 of resources required for such an attack would be very significant and
4409 likely only accessible to a limited number of attackers. An attacker would
4410 additionally need online access to an unpatched system using the target
4411 private key in a scenario with persistent DH parameters and a private
4412 key that is shared between multiple clients.
4414 This only affects processors that support the BMI1, BMI2 and ADX extensions
4415 like Intel Broadwell (5th generation) and later or AMD Ryzen.
4417 This issue was reported to OpenSSL by the OSS-Fuzz project.
4422 * Malformed X.509 IPAddressFamily could cause OOB read
4424 If an X.509 certificate has a malformed IPAddressFamily extension,
4425 OpenSSL could do a one-byte buffer overread. The most likely result
4426 would be an erroneous display of the certificate in text format.
4428 This issue was reported to OpenSSL by the OSS-Fuzz project.
4432 ### Changes between 1.0.2k and 1.0.2l [25 May 2017]
4434 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
4435 platform rather than 'mingw'.
4439 ### Changes between 1.0.2j and 1.0.2k [26 Jan 2017]
4441 * Truncated packet could crash via OOB read
4443 If one side of an SSL/TLS path is running on a 32-bit host and a specific
4444 cipher is being used, then a truncated packet can cause that host to
4445 perform an out-of-bounds read, usually resulting in a crash.
4447 This issue was reported to OpenSSL by Robert Święcki of Google.
4452 * BN_mod_exp may produce incorrect results on x86_64
4454 There is a carry propagating bug in the x86_64 Montgomery squaring
4455 procedure. No EC algorithms are affected. Analysis suggests that attacks
4456 against RSA and DSA as a result of this defect would be very difficult to
4457 perform and are not believed likely. Attacks against DH are considered just
4458 feasible (although very difficult) because most of the work necessary to
4459 deduce information about a private key may be performed offline. The amount
4460 of resources required for such an attack would be very significant and
4461 likely only accessible to a limited number of attackers. An attacker would
4462 additionally need online access to an unpatched system using the target
4463 private key in a scenario with persistent DH parameters and a private
4464 key that is shared between multiple clients. For example this can occur by
4465 default in OpenSSL DHE based SSL/TLS ciphersuites. Note: This issue is very
4466 similar to CVE-2015-3193 but must be treated as a separate problem.
4468 This issue was reported to OpenSSL by the OSS-Fuzz project.
4473 * Montgomery multiplication may produce incorrect results
4475 There is a carry propagating bug in the Broadwell-specific Montgomery
4476 multiplication procedure that handles input lengths divisible by, but
4477 longer than 256 bits. Analysis suggests that attacks against RSA, DSA
4478 and DH private keys are impossible. This is because the subroutine in
4479 question is not used in operations with the private key itself and an input
4480 of the attacker's direct choice. Otherwise the bug can manifest itself as
4481 transient authentication and key negotiation failures or reproducible
4482 erroneous outcome of public-key operations with specially crafted input.
4483 Among EC algorithms only Brainpool P-512 curves are affected and one
4484 presumably can attack ECDH key negotiation. Impact was not analyzed in
4485 detail, because pre-requisites for attack are considered unlikely. Namely
4486 multiple clients have to choose the curve in question and the server has to
4487 share the private key among them, neither of which is default behaviour.
4488 Even then only clients that chose the curve will be affected.
4490 This issue was publicly reported as transient failures and was not
4491 initially recognized as a security issue. Thanks to Richard Morgan for
4492 providing reproducible case.
4497 * OpenSSL now fails if it receives an unrecognised record type in TLS1.0
4498 or TLS1.1. Previously this only happened in SSLv3 and TLS1.2. This is to
4499 prevent issues where no progress is being made and the peer continually
4500 sends unrecognised record types, using up resources processing them.
4504 ### Changes between 1.0.2i and 1.0.2j [26 Sep 2016]
4506 * Missing CRL sanity check
4508 A bug fix which included a CRL sanity check was added to OpenSSL 1.1.0
4509 but was omitted from OpenSSL 1.0.2i. As a result any attempt to use
4510 CRLs in OpenSSL 1.0.2i will crash with a null pointer exception.
4512 This issue only affects the OpenSSL 1.0.2i
4517 ### Changes between 1.0.2h and 1.0.2i [22 Sep 2016]
4519 * OCSP Status Request extension unbounded memory growth
4521 A malicious client can send an excessively large OCSP Status Request
4522 extension. If that client continually requests renegotiation, sending a
4523 large OCSP Status Request extension each time, then there will be unbounded
4524 memory growth on the server. This will eventually lead to a Denial Of
4525 Service attack through memory exhaustion. Servers with a default
4526 configuration are vulnerable even if they do not support OCSP. Builds using
4527 the "no-ocsp" build time option are not affected.
4529 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4534 * In order to mitigate the SWEET32 attack, the DES ciphers were moved from
4537 This issue was reported to OpenSSL Karthikeyan Bhargavan and Gaetan
4543 * OOB write in MDC2_Update()
4545 An overflow can occur in MDC2_Update() either if called directly or
4546 through the EVP_DigestUpdate() function using MDC2. If an attacker
4547 is able to supply very large amounts of input data after a previous
4548 call to EVP_EncryptUpdate() with a partial block then a length check
4549 can overflow resulting in a heap corruption.
4551 The amount of data needed is comparable to SIZE_MAX which is impractical
4554 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4559 * Malformed SHA512 ticket DoS
4561 If a server uses SHA512 for TLS session ticket HMAC it is vulnerable to a
4562 DoS attack where a malformed ticket will result in an OOB read which will
4565 The use of SHA512 in TLS session tickets is comparatively rare as it requires
4566 a custom server callback and ticket lookup mechanism.
4568 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4573 * OOB write in BN_bn2dec()
4575 The function BN_bn2dec() does not check the return value of BN_div_word().
4576 This can cause an OOB write if an application uses this function with an
4577 overly large BIGNUM. This could be a problem if an overly large certificate
4578 or CRL is printed out from an untrusted source. TLS is not affected because
4579 record limits will reject an oversized certificate before it is parsed.
4581 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4586 * OOB read in TS_OBJ_print_bio()
4588 The function TS_OBJ_print_bio() misuses OBJ_obj2txt(): the return value is
4589 the total length the OID text representation would use and not the amount
4590 of data written. This will result in OOB reads when large OIDs are
4593 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4598 * Pointer arithmetic undefined behaviour
4600 Avoid some undefined pointer arithmetic
4602 A common idiom in the codebase is to check limits in the following manner:
4605 Where "p" points to some malloc'd data of SIZE bytes and
4608 "len" here could be from some externally supplied data (e.g. from a TLS
4611 The rules of C pointer arithmetic are such that "p + len" is only well
4612 defined where len <= SIZE. Therefore the above idiom is actually
4613 undefined behaviour.
4615 For example this could cause problems if some malloc implementation
4616 provides an address for "p" such that "p + len" actually overflows for
4617 values of len that are too big and therefore p + len < limit.
4619 This issue was reported to OpenSSL by Guido Vranken
4624 * Constant time flag not preserved in DSA signing
4626 Operations in the DSA signing algorithm should run in constant time in
4627 order to avoid side channel attacks. A flaw in the OpenSSL DSA
4628 implementation means that a non-constant time codepath is followed for
4629 certain operations. This has been demonstrated through a cache-timing
4630 attack to be sufficient for an attacker to recover the private DSA key.
4632 This issue was reported by César Pereida (Aalto University), Billy Brumley
4633 (Tampere University of Technology), and Yuval Yarom (The University of
4634 Adelaide and NICTA).
4639 * DTLS buffered message DoS
4641 In a DTLS connection where handshake messages are delivered out-of-order
4642 those messages that OpenSSL is not yet ready to process will be buffered
4643 for later use. Under certain circumstances, a flaw in the logic means that
4644 those messages do not get removed from the buffer even though the handshake
4645 has been completed. An attacker could force up to approx. 15 messages to
4646 remain in the buffer when they are no longer required. These messages will
4647 be cleared when the DTLS connection is closed. The default maximum size for
4648 a message is 100k. Therefore the attacker could force an additional 1500k
4649 to be consumed per connection. By opening many simulataneous connections an
4650 attacker could cause a DoS attack through memory exhaustion.
4652 This issue was reported to OpenSSL by Quan Luo.
4657 * DTLS replay protection DoS
4659 A flaw in the DTLS replay attack protection mechanism means that records
4660 that arrive for future epochs update the replay protection "window" before
4661 the MAC for the record has been validated. This could be exploited by an
4662 attacker by sending a record for the next epoch (which does not have to
4663 decrypt or have a valid MAC), with a very large sequence number. This means
4664 that all subsequent legitimate packets are dropped causing a denial of
4665 service for a specific DTLS connection.
4667 This issue was reported to OpenSSL by the OCAP audit team.
4672 * Certificate message OOB reads
4674 In OpenSSL 1.0.2 and earlier some missing message length checks can result
4675 in OOB reads of up to 2 bytes beyond an allocated buffer. There is a
4676 theoretical DoS risk but this has not been observed in practice on common
4679 The messages affected are client certificate, client certificate request
4680 and server certificate. As a result the attack can only be performed
4681 against a client or a server which enables client authentication.
4683 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4688 ### Changes between 1.0.2g and 1.0.2h [3 May 2016]
4690 * Prevent padding oracle in AES-NI CBC MAC check
4692 A MITM attacker can use a padding oracle attack to decrypt traffic
4693 when the connection uses an AES CBC cipher and the server support
4696 This issue was introduced as part of the fix for Lucky 13 padding
4697 attack ([CVE-2013-0169]). The padding check was rewritten to be in
4698 constant time by making sure that always the same bytes are read and
4699 compared against either the MAC or padding bytes. But it no longer
4700 checked that there was enough data to have both the MAC and padding
4703 This issue was reported by Juraj Somorovsky using TLS-Attacker.
4707 * Fix EVP_EncodeUpdate overflow
4709 An overflow can occur in the EVP_EncodeUpdate() function which is used for
4710 Base64 encoding of binary data. If an attacker is able to supply very large
4711 amounts of input data then a length check can overflow resulting in a heap
4714 Internally to OpenSSL the EVP_EncodeUpdate() function is primarily used by
4715 the `PEM_write_bio*` family of functions. These are mainly used within the
4716 OpenSSL command line applications, so any application which processes data
4717 from an untrusted source and outputs it as a PEM file should be considered
4718 vulnerable to this issue. User applications that call these APIs directly
4719 with large amounts of untrusted data may also be vulnerable.
4721 This issue was reported by Guido Vranken.
4726 * Fix EVP_EncryptUpdate overflow
4728 An overflow can occur in the EVP_EncryptUpdate() function. If an attacker
4729 is able to supply very large amounts of input data after a previous call to
4730 EVP_EncryptUpdate() with a partial block then a length check can overflow
4731 resulting in a heap corruption. Following an analysis of all OpenSSL
4732 internal usage of the EVP_EncryptUpdate() function all usage is one of two
4733 forms. The first form is where the EVP_EncryptUpdate() call is known to be
4734 the first called function after an EVP_EncryptInit(), and therefore that
4735 specific call must be safe. The second form is where the length passed to
4736 EVP_EncryptUpdate() can be seen from the code to be some small value and
4737 therefore there is no possibility of an overflow. Since all instances are
4738 one of these two forms, it is believed that there can be no overflows in
4739 internal code due to this problem. It should be noted that
4740 EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths.
4741 Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances
4742 of these calls have also been analysed too and it is believed there are no
4743 instances in internal usage where an overflow could occur.
4745 This issue was reported by Guido Vranken.
4750 * Prevent ASN.1 BIO excessive memory allocation
4752 When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio()
4753 a short invalid encoding can cause allocation of large amounts of memory
4754 potentially consuming excessive resources or exhausting memory.
4756 Any application parsing untrusted data through d2i BIO functions is
4757 affected. The memory based functions such as d2i_X509() are *not* affected.
4758 Since the memory based functions are used by the TLS library, TLS
4759 applications are not affected.
4761 This issue was reported by Brian Carpenter.
4768 ASN1 Strings that are over 1024 bytes can cause an overread in applications
4769 using the X509_NAME_oneline() function on EBCDIC systems. This could result
4770 in arbitrary stack data being returned in the buffer.
4772 This issue was reported by Guido Vranken.
4777 * Modify behavior of ALPN to invoke callback after SNI/servername
4778 callback, such that updates to the SSL_CTX affect ALPN.
4782 * Remove LOW from the DEFAULT cipher list. This removes singles DES from the
4787 * Only remove the SSLv2 methods with the no-ssl2-method option. When the
4788 methods are enabled and ssl2 is disabled the methods return NULL.
4792 ### Changes between 1.0.2f and 1.0.2g [1 Mar 2016]
4794 * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL.
4795 Builds that are not configured with "enable-weak-ssl-ciphers" will not
4796 provide any "EXPORT" or "LOW" strength ciphers.
4800 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2
4801 is by default disabled at build-time. Builds that are not configured with
4802 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used,
4803 users who want to negotiate SSLv2 via the version-flexible SSLv23_method()
4804 will need to explicitly call either of:
4806 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2);
4808 SSL_clear_options(ssl, SSL_OP_NO_SSLv2);
4810 as appropriate. Even if either of those is used, or the application
4811 explicitly uses the version-specific SSLv2_method() or its client and
4812 server variants, SSLv2 ciphers vulnerable to exhaustive search key
4813 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT
4814 ciphers, and SSLv2 56-bit DES are no longer available.
4819 * Fix a double-free in DSA code
4821 A double free bug was discovered when OpenSSL parses malformed DSA private
4822 keys and could lead to a DoS attack or memory corruption for applications
4823 that receive DSA private keys from untrusted sources. This scenario is
4826 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using
4832 * Disable SRP fake user seed to address a server memory leak.
4834 Add a new method SRP_VBASE_get1_by_user that handles the seed properly.
4836 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
4837 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user
4838 was changed to ignore the "fake user" SRP seed, even if the seed
4841 Users should use SRP_VBASE_get1_by_user instead. Note that in
4842 SRP_VBASE_get1_by_user, caller must free the returned value. Note
4843 also that even though configuring the SRP seed attempts to hide
4844 invalid usernames by continuing the handshake with fake
4845 credentials, this behaviour is not constant time and no strong
4846 guarantees are made that the handshake is indistinguishable from
4847 that of a valid user.
4852 * Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption
4854 In the BN_hex2bn function the number of hex digits is calculated using an
4855 int value `i`. Later `bn_expand` is called with a value of `i * 4`. For
4856 large values of `i` this can result in `bn_expand` not allocating any
4857 memory because `i * 4` is negative. This can leave the internal BIGNUM data
4858 field as NULL leading to a subsequent NULL ptr deref. For very large values
4859 of `i`, the calculation `i * 4` could be a positive value smaller than `i`.
4860 In this case memory is allocated to the internal BIGNUM data field, but it
4861 is insufficiently sized leading to heap corruption. A similar issue exists
4862 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn
4863 is ever called by user applications with very large untrusted hex/dec data.
4864 This is anticipated to be a rare occurrence.
4866 All OpenSSL internal usage of these functions use data that is not expected
4867 to be untrusted, e.g. config file data or application command line
4868 arguments. If user developed applications generate config file data based
4869 on untrusted data then it is possible that this could also lead to security
4870 consequences. This is also anticipated to be rare.
4872 This issue was reported to OpenSSL by Guido Vranken.
4877 * Fix memory issues in `BIO_*printf` functions
4879 The internal `fmtstr` function used in processing a "%s" format string in
4880 the `BIO_*printf` functions could overflow while calculating the length of a
4881 string and cause an OOB read when printing very long strings.
4883 Additionally the internal `doapr_outch` function can attempt to write to an
4884 OOB memory location (at an offset from the NULL pointer) in the event of a
4885 memory allocation failure. In 1.0.2 and below this could be caused where
4886 the size of a buffer to be allocated is greater than INT_MAX. E.g. this
4887 could be in processing a very long "%s" format string. Memory leaks can
4890 The first issue may mask the second issue dependent on compiler behaviour.
4891 These problems could enable attacks where large amounts of untrusted data
4892 is passed to the `BIO_*printf` functions. If applications use these functions
4893 in this way then they could be vulnerable. OpenSSL itself uses these
4894 functions when printing out human-readable dumps of ASN.1 data. Therefore
4895 applications that print this data could be vulnerable if the data is from
4896 untrusted sources. OpenSSL command line applications could also be
4897 vulnerable where they print out ASN.1 data, or if untrusted data is passed
4898 as command line arguments.
4900 Libssl is not considered directly vulnerable. Additionally certificates etc
4901 received via remote connections via libssl are also unlikely to be able to
4902 trigger these issues because of message size limits enforced within libssl.
4904 This issue was reported to OpenSSL Guido Vranken.
4909 * Side channel attack on modular exponentiation
4911 A side-channel attack was found which makes use of cache-bank conflicts on
4912 the Intel Sandy-Bridge microarchitecture which could lead to the recovery
4913 of RSA keys. The ability to exploit this issue is limited as it relies on
4914 an attacker who has control of code in a thread running on the same
4915 hyper-threaded core as the victim thread which is performing decryptions.
4917 This issue was reported to OpenSSL by Yuval Yarom, The University of
4918 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and
4919 Nadia Heninger, University of Pennsylvania with more information at
4920 <http://cachebleed.info>.
4925 * Change the `req` command to generate a 2048-bit RSA/DSA key by default,
4926 if no keysize is specified with default_bits. This fixes an
4927 omission in an earlier change that changed all RSA/DSA key generation
4928 commands to use 2048 bits by default.
4932 ### Changes between 1.0.2e and 1.0.2f [28 Jan 2016]
4934 * DH small subgroups
4936 Historically OpenSSL only ever generated DH parameters based on "safe"
4937 primes. More recently (in version 1.0.2) support was provided for
4938 generating X9.42 style parameter files such as those required for RFC 5114
4939 support. The primes used in such files may not be "safe". Where an
4940 application is using DH configured with parameters based on primes that are
4941 not "safe" then an attacker could use this fact to find a peer's private
4942 DH exponent. This attack requires that the attacker complete multiple
4943 handshakes in which the peer uses the same private DH exponent. For example
4944 this could be used to discover a TLS server's private DH exponent if it's
4945 reusing the private DH exponent or it's using a static DH ciphersuite.
4947 OpenSSL provides the option SSL_OP_SINGLE_DH_USE for ephemeral DH (DHE) in
4948 TLS. It is not on by default. If the option is not set then the server
4949 reuses the same private DH exponent for the life of the server process and
4950 would be vulnerable to this attack. It is believed that many popular
4951 applications do set this option and would therefore not be at risk.
4953 The fix for this issue adds an additional check where a "q" parameter is
4954 available (as is the case in X9.42 based parameters). This detects the
4955 only known attack, and is the only possible defense for static DH
4956 ciphersuites. This could have some performance impact.
4958 Additionally the SSL_OP_SINGLE_DH_USE option has been switched on by
4959 default and cannot be disabled. This could have some performance impact.
4961 This issue was reported to OpenSSL by Antonio Sanso (Adobe).
4966 * SSLv2 doesn't block disabled ciphers
4968 A malicious client can negotiate SSLv2 ciphers that have been disabled on
4969 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
4970 been disabled, provided that the SSLv2 protocol was not also disabled via
4973 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
4974 and Sebastian Schinzel.
4979 ### Changes between 1.0.2d and 1.0.2e [3 Dec 2015]
4981 * BN_mod_exp may produce incorrect results on x86_64
4983 There is a carry propagating bug in the x86_64 Montgomery squaring
4984 procedure. No EC algorithms are affected. Analysis suggests that attacks
4985 against RSA and DSA as a result of this defect would be very difficult to
4986 perform and are not believed likely. Attacks against DH are considered just
4987 feasible (although very difficult) because most of the work necessary to
4988 deduce information about a private key may be performed offline. The amount
4989 of resources required for such an attack would be very significant and
4990 likely only accessible to a limited number of attackers. An attacker would
4991 additionally need online access to an unpatched system using the target
4992 private key in a scenario with persistent DH parameters and a private
4993 key that is shared between multiple clients. For example this can occur by
4994 default in OpenSSL DHE based SSL/TLS ciphersuites.
4996 This issue was reported to OpenSSL by Hanno Böck.
5001 * Certificate verify crash with missing PSS parameter
5003 The signature verification routines will crash with a NULL pointer
5004 dereference if presented with an ASN.1 signature using the RSA PSS
5005 algorithm and absent mask generation function parameter. Since these
5006 routines are used to verify certificate signature algorithms this can be
5007 used to crash any certificate verification operation and exploited in a
5008 DoS attack. Any application which performs certificate verification is
5009 vulnerable including OpenSSL clients and servers which enable client
5012 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
5017 * X509_ATTRIBUTE memory leak
5019 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
5020 memory. This structure is used by the PKCS#7 and CMS routines so any
5021 application which reads PKCS#7 or CMS data from untrusted sources is
5022 affected. SSL/TLS is not affected.
5024 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
5030 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
5031 This changes the decoding behaviour for some invalid messages,
5032 though the change is mostly in the more lenient direction, and
5033 legacy behaviour is preserved as much as possible.
5037 * In DSA_generate_parameters_ex, if the provided seed is too short,
5040 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
5042 ### Changes between 1.0.2c and 1.0.2d [9 Jul 2015]
5044 * Alternate chains certificate forgery
5046 During certificate verification, OpenSSL will attempt to find an
5047 alternative certificate chain if the first attempt to build such a chain
5048 fails. An error in the implementation of this logic can mean that an
5049 attacker could cause certain checks on untrusted certificates to be
5050 bypassed, such as the CA flag, enabling them to use a valid leaf
5051 certificate to act as a CA and "issue" an invalid certificate.
5053 This issue was reported to OpenSSL by Adam Langley/David Benjamin
5058 ### Changes between 1.0.2b and 1.0.2c [12 Jun 2015]
5060 * Fix HMAC ABI incompatibility. The previous version introduced an ABI
5061 incompatibility in the handling of HMAC. The previous ABI has now been
5066 ### Changes between 1.0.2a and 1.0.2b [11 Jun 2015]
5068 * Malformed ECParameters causes infinite loop
5070 When processing an ECParameters structure OpenSSL enters an infinite loop
5071 if the curve specified is over a specially malformed binary polynomial
5074 This can be used to perform denial of service against any
5075 system which processes public keys, certificate requests or
5076 certificates. This includes TLS clients and TLS servers with
5077 client authentication enabled.
5079 This issue was reported to OpenSSL by Joseph Barr-Pixton.
5084 * Exploitable out-of-bounds read in X509_cmp_time
5086 X509_cmp_time does not properly check the length of the ASN1_TIME
5087 string and can read a few bytes out of bounds. In addition,
5088 X509_cmp_time accepts an arbitrary number of fractional seconds in the
5091 An attacker can use this to craft malformed certificates and CRLs of
5092 various sizes and potentially cause a segmentation fault, resulting in
5093 a DoS on applications that verify certificates or CRLs. TLS clients
5094 that verify CRLs are affected. TLS clients and servers with client
5095 authentication enabled may be affected if they use custom verification
5098 This issue was reported to OpenSSL by Robert Swiecki (Google), and
5099 independently by Hanno Böck.
5104 * PKCS7 crash with missing EnvelopedContent
5106 The PKCS#7 parsing code does not handle missing inner EncryptedContent
5107 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
5108 with missing content and trigger a NULL pointer dereference on parsing.
5110 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
5111 structures from untrusted sources are affected. OpenSSL clients and
5112 servers are not affected.
5114 This issue was reported to OpenSSL by Michal Zalewski (Google).
5119 * CMS verify infinite loop with unknown hash function
5121 When verifying a signedData message the CMS code can enter an infinite loop
5122 if presented with an unknown hash function OID. This can be used to perform
5123 denial of service against any system which verifies signedData messages using
5125 This issue was reported to OpenSSL by Johannes Bauer.
5130 * Race condition handling NewSessionTicket
5132 If a NewSessionTicket is received by a multi-threaded client when attempting to
5133 reuse a previous ticket then a race condition can occur potentially leading to
5134 a double free of the ticket data.
5139 * Only support 256-bit or stronger elliptic curves with the
5140 'ecdh_auto' setting (server) or by default (client). Of supported
5141 curves, prefer P-256 (both).
5145 ### Changes between 1.0.2 and 1.0.2a [19 Mar 2015]
5147 * ClientHello sigalgs DoS fix
5149 If a client connects to an OpenSSL 1.0.2 server and renegotiates with an
5150 invalid signature algorithms extension a NULL pointer dereference will
5151 occur. This can be exploited in a DoS attack against the server.
5153 This issue was was reported to OpenSSL by David Ramos of Stanford
5157 *Stephen Henson and Matt Caswell*
5159 * Multiblock corrupted pointer fix
5161 OpenSSL 1.0.2 introduced the "multiblock" performance improvement. This
5162 feature only applies on 64 bit x86 architecture platforms that support AES
5163 NI instructions. A defect in the implementation of "multiblock" can cause
5164 OpenSSL's internal write buffer to become incorrectly set to NULL when
5165 using non-blocking IO. Typically, when the user application is using a
5166 socket BIO for writing, this will only result in a failed connection.
5167 However if some other BIO is used then it is likely that a segmentation
5168 fault will be triggered, thus enabling a potential DoS attack.
5170 This issue was reported to OpenSSL by Daniel Danner and Rainer Mueller.
5175 * Segmentation fault in DTLSv1_listen fix
5177 The DTLSv1_listen function is intended to be stateless and processes the
5178 initial ClientHello from many peers. It is common for user code to loop
5179 over the call to DTLSv1_listen until a valid ClientHello is received with
5180 an associated cookie. A defect in the implementation of DTLSv1_listen means
5181 that state is preserved in the SSL object from one invocation to the next
5182 that can lead to a segmentation fault. Errors processing the initial
5183 ClientHello can trigger this scenario. An example of such an error could be
5184 that a DTLS1.0 only client is attempting to connect to a DTLS1.2 only
5187 This issue was reported to OpenSSL by Per Allansson.
5192 * Segmentation fault in ASN1_TYPE_cmp fix
5194 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
5195 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
5196 certificate signature algorithm consistency this can be used to crash any
5197 certificate verification operation and exploited in a DoS attack. Any
5198 application which performs certificate verification is vulnerable including
5199 OpenSSL clients and servers which enable client authentication.
5204 * Segmentation fault for invalid PSS parameters fix
5206 The signature verification routines will crash with a NULL pointer
5207 dereference if presented with an ASN.1 signature using the RSA PSS
5208 algorithm and invalid parameters. Since these routines are used to verify
5209 certificate signature algorithms 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.
5214 This issue was was reported to OpenSSL by Brian Carpenter.
5219 * ASN.1 structure reuse memory corruption fix
5221 Reusing a structure in ASN.1 parsing may allow an attacker to cause
5222 memory corruption via an invalid write. Such reuse is and has been
5223 strongly discouraged and is believed to be rare.
5225 Applications that parse structures containing CHOICE or ANY DEFINED BY
5226 components may be affected. Certificate parsing (d2i_X509 and related
5227 functions) are however not affected. OpenSSL clients and servers are
5233 * PKCS7 NULL pointer dereferences fix
5235 The PKCS#7 parsing code does not handle missing outer ContentInfo
5236 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
5237 missing content and trigger a NULL pointer dereference on parsing.
5239 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
5240 otherwise parse PKCS#7 structures from untrusted sources are
5241 affected. OpenSSL clients and servers are not affected.
5243 This issue was reported to OpenSSL by Michal Zalewski (Google).
5248 * DoS via reachable assert in SSLv2 servers fix
5250 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
5251 servers that both support SSLv2 and enable export cipher suites by sending
5252 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
5254 This issue was discovered by Sean Burford (Google) and Emilia Käsper
5255 (OpenSSL development team).
5260 * Empty CKE with client auth and DHE fix
5262 If client auth is used then a server can seg fault in the event of a DHE
5263 ciphersuite being selected and a zero length ClientKeyExchange message
5264 being sent by the client. This could be exploited in a DoS attack.
5269 * Handshake with unseeded PRNG fix
5271 Under certain conditions an OpenSSL 1.0.2 client can complete a handshake
5272 with an unseeded PRNG. The conditions are:
5273 - The client is on a platform where the PRNG has not been seeded
5274 automatically, and the user has not seeded manually
5275 - A protocol specific client method version has been used (i.e. not
5276 SSL_client_methodv23)
5277 - A ciphersuite is used that does not require additional random data from
5278 the PRNG beyond the initial ClientHello client random (e.g. PSK-RC4-SHA).
5280 If the handshake succeeds then the client random that has been used will
5281 have been generated from a PRNG with insufficient entropy and therefore the
5282 output may be predictable.
5284 For example using the following command with an unseeded openssl will
5285 succeed on an unpatched platform:
5287 openssl s_client -psk 1a2b3c4d -tls1_2 -cipher PSK-RC4-SHA
5292 * Use After Free following d2i_ECPrivatekey error fix
5294 A malformed EC private key file consumed via the d2i_ECPrivateKey function
5295 could cause a use after free condition. This, in turn, could cause a double
5296 free in several private key parsing functions (such as d2i_PrivateKey
5297 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
5298 for applications that receive EC private keys from untrusted
5299 sources. This scenario is considered rare.
5301 This issue was discovered by the BoringSSL project and fixed in their
5307 * X509_to_X509_REQ NULL pointer deref fix
5309 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
5310 the certificate key is invalid. This function is rarely used in practice.
5312 This issue was discovered by Brian Carpenter.
5317 * Removed the export ciphers from the DEFAULT ciphers
5321 ### Changes between 1.0.1l and 1.0.2 [22 Jan 2015]
5323 * Facilitate "universal" ARM builds targeting range of ARM ISAs, e.g.
5324 ARMv5 through ARMv8, as opposite to "locking" it to single one.
5325 So far those who have to target multiple platforms would compromise
5326 and argue that binary targeting say ARMv5 would still execute on
5327 ARMv8. "Universal" build resolves this compromise by providing
5328 near-optimal performance even on newer platforms.
5332 * Accelerated NIST P-256 elliptic curve implementation for x86_64
5333 (other platforms pending).
5335 *Shay Gueron & Vlad Krasnov (Intel Corp), Andy Polyakov*
5337 * Add support for the SignedCertificateTimestampList certificate and
5338 OCSP response extensions from RFC6962.
5342 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
5343 for corner cases. (Certain input points at infinity could lead to
5344 bogus results, with non-infinity inputs mapped to infinity too.)
5348 * Initial support for PowerISA 2.0.7, first implemented in POWER8.
5349 This covers AES, SHA256/512 and GHASH. "Initial" means that most
5350 common cases are optimized and there still is room for further
5351 improvements. Vector Permutation AES for Altivec is also added.
5355 * Add support for little-endian ppc64 Linux target.
5357 *Marcelo Cerri (IBM)*
5359 * Initial support for AMRv8 ISA crypto extensions. This covers AES,
5360 SHA1, SHA256 and GHASH. "Initial" means that most common cases
5361 are optimized and there still is room for further improvements.
5362 Both 32- and 64-bit modes are supported.
5364 *Andy Polyakov, Ard Biesheuvel (Linaro)*
5366 * Improved ARMv7 NEON support.
5370 * Support for SPARC Architecture 2011 crypto extensions, first
5371 implemented in SPARC T4. This covers AES, DES, Camellia, SHA1,
5372 SHA256/512, MD5, GHASH and modular exponentiation.
5374 *Andy Polyakov, David Miller*
5376 * Accelerated modular exponentiation for Intel processors, a.k.a.
5379 *Shay Gueron & Vlad Krasnov (Intel Corp)*
5381 * Support for new and upcoming Intel processors, including AVX2,
5382 BMI and SHA ISA extensions. This includes additional "stitched"
5383 implementations, AESNI-SHA256 and GCM, and multi-buffer support
5386 This work was sponsored by Intel Corp.
5390 * Support for DTLS 1.2. This adds two sets of DTLS methods: DTLS_*_method()
5391 supports both DTLS 1.2 and 1.0 and should use whatever version the peer
5392 supports and DTLSv1_2_*_method() which supports DTLS 1.2 only.
5396 * Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
5397 this fixes a limitation in previous versions of OpenSSL.
5401 * Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
5402 MGF1 digest and OAEP label.
5406 * Add EVP support for key wrapping algorithms, to avoid problems with
5407 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
5408 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
5409 algorithms and include tests cases.
5413 * Add functions to allocate and set the fields of an ECDSA_METHOD
5416 *Douglas E. Engert, Steve Henson*
5418 * New functions OPENSSL_gmtime_diff and ASN1_TIME_diff to find the
5419 difference in days and seconds between two tm or ASN1_TIME structures.
5423 * Add -rev test option to s_server to just reverse order of characters
5424 received by client and send back to server. Also prints an abbreviated
5425 summary of the connection parameters.
5429 * New option -brief for s_client and s_server to print out a brief summary
5430 of connection parameters.
5434 * Add callbacks for arbitrary TLS extensions.
5436 *Trevor Perrin <trevp@trevp.net> and Ben Laurie*
5438 * New option -crl_download in several openssl utilities to download CRLs
5439 from CRLDP extension in certificates.
5443 * New options -CRL and -CRLform for s_client and s_server for CRLs.
5447 * New function X509_CRL_diff to generate a delta CRL from the difference
5448 of two full CRLs. Add support to "crl" utility.
5452 * New functions to set lookup_crls function and to retrieve
5453 X509_STORE from X509_STORE_CTX.
5457 * Print out deprecated issuer and subject unique ID fields in
5462 * Extend OCSP I/O functions so they can be used for simple general purpose
5463 HTTP as well as OCSP. New wrapper function which can be used to download
5464 CRLs using the OCSP API.
5468 * Delegate command line handling in s_client/s_server to SSL_CONF APIs.
5472 * `SSL_CONF*` functions. These provide a common framework for application
5473 configuration using configuration files or command lines.
5477 * SSL/TLS tracing code. This parses out SSL/TLS records using the
5478 message callback and prints the results. Needs compile time option
5479 "enable-ssl-trace". New options to s_client and s_server to enable
5484 * New ctrl and macro to retrieve supported points extensions.
5485 Print out extension in s_server and s_client.
5489 * New functions to retrieve certificate signature and signature
5494 * Add functions to retrieve and manipulate the raw cipherlist sent by a
5499 * New Suite B modes for TLS code. These use and enforce the requirements
5500 of RFC6460: restrict ciphersuites, only permit Suite B algorithms and
5501 only use Suite B curves. The Suite B modes can be set by using the
5502 strings "SUITEB128", "SUITEB192" or "SUITEB128ONLY" for the cipherstring.
5506 * New chain verification flags for Suite B levels of security. Check
5507 algorithms are acceptable when flags are set in X509_verify_cert.
5511 * Make tls1_check_chain return a set of flags indicating checks passed
5512 by a certificate chain. Add additional tests to handle client
5513 certificates: checks for matching certificate type and issuer name
5518 * If an attempt is made to use a signature algorithm not in the peer
5519 preference list abort the handshake. If client has no suitable
5520 signature algorithms in response to a certificate request do not
5521 use the certificate.
5525 * If server EC tmp key is not in client preference list abort handshake.
5529 * Add support for certificate stores in CERT structure. This makes it
5530 possible to have different stores per SSL structure or one store in
5531 the parent SSL_CTX. Include distinct stores for certificate chain
5532 verification and chain building. New ctrl SSL_CTRL_BUILD_CERT_CHAIN
5533 to build and store a certificate chain in CERT structure: returning
5534 an error if the chain cannot be built: this will allow applications
5535 to test if a chain is correctly configured.
5537 Note: if the CERT based stores are not set then the parent SSL_CTX
5538 store is used to retain compatibility with existing behaviour.
5542 * New function ssl_set_client_disabled to set a ciphersuite disabled
5543 mask based on the current session, check mask when sending client
5544 hello and checking the requested ciphersuite.
5548 * New ctrls to retrieve and set certificate types in a certificate
5549 request message. Print out received values in s_client. If certificate
5550 types is not set with custom values set sensible values based on
5551 supported signature algorithms.
5555 * Support for distinct client and server supported signature algorithms.
5559 * Add certificate callback. If set this is called whenever a certificate
5560 is required by client or server. An application can decide which
5561 certificate chain to present based on arbitrary criteria: for example
5562 supported signature algorithms. Add very simple example to s_server.
5563 This fixes many of the problems and restrictions of the existing client
5564 certificate callback: for example you can now clear an existing
5565 certificate and specify the whole chain.
5569 * Add new "valid_flags" field to CERT_PKEY structure which determines what
5570 the certificate can be used for (if anything). Set valid_flags field
5571 in new tls1_check_chain function. Simplify ssl_set_cert_masks which used
5572 to have similar checks in it.
5574 Add new "cert_flags" field to CERT structure and include a "strict mode".
5575 This enforces some TLS certificate requirements (such as only permitting
5576 certificate signature algorithms contained in the supported algorithms
5577 extension) which some implementations ignore: this option should be used
5578 with caution as it could cause interoperability issues.
5582 * Update and tidy signature algorithm extension processing. Work out
5583 shared signature algorithms based on preferences and peer algorithms
5584 and print them out in s_client and s_server. Abort handshake if no
5585 shared signature algorithms.
5589 * Add new functions to allow customised supported signature algorithms
5590 for SSL and SSL_CTX structures. Add options to s_client and s_server
5595 * New function SSL_certs_clear() to delete all references to certificates
5596 from an SSL structure. Before this once a certificate had been added
5597 it couldn't be removed.
5601 * Integrate hostname, email address and IP address checking with certificate
5602 verification. New verify options supporting checking in openssl utility.
5606 * Fixes and wildcard matching support to hostname and email checking
5607 functions. Add manual page.
5609 *Florian Weimer (Red Hat Product Security Team)*
5611 * New functions to check a hostname email or IP address against a
5612 certificate. Add options x509 utility to print results of checks against
5617 * Fix OCSP checking.
5619 *Rob Stradling <rob.stradling@comodo.com> and Ben Laurie*
5621 * Initial experimental support for explicitly trusted non-root CAs.
5622 OpenSSL still tries to build a complete chain to a root but if an
5623 intermediate CA has a trust setting included that is used. The first
5624 setting is used: whether to trust (e.g., -addtrust option to the x509
5629 * Add -trusted_first option which attempts to find certificates in the
5630 trusted store even if an untrusted chain is also supplied.
5634 * MIPS assembly pack updates: support for MIPS32r2 and SmartMIPS ASE,
5635 platform support for Linux and Android.
5639 * Support for linux-x32, ILP32 environment in x86_64 framework.
5643 * Experimental multi-implementation support for FIPS capable OpenSSL.
5644 When in FIPS mode the approved implementations are used as normal,
5645 when not in FIPS mode the internal unapproved versions are used instead.
5646 This means that the FIPS capable OpenSSL isn't forced to use the
5647 (often lower performance) FIPS implementations outside FIPS mode.
5651 * Transparently support X9.42 DH parameters when calling
5652 PEM_read_bio_DHparameters. This means existing applications can handle
5653 the new parameter format automatically.
5657 * Initial experimental support for X9.42 DH parameter format: mainly
5658 to support use of 'q' parameter for RFC5114 parameters.
5662 * Add DH parameters from RFC5114 including test data to dhtest.
5666 * Support for automatic EC temporary key parameter selection. If enabled
5667 the most preferred EC parameters are automatically used instead of
5668 hardcoded fixed parameters. Now a server just has to call:
5669 SSL_CTX_set_ecdh_auto(ctx, 1) and the server will automatically
5670 support ECDH and use the most appropriate parameters.
5674 * Enhance and tidy EC curve and point format TLS extension code. Use
5675 static structures instead of allocation if default values are used.
5676 New ctrls to set curves we wish to support and to retrieve shared curves.
5677 Print out shared curves in s_server. New options to s_server and s_client
5678 to set list of supported curves.
5682 * New ctrls to retrieve supported signature algorithms and
5683 supported curve values as an array of NIDs. Extend openssl utility
5684 to print out received values.
5688 * Add new APIs EC_curve_nist2nid and EC_curve_nid2nist which convert
5689 between NIDs and the more common NIST names such as "P-256". Enhance
5690 ecparam utility and ECC method to recognise the NIST names for curves.
5694 * Enhance SSL/TLS certificate chain handling to support different
5695 chains for each certificate instead of one chain in the parent SSL_CTX.
5699 * Support for fixed DH ciphersuite client authentication: where both
5700 server and client use DH certificates with common parameters.
5704 * Support for fixed DH ciphersuites: those requiring DH server
5709 * New function i2d_re_X509_tbs for re-encoding the TBS portion of
5711 Note: Related 1.0.2-beta specific macros X509_get_cert_info,
5712 X509_CINF_set_modified, X509_CINF_get_issuer, X509_CINF_get_extensions and
5713 X509_CINF_get_signature were reverted post internal team review.
5718 ### Changes between 1.0.1t and 1.0.1u [22 Sep 2016]
5720 * OCSP Status Request extension unbounded memory growth
5722 A malicious client can send an excessively large OCSP Status Request
5723 extension. If that client continually requests renegotiation, sending a
5724 large OCSP Status Request extension each time, then there will be unbounded
5725 memory growth on the server. This will eventually lead to a Denial Of
5726 Service attack through memory exhaustion. Servers with a default
5727 configuration are vulnerable even if they do not support OCSP. Builds using
5728 the "no-ocsp" build time option are not affected.
5730 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5735 * In order to mitigate the SWEET32 attack, the DES ciphers were moved from
5738 This issue was reported to OpenSSL Karthikeyan Bhargavan and Gaetan
5744 * OOB write in MDC2_Update()
5746 An overflow can occur in MDC2_Update() either if called directly or
5747 through the EVP_DigestUpdate() function using MDC2. If an attacker
5748 is able to supply very large amounts of input data after a previous
5749 call to EVP_EncryptUpdate() with a partial block then a length check
5750 can overflow resulting in a heap corruption.
5752 The amount of data needed is comparable to SIZE_MAX which is impractical
5755 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5760 * Malformed SHA512 ticket DoS
5762 If a server uses SHA512 for TLS session ticket HMAC it is vulnerable to a
5763 DoS attack where a malformed ticket will result in an OOB read which will
5766 The use of SHA512 in TLS session tickets is comparatively rare as it requires
5767 a custom server callback and ticket lookup mechanism.
5769 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5774 * OOB write in BN_bn2dec()
5776 The function BN_bn2dec() does not check the return value of BN_div_word().
5777 This can cause an OOB write if an application uses this function with an
5778 overly large BIGNUM. This could be a problem if an overly large certificate
5779 or CRL is printed out from an untrusted source. TLS is not affected because
5780 record limits will reject an oversized certificate before it is parsed.
5782 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5787 * OOB read in TS_OBJ_print_bio()
5789 The function TS_OBJ_print_bio() misuses OBJ_obj2txt(): the return value is
5790 the total length the OID text representation would use and not the amount
5791 of data written. This will result in OOB reads when large OIDs are
5794 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5799 * Pointer arithmetic undefined behaviour
5801 Avoid some undefined pointer arithmetic
5803 A common idiom in the codebase is to check limits in the following manner:
5806 Where "p" points to some malloc'd data of SIZE bytes and
5809 "len" here could be from some externally supplied data (e.g. from a TLS
5812 The rules of C pointer arithmetic are such that "p + len" is only well
5813 defined where len <= SIZE. Therefore the above idiom is actually
5814 undefined behaviour.
5816 For example this could cause problems if some malloc implementation
5817 provides an address for "p" such that "p + len" actually overflows for
5818 values of len that are too big and therefore p + len < limit.
5820 This issue was reported to OpenSSL by Guido Vranken
5825 * Constant time flag not preserved in DSA signing
5827 Operations in the DSA signing algorithm should run in constant time in
5828 order to avoid side channel attacks. A flaw in the OpenSSL DSA
5829 implementation means that a non-constant time codepath is followed for
5830 certain operations. This has been demonstrated through a cache-timing
5831 attack to be sufficient for an attacker to recover the private DSA key.
5833 This issue was reported by César Pereida (Aalto University), Billy Brumley
5834 (Tampere University of Technology), and Yuval Yarom (The University of
5835 Adelaide and NICTA).
5840 * DTLS buffered message DoS
5842 In a DTLS connection where handshake messages are delivered out-of-order
5843 those messages that OpenSSL is not yet ready to process will be buffered
5844 for later use. Under certain circumstances, a flaw in the logic means that
5845 those messages do not get removed from the buffer even though the handshake
5846 has been completed. An attacker could force up to approx. 15 messages to
5847 remain in the buffer when they are no longer required. These messages will
5848 be cleared when the DTLS connection is closed. The default maximum size for
5849 a message is 100k. Therefore the attacker could force an additional 1500k
5850 to be consumed per connection. By opening many simulataneous connections an
5851 attacker could cause a DoS attack through memory exhaustion.
5853 This issue was reported to OpenSSL by Quan Luo.
5858 * DTLS replay protection DoS
5860 A flaw in the DTLS replay attack protection mechanism means that records
5861 that arrive for future epochs update the replay protection "window" before
5862 the MAC for the record has been validated. This could be exploited by an
5863 attacker by sending a record for the next epoch (which does not have to
5864 decrypt or have a valid MAC), with a very large sequence number. This means
5865 that all subsequent legitimate packets are dropped causing a denial of
5866 service for a specific DTLS connection.
5868 This issue was reported to OpenSSL by the OCAP audit team.
5873 * Certificate message OOB reads
5875 In OpenSSL 1.0.2 and earlier some missing message length checks can result
5876 in OOB reads of up to 2 bytes beyond an allocated buffer. There is a
5877 theoretical DoS risk but this has not been observed in practice on common
5880 The messages affected are client certificate, client certificate request
5881 and server certificate. As a result the attack can only be performed
5882 against a client or a server which enables client authentication.
5884 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5889 ### Changes between 1.0.1s and 1.0.1t [3 May 2016]
5891 * Prevent padding oracle in AES-NI CBC MAC check
5893 A MITM attacker can use a padding oracle attack to decrypt traffic
5894 when the connection uses an AES CBC cipher and the server support
5897 This issue was introduced as part of the fix for Lucky 13 padding
5898 attack ([CVE-2013-0169]). The padding check was rewritten to be in
5899 constant time by making sure that always the same bytes are read and
5900 compared against either the MAC or padding bytes. But it no longer
5901 checked that there was enough data to have both the MAC and padding
5904 This issue was reported by Juraj Somorovsky using TLS-Attacker.
5909 * Fix EVP_EncodeUpdate overflow
5911 An overflow can occur in the EVP_EncodeUpdate() function which is used for
5912 Base64 encoding of binary data. If an attacker is able to supply very large
5913 amounts of input data then a length check can overflow resulting in a heap
5916 Internally to OpenSSL the EVP_EncodeUpdate() function is primarly used by
5917 the `PEM_write_bio*` family of functions. These are mainly used within the
5918 OpenSSL command line applications, so any application which processes data
5919 from an untrusted source and outputs it as a PEM file should be considered
5920 vulnerable to this issue. User applications that call these APIs directly
5921 with large amounts of untrusted data may also be vulnerable.
5923 This issue was reported by Guido Vranken.
5928 * Fix EVP_EncryptUpdate overflow
5930 An overflow can occur in the EVP_EncryptUpdate() function. If an attacker
5931 is able to supply very large amounts of input data after a previous call to
5932 EVP_EncryptUpdate() with a partial block then a length check can overflow
5933 resulting in a heap corruption. Following an analysis of all OpenSSL
5934 internal usage of the EVP_EncryptUpdate() function all usage is one of two
5935 forms. The first form is where the EVP_EncryptUpdate() call is known to be
5936 the first called function after an EVP_EncryptInit(), and therefore that
5937 specific call must be safe. The second form is where the length passed to
5938 EVP_EncryptUpdate() can be seen from the code to be some small value and
5939 therefore there is no possibility of an overflow. Since all instances are
5940 one of these two forms, it is believed that there can be no overflows in
5941 internal code due to this problem. It should be noted that
5942 EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths.
5943 Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances
5944 of these calls have also been analysed too and it is believed there are no
5945 instances in internal usage where an overflow could occur.
5947 This issue was reported by Guido Vranken.
5952 * Prevent ASN.1 BIO excessive memory allocation
5954 When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio()
5955 a short invalid encoding can casuse allocation of large amounts of memory
5956 potentially consuming excessive resources or exhausting memory.
5958 Any application parsing untrusted data through d2i BIO functions is
5959 affected. The memory based functions such as d2i_X509() are *not* affected.
5960 Since the memory based functions are used by the TLS library, TLS
5961 applications are not affected.
5963 This issue was reported by Brian Carpenter.
5970 ASN1 Strings that are over 1024 bytes can cause an overread in applications
5971 using the X509_NAME_oneline() function on EBCDIC systems. This could result
5972 in arbitrary stack data being returned in the buffer.
5974 This issue was reported by Guido Vranken.
5979 * Modify behavior of ALPN to invoke callback after SNI/servername
5980 callback, such that updates to the SSL_CTX affect ALPN.
5984 * Remove LOW from the DEFAULT cipher list. This removes singles DES from the
5989 * Only remove the SSLv2 methods with the no-ssl2-method option. When the
5990 methods are enabled and ssl2 is disabled the methods return NULL.
5994 ### Changes between 1.0.1r and 1.0.1s [1 Mar 2016]
5996 * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL.
5997 Builds that are not configured with "enable-weak-ssl-ciphers" will not
5998 provide any "EXPORT" or "LOW" strength ciphers.
6002 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2
6003 is by default disabled at build-time. Builds that are not configured with
6004 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used,
6005 users who want to negotiate SSLv2 via the version-flexible SSLv23_method()
6006 will need to explicitly call either of:
6008 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2);
6010 SSL_clear_options(ssl, SSL_OP_NO_SSLv2);
6012 as appropriate. Even if either of those is used, or the application
6013 explicitly uses the version-specific SSLv2_method() or its client and
6014 server variants, SSLv2 ciphers vulnerable to exhaustive search key
6015 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT
6016 ciphers, and SSLv2 56-bit DES are no longer available.
6021 * Fix a double-free in DSA code
6023 A double free bug was discovered when OpenSSL parses malformed DSA private
6024 keys and could lead to a DoS attack or memory corruption for applications
6025 that receive DSA private keys from untrusted sources. This scenario is
6028 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using
6034 * Disable SRP fake user seed to address a server memory leak.
6036 Add a new method SRP_VBASE_get1_by_user that handles the seed properly.
6038 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
6039 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user
6040 was changed to ignore the "fake user" SRP seed, even if the seed
6043 Users should use SRP_VBASE_get1_by_user instead. Note that in
6044 SRP_VBASE_get1_by_user, caller must free the returned value. Note
6045 also that even though configuring the SRP seed attempts to hide
6046 invalid usernames by continuing the handshake with fake
6047 credentials, this behaviour is not constant time and no strong
6048 guarantees are made that the handshake is indistinguishable from
6049 that of a valid user.
6054 * Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption
6056 In the BN_hex2bn function the number of hex digits is calculated using an
6057 int value `i`. Later `bn_expand` is called with a value of `i * 4`. For
6058 large values of `i` this can result in `bn_expand` not allocating any
6059 memory because `i * 4` is negative. This can leave the internal BIGNUM data
6060 field as NULL leading to a subsequent NULL ptr deref. For very large values
6061 of `i`, the calculation `i * 4` could be a positive value smaller than `i`.
6062 In this case memory is allocated to the internal BIGNUM data field, but it
6063 is insufficiently sized leading to heap corruption. A similar issue exists
6064 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn
6065 is ever called by user applications with very large untrusted hex/dec data.
6066 This is anticipated to be a rare occurrence.
6068 All OpenSSL internal usage of these functions use data that is not expected
6069 to be untrusted, e.g. config file data or application command line
6070 arguments. If user developed applications generate config file data based
6071 on untrusted data then it is possible that this could also lead to security
6072 consequences. This is also anticipated to be rare.
6074 This issue was reported to OpenSSL by Guido Vranken.
6079 * Fix memory issues in `BIO_*printf` functions
6081 The internal `fmtstr` function used in processing a "%s" format string in
6082 the `BIO_*printf` functions could overflow while calculating the length of a
6083 string and cause an OOB read when printing very long strings.
6085 Additionally the internal `doapr_outch` function can attempt to write to an
6086 OOB memory location (at an offset from the NULL pointer) in the event of a
6087 memory allocation failure. In 1.0.2 and below this could be caused where
6088 the size of a buffer to be allocated is greater than INT_MAX. E.g. this
6089 could be in processing a very long "%s" format string. Memory leaks can
6092 The first issue may mask the second issue dependent on compiler behaviour.
6093 These problems could enable attacks where large amounts of untrusted data
6094 is passed to the `BIO_*printf` functions. If applications use these functions
6095 in this way then they could be vulnerable. OpenSSL itself uses these
6096 functions when printing out human-readable dumps of ASN.1 data. Therefore
6097 applications that print this data could be vulnerable if the data is from
6098 untrusted sources. OpenSSL command line applications could also be
6099 vulnerable where they print out ASN.1 data, or if untrusted data is passed
6100 as command line arguments.
6102 Libssl is not considered directly vulnerable. Additionally certificates etc
6103 received via remote connections via libssl are also unlikely to be able to
6104 trigger these issues because of message size limits enforced within libssl.
6106 This issue was reported to OpenSSL Guido Vranken.
6111 * Side channel attack on modular exponentiation
6113 A side-channel attack was found which makes use of cache-bank conflicts on
6114 the Intel Sandy-Bridge microarchitecture which could lead to the recovery
6115 of RSA keys. The ability to exploit this issue is limited as it relies on
6116 an attacker who has control of code in a thread running on the same
6117 hyper-threaded core as the victim thread which is performing decryptions.
6119 This issue was reported to OpenSSL by Yuval Yarom, The University of
6120 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and
6121 Nadia Heninger, University of Pennsylvania with more information at
6122 <http://cachebleed.info>.
6127 * Change the req command to generate a 2048-bit RSA/DSA key by default,
6128 if no keysize is specified with default_bits. This fixes an
6129 omission in an earlier change that changed all RSA/DSA key generation
6130 commands to use 2048 bits by default.
6134 ### Changes between 1.0.1q and 1.0.1r [28 Jan 2016]
6136 * Protection for DH small subgroup attacks
6138 As a precautionary measure the SSL_OP_SINGLE_DH_USE option has been
6139 switched on by default and cannot be disabled. This could have some
6144 * SSLv2 doesn't block disabled ciphers
6146 A malicious client can negotiate SSLv2 ciphers that have been disabled on
6147 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
6148 been disabled, provided that the SSLv2 protocol was not also disabled via
6151 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
6152 and Sebastian Schinzel.
6157 * Reject DH handshakes with parameters shorter than 1024 bits.
6161 ### Changes between 1.0.1p and 1.0.1q [3 Dec 2015]
6163 * Certificate verify crash with missing PSS parameter
6165 The signature verification routines will crash with a NULL pointer
6166 dereference if presented with an ASN.1 signature using the RSA PSS
6167 algorithm and absent mask generation function parameter. Since these
6168 routines are used to verify certificate signature algorithms this can be
6169 used to crash any certificate verification operation and exploited in a
6170 DoS attack. Any application which performs certificate verification is
6171 vulnerable including OpenSSL clients and servers which enable client
6174 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
6179 * X509_ATTRIBUTE memory leak
6181 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
6182 memory. This structure is used by the PKCS#7 and CMS routines so any
6183 application which reads PKCS#7 or CMS data from untrusted sources is
6184 affected. SSL/TLS is not affected.
6186 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
6192 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
6193 This changes the decoding behaviour for some invalid messages,
6194 though the change is mostly in the more lenient direction, and
6195 legacy behaviour is preserved as much as possible.
6199 * In DSA_generate_parameters_ex, if the provided seed is too short,
6200 use a random seed, as already documented.
6202 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
6204 ### Changes between 1.0.1o and 1.0.1p [9 Jul 2015]
6206 * Alternate chains certificate forgery
6208 During certificate verfification, OpenSSL will attempt to find an
6209 alternative certificate chain if the first attempt to build such a chain
6210 fails. An error in the implementation of this logic can mean that an
6211 attacker could cause certain checks on untrusted certificates to be
6212 bypassed, such as the CA flag, enabling them to use a valid leaf
6213 certificate to act as a CA and "issue" an invalid certificate.
6215 This issue was reported to OpenSSL by Adam Langley/David Benjamin
6221 * Race condition handling PSK identify hint
6223 If PSK identity hints are received by a multi-threaded client then
6224 the values are wrongly updated in the parent SSL_CTX structure. This can
6225 result in a race condition potentially leading to a double free of the
6231 ### Changes between 1.0.1n and 1.0.1o [12 Jun 2015]
6233 * Fix HMAC ABI incompatibility. The previous version introduced an ABI
6234 incompatibility in the handling of HMAC. The previous ABI has now been
6237 ### Changes between 1.0.1m and 1.0.1n [11 Jun 2015]
6239 * Malformed ECParameters causes infinite loop
6241 When processing an ECParameters structure OpenSSL enters an infinite loop
6242 if the curve specified is over a specially malformed binary polynomial
6245 This can be used to perform denial of service against any
6246 system which processes public keys, certificate requests or
6247 certificates. This includes TLS clients and TLS servers with
6248 client authentication enabled.
6250 This issue was reported to OpenSSL by Joseph Barr-Pixton.
6255 * Exploitable out-of-bounds read in X509_cmp_time
6257 X509_cmp_time does not properly check the length of the ASN1_TIME
6258 string and can read a few bytes out of bounds. In addition,
6259 X509_cmp_time accepts an arbitrary number of fractional seconds in the
6262 An attacker can use this to craft malformed certificates and CRLs of
6263 various sizes and potentially cause a segmentation fault, resulting in
6264 a DoS on applications that verify certificates or CRLs. TLS clients
6265 that verify CRLs are affected. TLS clients and servers with client
6266 authentication enabled may be affected if they use custom verification
6269 This issue was reported to OpenSSL by Robert Swiecki (Google), and
6270 independently by Hanno Böck.
6275 * PKCS7 crash with missing EnvelopedContent
6277 The PKCS#7 parsing code does not handle missing inner EncryptedContent
6278 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
6279 with missing content and trigger a NULL pointer dereference on parsing.
6281 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
6282 structures from untrusted sources are affected. OpenSSL clients and
6283 servers are not affected.
6285 This issue was reported to OpenSSL by Michal Zalewski (Google).
6290 * CMS verify infinite loop with unknown hash function
6292 When verifying a signedData message the CMS code can enter an infinite loop
6293 if presented with an unknown hash function OID. This can be used to perform
6294 denial of service against any system which verifies signedData messages using
6296 This issue was reported to OpenSSL by Johannes Bauer.
6301 * Race condition handling NewSessionTicket
6303 If a NewSessionTicket is received by a multi-threaded client when attempting to
6304 reuse a previous ticket then a race condition can occur potentially leading to
6305 a double free of the ticket data.
6310 * Reject DH handshakes with parameters shorter than 768 bits.
6312 *Kurt Roeckx and Emilia Kasper*
6314 * dhparam: generate 2048-bit parameters by default.
6316 *Kurt Roeckx and Emilia Kasper*
6318 ### Changes between 1.0.1l and 1.0.1m [19 Mar 2015]
6320 * Segmentation fault in ASN1_TYPE_cmp fix
6322 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
6323 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
6324 certificate signature algorithm consistency this can be used to crash any
6325 certificate verification operation and exploited in a DoS attack. Any
6326 application which performs certificate verification is vulnerable including
6327 OpenSSL clients and servers which enable client authentication.
6332 * ASN.1 structure reuse memory corruption fix
6334 Reusing a structure in ASN.1 parsing may allow an attacker to cause
6335 memory corruption via an invalid write. Such reuse is and has been
6336 strongly discouraged and is believed to be rare.
6338 Applications that parse structures containing CHOICE or ANY DEFINED BY
6339 components may be affected. Certificate parsing (d2i_X509 and related
6340 functions) are however not affected. OpenSSL clients and servers are
6346 * PKCS7 NULL pointer dereferences fix
6348 The PKCS#7 parsing code does not handle missing outer ContentInfo
6349 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
6350 missing content and trigger a NULL pointer dereference on parsing.
6352 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
6353 otherwise parse PKCS#7 structures from untrusted sources are
6354 affected. OpenSSL clients and servers are not affected.
6356 This issue was reported to OpenSSL by Michal Zalewski (Google).
6361 * DoS via reachable assert in SSLv2 servers fix
6363 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
6364 servers that both support SSLv2 and enable export cipher suites by sending
6365 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
6367 This issue was discovered by Sean Burford (Google) and Emilia Käsper
6368 (OpenSSL development team).
6373 * Use After Free following d2i_ECPrivatekey error fix
6375 A malformed EC private key file consumed via the d2i_ECPrivateKey function
6376 could cause a use after free condition. This, in turn, could cause a double
6377 free in several private key parsing functions (such as d2i_PrivateKey
6378 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
6379 for applications that receive EC private keys from untrusted
6380 sources. This scenario is considered rare.
6382 This issue was discovered by the BoringSSL project and fixed in their
6388 * X509_to_X509_REQ NULL pointer deref fix
6390 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
6391 the certificate key is invalid. This function is rarely used in practice.
6393 This issue was discovered by Brian Carpenter.
6398 * Removed the export ciphers from the DEFAULT ciphers
6402 ### Changes between 1.0.1k and 1.0.1l [15 Jan 2015]
6404 * Build fixes for the Windows and OpenVMS platforms
6406 *Matt Caswell and Richard Levitte*
6408 ### Changes between 1.0.1j and 1.0.1k [8 Jan 2015]
6410 * Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
6411 message can cause a segmentation fault in OpenSSL due to a NULL pointer
6412 dereference. This could lead to a Denial Of Service attack. Thanks to
6413 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
6418 * Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
6419 dtls1_buffer_record function under certain conditions. In particular this
6420 could occur if an attacker sent repeated DTLS records with the same
6421 sequence number but for the next epoch. The memory leak could be exploited
6422 by an attacker in a Denial of Service attack through memory exhaustion.
6423 Thanks to Chris Mueller for reporting this issue.
6428 * Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
6429 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
6430 method would be set to NULL which could later result in a NULL pointer
6431 dereference. Thanks to Frank Schmirler for reporting this issue.
6436 * Abort handshake if server key exchange message is omitted for ephemeral
6439 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
6440 reporting this issue.
6445 * Remove non-export ephemeral RSA code on client and server. This code
6446 violated the TLS standard by allowing the use of temporary RSA keys in
6447 non-export ciphersuites and could be used by a server to effectively
6448 downgrade the RSA key length used to a value smaller than the server
6449 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
6450 INRIA or reporting this issue.
6455 * Fixed issue where DH client certificates are accepted without verification.
6456 An OpenSSL server will accept a DH certificate for client authentication
6457 without the certificate verify message. This effectively allows a client to
6458 authenticate without the use of a private key. This only affects servers
6459 which trust a client certificate authority which issues certificates
6460 containing DH keys: these are extremely rare and hardly ever encountered.
6461 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
6467 * Ensure that the session ID context of an SSL is updated when its
6468 SSL_CTX is updated via SSL_set_SSL_CTX.
6470 The session ID context is typically set from the parent SSL_CTX,
6471 and can vary with the CTX.
6475 * Fix various certificate fingerprint issues.
6477 By using non-DER or invalid encodings outside the signed portion of a
6478 certificate the fingerprint can be changed without breaking the signature.
6479 Although no details of the signed portion of the certificate can be changed
6480 this can cause problems with some applications: e.g. those using the
6481 certificate fingerprint for blacklists.
6483 1. Reject signatures with non zero unused bits.
6485 If the BIT STRING containing the signature has non zero unused bits reject
6486 the signature. All current signature algorithms require zero unused bits.
6488 2. Check certificate algorithm consistency.
6490 Check the AlgorithmIdentifier inside TBS matches the one in the
6491 certificate signature. NB: this will result in signature failure
6492 errors for some broken certificates.
6494 Thanks to Konrad Kraszewski from Google for reporting this issue.
6496 3. Check DSA/ECDSA signatures use DER.
6498 Re-encode DSA/ECDSA signatures and compare with the original received
6499 signature. Return an error if there is a mismatch.
6501 This will reject various cases including garbage after signature
6502 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
6503 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
6504 (negative or with leading zeroes).
6506 Further analysis was conducted and fixes were developed by Stephen Henson
6507 of the OpenSSL core team.
6513 * Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
6514 results on some platforms, including x86_64. This bug occurs at random
6515 with a very low probability, and is not known to be exploitable in any
6516 way, though its exact impact is difficult to determine. Thanks to Pieter
6517 Wuille (Blockstream) who reported this issue and also suggested an initial
6518 fix. Further analysis was conducted by the OpenSSL development team and
6519 Adam Langley of Google. The final fix was developed by Andy Polyakov of
6520 the OpenSSL core team.
6525 * Do not resume sessions on the server if the negotiated protocol
6526 version does not match the session's version. Resuming with a different
6527 version, while not strictly forbidden by the RFC, is of questionable
6528 sanity and breaks all known clients.
6530 *David Benjamin, Emilia Käsper*
6532 * Tighten handling of the ChangeCipherSpec (CCS) message: reject
6533 early CCS messages during renegotiation. (Note that because
6534 renegotiation is encrypted, this early CCS was not exploitable.)
6538 * Tighten client-side session ticket handling during renegotiation:
6539 ensure that the client only accepts a session ticket if the server sends
6540 the extension anew in the ServerHello. Previously, a TLS client would
6541 reuse the old extension state and thus accept a session ticket if one was
6542 announced in the initial ServerHello.
6544 Similarly, ensure that the client requires a session ticket if one
6545 was advertised in the ServerHello. Previously, a TLS client would
6546 ignore a missing NewSessionTicket message.
6550 ### Changes between 1.0.1i and 1.0.1j [15 Oct 2014]
6554 A flaw in the DTLS SRTP extension parsing code allows an attacker, who
6555 sends a carefully crafted handshake message, to cause OpenSSL to fail
6556 to free up to 64k of memory causing a memory leak. This could be
6557 exploited in a Denial Of Service attack. This issue affects OpenSSL
6558 1.0.1 server implementations for both SSL/TLS and DTLS regardless of
6559 whether SRTP is used or configured. Implementations of OpenSSL that
6560 have been compiled with OPENSSL_NO_SRTP defined are not affected.
6562 The fix was developed by the OpenSSL team.
6567 * Session Ticket Memory Leak.
6569 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
6570 integrity of that ticket is first verified. In the event of a session
6571 ticket integrity check failing, OpenSSL will fail to free memory
6572 causing a memory leak. By sending a large number of invalid session
6573 tickets an attacker could exploit this issue in a Denial Of Service
6579 * Build option no-ssl3 is incomplete.
6581 When OpenSSL is configured with "no-ssl3" as a build option, servers
6582 could accept and complete a SSL 3.0 handshake, and clients could be
6583 configured to send them.
6586 *Akamai and the OpenSSL team*
6588 * Add support for TLS_FALLBACK_SCSV.
6589 Client applications doing fallback retries should call
6590 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
6593 *Adam Langley, Bodo Moeller*
6595 * Add additional DigestInfo checks.
6597 Re-encode DigestInto in DER and check against the original when
6598 verifying RSA signature: this will reject any improperly encoded
6599 DigestInfo structures.
6601 Note: this is a precautionary measure and no attacks are currently known.
6605 ### Changes between 1.0.1h and 1.0.1i [6 Aug 2014]
6607 * Fix SRP buffer overrun vulnerability. Invalid parameters passed to the
6608 SRP code can be overrun an internal buffer. Add sanity check that
6609 g, A, B < N to SRP code.
6611 Thanks to Sean Devlin and Watson Ladd of Cryptography Services, NCC
6612 Group for discovering this issue.
6617 * A flaw in the OpenSSL SSL/TLS server code causes the server to negotiate
6618 TLS 1.0 instead of higher protocol versions when the ClientHello message
6619 is badly fragmented. This allows a man-in-the-middle attacker to force a
6620 downgrade to TLS 1.0 even if both the server and the client support a
6621 higher protocol version, by modifying the client's TLS records.
6623 Thanks to David Benjamin and Adam Langley (Google) for discovering and
6624 researching this issue.
6629 * OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
6630 to a denial of service attack. A malicious server can crash the client
6631 with a null pointer dereference (read) by specifying an anonymous (EC)DH
6632 ciphersuite and sending carefully crafted handshake messages.
6634 Thanks to Felix Gröbert (Google) for discovering and researching this
6640 * By sending carefully crafted DTLS packets an attacker could cause openssl
6641 to leak memory. This can be exploited through a Denial of Service attack.
6642 Thanks to Adam Langley for discovering and researching this issue.
6647 * An attacker can force openssl to consume large amounts of memory whilst
6648 processing DTLS handshake messages. This can be exploited through a
6649 Denial of Service attack.
6650 Thanks to Adam Langley for discovering and researching this issue.
6655 * An attacker can force an error condition which causes openssl to crash
6656 whilst processing DTLS packets due to memory being freed twice. This
6657 can be exploited through a Denial of Service attack.
6658 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
6664 * If a multithreaded client connects to a malicious server using a resumed
6665 session and the server sends an ec point format extension it could write
6666 up to 255 bytes to freed memory.
6668 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
6674 * A malicious server can crash an OpenSSL client with a null pointer
6675 dereference (read) by specifying an SRP ciphersuite even though it was not
6676 properly negotiated with the client. This can be exploited through a
6677 Denial of Service attack.
6679 Thanks to Joonas Kuorilehto and Riku Hietamäki (Codenomicon) for
6680 discovering and researching this issue.
6685 * A flaw in OBJ_obj2txt may cause pretty printing functions such as
6686 X509_name_oneline, X509_name_print_ex et al. to leak some information
6687 from the stack. Applications may be affected if they echo pretty printing
6688 output to the attacker.
6690 Thanks to Ivan Fratric (Google) for discovering this issue.
6693 *Emilia Käsper, and Steve Henson*
6695 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
6696 for corner cases. (Certain input points at infinity could lead to
6697 bogus results, with non-infinity inputs mapped to infinity too.)
6701 ### Changes between 1.0.1g and 1.0.1h [5 Jun 2014]
6703 * Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
6704 handshake can force the use of weak keying material in OpenSSL
6705 SSL/TLS clients and servers.
6707 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
6708 researching this issue. ([CVE-2014-0224])
6710 *KIKUCHI Masashi, Steve Henson*
6712 * Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
6713 OpenSSL DTLS client the code can be made to recurse eventually crashing
6716 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
6719 *Imre Rad, Steve Henson*
6721 * Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
6722 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
6723 client or server. This is potentially exploitable to run arbitrary
6724 code on a vulnerable client or server.
6726 Thanks to Jüri Aedla for reporting this issue. ([CVE-2014-0195])
6728 *Jüri Aedla, Steve Henson*
6730 * Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
6731 are subject to a denial of service attack.
6733 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
6734 this issue. ([CVE-2014-3470])
6736 *Felix Gröbert, Ivan Fratric, Steve Henson*
6738 * Harmonize version and its documentation. -f flag is used to display
6741 *mancha <mancha1@zoho.com>*
6743 * Fix eckey_priv_encode so it immediately returns an error upon a failure
6744 in i2d_ECPrivateKey.
6746 *mancha <mancha1@zoho.com>*
6748 * Fix some double frees. These are not thought to be exploitable.
6750 *mancha <mancha1@zoho.com>*
6752 ### Changes between 1.0.1f and 1.0.1g [7 Apr 2014]
6754 * A missing bounds check in the handling of the TLS heartbeat extension
6755 can be used to reveal up to 64k of memory to a connected client or
6758 Thanks for Neel Mehta of Google Security for discovering this bug and to
6759 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
6760 preparing the fix ([CVE-2014-0160])
6762 *Adam Langley, Bodo Moeller*
6764 * Fix for the attack described in the paper "Recovering OpenSSL
6765 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
6766 by Yuval Yarom and Naomi Benger. Details can be obtained from:
6767 <http://eprint.iacr.org/2014/140>
6769 Thanks to Yuval Yarom and Naomi Benger for discovering this
6770 flaw and to Yuval Yarom for supplying a fix ([CVE-2014-0076])
6772 *Yuval Yarom and Naomi Benger*
6774 * TLS pad extension: draft-agl-tls-padding-03
6776 Workaround for the "TLS hang bug" (see FAQ and PR#2771): if the
6777 TLS client Hello record length value would otherwise be > 255 and
6778 less that 512 pad with a dummy extension containing zeroes so it
6779 is at least 512 bytes long.
6781 *Adam Langley, Steve Henson*
6783 ### Changes between 1.0.1e and 1.0.1f [6 Jan 2014]
6785 * Fix for TLS record tampering bug. A carefully crafted invalid
6786 handshake could crash OpenSSL with a NULL pointer exception.
6787 Thanks to Anton Johansson for reporting this issues.
6790 * Keep original DTLS digest and encryption contexts in retransmission
6791 structures so we can use the previous session parameters if they need
6792 to be resent. ([CVE-2013-6450])
6796 * Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
6797 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
6798 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
6799 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
6800 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
6801 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
6803 *Rob Stradling, Adam Langley*
6805 ### Changes between 1.0.1d and 1.0.1e [11 Feb 2013]
6807 * Correct fix for CVE-2013-0169. The original didn't work on AES-NI
6808 supporting platforms or when small records were transferred.
6810 *Andy Polyakov, Steve Henson*
6812 ### Changes between 1.0.1c and 1.0.1d [5 Feb 2013]
6814 * Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
6816 This addresses the flaw in CBC record processing discovered by
6817 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
6818 at: <http://www.isg.rhul.ac.uk/tls/>
6820 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
6821 Security Group at Royal Holloway, University of London
6822 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
6823 Emilia Käsper for the initial patch.
6826 *Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson*
6828 * Fix flaw in AESNI handling of TLS 1.2 and 1.1 records for CBC mode
6829 ciphersuites which can be exploited in a denial of service attack.
6830 Thanks go to and to Adam Langley <agl@chromium.org> for discovering
6831 and detecting this bug and to Wolfgang Ettlinger
6832 <wolfgang.ettlinger@gmail.com> for independently discovering this issue.
6837 * Return an error when checking OCSP signatures when key is NULL.
6838 This fixes a DoS attack. ([CVE-2013-0166])
6842 * Make openssl verify return errors.
6844 *Chris Palmer <palmer@google.com> and Ben Laurie*
6846 * Call OCSP Stapling callback after ciphersuite has been chosen, so
6847 the right response is stapled. Also change SSL_get_certificate()
6848 so it returns the certificate actually sent.
6849 See <http://rt.openssl.org/Ticket/Display.html?id=2836>.
6851 *Rob Stradling <rob.stradling@comodo.com>*
6853 * Fix possible deadlock when decoding public keys.
6857 * Don't use TLS 1.0 record version number in initial client hello
6862 ### Changes between 1.0.1b and 1.0.1c [10 May 2012]
6864 * Sanity check record length before skipping explicit IV in TLS
6865 1.2, 1.1 and DTLS to fix DoS attack.
6867 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
6868 fuzzing as a service testing platform.
6873 * Initialise tkeylen properly when encrypting CMS messages.
6874 Thanks to Solar Designer of Openwall for reporting this issue.
6878 * In FIPS mode don't try to use composite ciphers as they are not
6883 ### Changes between 1.0.1a and 1.0.1b [26 Apr 2012]
6885 * OpenSSL 1.0.0 sets SSL_OP_ALL to 0x80000FFFL and OpenSSL 1.0.1 and
6886 1.0.1a set SSL_OP_NO_TLSv1_1 to 0x00000400L which would unfortunately
6887 mean any application compiled against OpenSSL 1.0.0 headers setting
6888 SSL_OP_ALL would also set SSL_OP_NO_TLSv1_1, unintentionally disabling
6889 TLS 1.1 also. Fix this by changing the value of SSL_OP_NO_TLSv1_1 to
6890 0x10000000L Any application which was previously compiled against
6891 OpenSSL 1.0.1 or 1.0.1a headers and which cares about SSL_OP_NO_TLSv1_1
6892 will need to be recompiled as a result. Letting be results in
6893 inability to disable specifically TLS 1.1 and in client context,
6894 in unlike event, limit maximum offered version to TLS 1.0 [see below].
6898 * In order to ensure interoperability SSL_OP_NO_protocolX does not
6899 disable just protocol X, but all protocols above X *if* there are
6900 protocols *below* X still enabled. In more practical terms it means
6901 that if application wants to disable TLS1.0 in favor of TLS1.1 and
6902 above, it's not sufficient to pass `SSL_OP_NO_TLSv1`, one has to pass
6903 `SSL_OP_NO_TLSv1|SSL_OP_NO_SSLv3|SSL_OP_NO_SSLv2`. This applies to
6908 ### Changes between 1.0.1 and 1.0.1a [19 Apr 2012]
6910 * Check for potentially exploitable overflows in asn1_d2i_read_bio
6911 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
6912 in CRYPTO_realloc_clean.
6914 Thanks to Tavis Ormandy, Google Security Team, for discovering this
6915 issue and to Adam Langley <agl@chromium.org> for fixing it.
6918 *Adam Langley (Google), Tavis Ormandy, Google Security Team*
6920 * Don't allow TLS 1.2 SHA-256 ciphersuites in TLS 1.0, 1.1 connections.
6924 * Workarounds for some broken servers that "hang" if a client hello
6925 record length exceeds 255 bytes.
6927 1. Do not use record version number > TLS 1.0 in initial client
6928 hello: some (but not all) hanging servers will now work.
6929 2. If we set OPENSSL_MAX_TLS1_2_CIPHER_LENGTH this will truncate
6930 the number of ciphers sent in the client hello. This should be
6931 set to an even number, such as 50, for example by passing:
6932 -DOPENSSL_MAX_TLS1_2_CIPHER_LENGTH=50 to config or Configure.
6933 Most broken servers should now work.
6934 3. If all else fails setting OPENSSL_NO_TLS1_2_CLIENT will disable
6935 TLS 1.2 client support entirely.
6939 * Fix SEGV in Vector Permutation AES module observed in OpenSSH.
6943 ### Changes between 1.0.0h and 1.0.1 [14 Mar 2012]
6945 * Add compatibility with old MDC2 signatures which use an ASN1 OCTET
6946 STRING form instead of a DigestInfo.
6950 * The format used for MDC2 RSA signatures is inconsistent between EVP
6951 and the RSA_sign/RSA_verify functions. This was made more apparent when
6952 OpenSSL used RSA_sign/RSA_verify for some RSA signatures in particular
6953 those which went through EVP_PKEY_METHOD in 1.0.0 and later. Detect
6954 the correct format in RSA_verify so both forms transparently work.
6958 * Some servers which support TLS 1.0 can choke if we initially indicate
6959 support for TLS 1.2 and later renegotiate using TLS 1.0 in the RSA
6960 encrypted premaster secret. As a workaround use the maximum permitted
6961 client version in client hello, this should keep such servers happy
6962 and still work with previous versions of OpenSSL.
6966 * Add support for TLS/DTLS heartbeats.
6968 *Robin Seggelmann <seggelmann@fh-muenster.de>*
6970 * Add support for SCTP.
6972 *Robin Seggelmann <seggelmann@fh-muenster.de>*
6974 * Improved PRNG seeding for VOS.
6976 *Paul Green <Paul.Green@stratus.com>*
6978 * Extensive assembler packs updates, most notably:
6980 - x86[_64]: AES-NI, PCLMULQDQ, RDRAND support;
6981 - x86[_64]: SSSE3 support (SHA1, vector-permutation AES);
6982 - x86_64: bit-sliced AES implementation;
6983 - ARM: NEON support, contemporary platforms optimizations;
6984 - s390x: z196 support;
6985 - `*`: GHASH and GF(2^m) multiplication implementations;
6989 * Make TLS-SRP code conformant with RFC 5054 API cleanup
6990 (removal of unnecessary code)
6992 *Peter Sylvester <peter.sylvester@edelweb.fr>*
6994 * Add TLS key material exporter from RFC 5705.
6998 * Add DTLS-SRTP negotiation from RFC 5764.
7002 * Add Next Protocol Negotiation,
7003 <http://tools.ietf.org/html/draft-agl-tls-nextprotoneg-00>. Can be
7004 disabled with a no-npn flag to config or Configure. Code donated
7007 *Adam Langley <agl@google.com> and Ben Laurie*
7009 * Add optional 64-bit optimized implementations of elliptic curves NIST-P224,
7010 NIST-P256, NIST-P521, with constant-time single point multiplication on
7011 typical inputs. Compiler support for the nonstandard type `__uint128_t` is
7012 required to use this (present in gcc 4.4 and later, for 64-bit builds).
7013 Code made available under Apache License version 2.0.
7015 Specify "enable-ec_nistp_64_gcc_128" on the Configure (or config) command
7016 line to include this in your build of OpenSSL, and run "make depend" (or
7017 "make update"). This enables the following EC_METHODs:
7019 EC_GFp_nistp224_method()
7020 EC_GFp_nistp256_method()
7021 EC_GFp_nistp521_method()
7023 EC_GROUP_new_by_curve_name() will automatically use these (while
7024 EC_GROUP_new_curve_GFp() currently prefers the more flexible
7027 *Emilia Käsper, Adam Langley, Bodo Moeller (Google)*
7029 * Use type ossl_ssize_t instead of ssize_t which isn't available on
7030 all platforms. Move ssize_t definition from e_os.h to the public
7031 header file e_os2.h as it now appears in public header file cms.h
7035 * New -sigopt option to the ca, req and x509 utilities. Additional
7036 signature parameters can be passed using this option and in
7041 * Add RSA PSS signing function. This will generate and set the
7042 appropriate AlgorithmIdentifiers for PSS based on those in the
7043 corresponding EVP_MD_CTX structure. No application support yet.
7047 * Support for companion algorithm specific ASN1 signing routines.
7048 New function ASN1_item_sign_ctx() signs a pre-initialised
7049 EVP_MD_CTX structure and sets AlgorithmIdentifiers based on
7050 the appropriate parameters.
7054 * Add new algorithm specific ASN1 verification initialisation function
7055 to EVP_PKEY_ASN1_METHOD: this is not in EVP_PKEY_METHOD since the ASN1
7056 handling will be the same no matter what EVP_PKEY_METHOD is used.
7057 Add a PSS handler to support verification of PSS signatures: checked
7058 against a number of sample certificates.
7062 * Add signature printing for PSS. Add PSS OIDs.
7064 *Steve Henson, Martin Kaiser <lists@kaiser.cx>*
7066 * Add algorithm specific signature printing. An individual ASN1 method
7067 can now print out signatures instead of the standard hex dump.
7069 More complex signatures (e.g. PSS) can print out more meaningful
7070 information. Include DSA version that prints out the signature
7075 * Password based recipient info support for CMS library: implementing
7080 * Split password based encryption into PBES2 and PBKDF2 functions. This
7081 neatly separates the code into cipher and PBE sections and is required
7082 for some algorithms that split PBES2 into separate pieces (such as
7083 password based CMS).
7087 * Session-handling fixes:
7088 - Fix handling of connections that are resuming with a session ID,
7089 but also support Session Tickets.
7090 - Fix a bug that suppressed issuing of a new ticket if the client
7091 presented a ticket with an expired session.
7092 - Try to set the ticket lifetime hint to something reasonable.
7093 - Make tickets shorter by excluding irrelevant information.
7094 - On the client side, don't ignore renewed tickets.
7096 *Adam Langley, Bodo Moeller (Google)*
7098 * Fix PSK session representation.
7102 * Add RC4-MD5 and AESNI-SHA1 "stitched" implementations.
7104 This work was sponsored by Intel.
7108 * Add GCM support to TLS library. Some custom code is needed to split
7109 the IV between the fixed (from PRF) and explicit (from TLS record)
7110 portions. This adds all GCM ciphersuites supported by RFC5288 and
7111 RFC5289. Generalise some `AES*` cipherstrings to include GCM and
7112 add a special AESGCM string for GCM only.
7116 * Expand range of ctrls for AES GCM. Permit setting invocation
7117 field on decrypt and retrieval of invocation field only on encrypt.
7121 * Add HMAC ECC ciphersuites from RFC5289. Include SHA384 PRF support.
7122 As required by RFC5289 these ciphersuites cannot be used if for
7123 versions of TLS earlier than 1.2.
7127 * For FIPS capable OpenSSL interpret a NULL default public key method
7128 as unset and return the appropriate default but do *not* set the default.
7129 This means we can return the appropriate method in applications that
7130 switch between FIPS and non-FIPS modes.
7134 * Redirect HMAC and CMAC operations to FIPS module in FIPS mode. If an
7135 ENGINE is used then we cannot handle that in the FIPS module so we
7136 keep original code iff non-FIPS operations are allowed.
7140 * Add -attime option to openssl utilities.
7142 *Peter Eckersley <pde@eff.org>, Ben Laurie and Steve Henson*
7144 * Redirect DSA and DH operations to FIPS module in FIPS mode.
7148 * Redirect ECDSA and ECDH operations to FIPS module in FIPS mode. Also use
7149 FIPS EC methods unconditionally for now.
7153 * New build option no-ec2m to disable characteristic 2 code.
7157 * Backport libcrypto audit of return value checking from 1.1.0-dev; not
7158 all cases can be covered as some introduce binary incompatibilities.
7162 * Redirect RSA operations to FIPS module including keygen,
7163 encrypt, decrypt, sign and verify. Block use of non FIPS RSA methods.
7167 * Add similar low-level API blocking to ciphers.
7171 * low-level digest APIs are not approved in FIPS mode: any attempt
7172 to use these will cause a fatal error. Applications that *really* want
7173 to use them can use the `private_*` version instead.
7177 * Redirect cipher operations to FIPS module for FIPS builds.
7181 * Redirect digest operations to FIPS module for FIPS builds.
7185 * Update build system to add "fips" flag which will link in fipscanister.o
7186 for static and shared library builds embedding a signature if needed.
7190 * Output TLS supported curves in preference order instead of numerical
7191 order. This is currently hardcoded for the highest order curves first.
7192 This should be configurable so applications can judge speed vs strength.
7196 * Add TLS v1.2 server support for client authentication.
7200 * Add support for FIPS mode in ssl library: disable SSLv3, non-FIPS ciphers
7205 * Functions FIPS_mode_set() and FIPS_mode() which call the underlying
7206 FIPS modules versions.
7210 * Add TLS v1.2 client side support for client authentication. Keep cache
7211 of handshake records longer as we don't know the hash algorithm to use
7212 until after the certificate request message is received.
7216 * Initial TLS v1.2 client support. Add a default signature algorithms
7217 extension including all the algorithms we support. Parse new signature
7218 format in client key exchange. Relax some ECC signing restrictions for
7219 TLS v1.2 as indicated in RFC5246.
7223 * Add server support for TLS v1.2 signature algorithms extension. Switch
7224 to new signature format when needed using client digest preference.
7225 All server ciphersuites should now work correctly in TLS v1.2. No client
7226 support yet and no support for client certificates.
7230 * Initial TLS v1.2 support. Add new SHA256 digest to ssl code, switch
7231 to SHA256 for PRF when using TLS v1.2 and later. Add new SHA256 based
7232 ciphersuites. At present only RSA key exchange ciphersuites work with
7233 TLS v1.2. Add new option for TLS v1.2 replacing the old and obsolete
7234 SSL_OP_PKCS1_CHECK flags with SSL_OP_NO_TLSv1_2. New TLSv1.2 methods
7235 and version checking.
7239 * New option OPENSSL_NO_SSL_INTERN. If an application can be compiled
7240 with this defined it will not be affected by any changes to ssl internal
7241 structures. Add several utility functions to allow openssl application
7242 to work with OPENSSL_NO_SSL_INTERN defined.
7246 * A long standing patch to add support for SRP from EdelWeb (Peter
7247 Sylvester and Christophe Renou) was integrated.
7248 *Christophe Renou <christophe.renou@edelweb.fr>, Peter Sylvester
7249 <peter.sylvester@edelweb.fr>, Tom Wu <tjw@cs.stanford.edu>, and
7252 * Add functions to copy EVP_PKEY_METHOD and retrieve flags and id.
7256 * Permit abbreviated handshakes when renegotiating using the function
7257 SSL_renegotiate_abbreviated().
7259 *Robin Seggelmann <seggelmann@fh-muenster.de>*
7261 * Add call to ENGINE_register_all_complete() to
7262 ENGINE_load_builtin_engines(), so some implementations get used
7263 automatically instead of needing explicit application support.
7267 * Add support for TLS key exporter as described in RFC5705.
7269 *Robin Seggelmann <seggelmann@fh-muenster.de>, Steve Henson*
7271 * Initial TLSv1.1 support. Since TLSv1.1 is very similar to TLS v1.0 only
7272 a few changes are required:
7274 Add SSL_OP_NO_TLSv1_1 flag.
7275 Add TLSv1_1 methods.
7276 Update version checking logic to handle version 1.1.
7277 Add explicit IV handling (ported from DTLS code).
7278 Add command line options to s_client/s_server.
7285 ### Changes between 1.0.0s and 1.0.0t [3 Dec 2015]
7287 * X509_ATTRIBUTE memory leak
7289 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
7290 memory. This structure is used by the PKCS#7 and CMS routines so any
7291 application which reads PKCS#7 or CMS data from untrusted sources is
7292 affected. SSL/TLS is not affected.
7294 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
7300 * Race condition handling PSK identify hint
7302 If PSK identity hints are received by a multi-threaded client then
7303 the values are wrongly updated in the parent SSL_CTX structure. This can
7304 result in a race condition potentially leading to a double free of the
7310 ### Changes between 1.0.0r and 1.0.0s [11 Jun 2015]
7312 * Malformed ECParameters causes infinite loop
7314 When processing an ECParameters structure OpenSSL enters an infinite loop
7315 if the curve specified is over a specially malformed binary polynomial
7318 This can be used to perform denial of service against any
7319 system which processes public keys, certificate requests or
7320 certificates. This includes TLS clients and TLS servers with
7321 client authentication enabled.
7323 This issue was reported to OpenSSL by Joseph Barr-Pixton.
7328 * Exploitable out-of-bounds read in X509_cmp_time
7330 X509_cmp_time does not properly check the length of the ASN1_TIME
7331 string and can read a few bytes out of bounds. In addition,
7332 X509_cmp_time accepts an arbitrary number of fractional seconds in the
7335 An attacker can use this to craft malformed certificates and CRLs of
7336 various sizes and potentially cause a segmentation fault, resulting in
7337 a DoS on applications that verify certificates or CRLs. TLS clients
7338 that verify CRLs are affected. TLS clients and servers with client
7339 authentication enabled may be affected if they use custom verification
7342 This issue was reported to OpenSSL by Robert Swiecki (Google), and
7343 independently by Hanno Böck.
7348 * PKCS7 crash with missing EnvelopedContent
7350 The PKCS#7 parsing code does not handle missing inner EncryptedContent
7351 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
7352 with missing content and trigger a NULL pointer dereference on parsing.
7354 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
7355 structures from untrusted sources are affected. OpenSSL clients and
7356 servers are not affected.
7358 This issue was reported to OpenSSL by Michal Zalewski (Google).
7363 * CMS verify infinite loop with unknown hash function
7365 When verifying a signedData message the CMS code can enter an infinite loop
7366 if presented with an unknown hash function OID. This can be used to perform
7367 denial of service against any system which verifies signedData messages using
7369 This issue was reported to OpenSSL by Johannes Bauer.
7374 * Race condition handling NewSessionTicket
7376 If a NewSessionTicket is received by a multi-threaded client when attempting to
7377 reuse a previous ticket then a race condition can occur potentially leading to
7378 a double free of the ticket data.
7383 ### Changes between 1.0.0q and 1.0.0r [19 Mar 2015]
7385 * Segmentation fault in ASN1_TYPE_cmp fix
7387 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
7388 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
7389 certificate signature algorithm consistency this can be used to crash any
7390 certificate verification operation and exploited in a DoS attack. Any
7391 application which performs certificate verification is vulnerable including
7392 OpenSSL clients and servers which enable client authentication.
7397 * ASN.1 structure reuse memory corruption fix
7399 Reusing a structure in ASN.1 parsing may allow an attacker to cause
7400 memory corruption via an invalid write. Such reuse is and has been
7401 strongly discouraged and is believed to be rare.
7403 Applications that parse structures containing CHOICE or ANY DEFINED BY
7404 components may be affected. Certificate parsing (d2i_X509 and related
7405 functions) are however not affected. OpenSSL clients and servers are
7411 * PKCS7 NULL pointer dereferences fix
7413 The PKCS#7 parsing code does not handle missing outer ContentInfo
7414 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
7415 missing content and trigger a NULL pointer dereference on parsing.
7417 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
7418 otherwise parse PKCS#7 structures from untrusted sources are
7419 affected. OpenSSL clients and servers are not affected.
7421 This issue was reported to OpenSSL by Michal Zalewski (Google).
7426 * DoS via reachable assert in SSLv2 servers fix
7428 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
7429 servers that both support SSLv2 and enable export cipher suites by sending
7430 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
7432 This issue was discovered by Sean Burford (Google) and Emilia Käsper
7433 (OpenSSL development team).
7438 * Use After Free following d2i_ECPrivatekey error fix
7440 A malformed EC private key file consumed via the d2i_ECPrivateKey function
7441 could cause a use after free condition. This, in turn, could cause a double
7442 free in several private key parsing functions (such as d2i_PrivateKey
7443 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
7444 for applications that receive EC private keys from untrusted
7445 sources. This scenario is considered rare.
7447 This issue was discovered by the BoringSSL project and fixed in their
7453 * X509_to_X509_REQ NULL pointer deref fix
7455 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
7456 the certificate key is invalid. This function is rarely used in practice.
7458 This issue was discovered by Brian Carpenter.
7463 * Removed the export ciphers from the DEFAULT ciphers
7467 ### Changes between 1.0.0p and 1.0.0q [15 Jan 2015]
7469 * Build fixes for the Windows and OpenVMS platforms
7471 *Matt Caswell and Richard Levitte*
7473 ### Changes between 1.0.0o and 1.0.0p [8 Jan 2015]
7475 * Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
7476 message can cause a segmentation fault in OpenSSL due to a NULL pointer
7477 dereference. This could lead to a Denial Of Service attack. Thanks to
7478 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
7483 * Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
7484 dtls1_buffer_record function under certain conditions. In particular this
7485 could occur if an attacker sent repeated DTLS records with the same
7486 sequence number but for the next epoch. The memory leak could be exploited
7487 by an attacker in a Denial of Service attack through memory exhaustion.
7488 Thanks to Chris Mueller for reporting this issue.
7493 * Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
7494 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
7495 method would be set to NULL which could later result in a NULL pointer
7496 dereference. Thanks to Frank Schmirler for reporting this issue.
7501 * Abort handshake if server key exchange message is omitted for ephemeral
7504 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
7505 reporting this issue.
7510 * Remove non-export ephemeral RSA code on client and server. This code
7511 violated the TLS standard by allowing the use of temporary RSA keys in
7512 non-export ciphersuites and could be used by a server to effectively
7513 downgrade the RSA key length used to a value smaller than the server
7514 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
7515 INRIA or reporting this issue.
7520 * Fixed issue where DH client certificates are accepted without verification.
7521 An OpenSSL server will accept a DH certificate for client authentication
7522 without the certificate verify message. This effectively allows a client to
7523 authenticate without the use of a private key. This only affects servers
7524 which trust a client certificate authority which issues certificates
7525 containing DH keys: these are extremely rare and hardly ever encountered.
7526 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
7532 * Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
7533 results on some platforms, including x86_64. This bug occurs at random
7534 with a very low probability, and is not known to be exploitable in any
7535 way, though its exact impact is difficult to determine. Thanks to Pieter
7536 Wuille (Blockstream) who reported this issue and also suggested an initial
7537 fix. Further analysis was conducted by the OpenSSL development team and
7538 Adam Langley of Google. The final fix was developed by Andy Polyakov of
7539 the OpenSSL core team.
7544 * Fix various certificate fingerprint issues.
7546 By using non-DER or invalid encodings outside the signed portion of a
7547 certificate the fingerprint can be changed without breaking the signature.
7548 Although no details of the signed portion of the certificate can be changed
7549 this can cause problems with some applications: e.g. those using the
7550 certificate fingerprint for blacklists.
7552 1. Reject signatures with non zero unused bits.
7554 If the BIT STRING containing the signature has non zero unused bits reject
7555 the signature. All current signature algorithms require zero unused bits.
7557 2. Check certificate algorithm consistency.
7559 Check the AlgorithmIdentifier inside TBS matches the one in the
7560 certificate signature. NB: this will result in signature failure
7561 errors for some broken certificates.
7563 Thanks to Konrad Kraszewski from Google for reporting this issue.
7565 3. Check DSA/ECDSA signatures use DER.
7567 Reencode DSA/ECDSA signatures and compare with the original received
7568 signature. Return an error if there is a mismatch.
7570 This will reject various cases including garbage after signature
7571 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
7572 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
7573 (negative or with leading zeroes).
7575 Further analysis was conducted and fixes were developed by Stephen Henson
7576 of the OpenSSL core team.
7582 ### Changes between 1.0.0n and 1.0.0o [15 Oct 2014]
7584 * Session Ticket Memory Leak.
7586 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
7587 integrity of that ticket is first verified. In the event of a session
7588 ticket integrity check failing, OpenSSL will fail to free memory
7589 causing a memory leak. By sending a large number of invalid session
7590 tickets an attacker could exploit this issue in a Denial Of Service
7596 * Build option no-ssl3 is incomplete.
7598 When OpenSSL is configured with "no-ssl3" as a build option, servers
7599 could accept and complete a SSL 3.0 handshake, and clients could be
7600 configured to send them.
7603 *Akamai and the OpenSSL team*
7605 * Add support for TLS_FALLBACK_SCSV.
7606 Client applications doing fallback retries should call
7607 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
7610 *Adam Langley, Bodo Moeller*
7612 * Add additional DigestInfo checks.
7614 Reencode DigestInto in DER and check against the original when
7615 verifying RSA signature: this will reject any improperly encoded
7616 DigestInfo structures.
7618 Note: this is a precautionary measure and no attacks are currently known.
7622 ### Changes between 1.0.0m and 1.0.0n [6 Aug 2014]
7624 * OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
7625 to a denial of service attack. A malicious server can crash the client
7626 with a null pointer dereference (read) by specifying an anonymous (EC)DH
7627 ciphersuite and sending carefully crafted handshake messages.
7629 Thanks to Felix Gröbert (Google) for discovering and researching this
7635 * By sending carefully crafted DTLS packets an attacker could cause openssl
7636 to leak memory. This can be exploited through a Denial of Service attack.
7637 Thanks to Adam Langley for discovering and researching this issue.
7642 * An attacker can force openssl to consume large amounts of memory whilst
7643 processing DTLS handshake messages. This can be exploited through a
7644 Denial of Service attack.
7645 Thanks to Adam Langley for discovering and researching this issue.
7650 * An attacker can force an error condition which causes openssl to crash
7651 whilst processing DTLS packets due to memory being freed twice. This
7652 can be exploited through a Denial of Service attack.
7653 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
7659 * If a multithreaded client connects to a malicious server using a resumed
7660 session and the server sends an ec point format extension it could write
7661 up to 255 bytes to freed memory.
7663 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
7669 * A flaw in OBJ_obj2txt may cause pretty printing functions such as
7670 X509_name_oneline, X509_name_print_ex et al. to leak some information
7671 from the stack. Applications may be affected if they echo pretty printing
7672 output to the attacker.
7674 Thanks to Ivan Fratric (Google) for discovering this issue.
7677 *Emilia Käsper, and Steve Henson*
7679 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
7680 for corner cases. (Certain input points at infinity could lead to
7681 bogus results, with non-infinity inputs mapped to infinity too.)
7685 ### Changes between 1.0.0l and 1.0.0m [5 Jun 2014]
7687 * Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
7688 handshake can force the use of weak keying material in OpenSSL
7689 SSL/TLS clients and servers.
7691 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
7692 researching this issue. ([CVE-2014-0224])
7694 *KIKUCHI Masashi, Steve Henson*
7696 * Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
7697 OpenSSL DTLS client the code can be made to recurse eventually crashing
7700 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
7703 *Imre Rad, Steve Henson*
7705 * Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
7706 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
7707 client or server. This is potentially exploitable to run arbitrary
7708 code on a vulnerable client or server.
7710 Thanks to Jüri Aedla for reporting this issue. ([CVE-2014-0195])
7712 *Jüri Aedla, Steve Henson*
7714 * Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
7715 are subject to a denial of service attack.
7717 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
7718 this issue. ([CVE-2014-3470])
7720 *Felix Gröbert, Ivan Fratric, Steve Henson*
7722 * Harmonize version and its documentation. -f flag is used to display
7725 *mancha <mancha1@zoho.com>*
7727 * Fix eckey_priv_encode so it immediately returns an error upon a failure
7728 in i2d_ECPrivateKey.
7730 *mancha <mancha1@zoho.com>*
7732 * Fix some double frees. These are not thought to be exploitable.
7734 *mancha <mancha1@zoho.com>*
7736 * Fix for the attack described in the paper "Recovering OpenSSL
7737 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
7738 by Yuval Yarom and Naomi Benger. Details can be obtained from:
7739 <http://eprint.iacr.org/2014/140>
7741 Thanks to Yuval Yarom and Naomi Benger for discovering this
7742 flaw and to Yuval Yarom for supplying a fix ([CVE-2014-0076])
7744 *Yuval Yarom and Naomi Benger*
7746 ### Changes between 1.0.0k and 1.0.0l [6 Jan 2014]
7748 * Keep original DTLS digest and encryption contexts in retransmission
7749 structures so we can use the previous session parameters if they need
7750 to be resent. ([CVE-2013-6450])
7754 * Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
7755 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
7756 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
7757 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
7758 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
7759 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
7761 *Rob Stradling, Adam Langley*
7763 ### Changes between 1.0.0j and 1.0.0k [5 Feb 2013]
7765 * Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
7767 This addresses the flaw in CBC record processing discovered by
7768 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
7769 at: <http://www.isg.rhul.ac.uk/tls/>
7771 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
7772 Security Group at Royal Holloway, University of London
7773 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
7774 Emilia Käsper for the initial patch.
7777 *Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson*
7779 * Return an error when checking OCSP signatures when key is NULL.
7780 This fixes a DoS attack. ([CVE-2013-0166])
7784 * Call OCSP Stapling callback after ciphersuite has been chosen, so
7785 the right response is stapled. Also change SSL_get_certificate()
7786 so it returns the certificate actually sent.
7787 See <http://rt.openssl.org/Ticket/Display.html?id=2836>.
7788 (This is a backport)
7790 *Rob Stradling <rob.stradling@comodo.com>*
7792 * Fix possible deadlock when decoding public keys.
7796 ### Changes between 1.0.0i and 1.0.0j [10 May 2012]
7798 [NB: OpenSSL 1.0.0i and later 1.0.0 patch levels were released after
7801 * Sanity check record length before skipping explicit IV in DTLS
7804 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
7805 fuzzing as a service testing platform.
7810 * Initialise tkeylen properly when encrypting CMS messages.
7811 Thanks to Solar Designer of Openwall for reporting this issue.
7815 ### Changes between 1.0.0h and 1.0.0i [19 Apr 2012]
7817 * Check for potentially exploitable overflows in asn1_d2i_read_bio
7818 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
7819 in CRYPTO_realloc_clean.
7821 Thanks to Tavis Ormandy, Google Security Team, for discovering this
7822 issue and to Adam Langley <agl@chromium.org> for fixing it.
7825 *Adam Langley (Google), Tavis Ormandy, Google Security Team*
7827 ### Changes between 1.0.0g and 1.0.0h [12 Mar 2012]
7829 * Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness
7830 in CMS and PKCS7 code. When RSA decryption fails use a random key for
7831 content decryption and always return the same error. Note: this attack
7832 needs on average 2^20 messages so it only affects automated senders. The
7833 old behaviour can be re-enabled in the CMS code by setting the
7834 CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where
7835 an MMA defence is not necessary.
7836 Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering
7837 this issue. ([CVE-2012-0884])
7841 * Fix CVE-2011-4619: make sure we really are receiving a
7842 client hello before rejecting multiple SGC restarts. Thanks to
7843 Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug.
7847 ### Changes between 1.0.0f and 1.0.0g [18 Jan 2012]
7849 * Fix for DTLS DoS issue introduced by fix for CVE-2011-4109.
7850 Thanks to Antonio Martin, Enterprise Secure Access Research and
7851 Development, Cisco Systems, Inc. for discovering this bug and
7852 preparing a fix. ([CVE-2012-0050])
7856 ### Changes between 1.0.0e and 1.0.0f [4 Jan 2012]
7858 * Nadhem Alfardan and Kenny Paterson have discovered an extension
7859 of the Vaudenay padding oracle attack on CBC mode encryption
7860 which enables an efficient plaintext recovery attack against
7861 the OpenSSL implementation of DTLS. Their attack exploits timing
7862 differences arising during decryption processing. A research
7863 paper describing this attack can be found at:
7864 <http://www.isg.rhul.ac.uk/~kp/dtls.pdf>
7865 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
7866 Security Group at Royal Holloway, University of London
7867 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
7868 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
7869 for preparing the fix. ([CVE-2011-4108])
7871 *Robin Seggelmann, Michael Tuexen*
7873 * Clear bytes used for block padding of SSL 3.0 records.
7876 *Adam Langley (Google)*
7878 * Only allow one SGC handshake restart for SSL/TLS. Thanks to George
7879 Kadianakis <desnacked@gmail.com> for discovering this issue and
7880 Adam Langley for preparing the fix. ([CVE-2011-4619])
7882 *Adam Langley (Google)*
7884 * Check parameters are not NULL in GOST ENGINE. ([CVE-2012-0027])
7886 *Andrey Kulikov <amdeich@gmail.com>*
7888 * Prevent malformed RFC3779 data triggering an assertion failure.
7889 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
7890 and Rob Austein <sra@hactrn.net> for fixing it. ([CVE-2011-4577])
7892 *Rob Austein <sra@hactrn.net>*
7894 * Improved PRNG seeding for VOS.
7896 *Paul Green <Paul.Green@stratus.com>*
7898 * Fix ssl_ciph.c set-up race.
7900 *Adam Langley (Google)*
7902 * Fix spurious failures in ecdsatest.c.
7904 *Emilia Käsper (Google)*
7906 * Fix the BIO_f_buffer() implementation (which was mixing different
7907 interpretations of the `..._len` fields).
7909 *Adam Langley (Google)*
7911 * Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
7912 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
7913 threads won't reuse the same blinding coefficients.
7915 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
7916 lock to call BN_BLINDING_invert_ex, and avoids one use of
7917 BN_BLINDING_update for each BN_BLINDING structure (previously,
7918 the last update always remained unused).
7920 *Emilia Käsper (Google)*
7922 * In ssl3_clear, preserve s3->init_extra along with s3->rbuf.
7924 *Bob Buckholz (Google)*
7926 ### Changes between 1.0.0d and 1.0.0e [6 Sep 2011]
7928 * Fix bug where CRLs with nextUpdate in the past are sometimes accepted
7929 by initialising X509_STORE_CTX properly. ([CVE-2011-3207])
7931 *Kaspar Brand <ossl@velox.ch>*
7933 * Fix SSL memory handling for (EC)DH ciphersuites, in particular
7934 for multi-threaded use of ECDH. ([CVE-2011-3210])
7936 *Adam Langley (Google)*
7938 * Fix x509_name_ex_d2i memory leak on bad inputs.
7942 * Remove hard coded ecdsaWithSHA1 signature tests in ssl code and check
7943 signature public key algorithm by using OID xref utilities instead.
7944 Before this you could only use some ECC ciphersuites with SHA1 only.
7948 * Add protection against ECDSA timing attacks as mentioned in the paper
7949 by Billy Bob Brumley and Nicola Tuveri, see:
7950 <http://eprint.iacr.org/2011/232.pdf>
7952 *Billy Bob Brumley and Nicola Tuveri*
7954 ### Changes between 1.0.0c and 1.0.0d [8 Feb 2011]
7956 * Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
7958 *Neel Mehta, Adam Langley, Bodo Moeller (Google)*
7960 * Fix bug in string printing code: if *any* escaping is enabled we must
7961 escape the escape character (backslash) or the resulting string is
7966 ### Changes between 1.0.0b and 1.0.0c [2 Dec 2010]
7968 * Disable code workaround for ancient and obsolete Netscape browsers
7969 and servers: an attacker can use it in a ciphersuite downgrade attack.
7970 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
7974 * Fixed J-PAKE implementation error, originally discovered by
7975 Sebastien Martini, further info and confirmation from Stefan
7976 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
7980 ### Changes between 1.0.0a and 1.0.0b [16 Nov 2010]
7982 * Fix extension code to avoid race conditions which can result in a buffer
7983 overrun vulnerability: resumed sessions must not be modified as they can
7984 be shared by multiple threads. CVE-2010-3864
7988 * Fix WIN32 build system to correctly link an ENGINE directory into
7993 ### Changes between 1.0.0 and 1.0.0a [01 Jun 2010]
7995 * Check return value of int_rsa_verify in pkey_rsa_verifyrecover
7998 *Steve Henson, Peter-Michael Hager <hager@dortmund.net>*
8000 ### Changes between 0.9.8n and 1.0.0 [29 Mar 2010]
8002 * Add "missing" function EVP_CIPHER_CTX_copy(). This copies a cipher
8003 context. The operation can be customised via the ctrl mechanism in
8004 case ENGINEs want to include additional functionality.
8008 * Tolerate yet another broken PKCS#8 key format: private key value negative.
8012 * Add new -subject_hash_old and -issuer_hash_old options to x509 utility to
8013 output hashes compatible with older versions of OpenSSL.
8015 *Willy Weisz <weisz@vcpc.univie.ac.at>*
8017 * Fix compression algorithm handling: if resuming a session use the
8018 compression algorithm of the resumed session instead of determining
8019 it from client hello again. Don't allow server to change algorithm.
8023 * Add load_crls() function to commands tidying load_certs() too. Add option
8024 to verify utility to allow additional CRLs to be included.
8028 * Update OCSP request code to permit adding custom headers to the request:
8029 some responders need this.
8033 * The function EVP_PKEY_sign() returns <=0 on error: check return code
8036 *Julia Lawall <julia@diku.dk>*
8038 * Update verify callback code in `apps/s_cb.c` and `apps/verify.c`, it
8039 needlessly dereferenced structures, used obsolete functions and
8040 didn't handle all updated verify codes correctly.
8044 * Disable MD2 in the default configuration.
8048 * In BIO_pop() and BIO_push() use the ctrl argument (which was NULL) to
8049 indicate the initial BIO being pushed or popped. This makes it possible
8050 to determine whether the BIO is the one explicitly called or as a result
8051 of the ctrl being passed down the chain. Fix BIO_pop() and SSL BIOs so
8052 it handles reference counts correctly and doesn't zero out the I/O bio
8053 when it is not being explicitly popped. WARNING: applications which
8054 included workarounds for the old buggy behaviour will need to be modified
8055 or they could free up already freed BIOs.
8059 * Extend the uni2asc/asc2uni => OPENSSL_uni2asc/OPENSSL_asc2uni
8060 renaming to all platforms (within the 0.9.8 branch, this was
8061 done conditionally on Netware platforms to avoid a name clash).
8063 *Guenter <lists@gknw.net>*
8065 * Add ECDHE and PSK support to DTLS.
8067 *Michael Tuexen <tuexen@fh-muenster.de>*
8069 * Add CHECKED_STACK_OF macro to safestack.h, otherwise safestack can't
8074 * Add "missing" function EVP_MD_flags() (without this the only way to
8075 retrieve a digest flags is by accessing the structure directly. Update
8076 `EVP_MD_do_all*()` and `EVP_CIPHER_do_all*()` to include the name a digest
8077 or cipher is registered as in the "from" argument. Print out all
8078 registered digests in the dgst usage message instead of manually
8079 attempting to work them out.
8083 * If no SSLv2 ciphers are used don't use an SSLv2 compatible client hello:
8084 this allows the use of compression and extensions. Change default cipher
8085 string to remove SSLv2 ciphersuites. This effectively avoids ancient SSLv2
8086 by default unless an application cipher string requests it.
8090 * Alter match criteria in PKCS12_parse(). It used to try to use local
8091 key ids to find matching certificates and keys but some PKCS#12 files
8092 don't follow the (somewhat unwritten) rules and this strategy fails.
8093 Now just gather all certificates together and the first private key
8094 then look for the first certificate that matches the key.
8098 * Support use of registered digest and cipher names for dgst and cipher
8099 commands instead of having to add each one as a special case. So now
8106 openssl dgst -sha256 foo
8108 and this works for ENGINE based algorithms too.
8112 * Update Gost ENGINE to support parameter files.
8114 *Victor B. Wagner <vitus@cryptocom.ru>*
8116 * Support GeneralizedTime in ca utility.
8118 *Oliver Martin <oliver@volatilevoid.net>, Steve Henson*
8120 * Enhance the hash format used for certificate directory links. The new
8121 form uses the canonical encoding (meaning equivalent names will work
8122 even if they aren't identical) and uses SHA1 instead of MD5. This form
8123 is incompatible with the older format and as a result c_rehash should
8124 be used to rebuild symbolic links.
8128 * Make PKCS#8 the default write format for private keys, replacing the
8129 traditional format. This form is standardised, more secure and doesn't
8130 include an implicit MD5 dependency.
8134 * Add a $gcc_devteam_warn option to Configure. The idea is that any code
8135 committed to OpenSSL should pass this lot as a minimum.
8139 * Add session ticket override functionality for use by EAP-FAST.
8141 *Jouni Malinen <j@w1.fi>*
8143 * Modify HMAC functions to return a value. Since these can be implemented
8144 in an ENGINE errors can occur.
8148 * Type-checked OBJ_bsearch_ex.
8152 * Type-checked OBJ_bsearch. Also some constification necessitated
8153 by type-checking. Still to come: TXT_DB, bsearch(?),
8154 OBJ_bsearch_ex, qsort, CRYPTO_EX_DATA, ASN1_VALUE, ASN1_STRING,
8159 * New function OPENSSL_gmtime_adj() to add a specific number of days and
8160 seconds to a tm structure directly, instead of going through OS
8161 specific date routines. This avoids any issues with OS routines such
8162 as the year 2038 bug. New `*_adj()` functions for ASN1 time structures
8163 and X509_time_adj_ex() to cover the extended range. The existing
8164 X509_time_adj() is still usable and will no longer have any date issues.
8168 * Delta CRL support. New use deltas option which will attempt to locate
8169 and search any appropriate delta CRLs available.
8171 This work was sponsored by Google.
8175 * Support for CRLs partitioned by reason code. Reorganise CRL processing
8176 code and add additional score elements. Validate alternate CRL paths
8177 as part of the CRL checking and indicate a new error "CRL path validation
8178 error" in this case. Applications wanting additional details can use
8179 the verify callback and check the new "parent" field. If this is not
8180 NULL CRL path validation is taking place. Existing applications won't
8181 see this because it requires extended CRL support which is off by
8184 This work was sponsored by Google.
8188 * Support for freshest CRL extension.
8190 This work was sponsored by Google.
8194 * Initial indirect CRL support. Currently only supported in the CRLs
8195 passed directly and not via lookup. Process certificate issuer
8196 CRL entry extension and lookup CRL entries by bother issuer name
8197 and serial number. Check and process CRL issuer entry in IDP extension.
8199 This work was sponsored by Google.
8203 * Add support for distinct certificate and CRL paths. The CRL issuer
8204 certificate is validated separately in this case. Only enabled if
8205 an extended CRL support flag is set: this flag will enable additional
8206 CRL functionality in future.
8208 This work was sponsored by Google.
8212 * Add support for policy mappings extension.
8214 This work was sponsored by Google.
8218 * Fixes to pathlength constraint, self issued certificate handling,
8219 policy processing to align with RFC3280 and PKITS tests.
8221 This work was sponsored by Google.
8225 * Support for name constraints certificate extension. DN, email, DNS
8226 and URI types are currently supported.
8228 This work was sponsored by Google.
8232 * To cater for systems that provide a pointer-based thread ID rather
8233 than numeric, deprecate the current numeric thread ID mechanism and
8234 replace it with a structure and associated callback type. This
8235 mechanism allows a numeric "hash" to be extracted from a thread ID in
8236 either case, and on platforms where pointers are larger than 'long',
8237 mixing is done to help ensure the numeric 'hash' is usable even if it
8238 can't be guaranteed unique. The default mechanism is to use "&errno"
8239 as a pointer-based thread ID to distinguish between threads.
8241 Applications that want to provide their own thread IDs should now use
8242 CRYPTO_THREADID_set_callback() to register a callback that will call
8243 either CRYPTO_THREADID_set_numeric() or CRYPTO_THREADID_set_pointer().
8245 Note that ERR_remove_state() is now deprecated, because it is tied
8246 to the assumption that thread IDs are numeric. ERR_remove_state(0)
8247 to free the current thread's error state should be replaced by
8248 ERR_remove_thread_state(NULL).
8250 (This new approach replaces the functions CRYPTO_set_idptr_callback(),
8251 CRYPTO_get_idptr_callback(), and CRYPTO_thread_idptr() that existed in
8252 OpenSSL 0.9.9-dev between June 2006 and August 2008. Also, if an
8253 application was previously providing a numeric thread callback that
8254 was inappropriate for distinguishing threads, then uniqueness might
8255 have been obtained with &errno that happened immediately in the
8256 intermediate development versions of OpenSSL; this is no longer the
8257 case, the numeric thread callback will now override the automatic use
8260 *Geoff Thorpe, with help from Bodo Moeller*
8262 * Initial support for different CRL issuing certificates. This covers a
8263 simple case where the self issued certificates in the chain exist and
8264 the real CRL issuer is higher in the existing chain.
8266 This work was sponsored by Google.
8270 * Removed effectively defunct crypto/store from the build.
8274 * Revamp of STACK to provide stronger type-checking. Still to come:
8275 TXT_DB, bsearch(?), OBJ_bsearch, qsort, CRYPTO_EX_DATA, ASN1_VALUE,
8276 ASN1_STRING, CONF_VALUE.
8280 * Add a new SSL_MODE_RELEASE_BUFFERS mode flag to release unused buffer
8281 RAM on SSL connections. This option can save about 34k per idle SSL.
8285 * Revamp of LHASH to provide stronger type-checking. Still to come:
8286 STACK, TXT_DB, bsearch, qsort.
8290 * Initial support for Cryptographic Message Syntax (aka CMS) based
8291 on RFC3850, RFC3851 and RFC3852. New cms directory and cms utility,
8292 support for data, signedData, compressedData, digestedData and
8293 encryptedData, envelopedData types included. Scripts to check against
8294 RFC4134 examples draft and interop and consistency checks of many
8295 content types and variants.
8299 * Add options to enc utility to support use of zlib compression BIO.
8303 * Extend mk1mf to support importing of options and assembly language
8304 files from Configure script, currently only included in VC-WIN32.
8305 The assembly language rules can now optionally generate the source
8306 files from the associated perl scripts.
8310 * Implement remaining functionality needed to support GOST ciphersuites.
8311 Interop testing has been performed using CryptoPro implementations.
8313 *Victor B. Wagner <vitus@cryptocom.ru>*
8315 * s390x assembler pack.
8319 * ARMv4 assembler pack. ARMv4 refers to v4 and later ISA, not CPU
8324 * Implement Opaque PRF Input TLS extension as specified in
8325 draft-rescorla-tls-opaque-prf-input-00.txt. Since this is not an
8326 official specification yet and no extension type assignment by
8327 IANA exists, this extension (for now) will have to be explicitly
8328 enabled when building OpenSSL by providing the extension number
8329 to use. For example, specify an option
8331 -DTLSEXT_TYPE_opaque_prf_input=0x9527
8333 to the "config" or "Configure" script to enable the extension,
8334 assuming extension number 0x9527 (which is a completely arbitrary
8335 and unofficial assignment based on the MD5 hash of the Internet
8336 Draft). Note that by doing so, you potentially lose
8337 interoperability with other TLS implementations since these might
8338 be using the same extension number for other purposes.
8340 SSL_set_tlsext_opaque_prf_input(ssl, src, len) is used to set the
8341 opaque PRF input value to use in the handshake. This will create
8342 an internal copy of the length-'len' string at 'src', and will
8343 return non-zero for success.
8345 To get more control and flexibility, provide a callback function
8348 SSL_CTX_set_tlsext_opaque_prf_input_callback(ctx, cb)
8349 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg(ctx, arg)
8353 int (*cb)(SSL *, void *peerinput, size_t len, void *arg);
8356 Callback function 'cb' will be called in handshakes, and is
8357 expected to use SSL_set_tlsext_opaque_prf_input() as appropriate.
8358 Argument 'arg' is for application purposes (the value as given to
8359 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg() will directly
8360 be provided to the callback function). The callback function
8361 has to return non-zero to report success: usually 1 to use opaque
8362 PRF input just if possible, or 2 to enforce use of the opaque PRF
8363 input. In the latter case, the library will abort the handshake
8364 if opaque PRF input is not successfully negotiated.
8366 Arguments 'peerinput' and 'len' given to the callback function
8367 will always be NULL and 0 in the case of a client. A server will
8368 see the client's opaque PRF input through these variables if
8369 available (NULL and 0 otherwise). Note that if the server
8370 provides an opaque PRF input, the length must be the same as the
8371 length of the client's opaque PRF input.
8373 Note that the callback function will only be called when creating
8374 a new session (session resumption can resume whatever was
8375 previously negotiated), and will not be called in SSL 2.0
8376 handshakes; thus, SSL_CTX_set_options(ctx, SSL_OP_NO_SSLv2) or
8377 SSL_set_options(ssl, SSL_OP_NO_SSLv2) is especially recommended
8378 for applications that need to enforce opaque PRF input.
8382 * Update ssl code to support digests other than SHA1+MD5 for handshake
8385 *Victor B. Wagner <vitus@cryptocom.ru>*
8387 * Add RFC4507 support to OpenSSL. This includes the corrections in
8388 RFC4507bis. The encrypted ticket format is an encrypted encoded
8389 SSL_SESSION structure, that way new session features are automatically
8392 If a client application caches session in an SSL_SESSION structure
8393 support is transparent because tickets are now stored in the encoded
8396 The SSL_CTX structure automatically generates keys for ticket
8397 protection in servers so again support should be possible
8398 with no application modification.
8400 If a client or server wishes to disable RFC4507 support then the option
8401 SSL_OP_NO_TICKET can be set.
8403 Add a TLS extension debugging callback to allow the contents of any client
8404 or server extensions to be examined.
8406 This work was sponsored by Google.
8410 * Final changes to avoid use of pointer pointer casts in OpenSSL.
8411 OpenSSL should now compile cleanly on gcc 4.2
8413 *Peter Hartley <pdh@utter.chaos.org.uk>, Steve Henson*
8415 * Update SSL library to use new EVP_PKEY MAC API. Include generic MAC
8416 support including streaming MAC support: this is required for GOST
8417 ciphersuite support.
8419 *Victor B. Wagner <vitus@cryptocom.ru>, Steve Henson*
8421 * Add option -stream to use PKCS#7 streaming in smime utility. New
8422 function i2d_PKCS7_bio_stream() and PEM_write_PKCS7_bio_stream()
8423 to output in BER and PEM format.
8427 * Experimental support for use of HMAC via EVP_PKEY interface. This
8428 allows HMAC to be handled via the `EVP_DigestSign*()` interface. The
8429 EVP_PKEY "key" in this case is the HMAC key, potentially allowing
8430 ENGINE support for HMAC keys which are unextractable. New -mac and
8431 -macopt options to dgst utility.
8435 * New option -sigopt to dgst utility. Update dgst to use
8436 `EVP_Digest{Sign,Verify}*`. These two changes make it possible to use
8437 alternative signing parameters such as X9.31 or PSS in the dgst
8442 * Change ssl_cipher_apply_rule(), the internal function that does
8443 the work each time a ciphersuite string requests enabling
8444 ("foo+bar"), moving ("+foo+bar"), disabling ("-foo+bar", or
8445 removing ("!foo+bar") a class of ciphersuites: Now it maintains
8446 the order of disabled ciphersuites such that those ciphersuites
8447 that most recently went from enabled to disabled not only stay
8448 in order with respect to each other, but also have higher priority
8449 than other disabled ciphersuites the next time ciphersuites are
8452 This means that you can now say, e.g., "PSK:-PSK:HIGH" to enable
8453 the same ciphersuites as with "HIGH" alone, but in a specific
8454 order where the PSK ciphersuites come first (since they are the
8455 most recently disabled ciphersuites when "HIGH" is parsed).
8457 Also, change ssl_create_cipher_list() (using this new
8458 functionality) such that between otherwise identical
8459 ciphersuites, ephemeral ECDH is preferred over ephemeral DH in
8464 * Change ssl_create_cipher_list() so that it automatically
8465 arranges the ciphersuites in reasonable order before starting
8466 to process the rule string. Thus, the definition for "DEFAULT"
8467 (SSL_DEFAULT_CIPHER_LIST) now is just "ALL:!aNULL:!eNULL", but
8468 remains equivalent to `"AES:ALL:!aNULL:!eNULL:+aECDH:+kRSA:+RC4:@STRENGTH"`.
8469 This makes it much easier to arrive at a reasonable default order
8470 in applications for which anonymous ciphers are OK (meaning
8471 that you can't actually use DEFAULT).
8473 *Bodo Moeller; suggested by Victor Duchovni*
8475 * Split the SSL/TLS algorithm mask (as used for ciphersuite string
8476 processing) into multiple integers instead of setting
8477 "SSL_MKEY_MASK" bits, "SSL_AUTH_MASK" bits, "SSL_ENC_MASK",
8478 "SSL_MAC_MASK", and "SSL_SSL_MASK" bits all in a single integer.
8479 (These masks as well as the individual bit definitions are hidden
8480 away into the non-exported interface ssl/ssl_locl.h, so this
8481 change to the definition of the SSL_CIPHER structure shouldn't
8482 affect applications.) This give us more bits for each of these
8483 categories, so there is no longer a need to coagulate AES128 and
8484 AES256 into a single algorithm bit, and to coagulate Camellia128
8485 and Camellia256 into a single algorithm bit, which has led to all
8488 Thus, among other things, the kludge introduced in 0.9.7m and
8489 0.9.8e for masking out AES256 independently of AES128 or masking
8490 out Camellia256 independently of AES256 is not needed here in 0.9.9.
8492 With the change, we also introduce new ciphersuite aliases that
8493 so far were missing: "AES128", "AES256", "CAMELLIA128", and
8498 * Add support for dsa-with-SHA224 and dsa-with-SHA256.
8499 Use the leftmost N bytes of the signature input if the input is
8500 larger than the prime q (with N being the size in bytes of q).
8504 * Very *very* experimental PKCS#7 streaming encoder support. Nothing uses
8505 it yet and it is largely untested.
8509 * Add support for the ecdsa-with-SHA224/256/384/512 signature types.
8513 * Initial incomplete changes to avoid need for function casts in OpenSSL
8514 some compilers (gcc 4.2 and later) reject their use. Safestack is
8515 reimplemented. Update ASN1 to avoid use of legacy functions.
8519 * Win32/64 targets are linked with Winsock2.
8523 * Add an X509_CRL_METHOD structure to allow CRL processing to be redirected
8524 to external functions. This can be used to increase CRL handling
8525 efficiency especially when CRLs are very large by (for example) storing
8526 the CRL revoked certificates in a database.
8530 * Overhaul of by_dir code. Add support for dynamic loading of CRLs so
8531 new CRLs added to a directory can be used. New command line option
8532 -verify_return_error to s_client and s_server. This causes real errors
8533 to be returned by the verify callback instead of carrying on no matter
8534 what. This reflects the way a "real world" verify callback would behave.
8538 * GOST engine, supporting several GOST algorithms and public key formats.
8539 Kindly donated by Cryptocom.
8543 * Partial support for Issuing Distribution Point CRL extension. CRLs
8544 partitioned by DP are handled but no indirect CRL or reason partitioning
8545 (yet). Complete overhaul of CRL handling: now the most suitable CRL is
8546 selected via a scoring technique which handles IDP and AKID in CRLs.
8550 * New X509_STORE_CTX callbacks lookup_crls() and lookup_certs() which
8551 will ultimately be used for all verify operations: this will remove the
8552 X509_STORE dependency on certificate verification and allow alternative
8553 lookup methods. X509_STORE based implementations of these two callbacks.
8557 * Allow multiple CRLs to exist in an X509_STORE with matching issuer names.
8558 Modify get_crl() to find a valid (unexpired) CRL if possible.
8562 * New function X509_CRL_match() to check if two CRLs are identical. Normally
8563 this would be called X509_CRL_cmp() but that name is already used by
8564 a function that just compares CRL issuer names. Cache several CRL
8565 extensions in X509_CRL structure and cache CRLDP in X509.
8569 * Store a "canonical" representation of X509_NAME structure (ASN1 Name)
8570 this maps equivalent X509_NAME structures into a consistent structure.
8571 Name comparison can then be performed rapidly using memcmp().
8575 * Non-blocking OCSP request processing. Add -timeout option to ocsp
8580 * Allow digests to supply their own micalg string for S/MIME type using
8581 the ctrl EVP_MD_CTRL_MICALG.
8585 * During PKCS7 signing pass the PKCS7 SignerInfo structure to the
8586 EVP_PKEY_METHOD before and after signing via the EVP_PKEY_CTRL_PKCS7_SIGN
8587 ctrl. It can then customise the structure before and/or after signing
8592 * New function OBJ_add_sigid() to allow application defined signature OIDs
8593 to be added to OpenSSLs internal tables. New function OBJ_sigid_free()
8594 to free up any added signature OIDs.
8598 * New functions EVP_CIPHER_do_all(), EVP_CIPHER_do_all_sorted(),
8599 EVP_MD_do_all() and EVP_MD_do_all_sorted() to enumerate internal
8600 digest and cipher tables. New options added to openssl utility:
8601 list-message-digest-algorithms and list-cipher-algorithms.
8605 * Change the array representation of binary polynomials: the list
8606 of degrees of non-zero coefficients is now terminated with -1.
8607 Previously it was terminated with 0, which was also part of the
8608 value; thus, the array representation was not applicable to
8609 polynomials where t^0 has coefficient zero. This change makes
8610 the array representation useful in a more general context.
8614 * Various modifications and fixes to SSL/TLS cipher string
8615 handling. For ECC, the code now distinguishes between fixed ECDH
8616 with RSA certificates on the one hand and with ECDSA certificates
8617 on the other hand, since these are separate ciphersuites. The
8618 unused code for Fortezza ciphersuites has been removed.
8620 For consistency with EDH, ephemeral ECDH is now called "EECDH"
8621 (not "ECDHE"). For consistency with the code for DH
8622 certificates, use of ECDH certificates is now considered ECDH
8623 authentication, not RSA or ECDSA authentication (the latter is
8624 merely the CA's signing algorithm and not actively used in the
8627 The temporary ciphersuite alias "ECCdraft" is no longer
8628 available, and ECC ciphersuites are no longer excluded from "ALL"
8629 and "DEFAULT". The following aliases now exist for RFC 4492
8630 ciphersuites, most of these by analogy with the DH case:
8632 kECDHr - ECDH cert, signed with RSA
8633 kECDHe - ECDH cert, signed with ECDSA
8634 kECDH - ECDH cert (signed with either RSA or ECDSA)
8635 kEECDH - ephemeral ECDH
8636 ECDH - ECDH cert or ephemeral ECDH
8642 AECDH - anonymous ECDH
8643 EECDH - non-anonymous ephemeral ECDH (equivalent to "kEECDH:-AECDH")
8647 * Add additional S/MIME capabilities for AES and GOST ciphers if supported.
8648 Use correct micalg parameters depending on digest(s) in signed message.
8652 * Add engine support for EVP_PKEY_ASN1_METHOD. Add functions to process
8653 an ENGINE asn1 method. Support ENGINE lookups in the ASN1 code.
8657 * Initial engine support for EVP_PKEY_METHOD. New functions to permit
8658 an engine to register a method. Add ENGINE lookups for methods and
8659 functional reference processing.
8663 * New functions `EVP_Digest{Sign,Verify)*`. These are enhanced versions of
8664 `EVP_{Sign,Verify}*` which allow an application to customise the signature
8669 * New -resign option to smime utility. This adds one or more signers
8670 to an existing PKCS#7 signedData structure. Also -md option to use an
8671 alternative message digest algorithm for signing.
8675 * Tidy up PKCS#7 routines and add new functions to make it easier to
8676 create PKCS7 structures containing multiple signers. Update smime
8677 application to support multiple signers.
8681 * New -macalg option to pkcs12 utility to allow setting of an alternative
8686 * Initial support for PKCS#5 v2.0 PRFs other than default SHA1 HMAC.
8687 Reorganize PBE internals to lookup from a static table using NIDs,
8688 add support for HMAC PBE OID translation. Add a EVP_CIPHER ctrl:
8689 EVP_CTRL_PBE_PRF_NID this allows a cipher to specify an alternative
8690 PRF which will be automatically used with PBES2.
8694 * Replace the algorithm specific calls to generate keys in "req" with the
8699 * Update PKCS#7 enveloped data routines to use new API. This is now
8700 supported by any public key method supporting the encrypt operation. A
8701 ctrl is added to allow the public key algorithm to examine or modify
8702 the PKCS#7 RecipientInfo structure if it needs to: for RSA this is
8707 * Add a ctrl to asn1 method to allow a public key algorithm to express
8708 a default digest type to use. In most cases this will be SHA1 but some
8709 algorithms (such as GOST) need to specify an alternative digest. The
8710 return value indicates how strong the preference is 1 means optional and
8711 2 is mandatory (that is it is the only supported type). Modify
8712 ASN1_item_sign() to accept a NULL digest argument to indicate it should
8713 use the default md. Update openssl utilities to use the default digest
8714 type for signing if it is not explicitly indicated.
8718 * Use OID cross reference table in ASN1_sign() and ASN1_verify(). New
8719 EVP_MD flag EVP_MD_FLAG_PKEY_METHOD_SIGNATURE. This uses the relevant
8720 signing method from the key type. This effectively removes the link
8721 between digests and public key types.
8725 * Add an OID cross reference table and utility functions. Its purpose is to
8726 translate between signature OIDs such as SHA1WithrsaEncryption and SHA1,
8727 rsaEncryption. This will allow some of the algorithm specific hackery
8728 needed to use the correct OID to be removed.
8732 * Remove algorithm specific dependencies when setting PKCS7_SIGNER_INFO
8733 structures for PKCS7_sign(). They are now set up by the relevant public
8738 * Add provisional EC pkey method with support for ECDSA and ECDH.
8742 * Add support for key derivation (agreement) in the API, DH method and
8747 * Add DSA pkey method and DH pkey methods, extend DH ASN1 method to support
8748 public and private key formats. As a side effect these add additional
8749 command line functionality not previously available: DSA signatures can be
8750 generated and verified using pkeyutl and DH key support and generation in
8757 *Oliver Tappe <zooey@hirschkaefer.de>*
8759 * New make target "install_html_docs" installs HTML renditions of the
8762 *Oliver Tappe <zooey@hirschkaefer.de>*
8764 * New utility "genpkey" this is analogous to "genrsa" etc except it can
8765 generate keys for any algorithm. Extend and update EVP_PKEY_METHOD to
8766 support key and parameter generation and add initial key generation
8767 functionality for RSA.
8771 * Add functions for main EVP_PKEY_method operations. The undocumented
8772 functions `EVP_PKEY_{encrypt,decrypt}` have been renamed to
8773 `EVP_PKEY_{encrypt,decrypt}_old`.
8777 * Initial definitions for EVP_PKEY_METHOD. This will be a high level public
8778 key API, doesn't do much yet.
8782 * New function EVP_PKEY_asn1_get0_info() to retrieve information about
8783 public key algorithms. New option to openssl utility:
8784 "list-public-key-algorithms" to print out info.
8788 * Implement the Supported Elliptic Curves Extension for
8789 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
8793 * Don't free up OIDs in OBJ_cleanup() if they are in use by EVP_MD or
8794 EVP_CIPHER structures to avoid later problems in EVP_cleanup().
8798 * New utilities pkey and pkeyparam. These are similar to algorithm specific
8799 utilities such as rsa, dsa, dsaparam etc except they process any key
8804 * Transfer public key printing routines to EVP_PKEY_ASN1_METHOD. New
8805 functions EVP_PKEY_print_public(), EVP_PKEY_print_private(),
8806 EVP_PKEY_print_param() to print public key data from an EVP_PKEY
8811 * Initial support for pluggable public key ASN1.
8812 De-spaghettify the public key ASN1 handling. Move public and private
8813 key ASN1 handling to a new EVP_PKEY_ASN1_METHOD structure. Relocate
8814 algorithm specific handling to a single module within the relevant
8815 algorithm directory. Add functions to allow (near) opaque processing
8816 of public and private key structures.
8820 * Implement the Supported Point Formats Extension for
8821 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
8825 * Add initial support for RFC 4279 PSK TLS ciphersuites. Add members
8826 for the psk identity [hint] and the psk callback functions to the
8827 SSL_SESSION, SSL and SSL_CTX structure.
8830 PSK-RC4-SHA, PSK-3DES-EDE-CBC-SHA, PSK-AES128-CBC-SHA,
8834 SSL_CTX_use_psk_identity_hint
8835 SSL_get_psk_identity_hint
8836 SSL_get_psk_identity
8837 SSL_use_psk_identity_hint
8839 *Mika Kousa and Pasi Eronen of Nokia Corporation*
8841 * Add RFC 3161 compliant time stamp request creation, response generation
8842 and response verification functionality.
8844 *Zoltán Glózik <zglozik@opentsa.org>, The OpenTSA Project*
8846 * Add initial support for TLS extensions, specifically for the server_name
8847 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
8848 have new members for a host name. The SSL data structure has an
8849 additional member `SSL_CTX *initial_ctx` so that new sessions can be
8850 stored in that context to allow for session resumption, even after the
8851 SSL has been switched to a new SSL_CTX in reaction to a client's
8852 server_name extension.
8854 New functions (subject to change):
8856 SSL_get_servername()
8857 SSL_get_servername_type()
8860 New CTRL codes and macros (subject to change):
8862 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
8863 - SSL_CTX_set_tlsext_servername_callback()
8864 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
8865 - SSL_CTX_set_tlsext_servername_arg()
8866 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
8868 openssl s_client has a new '-servername ...' option.
8870 openssl s_server has new options '-servername_host ...', '-cert2 ...',
8871 '-key2 ...', '-servername_fatal' (subject to change). This allows
8872 testing the HostName extension for a specific single host name ('-cert'
8873 and '-key' remain fallbacks for handshakes without HostName
8874 negotiation). If the unrecognized_name alert has to be sent, this by
8875 default is a warning; it becomes fatal with the '-servername_fatal'
8878 *Peter Sylvester, Remy Allais, Christophe Renou*
8880 * Whirlpool hash implementation is added.
8884 * BIGNUM code on 64-bit SPARCv9 targets is switched from bn(64,64) to
8885 bn(64,32). Because of instruction set limitations it doesn't have
8886 any negative impact on performance. This was done mostly in order
8887 to make it possible to share assembler modules, such as bn_mul_mont
8888 implementations, between 32- and 64-bit builds without hassle.
8892 * Move code previously exiled into file crypto/ec/ec2_smpt.c
8893 to ec2_smpl.c, and no longer require the OPENSSL_EC_BIN_PT_COMP
8898 * New candidate for BIGNUM assembler implementation, bn_mul_mont,
8899 dedicated Montgomery multiplication procedure, is introduced.
8900 BN_MONT_CTX is modified to allow bn_mul_mont to reach for higher
8901 "64-bit" performance on certain 32-bit targets.
8905 * New option SSL_OP_NO_COMP to disable use of compression selectively
8906 in SSL structures. New SSL ctrl to set maximum send fragment size.
8907 Save memory by setting the I/O buffer sizes dynamically instead of
8908 using the maximum available value.
8912 * New option -V for 'openssl ciphers'. This prints the ciphersuite code
8913 in addition to the text details.
8917 * Very, very preliminary EXPERIMENTAL support for printing of general
8918 ASN1 structures. This currently produces rather ugly output and doesn't
8919 handle several customised structures at all.
8923 * Integrated support for PVK file format and some related formats such
8924 as MS PUBLICKEYBLOB and PRIVATEKEYBLOB. Command line switches to support
8925 these in the 'rsa' and 'dsa' utilities.
8929 * Support for PKCS#1 RSAPublicKey format on rsa utility command line.
8933 * Remove the ancient ASN1_METHOD code. This was only ever used in one
8934 place for the (very old) "NETSCAPE" format certificates which are now
8935 handled using new ASN1 code equivalents.
8939 * Let the TLSv1_method() etc. functions return a 'const' SSL_METHOD
8940 pointer and make the SSL_METHOD parameter in SSL_CTX_new,
8941 SSL_CTX_set_ssl_version and SSL_set_ssl_method 'const'.
8945 * Modify CRL distribution points extension code to print out previously
8946 unsupported fields. Enhance extension setting code to allow setting of
8951 * Add print and set support for Issuing Distribution Point CRL extension.
8955 * Change 'Configure' script to enable Camellia by default.
8962 ### Changes between 0.9.8m and 0.9.8n [24 Mar 2010]
8964 * When rejecting SSL/TLS records due to an incorrect version number, never
8965 update s->server with a new major version number. As of
8966 - OpenSSL 0.9.8m if 'short' is a 16-bit type,
8967 - OpenSSL 0.9.8f if 'short' is longer than 16 bits,
8968 the previous behavior could result in a read attempt at NULL when
8969 receiving specific incorrect SSL/TLS records once record payload
8970 protection is active. ([CVE-2010-0740])
8972 *Bodo Moeller, Adam Langley <agl@chromium.org>*
8974 * Fix for CVE-2010-0433 where some kerberos enabled versions of OpenSSL
8975 could be crashed if the relevant tables were not present (e.g. chrooted).
8977 *Tomas Hoger <thoger@redhat.com>*
8979 ### Changes between 0.9.8l and 0.9.8m [25 Feb 2010]
8981 * Always check bn_wexpand() return values for failure. ([CVE-2009-3245])
8983 *Martin Olsson, Neel Mehta*
8985 * Fix X509_STORE locking: Every 'objs' access requires a lock (to
8986 accommodate for stack sorting, always a write lock!).
8990 * On some versions of WIN32 Heap32Next is very slow. This can cause
8991 excessive delays in the RAND_poll(): over a minute. As a workaround
8992 include a time check in the inner Heap32Next loop too.
8996 * The code that handled flushing of data in SSL/TLS originally used the
8997 BIO_CTRL_INFO ctrl to see if any data was pending first. This caused
8998 the problem outlined in PR#1949. The fix suggested there however can
8999 trigger problems with buggy BIO_CTRL_WPENDING (e.g. some versions
9000 of Apache). So instead simplify the code to flush unconditionally.
9001 This should be fine since flushing with no data to flush is a no op.
9005 * Handle TLS versions 2.0 and later properly and correctly use the
9006 highest version of TLS/SSL supported. Although TLS >= 2.0 is some way
9007 off ancient servers have a habit of sticking around for a while...
9011 * Modify compression code so it frees up structures without using the
9012 ex_data callbacks. This works around a problem where some applications
9013 call CRYPTO_cleanup_all_ex_data() before application exit (e.g. when
9014 restarting) then use compression (e.g. SSL with compression) later.
9015 This results in significant per-connection memory leaks and
9016 has caused some security issues including CVE-2008-1678 and
9021 * Constify crypto/cast (i.e., <openssl/cast.h>): a CAST_KEY doesn't
9022 change when encrypting or decrypting.
9026 * Add option SSL_OP_LEGACY_SERVER_CONNECT which will allow clients to
9027 connect and renegotiate with servers which do not support RI.
9028 Until RI is more widely deployed this option is enabled by default.
9032 * Add "missing" ssl ctrls to clear options and mode.
9036 * If client attempts to renegotiate and doesn't support RI respond with
9037 a no_renegotiation alert as required by RFC5746. Some renegotiating
9038 TLS clients will continue a connection gracefully when they receive
9039 the alert. Unfortunately OpenSSL mishandled this alert and would hang
9040 waiting for a server hello which it will never receive. Now we treat a
9041 received no_renegotiation alert as a fatal error. This is because
9042 applications requesting a renegotiation might well expect it to succeed
9043 and would have no code in place to handle the server denying it so the
9044 only safe thing to do is to terminate the connection.
9048 * Add ctrl macro SSL_get_secure_renegotiation_support() which returns 1 if
9049 peer supports secure renegotiation and 0 otherwise. Print out peer
9050 renegotiation support in s_client/s_server.
9054 * Replace the highly broken and deprecated SPKAC certification method with
9055 the updated NID creation version. This should correctly handle UTF8.
9059 * Implement RFC5746. Re-enable renegotiation but require the extension
9060 as needed. Unfortunately, SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION
9061 turns out to be a bad idea. It has been replaced by
9062 SSL_OP_ALLOW_UNSAFE_LEGACY_RENEGOTIATION which can be set with
9063 SSL_CTX_set_options(). This is really not recommended unless you
9064 know what you are doing.
9066 *Eric Rescorla <ekr@networkresonance.com>, Ben Laurie, Steve Henson*
9068 * Fixes to stateless session resumption handling. Use initial_ctx when
9069 issuing and attempting to decrypt tickets in case it has changed during
9070 servername handling. Use a non-zero length session ID when attempting
9071 stateless session resumption: this makes it possible to determine if
9072 a resumption has occurred immediately after receiving server hello
9073 (several places in OpenSSL subtly assume this) instead of later in
9078 * The functions ENGINE_ctrl(), OPENSSL_isservice(),
9079 CMS_get1_RecipientRequest() and RAND_bytes() can return <=0 on error
9080 fixes for a few places where the return code is not checked
9083 *Julia Lawall <julia@diku.dk>*
9085 * Add --strict-warnings option to Configure script to include devteam
9086 warnings in other configurations.
9090 * Add support for --libdir option and LIBDIR variable in makefiles. This
9091 makes it possible to install openssl libraries in locations which
9092 have names other than "lib", for example "/usr/lib64" which some
9095 *Steve Henson, based on patch from Jeremy Utley*
9097 * Don't allow the use of leading 0x80 in OIDs. This is a violation of
9098 X690 8.9.12 and can produce some misleading textual output of OIDs.
9100 *Steve Henson, reported by Dan Kaminsky*
9102 * Delete MD2 from algorithm tables. This follows the recommendation in
9103 several standards that it is not used in new applications due to
9104 several cryptographic weaknesses. For binary compatibility reasons
9105 the MD2 API is still compiled in by default.
9109 * Add compression id to {d2i,i2d}_SSL_SESSION so it is correctly saved
9114 * Rename uni2asc and asc2uni functions to OPENSSL_uni2asc and
9115 OPENSSL_asc2uni conditionally on Netware platforms to avoid a name
9118 *Guenter <lists@gknw.net>*
9120 * Fix the server certificate chain building code to use X509_verify_cert(),
9121 it used to have an ad-hoc builder which was unable to cope with anything
9122 other than a simple chain.
9124 *David Woodhouse <dwmw2@infradead.org>, Steve Henson*
9126 * Don't check self signed certificate signatures in X509_verify_cert()
9127 by default (a flag can override this): it just wastes time without
9128 adding any security. As a useful side effect self signed root CAs
9129 with non-FIPS digests are now usable in FIPS mode.
9133 * In dtls1_process_out_of_seq_message() the check if the current message
9134 is already buffered was missing. For every new message was memory
9135 allocated, allowing an attacker to perform an denial of service attack
9136 with sending out of seq handshake messages until there is no memory
9137 left. Additionally every future message was buffered, even if the
9138 sequence number made no sense and would be part of another handshake.
9139 So only messages with sequence numbers less than 10 in advance will be
9140 buffered. ([CVE-2009-1378])
9142 *Robin Seggelmann, discovered by Daniel Mentz*
9144 * Records are buffered if they arrive with a future epoch to be
9145 processed after finishing the corresponding handshake. There is
9146 currently no limitation to this buffer allowing an attacker to perform
9147 a DOS attack with sending records with future epochs until there is no
9148 memory left. This patch adds the pqueue_size() function to determine
9149 the size of a buffer and limits the record buffer to 100 entries.
9152 *Robin Seggelmann, discovered by Daniel Mentz*
9154 * Keep a copy of frag->msg_header.frag_len so it can be used after the
9155 parent structure is freed. ([CVE-2009-1379])
9159 * Handle non-blocking I/O properly in SSL_shutdown() call.
9161 *Darryl Miles <darryl-mailinglists@netbauds.net>*
9163 * Add `2.5.4.*` OIDs
9165 *Ilya O. <vrghost@gmail.com>*
9167 ### Changes between 0.9.8k and 0.9.8l [5 Nov 2009]
9169 * Disable renegotiation completely - this fixes a severe security
9170 problem ([CVE-2009-3555]) at the cost of breaking all
9171 renegotiation. Renegotiation can be re-enabled by setting
9172 SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION in s3->flags at
9173 run-time. This is really not recommended unless you know what
9178 ### Changes between 0.9.8j and 0.9.8k [25 Mar 2009]
9180 * Don't set val to NULL when freeing up structures, it is freed up by
9181 underlying code. If `sizeof(void *) > sizeof(long)` this can result in
9182 zeroing past the valid field. ([CVE-2009-0789])
9184 *Paolo Ganci <Paolo.Ganci@AdNovum.CH>*
9186 * Fix bug where return value of CMS_SignerInfo_verify_content() was not
9187 checked correctly. This would allow some invalid signed attributes to
9188 appear to verify correctly. ([CVE-2009-0591])
9190 *Ivan Nestlerode <inestlerode@us.ibm.com>*
9192 * Reject UniversalString and BMPString types with invalid lengths. This
9193 prevents a crash in ASN1_STRING_print_ex() which assumes the strings have
9194 a legal length. ([CVE-2009-0590])
9198 * Set S/MIME signing as the default purpose rather than setting it
9199 unconditionally. This allows applications to override it at the store
9204 * Permit restricted recursion of ASN1 strings. This is needed in practice
9205 to handle some structures.
9209 * Improve efficiency of mem_gets: don't search whole buffer each time
9212 *Jeremy Shapiro <jnshapir@us.ibm.com>*
9214 * New -hex option for openssl rand.
9218 * Print out UTF8String and NumericString when parsing ASN1.
9222 * Support NumericString type for name components.
9226 * Allow CC in the environment to override the automatically chosen
9227 compiler. Note that nothing is done to ensure flags work with the
9232 ### Changes between 0.9.8i and 0.9.8j [07 Jan 2009]
9234 * Properly check EVP_VerifyFinal() and similar return values
9237 *Ben Laurie, Bodo Moeller, Google Security Team*
9239 * Enable TLS extensions by default.
9243 * Allow the CHIL engine to be loaded, whether the application is
9244 multithreaded or not. (This does not release the developer from the
9245 obligation to set up the dynamic locking callbacks.)
9247 *Sander Temme <sander@temme.net>*
9249 * Use correct exit code if there is an error in dgst command.
9251 *Steve Henson; problem pointed out by Roland Dirlewanger*
9253 * Tweak Configure so that you need to say "experimental-jpake" to enable
9254 JPAKE, and need to use -DOPENSSL_EXPERIMENTAL_JPAKE in applications.
9258 * Add experimental JPAKE support, including demo authentication in
9259 s_client and s_server.
9263 * Set the comparison function in v3_addr_canonize().
9265 *Rob Austein <sra@hactrn.net>*
9267 * Add support for XMPP STARTTLS in s_client.
9269 *Philip Paeps <philip@freebsd.org>*
9271 * Change the server-side SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG behavior
9272 to ensure that even with this option, only ciphersuites in the
9273 server's preference list will be accepted. (Note that the option
9274 applies only when resuming a session, so the earlier behavior was
9275 just about the algorithm choice for symmetric cryptography.)
9279 ### Changes between 0.9.8h and 0.9.8i [15 Sep 2008]
9281 * Fix NULL pointer dereference if a DTLS server received
9282 ChangeCipherSpec as first record ([CVE-2009-1386]).
9286 * Fix a state transition in s3_srvr.c and d1_srvr.c
9287 (was using SSL3_ST_CW_CLNT_HELLO_B, should be `..._ST_SW_SRVR_...`).
9291 * The fix in 0.9.8c that supposedly got rid of unsafe
9292 double-checked locking was incomplete for RSA blinding,
9293 addressing just one layer of what turns out to have been
9294 doubly unsafe triple-checked locking.
9296 So now fix this for real by retiring the MONT_HELPER macro
9297 in crypto/rsa/rsa_eay.c.
9299 *Bodo Moeller; problem pointed out by Marius Schilder*
9301 * Various precautionary measures:
9303 - Avoid size_t integer overflow in HASH_UPDATE (md32_common.h).
9305 - Avoid a buffer overflow in d2i_SSL_SESSION() (ssl_asn1.c).
9306 (NB: This would require knowledge of the secret session ticket key
9307 to exploit, in which case you'd be SOL either way.)
9309 - Change bn_nist.c so that it will properly handle input BIGNUMs
9310 outside the expected range.
9312 - Enforce the 'num' check in BN_div() (bn_div.c) for non-BN_DEBUG
9315 *Neel Mehta, Bodo Moeller*
9317 * Allow engines to be "soft loaded" - i.e. optionally don't die if
9318 the load fails. Useful for distros.
9320 *Ben Laurie and the FreeBSD team*
9322 * Add support for Local Machine Keyset attribute in PKCS#12 files.
9326 * Fix BN_GF2m_mod_arr() top-bit cleanup code.
9330 * Expand ENGINE to support engine supplied SSL client certificate functions.
9332 This work was sponsored by Logica.
9336 * Add CryptoAPI ENGINE to support use of RSA and DSA keys held in Windows
9337 keystores. Support for SSL/TLS client authentication too.
9338 Not compiled unless enable-capieng specified to Configure.
9340 This work was sponsored by Logica.
9344 * Fix bug in X509_ATTRIBUTE creation: don't set attribute using
9345 ASN1_TYPE_set1 if MBSTRING flag set. This bug would crash certain
9346 attribute creation routines such as certificate requests and PKCS#12
9351 ### Changes between 0.9.8g and 0.9.8h [28 May 2008]
9353 * Fix flaw if 'Server Key exchange message' is omitted from a TLS
9354 handshake which could lead to a client crash as found using the
9355 Codenomicon TLS test suite ([CVE-2008-1672])
9357 *Steve Henson, Mark Cox*
9359 * Fix double free in TLS server name extensions which could lead to
9360 a remote crash found by Codenomicon TLS test suite ([CVE-2008-0891])
9364 * Clear error queue in SSL_CTX_use_certificate_chain_file()
9366 Clear the error queue to ensure that error entries left from
9367 older function calls do not interfere with the correct operation.
9369 *Lutz Jaenicke, Erik de Castro Lopo*
9371 * Remove root CA certificates of commercial CAs:
9373 The OpenSSL project does not recommend any specific CA and does not
9374 have any policy with respect to including or excluding any CA.
9375 Therefore it does not make any sense to ship an arbitrary selection
9376 of root CA certificates with the OpenSSL software.
9380 * RSA OAEP patches to fix two separate invalid memory reads.
9381 The first one involves inputs when 'lzero' is greater than
9382 'SHA_DIGEST_LENGTH' (it would read about SHA_DIGEST_LENGTH bytes
9383 before the beginning of from). The second one involves inputs where
9384 the 'db' section contains nothing but zeroes (there is a one-byte
9385 invalid read after the end of 'db').
9387 *Ivan Nestlerode <inestlerode@us.ibm.com>*
9389 * Partial backport from 0.9.9-dev:
9391 Introduce bn_mul_mont (dedicated Montgomery multiplication
9392 procedure) as a candidate for BIGNUM assembler implementation.
9393 While 0.9.9-dev uses assembler for various architectures, only
9394 x86_64 is available by default here in the 0.9.8 branch, and
9395 32-bit x86 is available through a compile-time setting.
9397 To try the 32-bit x86 assembler implementation, use Configure
9398 option "enable-montasm" (which exists only for this backport).
9400 As "enable-montasm" for 32-bit x86 disclaims code stability
9401 anyway, in this constellation we activate additional code
9402 backported from 0.9.9-dev for further performance improvements,
9403 namely BN_from_montgomery_word. (To enable this otherwise,
9404 e.g. x86_64, try `-DMONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD`.)
9406 *Andy Polyakov (backport partially by Bodo Moeller)*
9408 * Add TLS session ticket callback. This allows an application to set
9409 TLS ticket cipher and HMAC keys rather than relying on hardcoded fixed
9410 values. This is useful for key rollover for example where several key
9411 sets may exist with different names.
9415 * Reverse ENGINE-internal logic for caching default ENGINE handles.
9416 This was broken until now in 0.9.8 releases, such that the only way
9417 a registered ENGINE could be used (assuming it initialises
9418 successfully on the host) was to explicitly set it as the default
9419 for the relevant algorithms. This is in contradiction with 0.9.7
9420 behaviour and the documentation. With this fix, when an ENGINE is
9421 registered into a given algorithm's table of implementations, the
9422 'uptodate' flag is reset so that auto-discovery will be used next
9423 time a new context for that algorithm attempts to select an
9426 *Ian Lister (tweaked by Geoff Thorpe)*
9428 * Backport of CMS code to OpenSSL 0.9.8. This differs from the 0.9.9
9429 implementation in the following ways:
9431 Lack of EVP_PKEY_ASN1_METHOD means algorithm parameters have to be
9434 Lack of BER streaming support means one pass streaming processing is
9435 only supported if data is detached: setting the streaming flag is
9436 ignored for embedded content.
9438 CMS support is disabled by default and must be explicitly enabled
9439 with the enable-cms configuration option.
9443 * Update the GMP engine glue to do direct copies between BIGNUM and
9444 mpz_t when openssl and GMP use the same limb size. Otherwise the
9445 existing "conversion via a text string export" trick is still used.
9447 *Paul Sheer <paulsheer@gmail.com>*
9449 * Zlib compression BIO. This is a filter BIO which compressed and
9450 uncompresses any data passed through it.
9454 * Add AES_wrap_key() and AES_unwrap_key() functions to implement
9455 RFC3394 compatible AES key wrapping.
9459 * Add utility functions to handle ASN1 structures. ASN1_STRING_set0():
9460 sets string data without copying. X509_ALGOR_set0() and
9461 X509_ALGOR_get0(): set and retrieve X509_ALGOR (AlgorithmIdentifier)
9462 data. Attribute function X509at_get0_data_by_OBJ(): retrieves data
9463 from an X509_ATTRIBUTE structure optionally checking it occurs only
9464 once. ASN1_TYPE_set1(): set and ASN1_TYPE structure copying supplied
9469 * Fix BN flag handling in RSA_eay_mod_exp() and BN_MONT_CTX_set()
9470 to get the expected BN_FLG_CONSTTIME behavior.
9472 *Bodo Moeller (Google)*
9476 - fixed wrong usage of ioctlsocket() when build for LIBC BSD sockets
9477 - fixed do_tests.pl to run the test suite with CLIB builds too (CLIB_OPT)
9478 - added some more tests to do_tests.pl
9479 - fixed RunningProcess usage so that it works with newer LIBC NDKs too
9480 - removed usage of BN_LLONG for CLIB builds to avoid runtime dependency
9481 - added new Configure targets netware-clib-bsdsock, netware-clib-gcc,
9482 netware-clib-bsdsock-gcc, netware-libc-bsdsock-gcc
9483 - various changes to netware.pl to enable gcc-cross builds on Win32
9485 - changed crypto/bio/b_sock.c to work with macro functions (CLIB BSD)
9486 - various changes to fix missing prototype warnings
9487 - fixed x86nasm.pl to create correct asm files for NASM COFF output
9488 - added AES, WHIRLPOOL and CPUID assembler code to build files
9489 - added missing AES assembler make rules to mk1mf.pl
9490 - fixed order of includes in `apps/ocsp.c` so that `e_os.h` settings apply
9492 *Guenter Knauf <eflash@gmx.net>*
9494 * Implement certificate status request TLS extension defined in RFC3546.
9495 A client can set the appropriate parameters and receive the encoded
9496 OCSP response via a callback. A server can query the supplied parameters
9497 and set the encoded OCSP response in the callback. Add simplified examples
9498 to s_client and s_server.
9502 ### Changes between 0.9.8f and 0.9.8g [19 Oct 2007]
9505 + Binary incompatibility of ssl_ctx_st structure
9506 + DTLS interoperation with non-compliant servers
9507 + Don't call get_session_cb() without proposed session
9508 + Fix ia64 assembler code
9510 *Andy Polyakov, Steve Henson*
9512 ### Changes between 0.9.8e and 0.9.8f [11 Oct 2007]
9514 * DTLS Handshake overhaul. There were longstanding issues with
9515 OpenSSL DTLS implementation, which were making it impossible for
9516 RFC 4347 compliant client to communicate with OpenSSL server.
9517 Unfortunately just fixing these incompatibilities would "cut off"
9518 pre-0.9.8f clients. To allow for hassle free upgrade post-0.9.8e
9519 server keeps tolerating non RFC compliant syntax. The opposite is
9520 not true, 0.9.8f client can not communicate with earlier server.
9521 This update even addresses CVE-2007-4995.
9525 * Changes to avoid need for function casts in OpenSSL: some compilers
9526 (gcc 4.2 and later) reject their use.
9527 *Kurt Roeckx <kurt@roeckx.be>, Peter Hartley <pdh@utter.chaos.org.uk>,
9530 * Add RFC4507 support to OpenSSL. This includes the corrections in
9531 RFC4507bis. The encrypted ticket format is an encrypted encoded
9532 SSL_SESSION structure, that way new session features are automatically
9535 If a client application caches session in an SSL_SESSION structure
9536 support is transparent because tickets are now stored in the encoded
9539 The SSL_CTX structure automatically generates keys for ticket
9540 protection in servers so again support should be possible
9541 with no application modification.
9543 If a client or server wishes to disable RFC4507 support then the option
9544 SSL_OP_NO_TICKET can be set.
9546 Add a TLS extension debugging callback to allow the contents of any client
9547 or server extensions to be examined.
9549 This work was sponsored by Google.
9553 * Add initial support for TLS extensions, specifically for the server_name
9554 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
9555 have new members for a host name. The SSL data structure has an
9556 additional member `SSL_CTX *initial_ctx` so that new sessions can be
9557 stored in that context to allow for session resumption, even after the
9558 SSL has been switched to a new SSL_CTX in reaction to a client's
9559 server_name extension.
9561 New functions (subject to change):
9563 SSL_get_servername()
9564 SSL_get_servername_type()
9567 New CTRL codes and macros (subject to change):
9569 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
9570 - SSL_CTX_set_tlsext_servername_callback()
9571 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
9572 - SSL_CTX_set_tlsext_servername_arg()
9573 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
9575 openssl s_client has a new '-servername ...' option.
9577 openssl s_server has new options '-servername_host ...', '-cert2 ...',
9578 '-key2 ...', '-servername_fatal' (subject to change). This allows
9579 testing the HostName extension for a specific single host name ('-cert'
9580 and '-key' remain fallbacks for handshakes without HostName
9581 negotiation). If the unrecognized_name alert has to be sent, this by
9582 default is a warning; it becomes fatal with the '-servername_fatal'
9585 *Peter Sylvester, Remy Allais, Christophe Renou, Steve Henson*
9587 * Add AES and SSE2 assembly language support to VC++ build.
9591 * Mitigate attack on final subtraction in Montgomery reduction.
9595 * Fix crypto/ec/ec_mult.c to work properly with scalars of value 0
9596 (which previously caused an internal error).
9600 * Squeeze another 10% out of IGE mode when in != out.
9604 * AES IGE mode speedup.
9606 *Dean Gaudet (Google)*
9608 * Add the Korean symmetric 128-bit cipher SEED (see
9609 <http://www.kisa.or.kr/kisa/seed/jsp/seed_eng.jsp>) and
9610 add SEED ciphersuites from RFC 4162:
9612 TLS_RSA_WITH_SEED_CBC_SHA = "SEED-SHA"
9613 TLS_DHE_DSS_WITH_SEED_CBC_SHA = "DHE-DSS-SEED-SHA"
9614 TLS_DHE_RSA_WITH_SEED_CBC_SHA = "DHE-RSA-SEED-SHA"
9615 TLS_DH_anon_WITH_SEED_CBC_SHA = "ADH-SEED-SHA"
9617 To minimize changes between patchlevels in the OpenSSL 0.9.8
9618 series, SEED remains excluded from compilation unless OpenSSL
9619 is configured with 'enable-seed'.
9621 *KISA, Bodo Moeller*
9623 * Mitigate branch prediction attacks, which can be practical if a
9624 single processor is shared, allowing a spy process to extract
9625 information. For detailed background information, see
9626 <http://eprint.iacr.org/2007/039> (O. Aciicmez, S. Gueron,
9627 J.-P. Seifert, "New Branch Prediction Vulnerabilities in OpenSSL
9628 and Necessary Software Countermeasures"). The core of the change
9629 are new versions BN_div_no_branch() and
9630 BN_mod_inverse_no_branch() of BN_div() and BN_mod_inverse(),
9631 respectively, which are slower, but avoid the security-relevant
9632 conditional branches. These are automatically called by BN_div()
9633 and BN_mod_inverse() if the flag BN_FLG_CONSTTIME is set for one
9634 of the input BIGNUMs. Also, BN_is_bit_set() has been changed to
9635 remove a conditional branch.
9637 BN_FLG_CONSTTIME is the new name for the previous
9638 BN_FLG_EXP_CONSTTIME flag, since it now affects more than just
9639 modular exponentiation. (Since OpenSSL 0.9.7h, setting this flag
9640 in the exponent causes BN_mod_exp_mont() to use the alternative
9641 implementation in BN_mod_exp_mont_consttime().) The old name
9642 remains as a deprecated alias.
9644 Similarly, RSA_FLAG_NO_EXP_CONSTTIME is replaced by a more general
9645 RSA_FLAG_NO_CONSTTIME flag since the RSA implementation now uses
9646 constant-time implementations for more than just exponentiation.
9647 Here too the old name is kept as a deprecated alias.
9649 BN_BLINDING_new() will now use BN_dup() for the modulus so that
9650 the BN_BLINDING structure gets an independent copy of the
9651 modulus. This means that the previous `BIGNUM *m` argument to
9652 BN_BLINDING_new() and to BN_BLINDING_create_param() now
9653 essentially becomes `const BIGNUM *m`, although we can't actually
9654 change this in the header file before 0.9.9. It allows
9655 RSA_setup_blinding() to use BN_with_flags() on the modulus to
9656 enable BN_FLG_CONSTTIME.
9658 *Matthew D Wood (Intel Corp)*
9660 * In the SSL/TLS server implementation, be strict about session ID
9661 context matching (which matters if an application uses a single
9662 external cache for different purposes). Previously,
9663 out-of-context reuse was forbidden only if SSL_VERIFY_PEER was
9664 set. This did ensure strict client verification, but meant that,
9665 with applications using a single external cache for quite
9666 different requirements, clients could circumvent ciphersuite
9667 restrictions for a given session ID context by starting a session
9668 in a different context.
9672 * Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
9673 a ciphersuite string such as "DEFAULT:RSA" cannot enable
9674 authentication-only ciphersuites.
9678 * Update the SSL_get_shared_ciphers() fix CVE-2006-3738 which was
9679 not complete and could lead to a possible single byte overflow
9680 ([CVE-2007-5135]) [Ben Laurie]
9682 ### Changes between 0.9.8d and 0.9.8e [23 Feb 2007]
9684 * Since AES128 and AES256 (and similarly Camellia128 and
9685 Camellia256) share a single mask bit in the logic of
9686 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
9687 kludge to work properly if AES128 is available and AES256 isn't
9688 (or if Camellia128 is available and Camellia256 isn't).
9692 * Fix the BIT STRING encoding generated by crypto/ec/ec_asn1.c
9693 (within i2d_ECPrivateKey, i2d_ECPKParameters, i2d_ECParameters):
9694 When a point or a seed is encoded in a BIT STRING, we need to
9695 prevent the removal of trailing zero bits to get the proper DER
9696 encoding. (By default, crypto/asn1/a_bitstr.c assumes the case
9697 of a NamedBitList, for which trailing 0 bits need to be removed.)
9701 * Have SSL/TLS server implementation tolerate "mismatched" record
9702 protocol version while receiving ClientHello even if the
9703 ClientHello is fragmented. (The server can't insist on the
9704 particular protocol version it has chosen before the ServerHello
9705 message has informed the client about his choice.)
9709 * Add RFC 3779 support.
9711 *Rob Austein for ARIN, Ben Laurie*
9713 * Load error codes if they are not already present instead of using a
9714 static variable. This allows them to be cleanly unloaded and reloaded.
9715 Improve header file function name parsing.
9719 * extend SMTP and IMAP protocol emulation in s_client to use EHLO
9720 or CAPABILITY handshake as required by RFCs.
9724 ### Changes between 0.9.8c and 0.9.8d [28 Sep 2006]
9726 * Introduce limits to prevent malicious keys being able to
9727 cause a denial of service. ([CVE-2006-2940])
9729 *Steve Henson, Bodo Moeller*
9731 * Fix ASN.1 parsing of certain invalid structures that can result
9732 in a denial of service. ([CVE-2006-2937]) [Steve Henson]
9734 * Fix buffer overflow in SSL_get_shared_ciphers() function.
9735 ([CVE-2006-3738]) [Tavis Ormandy and Will Drewry, Google Security Team]
9737 * Fix SSL client code which could crash if connecting to a
9738 malicious SSLv2 server. ([CVE-2006-4343])
9740 *Tavis Ormandy and Will Drewry, Google Security Team*
9742 * Since 0.9.8b, ciphersuite strings naming explicit ciphersuites
9743 match only those. Before that, "AES256-SHA" would be interpreted
9744 as a pattern and match "AES128-SHA" too (since AES128-SHA got
9745 the same strength classification in 0.9.7h) as we currently only
9746 have a single AES bit in the ciphersuite description bitmap.
9747 That change, however, also applied to ciphersuite strings such as
9748 "RC4-MD5" that intentionally matched multiple ciphersuites --
9749 namely, SSL 2.0 ciphersuites in addition to the more common ones
9750 from SSL 3.0/TLS 1.0.
9752 So we change the selection algorithm again: Naming an explicit
9753 ciphersuite selects this one ciphersuite, and any other similar
9754 ciphersuite (same bitmap) from *other* protocol versions.
9755 Thus, "RC4-MD5" again will properly select both the SSL 2.0
9756 ciphersuite and the SSL 3.0/TLS 1.0 ciphersuite.
9758 Since SSL 2.0 does not have any ciphersuites for which the
9759 128/256 bit distinction would be relevant, this works for now.
9760 The proper fix will be to use different bits for AES128 and
9761 AES256, which would have avoided the problems from the beginning;
9762 however, bits are scarce, so we can only do this in a new release
9763 (not just a patchlevel) when we can change the SSL_CIPHER
9764 definition to split the single 'unsigned long mask' bitmap into
9765 multiple values to extend the available space.
9769 ### Changes between 0.9.8b and 0.9.8c [05 Sep 2006]
9771 * Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
9772 ([CVE-2006-4339]) [Ben Laurie and Google Security Team]
9774 * Add AES IGE and biIGE modes.
9778 * Change the Unix randomness entropy gathering to use poll() when
9779 possible instead of select(), since the latter has some
9780 undesirable limitations.
9782 *Darryl Miles via Richard Levitte and Bodo Moeller*
9784 * Disable "ECCdraft" ciphersuites more thoroughly. Now special
9785 treatment in ssl/ssl_ciph.s makes sure that these ciphersuites
9786 cannot be implicitly activated as part of, e.g., the "AES" alias.
9787 However, please upgrade to OpenSSL 0.9.9[-dev] for
9788 non-experimental use of the ECC ciphersuites to get TLS extension
9789 support, which is required for curve and point format negotiation
9790 to avoid potential handshake problems.
9794 * Disable rogue ciphersuites:
9796 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
9797 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
9798 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
9800 The latter two were purportedly from
9801 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
9804 Also deactivate the remaining ciphersuites from
9805 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
9806 unofficial, and the ID has long expired.
9810 * Fix RSA blinding Heisenbug (problems sometimes occurred on
9811 dual-core machines) and other potential thread-safety issues.
9815 * Add the symmetric cipher Camellia (128-bit, 192-bit, 256-bit key
9816 versions), which is now available for royalty-free use
9817 (see <http://info.isl.ntt.co.jp/crypt/eng/info/chiteki.html>).
9818 Also, add Camellia TLS ciphersuites from RFC 4132.
9820 To minimize changes between patchlevels in the OpenSSL 0.9.8
9821 series, Camellia remains excluded from compilation unless OpenSSL
9822 is configured with 'enable-camellia'.
9826 * Disable the padding bug check when compression is in use. The padding
9827 bug check assumes the first packet is of even length, this is not
9828 necessarily true if compression is enabled and can result in false
9829 positives causing handshake failure. The actual bug test is ancient
9830 code so it is hoped that implementations will either have fixed it by
9831 now or any which still have the bug do not support compression.
9835 ### Changes between 0.9.8a and 0.9.8b [04 May 2006]
9837 * When applying a cipher rule check to see if string match is an explicit
9838 cipher suite and only match that one cipher suite if it is.
9842 * Link in manifests for VC++ if needed.
9844 *Austin Ziegler <halostatue@gmail.com>*
9846 * Update support for ECC-based TLS ciphersuites according to
9847 draft-ietf-tls-ecc-12.txt with proposed changes (but without
9848 TLS extensions, which are supported starting with the 0.9.9
9849 branch, not in the OpenSSL 0.9.8 branch).
9853 * New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support
9854 opaque EVP_CIPHER_CTX handling.
9858 * Fixes and enhancements to zlib compression code. We now only use
9859 "zlib1.dll" and use the default `__cdecl` calling convention on Win32
9860 to conform with the standards mentioned here:
9861 <http://www.zlib.net/DLL_FAQ.txt>
9862 Static zlib linking now works on Windows and the new --with-zlib-include
9863 --with-zlib-lib options to Configure can be used to supply the location
9864 of the headers and library. Gracefully handle case where zlib library
9869 * Several fixes and enhancements to the OID generation code. The old code
9870 sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't
9871 handle numbers larger than ULONG_MAX, truncated printing and had a
9872 non standard OBJ_obj2txt() behaviour.
9876 * Add support for building of engines under engine/ as shared libraries
9877 under VC++ build system.
9881 * Corrected the numerous bugs in the Win32 path splitter in DSO.
9882 Hopefully, we will not see any false combination of paths any more.
9886 ### Changes between 0.9.8 and 0.9.8a [11 Oct 2005]
9888 * Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
9889 (part of SSL_OP_ALL). This option used to disable the
9890 countermeasure against man-in-the-middle protocol-version
9891 rollback in the SSL 2.0 server implementation, which is a bad
9892 idea. ([CVE-2005-2969])
9894 *Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
9895 for Information Security, National Institute of Advanced Industrial
9896 Science and Technology [AIST], Japan)*
9898 * Add two function to clear and return the verify parameter flags.
9902 * Keep cipherlists sorted in the source instead of sorting them at
9903 runtime, thus removing the need for a lock.
9907 * Avoid some small subgroup attacks in Diffie-Hellman.
9909 *Nick Mathewson and Ben Laurie*
9911 * Add functions for well-known primes.
9915 * Extended Windows CE support.
9917 *Satoshi Nakamura and Andy Polyakov*
9919 * Initialize SSL_METHOD structures at compile time instead of during
9920 runtime, thus removing the need for a lock.
9924 * Make PKCS7_decrypt() work even if no certificate is supplied by
9925 attempting to decrypt each encrypted key in turn. Add support to
9930 ### Changes between 0.9.7h and 0.9.8 [05 Jul 2005]
9932 [NB: OpenSSL 0.9.7i and later 0.9.7 patch levels were released after
9935 * Add libcrypto.pc and libssl.pc for those who feel they need them.
9939 * Change CA.sh and CA.pl so they don't bundle the CSR and the private
9940 key into the same file any more.
9944 * Add initial support for Win64, both IA64 and AMD64/x64 flavors.
9948 * Add -utf8 command line and config file option to 'ca'.
9950 *Stefan <stf@udoma.org*
9952 * Removed the macro des_crypt(), as it seems to conflict with some
9953 libraries. Use DES_crypt().
9957 * Correct naming of the 'chil' and '4758cca' ENGINEs. This
9958 involves renaming the source and generated shared-libs for
9959 both. The engines will accept the corrected or legacy ids
9960 ('ncipher' and '4758_cca' respectively) when binding. NB,
9961 this only applies when building 'shared'.
9963 *Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe*
9965 * Add attribute functions to EVP_PKEY structure. Modify
9966 PKCS12_create() to recognize a CSP name attribute and
9967 use it. Make -CSP option work again in pkcs12 utility.
9971 * Add new functionality to the bn blinding code:
9972 - automatic re-creation of the BN_BLINDING parameters after
9973 a fixed number of uses (currently 32)
9974 - add new function for parameter creation
9975 - introduce flags to control the update behaviour of the
9976 BN_BLINDING parameters
9977 - hide BN_BLINDING structure
9978 Add a second BN_BLINDING slot to the RSA structure to improve
9979 performance when a single RSA object is shared among several
9984 * Add support for DTLS.
9986 *Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie*
9988 * Add support for DER encoded private keys (SSL_FILETYPE_ASN1)
9989 to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file()
9993 * Remove buggy and incomplete DH cert support from
9994 ssl/ssl_rsa.c and ssl/s3_both.c
9998 * Use SHA-1 instead of MD5 as the default digest algorithm for
9999 the `apps/openssl` commands.
10003 * Compile clean with "-Wall -Wmissing-prototypes
10004 -Wstrict-prototypes -Wmissing-declarations -Werror". Currently
10005 DEBUG_SAFESTACK must also be set.
10009 * Change ./Configure so that certain algorithms can be disabled by default.
10010 The new counterpiece to "no-xxx" is "enable-xxx".
10012 The patented RC5 and MDC2 algorithms will now be disabled unless
10013 "enable-rc5" and "enable-mdc2", respectively, are specified.
10015 (IDEA remains enabled despite being patented. This is because IDEA
10016 is frequently required for interoperability, and there is no license
10017 fee for non-commercial use. As before, "no-idea" can be used to
10018 avoid this algorithm.)
10022 * Add processing of proxy certificates (see RFC 3820). This work was
10023 sponsored by KTH (The Royal Institute of Technology in Stockholm) and
10024 EGEE (Enabling Grids for E-science in Europe).
10028 * RC4 performance overhaul on modern architectures/implementations, such
10029 as Intel P4, IA-64 and AMD64.
10033 * New utility extract-section.pl. This can be used specify an alternative
10034 section number in a pod file instead of having to treat each file as
10035 a separate case in Makefile. This can be done by adding two lines to the
10038 =for comment openssl_section:XXX
10040 The blank line is mandatory.
10044 * New arguments -certform, -keyform and -pass for s_client and s_server
10045 to allow alternative format key and certificate files and passphrase
10050 * New structure X509_VERIFY_PARAM which combines current verify parameters,
10051 update associated structures and add various utility functions.
10053 Add new policy related verify parameters, include policy checking in
10054 standard verify code. Enhance 'smime' application with extra parameters
10055 to support policy checking and print out.
10059 * Add a new engine to support VIA PadLock ACE extensions in the VIA C3
10060 Nehemiah processors. These extensions support AES encryption in hardware
10061 as well as RNG (though RNG support is currently disabled).
10063 *Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov*
10065 * Deprecate `BN_[get|set]_params()` functions (they were ignored internally).
10069 * New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented.
10071 *Andy Polyakov and a number of other people*
10073 * Improved PowerPC platform support. Most notably BIGNUM assembler
10074 implementation contributed by IBM.
10076 *Suresh Chari, Peter Waltenberg, Andy Polyakov*
10078 * The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public
10079 exponent rather than 'unsigned long'. There is a corresponding change to
10080 the new 'rsa_keygen' element of the RSA_METHOD structure.
10082 *Jelte Jansen, Geoff Thorpe*
10084 * Functionality for creating the initial serial number file is now
10085 moved from CA.pl to the 'ca' utility with a new option -create_serial.
10087 (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial
10088 number file to 1, which is bound to cause problems. To avoid
10089 the problems while respecting compatibility between different 0.9.7
10090 patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in
10091 CA.pl for serial number initialization. With the new release 0.9.8,
10092 we can fix the problem directly in the 'ca' utility.)
10096 * Reduced header interdependencies by declaring more opaque objects in
10097 ossl_typ.h. As a consequence, including some headers (eg. engine.h) will
10098 give fewer recursive includes, which could break lazy source code - so
10099 this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always,
10100 developers should define this symbol when building and using openssl to
10101 ensure they track the recommended behaviour, interfaces, [etc], but
10102 backwards-compatible behaviour prevails when this isn't defined.
10106 * New function X509_POLICY_NODE_print() which prints out policy nodes.
10110 * Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality.
10111 This will generate a random key of the appropriate length based on the
10112 cipher context. The EVP_CIPHER can provide its own random key generation
10113 routine to support keys of a specific form. This is used in the des and
10114 3des routines to generate a key of the correct parity. Update S/MIME
10115 code to use new functions and hence generate correct parity DES keys.
10116 Add EVP_CHECK_DES_KEY #define to return an error if the key is not
10117 valid (weak or incorrect parity).
10121 * Add a local set of CRLs that can be used by X509_verify_cert() as well
10122 as looking them up. This is useful when the verified structure may contain
10123 CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs
10124 present unless the new PKCS7_NO_CRL flag is asserted.
10128 * Extend ASN1 oid configuration module. It now additionally accepts the
10131 shortName = some long name, 1.2.3.4
10135 * Reimplemented the BN_CTX implementation. There is now no more static
10136 limitation on the number of variables it can handle nor the depth of the
10137 "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack
10138 information can now expand as required, and rather than having a single
10139 static array of bignums, BN_CTX now uses a linked-list of such arrays
10140 allowing it to expand on demand whilst maintaining the usefulness of
10141 BN_CTX's "bundling".
10145 * Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD
10146 to allow all RSA operations to function using a single BN_CTX.
10150 * Preliminary support for certificate policy evaluation and checking. This
10151 is initially intended to pass the tests outlined in "Conformance Testing
10152 of Relying Party Client Certificate Path Processing Logic" v1.07.
10156 * bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and
10157 remained unused and not that useful. A variety of other little bignum
10158 tweaks and fixes have also been made continuing on from the audit (see
10163 * Constify all or almost all d2i, c2i, s2i and r2i functions, along with
10164 associated ASN1, EVP and SSL functions and old ASN1 macros.
10168 * BN_zero() only needs to set 'top' and 'neg' to zero for correct results,
10169 and this should never fail. So the return value from the use of
10170 BN_set_word() (which can fail due to needless expansion) is now deprecated;
10171 if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro.
10175 * BN_CTX_get() should return zero-valued bignums, providing the same
10176 initialised value as BN_new().
10178 *Geoff Thorpe, suggested by Ulf Möller*
10180 * Support for inhibitAnyPolicy certificate extension.
10184 * An audit of the BIGNUM code is underway, for which debugging code is
10185 enabled when BN_DEBUG is defined. This makes stricter enforcements on what
10186 is considered valid when processing BIGNUMs, and causes execution to
10187 assert() when a problem is discovered. If BN_DEBUG_RAND is defined,
10188 further steps are taken to deliberately pollute unused data in BIGNUM
10189 structures to try and expose faulty code further on. For now, openssl will
10190 (in its default mode of operation) continue to tolerate the inconsistent
10191 forms that it has tolerated in the past, but authors and packagers should
10192 consider trying openssl and their own applications when compiled with
10193 these debugging symbols defined. It will help highlight potential bugs in
10194 their own code, and will improve the test coverage for OpenSSL itself. At
10195 some point, these tighter rules will become openssl's default to improve
10196 maintainability, though the assert()s and other overheads will remain only
10197 in debugging configurations. See bn.h for more details.
10199 *Geoff Thorpe, Nils Larsch, Ulf Möller*
10201 * BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure
10202 that can only be obtained through BN_CTX_new() (which implicitly
10203 initialises it). The presence of this function only made it possible
10204 to overwrite an existing structure (and cause memory leaks).
10208 * Because of the callback-based approach for implementing LHASH as a
10209 template type, lh_insert() adds opaque objects to hash-tables and
10210 lh_doall() or lh_doall_arg() are typically used with a destructor callback
10211 to clean up those corresponding objects before destroying the hash table
10212 (and losing the object pointers). So some over-zealous constifications in
10213 LHASH have been relaxed so that lh_insert() does not take (nor store) the
10214 objects as "const" and the `lh_doall[_arg]` callback wrappers are not
10215 prototyped to have "const" restrictions on the object pointers they are
10216 given (and so aren't required to cast them away any more).
10220 * The tmdiff.h API was so ugly and minimal that our own timing utility
10221 (speed) prefers to use its own implementation. The two implementations
10222 haven't been consolidated as yet (volunteers?) but the tmdiff API has had
10223 its object type properly exposed (MS_TM) instead of casting to/from
10224 `char *`. This may still change yet if someone realises MS_TM and
10226 aren't necessarily the greatest nomenclatures - but this is what was used
10227 internally to the implementation so I've used that for now.
10231 * Ensure that deprecated functions do not get compiled when
10232 OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of
10233 the self-tests were still using deprecated key-generation functions so
10234 these have been updated also.
10238 * Reorganise PKCS#7 code to separate the digest location functionality
10239 into PKCS7_find_digest(), digest addition into PKCS7_bio_add_digest().
10240 New function PKCS7_set_digest() to set the digest type for PKCS#7
10241 digestedData type. Add additional code to correctly generate the
10242 digestedData type and add support for this type in PKCS7 initialization
10247 * New function PKCS7_set0_type_other() this initializes a PKCS7
10248 structure of type "other".
10252 * Fix prime generation loop in crypto/bn/bn_prime.pl by making
10253 sure the loop does correctly stop and breaking ("division by zero")
10254 modulus operations are not performed. The (pre-generated) prime
10255 table crypto/bn/bn_prime.h was already correct, but it could not be
10256 re-generated on some platforms because of the "division by zero"
10257 situation in the script.
10259 *Ralf S. Engelschall*
10261 * Update support for ECC-based TLS ciphersuites according to
10262 draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with
10263 SHA-1 now is only used for "small" curves (where the
10264 representation of a field element takes up to 24 bytes); for
10265 larger curves, the field element resulting from ECDH is directly
10266 used as premaster secret.
10268 *Douglas Stebila (Sun Microsystems Laboratories)*
10270 * Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2
10271 curve secp160r1 to the tests.
10273 *Douglas Stebila (Sun Microsystems Laboratories)*
10275 * Add the possibility to load symbols globally with DSO.
10277 *Götz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte*
10279 * Add the functions ERR_set_mark() and ERR_pop_to_mark() for better
10280 control of the error stack.
10284 * Add support for STORE in ENGINE.
10288 * Add the STORE type. The intention is to provide a common interface
10289 to certificate and key stores, be they simple file-based stores, or
10290 HSM-type store, or LDAP stores, or...
10291 NOTE: The code is currently UNTESTED and isn't really used anywhere.
10295 * Add a generic structure called OPENSSL_ITEM. This can be used to
10296 pass a list of arguments to any function as well as provide a way
10297 for a function to pass data back to the caller.
10301 * Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup()
10302 works like BUF_strdup() but can be used to duplicate a portion of
10303 a string. The copy gets NUL-terminated. BUF_memdup() duplicates
10308 * Add the function sk_find_ex() which works like sk_find(), but will
10309 return an index to an element even if an exact match couldn't be
10310 found. The index is guaranteed to point at the element where the
10311 searched-for key would be inserted to preserve sorting order.
10315 * Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but
10316 takes an extra flags argument for optional functionality. Currently,
10317 the following flags are defined:
10319 OBJ_BSEARCH_VALUE_ON_NOMATCH
10320 This one gets OBJ_bsearch_ex() to return a pointer to the first
10321 element where the comparing function returns a negative or zero
10324 OBJ_BSEARCH_FIRST_VALUE_ON_MATCH
10325 This one gets OBJ_bsearch_ex() to return a pointer to the first
10326 element where the comparing function returns zero. This is useful
10327 if there are more than one element where the comparing function
10332 * Make it possible to create self-signed certificates with 'openssl ca'
10333 in such a way that the self-signed certificate becomes part of the
10334 CA database and uses the same mechanisms for serial number generation
10335 as all other certificate signing. The new flag '-selfsign' enables
10336 this functionality. Adapt CA.sh and CA.pl.in.
10340 * Add functionality to check the public key of a certificate request
10341 against a given private. This is useful to check that a certificate
10342 request can be signed by that key (self-signing).
10346 * Make it possible to have multiple active certificates with the same
10347 subject in the CA index file. This is done only if the keyword
10348 'unique_subject' is set to 'no' in the main CA section (default
10349 if 'CA_default') of the configuration file. The value is saved
10350 with the database itself in a separate index attribute file,
10351 named like the index file with '.attr' appended to the name.
10355 * Generate multi-valued AVAs using '+' notation in config files for
10360 * Support for nameConstraints certificate extension.
10364 * Support for policyConstraints certificate extension.
10368 * Support for policyMappings certificate extension.
10372 * Make sure the default DSA_METHOD implementation only uses its
10373 dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL,
10374 and change its own handlers to be NULL so as to remove unnecessary
10375 indirection. This lets alternative implementations fallback to the
10376 default implementation more easily.
10380 * Support for directoryName in GeneralName related extensions
10385 * Make it possible to link applications using Makefile.shared.
10386 Make that possible even when linking against static libraries!
10390 * Support for single pass processing for S/MIME signing. This now
10391 means that S/MIME signing can be done from a pipe, in addition
10392 cleartext signing (multipart/signed type) is effectively streaming
10393 and the signed data does not need to be all held in memory.
10395 This is done with a new flag PKCS7_STREAM. When this flag is set
10396 PKCS7_sign() only initializes the PKCS7 structure and the actual signing
10397 is done after the data is output (and digests calculated) in
10398 SMIME_write_PKCS7().
10402 * Add full support for -rpath/-R, both in shared libraries and
10403 applications, at least on the platforms where it's known how
10408 * In crypto/ec/ec_mult.c, implement fast point multiplication with
10409 precomputation, based on wNAF splitting: EC_GROUP_precompute_mult()
10410 will now compute a table of multiples of the generator that
10411 makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul()
10412 faster (notably in the case of a single point multiplication,
10413 scalar * generator).
10415 *Nils Larsch, Bodo Moeller*
10417 * IPv6 support for certificate extensions. The various extensions
10418 which use the IP:a.b.c.d can now take IPv6 addresses using the
10419 formats of RFC1884 2.2 . IPv6 addresses are now also displayed
10424 * Added an ENGINE that implements RSA by performing private key
10425 exponentiations with the GMP library. The conversions to and from
10426 GMP's mpz_t format aren't optimised nor are any montgomery forms
10427 cached, and on x86 it appears OpenSSL's own performance has caught up.
10428 However there are likely to be other architectures where GMP could
10429 provide a boost. This ENGINE is not built in by default, but it can be
10430 specified at Configure time and should be accompanied by the necessary
10431 linker additions, eg;
10432 ./config -DOPENSSL_USE_GMP -lgmp
10436 * "openssl engine" will not display ENGINE/DSO load failure errors when
10437 testing availability of engines with "-t" - the old behaviour is
10438 produced by increasing the feature's verbosity with "-tt".
10442 * ECDSA routines: under certain error conditions uninitialized BN objects
10443 could be freed. Solution: make sure initialization is performed early
10444 enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de>
10449 * Key-generation can now be implemented in RSA_METHOD, DSA_METHOD
10450 and DH_METHOD (eg. by ENGINE implementations) to override the normal
10451 software implementations. For DSA and DH, parameter generation can
10452 also be overridden by providing the appropriate method callbacks.
10456 * Change the "progress" mechanism used in key-generation and
10457 primality testing to functions that take a new BN_GENCB pointer in
10458 place of callback/argument pairs. The new API functions have `_ex`
10459 postfixes and the older functions are reimplemented as wrappers for
10460 the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide
10461 declarations of the old functions to help (graceful) attempts to
10462 migrate to the new functions. Also, the new key-generation API
10463 functions operate on a caller-supplied key-structure and return
10464 success/failure rather than returning a key or NULL - this is to
10465 help make "keygen" another member function of RSA_METHOD etc.
10467 Example for using the new callback interface:
10469 int (*my_callback)(int a, int b, BN_GENCB *cb) = ...;
10470 void *my_arg = ...;
10473 BN_GENCB_set(&my_cb, my_callback, my_arg);
10475 return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb);
10476 /* For the meaning of a, b in calls to my_callback(), see the
10477 * documentation of the function that calls the callback.
10478 * cb will point to my_cb; my_arg can be retrieved as cb->arg.
10479 * my_callback should return 1 if it wants BN_is_prime_ex()
10480 * to continue, or 0 to stop.
10485 * Change the ZLIB compression method to be stateful, and make it
10486 available to TLS with the number defined in
10487 draft-ietf-tls-compression-04.txt.
10491 * Add the ASN.1 structures and functions for CertificatePair, which
10492 is defined as follows (according to X.509_4thEditionDraftV6.pdf):
10494 CertificatePair ::= SEQUENCE {
10495 forward [0] Certificate OPTIONAL,
10496 reverse [1] Certificate OPTIONAL,
10497 -- at least one of the pair shall be present -- }
10499 Also implement the PEM functions to read and write certificate
10500 pairs, and defined the PEM tag as "CERTIFICATE PAIR".
10502 This needed to be defined, mostly for the sake of the LDAP
10503 attribute crossCertificatePair, but may prove useful elsewhere as
10508 * Make it possible to inhibit symlinking of shared libraries in
10509 Makefile.shared, for Cygwin's sake.
10513 * Extend the BIGNUM API by creating a function
10514 void BN_set_negative(BIGNUM *a, int neg);
10515 and a macro that behave like
10516 int BN_is_negative(const BIGNUM *a);
10518 to avoid the need to access 'a->neg' directly in applications.
10522 * Implement fast modular reduction for pseudo-Mersenne primes
10523 used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c).
10524 EC_GROUP_new_curve_GFp() will now automatically use this
10527 *Nils Larsch <nla@trustcenter.de>*
10529 * Add new lock type (CRYPTO_LOCK_BN).
10533 * Change the ENGINE framework to automatically load engines
10534 dynamically from specific directories unless they could be
10535 found to already be built in or loaded. Move all the
10536 current engines except for the cryptodev one to a new
10537 directory engines/.
10538 The engines in engines/ are built as shared libraries if
10539 the "shared" options was given to ./Configure or ./config.
10540 Otherwise, they are inserted in libcrypto.a.
10541 /usr/local/ssl/engines is the default directory for dynamic
10542 engines, but that can be overridden at configure time through
10543 the usual use of --prefix and/or --openssldir, and at run
10544 time with the environment variable OPENSSL_ENGINES.
10546 *Geoff Thorpe and Richard Levitte*
10548 * Add Makefile.shared, a helper makefile to build shared
10549 libraries. Adapt Makefile.org.
10553 * Add version info to Win32 DLLs.
10555 *Peter 'Luna' Runestig" <peter@runestig.com>*
10557 * Add new 'medium level' PKCS#12 API. Certificates and keys
10558 can be added using this API to created arbitrary PKCS#12
10559 files while avoiding the low-level API.
10561 New options to PKCS12_create(), key or cert can be NULL and
10562 will then be omitted from the output file. The encryption
10563 algorithm NIDs can be set to -1 for no encryption, the mac
10564 iteration count can be set to 0 to omit the mac.
10566 Enhance pkcs12 utility by making the -nokeys and -nocerts
10567 options work when creating a PKCS#12 file. New option -nomac
10568 to omit the mac, NONE can be set for an encryption algorithm.
10569 New code is modified to use the enhanced PKCS12_create()
10570 instead of the low-level API.
10574 * Extend ASN1 encoder to support indefinite length constructed
10575 encoding. This can output sequences tags and octet strings in
10576 this form. Modify pk7_asn1.c to support indefinite length
10577 encoding. This is experimental and needs additional code to
10578 be useful, such as an ASN1 bio and some enhanced streaming
10581 Extend template encode functionality so that tagging is passed
10582 down to the template encoder.
10586 * Let 'openssl req' fail if an argument to '-newkey' is not
10587 recognized instead of using RSA as a default.
10591 * Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt.
10592 As these are not official, they are not included in "ALL";
10593 the "ECCdraft" ciphersuite group alias can be used to select them.
10595 *Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)*
10597 * Add ECDH engine support.
10599 *Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)*
10601 * Add ECDH in new directory crypto/ecdh/.
10603 *Douglas Stebila (Sun Microsystems Laboratories)*
10605 * Let BN_rand_range() abort with an error after 100 iterations
10606 without success (which indicates a broken PRNG).
10610 * Change BN_mod_sqrt() so that it verifies that the input value
10611 is really the square of the return value. (Previously,
10612 BN_mod_sqrt would show GIGO behaviour.)
10616 * Add named elliptic curves over binary fields from X9.62, SECG,
10617 and WAP/WTLS; add OIDs that were still missing.
10619 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10621 * Extend the EC library for elliptic curves over binary fields
10622 (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/).
10625 EC_GF2m_simple_method
10629 EC_GROUP_new_curve_GF2m
10630 EC_GROUP_set_curve_GF2m
10631 EC_GROUP_get_curve_GF2m
10632 EC_POINT_set_affine_coordinates_GF2m
10633 EC_POINT_get_affine_coordinates_GF2m
10634 EC_POINT_set_compressed_coordinates_GF2m
10636 Point compression for binary fields is disabled by default for
10637 patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to
10640 As binary polynomials are represented as BIGNUMs, various members
10641 of the EC_GROUP and EC_POINT data structures can be shared
10642 between the implementations for prime fields and binary fields;
10643 the above `..._GF2m functions` (except for EX_GROUP_new_curve_GF2m)
10644 are essentially identical to their `..._GFp` counterparts.
10645 (For simplicity, the `..._GFp` prefix has been dropped from
10646 various internal method names.)
10648 An internal 'field_div' method (similar to 'field_mul' and
10649 'field_sqr') has been added; this is used only for binary fields.
10651 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10653 * Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult()
10654 through methods ('mul', 'precompute_mult').
10656 The generic implementations (now internally called 'ec_wNAF_mul'
10657 and 'ec_wNAF_precomputed_mult') remain the default if these
10658 methods are undefined.
10660 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10662 * New function EC_GROUP_get_degree, which is defined through
10663 EC_METHOD. For curves over prime fields, this returns the bit
10664 length of the modulus.
10666 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10668 * New functions EC_GROUP_dup, EC_POINT_dup.
10669 (These simply call ..._new and ..._copy).
10671 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10673 * Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c.
10674 Polynomials are represented as BIGNUMs (where the sign bit is not
10675 used) in the following functions [macros]:
10678 BN_GF2m_sub [= BN_GF2m_add]
10679 BN_GF2m_mod [wrapper for BN_GF2m_mod_arr]
10680 BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr]
10681 BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr]
10683 BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr]
10684 BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr]
10685 BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr]
10686 BN_GF2m_cmp [= BN_ucmp]
10688 (Note that only the 'mod' functions are actually for fields GF(2^m).
10689 BN_GF2m_add() is misnomer, but this is for the sake of consistency.)
10691 For some functions, an the irreducible polynomial defining a
10692 field can be given as an 'unsigned int[]' with strictly
10693 decreasing elements giving the indices of those bits that are set;
10694 i.e., p[] represents the polynomial
10695 f(t) = t^p[0] + t^p[1] + ... + t^p[k]
10697 p[0] > p[1] > ... > p[k] = 0.
10698 This applies to the following functions:
10701 BN_GF2m_mod_mul_arr
10702 BN_GF2m_mod_sqr_arr
10703 BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv]
10704 BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div]
10705 BN_GF2m_mod_exp_arr
10706 BN_GF2m_mod_sqrt_arr
10707 BN_GF2m_mod_solve_quad_arr
10711 Conversion can be performed by the following functions:
10716 bntest.c has additional tests for binary polynomial arithmetic.
10718 Two implementations for BN_GF2m_mod_div() are available.
10719 The default algorithm simply uses BN_GF2m_mod_inv() and
10720 BN_GF2m_mod_mul(). The alternative algorithm is compiled in only
10721 if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the
10722 copyright notice in crypto/bn/bn_gf2m.c before enabling it).
10724 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10726 * Add new error code 'ERR_R_DISABLED' that can be used when some
10727 functionality is disabled at compile-time.
10729 *Douglas Stebila <douglas.stebila@sun.com>*
10731 * Change default behaviour of 'openssl asn1parse' so that more
10732 information is visible when viewing, e.g., a certificate:
10734 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump'
10735 mode the content of non-printable OCTET STRINGs is output in a
10736 style similar to INTEGERs, but with '[HEX DUMP]' prepended to
10737 avoid the appearance of a printable string.
10739 *Nils Larsch <nla@trustcenter.de>*
10741 * Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access
10743 EC_GROUP_set_asn1_flag()
10744 EC_GROUP_get_asn1_flag()
10745 EC_GROUP_set_point_conversion_form()
10746 EC_GROUP_get_point_conversion_form()
10747 These control ASN1 encoding details:
10748 - Curves (i.e., groups) are encoded explicitly unless asn1_flag
10749 has been set to OPENSSL_EC_NAMED_CURVE.
10750 - Points are encoded in uncompressed form by default; options for
10751 asn1_for are as for point2oct, namely
10752 POINT_CONVERSION_COMPRESSED
10753 POINT_CONVERSION_UNCOMPRESSED
10754 POINT_CONVERSION_HYBRID
10756 Also add 'seed' and 'seed_len' members to EC_GROUP with access
10758 EC_GROUP_set_seed()
10759 EC_GROUP_get0_seed()
10760 EC_GROUP_get_seed_len()
10761 This is used only for ASN1 purposes (so far).
10763 *Nils Larsch <nla@trustcenter.de>*
10765 * Add 'field_type' member to EC_METHOD, which holds the NID
10766 of the appropriate field type OID. The new function
10767 EC_METHOD_get_field_type() returns this value.
10769 *Nils Larsch <nla@trustcenter.de>*
10772 EC_POINT_point2bn()
10773 EC_POINT_bn2point()
10774 EC_POINT_point2hex()
10775 EC_POINT_hex2point()
10776 providing useful interfaces to EC_POINT_point2oct() and
10777 EC_POINT_oct2point().
10779 *Nils Larsch <nla@trustcenter.de>*
10781 * Change internals of the EC library so that the functions
10782 EC_GROUP_set_generator()
10783 EC_GROUP_get_generator()
10784 EC_GROUP_get_order()
10785 EC_GROUP_get_cofactor()
10786 are implemented directly in crypto/ec/ec_lib.c and not dispatched
10787 to methods, which would lead to unnecessary code duplication when
10788 adding different types of curves.
10790 *Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller*
10792 * Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM
10793 arithmetic, and such that modified wNAFs are generated
10794 (which avoid length expansion in many cases).
10798 * Add a function EC_GROUP_check_discriminant() (defined via
10799 EC_METHOD) that verifies that the curve discriminant is non-zero.
10801 Add a function EC_GROUP_check() that makes some sanity tests
10802 on a EC_GROUP, its generator and order. This includes
10803 EC_GROUP_check_discriminant().
10805 *Nils Larsch <nla@trustcenter.de>*
10807 * Add ECDSA in new directory crypto/ecdsa/.
10809 Add applications 'openssl ecparam' and 'openssl ecdsa'
10810 (these are based on 'openssl dsaparam' and 'openssl dsa').
10812 ECDSA support is also included in various other files across the
10813 library. Most notably,
10814 - 'openssl req' now has a '-newkey ecdsa:file' option;
10815 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA;
10816 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and
10817 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make
10818 them suitable for ECDSA where domain parameters must be
10819 extracted before the specific public key;
10820 - ECDSA engine support has been added.
10822 *Nils Larsch <nla@trustcenter.de>*
10824 * Include some named elliptic curves, and add OIDs from X9.62,
10825 SECG, and WAP/WTLS. Each curve can be obtained from the new
10827 EC_GROUP_new_by_curve_name(),
10828 and the list of available named curves can be obtained with
10829 EC_get_builtin_curves().
10830 Also add a 'curve_name' member to EC_GROUP objects, which can be
10832 EC_GROUP_set_curve_name()
10833 EC_GROUP_get_curve_name()
10835 *Nils Larsch <larsch@trustcenter.de, Bodo Moeller*
10837 * Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
10838 was actually never needed) and in BN_mul(). The removal in BN_mul()
10839 required a small change in bn_mul_part_recursive() and the addition
10840 of the functions bn_cmp_part_words(), bn_sub_part_words() and
10841 bn_add_part_words(), which do the same thing as bn_cmp_words(),
10842 bn_sub_words() and bn_add_words() except they take arrays with
10847 ### Changes between 0.9.7l and 0.9.7m [23 Feb 2007]
10849 * Cleanse PEM buffers before freeing them since they may contain
10852 *Benjamin Bennett <ben@psc.edu>*
10854 * Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
10855 a ciphersuite string such as "DEFAULT:RSA" cannot enable
10856 authentication-only ciphersuites.
10860 * Since AES128 and AES256 share a single mask bit in the logic of
10861 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
10862 kludge to work properly if AES128 is available and AES256 isn't.
10866 * Expand security boundary to match 1.1.1 module.
10870 * Remove redundant features: hash file source, editing of test vectors
10871 modify fipsld to use external fips_premain.c signature.
10875 * New perl script mkfipsscr.pl to create shell scripts or batch files to
10876 run algorithm test programs.
10880 * Make algorithm test programs more tolerant of whitespace.
10884 * Have SSL/TLS server implementation tolerate "mismatched" record
10885 protocol version while receiving ClientHello even if the
10886 ClientHello is fragmented. (The server can't insist on the
10887 particular protocol version it has chosen before the ServerHello
10888 message has informed the client about his choice.)
10892 * Load error codes if they are not already present instead of using a
10893 static variable. This allows them to be cleanly unloaded and reloaded.
10897 ### Changes between 0.9.7k and 0.9.7l [28 Sep 2006]
10899 * Introduce limits to prevent malicious keys being able to
10900 cause a denial of service. ([CVE-2006-2940])
10902 *Steve Henson, Bodo Moeller*
10904 * Fix ASN.1 parsing of certain invalid structures that can result
10905 in a denial of service. ([CVE-2006-2937]) [Steve Henson]
10907 * Fix buffer overflow in SSL_get_shared_ciphers() function.
10908 ([CVE-2006-3738]) [Tavis Ormandy and Will Drewry, Google Security Team]
10910 * Fix SSL client code which could crash if connecting to a
10911 malicious SSLv2 server. ([CVE-2006-4343])
10913 *Tavis Ormandy and Will Drewry, Google Security Team*
10915 * Change ciphersuite string processing so that an explicit
10916 ciphersuite selects this one ciphersuite (so that "AES256-SHA"
10917 will no longer include "AES128-SHA"), and any other similar
10918 ciphersuite (same bitmap) from *other* protocol versions (so that
10919 "RC4-MD5" will still include both the SSL 2.0 ciphersuite and the
10920 SSL 3.0/TLS 1.0 ciphersuite). This is a backport combining
10921 changes from 0.9.8b and 0.9.8d.
10925 ### Changes between 0.9.7j and 0.9.7k [05 Sep 2006]
10927 * Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
10928 ([CVE-2006-4339]) [Ben Laurie and Google Security Team]
10930 * Change the Unix randomness entropy gathering to use poll() when
10931 possible instead of select(), since the latter has some
10932 undesirable limitations.
10934 *Darryl Miles via Richard Levitte and Bodo Moeller*
10936 * Disable rogue ciphersuites:
10938 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
10939 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
10940 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
10942 The latter two were purportedly from
10943 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
10946 Also deactivate the remaining ciphersuites from
10947 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
10948 unofficial, and the ID has long expired.
10952 * Fix RSA blinding Heisenbug (problems sometimes occurred on
10953 dual-core machines) and other potential thread-safety issues.
10957 ### Changes between 0.9.7i and 0.9.7j [04 May 2006]
10959 * Adapt fipsld and the build system to link against the validated FIPS
10960 module in FIPS mode.
10964 * Fixes for VC++ 2005 build under Windows.
10968 * Add new Windows build target VC-32-GMAKE for VC++. This uses GNU make
10969 from a Windows bash shell such as MSYS. It is autodetected from the
10970 "config" script when run from a VC++ environment. Modify standard VC++
10971 build to use fipscanister.o from the GNU make build.
10975 ### Changes between 0.9.7h and 0.9.7i [14 Oct 2005]
10977 * Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS.
10978 The value now differs depending on if you build for FIPS or not.
10979 BEWARE! A program linked with a shared FIPSed libcrypto can't be
10980 safely run with a non-FIPSed libcrypto, as it may crash because of
10981 the difference induced by this change.
10985 ### Changes between 0.9.7g and 0.9.7h [11 Oct 2005]
10987 * Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
10988 (part of SSL_OP_ALL). This option used to disable the
10989 countermeasure against man-in-the-middle protocol-version
10990 rollback in the SSL 2.0 server implementation, which is a bad
10991 idea. ([CVE-2005-2969])
10993 *Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
10994 for Information Security, National Institute of Advanced Industrial
10995 Science and Technology [AIST, Japan)]*
10997 * Minimal support for X9.31 signatures and PSS padding modes. This is
10998 mainly for FIPS compliance and not fully integrated at this stage.
11002 * For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform
11003 the exponentiation using a fixed-length exponent. (Otherwise,
11004 the information leaked through timing could expose the secret key
11005 after many signatures; cf. Bleichenbacher's attack on DSA with
11010 * Make a new fixed-window mod_exp implementation the default for
11011 RSA, DSA, and DH private-key operations so that the sequence of
11012 squares and multiplies and the memory access pattern are
11013 independent of the particular secret key. This will mitigate
11014 cache-timing and potential related attacks.
11016 BN_mod_exp_mont_consttime() is the new exponentiation implementation,
11017 and this is automatically used by BN_mod_exp_mont() if the new flag
11018 BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH
11019 will use this BN flag for private exponents unless the flag
11020 RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or
11021 DH_FLAG_NO_EXP_CONSTTIME, respectively, is set.
11023 *Matthew D Wood (Intel Corp), with some changes by Bodo Moeller*
11025 * Change the client implementation for SSLv23_method() and
11026 SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0
11027 Client Hello message format if the SSL_OP_NO_SSLv2 option is set.
11028 (Previously, the SSL 2.0 backwards compatible Client Hello
11029 message format would be used even with SSL_OP_NO_SSLv2.)
11033 * Add support for smime-type MIME parameter in S/MIME messages which some
11038 * New function BN_MONT_CTX_set_locked() to set montgomery parameters in
11039 a threadsafe manner. Modify rsa code to use new function and add calls
11040 to dsa and dh code (which had race conditions before).
11044 * Include the fixed error library code in the C error file definitions
11045 instead of fixing them up at runtime. This keeps the error code
11046 structures constant.
11050 ### Changes between 0.9.7f and 0.9.7g [11 Apr 2005]
11052 [NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after
11055 * Fixes for newer kerberos headers. NB: the casts are needed because
11056 the 'length' field is signed on one version and unsigned on another
11057 with no (?) obvious way to tell the difference, without these VC++
11058 complains. Also the "definition" of FAR (blank) is no longer included
11059 nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up
11060 some needed definitions.
11064 * Undo Cygwin change.
11068 * Added support for proxy certificates according to RFC 3820.
11069 Because they may be a security thread to unaware applications,
11070 they must be explicitly allowed in run-time. See
11071 docs/HOWTO/proxy_certificates.txt for further information.
11075 ### Changes between 0.9.7e and 0.9.7f [22 Mar 2005]
11077 * Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating
11078 server and client random values. Previously
11079 (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in
11080 less random data when sizeof(time_t) > 4 (some 64 bit platforms).
11082 This change has negligible security impact because:
11084 1. Server and client random values still have 24 bytes of pseudo random
11087 2. Server and client random values are sent in the clear in the initial
11090 3. The master secret is derived using the premaster secret (48 bytes in
11091 size for static RSA ciphersuites) as well as client server and random
11094 The OpenSSL team would like to thank the UK NISCC for bringing this issue
11097 *Stephen Henson, reported by UK NISCC*
11099 * Use Windows randomness collection on Cygwin.
11103 * Fix hang in EGD/PRNGD query when communication socket is closed
11104 prematurely by EGD/PRNGD.
11106 *Darren Tucker <dtucker@zip.com.au> via Lutz Jänicke, resolves #1014*
11108 * Prompt for pass phrases when appropriate for PKCS12 input format.
11112 * Back-port of selected performance improvements from development
11113 branch, as well as improved support for PowerPC platforms.
11117 * Add lots of checks for memory allocation failure, error codes to indicate
11118 failure and freeing up memory if a failure occurs.
11120 *Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson*
11122 * Add new -passin argument to dgst.
11126 * Perform some character comparisons of different types in X509_NAME_cmp:
11127 this is needed for some certificates that re-encode DNs into UTF8Strings
11128 (in violation of RFC3280) and can't or won't issue name rollover
11133 * Make an explicit check during certificate validation to see that
11134 the CA setting in each certificate on the chain is correct. As a
11135 side effect always do the following basic checks on extensions,
11136 not just when there's an associated purpose to the check:
11138 - if there is an unhandled critical extension (unless the user
11139 has chosen to ignore this fault)
11140 - if the path length has been exceeded (if one is set at all)
11141 - that certain extensions fit the associated purpose (if one has
11146 ### Changes between 0.9.7d and 0.9.7e [25 Oct 2004]
11148 * Avoid a race condition when CRLs are checked in a multi threaded
11149 environment. This would happen due to the reordering of the revoked
11150 entries during signature checking and serial number lookup. Now the
11151 encoding is cached and the serial number sort performed under a lock.
11152 Add new STACK function sk_is_sorted().
11156 * Add Delta CRL to the extension code.
11160 * Various fixes to s3_pkt.c so alerts are sent properly.
11162 *David Holmes <d.holmes@f5.com>*
11164 * Reduce the chances of duplicate issuer name and serial numbers (in
11165 violation of RFC3280) using the OpenSSL certificate creation utilities.
11166 This is done by creating a random 64 bit value for the initial serial
11167 number when a serial number file is created or when a self signed
11168 certificate is created using 'openssl req -x509'. The initial serial
11169 number file is created using 'openssl x509 -next_serial' in CA.pl
11170 rather than being initialized to 1.
11174 ### Changes between 0.9.7c and 0.9.7d [17 Mar 2004]
11176 * Fix null-pointer assignment in do_change_cipher_spec() revealed
11177 by using the Codenomicon TLS Test Tool ([CVE-2004-0079])
11179 *Joe Orton, Steve Henson*
11181 * Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites
11184 *Joe Orton, Steve Henson*
11186 * Make it possible to have multiple active certificates with the same
11187 subject in the CA index file. This is done only if the keyword
11188 'unique_subject' is set to 'no' in the main CA section (default
11189 if 'CA_default') of the configuration file. The value is saved
11190 with the database itself in a separate index attribute file,
11191 named like the index file with '.attr' appended to the name.
11195 * X509 verify fixes. Disable broken certificate workarounds when
11196 X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if
11197 keyUsage extension present. Don't accept CRLs with unhandled critical
11198 extensions: since verify currently doesn't process CRL extensions this
11199 rejects a CRL with *any* critical extensions. Add new verify error codes
11204 * When creating an OCSP nonce use an OCTET STRING inside the extnValue.
11205 A clarification of RFC2560 will require the use of OCTET STRINGs and
11206 some implementations cannot handle the current raw format. Since OpenSSL
11207 copies and compares OCSP nonces as opaque blobs without any attempt at
11208 parsing them this should not create any compatibility issues.
11212 * New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when
11213 calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without
11214 this HMAC (and other) operations are several times slower than OpenSSL
11219 * Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex().
11221 *Peter Sylvester <Peter.Sylvester@EdelWeb.fr>*
11223 * Use the correct content when signing type "other".
11227 ### Changes between 0.9.7b and 0.9.7c [30 Sep 2003]
11229 * Fix various bugs revealed by running the NISCC test suite:
11231 Stop out of bounds reads in the ASN1 code when presented with
11232 invalid tags (CVE-2003-0543 and CVE-2003-0544).
11234 Free up ASN1_TYPE correctly if ANY type is invalid ([CVE-2003-0545]).
11236 If verify callback ignores invalid public key errors don't try to check
11237 certificate signature with the NULL public key.
11241 * New -ignore_err option in ocsp application to stop the server
11242 exiting on the first error in a request.
11246 * In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
11247 if the server requested one: as stated in TLS 1.0 and SSL 3.0
11252 * In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
11253 extra data after the compression methods not only for TLS 1.0
11254 but also for SSL 3.0 (as required by the specification).
11256 *Bodo Moeller; problem pointed out by Matthias Loepfe*
11258 * Change X509_certificate_type() to mark the key as exported/exportable
11259 when it's 512 *bits* long, not 512 bytes.
11263 * Change AES_cbc_encrypt() so it outputs exact multiple of
11264 blocks during encryption.
11268 * Various fixes to base64 BIO and non blocking I/O. On write
11269 flushes were not handled properly if the BIO retried. On read
11270 data was not being buffered properly and had various logic bugs.
11271 This also affects blocking I/O when the data being decoded is a
11276 * Various S/MIME bugfixes and compatibility changes:
11277 output correct application/pkcs7 MIME type if
11278 PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures.
11279 Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening
11280 of files as .eml work). Correctly handle very long lines in MIME
11285 ### Changes between 0.9.7a and 0.9.7b [10 Apr 2003]
11287 * Countermeasure against the Klima-Pokorny-Rosa extension of
11288 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
11289 a protocol version number mismatch like a decryption error
11290 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
11294 * Turn on RSA blinding by default in the default implementation
11295 to avoid a timing attack. Applications that don't want it can call
11296 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
11297 They would be ill-advised to do so in most cases.
11299 *Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller*
11301 * Change RSA blinding code so that it works when the PRNG is not
11302 seeded (in this case, the secret RSA exponent is abused as
11303 an unpredictable seed -- if it is not unpredictable, there
11304 is no point in blinding anyway). Make RSA blinding thread-safe
11305 by remembering the creator's thread ID in rsa->blinding and
11306 having all other threads use local one-time blinding factors
11307 (this requires more computation than sharing rsa->blinding, but
11308 avoids excessive locking; and if an RSA object is not shared
11309 between threads, blinding will still be very fast).
11313 * Fixed a typo bug that would cause ENGINE_set_default() to set an
11314 ENGINE as defaults for all supported algorithms irrespective of
11315 the 'flags' parameter. 'flags' is now honoured, so applications
11316 should make sure they are passing it correctly.
11320 * Target "mingw" now allows native Windows code to be generated in
11321 the Cygwin environment as well as with the MinGW compiler.
11325 ### Changes between 0.9.7 and 0.9.7a [19 Feb 2003]
11327 * In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
11328 via timing by performing a MAC computation even if incorrect
11329 block cipher padding has been found. This is a countermeasure
11330 against active attacks where the attacker has to distinguish
11331 between bad padding and a MAC verification error. ([CVE-2003-0078])
11333 *Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
11334 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
11335 Martin Vuagnoux (EPFL, Ilion)*
11337 * Make the no-err option work as intended. The intention with no-err
11338 is not to have the whole error stack handling routines removed from
11339 libcrypto, it's only intended to remove all the function name and
11340 reason texts, thereby removing some of the footprint that may not
11341 be interesting if those errors aren't displayed anyway.
11343 NOTE: it's still possible for any application or module to have its
11344 own set of error texts inserted. The routines are there, just not
11345 used by default when no-err is given.
11349 * Add support for FreeBSD on IA64.
11351 *dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454*
11353 * Adjust DES_cbc_cksum() so it returns the same value as the MIT
11354 Kerberos function mit_des_cbc_cksum(). Before this change,
11355 the value returned by DES_cbc_cksum() was like the one from
11356 mit_des_cbc_cksum(), except the bytes were swapped.
11358 *Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte*
11360 * Allow an application to disable the automatic SSL chain building.
11361 Before this a rather primitive chain build was always performed in
11362 ssl3_output_cert_chain(): an application had no way to send the
11363 correct chain if the automatic operation produced an incorrect result.
11365 Now the chain builder is disabled if either:
11367 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert().
11369 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set.
11371 The reasoning behind this is that an application would not want the
11372 auto chain building to take place if extra chain certificates are
11373 present and it might also want a means of sending no additional
11374 certificates (for example the chain has two certificates and the
11379 * Add the possibility to build without the ENGINE framework.
11381 *Steven Reddie <smr@essemer.com.au> via Richard Levitte*
11383 * Under Win32 gmtime() can return NULL: check return value in
11384 OPENSSL_gmtime(). Add error code for case where gmtime() fails.
11388 * DSA routines: under certain error conditions uninitialized BN objects
11389 could be freed. Solution: make sure initialization is performed early
11390 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>,
11391 Nils Larsch <nla@trustcenter.de> via PR#459)
11395 * Another fix for SSLv2 session ID handling: the session ID was incorrectly
11396 checked on reconnect on the client side, therefore session resumption
11397 could still fail with a "ssl session id is different" error. This
11398 behaviour is masked when SSL_OP_ALL is used due to
11399 SSL_OP_MICROSOFT_SESS_ID_BUG being set.
11400 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
11401 followup to PR #377.
11405 * IA-32 assembler support enhancements: unified ELF targets, support
11406 for SCO/Caldera platforms, fix for Cygwin shared build.
11410 * Add support for FreeBSD on sparc64. As a consequence, support for
11411 FreeBSD on non-x86 processors is separate from x86 processors on
11412 the config script, much like the NetBSD support.
11414 *Richard Levitte & Kris Kennaway <kris@obsecurity.org>*
11416 ### Changes between 0.9.6h and 0.9.7 [31 Dec 2002]
11418 [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after
11421 * Fix session ID handling in SSLv2 client code: the SERVER FINISHED
11422 code (06) was taken as the first octet of the session ID and the last
11423 octet was ignored consequently. As a result SSLv2 client side session
11424 caching could not have worked due to the session ID mismatch between
11426 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
11431 * Change the declaration of needed Kerberos libraries to use EX_LIBS
11432 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is
11437 * The hw_ncipher.c engine requires dynamic locks. Unfortunately, it
11438 seems that in spite of existing for more than a year, many application
11439 author have done nothing to provide the necessary callbacks, which
11440 means that this particular engine will not work properly anywhere.
11441 This is a very unfortunate situation which forces us, in the name
11442 of usability, to give the hw_ncipher.c a static lock, which is part
11444 NOTE: This is for the 0.9.7 series ONLY. This hack will never
11445 appear in 0.9.8 or later. We EXPECT application authors to have
11446 dealt properly with this when 0.9.8 is released (unless we actually
11447 make such changes in the libcrypto locking code that changes will
11448 have to be made anyway).
11452 * In asn1_d2i_read_bio() repeatedly call BIO_read() until all content
11453 octets have been read, EOF or an error occurs. Without this change
11454 some truncated ASN1 structures will not produce an error.
11458 * Disable Heimdal support, since it hasn't been fully implemented.
11459 Still give the possibility to force the use of Heimdal, but with
11460 warnings and a request that patches get sent to openssl-dev.
11464 * Add the VC-CE target, introduce the WINCE sysname, and add
11465 INSTALL.WCE and appropriate conditionals to make it build.
11467 *Steven Reddie <smr@essemer.com.au> via Richard Levitte*
11469 * Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and
11470 cygssl-x.y.z.dll, where x, y and z are the major, minor and
11471 edit numbers of the version.
11473 *Corinna Vinschen <vinschen@redhat.com> and Richard Levitte*
11475 * Introduce safe string copy and catenation functions
11476 (BUF_strlcpy() and BUF_strlcat()).
11478 *Ben Laurie (CHATS) and Richard Levitte*
11480 * Avoid using fixed-size buffers for one-line DNs.
11482 *Ben Laurie (CHATS)*
11484 * Add BUF_MEM_grow_clean() to avoid information leakage when
11485 resizing buffers containing secrets, and use where appropriate.
11487 *Ben Laurie (CHATS)*
11489 * Avoid using fixed size buffers for configuration file location.
11491 *Ben Laurie (CHATS)*
11493 * Avoid filename truncation for various CA files.
11495 *Ben Laurie (CHATS)*
11497 * Use sizeof in preference to magic numbers.
11499 *Ben Laurie (CHATS)*
11501 * Avoid filename truncation in cert requests.
11503 *Ben Laurie (CHATS)*
11505 * Add assertions to check for (supposedly impossible) buffer
11508 *Ben Laurie (CHATS)*
11510 * Don't cache truncated DNS entries in the local cache (this could
11511 potentially lead to a spoofing attack).
11513 *Ben Laurie (CHATS)*
11515 * Fix various buffers to be large enough for hex/decimal
11516 representations in a platform independent manner.
11518 *Ben Laurie (CHATS)*
11520 * Add CRYPTO_realloc_clean() to avoid information leakage when
11521 resizing buffers containing secrets, and use where appropriate.
11523 *Ben Laurie (CHATS)*
11525 * Add BIO_indent() to avoid much slightly worrying code to do
11528 *Ben Laurie (CHATS)*
11530 * Convert sprintf()/BIO_puts() to BIO_printf().
11532 *Ben Laurie (CHATS)*
11534 * buffer_gets() could terminate with the buffer only half
11537 *Ben Laurie (CHATS)*
11539 * Add assertions to prevent user-supplied crypto functions from
11540 overflowing internal buffers by having large block sizes, etc.
11542 *Ben Laurie (CHATS)*
11544 * New OPENSSL_assert() macro (similar to assert(), but enabled
11547 *Ben Laurie (CHATS)*
11549 * Eliminate unused copy of key in RC4.
11551 *Ben Laurie (CHATS)*
11553 * Eliminate unused and incorrectly sized buffers for IV in pem.h.
11555 *Ben Laurie (CHATS)*
11557 * Fix off-by-one error in EGD path.
11559 *Ben Laurie (CHATS)*
11561 * If RANDFILE path is too long, ignore instead of truncating.
11563 *Ben Laurie (CHATS)*
11565 * Eliminate unused and incorrectly sized X.509 structure
11568 *Ben Laurie (CHATS)*
11570 * Eliminate unused and dangerous function knumber().
11572 *Ben Laurie (CHATS)*
11574 * Eliminate unused and dangerous structure, KSSL_ERR.
11576 *Ben Laurie (CHATS)*
11578 * Protect against overlong session ID context length in an encoded
11579 session object. Since these are local, this does not appear to be
11582 *Ben Laurie (CHATS)*
11584 * Change from security patch (see 0.9.6e below) that did not affect
11585 the 0.9.6 release series:
11587 Remote buffer overflow in SSL3 protocol - an attacker could
11588 supply an oversized master key in Kerberos-enabled versions.
11591 *Ben Laurie (CHATS)*
11593 * Change the SSL kerb5 codes to match RFC 2712.
11597 * Make -nameopt work fully for req and add -reqopt switch.
11599 *Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson*
11601 * The "block size" for block ciphers in CFB and OFB mode should be 1.
11603 *Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>*
11605 * Make sure tests can be performed even if the corresponding algorithms
11606 have been removed entirely. This was also the last step to make
11607 OpenSSL compilable with DJGPP under all reasonable conditions.
11609 *Richard Levitte, Doug Kaufman <dkaufman@rahul.net>*
11611 * Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT
11612 to allow version independent disabling of normally unselected ciphers,
11613 which may be activated as a side-effect of selecting a single cipher.
11615 (E.g., cipher list string "RSA" enables ciphersuites that are left
11616 out of "ALL" because they do not provide symmetric encryption.
11617 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.)
11619 *Lutz Jaenicke, Bodo Moeller*
11621 * Add appropriate support for separate platform-dependent build
11622 directories. The recommended way to make a platform-dependent
11623 build directory is the following (tested on Linux), maybe with
11626 # Place yourself outside of the OpenSSL source tree. In
11627 # this example, the environment variable OPENSSL_SOURCE
11628 # is assumed to contain the absolute OpenSSL source directory.
11629 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`"
11630 cd objtree/"`uname -s`-`uname -r`-`uname -m`"
11631 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do
11632 mkdir -p `dirname $F`
11633 ln -s $OPENSSL_SOURCE/$F $F
11636 To be absolutely sure not to disturb the source tree, a "make clean"
11637 is a good thing. If it isn't successful, don't worry about it,
11638 it probably means the source directory is very clean.
11642 * Make sure any ENGINE control commands make local copies of string
11643 pointers passed to them whenever necessary. Otherwise it is possible
11644 the caller may have overwritten (or deallocated) the original string
11645 data when a later ENGINE operation tries to use the stored values.
11647 *Götz Babin-Ebell <babinebell@trustcenter.de>*
11649 * Improve diagnostics in file reading and command-line digests.
11651 *Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>*
11653 * Add AES modes CFB and OFB to the object database. Correct an
11654 error in AES-CFB decryption.
11658 * Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this
11659 allows existing EVP_CIPHER_CTX structures to be reused after
11660 calling `EVP_*Final()`. This behaviour is used by encryption
11661 BIOs and some applications. This has the side effect that
11662 applications must explicitly clean up cipher contexts with
11663 EVP_CIPHER_CTX_cleanup() or they will leak memory.
11667 * Check the values of dna and dnb in bn_mul_recursive before calling
11668 bn_mul_comba (a non zero value means the a or b arrays do not contain
11669 n2 elements) and fallback to bn_mul_normal if either is not zero.
11673 * Fix escaping of non-ASCII characters when using the -subj option
11674 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>)
11678 * Make object definitions compliant to LDAP (RFC2256): SN is the short
11679 form for "surname", serialNumber has no short form.
11680 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798;
11681 therefore remove "mail" short name for "internet 7".
11682 The OID for unique identifiers in X509 certificates is
11683 x500UniqueIdentifier, not uniqueIdentifier.
11684 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>)
11688 * Add an "init" command to the ENGINE config module and auto initialize
11689 ENGINEs. Without any "init" command the ENGINE will be initialized
11690 after all ctrl commands have been executed on it. If init=1 the
11691 ENGINE is initialized at that point (ctrls before that point are run
11692 on the uninitialized ENGINE and after on the initialized one). If
11693 init=0 then the ENGINE will not be initialized at all.
11697 * Fix the 'app_verify_callback' interface so that the user-defined
11698 argument is actually passed to the callback: In the
11699 SSL_CTX_set_cert_verify_callback() prototype, the callback
11700 declaration has been changed from
11703 int (*cb)(X509_STORE_CTX *,void *);
11704 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
11705 i=s->ctx->app_verify_callback(&ctx)
11706 has been changed into
11707 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
11709 To update applications using SSL_CTX_set_cert_verify_callback(),
11710 a dummy argument can be added to their callback functions.
11712 *D. K. Smetters <smetters@parc.xerox.com>*
11714 * Added the '4758cca' ENGINE to support IBM 4758 cards.
11716 *Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe*
11718 * Add and OPENSSL_LOAD_CONF define which will cause
11719 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
11720 This allows older applications to transparently support certain
11721 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
11722 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
11723 load the config file and OPENSSL_add_all_algorithms_conf() which will
11724 always load it have also been added.
11728 * Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
11729 Adjust NIDs and EVP layer.
11731 *Stephen Sprunk <stephen@sprunk.org> and Richard Levitte*
11733 * Config modules support in openssl utility.
11735 Most commands now load modules from the config file,
11736 though in a few (such as version) this isn't done
11737 because it couldn't be used for anything.
11739 In the case of ca and req the config file used is
11740 the same as the utility itself: that is the -config
11741 command line option can be used to specify an
11746 * Move default behaviour from OPENSSL_config(). If appname is NULL
11747 use "openssl_conf" if filename is NULL use default openssl config file.
11751 * Add an argument to OPENSSL_config() to allow the use of an alternative
11752 config section name. Add a new flag to tolerate a missing config file
11753 and move code to CONF_modules_load_file().
11757 * Support for crypto accelerator cards from Accelerated Encryption
11758 Processing, www.aep.ie. (Use engine 'aep')
11759 The support was copied from 0.9.6c [engine] and adapted/corrected
11760 to work with the new engine framework.
11762 *AEP Inc. and Richard Levitte*
11764 * Support for SureWare crypto accelerator cards from Baltimore
11765 Technologies. (Use engine 'sureware')
11766 The support was copied from 0.9.6c [engine] and adapted
11767 to work with the new engine framework.
11771 * Have the CHIL engine fork-safe (as defined by nCipher) and actually
11772 make the newer ENGINE framework commands for the CHIL engine work.
11774 *Toomas Kiisk <vix@cyber.ee> and Richard Levitte*
11776 * Make it possible to produce shared libraries on ReliantUNIX.
11778 *Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte*
11780 * Add the configuration target debug-linux-ppro.
11781 Make 'openssl rsa' use the general key loading routines
11782 implemented in `apps.c`, and make those routines able to
11783 handle the key format FORMAT_NETSCAPE and the variant
11786 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
11788 * Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
11790 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
11792 * Add -keyform to rsautl, and document -engine.
11794 *Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>*
11796 * Change BIO_new_file (crypto/bio/bss_file.c) to use new
11797 BIO_R_NO_SUCH_FILE error code rather than the generic
11798 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
11802 * Add new functions
11803 ERR_peek_last_error
11804 ERR_peek_last_error_line
11805 ERR_peek_last_error_line_data.
11806 These are similar to
11808 ERR_peek_error_line
11809 ERR_peek_error_line_data,
11810 but report on the latest error recorded rather than the first one
11811 still in the error queue.
11813 *Ben Laurie, Bodo Moeller*
11815 * default_algorithms option in ENGINE config module. This allows things
11817 default_algorithms = ALL
11818 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
11822 * Preliminary ENGINE config module.
11826 * New experimental application configuration code.
11830 * Change the AES code to follow the same name structure as all other
11831 symmetric ciphers, and behave the same way. Move everything to
11832 the directory crypto/aes, thereby obsoleting crypto/rijndael.
11834 *Stephen Sprunk <stephen@sprunk.org> and Richard Levitte*
11836 * SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
11838 *Ben Laurie and Theo de Raadt*
11840 * Add option to output public keys in req command.
11842 *Massimiliano Pala madwolf@openca.org*
11844 * Use wNAFs in EC_POINTs_mul() for improved efficiency
11845 (up to about 10% better than before for P-192 and P-224).
11849 * New functions/macros
11851 SSL_CTX_set_msg_callback(ctx, cb)
11852 SSL_CTX_set_msg_callback_arg(ctx, arg)
11853 SSL_set_msg_callback(ssl, cb)
11854 SSL_set_msg_callback_arg(ssl, arg)
11856 to request calling a callback function
11858 void cb(int write_p, int version, int content_type,
11859 const void *buf, size_t len, SSL *ssl, void *arg)
11861 whenever a protocol message has been completely received
11862 (write_p == 0) or sent (write_p == 1). Here 'version' is the
11863 protocol version according to which the SSL library interprets
11864 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
11865 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
11866 the content type as defined in the SSL 3.0/TLS 1.0 protocol
11867 specification (change_cipher_spec(20), alert(21), handshake(22)).
11868 'buf' and 'len' point to the actual message, 'ssl' to the
11869 SSL object, and 'arg' is the application-defined value set by
11870 SSL[_CTX]_set_msg_callback_arg().
11872 'openssl s_client' and 'openssl s_server' have new '-msg' options
11873 to enable a callback that displays all protocol messages.
11877 * Change the shared library support so shared libraries are built as
11878 soon as the corresponding static library is finished, and thereby get
11879 openssl and the test programs linked against the shared library.
11880 This still only happens when the keyword "shard" has been given to
11881 the configuration scripts.
11883 NOTE: shared library support is still an experimental thing, and
11884 backward binary compatibility is still not guaranteed.
11886 *"Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte*
11888 * Add support for Subject Information Access extension.
11890 *Peter Sylvester <Peter.Sylvester@EdelWeb.fr>*
11892 * Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
11893 additional bytes when new memory had to be allocated, not just
11894 when reusing an existing buffer.
11898 * New command line and configuration option 'utf8' for the req command.
11899 This allows field values to be specified as UTF8 strings.
11903 * Add -multi and -mr options to "openssl speed" - giving multiple parallel
11904 runs for the former and machine-readable output for the latter.
11908 * Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
11909 of the e-mail address in the DN (i.e., it will go into a certificate
11910 extension only). The new configuration file option 'email_in_dn = no'
11911 has the same effect.
11913 *Massimiliano Pala madwolf@openca.org*
11915 * Change all functions with names starting with `des_` to be starting
11916 with `DES_` instead. Add wrappers that are compatible with libdes,
11917 but are named `_ossl_old_des_*`. Finally, add macros that map the
11918 `des_*` symbols to the corresponding `_ossl_old_des_*` if libdes
11919 compatibility is desired. If OpenSSL 0.9.6c compatibility is
11920 desired, the `des_*` symbols will be mapped to `DES_*`, with one
11923 Since we provide two compatibility mappings, the user needs to
11924 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
11925 compatibility is desired. The default (i.e., when that macro
11926 isn't defined) is OpenSSL 0.9.6c compatibility.
11928 There are also macros that enable and disable the support of old
11929 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
11930 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
11931 are defined, the default will apply: to support the old des routines.
11933 In either case, one must include openssl/des.h to get the correct
11934 definitions. Do not try to just include openssl/des_old.h, that
11937 NOTE: This is a major break of an old API into a new one. Software
11938 authors are encouraged to switch to the `DES_` style functions. Some
11939 time in the future, des_old.h and the libdes compatibility functions
11940 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
11941 default), and then completely removed.
11945 * Test for certificates which contain unsupported critical extensions.
11946 If such a certificate is found during a verify operation it is
11947 rejected by default: this behaviour can be overridden by either
11948 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
11949 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
11950 X509_supported_extension() has also been added which returns 1 if a
11951 particular extension is supported.
11955 * Modify the behaviour of EVP cipher functions in similar way to digests
11956 to retain compatibility with existing code.
11960 * Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
11961 compatibility with existing code. In particular the 'ctx' parameter does
11962 not have to be to be initialized before the call to EVP_DigestInit() and
11963 it is tidied up after a call to EVP_DigestFinal(). New function
11964 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
11965 EVP_MD_CTX_copy() changed to not require the destination to be
11966 initialized valid and new function EVP_MD_CTX_copy_ex() added which
11967 requires the destination to be valid.
11969 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
11970 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
11974 * Change ssl3_get_message (ssl/s3_both.c) and the functions using it
11975 so that complete 'Handshake' protocol structures are kept in memory
11976 instead of overwriting 'msg_type' and 'length' with 'body' data.
11980 * Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
11982 *Massimo Santin via Richard Levitte*
11984 * Major restructuring to the underlying ENGINE code. This includes
11985 reduction of linker bloat, separation of pure "ENGINE" manipulation
11986 (initialisation, etc) from functionality dealing with implementations
11987 of specific crypto interfaces. This change also introduces integrated
11988 support for symmetric ciphers and digest implementations - so ENGINEs
11989 can now accelerate these by providing EVP_CIPHER and EVP_MD
11990 implementations of their own. This is detailed in
11991 [crypto/engine/README.md](crypto/engine/README.md)
11992 as it couldn't be adequately described here. However, there are a few
11993 API changes worth noting - some RSA, DSA, DH, and RAND functions that
11994 were changed in the original introduction of ENGINE code have now
11995 reverted back - the hooking from this code to ENGINE is now a good
11996 deal more passive and at run-time, operations deal directly with
11997 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
11998 dereferencing through an ENGINE pointer any more. Also, the ENGINE
11999 functions dealing with `BN_MOD_EXP[_CRT]` handlers have been removed -
12000 they were not being used by the framework as there is no concept of a
12001 BIGNUM_METHOD and they could not be generalised to the new
12002 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
12003 ENGINE_cpy() has been removed as it cannot be consistently defined in
12008 * Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
12012 * Change mkdef.pl to sort symbols that get the same entry number,
12013 and make sure the automatically generated functions `ERR_load_*`
12014 become part of libeay.num as well.
12018 * New function SSL_renegotiate_pending(). This returns true once
12019 renegotiation has been requested (either SSL_renegotiate() call
12020 or HelloRequest/ClientHello received from the peer) and becomes
12021 false once a handshake has been completed.
12022 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
12023 sends a HelloRequest, but does not ensure that a handshake takes
12024 place. SSL_renegotiate_pending() is useful for checking if the
12025 client has followed the request.)
12029 * New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
12030 By default, clients may request session resumption even during
12031 renegotiation (if session ID contexts permit); with this option,
12032 session resumption is possible only in the first handshake.
12034 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes
12035 more bits available for options that should not be part of
12036 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION).
12040 * Add some demos for certificate and certificate request creation.
12044 * Make maximum certificate chain size accepted from the peer application
12045 settable (`SSL*_get/set_max_cert_list()`), as proposed by
12046 "Douglas E. Engert" <deengert@anl.gov>.
12050 * Add support for shared libraries for Unixware-7
12051 (Boyd Lynn Gerber <gerberb@zenez.com>).
12055 * Add a "destroy" handler to ENGINEs that allows structural cleanup to
12056 be done prior to destruction. Use this to unload error strings from
12057 ENGINEs that load their own error strings. NB: This adds two new API
12058 functions to "get" and "set" this destroy handler in an ENGINE.
12062 * Alter all existing ENGINE implementations (except "openssl" and
12063 "openbsd") to dynamically instantiate their own error strings. This
12064 makes them more flexible to be built both as statically-linked ENGINEs
12065 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
12066 Also, add stub code to each that makes building them as self-contained
12067 shared-libraries easier (see [README-Engine.md](README-Engine.md)).
12071 * Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
12072 implementations into applications that are completely implemented in
12073 self-contained shared-libraries. The "dynamic" ENGINE exposes control
12074 commands that can be used to configure what shared-library to load and
12075 to control aspects of the way it is handled. Also, made an update to
12076 the [README-Engine.md](README-Engine.md) file
12077 that brings its information up-to-date and
12078 provides some information and instructions on the "dynamic" ENGINE
12079 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
12083 * Make it possible to unload ranges of ERR strings with a new
12084 "ERR_unload_strings" function.
12088 * Add a copy() function to EVP_MD.
12092 * Make EVP_MD routines take a context pointer instead of just the
12093 md_data void pointer.
12097 * Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
12098 that the digest can only process a single chunk of data
12099 (typically because it is provided by a piece of
12100 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
12101 is only going to provide a single chunk of data, and hence the
12102 framework needn't accumulate the data for oneshot drivers.
12106 * As with "ERR", make it possible to replace the underlying "ex_data"
12107 functions. This change also alters the storage and management of global
12108 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
12109 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
12110 index counters. The API functions that use this state have been changed
12111 to take a "class_index" rather than pointers to the class's local STACK
12112 and counter, and there is now an API function to dynamically create new
12113 classes. This centralisation allows us to (a) plug a lot of the
12114 thread-safety problems that existed, and (b) makes it possible to clean
12115 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
12116 such data would previously have always leaked in application code and
12117 workarounds were in place to make the memory debugging turn a blind eye
12118 to it. Application code that doesn't use this new function will still
12119 leak as before, but their memory debugging output will announce it now
12120 rather than letting it slide.
12122 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
12123 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
12124 has a return value to indicate success or failure.
12128 * Make it possible to replace the underlying "ERR" functions such that the
12129 global state (2 LHASH tables and 2 locks) is only used by the "default"
12130 implementation. This change also adds two functions to "get" and "set"
12131 the implementation prior to it being automatically set the first time
12132 any other ERR function takes place. Ie. an application can call "get",
12133 pass the return value to a module it has just loaded, and that module
12134 can call its own "set" function using that value. This means the
12135 module's "ERR" operations will use (and modify) the error state in the
12136 application and not in its own statically linked copy of OpenSSL code.
12140 * Give DH, DSA, and RSA types their own `*_up_ref()` function to increment
12141 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
12142 the operation, and provides a more encapsulated way for external code
12143 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
12144 to use these functions rather than manually incrementing the counts.
12146 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
12150 * Add EVP test program.
12154 * Add symmetric cipher support to ENGINE. Expect the API to change!
12158 * New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
12159 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
12160 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
12161 These allow a CRL to be built without having to access X509_CRL fields
12162 directly. Modify 'ca' application to use new functions.
12166 * Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
12167 bug workarounds. Rollback attack detection is a security feature.
12168 The problem will only arise on OpenSSL servers when TLSv1 is not
12169 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
12170 Software authors not wanting to support TLSv1 will have special reasons
12171 for their choice and can explicitly enable this option.
12173 *Bodo Moeller, Lutz Jaenicke*
12175 * Rationalise EVP so it can be extended: don't include a union of
12176 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX
12177 (similar to those existing for EVP_CIPHER_CTX).
12182 EVP_MD_CTX_init(&md); /* new function call */
12183 EVP_DigestInit(&md, EVP_sha1());
12184 EVP_DigestUpdate(&md, in, len);
12185 EVP_DigestFinal(&md, out, NULL);
12186 EVP_MD_CTX_cleanup(&md); /* new function call */
12190 * Make DES key schedule conform to the usual scheme, as well as
12191 correcting its structure. This means that calls to DES functions
12192 now have to pass a pointer to a des_key_schedule instead of a
12193 plain des_key_schedule (which was actually always a pointer
12196 des_key_schedule ks;
12198 des_set_key_checked(..., &ks);
12199 des_ncbc_encrypt(..., &ks, ...);
12201 (Note that a later change renames 'des_...' into 'DES_...'.)
12205 * Initial reduction of linker bloat: the use of some functions, such as
12206 PEM causes large amounts of unused functions to be linked in due to
12207 poor organisation. For example pem_all.c contains every PEM function
12208 which has a knock on effect of linking in large amounts of (unused)
12209 ASN1 code. Grouping together similar functions and splitting unrelated
12210 functions prevents this.
12214 * Cleanup of EVP macros.
12218 * Change historical references to `{NID,SN,LN}_des_ede` and ede3 to add the
12219 correct `_ecb suffix`.
12223 * Add initial OCSP responder support to ocsp application. The
12224 revocation information is handled using the text based index
12225 use by the ca application. The responder can either handle
12226 requests generated internally, supplied in files (for example
12227 via a CGI script) or using an internal minimal server.
12231 * Add configuration choices to get zlib compression for TLS.
12235 * Changes to Kerberos SSL for RFC 2712 compliance:
12236 1. Implemented real KerberosWrapper, instead of just using
12237 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
12238 2. Implemented optional authenticator field of KerberosWrapper.
12240 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
12241 and authenticator structs; see crypto/krb5/.
12243 Generalized Kerberos calls to support multiple Kerberos libraries.
12244 *Vern Staats <staatsvr@asc.hpc.mil>, Jeffrey Altman <jaltman@columbia.edu>
12245 via Richard Levitte*
12247 * Cause 'openssl speed' to use fully hard-coded DSA keys as it
12248 already does with RSA. testdsa.h now has 'priv_key/pub_key'
12249 values for each of the key sizes rather than having just
12250 parameters (and 'speed' generating keys each time).
12254 * Speed up EVP routines.
12257 pe 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
12258 s-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
12259 s-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
12260 s-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
12262 s-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
12263 s-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
12264 s-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
12267 s-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
12269 s-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
12273 * Added the OS2-EMX target.
12275 *"Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte*
12277 * Rewrite commands to use `NCONF` routines instead of the old `CONF`.
12278 New functions to support `NCONF` routines in extension code.
12279 New function `CONF_set_nconf()`
12280 to allow functions which take an `NCONF` to also handle the old `LHASH`
12281 structure: this means that the old `CONF` compatible routines can be
12282 retained (in particular w.rt. extensions) without having to duplicate the
12283 code. New function `X509V3_add_ext_nconf_sk()` to add extensions to a stack.
12287 * Enhance the general user interface with mechanisms for inner control
12288 and with possibilities to have yes/no kind of prompts.
12292 * Change all calls to low-level digest routines in the library and
12293 applications to use EVP. Add missing calls to HMAC_cleanup() and
12294 don't assume HMAC_CTX can be copied using memcpy().
12296 *Verdon Walker <VWalker@novell.com>, Steve Henson*
12298 * Add the possibility to control engines through control names but with
12299 arbitrary arguments instead of just a string.
12300 Change the key loaders to take a UI_METHOD instead of a callback
12301 function pointer. NOTE: this breaks binary compatibility with earlier
12302 versions of OpenSSL [engine].
12303 Adapt the nCipher code for these new conditions and add a card insertion
12308 * Enhance the general user interface with mechanisms to better support
12309 dialog box interfaces, application-defined prompts, the possibility
12310 to use defaults (for example default passwords from somewhere else)
12311 and interrupts/cancellations.
12315 * Tidy up PKCS#12 attribute handling. Add support for the CSP name
12316 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
12320 * Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
12321 tidy up some unnecessarily weird code in 'sk_new()').
12323 *Geoff, reported by Diego Tartara <dtartara@novamens.com>*
12325 * Change the key loading routines for ENGINEs to use the same kind
12326 callback (pem_password_cb) as all other routines that need this
12331 * Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
12332 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
12333 than this minimum value is recommended.
12337 * New random seeder for OpenVMS, using the system process statistics
12338 that are easily reachable.
12342 * Windows apparently can't transparently handle global
12343 variables defined in DLLs. Initialisations such as:
12345 const ASN1_ITEM *it = &ASN1_INTEGER_it;
12347 won't compile. This is used by the any applications that need to
12348 declare their own ASN1 modules. This was fixed by adding the option
12349 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
12350 needed for static libraries under Win32.
12354 * New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
12355 setting of purpose and trust fields. New X509_STORE trust and
12356 purpose functions and tidy up setting in other SSL functions.
12360 * Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
12361 structure. These are inherited by X509_STORE_CTX when it is
12362 initialised. This allows various defaults to be set in the
12363 X509_STORE structure (such as flags for CRL checking and custom
12364 purpose or trust settings) for functions which only use X509_STORE_CTX
12365 internally such as S/MIME.
12367 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
12368 trust settings if they are not set in X509_STORE. This allows X509_STORE
12369 purposes and trust (in S/MIME for example) to override any set by default.
12371 Add command line options for CRL checking to smime, s_client and s_server
12376 * Initial CRL based revocation checking. If the CRL checking flag(s)
12377 are set then the CRL is looked up in the X509_STORE structure and
12378 its validity and signature checked, then if the certificate is found
12379 in the CRL the verify fails with a revoked error.
12381 Various new CRL related callbacks added to X509_STORE_CTX structure.
12383 Command line options added to 'verify' application to support this.
12385 This needs some additional work, such as being able to handle multiple
12386 CRLs with different times, extension based lookup (rather than just
12387 by subject name) and ultimately more complete V2 CRL extension
12392 * Add a general user interface API (crypto/ui/). This is designed
12393 to replace things like des_read_password and friends (backward
12394 compatibility functions using this new API are provided).
12395 The purpose is to remove prompting functions from the DES code
12396 section as well as provide for prompting through dialog boxes in
12397 a window system and the like.
12401 * Add "ex_data" support to ENGINE so implementations can add state at a
12402 per-structure level rather than having to store it globally.
12406 * Make it possible for ENGINE structures to be copied when retrieved by
12407 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
12408 This causes the "original" ENGINE structure to act like a template,
12409 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
12410 operational state can be localised to each ENGINE structure, despite the
12411 fact they all share the same "methods". New ENGINE structures returned in
12412 this case have no functional references and the return value is the single
12413 structural reference. This matches the single structural reference returned
12414 by ENGINE_by_id() normally, when it is incremented on the pre-existing
12419 * Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
12420 needs to match any other type at all we need to manually clear the
12425 * Changes to the "openssl engine" utility to include;
12426 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
12427 about an ENGINE's available control commands.
12428 - executing control commands from command line arguments using the
12429 '-pre' and '-post' switches. '-post' is only used if '-t' is
12430 specified and the ENGINE is successfully initialised. The syntax for
12431 the individual commands are colon-separated, for example;
12432 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
12436 * New dynamic control command support for ENGINEs. ENGINEs can now
12437 declare their own commands (numbers), names (strings), descriptions,
12438 and input types for run-time discovery by calling applications. A
12439 subset of these commands are implicitly classed as "executable"
12440 depending on their input type, and only these can be invoked through
12441 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
12442 can be based on user input, config files, etc). The distinction is
12443 that "executable" commands cannot return anything other than a boolean
12444 result and can only support numeric or string input, whereas some
12445 discoverable commands may only be for direct use through
12446 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
12447 pointers, or other custom uses. The "executable" commands are to
12448 support parameterisations of ENGINE behaviour that can be
12449 unambiguously defined by ENGINEs and used consistently across any
12450 OpenSSL-based application. Commands have been added to all the
12451 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
12452 control over shared-library paths without source code alterations.
12456 * Changed all ENGINE implementations to dynamically allocate their
12457 ENGINEs rather than declaring them statically. Apart from this being
12458 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
12459 this also allows the implementations to compile without using the
12460 internal engine_int.h header.
12464 * Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
12465 'const' value. Any code that should be able to modify a RAND_METHOD
12466 should already have non-const pointers to it (ie. they should only
12467 modify their own ones).
12471 * Made a variety of little tweaks to the ENGINE code.
12472 - "atalla" and "ubsec" string definitions were moved from header files
12473 to C code. "nuron" string definitions were placed in variables
12474 rather than hard-coded - allowing parameterisation of these values
12475 later on via ctrl() commands.
12476 - Removed unused "#if 0"'d code.
12477 - Fixed engine list iteration code so it uses ENGINE_free() to release
12478 structural references.
12479 - Constified the RAND_METHOD element of ENGINE structures.
12480 - Constified various get/set functions as appropriate and added
12481 missing functions (including a catch-all ENGINE_cpy that duplicates
12482 all ENGINE values onto a new ENGINE except reference counts/state).
12483 - Removed NULL parameter checks in get/set functions. Setting a method
12484 or function to NULL is a way of cancelling out a previously set
12485 value. Passing a NULL ENGINE parameter is just plain stupid anyway
12486 and doesn't justify the extra error symbols and code.
12487 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
12488 flags from engine_int.h to engine.h.
12489 - Changed prototypes for ENGINE handler functions (init(), finish(),
12490 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
12494 * Implement binary inversion algorithm for BN_mod_inverse in addition
12495 to the algorithm using long division. The binary algorithm can be
12496 used only if the modulus is odd. On 32-bit systems, it is faster
12497 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
12498 roughly 5-15% for 256-bit moduli), so we use it only for moduli
12499 up to 450 bits. In 64-bit environments, the binary algorithm
12500 appears to be advantageous for much longer moduli; here we use it
12501 for moduli up to 2048 bits.
12505 * Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
12506 could not support the combine flag in choice fields.
12510 * Add a 'copy_extensions' option to the 'ca' utility. This copies
12511 extensions from a certificate request to the certificate.
12515 * Allow multiple 'certopt' and 'nameopt' options to be separated
12516 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
12517 file: this allows the display of the certificate about to be
12518 signed to be customised, to allow certain fields to be included
12519 or excluded and extension details. The old system didn't display
12520 multicharacter strings properly, omitted fields not in the policy
12521 and couldn't display additional details such as extensions.
12525 * Function EC_POINTs_mul for multiple scalar multiplication
12526 of an arbitrary number of elliptic curve points
12527 \sum scalars[i]*points[i],
12528 optionally including the generator defined for the EC_GROUP:
12529 scalar*generator + \sum scalars[i]*points[i].
12531 EC_POINT_mul is a simple wrapper function for the typical case
12532 that the point list has just one item (besides the optional
12537 * First EC_METHODs for curves over GF(p):
12539 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
12540 operations and provides various method functions that can also
12541 operate with faster implementations of modular arithmetic.
12543 EC_GFp_mont_method() reuses most functions that are part of
12544 EC_GFp_simple_method, but uses Montgomery arithmetic.
12546 *Bodo Moeller; point addition and point doubling
12547 implementation directly derived from source code provided by
12548 Lenka Fibikova <fibikova@exp-math.uni-essen.de>*
12550 * Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
12551 crypto/ec/ec_lib.c):
12553 Curves are EC_GROUP objects (with an optional group generator)
12554 based on EC_METHODs that are built into the library.
12556 Points are EC_POINT objects based on EC_GROUP objects.
12558 Most of the framework would be able to handle curves over arbitrary
12559 finite fields, but as there are no obvious types for fields other
12560 than GF(p), some functions are limited to that for now.
12564 * Add the -HTTP option to s_server. It is similar to -WWW, but requires
12565 that the file contains a complete HTTP response.
12569 * Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
12570 change the def and num file printf format specifier from "%-40sXXX"
12571 to "%-39s XXX". The latter will always guarantee a space after the
12572 field while the former will cause them to run together if the field
12573 is 40 of more characters long.
12577 * Constify the cipher and digest 'method' functions and structures
12578 and modify related functions to take constant EVP_MD and EVP_CIPHER
12583 * Hide BN_CTX structure details in bn_lcl.h instead of publishing them
12584 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
12588 * Modify `EVP_Digest*()` routines so they now return values. Although the
12589 internal software routines can never fail additional hardware versions
12594 * Clean up crypto/err/err.h and change some error codes to avoid conflicts:
12596 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
12597 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
12600 ERR_R_NESTED_ASN1_ERROR
12602 ERR_R_MISSING_ASN1_EOS
12603 were 4 .. 9, conflicting with
12604 ERR_LIB_RSA (= ERR_R_RSA_LIB)
12606 ERR_LIB_PEM (= ERR_R_PEM_LIB).
12607 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
12609 Add new error code 'ERR_R_INTERNAL_ERROR'.
12613 * Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
12618 * New option '-subj arg' for 'openssl req' and 'openssl ca'. This
12619 sets the subject name for a new request or supersedes the
12620 subject name in a given request. Formats that can be parsed are
12621 'CN=Some Name, OU=myOU, C=IT'
12623 'CN=Some Name/OU=myOU/C=IT'.
12625 Add options '-batch' and '-verbose' to 'openssl req'.
12627 *Massimiliano Pala <madwolf@hackmasters.net>*
12629 * Introduce the possibility to access global variables through
12630 functions on platform were that's the best way to handle exporting
12631 global variables in shared libraries. To enable this functionality,
12632 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
12633 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
12634 is normally done by Configure or something similar).
12636 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
12637 in the source file (foo.c) like this:
12639 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
12640 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
12642 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
12643 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
12645 OPENSSL_DECLARE_GLOBAL(int,foo);
12646 #define foo OPENSSL_GLOBAL_REF(foo)
12647 OPENSSL_DECLARE_GLOBAL(double,bar);
12648 #define bar OPENSSL_GLOBAL_REF(bar)
12650 The #defines are very important, and therefore so is including the
12651 header file everywhere where the defined globals are used.
12653 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
12654 of ASN.1 items, but that structure is a bit different.
12656 The largest change is in util/mkdef.pl which has been enhanced with
12657 better and easier to understand logic to choose which symbols should
12658 go into the Windows .def files as well as a number of fixes and code
12659 cleanup (among others, algorithm keywords are now sorted
12660 lexicographically to avoid constant rewrites).
12664 * In BN_div() keep a copy of the sign of 'num' before writing the
12665 result to 'rm' because if rm==num the value will be overwritten
12666 and produce the wrong result if 'num' is negative: this caused
12667 problems with BN_mod() and BN_nnmod().
12671 * Function OCSP_request_verify(). This checks the signature on an
12672 OCSP request and verifies the signer certificate. The signer
12673 certificate is just checked for a generic purpose and OCSP request
12678 * Add OCSP_check_validity() function to check the validity of OCSP
12679 responses. OCSP responses are prepared in real time and may only
12680 be a few seconds old. Simply checking that the current time lies
12681 between thisUpdate and nextUpdate max reject otherwise valid responses
12682 caused by either OCSP responder or client clock inaccuracy. Instead
12683 we allow thisUpdate and nextUpdate to fall within a certain period of
12684 the current time. The age of the response can also optionally be
12685 checked. Two new options -validity_period and -status_age added to
12690 * If signature or public key algorithm is unrecognized print out its
12691 OID rather that just UNKNOWN.
12695 * Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
12696 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
12697 ID to be generated from the issuer certificate alone which can then be
12698 passed to OCSP_id_issuer_cmp().
12702 * New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
12703 ASN1 modules to export functions returning ASN1_ITEM pointers
12704 instead of the ASN1_ITEM structures themselves. This adds several
12705 new macros which allow the underlying ASN1 function/structure to
12706 be accessed transparently. As a result code should not use ASN1_ITEM
12707 references directly (such as &X509_it) but instead use the relevant
12708 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
12709 use of the new ASN1 code on platforms where exporting structures
12710 is problematical (for example in shared libraries) but exporting
12711 functions returning pointers to structures is not.
12715 * Add support for overriding the generation of SSL/TLS session IDs.
12716 These callbacks can be registered either in an SSL_CTX or per SSL.
12717 The purpose of this is to allow applications to control, if they wish,
12718 the arbitrary values chosen for use as session IDs, particularly as it
12719 can be useful for session caching in multiple-server environments. A
12720 command-line switch for testing this (and any client code that wishes
12721 to use such a feature) has been added to "s_server".
12723 *Geoff Thorpe, Lutz Jaenicke*
12725 * Modify mkdef.pl to recognise and parse preprocessor conditionals
12726 of the form `#if defined(...) || defined(...) || ...` and
12727 `#if !defined(...) && !defined(...) && ...`. This also avoids
12728 the growing number of special cases it was previously handling.
12732 * Make all configuration macros available for application by making
12733 sure they are available in opensslconf.h, by giving them names starting
12734 with `OPENSSL_` to avoid conflicts with other packages and by making
12735 sure e_os2.h will cover all platform-specific cases together with
12737 Additionally, it is now possible to define configuration/platform-
12738 specific names (called "system identities"). In the C code, these
12739 are prefixed with `OPENSSL_SYSNAME_`. e_os2.h will create another
12740 macro with the name beginning with `OPENSSL_SYS_`, which is determined
12741 from `OPENSSL_SYSNAME_*` or compiler-specific macros depending on
12746 * New option -set_serial to 'req' and 'x509' this allows the serial
12747 number to use to be specified on the command line. Previously self
12748 signed certificates were hard coded with serial number 0 and the
12749 CA options of 'x509' had to use a serial number in a file which was
12754 * New options to 'ca' utility to support V2 CRL entry extensions.
12755 Currently CRL reason, invalidity date and hold instruction are
12756 supported. Add new CRL extensions to V3 code and some new objects.
12760 * New function EVP_CIPHER_CTX_set_padding() this is used to
12761 disable standard block padding (aka PKCS#5 padding) in the EVP
12762 API, which was previously mandatory. This means that the data is
12763 not padded in any way and so the total length much be a multiple
12764 of the block size, otherwise an error occurs.
12768 * Initial (incomplete) OCSP SSL support.
12772 * New function OCSP_parse_url(). This splits up a URL into its host,
12773 port and path components: primarily to parse OCSP URLs. New -url
12774 option to ocsp utility.
12778 * New nonce behavior. The return value of OCSP_check_nonce() now
12779 reflects the various checks performed. Applications can decide
12780 whether to tolerate certain situations such as an absent nonce
12781 in a response when one was present in a request: the ocsp application
12782 just prints out a warning. New function OCSP_add1_basic_nonce()
12783 this is to allow responders to include a nonce in a response even if
12784 the request is nonce-less.
12788 * Disable stdin buffering in `load_cert()` (`apps/apps.c`) so that no certs are
12789 skipped when using openssl x509 multiple times on a single input file,
12790 e.g. `(openssl x509 -out cert1; openssl x509 -out cert2) <certs`.
12794 * Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
12795 set string type: to handle setting ASN1_TIME structures. Fix ca
12796 utility to correctly initialize revocation date of CRLs.
12800 * New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
12801 the clients preferred ciphersuites and rather use its own preferences.
12802 Should help to work around M$ SGC (Server Gated Cryptography) bug in
12803 Internet Explorer by ensuring unchanged hash method during stepup.
12804 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
12808 * Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
12809 to aes and add a new 'exist' option to print out symbols that don't
12814 * Additional options to ocsp utility to allow flags to be set and
12815 additional certificates supplied.
12819 * Add the option -VAfile to 'openssl ocsp', so the user can give the
12820 OCSP client a number of certificate to only verify the response
12825 * Update Rijndael code to version 3.0 and change EVP AES ciphers to
12826 handle the new API. Currently only ECB, CBC modes supported. Add new
12829 Add TLS AES ciphersuites as described in RFC3268, "Advanced
12830 Encryption Standard (AES) Ciphersuites for Transport Layer
12831 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were
12832 not enabled by default and were not part of the "ALL" ciphersuite
12833 alias because they were not yet official; they could be
12834 explicitly requested by specifying the "AESdraft" ciphersuite
12835 group alias. In the final release of OpenSSL 0.9.7, the group
12836 alias is called "AES" and is part of "ALL".)
12838 *Ben Laurie, Steve Henson, Bodo Moeller*
12840 * New function OCSP_copy_nonce() to copy nonce value (if present) from
12841 request to response.
12845 * Functions for OCSP responders. OCSP_request_onereq_count(),
12846 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
12847 extract information from a certificate request. OCSP_response_create()
12848 creates a response and optionally adds a basic response structure.
12849 OCSP_basic_add1_status() adds a complete single response to a basic
12850 response and returns the OCSP_SINGLERESP structure just added (to allow
12851 extensions to be included for example). OCSP_basic_add1_cert() adds a
12852 certificate to a basic response and OCSP_basic_sign() signs a basic
12853 response with various flags. New helper functions ASN1_TIME_check()
12854 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
12855 (converts ASN1_TIME to GeneralizedTime).
12859 * Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
12860 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
12861 structure from a certificate. X509_pubkey_digest() digests the public_key
12862 contents: this is used in various key identifiers.
12866 * Make sk_sort() tolerate a NULL argument.
12868 *Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>*
12870 * New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
12871 passed by the function are trusted implicitly. If any of them signed the
12872 response then it is assumed to be valid and is not verified.
12876 * In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
12877 to data. This was previously part of the PKCS7 ASN1 code. This
12878 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
12879 *Steve Henson, reported by Kenneth R. Robinette
12880 <support@securenetterm.com>*
12882 * Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
12883 routines: without these tracing memory leaks is very painful.
12884 Fix leaks in PKCS12 and PKCS7 routines.
12888 * Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
12889 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
12890 effectively meant GeneralizedTime would never be used. Now it
12891 is initialised to -1 but X509_time_adj() now has to check the value
12892 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
12893 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
12894 *Steve Henson, reported by Kenneth R. Robinette
12895 <support@securenetterm.com>*
12897 * Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
12898 result in a zero length in the ASN1_INTEGER structure which was
12899 not consistent with the structure when d2i_ASN1_INTEGER() was used
12900 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
12901 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
12902 where it did not print out a minus for negative ASN1_INTEGER.
12906 * Add summary printout to ocsp utility. The various functions which
12907 convert status values to strings have been renamed to:
12908 OCSP_response_status_str(), OCSP_cert_status_str() and
12909 OCSP_crl_reason_str() and are no longer static. New options
12910 to verify nonce values and to disable verification. OCSP response
12911 printout format cleaned up.
12915 * Add additional OCSP certificate checks. These are those specified
12916 in RFC2560. This consists of two separate checks: the CA of the
12917 certificate being checked must either be the OCSP signer certificate
12918 or the issuer of the OCSP signer certificate. In the latter case the
12919 OCSP signer certificate must contain the OCSP signing extended key
12920 usage. This check is performed by attempting to match the OCSP
12921 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
12922 in the OCSP_CERTID structures of the response.
12926 * Initial OCSP certificate verification added to OCSP_basic_verify()
12927 and related routines. This uses the standard OpenSSL certificate
12928 verify routines to perform initial checks (just CA validity) and
12929 to obtain the certificate chain. Then additional checks will be
12930 performed on the chain. Currently the root CA is checked to see
12931 if it is explicitly trusted for OCSP signing. This is used to set
12932 a root CA as a global signing root: that is any certificate that
12933 chains to that CA is an acceptable OCSP signing certificate.
12937 * New '-extfile ...' option to 'openssl ca' for reading X.509v3
12938 extensions from a separate configuration file.
12939 As when reading extensions from the main configuration file,
12940 the '-extensions ...' option may be used for specifying the
12943 *Massimiliano Pala <madwolf@comune.modena.it>*
12945 * New OCSP utility. Allows OCSP requests to be generated or
12946 read. The request can be sent to a responder and the output
12947 parsed, outputted or printed in text form. Not complete yet:
12948 still needs to check the OCSP response validity.
12952 * New subcommands for 'openssl ca':
12953 `openssl ca -status <serial>` prints the status of the cert with
12954 the given serial number (according to the index file).
12955 `openssl ca -updatedb` updates the expiry status of certificates
12958 *Massimiliano Pala <madwolf@comune.modena.it>*
12960 * New '-newreq-nodes' command option to CA.pl. This is like
12961 '-newreq', but calls 'openssl req' with the '-nodes' option
12962 so that the resulting key is not encrypted.
12964 *Damien Miller <djm@mindrot.org>*
12966 * New configuration for the GNU Hurd.
12968 *Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte*
12970 * Initial code to implement OCSP basic response verify. This
12971 is currently incomplete. Currently just finds the signer's
12972 certificate and verifies the signature on the response.
12976 * New SSLeay_version code SSLEAY_DIR to determine the compiled-in
12977 value of OPENSSLDIR. This is available via the new '-d' option
12978 to 'openssl version', and is also included in 'openssl version -a'.
12982 * Allowing defining memory allocation callbacks that will be given
12983 file name and line number information in additional arguments
12984 (a `const char*` and an int). The basic functionality remains, as
12985 well as the original possibility to just replace malloc(),
12986 realloc() and free() by functions that do not know about these
12987 additional arguments. To register and find out the current
12988 settings for extended allocation functions, the following
12989 functions are provided:
12991 CRYPTO_set_mem_ex_functions
12992 CRYPTO_set_locked_mem_ex_functions
12993 CRYPTO_get_mem_ex_functions
12994 CRYPTO_get_locked_mem_ex_functions
12996 These work the same way as CRYPTO_set_mem_functions and friends.
12997 `CRYPTO_get_[locked_]mem_functions` now writes 0 where such an
12998 extended allocation function is enabled.
12999 Similarly, `CRYPTO_get_[locked_]mem_ex_functions` writes 0 where
13000 a conventional allocation function is enabled.
13002 *Richard Levitte, Bodo Moeller*
13004 * Finish off removing the remaining LHASH function pointer casts.
13005 There should no longer be any prototype-casting required when using
13006 the LHASH abstraction, and any casts that remain are "bugs". See
13007 the callback types and macros at the head of lhash.h for details
13008 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
13012 * Add automatic query of EGD sockets in RAND_poll() for the unix variant.
13013 If /dev/[u]random devices are not available or do not return enough
13014 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
13016 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
13017 /etc/entropy will be queried once each in this sequence, querying stops
13018 when enough entropy was collected without querying more sockets.
13022 * Change the Unix RAND_poll() variant to be able to poll several
13023 random devices, as specified by DEVRANDOM, until a sufficient amount
13024 of data has been collected. We spend at most 10 ms on each file
13025 (select timeout) and read in non-blocking mode. DEVRANDOM now
13026 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
13027 (previously it was just the string "/dev/urandom"), so on typical
13028 platforms the 10 ms delay will never occur.
13029 Also separate out the Unix variant to its own file, rand_unix.c.
13030 For VMS, there's a currently-empty rand_vms.c.
13034 * Move OCSP client related routines to ocsp_cl.c. These
13035 provide utility functions which an application needing
13036 to issue a request to an OCSP responder and analyse the
13037 response will typically need: as opposed to those which an
13038 OCSP responder itself would need which will be added later.
13040 OCSP_request_sign() signs an OCSP request with an API similar
13041 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
13042 response. OCSP_response_get1_basic() extracts basic response
13043 from response. OCSP_resp_find_status(): finds and extracts status
13044 information from an OCSP_CERTID structure (which will be created
13045 when the request structure is built). These are built from lower
13046 level functions which work on OCSP_SINGLERESP structures but
13047 won't normally be used unless the application wishes to examine
13048 extensions in the OCSP response for example.
13050 Replace nonce routines with a pair of functions.
13051 OCSP_request_add1_nonce() adds a nonce value and optionally
13052 generates a random value. OCSP_check_nonce() checks the
13053 validity of the nonce in an OCSP response.
13057 * Change function OCSP_request_add() to OCSP_request_add0_id().
13058 This doesn't copy the supplied OCSP_CERTID and avoids the
13059 need to free up the newly created id. Change return type
13060 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
13061 This can then be used to add extensions to the request.
13062 Deleted OCSP_request_new(), since most of its functionality
13063 is now in OCSP_REQUEST_new() (and the case insensitive name
13064 clash) apart from the ability to set the request name which
13065 will be added elsewhere.
13069 * Update OCSP API. Remove obsolete extensions argument from
13070 various functions. Extensions are now handled using the new
13071 OCSP extension code. New simple OCSP HTTP function which
13072 can be used to send requests and parse the response.
13076 * Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
13077 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
13078 uses the special reorder version of SET OF to sort the attributes
13079 and reorder them to match the encoded order. This resolves a long
13080 standing problem: a verify on a PKCS7 structure just after signing
13081 it used to fail because the attribute order did not match the
13082 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
13083 it uses the received order. This is necessary to tolerate some broken
13084 software that does not order SET OF. This is handled by encoding
13085 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
13086 to produce the required SET OF.
13090 * Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
13091 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
13092 files to get correct declarations of the ASN.1 item variables.
13096 * Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
13097 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
13098 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
13099 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
13100 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
13101 ASN1_ITEM and no wrapper functions.
13105 * New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
13106 replace the old function pointer based I/O routines. Change most of
13107 the `*_d2i_bio()` and `*_d2i_fp()` functions to use these.
13111 * Enhance mkdef.pl to be more accepting about spacing in C preprocessor
13112 lines, recognize more "algorithms" that can be deselected, and make
13113 it complain about algorithm deselection that isn't recognised.
13117 * New ASN1 functions to handle dup, sign, verify, digest, pack and
13118 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
13119 to use new functions. Add NO_ASN1_OLD which can be set to remove
13120 some old style ASN1 functions: this can be used to determine if old
13121 code will still work when these eventually go away.
13125 * New extension functions for OCSP structures, these follow the
13126 same conventions as certificates and CRLs.
13130 * New function X509V3_add1_i2d(). This automatically encodes and
13131 adds an extension. Its behaviour can be customised with various
13132 flags to append, replace or delete. Various wrappers added for
13133 certificates and CRLs.
13137 * Fix to avoid calling the underlying ASN1 print routine when
13138 an extension cannot be parsed. Correct a typo in the
13139 OCSP_SERVICELOC extension. Tidy up print OCSP format.
13143 * Make mkdef.pl parse some of the ASN1 macros and add appropriate
13144 entries for variables.
13148 * Add functionality to `apps/openssl.c` for detecting locking
13149 problems: As the program is single-threaded, all we have
13150 to do is register a locking callback using an array for
13151 storing which locks are currently held by the program.
13155 * Use a lock around the call to CRYPTO_get_ex_new_index() in
13156 SSL_get_ex_data_X509_STORE_idx(), which is used in
13157 ssl_verify_cert_chain() and thus can be called at any time
13158 during TLS/SSL handshakes so that thread-safety is essential.
13159 Unfortunately, the ex_data design is not at all suited
13160 for multi-threaded use, so it probably should be abolished.
13164 * Added Broadcom "ubsec" ENGINE to OpenSSL.
13166 *Broadcom, tweaked and integrated by Geoff Thorpe*
13168 * Move common extension printing code to new function
13169 X509V3_print_extensions(). Reorganise OCSP print routines and
13170 implement some needed OCSP ASN1 functions. Add OCSP extensions.
13174 * New function X509_signature_print() to remove duplication in some
13179 * Add a special meaning when SET OF and SEQUENCE OF flags are both
13180 set (this was treated exactly the same as SET OF previously). This
13181 is used to reorder the STACK representing the structure to match the
13182 encoding. This will be used to get round a problem where a PKCS7
13183 structure which was signed could not be verified because the STACK
13184 order did not reflect the encoded order.
13188 * Reimplement the OCSP ASN1 module using the new code.
13192 * Update the X509V3 code to permit the use of an ASN1_ITEM structure
13193 for its ASN1 operations. The old style function pointers still exist
13194 for now but they will eventually go away.
13198 * Merge in replacement ASN1 code from the ASN1 branch. This almost
13199 completely replaces the old ASN1 functionality with a table driven
13200 encoder and decoder which interprets an ASN1_ITEM structure describing
13201 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
13202 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
13203 has also been converted to the new form.
13207 * Change BN_mod_exp_recp so that negative moduli are tolerated
13208 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
13209 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
13210 for negative moduli.
13214 * Fix BN_uadd and BN_usub: Always return non-negative results instead
13215 of not touching the result's sign bit.
13219 * BN_div bugfix: If the result is 0, the sign (res->neg) must not be
13224 * Changed the LHASH code to use prototypes for callbacks, and created
13225 macros to declare and implement thin (optionally static) functions
13226 that provide type-safety and avoid function pointer casting for the
13227 type-specific callbacks.
13231 * Added Kerberos Cipher Suites to be used with TLS, as written in
13233 *Veers Staats <staatsvr@asc.hpc.mil>,
13234 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte*
13236 * Reformat the FAQ so the different questions and answers can be divided
13237 in sections depending on the subject.
13241 * Have the zlib compression code load ZLIB.DLL dynamically under
13246 * New function BN_mod_sqrt for computing square roots modulo a prime
13247 (using the probabilistic Tonelli-Shanks algorithm unless
13248 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
13249 be handled deterministically).
13251 *Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller*
13253 * Make BN_mod_inverse faster by explicitly handling small quotients
13254 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
13255 512 bits], about 30% for larger ones [1024 or 2048 bits].)
13259 * New function BN_kronecker.
13263 * Fix BN_gcd so that it works on negative inputs; the result is
13264 positive unless both parameters are zero.
13265 Previously something reasonably close to an infinite loop was
13266 possible because numbers could be growing instead of shrinking
13267 in the implementation of Euclid's algorithm.
13271 * Fix BN_is_word() and BN_is_one() macros to take into account the
13272 sign of the number in question.
13274 Fix BN_is_word(a,w) to work correctly for w == 0.
13276 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
13277 because its test if the absolute value of 'a' equals 'w'.
13278 Note that BN_abs_is_word does *not* handle w == 0 reliably;
13279 it exists mostly for use in the implementations of BN_is_zero(),
13280 BN_is_one(), and BN_is_word().
13284 * New function BN_swap.
13288 * Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
13289 the exponentiation functions are more likely to produce reasonable
13290 results on negative inputs.
13294 * Change BN_mod_mul so that the result is always non-negative.
13295 Previously, it could be negative if one of the factors was negative;
13296 I don't think anyone really wanted that behaviour.
13300 * Move `BN_mod_...` functions into new file `crypto/bn/bn_mod.c`
13301 (except for exponentiation, which stays in `crypto/bn/bn_exp.c`,
13302 and `BN_mod_mul_reciprocal`, which stays in `crypto/bn/bn_recp.c`)
13303 and add new functions:
13312 BN_mod_lshift1_quick
13314 BN_mod_lshift_quick
13316 These functions always generate non-negative results.
13318 `BN_nnmod` otherwise is `like BN_mod` (if `BN_mod` computes a remainder `r`
13319 such that `|m| < r < 0`, `BN_nnmod` will output `rem + |m|` instead).
13321 `BN_mod_XXX_quick(r, a, [b,] m)` generates the same result as
13322 `BN_mod_XXX(r, a, [b,] m, ctx)`, but requires that `a` [and `b`]
13323 be reduced modulo `m`.
13325 *Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller*
13328 The following entry accidentally appeared in the CHANGES file
13329 distributed with OpenSSL 0.9.7. The modifications described in
13330 it do *not* apply to OpenSSL 0.9.7.
13332 * Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
13333 was actually never needed) and in BN_mul(). The removal in BN_mul()
13334 required a small change in bn_mul_part_recursive() and the addition
13335 of the functions bn_cmp_part_words(), bn_sub_part_words() and
13336 bn_add_part_words(), which do the same thing as bn_cmp_words(),
13337 bn_sub_words() and bn_add_words() except they take arrays with
13343 * In 'openssl passwd', verify passwords read from the terminal
13344 unless the '-salt' option is used (which usually means that
13345 verification would just waste user's time since the resulting
13346 hash is going to be compared with some given password hash)
13347 or the new '-noverify' option is used.
13349 This is an incompatible change, but it does not affect
13350 non-interactive use of 'openssl passwd' (passwords on the command
13351 line, '-stdin' option, '-in ...' option) and thus should not
13352 cause any problems.
13356 * Remove all references to RSAref, since there's no more need for it.
13360 * Make DSO load along a path given through an environment variable
13361 (SHLIB_PATH) with shl_load().
13365 * Constify the ENGINE code as a result of BIGNUM constification.
13366 Also constify the RSA code and most things related to it. In a
13367 few places, most notable in the depth of the ASN.1 code, ugly
13368 casts back to non-const were required (to be solved at a later
13373 * Make it so the openssl application has all engines loaded by default.
13377 * Constify the BIGNUM routines a little more.
13381 * Add the following functions:
13383 ENGINE_load_cswift()
13385 ENGINE_load_atalla()
13386 ENGINE_load_nuron()
13387 ENGINE_load_builtin_engines()
13389 That way, an application can itself choose if external engines that
13390 are built-in in OpenSSL shall ever be used or not. The benefit is
13391 that applications won't have to be linked with libdl or other dso
13392 libraries unless it's really needed.
13394 Changed 'openssl engine' to load all engines on demand.
13395 Changed the engine header files to avoid the duplication of some
13396 declarations (they differed!).
13400 * 'openssl engine' can now list capabilities.
13404 * Better error reporting in 'openssl engine'.
13408 * Never call load_dh_param(NULL) in s_server.
13412 * Add engine application. It can currently list engines by name and
13413 identity, and test if they are actually available.
13417 * Improve RPM specification file by forcing symbolic linking and making
13418 sure the installed documentation is also owned by root.root.
13420 *Damien Miller <djm@mindrot.org>*
13422 * Give the OpenSSL applications more possibilities to make use of
13423 keys (public as well as private) handled by engines.
13427 * Add OCSP code that comes from CertCo.
13431 * Add VMS support for the Rijndael code.
13435 * Added untested support for Nuron crypto accelerator.
13439 * Add support for external cryptographic devices. This code was
13440 previously distributed separately as the "engine" branch.
13442 *Geoff Thorpe, Richard Levitte*
13444 * Rework the filename-translation in the DSO code. It is now possible to
13445 have far greater control over how a "name" is turned into a filename
13446 depending on the operating environment and any oddities about the
13447 different shared library filenames on each system.
13451 * Support threads on FreeBSD-elf in Configure.
13455 * Fix for SHA1 assembly problem with MASM: it produces
13456 warnings about corrupt line number information when assembling
13457 with debugging information. This is caused by the overlapping
13460 *Bernd Matthes <mainbug@celocom.de>, Steve Henson*
13463 NCONF_get_number() has no error checking at all. As a replacement,
13464 NCONF_get_number_e() is defined (`_e` for "error checking") and is
13465 promoted strongly. The old NCONF_get_number is kept around for
13466 binary backward compatibility.
13467 Make it possible for methods to load from something other than a BIO,
13468 by providing a function pointer that is given a name instead of a BIO.
13469 For example, this could be used to load configuration data from an
13474 * Fix for non blocking accept BIOs. Added new I/O special reason
13475 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
13476 with non blocking I/O was not possible because no retry code was
13477 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
13482 * Added the beginnings of Rijndael support.
13486 * Fix for bug in DirectoryString mask setting. Add support for
13487 X509_NAME_print_ex() in 'req' and X509_print_ex() function
13488 to allow certificate printing to more controllable, additional
13489 'certopt' option to 'x509' to allow new printing options to be
13494 * Clean old EAY MD5 hack from e_os.h.
13498 ### Changes between 0.9.6l and 0.9.6m [17 Mar 2004]
13500 * Fix null-pointer assignment in do_change_cipher_spec() revealed
13501 by using the Codenomicon TLS Test Tool ([CVE-2004-0079])
13503 *Joe Orton, Steve Henson*
13505 ### Changes between 0.9.6k and 0.9.6l [04 Nov 2003]
13507 * Fix additional bug revealed by the NISCC test suite:
13509 Stop bug triggering large recursion when presented with
13510 certain ASN.1 tags ([CVE-2003-0851])
13514 ### Changes between 0.9.6j and 0.9.6k [30 Sep 2003]
13516 * Fix various bugs revealed by running the NISCC test suite:
13518 Stop out of bounds reads in the ASN1 code when presented with
13519 invalid tags (CVE-2003-0543 and CVE-2003-0544).
13521 If verify callback ignores invalid public key errors don't try to check
13522 certificate signature with the NULL public key.
13526 * In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
13527 if the server requested one: as stated in TLS 1.0 and SSL 3.0
13532 * In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
13533 extra data after the compression methods not only for TLS 1.0
13534 but also for SSL 3.0 (as required by the specification).
13536 *Bodo Moeller; problem pointed out by Matthias Loepfe*
13538 * Change X509_certificate_type() to mark the key as exported/exportable
13539 when it's 512 *bits* long, not 512 bytes.
13543 ### Changes between 0.9.6i and 0.9.6j [10 Apr 2003]
13545 * Countermeasure against the Klima-Pokorny-Rosa extension of
13546 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
13547 a protocol version number mismatch like a decryption error
13548 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
13552 * Turn on RSA blinding by default in the default implementation
13553 to avoid a timing attack. Applications that don't want it can call
13554 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
13555 They would be ill-advised to do so in most cases.
13557 *Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller*
13559 * Change RSA blinding code so that it works when the PRNG is not
13560 seeded (in this case, the secret RSA exponent is abused as
13561 an unpredictable seed -- if it is not unpredictable, there
13562 is no point in blinding anyway). Make RSA blinding thread-safe
13563 by remembering the creator's thread ID in rsa->blinding and
13564 having all other threads use local one-time blinding factors
13565 (this requires more computation than sharing rsa->blinding, but
13566 avoids excessive locking; and if an RSA object is not shared
13567 between threads, blinding will still be very fast).
13571 ### Changes between 0.9.6h and 0.9.6i [19 Feb 2003]
13573 * In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
13574 via timing by performing a MAC computation even if incorrect
13575 block cipher padding has been found. This is a countermeasure
13576 against active attacks where the attacker has to distinguish
13577 between bad padding and a MAC verification error. ([CVE-2003-0078])
13579 *Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
13580 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
13581 Martin Vuagnoux (EPFL, Ilion)*
13583 ### Changes between 0.9.6g and 0.9.6h [5 Dec 2002]
13585 * New function OPENSSL_cleanse(), which is used to cleanse a section of
13586 memory from its contents. This is done with a counter that will
13587 place alternating values in each byte. This can be used to solve
13588 two issues: 1) the removal of calls to memset() by highly optimizing
13589 compilers, and 2) cleansing with other values than 0, since those can
13590 be read through on certain media, for example a swap space on disk.
13594 * Bugfix: client side session caching did not work with external caching,
13595 because the session->cipher setting was not restored when reloading
13596 from the external cache. This problem was masked, when
13597 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set.
13598 (Found by Steve Haslam <steve@araqnid.ddts.net>.)
13602 * Fix client_certificate (ssl/s2_clnt.c): The permissible total
13603 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33.
13605 *Zeev Lieber <zeev-l@yahoo.com>*
13607 * Undo an undocumented change introduced in 0.9.6e which caused
13608 repeated calls to OpenSSL_add_all_ciphers() and
13609 OpenSSL_add_all_digests() to be ignored, even after calling
13614 * Change the default configuration reader to deal with last line not
13615 being properly terminated.
13619 * Change X509_NAME_cmp() so it applies the special rules on handling
13620 DN values that are of type PrintableString, as well as RDNs of type
13621 emailAddress where the value has the type ia5String.
13623 *stefank@valicert.com via Richard Levitte*
13625 * Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half
13626 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently
13627 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be
13628 the bitwise-OR of the two for use by the majority of applications
13629 wanting this behaviour, and update the docs. The documented
13630 behaviour and actual behaviour were inconsistent and had been
13631 changing anyway, so this is more a bug-fix than a behavioural
13634 *Geoff Thorpe, diagnosed by Nadav Har'El*
13636 * Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c
13637 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes).
13641 * Fix initialization code race conditions in
13642 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(),
13643 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(),
13644 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(),
13645 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(),
13646 ssl2_get_cipher_by_char(),
13647 ssl3_get_cipher_by_char().
13649 *Patrick McCormick <patrick@tellme.com>, Bodo Moeller*
13651 * Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after
13652 the cached sessions are flushed, as the remove_cb() might use ex_data
13653 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com>
13654 (see [openssl.org #212]).
13656 *Geoff Thorpe, Lutz Jaenicke*
13658 * Fix typo in OBJ_txt2obj which incorrectly passed the content
13659 length, instead of the encoding length to d2i_ASN1_OBJECT.
13663 ### Changes between 0.9.6f and 0.9.6g [9 Aug 2002]
13665 * [In 0.9.6g-engine release:]
13666 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use `_stdcall`).
13668 *Lynn Gazis <lgazis@rainbow.com>*
13670 ### Changes between 0.9.6e and 0.9.6f [8 Aug 2002]
13672 * Fix ASN1 checks. Check for overflow by comparing with LONG_MAX
13673 and get fix the header length calculation.
13674 *Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>,
13675 Alon Kantor <alonk@checkpoint.com> (and others), Steve Henson*
13677 * Use proper error handling instead of 'assertions' in buffer
13678 overflow checks added in 0.9.6e. This prevents DoS (the
13679 assertions could call abort()).
13681 *Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller*
13683 ### Changes between 0.9.6d and 0.9.6e [30 Jul 2002]
13685 * Add various sanity checks to asn1_get_length() to reject
13686 the ASN1 length bytes if they exceed sizeof(long), will appear
13687 negative or the content length exceeds the length of the
13690 *Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>*
13692 * Fix cipher selection routines: ciphers without encryption had no flags
13693 for the cipher strength set and where therefore not handled correctly
13694 by the selection routines (PR #130).
13698 * Fix EVP_dsa_sha macro.
13703 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS
13704 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure
13705 that was added in OpenSSL 0.9.6d.
13707 As the countermeasure turned out to be incompatible with some
13708 broken SSL implementations, the new option is part of SSL_OP_ALL.
13709 SSL_OP_ALL is usually employed when compatibility with weird SSL
13710 implementations is desired (e.g. '-bugs' option to 's_client' and
13711 's_server'), so the new option is automatically set in many
13716 * Changes in security patch:
13718 Changes marked "(CHATS)" were sponsored by the Defense Advanced
13719 Research Projects Agency (DARPA) and Air Force Research Laboratory,
13720 Air Force Materiel Command, USAF, under agreement number
13723 * Add various sanity checks to asn1_get_length() to reject
13724 the ASN1 length bytes if they exceed sizeof(long), will appear
13725 negative or the content length exceeds the length of the
13726 supplied buffer. ([CVE-2002-0659])
13728 *Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>*
13730 * Assertions for various potential buffer overflows, not known to
13731 happen in practice.
13733 *Ben Laurie (CHATS)*
13735 * Various temporary buffers to hold ASCII versions of integers were
13736 too small for 64 bit platforms. ([CVE-2002-0655])
13737 *Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>*
13739 * Remote buffer overflow in SSL3 protocol - an attacker could
13740 supply an oversized session ID to a client. ([CVE-2002-0656])
13742 *Ben Laurie (CHATS)*
13744 * Remote buffer overflow in SSL2 protocol - an attacker could
13745 supply an oversized client master key. ([CVE-2002-0656])
13747 *Ben Laurie (CHATS)*
13749 ### Changes between 0.9.6c and 0.9.6d [9 May 2002]
13751 * Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
13752 encoded as NULL) with id-dsa-with-sha1.
13754 *Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller*
13756 * Check various `X509_...()` return values in `apps/req.c`.
13758 *Nils Larsch <nla@trustcenter.de>*
13760 * Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
13761 an end-of-file condition would erroneously be flagged, when the CRLF
13762 was just at the end of a processed block. The bug was discovered when
13763 processing data through a buffering memory BIO handing the data to a
13764 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
13765 <ptsekov@syntrex.com> and Nedelcho Stanev.
13769 * Implement a countermeasure against a vulnerability recently found
13770 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
13771 before application data chunks to avoid the use of known IVs
13772 with data potentially chosen by the attacker.
13776 * Fix length checks in ssl3_get_client_hello().
13780 * TLS/SSL library bugfix: use s->s3->in_read_app_data differently
13781 to prevent ssl3_read_internal() from incorrectly assuming that
13782 ssl3_read_bytes() found application data while handshake
13783 processing was enabled when in fact s->s3->in_read_app_data was
13784 merely automatically cleared during the initial handshake.
13786 *Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>*
13788 * Fix object definitions for Private and Enterprise: they were not
13789 recognized in their shortname (=lowercase) representation. Extend
13790 obj_dat.pl to issue an error when using undefined keywords instead
13791 of silently ignoring the problem (Svenning Sorensen
13792 <sss@sss.dnsalias.net>).
13796 * Fix DH_generate_parameters() so that it works for 'non-standard'
13797 generators, i.e. generators other than 2 and 5. (Previously, the
13798 code did not properly initialise the 'add' and 'rem' values to
13799 BN_generate_prime().)
13801 In the new general case, we do not insist that 'generator' is
13802 actually a primitive root: This requirement is rather pointless;
13803 a generator of the order-q subgroup is just as good, if not
13808 * Map new X509 verification errors to alerts. Discovered and submitted by
13809 Tom Wu <tom@arcot.com>.
13813 * Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
13814 returning non-zero before the data has been completely received
13815 when using non-blocking I/O.
13817 *Bodo Moeller; problem pointed out by John Hughes*
13819 * Some of the ciphers missed the strength entry (SSL_LOW etc).
13821 *Ben Laurie, Lutz Jaenicke*
13823 * Fix bug in SSL_clear(): bad sessions were not removed (found by
13824 Yoram Zahavi <YoramZ@gilian.com>).
13828 * Add information about CygWin 1.3 and on, and preserve proper
13829 configuration for the versions before that.
13831 *Corinna Vinschen <vinschen@redhat.com> and Richard Levitte*
13833 * Make removal from session cache (SSL_CTX_remove_session()) more robust:
13834 check whether we deal with a copy of a session and do not delete from
13835 the cache in this case. Problem reported by "Izhar Shoshani Levi"
13836 <izhar@checkpoint.com>.
13840 * Do not store session data into the internal session cache, if it
13841 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
13842 flag is set). Proposed by Aslam <aslam@funk.com>.
13846 * Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
13851 * [In 0.9.6d-engine release:]
13852 Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
13854 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
13856 * Add the configuration target linux-s390x.
13858 *Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte*
13860 * The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
13861 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
13862 variable as an indication that a ClientHello message has been
13863 received. As the flag value will be lost between multiple
13864 invocations of ssl3_accept when using non-blocking I/O, the
13865 function may not be aware that a handshake has actually taken
13866 place, thus preventing a new session from being added to the
13869 To avoid this problem, we now set s->new_session to 2 instead of
13870 using a local variable.
13872 *Lutz Jaenicke, Bodo Moeller*
13874 * Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
13875 if the SSL_R_LENGTH_MISMATCH error is detected.
13877 *Geoff Thorpe, Bodo Moeller*
13879 * New 'shared_ldflag' column in Configure platform table.
13883 * Fix EVP_CIPHER_mode macro.
13885 *"Dan S. Camper" <dan@bti.net>*
13887 * Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
13888 type, we must throw them away by setting rr->length to 0.
13890 *D P Chang <dpc@qualys.com>*
13892 ### Changes between 0.9.6b and 0.9.6c [21 dec 2001]
13894 * Fix BN_rand_range bug pointed out by Dominikus Scherkl
13895 <Dominikus.Scherkl@biodata.com>. (The previous implementation
13896 worked incorrectly for those cases where range = `10..._2` and
13897 `3*range` is two bits longer than range.)
13901 * Only add signing time to PKCS7 structures if it is not already
13906 * Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
13907 OBJ_ld_ce should be OBJ_id_ce.
13908 Also some ip-pda OIDs in crypto/objects/objects.txt were
13909 incorrect (cf. RFC 3039).
13911 *Matt Cooper, Frederic Giudicelli, Bodo Moeller*
13913 * Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
13914 returns early because it has nothing to do.
13916 *Andy Schneider <andy.schneider@bjss.co.uk>*
13918 * [In 0.9.6c-engine release:]
13919 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
13921 *Andy Schneider <andy.schneider@bjss.co.uk>*
13923 * [In 0.9.6c-engine release:]
13924 Add support for Cryptographic Appliance's keyserver technology.
13925 (Use engine 'keyclient')
13927 *Cryptographic Appliances and Geoff Thorpe*
13929 * Add a configuration entry for OS/390 Unix. The C compiler 'c89'
13930 is called via tools/c89.sh because arguments have to be
13931 rearranged (all '-L' options must appear before the first object
13934 *Richard Shapiro <rshapiro@abinitio.com>*
13936 * [In 0.9.6c-engine release:]
13937 Add support for Broadcom crypto accelerator cards, backported
13940 *Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox*
13942 * [In 0.9.6c-engine release:]
13943 Add support for SureWare crypto accelerator cards from
13944 Baltimore Technologies. (Use engine 'sureware')
13946 *Baltimore Technologies and Mark Cox*
13948 * [In 0.9.6c-engine release:]
13949 Add support for crypto accelerator cards from Accelerated
13950 Encryption Processing, www.aep.ie. (Use engine 'aep')
13952 *AEP Inc. and Mark Cox*
13954 * Add a configuration entry for gcc on UnixWare.
13956 *Gary Benson <gbenson@redhat.com>*
13958 * Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
13959 messages are stored in a single piece (fixed-length part and
13960 variable-length part combined) and fix various bugs found on the way.
13964 * Disable caching in BIO_gethostbyname(), directly use gethostbyname()
13965 instead. BIO_gethostbyname() does not know what timeouts are
13966 appropriate, so entries would stay in cache even when they have
13968 *Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>*
13970 * Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
13971 faced with a pathologically small ClientHello fragment that does
13972 not contain client_version: Instead of aborting with an error,
13973 simply choose the highest available protocol version (i.e.,
13974 TLS 1.0 unless it is disabled). In practice, ClientHello
13975 messages are never sent like this, but this change gives us
13976 strictly correct behaviour at least for TLS.
13980 * Fix SSL handshake functions and SSL_clear() such that SSL_clear()
13981 never resets s->method to s->ctx->method when called from within
13982 one of the SSL handshake functions.
13984 *Bodo Moeller; problem pointed out by Niko Baric*
13986 * In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
13987 (sent using the client's version number) if client_version is
13988 smaller than the protocol version in use. Also change
13989 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
13990 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
13991 the client will at least see that alert.
13995 * Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
14000 * Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
14001 client receives HelloRequest while in a handshake.
14003 *Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>*
14005 * Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
14006 should end in 'break', not 'goto end' which circumvents various
14007 cleanups done in state SSL_ST_OK. But session related stuff
14008 must be disabled for SSL_ST_OK in the case that we just sent a
14011 Also avoid some overhead by not calling ssl_init_wbio_buffer()
14012 before just sending a HelloRequest.
14014 *Bodo Moeller, Eric Rescorla <ekr@rtfm.com>*
14016 * Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
14017 reveal whether illegal block cipher padding was found or a MAC
14018 verification error occurred. (Neither SSLerr() codes nor alerts
14019 are directly visible to potential attackers, but the information
14020 may leak via logfiles.)
14022 Similar changes are not required for the SSL 2.0 implementation
14023 because the number of padding bytes is sent in clear for SSL 2.0,
14024 and the extra bytes are just ignored. However ssl/s2_pkt.c
14025 failed to verify that the purported number of padding bytes is in
14030 * Add OpenUNIX-8 support including shared libraries
14031 (Boyd Lynn Gerber <gerberb@zenez.com>).
14035 * Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
14036 'wristwatch attack' using huge encoding parameters (cf.
14037 James H. Manger's CRYPTO 2001 paper). Note that the
14038 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
14039 encoding parameters and hence was not vulnerable.
14043 * BN_sqr() bug fix.
14045 *Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>*
14047 * Rabin-Miller test analyses assume uniformly distributed witnesses,
14048 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
14049 followed by modular reduction.
14051 *Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>*
14053 * Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
14054 equivalent based on BN_pseudo_rand() instead of BN_rand().
14058 * s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
14059 This function was broken, as the check for a new client hello message
14060 to handle SGC did not allow these large messages.
14061 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
14065 * Add alert descriptions for TLSv1 to `SSL_alert_desc_string[_long]()`.
14069 * Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
14070 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
14074 * Rework the configuration and shared library support for Tru64 Unix.
14075 The configuration part makes use of modern compiler features and
14076 still retains old compiler behavior for those that run older versions
14077 of the OS. The shared library support part includes a variant that
14078 uses the RPATH feature, and is available through the special
14079 configuration target "alpha-cc-rpath", which will never be selected
14082 *Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte*
14084 * In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
14085 with the same message size as in ssl3_get_certificate_request().
14086 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
14087 messages might inadvertently be reject as too long.
14089 *Petr Lampa <lampa@fee.vutbr.cz>*
14091 * Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
14095 * Modified SSL library such that the verify_callback that has been set
14096 specifically for an SSL object with SSL_set_verify() is actually being
14097 used. Before the change, a verify_callback set with this function was
14098 ignored and the verify_callback() set in the SSL_CTX at the time of
14099 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
14100 to allow the necessary settings.
14104 * Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
14105 explicitly to NULL, as at least on Solaris 8 this seems not always to be
14106 done automatically (in contradiction to the requirements of the C
14107 standard). This made problems when used from OpenSSH.
14111 * In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
14112 dh->length and always used
14114 BN_rand_range(priv_key, dh->p).
14116 BN_rand_range() is not necessary for Diffie-Hellman, and this
14117 specific range makes Diffie-Hellman unnecessarily inefficient if
14118 dh->length (recommended exponent length) is much smaller than the
14119 length of dh->p. We could use BN_rand_range() if the order of
14120 the subgroup was stored in the DH structure, but we only have
14125 BN_rand(priv_key, l, ...)
14127 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
14134 RSA_eay_public_encrypt
14135 RSA_eay_private_decrypt
14136 RSA_eay_private_encrypt (signing)
14137 RSA_eay_public_decrypt (signature verification)
14139 (default implementations for RSA_public_encrypt,
14140 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
14141 always reject numbers >= n.
14145 * In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
14146 to synchronize access to 'locking_thread'. This is necessary on
14147 systems where access to 'locking_thread' (an 'unsigned long'
14148 variable) is not atomic.
14152 * In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
14153 *before* setting the 'crypto_lock_rand' flag. The previous code had
14154 a race condition if 0 is a valid thread ID.
14156 *Travis Vitek <vitek@roguewave.com>*
14158 * Add support for shared libraries under Irix.
14160 *Albert Chin-A-Young <china@thewrittenword.com>*
14162 * Add configuration option to build on Linux on both big-endian and
14163 little-endian MIPS.
14165 *Ralf Baechle <ralf@uni-koblenz.de>*
14167 * Add the possibility to create shared libraries on HP-UX.
14171 ### Changes between 0.9.6a and 0.9.6b [9 Jul 2001]
14173 * Change ssleay_rand_bytes (crypto/rand/md_rand.c)
14174 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
14175 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
14176 PRNG state recovery was possible based on the output of
14177 one PRNG request appropriately sized to gain knowledge on
14178 'md' followed by enough consecutive 1-byte PRNG requests
14179 to traverse all of 'state'.
14181 1. When updating 'md_local' (the current thread's copy of 'md')
14182 during PRNG output generation, hash all of the previous
14183 'md_local' value, not just the half used for PRNG output.
14185 2. Make the number of bytes from 'state' included into the hash
14186 independent from the number of PRNG bytes requested.
14188 The first measure alone would be sufficient to avoid
14189 Markku-Juhani's attack. (Actually it had never occurred
14190 to me that the half of 'md_local' used for chaining was the
14191 half from which PRNG output bytes were taken -- I had always
14192 assumed that the secret half would be used.) The second
14193 measure makes sure that additional data from 'state' is never
14194 mixed into 'md_local' in small portions; this heuristically
14195 further strengthens the PRNG.
14199 * Fix crypto/bn/asm/mips3.s.
14203 * When only the key is given to "enc", the IV is undefined. Print out
14204 an error message in this case.
14208 * Handle special case when X509_NAME is empty in X509 printing routines.
14212 * In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
14213 positive and less than q.
14217 * Don't change `*pointer` in CRYPTO_add_lock() is add_lock_callback is
14218 used: it isn't thread safe and the add_lock_callback should handle
14221 *Paul Rose <Paul.Rose@bridge.com>*
14223 * Verify that incoming data obeys the block size in
14224 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
14230 *Ulf Möller, Bodo Möller*
14232 * The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
14233 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
14234 when fixing the server behaviour for backwards-compatible 'client
14235 hello' messages. (Note that the attack is impractical against
14236 SSL 3.0 and TLS 1.0 anyway because length and version checking
14237 means that the probability of guessing a valid ciphertext is
14238 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
14241 Before 0.9.5, the countermeasure (hide the error by generating a
14242 random 'decryption result') did not work properly because
14243 ERR_clear_error() was missing, meaning that SSL_get_error() would
14244 detect the supposedly ignored error.
14246 Both problems are now fixed.
14250 * In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
14251 (previously it was 1024).
14255 * Fix for compatibility mode trust settings: ignore trust settings
14256 unless some valid trust or reject settings are present.
14260 * Fix for blowfish EVP: its a variable length cipher.
14264 * Fix various bugs related to DSA S/MIME verification. Handle missing
14265 parameters in DSA public key structures and return an error in the
14266 DSA routines if parameters are absent.
14270 * In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
14271 in the current directory if neither $RANDFILE nor $HOME was set.
14272 RAND_file_name() in 0.9.6a returned NULL in this case. This has
14273 caused some confusion to Windows users who haven't defined $HOME.
14274 Thus RAND_file_name() is changed again: e_os.h can define a
14275 DEFAULT_HOME, which will be used if $HOME is not set.
14276 For Windows, we use "C:"; on other platforms, we still require
14277 environment variables.
14279 * Move 'if (!initialized) RAND_poll()' into regions protected by
14280 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
14281 having multiple threads call RAND_poll() concurrently.
14285 * In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
14286 combination of a flag and a thread ID variable.
14287 Otherwise while one thread is in ssleay_rand_bytes (which sets the
14288 flag), *other* threads can enter ssleay_add_bytes without obeying
14289 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
14290 that they do not hold after the first thread unsets add_do_not_lock).
14294 * Change bctest again: '-x' expressions are not available in all
14295 versions of 'test'.
14299 ### Changes between 0.9.6 and 0.9.6a [5 Apr 2001]
14301 * Fix a couple of memory leaks in PKCS7_dataDecode()
14303 *Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>*
14305 * Change Configure and Makefiles to provide EXE_EXT, which will contain
14306 the default extension for executables, if any. Also, make the perl
14307 scripts that use symlink() to test if it really exists and use "cp"
14308 if it doesn't. All this made OpenSSL compilable and installable in
14313 * Fix for asn1_GetSequence() for indefinite length constructed data.
14314 If SEQUENCE is length is indefinite just set c->slen to the total
14315 amount of data available.
14317 *Steve Henson, reported by shige@FreeBSD.org*
14319 *This change does not apply to 0.9.7.*
14321 * Change bctest to avoid here-documents inside command substitution
14322 (workaround for FreeBSD /bin/sh bug).
14323 For compatibility with Ultrix, avoid shell functions (introduced
14324 in the bctest version that searches along $PATH).
14328 * Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
14329 with des_encrypt() defined on some operating systems, like Solaris
14334 * Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
14335 On the Importance of Eliminating Errors in Cryptographic
14336 Computations, J. Cryptology 14 (2001) 2, 101-119,
14337 <http://theory.stanford.edu/~dabo/papers/faults.ps.gz>).
14341 * MIPS assembler BIGNUM division bug fix.
14345 * Disabled incorrect Alpha assembler code.
14349 * Fix PKCS#7 decode routines so they correctly update the length
14350 after reading an EOC for the EXPLICIT tag.
14354 *This change does not apply to 0.9.7.*
14356 * Fix bug in PKCS#12 key generation routines. This was triggered
14357 if a 3DES key was generated with a 0 initial byte. Include
14358 PKCS12_BROKEN_KEYGEN compilation option to retain the old
14359 (but broken) behaviour.
14363 * Enhance bctest to search for a working bc along $PATH and print
14366 *Tim Rice <tim@multitalents.net> via Richard Levitte*
14368 * Fix memory leaks in err.c: free err_data string if necessary;
14369 don't write to the wrong index in ERR_set_error_data.
14373 * Implement ssl23_peek (analogous to ssl23_read), which previously
14378 * Replace rdtsc with `_emit` statements for VC++ version 5.
14380 *Jeremy Cooper <jeremy@baymoo.org>*
14382 * Make it possible to reuse SSLv2 sessions.
14386 * In copy_email() check for >= 0 as a return value for
14387 X509_NAME_get_index_by_NID() since 0 is a valid index.
14389 *Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>*
14391 * Avoid coredump with unsupported or invalid public keys by checking if
14392 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
14393 PKCS7_verify() fails with non detached data.
14397 * Don't use getenv in library functions when run as setuid/setgid.
14398 New function OPENSSL_issetugid().
14402 * Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
14403 due to incorrect handling of multi-threading:
14405 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
14407 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
14409 3. Count how many times MemCheck_off() has been called so that
14410 nested use can be treated correctly. This also avoids
14411 inband-signalling in the previous code (which relied on the
14412 assumption that thread ID 0 is impossible).
14416 * Add "-rand" option also to s_client and s_server.
14420 * Fix CPU detection on Irix 6.x.
14421 *Kurt Hockenbury <khockenb@stevens-tech.edu> and
14422 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>*
14424 * Fix X509_NAME bug which produced incorrect encoding if X509_NAME
14429 *This change does not apply to 0.9.7.*
14431 * Use the cached encoding of an X509_NAME structure rather than
14432 copying it. This is apparently the reason for the libsafe "errors"
14433 but the code is actually correct.
14437 * Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
14438 Bleichenbacher's DSA attack.
14439 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
14440 to be set and top=0 forces the highest bit to be set; top=-1 is new
14441 and leaves the highest bit random.
14443 *Ulf Moeller, Bodo Moeller*
14445 * In the `NCONF_...`-based implementations for `CONF_...` queries
14446 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
14447 a temporary CONF structure with the data component set to NULL
14448 (which gives segmentation faults in lh_retrieve).
14449 Instead, use NULL for the CONF pointer in CONF_get_string and
14450 CONF_get_number (which may use environment variables) and directly
14451 return NULL from CONF_get_section.
14455 * Fix potential buffer overrun for EBCDIC.
14459 * Tolerate nonRepudiation as being valid for S/MIME signing and certSign
14460 keyUsage if basicConstraints absent for a CA.
14464 * Make SMIME_write_PKCS7() write mail header values with a format that
14465 is more generally accepted (no spaces before the semicolon), since
14466 some programs can't parse those values properly otherwise. Also make
14467 sure BIO's that break lines after each write do not create invalid
14472 * Make the CRL encoding routines work with empty SEQUENCE OF. The
14473 macros previously used would not encode an empty SEQUENCE OF
14474 and break the signature.
14478 *This change does not apply to 0.9.7.*
14480 * Zero the premaster secret after deriving the master secret in
14485 * Add some EVP_add_digest_alias registrations (as found in
14486 OpenSSL_add_all_digests()) to SSL_library_init()
14487 aka OpenSSL_add_ssl_algorithms(). This provides improved
14488 compatibility with peers using X.509 certificates
14489 with unconventional AlgorithmIdentifier OIDs.
14493 * Fix for Irix with NO_ASM.
14495 *"Bruce W. Forsberg" <bruce.forsberg@baesystems.com>*
14497 * ./config script fixes.
14499 *Ulf Moeller, Richard Levitte*
14501 * Fix 'openssl passwd -1'.
14505 * Change PKCS12_key_gen_asc() so it can cope with non null
14506 terminated strings whose length is passed in the passlen
14507 parameter, for example from PEM callbacks. This was done
14508 by adding an extra length parameter to asc2uni().
14510 *Steve Henson, reported by <oddissey@samsung.co.kr>*
14512 * Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
14513 call failed, free the DSA structure.
14517 * Fix to uni2asc() to cope with zero length Unicode strings.
14518 These are present in some PKCS#12 files.
14522 * Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
14523 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
14524 when writing a 32767 byte record.
14526 *Bodo Moeller; problem reported by Eric Day <eday@concentric.net>*
14528 * In `RSA_eay_public_{en,ed}crypt` and RSA_eay_mod_exp (rsa_eay.c),
14529 obtain lock CRYPTO_LOCK_RSA before setting `rsa->_method_mod_{n,p,q}`.
14531 (RSA objects have a reference count access to which is protected
14532 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
14533 so they are meant to be shared between threads.)
14534 *Bodo Moeller, Geoff Thorpe; original patch submitted by
14535 "Reddie, Steven" <Steven.Reddie@ca.com>*
14537 * Fix a deadlock in CRYPTO_mem_leaks().
14541 * Use better test patterns in bntest.
14545 * rand_win.c fix for Borland C.
14549 * BN_rshift bugfix for n == 0.
14553 * Add a 'bctest' script that checks for some known 'bc' bugs
14554 so that 'make test' does not abort just because 'bc' is broken.
14558 * Store verify_result within SSL_SESSION also for client side to
14559 avoid potential security hole. (Re-used sessions on the client side
14560 always resulted in verify_result==X509_V_OK, not using the original
14561 result of the server certificate verification.)
14565 * Fix ssl3_pending: If the record in s->s3->rrec is not of type
14566 SSL3_RT_APPLICATION_DATA, return 0.
14567 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
14572 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
14573 releases, have been re-implemented by renaming the previous
14574 implementations of ssl2_read and ssl3_read to ssl2_read_internal
14575 and ssl3_read_internal, respectively, and adding 'peek' parameters
14576 to them. The new ssl[23]_{read,peek} functions are calls to
14577 ssl[23]_read_internal with the 'peek' flag set appropriately.
14578 A 'peek' parameter has also been added to ssl3_read_bytes, which
14579 does the actual work for ssl3_read_internal.
14583 * Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
14584 the method-specific "init()" handler. Also clean up ex_data after
14585 calling the method-specific "finish()" handler. Previously, this was
14586 happening the other way round.
14590 * Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
14591 The previous value, 12, was not always sufficient for BN_mod_exp().
14595 * Make sure that shared libraries get the internal name engine with
14596 the full version number and not just 0. This should mark the
14597 shared libraries as not backward compatible. Of course, this should
14598 be changed again when we can guarantee backward binary compatibility.
14602 * Fix typo in get_cert_by_subject() in by_dir.c
14604 *Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>*
14606 * Rework the system to generate shared libraries:
14608 - Make note of the expected extension for the shared libraries and
14609 if there is a need for symbolic links from for example libcrypto.so.0
14610 to libcrypto.so.0.9.7. There is extended info in Configure for
14613 - Make as few rebuilds of the shared libraries as possible.
14615 - Still avoid linking the OpenSSL programs with the shared libraries.
14617 - When installing, install the shared libraries separately from the
14622 * Fix SSL_CTX_set_read_ahead macro to actually use its argument.
14624 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
14625 and not in SSL_clear because the latter is also used by the
14626 accept/connect functions; previously, the settings made by
14627 SSL_set_read_ahead would be lost during the handshake.
14629 *Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>*
14631 * Correct util/mkdef.pl to be selective about disabled algorithms.
14632 Previously, it would create entries for disabled algorithms no
14637 * Added several new manual pages for SSL_* function.
14641 ### Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
14643 * In ssl23_get_client_hello, generate an error message when faced
14644 with an initial SSL 3.0/TLS record that is too small to contain the
14645 first two bytes of the ClientHello message, i.e. client_version.
14646 (Note that this is a pathologic case that probably has never happened
14647 in real life.) The previous approach was to use the version number
14648 from the record header as a substitute; but our protocol choice
14649 should not depend on that one because it is not authenticated
14650 by the Finished messages.
14654 * More robust randomness gathering functions for Windows.
14656 *Jeffrey Altman <jaltman@columbia.edu>*
14658 * For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
14659 not set then we don't setup the error code for issuer check errors
14660 to avoid possibly overwriting other errors which the callback does
14661 handle. If an application does set the flag then we assume it knows
14662 what it is doing and can handle the new informational codes
14667 * Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
14668 a general "ANY" type, as such it should be able to decode anything
14669 including tagged types. However it didn't check the class so it would
14670 wrongly interpret tagged types in the same way as their universal
14671 counterpart and unknown types were just rejected. Changed so that the
14672 tagged and unknown types are handled in the same way as a SEQUENCE:
14673 that is the encoding is stored intact. There is also a new type
14674 "V_ASN1_OTHER" which is used when the class is not universal, in this
14675 case we have no idea what the actual type is so we just lump them all
14680 * On VMS, stdout may very well lead to a file that is written to
14681 in a record-oriented fashion. That means that every write() will
14682 write a separate record, which will be read separately by the
14683 programs trying to read from it. This can be very confusing.
14685 The solution is to put a BIO filter in the way that will buffer
14686 text until a linefeed is reached, and then write everything a
14687 line at a time, so every record written will be an actual line,
14688 not chunks of lines and not (usually doesn't happen, but I've
14689 seen it once) several lines in one record. BIO_f_linebuffer() is
14692 Currently, it's a VMS-only method, because that's where it has
14693 been tested well enough.
14697 * Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
14698 it can return incorrect results.
14699 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
14700 but it was in 0.9.6-beta[12].)
14704 * Disable the check for content being present when verifying detached
14705 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
14706 include zero length content when signing messages.
14710 * New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
14711 BIO_ctrl (for BIO pairs).
14715 * Add DSO method for VMS.
14719 * Bug fix: Montgomery multiplication could produce results with the
14724 * Add RPM specification openssl.spec and modify it to build three
14725 packages. The default package contains applications, application
14726 documentation and run-time libraries. The devel package contains
14727 include files, static libraries and function documentation. The
14728 doc package contains the contents of the doc directory. The original
14729 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
14733 * Add a large number of documentation files for many SSL routines.
14735 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>*
14737 * Add a configuration entry for Sony News 4.
14739 *NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>*
14741 * Don't set the two most significant bits to one when generating a
14742 random number < q in the DSA library.
14746 * New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
14747 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
14748 the underlying transport is blocking) if a handshake took place.
14749 (The default behaviour is needed by applications such as s_client
14750 and s_server that use select() to determine when to use SSL_read;
14751 but for applications that know in advance when to expect data, it
14752 just makes things more complicated.)
14756 * Add RAND_egd_bytes(), which gives control over the number of bytes read
14761 * Add a few more EBCDIC conditionals that make `req` and `x509`
14762 work better on such systems.
14764 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
14766 * Add two demo programs for PKCS12_parse() and PKCS12_create().
14767 Update PKCS12_parse() so it copies the friendlyName and the
14768 keyid to the certificates aux info.
14772 * Fix bug in PKCS7_verify() which caused an infinite loop
14773 if there was more than one signature.
14775 *Sven Uszpelkat <su@celocom.de>*
14777 * Major change in util/mkdef.pl to include extra information
14778 about each symbol, as well as presenting variables as well
14779 as functions. This change means that there's n more need
14780 to rebuild the .num files when some algorithms are excluded.
14784 * Allow the verify time to be set by an application,
14785 rather than always using the current time.
14789 * Phase 2 verify code reorganisation. The certificate
14790 verify code now looks up an issuer certificate by a
14791 number of criteria: subject name, authority key id
14792 and key usage. It also verifies self signed certificates
14793 by the same criteria. The main comparison function is
14794 X509_check_issued() which performs these checks.
14796 Lot of changes were necessary in order to support this
14797 without completely rewriting the lookup code.
14799 Authority and subject key identifier are now cached.
14801 The LHASH 'certs' is X509_STORE has now been replaced
14802 by a STACK_OF(X509_OBJECT). This is mainly because an
14803 LHASH can't store or retrieve multiple objects with
14804 the same hash value.
14806 As a result various functions (which were all internal
14807 use only) have changed to handle the new X509_STORE
14808 structure. This will break anything that messed round
14809 with X509_STORE internally.
14811 The functions X509_STORE_add_cert() now checks for an
14812 exact match, rather than just subject name.
14814 The X509_STORE API doesn't directly support the retrieval
14815 of multiple certificates matching a given criteria, however
14816 this can be worked round by performing a lookup first
14817 (which will fill the cache with candidate certificates)
14818 and then examining the cache for matches. This is probably
14819 the best we can do without throwing out X509_LOOKUP
14820 entirely (maybe later...).
14822 The X509_VERIFY_CTX structure has been enhanced considerably.
14824 All certificate lookup operations now go via a get_issuer()
14825 callback. Although this currently uses an X509_STORE it
14826 can be replaced by custom lookups. This is a simple way
14827 to bypass the X509_STORE hackery necessary to make this
14828 work and makes it possible to use more efficient techniques
14829 in future. A very simple version which uses a simple
14830 STACK for its trusted certificate store is also provided
14831 using X509_STORE_CTX_trusted_stack().
14833 The verify_cb() and verify() callbacks now have equivalents
14834 in the X509_STORE_CTX structure.
14836 X509_STORE_CTX also has a 'flags' field which can be used
14837 to customise the verify behaviour.
14841 * Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
14842 excludes S/MIME capabilities.
14846 * When a certificate request is read in keep a copy of the
14847 original encoding of the signed data and use it when outputting
14848 again. Signatures then use the original encoding rather than
14849 a decoded, encoded version which may cause problems if the
14850 request is improperly encoded.
14854 * For consistency with other BIO_puts implementations, call
14855 buffer_write(b, ...) directly in buffer_puts instead of calling
14858 In BIO_puts, increment b->num_write as in BIO_write.
14860 *Peter.Sylvester@EdelWeb.fr*
14862 * Fix BN_mul_word for the case where the word is 0. (We have to use
14863 BN_zero, we may not return a BIGNUM with an array consisting of
14864 words set to zero.)
14868 * Avoid calling abort() from within the library when problems are
14869 detected, except if preprocessor symbols have been defined
14870 (such as REF_CHECK, BN_DEBUG etc.).
14874 * New openssl application 'rsautl'. This utility can be
14875 used for low-level RSA operations. DER public key
14876 BIO/fp routines also added.
14880 * New Configure entry and patches for compiling on QNX 4.
14882 *Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>*
14884 * A demo state-machine implementation was sponsored by
14885 Nuron (<http://www.nuron.com/>) and is now available in
14886 demos/state_machine.
14890 * New options added to the 'dgst' utility for signature
14891 generation and verification.
14895 * Unrecognized PKCS#7 content types are now handled via a
14896 catch all ASN1_TYPE structure. This allows unsupported
14897 types to be stored as a "blob" and an application can
14898 encode and decode it manually.
14902 * Fix various signed/unsigned issues to make a_strex.c
14903 compile under VC++.
14905 *Oscar Jacobsson <oscar.jacobsson@celocom.com>*
14907 * ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
14908 length if passed a buffer. ASN1_INTEGER_to_BN failed
14909 if passed a NULL BN and its argument was negative.
14911 *Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>*
14913 * Modification to PKCS#7 encoding routines to output definite
14914 length encoding. Since currently the whole structures are in
14915 memory there's not real point in using indefinite length
14916 constructed encoding. However if OpenSSL is compiled with
14917 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
14921 * Added BIO_vprintf() and BIO_vsnprintf().
14925 * Added more prefixes to parse for in the strings written
14926 through a logging bio, to cover all the levels that are available
14927 through syslog. The prefixes are now:
14929 PANIC, EMERG, EMR => LOG_EMERG
14930 ALERT, ALR => LOG_ALERT
14931 CRIT, CRI => LOG_CRIT
14932 ERROR, ERR => LOG_ERR
14933 WARNING, WARN, WAR => LOG_WARNING
14934 NOTICE, NOTE, NOT => LOG_NOTICE
14935 INFO, INF => LOG_INFO
14936 DEBUG, DBG => LOG_DEBUG
14938 and as before, if none of those prefixes are present at the
14939 beginning of the string, LOG_ERR is chosen.
14941 On Win32, the `LOG_*` levels are mapped according to this:
14943 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
14944 LOG_WARNING => EVENTLOG_WARNING_TYPE
14945 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
14949 * Made it possible to reconfigure with just the configuration
14950 argument "reconf" or "reconfigure". The command line arguments
14951 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
14952 and are retrieved from there when reconfiguring.
14958 *Assar Westerlund <assar@sics.se>, Richard Levitte*
14960 * Add the arguments -CAfile and -CApath to the pkcs12 utility.
14964 * The obj_dat.pl script was messing up the sorting of object
14965 names. The reason was that it compared the quoted version
14966 of strings as a result "OCSP" > "OCSP Signing" because
14967 " > SPACE. Changed script to store unquoted versions of
14968 names and add quotes on output. It was also omitting some
14969 names from the lookup table if they were given a default
14970 value (that is if SN is missing it is given the same
14971 value as LN and vice versa), these are now added on the
14972 grounds that if an object has a name we should be able to
14973 look it up. Finally added warning output when duplicate
14974 short or long names are found.
14978 * Changes needed for Tandem NSK.
14980 *Scott Uroff <scott@xypro.com>*
14982 * Fix SSL 2.0 rollback checking: Due to an off-by-one error in
14983 RSA_padding_check_SSLv23(), special padding was never detected
14984 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
14985 version rollback attacks was not effective.
14987 In s23_clnt.c, don't use special rollback-attack detection padding
14988 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
14989 client; similarly, in s23_srvr.c, don't do the rollback check if
14990 SSL 2.0 is the only protocol enabled in the server.
14994 * Make it possible to get hexdumps of unprintable data with 'openssl
14995 asn1parse'. By implication, the functions ASN1_parse_dump() and
14996 BIO_dump_indent() are added.
15000 * New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
15001 these print out strings and name structures based on various
15002 flags including RFC2253 support and proper handling of
15003 multibyte characters. Added options to the 'x509' utility
15004 to allow the various flags to be set.
15008 * Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
15009 Also change the functions X509_cmp_current_time() and
15010 X509_gmtime_adj() work with an ASN1_TIME structure,
15011 this will enable certificates using GeneralizedTime in validity
15012 dates to be checked.
15016 * Make the NEG_PUBKEY_BUG code (which tolerates invalid
15017 negative public key encodings) on by default,
15018 NO_NEG_PUBKEY_BUG can be set to disable it.
15022 * New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
15023 content octets. An i2c_ASN1_OBJECT is unnecessary because
15024 the encoding can be trivially obtained from the structure.
15028 * crypto/err.c locking bugfix: Use write locks (`CRYPTO_w_[un]lock`),
15029 not read locks (`CRYPTO_r_[un]lock`).
15033 * A first attempt at creating official support for shared
15034 libraries through configuration. I've kept it so the
15035 default is static libraries only, and the OpenSSL programs
15036 are always statically linked for now, but there are
15037 preparations for dynamic linking in place.
15038 This has been tested on Linux and Tru64.
15042 * Randomness polling function for Win9x, as described in:
15043 Peter Gutmann, Software Generation of Practically Strong
15048 * Fix so PRNG is seeded in req if using an already existing
15053 * New options to smime application. -inform and -outform
15054 allow alternative formats for the S/MIME message including
15055 PEM and DER. The -content option allows the content to be
15056 specified separately. This should allow things like Netscape
15057 form signing output easier to verify.
15061 * Fix the ASN1 encoding of tags using the 'long form'.
15065 * New ASN1 functions, `i2c_*` and `c2i_*` for INTEGER and BIT
15066 STRING types. These convert content octets to and from the
15067 underlying type. The actual tag and length octets are
15068 already assumed to have been read in and checked. These
15069 are needed because all other string types have virtually
15070 identical handling apart from the tag. By having versions
15071 of the ASN1 functions that just operate on content octets
15072 IMPLICIT tagging can be handled properly. It also allows
15073 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
15074 and ASN1_INTEGER are identical apart from the tag.
15078 * Change the handling of OID objects as follows:
15080 - New object identifiers are inserted in objects.txt, following
15081 the syntax given in [crypto/objects/README.md](crypto/objects/README.md).
15082 - objects.pl is used to process obj_mac.num and create a new
15084 - obj_dat.pl is used to create a new obj_dat.h, using the data in
15087 This is currently kind of a hack, and the perl code in objects.pl
15088 isn't very elegant, but it works as I intended. The simplest way
15089 to check that it worked correctly is to look in obj_dat.h and
15090 check the array nid_objs and make sure the objects haven't moved
15091 around (this is important!). Additions are OK, as well as
15092 consistent name changes.
15096 * Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
15100 * Addition of the command line parameter '-rand file' to 'openssl req'.
15101 The given file adds to whatever has already been seeded into the
15102 random pool through the RANDFILE configuration file option or
15103 environment variable, or the default random state file.
15107 * mkstack.pl now sorts each macro group into lexical order.
15108 Previously the output order depended on the order the files
15109 appeared in the directory, resulting in needless rewriting
15114 * Patches to make OpenSSL compile under Win32 again. Mostly
15115 work arounds for the VC++ problem that it treats func() as
15116 func(void). Also stripped out the parts of mkdef.pl that
15117 added extra typesafe functions: these no longer exist.
15121 * Reorganisation of the stack code. The macros are now all
15122 collected in safestack.h . Each macro is defined in terms of
15123 a "stack macro" of the form `SKM_<name>(type, a, b)`. The
15124 DEBUG_SAFESTACK is now handled in terms of function casts,
15125 this has the advantage of retaining type safety without the
15126 use of additional functions. If DEBUG_SAFESTACK is not defined
15127 then the non typesafe macros are used instead. Also modified the
15128 mkstack.pl script to handle the new form. Needs testing to see
15129 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
15130 the default if no major problems. Similar behaviour for ASN1_SET_OF
15131 and PKCS12_STACK_OF.
15135 * When some versions of IIS use the 'NET' form of private key the
15136 key derivation algorithm is different. Normally MD5(password) is
15137 used as a 128 bit RC4 key. In the modified case
15138 MD5(MD5(password) + "SGCKEYSALT") is used instead. Added some
15139 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
15140 as the old Netscape_RSA functions except they have an additional
15141 'sgckey' parameter which uses the modified algorithm. Also added
15142 an -sgckey command line option to the rsa utility. Thanks to
15143 Adrian Peck <bertie@ncipher.com> for posting details of the modified
15144 algorithm to openssl-dev.
15148 * The evp_local.h macros were using 'c.##kname' which resulted in
15149 invalid expansion on some systems (SCO 5.0.5 for example).
15150 Corrected to 'c.kname'.
15152 *Phillip Porch <root@theporch.com>*
15154 * New X509_get1_email() and X509_REQ_get1_email() functions that return
15155 a STACK of email addresses from a certificate or request, these look
15156 in the subject name and the subject alternative name extensions and
15157 omit any duplicate addresses.
15161 * Re-implement BN_mod_exp2_mont using independent (and larger) windows.
15162 This makes DSA verification about 2 % faster.
15166 * Increase maximum window size in `BN_mod_exp_...` to 6 bits instead of 5
15167 (meaning that now 2^5 values will be precomputed, which is only 4 KB
15168 plus overhead for 1024 bit moduli).
15169 This makes exponentiations about 0.5 % faster for 1024 bit
15170 exponents (as measured by "openssl speed rsa2048").
15174 * Rename memory handling macros to avoid conflicts with other
15176 Malloc => OPENSSL_malloc
15177 Malloc_locked => OPENSSL_malloc_locked
15178 Realloc => OPENSSL_realloc
15179 Free => OPENSSL_free
15183 * New function BN_mod_exp_mont_word for small bases (roughly 15%
15184 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
15188 * CygWin32 support.
15190 *John Jarvie <jjarvie@newsguy.com>*
15192 * The type-safe stack code has been rejigged. It is now only compiled
15193 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
15194 by default all type-specific stack functions are "#define"d back to
15195 standard stack functions. This results in more streamlined output
15196 but retains the type-safety checking possibilities of the original
15201 * The STACK code has been cleaned up, and certain type declarations
15202 that didn't make a lot of sense have been brought in line. This has
15203 also involved a cleanup of sorts in safestack.h to more correctly
15204 map type-safe stack functions onto their plain stack counterparts.
15205 This work has also resulted in a variety of "const"ifications of
15206 lots of the code, especially `_cmp` operations which should normally
15207 be prototyped with "const" parameters anyway.
15211 * When generating bytes for the first time in md_rand.c, 'stir the pool'
15212 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
15213 (The PRNG state consists of two parts, the large pool 'state' and 'md',
15214 where all of 'md' is used each time the PRNG is used, but 'state'
15215 is used only indexed by a cyclic counter. As entropy may not be
15216 well distributed from the beginning, 'md' is important as a
15217 chaining variable. However, the output function chains only half
15218 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
15219 all of 'md', and seeding with STATE_SIZE dummy bytes will result
15220 in all of 'state' being rewritten, with the new values depending
15221 on virtually all of 'md'. This overcomes the 80 bit limitation.)
15225 * In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
15226 the handshake is continued after ssl_verify_cert_chain();
15227 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
15228 can lead to 'unexplainable' connection aborts later.
15230 *Bodo Moeller; problem tracked down by Lutz Jaenicke*
15232 * Major EVP API cipher revision.
15233 Add hooks for extra EVP features. This allows various cipher
15234 parameters to be set in the EVP interface. Support added for variable
15235 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
15236 setting of RC2 and RC5 parameters.
15238 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
15241 Remove lots of duplicated code from the EVP library. For example *every*
15242 cipher init() function handles the 'iv' in the same way according to the
15243 cipher mode. They also all do nothing if the 'key' parameter is NULL and
15244 for CFB and OFB modes they zero ctx->num.
15246 New functionality allows removal of S/MIME code RC2 hack.
15248 Most of the routines have the same form and so can be declared in terms
15251 By shifting this to the top level EVP_CipherInit() it can be removed from
15252 all individual ciphers. If the cipher wants to handle IVs or keys
15253 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
15256 Change lots of functions like EVP_EncryptUpdate() to now return a
15257 value: although software versions of the algorithms cannot fail
15258 any installed hardware versions can.
15262 * Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
15263 this option is set, tolerate broken clients that send the negotiated
15264 protocol version number instead of the requested protocol version
15269 * Call dh_tmp_cb (set by `..._TMP_DH_CB`) with correct 'is_export' flag;
15270 i.e. non-zero for export ciphersuites, zero otherwise.
15271 Previous versions had this flag inverted, inconsistent with
15272 rsa_tmp_cb (..._TMP_RSA_CB).
15274 *Bodo Moeller; problem reported by Amit Chopra*
15276 * Add missing DSA library text string. Work around for some IIS
15277 key files with invalid SEQUENCE encoding.
15281 * Add a document (doc/standards.txt) that list all kinds of standards
15282 and so on that are implemented in OpenSSL.
15286 * Enhance c_rehash script. Old version would mishandle certificates
15287 with the same subject name hash and wouldn't handle CRLs at all.
15288 Added -fingerprint option to crl utility, to support new c_rehash
15293 * Eliminate non-ANSI declarations in crypto.h and stack.h.
15297 * Fix for SSL server purpose checking. Server checking was
15298 rejecting certificates which had extended key usage present
15299 but no ssl client purpose.
15301 *Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>*
15303 * Make PKCS#12 code work with no password. The PKCS#12 spec
15304 is a little unclear about how a blank password is handled.
15305 Since the password in encoded as a BMPString with terminating
15306 double NULL a zero length password would end up as just the
15307 double NULL. However no password at all is different and is
15308 handled differently in the PKCS#12 key generation code. NS
15309 treats a blank password as zero length. MSIE treats it as no
15310 password on export: but it will try both on import. We now do
15311 the same: PKCS12_parse() tries zero length and no password if
15312 the password is set to "" or NULL (NULL is now a valid password:
15313 it wasn't before) as does the pkcs12 application.
15317 * Bugfixes in `apps/x509.c`: Avoid a memory leak; and don't use
15318 perror when PEM_read_bio_X509_REQ fails, the error message must
15319 be obtained from the error queue.
15323 * Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
15324 it in ERR_remove_state if appropriate, and change ERR_get_state
15325 accordingly to avoid race conditions (this is necessary because
15326 thread_hash is no longer constant once set).
15330 * Bugfix for linux-elf makefile.one.
15334 * RSA_get_default_method() will now cause a default
15335 RSA_METHOD to be chosen if one doesn't exist already.
15336 Previously this was only set during a call to RSA_new()
15337 or RSA_new_method(NULL) meaning it was possible for
15338 RSA_get_default_method() to return NULL.
15342 * Added native name translation to the existing DSO code
15343 that will convert (if the flag to do so is set) filenames
15344 that are sufficiently small and have no path information
15345 into a canonical native form. Eg. "blah" converted to
15346 "libblah.so" or "blah.dll" etc.
15350 * New function ERR_error_string_n(e, buf, len) which is like
15351 ERR_error_string(e, buf), but writes at most 'len' bytes
15352 including the 0 terminator. For ERR_error_string_n, 'buf'
15355 *Damien Miller <djm@mindrot.org>, Bodo Moeller*
15357 * CONF library reworked to become more general. A new CONF
15358 configuration file reader "class" is implemented as well as a
15359 new functions (`NCONF_*`, for "New CONF") to handle it. The now
15360 old `CONF_*` functions are still there, but are reimplemented to
15361 work in terms of the new functions. Also, a set of functions
15362 to handle the internal storage of the configuration data is
15363 provided to make it easier to write new configuration file
15364 reader "classes" (I can definitely see something reading a
15365 configuration file in XML format, for example), called `_CONF_*`,
15366 or "the configuration storage API"...
15368 The new configuration file reading functions are:
15370 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
15371 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
15373 NCONF_default, NCONF_WIN32
15375 NCONF_dump_fp, NCONF_dump_bio
15377 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
15378 NCONF_new creates a new CONF object. This works in the same way
15379 as other interfaces in OpenSSL, like the BIO interface.
15380 `NCONF_dump_*` dump the internal storage of the configuration file,
15381 which is useful for debugging. All other functions take the same
15382 arguments as the old `CONF_*` functions with the exception of the
15383 first that must be a `CONF *` instead of a `LHASH *`.
15385 To make it easier to use the new classes with the old `CONF_*` functions,
15386 the function CONF_set_default_method is provided.
15390 * Add '-tls1' option to 'openssl ciphers', which was already
15391 mentioned in the documentation but had not been implemented.
15392 (This option is not yet really useful because even the additional
15393 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
15397 * Initial DSO code added into libcrypto for letting OpenSSL (and
15398 OpenSSL-based applications) load shared libraries and bind to
15399 them in a portable way.
15401 *Geoff Thorpe, with contributions from Richard Levitte*
15403 ### Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
15405 * Make sure _lrotl and _lrotr are only used with MSVC.
15407 * Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
15408 (the default implementation of RAND_status).
15410 * Rename openssl x509 option '-crlext', which was added in 0.9.5,
15411 to '-clrext' (= clear extensions), as intended and documented.
15412 *Bodo Moeller; inconsistency pointed out by Michael Attili
15413 <attili@amaxo.com>*
15415 * Fix for HMAC. It wasn't zeroing the rest of the block if the key length
15416 was larger than the MD block size.
15418 *Steve Henson, pointed out by Yost William <YostW@tce.com>*
15420 * Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
15421 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
15422 using the passed key: if the passed key was a private key the result
15423 of X509_print(), for example, would be to print out all the private key
15428 * des_quad_cksum() byte order bug fix.
15429 *Ulf Möller, using the problem description in krb4-0.9.7, where
15430 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>*
15432 * Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
15435 *Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>*
15437 * For easily testing in shell scripts whether some command
15438 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
15439 returns with exit code 0 iff no command of the given name is available.
15440 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
15441 the output goes to stdout and nothing is printed to stderr.
15442 Additional arguments are always ignored.
15444 Since for each cipher there is a command of the same name,
15445 the 'no-cipher' compilation switches can be tested this way.
15447 ('openssl no-XXX' is not able to detect pseudo-commands such
15448 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
15452 * Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
15456 * For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
15457 is set; it will be thrown away anyway because each handshake creates
15459 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
15460 to parameters -- in previous versions (since OpenSSL 0.9.3) the
15461 'default key' from SSL_CTX_set_tmp_dh would always be lost, meaning
15462 you effectively got SSL_OP_SINGLE_DH_USE when using this macro.
15466 * New s_client option -ign_eof: EOF at stdin is ignored, and
15467 'Q' and 'R' lose their special meanings (quit/renegotiate).
15468 This is part of what -quiet does; unlike -quiet, -ign_eof
15469 does not suppress any output.
15473 * Add compatibility options to the purpose and trust code. The
15474 purpose X509_PURPOSE_ANY is "any purpose" which automatically
15475 accepts a certificate or CA, this was the previous behaviour,
15476 with all the associated security issues.
15478 X509_TRUST_COMPAT is the old trust behaviour: only and
15479 automatically trust self signed roots in certificate store. A
15480 new trust setting X509_TRUST_DEFAULT is used to specify that
15481 a purpose has no associated trust setting and it should instead
15482 use the value in the default purpose.
15486 * Fix the PKCS#8 DSA private key code so it decodes keys again
15487 and fix a memory leak.
15491 * In util/mkerr.pl (which implements 'make errors'), preserve
15492 reason strings from the previous version of the .c file, as
15493 the default to have only downcase letters (and digits) in
15494 automatically generated reasons codes is not always appropriate.
15498 * In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
15499 using strerror. Previously, ERR_reason_error_string() returned
15500 library names as reason strings for SYSerr; but SYSerr is a special
15501 case where small numbers are errno values, not library numbers.
15505 * Add '-dsaparam' option to 'openssl dhparam' application. This
15506 converts DSA parameters into DH parameters. (When creating parameters,
15507 DSA_generate_parameters is used.)
15511 * Include 'length' (recommended exponent length) in C code generated
15512 by 'openssl dhparam -C'.
15516 * The second argument to set_label in perlasm was already being used
15517 so couldn't be used as a "file scope" flag. Moved to third argument
15522 * In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
15523 instead of RAND_bytes for encryption IVs and salts.
15527 * Include RAND_status() into RAND_METHOD instead of implementing
15528 it only for md_rand.c Otherwise replacing the PRNG by calling
15529 RAND_set_rand_method would be impossible.
15533 * Don't let DSA_generate_key() enter an infinite loop if the random
15534 number generation fails.
15538 * New 'rand' application for creating pseudo-random output.
15542 * Added configuration support for Linux/IA64
15544 *Rolf Haberrecker <rolf@suse.de>*
15546 * Assembler module support for Mingw32.
15550 * Shared library support for HPUX (in shlib/).
15552 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous*
15554 * Shared library support for Solaris gcc.
15556 *Lutz Behnke <behnke@trustcenter.de>*
15558 ### Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
15560 * PKCS7_encrypt() was adding text MIME headers twice because they
15561 were added manually and by SMIME_crlf_copy().
15565 * In bntest.c don't call BN_rand with zero bits argument.
15567 *Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>*
15569 * BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
15570 case was implemented. This caused BN_div_recp() to fail occasionally.
15574 * Add an optional second argument to the set_label() in the perl
15575 assembly language builder. If this argument exists and is set
15576 to 1 it signals that the assembler should use a symbol whose
15577 scope is the entire file, not just the current function. This
15578 is needed with MASM which uses the format label:: for this scope.
15580 *Steve Henson, pointed out by Peter Runestig <peter@runestig.com>*
15582 * Change the ASN1 types so they are typedefs by default. Before
15583 almost all types were #define'd to ASN1_STRING which was causing
15584 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
15589 * Change names of new functions to the new get1/get0 naming
15590 convention: After 'get1', the caller owns a reference count
15591 and has to call `..._free`; 'get0' returns a pointer to some
15592 data structure without incrementing reference counters.
15593 (Some of the existing 'get' functions increment a reference
15594 counter, some don't.)
15595 Similarly, 'set1' and 'add1' functions increase reference
15596 counters or duplicate objects.
15600 * Allow for the possibility of temp RSA key generation failure:
15601 the code used to assume it always worked and crashed on failure.
15605 * Fix potential buffer overrun problem in BIO_printf().
15606 *Ulf Möller, using public domain code by Patrick Powell; problem
15607 pointed out by David Sacerdote <das33@cornell.edu>*
15609 * Support EGD <http://www.lothar.com/tech/crypto/>. New functions
15610 RAND_egd() and RAND_status(). In the command line application,
15611 the EGD socket can be specified like a seed file using RANDFILE
15616 * Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
15617 Some CAs (e.g. Verisign) distribute certificates in this form.
15621 * Remove the SSL_ALLOW_ADH compile option and set the default cipher
15622 list to exclude them. This means that no special compilation option
15623 is needed to use anonymous DH: it just needs to be included in the
15628 * Change the EVP_MD_CTX_type macro so its meaning consistent with
15629 EVP_MD_type. The old functionality is available in a new macro called
15630 EVP_MD_md(). Change code that uses it and update docs.
15634 * `..._ctrl` functions now have corresponding `..._callback_ctrl` functions
15635 where the `void *` argument is replaced by a function pointer argument.
15636 Previously `void *` was abused to point to functions, which works on
15637 many platforms, but is not correct. As these functions are usually
15638 called by macros defined in OpenSSL header files, most source code
15639 should work without changes.
15643 * `<openssl/opensslconf.h>` (which is created by Configure) now contains
15644 sections with information on -D... compiler switches used for
15645 compiling the library so that applications can see them. To enable
15646 one of these sections, a pre-processor symbol `OPENSSL_..._DEFINES`
15647 must be defined. E.g.,
15648 #define OPENSSL_ALGORITHM_DEFINES
15649 #include <openssl/opensslconf.h>
15650 defines all pertinent `NO_<algo>` symbols, such as NO_IDEA, NO_RSA, etc.
15652 *Richard Levitte, Ulf and Bodo Möller*
15654 * Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
15659 * Change the 'other' type in certificate aux info to a STACK_OF
15660 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
15661 the required ASN1 format: arbitrary types determined by an OID.
15665 * Add some PEM_write_X509_REQ_NEW() functions and a command line
15666 argument to 'req'. This is not because the function is newer or
15667 better than others it just uses the work 'NEW' in the certificate
15668 request header lines. Some software needs this.
15672 * Reorganise password command line arguments: now passwords can be
15673 obtained from various sources. Delete the PEM_cb function and make
15674 it the default behaviour: i.e. if the callback is NULL and the
15675 usrdata argument is not NULL interpret it as a null terminated pass
15676 phrase. If usrdata and the callback are NULL then the pass phrase
15677 is prompted for as usual.
15681 * Add support for the Compaq Atalla crypto accelerator. If it is installed,
15682 the support is automatically enabled. The resulting binaries will
15683 autodetect the card and use it if present.
15685 *Ben Laurie and Compaq Inc.*
15687 * Work around for Netscape hang bug. This sends certificate request
15688 and server done in one record. Since this is perfectly legal in the
15689 SSL/TLS protocol it isn't a "bug" option and is on by default. See
15690 the bugs/SSLv3 entry for more info.
15694 * HP-UX tune-up: new unified configs, HP C compiler bug workaround.
15698 * Add -rand argument to smime and pkcs12 applications and read/write
15703 * New 'passwd' tool for crypt(3) and apr1 password hashes.
15707 * Add command line password options to the remaining applications.
15711 * Bug fix for BN_div_recp() for numerators with an even number of
15716 * More tests in bntest.c, and changed test_bn output.
15720 * ./config recognizes MacOS X now.
15724 * Bug fix for BN_div() when the first words of num and divisor are
15725 equal (it gave wrong results if `(rem=(n1-q*d0)&BN_MASK2) < d0)`.
15729 * Add support for various broken PKCS#8 formats, and command line
15730 options to produce them.
15734 * New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
15735 get temporary BIGNUMs from a BN_CTX.
15739 * Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
15744 * Change the `SSLeay_add_all_*()` functions to `OpenSSL_add_all_*()` and
15745 include a #define from the old name to the new. The original intent
15746 was that statically linked binaries could for example just call
15747 SSLeay_add_all_ciphers() to just add ciphers to the table and not
15748 link with digests. This never worked because SSLeay_add_all_digests()
15749 and SSLeay_add_all_ciphers() were in the same source file so calling
15750 one would link with the other. They are now in separate source files.
15754 * Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
15758 * Use a less unusual form of the Miller-Rabin primality test (it used
15759 a binary algorithm for exponentiation integrated into the Miller-Rabin
15760 loop, our standard modexp algorithms are faster).
15764 * Support for the EBCDIC character set completed.
15766 *Martin Kraemer <Martin.Kraemer@Mch.SNI.De>*
15768 * Source code cleanups: use const where appropriate, eliminate casts,
15769 use `void *` instead of `char *` in lhash.
15773 * Bugfix: ssl3_send_server_key_exchange was not restartable
15774 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
15775 this the server could overwrite ephemeral keys that the client
15780 * Turn DSA_is_prime into a macro that calls BN_is_prime,
15781 using 50 iterations of the Rabin-Miller test.
15783 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
15784 iterations of the Rabin-Miller test as required by the appendix
15785 to FIPS PUB 186[-1]) instead of DSA_is_prime.
15786 As BN_is_prime_fasttest includes trial division, DSA parameter
15787 generation becomes much faster.
15789 This implies a change for the callback functions in DSA_is_prime
15790 and DSA_generate_parameters: The callback function is called once
15791 for each positive witness in the Rabin-Miller test, not just
15792 occasionally in the inner loop; and the parameters to the
15793 callback function now provide an iteration count for the outer
15794 loop rather than for the current invocation of the inner loop.
15795 DSA_generate_parameters additionally can call the callback
15796 function with an 'iteration count' of -1, meaning that a
15797 candidate has passed the trial division test (when q is generated
15798 from an application-provided seed, trial division is skipped).
15802 * New function BN_is_prime_fasttest that optionally does trial
15803 division before starting the Rabin-Miller test and has
15804 an additional BN_CTX * argument (whereas BN_is_prime always
15805 has to allocate at least one BN_CTX).
15806 'callback(1, -1, cb_arg)' is called when a number has passed the
15807 trial division stage.
15811 * Fix for bug in CRL encoding. The validity dates weren't being handled
15816 * New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
15820 * New function BN_pseudo_rand().
15824 * Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
15825 bignum version of BN_from_montgomery() with the working code from
15826 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
15831 * Avoid a race condition in s2_clnt.c (function get_server_hello) that
15832 made it impossible to use the same SSL_SESSION data structure in
15833 SSL2 clients in multiple threads.
15837 * The return value of RAND_load_file() no longer counts bytes obtained
15838 by stat(). RAND_load_file(..., -1) is new and uses the complete file
15839 to seed the PRNG (previously an explicit byte count was required).
15841 *Ulf Möller, Bodo Möller*
15843 * Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
15844 used `char *` instead of `void *` and had casts all over the place.
15848 * Make BN_generate_prime() return NULL on error if ret!=NULL.
15852 * Retain source code compatibility for BN_prime_checks macro:
15853 BN_is_prime(..., BN_prime_checks, ...) now uses
15854 BN_prime_checks_for_size to determine the appropriate number of
15855 Rabin-Miller iterations.
15859 * Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
15860 DH_CHECK_P_NOT_SAFE_PRIME.
15861 (Check if this is true? OpenPGP calls them "strong".)
15865 * Merge the functionality of "dh" and "gendh" programs into a new program
15866 "dhparam". The old programs are retained for now but will handle DH keys
15867 (instead of parameters) in future.
15871 * Make the ciphers, s_server and s_client programs check the return values
15872 when a new cipher list is set.
15876 * Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
15877 ciphers. Before when the 56bit ciphers were enabled the sorting was
15880 The syntax for the cipher sorting has been extended to support sorting by
15881 cipher-strength (using the strength_bits hard coded in the tables).
15882 The new command is `@STRENGTH` (see also `doc/apps/ciphers.pod`).
15884 Fix a bug in the cipher-command parser: when supplying a cipher command
15885 string with an "undefined" symbol (neither command nor alphanumeric
15886 *A-Za-z0-9*, ssl_set_cipher_list used to hang in an endless loop. Now
15887 an error is flagged.
15889 Due to the strength-sorting extension, the code of the
15890 ssl_create_cipher_list() function was completely rearranged. I hope that
15891 the readability was also increased :-)
15893 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>*
15895 * Minor change to 'x509' utility. The -CAcreateserial option now uses 1
15896 for the first serial number and places 2 in the serial number file. This
15897 avoids problems when the root CA is created with serial number zero and
15898 the first user certificate has the same issuer name and serial number
15903 * Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
15904 the new code. Add documentation for this stuff.
15908 * Changes to X509_ATTRIBUTE utilities. These have been renamed from
15909 `X509_*()` to `X509at_*()` on the grounds that they don't handle X509
15910 structures and behave in an analogous way to the X509v3 functions:
15911 they shouldn't be called directly but wrapper functions should be used
15914 So we also now have some wrapper functions that call the X509at functions
15915 when passed certificate requests. (TO DO: similar things can be done with
15916 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
15917 things. Some of these need some d2i or i2d and print functionality
15918 because they handle more complex structures.)
15922 * Add missing #ifndefs that caused missing symbols when building libssl
15923 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
15924 NO_RSA in `ssl/s2*.c`.
15926 *Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller*
15928 * Precautions against using the PRNG uninitialized: RAND_bytes() now
15929 has a return value which indicates the quality of the random data
15930 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
15931 error queue. New function RAND_pseudo_bytes() generates output that is
15932 guaranteed to be unique but not unpredictable. RAND_add is like
15933 RAND_seed, but takes an extra argument for an entropy estimate
15934 (RAND_seed always assumes full entropy).
15938 * Do more iterations of Rabin-Miller probable prime test (specifically,
15939 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
15940 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
15941 in crypto/bn/bn_prime.c for the complete table). This guarantees a
15942 false-positive rate of at most 2^-80 for random input.
15946 * Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
15950 * New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
15951 in the 0.9.5 release), this returns the chain
15952 from an X509_CTX structure with a dup of the stack and all
15953 the X509 reference counts upped: so the stack will exist
15954 after X509_CTX_cleanup() has been called. Modify pkcs12.c
15957 Also make SSL_SESSION_print() print out the verify return
15962 * Add manpage for the pkcs12 command. Also change the default
15963 behaviour so MAC iteration counts are used unless the new
15964 -nomaciter option is used. This improves file security and
15965 only older versions of MSIE (4.0 for example) need it.
15969 * Honor the no-xxx Configure options when creating .DEF files.
15973 * Add PKCS#10 attributes to field table: challengePassword,
15974 unstructuredName and unstructuredAddress. These are taken from
15975 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
15976 international characters are used.
15978 More changes to X509_ATTRIBUTE code: allow the setting of types
15979 based on strings. Remove the 'loc' parameter when adding
15980 attributes because these will be a SET OF encoding which is sorted
15985 * Initial changes to the 'req' utility to allow request generation
15986 automation. This will allow an application to just generate a template
15987 file containing all the field values and have req construct the
15990 Initial support for X509_ATTRIBUTE handling. Stacks of these are
15991 used all over the place including certificate requests and PKCS#7
15992 structures. They are currently handled manually where necessary with
15993 some primitive wrappers for PKCS#7. The new functions behave in a
15994 manner analogous to the X509 extension functions: they allow
15995 attributes to be looked up by NID and added.
15997 Later something similar to the X509V3 code would be desirable to
15998 automatically handle the encoding, decoding and printing of the
15999 more complex types. The string types like challengePassword can
16000 be handled by the string table functions.
16002 Also modified the multi byte string table handling. Now there is
16003 a 'global mask' which masks out certain types. The table itself
16004 can use the flag STABLE_NO_MASK to ignore the mask setting: this
16005 is useful when for example there is only one permissible type
16006 (as in countryName) and using the mask might result in no valid
16011 * Clean up 'Finished' handling, and add functions SSL_get_finished and
16012 SSL_get_peer_finished to allow applications to obtain the latest
16013 Finished messages sent to the peer or expected from the peer,
16014 respectively. (SSL_get_peer_finished is usually the Finished message
16015 actually received from the peer, otherwise the protocol will be aborted.)
16017 As the Finished message are message digests of the complete handshake
16018 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
16019 be used for external authentication procedures when the authentication
16020 provided by SSL/TLS is not desired or is not enough.
16024 * Enhanced support for Alpha Linux is added. Now ./config checks if
16025 the host supports BWX extension and if Compaq C is present on the
16026 $PATH. Just exploiting of the BWX extension results in 20-30%
16027 performance kick for some algorithms, e.g. DES and RC4 to mention
16028 a couple. Compaq C in turn generates ~20% faster code for MD5 and
16033 * Add support for MS "fast SGC". This is arguably a violation of the
16034 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
16035 weak crypto and after checking the certificate is SGC a second one
16036 with strong crypto. MS SGC stops the first handshake after receiving
16037 the server certificate message and sends a second client hello. Since
16038 a server will typically do all the time consuming operations before
16039 expecting any further messages from the client (server key exchange
16040 is the most expensive) there is little difference between the two.
16042 To get OpenSSL to support MS SGC we have to permit a second client
16043 hello message after we have sent server done. In addition we have to
16044 reset the MAC if we do get this second client hello.
16048 * Add a function 'd2i_AutoPrivateKey()' this will automatically decide
16049 if a DER encoded private key is RSA or DSA traditional format. Changed
16050 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
16051 format DER encoded private key. Newer code should use PKCS#8 format which
16052 has the key type encoded in the ASN1 structure. Added DER private key
16053 support to pkcs8 application.
16057 * SSL 3/TLS 1 servers now don't request certificates when an anonymous
16058 ciphersuites has been selected (as required by the SSL 3/TLS 1
16059 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
16060 is set, we interpret this as a request to violate the specification
16061 (the worst that can happen is a handshake failure, and 'correct'
16062 behaviour would result in a handshake failure anyway).
16066 * In SSL_CTX_add_session, take into account that there might be multiple
16067 SSL_SESSION structures with the same session ID (e.g. when two threads
16068 concurrently obtain them from an external cache).
16069 The internal cache can handle only one SSL_SESSION with a given ID,
16070 so if there's a conflict, we now throw out the old one to achieve
16075 * Add OIDs for idea and blowfish in CBC mode. This will allow both
16076 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
16077 some routines that use cipher OIDs: some ciphers do not have OIDs
16078 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
16083 * Simplify the trust setting structure and code. Now we just have
16084 two sequences of OIDs for trusted and rejected settings. These will
16085 typically have values the same as the extended key usage extension
16086 and any application specific purposes.
16088 The trust checking code now has a default behaviour: it will just
16089 check for an object with the same NID as the passed id. Functions can
16090 be provided to override either the default behaviour or the behaviour
16091 for a given id. SSL client, server and email already have functions
16092 in place for compatibility: they check the NID and also return "trusted"
16093 if the certificate is self signed.
16097 * Add d2i,i2d bio/fp functions for PrivateKey: these convert the
16098 traditional format into an EVP_PKEY structure.
16102 * Add a password callback function PEM_cb() which either prompts for
16103 a password if usr_data is NULL or otherwise assumes it is a null
16104 terminated password. Allow passwords to be passed on command line
16105 environment or config files in a few more utilities.
16109 * Add a bunch of DER and PEM functions to handle PKCS#8 format private
16110 keys. Add some short names for PKCS#8 PBE algorithms and allow them
16111 to be specified on the command line for the pkcs8 and pkcs12 utilities.
16112 Update documentation.
16116 * Support for ASN1 "NULL" type. This could be handled before by using
16117 ASN1_TYPE but there wasn't any function that would try to read a NULL
16118 and produce an error if it couldn't. For compatibility we also have
16119 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
16120 don't allocate anything because they don't need to.
16124 * Initial support for MacOS is now provided. Examine INSTALL.MacOS
16127 *Andy Polyakov, Roy Woods <roy@centicsystems.ca>*
16129 * Rebuild of the memory allocation routines used by OpenSSL code and
16130 possibly others as well. The purpose is to make an interface that
16131 provide hooks so anyone can build a separate set of allocation and
16132 deallocation routines to be used by OpenSSL, for example memory
16133 pool implementations, or something else, which was previously hard
16134 since Malloc(), Realloc() and Free() were defined as macros having
16135 the values malloc, realloc and free, respectively (except for Win32
16136 compilations). The same is provided for memory debugging code.
16137 OpenSSL already comes with functionality to find memory leaks, but
16138 this gives people a chance to debug other memory problems.
16140 With these changes, a new set of functions and macros have appeared:
16142 CRYPTO_set_mem_debug_functions() [F]
16143 CRYPTO_get_mem_debug_functions() [F]
16144 CRYPTO_dbg_set_options() [F]
16145 CRYPTO_dbg_get_options() [F]
16146 CRYPTO_malloc_debug_init() [M]
16148 The memory debug functions are NULL by default, unless the library
16149 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
16150 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
16151 gives the standard debugging functions that come with OpenSSL) or
16152 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
16153 provided by the library user) must be used. When the standard
16154 debugging functions are used, CRYPTO_dbg_set_options can be used to
16155 request additional information:
16156 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
16157 the CRYPTO_MDEBUG_xxx macro when compiling the library.
16159 Also, things like CRYPTO_set_mem_functions will always give the
16160 expected result (the new set of functions is used for allocation
16161 and deallocation) at all times, regardless of platform and compiler
16164 To finish it up, some functions that were never use in any other
16165 way than through macros have a new API and new semantic:
16167 CRYPTO_dbg_malloc()
16168 CRYPTO_dbg_realloc()
16171 All macros of value have retained their old syntax.
16173 *Richard Levitte and Bodo Moeller*
16175 * Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
16176 ordering of SMIMECapabilities wasn't in "strength order" and there
16177 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
16182 * Some ASN1 types with illegal zero length encoding (INTEGER,
16183 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
16185 *Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson*
16187 * Merge in my S/MIME library for OpenSSL. This provides a simple
16188 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
16189 functionality to handle multipart/signed properly) and a utility
16190 called 'smime' to call all this stuff. This is based on code I
16191 originally wrote for Celo who have kindly allowed it to be
16192 included in OpenSSL.
16196 * Add variants des_set_key_checked and des_set_key_unchecked of
16197 des_set_key (aka des_key_sched). Global variable des_check_key
16198 decides which of these is called by des_set_key; this way
16199 des_check_key behaves as it always did, but applications and
16200 the library itself, which was buggy for des_check_key == 1,
16201 have a cleaner way to pick the version they need.
16205 * New function PKCS12_newpass() which changes the password of a
16210 * Modify X509_TRUST and X509_PURPOSE so it also uses a static and
16211 dynamic mix. In both cases the ids can be used as an index into the
16212 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
16213 functions so they accept a list of the field values and the
16214 application doesn't need to directly manipulate the X509_TRUST
16219 * Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
16224 * Modify the way the V3 extension code looks up extensions. This now
16225 works in a similar way to the object code: we have some "standard"
16226 extensions in a static table which is searched with OBJ_bsearch()
16227 and the application can add dynamic ones if needed. The file
16228 crypto/x509v3/ext_dat.h now has the info: this file needs to be
16229 updated whenever a new extension is added to the core code and kept
16230 in ext_nid order. There is a simple program 'tabtest.c' which checks
16231 this. New extensions are not added too often so this file can readily
16232 be maintained manually.
16234 There are two big advantages in doing things this way. The extensions
16235 can be looked up immediately and no longer need to be "added" using
16236 X509V3_add_standard_extensions(): this function now does nothing.
16237 Side note: I get *lots* of email saying the extension code doesn't
16238 work because people forget to call this function.
16239 Also no dynamic allocation is done unless new extensions are added:
16240 so if we don't add custom extensions there is no need to call
16241 X509V3_EXT_cleanup().
16245 * Modify enc utility's salting as follows: make salting the default. Add a
16246 magic header, so unsalted files fail gracefully instead of just decrypting
16247 to garbage. This is because not salting is a big security hole, so people
16248 should be discouraged from doing it.
16252 * Fixes and enhancements to the 'x509' utility. It allowed a message
16253 digest to be passed on the command line but it only used this
16254 parameter when signing a certificate. Modified so all relevant
16255 operations are affected by the digest parameter including the
16256 -fingerprint and -x509toreq options. Also -x509toreq choked if a
16257 DSA key was used because it didn't fix the digest.
16261 * Initial certificate chain verify code. Currently tests the untrusted
16262 certificates for consistency with the verify purpose (which is set
16263 when the X509_STORE_CTX structure is set up) and checks the pathlength.
16265 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
16266 this is because it will reject chains with invalid extensions whereas
16267 every previous version of OpenSSL and SSLeay made no checks at all.
16269 Trust code: checks the root CA for the relevant trust settings. Trust
16270 settings have an initial value consistent with the verify purpose: e.g.
16271 if the verify purpose is for SSL client use it expects the CA to be
16272 trusted for SSL client use. However the default value can be changed to
16273 permit custom trust settings: one example of this would be to only trust
16274 certificates from a specific "secure" set of CAs.
16276 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
16277 which should be used for version portability: especially since the
16278 verify structure is likely to change more often now.
16280 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
16281 to set them. If not set then assume SSL clients will verify SSL servers
16284 Two new options to the verify program: -untrusted allows a set of
16285 untrusted certificates to be passed in and -purpose which sets the
16286 intended purpose of the certificate. If a purpose is set then the
16287 new chain verify code is used to check extension consistency.
16291 * Support for the authority information access extension.
16295 * Modify RSA and DSA PEM read routines to transparently handle
16296 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
16297 public keys in a format compatible with certificate
16298 SubjectPublicKeyInfo structures. Unfortunately there were already
16299 functions called *_PublicKey_* which used various odd formats so
16300 these are retained for compatibility: however the DSA variants were
16301 never in a public release so they have been deleted. Changed dsa/rsa
16302 utilities to handle the new format: note no releases ever handled public
16303 keys so we should be OK.
16305 The primary motivation for this change is to avoid the same fiasco
16306 that dogs private keys: there are several incompatible private key
16307 formats some of which are standard and some OpenSSL specific and
16308 require various evil hacks to allow partial transparent handling and
16309 even then it doesn't work with DER formats. Given the option anything
16310 other than PKCS#8 should be dumped: but the other formats have to
16311 stay in the name of compatibility.
16313 With public keys and the benefit of hindsight one standard format
16314 is used which works with EVP_PKEY, RSA or DSA structures: though
16315 it clearly returns an error if you try to read the wrong kind of key.
16317 Added a -pubkey option to the 'x509' utility to output the public key.
16318 Also rename the `EVP_PKEY_get_*()` to `EVP_PKEY_rget_*()`
16319 (renamed to `EVP_PKEY_get1_*()` in the OpenSSL 0.9.5 release) and add
16320 `EVP_PKEY_rset_*()` functions (renamed to `EVP_PKEY_set1_*()`)
16321 that do the same as the `EVP_PKEY_assign_*()` except they up the
16322 reference count of the added key (they don't "swallow" the
16327 * Fixes to crypto/x509/by_file.c the code to read in certificates and
16328 CRLs would fail if the file contained no certificates or no CRLs:
16329 added a new function to read in both types and return the number
16330 read: this means that if none are read it will be an error. The
16331 DER versions of the certificate and CRL reader would always fail
16332 because it isn't possible to mix certificates and CRLs in DER format
16333 without choking one or the other routine. Changed this to just read
16334 a certificate: this is the best we can do. Also modified the code
16335 in `apps/verify.c` to take notice of return codes: it was previously
16336 attempting to read in certificates from NULL pointers and ignoring
16337 any errors: this is one reason why the cert and CRL reader seemed
16338 to work. It doesn't check return codes from the default certificate
16339 routines: these may well fail if the certificates aren't installed.
16343 * Code to support otherName option in GeneralName.
16347 * First update to verify code. Change the verify utility
16348 so it warns if it is passed a self signed certificate:
16349 for consistency with the normal behaviour. X509_verify
16350 has been modified to it will now verify a self signed
16351 certificate if *exactly* the same certificate appears
16352 in the store: it was previously impossible to trust a
16353 single self signed certificate. This means that:
16354 openssl verify ss.pem
16355 now gives a warning about a self signed certificate but
16356 openssl verify -CAfile ss.pem ss.pem
16361 * For servers, store verify_result in SSL_SESSION data structure
16362 (and add it to external session representation).
16363 This is needed when client certificate verifications fails,
16364 but an application-provided verification callback (set by
16365 SSL_CTX_set_cert_verify_callback) allows accepting the session
16366 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
16367 but returns 1): When the session is reused, we have to set
16368 ssl->verify_result to the appropriate error code to avoid
16371 *Bodo Moeller, problem pointed out by Lutz Jaenicke*
16373 * Fix a bug in the new PKCS#7 code: it didn't consider the
16374 case in PKCS7_dataInit() where the signed PKCS7 structure
16375 didn't contain any existing data because it was being created.
16377 *Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson*
16379 * Add a salt to the key derivation routines in enc.c. This
16380 forms the first 8 bytes of the encrypted file. Also add a
16381 -S option to allow a salt to be input on the command line.
16385 * New function X509_cmp(). Oddly enough there wasn't a function
16386 to compare two certificates. We do this by working out the SHA1
16387 hash and comparing that. X509_cmp() will be needed by the trust
16392 * SSL_get1_session() is like SSL_get_session(), but increments
16393 the reference count in the SSL_SESSION returned.
16395 *Geoff Thorpe <geoff@eu.c2.net>*
16397 * Fix for 'req': it was adding a null to request attributes.
16398 Also change the X509_LOOKUP and X509_INFO code to handle
16399 certificate auxiliary information.
16403 * Add support for 40 and 64 bit RC2 and RC4 algorithms: document
16408 * Add the possibility to add extra information to the memory leak
16409 detecting output, to form tracebacks, showing from where each
16410 allocation was originated: CRYPTO_push_info("constant string") adds
16411 the string plus current file name and line number to a per-thread
16412 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
16413 is like calling CYRPTO_pop_info() until the stack is empty.
16414 Also updated memory leak detection code to be multi-thread-safe.
16418 * Add options -text and -noout to pkcs7 utility and delete the
16419 encryption options which never did anything. Update docs.
16423 * Add options to some of the utilities to allow the pass phrase
16424 to be included on either the command line (not recommended on
16425 OSes like Unix) or read from the environment. Update the
16426 manpages and fix a few bugs.
16430 * Add a few manpages for some of the openssl commands.
16434 * Fix the -revoke option in ca. It was freeing up memory twice,
16435 leaking and not finding already revoked certificates.
16439 * Extensive changes to support certificate auxiliary information.
16440 This involves the use of X509_CERT_AUX structure and X509_AUX
16441 functions. An X509_AUX function such as PEM_read_X509_AUX()
16442 can still read in a certificate file in the usual way but it
16443 will also read in any additional "auxiliary information". By
16444 doing things this way a fair degree of compatibility can be
16445 retained: existing certificates can have this information added
16446 using the new 'x509' options.
16448 Current auxiliary information includes an "alias" and some trust
16449 settings. The trust settings will ultimately be used in enhanced
16450 certificate chain verification routines: currently a certificate
16451 can only be trusted if it is self signed and then it is trusted
16456 * Fix assembler for Alpha (tested only on DEC OSF not Linux or `*BSD`).
16457 The problem was that one of the replacement routines had not been working
16458 since SSLeay releases. For now the offending routine has been replaced
16459 with non-optimised assembler. Even so, this now gives around 95%
16460 performance improvement for 1024 bit RSA signs.
16464 * Hack to fix PKCS#7 decryption when used with some unorthodox RC2
16465 handling. Most clients have the effective key size in bits equal to
16466 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
16467 A few however don't do this and instead use the size of the decrypted key
16468 to determine the RC2 key length and the AlgorithmIdentifier to determine
16469 the effective key length. In this case the effective key length can still
16470 be 40 bits but the key length can be 168 bits for example. This is fixed
16471 by manually forcing an RC2 key into the EVP_PKEY structure because the
16472 EVP code can't currently handle unusual RC2 key sizes: it always assumes
16473 the key length and effective key length are equal.
16477 * Add a bunch of functions that should simplify the creation of
16478 X509_NAME structures. Now you should be able to do:
16479 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
16480 and have it automatically work out the correct field type and fill in
16481 the structures. The more adventurous can try:
16482 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
16483 and it will (hopefully) work out the correct multibyte encoding.
16487 * Change the 'req' utility to use the new field handling and multibyte
16488 copy routines. Before the DN field creation was handled in an ad hoc
16489 way in req, ca, and x509 which was rather broken and didn't support
16490 BMPStrings or UTF8Strings. Since some software doesn't implement
16491 BMPStrings or UTF8Strings yet, they can be enabled using the config file
16492 using the dirstring_type option. See the new comment in the default
16493 openssl.cnf for more info.
16497 * Make crypto/rand/md_rand.c more robust:
16498 - Assure unique random numbers after fork().
16499 - Make sure that concurrent threads access the global counter and
16500 md serializably so that we never lose entropy in them
16501 or use exactly the same state in multiple threads.
16502 Access to the large state is not always serializable because
16503 the additional locking could be a performance killer, and
16504 md should be large enough anyway.
16508 * New file `apps/app_rand.c` with commonly needed functionality
16509 for handling the random seed file.
16511 Use the random seed file in some applications that previously did not:
16513 dsaparam -genkey (which also ignored its '-rand' option),
16516 x509 (when signing).
16517 Except on systems with /dev/urandom, it is crucial to have a random
16518 seed file at least for key creation, DSA signing, and for DH exchanges;
16519 for RSA signatures we could do without one.
16521 gendh and gendsa (unlike genrsa) used to read only the first byte
16522 of each file listed in the '-rand' option. The function as previously
16523 found in genrsa is now in app_rand.c and is used by all programs
16524 that support '-rand'.
16528 * In RAND_write_file, use mode 0600 for creating files;
16529 don't just chmod when it may be too late.
16533 * Report an error from X509_STORE_load_locations
16534 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
16538 * New function ASN1_mbstring_copy() this copies a string in either
16539 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
16540 into an ASN1_STRING type. A mask of permissible types is passed
16541 and it chooses the "minimal" type to use or an error if not type
16546 * Add function equivalents to the various macros in asn1.h. The old
16547 macros are retained with an `M_` prefix. Code inside the library can
16548 use the `M_` macros. External code (including the openssl utility)
16549 should *NOT* in order to be "shared library friendly".
16553 * Add various functions that can check a certificate's extensions
16554 to see if it usable for various purposes such as SSL client,
16555 server or S/MIME and CAs of these types. This is currently
16556 VERY EXPERIMENTAL but will ultimately be used for certificate chain
16557 verification. Also added a -purpose flag to x509 utility to
16558 print out all the purposes.
16562 * Add a CRYPTO_EX_DATA to X509 certificate structure and associated
16567 * New `X509V3_{X509,CRL,REVOKED}_get_d2i()` functions. These will search
16568 for, obtain and decode and extension and obtain its critical flag.
16569 This allows all the necessary extension code to be handled in a
16570 single function call.
16574 * RC4 tune-up featuring 30-40% performance improvement on most RISC
16575 platforms. See crypto/rc4/rc4_enc.c for further details.
16579 * New -noout option to asn1parse. This causes no output to be produced
16580 its main use is when combined with -strparse and -out to extract data
16581 from a file (which may not be in ASN.1 format).
16585 * Fix for pkcs12 program. It was hashing an invalid certificate pointer
16586 when producing the local key id.
16588 *Richard Levitte <levitte@stacken.kth.se>*
16590 * New option -dhparam in s_server. This allows a DH parameter file to be
16591 stated explicitly. If it is not stated then it tries the first server
16592 certificate file. The previous behaviour hard coded the filename
16597 * Add -pubin and -pubout options to the rsa and dsa commands. These allow
16598 a public key to be input or output. For example:
16599 openssl rsa -in key.pem -pubout -out pubkey.pem
16600 Also added necessary DSA public key functions to handle this.
16604 * Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
16605 in the message. This was handled by allowing
16606 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
16608 *Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>*
16610 * Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
16611 to the end of the strings whereas this didn't. This would cause problems
16612 if strings read with d2i_ASN1_bytes() were later modified.
16614 *Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>*
16616 * Fix for base64 decode bug. When a base64 bio reads only one line of
16617 data and it contains EOF it will end up returning an error. This is
16618 caused by input 46 bytes long. The cause is due to the way base64
16619 BIOs find the start of base64 encoded data. They do this by trying a
16620 trial decode on each line until they find one that works. When they
16621 do a flag is set and it starts again knowing it can pass all the
16622 data directly through the decoder. Unfortunately it doesn't reset
16623 the context it uses. This means that if EOF is reached an attempt
16624 is made to pass two EOFs through the context and this causes the
16625 resulting error. This can also cause other problems as well. As is
16626 usual with these problems it takes *ages* to find and the fix is
16627 trivial: move one line.
16629 *Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer)*
16631 * Ugly workaround to get s_client and s_server working under Windows. The
16632 old code wouldn't work because it needed to select() on sockets and the
16633 tty (for keypresses and to see if data could be written). Win32 only
16634 supports select() on sockets so we select() with a 1s timeout on the
16635 sockets and then see if any characters are waiting to be read, if none
16636 are present then we retry, we also assume we can always write data to
16637 the tty. This isn't nice because the code then blocks until we've
16638 received a complete line of data and it is effectively polling the
16639 keyboard at 1s intervals: however it's quite a bit better than not
16640 working at all :-) A dedicated Windows application might handle this
16641 with an event loop for example.
16645 * Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
16646 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
16647 will be called when RSA_sign() and RSA_verify() are used. This is useful
16648 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
16649 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
16650 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
16651 This necessitated the support of an extra signature type NID_md5_sha1
16652 for SSL signatures and modifications to the SSL library to use it instead
16653 of calling RSA_public_decrypt() and RSA_private_encrypt().
16657 * Add new -verify -CAfile and -CApath options to the crl program, these
16658 will lookup a CRL issuers certificate and verify the signature in a
16659 similar way to the verify program. Tidy up the crl program so it
16660 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
16661 less strict. It will now permit CRL extensions even if it is not
16662 a V2 CRL: this will allow it to tolerate some broken CRLs.
16666 * Initialize all non-automatic variables each time one of the openssl
16667 sub-programs is started (this is necessary as they may be started
16668 multiple times from the "OpenSSL>" prompt).
16670 *Lennart Bang, Bodo Moeller*
16672 * Preliminary compilation option RSA_NULL which disables RSA crypto without
16673 removing all other RSA functionality (this is what NO_RSA does). This
16674 is so (for example) those in the US can disable those operations covered
16675 by the RSA patent while allowing storage and parsing of RSA keys and RSA
16680 * Non-copying interface to BIO pairs.
16681 (still largely untested)
16685 * New function ASN1_tag2str() to convert an ASN1 tag to a descriptive
16686 ASCII string. This was handled independently in various places before.
16690 * New functions UTF8_getc() and UTF8_putc() that parse and generate
16691 UTF8 strings a character at a time.
16695 * Use client_version from client hello to select the protocol
16696 (s23_srvr.c) and for RSA client key exchange verification
16697 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
16701 * Add various utility functions to handle SPKACs, these were previously
16702 handled by poking round in the structure internals. Added new function
16703 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
16704 print, verify and generate SPKACs. Based on an original idea from
16705 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
16709 * RIPEMD160 is operational on all platforms and is back in 'make test'.
16713 * Allow the config file extension section to be overwritten on the
16714 command line. Based on an original idea from Massimiliano Pala
16715 <madwolf@comune.modena.it>. The new option is called -extensions
16716 and can be applied to ca, req and x509. Also -reqexts to override
16717 the request extensions in req and -crlexts to override the crl extensions
16722 * Add new feature to the SPKAC handling in ca. Now you can include
16723 the same field multiple times by preceding it by "XXXX." for example:
16726 this is the same syntax as used in the req config file.
16730 * Allow certificate extensions to be added to certificate requests. These
16731 are specified in a 'req_extensions' option of the req section of the
16732 config file. They can be printed out with the -text option to req but
16733 are otherwise ignored at present.
16737 * Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
16738 data read consists of only the final block it would not decrypted because
16739 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
16740 A misplaced 'break' also meant the decrypted final block might not be
16741 copied until the next read.
16745 * Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
16746 a few extra parameters to the DH structure: these will be useful if
16747 for example we want the value of 'q' or implement X9.42 DH.
16751 * Initial support for DSA_METHOD. This is based on the RSA_METHOD and
16752 provides hooks that allow the default DSA functions or functions on a
16753 "per key" basis to be replaced. This allows hardware acceleration and
16754 hardware key storage to be handled without major modification to the
16755 library. Also added low-level modexp hooks and CRYPTO_EX structure and
16756 associated functions.
16760 * Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
16761 as "read only": it can't be written to and the buffer it points to will
16762 not be freed. Reading from a read only BIO is much more efficient than
16763 a normal memory BIO. This was added because there are several times when
16764 an area of memory needs to be read from a BIO. The previous method was
16765 to create a memory BIO and write the data to it, this results in two
16766 copies of the data and an O(n^2) reading algorithm. There is a new
16767 function BIO_new_mem_buf() which creates a read only memory BIO from
16768 an area of memory. Also modified the PKCS#7 routines to use read only
16773 * Bugfix: ssl23_get_client_hello did not work properly when called in
16774 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
16775 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
16776 but a retry condition occurred while trying to read the rest.
16780 * The PKCS7_ENC_CONTENT_new() function was setting the content type as
16781 NID_pkcs7_encrypted by default: this was wrong since this should almost
16782 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
16783 the encrypted data type: this is a more sensible place to put it and it
16784 allows the PKCS#12 code to be tidied up that duplicated this
16789 * Changed obj_dat.pl script so it takes its input and output files on
16790 the command line. This should avoid shell escape redirection problems
16795 * Initial support for certificate extension requests, these are included
16796 in things like Xenroll certificate requests. Included functions to allow
16797 extensions to be obtained and added.
16801 * -crlf option to s_client and s_server for sending newlines as
16802 CRLF (as required by many protocols).
16806 ### Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
16808 * Install libRSAglue.a when OpenSSL is built with RSAref.
16810 *Ralf S. Engelschall*
16812 * A few more `#ifndef NO_FP_API / #endif` pairs for consistency.
16814 *Andrija Antonijevic <TheAntony2@bigfoot.com>*
16816 * Fix -startdate and -enddate (which was missing) arguments to 'ca'
16821 * New function DSA_dup_DH, which duplicates DSA parameters/keys as
16822 DH parameters/keys (q is lost during that conversion, but the resulting
16823 DH parameters contain its length).
16825 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
16826 much faster than DH_generate_parameters (which creates parameters
16827 where `p = 2*q + 1`), and also the smaller q makes DH computations
16828 much more efficient (160-bit exponentiation instead of 1024-bit
16829 exponentiation); so this provides a convenient way to support DHE
16830 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
16831 utter importance to use
16832 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
16834 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
16835 when such DH parameters are used, because otherwise small subgroup
16836 attacks may become possible!
16840 * Avoid memory leak in i2d_DHparams.
16844 * Allow the -k option to be used more than once in the enc program:
16845 this allows the same encrypted message to be read by multiple recipients.
16849 * New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
16850 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
16851 it will always use the numerical form of the OID, even if it has a short
16856 * Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
16857 method only got called if p,q,dmp1,dmq1,iqmp components were present,
16858 otherwise bn_mod_exp was called. In the case of hardware keys for example
16859 no private key components need be present and it might store extra data
16860 in the RSA structure, which cannot be accessed from bn_mod_exp.
16861 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
16862 private key operations.
16866 * Added support for SPARC Linux.
16870 * pem_password_cb function type incompatibly changed from
16871 typedef int pem_password_cb(char *buf, int size, int rwflag);
16873 ....(char *buf, int size, int rwflag, void *userdata);
16874 so that applications can pass data to their callbacks:
16875 The `PEM[_ASN1]_{read,write}...` functions and macros now take an
16876 additional void * argument, which is just handed through whenever
16877 the password callback is called.
16879 *Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller*
16881 New function SSL_CTX_set_default_passwd_cb_userdata.
16883 Compatibility note: As many C implementations push function arguments
16884 onto the stack in reverse order, the new library version is likely to
16885 interoperate with programs that have been compiled with the old
16886 pem_password_cb definition (PEM_whatever takes some data that
16887 happens to be on the stack as its last argument, and the callback
16888 just ignores this garbage); but there is no guarantee whatsoever that
16891 * The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
16892 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
16893 problems not only on Windows, but also on some Unix platforms.
16894 To avoid problematic command lines, these definitions are now in an
16895 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
16896 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
16900 * MIPS III/IV assembler module is reimplemented.
16904 * More DES library cleanups: remove references to srand/rand and
16905 delete an unused file.
16909 * Add support for the free Netwide assembler (NASM) under Win32,
16910 since not many people have MASM (ml) and it can be hard to obtain.
16911 This is currently experimental but it seems to work OK and pass all
16912 the tests. Check out INSTALL.W32 for info.
16916 * Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
16917 without temporary keys kept an extra copy of the server key,
16918 and connections with temporary keys did not free everything in case
16923 * New function RSA_check_key and new openssl rsa option -check
16924 for verifying the consistency of RSA keys.
16926 *Ulf Moeller, Bodo Moeller*
16928 * Various changes to make Win32 compile work:
16929 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
16930 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
16931 comparison" warnings.
16932 3. Add `sk_<TYPE>_sort` to DEF file generator and do make update.
16936 * Add a debugging option to PKCS#5 v2 key generation function: when
16937 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
16938 derived keys are printed to stderr.
16942 * Copy the flags in ASN1_STRING_dup().
16944 *Roman E. Pavlov <pre@mo.msk.ru>*
16946 * The x509 application mishandled signing requests containing DSA
16947 keys when the signing key was also DSA and the parameters didn't match.
16949 It was supposed to omit the parameters when they matched the signing key:
16950 the verifying software was then supposed to automatically use the CA's
16951 parameters if they were absent from the end user certificate.
16953 Omitting parameters is no longer recommended. The test was also
16954 the wrong way round! This was probably due to unusual behaviour in
16955 EVP_cmp_parameters() which returns 1 if the parameters match.
16956 This meant that parameters were omitted when they *didn't* match and
16957 the certificate was useless. Certificates signed with 'ca' didn't have
16960 *Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>*
16962 * Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
16963 The interface is as follows:
16964 Applications can use
16965 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
16966 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
16967 "off" is now the default.
16968 The library internally uses
16969 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
16970 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
16971 to disable memory-checking temporarily.
16973 Some inconsistent states that previously were possible (and were
16974 even the default) are now avoided.
16976 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
16977 with each memory chunk allocated; this is occasionally more helpful
16978 than just having a counter.
16980 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
16982 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
16987 * Introduce "mode" for SSL structures (with defaults in SSL_CTX),
16988 which largely parallels "options", but is for changing API behaviour,
16989 whereas "options" are about protocol behaviour.
16990 Initial "mode" flags are:
16992 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
16993 a single record has been written.
16994 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
16995 retries use the same buffer location.
16996 (But all of the contents must be
17001 * Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
17004 * Fix problems with no-hmac etc.
17006 *Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>*
17008 * New functions RSA_get_default_method(), RSA_set_method() and
17009 RSA_get_method(). These allows replacement of RSA_METHODs without having
17010 to mess around with the internals of an RSA structure.
17014 * Fix memory leaks in DSA_do_sign and DSA_is_prime.
17015 Also really enable memory leak checks in openssl.c and in some
17018 *Chad C. Mulligan, Bodo Moeller*
17020 * Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
17021 up the length of negative integers. This has now been simplified to just
17022 store the length when it is first determined and use it later, rather
17023 than trying to keep track of where data is copied and updating it to
17025 *Steve Henson, reported by Brien Wheeler <bwheeler@authentica-security.com>*
17027 * Add a new function PKCS7_signatureVerify. This allows the verification
17028 of a PKCS#7 signature but with the signing certificate passed to the
17029 function itself. This contrasts with PKCS7_dataVerify which assumes the
17030 certificate is present in the PKCS#7 structure. This isn't always the
17031 case: certificates can be omitted from a PKCS#7 structure and be
17032 distributed by "out of band" means (such as a certificate database).
17036 * Complete the `PEM_*` macros with DECLARE_PEM versions to replace the
17037 function prototypes in pem.h, also change util/mkdef.pl to add the
17038 necessary function names.
17042 * mk1mf.pl (used by Windows builds) did not properly read the
17043 options set by Configure in the top level Makefile, and Configure
17044 was not even able to write more than one option correctly.
17045 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
17049 * New functions CONF_load_bio() and CONF_load_fp() to allow a config
17050 file to be loaded from a BIO or FILE pointer. The BIO version will
17051 for example allow memory BIOs to contain config info.
17055 * New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
17056 Whoever hopes to achieve shared-library compatibility across versions
17057 must use this, not the compile-time macro.
17058 (Exercise 0.9.4: Which is the minimum library version required by
17060 Note: All this applies only to multi-threaded programs, others don't
17065 * Add missing case to s3_clnt.c state machine -- one of the new SSL tests
17066 through a BIO pair triggered the default case, i.e.
17067 SSLerr(...,SSL_R_UNKNOWN_STATE).
17071 * New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
17072 can use the SSL library even if none of the specific BIOs is
17077 * Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
17078 for the encoded length.
17080 *Jeon KyoungHo <khjeon@sds.samsung.co.kr>*
17082 * Add initial documentation of the X509V3 functions.
17086 * Add a new pair of functions PEM_write_PKCS8PrivateKey() and
17087 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
17088 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
17089 secure PKCS#8 private key format with a high iteration count.
17093 * Fix determination of Perl interpreter: A perl or perl5
17094 *directory* in $PATH was also accepted as the interpreter.
17096 *Ralf S. Engelschall*
17098 * Fix demos/sign/sign.c: well there wasn't anything strictly speaking
17099 wrong with it but it was very old and did things like calling
17100 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
17101 unusual formatting.
17105 * Fix demos/selfsign.c: it used obsolete and deleted functions, changed
17106 to use the new extension code.
17110 * Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
17111 with macros. This should make it easier to change their form, add extra
17112 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
17117 * Add to configuration table a new entry that can specify an alternative
17118 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
17119 according to Mark Crispin <MRC@Panda.COM>.
17123 * DES CBC did not update the IV. Weird.
17127 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
17128 Changing the behaviour of the former might break existing programs --
17129 where IV updating is needed, des_ncbc_encrypt can be used.
17132 * When bntest is run from "make test" it drives bc to check its
17133 calculations, as well as internally checking them. If an internal check
17134 fails, it needs to cause bc to give a non-zero result or make test carries
17135 on without noticing the failure. Fixed.
17139 * DES library cleanups.
17143 * Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
17144 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
17145 ciphers. NOTE: although the key derivation function has been verified
17146 against some published test vectors it has not been extensively tested
17147 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
17152 * Instead of "mkdir -p", which is not fully portable, use new
17153 Perl script "util/mkdir-p.pl".
17157 * Rewrite the way password based encryption (PBE) is handled. It used to
17158 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
17159 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
17160 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
17161 the 'parameter' field of the AlgorithmIdentifier is passed to the
17162 underlying key generation function so it must do its own ASN1 parsing.
17163 This has also changed the EVP_PBE_CipherInit() function which now has a
17164 'parameter' argument instead of literal salt and iteration count values
17165 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
17169 * Support for PKCS#5 v1.5 compatible password based encryption algorithms
17170 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
17171 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
17172 KEY" because this clashed with PKCS#8 unencrypted string. Since this
17173 value was just used as a "magic string" and not used directly its
17174 value doesn't matter.
17178 * Introduce some semblance of const correctness to BN. Shame C doesn't
17183 * "linux-sparc64" configuration (ultrapenguin).
17185 *Ray Miller <ray.miller@oucs.ox.ac.uk>*
17186 "linux-sparc" configuration.
17188 *Christian Forster <fo@hawo.stw.uni-erlangen.de>*
17190 * config now generates no-xxx options for missing ciphers.
17194 * Support the EBCDIC character set (work in progress).
17195 File ebcdic.c not yet included because it has a different license.
17197 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
17199 * Support BS2000/OSD-POSIX.
17201 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
17203 * Make callbacks for key generation use `void *` instead of `char *`.
17207 * Make S/MIME samples compile (not yet tested).
17211 * Additional typesafe stacks.
17215 * New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
17219 ### Changes between 0.9.3 and 0.9.3a [29 May 1999]
17221 * New configuration variant "sco5-gcc".
17223 * Updated some demos.
17225 *Sean O Riordain, Wade Scholine*
17227 * Add missing BIO_free at exit of pkcs12 application.
17231 * Fix memory leak in conf.c.
17235 * Updates for Win32 to assembler version of MD5.
17239 * Set #! path to perl in `apps/der_chop` to where we found it
17240 instead of using a fixed path.
17244 * SHA library changes for irix64-mips4-cc.
17248 * Improvements for VMS support.
17252 ### Changes between 0.9.2b and 0.9.3 [24 May 1999]
17254 * Bignum library bug fix. IRIX 6 passes "make test" now!
17255 This also avoids the problems with SC4.2 and unpatched SC5.
17257 *Andy Polyakov <appro@fy.chalmers.se>*
17259 * New functions sk_num, sk_value and sk_set to replace the previous macros.
17260 These are required because of the typesafe stack would otherwise break
17261 existing code. If old code used a structure member which used to be STACK
17262 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
17263 sk_num or sk_value it would produce an error because the num, data members
17264 are not present in STACK_OF. Now it just produces a warning. sk_set
17265 replaces the old method of assigning a value to sk_value
17266 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
17267 that does this will no longer work (and should use sk_set instead) but
17268 this could be regarded as a "questionable" behaviour anyway.
17272 * Fix most of the other PKCS#7 bugs. The "experimental" code can now
17273 correctly handle encrypted S/MIME data.
17277 * Change type of various DES function arguments from des_cblock
17278 (which means, in function argument declarations, pointer to char)
17279 to des_cblock * (meaning pointer to array with 8 char elements),
17280 which allows the compiler to do more typechecking; it was like
17281 that back in SSLeay, but with lots of ugly casts.
17283 Introduce new type const_des_cblock.
17287 * Reorganise the PKCS#7 library and get rid of some of the more obvious
17288 problems: find RecipientInfo structure that matches recipient certificate
17289 and initialise the ASN1 structures properly based on passed cipher.
17293 * Belatedly make the BN tests actually check the results.
17297 * Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
17298 to and from BNs: it was completely broken. New compilation option
17299 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
17300 key elements as negative integers.
17304 * Reorganize and speed up MD5.
17306 *Andy Polyakov <appro@fy.chalmers.se>*
17310 *Richard Levitte <richard@levitte.org>*
17312 * New option -out to asn1parse to allow the parsed structure to be
17313 output to a file. This is most useful when combined with the -strparse
17314 option to examine the output of things like OCTET STRINGS.
17318 * Make SSL library a little more fool-proof by not requiring any longer
17319 that `SSL_set_{accept,connect}_state` be called before
17320 `SSL_{accept,connect}` may be used (`SSL_set_..._state` is omitted
17321 in many applications because usually everything *appeared* to work as
17322 intended anyway -- now it really works as intended).
17326 * Move openssl.cnf out of lib/.
17330 * Fix various things to let OpenSSL even pass "egcc -pipe -O2 -Wall
17331 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
17332 -Wmissing-declarations -Wnested-externs -Winline" with EGCS 1.1.2+
17334 *Ralf S. Engelschall*
17336 * Various fixes to the EVP and PKCS#7 code. It may now be able to
17337 handle PKCS#7 enveloped data properly.
17339 *Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve*
17341 * Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
17342 copying pointers. The cert_st handling is changed by this in
17343 various ways (and thus what used to be known as ctx->default_cert
17344 is now called ctx->cert, since we don't resort to `s->ctx->[default_]cert`
17345 any longer when s->cert does not give us what we need).
17346 ssl_cert_instantiate becomes obsolete by this change.
17347 As soon as we've got the new code right (possibly it already is?),
17348 we have solved a couple of bugs of the earlier code where s->cert
17349 was used as if it could not have been shared with other SSL structures.
17351 Note that using the SSL API in certain dirty ways now will result
17352 in different behaviour than observed with earlier library versions:
17353 Changing settings for an `SSL_CTX *ctx` after having done s = SSL_new(ctx)
17354 does not influence s as it used to.
17356 In order to clean up things more thoroughly, inside SSL_SESSION
17357 we don't use CERT any longer, but a new structure SESS_CERT
17358 that holds per-session data (if available); currently, this is
17359 the peer's certificate chain and, for clients, the server's certificate
17360 and temporary key. CERT holds only those values that can have
17361 meaningful defaults in an SSL_CTX.
17365 * New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
17366 from the internal representation. Various PKCS#7 fixes: remove some
17367 evil casts and set the enc_dig_alg field properly based on the signing
17372 * Allow PKCS#12 password to be set from the command line or the
17373 environment. Let 'ca' get its config file name from the environment
17374 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
17379 * Allow certificate policies extension to use an IA5STRING for the
17380 organization field. This is contrary to the PKIX definition but
17381 VeriSign uses it and IE5 only recognises this form. Document 'x509'
17386 * Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
17387 without disallowing inline assembler and the like for non-pedantic builds.
17391 * Support Borland C++ builder.
17393 *Janez Jere <jj@void.si>, modified by Ulf Möller*
17399 * SHA-1 cleanups and performance enhancements.
17401 *Andy Polyakov <appro@fy.chalmers.se>*
17403 * Sparc v8plus assembler for the bignum library.
17405 *Andy Polyakov <appro@fy.chalmers.se>*
17407 * Accept any -xxx and +xxx compiler options in Configure.
17411 * Update HPUX configuration.
17415 * Add missing `sk_<type>_unshift()` function to safestack.h
17417 *Ralf S. Engelschall*
17419 * New function SSL_CTX_use_certificate_chain_file that sets the
17420 "extra_cert"s in addition to the certificate. (This makes sense
17421 only for "PEM" format files, as chains as a whole are not
17426 * Support verify_depth from the SSL API.
17427 x509_vfy.c had what can be considered an off-by-one-error:
17428 Its depth (which was not part of the external interface)
17429 was actually counting the number of certificates in a chain;
17430 now it really counts the depth.
17434 * Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
17435 instead of X509err, which often resulted in confusing error
17436 messages since the error codes are not globally unique
17437 (e.g. an alleged error in ssl3_accept when a certificate
17438 didn't match the private key).
17440 * New function SSL_CTX_set_session_id_context that allows to set a default
17441 value (so that you don't need SSL_set_session_id_context for each
17442 connection using the SSL_CTX).
17446 * OAEP decoding bug fix.
17450 * Support INSTALL_PREFIX for package builders, as proposed by
17455 * New Configure options "threads" and "no-threads". For systems
17456 where the proper compiler options are known (currently Solaris
17457 and Linux), "threads" is the default.
17461 * New script util/mklink.pl as a faster substitute for util/mklink.sh.
17465 * Install various scripts to $(OPENSSLDIR)/misc, not to
17466 $(INSTALLTOP)/bin -- they shouldn't clutter directories
17467 such as /usr/local/bin.
17471 * "make linux-shared" to build shared libraries.
17473 *Niels Poppe <niels@netbox.org>*
17475 * New Configure option `no-<cipher>` (rsa, idea, rc5, ...).
17479 * Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
17480 extension adding in x509 utility.
17484 * Remove NOPROTO sections and error code comments.
17488 * Partial rewrite of the DEF file generator to now parse the ANSI
17493 * New Configure options --prefix=DIR and --openssldir=DIR.
17497 * Complete rewrite of the error code script(s). It is all now handled
17498 by one script at the top level which handles error code gathering,
17499 header rewriting and C source file generation. It should be much better
17500 than the old method: it now uses a modified version of Ulf's parser to
17501 read the ANSI prototypes in all header files (thus the old K&R definitions
17502 aren't needed for error creation any more) and do a better job of
17503 translating function codes into names. The old 'ASN1 error code embedded
17504 in a comment' is no longer necessary and it doesn't use .err files which
17505 have now been deleted. Also the error code call doesn't have to appear all
17506 on one line (which resulted in some large lines...).
17510 * Change #include filenames from `<foo.h>` to `<openssl/foo.h>`.
17514 * Change behaviour of ssl2_read when facing length-0 packets: Don't return
17515 0 (which usually indicates a closed connection), but continue reading.
17519 * Fix some race conditions.
17523 * Add support for CRL distribution points extension. Add Certificate
17524 Policies and CRL distribution points documentation.
17528 * Move the autogenerated header file parts to crypto/opensslconf.h.
17532 * Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
17533 8 of keying material. Merlin has also confirmed interop with this fix
17534 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
17536 *Merlin Hughes <merlin@baltimore.ie>*
17538 * Fix lots of warnings.
17540 *Richard Levitte <levitte@stacken.kth.se>*
17542 * In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
17543 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
17545 *Richard Levitte <levitte@stacken.kth.se>*
17547 * Fix problems with sizeof(long) == 8.
17549 *Andy Polyakov <appro@fy.chalmers.se>*
17551 * Change functions to ANSI C.
17555 * Fix typos in error codes.
17557 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller*
17559 * Remove defunct assembler files from Configure.
17563 * SPARC v8 assembler BIGNUM implementation.
17565 *Andy Polyakov <appro@fy.chalmers.se>*
17567 * Support for Certificate Policies extension: both print and set.
17568 Various additions to support the r2i method this uses.
17572 * A lot of constification, and fix a bug in X509_NAME_oneline() that could
17573 return a const string when you are expecting an allocated buffer.
17577 * Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
17578 types DirectoryString and DisplayText.
17582 * Add code to allow r2i extensions to access the configuration database,
17583 add an LHASH database driver and add several ctx helper functions.
17587 * Fix an evil bug in bn_expand2() which caused various BN functions to
17588 fail when they extended the size of a BIGNUM.
17592 * Various utility functions to handle SXNet extension. Modify mkdef.pl to
17593 support typesafe stack.
17597 * Fix typo in SSL_[gs]et_options().
17599 *Nils Frostberg <nils@medcom.se>*
17601 * Delete various functions and files that belonged to the (now obsolete)
17602 old X509V3 handling code.
17606 * New Configure option "rsaref".
17610 * Don't auto-generate pem.h.
17614 * Introduce type-safe ASN.1 SETs.
17618 * Convert various additional casted stacks to type-safe STACK_OF() variants.
17620 *Ben Laurie, Ralf S. Engelschall, Steve Henson*
17622 * Introduce type-safe STACKs. This will almost certainly break lots of code
17623 that links with OpenSSL (well at least cause lots of warnings), but fear
17624 not: the conversion is trivial, and it eliminates loads of evil casts. A
17625 few STACKed things have been converted already. Feel free to convert more.
17626 In the fullness of time, I'll do away with the STACK type altogether.
17630 * Add `openssl ca -revoke <certfile>` facility which revokes a certificate
17631 specified in `<certfile>` by updating the entry in the index.txt file.
17632 This way one no longer has to edit the index.txt file manually for
17633 revoking a certificate. The -revoke option does the gory details now.
17635 *Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall*
17637 * Fix `openssl crl -noout -text` combination where `-noout` killed the
17638 `-text` option at all and this way the `-noout -text` combination was
17639 inconsistent in `openssl crl` with the friends in `openssl x509|rsa|dsa`.
17641 *Ralf S. Engelschall*
17643 * Make sure a corresponding plain text error message exists for the
17644 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
17645 verify callback function determined that a certificate was revoked.
17647 *Ralf S. Engelschall*
17649 * Bugfix: In test/testenc, don't test `openssl <cipher>` for
17650 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
17651 all available ciphers including rc5, which was forgotten until now.
17652 In order to let the testing shell script know which algorithms
17653 are available, a new (up to now undocumented) command
17654 `openssl list-cipher-commands` is used.
17658 * Bugfix: s_client occasionally would sleep in select() when
17659 it should have checked SSL_pending() first.
17663 * New functions DSA_do_sign and DSA_do_verify to provide access to
17664 the raw DSA values prior to ASN.1 encoding.
17668 * Tweaks to Configure
17670 *Niels Poppe <niels@netbox.org>*
17672 * Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
17677 * New variables $(RANLIB) and $(PERL) in the Makefiles.
17681 * New config option to avoid instructions that are illegal on the 80386.
17682 The default code is faster, but requires at least a 486.
17686 * Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
17687 SSL2_SERVER_VERSION (not used at all) macros, which are now the
17688 same as SSL2_VERSION anyway.
17692 * New "-showcerts" option for s_client.
17696 * Still more PKCS#12 integration. Add pkcs12 application to openssl
17697 application. Various cleanups and fixes.
17701 * More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
17702 modify error routines to work internally. Add error codes and PBE init
17703 to library startup routines.
17707 * Further PKCS#12 integration. Added password based encryption, PKCS#8 and
17708 packing functions to asn1 and evp. Changed function names and error
17709 codes along the way.
17713 * PKCS12 integration: and so it begins... First of several patches to
17714 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
17715 objects to objects.h
17719 * Add a new 'indent' option to some X509V3 extension code. Initial ASN1
17720 and display support for Thawte strong extranet extension.
17724 * Add LinuxPPC support.
17726 *Jeff Dubrule <igor@pobox.org>*
17728 * Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
17729 bn_div_words in alpha.s.
17731 *Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie*
17733 * Make sure the RSA OAEP test is skipped under -DRSAref because
17734 OAEP isn't supported when OpenSSL is built with RSAref.
17736 *Ulf Moeller <ulf@fitug.de>*
17738 * Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
17739 so they no longer are missing under -DNOPROTO.
17741 *Soren S. Jorvang <soren@t.dk>*
17743 ### Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
17745 * Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
17746 doesn't work when the session is reused. Coming soon!
17750 * Fix a security hole, that allows sessions to be reused in the wrong
17751 context thus bypassing client cert protection! All software that uses
17752 client certs and session caches in multiple contexts NEEDS PATCHING to
17753 allow session reuse! A fuller solution is in the works.
17755 *Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)*
17757 * Some more source tree cleanups (removed obsolete files
17758 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
17759 permission on "config" script to be executable) and a fix for the INSTALL
17762 *Ulf Moeller <ulf@fitug.de>*
17764 * Remove some legacy and erroneous uses of malloc, free instead of
17767 *Lennart Bang <lob@netstream.se>, with minor changes by Steve*
17769 * Make rsa_oaep_test return non-zero on error.
17771 *Ulf Moeller <ulf@fitug.de>*
17773 * Add support for native Solaris shared libraries. Configure
17774 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
17775 if someone would make that last step automatic.
17777 *Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>*
17779 * ctx_size was not built with the right compiler during "make links". Fixed.
17783 * Change the meaning of 'ALL' in the cipher list. It now means "everything
17784 except NULL ciphers". This means the default cipher list will no longer
17785 enable NULL ciphers. They need to be specifically enabled e.g. with
17786 the string "DEFAULT:eNULL".
17790 * Fix to RSA private encryption routines: if p < q then it would
17791 occasionally produce an invalid result. This will only happen with
17792 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
17796 * Be less restrictive and allow also `perl util/perlpath.pl
17797 /path/to/bin/perl` in addition to `perl util/perlpath.pl /path/to/bin`,
17798 because this way one can also use an interpreter named `perl5` (which is
17799 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
17800 installed as `perl`).
17802 *Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
17804 * Let util/clean-depend.pl work also with older Perl 5.00x versions.
17806 *Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
17808 * Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
17809 advapi32.lib to Win32 build and change the pem test comparison
17810 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
17811 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
17812 and crypto/des/ede_cbcm_enc.c.
17816 * DES quad checksum was broken on big-endian architectures. Fixed.
17820 * Comment out two functions in bio.h that aren't implemented. Fix up the
17821 Win32 test batch file so it (might) work again. The Win32 test batch file
17822 is horrible: I feel ill....
17826 * Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
17827 in e_os.h. Audit of header files to check ANSI and non ANSI
17828 sections: 10 functions were absent from non ANSI section and not exported
17829 from Windows DLLs. Fixed up libeay.num for new functions.
17833 * Make `openssl version` output lines consistent.
17835 *Ralf S. Engelschall*
17837 * Fix Win32 symbol export lists for BIO functions: Added
17838 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
17839 to ms/libeay{16,32}.def.
17841 *Ralf S. Engelschall*
17843 * Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
17844 fine under Unix and passes some trivial tests I've now added. But the
17845 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
17846 added to make sure no one expects that this stuff really works in the
17847 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
17848 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
17851 *Ralf S. Engelschall*
17853 * Fix the generation of two part addresses in perl.
17855 *Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie*
17857 * Add config entry for Linux on MIPS.
17859 *John Tobey <jtobey@channel1.com>*
17861 * Make links whenever Configure is run, unless we are on Windoze.
17865 * Permit extensions to be added to CRLs using crl_section in openssl.cnf.
17866 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
17871 * Add a useful kludge to allow package maintainers to specify compiler and
17872 other platforms details on the command line without having to patch the
17873 Configure script every time: One now can use
17874 `perl Configure <id>:<details>`,
17875 i.e. platform ids are allowed to have details appended
17876 to them (separated by colons). This is treated as there would be a static
17877 pre-configured entry in Configure's %table under key `<id>` with value
17878 `<details>` and `perl Configure <id>` is called. So, when you want to
17879 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
17880 assembler stuff you can use `perl Configure "FreeBSD-elf:pgcc:-O6:::"`
17881 now, which overrides the FreeBSD-elf entry on-the-fly.
17883 *Ralf S. Engelschall*
17885 * Disable new TLS1 ciphersuites by default: they aren't official yet.
17889 * Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
17890 on the `perl Configure ...` command line. This way one can compile
17891 OpenSSL libraries with Position Independent Code (PIC) which is needed
17892 for linking it into DSOs.
17894 *Ralf S. Engelschall*
17896 * Remarkably, export ciphers were totally broken and no-one had noticed!
17901 * Cleaned up the LICENSE document: The official contact for any license
17902 questions now is the OpenSSL core team under openssl-core@openssl.org.
17903 And add a paragraph about the dual-license situation to make sure people
17904 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
17905 to the OpenSSL toolkit.
17907 *Ralf S. Engelschall*
17909 * General source tree makefile cleanups: Made `making xxx in yyy...`
17910 display consistent in the source tree and replaced `/bin/rm` by `rm`.
17911 Additionally cleaned up the `make links` target: Remove unnecessary
17912 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
17913 to speed processing and no longer clutter the display with confusing
17914 stuff. Instead only the actually done links are displayed.
17916 *Ralf S. Engelschall*
17918 * Permit null encryption ciphersuites, used for authentication only. It used
17919 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
17920 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
17925 * Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
17926 signed attributes when verifying signatures (this would break them),
17927 the detached data encoding was wrong and public keys obtained using
17928 X509_get_pubkey() weren't freed.
17932 * Add text documentation for the BUFFER functions. Also added a work around
17933 to a Win95 console bug. This was triggered by the password read stuff: the
17934 last character typed gets carried over to the next fread(). If you were
17935 generating a new cert request using 'req' for example then the last
17936 character of the passphrase would be CR which would then enter the first
17941 * Added the new 'Includes OpenSSL Cryptography Software' button as
17942 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
17943 button and can be used by applications based on OpenSSL to show the
17944 relationship to the OpenSSL project.
17946 *Ralf S. Engelschall*
17948 * Remove confusing variables in function signatures in files
17949 ssl/ssl_lib.c and ssl/ssl.h.
17951 *Lennart Bong <lob@kulthea.stacken.kth.se>*
17953 * Don't install bss_file.c under PREFIX/include/
17955 *Lennart Bong <lob@kulthea.stacken.kth.se>*
17957 * Get the Win32 compile working again. Modify mkdef.pl so it can handle
17958 functions that return function pointers and has support for NT specific
17959 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
17960 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
17961 unsigned to signed types: this was killing the Win32 compile.
17965 * Add new certificate file to stack functions,
17966 SSL_add_dir_cert_subjects_to_stack() and
17967 SSL_add_file_cert_subjects_to_stack(). These largely supplant
17968 SSL_load_client_CA_file(), and can be used to add multiple certs easily
17969 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
17970 This means that Apache-SSL and similar packages don't have to mess around
17971 to add as many CAs as they want to the preferred list.
17975 * Experiment with doxygen documentation. Currently only partially applied to
17977 See <http://www.stack.nl/~dimitri/doxygen/index.html>, and run doxygen with
17978 openssl.doxy as the configuration file.
17982 * Get rid of remaining C++-style comments which strict C compilers hate.
17984 *Ralf S. Engelschall, pointed out by Carlos Amengual*
17986 * Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
17987 compiled in by default: it has problems with large keys.
17991 * Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
17992 DH private keys and/or callback functions which directly correspond to
17993 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
17994 is needed for applications which have to configure certificates on a
17995 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
17997 For the RSA certificate situation is makes no difference, but
17998 for the DSA certificate situation this fixes the "no shared cipher"
17999 problem where the OpenSSL cipher selection procedure failed because the
18000 temporary keys were not overtaken from the context and the API provided
18001 no way to reconfigure them.
18002 The new functions now let applications reconfigure the stuff and they
18003 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
18004 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
18005 non-public-API function ssl_cert_instantiate() is used as a helper
18006 function and also to reduce code redundancy inside ssl_rsa.c.
18008 *Ralf S. Engelschall*
18010 * Move s_server -dcert and -dkey options out of the undocumented feature
18011 area because they are useful for the DSA situation and should be
18012 recognized by the users.
18014 *Ralf S. Engelschall*
18016 * Fix the cipher decision scheme for export ciphers: the export bits are
18017 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
18018 SSL_EXP_MASK. So, the original variable has to be used instead of the
18019 already masked variable.
18021 *Richard Levitte <levitte@stacken.kth.se>*
18023 * Fix `port` variable from `int` to `unsigned int` in crypto/bio/b_sock.c
18025 *Richard Levitte <levitte@stacken.kth.se>*
18027 * Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
18028 from `int` to `unsigned int` because it is a length and initialized by
18029 EVP_DigestFinal() which expects an `unsigned int *`.
18031 *Richard Levitte <levitte@stacken.kth.se>*
18033 * Don't hard-code path to Perl interpreter on shebang line of Configure
18034 script. Instead use the usual Shell->Perl transition trick.
18036 *Ralf S. Engelschall*
18038 * Make `openssl x509 -noout -modulus`' functional also for DSA certificates
18039 (in addition to RSA certificates) to match the behaviour of `openssl dsa
18040 -noout -modulus` as it's already the case for `openssl rsa -noout
18041 -modulus`. For RSA the -modulus is the real "modulus" while for DSA
18042 currently the public key is printed (a decision which was already done by
18043 `openssl dsa -modulus` in the past) which serves a similar purpose.
18044 Additionally the NO_RSA no longer completely removes the whole -modulus
18045 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
18048 *Ralf S. Engelschall*
18050 * Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
18051 BIO. See the source (crypto/evp/bio_ok.c) for more info.
18053 *Arne Ansper <arne@ats.cyber.ee>*
18055 * Dump the old yucky req code that tried (and failed) to allow raw OIDs
18056 to be added. Now both 'req' and 'ca' can use new objects defined in the
18061 * Add cool BIO that does syslog (or event log on NT).
18063 *Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie*
18065 * Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
18066 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
18067 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
18068 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
18072 * Add preliminary config info for new extension code.
18076 * Make RSA_NO_PADDING really use no padding.
18078 *Ulf Moeller <ulf@fitug.de>*
18080 * Generate errors when private/public key check is done.
18084 * Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
18085 for some CRL extensions and new objects added.
18089 * Really fix the ASN1 IMPLICIT bug this time... Partial support for private
18090 key usage extension and fuller support for authority key id.
18094 * Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
18095 padding method for RSA, which is recommended for new applications in PKCS
18096 #1 v2.0 (RFC 2437, October 1998).
18097 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
18098 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
18099 against Bleichbacher's attack on RSA.
18100 *Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
18103 * Updates to the new SSL compression code
18105 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
18107 * Fix so that the version number in the master secret, when passed
18108 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
18109 (because the server will not accept higher), that the version number
18110 is 0x03,0x01, not 0x03,0x00
18112 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
18114 * Run extensive memory leak checks on SSL commands. Fixed *lots* of memory
18115 leaks in `ssl/` relating to new `X509_get_pubkey()` behaviour. Also fixes
18116 in `apps/` and an unrelated leak in `crypto/dsa/dsa_vrf.c`.
18120 * Support for RAW extensions where an arbitrary extension can be
18121 created by including its DER encoding. See `apps/openssl.cnf` for
18126 * Make sure latest Perl versions don't interpret some generated C array
18127 code as Perl array code in the crypto/err/err_genc.pl script.
18129 *Lars Weber <3weber@informatik.uni-hamburg.de>*
18131 * Modify ms/do_ms.bat to not generate assembly language makefiles since
18132 not many people have the assembler. Various Win32 compilation fixes and
18133 update to the INSTALL.W32 file with (hopefully) more accurate Win32
18134 build instructions.
18138 * Modify configure script 'Configure' to automatically create crypto/date.h
18139 file under Win32 and also build pem.h from pem.org. New script
18140 util/mkfiles.pl to create the MINFO file on environments that can't do a
18141 'make files': perl util/mkfiles.pl >MINFO should work.
18145 * Major rework of DES function declarations, in the pursuit of correctness
18146 and purity. As a result, many evil casts evaporated, and some weirdness,
18147 too. You may find this causes warnings in your code. Zapping your evil
18148 casts will probably fix them. Mostly.
18152 * Fix for a typo in asn1.h. Bug fix to object creation script
18153 obj_dat.pl. It considered a zero in an object definition to mean
18154 "end of object": none of the objects in objects.h have any zeros
18155 so it wasn't spotted.
18157 *Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>*
18159 * Add support for Triple DES Cipher Block Chaining with Output Feedback
18160 Masking (CBCM). In the absence of test vectors, the best I have been able
18161 to do is check that the decrypt undoes the encrypt, so far. Send me test
18162 vectors if you have them.
18166 * Correct calculation of key length for export ciphers (too much space was
18167 allocated for null ciphers). This has not been tested!
18171 * Modifications to the mkdef.pl for Win32 DEF file creation. The usage
18172 message is now correct (it understands "crypto" and "ssl" on its
18173 command line). There is also now an "update" option. This will update
18174 the util/ssleay.num and util/libeay.num files with any new functions.
18176 perl util/mkdef.pl crypto ssl update
18177 it will update them.
18181 * Overhauled the Perl interface:
18182 - ported BN stuff to OpenSSL's different BN library
18183 - made the perl/ source tree CVS-aware
18184 - renamed the package from SSLeay to OpenSSL (the files still contain
18185 their history because I've copied them in the repository)
18186 - removed obsolete files (the test scripts will be replaced
18187 by better Test::Harness variants in the future)
18189 *Ralf S. Engelschall*
18191 * First cut for a very conservative source tree cleanup:
18192 1. merge various obsolete readme texts into doc/ssleay.txt
18193 where we collect the old documents and readme texts.
18194 2. remove the first part of files where I'm already sure that we no
18195 longer need them because of three reasons: either they are just temporary
18196 files which were left by Eric or they are preserved original files where
18197 I've verified that the diff is also available in the CVS via "cvs diff
18198 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
18199 the crypto/md/ stuff).
18201 *Ralf S. Engelschall*
18203 * More extension code. Incomplete support for subject and issuer alt
18204 name, issuer and authority key id. Change the i2v function parameters
18205 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
18206 what that's for :-) Fix to ASN1 macro which messed up
18207 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
18211 * Preliminary support for ENUMERATED type. This is largely copied from the
18216 * Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
18218 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
18220 * Make sure `make rehash` target really finds the `openssl` program.
18222 *Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
18224 * Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
18225 like to hear about it if this slows down other processors.
18229 * Add CygWin32 platform information to Configure script.
18231 *Alan Batie <batie@aahz.jf.intel.com>*
18233 * Fixed ms/32all.bat script: `no_asm` -> `no-asm`
18235 *Rainer W. Gerling <gerling@mpg-gv.mpg.de>*
18237 * New program nseq to manipulate netscape certificate sequences
18241 * Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
18246 * Fixes to BN code. Previously the default was to define BN_RECURSION
18247 but the BN code had some problems that would cause failures when
18248 doing certificate verification and some other functions.
18250 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
18252 * Add ASN1 and PEM code to support netscape certificate sequences.
18256 * Add ASN1 and PEM code to support netscape certificate sequences.
18260 * Add several PKIX and private extended key usage OIDs.
18264 * Modify the 'ca' program to handle the new extension code. Modify
18265 openssl.cnf for new extension format, add comments.
18269 * More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
18270 and add a sample to openssl.cnf so req -x509 now adds appropriate
18275 * Continued X509 V3 changes. Add to other makefiles, integrate with the
18276 error code, add initial support to X509_print() and x509 application.
18280 * Takes a deep breath and start adding X509 V3 extension support code. Add
18281 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
18282 stuff is currently isolated and isn't even compiled yet.
18286 * Continuing patches for GeneralizedTime. Fix up certificate and CRL
18287 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
18288 Removed the versions check from X509 routines when loading extensions:
18289 this allows certain broken certificates that don't set the version
18290 properly to be processed.
18294 * Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
18295 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
18296 can still be regenerated with "make depend".
18300 * Spelling mistake in C version of CAST-128.
18302 *Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>*
18304 * Changes to the error generation code. The perl script err-code.pl
18305 now reads in the old error codes and retains the old numbers, only
18306 adding new ones if necessary. It also only changes the .err files if new
18307 codes are added. The makefiles have been modified to only insert errors
18308 when needed (to avoid needlessly modifying header files). This is done
18309 by only inserting errors if the .err file is newer than the auto generated
18310 C file. To rebuild all the error codes from scratch (the old behaviour)
18311 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
18312 or delete all the .err files.
18316 * CAST-128 was incorrectly implemented for short keys. The C version has
18317 been fixed, but is untested. The assembler versions are also fixed, but
18318 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
18319 to regenerate it if needed.
18320 *Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
18321 Hagino <itojun@kame.net>*
18323 * File was opened incorrectly in randfile.c.
18325 *Ulf Möller <ulf@fitug.de>*
18327 * Beginning of support for GeneralizedTime. d2i, i2d, check and print
18328 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
18329 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
18330 al: it's just almost always a UTCTime. Note this patch adds new error
18331 codes so do a "make errors" if there are problems.
18335 * Correct Linux 1 recognition in config.
18337 *Ulf Möller <ulf@fitug.de>*
18339 * Remove pointless MD5 hash when using DSA keys in ca.
18341 *Anonymous <nobody@replay.com>*
18343 * Generate an error if given an empty string as a cert directory. Also
18344 generate an error if handed NULL (previously returned 0 to indicate an
18345 error, but didn't set one).
18347 *Ben Laurie, reported by Anonymous <nobody@replay.com>*
18349 * Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
18353 * Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
18354 parameters. This was causing a warning which killed off the Win32 compile.
18358 * Remove C++ style comments from crypto/bn/bn_local.h.
18360 *Neil Costigan <neil.costigan@celocom.com>*
18362 * The function OBJ_txt2nid was broken. It was supposed to return a nid
18363 based on a text string, looking up short and long names and finally
18364 "dot" format. The "dot" format stuff didn't work. Added new function
18365 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
18366 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
18367 OID is not part of the table.
18371 * Add prototypes to X509 lookup/verify methods, fixing a bug in
18372 X509_LOOKUP_by_alias().
18376 * Sort openssl functions by name.
18380 * Get the `gendsa` command working and add it to the `list` command. Remove
18381 encryption from sample DSA keys (in case anyone is interested the password
18386 * Make *all* `*_free` functions accept a NULL pointer.
18388 *Frans Heymans <fheymans@isaserver.be>*
18390 * If a DH key is generated in s3_srvr.c, don't blow it by trying to use
18393 *Anonymous <nobody@replay.com>*
18395 * s_server should send the CAfile as acceptable CAs, not its own cert.
18397 *Bodo Moeller <3moeller@informatik.uni-hamburg.de>*
18399 * Don't blow it for numeric `-newkey` arguments to `apps/req`.
18401 *Bodo Moeller <3moeller@informatik.uni-hamburg.de>*
18403 * Temp key "for export" tests were wrong in s3_srvr.c.
18405 *Anonymous <nobody@replay.com>*
18407 * Add prototype for temp key callback functions
18408 SSL_CTX_set_tmp_{rsa,dh}_callback().
18412 * Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
18413 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
18417 * X509_name_add_entry() freed the wrong thing after an error.
18419 *Arne Ansper <arne@ats.cyber.ee>*
18421 * rsa_eay.c would attempt to free a NULL context.
18423 *Arne Ansper <arne@ats.cyber.ee>*
18425 * BIO_s_socket() had a broken should_retry() on Windoze.
18427 *Arne Ansper <arne@ats.cyber.ee>*
18429 * BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
18431 *Arne Ansper <arne@ats.cyber.ee>*
18433 * Make sure the already existing X509_STORE->depth variable is initialized
18434 in X509_STORE_new(), but document the fact that this variable is still
18435 unused in the certificate verification process.
18437 *Ralf S. Engelschall*
18439 * Fix the various library and `apps/` files to free up pkeys obtained from
18440 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
18444 * Fix reference counting in X509_PUBKEY_get(). This makes
18445 demos/maurice/example2.c work, amongst others, probably.
18447 *Steve Henson and Ben Laurie*
18449 * First cut of a cleanup for `apps/`. First the `ssleay` program is now named
18450 `openssl` and second, the shortcut symlinks for the `openssl <command>`
18451 are no longer created. This way we have a single and consistent command
18452 line interface `openssl <command>`, similar to `cvs <command>`.
18454 *Ralf S. Engelschall, Paul Sutton and Ben Laurie*
18456 * ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
18457 BIT STRING wrapper always have zero unused bits.
18461 * Add CA.pl, perl version of CA.sh, add extended key usage OID.
18465 * Make the top-level INSTALL documentation easier to understand.
18469 * Makefiles updated to exit if an error occurs in a sub-directory
18470 make (including if user presses ^C) [Paul Sutton]
18472 * Make Montgomery context stuff explicit in RSA data structure.
18476 * Fix build order of pem and err to allow for generated pem.h.
18480 * Fix renumbering bug in X509_NAME_delete_entry().
18484 * Enhanced the err-ins.pl script so it makes the error library number
18485 global and can add a library name. This is needed for external ASN1 and
18486 other error libraries.
18490 * Fixed sk_insert which never worked properly.
18494 * Fix ASN1 macros so they can handle indefinite length constructed
18495 EXPLICIT tags. Some non standard certificates use these: they can now
18500 * Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
18501 into a single doc/ssleay.txt bundle. This way the information is still
18502 preserved but no longer messes up this directory. Now it's new room for
18503 the new set of documentation files.
18505 *Ralf S. Engelschall*
18507 * SETs were incorrectly DER encoded. This was a major pain, because they
18508 shared code with SEQUENCEs, which aren't coded the same. This means that
18509 almost everything to do with SETs or SEQUENCEs has either changed name or
18510 number of arguments.
18512 *Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>*
18514 * Fix test data to work with the above.
18518 * Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
18519 was already fixed by Eric for 0.9.1 it seems.
18521 *Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>*
18523 * Autodetect FreeBSD3.
18527 * Fix various bugs in Configure. This affects the following platforms:
18531 unixware-2.0-pentium
18536 * Eliminate generated files from CVS. Reorder tests to regenerate files
18537 before they are needed.
18541 * Generate Makefile.ssl from Makefile.org (to keep CVS happy).
18545 ### Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
18547 * Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
18548 changed SSLeay to OpenSSL in version strings.
18550 *Ralf S. Engelschall*
18552 * Some fixups to the top-level documents.
18556 * Fixed the nasty bug where rsaref.h was not found under compile-time
18557 because the symlink to include/ was missing.
18559 *Ralf S. Engelschall*
18561 * Incorporated the popular no-RSA/DSA-only patches
18562 which allow to compile a RSA-free SSLeay.
18564 *Andrew Cooke / Interrader Ldt., Ralf S. Engelschall*
18566 * Fixed nasty rehash problem under `make -f Makefile.ssl links`
18567 when "ssleay" is still not found.
18569 *Ralf S. Engelschall*
18571 * Added more platforms to Configure: Cray T3E, HPUX 11,
18573 *Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>*
18575 * Updated the README file.
18577 *Ralf S. Engelschall*
18579 * Added various .cvsignore files in the CVS repository subdirs
18580 to make a "cvs update" really silent.
18582 *Ralf S. Engelschall*
18584 * Recompiled the error-definition header files and added
18585 missing symbols to the Win32 linker tables.
18587 *Ralf S. Engelschall*
18589 * Cleaned up the top-level documents;
18590 o new files: CHANGES and LICENSE
18591 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
18592 o merged COPYRIGHT into LICENSE
18593 o removed obsolete TODO file
18594 o renamed MICROSOFT to INSTALL.W32
18596 *Ralf S. Engelschall*
18598 * Removed dummy files from the 0.9.1b source tree:
18599 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
18600 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
18601 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
18602 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
18603 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
18605 *Ralf S. Engelschall*
18607 * Added various platform portability fixes.
18611 * The Genesis of the OpenSSL rpject:
18612 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
18613 Young and Tim J. Hudson created while they were working for C2Net until
18616 *The OpenSSL Project*
18618 ### Changes between 0.9.0b and 0.9.1b [not released]
18620 * Updated a few CA certificates under certs/
18624 * Changed some BIGNUM api stuff.
18628 * Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
18629 DGUX x86, Linux Alpha, etc.
18633 * New COMP library [crypto/comp/] for SSL Record Layer Compression:
18634 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
18639 * Add -strparse option to asn1pars program which parses nested
18642 *Dr Stephen Henson <shenson@bigfoot.com>*
18644 * Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
18648 * DSA fix for "ca" program.
18652 * Added "-genkey" option to "dsaparam" program.
18656 * Added RIPE MD160 (rmd160) message digest.
18660 * Added -a (all) option to "ssleay version" command.
18664 * Added PLATFORM define which is the id given to Configure.
18668 * Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
18672 * Extended the ASN.1 parser routines.
18676 * Extended BIO routines to support REUSEADDR, seek, tell, etc.
18680 * Added a BN_CTX to the BN library.
18684 * Fixed the weak key values in DES library
18688 * Changed API in EVP library for cipher aliases.
18692 * Added support for RC2/64bit cipher.
18696 * Converted the lhash library to the crypto/mem.c functions.
18700 * Added more recognized ASN.1 object ids.
18704 * Added more RSA padding checks for SSL/TLS.
18708 * Added BIO proxy/filter functionality.
18712 * Added extra_certs to SSL_CTX which can be used
18713 send extra CA certificates to the client in the CA cert chain sending
18714 process. It can be configured with SSL_CTX_add_extra_chain_cert().
18718 * Now Fortezza is denied in the authentication phase because
18719 this is key exchange mechanism is not supported by SSLeay at all.
18723 * Additional PKCS1 checks.
18727 * Support the string "TLSv1" for all TLS v1 ciphers.
18731 * Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
18732 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
18736 * Fixed a few memory leaks.
18740 * Fixed various code and comment typos.
18744 * A minor bug in ssl/s3_clnt.c where there would always be 4 0
18745 bytes sent in the client random.
18747 *Edward Bishop <ebishop@spyglass.com>*
18751 [CVE-2020-1971]: https://www.openssl.org/news/vulnerabilities.html#CVE-2020-1971
18752 [CVE-2020-1967]: https://www.openssl.org/news/vulnerabilities.html#CVE-2020-1967
18753 [CVE-2019-1563]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1563
18754 [CVE-2019-1559]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1559
18755 [CVE-2019-1552]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1552
18756 [CVE-2019-1551]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1551
18757 [CVE-2019-1549]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1549
18758 [CVE-2019-1547]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1547
18759 [CVE-2019-1543]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1543
18760 [CVE-2018-5407]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-5407
18761 [CVE-2018-0739]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0739
18762 [CVE-2018-0737]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0737
18763 [CVE-2018-0735]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0735
18764 [CVE-2018-0734]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0734
18765 [CVE-2018-0733]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0733
18766 [CVE-2018-0732]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0732
18767 [CVE-2017-3738]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3738
18768 [CVE-2017-3737]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3737
18769 [CVE-2017-3736]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3736
18770 [CVE-2017-3735]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3735
18771 [CVE-2017-3733]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3733
18772 [CVE-2017-3732]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3732
18773 [CVE-2017-3731]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3731
18774 [CVE-2017-3730]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3730
18775 [CVE-2016-7055]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7055
18776 [CVE-2016-7054]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7054
18777 [CVE-2016-7053]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7053
18778 [CVE-2016-7052]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7052
18779 [CVE-2016-6309]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6309
18780 [CVE-2016-6308]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6308
18781 [CVE-2016-6307]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6307
18782 [CVE-2016-6306]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6306
18783 [CVE-2016-6305]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6305
18784 [CVE-2016-6304]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6304
18785 [CVE-2016-6303]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6303
18786 [CVE-2016-6302]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6302
18787 [CVE-2016-2183]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2183
18788 [CVE-2016-2182]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2182
18789 [CVE-2016-2181]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2181
18790 [CVE-2016-2180]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2180
18791 [CVE-2016-2179]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2179
18792 [CVE-2016-2178]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2178
18793 [CVE-2016-2177]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2177
18794 [CVE-2016-2176]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2176
18795 [CVE-2016-2109]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2109
18796 [CVE-2016-2107]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2107
18797 [CVE-2016-2106]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2106
18798 [CVE-2016-2105]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2105
18799 [CVE-2016-0800]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0800
18800 [CVE-2016-0799]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0799
18801 [CVE-2016-0798]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0798
18802 [CVE-2016-0797]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0797
18803 [CVE-2016-0705]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0705
18804 [CVE-2016-0702]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0702
18805 [CVE-2016-0701]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0701
18806 [CVE-2015-3197]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3197
18807 [CVE-2015-3196]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3196
18808 [CVE-2015-3195]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3195
18809 [CVE-2015-3194]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3194
18810 [CVE-2015-3193]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3193
18811 [CVE-2015-1793]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1793
18812 [CVE-2015-1792]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1792
18813 [CVE-2015-1791]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1791
18814 [CVE-2015-1790]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1790
18815 [CVE-2015-1789]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1789
18816 [CVE-2015-1788]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1788
18817 [CVE-2015-1787]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1787
18818 [CVE-2015-0293]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0293
18819 [CVE-2015-0291]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0291
18820 [CVE-2015-0290]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0290
18821 [CVE-2015-0289]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0289
18822 [CVE-2015-0288]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0288
18823 [CVE-2015-0287]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0287
18824 [CVE-2015-0286]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0286
18825 [CVE-2015-0285]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0285
18826 [CVE-2015-0209]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0209
18827 [CVE-2015-0208]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0208
18828 [CVE-2015-0207]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0207
18829 [CVE-2015-0206]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0206
18830 [CVE-2015-0205]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0205
18831 [CVE-2015-0204]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0204
18832 [CVE-2014-8275]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-8275
18833 [CVE-2014-5139]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-5139
18834 [CVE-2014-3572]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3572
18835 [CVE-2014-3571]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3571
18836 [CVE-2014-3570]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3570
18837 [CVE-2014-3569]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3569
18838 [CVE-2014-3568]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3568
18839 [CVE-2014-3567]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3567
18840 [CVE-2014-3566]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3566
18841 [CVE-2014-3513]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3513
18842 [CVE-2014-3512]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3512
18843 [CVE-2014-3511]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3511
18844 [CVE-2014-3510]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3510
18845 [CVE-2014-3509]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3509
18846 [CVE-2014-3508]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3508
18847 [CVE-2014-3507]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3507
18848 [CVE-2014-3506]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3506
18849 [CVE-2014-3505]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3505
18850 [CVE-2014-3470]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3470
18851 [CVE-2014-0224]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0224
18852 [CVE-2014-0221]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0221
18853 [CVE-2014-0195]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0195
18854 [CVE-2014-0160]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0160
18855 [CVE-2014-0076]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0076
18856 [CVE-2013-6450]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-6450
18857 [CVE-2013-4353]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-4353
18858 [CVE-2013-0169]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-0169
18859 [CVE-2013-0166]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-0166
18860 [CVE-2012-2686]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2686
18861 [CVE-2012-2333]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2333
18862 [CVE-2012-2110]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2110
18863 [CVE-2012-0884]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0884
18864 [CVE-2012-0050]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0050
18865 [CVE-2012-0027]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0027
18866 [CVE-2011-4619]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4619
18867 [CVE-2011-4577]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4577
18868 [CVE-2011-4576]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4576
18869 [CVE-2011-4109]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4109
18870 [CVE-2011-4108]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4108
18871 [CVE-2011-3210]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-3210
18872 [CVE-2011-3207]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-3207
18873 [CVE-2011-0014]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-0014
18874 [CVE-2010-4252]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-4252
18875 [CVE-2010-4180]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-4180
18876 [CVE-2010-3864]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-3864
18877 [CVE-2010-1633]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-1633
18878 [CVE-2010-0740]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-0740
18879 [CVE-2010-0433]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-0433
18880 [CVE-2009-4355]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-4355
18881 [CVE-2009-3555]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-3555
18882 [CVE-2009-3245]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-3245
18883 [CVE-2009-1386]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1386
18884 [CVE-2009-1379]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1379
18885 [CVE-2009-1378]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1378
18886 [CVE-2009-1377]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1377
18887 [CVE-2009-0789]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0789
18888 [CVE-2009-0591]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0591
18889 [CVE-2009-0590]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0590
18890 [CVE-2008-5077]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-5077
18891 [CVE-2008-1678]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-1678
18892 [CVE-2008-1672]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-1672
18893 [CVE-2008-0891]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-0891
18894 [CVE-2007-5135]: https://www.openssl.org/news/vulnerabilities.html#CVE-2007-5135
18895 [CVE-2007-4995]: https://www.openssl.org/news/vulnerabilities.html#CVE-2007-4995
18896 [CVE-2006-4343]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-4343
18897 [CVE-2006-4339]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-4339
18898 [CVE-2006-3738]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-3738
18899 [CVE-2006-2940]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-2940
18900 [CVE-2006-2937]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-2937
18901 [CVE-2005-2969]: https://www.openssl.org/news/vulnerabilities.html#CVE-2005-2969
18902 [CVE-2004-0112]: https://www.openssl.org/news/vulnerabilities.html#CVE-2004-0112
18903 [CVE-2004-0079]: https://www.openssl.org/news/vulnerabilities.html#CVE-2004-0079
18904 [CVE-2003-0851]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0851
18905 [CVE-2003-0545]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0545
18906 [CVE-2003-0544]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0544
18907 [CVE-2003-0543]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0543
18908 [CVE-2003-0078]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0078
18909 [CVE-2002-0659]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0659
18910 [CVE-2002-0657]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0657
18911 [CVE-2002-0656]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0656
18912 [CVE-2002-0655]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0655