4 This is a high-level summary of the most important changes.
5 For a full list of changes, see the [git commit log][log] and
6 pick the appropriate release branch.
8 [log]: https://github.com/openssl/openssl/commits/
13 - [OpenSSL 3.0](#openssl-30)
14 - [OpenSSL 1.1.1](#openssl-111)
15 - [OpenSSL 1.1.0](#openssl-110)
16 - [OpenSSL 1.0.2](#openssl-102)
17 - [OpenSSL 1.0.1](#openssl-101)
18 - [OpenSSL 1.0.0](#openssl-100)
19 - [OpenSSL 0.9.x](#openssl-09x)
24 For OpenSSL 3.0 a [Migration guide][] has been added, so the CHANGES entries
25 listed here are only a brief description.
26 The migration guide contains more detailed information related to new features,
27 breaking changes, and mappings for the large list of deprecated functions.
29 [Migration guide]: https://github.com/openssl/openssl/tree/master/doc/man7/migration_guide.pod
31 ### Changes between 1.1.1 and 3.0 [xx XXX xxxx]
33 * The ERR_GET_FUNC() function was removed. With the loss of meaningful
34 function codes, this function can only cause problems for calling
39 * Add a configurable flag to output date formats as ISO 8601. Does not
40 change the default date format.
44 * Version of MSVC earlier than 1300 could get link warnings, which could
45 be suppressed if the undocumented -DI_CAN_LIVE_WITH_LNK4049 was set.
46 Support for this flag has been removed.
50 * Rework and make DEBUG macros consistent. Remove unused -DCONF_DEBUG,
51 -DBN_CTX_DEBUG, and REF_PRINT. Add a new tracing category and use it for
52 printing reference counts. Rename -DDEBUG_UNUSED to -DUNUSED_RESULT_DEBUG
53 Fix BN_DEBUG_RAND so it compiles and, when set, force DEBUG_RAND to be set
54 also. Rename engine_debug_ref to be ENGINE_REF_PRINT also for consistency.
58 * The signatures of the functions to get and set options on SSL and
59 SSL_CTX objects changed from "unsigned long" to "uint64_t" type.
60 Some source code changes may be required.
64 * The public definitions of conf_method_st and conf_st have been
65 deprecated. They will be made opaque in a future release.
67 *Rich Salz and Tomáš Mráz*
69 * Client-initiated renegotiation is disabled by default. To allow it, use
70 the -client_renegotiation option, the SSL_OP_ALLOW_CLIENT_RENEGOTIATION
71 flag, or the "ClientRenegotiation" config parameter as appropriate.
75 * Add "abspath" and "includedir" pragma's to config files, to prevent,
76 or modify relative pathname inclusion.
80 * OpenSSL includes a cryptographic module that is intended to be FIPS 140-2
81 validated. Please consult the README-FIPS and
82 README-PROVIDERS files, as well as the migration guide.
84 *OpenSSL team members and many third party contributors*
86 * For the key types DH and DHX the allowed settable parameters are now different.
90 * The openssl commands that read keys, certificates, and CRLs now
91 automatically detect the PEM or DER format of the input files.
93 *David von Oheimb, Richard Levitte, and Tomáš Mráz*
95 * Added enhanced PKCS#12 APIs which accept a library context.
99 * The default manual page suffix ($MANSUFFIX) has been changed to "ossl"
103 * Added support for Kernel TLS (KTLS).
105 *Boris Pismenny, John Baldwin and Andrew Gallatin*
107 * Support for RFC 5746 secure renegotiation is now required by default for
108 SSL or TLS connections to succeed.
112 * The signature of the `copy` functional parameter of the
113 EVP_PKEY_meth_set_copy() function has changed so its `src` argument is
114 now `const EVP_PKEY_CTX *` instead of `EVP_PKEY_CTX *`. Similarly
115 the signature of the `pub_decode` functional parameter of the
116 EVP_PKEY_asn1_set_public() function has changed so its `pub` argument is
117 now `const X509_PUBKEY *` instead of `X509_PUBKEY *`.
121 * The error return values from some control calls (ctrl) have changed.
125 * A public key check is now performed during EVP_PKEY_derive_set_peer().
129 * Many functions in the EVP_ namespace that are getters of values from
130 implementations or contexts were renamed to include get or get0 in their
131 names. Old names are provided as macro aliases for compatibility and
136 * The EVP_PKEY_CTRL_PKCS7_ENCRYPT, EVP_PKEY_CTRL_PKCS7_DECRYPT,
137 EVP_PKEY_CTRL_PKCS7_SIGN, EVP_PKEY_CTRL_CMS_ENCRYPT,
138 EVP_PKEY_CTRL_CMS_DECRYPT, and EVP_PKEY_CTRL_CMS_SIGN control operations
143 * The EVP_PKEY_public_check() and EVP_PKEY_param_check() functions now work for
146 * The output from the command line applications may have minor
151 * The output from numerous "printing" may have minor changes.
155 * Windows thread synchronization uses read/write primitives (SRWLock) when
156 supported by the OS, otherwise CriticalSection continues to be used.
160 * Add filter BIO BIO_f_readbuffer() that allows BIO_tell() and BIO_seek() to
161 work on read only BIO source/sinks that do not support these functions.
162 This allows piping or redirection of a file BIO using stdin to be buffered
163 into memory. This is used internally in OSSL_DECODER_from_bio().
167 * OSSL_STORE_INFO_get_type() may now return an additional value. In 1.1.1
168 this function would return one of the values OSSL_STORE_INFO_NAME,
169 OSSL_STORE_INFO_PKEY, OSSL_STORE_INFO_PARAMS, OSSL_STORE_INFO_CERT or
170 OSSL_STORE_INFO_CRL. Decoded public keys would previously have been reported
171 as type OSSL_STORE_INFO_PKEY in 1.1.1. In 3.0 decoded public keys are now
172 reported as having the new type OSSL_STORE_INFO_PUBKEY. Applications
173 using this function should be amended to handle the changed return value.
177 * Improved adherence to Enhanced Security Services (ESS, RFC 2634 and RFC 5035)
178 for the TSP and CMS Advanced Electronic Signatures (CAdES) implementations.
179 As required by RFC 5035 check both ESSCertID and ESSCertIDv2 if both present.
180 Correct the semantics of checking the validation chain in case ESSCertID{,v2}
181 contains more than one certificate identifier: This means that all
182 certificates referenced there MUST be part of the validation chain.
186 * The implementation of older EVP ciphers related to CAST, IDEA, SEED, RC2, RC4,
187 RC5, DESX and DES have been moved to the legacy provider.
191 * The implementation of the EVP digests MD2, MD4, MDC2, WHIRLPOOL and
192 RIPEMD-160 have been moved to the legacy provider.
196 * The deprecated function EVP_PKEY_get0() now returns NULL being called for a
201 * The deprecated functions EVP_PKEY_get0_RSA(),
202 EVP_PKEY_get0_DSA(), EVP_PKEY_get0_EC_KEY(), EVP_PKEY_get0_DH(),
203 EVP_PKEY_get0_hmac(), EVP_PKEY_get0_poly1305() and EVP_PKEY_get0_siphash() as
204 well as the similarly named "get1" functions behave differently in
209 * A number of functions handling low-level keys or engines were deprecated
210 including EVP_PKEY_set1_engine(), EVP_PKEY_get0_engine(), EVP_PKEY_assign(),
211 EVP_PKEY_get0(), EVP_PKEY_get0_hmac(), EVP_PKEY_get0_poly1305() and
212 EVP_PKEY_get0_siphash().
216 * PKCS#5 PBKDF1 key derivation has been moved from PKCS5_PBE_keyivgen() into
217 the legacy crypto provider as an EVP_KDF. Applications requiring this KDF
218 will need to load the legacy crypto provider. This includes these PBE
219 algorithms which use this KDF:
220 - NID_pbeWithMD2AndDES_CBC
221 - NID_pbeWithMD5AndDES_CBC
222 - NID_pbeWithSHA1AndRC2_CBC
223 - NID_pbeWithMD2AndRC2_CBC
224 - NID_pbeWithMD5AndRC2_CBC
225 - NID_pbeWithSHA1AndDES_CBC
229 * Deprecated obsolete BIO_set_callback(), BIO_get_callback(), and
230 BIO_debug_callback() functions.
234 * Deprecated obsolete EVP_PKEY_CTX_get0_dh_kdf_ukm() and
235 EVP_PKEY_CTX_get0_ecdh_kdf_ukm() functions.
239 * The RAND_METHOD APIs have been deprecated.
243 * The SRP APIs have been deprecated.
247 * Add a compile time option to prevent the caching of provider fetched
248 algorithms. This is enabled by including the no-cached-fetch option
249 at configuration time.
253 * pkcs12 now uses defaults of PBKDF2, AES and SHA-256, with a MAC iteration
254 count of PKCS12_DEFAULT_ITER.
256 *Tomáš Mráz and Sahana Prasad*
258 * The openssl speed command does not use low-level API calls anymore.
262 * Parallel dual-prime 1024-bit modular exponentiation for AVX512_IFMA
265 *Ilya Albrekht, Sergey Kirillov, Andrey Matyukov (Intel Corp)*
267 * Combining the Configure options no-ec and no-dh no longer disables TLSv1.3.
271 * Implemented support for fully "pluggable" TLSv1.3 groups. This means that
272 providers may supply their own group implementations (using either the "key
273 exchange" or the "key encapsulation" methods) which will automatically be
274 detected and used by libssl.
276 *Matt Caswell, Nicola Tuveri*
278 * The undocumented function X509_certificate_type() has been deprecated;
282 * Deprecated the obsolete BN_pseudo_rand() and BN_pseudo_rand_range().
286 * Removed RSA padding mode for SSLv23 (which was only used for
287 SSLv2). This includes the functions RSA_padding_check_SSLv23() and
288 RSA_padding_add_SSLv23() and the `-ssl` option in the deprecated
293 * Deprecated the obsolete X9.31 RSA key generation related functions.
295 * While a callback function set via `SSL_CTX_set_cert_verify_callback()`
296 is not allowed to return a value > 1, this is no more taken as failure.
298 *Viktor Dukhovni and David von Oheimb*
300 * Deprecated the obsolete X9.31 RSA key generation related functions
301 BN_X931_generate_Xpq(), BN_X931_derive_prime_ex(), and
302 BN_X931_generate_prime_ex().
306 * The default key generation method for the regular 2-prime RSA keys was
307 changed to the FIPS 186-4 B.3.6 method.
311 * Deprecated the BN_is_prime_ex() and BN_is_prime_fasttest_ex() functions.
315 * Deprecated EVP_MD_CTX_set_update_fn() and EVP_MD_CTX_update_fn().
319 * Deprecated the type OCSP_REQ_CTX and the functions OCSP_REQ_CTX_*() and
320 replaced with OSSL_HTTP_REQ_CTX and the functions OSSL_HTTP_REQ_CTX_*().
322 *Rich Salz, Richard Levitte, and David von Oheimb*
324 * Deprecated `X509_http_nbio()` and `X509_CRL_http_nbio()`.
328 * Deprecated `OCSP_parse_url()`.
332 * Validation of SM2 keys has been separated from the validation of regular EC
337 * Behavior of the `pkey` app is changed, when using the `-check` or `-pubcheck`
338 switches: a validation failure triggers an early exit, returning a failure
339 exit status to the parent process.
343 * Changed behavior of SSL_CTX_set_ciphersuites() and SSL_set_ciphersuites()
344 to ignore unknown ciphers.
348 * The `-cipher-commands` and `-digest-commands` options
349 of the command line utility `list` have been deprecated.
350 Instead use the `-cipher-algorithms` and `-digest-algorithms` options.
354 * Added convenience functions for generating asymmetric key pairs:
355 The 'quick' one-shot (yet somewhat limited) function L<EVP_PKEY_Q_keygen(3)>
356 and macros for the most common cases: <EVP_RSA_gen(3)> and L<EVP_EC_gen(3)>.
360 * All of the low level EC_KEY functions have been deprecated.
362 *Shane Lontis, Paul Dale, Richard Levitte, and Tomáš Mráz*
364 * Deprecated all the libcrypto and libssl error string loading
369 * The functions SSL_CTX_set_tmp_dh_callback and SSL_set_tmp_dh_callback, as
370 well as the macros SSL_CTX_set_tmp_dh() and SSL_set_tmp_dh() have been
375 * The `-crypt` option to the `passwd` command line tool has been removed.
379 * The -C option to the `x509`, `dhparam`, `dsaparam`, and `ecparam` commands
384 * Add support for AES Key Wrap inverse ciphers to the EVP layer.
388 * Deprecated EVP_PKEY_set1_tls_encodedpoint() and
389 EVP_PKEY_get1_tls_encodedpoint().
393 * The security callback, which can be customised by application code, supports
394 the security operation SSL_SECOP_TMP_DH. One location of the "other" parameter
395 was incorrectly passing a DH object. It now passed an EVP_PKEY in all cases.
399 * Add PKCS7_get_octet_string() and PKCS7_type_is_other() to the public
400 interface. Their functionality remains unchanged.
404 * Added new option for 'openssl list', '-providers', which will display the
405 list of loaded providers, their names, version and status. It optionally
406 displays their gettable parameters.
410 * Removed EVP_PKEY_set_alias_type().
414 * Deprecated `EVP_PKEY_CTX_set_rsa_keygen_pubexp()` and introduced
415 `EVP_PKEY_CTX_set1_rsa_keygen_pubexp()`, which is now preferred.
419 * Changed all "STACK" functions to be macros instead of inline functions. Macro
420 parameters are still checked for type safety at compile time via helper
425 * Remove the RAND_DRBG API
427 *Paul Dale and Matthias St. Pierre*
429 * Allow `SSL_set1_host()` and `SSL_add1_host()` to take IP literal addresses
430 as well as actual hostnames.
434 * The 'MinProtocol' and 'MaxProtocol' configuration commands now silently
435 ignore TLS protocol version bounds when configuring DTLS-based contexts, and
436 conversely, silently ignore DTLS protocol version bounds when configuring
437 TLS-based contexts. The commands can be repeated to set bounds of both
438 types. The same applies with the corresponding "min_protocol" and
439 "max_protocol" command-line switches, in case some application uses both TLS
442 SSL_CTX instances that are created for a fixed protocol version (e.g.
443 `TLSv1_server_method()`) also silently ignore version bounds. Previously
444 attempts to apply bounds to these protocol versions would result in an
445 error. Now only the "version-flexible" SSL_CTX instances are subject to
446 limits in configuration files in command-line options.
450 * Deprecated the `ENGINE` API. Engines should be replaced with providers
455 * Reworked the recorded ERR codes to make better space for system errors.
456 To distinguish them, the macro `ERR_SYSTEM_ERROR()` indicates if the
457 given code is a system error (true) or an OpenSSL error (false).
461 * Reworked the test perl framework to better allow parallel testing.
463 *Nicola Tuveri and David von Oheimb*
465 * Added ciphertext stealing algorithms AES-128-CBC-CTS, AES-192-CBC-CTS and
466 AES-256-CBC-CTS to the providers. CS1, CS2 and CS3 variants are supported.
470 * 'Configure' has been changed to figure out the configuration target if
471 none is given on the command line. Consequently, the 'config' script is
472 now only a mere wrapper. All documentation is changed to only mention
475 *Rich Salz and Richard Levitte*
477 * Added a library context `OSSL_LIB_CTX` that applications as well as
478 other libraries can use to form a separate context within which
479 libcrypto operations are performed.
483 * Handshake now fails if Extended Master Secret extension is dropped
488 * Dropped interactive mode from the `openssl` program.
492 * Deprecated `EVP_PKEY_cmp()` and `EVP_PKEY_cmp_parameters()`.
494 *David von Oheimb and Shane Lontis*
496 * Deprecated `EC_METHOD_get_field_type()`.
500 * Deprecated EC_GFp_simple_method(), EC_GFp_mont_method(),
501 EC_GF2m_simple_method(), EC_GFp_nist_method(), EC_GFp_nistp224_method()
502 EC_GFp_nistp256_method(), and EC_GFp_nistp521_method().
506 * Deprecated EC_GROUP_new(), EC_GROUP_method_of(), and EC_POINT_method_of().
510 * Add CAdES-BES signature verification support, mostly derived
511 from ESSCertIDv2 TS (RFC 5816) contribution by Marek Klein.
513 *Filipe Raimundo da Silva*
515 * Add CAdES-BES signature scheme and attributes support (RFC 5126) to CMS API.
519 * Added the AuthEnvelopedData content type structure (RFC 5083) with AES-GCM
520 parameter (RFC 5084) for the Cryptographic Message Syntax (CMS).
524 * Deprecated EC_POINT_make_affine() and EC_POINTs_make_affine().
528 * Deprecated EC_GROUP_precompute_mult(), EC_GROUP_have_precompute_mult(), and
529 EC_KEY_precompute_mult().
533 * Deprecated EC_POINTs_mul().
537 * Removed FIPS_mode() and FIPS_mode_set().
541 * The SSL option SSL_OP_IGNORE_UNEXPECTED_EOF is introduced.
545 * Deprecated EC_POINT_set_Jprojective_coordinates_GFp() and
546 EC_POINT_get_Jprojective_coordinates_GFp().
550 * Added OSSL_PARAM_BLD to the public interface. This allows OSSL_PARAM
551 arrays to be more easily constructed via a series of utility functions.
552 Create a parameter builder using OSSL_PARAM_BLD_new(), add parameters using
553 the various push functions and finally convert to a passable OSSL_PARAM
554 array using OSSL_PARAM_BLD_to_param().
558 * The security strength of SHA1 and MD5 based signatures in TLS has been
563 * Added EVP_PKEY_set_type_by_keymgmt(), to initialise an EVP_PKEY to
564 contain a provider side internal key.
568 * ASN1_verify(), ASN1_digest() and ASN1_sign() have been deprecated.
572 * Project text documents not yet having a proper file name extension
573 (`HACKING`, `LICENSE`, `NOTES*`, `README*`, `VERSION`) have been renamed to
574 `*.md` as far as reasonable, else `*.txt`, for better use with file managers.
578 * The main project documents (README, NEWS, CHANGES, INSTALL, SUPPORT)
579 have been converted to Markdown with the goal to produce documents
580 which not only look pretty when viewed online in the browser, but
581 remain well readable inside a plain text editor.
583 To achieve this goal, a 'minimalistic' Markdown style has been applied
584 which avoids formatting elements that interfere too much with the
585 reading flow in the text file. For example, it
587 * avoids [ATX headings][] and uses [setext headings][] instead
588 (which works for `<h1>` and `<h2>` headings only).
589 * avoids [inline links][] and uses [reference links][] instead.
590 * avoids [fenced code blocks][] and uses [indented code blocks][] instead.
592 [ATX headings]: https://github.github.com/gfm/#atx-headings
593 [setext headings]: https://github.github.com/gfm/#setext-headings
594 [inline links]: https://github.github.com/gfm/#inline-link
595 [reference links]: https://github.github.com/gfm/#reference-link
596 [fenced code blocks]: https://github.github.com/gfm/#fenced-code-blocks
597 [indented code blocks]: https://github.github.com/gfm/#indented-code-blocks
599 *Matthias St. Pierre*
601 * The test suite is changed to preserve results of each test recipe.
602 A new directory test-runs/ with subdirectories named like the
603 test recipes are created in the build tree for this purpose.
607 * Added an implementation of CMP and CRMF (RFC 4210, RFC 4211 RFC 6712).
608 This adds `crypto/cmp/`, `crpyto/crmf/`, `apps/cmp.c`, and `test/cmp_*`.
609 See L<openssl-cmp(1)> and L<OSSL_CMP_exec_IR_ses(3)> as starting points.
611 *David von Oheimb, Martin Peylo*
613 * Generalized the HTTP client code from `crypto/ocsp/` into `crpyto/http/`.
614 It supports arbitrary request and response content types, GET redirection,
615 TLS, connections via HTTP(S) proxies, connections and exchange via
616 user-defined BIOs (allowing implicit connections), persistent connections,
617 and timeout checks. See L<OSSL_HTTP_transfer(3)> etc. for details.
618 The legacy OCSP-focused (and only partly documented) API
619 is retained for backward compatibility, while most of it is deprecated.
623 * Added `util/check-format.pl`, a tool for checking adherence to the
624 OpenSSL coding style <https://www.openssl.org/policies/codingstyle.html>.
625 The checks performed are incomplete and yield some false positives.
626 Still the tool should be useful for detecting most typical glitches.
630 * `BIO_do_connect()` and `BIO_do_handshake()` have been extended:
631 If domain name resolution yields multiple IP addresses all of them are tried
632 after `connect()` failures.
636 * All of the low level RSA functions have been deprecated.
640 * X509 certificates signed using SHA1 are no longer allowed at security
645 * The command line utilities dhparam, dsa, gendsa and dsaparam have been
646 modified to use PKEY APIs. These commands are now in maintenance mode
647 and no new features will be added to them.
651 * The command line utility rsautl has been deprecated.
655 * The command line utilities genrsa and rsa have been modified to use PKEY
656 APIs. They now write PKCS#8 keys by default. These commands are now in
657 maintenance mode and no new features will be added to them.
661 * All of the low level DH functions have been deprecated.
663 *Paul Dale and Matt Caswell*
665 * All of the low level DSA functions have been deprecated.
669 * Reworked the treatment of EC EVP_PKEYs with the SM2 curve to
670 automatically become EVP_PKEY_SM2 rather than EVP_PKEY_EC.
674 * Deprecated low level ECDH and ECDSA functions.
678 * Deprecated EVP_PKEY_decrypt_old() and EVP_PKEY_encrypt_old().
682 * Enhanced the documentation of EVP_PKEY_get_size(), EVP_PKEY_get_bits()
683 and EVP_PKEY_get_security_bits(). Especially EVP_PKEY_get_size() needed
684 a new formulation to include all the things it can be used for,
685 as well as words of caution.
689 * The SSL_CTX_set_tlsext_ticket_key_cb(3) function has been deprecated.
693 * All of the low level HMAC functions have been deprecated.
695 *Paul Dale and David von Oheimb*
697 * Over two thousand fixes were made to the documentation, including:
698 - Common options (such as -rand/-writerand, TLS version control, etc)
699 were refactored and point to newly-enhanced descriptions in openssl.pod.
700 - Added style conformance for all options (with help from Richard Levitte),
701 documented all reported missing options, added a CI build to check
702 that all options are documented and that no unimplemented options
704 - Documented some internals, such as all use of environment variables.
705 - Addressed all internal broken L<> references.
709 * All of the low level CMAC functions have been deprecated.
713 * The low-level MD2, MD4, MD5, MDC2, RIPEMD160 and Whirlpool digest
714 functions have been deprecated.
716 *Paul Dale and David von Oheimb*
718 * Corrected the documentation of the return values from the `EVP_DigestSign*`
719 set of functions. The documentation mentioned negative values for some
720 errors, but this was never the case, so the mention of negative values
723 Code that followed the documentation and thereby check with something
724 like `EVP_DigestSignInit(...) <= 0` will continue to work undisturbed.
728 * All of the low level cipher functions have been deprecated.
730 *Matt Caswell and Paul Dale*
732 * Removed include/openssl/opensslconf.h.in and replaced it with
733 include/openssl/configuration.h.in, which differs in not including
734 <openssl/macros.h>. A short header include/openssl/opensslconf.h
735 was added to include both.
737 This allows internal hacks where one might need to modify the set
738 of configured macros, for example this if deprecated symbols are
739 still supposed to be available internally:
741 #include <openssl/configuration.h>
743 #undef OPENSSL_NO_DEPRECATED
744 #define OPENSSL_SUPPRESS_DEPRECATED
746 #include <openssl/macros.h>
748 This should not be used by applications that use the exported
749 symbols, as that will lead to linking errors.
753 * Fixed an overflow bug in the x64_64 Montgomery squaring procedure
754 used in exponentiation with 512-bit moduli. No EC algorithms are
755 affected. Analysis suggests that attacks against 2-prime RSA1024,
756 3-prime RSA1536, and DSA1024 as a result of this defect would be very
757 difficult to perform and are not believed likely. Attacks against DH512
758 are considered just feasible. However, for an attack the target would
759 have to re-use the DH512 private key, which is not recommended anyway.
760 Also applications directly using the low-level API BN_mod_exp may be
761 affected if they use BN_FLG_CONSTTIME.
766 * Most memory-debug features have been deprecated, and the functionality
767 replaced with no-ops.
771 * Added documentation for the STACK API.
775 * Introduced a new method type and API, OSSL_ENCODER, to
776 represent generic encoders.
780 * Added a .pragma directive to the syntax of configuration files, to
781 allow varying behavior in a supported and predictable manner.
782 Currently added pragma:
786 This allows dollar signs to be a keyword character unless it's
787 followed by a opening brace or parenthesis. This is useful for
788 platforms where dollar signs are commonly used in names, such as
789 volume names and system directory names on VMS.
793 * Added functionality to create an EVP_PKEY from user data.
797 * Change the interpretation of the '--api' configuration option to
798 mean that this is a desired API compatibility level with no
799 further meaning. The previous interpretation, that this would
800 also mean to remove all deprecated symbols up to and including
801 the given version, no requires that 'no-deprecated' is also used
802 in the configuration.
804 When building applications, the desired API compatibility level
805 can be set with the OPENSSL_API_COMPAT macro like before. For
806 API compatibility version below 3.0, the old style numerical
807 value is valid as before, such as -DOPENSSL_API_COMPAT=0x10100000L.
808 For version 3.0 and on, the value is expected to be the decimal
809 value calculated from the major and minor version like this:
811 MAJOR * 10000 + MINOR * 100
815 -DOPENSSL_API_COMPAT=30000 For 3.0
816 -DOPENSSL_API_COMPAT=30200 For 3.2
818 To hide declarations that are deprecated up to and including the
819 given API compatibility level, -DOPENSSL_NO_DEPRECATED must be
820 given when building the application as well.
824 * Added the X509_LOOKUP_METHOD called X509_LOOKUP_store, to allow
825 access to certificate and CRL stores via URIs and OSSL_STORE
828 This adds the following functions:
830 - X509_LOOKUP_store()
831 - X509_STORE_load_file()
832 - X509_STORE_load_path()
833 - X509_STORE_load_store()
834 - SSL_add_store_cert_subjects_to_stack()
835 - SSL_CTX_set_default_verify_store()
836 - SSL_CTX_load_verify_file()
837 - SSL_CTX_load_verify_dir()
838 - SSL_CTX_load_verify_store()
842 * Added a new method to gather entropy on VMS, based on SYS$GET_ENTROPY.
843 The presence of this system service is determined at run-time.
847 * Added functionality to create an EVP_PKEY context based on data
848 for methods from providers. This takes an algorithm name and a
849 property query string and simply stores them, with the intent
850 that any operation that uses this context will use those strings
851 to fetch the needed methods implicitly, thereby making the port
852 of application written for pre-3.0 OpenSSL easier.
856 * The undocumented function NCONF_WIN32() has been deprecated; for
857 conversion details see the HISTORY section of doc/man5/config.pod
861 * Introduced the new functions EVP_DigestSignInit_ex() and
862 EVP_DigestVerifyInit_ex(). The macros EVP_DigestSignUpdate() and
863 EVP_DigestVerifyUpdate() have been converted to functions. See the man
864 pages for further details.
868 * Over two thousand fixes were made to the documentation, including:
869 adding missing command flags, better style conformance, documentation
872 *Rich Salz, Richard Levitte*
874 * s390x assembly pack: add hardware-support for P-256, P-384, P-521,
875 X25519, X448, Ed25519 and Ed448.
879 * Print all values for a PKCS#12 attribute with 'openssl pkcs12', not just
884 * Deprecated the public definition of `ERR_STATE` as well as the function
885 `ERR_get_state()`. This is done in preparation of making `ERR_STATE` an
890 * Added ERR functionality to give callers access to the stored function
891 names that have replaced the older function code based functions.
893 New functions are ERR_peek_error_func(), ERR_peek_last_error_func(),
894 ERR_peek_error_data(), ERR_peek_last_error_data(), ERR_get_error_all(),
895 ERR_peek_error_all() and ERR_peek_last_error_all().
897 Deprecate ERR functions ERR_get_error_line(), ERR_get_error_line_data(),
898 ERR_peek_error_line_data(), ERR_peek_last_error_line_data() and
899 ERR_func_error_string().
903 * Extended testing to be verbose for failing tests only. The make variables
904 VERBOSE_FAILURE or VF can be used to enable this:
906 $ make VF=1 test # Unix
907 $ mms /macro=(VF=1) test ! OpenVMS
908 $ nmake VF=1 test # Windows
912 * Added the `-copy_extensions` option to the `x509` command for use with
913 `-req` and `-x509toreq`. When given with the `copy` or `copyall` argument,
914 all extensions in the request are copied to the certificate or vice versa.
916 *David von Oheimb*, *Kirill Stefanenkov <kirill_stefanenkov@rambler.ru>*
918 * Added the `-copy_extensions` option to the `req` command for use with
919 `-x509`. When given with the `copy` or `copyall` argument,
920 all extensions in the certification request are copied to the certificate.
924 * The `x509`, `req`, and `ca` commands now make sure that X.509v3 certificates
925 they generate are by default RFC 5280 compliant in the following sense:
926 There is a subjectKeyIdentifier extension with a hash value of the public key
927 and for not self-signed certs there is an authorityKeyIdentifier extension
928 with a keyIdentifier field or issuer information identifying the signing key.
929 This is done unless some configuration overrides the new default behavior,
930 such as `subjectKeyIdentifier = none` and `authorityKeyIdentifier = none`.
934 * Added several checks to `X509_verify_cert()` according to requirements in
935 RFC 5280 in case `X509_V_FLAG_X509_STRICT` is set
936 (which may be done by using the CLI option `-x509_strict`):
937 * The basicConstraints of CA certificates must be marked critical.
938 * CA certificates must explicitly include the keyUsage extension.
939 * If a pathlenConstraint is given the key usage keyCertSign must be allowed.
940 * The issuer name of any certificate must not be empty.
941 * The subject name of CA certs, certs with keyUsage crlSign,
942 and certs without subjectAlternativeName must not be empty.
943 * If a subjectAlternativeName extension is given it must not be empty.
944 * The signatureAlgorithm field and the cert signature must be consistent.
945 * Any given authorityKeyIdentifier and any given subjectKeyIdentifier
946 must not be marked critical.
947 * The authorityKeyIdentifier must be given for X.509v3 certs
948 unless they are self-signed.
949 * The subjectKeyIdentifier must be given for all X.509v3 CA certs.
953 * Certificate verification using `X509_verify_cert()` meanwhile rejects EC keys
954 with explicit curve parameters (specifiedCurve) as required by RFC 5480.
958 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
959 used even when parsing explicit parameters, when loading a encoded key
960 or calling `EC_GROUP_new_from_ecpkparameters()`/
961 `EC_GROUP_new_from_ecparameters()`.
962 This prevents bypass of security hardening and performance gains,
963 especially for curves with specialized EC_METHODs.
964 By default, if a key encoded with explicit parameters is loaded and later
965 encoded, the output is still encoded with explicit parameters, even if
966 internally a "named" EC_GROUP is used for computation.
970 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
971 this change, EC_GROUP_set_generator would accept order and/or cofactor as
972 NULL. After this change, only the cofactor parameter can be NULL. It also
973 does some minimal sanity checks on the passed order.
978 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
979 An attack is simple, if the first CMS_recipientInfo is valid but the
980 second CMS_recipientInfo is chosen ciphertext. If the second
981 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
982 encryption key will be replaced by garbage, and the message cannot be
983 decoded, but if the RSA decryption fails, the correct encryption key is
984 used and the recipient will not notice the attack.
985 As a work around for this potential attack the length of the decrypted
986 key must be equal to the cipher default key length, in case the
987 certifiate is not given and all recipientInfo are tried out.
988 The old behaviour can be re-enabled in the CMS code by setting the
989 CMS_DEBUG_DECRYPT flag.
993 * Early start up entropy quality from the DEVRANDOM seed source has been
994 improved for older Linux systems. The RAND subsystem will wait for
995 /dev/random to be producing output before seeding from /dev/urandom.
996 The seeded state is stored for future library initialisations using
997 a system global shared memory segment. The shared memory identifier
998 can be configured by defining OPENSSL_RAND_SEED_DEVRANDOM_SHM_ID to
999 the desired value. The default identifier is 114.
1003 * Revised BN_generate_prime_ex to not avoid factors 2..17863 in p-1
1004 when primes for RSA keys are computed.
1005 Since we previously always generated primes == 2 (mod 3) for RSA keys,
1006 the 2-prime and 3-prime RSA modules were easy to distinguish, since
1007 `N = p*q = 1 (mod 3)`, but `N = p*q*r = 2 (mod 3)`. Therefore fingerprinting
1008 2-prime vs. 3-prime RSA keys was possible by computing N mod 3.
1009 This avoids possible fingerprinting of newly generated RSA modules.
1013 * Correct the extended master secret constant on EBCDIC systems. Without this
1014 fix TLS connections between an EBCDIC system and a non-EBCDIC system that
1015 negotiate EMS will fail. Unfortunately this also means that TLS connections
1016 between EBCDIC systems with this fix, and EBCDIC systems without this
1017 fix will fail if they negotiate EMS.
1021 * Changed the library initialisation so that the config file is now loaded
1022 by default. This was already the case for libssl. It now occurs for both
1023 libcrypto and libssl. Use the OPENSSL_INIT_NO_LOAD_CONFIG option to
1024 `OPENSSL_init_crypto()` to suppress automatic loading of a config file.
1028 * Introduced new error raising macros, `ERR_raise()` and `ERR_raise_data()`,
1029 where the former acts as a replacement for `ERR_put_error()`, and the
1030 latter replaces the combination `ERR_put_error()` + `ERR_add_error_data()`.
1031 `ERR_raise_data()` adds more flexibility by taking a format string and
1032 an arbitrary number of arguments following it, to be processed with
1037 * Introduced a new function, `OSSL_PROVIDER_available()`, which can be used
1038 to check if a named provider is loaded and available. When called, it
1039 will also activate all fallback providers if such are still present.
1043 * Enforce a minimum DH modulus size of 512 bits.
1047 * Changed DH parameters to generate the order q subgroup instead of 2q.
1048 Previously generated DH parameters are still accepted by DH_check
1049 but DH_generate_key works around that by clearing bit 0 of the
1050 private key for those. This avoids leaking bit 0 of the private key.
1054 * Significantly reduce secure memory usage by the randomness pools.
1058 * `{CRYPTO,OPENSSL}_mem_debug_{push,pop}` are now no-ops and have been
1063 * A new type, EVP_KEYEXCH, has been introduced to represent key exchange
1064 algorithms. An implementation of a key exchange algorithm can be obtained
1065 by using the function EVP_KEYEXCH_fetch(). An EVP_KEYEXCH algorithm can be
1066 used in a call to EVP_PKEY_derive_init_ex() which works in a similar way to
1067 the older EVP_PKEY_derive_init() function. See the man pages for the new
1068 functions for further details.
1072 * The EVP_PKEY_CTX_set_dh_pad() macro has now been converted to a function.
1076 * Removed the function names from error messages and deprecated the
1081 * Removed NextStep support and the macro OPENSSL_UNISTD
1085 * Removed DES_check_key. Also removed OPENSSL_IMPLEMENT_GLOBAL,
1086 OPENSSL_GLOBAL_REF, OPENSSL_DECLARE_GLOBAL.
1087 Also removed "export var as function" capability; we do not export
1088 variables, only functions.
1092 * RC5_32_set_key has been changed to return an int type, with 0 indicating
1093 an error and 1 indicating success. In previous versions of OpenSSL this
1094 was a void type. If a key was set longer than the maximum possible this
1099 * Support SM2 signing and verification schemes with X509 certificate.
1103 * Use SHA256 as the default digest for TS query in the `ts` app.
1107 * Change PBKDF2 to conform to SP800-132 instead of the older PKCS5 RFC2898.
1111 * Default cipher lists/suites are now available via a function, the
1112 #defines are deprecated.
1116 * Add target VC-WIN32-UWP, VC-WIN64A-UWP, VC-WIN32-ARM-UWP and
1117 VC-WIN64-ARM-UWP in Windows OneCore target for making building libraries
1118 for Windows Store apps easier. Also, the "no-uplink" option has been added.
1122 * Join the directories crypto/x509 and crypto/x509v3
1126 * Added command 'openssl kdf' that uses the EVP_KDF API.
1130 * Added command 'openssl mac' that uses the EVP_MAC API.
1134 * Added OPENSSL_info() to get diverse built-in OpenSSL data, such
1135 as default directories. Also added the command 'openssl info'
1136 for scripting purposes.
1140 * The functions AES_ige_encrypt() and AES_bi_ige_encrypt() have been
1145 * Add prediction resistance to the DRBG reseeding process.
1149 * Limit the number of blocks in a data unit for AES-XTS to 2^20 as
1150 mandated by IEEE Std 1619-2018.
1154 * Added newline escaping functionality to a filename when using openssl dgst.
1155 This output format is to replicate the output format found in the `*sum`
1156 checksum programs. This aims to preserve backward compatibility.
1158 *Matt Eaton, Richard Levitte, and Paul Dale*
1160 * Removed the heartbeat message in DTLS feature, as it has very
1161 little usage and doesn't seem to fulfill a valuable purpose.
1162 The configuration option is now deprecated.
1166 * Changed the output of 'openssl {digestname} < file' to display the
1167 digest name in its output.
1171 * Added a new generic trace API which provides support for enabling
1172 instrumentation through trace output.
1174 *Richard Levitte & Matthias St. Pierre*
1176 * Added build tests for C++. These are generated files that only do one
1177 thing, to include one public OpenSSL head file each. This tests that
1178 the public header files can be usefully included in a C++ application.
1180 This test isn't enabled by default. It can be enabled with the option
1181 'enable-buildtest-c++'.
1185 * Add Single Step KDF (EVP_KDF_SS) to EVP_KDF.
1189 * Add KMAC to EVP_MAC.
1193 * Added property based algorithm implementation selection framework to
1198 * Added SCA hardening for modular field inversion in EC_GROUP through
1199 a new dedicated field_inv() pointer in EC_METHOD.
1200 This also addresses a leakage affecting conversions from projective
1201 to affine coordinates.
1203 *Billy Bob Brumley, Nicola Tuveri*
1205 * Added EVP_KDF, an EVP layer KDF API, to simplify adding KDF and PRF
1206 implementations. This includes an EVP_PKEY to EVP_KDF bridge for
1207 those algorithms that were already supported through the EVP_PKEY API
1208 (scrypt, TLS1 PRF and HKDF). The low-level KDF functions for PBKDF2
1209 and scrypt are now wrappers that call EVP_KDF.
1213 * Build devcrypto engine as a dynamic engine.
1215 *Eneas U de Queiroz*
1217 * Add keyed BLAKE2 to EVP_MAC.
1221 * Fix a bug in the computation of the endpoint-pair shared secret used
1222 by DTLS over SCTP. This breaks interoperability with older versions
1223 of OpenSSL like OpenSSL 1.1.0 and OpenSSL 1.0.2. There is a runtime
1224 switch SSL_MODE_DTLS_SCTP_LABEL_LENGTH_BUG (off by default) enabling
1225 interoperability with such broken implementations. However, enabling
1226 this switch breaks interoperability with correct implementations.
1228 * Fix a use after free bug in d2i_X509_PUBKEY when overwriting a
1229 re-used X509_PUBKEY object if the second PUBKEY is malformed.
1233 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
1237 * Changed the license to the Apache License v2.0.
1241 * Switch to a new version scheme using three numbers MAJOR.MINOR.PATCH.
1243 - Major releases (indicated by incrementing the MAJOR release number)
1244 may introduce incompatible API/ABI changes.
1245 - Minor releases (indicated by incrementing the MINOR release number)
1246 may introduce new features but retain API/ABI compatibility.
1247 - Patch releases (indicated by incrementing the PATCH number)
1248 are intended for bug fixes and other improvements of existing
1249 features only (like improving performance or adding documentation)
1250 and retain API/ABI compatibility.
1254 * Add support for RFC5297 SIV mode (siv128), including AES-SIV.
1258 * Remove the 'dist' target and add a tarball building script. The
1259 'dist' target has fallen out of use, and it shouldn't be
1260 necessary to configure just to create a source distribution.
1264 * Recreate the OS390-Unix config target. It no longer relies on a
1265 special script like it did for OpenSSL pre-1.1.0.
1269 * Instead of having the source directories listed in Configure, add
1270 a 'build.info' keyword SUBDIRS to indicate what sub-directories to
1275 * Add GMAC to EVP_MAC.
1279 * Ported the HMAC, CMAC and SipHash EVP_PKEY_METHODs to EVP_MAC.
1283 * Added EVP_MAC, an EVP layer MAC API, to simplify adding MAC
1284 implementations. This includes a generic EVP_PKEY to EVP_MAC bridge,
1285 to facilitate the continued use of MACs through raw private keys in
1286 functionality such as `EVP_DigestSign*` and `EVP_DigestVerify*`.
1290 * Deprecate ECDH_KDF_X9_62().
1294 * Added EVP_PKEY_ECDH_KDF_X9_63 and ecdh_KDF_X9_63() as replacements for
1295 the EVP_PKEY_ECDH_KDF_X9_62 KDF type and ECDH_KDF_X9_62(). The old names
1296 are retained for backwards compatibility.
1300 * AES-XTS mode now enforces that its two keys are different to mitigate
1301 the attacked described in "Efficient Instantiations of Tweakable
1302 Blockciphers and Refinements to Modes OCB and PMAC" by Phillip Rogaway.
1303 Details of this attack can be obtained from:
1304 <http://web.cs.ucdavis.edu/%7Erogaway/papers/offsets.pdf>
1308 * Rename the object files, i.e. give them other names than in previous
1309 versions. Their names now include the name of the final product, as
1310 well as its type mnemonic (bin, lib, shlib).
1314 * Added new option for 'openssl list', '-objects', which will display the
1315 list of built in objects, i.e. OIDs with names.
1319 * Added the options `-crl_lastupdate` and `-crl_nextupdate` to `openssl ca`,
1320 allowing the `lastUpdate` and `nextUpdate` fields in the generated CRL to
1325 * Added support for Linux Kernel TLS data-path. The Linux Kernel data-path
1326 improves application performance by removing data copies and providing
1327 applications with zero-copy system calls such as sendfile and splice.
1331 * The SSL option SSL_OP_CLEANSE_PLAINTEXT is introduced.
1335 * `PKCS12_parse` now maintains the order of the parsed certificates
1336 when outputting them via `*ca` (rather than reversing it).
1340 * Deprecated pthread fork support methods.
1344 * Added support for FFDHE key exchange in TLS 1.3.
1351 ### Changes between 1.1.1j and 1.1.1k [xx XXX xxxx]
1353 * Fixed a problem with verifying a certificate chain when using the
1354 X509_V_FLAG_X509_STRICT flag. This flag enables additional security checks of
1355 the certificates present in a certificate chain. It is not set by default.
1357 Starting from OpenSSL version 1.1.1h a check to disallow certificates in
1358 the chain that have explicitly encoded elliptic curve parameters was added
1359 as an additional strict check.
1361 An error in the implementation of this check meant that the result of a
1362 previous check to confirm that certificates in the chain are valid CA
1363 certificates was overwritten. This effectively bypasses the check
1364 that non-CA certificates must not be able to issue other certificates.
1366 If a "purpose" has been configured then there is a subsequent opportunity
1367 for checks that the certificate is a valid CA. All of the named "purpose"
1368 values implemented in libcrypto perform this check. Therefore, where
1369 a purpose is set the certificate chain will still be rejected even when the
1370 strict flag has been used. A purpose is set by default in libssl client and
1371 server certificate verification routines, but it can be overridden or
1372 removed by an application.
1374 In order to be affected, an application must explicitly set the
1375 X509_V_FLAG_X509_STRICT verification flag and either not set a purpose
1376 for the certificate verification or, in the case of TLS client or server
1377 applications, override the default purpose.
1382 * Fixed an issue where an OpenSSL TLS server may crash if sent a maliciously
1383 crafted renegotiation ClientHello message from a client. If a TLSv1.2
1384 renegotiation ClientHello omits the signature_algorithms extension (where it
1385 was present in the initial ClientHello), but includes a
1386 signature_algorithms_cert extension then a NULL pointer dereference will
1387 result, leading to a crash and a denial of service attack.
1389 A server is only vulnerable if it has TLSv1.2 and renegotiation enabled
1390 (which is the default configuration). OpenSSL TLS clients are not impacted by
1394 *Peter Kästle and Samuel Sapalski*
1396 ### Changes between 1.1.1i and 1.1.1j [16 Feb 2021]
1398 * Fixed the X509_issuer_and_serial_hash() function. It attempts to
1399 create a unique hash value based on the issuer and serial number data
1400 contained within an X509 certificate. However it was failing to correctly
1401 handle any errors that may occur while parsing the issuer field (which might
1402 occur if the issuer field is maliciously constructed). This may subsequently
1403 result in a NULL pointer deref and a crash leading to a potential denial of
1409 * Fixed the RSA_padding_check_SSLv23() function and the RSA_SSLV23_PADDING
1410 padding mode to correctly check for rollback attacks. This is considered a
1411 bug in OpenSSL 1.1.1 because it does not support SSLv2. In 1.0.2 this is
1416 Fixed the EVP_CipherUpdate, EVP_EncryptUpdate and EVP_DecryptUpdate
1417 functions. Previously they could overflow the output length argument in some
1418 cases where the input length is close to the maximum permissable length for
1419 an integer on the platform. In such cases the return value from the function
1420 call would be 1 (indicating success), but the output length value would be
1421 negative. This could cause applications to behave incorrectly or crash.
1426 * Fixed SRP_Calc_client_key so that it runs in constant time. The previous
1427 implementation called BN_mod_exp without setting BN_FLG_CONSTTIME. This
1428 could be exploited in a side channel attack to recover the password. Since
1429 the attack is local host only this is outside of the current OpenSSL
1430 threat model and therefore no CVE is assigned.
1432 Thanks to Mohammed Sabt and Daniel De Almeida Braga for reporting this
1437 ### Changes between 1.1.1h and 1.1.1i [8 Dec 2020]
1439 * Fixed NULL pointer deref in the GENERAL_NAME_cmp function
1440 This function could crash if both GENERAL_NAMEs contain an EDIPARTYNAME.
1441 If an attacker can control both items being compared then this could lead
1442 to a possible denial of service attack. OpenSSL itself uses the
1443 GENERAL_NAME_cmp function for two purposes:
1444 1) Comparing CRL distribution point names between an available CRL and a
1445 CRL distribution point embedded in an X509 certificate
1446 2) When verifying that a timestamp response token signer matches the
1447 timestamp authority name (exposed via the API functions
1448 TS_RESP_verify_response and TS_RESP_verify_token)
1453 ### Changes between 1.1.1g and 1.1.1h [22 Sep 2020]
1455 * Certificates with explicit curve parameters are now disallowed in
1456 verification chains if the X509_V_FLAG_X509_STRICT flag is used.
1460 * The 'MinProtocol' and 'MaxProtocol' configuration commands now silently
1461 ignore TLS protocol version bounds when configuring DTLS-based contexts, and
1462 conversely, silently ignore DTLS protocol version bounds when configuring
1463 TLS-based contexts. The commands can be repeated to set bounds of both
1464 types. The same applies with the corresponding "min_protocol" and
1465 "max_protocol" command-line switches, in case some application uses both TLS
1468 SSL_CTX instances that are created for a fixed protocol version (e.g.
1469 TLSv1_server_method()) also silently ignore version bounds. Previously
1470 attempts to apply bounds to these protocol versions would result in an
1471 error. Now only the "version-flexible" SSL_CTX instances are subject to
1472 limits in configuration files in command-line options.
1476 * Handshake now fails if Extended Master Secret extension is dropped
1481 * The Oracle Developer Studio compiler will start reporting deprecated APIs
1483 ### Changes between 1.1.1f and 1.1.1g [21 Apr 2020]
1485 * Fixed segmentation fault in SSL_check_chain()
1486 Server or client applications that call the SSL_check_chain() function
1487 during or after a TLS 1.3 handshake may crash due to a NULL pointer
1488 dereference as a result of incorrect handling of the
1489 "signature_algorithms_cert" TLS extension. The crash occurs if an invalid
1490 or unrecognised signature algorithm is received from the peer. This could
1491 be exploited by a malicious peer in a Denial of Service attack.
1496 * Added AES consttime code for no-asm configurations
1497 an optional constant time support for AES was added
1498 when building openssl for no-asm.
1499 Enable with: ./config no-asm -DOPENSSL_AES_CONST_TIME
1500 Disable with: ./config no-asm -DOPENSSL_NO_AES_CONST_TIME
1501 At this time this feature is by default disabled.
1502 It will be enabled by default in 3.0.
1506 ### Changes between 1.1.1e and 1.1.1f [31 Mar 2020]
1508 * Revert the change of EOF detection while reading in libssl to avoid
1509 regressions in applications depending on the current way of reporting
1510 the EOF. As the existing method is not fully accurate the change to
1511 reporting the EOF via SSL_ERROR_SSL is kept on the current development
1512 branch and will be present in the 3.0 release.
1516 * Revised BN_generate_prime_ex to not avoid factors 3..17863 in p-1
1517 when primes for RSA keys are computed.
1518 Since we previously always generated primes == 2 (mod 3) for RSA keys,
1519 the 2-prime and 3-prime RSA modules were easy to distinguish, since
1520 N = p*q = 1 (mod 3), but N = p*q*r = 2 (mod 3). Therefore fingerprinting
1521 2-prime vs. 3-prime RSA keys was possible by computing N mod 3.
1522 This avoids possible fingerprinting of newly generated RSA modules.
1526 ### Changes between 1.1.1d and 1.1.1e [17 Mar 2020]
1528 * Properly detect EOF while reading in libssl. Previously if we hit an EOF
1529 while reading in libssl then we would report an error back to the
1530 application (SSL_ERROR_SYSCALL) but errno would be 0. We now add
1531 an error to the stack (which means we instead return SSL_ERROR_SSL) and
1532 therefore give a hint as to what went wrong.
1536 * Check that ed25519 and ed448 are allowed by the security level. Previously
1537 signature algorithms not using an MD were not being checked that they were
1538 allowed by the security level.
1542 * Fixed SSL_get_servername() behaviour. The behaviour of SSL_get_servername()
1543 was not quite right. The behaviour was not consistent between resumption
1544 and normal handshakes, and also not quite consistent with historical
1545 behaviour. The behaviour in various scenarios has been clarified and
1546 it has been updated to make it match historical behaviour as closely as
1551 * *[VMS only]* The header files that the VMS compilers include automatically,
1552 `__DECC_INCLUDE_PROLOGUE.H` and `__DECC_INCLUDE_EPILOGUE.H`, use pragmas
1553 that the C++ compiler doesn't understand. This is a shortcoming in the
1554 compiler, but can be worked around with `__cplusplus` guards.
1556 C++ applications that use OpenSSL libraries must be compiled using the
1557 qualifier `/NAMES=(AS_IS,SHORTENED)` to be able to use all the OpenSSL
1558 functions. Otherwise, only functions with symbols of less than 31
1559 characters can be used, as the linker will not be able to successfully
1560 resolve symbols with longer names.
1564 * Added a new method to gather entropy on VMS, based on SYS$GET_ENTROPY.
1565 The presence of this system service is determined at run-time.
1569 * Added newline escaping functionality to a filename when using openssl dgst.
1570 This output format is to replicate the output format found in the `*sum`
1571 checksum programs. This aims to preserve backward compatibility.
1573 *Matt Eaton, Richard Levitte, and Paul Dale*
1575 * Print all values for a PKCS#12 attribute with 'openssl pkcs12', not just
1580 ### Changes between 1.1.1c and 1.1.1d [10 Sep 2019]
1582 * Fixed a fork protection issue. OpenSSL 1.1.1 introduced a rewritten random
1583 number generator (RNG). This was intended to include protection in the
1584 event of a fork() system call in order to ensure that the parent and child
1585 processes did not share the same RNG state. However this protection was not
1586 being used in the default case.
1588 A partial mitigation for this issue is that the output from a high
1589 precision timer is mixed into the RNG state so the likelihood of a parent
1590 and child process sharing state is significantly reduced.
1592 If an application already calls OPENSSL_init_crypto() explicitly using
1593 OPENSSL_INIT_ATFORK then this problem does not occur at all.
1596 *Matthias St. Pierre*
1598 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
1599 used even when parsing explicit parameters, when loading a encoded key
1600 or calling `EC_GROUP_new_from_ecpkparameters()`/
1601 `EC_GROUP_new_from_ecparameters()`.
1602 This prevents bypass of security hardening and performance gains,
1603 especially for curves with specialized EC_METHODs.
1604 By default, if a key encoded with explicit parameters is loaded and later
1605 encoded, the output is still encoded with explicit parameters, even if
1606 internally a "named" EC_GROUP is used for computation.
1610 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
1611 this change, EC_GROUP_set_generator would accept order and/or cofactor as
1612 NULL. After this change, only the cofactor parameter can be NULL. It also
1613 does some minimal sanity checks on the passed order.
1618 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
1619 An attack is simple, if the first CMS_recipientInfo is valid but the
1620 second CMS_recipientInfo is chosen ciphertext. If the second
1621 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
1622 encryption key will be replaced by garbage, and the message cannot be
1623 decoded, but if the RSA decryption fails, the correct encryption key is
1624 used and the recipient will not notice the attack.
1625 As a work around for this potential attack the length of the decrypted
1626 key must be equal to the cipher default key length, in case the
1627 certifiate is not given and all recipientInfo are tried out.
1628 The old behaviour can be re-enabled in the CMS code by setting the
1629 CMS_DEBUG_DECRYPT flag.
1634 * Early start up entropy quality from the DEVRANDOM seed source has been
1635 improved for older Linux systems. The RAND subsystem will wait for
1636 /dev/random to be producing output before seeding from /dev/urandom.
1637 The seeded state is stored for future library initialisations using
1638 a system global shared memory segment. The shared memory identifier
1639 can be configured by defining OPENSSL_RAND_SEED_DEVRANDOM_SHM_ID to
1640 the desired value. The default identifier is 114.
1644 * Correct the extended master secret constant on EBCDIC systems. Without this
1645 fix TLS connections between an EBCDIC system and a non-EBCDIC system that
1646 negotiate EMS will fail. Unfortunately this also means that TLS connections
1647 between EBCDIC systems with this fix, and EBCDIC systems without this
1648 fix will fail if they negotiate EMS.
1652 * Use Windows installation paths in the mingw builds
1654 Mingw isn't a POSIX environment per se, which means that Windows
1655 paths should be used for installation.
1660 * Changed DH_check to accept parameters with order q and 2q subgroups.
1661 With order 2q subgroups the bit 0 of the private key is not secret
1662 but DH_generate_key works around that by clearing bit 0 of the
1663 private key for those. This avoids leaking bit 0 of the private key.
1667 * Significantly reduce secure memory usage by the randomness pools.
1671 * Revert the DEVRANDOM_WAIT feature for Linux systems
1673 The DEVRANDOM_WAIT feature added a select() call to wait for the
1674 /dev/random device to become readable before reading from the
1675 /dev/urandom device.
1677 It turned out that this change had negative side effects on
1678 performance which were not acceptable. After some discussion it
1679 was decided to revert this feature and leave it up to the OS
1680 resp. the platform maintainer to ensure a proper initialization
1681 during early boot time.
1683 *Matthias St. Pierre*
1685 ### Changes between 1.1.1b and 1.1.1c [28 May 2019]
1687 * Add build tests for C++. These are generated files that only do one
1688 thing, to include one public OpenSSL head file each. This tests that
1689 the public header files can be usefully included in a C++ application.
1691 This test isn't enabled by default. It can be enabled with the option
1692 'enable-buildtest-c++'.
1696 * Enable SHA3 pre-hashing for ECDSA and DSA.
1700 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
1701 This changes the size when using the `genpkey` command when no size is given.
1702 It fixes an omission in earlier changes that changed all RSA, DSA and DH
1703 generation commands to use 2048 bits by default.
1707 * Reorganize the manual pages to consistently have RETURN VALUES,
1708 EXAMPLES, SEE ALSO and HISTORY come in that order, and adjust
1709 util/fix-doc-nits accordingly.
1711 *Paul Yang, Joshua Lock*
1713 * Add the missing accessor EVP_PKEY_get0_engine()
1717 * Have commands like `s_client` and `s_server` output the signature scheme
1718 along with other cipher suite parameters when debugging.
1722 * Make OPENSSL_config() error agnostic again.
1726 * Do the error handling in RSA decryption constant time.
1730 * Prevent over long nonces in ChaCha20-Poly1305.
1732 ChaCha20-Poly1305 is an AEAD cipher, and requires a unique nonce input
1733 for every encryption operation. RFC 7539 specifies that the nonce value
1734 (IV) should be 96 bits (12 bytes). OpenSSL allows a variable nonce length
1735 and front pads the nonce with 0 bytes if it is less than 12
1736 bytes. However it also incorrectly allows a nonce to be set of up to 16
1737 bytes. In this case only the last 12 bytes are significant and any
1738 additional leading bytes are ignored.
1740 It is a requirement of using this cipher that nonce values are
1741 unique. Messages encrypted using a reused nonce value are susceptible to
1742 serious confidentiality and integrity attacks. If an application changes
1743 the default nonce length to be longer than 12 bytes and then makes a
1744 change to the leading bytes of the nonce expecting the new value to be a
1745 new unique nonce then such an application could inadvertently encrypt
1746 messages with a reused nonce.
1748 Additionally the ignored bytes in a long nonce are not covered by the
1749 integrity guarantee of this cipher. Any application that relies on the
1750 integrity of these ignored leading bytes of a long nonce may be further
1751 affected. Any OpenSSL internal use of this cipher, including in SSL/TLS,
1752 is safe because no such use sets such a long nonce value. However user
1753 applications that use this cipher directly and set a non-default nonce
1754 length to be longer than 12 bytes may be vulnerable.
1756 This issue was reported to OpenSSL on 16th of March 2019 by Joran Dirk
1762 * Add DEVRANDOM_WAIT feature for Linux systems
1764 On older Linux systems where the getrandom() system call is not available,
1765 OpenSSL normally uses the /dev/urandom device for seeding its CSPRNG.
1766 Contrary to getrandom(), the /dev/urandom device will not block during
1767 early boot when the kernel CSPRNG has not been seeded yet.
1769 To mitigate this known weakness, use select() to wait for /dev/random to
1770 become readable before reading from /dev/urandom.
1772 * Ensure that SM2 only uses SM3 as digest algorithm
1776 ### Changes between 1.1.1a and 1.1.1b [26 Feb 2019]
1778 * Change the info callback signals for the start and end of a post-handshake
1779 message exchange in TLSv1.3. In 1.1.1/1.1.1a we used SSL_CB_HANDSHAKE_START
1780 and SSL_CB_HANDSHAKE_DONE. Experience has shown that many applications get
1781 confused by this and assume that a TLSv1.2 renegotiation has started. This
1782 can break KeyUpdate handling. Instead we no longer signal the start and end
1783 of a post handshake message exchange (although the messages themselves are
1784 still signalled). This could break some applications that were expecting
1785 the old signals. However without this KeyUpdate is not usable for many
1790 ### Changes between 1.1.1 and 1.1.1a [20 Nov 2018]
1792 * Timing vulnerability in DSA signature generation
1794 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
1795 timing side channel attack. An attacker could use variations in the signing
1796 algorithm to recover the private key.
1798 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
1803 * Timing vulnerability in ECDSA signature generation
1805 The OpenSSL ECDSA signature algorithm has been shown to be vulnerable to a
1806 timing side channel attack. An attacker could use variations in the signing
1807 algorithm to recover the private key.
1809 This issue was reported to OpenSSL on 25th October 2018 by Samuel Weiser.
1814 * Fixed the issue that RAND_add()/RAND_seed() silently discards random input
1815 if its length exceeds 4096 bytes. The limit has been raised to a buffer size
1816 of two gigabytes and the error handling improved.
1818 This issue was reported to OpenSSL by Dr. Falko Strenzke. It has been
1819 categorized as a normal bug, not a security issue, because the DRBG reseeds
1820 automatically and is fully functional even without additional randomness
1821 provided by the application.
1823 ### Changes between 1.1.0i and 1.1.1 [11 Sep 2018]
1825 * Add a new ClientHello callback. Provides a callback interface that gives
1826 the application the ability to adjust the nascent SSL object at the
1827 earliest stage of ClientHello processing, immediately after extensions have
1828 been collected but before they have been processed. In particular, this
1829 callback can adjust the supported TLS versions in response to the contents
1834 * Add SM2 base algorithm support.
1838 * s390x assembly pack: add (improved) hardware-support for the following
1839 cryptographic primitives: sha3, shake, aes-gcm, aes-ccm, aes-ctr, aes-ofb,
1840 aes-cfb/cfb8, aes-ecb.
1844 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
1845 parameter is no longer accepted, as it leads to a corrupt table. NULL
1846 pem_str is reserved for alias entries only.
1850 * Use the new ec_scalar_mul_ladder scaffold to implement a specialized ladder
1851 step for prime curves. The new implementation is based on formulae from
1852 differential addition-and-doubling in homogeneous projective coordinates
1853 from Izu-Takagi "A fast parallel elliptic curve multiplication resistant
1854 against side channel attacks" and Brier-Joye "Weierstrass Elliptic Curves
1855 and Side-Channel Attacks" Eq. (8) for y-coordinate recovery, modified
1856 to work in projective coordinates.
1858 *Billy Bob Brumley, Nicola Tuveri*
1860 * Change generating and checking of primes so that the error rate of not
1861 being prime depends on the intended use based on the size of the input.
1862 For larger primes this will result in more rounds of Miller-Rabin.
1863 The maximal error rate for primes with more than 1080 bits is lowered
1866 *Kurt Roeckx, Annie Yousar*
1868 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
1872 * The 'tsget' script is renamed to 'tsget.pl', to avoid confusion when
1873 moving between systems, and to avoid confusion when a Windows build is
1874 done with mingw vs with MSVC. For POSIX installs, there's still a
1875 symlink or copy named 'tsget' to avoid that confusion as well.
1879 * Revert blinding in ECDSA sign and instead make problematic addition
1880 length-invariant. Switch even to fixed-length Montgomery multiplication.
1884 * Use the new ec_scalar_mul_ladder scaffold to implement a specialized ladder
1885 step for binary curves. The new implementation is based on formulae from
1886 differential addition-and-doubling in mixed Lopez-Dahab projective
1887 coordinates, modified to independently blind the operands.
1889 *Billy Bob Brumley, Sohaib ul Hassan, Nicola Tuveri*
1891 * Add a scaffold to optionally enhance the Montgomery ladder implementation
1892 for `ec_scalar_mul_ladder` (formerly `ec_mul_consttime`) allowing
1893 EC_METHODs to implement their own specialized "ladder step", to take
1894 advantage of more favorable coordinate systems or more efficient
1895 differential addition-and-doubling algorithms.
1897 *Billy Bob Brumley, Sohaib ul Hassan, Nicola Tuveri*
1899 * Modified the random device based seed sources to keep the relevant
1900 file descriptors open rather than reopening them on each access.
1901 This allows such sources to operate in a chroot() jail without
1902 the associated device nodes being available. This behaviour can be
1903 controlled using RAND_keep_random_devices_open().
1907 * Numerous side-channel attack mitigations have been applied. This may have
1908 performance impacts for some algorithms for the benefit of improved
1909 security. Specific changes are noted in this change log by their respective
1914 * AIX shared library support overhaul. Switch to AIX "natural" way of
1915 handling shared libraries, which means collecting shared objects of
1916 different versions and bitnesses in one common archive. This allows to
1917 mitigate conflict between 1.0 and 1.1 side-by-side installations. It
1918 doesn't affect the way 3rd party applications are linked, only how
1919 multi-version installation is managed.
1923 * Make ec_group_do_inverse_ord() more robust and available to other
1924 EC cryptosystems, so that irrespective of BN_FLG_CONSTTIME, SCA
1925 mitigations are applied to the fallback BN_mod_inverse().
1926 When using this function rather than BN_mod_inverse() directly, new
1927 EC cryptosystem implementations are then safer-by-default.
1931 * Add coordinate blinding for EC_POINT and implement projective
1932 coordinate blinding for generic prime curves as a countermeasure to
1933 chosen point SCA attacks.
1935 *Sohaib ul Hassan, Nicola Tuveri, Billy Bob Brumley*
1937 * Add blinding to ECDSA and DSA signatures to protect against side channel
1938 attacks discovered by Keegan Ryan (NCC Group).
1942 * Enforce checking in the `pkeyutl` command to ensure that the input
1943 length does not exceed the maximum supported digest length when performing
1944 a sign, verify or verifyrecover operation.
1948 * SSL_MODE_AUTO_RETRY is enabled by default. Applications that use blocking
1949 I/O in combination with something like select() or poll() will hang. This
1950 can be turned off again using SSL_CTX_clear_mode().
1951 Many applications do not properly handle non-application data records, and
1952 TLS 1.3 sends more of such records. Setting SSL_MODE_AUTO_RETRY works
1953 around the problems in those applications, but can also break some.
1954 It's recommended to read the manpages about SSL_read(), SSL_write(),
1955 SSL_get_error(), SSL_shutdown(), SSL_CTX_set_mode() and
1956 SSL_CTX_set_read_ahead() again.
1960 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
1961 now allow empty (zero character) pass phrases.
1965 * Apply blinding to binary field modular inversion and remove patent
1966 pending (OPENSSL_SUN_GF2M_DIV) BN_GF2m_mod_div implementation.
1970 * Deprecate ec2_mult.c and unify scalar multiplication code paths for
1971 binary and prime elliptic curves.
1975 * Remove ECDSA nonce padding: EC_POINT_mul is now responsible for
1976 constant time fixed point multiplication.
1980 * Revise elliptic curve scalar multiplication with timing attack
1981 defenses: ec_wNAF_mul redirects to a constant time implementation
1982 when computing fixed point and variable point multiplication (which
1983 in OpenSSL are mostly used with secret scalars in keygen, sign,
1984 ECDH derive operations).
1985 *Billy Bob Brumley, Nicola Tuveri, Cesar Pereida García,
1988 * Updated CONTRIBUTING
1992 * Updated DRBG / RAND to request nonce and additional low entropy
1993 randomness from the system.
1995 *Matthias St. Pierre*
1997 * Updated 'openssl rehash' to use OpenSSL consistent default.
2001 * Moved the load of the ssl_conf module to libcrypto, which helps
2002 loading engines that libssl uses before libssl is initialised.
2006 * Added EVP_PKEY_sign() and EVP_PKEY_verify() for EdDSA
2010 * Fixed X509_NAME_ENTRY_set to get multi-valued RDNs right in all cases.
2012 *Ingo Schwarze, Rich Salz*
2014 * Added output of accepting IP address and port for 'openssl s_server'
2018 * Added a new API for TLSv1.3 ciphersuites:
2019 SSL_CTX_set_ciphersuites()
2020 SSL_set_ciphersuites()
2024 * Memory allocation failures consistently add an error to the error
2029 * Don't use OPENSSL_ENGINES and OPENSSL_CONF environment values
2030 in libcrypto when run as setuid/setgid.
2034 * Load any config file by default when libssl is used.
2038 * Added new public header file <openssl/rand_drbg.h> and documentation
2039 for the RAND_DRBG API. See manual page RAND_DRBG(7) for an overview.
2041 *Matthias St. Pierre*
2043 * QNX support removed (cannot find contributors to get their approval
2044 for the license change).
2048 * TLSv1.3 replay protection for early data has been implemented. See the
2049 SSL_read_early_data() man page for further details.
2053 * Separated TLSv1.3 ciphersuite configuration out from TLSv1.2 ciphersuite
2054 configuration. TLSv1.3 ciphersuites are not compatible with TLSv1.2 and
2055 below. Similarly TLSv1.2 ciphersuites are not compatible with TLSv1.3.
2056 In order to avoid issues where legacy TLSv1.2 ciphersuite configuration
2057 would otherwise inadvertently disable all TLSv1.3 ciphersuites the
2058 configuration has been separated out. See the ciphers man page or the
2059 SSL_CTX_set_ciphersuites() man page for more information.
2063 * On POSIX (BSD, Linux, ...) systems the ocsp(1) command running
2064 in responder mode now supports the new "-multi" option, which
2065 spawns the specified number of child processes to handle OCSP
2066 requests. The "-timeout" option now also limits the OCSP
2067 responder's patience to wait to receive the full client request
2068 on a newly accepted connection. Child processes are respawned
2069 as needed, and the CA index file is automatically reloaded
2070 when changed. This makes it possible to run the "ocsp" responder
2071 as a long-running service, making the OpenSSL CA somewhat more
2072 feature-complete. In this mode, most diagnostic messages logged
2073 after entering the event loop are logged via syslog(3) rather than
2078 * Added support for X448 and Ed448. Heavily based on original work by
2083 * Extend OSSL_STORE with capabilities to search and to narrow the set of
2084 objects loaded. This adds the functions OSSL_STORE_expect() and
2085 OSSL_STORE_find() as well as needed tools to construct searches and
2086 get the search data out of them.
2090 * Support for TLSv1.3 added. Note that users upgrading from an earlier
2091 version of OpenSSL should review their configuration settings to ensure
2092 that they are still appropriate for TLSv1.3. For further information see:
2093 <https://wiki.openssl.org/index.php/TLS1.3>
2097 * Grand redesign of the OpenSSL random generator
2099 The default RAND method now utilizes an AES-CTR DRBG according to
2100 NIST standard SP 800-90Ar1. The new random generator is essentially
2101 a port of the default random generator from the OpenSSL FIPS 2.0
2102 object module. It is a hybrid deterministic random bit generator
2103 using an AES-CTR bit stream and which seeds and reseeds itself
2104 automatically using trusted system entropy sources.
2106 Some of its new features are:
2107 - Support for multiple DRBG instances with seed chaining.
2108 - The default RAND method makes use of a DRBG.
2109 - There is a public and private DRBG instance.
2110 - The DRBG instances are fork-safe.
2111 - Keep all global DRBG instances on the secure heap if it is enabled.
2112 - The public and private DRBG instance are per thread for lock free
2115 *Paul Dale, Benjamin Kaduk, Kurt Roeckx, Rich Salz, Matthias St. Pierre*
2117 * Changed Configure so it only says what it does and doesn't dump
2118 so much data. Instead, ./configdata.pm should be used as a script
2119 to display all sorts of configuration data.
2123 * Added processing of "make variables" to Configure.
2127 * Added SHA512/224 and SHA512/256 algorithm support.
2131 * The last traces of Netware support, first removed in 1.1.0, have
2136 * Get rid of Makefile.shared, and in the process, make the processing
2137 of certain files (rc.obj, or the .def/.map/.opt files produced from
2138 the ordinal files) more visible and hopefully easier to trace and
2139 debug (or make silent).
2143 * Make it possible to have environment variable assignments as
2144 arguments to config / Configure.
2148 * Add multi-prime RSA (RFC 8017) support.
2152 * Add SM3 implemented according to GB/T 32905-2016
2153 *Jack Lloyd <jack.lloyd@ribose.com>,*
2154 *Ronald Tse <ronald.tse@ribose.com>,*
2155 *Erick Borsboom <erick.borsboom@ribose.com>*
2157 * Add 'Maximum Fragment Length' TLS extension negotiation and support
2158 as documented in RFC6066.
2159 Based on a patch from Tomasz Moń
2161 *Filipe Raimundo da Silva*
2163 * Add SM4 implemented according to GB/T 32907-2016.
2164 *Jack Lloyd <jack.lloyd@ribose.com>,*
2165 *Ronald Tse <ronald.tse@ribose.com>,*
2166 *Erick Borsboom <erick.borsboom@ribose.com>*
2168 * Reimplement -newreq-nodes and ERR_error_string_n; the
2169 original author does not agree with the license change.
2173 * Add ARIA AEAD TLS support.
2177 * Some macro definitions to support VS6 have been removed. Visual
2178 Studio 6 has not worked since 1.1.0
2182 * Add ERR_clear_last_mark(), to allow callers to clear the last mark
2183 without clearing the errors.
2187 * Add "atfork" functions. If building on a system that without
2188 pthreads, see doc/man3/OPENSSL_fork_prepare.pod for application
2189 requirements. The RAND facility now uses/requires this.
2197 * The UI API becomes a permanent and integral part of libcrypto, i.e.
2198 not possible to disable entirely. However, it's still possible to
2199 disable the console reading UI method, UI_OpenSSL() (use UI_null()
2202 To disable, configure with 'no-ui-console'. 'no-ui' is still
2203 possible to use as an alias. Check at compile time with the
2204 macro OPENSSL_NO_UI_CONSOLE. The macro OPENSSL_NO_UI is still
2205 possible to check and is an alias for OPENSSL_NO_UI_CONSOLE.
2209 * Add a STORE module, which implements a uniform and URI based reader of
2210 stores that can contain keys, certificates, CRLs and numerous other
2211 objects. The main API is loosely based on a few stdio functions,
2212 and includes OSSL_STORE_open, OSSL_STORE_load, OSSL_STORE_eof,
2213 OSSL_STORE_error and OSSL_STORE_close.
2214 The implementation uses backends called "loaders" to implement arbitrary
2215 URI schemes. There is one built in "loader" for the 'file' scheme.
2219 * Add devcrypto engine. This has been implemented against cryptodev-linux,
2220 then adjusted to work on FreeBSD 8.4 as well.
2221 Enable by configuring with 'enable-devcryptoeng'. This is done by default
2222 on BSD implementations, as cryptodev.h is assumed to exist on all of them.
2226 * Module names can prefixed with OSSL_ or OPENSSL_. This affects
2227 util/mkerr.pl, which is adapted to allow those prefixes, leading to
2228 error code calls like this:
2230 OSSL_FOOerr(OSSL_FOO_F_SOMETHING, OSSL_FOO_R_WHATEVER);
2232 With this change, we claim the namespaces OSSL and OPENSSL in a manner
2233 that can be encoded in C. For the foreseeable future, this will only
2236 *Richard Levitte and Tim Hudson*
2238 * Removed BSD cryptodev engine.
2242 * Add a build target 'build_all_generated', to build all generated files
2243 and only that. This can be used to prepare everything that requires
2244 things like perl for a system that lacks perl and then move everything
2245 to that system and do the rest of the build there.
2249 * In the UI interface, make it possible to duplicate the user data. This
2250 can be used by engines that need to retain the data for a longer time
2251 than just the call where this user data is passed.
2255 * Ignore the '-named_curve auto' value for compatibility of applications
2258 *Tomáš Mráz <tmraz@fedoraproject.org>*
2260 * Fragmented SSL/TLS alerts are no longer accepted. An alert message is 2
2261 bytes long. In theory it is permissible in SSLv3 - TLSv1.2 to fragment such
2262 alerts across multiple records (some of which could be empty). In practice
2263 it make no sense to send an empty alert record, or to fragment one. TLSv1.3
2264 prohibits this altogether and other libraries (BoringSSL, NSS) do not
2265 support this at all. Supporting it adds significant complexity to the
2266 record layer, and its removal is unlikely to cause interoperability
2271 * Add the ASN.1 types INT32, UINT32, INT64, UINT64 and variants prefixed
2272 with Z. These are meant to replace LONG and ZLONG and to be size safe.
2273 The use of LONG and ZLONG is discouraged and scheduled for deprecation
2278 * Add the 'z' and 'j' modifiers to BIO_printf() et al formatting string,
2279 'z' is to be used for [s]size_t, and 'j' - with [u]int64_t.
2281 *Richard Levitte, Andy Polyakov*
2283 * Add EC_KEY_get0_engine(), which does for EC_KEY what RSA_get0_engine()
2288 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
2289 platform rather than 'mingw'.
2293 * The functions X509_STORE_add_cert and X509_STORE_add_crl return
2294 success if they are asked to add an object which already exists
2295 in the store. This change cascades to other functions which load
2296 certificates and CRLs.
2300 * x86_64 assembly pack: annotate code with DWARF CFI directives to
2301 facilitate stack unwinding even from assembly subroutines.
2305 * Remove VAX C specific definitions of OPENSSL_EXPORT, OPENSSL_EXTERN.
2306 Also remove OPENSSL_GLOBAL entirely, as it became a no-op.
2310 * Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
2311 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
2312 which is the minimum version we support.
2316 * Certificate time validation (X509_cmp_time) enforces stricter
2317 compliance with RFC 5280. Fractional seconds and timezone offsets
2318 are no longer allowed.
2322 * Add support for ARIA
2326 * s_client will now send the Server Name Indication (SNI) extension by
2327 default unless the new "-noservername" option is used. The server name is
2328 based on the host provided to the "-connect" option unless overridden by
2329 using "-servername".
2333 * Add support for SipHash
2337 * OpenSSL now fails if it receives an unrecognised record type in TLS1.0
2338 or TLS1.1. Previously this only happened in SSLv3 and TLS1.2. This is to
2339 prevent issues where no progress is being made and the peer continually
2340 sends unrecognised record types, using up resources processing them.
2344 * 'openssl passwd' can now produce SHA256 and SHA512 based output,
2345 using the algorithm defined in
2346 <https://www.akkadia.org/drepper/SHA-crypt.txt>
2350 * Heartbeat support has been removed; the ABI is changed for now.
2352 *Richard Levitte, Rich Salz*
2354 * Support for SSL_OP_NO_ENCRYPT_THEN_MAC in SSL_CONF_cmd.
2358 * The RSA "null" method, which was partially supported to avoid patent
2359 issues, has been replaced to always returns NULL.
2366 ### Changes between 1.1.0k and 1.1.0l [10 Sep 2019]
2368 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
2369 used even when parsing explicit parameters, when loading a encoded key
2370 or calling `EC_GROUP_new_from_ecpkparameters()`/
2371 `EC_GROUP_new_from_ecparameters()`.
2372 This prevents bypass of security hardening and performance gains,
2373 especially for curves with specialized EC_METHODs.
2374 By default, if a key encoded with explicit parameters is loaded and later
2375 encoded, the output is still encoded with explicit parameters, even if
2376 internally a "named" EC_GROUP is used for computation.
2380 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
2381 this change, EC_GROUP_set_generator would accept order and/or cofactor as
2382 NULL. After this change, only the cofactor parameter can be NULL. It also
2383 does some minimal sanity checks on the passed order.
2388 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
2389 An attack is simple, if the first CMS_recipientInfo is valid but the
2390 second CMS_recipientInfo is chosen ciphertext. If the second
2391 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
2392 encryption key will be replaced by garbage, and the message cannot be
2393 decoded, but if the RSA decryption fails, the correct encryption key is
2394 used and the recipient will not notice the attack.
2395 As a work around for this potential attack the length of the decrypted
2396 key must be equal to the cipher default key length, in case the
2397 certifiate is not given and all recipientInfo are tried out.
2398 The old behaviour can be re-enabled in the CMS code by setting the
2399 CMS_DEBUG_DECRYPT flag.
2404 * Use Windows installation paths in the mingw builds
2406 Mingw isn't a POSIX environment per se, which means that Windows
2407 paths should be used for installation.
2412 ### Changes between 1.1.0j and 1.1.0k [28 May 2019]
2414 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
2415 This changes the size when using the `genpkey` command when no size is given.
2416 It fixes an omission in earlier changes that changed all RSA, DSA and DH
2417 generation commands to use 2048 bits by default.
2421 * Prevent over long nonces in ChaCha20-Poly1305.
2423 ChaCha20-Poly1305 is an AEAD cipher, and requires a unique nonce input
2424 for every encryption operation. RFC 7539 specifies that the nonce value
2425 (IV) should be 96 bits (12 bytes). OpenSSL allows a variable nonce length
2426 and front pads the nonce with 0 bytes if it is less than 12
2427 bytes. However it also incorrectly allows a nonce to be set of up to 16
2428 bytes. In this case only the last 12 bytes are significant and any
2429 additional leading bytes are ignored.
2431 It is a requirement of using this cipher that nonce values are
2432 unique. Messages encrypted using a reused nonce value are susceptible to
2433 serious confidentiality and integrity attacks. If an application changes
2434 the default nonce length to be longer than 12 bytes and then makes a
2435 change to the leading bytes of the nonce expecting the new value to be a
2436 new unique nonce then such an application could inadvertently encrypt
2437 messages with a reused nonce.
2439 Additionally the ignored bytes in a long nonce are not covered by the
2440 integrity guarantee of this cipher. Any application that relies on the
2441 integrity of these ignored leading bytes of a long nonce may be further
2442 affected. Any OpenSSL internal use of this cipher, including in SSL/TLS,
2443 is safe because no such use sets such a long nonce value. However user
2444 applications that use this cipher directly and set a non-default nonce
2445 length to be longer than 12 bytes may be vulnerable.
2447 This issue was reported to OpenSSL on 16th of March 2019 by Joran Dirk
2453 * Added SCA hardening for modular field inversion in EC_GROUP through
2454 a new dedicated field_inv() pointer in EC_METHOD.
2455 This also addresses a leakage affecting conversions from projective
2456 to affine coordinates.
2458 *Billy Bob Brumley, Nicola Tuveri*
2460 * Fix a use after free bug in d2i_X509_PUBKEY when overwriting a
2461 re-used X509_PUBKEY object if the second PUBKEY is malformed.
2465 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
2469 * Remove the 'dist' target and add a tarball building script. The
2470 'dist' target has fallen out of use, and it shouldn't be
2471 necessary to configure just to create a source distribution.
2475 ### Changes between 1.1.0i and 1.1.0j [20 Nov 2018]
2477 * Timing vulnerability in DSA signature generation
2479 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
2480 timing side channel attack. An attacker could use variations in the signing
2481 algorithm to recover the private key.
2483 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
2488 * Timing vulnerability in ECDSA signature generation
2490 The OpenSSL ECDSA signature algorithm has been shown to be vulnerable to a
2491 timing side channel attack. An attacker could use variations in the signing
2492 algorithm to recover the private key.
2494 This issue was reported to OpenSSL on 25th October 2018 by Samuel Weiser.
2499 * Add coordinate blinding for EC_POINT and implement projective
2500 coordinate blinding for generic prime curves as a countermeasure to
2501 chosen point SCA attacks.
2503 *Sohaib ul Hassan, Nicola Tuveri, Billy Bob Brumley*
2505 ### Changes between 1.1.0h and 1.1.0i [14 Aug 2018]
2507 * Client DoS due to large DH parameter
2509 During key agreement in a TLS handshake using a DH(E) based ciphersuite a
2510 malicious server can send a very large prime value to the client. This will
2511 cause the client to spend an unreasonably long period of time generating a
2512 key for this prime resulting in a hang until the client has finished. This
2513 could be exploited in a Denial Of Service attack.
2515 This issue was reported to OpenSSL on 5th June 2018 by Guido Vranken
2520 * Cache timing vulnerability in RSA Key Generation
2522 The OpenSSL RSA Key generation algorithm has been shown to be vulnerable to
2523 a cache timing side channel attack. An attacker with sufficient access to
2524 mount cache timing attacks during the RSA key generation process could
2525 recover the private key.
2527 This issue was reported to OpenSSL on 4th April 2018 by Alejandro Cabrera
2528 Aldaya, Billy Brumley, Cesar Pereida Garcia and Luis Manuel Alvarez Tapia.
2533 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
2534 parameter is no longer accepted, as it leads to a corrupt table. NULL
2535 pem_str is reserved for alias entries only.
2539 * Revert blinding in ECDSA sign and instead make problematic addition
2540 length-invariant. Switch even to fixed-length Montgomery multiplication.
2544 * Change generating and checking of primes so that the error rate of not
2545 being prime depends on the intended use based on the size of the input.
2546 For larger primes this will result in more rounds of Miller-Rabin.
2547 The maximal error rate for primes with more than 1080 bits is lowered
2550 *Kurt Roeckx, Annie Yousar*
2552 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
2556 * Add blinding to ECDSA and DSA signatures to protect against side channel
2557 attacks discovered by Keegan Ryan (NCC Group).
2561 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
2562 now allow empty (zero character) pass phrases.
2566 * Certificate time validation (X509_cmp_time) enforces stricter
2567 compliance with RFC 5280. Fractional seconds and timezone offsets
2568 are no longer allowed.
2572 * Fixed a text canonicalisation bug in CMS
2574 Where a CMS detached signature is used with text content the text goes
2575 through a canonicalisation process first prior to signing or verifying a
2576 signature. This process strips trailing space at the end of lines, converts
2577 line terminators to CRLF and removes additional trailing line terminators
2578 at the end of a file. A bug in the canonicalisation process meant that
2579 some characters, such as form-feed, were incorrectly treated as whitespace
2580 and removed. This is contrary to the specification (RFC5485). This fix
2581 could mean that detached text data signed with an earlier version of
2582 OpenSSL 1.1.0 may fail to verify using the fixed version, or text data
2583 signed with a fixed OpenSSL may fail to verify with an earlier version of
2584 OpenSSL 1.1.0. A workaround is to only verify the canonicalised text data
2585 and use the "-binary" flag (for the "cms" command line application) or set
2586 the SMIME_BINARY/PKCS7_BINARY/CMS_BINARY flags (if using CMS_verify()).
2590 ### Changes between 1.1.0g and 1.1.0h [27 Mar 2018]
2592 * Constructed ASN.1 types with a recursive definition could exceed the stack
2594 Constructed ASN.1 types with a recursive definition (such as can be found
2595 in PKCS7) could eventually exceed the stack given malicious input with
2596 excessive recursion. This could result in a Denial Of Service attack. There
2597 are no such structures used within SSL/TLS that come from untrusted sources
2598 so this is considered safe.
2600 This issue was reported to OpenSSL on 4th January 2018 by the OSS-fuzz
2606 * Incorrect CRYPTO_memcmp on HP-UX PA-RISC
2608 Because of an implementation bug the PA-RISC CRYPTO_memcmp function is
2609 effectively reduced to only comparing the least significant bit of each
2610 byte. This allows an attacker to forge messages that would be considered as
2611 authenticated in an amount of tries lower than that guaranteed by the
2612 security claims of the scheme. The module can only be compiled by the
2613 HP-UX assembler, so that only HP-UX PA-RISC targets are affected.
2615 This issue was reported to OpenSSL on 2nd March 2018 by Peter Waltenberg
2621 * Add a build target 'build_all_generated', to build all generated files
2622 and only that. This can be used to prepare everything that requires
2623 things like perl for a system that lacks perl and then move everything
2624 to that system and do the rest of the build there.
2628 * Backport SSL_OP_NO_RENGOTIATION
2630 OpenSSL 1.0.2 and below had the ability to disable renegotiation using the
2631 (undocumented) SSL3_FLAGS_NO_RENEGOTIATE_CIPHERS flag. Due to the opacity
2632 changes this is no longer possible in 1.1.0. Therefore the new
2633 SSL_OP_NO_RENEGOTIATION option from 1.1.1-dev has been backported to
2634 1.1.0 to provide equivalent functionality.
2636 Note that if an application built against 1.1.0h headers (or above) is run
2637 using an older version of 1.1.0 (prior to 1.1.0h) then the option will be
2638 accepted but nothing will happen, i.e. renegotiation will not be prevented.
2642 * Removed the OS390-Unix config target. It relied on a script that doesn't
2647 * rsaz_1024_mul_avx2 overflow bug on x86_64
2649 There is an overflow bug in the AVX2 Montgomery multiplication procedure
2650 used in exponentiation with 1024-bit moduli. No EC algorithms are affected.
2651 Analysis suggests that attacks against RSA and DSA as a result of this
2652 defect would be very difficult to perform and are not believed likely.
2653 Attacks against DH1024 are considered just feasible, because most of the
2654 work necessary to deduce information about a private key may be performed
2655 offline. The amount of resources required for such an attack would be
2656 significant. However, for an attack on TLS to be meaningful, the server
2657 would have to share the DH1024 private key among multiple clients, which is
2658 no longer an option since CVE-2016-0701.
2660 This only affects processors that support the AVX2 but not ADX extensions
2661 like Intel Haswell (4th generation).
2663 This issue was reported to OpenSSL by David Benjamin (Google). The issue
2664 was originally found via the OSS-Fuzz project.
2669 ### Changes between 1.1.0f and 1.1.0g [2 Nov 2017]
2671 * bn_sqrx8x_internal carry bug on x86_64
2673 There is a carry propagating bug in the x86_64 Montgomery squaring
2674 procedure. No EC algorithms are affected. Analysis suggests that attacks
2675 against RSA and DSA as a result of this defect would be very difficult to
2676 perform and are not believed likely. Attacks against DH are considered just
2677 feasible (although very difficult) because most of the work necessary to
2678 deduce information about a private key may be performed offline. The amount
2679 of resources required for such an attack would be very significant and
2680 likely only accessible to a limited number of attackers. An attacker would
2681 additionally need online access to an unpatched system using the target
2682 private key in a scenario with persistent DH parameters and a private
2683 key that is shared between multiple clients.
2685 This only affects processors that support the BMI1, BMI2 and ADX extensions
2686 like Intel Broadwell (5th generation) and later or AMD Ryzen.
2688 This issue was reported to OpenSSL by the OSS-Fuzz project.
2693 * Malformed X.509 IPAddressFamily could cause OOB read
2695 If an X.509 certificate has a malformed IPAddressFamily extension,
2696 OpenSSL could do a one-byte buffer overread. The most likely result
2697 would be an erroneous display of the certificate in text format.
2699 This issue was reported to OpenSSL by the OSS-Fuzz project.
2704 ### Changes between 1.1.0e and 1.1.0f [25 May 2017]
2706 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
2707 platform rather than 'mingw'.
2711 * Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
2712 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
2713 which is the minimum version we support.
2717 ### Changes between 1.1.0d and 1.1.0e [16 Feb 2017]
2719 * Encrypt-Then-Mac renegotiation crash
2721 During a renegotiation handshake if the Encrypt-Then-Mac extension is
2722 negotiated where it was not in the original handshake (or vice-versa) then
2723 this can cause OpenSSL to crash (dependant on ciphersuite). Both clients
2724 and servers are affected.
2726 This issue was reported to OpenSSL by Joe Orton (Red Hat).
2731 ### Changes between 1.1.0c and 1.1.0d [26 Jan 2017]
2733 * Truncated packet could crash via OOB read
2735 If one side of an SSL/TLS path is running on a 32-bit host and a specific
2736 cipher is being used, then a truncated packet can cause that host to
2737 perform an out-of-bounds read, usually resulting in a crash.
2739 This issue was reported to OpenSSL by Robert Święcki of Google.
2744 * Bad (EC)DHE parameters cause a client crash
2746 If a malicious server supplies bad parameters for a DHE or ECDHE key
2747 exchange then this can result in the client attempting to dereference a
2748 NULL pointer leading to a client crash. This could be exploited in a Denial
2751 This issue was reported to OpenSSL by Guido Vranken.
2756 * BN_mod_exp may produce incorrect results on x86_64
2758 There is a carry propagating bug in the x86_64 Montgomery squaring
2759 procedure. No EC algorithms are affected. Analysis suggests that attacks
2760 against RSA and DSA as a result of this defect would be very difficult to
2761 perform and are not believed likely. Attacks against DH are considered just
2762 feasible (although very difficult) because most of the work necessary to
2763 deduce information about a private key may be performed offline. The amount
2764 of resources required for such an attack would be very significant and
2765 likely only accessible to a limited number of attackers. An attacker would
2766 additionally need online access to an unpatched system using the target
2767 private key in a scenario with persistent DH parameters and a private
2768 key that is shared between multiple clients. For example this can occur by
2769 default in OpenSSL DHE based SSL/TLS ciphersuites. Note: This issue is very
2770 similar to CVE-2015-3193 but must be treated as a separate problem.
2772 This issue was reported to OpenSSL by the OSS-Fuzz project.
2777 ### Changes between 1.1.0b and 1.1.0c [10 Nov 2016]
2779 * ChaCha20/Poly1305 heap-buffer-overflow
2781 TLS connections using `*-CHACHA20-POLY1305` ciphersuites are susceptible to
2782 a DoS attack by corrupting larger payloads. This can result in an OpenSSL
2783 crash. This issue is not considered to be exploitable beyond a DoS.
2785 This issue was reported to OpenSSL by Robert Święcki (Google Security Team)
2790 * CMS Null dereference
2792 Applications parsing invalid CMS structures can crash with a NULL pointer
2793 dereference. This is caused by a bug in the handling of the ASN.1 CHOICE
2794 type in OpenSSL 1.1.0 which can result in a NULL value being passed to the
2795 structure callback if an attempt is made to free certain invalid encodings.
2796 Only CHOICE structures using a callback which do not handle NULL value are
2799 This issue was reported to OpenSSL by Tyler Nighswander of ForAllSecure.
2804 * Montgomery multiplication may produce incorrect results
2806 There is a carry propagating bug in the Broadwell-specific Montgomery
2807 multiplication procedure that handles input lengths divisible by, but
2808 longer than 256 bits. Analysis suggests that attacks against RSA, DSA
2809 and DH private keys are impossible. This is because the subroutine in
2810 question is not used in operations with the private key itself and an input
2811 of the attacker's direct choice. Otherwise the bug can manifest itself as
2812 transient authentication and key negotiation failures or reproducible
2813 erroneous outcome of public-key operations with specially crafted input.
2814 Among EC algorithms only Brainpool P-512 curves are affected and one
2815 presumably can attack ECDH key negotiation. Impact was not analyzed in
2816 detail, because pre-requisites for attack are considered unlikely. Namely
2817 multiple clients have to choose the curve in question and the server has to
2818 share the private key among them, neither of which is default behaviour.
2819 Even then only clients that chose the curve will be affected.
2821 This issue was publicly reported as transient failures and was not
2822 initially recognized as a security issue. Thanks to Richard Morgan for
2823 providing reproducible case.
2828 * Removed automatic addition of RPATH in shared libraries and executables,
2829 as this was a remainder from OpenSSL 1.0.x and isn't needed any more.
2833 ### Changes between 1.1.0a and 1.1.0b [26 Sep 2016]
2835 * Fix Use After Free for large message sizes
2837 The patch applied to address CVE-2016-6307 resulted in an issue where if a
2838 message larger than approx 16k is received then the underlying buffer to
2839 store the incoming message is reallocated and moved. Unfortunately a
2840 dangling pointer to the old location is left which results in an attempt to
2841 write to the previously freed location. This is likely to result in a
2842 crash, however it could potentially lead to execution of arbitrary code.
2844 This issue only affects OpenSSL 1.1.0a.
2846 This issue was reported to OpenSSL by Robert Święcki.
2851 ### Changes between 1.1.0 and 1.1.0a [22 Sep 2016]
2853 * OCSP Status Request extension unbounded memory growth
2855 A malicious client can send an excessively large OCSP Status Request
2856 extension. If that client continually requests renegotiation, sending a
2857 large OCSP Status Request extension each time, then there will be unbounded
2858 memory growth on the server. This will eventually lead to a Denial Of
2859 Service attack through memory exhaustion. Servers with a default
2860 configuration are vulnerable even if they do not support OCSP. Builds using
2861 the "no-ocsp" build time option are not affected.
2863 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
2868 * SSL_peek() hang on empty record
2870 OpenSSL 1.1.0 SSL/TLS will hang during a call to SSL_peek() if the peer
2871 sends an empty record. This could be exploited by a malicious peer in a
2872 Denial Of Service attack.
2874 This issue was reported to OpenSSL by Alex Gaynor.
2879 * Excessive allocation of memory in tls_get_message_header() and
2880 dtls1_preprocess_fragment()
2882 A (D)TLS message includes 3 bytes for its length in the header for the
2883 message. This would allow for messages up to 16Mb in length. Messages of
2884 this length are excessive and OpenSSL includes a check to ensure that a
2885 peer is sending reasonably sized messages in order to avoid too much memory
2886 being consumed to service a connection. A flaw in the logic of version
2887 1.1.0 means that memory for the message is allocated too early, prior to
2888 the excessive message length check. Due to way memory is allocated in
2889 OpenSSL this could mean an attacker could force up to 21Mb to be allocated
2890 to service a connection. This could lead to a Denial of Service through
2891 memory exhaustion. However, the excessive message length check still takes
2892 place, and this would cause the connection to immediately fail. Assuming
2893 that the application calls SSL_free() on the failed connection in a timely
2894 manner then the 21Mb of allocated memory will then be immediately freed
2895 again. Therefore the excessive memory allocation will be transitory in
2896 nature. This then means that there is only a security impact if:
2898 1) The application does not call SSL_free() in a timely manner in the event
2899 that the connection fails
2901 2) The application is working in a constrained environment where there is
2902 very little free memory
2904 3) The attacker initiates multiple connection attempts such that there are
2905 multiple connections in a state where memory has been allocated for the
2906 connection; SSL_free() has not yet been called; and there is insufficient
2907 memory to service the multiple requests.
2909 Except in the instance of (1) above any Denial Of Service is likely to be
2910 transitory because as soon as the connection fails the memory is
2911 subsequently freed again in the SSL_free() call. However there is an
2912 increased risk during this period of application crashes due to the lack of
2913 memory - which would then mean a more serious Denial of Service.
2915 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
2916 (CVE-2016-6307 and CVE-2016-6308)
2920 * solaris-x86-cc, i.e. 32-bit configuration with vendor compiler,
2921 had to be removed. Primary reason is that vendor assembler can't
2922 assemble our modules with -KPIC flag. As result it, assembly
2923 support, was not even available as option. But its lack means
2924 lack of side-channel resistant code, which is incompatible with
2925 security by todays standards. Fortunately gcc is readily available
2926 prepackaged option, which we firmly point at...
2930 ### Changes between 1.0.2h and 1.1.0 [25 Aug 2016]
2932 * Windows command-line tool supports UTF-8 opt-in option for arguments
2933 and console input. Setting OPENSSL_WIN32_UTF8 environment variable
2934 (to any value) allows Windows user to access PKCS#12 file generated
2935 with Windows CryptoAPI and protected with non-ASCII password, as well
2936 as files generated under UTF-8 locale on Linux also protected with
2941 * To mitigate the SWEET32 attack ([CVE-2016-2183]), 3DES cipher suites
2942 have been disabled by default and removed from DEFAULT, just like RC4.
2943 See the RC4 item below to re-enable both.
2947 * The method for finding the storage location for the Windows RAND seed file
2948 has changed. First we check %RANDFILE%. If that is not set then we check
2949 the directories %HOME%, %USERPROFILE% and %SYSTEMROOT% in that order. If
2950 all else fails we fall back to C:\.
2954 * The EVP_EncryptUpdate() function has had its return type changed from void
2955 to int. A return of 0 indicates and error while a return of 1 indicates
2960 * The flags RSA_FLAG_NO_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME and
2961 DH_FLAG_NO_EXP_CONSTTIME which previously provided the ability to switch
2962 off the constant time implementation for RSA, DSA and DH have been made
2963 no-ops and deprecated.
2967 * Windows RAND implementation was simplified to only get entropy by
2968 calling CryptGenRandom(). Various other RAND-related tickets
2971 *Joseph Wylie Yandle, Rich Salz*
2973 * The stack and lhash API's were renamed to start with `OPENSSL_SK_`
2974 and `OPENSSL_LH_`, respectively. The old names are available
2975 with API compatibility. They new names are now completely documented.
2979 * Unify TYPE_up_ref(obj) methods signature.
2980 SSL_CTX_up_ref(), SSL_up_ref(), X509_up_ref(), EVP_PKEY_up_ref(),
2981 X509_CRL_up_ref(), X509_OBJECT_up_ref_count() methods are now returning an
2982 int (instead of void) like all others TYPE_up_ref() methods.
2983 So now these methods also check the return value of CRYPTO_atomic_add(),
2984 and the validity of object reference counter.
2986 *fdasilvayy@gmail.com*
2988 * With Windows Visual Studio builds, the .pdb files are installed
2989 alongside the installed libraries and executables. For a static
2990 library installation, ossl_static.pdb is the associate compiler
2991 generated .pdb file to be used when linking programs.
2995 * Remove openssl.spec. Packaging files belong with the packagers.
2999 * Automatic Darwin/OSX configuration has had a refresh, it will now
3000 recognise x86_64 architectures automatically. You can still decide
3001 to build for a different bitness with the environment variable
3002 KERNEL_BITS (can be 32 or 64), for example:
3004 KERNEL_BITS=32 ./config
3008 * Change default algorithms in pkcs8 utility to use PKCS#5 v2.0,
3009 256 bit AES and HMAC with SHA256.
3013 * Remove support for MIPS o32 ABI on IRIX (and IRIX only).
3017 * Triple-DES ciphers have been moved from HIGH to MEDIUM.
3021 * To enable users to have their own config files and build file templates,
3022 Configure looks in the directory indicated by the environment variable
3023 OPENSSL_LOCAL_CONFIG_DIR as well as the in-source Configurations/
3024 directory. On VMS, OPENSSL_LOCAL_CONFIG_DIR is expected to be a logical
3025 name and is used as is.
3029 * The following datatypes were made opaque: X509_OBJECT, X509_STORE_CTX,
3030 X509_STORE, X509_LOOKUP, and X509_LOOKUP_METHOD. The unused type
3031 X509_CERT_FILE_CTX was removed.
3035 * "shared" builds are now the default. To create only static libraries use
3036 the "no-shared" Configure option.
3040 * Remove the no-aes, no-hmac, no-rsa, no-sha and no-md5 Configure options.
3041 All of these option have not worked for some while and are fundamental
3046 * Make various cleanup routines no-ops and mark them as deprecated. Most
3047 global cleanup functions are no longer required because they are handled
3048 via auto-deinit (see OPENSSL_init_crypto and OPENSSL_init_ssl man pages).
3049 Explicitly de-initing can cause problems (e.g. where a library that uses
3050 OpenSSL de-inits, but an application is still using it). The affected
3051 functions are CONF_modules_free(), ENGINE_cleanup(), OBJ_cleanup(),
3052 EVP_cleanup(), BIO_sock_cleanup(), CRYPTO_cleanup_all_ex_data(),
3053 RAND_cleanup(), SSL_COMP_free_compression_methods(), ERR_free_strings() and
3054 COMP_zlib_cleanup().
3058 * --strict-warnings no longer enables runtime debugging options
3059 such as REF_DEBUG. Instead, debug options are automatically
3060 enabled with '--debug' builds.
3062 *Andy Polyakov, Emilia Käsper*
3064 * Made DH and DH_METHOD opaque. The structures for managing DH objects
3065 have been moved out of the public header files. New functions for managing
3066 these have been added.
3070 * Made RSA and RSA_METHOD opaque. The structures for managing RSA
3071 objects have been moved out of the public header files. New
3072 functions for managing these have been added.
3076 * Made DSA and DSA_METHOD opaque. The structures for managing DSA objects
3077 have been moved out of the public header files. New functions for managing
3078 these have been added.
3082 * Made BIO and BIO_METHOD opaque. The structures for managing BIOs have been
3083 moved out of the public header files. New functions for managing these
3088 * Removed no-rijndael as a config option. Rijndael is an old name for AES.
3092 * Removed the mk1mf build scripts.
3096 * Headers are now wrapped, if necessary, with OPENSSL_NO_xxx, so
3097 it is always safe to #include a header now.
3101 * Removed the aged BC-32 config and all its supporting scripts
3105 * Removed support for Ultrix, Netware, and OS/2.
3109 * Add support for HKDF.
3111 *Alessandro Ghedini*
3113 * Add support for blake2b and blake2s
3117 * Added support for "pipelining". Ciphers that have the
3118 EVP_CIPH_FLAG_PIPELINE flag set have a capability to process multiple
3119 encryptions/decryptions simultaneously. There are currently no built-in
3120 ciphers with this property but the expectation is that engines will be able
3121 to offer it to significantly improve throughput. Support has been extended
3122 into libssl so that multiple records for a single connection can be
3123 processed in one go (for >=TLS 1.1).
3127 * Added the AFALG engine. This is an async capable engine which is able to
3128 offload work to the Linux kernel. In this initial version it only supports
3129 AES128-CBC. The kernel must be version 4.1.0 or greater.
3133 * OpenSSL now uses a new threading API. It is no longer necessary to
3134 set locking callbacks to use OpenSSL in a multi-threaded environment. There
3135 are two supported threading models: pthreads and windows threads. It is
3136 also possible to configure OpenSSL at compile time for "no-threads". The
3137 old threading API should no longer be used. The functions have been
3138 replaced with "no-op" compatibility macros.
3140 *Alessandro Ghedini, Matt Caswell*
3142 * Modify behavior of ALPN to invoke callback after SNI/servername
3143 callback, such that updates to the SSL_CTX affect ALPN.
3147 * Add SSL_CIPHER queries for authentication and key-exchange.
3151 * Changes to the DEFAULT cipherlist:
3152 - Prefer (EC)DHE handshakes over plain RSA.
3153 - Prefer AEAD ciphers over legacy ciphers.
3154 - Prefer ECDSA over RSA when both certificates are available.
3155 - Prefer TLSv1.2 ciphers/PRF.
3156 - Remove DSS, SEED, IDEA, CAMELLIA, and AES-CCM from the
3161 * Change the ECC default curve list to be this, in order: x25519,
3162 secp256r1, secp521r1, secp384r1.
3166 * RC4 based libssl ciphersuites are now classed as "weak" ciphers and are
3167 disabled by default. They can be re-enabled using the
3168 enable-weak-ssl-ciphers option to Configure.
3172 * If the server has ALPN configured, but supports no protocols that the
3173 client advertises, send a fatal "no_application_protocol" alert.
3174 This behaviour is SHALL in RFC 7301, though it isn't universally
3175 implemented by other servers.
3179 * Add X25519 support.
3180 Add ASN.1 and EVP_PKEY methods for X25519. This includes support
3181 for public and private key encoding using the format documented in
3182 draft-ietf-curdle-pkix-02. The corresponding EVP_PKEY method supports
3183 key generation and key derivation.
3185 TLS support complies with draft-ietf-tls-rfc4492bis-08 and uses
3190 * Deprecate SRP_VBASE_get_by_user.
3191 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
3192 In order to fix an unavoidable memory leak ([CVE-2016-0798]),
3193 SRP_VBASE_get_by_user was changed to ignore the "fake user" SRP
3194 seed, even if the seed is configured.
3196 Users should use SRP_VBASE_get1_by_user instead. Note that in
3197 SRP_VBASE_get1_by_user, caller must free the returned value. Note
3198 also that even though configuring the SRP seed attempts to hide
3199 invalid usernames by continuing the handshake with fake
3200 credentials, this behaviour is not constant time and no strong
3201 guarantees are made that the handshake is indistinguishable from
3202 that of a valid user.
3206 * Configuration change; it's now possible to build dynamic engines
3207 without having to build shared libraries and vice versa. This
3208 only applies to the engines in `engines/`, those in `crypto/engine/`
3209 will always be built into libcrypto (i.e. "static").
3211 Building dynamic engines is enabled by default; to disable, use
3212 the configuration option "disable-dynamic-engine".
3214 The only requirements for building dynamic engines are the
3215 presence of the DSO module and building with position independent
3216 code, so they will also automatically be disabled if configuring
3217 with "disable-dso" or "disable-pic".
3219 The macros OPENSSL_NO_STATIC_ENGINE and OPENSSL_NO_DYNAMIC_ENGINE
3220 are also taken away from openssl/opensslconf.h, as they are
3225 * Configuration change; if there is a known flag to compile
3226 position independent code, it will always be applied on the
3227 libcrypto and libssl object files, and never on the application
3228 object files. This means other libraries that use routines from
3229 libcrypto / libssl can be made into shared libraries regardless
3230 of how OpenSSL was configured.
3232 If this isn't desirable, the configuration options "disable-pic"
3233 or "no-pic" can be used to disable the use of PIC. This will
3234 also disable building shared libraries and dynamic engines.
3238 * Removed JPAKE code. It was experimental and has no wide use.
3242 * The INSTALL_PREFIX Makefile variable has been renamed to
3243 DESTDIR. That makes for less confusion on what this variable
3244 is for. Also, the configuration option --install_prefix is
3249 * Heartbeat for TLS has been removed and is disabled by default
3250 for DTLS; configure with enable-heartbeats. Code that uses the
3251 old #define's might need to be updated.
3253 *Emilia Käsper, Rich Salz*
3255 * Rename REF_CHECK to REF_DEBUG.
3259 * New "unified" build system
3261 The "unified" build system is aimed to be a common system for all
3262 platforms we support. With it comes new support for VMS.
3264 This system builds supports building in a different directory tree
3265 than the source tree. It produces one Makefile (for unix family
3266 or lookalikes), or one descrip.mms (for VMS).
3268 The source of information to make the Makefile / descrip.mms is
3269 small files called 'build.info', holding the necessary
3270 information for each directory with source to compile, and a
3271 template in Configurations, like unix-Makefile.tmpl or
3274 With this change, the library names were also renamed on Windows
3275 and on VMS. They now have names that are closer to the standard
3276 on Unix, and include the major version number, and in certain
3277 cases, the architecture they are built for. See "Notes on shared
3278 libraries" in INSTALL.
3280 We rely heavily on the perl module Text::Template.
3284 * Added support for auto-initialisation and de-initialisation of the library.
3285 OpenSSL no longer requires explicit init or deinit routines to be called,
3286 except in certain circumstances. See the OPENSSL_init_crypto() and
3287 OPENSSL_init_ssl() man pages for further information.
3291 * The arguments to the DTLSv1_listen function have changed. Specifically the
3292 "peer" argument is now expected to be a BIO_ADDR object.
3294 * Rewrite of BIO networking library. The BIO library lacked consistent
3295 support of IPv6, and adding it required some more extensive
3296 modifications. This introduces the BIO_ADDR and BIO_ADDRINFO types,
3297 which hold all types of addresses and chains of address information.
3298 It also introduces a new API, with functions like BIO_socket,
3299 BIO_connect, BIO_listen, BIO_lookup and a rewrite of BIO_accept.
3300 The source/sink BIOs BIO_s_connect, BIO_s_accept and BIO_s_datagram
3301 have been adapted accordingly.
3305 * RSA_padding_check_PKCS1_type_1 now accepts inputs with and without
3310 * CRIME protection: disable compression by default, even if OpenSSL is
3311 compiled with zlib enabled. Applications can still enable compression
3312 by calling SSL_CTX_clear_options(ctx, SSL_OP_NO_COMPRESSION), or by
3313 using the SSL_CONF library to configure compression.
3317 * The signature of the session callback configured with
3318 SSL_CTX_sess_set_get_cb was changed. The read-only input buffer
3319 was explicitly marked as `const unsigned char*` instead of
3324 * Always DPURIFY. Remove the use of uninitialized memory in the
3325 RNG, and other conditional uses of DPURIFY. This makes -DPURIFY a no-op.
3329 * Removed many obsolete configuration items, including
3330 DES_PTR, DES_RISC1, DES_RISC2, DES_INT
3331 MD2_CHAR, MD2_INT, MD2_LONG
3333 IDEA_SHORT, IDEA_LONG
3334 RC2_SHORT, RC2_LONG, RC4_LONG, RC4_CHUNK, RC4_INDEX
3336 *Rich Salz, with advice from Andy Polyakov*
3338 * Many BN internals have been moved to an internal header file.
3340 *Rich Salz with help from Andy Polyakov*
3342 * Configuration and writing out the results from it has changed.
3343 Files such as Makefile include/openssl/opensslconf.h and are now
3344 produced through general templates, such as Makefile.in and
3345 crypto/opensslconf.h.in and some help from the perl module
3348 Also, the center of configuration information is no longer
3349 Makefile. Instead, Configure produces a perl module in
3350 configdata.pm which holds most of the config data (in the hash
3351 table %config), the target data that comes from the target
3352 configuration in one of the `Configurations/*.conf` files (in
3357 * To clarify their intended purposes, the Configure options
3358 --prefix and --openssldir change their semantics, and become more
3359 straightforward and less interdependent.
3361 --prefix shall be used exclusively to give the location INSTALLTOP
3362 where programs, scripts, libraries, include files and manuals are
3363 going to be installed. The default is now /usr/local.
3365 --openssldir shall be used exclusively to give the default
3366 location OPENSSLDIR where certificates, private keys, CRLs are
3367 managed. This is also where the default openssl.cnf gets
3369 If the directory given with this option is a relative path, the
3370 values of both the --prefix value and the --openssldir value will
3371 be combined to become OPENSSLDIR.
3372 The default for --openssldir is INSTALLTOP/ssl.
3374 Anyone who uses --openssldir to specify where OpenSSL is to be
3375 installed MUST change to use --prefix instead.
3379 * The GOST engine was out of date and therefore it has been removed. An up
3380 to date GOST engine is now being maintained in an external repository.
3381 See: <https://wiki.openssl.org/index.php/Binaries>. Libssl still retains
3382 support for GOST ciphersuites (these are only activated if a GOST engine
3387 * EGD is no longer supported by default; use enable-egd when
3390 *Ben Kaduk and Rich Salz*
3392 * The distribution now has Makefile.in files, which are used to
3393 create Makefile's when Configure is run. *Configure must be run
3394 before trying to build now.*
3398 * The return value for SSL_CIPHER_description() for error conditions
3403 * Support for RFC6698/RFC7671 DANE TLSA peer authentication.
3405 Obtaining and performing DNSSEC validation of TLSA records is
3406 the application's responsibility. The application provides
3407 the TLSA records of its choice to OpenSSL, and these are then
3408 used to authenticate the peer.
3410 The TLSA records need not even come from DNS. They can, for
3411 example, be used to implement local end-entity certificate or
3412 trust-anchor "pinning", where the "pin" data takes the form
3413 of TLSA records, which can augment or replace verification
3414 based on the usual WebPKI public certification authorities.
3418 * Revert default OPENSSL_NO_DEPRECATED setting. Instead OpenSSL
3419 continues to support deprecated interfaces in default builds.
3420 However, applications are strongly advised to compile their
3421 source files with -DOPENSSL_API_COMPAT=0x10100000L, which hides
3422 the declarations of all interfaces deprecated in 0.9.8, 1.0.0
3423 or the 1.1.0 releases.
3425 In environments in which all applications have been ported to
3426 not use any deprecated interfaces OpenSSL's Configure script
3427 should be used with the --api=1.1.0 option to entirely remove
3428 support for the deprecated features from the library and
3429 unconditionally disable them in the installed headers.
3430 Essentially the same effect can be achieved with the "no-deprecated"
3431 argument to Configure, except that this will always restrict
3432 the build to just the latest API, rather than a fixed API
3435 As applications are ported to future revisions of the API,
3436 they should update their compile-time OPENSSL_API_COMPAT define
3437 accordingly, but in most cases should be able to continue to
3438 compile with later releases.
3440 The OPENSSL_API_COMPAT versions for 1.0.0, and 0.9.8 are
3441 0x10000000L and 0x00908000L, respectively. However those
3442 versions did not support the OPENSSL_API_COMPAT feature, and
3443 so applications are not typically tested for explicit support
3444 of just the undeprecated features of either release.
3448 * Add support for setting the minimum and maximum supported protocol.
3449 It can bet set via the SSL_set_min_proto_version() and
3450 SSL_set_max_proto_version(), or via the SSL_CONF's MinProtocol and
3451 MaxProtocol. It's recommended to use the new APIs to disable
3452 protocols instead of disabling individual protocols using
3453 SSL_set_options() or SSL_CONF's Protocol. This change also
3454 removes support for disabling TLS 1.2 in the OpenSSL TLS
3455 client at compile time by defining OPENSSL_NO_TLS1_2_CLIENT.
3459 * Support for ChaCha20 and Poly1305 added to libcrypto and libssl.
3463 * New EC_KEY_METHOD, this replaces the older ECDSA_METHOD and ECDH_METHOD
3464 and integrates ECDSA and ECDH functionality into EC. Implementations can
3465 now redirect key generation and no longer need to convert to or from
3468 Note: the ecdsa.h and ecdh.h headers are now no longer needed and just
3469 include the ec.h header file instead.
3473 * Remove support for all 40 and 56 bit ciphers. This includes all the export
3474 ciphers who are no longer supported and drops support the ephemeral RSA key
3475 exchange. The LOW ciphers currently doesn't have any ciphers in it.
3479 * Made EVP_MD_CTX, EVP_MD, EVP_CIPHER_CTX, EVP_CIPHER and HMAC_CTX
3480 opaque. For HMAC_CTX, the following constructors and destructors
3483 HMAC_CTX *HMAC_CTX_new(void);
3484 void HMAC_CTX_free(HMAC_CTX *ctx);
3486 For EVP_MD and EVP_CIPHER, complete APIs to create, fill and
3487 destroy such methods has been added. See EVP_MD_meth_new(3) and
3488 EVP_CIPHER_meth_new(3) for documentation.
3491 1) `EVP_MD_CTX_cleanup()`, `EVP_CIPHER_CTX_cleanup()` and
3492 `HMAC_CTX_cleanup()` were removed. `HMAC_CTX_reset()` and
3493 `EVP_MD_CTX_reset()` should be called instead to reinitialise
3494 an already created structure.
3495 2) For consistency with the majority of our object creators and
3496 destructors, `EVP_MD_CTX_(create|destroy)` were renamed to
3497 `EVP_MD_CTX_(new|free)`. The old names are retained as macros
3498 for deprecated builds.
3502 * Added ASYNC support. Libcrypto now includes the async sub-library to enable
3503 cryptographic operations to be performed asynchronously as long as an
3504 asynchronous capable engine is used. See the ASYNC_start_job() man page for
3505 further details. Libssl has also had this capability integrated with the
3506 introduction of the new mode SSL_MODE_ASYNC and associated error
3507 SSL_ERROR_WANT_ASYNC. See the SSL_CTX_set_mode() and SSL_get_error() man
3508 pages. This work was developed in partnership with Intel Corp.
3512 * SSL_{CTX_}set_ecdh_auto() has been removed and ECDH is support is
3513 always enabled now. If you want to disable the support you should
3514 exclude it using the list of supported ciphers. This also means that the
3515 "-no_ecdhe" option has been removed from s_server.
3519 * SSL_{CTX}_set_tmp_ecdh() which can set 1 EC curve now internally calls
3520 SSL_{CTX_}set1_curves() which can set a list.
3524 * Remove support for SSL_{CTX_}set_tmp_ecdh_callback(). You should set the
3525 curve you want to support using SSL_{CTX_}set1_curves().
3529 * State machine rewrite. The state machine code has been significantly
3530 refactored in order to remove much duplication of code and solve issues
3531 with the old code (see [ssl/statem/README.md](ssl/statem/README.md) for
3532 further details). This change does have some associated API changes.
3533 Notably the SSL_state() function has been removed and replaced by
3534 SSL_get_state which now returns an "OSSL_HANDSHAKE_STATE" instead of an int.
3535 SSL_set_state() has been removed altogether. The previous handshake states
3536 defined in ssl.h and ssl3.h have also been removed.
3540 * All instances of the string "ssleay" in the public API were replaced
3541 with OpenSSL (case-matching; e.g., OPENSSL_VERSION for #define's)
3542 Some error codes related to internal RSA_eay API's were renamed.
3546 * The demo files in crypto/threads were moved to demo/threads.
3550 * Removed obsolete engines: 4758cca, aep, atalla, cswift, nuron, gmp,
3553 *Matt Caswell, Rich Salz*
3555 * New ASN.1 embed macro.
3557 New ASN.1 macro ASN1_EMBED. This is the same as ASN1_SIMPLE except the
3558 structure is not allocated: it is part of the parent. That is instead of
3566 This reduces memory fragmentation and make it impossible to accidentally
3567 set a mandatory field to NULL.
3569 This currently only works for some fields specifically a SEQUENCE, CHOICE,
3570 or ASN1_STRING type which is part of a parent SEQUENCE. Since it is
3571 equivalent to ASN1_SIMPLE it cannot be tagged, OPTIONAL, SET OF or
3576 * Remove EVP_CHECK_DES_KEY, a compile-time option that never compiled.
3580 * Removed DES and RC4 ciphersuites from DEFAULT. Also removed RC2 although
3581 in 1.0.2 EXPORT was already removed and the only RC2 ciphersuite is also
3582 an EXPORT one. COMPLEMENTOFDEFAULT has been updated accordingly to add
3583 DES and RC4 ciphersuites.
3587 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
3588 This changes the decoding behaviour for some invalid messages,
3589 though the change is mostly in the more lenient direction, and
3590 legacy behaviour is preserved as much as possible.
3594 * Fix no-stdio build.
3595 *David Woodhouse <David.Woodhouse@intel.com> and also*
3596 *Ivan Nestlerode <ivan.nestlerode@sonos.com>*
3598 * New testing framework
3599 The testing framework has been largely rewritten and is now using
3600 perl and the perl modules Test::Harness and an extended variant of
3601 Test::More called OpenSSL::Test to do its work. All test scripts in
3602 test/ have been rewritten into test recipes, and all direct calls to
3603 executables in test/Makefile have become individual recipes using the
3604 simplified testing OpenSSL::Test::Simple.
3606 For documentation on our testing modules, do:
3608 perldoc test/testlib/OpenSSL/Test/Simple.pm
3609 perldoc test/testlib/OpenSSL/Test.pm
3613 * Revamped memory debug; only -DCRYPTO_MDEBUG and -DCRYPTO_MDEBUG_ABORT
3614 are used; the latter aborts on memory leaks (usually checked on exit).
3615 Some undocumented "set malloc, etc., hooks" functions were removed
3616 and others were changed. All are now documented.
3620 * In DSA_generate_parameters_ex, if the provided seed is too short,
3623 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
3625 * Rewrite PSK to support ECDHE_PSK, DHE_PSK and RSA_PSK. Add ciphersuites
3626 from RFC4279, RFC4785, RFC5487, RFC5489.
3628 Thanks to Christian J. Dietrich and Giuseppe D'Angelo for the
3629 original RSA_PSK patch.
3633 * Dropped support for the SSL3_FLAGS_DELAY_CLIENT_FINISHED flag. This SSLeay
3634 era flag was never set throughout the codebase (only read). Also removed
3635 SSL3_FLAGS_POP_BUFFER which was only used if
3636 SSL3_FLAGS_DELAY_CLIENT_FINISHED was also set.
3640 * Changed the default name options in the "ca", "crl", "req" and "x509"
3641 to be "oneline" instead of "compat".
3645 * Remove SSL_OP_TLS_BLOCK_PADDING_BUG. This is SSLeay legacy, we're
3646 not aware of clients that still exhibit this bug, and the workaround
3647 hasn't been working properly for a while.
3651 * The return type of BIO_number_read() and BIO_number_written() as well as
3652 the corresponding num_read and num_write members in the BIO structure has
3653 changed from unsigned long to uint64_t. On platforms where an unsigned
3654 long is 32 bits (e.g. Windows) these counters could overflow if >4Gb is
3659 * Given the pervasive nature of TLS extensions it is inadvisable to run
3660 OpenSSL without support for them. It also means that maintaining
3661 the OPENSSL_NO_TLSEXT option within the code is very invasive (and probably
3662 not well tested). Therefore the OPENSSL_NO_TLSEXT option has been removed.
3666 * Removed support for the two export grade static DH ciphersuites
3667 EXP-DH-RSA-DES-CBC-SHA and EXP-DH-DSS-DES-CBC-SHA. These two ciphersuites
3668 were newly added (along with a number of other static DH ciphersuites) to
3669 1.0.2. However the two export ones have *never* worked since they were
3670 introduced. It seems strange in any case to be adding new export
3671 ciphersuites, and given "logjam" it also does not seem correct to fix them.
3675 * Version negotiation has been rewritten. In particular SSLv23_method(),
3676 SSLv23_client_method() and SSLv23_server_method() have been deprecated,
3677 and turned into macros which simply call the new preferred function names
3678 TLS_method(), TLS_client_method() and TLS_server_method(). All new code
3679 should use the new names instead. Also as part of this change the ssl23.h
3680 header file has been removed.
3684 * Support for Kerberos ciphersuites in TLS (RFC2712) has been removed. This
3685 code and the associated standard is no longer considered fit-for-purpose.
3689 * RT2547 was closed. When generating a private key, try to make the
3690 output file readable only by the owner. This behavior change might
3691 be noticeable when interacting with other software.
3693 * Documented all exdata functions. Added CRYPTO_free_ex_index.
3698 * Added HTTP GET support to the ocsp command.
3702 * Changed default digest for the dgst and enc commands from MD5 to
3707 * RAND_pseudo_bytes has been deprecated. Users should use RAND_bytes instead.
3711 * Added support for TLS extended master secret from
3712 draft-ietf-tls-session-hash-03.txt. Thanks for Alfredo Pironti for an
3713 initial patch which was a great help during development.
3717 * All libssl internal structures have been removed from the public header
3718 files, and the OPENSSL_NO_SSL_INTERN option has been removed (since it is
3719 now redundant). Users should not attempt to access internal structures
3720 directly. Instead they should use the provided API functions.
3724 * config has been changed so that by default OPENSSL_NO_DEPRECATED is used.
3725 Access to deprecated functions can be re-enabled by running config with
3726 "enable-deprecated". In addition applications wishing to use deprecated
3727 functions must define OPENSSL_USE_DEPRECATED. Note that this new behaviour
3728 will, by default, disable some transitive includes that previously existed
3729 in the header files (e.g. ec.h will no longer, by default, include bn.h)
3733 * Added support for OCB mode. OpenSSL has been granted a patent license
3734 compatible with the OpenSSL license for use of OCB. Details are available
3735 at <https://www.openssl.org/source/OCB-patent-grant-OpenSSL.pdf>. Support
3736 for OCB can be removed by calling config with no-ocb.
3740 * SSLv2 support has been removed. It still supports receiving a SSLv2
3741 compatible client hello.
3745 * Increased the minimal RSA keysize from 256 to 512 bits [Rich Salz],
3746 done while fixing the error code for the key-too-small case.
3748 *Annie Yousar <a.yousar@informatik.hu-berlin.de>*
3750 * CA.sh has been removed; use CA.pl instead.
3754 * Removed old DES API.
3758 * Remove various unsupported platforms:
3764 Sinix/ReliantUNIX RM400
3769 16-bit platforms such as WIN16
3773 * Clean up OPENSSL_NO_xxx #define's
3774 - Use setbuf() and remove OPENSSL_NO_SETVBUF_IONBF
3775 - Rename OPENSSL_SYSNAME_xxx to OPENSSL_SYS_xxx
3776 - OPENSSL_NO_EC{DH,DSA} merged into OPENSSL_NO_EC
3777 - OPENSSL_NO_RIPEMD160, OPENSSL_NO_RIPEMD merged into OPENSSL_NO_RMD160
3778 - OPENSSL_NO_FP_API merged into OPENSSL_NO_STDIO
3779 - Remove OPENSSL_NO_BIO OPENSSL_NO_BUFFER OPENSSL_NO_CHAIN_VERIFY
3780 OPENSSL_NO_EVP OPENSSL_NO_FIPS_ERR OPENSSL_NO_HASH_COMP
3781 OPENSSL_NO_LHASH OPENSSL_NO_OBJECT OPENSSL_NO_SPEED OPENSSL_NO_STACK
3782 OPENSSL_NO_X509 OPENSSL_NO_X509_VERIFY
3783 - Remove MS_STATIC; it's a relic from platforms <32 bits.
3787 * Cleaned up dead code
3788 Remove all but one '#ifdef undef' which is to be looked at.
3792 * Clean up calling of xxx_free routines.
3793 Just like free(), fix most of the xxx_free routines to accept
3794 NULL. Remove the non-null checks from callers. Save much code.
3798 * Add secure heap for storage of private keys (when possible).
3799 Add BIO_s_secmem(), CBIGNUM, etc.
3800 Contributed by Akamai Technologies under our Corporate CLA.
3804 * Experimental support for a new, fast, unbiased prime candidate generator,
3805 bn_probable_prime_dh_coprime(). Not currently used by any prime generator.
3807 *Felix Laurie von Massenbach <felix@erbridge.co.uk>*
3809 * New output format NSS in the sess_id command line tool. This allows
3810 exporting the session id and the master key in NSS keylog format.
3812 *Martin Kaiser <martin@kaiser.cx>*
3814 * Harmonize version and its documentation. -f flag is used to display
3817 *mancha <mancha1@zoho.com>*
3819 * Fix eckey_priv_encode so it immediately returns an error upon a failure
3820 in i2d_ECPrivateKey. Thanks to Ted Unangst for feedback on this issue.
3822 *mancha <mancha1@zoho.com>*
3824 * Fix some double frees. These are not thought to be exploitable.
3826 *mancha <mancha1@zoho.com>*
3828 * A missing bounds check in the handling of the TLS heartbeat extension
3829 can be used to reveal up to 64k of memory to a connected client or
3832 Thanks for Neel Mehta of Google Security for discovering this bug and to
3833 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
3834 preparing the fix ([CVE-2014-0160])
3836 *Adam Langley, Bodo Moeller*
3838 * Fix for the attack described in the paper "Recovering OpenSSL
3839 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
3840 by Yuval Yarom and Naomi Benger. Details can be obtained from:
3841 <http://eprint.iacr.org/2014/140>
3843 Thanks to Yuval Yarom and Naomi Benger for discovering this
3844 flaw and to Yuval Yarom for supplying a fix ([CVE-2014-0076])
3846 *Yuval Yarom and Naomi Benger*
3848 * Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
3849 this fixes a limitation in previous versions of OpenSSL.
3853 * Experimental encrypt-then-mac support.
3855 Experimental support for encrypt then mac from
3856 draft-gutmann-tls-encrypt-then-mac-02.txt
3858 To enable it set the appropriate extension number (0x42 for the test
3859 server) using e.g. -DTLSEXT_TYPE_encrypt_then_mac=0x42
3861 For non-compliant peers (i.e. just about everything) this should have no
3864 WARNING: EXPERIMENTAL, SUBJECT TO CHANGE.
3868 * Add EVP support for key wrapping algorithms, to avoid problems with
3869 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
3870 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
3871 algorithms and include tests cases.
3875 * Extend CMS code to support RSA-PSS signatures and RSA-OAEP for
3880 * Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
3881 MGF1 digest and OAEP label.
3885 * Make openssl verify return errors.
3887 *Chris Palmer <palmer@google.com> and Ben Laurie*
3889 * New function ASN1_TIME_diff to calculate the difference between two
3890 ASN1_TIME structures or one structure and the current time.
3894 * Update fips_test_suite to support multiple command line options. New
3895 test to induce all self test errors in sequence and check expected
3900 * Add FIPS_{rsa,dsa,ecdsa}_{sign,verify} functions which digest and
3901 sign or verify all in one operation.
3905 * Add fips_algvs: a multicall fips utility incorporating all the algorithm
3906 test programs and fips_test_suite. Includes functionality to parse
3907 the minimal script output of fipsalgest.pl directly.
3911 * Add authorisation parameter to FIPS_module_mode_set().
3915 * Add FIPS selftest for ECDH algorithm using P-224 and B-233 curves.
3919 * Use separate DRBG fields for internal and external flags. New function
3920 FIPS_drbg_health_check() to perform on demand health checking. Add
3921 generation tests to fips_test_suite with reduced health check interval to
3922 demonstrate periodic health checking. Add "nodh" option to
3923 fips_test_suite to skip very slow DH test.
3927 * New function FIPS_get_cipherbynid() to lookup FIPS supported ciphers
3932 * More extensive health check for DRBG checking many more failure modes.
3933 New function FIPS_selftest_drbg_all() to handle every possible DRBG
3934 combination: call this in fips_test_suite.
3938 * Add support for canonical generation of DSA parameter 'g'. See
3941 * Add support for HMAC DRBG from SP800-90. Update DRBG algorithm test and
3942 POST to handle HMAC cases.
3946 * Add functions FIPS_module_version() and FIPS_module_version_text()
3947 to return numerical and string versions of the FIPS module number.
3951 * Rename FIPS_mode_set and FIPS_mode to FIPS_module_mode_set and
3952 FIPS_module_mode. FIPS_mode and FIPS_mode_set will be implemented
3953 outside the validated module in the FIPS capable OpenSSL.
3957 * Minor change to DRBG entropy callback semantics. In some cases
3958 there is no multiple of the block length between min_len and
3959 max_len. Allow the callback to return more than max_len bytes
3960 of entropy but discard any extra: it is the callback's responsibility
3961 to ensure that the extra data discarded does not impact the
3962 requested amount of entropy.
3966 * Add PRNG security strength checks to RSA, DSA and ECDSA using
3967 information in FIPS186-3, SP800-57 and SP800-131A.
3971 * CCM support via EVP. Interface is very similar to GCM case except we
3972 must supply all data in one chunk (i.e. no update, final) and the
3973 message length must be supplied if AAD is used. Add algorithm test
3978 * Initial version of POST overhaul. Add POST callback to allow the status
3979 of POST to be monitored and/or failures induced. Modify fips_test_suite
3980 to use callback. Always run all selftests even if one fails.
3984 * XTS support including algorithm test driver in the fips_gcmtest program.
3985 Note: this does increase the maximum key length from 32 to 64 bytes but
3986 there should be no binary compatibility issues as existing applications
3987 will never use XTS mode.
3991 * Extensive reorganisation of FIPS PRNG behaviour. Remove all dependencies
3992 to OpenSSL RAND code and replace with a tiny FIPS RAND API which also
3993 performs algorithm blocking for unapproved PRNG types. Also do not
3994 set PRNG type in FIPS_mode_set(): leave this to the application.
3995 Add default OpenSSL DRBG handling: sets up FIPS PRNG and seeds with
3996 the standard OpenSSL PRNG: set additional data to a date time vector.
4000 * Rename old X9.31 PRNG functions of the form `FIPS_rand*` to `FIPS_x931*`.
4001 This shouldn't present any incompatibility problems because applications
4002 shouldn't be using these directly and any that are will need to rethink
4003 anyway as the X9.31 PRNG is now deprecated by FIPS 140-2
4007 * Extensive self tests and health checking required by SP800-90 DRBG.
4008 Remove strength parameter from FIPS_drbg_instantiate and always
4009 instantiate at maximum supported strength.
4013 * Add ECDH code to fips module and fips_ecdhvs for primitives only testing.
4017 * New algorithm test program fips_dhvs to handle DH primitives only testing.
4021 * New function DH_compute_key_padded() to compute a DH key and pad with
4022 leading zeroes if needed: this complies with SP800-56A et al.
4026 * Initial implementation of SP800-90 DRBGs for Hash and CTR. Not used by
4027 anything, incomplete, subject to change and largely untested at present.
4031 * Modify fipscanisteronly build option to only build the necessary object
4032 files by filtering FIPS_EX_OBJ through a perl script in crypto/Makefile.
4036 * Add experimental option FIPSSYMS to give all symbols in
4037 fipscanister.o and FIPS or fips prefix. This will avoid
4038 conflicts with future versions of OpenSSL. Add perl script
4039 util/fipsas.pl to preprocess assembly language source files
4040 and rename any affected symbols.
4044 * Add selftest checks and algorithm block of non-fips algorithms in
4045 FIPS mode. Remove DES2 from selftests.
4049 * Add ECDSA code to fips module. Add tiny fips_ecdsa_check to just
4050 return internal method without any ENGINE dependencies. Add new
4051 tiny fips sign and verify functions.
4055 * New build option no-ec2m to disable characteristic 2 code.
4059 * New build option "fipscanisteronly". This only builds fipscanister.o
4060 and (currently) associated fips utilities. Uses the file Makefile.fips
4061 instead of Makefile.org as the prototype.
4065 * Add some FIPS mode restrictions to GCM. Add internal IV generator.
4066 Update fips_gcmtest to use IV generator.
4070 * Initial, experimental EVP support for AES-GCM. AAD can be input by
4071 setting output buffer to NULL. The `*Final` function must be
4072 called although it will not retrieve any additional data. The tag
4073 can be set or retrieved with a ctrl. The IV length is by default 12
4074 bytes (96 bits) but can be set to an alternative value. If the IV
4075 length exceeds the maximum IV length (currently 16 bytes) it cannot be
4080 * New flag in ciphers: EVP_CIPH_FLAG_CUSTOM_CIPHER. This means the
4081 underlying do_cipher function handles all cipher semantics itself
4082 including padding and finalisation. This is useful if (for example)
4083 an ENGINE cipher handles block padding itself. The behaviour of
4084 do_cipher is subtly changed if this flag is set: the return value
4085 is the number of characters written to the output buffer (zero is
4086 no longer an error code) or a negative error code. Also if the
4087 input buffer is NULL and length 0 finalisation should be performed.
4091 * If a candidate issuer certificate is already part of the constructed
4092 path ignore it: new debug notification X509_V_ERR_PATH_LOOP for this case.
4096 * Improve forward-security support: add functions
4098 void SSL_CTX_set_not_resumable_session_callback(
4099 SSL_CTX *ctx, int (*cb)(SSL *ssl, int is_forward_secure))
4100 void SSL_set_not_resumable_session_callback(
4101 SSL *ssl, int (*cb)(SSL *ssl, int is_forward_secure))
4103 for use by SSL/TLS servers; the callback function will be called whenever a
4104 new session is created, and gets to decide whether the session may be
4105 cached to make it resumable (return 0) or not (return 1). (As by the
4106 SSL/TLS protocol specifications, the session_id sent by the server will be
4107 empty to indicate that the session is not resumable; also, the server will
4108 not generate RFC 4507 (RFC 5077) session tickets.)
4110 A simple reasonable callback implementation is to return is_forward_secure.
4111 This parameter will be set to 1 or 0 depending on the ciphersuite selected
4112 by the SSL/TLS server library, indicating whether it can provide forward
4115 *Emilia Käsper <emilia.kasper@esat.kuleuven.be> (Google)*
4117 * New -verify_name option in command line utilities to set verification
4122 * Initial CMAC implementation. WARNING: EXPERIMENTAL, API MAY CHANGE.
4123 Add CMAC pkey methods.
4127 * Experimental renegotiation in s_server -www mode. If the client
4128 browses /reneg connection is renegotiated. If /renegcert it is
4129 renegotiated requesting a certificate.
4133 * Add an "external" session cache for debugging purposes to s_server. This
4134 should help trace issues which normally are only apparent in deployed
4135 multi-process servers.
4139 * Extensive audit of libcrypto with DEBUG_UNUSED. Fix many cases where
4140 return value is ignored. NB. The functions RAND_add(), RAND_seed(),
4141 BIO_set_cipher() and some obscure PEM functions were changed so they
4142 can now return an error. The RAND changes required a change to the
4143 RAND_METHOD structure.
4147 * New macro `__owur` for "OpenSSL Warn Unused Result". This makes use of
4148 a gcc attribute to warn if the result of a function is ignored. This
4149 is enable if DEBUG_UNUSED is set. Add to several functions in evp.h
4150 whose return value is often ignored.
4154 * New -noct, -requestct, -requirect and -ctlogfile options for s_client.
4155 These allow SCTs (signed certificate timestamps) to be requested and
4156 validated when establishing a connection.
4158 *Rob Percival <robpercival@google.com>*
4163 ### Changes between 1.0.2s and 1.0.2t [10 Sep 2019]
4165 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
4166 used even when parsing explicit parameters, when loading a encoded key
4167 or calling `EC_GROUP_new_from_ecpkparameters()`/
4168 `EC_GROUP_new_from_ecparameters()`.
4169 This prevents bypass of security hardening and performance gains,
4170 especially for curves with specialized EC_METHODs.
4171 By default, if a key encoded with explicit parameters is loaded and later
4172 encoded, the output is still encoded with explicit parameters, even if
4173 internally a "named" EC_GROUP is used for computation.
4177 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
4178 this change, EC_GROUP_set_generator would accept order and/or cofactor as
4179 NULL. After this change, only the cofactor parameter can be NULL. It also
4180 does some minimal sanity checks on the passed order.
4185 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
4186 An attack is simple, if the first CMS_recipientInfo is valid but the
4187 second CMS_recipientInfo is chosen ciphertext. If the second
4188 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
4189 encryption key will be replaced by garbage, and the message cannot be
4190 decoded, but if the RSA decryption fails, the correct encryption key is
4191 used and the recipient will not notice the attack.
4192 As a work around for this potential attack the length of the decrypted
4193 key must be equal to the cipher default key length, in case the
4194 certifiate is not given and all recipientInfo are tried out.
4195 The old behaviour can be re-enabled in the CMS code by setting the
4196 CMS_DEBUG_DECRYPT flag.
4201 * Document issue with installation paths in diverse Windows builds
4203 '/usr/local/ssl' is an unsafe prefix for location to install OpenSSL
4204 binaries and run-time config file.
4209 ### Changes between 1.0.2r and 1.0.2s [28 May 2019]
4211 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
4212 This changes the size when using the `genpkey` command when no size is given.
4213 It fixes an omission in earlier changes that changed all RSA, DSA and DH
4214 generation commands to use 2048 bits by default.
4218 * Add FIPS support for Android Arm 64-bit
4220 Support for Android Arm 64-bit was added to the OpenSSL FIPS Object
4221 Module in Version 2.0.10. For some reason, the corresponding target
4222 'android64-aarch64' was missing OpenSSL 1.0.2, whence it could not be
4223 built with FIPS support on Android Arm 64-bit. This omission has been
4226 *Matthias St. Pierre*
4228 ### Changes between 1.0.2q and 1.0.2r [26 Feb 2019]
4230 * 0-byte record padding oracle
4232 If an application encounters a fatal protocol error and then calls
4233 SSL_shutdown() twice (once to send a close_notify, and once to receive one)
4234 then OpenSSL can respond differently to the calling application if a 0 byte
4235 record is received with invalid padding compared to if a 0 byte record is
4236 received with an invalid MAC. If the application then behaves differently
4237 based on that in a way that is detectable to the remote peer, then this
4238 amounts to a padding oracle that could be used to decrypt data.
4240 In order for this to be exploitable "non-stitched" ciphersuites must be in
4241 use. Stitched ciphersuites are optimised implementations of certain
4242 commonly used ciphersuites. Also the application must call SSL_shutdown()
4243 twice even if a protocol error has occurred (applications should not do
4244 this but some do anyway).
4246 This issue was discovered by Juraj Somorovsky, Robert Merget and Nimrod
4247 Aviram, with additional investigation by Steven Collison and Andrew
4248 Hourselt. It was reported to OpenSSL on 10th December 2018.
4253 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
4257 ### Changes between 1.0.2p and 1.0.2q [20 Nov 2018]
4259 * Microarchitecture timing vulnerability in ECC scalar multiplication
4261 OpenSSL ECC scalar multiplication, used in e.g. ECDSA and ECDH, has been
4262 shown to be vulnerable to a microarchitecture timing side channel attack.
4263 An attacker with sufficient access to mount local timing attacks during
4264 ECDSA signature generation could recover the private key.
4266 This issue was reported to OpenSSL on 26th October 2018 by Alejandro
4267 Cabrera Aldaya, Billy Brumley, Sohaib ul Hassan, Cesar Pereida Garcia and
4273 * Timing vulnerability in DSA signature generation
4275 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
4276 timing side channel attack. An attacker could use variations in the signing
4277 algorithm to recover the private key.
4279 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
4284 * Resolve a compatibility issue in EC_GROUP handling with the FIPS Object
4285 Module, accidentally introduced while backporting security fixes from the
4286 development branch and hindering the use of ECC in FIPS mode.
4290 ### Changes between 1.0.2o and 1.0.2p [14 Aug 2018]
4292 * Client DoS due to large DH parameter
4294 During key agreement in a TLS handshake using a DH(E) based ciphersuite a
4295 malicious server can send a very large prime value to the client. This will
4296 cause the client to spend an unreasonably long period of time generating a
4297 key for this prime resulting in a hang until the client has finished. This
4298 could be exploited in a Denial Of Service attack.
4300 This issue was reported to OpenSSL on 5th June 2018 by Guido Vranken
4305 * Cache timing vulnerability in RSA Key Generation
4307 The OpenSSL RSA Key generation algorithm has been shown to be vulnerable to
4308 a cache timing side channel attack. An attacker with sufficient access to
4309 mount cache timing attacks during the RSA key generation process could
4310 recover the private key.
4312 This issue was reported to OpenSSL on 4th April 2018 by Alejandro Cabrera
4313 Aldaya, Billy Brumley, Cesar Pereida Garcia and Luis Manuel Alvarez Tapia.
4318 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
4319 parameter is no longer accepted, as it leads to a corrupt table. NULL
4320 pem_str is reserved for alias entries only.
4324 * Revert blinding in ECDSA sign and instead make problematic addition
4325 length-invariant. Switch even to fixed-length Montgomery multiplication.
4329 * Change generating and checking of primes so that the error rate of not
4330 being prime depends on the intended use based on the size of the input.
4331 For larger primes this will result in more rounds of Miller-Rabin.
4332 The maximal error rate for primes with more than 1080 bits is lowered
4335 *Kurt Roeckx, Annie Yousar*
4337 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
4341 * Add blinding to ECDSA and DSA signatures to protect against side channel
4342 attacks discovered by Keegan Ryan (NCC Group).
4346 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
4347 now allow empty (zero character) pass phrases.
4351 * Certificate time validation (X509_cmp_time) enforces stricter
4352 compliance with RFC 5280. Fractional seconds and timezone offsets
4353 are no longer allowed.
4357 ### Changes between 1.0.2n and 1.0.2o [27 Mar 2018]
4359 * Constructed ASN.1 types with a recursive definition could exceed the stack
4361 Constructed ASN.1 types with a recursive definition (such as can be found
4362 in PKCS7) could eventually exceed the stack given malicious input with
4363 excessive recursion. This could result in a Denial Of Service attack. There
4364 are no such structures used within SSL/TLS that come from untrusted sources
4365 so this is considered safe.
4367 This issue was reported to OpenSSL on 4th January 2018 by the OSS-fuzz
4373 ### Changes between 1.0.2m and 1.0.2n [7 Dec 2017]
4375 * Read/write after SSL object in error state
4377 OpenSSL 1.0.2 (starting from version 1.0.2b) introduced an "error state"
4378 mechanism. The intent was that if a fatal error occurred during a handshake
4379 then OpenSSL would move into the error state and would immediately fail if
4380 you attempted to continue the handshake. This works as designed for the
4381 explicit handshake functions (SSL_do_handshake(), SSL_accept() and
4382 SSL_connect()), however due to a bug it does not work correctly if
4383 SSL_read() or SSL_write() is called directly. In that scenario, if the
4384 handshake fails then a fatal error will be returned in the initial function
4385 call. If SSL_read()/SSL_write() is subsequently called by the application
4386 for the same SSL object then it will succeed and the data is passed without
4387 being decrypted/encrypted directly from the SSL/TLS record layer.
4389 In order to exploit this issue an application bug would have to be present
4390 that resulted in a call to SSL_read()/SSL_write() being issued after having
4391 already received a fatal error.
4393 This issue was reported to OpenSSL by David Benjamin (Google).
4398 * rsaz_1024_mul_avx2 overflow bug on x86_64
4400 There is an overflow bug in the AVX2 Montgomery multiplication procedure
4401 used in exponentiation with 1024-bit moduli. No EC algorithms are affected.
4402 Analysis suggests that attacks against RSA and DSA as a result of this
4403 defect would be very difficult to perform and are not believed likely.
4404 Attacks against DH1024 are considered just feasible, because most of the
4405 work necessary to deduce information about a private key may be performed
4406 offline. The amount of resources required for such an attack would be
4407 significant. However, for an attack on TLS to be meaningful, the server
4408 would have to share the DH1024 private key among multiple clients, which is
4409 no longer an option since CVE-2016-0701.
4411 This only affects processors that support the AVX2 but not ADX extensions
4412 like Intel Haswell (4th generation).
4414 This issue was reported to OpenSSL by David Benjamin (Google). The issue
4415 was originally found via the OSS-Fuzz project.
4420 ### Changes between 1.0.2l and 1.0.2m [2 Nov 2017]
4422 * bn_sqrx8x_internal carry bug on x86_64
4424 There is a carry propagating bug in the x86_64 Montgomery squaring
4425 procedure. No EC algorithms are affected. Analysis suggests that attacks
4426 against RSA and DSA as a result of this defect would be very difficult to
4427 perform and are not believed likely. Attacks against DH are considered just
4428 feasible (although very difficult) because most of the work necessary to
4429 deduce information about a private key may be performed offline. The amount
4430 of resources required for such an attack would be very significant and
4431 likely only accessible to a limited number of attackers. An attacker would
4432 additionally need online access to an unpatched system using the target
4433 private key in a scenario with persistent DH parameters and a private
4434 key that is shared between multiple clients.
4436 This only affects processors that support the BMI1, BMI2 and ADX extensions
4437 like Intel Broadwell (5th generation) and later or AMD Ryzen.
4439 This issue was reported to OpenSSL by the OSS-Fuzz project.
4444 * Malformed X.509 IPAddressFamily could cause OOB read
4446 If an X.509 certificate has a malformed IPAddressFamily extension,
4447 OpenSSL could do a one-byte buffer overread. The most likely result
4448 would be an erroneous display of the certificate in text format.
4450 This issue was reported to OpenSSL by the OSS-Fuzz project.
4454 ### Changes between 1.0.2k and 1.0.2l [25 May 2017]
4456 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
4457 platform rather than 'mingw'.
4461 ### Changes between 1.0.2j and 1.0.2k [26 Jan 2017]
4463 * Truncated packet could crash via OOB read
4465 If one side of an SSL/TLS path is running on a 32-bit host and a specific
4466 cipher is being used, then a truncated packet can cause that host to
4467 perform an out-of-bounds read, usually resulting in a crash.
4469 This issue was reported to OpenSSL by Robert Święcki of Google.
4474 * BN_mod_exp may produce incorrect results on x86_64
4476 There is a carry propagating bug in the x86_64 Montgomery squaring
4477 procedure. No EC algorithms are affected. Analysis suggests that attacks
4478 against RSA and DSA as a result of this defect would be very difficult to
4479 perform and are not believed likely. Attacks against DH are considered just
4480 feasible (although very difficult) because most of the work necessary to
4481 deduce information about a private key may be performed offline. The amount
4482 of resources required for such an attack would be very significant and
4483 likely only accessible to a limited number of attackers. An attacker would
4484 additionally need online access to an unpatched system using the target
4485 private key in a scenario with persistent DH parameters and a private
4486 key that is shared between multiple clients. For example this can occur by
4487 default in OpenSSL DHE based SSL/TLS ciphersuites. Note: This issue is very
4488 similar to CVE-2015-3193 but must be treated as a separate problem.
4490 This issue was reported to OpenSSL by the OSS-Fuzz project.
4495 * Montgomery multiplication may produce incorrect results
4497 There is a carry propagating bug in the Broadwell-specific Montgomery
4498 multiplication procedure that handles input lengths divisible by, but
4499 longer than 256 bits. Analysis suggests that attacks against RSA, DSA
4500 and DH private keys are impossible. This is because the subroutine in
4501 question is not used in operations with the private key itself and an input
4502 of the attacker's direct choice. Otherwise the bug can manifest itself as
4503 transient authentication and key negotiation failures or reproducible
4504 erroneous outcome of public-key operations with specially crafted input.
4505 Among EC algorithms only Brainpool P-512 curves are affected and one
4506 presumably can attack ECDH key negotiation. Impact was not analyzed in
4507 detail, because pre-requisites for attack are considered unlikely. Namely
4508 multiple clients have to choose the curve in question and the server has to
4509 share the private key among them, neither of which is default behaviour.
4510 Even then only clients that chose the curve will be affected.
4512 This issue was publicly reported as transient failures and was not
4513 initially recognized as a security issue. Thanks to Richard Morgan for
4514 providing reproducible case.
4519 * OpenSSL now fails if it receives an unrecognised record type in TLS1.0
4520 or TLS1.1. Previously this only happened in SSLv3 and TLS1.2. This is to
4521 prevent issues where no progress is being made and the peer continually
4522 sends unrecognised record types, using up resources processing them.
4526 ### Changes between 1.0.2i and 1.0.2j [26 Sep 2016]
4528 * Missing CRL sanity check
4530 A bug fix which included a CRL sanity check was added to OpenSSL 1.1.0
4531 but was omitted from OpenSSL 1.0.2i. As a result any attempt to use
4532 CRLs in OpenSSL 1.0.2i will crash with a null pointer exception.
4534 This issue only affects the OpenSSL 1.0.2i
4539 ### Changes between 1.0.2h and 1.0.2i [22 Sep 2016]
4541 * OCSP Status Request extension unbounded memory growth
4543 A malicious client can send an excessively large OCSP Status Request
4544 extension. If that client continually requests renegotiation, sending a
4545 large OCSP Status Request extension each time, then there will be unbounded
4546 memory growth on the server. This will eventually lead to a Denial Of
4547 Service attack through memory exhaustion. Servers with a default
4548 configuration are vulnerable even if they do not support OCSP. Builds using
4549 the "no-ocsp" build time option are not affected.
4551 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4556 * In order to mitigate the SWEET32 attack, the DES ciphers were moved from
4559 This issue was reported to OpenSSL Karthikeyan Bhargavan and Gaetan
4565 * OOB write in MDC2_Update()
4567 An overflow can occur in MDC2_Update() either if called directly or
4568 through the EVP_DigestUpdate() function using MDC2. If an attacker
4569 is able to supply very large amounts of input data after a previous
4570 call to EVP_EncryptUpdate() with a partial block then a length check
4571 can overflow resulting in a heap corruption.
4573 The amount of data needed is comparable to SIZE_MAX which is impractical
4576 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4581 * Malformed SHA512 ticket DoS
4583 If a server uses SHA512 for TLS session ticket HMAC it is vulnerable to a
4584 DoS attack where a malformed ticket will result in an OOB read which will
4587 The use of SHA512 in TLS session tickets is comparatively rare as it requires
4588 a custom server callback and ticket lookup mechanism.
4590 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4595 * OOB write in BN_bn2dec()
4597 The function BN_bn2dec() does not check the return value of BN_div_word().
4598 This can cause an OOB write if an application uses this function with an
4599 overly large BIGNUM. This could be a problem if an overly large certificate
4600 or CRL is printed out from an untrusted source. TLS is not affected because
4601 record limits will reject an oversized certificate before it is parsed.
4603 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4608 * OOB read in TS_OBJ_print_bio()
4610 The function TS_OBJ_print_bio() misuses OBJ_obj2txt(): the return value is
4611 the total length the OID text representation would use and not the amount
4612 of data written. This will result in OOB reads when large OIDs are
4615 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4620 * Pointer arithmetic undefined behaviour
4622 Avoid some undefined pointer arithmetic
4624 A common idiom in the codebase is to check limits in the following manner:
4627 Where "p" points to some malloc'd data of SIZE bytes and
4630 "len" here could be from some externally supplied data (e.g. from a TLS
4633 The rules of C pointer arithmetic are such that "p + len" is only well
4634 defined where len <= SIZE. Therefore the above idiom is actually
4635 undefined behaviour.
4637 For example this could cause problems if some malloc implementation
4638 provides an address for "p" such that "p + len" actually overflows for
4639 values of len that are too big and therefore p + len < limit.
4641 This issue was reported to OpenSSL by Guido Vranken
4646 * Constant time flag not preserved in DSA signing
4648 Operations in the DSA signing algorithm should run in constant time in
4649 order to avoid side channel attacks. A flaw in the OpenSSL DSA
4650 implementation means that a non-constant time codepath is followed for
4651 certain operations. This has been demonstrated through a cache-timing
4652 attack to be sufficient for an attacker to recover the private DSA key.
4654 This issue was reported by César Pereida (Aalto University), Billy Brumley
4655 (Tampere University of Technology), and Yuval Yarom (The University of
4656 Adelaide and NICTA).
4661 * DTLS buffered message DoS
4663 In a DTLS connection where handshake messages are delivered out-of-order
4664 those messages that OpenSSL is not yet ready to process will be buffered
4665 for later use. Under certain circumstances, a flaw in the logic means that
4666 those messages do not get removed from the buffer even though the handshake
4667 has been completed. An attacker could force up to approx. 15 messages to
4668 remain in the buffer when they are no longer required. These messages will
4669 be cleared when the DTLS connection is closed. The default maximum size for
4670 a message is 100k. Therefore the attacker could force an additional 1500k
4671 to be consumed per connection. By opening many simulataneous connections an
4672 attacker could cause a DoS attack through memory exhaustion.
4674 This issue was reported to OpenSSL by Quan Luo.
4679 * DTLS replay protection DoS
4681 A flaw in the DTLS replay attack protection mechanism means that records
4682 that arrive for future epochs update the replay protection "window" before
4683 the MAC for the record has been validated. This could be exploited by an
4684 attacker by sending a record for the next epoch (which does not have to
4685 decrypt or have a valid MAC), with a very large sequence number. This means
4686 that all subsequent legitimate packets are dropped causing a denial of
4687 service for a specific DTLS connection.
4689 This issue was reported to OpenSSL by the OCAP audit team.
4694 * Certificate message OOB reads
4696 In OpenSSL 1.0.2 and earlier some missing message length checks can result
4697 in OOB reads of up to 2 bytes beyond an allocated buffer. There is a
4698 theoretical DoS risk but this has not been observed in practice on common
4701 The messages affected are client certificate, client certificate request
4702 and server certificate. As a result the attack can only be performed
4703 against a client or a server which enables client authentication.
4705 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4710 ### Changes between 1.0.2g and 1.0.2h [3 May 2016]
4712 * Prevent padding oracle in AES-NI CBC MAC check
4714 A MITM attacker can use a padding oracle attack to decrypt traffic
4715 when the connection uses an AES CBC cipher and the server support
4718 This issue was introduced as part of the fix for Lucky 13 padding
4719 attack ([CVE-2013-0169]). The padding check was rewritten to be in
4720 constant time by making sure that always the same bytes are read and
4721 compared against either the MAC or padding bytes. But it no longer
4722 checked that there was enough data to have both the MAC and padding
4725 This issue was reported by Juraj Somorovsky using TLS-Attacker.
4729 * Fix EVP_EncodeUpdate overflow
4731 An overflow can occur in the EVP_EncodeUpdate() function which is used for
4732 Base64 encoding of binary data. If an attacker is able to supply very large
4733 amounts of input data then a length check can overflow resulting in a heap
4736 Internally to OpenSSL the EVP_EncodeUpdate() function is primarily used by
4737 the `PEM_write_bio*` family of functions. These are mainly used within the
4738 OpenSSL command line applications, so any application which processes data
4739 from an untrusted source and outputs it as a PEM file should be considered
4740 vulnerable to this issue. User applications that call these APIs directly
4741 with large amounts of untrusted data may also be vulnerable.
4743 This issue was reported by Guido Vranken.
4748 * Fix EVP_EncryptUpdate overflow
4750 An overflow can occur in the EVP_EncryptUpdate() function. If an attacker
4751 is able to supply very large amounts of input data after a previous call to
4752 EVP_EncryptUpdate() with a partial block then a length check can overflow
4753 resulting in a heap corruption. Following an analysis of all OpenSSL
4754 internal usage of the EVP_EncryptUpdate() function all usage is one of two
4755 forms. The first form is where the EVP_EncryptUpdate() call is known to be
4756 the first called function after an EVP_EncryptInit(), and therefore that
4757 specific call must be safe. The second form is where the length passed to
4758 EVP_EncryptUpdate() can be seen from the code to be some small value and
4759 therefore there is no possibility of an overflow. Since all instances are
4760 one of these two forms, it is believed that there can be no overflows in
4761 internal code due to this problem. It should be noted that
4762 EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths.
4763 Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances
4764 of these calls have also been analysed too and it is believed there are no
4765 instances in internal usage where an overflow could occur.
4767 This issue was reported by Guido Vranken.
4772 * Prevent ASN.1 BIO excessive memory allocation
4774 When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio()
4775 a short invalid encoding can cause allocation of large amounts of memory
4776 potentially consuming excessive resources or exhausting memory.
4778 Any application parsing untrusted data through d2i BIO functions is
4779 affected. The memory based functions such as d2i_X509() are *not* affected.
4780 Since the memory based functions are used by the TLS library, TLS
4781 applications are not affected.
4783 This issue was reported by Brian Carpenter.
4790 ASN1 Strings that are over 1024 bytes can cause an overread in applications
4791 using the X509_NAME_oneline() function on EBCDIC systems. This could result
4792 in arbitrary stack data being returned in the buffer.
4794 This issue was reported by Guido Vranken.
4799 * Modify behavior of ALPN to invoke callback after SNI/servername
4800 callback, such that updates to the SSL_CTX affect ALPN.
4804 * Remove LOW from the DEFAULT cipher list. This removes singles DES from the
4809 * Only remove the SSLv2 methods with the no-ssl2-method option. When the
4810 methods are enabled and ssl2 is disabled the methods return NULL.
4814 ### Changes between 1.0.2f and 1.0.2g [1 Mar 2016]
4816 * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL.
4817 Builds that are not configured with "enable-weak-ssl-ciphers" will not
4818 provide any "EXPORT" or "LOW" strength ciphers.
4822 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2
4823 is by default disabled at build-time. Builds that are not configured with
4824 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used,
4825 users who want to negotiate SSLv2 via the version-flexible SSLv23_method()
4826 will need to explicitly call either of:
4828 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2);
4830 SSL_clear_options(ssl, SSL_OP_NO_SSLv2);
4832 as appropriate. Even if either of those is used, or the application
4833 explicitly uses the version-specific SSLv2_method() or its client and
4834 server variants, SSLv2 ciphers vulnerable to exhaustive search key
4835 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT
4836 ciphers, and SSLv2 56-bit DES are no longer available.
4841 * Fix a double-free in DSA code
4843 A double free bug was discovered when OpenSSL parses malformed DSA private
4844 keys and could lead to a DoS attack or memory corruption for applications
4845 that receive DSA private keys from untrusted sources. This scenario is
4848 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using
4854 * Disable SRP fake user seed to address a server memory leak.
4856 Add a new method SRP_VBASE_get1_by_user that handles the seed properly.
4858 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
4859 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user
4860 was changed to ignore the "fake user" SRP seed, even if the seed
4863 Users should use SRP_VBASE_get1_by_user instead. Note that in
4864 SRP_VBASE_get1_by_user, caller must free the returned value. Note
4865 also that even though configuring the SRP seed attempts to hide
4866 invalid usernames by continuing the handshake with fake
4867 credentials, this behaviour is not constant time and no strong
4868 guarantees are made that the handshake is indistinguishable from
4869 that of a valid user.
4874 * Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption
4876 In the BN_hex2bn function the number of hex digits is calculated using an
4877 int value `i`. Later `bn_expand` is called with a value of `i * 4`. For
4878 large values of `i` this can result in `bn_expand` not allocating any
4879 memory because `i * 4` is negative. This can leave the internal BIGNUM data
4880 field as NULL leading to a subsequent NULL ptr deref. For very large values
4881 of `i`, the calculation `i * 4` could be a positive value smaller than `i`.
4882 In this case memory is allocated to the internal BIGNUM data field, but it
4883 is insufficiently sized leading to heap corruption. A similar issue exists
4884 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn
4885 is ever called by user applications with very large untrusted hex/dec data.
4886 This is anticipated to be a rare occurrence.
4888 All OpenSSL internal usage of these functions use data that is not expected
4889 to be untrusted, e.g. config file data or application command line
4890 arguments. If user developed applications generate config file data based
4891 on untrusted data then it is possible that this could also lead to security
4892 consequences. This is also anticipated to be rare.
4894 This issue was reported to OpenSSL by Guido Vranken.
4899 * Fix memory issues in `BIO_*printf` functions
4901 The internal `fmtstr` function used in processing a "%s" format string in
4902 the `BIO_*printf` functions could overflow while calculating the length of a
4903 string and cause an OOB read when printing very long strings.
4905 Additionally the internal `doapr_outch` function can attempt to write to an
4906 OOB memory location (at an offset from the NULL pointer) in the event of a
4907 memory allocation failure. In 1.0.2 and below this could be caused where
4908 the size of a buffer to be allocated is greater than INT_MAX. E.g. this
4909 could be in processing a very long "%s" format string. Memory leaks can
4912 The first issue may mask the second issue dependent on compiler behaviour.
4913 These problems could enable attacks where large amounts of untrusted data
4914 is passed to the `BIO_*printf` functions. If applications use these functions
4915 in this way then they could be vulnerable. OpenSSL itself uses these
4916 functions when printing out human-readable dumps of ASN.1 data. Therefore
4917 applications that print this data could be vulnerable if the data is from
4918 untrusted sources. OpenSSL command line applications could also be
4919 vulnerable where they print out ASN.1 data, or if untrusted data is passed
4920 as command line arguments.
4922 Libssl is not considered directly vulnerable. Additionally certificates etc
4923 received via remote connections via libssl are also unlikely to be able to
4924 trigger these issues because of message size limits enforced within libssl.
4926 This issue was reported to OpenSSL Guido Vranken.
4931 * Side channel attack on modular exponentiation
4933 A side-channel attack was found which makes use of cache-bank conflicts on
4934 the Intel Sandy-Bridge microarchitecture which could lead to the recovery
4935 of RSA keys. The ability to exploit this issue is limited as it relies on
4936 an attacker who has control of code in a thread running on the same
4937 hyper-threaded core as the victim thread which is performing decryptions.
4939 This issue was reported to OpenSSL by Yuval Yarom, The University of
4940 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and
4941 Nadia Heninger, University of Pennsylvania with more information at
4942 <http://cachebleed.info>.
4947 * Change the `req` command to generate a 2048-bit RSA/DSA key by default,
4948 if no keysize is specified with default_bits. This fixes an
4949 omission in an earlier change that changed all RSA/DSA key generation
4950 commands to use 2048 bits by default.
4954 ### Changes between 1.0.2e and 1.0.2f [28 Jan 2016]
4956 * DH small subgroups
4958 Historically OpenSSL only ever generated DH parameters based on "safe"
4959 primes. More recently (in version 1.0.2) support was provided for
4960 generating X9.42 style parameter files such as those required for RFC 5114
4961 support. The primes used in such files may not be "safe". Where an
4962 application is using DH configured with parameters based on primes that are
4963 not "safe" then an attacker could use this fact to find a peer's private
4964 DH exponent. This attack requires that the attacker complete multiple
4965 handshakes in which the peer uses the same private DH exponent. For example
4966 this could be used to discover a TLS server's private DH exponent if it's
4967 reusing the private DH exponent or it's using a static DH ciphersuite.
4969 OpenSSL provides the option SSL_OP_SINGLE_DH_USE for ephemeral DH (DHE) in
4970 TLS. It is not on by default. If the option is not set then the server
4971 reuses the same private DH exponent for the life of the server process and
4972 would be vulnerable to this attack. It is believed that many popular
4973 applications do set this option and would therefore not be at risk.
4975 The fix for this issue adds an additional check where a "q" parameter is
4976 available (as is the case in X9.42 based parameters). This detects the
4977 only known attack, and is the only possible defense for static DH
4978 ciphersuites. This could have some performance impact.
4980 Additionally the SSL_OP_SINGLE_DH_USE option has been switched on by
4981 default and cannot be disabled. This could have some performance impact.
4983 This issue was reported to OpenSSL by Antonio Sanso (Adobe).
4988 * SSLv2 doesn't block disabled ciphers
4990 A malicious client can negotiate SSLv2 ciphers that have been disabled on
4991 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
4992 been disabled, provided that the SSLv2 protocol was not also disabled via
4995 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
4996 and Sebastian Schinzel.
5001 ### Changes between 1.0.2d and 1.0.2e [3 Dec 2015]
5003 * BN_mod_exp may produce incorrect results on x86_64
5005 There is a carry propagating bug in the x86_64 Montgomery squaring
5006 procedure. No EC algorithms are affected. Analysis suggests that attacks
5007 against RSA and DSA as a result of this defect would be very difficult to
5008 perform and are not believed likely. Attacks against DH are considered just
5009 feasible (although very difficult) because most of the work necessary to
5010 deduce information about a private key may be performed offline. The amount
5011 of resources required for such an attack would be very significant and
5012 likely only accessible to a limited number of attackers. An attacker would
5013 additionally need online access to an unpatched system using the target
5014 private key in a scenario with persistent DH parameters and a private
5015 key that is shared between multiple clients. For example this can occur by
5016 default in OpenSSL DHE based SSL/TLS ciphersuites.
5018 This issue was reported to OpenSSL by Hanno Böck.
5023 * Certificate verify crash with missing PSS parameter
5025 The signature verification routines will crash with a NULL pointer
5026 dereference if presented with an ASN.1 signature using the RSA PSS
5027 algorithm and absent mask generation function parameter. Since these
5028 routines are used to verify certificate signature algorithms this can be
5029 used to crash any certificate verification operation and exploited in a
5030 DoS attack. Any application which performs certificate verification is
5031 vulnerable including OpenSSL clients and servers which enable client
5034 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
5039 * X509_ATTRIBUTE memory leak
5041 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
5042 memory. This structure is used by the PKCS#7 and CMS routines so any
5043 application which reads PKCS#7 or CMS data from untrusted sources is
5044 affected. SSL/TLS is not affected.
5046 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
5052 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
5053 This changes the decoding behaviour for some invalid messages,
5054 though the change is mostly in the more lenient direction, and
5055 legacy behaviour is preserved as much as possible.
5059 * In DSA_generate_parameters_ex, if the provided seed is too short,
5062 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
5064 ### Changes between 1.0.2c and 1.0.2d [9 Jul 2015]
5066 * Alternate chains certificate forgery
5068 During certificate verification, OpenSSL will attempt to find an
5069 alternative certificate chain if the first attempt to build such a chain
5070 fails. An error in the implementation of this logic can mean that an
5071 attacker could cause certain checks on untrusted certificates to be
5072 bypassed, such as the CA flag, enabling them to use a valid leaf
5073 certificate to act as a CA and "issue" an invalid certificate.
5075 This issue was reported to OpenSSL by Adam Langley/David Benjamin
5080 ### Changes between 1.0.2b and 1.0.2c [12 Jun 2015]
5082 * Fix HMAC ABI incompatibility. The previous version introduced an ABI
5083 incompatibility in the handling of HMAC. The previous ABI has now been
5088 ### Changes between 1.0.2a and 1.0.2b [11 Jun 2015]
5090 * Malformed ECParameters causes infinite loop
5092 When processing an ECParameters structure OpenSSL enters an infinite loop
5093 if the curve specified is over a specially malformed binary polynomial
5096 This can be used to perform denial of service against any
5097 system which processes public keys, certificate requests or
5098 certificates. This includes TLS clients and TLS servers with
5099 client authentication enabled.
5101 This issue was reported to OpenSSL by Joseph Barr-Pixton.
5106 * Exploitable out-of-bounds read in X509_cmp_time
5108 X509_cmp_time does not properly check the length of the ASN1_TIME
5109 string and can read a few bytes out of bounds. In addition,
5110 X509_cmp_time accepts an arbitrary number of fractional seconds in the
5113 An attacker can use this to craft malformed certificates and CRLs of
5114 various sizes and potentially cause a segmentation fault, resulting in
5115 a DoS on applications that verify certificates or CRLs. TLS clients
5116 that verify CRLs are affected. TLS clients and servers with client
5117 authentication enabled may be affected if they use custom verification
5120 This issue was reported to OpenSSL by Robert Swiecki (Google), and
5121 independently by Hanno Böck.
5126 * PKCS7 crash with missing EnvelopedContent
5128 The PKCS#7 parsing code does not handle missing inner EncryptedContent
5129 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
5130 with missing content and trigger a NULL pointer dereference on parsing.
5132 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
5133 structures from untrusted sources are affected. OpenSSL clients and
5134 servers are not affected.
5136 This issue was reported to OpenSSL by Michal Zalewski (Google).
5141 * CMS verify infinite loop with unknown hash function
5143 When verifying a signedData message the CMS code can enter an infinite loop
5144 if presented with an unknown hash function OID. This can be used to perform
5145 denial of service against any system which verifies signedData messages using
5147 This issue was reported to OpenSSL by Johannes Bauer.
5152 * Race condition handling NewSessionTicket
5154 If a NewSessionTicket is received by a multi-threaded client when attempting to
5155 reuse a previous ticket then a race condition can occur potentially leading to
5156 a double free of the ticket data.
5161 * Only support 256-bit or stronger elliptic curves with the
5162 'ecdh_auto' setting (server) or by default (client). Of supported
5163 curves, prefer P-256 (both).
5167 ### Changes between 1.0.2 and 1.0.2a [19 Mar 2015]
5169 * ClientHello sigalgs DoS fix
5171 If a client connects to an OpenSSL 1.0.2 server and renegotiates with an
5172 invalid signature algorithms extension a NULL pointer dereference will
5173 occur. This can be exploited in a DoS attack against the server.
5175 This issue was was reported to OpenSSL by David Ramos of Stanford
5179 *Stephen Henson and Matt Caswell*
5181 * Multiblock corrupted pointer fix
5183 OpenSSL 1.0.2 introduced the "multiblock" performance improvement. This
5184 feature only applies on 64 bit x86 architecture platforms that support AES
5185 NI instructions. A defect in the implementation of "multiblock" can cause
5186 OpenSSL's internal write buffer to become incorrectly set to NULL when
5187 using non-blocking IO. Typically, when the user application is using a
5188 socket BIO for writing, this will only result in a failed connection.
5189 However if some other BIO is used then it is likely that a segmentation
5190 fault will be triggered, thus enabling a potential DoS attack.
5192 This issue was reported to OpenSSL by Daniel Danner and Rainer Mueller.
5197 * Segmentation fault in DTLSv1_listen fix
5199 The DTLSv1_listen function is intended to be stateless and processes the
5200 initial ClientHello from many peers. It is common for user code to loop
5201 over the call to DTLSv1_listen until a valid ClientHello is received with
5202 an associated cookie. A defect in the implementation of DTLSv1_listen means
5203 that state is preserved in the SSL object from one invocation to the next
5204 that can lead to a segmentation fault. Errors processing the initial
5205 ClientHello can trigger this scenario. An example of such an error could be
5206 that a DTLS1.0 only client is attempting to connect to a DTLS1.2 only
5209 This issue was reported to OpenSSL by Per Allansson.
5214 * Segmentation fault in ASN1_TYPE_cmp fix
5216 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
5217 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
5218 certificate signature algorithm consistency this can be used to crash any
5219 certificate verification operation and exploited in a DoS attack. Any
5220 application which performs certificate verification is vulnerable including
5221 OpenSSL clients and servers which enable client authentication.
5226 * Segmentation fault for invalid PSS parameters fix
5228 The signature verification routines will crash with a NULL pointer
5229 dereference if presented with an ASN.1 signature using the RSA PSS
5230 algorithm and invalid parameters. Since these routines are used to verify
5231 certificate signature algorithms this can be used to crash any
5232 certificate verification operation and exploited in a DoS attack. Any
5233 application which performs certificate verification is vulnerable including
5234 OpenSSL clients and servers which enable client authentication.
5236 This issue was was reported to OpenSSL by Brian Carpenter.
5241 * ASN.1 structure reuse memory corruption fix
5243 Reusing a structure in ASN.1 parsing may allow an attacker to cause
5244 memory corruption via an invalid write. Such reuse is and has been
5245 strongly discouraged and is believed to be rare.
5247 Applications that parse structures containing CHOICE or ANY DEFINED BY
5248 components may be affected. Certificate parsing (d2i_X509 and related
5249 functions) are however not affected. OpenSSL clients and servers are
5255 * PKCS7 NULL pointer dereferences fix
5257 The PKCS#7 parsing code does not handle missing outer ContentInfo
5258 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
5259 missing content and trigger a NULL pointer dereference on parsing.
5261 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
5262 otherwise parse PKCS#7 structures from untrusted sources are
5263 affected. OpenSSL clients and servers are not affected.
5265 This issue was reported to OpenSSL by Michal Zalewski (Google).
5270 * DoS via reachable assert in SSLv2 servers fix
5272 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
5273 servers that both support SSLv2 and enable export cipher suites by sending
5274 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
5276 This issue was discovered by Sean Burford (Google) and Emilia Käsper
5277 (OpenSSL development team).
5282 * Empty CKE with client auth and DHE fix
5284 If client auth is used then a server can seg fault in the event of a DHE
5285 ciphersuite being selected and a zero length ClientKeyExchange message
5286 being sent by the client. This could be exploited in a DoS attack.
5291 * Handshake with unseeded PRNG fix
5293 Under certain conditions an OpenSSL 1.0.2 client can complete a handshake
5294 with an unseeded PRNG. The conditions are:
5295 - The client is on a platform where the PRNG has not been seeded
5296 automatically, and the user has not seeded manually
5297 - A protocol specific client method version has been used (i.e. not
5298 SSL_client_methodv23)
5299 - A ciphersuite is used that does not require additional random data from
5300 the PRNG beyond the initial ClientHello client random (e.g. PSK-RC4-SHA).
5302 If the handshake succeeds then the client random that has been used will
5303 have been generated from a PRNG with insufficient entropy and therefore the
5304 output may be predictable.
5306 For example using the following command with an unseeded openssl will
5307 succeed on an unpatched platform:
5309 openssl s_client -psk 1a2b3c4d -tls1_2 -cipher PSK-RC4-SHA
5314 * Use After Free following d2i_ECPrivatekey error fix
5316 A malformed EC private key file consumed via the d2i_ECPrivateKey function
5317 could cause a use after free condition. This, in turn, could cause a double
5318 free in several private key parsing functions (such as d2i_PrivateKey
5319 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
5320 for applications that receive EC private keys from untrusted
5321 sources. This scenario is considered rare.
5323 This issue was discovered by the BoringSSL project and fixed in their
5329 * X509_to_X509_REQ NULL pointer deref fix
5331 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
5332 the certificate key is invalid. This function is rarely used in practice.
5334 This issue was discovered by Brian Carpenter.
5339 * Removed the export ciphers from the DEFAULT ciphers
5343 ### Changes between 1.0.1l and 1.0.2 [22 Jan 2015]
5345 * Facilitate "universal" ARM builds targeting range of ARM ISAs, e.g.
5346 ARMv5 through ARMv8, as opposite to "locking" it to single one.
5347 So far those who have to target multiple platforms would compromise
5348 and argue that binary targeting say ARMv5 would still execute on
5349 ARMv8. "Universal" build resolves this compromise by providing
5350 near-optimal performance even on newer platforms.
5354 * Accelerated NIST P-256 elliptic curve implementation for x86_64
5355 (other platforms pending).
5357 *Shay Gueron & Vlad Krasnov (Intel Corp), Andy Polyakov*
5359 * Add support for the SignedCertificateTimestampList certificate and
5360 OCSP response extensions from RFC6962.
5364 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
5365 for corner cases. (Certain input points at infinity could lead to
5366 bogus results, with non-infinity inputs mapped to infinity too.)
5370 * Initial support for PowerISA 2.0.7, first implemented in POWER8.
5371 This covers AES, SHA256/512 and GHASH. "Initial" means that most
5372 common cases are optimized and there still is room for further
5373 improvements. Vector Permutation AES for Altivec is also added.
5377 * Add support for little-endian ppc64 Linux target.
5379 *Marcelo Cerri (IBM)*
5381 * Initial support for AMRv8 ISA crypto extensions. This covers AES,
5382 SHA1, SHA256 and GHASH. "Initial" means that most common cases
5383 are optimized and there still is room for further improvements.
5384 Both 32- and 64-bit modes are supported.
5386 *Andy Polyakov, Ard Biesheuvel (Linaro)*
5388 * Improved ARMv7 NEON support.
5392 * Support for SPARC Architecture 2011 crypto extensions, first
5393 implemented in SPARC T4. This covers AES, DES, Camellia, SHA1,
5394 SHA256/512, MD5, GHASH and modular exponentiation.
5396 *Andy Polyakov, David Miller*
5398 * Accelerated modular exponentiation for Intel processors, a.k.a.
5401 *Shay Gueron & Vlad Krasnov (Intel Corp)*
5403 * Support for new and upcoming Intel processors, including AVX2,
5404 BMI and SHA ISA extensions. This includes additional "stitched"
5405 implementations, AESNI-SHA256 and GCM, and multi-buffer support
5408 This work was sponsored by Intel Corp.
5412 * Support for DTLS 1.2. This adds two sets of DTLS methods: DTLS_*_method()
5413 supports both DTLS 1.2 and 1.0 and should use whatever version the peer
5414 supports and DTLSv1_2_*_method() which supports DTLS 1.2 only.
5418 * Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
5419 this fixes a limitation in previous versions of OpenSSL.
5423 * Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
5424 MGF1 digest and OAEP label.
5428 * Add EVP support for key wrapping algorithms, to avoid problems with
5429 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
5430 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
5431 algorithms and include tests cases.
5435 * Add functions to allocate and set the fields of an ECDSA_METHOD
5438 *Douglas E. Engert, Steve Henson*
5440 * New functions OPENSSL_gmtime_diff and ASN1_TIME_diff to find the
5441 difference in days and seconds between two tm or ASN1_TIME structures.
5445 * Add -rev test option to s_server to just reverse order of characters
5446 received by client and send back to server. Also prints an abbreviated
5447 summary of the connection parameters.
5451 * New option -brief for s_client and s_server to print out a brief summary
5452 of connection parameters.
5456 * Add callbacks for arbitrary TLS extensions.
5458 *Trevor Perrin <trevp@trevp.net> and Ben Laurie*
5460 * New option -crl_download in several openssl utilities to download CRLs
5461 from CRLDP extension in certificates.
5465 * New options -CRL and -CRLform for s_client and s_server for CRLs.
5469 * New function X509_CRL_diff to generate a delta CRL from the difference
5470 of two full CRLs. Add support to "crl" utility.
5474 * New functions to set lookup_crls function and to retrieve
5475 X509_STORE from X509_STORE_CTX.
5479 * Print out deprecated issuer and subject unique ID fields in
5484 * Extend OCSP I/O functions so they can be used for simple general purpose
5485 HTTP as well as OCSP. New wrapper function which can be used to download
5486 CRLs using the OCSP API.
5490 * Delegate command line handling in s_client/s_server to SSL_CONF APIs.
5494 * `SSL_CONF*` functions. These provide a common framework for application
5495 configuration using configuration files or command lines.
5499 * SSL/TLS tracing code. This parses out SSL/TLS records using the
5500 message callback and prints the results. Needs compile time option
5501 "enable-ssl-trace". New options to s_client and s_server to enable
5506 * New ctrl and macro to retrieve supported points extensions.
5507 Print out extension in s_server and s_client.
5511 * New functions to retrieve certificate signature and signature
5516 * Add functions to retrieve and manipulate the raw cipherlist sent by a
5521 * New Suite B modes for TLS code. These use and enforce the requirements
5522 of RFC6460: restrict ciphersuites, only permit Suite B algorithms and
5523 only use Suite B curves. The Suite B modes can be set by using the
5524 strings "SUITEB128", "SUITEB192" or "SUITEB128ONLY" for the cipherstring.
5528 * New chain verification flags for Suite B levels of security. Check
5529 algorithms are acceptable when flags are set in X509_verify_cert.
5533 * Make tls1_check_chain return a set of flags indicating checks passed
5534 by a certificate chain. Add additional tests to handle client
5535 certificates: checks for matching certificate type and issuer name
5540 * If an attempt is made to use a signature algorithm not in the peer
5541 preference list abort the handshake. If client has no suitable
5542 signature algorithms in response to a certificate request do not
5543 use the certificate.
5547 * If server EC tmp key is not in client preference list abort handshake.
5551 * Add support for certificate stores in CERT structure. This makes it
5552 possible to have different stores per SSL structure or one store in
5553 the parent SSL_CTX. Include distinct stores for certificate chain
5554 verification and chain building. New ctrl SSL_CTRL_BUILD_CERT_CHAIN
5555 to build and store a certificate chain in CERT structure: returning
5556 an error if the chain cannot be built: this will allow applications
5557 to test if a chain is correctly configured.
5559 Note: if the CERT based stores are not set then the parent SSL_CTX
5560 store is used to retain compatibility with existing behaviour.
5564 * New function ssl_set_client_disabled to set a ciphersuite disabled
5565 mask based on the current session, check mask when sending client
5566 hello and checking the requested ciphersuite.
5570 * New ctrls to retrieve and set certificate types in a certificate
5571 request message. Print out received values in s_client. If certificate
5572 types is not set with custom values set sensible values based on
5573 supported signature algorithms.
5577 * Support for distinct client and server supported signature algorithms.
5581 * Add certificate callback. If set this is called whenever a certificate
5582 is required by client or server. An application can decide which
5583 certificate chain to present based on arbitrary criteria: for example
5584 supported signature algorithms. Add very simple example to s_server.
5585 This fixes many of the problems and restrictions of the existing client
5586 certificate callback: for example you can now clear an existing
5587 certificate and specify the whole chain.
5591 * Add new "valid_flags" field to CERT_PKEY structure which determines what
5592 the certificate can be used for (if anything). Set valid_flags field
5593 in new tls1_check_chain function. Simplify ssl_set_cert_masks which used
5594 to have similar checks in it.
5596 Add new "cert_flags" field to CERT structure and include a "strict mode".
5597 This enforces some TLS certificate requirements (such as only permitting
5598 certificate signature algorithms contained in the supported algorithms
5599 extension) which some implementations ignore: this option should be used
5600 with caution as it could cause interoperability issues.
5604 * Update and tidy signature algorithm extension processing. Work out
5605 shared signature algorithms based on preferences and peer algorithms
5606 and print them out in s_client and s_server. Abort handshake if no
5607 shared signature algorithms.
5611 * Add new functions to allow customised supported signature algorithms
5612 for SSL and SSL_CTX structures. Add options to s_client and s_server
5617 * New function SSL_certs_clear() to delete all references to certificates
5618 from an SSL structure. Before this once a certificate had been added
5619 it couldn't be removed.
5623 * Integrate hostname, email address and IP address checking with certificate
5624 verification. New verify options supporting checking in openssl utility.
5628 * Fixes and wildcard matching support to hostname and email checking
5629 functions. Add manual page.
5631 *Florian Weimer (Red Hat Product Security Team)*
5633 * New functions to check a hostname email or IP address against a
5634 certificate. Add options x509 utility to print results of checks against
5639 * Fix OCSP checking.
5641 *Rob Stradling <rob.stradling@comodo.com> and Ben Laurie*
5643 * Initial experimental support for explicitly trusted non-root CAs.
5644 OpenSSL still tries to build a complete chain to a root but if an
5645 intermediate CA has a trust setting included that is used. The first
5646 setting is used: whether to trust (e.g., -addtrust option to the x509
5651 * Add -trusted_first option which attempts to find certificates in the
5652 trusted store even if an untrusted chain is also supplied.
5656 * MIPS assembly pack updates: support for MIPS32r2 and SmartMIPS ASE,
5657 platform support for Linux and Android.
5661 * Support for linux-x32, ILP32 environment in x86_64 framework.
5665 * Experimental multi-implementation support for FIPS capable OpenSSL.
5666 When in FIPS mode the approved implementations are used as normal,
5667 when not in FIPS mode the internal unapproved versions are used instead.
5668 This means that the FIPS capable OpenSSL isn't forced to use the
5669 (often lower performance) FIPS implementations outside FIPS mode.
5673 * Transparently support X9.42 DH parameters when calling
5674 PEM_read_bio_DHparameters. This means existing applications can handle
5675 the new parameter format automatically.
5679 * Initial experimental support for X9.42 DH parameter format: mainly
5680 to support use of 'q' parameter for RFC5114 parameters.
5684 * Add DH parameters from RFC5114 including test data to dhtest.
5688 * Support for automatic EC temporary key parameter selection. If enabled
5689 the most preferred EC parameters are automatically used instead of
5690 hardcoded fixed parameters. Now a server just has to call:
5691 SSL_CTX_set_ecdh_auto(ctx, 1) and the server will automatically
5692 support ECDH and use the most appropriate parameters.
5696 * Enhance and tidy EC curve and point format TLS extension code. Use
5697 static structures instead of allocation if default values are used.
5698 New ctrls to set curves we wish to support and to retrieve shared curves.
5699 Print out shared curves in s_server. New options to s_server and s_client
5700 to set list of supported curves.
5704 * New ctrls to retrieve supported signature algorithms and
5705 supported curve values as an array of NIDs. Extend openssl utility
5706 to print out received values.
5710 * Add new APIs EC_curve_nist2nid and EC_curve_nid2nist which convert
5711 between NIDs and the more common NIST names such as "P-256". Enhance
5712 ecparam utility and ECC method to recognise the NIST names for curves.
5716 * Enhance SSL/TLS certificate chain handling to support different
5717 chains for each certificate instead of one chain in the parent SSL_CTX.
5721 * Support for fixed DH ciphersuite client authentication: where both
5722 server and client use DH certificates with common parameters.
5726 * Support for fixed DH ciphersuites: those requiring DH server
5731 * New function i2d_re_X509_tbs for re-encoding the TBS portion of
5733 Note: Related 1.0.2-beta specific macros X509_get_cert_info,
5734 X509_CINF_set_modified, X509_CINF_get_issuer, X509_CINF_get_extensions and
5735 X509_CINF_get_signature were reverted post internal team review.
5740 ### Changes between 1.0.1t and 1.0.1u [22 Sep 2016]
5742 * OCSP Status Request extension unbounded memory growth
5744 A malicious client can send an excessively large OCSP Status Request
5745 extension. If that client continually requests renegotiation, sending a
5746 large OCSP Status Request extension each time, then there will be unbounded
5747 memory growth on the server. This will eventually lead to a Denial Of
5748 Service attack through memory exhaustion. Servers with a default
5749 configuration are vulnerable even if they do not support OCSP. Builds using
5750 the "no-ocsp" build time option are not affected.
5752 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5757 * In order to mitigate the SWEET32 attack, the DES ciphers were moved from
5760 This issue was reported to OpenSSL Karthikeyan Bhargavan and Gaetan
5766 * OOB write in MDC2_Update()
5768 An overflow can occur in MDC2_Update() either if called directly or
5769 through the EVP_DigestUpdate() function using MDC2. If an attacker
5770 is able to supply very large amounts of input data after a previous
5771 call to EVP_EncryptUpdate() with a partial block then a length check
5772 can overflow resulting in a heap corruption.
5774 The amount of data needed is comparable to SIZE_MAX which is impractical
5777 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5782 * Malformed SHA512 ticket DoS
5784 If a server uses SHA512 for TLS session ticket HMAC it is vulnerable to a
5785 DoS attack where a malformed ticket will result in an OOB read which will
5788 The use of SHA512 in TLS session tickets is comparatively rare as it requires
5789 a custom server callback and ticket lookup mechanism.
5791 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5796 * OOB write in BN_bn2dec()
5798 The function BN_bn2dec() does not check the return value of BN_div_word().
5799 This can cause an OOB write if an application uses this function with an
5800 overly large BIGNUM. This could be a problem if an overly large certificate
5801 or CRL is printed out from an untrusted source. TLS is not affected because
5802 record limits will reject an oversized certificate before it is parsed.
5804 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5809 * OOB read in TS_OBJ_print_bio()
5811 The function TS_OBJ_print_bio() misuses OBJ_obj2txt(): the return value is
5812 the total length the OID text representation would use and not the amount
5813 of data written. This will result in OOB reads when large OIDs are
5816 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5821 * Pointer arithmetic undefined behaviour
5823 Avoid some undefined pointer arithmetic
5825 A common idiom in the codebase is to check limits in the following manner:
5828 Where "p" points to some malloc'd data of SIZE bytes and
5831 "len" here could be from some externally supplied data (e.g. from a TLS
5834 The rules of C pointer arithmetic are such that "p + len" is only well
5835 defined where len <= SIZE. Therefore the above idiom is actually
5836 undefined behaviour.
5838 For example this could cause problems if some malloc implementation
5839 provides an address for "p" such that "p + len" actually overflows for
5840 values of len that are too big and therefore p + len < limit.
5842 This issue was reported to OpenSSL by Guido Vranken
5847 * Constant time flag not preserved in DSA signing
5849 Operations in the DSA signing algorithm should run in constant time in
5850 order to avoid side channel attacks. A flaw in the OpenSSL DSA
5851 implementation means that a non-constant time codepath is followed for
5852 certain operations. This has been demonstrated through a cache-timing
5853 attack to be sufficient for an attacker to recover the private DSA key.
5855 This issue was reported by César Pereida (Aalto University), Billy Brumley
5856 (Tampere University of Technology), and Yuval Yarom (The University of
5857 Adelaide and NICTA).
5862 * DTLS buffered message DoS
5864 In a DTLS connection where handshake messages are delivered out-of-order
5865 those messages that OpenSSL is not yet ready to process will be buffered
5866 for later use. Under certain circumstances, a flaw in the logic means that
5867 those messages do not get removed from the buffer even though the handshake
5868 has been completed. An attacker could force up to approx. 15 messages to
5869 remain in the buffer when they are no longer required. These messages will
5870 be cleared when the DTLS connection is closed. The default maximum size for
5871 a message is 100k. Therefore the attacker could force an additional 1500k
5872 to be consumed per connection. By opening many simulataneous connections an
5873 attacker could cause a DoS attack through memory exhaustion.
5875 This issue was reported to OpenSSL by Quan Luo.
5880 * DTLS replay protection DoS
5882 A flaw in the DTLS replay attack protection mechanism means that records
5883 that arrive for future epochs update the replay protection "window" before
5884 the MAC for the record has been validated. This could be exploited by an
5885 attacker by sending a record for the next epoch (which does not have to
5886 decrypt or have a valid MAC), with a very large sequence number. This means
5887 that all subsequent legitimate packets are dropped causing a denial of
5888 service for a specific DTLS connection.
5890 This issue was reported to OpenSSL by the OCAP audit team.
5895 * Certificate message OOB reads
5897 In OpenSSL 1.0.2 and earlier some missing message length checks can result
5898 in OOB reads of up to 2 bytes beyond an allocated buffer. There is a
5899 theoretical DoS risk but this has not been observed in practice on common
5902 The messages affected are client certificate, client certificate request
5903 and server certificate. As a result the attack can only be performed
5904 against a client or a server which enables client authentication.
5906 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5911 ### Changes between 1.0.1s and 1.0.1t [3 May 2016]
5913 * Prevent padding oracle in AES-NI CBC MAC check
5915 A MITM attacker can use a padding oracle attack to decrypt traffic
5916 when the connection uses an AES CBC cipher and the server support
5919 This issue was introduced as part of the fix for Lucky 13 padding
5920 attack ([CVE-2013-0169]). The padding check was rewritten to be in
5921 constant time by making sure that always the same bytes are read and
5922 compared against either the MAC or padding bytes. But it no longer
5923 checked that there was enough data to have both the MAC and padding
5926 This issue was reported by Juraj Somorovsky using TLS-Attacker.
5931 * Fix EVP_EncodeUpdate overflow
5933 An overflow can occur in the EVP_EncodeUpdate() function which is used for
5934 Base64 encoding of binary data. If an attacker is able to supply very large
5935 amounts of input data then a length check can overflow resulting in a heap
5938 Internally to OpenSSL the EVP_EncodeUpdate() function is primarly used by
5939 the `PEM_write_bio*` family of functions. These are mainly used within the
5940 OpenSSL command line applications, so any application which processes data
5941 from an untrusted source and outputs it as a PEM file should be considered
5942 vulnerable to this issue. User applications that call these APIs directly
5943 with large amounts of untrusted data may also be vulnerable.
5945 This issue was reported by Guido Vranken.
5950 * Fix EVP_EncryptUpdate overflow
5952 An overflow can occur in the EVP_EncryptUpdate() function. If an attacker
5953 is able to supply very large amounts of input data after a previous call to
5954 EVP_EncryptUpdate() with a partial block then a length check can overflow
5955 resulting in a heap corruption. Following an analysis of all OpenSSL
5956 internal usage of the EVP_EncryptUpdate() function all usage is one of two
5957 forms. The first form is where the EVP_EncryptUpdate() call is known to be
5958 the first called function after an EVP_EncryptInit(), and therefore that
5959 specific call must be safe. The second form is where the length passed to
5960 EVP_EncryptUpdate() can be seen from the code to be some small value and
5961 therefore there is no possibility of an overflow. Since all instances are
5962 one of these two forms, it is believed that there can be no overflows in
5963 internal code due to this problem. It should be noted that
5964 EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths.
5965 Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances
5966 of these calls have also been analysed too and it is believed there are no
5967 instances in internal usage where an overflow could occur.
5969 This issue was reported by Guido Vranken.
5974 * Prevent ASN.1 BIO excessive memory allocation
5976 When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio()
5977 a short invalid encoding can casuse allocation of large amounts of memory
5978 potentially consuming excessive resources or exhausting memory.
5980 Any application parsing untrusted data through d2i BIO functions is
5981 affected. The memory based functions such as d2i_X509() are *not* affected.
5982 Since the memory based functions are used by the TLS library, TLS
5983 applications are not affected.
5985 This issue was reported by Brian Carpenter.
5992 ASN1 Strings that are over 1024 bytes can cause an overread in applications
5993 using the X509_NAME_oneline() function on EBCDIC systems. This could result
5994 in arbitrary stack data being returned in the buffer.
5996 This issue was reported by Guido Vranken.
6001 * Modify behavior of ALPN to invoke callback after SNI/servername
6002 callback, such that updates to the SSL_CTX affect ALPN.
6006 * Remove LOW from the DEFAULT cipher list. This removes singles DES from the
6011 * Only remove the SSLv2 methods with the no-ssl2-method option. When the
6012 methods are enabled and ssl2 is disabled the methods return NULL.
6016 ### Changes between 1.0.1r and 1.0.1s [1 Mar 2016]
6018 * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL.
6019 Builds that are not configured with "enable-weak-ssl-ciphers" will not
6020 provide any "EXPORT" or "LOW" strength ciphers.
6024 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2
6025 is by default disabled at build-time. Builds that are not configured with
6026 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used,
6027 users who want to negotiate SSLv2 via the version-flexible SSLv23_method()
6028 will need to explicitly call either of:
6030 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2);
6032 SSL_clear_options(ssl, SSL_OP_NO_SSLv2);
6034 as appropriate. Even if either of those is used, or the application
6035 explicitly uses the version-specific SSLv2_method() or its client and
6036 server variants, SSLv2 ciphers vulnerable to exhaustive search key
6037 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT
6038 ciphers, and SSLv2 56-bit DES are no longer available.
6043 * Fix a double-free in DSA code
6045 A double free bug was discovered when OpenSSL parses malformed DSA private
6046 keys and could lead to a DoS attack or memory corruption for applications
6047 that receive DSA private keys from untrusted sources. This scenario is
6050 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using
6056 * Disable SRP fake user seed to address a server memory leak.
6058 Add a new method SRP_VBASE_get1_by_user that handles the seed properly.
6060 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
6061 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user
6062 was changed to ignore the "fake user" SRP seed, even if the seed
6065 Users should use SRP_VBASE_get1_by_user instead. Note that in
6066 SRP_VBASE_get1_by_user, caller must free the returned value. Note
6067 also that even though configuring the SRP seed attempts to hide
6068 invalid usernames by continuing the handshake with fake
6069 credentials, this behaviour is not constant time and no strong
6070 guarantees are made that the handshake is indistinguishable from
6071 that of a valid user.
6076 * Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption
6078 In the BN_hex2bn function the number of hex digits is calculated using an
6079 int value `i`. Later `bn_expand` is called with a value of `i * 4`. For
6080 large values of `i` this can result in `bn_expand` not allocating any
6081 memory because `i * 4` is negative. This can leave the internal BIGNUM data
6082 field as NULL leading to a subsequent NULL ptr deref. For very large values
6083 of `i`, the calculation `i * 4` could be a positive value smaller than `i`.
6084 In this case memory is allocated to the internal BIGNUM data field, but it
6085 is insufficiently sized leading to heap corruption. A similar issue exists
6086 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn
6087 is ever called by user applications with very large untrusted hex/dec data.
6088 This is anticipated to be a rare occurrence.
6090 All OpenSSL internal usage of these functions use data that is not expected
6091 to be untrusted, e.g. config file data or application command line
6092 arguments. If user developed applications generate config file data based
6093 on untrusted data then it is possible that this could also lead to security
6094 consequences. This is also anticipated to be rare.
6096 This issue was reported to OpenSSL by Guido Vranken.
6101 * Fix memory issues in `BIO_*printf` functions
6103 The internal `fmtstr` function used in processing a "%s" format string in
6104 the `BIO_*printf` functions could overflow while calculating the length of a
6105 string and cause an OOB read when printing very long strings.
6107 Additionally the internal `doapr_outch` function can attempt to write to an
6108 OOB memory location (at an offset from the NULL pointer) in the event of a
6109 memory allocation failure. In 1.0.2 and below this could be caused where
6110 the size of a buffer to be allocated is greater than INT_MAX. E.g. this
6111 could be in processing a very long "%s" format string. Memory leaks can
6114 The first issue may mask the second issue dependent on compiler behaviour.
6115 These problems could enable attacks where large amounts of untrusted data
6116 is passed to the `BIO_*printf` functions. If applications use these functions
6117 in this way then they could be vulnerable. OpenSSL itself uses these
6118 functions when printing out human-readable dumps of ASN.1 data. Therefore
6119 applications that print this data could be vulnerable if the data is from
6120 untrusted sources. OpenSSL command line applications could also be
6121 vulnerable where they print out ASN.1 data, or if untrusted data is passed
6122 as command line arguments.
6124 Libssl is not considered directly vulnerable. Additionally certificates etc
6125 received via remote connections via libssl are also unlikely to be able to
6126 trigger these issues because of message size limits enforced within libssl.
6128 This issue was reported to OpenSSL Guido Vranken.
6133 * Side channel attack on modular exponentiation
6135 A side-channel attack was found which makes use of cache-bank conflicts on
6136 the Intel Sandy-Bridge microarchitecture which could lead to the recovery
6137 of RSA keys. The ability to exploit this issue is limited as it relies on
6138 an attacker who has control of code in a thread running on the same
6139 hyper-threaded core as the victim thread which is performing decryptions.
6141 This issue was reported to OpenSSL by Yuval Yarom, The University of
6142 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and
6143 Nadia Heninger, University of Pennsylvania with more information at
6144 <http://cachebleed.info>.
6149 * Change the req command to generate a 2048-bit RSA/DSA key by default,
6150 if no keysize is specified with default_bits. This fixes an
6151 omission in an earlier change that changed all RSA/DSA key generation
6152 commands to use 2048 bits by default.
6156 ### Changes between 1.0.1q and 1.0.1r [28 Jan 2016]
6158 * Protection for DH small subgroup attacks
6160 As a precautionary measure the SSL_OP_SINGLE_DH_USE option has been
6161 switched on by default and cannot be disabled. This could have some
6166 * SSLv2 doesn't block disabled ciphers
6168 A malicious client can negotiate SSLv2 ciphers that have been disabled on
6169 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
6170 been disabled, provided that the SSLv2 protocol was not also disabled via
6173 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
6174 and Sebastian Schinzel.
6179 * Reject DH handshakes with parameters shorter than 1024 bits.
6183 ### Changes between 1.0.1p and 1.0.1q [3 Dec 2015]
6185 * Certificate verify crash with missing PSS parameter
6187 The signature verification routines will crash with a NULL pointer
6188 dereference if presented with an ASN.1 signature using the RSA PSS
6189 algorithm and absent mask generation function parameter. Since these
6190 routines are used to verify certificate signature algorithms this can be
6191 used to crash any certificate verification operation and exploited in a
6192 DoS attack. Any application which performs certificate verification is
6193 vulnerable including OpenSSL clients and servers which enable client
6196 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
6201 * X509_ATTRIBUTE memory leak
6203 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
6204 memory. This structure is used by the PKCS#7 and CMS routines so any
6205 application which reads PKCS#7 or CMS data from untrusted sources is
6206 affected. SSL/TLS is not affected.
6208 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
6214 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
6215 This changes the decoding behaviour for some invalid messages,
6216 though the change is mostly in the more lenient direction, and
6217 legacy behaviour is preserved as much as possible.
6221 * In DSA_generate_parameters_ex, if the provided seed is too short,
6222 use a random seed, as already documented.
6224 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
6226 ### Changes between 1.0.1o and 1.0.1p [9 Jul 2015]
6228 * Alternate chains certificate forgery
6230 During certificate verfification, OpenSSL will attempt to find an
6231 alternative certificate chain if the first attempt to build such a chain
6232 fails. An error in the implementation of this logic can mean that an
6233 attacker could cause certain checks on untrusted certificates to be
6234 bypassed, such as the CA flag, enabling them to use a valid leaf
6235 certificate to act as a CA and "issue" an invalid certificate.
6237 This issue was reported to OpenSSL by Adam Langley/David Benjamin
6243 * Race condition handling PSK identify hint
6245 If PSK identity hints are received by a multi-threaded client then
6246 the values are wrongly updated in the parent SSL_CTX structure. This can
6247 result in a race condition potentially leading to a double free of the
6253 ### Changes between 1.0.1n and 1.0.1o [12 Jun 2015]
6255 * Fix HMAC ABI incompatibility. The previous version introduced an ABI
6256 incompatibility in the handling of HMAC. The previous ABI has now been
6259 ### Changes between 1.0.1m and 1.0.1n [11 Jun 2015]
6261 * Malformed ECParameters causes infinite loop
6263 When processing an ECParameters structure OpenSSL enters an infinite loop
6264 if the curve specified is over a specially malformed binary polynomial
6267 This can be used to perform denial of service against any
6268 system which processes public keys, certificate requests or
6269 certificates. This includes TLS clients and TLS servers with
6270 client authentication enabled.
6272 This issue was reported to OpenSSL by Joseph Barr-Pixton.
6277 * Exploitable out-of-bounds read in X509_cmp_time
6279 X509_cmp_time does not properly check the length of the ASN1_TIME
6280 string and can read a few bytes out of bounds. In addition,
6281 X509_cmp_time accepts an arbitrary number of fractional seconds in the
6284 An attacker can use this to craft malformed certificates and CRLs of
6285 various sizes and potentially cause a segmentation fault, resulting in
6286 a DoS on applications that verify certificates or CRLs. TLS clients
6287 that verify CRLs are affected. TLS clients and servers with client
6288 authentication enabled may be affected if they use custom verification
6291 This issue was reported to OpenSSL by Robert Swiecki (Google), and
6292 independently by Hanno Böck.
6297 * PKCS7 crash with missing EnvelopedContent
6299 The PKCS#7 parsing code does not handle missing inner EncryptedContent
6300 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
6301 with missing content and trigger a NULL pointer dereference on parsing.
6303 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
6304 structures from untrusted sources are affected. OpenSSL clients and
6305 servers are not affected.
6307 This issue was reported to OpenSSL by Michal Zalewski (Google).
6312 * CMS verify infinite loop with unknown hash function
6314 When verifying a signedData message the CMS code can enter an infinite loop
6315 if presented with an unknown hash function OID. This can be used to perform
6316 denial of service against any system which verifies signedData messages using
6318 This issue was reported to OpenSSL by Johannes Bauer.
6323 * Race condition handling NewSessionTicket
6325 If a NewSessionTicket is received by a multi-threaded client when attempting to
6326 reuse a previous ticket then a race condition can occur potentially leading to
6327 a double free of the ticket data.
6332 * Reject DH handshakes with parameters shorter than 768 bits.
6334 *Kurt Roeckx and Emilia Kasper*
6336 * dhparam: generate 2048-bit parameters by default.
6338 *Kurt Roeckx and Emilia Kasper*
6340 ### Changes between 1.0.1l and 1.0.1m [19 Mar 2015]
6342 * Segmentation fault in ASN1_TYPE_cmp fix
6344 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
6345 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
6346 certificate signature algorithm consistency this can be used to crash any
6347 certificate verification operation and exploited in a DoS attack. Any
6348 application which performs certificate verification is vulnerable including
6349 OpenSSL clients and servers which enable client authentication.
6354 * ASN.1 structure reuse memory corruption fix
6356 Reusing a structure in ASN.1 parsing may allow an attacker to cause
6357 memory corruption via an invalid write. Such reuse is and has been
6358 strongly discouraged and is believed to be rare.
6360 Applications that parse structures containing CHOICE or ANY DEFINED BY
6361 components may be affected. Certificate parsing (d2i_X509 and related
6362 functions) are however not affected. OpenSSL clients and servers are
6368 * PKCS7 NULL pointer dereferences fix
6370 The PKCS#7 parsing code does not handle missing outer ContentInfo
6371 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
6372 missing content and trigger a NULL pointer dereference on parsing.
6374 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
6375 otherwise parse PKCS#7 structures from untrusted sources are
6376 affected. OpenSSL clients and servers are not affected.
6378 This issue was reported to OpenSSL by Michal Zalewski (Google).
6383 * DoS via reachable assert in SSLv2 servers fix
6385 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
6386 servers that both support SSLv2 and enable export cipher suites by sending
6387 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
6389 This issue was discovered by Sean Burford (Google) and Emilia Käsper
6390 (OpenSSL development team).
6395 * Use After Free following d2i_ECPrivatekey error fix
6397 A malformed EC private key file consumed via the d2i_ECPrivateKey function
6398 could cause a use after free condition. This, in turn, could cause a double
6399 free in several private key parsing functions (such as d2i_PrivateKey
6400 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
6401 for applications that receive EC private keys from untrusted
6402 sources. This scenario is considered rare.
6404 This issue was discovered by the BoringSSL project and fixed in their
6410 * X509_to_X509_REQ NULL pointer deref fix
6412 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
6413 the certificate key is invalid. This function is rarely used in practice.
6415 This issue was discovered by Brian Carpenter.
6420 * Removed the export ciphers from the DEFAULT ciphers
6424 ### Changes between 1.0.1k and 1.0.1l [15 Jan 2015]
6426 * Build fixes for the Windows and OpenVMS platforms
6428 *Matt Caswell and Richard Levitte*
6430 ### Changes between 1.0.1j and 1.0.1k [8 Jan 2015]
6432 * Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
6433 message can cause a segmentation fault in OpenSSL due to a NULL pointer
6434 dereference. This could lead to a Denial Of Service attack. Thanks to
6435 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
6440 * Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
6441 dtls1_buffer_record function under certain conditions. In particular this
6442 could occur if an attacker sent repeated DTLS records with the same
6443 sequence number but for the next epoch. The memory leak could be exploited
6444 by an attacker in a Denial of Service attack through memory exhaustion.
6445 Thanks to Chris Mueller for reporting this issue.
6450 * Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
6451 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
6452 method would be set to NULL which could later result in a NULL pointer
6453 dereference. Thanks to Frank Schmirler for reporting this issue.
6458 * Abort handshake if server key exchange message is omitted for ephemeral
6461 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
6462 reporting this issue.
6467 * Remove non-export ephemeral RSA code on client and server. This code
6468 violated the TLS standard by allowing the use of temporary RSA keys in
6469 non-export ciphersuites and could be used by a server to effectively
6470 downgrade the RSA key length used to a value smaller than the server
6471 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
6472 INRIA or reporting this issue.
6477 * Fixed issue where DH client certificates are accepted without verification.
6478 An OpenSSL server will accept a DH certificate for client authentication
6479 without the certificate verify message. This effectively allows a client to
6480 authenticate without the use of a private key. This only affects servers
6481 which trust a client certificate authority which issues certificates
6482 containing DH keys: these are extremely rare and hardly ever encountered.
6483 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
6489 * Ensure that the session ID context of an SSL is updated when its
6490 SSL_CTX is updated via SSL_set_SSL_CTX.
6492 The session ID context is typically set from the parent SSL_CTX,
6493 and can vary with the CTX.
6497 * Fix various certificate fingerprint issues.
6499 By using non-DER or invalid encodings outside the signed portion of a
6500 certificate the fingerprint can be changed without breaking the signature.
6501 Although no details of the signed portion of the certificate can be changed
6502 this can cause problems with some applications: e.g. those using the
6503 certificate fingerprint for blacklists.
6505 1. Reject signatures with non zero unused bits.
6507 If the BIT STRING containing the signature has non zero unused bits reject
6508 the signature. All current signature algorithms require zero unused bits.
6510 2. Check certificate algorithm consistency.
6512 Check the AlgorithmIdentifier inside TBS matches the one in the
6513 certificate signature. NB: this will result in signature failure
6514 errors for some broken certificates.
6516 Thanks to Konrad Kraszewski from Google for reporting this issue.
6518 3. Check DSA/ECDSA signatures use DER.
6520 Re-encode DSA/ECDSA signatures and compare with the original received
6521 signature. Return an error if there is a mismatch.
6523 This will reject various cases including garbage after signature
6524 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
6525 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
6526 (negative or with leading zeroes).
6528 Further analysis was conducted and fixes were developed by Stephen Henson
6529 of the OpenSSL core team.
6535 * Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
6536 results on some platforms, including x86_64. This bug occurs at random
6537 with a very low probability, and is not known to be exploitable in any
6538 way, though its exact impact is difficult to determine. Thanks to Pieter
6539 Wuille (Blockstream) who reported this issue and also suggested an initial
6540 fix. Further analysis was conducted by the OpenSSL development team and
6541 Adam Langley of Google. The final fix was developed by Andy Polyakov of
6542 the OpenSSL core team.
6547 * Do not resume sessions on the server if the negotiated protocol
6548 version does not match the session's version. Resuming with a different
6549 version, while not strictly forbidden by the RFC, is of questionable
6550 sanity and breaks all known clients.
6552 *David Benjamin, Emilia Käsper*
6554 * Tighten handling of the ChangeCipherSpec (CCS) message: reject
6555 early CCS messages during renegotiation. (Note that because
6556 renegotiation is encrypted, this early CCS was not exploitable.)
6560 * Tighten client-side session ticket handling during renegotiation:
6561 ensure that the client only accepts a session ticket if the server sends
6562 the extension anew in the ServerHello. Previously, a TLS client would
6563 reuse the old extension state and thus accept a session ticket if one was
6564 announced in the initial ServerHello.
6566 Similarly, ensure that the client requires a session ticket if one
6567 was advertised in the ServerHello. Previously, a TLS client would
6568 ignore a missing NewSessionTicket message.
6572 ### Changes between 1.0.1i and 1.0.1j [15 Oct 2014]
6576 A flaw in the DTLS SRTP extension parsing code allows an attacker, who
6577 sends a carefully crafted handshake message, to cause OpenSSL to fail
6578 to free up to 64k of memory causing a memory leak. This could be
6579 exploited in a Denial Of Service attack. This issue affects OpenSSL
6580 1.0.1 server implementations for both SSL/TLS and DTLS regardless of
6581 whether SRTP is used or configured. Implementations of OpenSSL that
6582 have been compiled with OPENSSL_NO_SRTP defined are not affected.
6584 The fix was developed by the OpenSSL team.
6589 * Session Ticket Memory Leak.
6591 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
6592 integrity of that ticket is first verified. In the event of a session
6593 ticket integrity check failing, OpenSSL will fail to free memory
6594 causing a memory leak. By sending a large number of invalid session
6595 tickets an attacker could exploit this issue in a Denial Of Service
6601 * Build option no-ssl3 is incomplete.
6603 When OpenSSL is configured with "no-ssl3" as a build option, servers
6604 could accept and complete a SSL 3.0 handshake, and clients could be
6605 configured to send them.
6608 *Akamai and the OpenSSL team*
6610 * Add support for TLS_FALLBACK_SCSV.
6611 Client applications doing fallback retries should call
6612 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
6615 *Adam Langley, Bodo Moeller*
6617 * Add additional DigestInfo checks.
6619 Re-encode DigestInto in DER and check against the original when
6620 verifying RSA signature: this will reject any improperly encoded
6621 DigestInfo structures.
6623 Note: this is a precautionary measure and no attacks are currently known.
6627 ### Changes between 1.0.1h and 1.0.1i [6 Aug 2014]
6629 * Fix SRP buffer overrun vulnerability. Invalid parameters passed to the
6630 SRP code can be overrun an internal buffer. Add sanity check that
6631 g, A, B < N to SRP code.
6633 Thanks to Sean Devlin and Watson Ladd of Cryptography Services, NCC
6634 Group for discovering this issue.
6639 * A flaw in the OpenSSL SSL/TLS server code causes the server to negotiate
6640 TLS 1.0 instead of higher protocol versions when the ClientHello message
6641 is badly fragmented. This allows a man-in-the-middle attacker to force a
6642 downgrade to TLS 1.0 even if both the server and the client support a
6643 higher protocol version, by modifying the client's TLS records.
6645 Thanks to David Benjamin and Adam Langley (Google) for discovering and
6646 researching this issue.
6651 * OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
6652 to a denial of service attack. A malicious server can crash the client
6653 with a null pointer dereference (read) by specifying an anonymous (EC)DH
6654 ciphersuite and sending carefully crafted handshake messages.
6656 Thanks to Felix Gröbert (Google) for discovering and researching this
6662 * By sending carefully crafted DTLS packets an attacker could cause openssl
6663 to leak memory. This can be exploited through a Denial of Service attack.
6664 Thanks to Adam Langley for discovering and researching this issue.
6669 * An attacker can force openssl to consume large amounts of memory whilst
6670 processing DTLS handshake messages. This can be exploited through a
6671 Denial of Service attack.
6672 Thanks to Adam Langley for discovering and researching this issue.
6677 * An attacker can force an error condition which causes openssl to crash
6678 whilst processing DTLS packets due to memory being freed twice. This
6679 can be exploited through a Denial of Service attack.
6680 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
6686 * If a multithreaded client connects to a malicious server using a resumed
6687 session and the server sends an ec point format extension it could write
6688 up to 255 bytes to freed memory.
6690 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
6696 * A malicious server can crash an OpenSSL client with a null pointer
6697 dereference (read) by specifying an SRP ciphersuite even though it was not
6698 properly negotiated with the client. This can be exploited through a
6699 Denial of Service attack.
6701 Thanks to Joonas Kuorilehto and Riku Hietamäki (Codenomicon) for
6702 discovering and researching this issue.
6707 * A flaw in OBJ_obj2txt may cause pretty printing functions such as
6708 X509_name_oneline, X509_name_print_ex et al. to leak some information
6709 from the stack. Applications may be affected if they echo pretty printing
6710 output to the attacker.
6712 Thanks to Ivan Fratric (Google) for discovering this issue.
6715 *Emilia Käsper, and Steve Henson*
6717 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
6718 for corner cases. (Certain input points at infinity could lead to
6719 bogus results, with non-infinity inputs mapped to infinity too.)
6723 ### Changes between 1.0.1g and 1.0.1h [5 Jun 2014]
6725 * Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
6726 handshake can force the use of weak keying material in OpenSSL
6727 SSL/TLS clients and servers.
6729 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
6730 researching this issue. ([CVE-2014-0224])
6732 *KIKUCHI Masashi, Steve Henson*
6734 * Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
6735 OpenSSL DTLS client the code can be made to recurse eventually crashing
6738 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
6741 *Imre Rad, Steve Henson*
6743 * Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
6744 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
6745 client or server. This is potentially exploitable to run arbitrary
6746 code on a vulnerable client or server.
6748 Thanks to Jüri Aedla for reporting this issue. ([CVE-2014-0195])
6750 *Jüri Aedla, Steve Henson*
6752 * Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
6753 are subject to a denial of service attack.
6755 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
6756 this issue. ([CVE-2014-3470])
6758 *Felix Gröbert, Ivan Fratric, Steve Henson*
6760 * Harmonize version and its documentation. -f flag is used to display
6763 *mancha <mancha1@zoho.com>*
6765 * Fix eckey_priv_encode so it immediately returns an error upon a failure
6766 in i2d_ECPrivateKey.
6768 *mancha <mancha1@zoho.com>*
6770 * Fix some double frees. These are not thought to be exploitable.
6772 *mancha <mancha1@zoho.com>*
6774 ### Changes between 1.0.1f and 1.0.1g [7 Apr 2014]
6776 * A missing bounds check in the handling of the TLS heartbeat extension
6777 can be used to reveal up to 64k of memory to a connected client or
6780 Thanks for Neel Mehta of Google Security for discovering this bug and to
6781 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
6782 preparing the fix ([CVE-2014-0160])
6784 *Adam Langley, Bodo Moeller*
6786 * Fix for the attack described in the paper "Recovering OpenSSL
6787 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
6788 by Yuval Yarom and Naomi Benger. Details can be obtained from:
6789 <http://eprint.iacr.org/2014/140>
6791 Thanks to Yuval Yarom and Naomi Benger for discovering this
6792 flaw and to Yuval Yarom for supplying a fix ([CVE-2014-0076])
6794 *Yuval Yarom and Naomi Benger*
6796 * TLS pad extension: draft-agl-tls-padding-03
6798 Workaround for the "TLS hang bug" (see FAQ and PR#2771): if the
6799 TLS client Hello record length value would otherwise be > 255 and
6800 less that 512 pad with a dummy extension containing zeroes so it
6801 is at least 512 bytes long.
6803 *Adam Langley, Steve Henson*
6805 ### Changes between 1.0.1e and 1.0.1f [6 Jan 2014]
6807 * Fix for TLS record tampering bug. A carefully crafted invalid
6808 handshake could crash OpenSSL with a NULL pointer exception.
6809 Thanks to Anton Johansson for reporting this issues.
6812 * Keep original DTLS digest and encryption contexts in retransmission
6813 structures so we can use the previous session parameters if they need
6814 to be resent. ([CVE-2013-6450])
6818 * Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
6819 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
6820 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
6821 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
6822 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
6823 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
6825 *Rob Stradling, Adam Langley*
6827 ### Changes between 1.0.1d and 1.0.1e [11 Feb 2013]
6829 * Correct fix for CVE-2013-0169. The original didn't work on AES-NI
6830 supporting platforms or when small records were transferred.
6832 *Andy Polyakov, Steve Henson*
6834 ### Changes between 1.0.1c and 1.0.1d [5 Feb 2013]
6836 * Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
6838 This addresses the flaw in CBC record processing discovered by
6839 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
6840 at: <http://www.isg.rhul.ac.uk/tls/>
6842 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
6843 Security Group at Royal Holloway, University of London
6844 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
6845 Emilia Käsper for the initial patch.
6848 *Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson*
6850 * Fix flaw in AESNI handling of TLS 1.2 and 1.1 records for CBC mode
6851 ciphersuites which can be exploited in a denial of service attack.
6852 Thanks go to and to Adam Langley <agl@chromium.org> for discovering
6853 and detecting this bug and to Wolfgang Ettlinger
6854 <wolfgang.ettlinger@gmail.com> for independently discovering this issue.
6859 * Return an error when checking OCSP signatures when key is NULL.
6860 This fixes a DoS attack. ([CVE-2013-0166])
6864 * Make openssl verify return errors.
6866 *Chris Palmer <palmer@google.com> and Ben Laurie*
6868 * Call OCSP Stapling callback after ciphersuite has been chosen, so
6869 the right response is stapled. Also change SSL_get_certificate()
6870 so it returns the certificate actually sent.
6871 See <http://rt.openssl.org/Ticket/Display.html?id=2836>.
6873 *Rob Stradling <rob.stradling@comodo.com>*
6875 * Fix possible deadlock when decoding public keys.
6879 * Don't use TLS 1.0 record version number in initial client hello
6884 ### Changes between 1.0.1b and 1.0.1c [10 May 2012]
6886 * Sanity check record length before skipping explicit IV in TLS
6887 1.2, 1.1 and DTLS to fix DoS attack.
6889 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
6890 fuzzing as a service testing platform.
6895 * Initialise tkeylen properly when encrypting CMS messages.
6896 Thanks to Solar Designer of Openwall for reporting this issue.
6900 * In FIPS mode don't try to use composite ciphers as they are not
6905 ### Changes between 1.0.1a and 1.0.1b [26 Apr 2012]
6907 * OpenSSL 1.0.0 sets SSL_OP_ALL to 0x80000FFFL and OpenSSL 1.0.1 and
6908 1.0.1a set SSL_OP_NO_TLSv1_1 to 0x00000400L which would unfortunately
6909 mean any application compiled against OpenSSL 1.0.0 headers setting
6910 SSL_OP_ALL would also set SSL_OP_NO_TLSv1_1, unintentionally disabling
6911 TLS 1.1 also. Fix this by changing the value of SSL_OP_NO_TLSv1_1 to
6912 0x10000000L Any application which was previously compiled against
6913 OpenSSL 1.0.1 or 1.0.1a headers and which cares about SSL_OP_NO_TLSv1_1
6914 will need to be recompiled as a result. Letting be results in
6915 inability to disable specifically TLS 1.1 and in client context,
6916 in unlike event, limit maximum offered version to TLS 1.0 [see below].
6920 * In order to ensure interoperability SSL_OP_NO_protocolX does not
6921 disable just protocol X, but all protocols above X *if* there are
6922 protocols *below* X still enabled. In more practical terms it means
6923 that if application wants to disable TLS1.0 in favor of TLS1.1 and
6924 above, it's not sufficient to pass `SSL_OP_NO_TLSv1`, one has to pass
6925 `SSL_OP_NO_TLSv1|SSL_OP_NO_SSLv3|SSL_OP_NO_SSLv2`. This applies to
6930 ### Changes between 1.0.1 and 1.0.1a [19 Apr 2012]
6932 * Check for potentially exploitable overflows in asn1_d2i_read_bio
6933 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
6934 in CRYPTO_realloc_clean.
6936 Thanks to Tavis Ormandy, Google Security Team, for discovering this
6937 issue and to Adam Langley <agl@chromium.org> for fixing it.
6940 *Adam Langley (Google), Tavis Ormandy, Google Security Team*
6942 * Don't allow TLS 1.2 SHA-256 ciphersuites in TLS 1.0, 1.1 connections.
6946 * Workarounds for some broken servers that "hang" if a client hello
6947 record length exceeds 255 bytes.
6949 1. Do not use record version number > TLS 1.0 in initial client
6950 hello: some (but not all) hanging servers will now work.
6951 2. If we set OPENSSL_MAX_TLS1_2_CIPHER_LENGTH this will truncate
6952 the number of ciphers sent in the client hello. This should be
6953 set to an even number, such as 50, for example by passing:
6954 -DOPENSSL_MAX_TLS1_2_CIPHER_LENGTH=50 to config or Configure.
6955 Most broken servers should now work.
6956 3. If all else fails setting OPENSSL_NO_TLS1_2_CLIENT will disable
6957 TLS 1.2 client support entirely.
6961 * Fix SEGV in Vector Permutation AES module observed in OpenSSH.
6965 ### Changes between 1.0.0h and 1.0.1 [14 Mar 2012]
6967 * Add compatibility with old MDC2 signatures which use an ASN1 OCTET
6968 STRING form instead of a DigestInfo.
6972 * The format used for MDC2 RSA signatures is inconsistent between EVP
6973 and the RSA_sign/RSA_verify functions. This was made more apparent when
6974 OpenSSL used RSA_sign/RSA_verify for some RSA signatures in particular
6975 those which went through EVP_PKEY_METHOD in 1.0.0 and later. Detect
6976 the correct format in RSA_verify so both forms transparently work.
6980 * Some servers which support TLS 1.0 can choke if we initially indicate
6981 support for TLS 1.2 and later renegotiate using TLS 1.0 in the RSA
6982 encrypted premaster secret. As a workaround use the maximum permitted
6983 client version in client hello, this should keep such servers happy
6984 and still work with previous versions of OpenSSL.
6988 * Add support for TLS/DTLS heartbeats.
6990 *Robin Seggelmann <seggelmann@fh-muenster.de>*
6992 * Add support for SCTP.
6994 *Robin Seggelmann <seggelmann@fh-muenster.de>*
6996 * Improved PRNG seeding for VOS.
6998 *Paul Green <Paul.Green@stratus.com>*
7000 * Extensive assembler packs updates, most notably:
7002 - x86[_64]: AES-NI, PCLMULQDQ, RDRAND support;
7003 - x86[_64]: SSSE3 support (SHA1, vector-permutation AES);
7004 - x86_64: bit-sliced AES implementation;
7005 - ARM: NEON support, contemporary platforms optimizations;
7006 - s390x: z196 support;
7007 - `*`: GHASH and GF(2^m) multiplication implementations;
7011 * Make TLS-SRP code conformant with RFC 5054 API cleanup
7012 (removal of unnecessary code)
7014 *Peter Sylvester <peter.sylvester@edelweb.fr>*
7016 * Add TLS key material exporter from RFC 5705.
7020 * Add DTLS-SRTP negotiation from RFC 5764.
7024 * Add Next Protocol Negotiation,
7025 <http://tools.ietf.org/html/draft-agl-tls-nextprotoneg-00>. Can be
7026 disabled with a no-npn flag to config or Configure. Code donated
7029 *Adam Langley <agl@google.com> and Ben Laurie*
7031 * Add optional 64-bit optimized implementations of elliptic curves NIST-P224,
7032 NIST-P256, NIST-P521, with constant-time single point multiplication on
7033 typical inputs. Compiler support for the nonstandard type `__uint128_t` is
7034 required to use this (present in gcc 4.4 and later, for 64-bit builds).
7035 Code made available under Apache License version 2.0.
7037 Specify "enable-ec_nistp_64_gcc_128" on the Configure (or config) command
7038 line to include this in your build of OpenSSL, and run "make depend" (or
7039 "make update"). This enables the following EC_METHODs:
7041 EC_GFp_nistp224_method()
7042 EC_GFp_nistp256_method()
7043 EC_GFp_nistp521_method()
7045 EC_GROUP_new_by_curve_name() will automatically use these (while
7046 EC_GROUP_new_curve_GFp() currently prefers the more flexible
7049 *Emilia Käsper, Adam Langley, Bodo Moeller (Google)*
7051 * Use type ossl_ssize_t instead of ssize_t which isn't available on
7052 all platforms. Move ssize_t definition from e_os.h to the public
7053 header file e_os2.h as it now appears in public header file cms.h
7057 * New -sigopt option to the ca, req and x509 utilities. Additional
7058 signature parameters can be passed using this option and in
7063 * Add RSA PSS signing function. This will generate and set the
7064 appropriate AlgorithmIdentifiers for PSS based on those in the
7065 corresponding EVP_MD_CTX structure. No application support yet.
7069 * Support for companion algorithm specific ASN1 signing routines.
7070 New function ASN1_item_sign_ctx() signs a pre-initialised
7071 EVP_MD_CTX structure and sets AlgorithmIdentifiers based on
7072 the appropriate parameters.
7076 * Add new algorithm specific ASN1 verification initialisation function
7077 to EVP_PKEY_ASN1_METHOD: this is not in EVP_PKEY_METHOD since the ASN1
7078 handling will be the same no matter what EVP_PKEY_METHOD is used.
7079 Add a PSS handler to support verification of PSS signatures: checked
7080 against a number of sample certificates.
7084 * Add signature printing for PSS. Add PSS OIDs.
7086 *Steve Henson, Martin Kaiser <lists@kaiser.cx>*
7088 * Add algorithm specific signature printing. An individual ASN1 method
7089 can now print out signatures instead of the standard hex dump.
7091 More complex signatures (e.g. PSS) can print out more meaningful
7092 information. Include DSA version that prints out the signature
7097 * Password based recipient info support for CMS library: implementing
7102 * Split password based encryption into PBES2 and PBKDF2 functions. This
7103 neatly separates the code into cipher and PBE sections and is required
7104 for some algorithms that split PBES2 into separate pieces (such as
7105 password based CMS).
7109 * Session-handling fixes:
7110 - Fix handling of connections that are resuming with a session ID,
7111 but also support Session Tickets.
7112 - Fix a bug that suppressed issuing of a new ticket if the client
7113 presented a ticket with an expired session.
7114 - Try to set the ticket lifetime hint to something reasonable.
7115 - Make tickets shorter by excluding irrelevant information.
7116 - On the client side, don't ignore renewed tickets.
7118 *Adam Langley, Bodo Moeller (Google)*
7120 * Fix PSK session representation.
7124 * Add RC4-MD5 and AESNI-SHA1 "stitched" implementations.
7126 This work was sponsored by Intel.
7130 * Add GCM support to TLS library. Some custom code is needed to split
7131 the IV between the fixed (from PRF) and explicit (from TLS record)
7132 portions. This adds all GCM ciphersuites supported by RFC5288 and
7133 RFC5289. Generalise some `AES*` cipherstrings to include GCM and
7134 add a special AESGCM string for GCM only.
7138 * Expand range of ctrls for AES GCM. Permit setting invocation
7139 field on decrypt and retrieval of invocation field only on encrypt.
7143 * Add HMAC ECC ciphersuites from RFC5289. Include SHA384 PRF support.
7144 As required by RFC5289 these ciphersuites cannot be used if for
7145 versions of TLS earlier than 1.2.
7149 * For FIPS capable OpenSSL interpret a NULL default public key method
7150 as unset and return the appropriate default but do *not* set the default.
7151 This means we can return the appropriate method in applications that
7152 switch between FIPS and non-FIPS modes.
7156 * Redirect HMAC and CMAC operations to FIPS module in FIPS mode. If an
7157 ENGINE is used then we cannot handle that in the FIPS module so we
7158 keep original code iff non-FIPS operations are allowed.
7162 * Add -attime option to openssl utilities.
7164 *Peter Eckersley <pde@eff.org>, Ben Laurie and Steve Henson*
7166 * Redirect DSA and DH operations to FIPS module in FIPS mode.
7170 * Redirect ECDSA and ECDH operations to FIPS module in FIPS mode. Also use
7171 FIPS EC methods unconditionally for now.
7175 * New build option no-ec2m to disable characteristic 2 code.
7179 * Backport libcrypto audit of return value checking from 1.1.0-dev; not
7180 all cases can be covered as some introduce binary incompatibilities.
7184 * Redirect RSA operations to FIPS module including keygen,
7185 encrypt, decrypt, sign and verify. Block use of non FIPS RSA methods.
7189 * Add similar low-level API blocking to ciphers.
7193 * low-level digest APIs are not approved in FIPS mode: any attempt
7194 to use these will cause a fatal error. Applications that *really* want
7195 to use them can use the `private_*` version instead.
7199 * Redirect cipher operations to FIPS module for FIPS builds.
7203 * Redirect digest operations to FIPS module for FIPS builds.
7207 * Update build system to add "fips" flag which will link in fipscanister.o
7208 for static and shared library builds embedding a signature if needed.
7212 * Output TLS supported curves in preference order instead of numerical
7213 order. This is currently hardcoded for the highest order curves first.
7214 This should be configurable so applications can judge speed vs strength.
7218 * Add TLS v1.2 server support for client authentication.
7222 * Add support for FIPS mode in ssl library: disable SSLv3, non-FIPS ciphers
7227 * Functions FIPS_mode_set() and FIPS_mode() which call the underlying
7228 FIPS modules versions.
7232 * Add TLS v1.2 client side support for client authentication. Keep cache
7233 of handshake records longer as we don't know the hash algorithm to use
7234 until after the certificate request message is received.
7238 * Initial TLS v1.2 client support. Add a default signature algorithms
7239 extension including all the algorithms we support. Parse new signature
7240 format in client key exchange. Relax some ECC signing restrictions for
7241 TLS v1.2 as indicated in RFC5246.
7245 * Add server support for TLS v1.2 signature algorithms extension. Switch
7246 to new signature format when needed using client digest preference.
7247 All server ciphersuites should now work correctly in TLS v1.2. No client
7248 support yet and no support for client certificates.
7252 * Initial TLS v1.2 support. Add new SHA256 digest to ssl code, switch
7253 to SHA256 for PRF when using TLS v1.2 and later. Add new SHA256 based
7254 ciphersuites. At present only RSA key exchange ciphersuites work with
7255 TLS v1.2. Add new option for TLS v1.2 replacing the old and obsolete
7256 SSL_OP_PKCS1_CHECK flags with SSL_OP_NO_TLSv1_2. New TLSv1.2 methods
7257 and version checking.
7261 * New option OPENSSL_NO_SSL_INTERN. If an application can be compiled
7262 with this defined it will not be affected by any changes to ssl internal
7263 structures. Add several utility functions to allow openssl application
7264 to work with OPENSSL_NO_SSL_INTERN defined.
7268 * A long standing patch to add support for SRP from EdelWeb (Peter
7269 Sylvester and Christophe Renou) was integrated.
7270 *Christophe Renou <christophe.renou@edelweb.fr>, Peter Sylvester
7271 <peter.sylvester@edelweb.fr>, Tom Wu <tjw@cs.stanford.edu>, and
7274 * Add functions to copy EVP_PKEY_METHOD and retrieve flags and id.
7278 * Permit abbreviated handshakes when renegotiating using the function
7279 SSL_renegotiate_abbreviated().
7281 *Robin Seggelmann <seggelmann@fh-muenster.de>*
7283 * Add call to ENGINE_register_all_complete() to
7284 ENGINE_load_builtin_engines(), so some implementations get used
7285 automatically instead of needing explicit application support.
7289 * Add support for TLS key exporter as described in RFC5705.
7291 *Robin Seggelmann <seggelmann@fh-muenster.de>, Steve Henson*
7293 * Initial TLSv1.1 support. Since TLSv1.1 is very similar to TLS v1.0 only
7294 a few changes are required:
7296 Add SSL_OP_NO_TLSv1_1 flag.
7297 Add TLSv1_1 methods.
7298 Update version checking logic to handle version 1.1.
7299 Add explicit IV handling (ported from DTLS code).
7300 Add command line options to s_client/s_server.
7307 ### Changes between 1.0.0s and 1.0.0t [3 Dec 2015]
7309 * X509_ATTRIBUTE memory leak
7311 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
7312 memory. This structure is used by the PKCS#7 and CMS routines so any
7313 application which reads PKCS#7 or CMS data from untrusted sources is
7314 affected. SSL/TLS is not affected.
7316 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
7322 * Race condition handling PSK identify hint
7324 If PSK identity hints are received by a multi-threaded client then
7325 the values are wrongly updated in the parent SSL_CTX structure. This can
7326 result in a race condition potentially leading to a double free of the
7332 ### Changes between 1.0.0r and 1.0.0s [11 Jun 2015]
7334 * Malformed ECParameters causes infinite loop
7336 When processing an ECParameters structure OpenSSL enters an infinite loop
7337 if the curve specified is over a specially malformed binary polynomial
7340 This can be used to perform denial of service against any
7341 system which processes public keys, certificate requests or
7342 certificates. This includes TLS clients and TLS servers with
7343 client authentication enabled.
7345 This issue was reported to OpenSSL by Joseph Barr-Pixton.
7350 * Exploitable out-of-bounds read in X509_cmp_time
7352 X509_cmp_time does not properly check the length of the ASN1_TIME
7353 string and can read a few bytes out of bounds. In addition,
7354 X509_cmp_time accepts an arbitrary number of fractional seconds in the
7357 An attacker can use this to craft malformed certificates and CRLs of
7358 various sizes and potentially cause a segmentation fault, resulting in
7359 a DoS on applications that verify certificates or CRLs. TLS clients
7360 that verify CRLs are affected. TLS clients and servers with client
7361 authentication enabled may be affected if they use custom verification
7364 This issue was reported to OpenSSL by Robert Swiecki (Google), and
7365 independently by Hanno Böck.
7370 * PKCS7 crash with missing EnvelopedContent
7372 The PKCS#7 parsing code does not handle missing inner EncryptedContent
7373 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
7374 with missing content and trigger a NULL pointer dereference on parsing.
7376 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
7377 structures from untrusted sources are affected. OpenSSL clients and
7378 servers are not affected.
7380 This issue was reported to OpenSSL by Michal Zalewski (Google).
7385 * CMS verify infinite loop with unknown hash function
7387 When verifying a signedData message the CMS code can enter an infinite loop
7388 if presented with an unknown hash function OID. This can be used to perform
7389 denial of service against any system which verifies signedData messages using
7391 This issue was reported to OpenSSL by Johannes Bauer.
7396 * Race condition handling NewSessionTicket
7398 If a NewSessionTicket is received by a multi-threaded client when attempting to
7399 reuse a previous ticket then a race condition can occur potentially leading to
7400 a double free of the ticket data.
7405 ### Changes between 1.0.0q and 1.0.0r [19 Mar 2015]
7407 * Segmentation fault in ASN1_TYPE_cmp fix
7409 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
7410 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
7411 certificate signature algorithm consistency this can be used to crash any
7412 certificate verification operation and exploited in a DoS attack. Any
7413 application which performs certificate verification is vulnerable including
7414 OpenSSL clients and servers which enable client authentication.
7419 * ASN.1 structure reuse memory corruption fix
7421 Reusing a structure in ASN.1 parsing may allow an attacker to cause
7422 memory corruption via an invalid write. Such reuse is and has been
7423 strongly discouraged and is believed to be rare.
7425 Applications that parse structures containing CHOICE or ANY DEFINED BY
7426 components may be affected. Certificate parsing (d2i_X509 and related
7427 functions) are however not affected. OpenSSL clients and servers are
7433 * PKCS7 NULL pointer dereferences fix
7435 The PKCS#7 parsing code does not handle missing outer ContentInfo
7436 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
7437 missing content and trigger a NULL pointer dereference on parsing.
7439 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
7440 otherwise parse PKCS#7 structures from untrusted sources are
7441 affected. OpenSSL clients and servers are not affected.
7443 This issue was reported to OpenSSL by Michal Zalewski (Google).
7448 * DoS via reachable assert in SSLv2 servers fix
7450 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
7451 servers that both support SSLv2 and enable export cipher suites by sending
7452 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
7454 This issue was discovered by Sean Burford (Google) and Emilia Käsper
7455 (OpenSSL development team).
7460 * Use After Free following d2i_ECPrivatekey error fix
7462 A malformed EC private key file consumed via the d2i_ECPrivateKey function
7463 could cause a use after free condition. This, in turn, could cause a double
7464 free in several private key parsing functions (such as d2i_PrivateKey
7465 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
7466 for applications that receive EC private keys from untrusted
7467 sources. This scenario is considered rare.
7469 This issue was discovered by the BoringSSL project and fixed in their
7475 * X509_to_X509_REQ NULL pointer deref fix
7477 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
7478 the certificate key is invalid. This function is rarely used in practice.
7480 This issue was discovered by Brian Carpenter.
7485 * Removed the export ciphers from the DEFAULT ciphers
7489 ### Changes between 1.0.0p and 1.0.0q [15 Jan 2015]
7491 * Build fixes for the Windows and OpenVMS platforms
7493 *Matt Caswell and Richard Levitte*
7495 ### Changes between 1.0.0o and 1.0.0p [8 Jan 2015]
7497 * Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
7498 message can cause a segmentation fault in OpenSSL due to a NULL pointer
7499 dereference. This could lead to a Denial Of Service attack. Thanks to
7500 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
7505 * Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
7506 dtls1_buffer_record function under certain conditions. In particular this
7507 could occur if an attacker sent repeated DTLS records with the same
7508 sequence number but for the next epoch. The memory leak could be exploited
7509 by an attacker in a Denial of Service attack through memory exhaustion.
7510 Thanks to Chris Mueller for reporting this issue.
7515 * Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
7516 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
7517 method would be set to NULL which could later result in a NULL pointer
7518 dereference. Thanks to Frank Schmirler for reporting this issue.
7523 * Abort handshake if server key exchange message is omitted for ephemeral
7526 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
7527 reporting this issue.
7532 * Remove non-export ephemeral RSA code on client and server. This code
7533 violated the TLS standard by allowing the use of temporary RSA keys in
7534 non-export ciphersuites and could be used by a server to effectively
7535 downgrade the RSA key length used to a value smaller than the server
7536 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
7537 INRIA or reporting this issue.
7542 * Fixed issue where DH client certificates are accepted without verification.
7543 An OpenSSL server will accept a DH certificate for client authentication
7544 without the certificate verify message. This effectively allows a client to
7545 authenticate without the use of a private key. This only affects servers
7546 which trust a client certificate authority which issues certificates
7547 containing DH keys: these are extremely rare and hardly ever encountered.
7548 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
7554 * Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
7555 results on some platforms, including x86_64. This bug occurs at random
7556 with a very low probability, and is not known to be exploitable in any
7557 way, though its exact impact is difficult to determine. Thanks to Pieter
7558 Wuille (Blockstream) who reported this issue and also suggested an initial
7559 fix. Further analysis was conducted by the OpenSSL development team and
7560 Adam Langley of Google. The final fix was developed by Andy Polyakov of
7561 the OpenSSL core team.
7566 * Fix various certificate fingerprint issues.
7568 By using non-DER or invalid encodings outside the signed portion of a
7569 certificate the fingerprint can be changed without breaking the signature.
7570 Although no details of the signed portion of the certificate can be changed
7571 this can cause problems with some applications: e.g. those using the
7572 certificate fingerprint for blacklists.
7574 1. Reject signatures with non zero unused bits.
7576 If the BIT STRING containing the signature has non zero unused bits reject
7577 the signature. All current signature algorithms require zero unused bits.
7579 2. Check certificate algorithm consistency.
7581 Check the AlgorithmIdentifier inside TBS matches the one in the
7582 certificate signature. NB: this will result in signature failure
7583 errors for some broken certificates.
7585 Thanks to Konrad Kraszewski from Google for reporting this issue.
7587 3. Check DSA/ECDSA signatures use DER.
7589 Reencode DSA/ECDSA signatures and compare with the original received
7590 signature. Return an error if there is a mismatch.
7592 This will reject various cases including garbage after signature
7593 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
7594 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
7595 (negative or with leading zeroes).
7597 Further analysis was conducted and fixes were developed by Stephen Henson
7598 of the OpenSSL core team.
7604 ### Changes between 1.0.0n and 1.0.0o [15 Oct 2014]
7606 * Session Ticket Memory Leak.
7608 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
7609 integrity of that ticket is first verified. In the event of a session
7610 ticket integrity check failing, OpenSSL will fail to free memory
7611 causing a memory leak. By sending a large number of invalid session
7612 tickets an attacker could exploit this issue in a Denial Of Service
7618 * Build option no-ssl3 is incomplete.
7620 When OpenSSL is configured with "no-ssl3" as a build option, servers
7621 could accept and complete a SSL 3.0 handshake, and clients could be
7622 configured to send them.
7625 *Akamai and the OpenSSL team*
7627 * Add support for TLS_FALLBACK_SCSV.
7628 Client applications doing fallback retries should call
7629 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
7632 *Adam Langley, Bodo Moeller*
7634 * Add additional DigestInfo checks.
7636 Reencode DigestInto in DER and check against the original when
7637 verifying RSA signature: this will reject any improperly encoded
7638 DigestInfo structures.
7640 Note: this is a precautionary measure and no attacks are currently known.
7644 ### Changes between 1.0.0m and 1.0.0n [6 Aug 2014]
7646 * OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
7647 to a denial of service attack. A malicious server can crash the client
7648 with a null pointer dereference (read) by specifying an anonymous (EC)DH
7649 ciphersuite and sending carefully crafted handshake messages.
7651 Thanks to Felix Gröbert (Google) for discovering and researching this
7657 * By sending carefully crafted DTLS packets an attacker could cause openssl
7658 to leak memory. This can be exploited through a Denial of Service attack.
7659 Thanks to Adam Langley for discovering and researching this issue.
7664 * An attacker can force openssl to consume large amounts of memory whilst
7665 processing DTLS handshake messages. This can be exploited through a
7666 Denial of Service attack.
7667 Thanks to Adam Langley for discovering and researching this issue.
7672 * An attacker can force an error condition which causes openssl to crash
7673 whilst processing DTLS packets due to memory being freed twice. This
7674 can be exploited through a Denial of Service attack.
7675 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
7681 * If a multithreaded client connects to a malicious server using a resumed
7682 session and the server sends an ec point format extension it could write
7683 up to 255 bytes to freed memory.
7685 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
7691 * A flaw in OBJ_obj2txt may cause pretty printing functions such as
7692 X509_name_oneline, X509_name_print_ex et al. to leak some information
7693 from the stack. Applications may be affected if they echo pretty printing
7694 output to the attacker.
7696 Thanks to Ivan Fratric (Google) for discovering this issue.
7699 *Emilia Käsper, and Steve Henson*
7701 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
7702 for corner cases. (Certain input points at infinity could lead to
7703 bogus results, with non-infinity inputs mapped to infinity too.)
7707 ### Changes between 1.0.0l and 1.0.0m [5 Jun 2014]
7709 * Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
7710 handshake can force the use of weak keying material in OpenSSL
7711 SSL/TLS clients and servers.
7713 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
7714 researching this issue. ([CVE-2014-0224])
7716 *KIKUCHI Masashi, Steve Henson*
7718 * Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
7719 OpenSSL DTLS client the code can be made to recurse eventually crashing
7722 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
7725 *Imre Rad, Steve Henson*
7727 * Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
7728 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
7729 client or server. This is potentially exploitable to run arbitrary
7730 code on a vulnerable client or server.
7732 Thanks to Jüri Aedla for reporting this issue. ([CVE-2014-0195])
7734 *Jüri Aedla, Steve Henson*
7736 * Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
7737 are subject to a denial of service attack.
7739 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
7740 this issue. ([CVE-2014-3470])
7742 *Felix Gröbert, Ivan Fratric, Steve Henson*
7744 * Harmonize version and its documentation. -f flag is used to display
7747 *mancha <mancha1@zoho.com>*
7749 * Fix eckey_priv_encode so it immediately returns an error upon a failure
7750 in i2d_ECPrivateKey.
7752 *mancha <mancha1@zoho.com>*
7754 * Fix some double frees. These are not thought to be exploitable.
7756 *mancha <mancha1@zoho.com>*
7758 * Fix for the attack described in the paper "Recovering OpenSSL
7759 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
7760 by Yuval Yarom and Naomi Benger. Details can be obtained from:
7761 <http://eprint.iacr.org/2014/140>
7763 Thanks to Yuval Yarom and Naomi Benger for discovering this
7764 flaw and to Yuval Yarom for supplying a fix ([CVE-2014-0076])
7766 *Yuval Yarom and Naomi Benger*
7768 ### Changes between 1.0.0k and 1.0.0l [6 Jan 2014]
7770 * Keep original DTLS digest and encryption contexts in retransmission
7771 structures so we can use the previous session parameters if they need
7772 to be resent. ([CVE-2013-6450])
7776 * Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
7777 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
7778 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
7779 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
7780 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
7781 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
7783 *Rob Stradling, Adam Langley*
7785 ### Changes between 1.0.0j and 1.0.0k [5 Feb 2013]
7787 * Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
7789 This addresses the flaw in CBC record processing discovered by
7790 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
7791 at: <http://www.isg.rhul.ac.uk/tls/>
7793 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
7794 Security Group at Royal Holloway, University of London
7795 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
7796 Emilia Käsper for the initial patch.
7799 *Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson*
7801 * Return an error when checking OCSP signatures when key is NULL.
7802 This fixes a DoS attack. ([CVE-2013-0166])
7806 * Call OCSP Stapling callback after ciphersuite has been chosen, so
7807 the right response is stapled. Also change SSL_get_certificate()
7808 so it returns the certificate actually sent.
7809 See <http://rt.openssl.org/Ticket/Display.html?id=2836>.
7810 (This is a backport)
7812 *Rob Stradling <rob.stradling@comodo.com>*
7814 * Fix possible deadlock when decoding public keys.
7818 ### Changes between 1.0.0i and 1.0.0j [10 May 2012]
7820 [NB: OpenSSL 1.0.0i and later 1.0.0 patch levels were released after
7823 * Sanity check record length before skipping explicit IV in DTLS
7826 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
7827 fuzzing as a service testing platform.
7832 * Initialise tkeylen properly when encrypting CMS messages.
7833 Thanks to Solar Designer of Openwall for reporting this issue.
7837 ### Changes between 1.0.0h and 1.0.0i [19 Apr 2012]
7839 * Check for potentially exploitable overflows in asn1_d2i_read_bio
7840 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
7841 in CRYPTO_realloc_clean.
7843 Thanks to Tavis Ormandy, Google Security Team, for discovering this
7844 issue and to Adam Langley <agl@chromium.org> for fixing it.
7847 *Adam Langley (Google), Tavis Ormandy, Google Security Team*
7849 ### Changes between 1.0.0g and 1.0.0h [12 Mar 2012]
7851 * Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness
7852 in CMS and PKCS7 code. When RSA decryption fails use a random key for
7853 content decryption and always return the same error. Note: this attack
7854 needs on average 2^20 messages so it only affects automated senders. The
7855 old behaviour can be re-enabled in the CMS code by setting the
7856 CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where
7857 an MMA defence is not necessary.
7858 Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering
7859 this issue. ([CVE-2012-0884])
7863 * Fix CVE-2011-4619: make sure we really are receiving a
7864 client hello before rejecting multiple SGC restarts. Thanks to
7865 Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug.
7869 ### Changes between 1.0.0f and 1.0.0g [18 Jan 2012]
7871 * Fix for DTLS DoS issue introduced by fix for CVE-2011-4109.
7872 Thanks to Antonio Martin, Enterprise Secure Access Research and
7873 Development, Cisco Systems, Inc. for discovering this bug and
7874 preparing a fix. ([CVE-2012-0050])
7878 ### Changes between 1.0.0e and 1.0.0f [4 Jan 2012]
7880 * Nadhem Alfardan and Kenny Paterson have discovered an extension
7881 of the Vaudenay padding oracle attack on CBC mode encryption
7882 which enables an efficient plaintext recovery attack against
7883 the OpenSSL implementation of DTLS. Their attack exploits timing
7884 differences arising during decryption processing. A research
7885 paper describing this attack can be found at:
7886 <http://www.isg.rhul.ac.uk/~kp/dtls.pdf>
7887 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
7888 Security Group at Royal Holloway, University of London
7889 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
7890 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
7891 for preparing the fix. ([CVE-2011-4108])
7893 *Robin Seggelmann, Michael Tuexen*
7895 * Clear bytes used for block padding of SSL 3.0 records.
7898 *Adam Langley (Google)*
7900 * Only allow one SGC handshake restart for SSL/TLS. Thanks to George
7901 Kadianakis <desnacked@gmail.com> for discovering this issue and
7902 Adam Langley for preparing the fix. ([CVE-2011-4619])
7904 *Adam Langley (Google)*
7906 * Check parameters are not NULL in GOST ENGINE. ([CVE-2012-0027])
7908 *Andrey Kulikov <amdeich@gmail.com>*
7910 * Prevent malformed RFC3779 data triggering an assertion failure.
7911 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
7912 and Rob Austein <sra@hactrn.net> for fixing it. ([CVE-2011-4577])
7914 *Rob Austein <sra@hactrn.net>*
7916 * Improved PRNG seeding for VOS.
7918 *Paul Green <Paul.Green@stratus.com>*
7920 * Fix ssl_ciph.c set-up race.
7922 *Adam Langley (Google)*
7924 * Fix spurious failures in ecdsatest.c.
7926 *Emilia Käsper (Google)*
7928 * Fix the BIO_f_buffer() implementation (which was mixing different
7929 interpretations of the `..._len` fields).
7931 *Adam Langley (Google)*
7933 * Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
7934 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
7935 threads won't reuse the same blinding coefficients.
7937 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
7938 lock to call BN_BLINDING_invert_ex, and avoids one use of
7939 BN_BLINDING_update for each BN_BLINDING structure (previously,
7940 the last update always remained unused).
7942 *Emilia Käsper (Google)*
7944 * In ssl3_clear, preserve s3->init_extra along with s3->rbuf.
7946 *Bob Buckholz (Google)*
7948 ### Changes between 1.0.0d and 1.0.0e [6 Sep 2011]
7950 * Fix bug where CRLs with nextUpdate in the past are sometimes accepted
7951 by initialising X509_STORE_CTX properly. ([CVE-2011-3207])
7953 *Kaspar Brand <ossl@velox.ch>*
7955 * Fix SSL memory handling for (EC)DH ciphersuites, in particular
7956 for multi-threaded use of ECDH. ([CVE-2011-3210])
7958 *Adam Langley (Google)*
7960 * Fix x509_name_ex_d2i memory leak on bad inputs.
7964 * Remove hard coded ecdsaWithSHA1 signature tests in ssl code and check
7965 signature public key algorithm by using OID xref utilities instead.
7966 Before this you could only use some ECC ciphersuites with SHA1 only.
7970 * Add protection against ECDSA timing attacks as mentioned in the paper
7971 by Billy Bob Brumley and Nicola Tuveri, see:
7972 <http://eprint.iacr.org/2011/232.pdf>
7974 *Billy Bob Brumley and Nicola Tuveri*
7976 ### Changes between 1.0.0c and 1.0.0d [8 Feb 2011]
7978 * Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
7980 *Neel Mehta, Adam Langley, Bodo Moeller (Google)*
7982 * Fix bug in string printing code: if *any* escaping is enabled we must
7983 escape the escape character (backslash) or the resulting string is
7988 ### Changes between 1.0.0b and 1.0.0c [2 Dec 2010]
7990 * Disable code workaround for ancient and obsolete Netscape browsers
7991 and servers: an attacker can use it in a ciphersuite downgrade attack.
7992 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
7996 * Fixed J-PAKE implementation error, originally discovered by
7997 Sebastien Martini, further info and confirmation from Stefan
7998 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
8002 ### Changes between 1.0.0a and 1.0.0b [16 Nov 2010]
8004 * Fix extension code to avoid race conditions which can result in a buffer
8005 overrun vulnerability: resumed sessions must not be modified as they can
8006 be shared by multiple threads. CVE-2010-3864
8010 * Fix WIN32 build system to correctly link an ENGINE directory into
8015 ### Changes between 1.0.0 and 1.0.0a [01 Jun 2010]
8017 * Check return value of int_rsa_verify in pkey_rsa_verifyrecover
8020 *Steve Henson, Peter-Michael Hager <hager@dortmund.net>*
8022 ### Changes between 0.9.8n and 1.0.0 [29 Mar 2010]
8024 * Add "missing" function EVP_CIPHER_CTX_copy(). This copies a cipher
8025 context. The operation can be customised via the ctrl mechanism in
8026 case ENGINEs want to include additional functionality.
8030 * Tolerate yet another broken PKCS#8 key format: private key value negative.
8034 * Add new -subject_hash_old and -issuer_hash_old options to x509 utility to
8035 output hashes compatible with older versions of OpenSSL.
8037 *Willy Weisz <weisz@vcpc.univie.ac.at>*
8039 * Fix compression algorithm handling: if resuming a session use the
8040 compression algorithm of the resumed session instead of determining
8041 it from client hello again. Don't allow server to change algorithm.
8045 * Add load_crls() function to commands tidying load_certs() too. Add option
8046 to verify utility to allow additional CRLs to be included.
8050 * Update OCSP request code to permit adding custom headers to the request:
8051 some responders need this.
8055 * The function EVP_PKEY_sign() returns <=0 on error: check return code
8058 *Julia Lawall <julia@diku.dk>*
8060 * Update verify callback code in `apps/s_cb.c` and `apps/verify.c`, it
8061 needlessly dereferenced structures, used obsolete functions and
8062 didn't handle all updated verify codes correctly.
8066 * Disable MD2 in the default configuration.
8070 * In BIO_pop() and BIO_push() use the ctrl argument (which was NULL) to
8071 indicate the initial BIO being pushed or popped. This makes it possible
8072 to determine whether the BIO is the one explicitly called or as a result
8073 of the ctrl being passed down the chain. Fix BIO_pop() and SSL BIOs so
8074 it handles reference counts correctly and doesn't zero out the I/O bio
8075 when it is not being explicitly popped. WARNING: applications which
8076 included workarounds for the old buggy behaviour will need to be modified
8077 or they could free up already freed BIOs.
8081 * Extend the uni2asc/asc2uni => OPENSSL_uni2asc/OPENSSL_asc2uni
8082 renaming to all platforms (within the 0.9.8 branch, this was
8083 done conditionally on Netware platforms to avoid a name clash).
8085 *Guenter <lists@gknw.net>*
8087 * Add ECDHE and PSK support to DTLS.
8089 *Michael Tuexen <tuexen@fh-muenster.de>*
8091 * Add CHECKED_STACK_OF macro to safestack.h, otherwise safestack can't
8096 * Add "missing" function EVP_MD_flags() (without this the only way to
8097 retrieve a digest flags is by accessing the structure directly. Update
8098 `EVP_MD_do_all*()` and `EVP_CIPHER_do_all*()` to include the name a digest
8099 or cipher is registered as in the "from" argument. Print out all
8100 registered digests in the dgst usage message instead of manually
8101 attempting to work them out.
8105 * If no SSLv2 ciphers are used don't use an SSLv2 compatible client hello:
8106 this allows the use of compression and extensions. Change default cipher
8107 string to remove SSLv2 ciphersuites. This effectively avoids ancient SSLv2
8108 by default unless an application cipher string requests it.
8112 * Alter match criteria in PKCS12_parse(). It used to try to use local
8113 key ids to find matching certificates and keys but some PKCS#12 files
8114 don't follow the (somewhat unwritten) rules and this strategy fails.
8115 Now just gather all certificates together and the first private key
8116 then look for the first certificate that matches the key.
8120 * Support use of registered digest and cipher names for dgst and cipher
8121 commands instead of having to add each one as a special case. So now
8128 openssl dgst -sha256 foo
8130 and this works for ENGINE based algorithms too.
8134 * Update Gost ENGINE to support parameter files.
8136 *Victor B. Wagner <vitus@cryptocom.ru>*
8138 * Support GeneralizedTime in ca utility.
8140 *Oliver Martin <oliver@volatilevoid.net>, Steve Henson*
8142 * Enhance the hash format used for certificate directory links. The new
8143 form uses the canonical encoding (meaning equivalent names will work
8144 even if they aren't identical) and uses SHA1 instead of MD5. This form
8145 is incompatible with the older format and as a result c_rehash should
8146 be used to rebuild symbolic links.
8150 * Make PKCS#8 the default write format for private keys, replacing the
8151 traditional format. This form is standardised, more secure and doesn't
8152 include an implicit MD5 dependency.
8156 * Add a $gcc_devteam_warn option to Configure. The idea is that any code
8157 committed to OpenSSL should pass this lot as a minimum.
8161 * Add session ticket override functionality for use by EAP-FAST.
8163 *Jouni Malinen <j@w1.fi>*
8165 * Modify HMAC functions to return a value. Since these can be implemented
8166 in an ENGINE errors can occur.
8170 * Type-checked OBJ_bsearch_ex.
8174 * Type-checked OBJ_bsearch. Also some constification necessitated
8175 by type-checking. Still to come: TXT_DB, bsearch(?),
8176 OBJ_bsearch_ex, qsort, CRYPTO_EX_DATA, ASN1_VALUE, ASN1_STRING,
8181 * New function OPENSSL_gmtime_adj() to add a specific number of days and
8182 seconds to a tm structure directly, instead of going through OS
8183 specific date routines. This avoids any issues with OS routines such
8184 as the year 2038 bug. New `*_adj()` functions for ASN1 time structures
8185 and X509_time_adj_ex() to cover the extended range. The existing
8186 X509_time_adj() is still usable and will no longer have any date issues.
8190 * Delta CRL support. New use deltas option which will attempt to locate
8191 and search any appropriate delta CRLs available.
8193 This work was sponsored by Google.
8197 * Support for CRLs partitioned by reason code. Reorganise CRL processing
8198 code and add additional score elements. Validate alternate CRL paths
8199 as part of the CRL checking and indicate a new error "CRL path validation
8200 error" in this case. Applications wanting additional details can use
8201 the verify callback and check the new "parent" field. If this is not
8202 NULL CRL path validation is taking place. Existing applications won't
8203 see this because it requires extended CRL support which is off by
8206 This work was sponsored by Google.
8210 * Support for freshest CRL extension.
8212 This work was sponsored by Google.
8216 * Initial indirect CRL support. Currently only supported in the CRLs
8217 passed directly and not via lookup. Process certificate issuer
8218 CRL entry extension and lookup CRL entries by bother issuer name
8219 and serial number. Check and process CRL issuer entry in IDP extension.
8221 This work was sponsored by Google.
8225 * Add support for distinct certificate and CRL paths. The CRL issuer
8226 certificate is validated separately in this case. Only enabled if
8227 an extended CRL support flag is set: this flag will enable additional
8228 CRL functionality in future.
8230 This work was sponsored by Google.
8234 * Add support for policy mappings extension.
8236 This work was sponsored by Google.
8240 * Fixes to pathlength constraint, self issued certificate handling,
8241 policy processing to align with RFC3280 and PKITS tests.
8243 This work was sponsored by Google.
8247 * Support for name constraints certificate extension. DN, email, DNS
8248 and URI types are currently supported.
8250 This work was sponsored by Google.
8254 * To cater for systems that provide a pointer-based thread ID rather
8255 than numeric, deprecate the current numeric thread ID mechanism and
8256 replace it with a structure and associated callback type. This
8257 mechanism allows a numeric "hash" to be extracted from a thread ID in
8258 either case, and on platforms where pointers are larger than 'long',
8259 mixing is done to help ensure the numeric 'hash' is usable even if it
8260 can't be guaranteed unique. The default mechanism is to use "&errno"
8261 as a pointer-based thread ID to distinguish between threads.
8263 Applications that want to provide their own thread IDs should now use
8264 CRYPTO_THREADID_set_callback() to register a callback that will call
8265 either CRYPTO_THREADID_set_numeric() or CRYPTO_THREADID_set_pointer().
8267 Note that ERR_remove_state() is now deprecated, because it is tied
8268 to the assumption that thread IDs are numeric. ERR_remove_state(0)
8269 to free the current thread's error state should be replaced by
8270 ERR_remove_thread_state(NULL).
8272 (This new approach replaces the functions CRYPTO_set_idptr_callback(),
8273 CRYPTO_get_idptr_callback(), and CRYPTO_thread_idptr() that existed in
8274 OpenSSL 0.9.9-dev between June 2006 and August 2008. Also, if an
8275 application was previously providing a numeric thread callback that
8276 was inappropriate for distinguishing threads, then uniqueness might
8277 have been obtained with &errno that happened immediately in the
8278 intermediate development versions of OpenSSL; this is no longer the
8279 case, the numeric thread callback will now override the automatic use
8282 *Geoff Thorpe, with help from Bodo Moeller*
8284 * Initial support for different CRL issuing certificates. This covers a
8285 simple case where the self issued certificates in the chain exist and
8286 the real CRL issuer is higher in the existing chain.
8288 This work was sponsored by Google.
8292 * Removed effectively defunct crypto/store from the build.
8296 * Revamp of STACK to provide stronger type-checking. Still to come:
8297 TXT_DB, bsearch(?), OBJ_bsearch, qsort, CRYPTO_EX_DATA, ASN1_VALUE,
8298 ASN1_STRING, CONF_VALUE.
8302 * Add a new SSL_MODE_RELEASE_BUFFERS mode flag to release unused buffer
8303 RAM on SSL connections. This option can save about 34k per idle SSL.
8307 * Revamp of LHASH to provide stronger type-checking. Still to come:
8308 STACK, TXT_DB, bsearch, qsort.
8312 * Initial support for Cryptographic Message Syntax (aka CMS) based
8313 on RFC3850, RFC3851 and RFC3852. New cms directory and cms utility,
8314 support for data, signedData, compressedData, digestedData and
8315 encryptedData, envelopedData types included. Scripts to check against
8316 RFC4134 examples draft and interop and consistency checks of many
8317 content types and variants.
8321 * Add options to enc utility to support use of zlib compression BIO.
8325 * Extend mk1mf to support importing of options and assembly language
8326 files from Configure script, currently only included in VC-WIN32.
8327 The assembly language rules can now optionally generate the source
8328 files from the associated perl scripts.
8332 * Implement remaining functionality needed to support GOST ciphersuites.
8333 Interop testing has been performed using CryptoPro implementations.
8335 *Victor B. Wagner <vitus@cryptocom.ru>*
8337 * s390x assembler pack.
8341 * ARMv4 assembler pack. ARMv4 refers to v4 and later ISA, not CPU
8346 * Implement Opaque PRF Input TLS extension as specified in
8347 draft-rescorla-tls-opaque-prf-input-00.txt. Since this is not an
8348 official specification yet and no extension type assignment by
8349 IANA exists, this extension (for now) will have to be explicitly
8350 enabled when building OpenSSL by providing the extension number
8351 to use. For example, specify an option
8353 -DTLSEXT_TYPE_opaque_prf_input=0x9527
8355 to the "config" or "Configure" script to enable the extension,
8356 assuming extension number 0x9527 (which is a completely arbitrary
8357 and unofficial assignment based on the MD5 hash of the Internet
8358 Draft). Note that by doing so, you potentially lose
8359 interoperability with other TLS implementations since these might
8360 be using the same extension number for other purposes.
8362 SSL_set_tlsext_opaque_prf_input(ssl, src, len) is used to set the
8363 opaque PRF input value to use in the handshake. This will create
8364 an internal copy of the length-'len' string at 'src', and will
8365 return non-zero for success.
8367 To get more control and flexibility, provide a callback function
8370 SSL_CTX_set_tlsext_opaque_prf_input_callback(ctx, cb)
8371 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg(ctx, arg)
8375 int (*cb)(SSL *, void *peerinput, size_t len, void *arg);
8378 Callback function 'cb' will be called in handshakes, and is
8379 expected to use SSL_set_tlsext_opaque_prf_input() as appropriate.
8380 Argument 'arg' is for application purposes (the value as given to
8381 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg() will directly
8382 be provided to the callback function). The callback function
8383 has to return non-zero to report success: usually 1 to use opaque
8384 PRF input just if possible, or 2 to enforce use of the opaque PRF
8385 input. In the latter case, the library will abort the handshake
8386 if opaque PRF input is not successfully negotiated.
8388 Arguments 'peerinput' and 'len' given to the callback function
8389 will always be NULL and 0 in the case of a client. A server will
8390 see the client's opaque PRF input through these variables if
8391 available (NULL and 0 otherwise). Note that if the server
8392 provides an opaque PRF input, the length must be the same as the
8393 length of the client's opaque PRF input.
8395 Note that the callback function will only be called when creating
8396 a new session (session resumption can resume whatever was
8397 previously negotiated), and will not be called in SSL 2.0
8398 handshakes; thus, SSL_CTX_set_options(ctx, SSL_OP_NO_SSLv2) or
8399 SSL_set_options(ssl, SSL_OP_NO_SSLv2) is especially recommended
8400 for applications that need to enforce opaque PRF input.
8404 * Update ssl code to support digests other than SHA1+MD5 for handshake
8407 *Victor B. Wagner <vitus@cryptocom.ru>*
8409 * Add RFC4507 support to OpenSSL. This includes the corrections in
8410 RFC4507bis. The encrypted ticket format is an encrypted encoded
8411 SSL_SESSION structure, that way new session features are automatically
8414 If a client application caches session in an SSL_SESSION structure
8415 support is transparent because tickets are now stored in the encoded
8418 The SSL_CTX structure automatically generates keys for ticket
8419 protection in servers so again support should be possible
8420 with no application modification.
8422 If a client or server wishes to disable RFC4507 support then the option
8423 SSL_OP_NO_TICKET can be set.
8425 Add a TLS extension debugging callback to allow the contents of any client
8426 or server extensions to be examined.
8428 This work was sponsored by Google.
8432 * Final changes to avoid use of pointer pointer casts in OpenSSL.
8433 OpenSSL should now compile cleanly on gcc 4.2
8435 *Peter Hartley <pdh@utter.chaos.org.uk>, Steve Henson*
8437 * Update SSL library to use new EVP_PKEY MAC API. Include generic MAC
8438 support including streaming MAC support: this is required for GOST
8439 ciphersuite support.
8441 *Victor B. Wagner <vitus@cryptocom.ru>, Steve Henson*
8443 * Add option -stream to use PKCS#7 streaming in smime utility. New
8444 function i2d_PKCS7_bio_stream() and PEM_write_PKCS7_bio_stream()
8445 to output in BER and PEM format.
8449 * Experimental support for use of HMAC via EVP_PKEY interface. This
8450 allows HMAC to be handled via the `EVP_DigestSign*()` interface. The
8451 EVP_PKEY "key" in this case is the HMAC key, potentially allowing
8452 ENGINE support for HMAC keys which are unextractable. New -mac and
8453 -macopt options to dgst utility.
8457 * New option -sigopt to dgst utility. Update dgst to use
8458 `EVP_Digest{Sign,Verify}*`. These two changes make it possible to use
8459 alternative signing parameters such as X9.31 or PSS in the dgst
8464 * Change ssl_cipher_apply_rule(), the internal function that does
8465 the work each time a ciphersuite string requests enabling
8466 ("foo+bar"), moving ("+foo+bar"), disabling ("-foo+bar", or
8467 removing ("!foo+bar") a class of ciphersuites: Now it maintains
8468 the order of disabled ciphersuites such that those ciphersuites
8469 that most recently went from enabled to disabled not only stay
8470 in order with respect to each other, but also have higher priority
8471 than other disabled ciphersuites the next time ciphersuites are
8474 This means that you can now say, e.g., "PSK:-PSK:HIGH" to enable
8475 the same ciphersuites as with "HIGH" alone, but in a specific
8476 order where the PSK ciphersuites come first (since they are the
8477 most recently disabled ciphersuites when "HIGH" is parsed).
8479 Also, change ssl_create_cipher_list() (using this new
8480 functionality) such that between otherwise identical
8481 ciphersuites, ephemeral ECDH is preferred over ephemeral DH in
8486 * Change ssl_create_cipher_list() so that it automatically
8487 arranges the ciphersuites in reasonable order before starting
8488 to process the rule string. Thus, the definition for "DEFAULT"
8489 (SSL_DEFAULT_CIPHER_LIST) now is just "ALL:!aNULL:!eNULL", but
8490 remains equivalent to `"AES:ALL:!aNULL:!eNULL:+aECDH:+kRSA:+RC4:@STRENGTH"`.
8491 This makes it much easier to arrive at a reasonable default order
8492 in applications for which anonymous ciphers are OK (meaning
8493 that you can't actually use DEFAULT).
8495 *Bodo Moeller; suggested by Victor Duchovni*
8497 * Split the SSL/TLS algorithm mask (as used for ciphersuite string
8498 processing) into multiple integers instead of setting
8499 "SSL_MKEY_MASK" bits, "SSL_AUTH_MASK" bits, "SSL_ENC_MASK",
8500 "SSL_MAC_MASK", and "SSL_SSL_MASK" bits all in a single integer.
8501 (These masks as well as the individual bit definitions are hidden
8502 away into the non-exported interface ssl/ssl_locl.h, so this
8503 change to the definition of the SSL_CIPHER structure shouldn't
8504 affect applications.) This give us more bits for each of these
8505 categories, so there is no longer a need to coagulate AES128 and
8506 AES256 into a single algorithm bit, and to coagulate Camellia128
8507 and Camellia256 into a single algorithm bit, which has led to all
8510 Thus, among other things, the kludge introduced in 0.9.7m and
8511 0.9.8e for masking out AES256 independently of AES128 or masking
8512 out Camellia256 independently of AES256 is not needed here in 0.9.9.
8514 With the change, we also introduce new ciphersuite aliases that
8515 so far were missing: "AES128", "AES256", "CAMELLIA128", and
8520 * Add support for dsa-with-SHA224 and dsa-with-SHA256.
8521 Use the leftmost N bytes of the signature input if the input is
8522 larger than the prime q (with N being the size in bytes of q).
8526 * Very *very* experimental PKCS#7 streaming encoder support. Nothing uses
8527 it yet and it is largely untested.
8531 * Add support for the ecdsa-with-SHA224/256/384/512 signature types.
8535 * Initial incomplete changes to avoid need for function casts in OpenSSL
8536 some compilers (gcc 4.2 and later) reject their use. Safestack is
8537 reimplemented. Update ASN1 to avoid use of legacy functions.
8541 * Win32/64 targets are linked with Winsock2.
8545 * Add an X509_CRL_METHOD structure to allow CRL processing to be redirected
8546 to external functions. This can be used to increase CRL handling
8547 efficiency especially when CRLs are very large by (for example) storing
8548 the CRL revoked certificates in a database.
8552 * Overhaul of by_dir code. Add support for dynamic loading of CRLs so
8553 new CRLs added to a directory can be used. New command line option
8554 -verify_return_error to s_client and s_server. This causes real errors
8555 to be returned by the verify callback instead of carrying on no matter
8556 what. This reflects the way a "real world" verify callback would behave.
8560 * GOST engine, supporting several GOST algorithms and public key formats.
8561 Kindly donated by Cryptocom.
8565 * Partial support for Issuing Distribution Point CRL extension. CRLs
8566 partitioned by DP are handled but no indirect CRL or reason partitioning
8567 (yet). Complete overhaul of CRL handling: now the most suitable CRL is
8568 selected via a scoring technique which handles IDP and AKID in CRLs.
8572 * New X509_STORE_CTX callbacks lookup_crls() and lookup_certs() which
8573 will ultimately be used for all verify operations: this will remove the
8574 X509_STORE dependency on certificate verification and allow alternative
8575 lookup methods. X509_STORE based implementations of these two callbacks.
8579 * Allow multiple CRLs to exist in an X509_STORE with matching issuer names.
8580 Modify get_crl() to find a valid (unexpired) CRL if possible.
8584 * New function X509_CRL_match() to check if two CRLs are identical. Normally
8585 this would be called X509_CRL_cmp() but that name is already used by
8586 a function that just compares CRL issuer names. Cache several CRL
8587 extensions in X509_CRL structure and cache CRLDP in X509.
8591 * Store a "canonical" representation of X509_NAME structure (ASN1 Name)
8592 this maps equivalent X509_NAME structures into a consistent structure.
8593 Name comparison can then be performed rapidly using memcmp().
8597 * Non-blocking OCSP request processing. Add -timeout option to ocsp
8602 * Allow digests to supply their own micalg string for S/MIME type using
8603 the ctrl EVP_MD_CTRL_MICALG.
8607 * During PKCS7 signing pass the PKCS7 SignerInfo structure to the
8608 EVP_PKEY_METHOD before and after signing via the EVP_PKEY_CTRL_PKCS7_SIGN
8609 ctrl. It can then customise the structure before and/or after signing
8614 * New function OBJ_add_sigid() to allow application defined signature OIDs
8615 to be added to OpenSSLs internal tables. New function OBJ_sigid_free()
8616 to free up any added signature OIDs.
8620 * New functions EVP_CIPHER_do_all(), EVP_CIPHER_do_all_sorted(),
8621 EVP_MD_do_all() and EVP_MD_do_all_sorted() to enumerate internal
8622 digest and cipher tables. New options added to openssl utility:
8623 list-message-digest-algorithms and list-cipher-algorithms.
8627 * Change the array representation of binary polynomials: the list
8628 of degrees of non-zero coefficients is now terminated with -1.
8629 Previously it was terminated with 0, which was also part of the
8630 value; thus, the array representation was not applicable to
8631 polynomials where t^0 has coefficient zero. This change makes
8632 the array representation useful in a more general context.
8636 * Various modifications and fixes to SSL/TLS cipher string
8637 handling. For ECC, the code now distinguishes between fixed ECDH
8638 with RSA certificates on the one hand and with ECDSA certificates
8639 on the other hand, since these are separate ciphersuites. The
8640 unused code for Fortezza ciphersuites has been removed.
8642 For consistency with EDH, ephemeral ECDH is now called "EECDH"
8643 (not "ECDHE"). For consistency with the code for DH
8644 certificates, use of ECDH certificates is now considered ECDH
8645 authentication, not RSA or ECDSA authentication (the latter is
8646 merely the CA's signing algorithm and not actively used in the
8649 The temporary ciphersuite alias "ECCdraft" is no longer
8650 available, and ECC ciphersuites are no longer excluded from "ALL"
8651 and "DEFAULT". The following aliases now exist for RFC 4492
8652 ciphersuites, most of these by analogy with the DH case:
8654 kECDHr - ECDH cert, signed with RSA
8655 kECDHe - ECDH cert, signed with ECDSA
8656 kECDH - ECDH cert (signed with either RSA or ECDSA)
8657 kEECDH - ephemeral ECDH
8658 ECDH - ECDH cert or ephemeral ECDH
8664 AECDH - anonymous ECDH
8665 EECDH - non-anonymous ephemeral ECDH (equivalent to "kEECDH:-AECDH")
8669 * Add additional S/MIME capabilities for AES and GOST ciphers if supported.
8670 Use correct micalg parameters depending on digest(s) in signed message.
8674 * Add engine support for EVP_PKEY_ASN1_METHOD. Add functions to process
8675 an ENGINE asn1 method. Support ENGINE lookups in the ASN1 code.
8679 * Initial engine support for EVP_PKEY_METHOD. New functions to permit
8680 an engine to register a method. Add ENGINE lookups for methods and
8681 functional reference processing.
8685 * New functions `EVP_Digest{Sign,Verify)*`. These are enhanced versions of
8686 `EVP_{Sign,Verify}*` which allow an application to customise the signature
8691 * New -resign option to smime utility. This adds one or more signers
8692 to an existing PKCS#7 signedData structure. Also -md option to use an
8693 alternative message digest algorithm for signing.
8697 * Tidy up PKCS#7 routines and add new functions to make it easier to
8698 create PKCS7 structures containing multiple signers. Update smime
8699 application to support multiple signers.
8703 * New -macalg option to pkcs12 utility to allow setting of an alternative
8708 * Initial support for PKCS#5 v2.0 PRFs other than default SHA1 HMAC.
8709 Reorganize PBE internals to lookup from a static table using NIDs,
8710 add support for HMAC PBE OID translation. Add a EVP_CIPHER ctrl:
8711 EVP_CTRL_PBE_PRF_NID this allows a cipher to specify an alternative
8712 PRF which will be automatically used with PBES2.
8716 * Replace the algorithm specific calls to generate keys in "req" with the
8721 * Update PKCS#7 enveloped data routines to use new API. This is now
8722 supported by any public key method supporting the encrypt operation. A
8723 ctrl is added to allow the public key algorithm to examine or modify
8724 the PKCS#7 RecipientInfo structure if it needs to: for RSA this is
8729 * Add a ctrl to asn1 method to allow a public key algorithm to express
8730 a default digest type to use. In most cases this will be SHA1 but some
8731 algorithms (such as GOST) need to specify an alternative digest. The
8732 return value indicates how strong the preference is 1 means optional and
8733 2 is mandatory (that is it is the only supported type). Modify
8734 ASN1_item_sign() to accept a NULL digest argument to indicate it should
8735 use the default md. Update openssl utilities to use the default digest
8736 type for signing if it is not explicitly indicated.
8740 * Use OID cross reference table in ASN1_sign() and ASN1_verify(). New
8741 EVP_MD flag EVP_MD_FLAG_PKEY_METHOD_SIGNATURE. This uses the relevant
8742 signing method from the key type. This effectively removes the link
8743 between digests and public key types.
8747 * Add an OID cross reference table and utility functions. Its purpose is to
8748 translate between signature OIDs such as SHA1WithrsaEncryption and SHA1,
8749 rsaEncryption. This will allow some of the algorithm specific hackery
8750 needed to use the correct OID to be removed.
8754 * Remove algorithm specific dependencies when setting PKCS7_SIGNER_INFO
8755 structures for PKCS7_sign(). They are now set up by the relevant public
8760 * Add provisional EC pkey method with support for ECDSA and ECDH.
8764 * Add support for key derivation (agreement) in the API, DH method and
8769 * Add DSA pkey method and DH pkey methods, extend DH ASN1 method to support
8770 public and private key formats. As a side effect these add additional
8771 command line functionality not previously available: DSA signatures can be
8772 generated and verified using pkeyutl and DH key support and generation in
8779 *Oliver Tappe <zooey@hirschkaefer.de>*
8781 * New make target "install_html_docs" installs HTML renditions of the
8784 *Oliver Tappe <zooey@hirschkaefer.de>*
8786 * New utility "genpkey" this is analogous to "genrsa" etc except it can
8787 generate keys for any algorithm. Extend and update EVP_PKEY_METHOD to
8788 support key and parameter generation and add initial key generation
8789 functionality for RSA.
8793 * Add functions for main EVP_PKEY_method operations. The undocumented
8794 functions `EVP_PKEY_{encrypt,decrypt}` have been renamed to
8795 `EVP_PKEY_{encrypt,decrypt}_old`.
8799 * Initial definitions for EVP_PKEY_METHOD. This will be a high level public
8800 key API, doesn't do much yet.
8804 * New function EVP_PKEY_asn1_get0_info() to retrieve information about
8805 public key algorithms. New option to openssl utility:
8806 "list-public-key-algorithms" to print out info.
8810 * Implement the Supported Elliptic Curves Extension for
8811 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
8815 * Don't free up OIDs in OBJ_cleanup() if they are in use by EVP_MD or
8816 EVP_CIPHER structures to avoid later problems in EVP_cleanup().
8820 * New utilities pkey and pkeyparam. These are similar to algorithm specific
8821 utilities such as rsa, dsa, dsaparam etc except they process any key
8826 * Transfer public key printing routines to EVP_PKEY_ASN1_METHOD. New
8827 functions EVP_PKEY_print_public(), EVP_PKEY_print_private(),
8828 EVP_PKEY_print_param() to print public key data from an EVP_PKEY
8833 * Initial support for pluggable public key ASN1.
8834 De-spaghettify the public key ASN1 handling. Move public and private
8835 key ASN1 handling to a new EVP_PKEY_ASN1_METHOD structure. Relocate
8836 algorithm specific handling to a single module within the relevant
8837 algorithm directory. Add functions to allow (near) opaque processing
8838 of public and private key structures.
8842 * Implement the Supported Point Formats Extension for
8843 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
8847 * Add initial support for RFC 4279 PSK TLS ciphersuites. Add members
8848 for the psk identity [hint] and the psk callback functions to the
8849 SSL_SESSION, SSL and SSL_CTX structure.
8852 PSK-RC4-SHA, PSK-3DES-EDE-CBC-SHA, PSK-AES128-CBC-SHA,
8856 SSL_CTX_use_psk_identity_hint
8857 SSL_get_psk_identity_hint
8858 SSL_get_psk_identity
8859 SSL_use_psk_identity_hint
8861 *Mika Kousa and Pasi Eronen of Nokia Corporation*
8863 * Add RFC 3161 compliant time stamp request creation, response generation
8864 and response verification functionality.
8866 *Zoltán Glózik <zglozik@opentsa.org>, The OpenTSA Project*
8868 * Add initial support for TLS extensions, specifically for the server_name
8869 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
8870 have new members for a host name. The SSL data structure has an
8871 additional member `SSL_CTX *initial_ctx` so that new sessions can be
8872 stored in that context to allow for session resumption, even after the
8873 SSL has been switched to a new SSL_CTX in reaction to a client's
8874 server_name extension.
8876 New functions (subject to change):
8878 SSL_get_servername()
8879 SSL_get_servername_type()
8882 New CTRL codes and macros (subject to change):
8884 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
8885 - SSL_CTX_set_tlsext_servername_callback()
8886 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
8887 - SSL_CTX_set_tlsext_servername_arg()
8888 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
8890 openssl s_client has a new '-servername ...' option.
8892 openssl s_server has new options '-servername_host ...', '-cert2 ...',
8893 '-key2 ...', '-servername_fatal' (subject to change). This allows
8894 testing the HostName extension for a specific single host name ('-cert'
8895 and '-key' remain fallbacks for handshakes without HostName
8896 negotiation). If the unrecognized_name alert has to be sent, this by
8897 default is a warning; it becomes fatal with the '-servername_fatal'
8900 *Peter Sylvester, Remy Allais, Christophe Renou*
8902 * Whirlpool hash implementation is added.
8906 * BIGNUM code on 64-bit SPARCv9 targets is switched from bn(64,64) to
8907 bn(64,32). Because of instruction set limitations it doesn't have
8908 any negative impact on performance. This was done mostly in order
8909 to make it possible to share assembler modules, such as bn_mul_mont
8910 implementations, between 32- and 64-bit builds without hassle.
8914 * Move code previously exiled into file crypto/ec/ec2_smpt.c
8915 to ec2_smpl.c, and no longer require the OPENSSL_EC_BIN_PT_COMP
8920 * New candidate for BIGNUM assembler implementation, bn_mul_mont,
8921 dedicated Montgomery multiplication procedure, is introduced.
8922 BN_MONT_CTX is modified to allow bn_mul_mont to reach for higher
8923 "64-bit" performance on certain 32-bit targets.
8927 * New option SSL_OP_NO_COMP to disable use of compression selectively
8928 in SSL structures. New SSL ctrl to set maximum send fragment size.
8929 Save memory by setting the I/O buffer sizes dynamically instead of
8930 using the maximum available value.
8934 * New option -V for 'openssl ciphers'. This prints the ciphersuite code
8935 in addition to the text details.
8939 * Very, very preliminary EXPERIMENTAL support for printing of general
8940 ASN1 structures. This currently produces rather ugly output and doesn't
8941 handle several customised structures at all.
8945 * Integrated support for PVK file format and some related formats such
8946 as MS PUBLICKEYBLOB and PRIVATEKEYBLOB. Command line switches to support
8947 these in the 'rsa' and 'dsa' utilities.
8951 * Support for PKCS#1 RSAPublicKey format on rsa utility command line.
8955 * Remove the ancient ASN1_METHOD code. This was only ever used in one
8956 place for the (very old) "NETSCAPE" format certificates which are now
8957 handled using new ASN1 code equivalents.
8961 * Let the TLSv1_method() etc. functions return a 'const' SSL_METHOD
8962 pointer and make the SSL_METHOD parameter in SSL_CTX_new,
8963 SSL_CTX_set_ssl_version and SSL_set_ssl_method 'const'.
8967 * Modify CRL distribution points extension code to print out previously
8968 unsupported fields. Enhance extension setting code to allow setting of
8973 * Add print and set support for Issuing Distribution Point CRL extension.
8977 * Change 'Configure' script to enable Camellia by default.
8984 ### Changes between 0.9.8m and 0.9.8n [24 Mar 2010]
8986 * When rejecting SSL/TLS records due to an incorrect version number, never
8987 update s->server with a new major version number. As of
8988 - OpenSSL 0.9.8m if 'short' is a 16-bit type,
8989 - OpenSSL 0.9.8f if 'short' is longer than 16 bits,
8990 the previous behavior could result in a read attempt at NULL when
8991 receiving specific incorrect SSL/TLS records once record payload
8992 protection is active. ([CVE-2010-0740])
8994 *Bodo Moeller, Adam Langley <agl@chromium.org>*
8996 * Fix for CVE-2010-0433 where some kerberos enabled versions of OpenSSL
8997 could be crashed if the relevant tables were not present (e.g. chrooted).
8999 *Tomas Hoger <thoger@redhat.com>*
9001 ### Changes between 0.9.8l and 0.9.8m [25 Feb 2010]
9003 * Always check bn_wexpand() return values for failure. ([CVE-2009-3245])
9005 *Martin Olsson, Neel Mehta*
9007 * Fix X509_STORE locking: Every 'objs' access requires a lock (to
9008 accommodate for stack sorting, always a write lock!).
9012 * On some versions of WIN32 Heap32Next is very slow. This can cause
9013 excessive delays in the RAND_poll(): over a minute. As a workaround
9014 include a time check in the inner Heap32Next loop too.
9018 * The code that handled flushing of data in SSL/TLS originally used the
9019 BIO_CTRL_INFO ctrl to see if any data was pending first. This caused
9020 the problem outlined in PR#1949. The fix suggested there however can
9021 trigger problems with buggy BIO_CTRL_WPENDING (e.g. some versions
9022 of Apache). So instead simplify the code to flush unconditionally.
9023 This should be fine since flushing with no data to flush is a no op.
9027 * Handle TLS versions 2.0 and later properly and correctly use the
9028 highest version of TLS/SSL supported. Although TLS >= 2.0 is some way
9029 off ancient servers have a habit of sticking around for a while...
9033 * Modify compression code so it frees up structures without using the
9034 ex_data callbacks. This works around a problem where some applications
9035 call CRYPTO_cleanup_all_ex_data() before application exit (e.g. when
9036 restarting) then use compression (e.g. SSL with compression) later.
9037 This results in significant per-connection memory leaks and
9038 has caused some security issues including CVE-2008-1678 and
9043 * Constify crypto/cast (i.e., <openssl/cast.h>): a CAST_KEY doesn't
9044 change when encrypting or decrypting.
9048 * Add option SSL_OP_LEGACY_SERVER_CONNECT which will allow clients to
9049 connect and renegotiate with servers which do not support RI.
9050 Until RI is more widely deployed this option is enabled by default.
9054 * Add "missing" ssl ctrls to clear options and mode.
9058 * If client attempts to renegotiate and doesn't support RI respond with
9059 a no_renegotiation alert as required by RFC5746. Some renegotiating
9060 TLS clients will continue a connection gracefully when they receive
9061 the alert. Unfortunately OpenSSL mishandled this alert and would hang
9062 waiting for a server hello which it will never receive. Now we treat a
9063 received no_renegotiation alert as a fatal error. This is because
9064 applications requesting a renegotiation might well expect it to succeed
9065 and would have no code in place to handle the server denying it so the
9066 only safe thing to do is to terminate the connection.
9070 * Add ctrl macro SSL_get_secure_renegotiation_support() which returns 1 if
9071 peer supports secure renegotiation and 0 otherwise. Print out peer
9072 renegotiation support in s_client/s_server.
9076 * Replace the highly broken and deprecated SPKAC certification method with
9077 the updated NID creation version. This should correctly handle UTF8.
9081 * Implement RFC5746. Re-enable renegotiation but require the extension
9082 as needed. Unfortunately, SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION
9083 turns out to be a bad idea. It has been replaced by
9084 SSL_OP_ALLOW_UNSAFE_LEGACY_RENEGOTIATION which can be set with
9085 SSL_CTX_set_options(). This is really not recommended unless you
9086 know what you are doing.
9088 *Eric Rescorla <ekr@networkresonance.com>, Ben Laurie, Steve Henson*
9090 * Fixes to stateless session resumption handling. Use initial_ctx when
9091 issuing and attempting to decrypt tickets in case it has changed during
9092 servername handling. Use a non-zero length session ID when attempting
9093 stateless session resumption: this makes it possible to determine if
9094 a resumption has occurred immediately after receiving server hello
9095 (several places in OpenSSL subtly assume this) instead of later in
9100 * The functions ENGINE_ctrl(), OPENSSL_isservice(),
9101 CMS_get1_RecipientRequest() and RAND_bytes() can return <=0 on error
9102 fixes for a few places where the return code is not checked
9105 *Julia Lawall <julia@diku.dk>*
9107 * Add --strict-warnings option to Configure script to include devteam
9108 warnings in other configurations.
9112 * Add support for --libdir option and LIBDIR variable in makefiles. This
9113 makes it possible to install openssl libraries in locations which
9114 have names other than "lib", for example "/usr/lib64" which some
9117 *Steve Henson, based on patch from Jeremy Utley*
9119 * Don't allow the use of leading 0x80 in OIDs. This is a violation of
9120 X690 8.9.12 and can produce some misleading textual output of OIDs.
9122 *Steve Henson, reported by Dan Kaminsky*
9124 * Delete MD2 from algorithm tables. This follows the recommendation in
9125 several standards that it is not used in new applications due to
9126 several cryptographic weaknesses. For binary compatibility reasons
9127 the MD2 API is still compiled in by default.
9131 * Add compression id to {d2i,i2d}_SSL_SESSION so it is correctly saved
9136 * Rename uni2asc and asc2uni functions to OPENSSL_uni2asc and
9137 OPENSSL_asc2uni conditionally on Netware platforms to avoid a name
9140 *Guenter <lists@gknw.net>*
9142 * Fix the server certificate chain building code to use X509_verify_cert(),
9143 it used to have an ad-hoc builder which was unable to cope with anything
9144 other than a simple chain.
9146 *David Woodhouse <dwmw2@infradead.org>, Steve Henson*
9148 * Don't check self signed certificate signatures in X509_verify_cert()
9149 by default (a flag can override this): it just wastes time without
9150 adding any security. As a useful side effect self signed root CAs
9151 with non-FIPS digests are now usable in FIPS mode.
9155 * In dtls1_process_out_of_seq_message() the check if the current message
9156 is already buffered was missing. For every new message was memory
9157 allocated, allowing an attacker to perform an denial of service attack
9158 with sending out of seq handshake messages until there is no memory
9159 left. Additionally every future message was buffered, even if the
9160 sequence number made no sense and would be part of another handshake.
9161 So only messages with sequence numbers less than 10 in advance will be
9162 buffered. ([CVE-2009-1378])
9164 *Robin Seggelmann, discovered by Daniel Mentz*
9166 * Records are buffered if they arrive with a future epoch to be
9167 processed after finishing the corresponding handshake. There is
9168 currently no limitation to this buffer allowing an attacker to perform
9169 a DOS attack with sending records with future epochs until there is no
9170 memory left. This patch adds the pqueue_size() function to determine
9171 the size of a buffer and limits the record buffer to 100 entries.
9174 *Robin Seggelmann, discovered by Daniel Mentz*
9176 * Keep a copy of frag->msg_header.frag_len so it can be used after the
9177 parent structure is freed. ([CVE-2009-1379])
9181 * Handle non-blocking I/O properly in SSL_shutdown() call.
9183 *Darryl Miles <darryl-mailinglists@netbauds.net>*
9185 * Add `2.5.4.*` OIDs
9187 *Ilya O. <vrghost@gmail.com>*
9189 ### Changes between 0.9.8k and 0.9.8l [5 Nov 2009]
9191 * Disable renegotiation completely - this fixes a severe security
9192 problem ([CVE-2009-3555]) at the cost of breaking all
9193 renegotiation. Renegotiation can be re-enabled by setting
9194 SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION in s3->flags at
9195 run-time. This is really not recommended unless you know what
9200 ### Changes between 0.9.8j and 0.9.8k [25 Mar 2009]
9202 * Don't set val to NULL when freeing up structures, it is freed up by
9203 underlying code. If `sizeof(void *) > sizeof(long)` this can result in
9204 zeroing past the valid field. ([CVE-2009-0789])
9206 *Paolo Ganci <Paolo.Ganci@AdNovum.CH>*
9208 * Fix bug where return value of CMS_SignerInfo_verify_content() was not
9209 checked correctly. This would allow some invalid signed attributes to
9210 appear to verify correctly. ([CVE-2009-0591])
9212 *Ivan Nestlerode <inestlerode@us.ibm.com>*
9214 * Reject UniversalString and BMPString types with invalid lengths. This
9215 prevents a crash in ASN1_STRING_print_ex() which assumes the strings have
9216 a legal length. ([CVE-2009-0590])
9220 * Set S/MIME signing as the default purpose rather than setting it
9221 unconditionally. This allows applications to override it at the store
9226 * Permit restricted recursion of ASN1 strings. This is needed in practice
9227 to handle some structures.
9231 * Improve efficiency of mem_gets: don't search whole buffer each time
9234 *Jeremy Shapiro <jnshapir@us.ibm.com>*
9236 * New -hex option for openssl rand.
9240 * Print out UTF8String and NumericString when parsing ASN1.
9244 * Support NumericString type for name components.
9248 * Allow CC in the environment to override the automatically chosen
9249 compiler. Note that nothing is done to ensure flags work with the
9254 ### Changes between 0.9.8i and 0.9.8j [07 Jan 2009]
9256 * Properly check EVP_VerifyFinal() and similar return values
9259 *Ben Laurie, Bodo Moeller, Google Security Team*
9261 * Enable TLS extensions by default.
9265 * Allow the CHIL engine to be loaded, whether the application is
9266 multithreaded or not. (This does not release the developer from the
9267 obligation to set up the dynamic locking callbacks.)
9269 *Sander Temme <sander@temme.net>*
9271 * Use correct exit code if there is an error in dgst command.
9273 *Steve Henson; problem pointed out by Roland Dirlewanger*
9275 * Tweak Configure so that you need to say "experimental-jpake" to enable
9276 JPAKE, and need to use -DOPENSSL_EXPERIMENTAL_JPAKE in applications.
9280 * Add experimental JPAKE support, including demo authentication in
9281 s_client and s_server.
9285 * Set the comparison function in v3_addr_canonize().
9287 *Rob Austein <sra@hactrn.net>*
9289 * Add support for XMPP STARTTLS in s_client.
9291 *Philip Paeps <philip@freebsd.org>*
9293 * Change the server-side SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG behavior
9294 to ensure that even with this option, only ciphersuites in the
9295 server's preference list will be accepted. (Note that the option
9296 applies only when resuming a session, so the earlier behavior was
9297 just about the algorithm choice for symmetric cryptography.)
9301 ### Changes between 0.9.8h and 0.9.8i [15 Sep 2008]
9303 * Fix NULL pointer dereference if a DTLS server received
9304 ChangeCipherSpec as first record ([CVE-2009-1386]).
9308 * Fix a state transition in s3_srvr.c and d1_srvr.c
9309 (was using SSL3_ST_CW_CLNT_HELLO_B, should be `..._ST_SW_SRVR_...`).
9313 * The fix in 0.9.8c that supposedly got rid of unsafe
9314 double-checked locking was incomplete for RSA blinding,
9315 addressing just one layer of what turns out to have been
9316 doubly unsafe triple-checked locking.
9318 So now fix this for real by retiring the MONT_HELPER macro
9319 in crypto/rsa/rsa_eay.c.
9321 *Bodo Moeller; problem pointed out by Marius Schilder*
9323 * Various precautionary measures:
9325 - Avoid size_t integer overflow in HASH_UPDATE (md32_common.h).
9327 - Avoid a buffer overflow in d2i_SSL_SESSION() (ssl_asn1.c).
9328 (NB: This would require knowledge of the secret session ticket key
9329 to exploit, in which case you'd be SOL either way.)
9331 - Change bn_nist.c so that it will properly handle input BIGNUMs
9332 outside the expected range.
9334 - Enforce the 'num' check in BN_div() (bn_div.c) for non-BN_DEBUG
9337 *Neel Mehta, Bodo Moeller*
9339 * Allow engines to be "soft loaded" - i.e. optionally don't die if
9340 the load fails. Useful for distros.
9342 *Ben Laurie and the FreeBSD team*
9344 * Add support for Local Machine Keyset attribute in PKCS#12 files.
9348 * Fix BN_GF2m_mod_arr() top-bit cleanup code.
9352 * Expand ENGINE to support engine supplied SSL client certificate functions.
9354 This work was sponsored by Logica.
9358 * Add CryptoAPI ENGINE to support use of RSA and DSA keys held in Windows
9359 keystores. Support for SSL/TLS client authentication too.
9360 Not compiled unless enable-capieng specified to Configure.
9362 This work was sponsored by Logica.
9366 * Fix bug in X509_ATTRIBUTE creation: don't set attribute using
9367 ASN1_TYPE_set1 if MBSTRING flag set. This bug would crash certain
9368 attribute creation routines such as certificate requests and PKCS#12
9373 ### Changes between 0.9.8g and 0.9.8h [28 May 2008]
9375 * Fix flaw if 'Server Key exchange message' is omitted from a TLS
9376 handshake which could lead to a client crash as found using the
9377 Codenomicon TLS test suite ([CVE-2008-1672])
9379 *Steve Henson, Mark Cox*
9381 * Fix double free in TLS server name extensions which could lead to
9382 a remote crash found by Codenomicon TLS test suite ([CVE-2008-0891])
9386 * Clear error queue in SSL_CTX_use_certificate_chain_file()
9388 Clear the error queue to ensure that error entries left from
9389 older function calls do not interfere with the correct operation.
9391 *Lutz Jaenicke, Erik de Castro Lopo*
9393 * Remove root CA certificates of commercial CAs:
9395 The OpenSSL project does not recommend any specific CA and does not
9396 have any policy with respect to including or excluding any CA.
9397 Therefore it does not make any sense to ship an arbitrary selection
9398 of root CA certificates with the OpenSSL software.
9402 * RSA OAEP patches to fix two separate invalid memory reads.
9403 The first one involves inputs when 'lzero' is greater than
9404 'SHA_DIGEST_LENGTH' (it would read about SHA_DIGEST_LENGTH bytes
9405 before the beginning of from). The second one involves inputs where
9406 the 'db' section contains nothing but zeroes (there is a one-byte
9407 invalid read after the end of 'db').
9409 *Ivan Nestlerode <inestlerode@us.ibm.com>*
9411 * Partial backport from 0.9.9-dev:
9413 Introduce bn_mul_mont (dedicated Montgomery multiplication
9414 procedure) as a candidate for BIGNUM assembler implementation.
9415 While 0.9.9-dev uses assembler for various architectures, only
9416 x86_64 is available by default here in the 0.9.8 branch, and
9417 32-bit x86 is available through a compile-time setting.
9419 To try the 32-bit x86 assembler implementation, use Configure
9420 option "enable-montasm" (which exists only for this backport).
9422 As "enable-montasm" for 32-bit x86 disclaims code stability
9423 anyway, in this constellation we activate additional code
9424 backported from 0.9.9-dev for further performance improvements,
9425 namely BN_from_montgomery_word. (To enable this otherwise,
9426 e.g. x86_64, try `-DMONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD`.)
9428 *Andy Polyakov (backport partially by Bodo Moeller)*
9430 * Add TLS session ticket callback. This allows an application to set
9431 TLS ticket cipher and HMAC keys rather than relying on hardcoded fixed
9432 values. This is useful for key rollover for example where several key
9433 sets may exist with different names.
9437 * Reverse ENGINE-internal logic for caching default ENGINE handles.
9438 This was broken until now in 0.9.8 releases, such that the only way
9439 a registered ENGINE could be used (assuming it initialises
9440 successfully on the host) was to explicitly set it as the default
9441 for the relevant algorithms. This is in contradiction with 0.9.7
9442 behaviour and the documentation. With this fix, when an ENGINE is
9443 registered into a given algorithm's table of implementations, the
9444 'uptodate' flag is reset so that auto-discovery will be used next
9445 time a new context for that algorithm attempts to select an
9448 *Ian Lister (tweaked by Geoff Thorpe)*
9450 * Backport of CMS code to OpenSSL 0.9.8. This differs from the 0.9.9
9451 implementation in the following ways:
9453 Lack of EVP_PKEY_ASN1_METHOD means algorithm parameters have to be
9456 Lack of BER streaming support means one pass streaming processing is
9457 only supported if data is detached: setting the streaming flag is
9458 ignored for embedded content.
9460 CMS support is disabled by default and must be explicitly enabled
9461 with the enable-cms configuration option.
9465 * Update the GMP engine glue to do direct copies between BIGNUM and
9466 mpz_t when openssl and GMP use the same limb size. Otherwise the
9467 existing "conversion via a text string export" trick is still used.
9469 *Paul Sheer <paulsheer@gmail.com>*
9471 * Zlib compression BIO. This is a filter BIO which compressed and
9472 uncompresses any data passed through it.
9476 * Add AES_wrap_key() and AES_unwrap_key() functions to implement
9477 RFC3394 compatible AES key wrapping.
9481 * Add utility functions to handle ASN1 structures. ASN1_STRING_set0():
9482 sets string data without copying. X509_ALGOR_set0() and
9483 X509_ALGOR_get0(): set and retrieve X509_ALGOR (AlgorithmIdentifier)
9484 data. Attribute function X509at_get0_data_by_OBJ(): retrieves data
9485 from an X509_ATTRIBUTE structure optionally checking it occurs only
9486 once. ASN1_TYPE_set1(): set and ASN1_TYPE structure copying supplied
9491 * Fix BN flag handling in RSA_eay_mod_exp() and BN_MONT_CTX_set()
9492 to get the expected BN_FLG_CONSTTIME behavior.
9494 *Bodo Moeller (Google)*
9498 - fixed wrong usage of ioctlsocket() when build for LIBC BSD sockets
9499 - fixed do_tests.pl to run the test suite with CLIB builds too (CLIB_OPT)
9500 - added some more tests to do_tests.pl
9501 - fixed RunningProcess usage so that it works with newer LIBC NDKs too
9502 - removed usage of BN_LLONG for CLIB builds to avoid runtime dependency
9503 - added new Configure targets netware-clib-bsdsock, netware-clib-gcc,
9504 netware-clib-bsdsock-gcc, netware-libc-bsdsock-gcc
9505 - various changes to netware.pl to enable gcc-cross builds on Win32
9507 - changed crypto/bio/b_sock.c to work with macro functions (CLIB BSD)
9508 - various changes to fix missing prototype warnings
9509 - fixed x86nasm.pl to create correct asm files for NASM COFF output
9510 - added AES, WHIRLPOOL and CPUID assembler code to build files
9511 - added missing AES assembler make rules to mk1mf.pl
9512 - fixed order of includes in `apps/ocsp.c` so that `e_os.h` settings apply
9514 *Guenter Knauf <eflash@gmx.net>*
9516 * Implement certificate status request TLS extension defined in RFC3546.
9517 A client can set the appropriate parameters and receive the encoded
9518 OCSP response via a callback. A server can query the supplied parameters
9519 and set the encoded OCSP response in the callback. Add simplified examples
9520 to s_client and s_server.
9524 ### Changes between 0.9.8f and 0.9.8g [19 Oct 2007]
9527 + Binary incompatibility of ssl_ctx_st structure
9528 + DTLS interoperation with non-compliant servers
9529 + Don't call get_session_cb() without proposed session
9530 + Fix ia64 assembler code
9532 *Andy Polyakov, Steve Henson*
9534 ### Changes between 0.9.8e and 0.9.8f [11 Oct 2007]
9536 * DTLS Handshake overhaul. There were longstanding issues with
9537 OpenSSL DTLS implementation, which were making it impossible for
9538 RFC 4347 compliant client to communicate with OpenSSL server.
9539 Unfortunately just fixing these incompatibilities would "cut off"
9540 pre-0.9.8f clients. To allow for hassle free upgrade post-0.9.8e
9541 server keeps tolerating non RFC compliant syntax. The opposite is
9542 not true, 0.9.8f client can not communicate with earlier server.
9543 This update even addresses CVE-2007-4995.
9547 * Changes to avoid need for function casts in OpenSSL: some compilers
9548 (gcc 4.2 and later) reject their use.
9549 *Kurt Roeckx <kurt@roeckx.be>, Peter Hartley <pdh@utter.chaos.org.uk>,
9552 * Add RFC4507 support to OpenSSL. This includes the corrections in
9553 RFC4507bis. The encrypted ticket format is an encrypted encoded
9554 SSL_SESSION structure, that way new session features are automatically
9557 If a client application caches session in an SSL_SESSION structure
9558 support is transparent because tickets are now stored in the encoded
9561 The SSL_CTX structure automatically generates keys for ticket
9562 protection in servers so again support should be possible
9563 with no application modification.
9565 If a client or server wishes to disable RFC4507 support then the option
9566 SSL_OP_NO_TICKET can be set.
9568 Add a TLS extension debugging callback to allow the contents of any client
9569 or server extensions to be examined.
9571 This work was sponsored by Google.
9575 * Add initial support for TLS extensions, specifically for the server_name
9576 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
9577 have new members for a host name. The SSL data structure has an
9578 additional member `SSL_CTX *initial_ctx` so that new sessions can be
9579 stored in that context to allow for session resumption, even after the
9580 SSL has been switched to a new SSL_CTX in reaction to a client's
9581 server_name extension.
9583 New functions (subject to change):
9585 SSL_get_servername()
9586 SSL_get_servername_type()
9589 New CTRL codes and macros (subject to change):
9591 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
9592 - SSL_CTX_set_tlsext_servername_callback()
9593 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
9594 - SSL_CTX_set_tlsext_servername_arg()
9595 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
9597 openssl s_client has a new '-servername ...' option.
9599 openssl s_server has new options '-servername_host ...', '-cert2 ...',
9600 '-key2 ...', '-servername_fatal' (subject to change). This allows
9601 testing the HostName extension for a specific single host name ('-cert'
9602 and '-key' remain fallbacks for handshakes without HostName
9603 negotiation). If the unrecognized_name alert has to be sent, this by
9604 default is a warning; it becomes fatal with the '-servername_fatal'
9607 *Peter Sylvester, Remy Allais, Christophe Renou, Steve Henson*
9609 * Add AES and SSE2 assembly language support to VC++ build.
9613 * Mitigate attack on final subtraction in Montgomery reduction.
9617 * Fix crypto/ec/ec_mult.c to work properly with scalars of value 0
9618 (which previously caused an internal error).
9622 * Squeeze another 10% out of IGE mode when in != out.
9626 * AES IGE mode speedup.
9628 *Dean Gaudet (Google)*
9630 * Add the Korean symmetric 128-bit cipher SEED (see
9631 <http://www.kisa.or.kr/kisa/seed/jsp/seed_eng.jsp>) and
9632 add SEED ciphersuites from RFC 4162:
9634 TLS_RSA_WITH_SEED_CBC_SHA = "SEED-SHA"
9635 TLS_DHE_DSS_WITH_SEED_CBC_SHA = "DHE-DSS-SEED-SHA"
9636 TLS_DHE_RSA_WITH_SEED_CBC_SHA = "DHE-RSA-SEED-SHA"
9637 TLS_DH_anon_WITH_SEED_CBC_SHA = "ADH-SEED-SHA"
9639 To minimize changes between patchlevels in the OpenSSL 0.9.8
9640 series, SEED remains excluded from compilation unless OpenSSL
9641 is configured with 'enable-seed'.
9643 *KISA, Bodo Moeller*
9645 * Mitigate branch prediction attacks, which can be practical if a
9646 single processor is shared, allowing a spy process to extract
9647 information. For detailed background information, see
9648 <http://eprint.iacr.org/2007/039> (O. Aciicmez, S. Gueron,
9649 J.-P. Seifert, "New Branch Prediction Vulnerabilities in OpenSSL
9650 and Necessary Software Countermeasures"). The core of the change
9651 are new versions BN_div_no_branch() and
9652 BN_mod_inverse_no_branch() of BN_div() and BN_mod_inverse(),
9653 respectively, which are slower, but avoid the security-relevant
9654 conditional branches. These are automatically called by BN_div()
9655 and BN_mod_inverse() if the flag BN_FLG_CONSTTIME is set for one
9656 of the input BIGNUMs. Also, BN_is_bit_set() has been changed to
9657 remove a conditional branch.
9659 BN_FLG_CONSTTIME is the new name for the previous
9660 BN_FLG_EXP_CONSTTIME flag, since it now affects more than just
9661 modular exponentiation. (Since OpenSSL 0.9.7h, setting this flag
9662 in the exponent causes BN_mod_exp_mont() to use the alternative
9663 implementation in BN_mod_exp_mont_consttime().) The old name
9664 remains as a deprecated alias.
9666 Similarly, RSA_FLAG_NO_EXP_CONSTTIME is replaced by a more general
9667 RSA_FLAG_NO_CONSTTIME flag since the RSA implementation now uses
9668 constant-time implementations for more than just exponentiation.
9669 Here too the old name is kept as a deprecated alias.
9671 BN_BLINDING_new() will now use BN_dup() for the modulus so that
9672 the BN_BLINDING structure gets an independent copy of the
9673 modulus. This means that the previous `BIGNUM *m` argument to
9674 BN_BLINDING_new() and to BN_BLINDING_create_param() now
9675 essentially becomes `const BIGNUM *m`, although we can't actually
9676 change this in the header file before 0.9.9. It allows
9677 RSA_setup_blinding() to use BN_with_flags() on the modulus to
9678 enable BN_FLG_CONSTTIME.
9680 *Matthew D Wood (Intel Corp)*
9682 * In the SSL/TLS server implementation, be strict about session ID
9683 context matching (which matters if an application uses a single
9684 external cache for different purposes). Previously,
9685 out-of-context reuse was forbidden only if SSL_VERIFY_PEER was
9686 set. This did ensure strict client verification, but meant that,
9687 with applications using a single external cache for quite
9688 different requirements, clients could circumvent ciphersuite
9689 restrictions for a given session ID context by starting a session
9690 in a different context.
9694 * Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
9695 a ciphersuite string such as "DEFAULT:RSA" cannot enable
9696 authentication-only ciphersuites.
9700 * Update the SSL_get_shared_ciphers() fix CVE-2006-3738 which was
9701 not complete and could lead to a possible single byte overflow
9702 ([CVE-2007-5135]) [Ben Laurie]
9704 ### Changes between 0.9.8d and 0.9.8e [23 Feb 2007]
9706 * Since AES128 and AES256 (and similarly Camellia128 and
9707 Camellia256) share a single mask bit in the logic of
9708 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
9709 kludge to work properly if AES128 is available and AES256 isn't
9710 (or if Camellia128 is available and Camellia256 isn't).
9714 * Fix the BIT STRING encoding generated by crypto/ec/ec_asn1.c
9715 (within i2d_ECPrivateKey, i2d_ECPKParameters, i2d_ECParameters):
9716 When a point or a seed is encoded in a BIT STRING, we need to
9717 prevent the removal of trailing zero bits to get the proper DER
9718 encoding. (By default, crypto/asn1/a_bitstr.c assumes the case
9719 of a NamedBitList, for which trailing 0 bits need to be removed.)
9723 * Have SSL/TLS server implementation tolerate "mismatched" record
9724 protocol version while receiving ClientHello even if the
9725 ClientHello is fragmented. (The server can't insist on the
9726 particular protocol version it has chosen before the ServerHello
9727 message has informed the client about his choice.)
9731 * Add RFC 3779 support.
9733 *Rob Austein for ARIN, Ben Laurie*
9735 * Load error codes if they are not already present instead of using a
9736 static variable. This allows them to be cleanly unloaded and reloaded.
9737 Improve header file function name parsing.
9741 * extend SMTP and IMAP protocol emulation in s_client to use EHLO
9742 or CAPABILITY handshake as required by RFCs.
9746 ### Changes between 0.9.8c and 0.9.8d [28 Sep 2006]
9748 * Introduce limits to prevent malicious keys being able to
9749 cause a denial of service. ([CVE-2006-2940])
9751 *Steve Henson, Bodo Moeller*
9753 * Fix ASN.1 parsing of certain invalid structures that can result
9754 in a denial of service. ([CVE-2006-2937]) [Steve Henson]
9756 * Fix buffer overflow in SSL_get_shared_ciphers() function.
9757 ([CVE-2006-3738]) [Tavis Ormandy and Will Drewry, Google Security Team]
9759 * Fix SSL client code which could crash if connecting to a
9760 malicious SSLv2 server. ([CVE-2006-4343])
9762 *Tavis Ormandy and Will Drewry, Google Security Team*
9764 * Since 0.9.8b, ciphersuite strings naming explicit ciphersuites
9765 match only those. Before that, "AES256-SHA" would be interpreted
9766 as a pattern and match "AES128-SHA" too (since AES128-SHA got
9767 the same strength classification in 0.9.7h) as we currently only
9768 have a single AES bit in the ciphersuite description bitmap.
9769 That change, however, also applied to ciphersuite strings such as
9770 "RC4-MD5" that intentionally matched multiple ciphersuites --
9771 namely, SSL 2.0 ciphersuites in addition to the more common ones
9772 from SSL 3.0/TLS 1.0.
9774 So we change the selection algorithm again: Naming an explicit
9775 ciphersuite selects this one ciphersuite, and any other similar
9776 ciphersuite (same bitmap) from *other* protocol versions.
9777 Thus, "RC4-MD5" again will properly select both the SSL 2.0
9778 ciphersuite and the SSL 3.0/TLS 1.0 ciphersuite.
9780 Since SSL 2.0 does not have any ciphersuites for which the
9781 128/256 bit distinction would be relevant, this works for now.
9782 The proper fix will be to use different bits for AES128 and
9783 AES256, which would have avoided the problems from the beginning;
9784 however, bits are scarce, so we can only do this in a new release
9785 (not just a patchlevel) when we can change the SSL_CIPHER
9786 definition to split the single 'unsigned long mask' bitmap into
9787 multiple values to extend the available space.
9791 ### Changes between 0.9.8b and 0.9.8c [05 Sep 2006]
9793 * Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
9794 ([CVE-2006-4339]) [Ben Laurie and Google Security Team]
9796 * Add AES IGE and biIGE modes.
9800 * Change the Unix randomness entropy gathering to use poll() when
9801 possible instead of select(), since the latter has some
9802 undesirable limitations.
9804 *Darryl Miles via Richard Levitte and Bodo Moeller*
9806 * Disable "ECCdraft" ciphersuites more thoroughly. Now special
9807 treatment in ssl/ssl_ciph.s makes sure that these ciphersuites
9808 cannot be implicitly activated as part of, e.g., the "AES" alias.
9809 However, please upgrade to OpenSSL 0.9.9[-dev] for
9810 non-experimental use of the ECC ciphersuites to get TLS extension
9811 support, which is required for curve and point format negotiation
9812 to avoid potential handshake problems.
9816 * Disable rogue ciphersuites:
9818 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
9819 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
9820 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
9822 The latter two were purportedly from
9823 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
9826 Also deactivate the remaining ciphersuites from
9827 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
9828 unofficial, and the ID has long expired.
9832 * Fix RSA blinding Heisenbug (problems sometimes occurred on
9833 dual-core machines) and other potential thread-safety issues.
9837 * Add the symmetric cipher Camellia (128-bit, 192-bit, 256-bit key
9838 versions), which is now available for royalty-free use
9839 (see <http://info.isl.ntt.co.jp/crypt/eng/info/chiteki.html>).
9840 Also, add Camellia TLS ciphersuites from RFC 4132.
9842 To minimize changes between patchlevels in the OpenSSL 0.9.8
9843 series, Camellia remains excluded from compilation unless OpenSSL
9844 is configured with 'enable-camellia'.
9848 * Disable the padding bug check when compression is in use. The padding
9849 bug check assumes the first packet is of even length, this is not
9850 necessarily true if compression is enabled and can result in false
9851 positives causing handshake failure. The actual bug test is ancient
9852 code so it is hoped that implementations will either have fixed it by
9853 now or any which still have the bug do not support compression.
9857 ### Changes between 0.9.8a and 0.9.8b [04 May 2006]
9859 * When applying a cipher rule check to see if string match is an explicit
9860 cipher suite and only match that one cipher suite if it is.
9864 * Link in manifests for VC++ if needed.
9866 *Austin Ziegler <halostatue@gmail.com>*
9868 * Update support for ECC-based TLS ciphersuites according to
9869 draft-ietf-tls-ecc-12.txt with proposed changes (but without
9870 TLS extensions, which are supported starting with the 0.9.9
9871 branch, not in the OpenSSL 0.9.8 branch).
9875 * New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support
9876 opaque EVP_CIPHER_CTX handling.
9880 * Fixes and enhancements to zlib compression code. We now only use
9881 "zlib1.dll" and use the default `__cdecl` calling convention on Win32
9882 to conform with the standards mentioned here:
9883 <http://www.zlib.net/DLL_FAQ.txt>
9884 Static zlib linking now works on Windows and the new --with-zlib-include
9885 --with-zlib-lib options to Configure can be used to supply the location
9886 of the headers and library. Gracefully handle case where zlib library
9891 * Several fixes and enhancements to the OID generation code. The old code
9892 sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't
9893 handle numbers larger than ULONG_MAX, truncated printing and had a
9894 non standard OBJ_obj2txt() behaviour.
9898 * Add support for building of engines under engine/ as shared libraries
9899 under VC++ build system.
9903 * Corrected the numerous bugs in the Win32 path splitter in DSO.
9904 Hopefully, we will not see any false combination of paths any more.
9908 ### Changes between 0.9.8 and 0.9.8a [11 Oct 2005]
9910 * Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
9911 (part of SSL_OP_ALL). This option used to disable the
9912 countermeasure against man-in-the-middle protocol-version
9913 rollback in the SSL 2.0 server implementation, which is a bad
9914 idea. ([CVE-2005-2969])
9916 *Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
9917 for Information Security, National Institute of Advanced Industrial
9918 Science and Technology [AIST], Japan)*
9920 * Add two function to clear and return the verify parameter flags.
9924 * Keep cipherlists sorted in the source instead of sorting them at
9925 runtime, thus removing the need for a lock.
9929 * Avoid some small subgroup attacks in Diffie-Hellman.
9931 *Nick Mathewson and Ben Laurie*
9933 * Add functions for well-known primes.
9937 * Extended Windows CE support.
9939 *Satoshi Nakamura and Andy Polyakov*
9941 * Initialize SSL_METHOD structures at compile time instead of during
9942 runtime, thus removing the need for a lock.
9946 * Make PKCS7_decrypt() work even if no certificate is supplied by
9947 attempting to decrypt each encrypted key in turn. Add support to
9952 ### Changes between 0.9.7h and 0.9.8 [05 Jul 2005]
9954 [NB: OpenSSL 0.9.7i and later 0.9.7 patch levels were released after
9957 * Add libcrypto.pc and libssl.pc for those who feel they need them.
9961 * Change CA.sh and CA.pl so they don't bundle the CSR and the private
9962 key into the same file any more.
9966 * Add initial support for Win64, both IA64 and AMD64/x64 flavors.
9970 * Add -utf8 command line and config file option to 'ca'.
9972 *Stefan <stf@udoma.org*
9974 * Removed the macro des_crypt(), as it seems to conflict with some
9975 libraries. Use DES_crypt().
9979 * Correct naming of the 'chil' and '4758cca' ENGINEs. This
9980 involves renaming the source and generated shared-libs for
9981 both. The engines will accept the corrected or legacy ids
9982 ('ncipher' and '4758_cca' respectively) when binding. NB,
9983 this only applies when building 'shared'.
9985 *Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe*
9987 * Add attribute functions to EVP_PKEY structure. Modify
9988 PKCS12_create() to recognize a CSP name attribute and
9989 use it. Make -CSP option work again in pkcs12 utility.
9993 * Add new functionality to the bn blinding code:
9994 - automatic re-creation of the BN_BLINDING parameters after
9995 a fixed number of uses (currently 32)
9996 - add new function for parameter creation
9997 - introduce flags to control the update behaviour of the
9998 BN_BLINDING parameters
9999 - hide BN_BLINDING structure
10000 Add a second BN_BLINDING slot to the RSA structure to improve
10001 performance when a single RSA object is shared among several
10006 * Add support for DTLS.
10008 *Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie*
10010 * Add support for DER encoded private keys (SSL_FILETYPE_ASN1)
10011 to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file()
10015 * Remove buggy and incomplete DH cert support from
10016 ssl/ssl_rsa.c and ssl/s3_both.c
10020 * Use SHA-1 instead of MD5 as the default digest algorithm for
10021 the `apps/openssl` commands.
10025 * Compile clean with "-Wall -Wmissing-prototypes
10026 -Wstrict-prototypes -Wmissing-declarations -Werror". Currently
10027 DEBUG_SAFESTACK must also be set.
10031 * Change ./Configure so that certain algorithms can be disabled by default.
10032 The new counterpiece to "no-xxx" is "enable-xxx".
10034 The patented RC5 and MDC2 algorithms will now be disabled unless
10035 "enable-rc5" and "enable-mdc2", respectively, are specified.
10037 (IDEA remains enabled despite being patented. This is because IDEA
10038 is frequently required for interoperability, and there is no license
10039 fee for non-commercial use. As before, "no-idea" can be used to
10040 avoid this algorithm.)
10044 * Add processing of proxy certificates (see RFC 3820). This work was
10045 sponsored by KTH (The Royal Institute of Technology in Stockholm) and
10046 EGEE (Enabling Grids for E-science in Europe).
10050 * RC4 performance overhaul on modern architectures/implementations, such
10051 as Intel P4, IA-64 and AMD64.
10055 * New utility extract-section.pl. This can be used specify an alternative
10056 section number in a pod file instead of having to treat each file as
10057 a separate case in Makefile. This can be done by adding two lines to the
10060 =for comment openssl_section:XXX
10062 The blank line is mandatory.
10066 * New arguments -certform, -keyform and -pass for s_client and s_server
10067 to allow alternative format key and certificate files and passphrase
10072 * New structure X509_VERIFY_PARAM which combines current verify parameters,
10073 update associated structures and add various utility functions.
10075 Add new policy related verify parameters, include policy checking in
10076 standard verify code. Enhance 'smime' application with extra parameters
10077 to support policy checking and print out.
10081 * Add a new engine to support VIA PadLock ACE extensions in the VIA C3
10082 Nehemiah processors. These extensions support AES encryption in hardware
10083 as well as RNG (though RNG support is currently disabled).
10085 *Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov*
10087 * Deprecate `BN_[get|set]_params()` functions (they were ignored internally).
10091 * New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented.
10093 *Andy Polyakov and a number of other people*
10095 * Improved PowerPC platform support. Most notably BIGNUM assembler
10096 implementation contributed by IBM.
10098 *Suresh Chari, Peter Waltenberg, Andy Polyakov*
10100 * The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public
10101 exponent rather than 'unsigned long'. There is a corresponding change to
10102 the new 'rsa_keygen' element of the RSA_METHOD structure.
10104 *Jelte Jansen, Geoff Thorpe*
10106 * Functionality for creating the initial serial number file is now
10107 moved from CA.pl to the 'ca' utility with a new option -create_serial.
10109 (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial
10110 number file to 1, which is bound to cause problems. To avoid
10111 the problems while respecting compatibility between different 0.9.7
10112 patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in
10113 CA.pl for serial number initialization. With the new release 0.9.8,
10114 we can fix the problem directly in the 'ca' utility.)
10118 * Reduced header interdependencies by declaring more opaque objects in
10119 ossl_typ.h. As a consequence, including some headers (eg. engine.h) will
10120 give fewer recursive includes, which could break lazy source code - so
10121 this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always,
10122 developers should define this symbol when building and using openssl to
10123 ensure they track the recommended behaviour, interfaces, [etc], but
10124 backwards-compatible behaviour prevails when this isn't defined.
10128 * New function X509_POLICY_NODE_print() which prints out policy nodes.
10132 * Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality.
10133 This will generate a random key of the appropriate length based on the
10134 cipher context. The EVP_CIPHER can provide its own random key generation
10135 routine to support keys of a specific form. This is used in the des and
10136 3des routines to generate a key of the correct parity. Update S/MIME
10137 code to use new functions and hence generate correct parity DES keys.
10138 Add EVP_CHECK_DES_KEY #define to return an error if the key is not
10139 valid (weak or incorrect parity).
10143 * Add a local set of CRLs that can be used by X509_verify_cert() as well
10144 as looking them up. This is useful when the verified structure may contain
10145 CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs
10146 present unless the new PKCS7_NO_CRL flag is asserted.
10150 * Extend ASN1 oid configuration module. It now additionally accepts the
10153 shortName = some long name, 1.2.3.4
10157 * Reimplemented the BN_CTX implementation. There is now no more static
10158 limitation on the number of variables it can handle nor the depth of the
10159 "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack
10160 information can now expand as required, and rather than having a single
10161 static array of bignums, BN_CTX now uses a linked-list of such arrays
10162 allowing it to expand on demand whilst maintaining the usefulness of
10163 BN_CTX's "bundling".
10167 * Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD
10168 to allow all RSA operations to function using a single BN_CTX.
10172 * Preliminary support for certificate policy evaluation and checking. This
10173 is initially intended to pass the tests outlined in "Conformance Testing
10174 of Relying Party Client Certificate Path Processing Logic" v1.07.
10178 * bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and
10179 remained unused and not that useful. A variety of other little bignum
10180 tweaks and fixes have also been made continuing on from the audit (see
10185 * Constify all or almost all d2i, c2i, s2i and r2i functions, along with
10186 associated ASN1, EVP and SSL functions and old ASN1 macros.
10190 * BN_zero() only needs to set 'top' and 'neg' to zero for correct results,
10191 and this should never fail. So the return value from the use of
10192 BN_set_word() (which can fail due to needless expansion) is now deprecated;
10193 if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro.
10197 * BN_CTX_get() should return zero-valued bignums, providing the same
10198 initialised value as BN_new().
10200 *Geoff Thorpe, suggested by Ulf Möller*
10202 * Support for inhibitAnyPolicy certificate extension.
10206 * An audit of the BIGNUM code is underway, for which debugging code is
10207 enabled when BN_DEBUG is defined. This makes stricter enforcements on what
10208 is considered valid when processing BIGNUMs, and causes execution to
10209 assert() when a problem is discovered. If BN_DEBUG_RAND is defined,
10210 further steps are taken to deliberately pollute unused data in BIGNUM
10211 structures to try and expose faulty code further on. For now, openssl will
10212 (in its default mode of operation) continue to tolerate the inconsistent
10213 forms that it has tolerated in the past, but authors and packagers should
10214 consider trying openssl and their own applications when compiled with
10215 these debugging symbols defined. It will help highlight potential bugs in
10216 their own code, and will improve the test coverage for OpenSSL itself. At
10217 some point, these tighter rules will become openssl's default to improve
10218 maintainability, though the assert()s and other overheads will remain only
10219 in debugging configurations. See bn.h for more details.
10221 *Geoff Thorpe, Nils Larsch, Ulf Möller*
10223 * BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure
10224 that can only be obtained through BN_CTX_new() (which implicitly
10225 initialises it). The presence of this function only made it possible
10226 to overwrite an existing structure (and cause memory leaks).
10230 * Because of the callback-based approach for implementing LHASH as a
10231 template type, lh_insert() adds opaque objects to hash-tables and
10232 lh_doall() or lh_doall_arg() are typically used with a destructor callback
10233 to clean up those corresponding objects before destroying the hash table
10234 (and losing the object pointers). So some over-zealous constifications in
10235 LHASH have been relaxed so that lh_insert() does not take (nor store) the
10236 objects as "const" and the `lh_doall[_arg]` callback wrappers are not
10237 prototyped to have "const" restrictions on the object pointers they are
10238 given (and so aren't required to cast them away any more).
10242 * The tmdiff.h API was so ugly and minimal that our own timing utility
10243 (speed) prefers to use its own implementation. The two implementations
10244 haven't been consolidated as yet (volunteers?) but the tmdiff API has had
10245 its object type properly exposed (MS_TM) instead of casting to/from
10246 `char *`. This may still change yet if someone realises MS_TM and
10248 aren't necessarily the greatest nomenclatures - but this is what was used
10249 internally to the implementation so I've used that for now.
10253 * Ensure that deprecated functions do not get compiled when
10254 OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of
10255 the self-tests were still using deprecated key-generation functions so
10256 these have been updated also.
10260 * Reorganise PKCS#7 code to separate the digest location functionality
10261 into PKCS7_find_digest(), digest addition into PKCS7_bio_add_digest().
10262 New function PKCS7_set_digest() to set the digest type for PKCS#7
10263 digestedData type. Add additional code to correctly generate the
10264 digestedData type and add support for this type in PKCS7 initialization
10269 * New function PKCS7_set0_type_other() this initializes a PKCS7
10270 structure of type "other".
10274 * Fix prime generation loop in crypto/bn/bn_prime.pl by making
10275 sure the loop does correctly stop and breaking ("division by zero")
10276 modulus operations are not performed. The (pre-generated) prime
10277 table crypto/bn/bn_prime.h was already correct, but it could not be
10278 re-generated on some platforms because of the "division by zero"
10279 situation in the script.
10281 *Ralf S. Engelschall*
10283 * Update support for ECC-based TLS ciphersuites according to
10284 draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with
10285 SHA-1 now is only used for "small" curves (where the
10286 representation of a field element takes up to 24 bytes); for
10287 larger curves, the field element resulting from ECDH is directly
10288 used as premaster secret.
10290 *Douglas Stebila (Sun Microsystems Laboratories)*
10292 * Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2
10293 curve secp160r1 to the tests.
10295 *Douglas Stebila (Sun Microsystems Laboratories)*
10297 * Add the possibility to load symbols globally with DSO.
10299 *Götz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte*
10301 * Add the functions ERR_set_mark() and ERR_pop_to_mark() for better
10302 control of the error stack.
10306 * Add support for STORE in ENGINE.
10310 * Add the STORE type. The intention is to provide a common interface
10311 to certificate and key stores, be they simple file-based stores, or
10312 HSM-type store, or LDAP stores, or...
10313 NOTE: The code is currently UNTESTED and isn't really used anywhere.
10317 * Add a generic structure called OPENSSL_ITEM. This can be used to
10318 pass a list of arguments to any function as well as provide a way
10319 for a function to pass data back to the caller.
10323 * Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup()
10324 works like BUF_strdup() but can be used to duplicate a portion of
10325 a string. The copy gets NUL-terminated. BUF_memdup() duplicates
10330 * Add the function sk_find_ex() which works like sk_find(), but will
10331 return an index to an element even if an exact match couldn't be
10332 found. The index is guaranteed to point at the element where the
10333 searched-for key would be inserted to preserve sorting order.
10337 * Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but
10338 takes an extra flags argument for optional functionality. Currently,
10339 the following flags are defined:
10341 OBJ_BSEARCH_VALUE_ON_NOMATCH
10342 This one gets OBJ_bsearch_ex() to return a pointer to the first
10343 element where the comparing function returns a negative or zero
10346 OBJ_BSEARCH_FIRST_VALUE_ON_MATCH
10347 This one gets OBJ_bsearch_ex() to return a pointer to the first
10348 element where the comparing function returns zero. This is useful
10349 if there are more than one element where the comparing function
10354 * Make it possible to create self-signed certificates with 'openssl ca'
10355 in such a way that the self-signed certificate becomes part of the
10356 CA database and uses the same mechanisms for serial number generation
10357 as all other certificate signing. The new flag '-selfsign' enables
10358 this functionality. Adapt CA.sh and CA.pl.in.
10362 * Add functionality to check the public key of a certificate request
10363 against a given private. This is useful to check that a certificate
10364 request can be signed by that key (self-signing).
10368 * Make it possible to have multiple active certificates with the same
10369 subject in the CA index file. This is done only if the keyword
10370 'unique_subject' is set to 'no' in the main CA section (default
10371 if 'CA_default') of the configuration file. The value is saved
10372 with the database itself in a separate index attribute file,
10373 named like the index file with '.attr' appended to the name.
10377 * Generate multi-valued AVAs using '+' notation in config files for
10382 * Support for nameConstraints certificate extension.
10386 * Support for policyConstraints certificate extension.
10390 * Support for policyMappings certificate extension.
10394 * Make sure the default DSA_METHOD implementation only uses its
10395 dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL,
10396 and change its own handlers to be NULL so as to remove unnecessary
10397 indirection. This lets alternative implementations fallback to the
10398 default implementation more easily.
10402 * Support for directoryName in GeneralName related extensions
10407 * Make it possible to link applications using Makefile.shared.
10408 Make that possible even when linking against static libraries!
10412 * Support for single pass processing for S/MIME signing. This now
10413 means that S/MIME signing can be done from a pipe, in addition
10414 cleartext signing (multipart/signed type) is effectively streaming
10415 and the signed data does not need to be all held in memory.
10417 This is done with a new flag PKCS7_STREAM. When this flag is set
10418 PKCS7_sign() only initializes the PKCS7 structure and the actual signing
10419 is done after the data is output (and digests calculated) in
10420 SMIME_write_PKCS7().
10424 * Add full support for -rpath/-R, both in shared libraries and
10425 applications, at least on the platforms where it's known how
10430 * In crypto/ec/ec_mult.c, implement fast point multiplication with
10431 precomputation, based on wNAF splitting: EC_GROUP_precompute_mult()
10432 will now compute a table of multiples of the generator that
10433 makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul()
10434 faster (notably in the case of a single point multiplication,
10435 scalar * generator).
10437 *Nils Larsch, Bodo Moeller*
10439 * IPv6 support for certificate extensions. The various extensions
10440 which use the IP:a.b.c.d can now take IPv6 addresses using the
10441 formats of RFC1884 2.2 . IPv6 addresses are now also displayed
10446 * Added an ENGINE that implements RSA by performing private key
10447 exponentiations with the GMP library. The conversions to and from
10448 GMP's mpz_t format aren't optimised nor are any montgomery forms
10449 cached, and on x86 it appears OpenSSL's own performance has caught up.
10450 However there are likely to be other architectures where GMP could
10451 provide a boost. This ENGINE is not built in by default, but it can be
10452 specified at Configure time and should be accompanied by the necessary
10453 linker additions, eg;
10454 ./config -DOPENSSL_USE_GMP -lgmp
10458 * "openssl engine" will not display ENGINE/DSO load failure errors when
10459 testing availability of engines with "-t" - the old behaviour is
10460 produced by increasing the feature's verbosity with "-tt".
10464 * ECDSA routines: under certain error conditions uninitialized BN objects
10465 could be freed. Solution: make sure initialization is performed early
10466 enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de>
10471 * Key-generation can now be implemented in RSA_METHOD, DSA_METHOD
10472 and DH_METHOD (eg. by ENGINE implementations) to override the normal
10473 software implementations. For DSA and DH, parameter generation can
10474 also be overridden by providing the appropriate method callbacks.
10478 * Change the "progress" mechanism used in key-generation and
10479 primality testing to functions that take a new BN_GENCB pointer in
10480 place of callback/argument pairs. The new API functions have `_ex`
10481 postfixes and the older functions are reimplemented as wrappers for
10482 the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide
10483 declarations of the old functions to help (graceful) attempts to
10484 migrate to the new functions. Also, the new key-generation API
10485 functions operate on a caller-supplied key-structure and return
10486 success/failure rather than returning a key or NULL - this is to
10487 help make "keygen" another member function of RSA_METHOD etc.
10489 Example for using the new callback interface:
10491 int (*my_callback)(int a, int b, BN_GENCB *cb) = ...;
10492 void *my_arg = ...;
10495 BN_GENCB_set(&my_cb, my_callback, my_arg);
10497 return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb);
10498 /* For the meaning of a, b in calls to my_callback(), see the
10499 * documentation of the function that calls the callback.
10500 * cb will point to my_cb; my_arg can be retrieved as cb->arg.
10501 * my_callback should return 1 if it wants BN_is_prime_ex()
10502 * to continue, or 0 to stop.
10507 * Change the ZLIB compression method to be stateful, and make it
10508 available to TLS with the number defined in
10509 draft-ietf-tls-compression-04.txt.
10513 * Add the ASN.1 structures and functions for CertificatePair, which
10514 is defined as follows (according to X.509_4thEditionDraftV6.pdf):
10516 CertificatePair ::= SEQUENCE {
10517 forward [0] Certificate OPTIONAL,
10518 reverse [1] Certificate OPTIONAL,
10519 -- at least one of the pair shall be present -- }
10521 Also implement the PEM functions to read and write certificate
10522 pairs, and defined the PEM tag as "CERTIFICATE PAIR".
10524 This needed to be defined, mostly for the sake of the LDAP
10525 attribute crossCertificatePair, but may prove useful elsewhere as
10530 * Make it possible to inhibit symlinking of shared libraries in
10531 Makefile.shared, for Cygwin's sake.
10535 * Extend the BIGNUM API by creating a function
10536 void BN_set_negative(BIGNUM *a, int neg);
10537 and a macro that behave like
10538 int BN_is_negative(const BIGNUM *a);
10540 to avoid the need to access 'a->neg' directly in applications.
10544 * Implement fast modular reduction for pseudo-Mersenne primes
10545 used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c).
10546 EC_GROUP_new_curve_GFp() will now automatically use this
10549 *Nils Larsch <nla@trustcenter.de>*
10551 * Add new lock type (CRYPTO_LOCK_BN).
10555 * Change the ENGINE framework to automatically load engines
10556 dynamically from specific directories unless they could be
10557 found to already be built in or loaded. Move all the
10558 current engines except for the cryptodev one to a new
10559 directory engines/.
10560 The engines in engines/ are built as shared libraries if
10561 the "shared" options was given to ./Configure or ./config.
10562 Otherwise, they are inserted in libcrypto.a.
10563 /usr/local/ssl/engines is the default directory for dynamic
10564 engines, but that can be overridden at configure time through
10565 the usual use of --prefix and/or --openssldir, and at run
10566 time with the environment variable OPENSSL_ENGINES.
10568 *Geoff Thorpe and Richard Levitte*
10570 * Add Makefile.shared, a helper makefile to build shared
10571 libraries. Adapt Makefile.org.
10575 * Add version info to Win32 DLLs.
10577 *Peter 'Luna' Runestig" <peter@runestig.com>*
10579 * Add new 'medium level' PKCS#12 API. Certificates and keys
10580 can be added using this API to created arbitrary PKCS#12
10581 files while avoiding the low-level API.
10583 New options to PKCS12_create(), key or cert can be NULL and
10584 will then be omitted from the output file. The encryption
10585 algorithm NIDs can be set to -1 for no encryption, the mac
10586 iteration count can be set to 0 to omit the mac.
10588 Enhance pkcs12 utility by making the -nokeys and -nocerts
10589 options work when creating a PKCS#12 file. New option -nomac
10590 to omit the mac, NONE can be set for an encryption algorithm.
10591 New code is modified to use the enhanced PKCS12_create()
10592 instead of the low-level API.
10596 * Extend ASN1 encoder to support indefinite length constructed
10597 encoding. This can output sequences tags and octet strings in
10598 this form. Modify pk7_asn1.c to support indefinite length
10599 encoding. This is experimental and needs additional code to
10600 be useful, such as an ASN1 bio and some enhanced streaming
10603 Extend template encode functionality so that tagging is passed
10604 down to the template encoder.
10608 * Let 'openssl req' fail if an argument to '-newkey' is not
10609 recognized instead of using RSA as a default.
10613 * Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt.
10614 As these are not official, they are not included in "ALL";
10615 the "ECCdraft" ciphersuite group alias can be used to select them.
10617 *Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)*
10619 * Add ECDH engine support.
10621 *Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)*
10623 * Add ECDH in new directory crypto/ecdh/.
10625 *Douglas Stebila (Sun Microsystems Laboratories)*
10627 * Let BN_rand_range() abort with an error after 100 iterations
10628 without success (which indicates a broken PRNG).
10632 * Change BN_mod_sqrt() so that it verifies that the input value
10633 is really the square of the return value. (Previously,
10634 BN_mod_sqrt would show GIGO behaviour.)
10638 * Add named elliptic curves over binary fields from X9.62, SECG,
10639 and WAP/WTLS; add OIDs that were still missing.
10641 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10643 * Extend the EC library for elliptic curves over binary fields
10644 (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/).
10647 EC_GF2m_simple_method
10651 EC_GROUP_new_curve_GF2m
10652 EC_GROUP_set_curve_GF2m
10653 EC_GROUP_get_curve_GF2m
10654 EC_POINT_set_affine_coordinates_GF2m
10655 EC_POINT_get_affine_coordinates_GF2m
10656 EC_POINT_set_compressed_coordinates_GF2m
10658 Point compression for binary fields is disabled by default for
10659 patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to
10662 As binary polynomials are represented as BIGNUMs, various members
10663 of the EC_GROUP and EC_POINT data structures can be shared
10664 between the implementations for prime fields and binary fields;
10665 the above `..._GF2m functions` (except for EX_GROUP_new_curve_GF2m)
10666 are essentially identical to their `..._GFp` counterparts.
10667 (For simplicity, the `..._GFp` prefix has been dropped from
10668 various internal method names.)
10670 An internal 'field_div' method (similar to 'field_mul' and
10671 'field_sqr') has been added; this is used only for binary fields.
10673 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10675 * Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult()
10676 through methods ('mul', 'precompute_mult').
10678 The generic implementations (now internally called 'ec_wNAF_mul'
10679 and 'ec_wNAF_precomputed_mult') remain the default if these
10680 methods are undefined.
10682 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10684 * New function EC_GROUP_get_degree, which is defined through
10685 EC_METHOD. For curves over prime fields, this returns the bit
10686 length of the modulus.
10688 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10690 * New functions EC_GROUP_dup, EC_POINT_dup.
10691 (These simply call ..._new and ..._copy).
10693 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10695 * Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c.
10696 Polynomials are represented as BIGNUMs (where the sign bit is not
10697 used) in the following functions [macros]:
10700 BN_GF2m_sub [= BN_GF2m_add]
10701 BN_GF2m_mod [wrapper for BN_GF2m_mod_arr]
10702 BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr]
10703 BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr]
10705 BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr]
10706 BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr]
10707 BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr]
10708 BN_GF2m_cmp [= BN_ucmp]
10710 (Note that only the 'mod' functions are actually for fields GF(2^m).
10711 BN_GF2m_add() is misnomer, but this is for the sake of consistency.)
10713 For some functions, an the irreducible polynomial defining a
10714 field can be given as an 'unsigned int[]' with strictly
10715 decreasing elements giving the indices of those bits that are set;
10716 i.e., p[] represents the polynomial
10717 f(t) = t^p[0] + t^p[1] + ... + t^p[k]
10719 p[0] > p[1] > ... > p[k] = 0.
10720 This applies to the following functions:
10723 BN_GF2m_mod_mul_arr
10724 BN_GF2m_mod_sqr_arr
10725 BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv]
10726 BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div]
10727 BN_GF2m_mod_exp_arr
10728 BN_GF2m_mod_sqrt_arr
10729 BN_GF2m_mod_solve_quad_arr
10733 Conversion can be performed by the following functions:
10738 bntest.c has additional tests for binary polynomial arithmetic.
10740 Two implementations for BN_GF2m_mod_div() are available.
10741 The default algorithm simply uses BN_GF2m_mod_inv() and
10742 BN_GF2m_mod_mul(). The alternative algorithm is compiled in only
10743 if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the
10744 copyright notice in crypto/bn/bn_gf2m.c before enabling it).
10746 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10748 * Add new error code 'ERR_R_DISABLED' that can be used when some
10749 functionality is disabled at compile-time.
10751 *Douglas Stebila <douglas.stebila@sun.com>*
10753 * Change default behaviour of 'openssl asn1parse' so that more
10754 information is visible when viewing, e.g., a certificate:
10756 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump'
10757 mode the content of non-printable OCTET STRINGs is output in a
10758 style similar to INTEGERs, but with '[HEX DUMP]' prepended to
10759 avoid the appearance of a printable string.
10761 *Nils Larsch <nla@trustcenter.de>*
10763 * Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access
10765 EC_GROUP_set_asn1_flag()
10766 EC_GROUP_get_asn1_flag()
10767 EC_GROUP_set_point_conversion_form()
10768 EC_GROUP_get_point_conversion_form()
10769 These control ASN1 encoding details:
10770 - Curves (i.e., groups) are encoded explicitly unless asn1_flag
10771 has been set to OPENSSL_EC_NAMED_CURVE.
10772 - Points are encoded in uncompressed form by default; options for
10773 asn1_for are as for point2oct, namely
10774 POINT_CONVERSION_COMPRESSED
10775 POINT_CONVERSION_UNCOMPRESSED
10776 POINT_CONVERSION_HYBRID
10778 Also add 'seed' and 'seed_len' members to EC_GROUP with access
10780 EC_GROUP_set_seed()
10781 EC_GROUP_get0_seed()
10782 EC_GROUP_get_seed_len()
10783 This is used only for ASN1 purposes (so far).
10785 *Nils Larsch <nla@trustcenter.de>*
10787 * Add 'field_type' member to EC_METHOD, which holds the NID
10788 of the appropriate field type OID. The new function
10789 EC_METHOD_get_field_type() returns this value.
10791 *Nils Larsch <nla@trustcenter.de>*
10794 EC_POINT_point2bn()
10795 EC_POINT_bn2point()
10796 EC_POINT_point2hex()
10797 EC_POINT_hex2point()
10798 providing useful interfaces to EC_POINT_point2oct() and
10799 EC_POINT_oct2point().
10801 *Nils Larsch <nla@trustcenter.de>*
10803 * Change internals of the EC library so that the functions
10804 EC_GROUP_set_generator()
10805 EC_GROUP_get_generator()
10806 EC_GROUP_get_order()
10807 EC_GROUP_get_cofactor()
10808 are implemented directly in crypto/ec/ec_lib.c and not dispatched
10809 to methods, which would lead to unnecessary code duplication when
10810 adding different types of curves.
10812 *Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller*
10814 * Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM
10815 arithmetic, and such that modified wNAFs are generated
10816 (which avoid length expansion in many cases).
10820 * Add a function EC_GROUP_check_discriminant() (defined via
10821 EC_METHOD) that verifies that the curve discriminant is non-zero.
10823 Add a function EC_GROUP_check() that makes some sanity tests
10824 on a EC_GROUP, its generator and order. This includes
10825 EC_GROUP_check_discriminant().
10827 *Nils Larsch <nla@trustcenter.de>*
10829 * Add ECDSA in new directory crypto/ecdsa/.
10831 Add applications 'openssl ecparam' and 'openssl ecdsa'
10832 (these are based on 'openssl dsaparam' and 'openssl dsa').
10834 ECDSA support is also included in various other files across the
10835 library. Most notably,
10836 - 'openssl req' now has a '-newkey ecdsa:file' option;
10837 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA;
10838 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and
10839 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make
10840 them suitable for ECDSA where domain parameters must be
10841 extracted before the specific public key;
10842 - ECDSA engine support has been added.
10844 *Nils Larsch <nla@trustcenter.de>*
10846 * Include some named elliptic curves, and add OIDs from X9.62,
10847 SECG, and WAP/WTLS. Each curve can be obtained from the new
10849 EC_GROUP_new_by_curve_name(),
10850 and the list of available named curves can be obtained with
10851 EC_get_builtin_curves().
10852 Also add a 'curve_name' member to EC_GROUP objects, which can be
10854 EC_GROUP_set_curve_name()
10855 EC_GROUP_get_curve_name()
10857 *Nils Larsch <larsch@trustcenter.de, Bodo Moeller*
10859 * Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
10860 was actually never needed) and in BN_mul(). The removal in BN_mul()
10861 required a small change in bn_mul_part_recursive() and the addition
10862 of the functions bn_cmp_part_words(), bn_sub_part_words() and
10863 bn_add_part_words(), which do the same thing as bn_cmp_words(),
10864 bn_sub_words() and bn_add_words() except they take arrays with
10869 ### Changes between 0.9.7l and 0.9.7m [23 Feb 2007]
10871 * Cleanse PEM buffers before freeing them since they may contain
10874 *Benjamin Bennett <ben@psc.edu>*
10876 * Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
10877 a ciphersuite string such as "DEFAULT:RSA" cannot enable
10878 authentication-only ciphersuites.
10882 * Since AES128 and AES256 share a single mask bit in the logic of
10883 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
10884 kludge to work properly if AES128 is available and AES256 isn't.
10888 * Expand security boundary to match 1.1.1 module.
10892 * Remove redundant features: hash file source, editing of test vectors
10893 modify fipsld to use external fips_premain.c signature.
10897 * New perl script mkfipsscr.pl to create shell scripts or batch files to
10898 run algorithm test programs.
10902 * Make algorithm test programs more tolerant of whitespace.
10906 * Have SSL/TLS server implementation tolerate "mismatched" record
10907 protocol version while receiving ClientHello even if the
10908 ClientHello is fragmented. (The server can't insist on the
10909 particular protocol version it has chosen before the ServerHello
10910 message has informed the client about his choice.)
10914 * Load error codes if they are not already present instead of using a
10915 static variable. This allows them to be cleanly unloaded and reloaded.
10919 ### Changes between 0.9.7k and 0.9.7l [28 Sep 2006]
10921 * Introduce limits to prevent malicious keys being able to
10922 cause a denial of service. ([CVE-2006-2940])
10924 *Steve Henson, Bodo Moeller*
10926 * Fix ASN.1 parsing of certain invalid structures that can result
10927 in a denial of service. ([CVE-2006-2937]) [Steve Henson]
10929 * Fix buffer overflow in SSL_get_shared_ciphers() function.
10930 ([CVE-2006-3738]) [Tavis Ormandy and Will Drewry, Google Security Team]
10932 * Fix SSL client code which could crash if connecting to a
10933 malicious SSLv2 server. ([CVE-2006-4343])
10935 *Tavis Ormandy and Will Drewry, Google Security Team*
10937 * Change ciphersuite string processing so that an explicit
10938 ciphersuite selects this one ciphersuite (so that "AES256-SHA"
10939 will no longer include "AES128-SHA"), and any other similar
10940 ciphersuite (same bitmap) from *other* protocol versions (so that
10941 "RC4-MD5" will still include both the SSL 2.0 ciphersuite and the
10942 SSL 3.0/TLS 1.0 ciphersuite). This is a backport combining
10943 changes from 0.9.8b and 0.9.8d.
10947 ### Changes between 0.9.7j and 0.9.7k [05 Sep 2006]
10949 * Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
10950 ([CVE-2006-4339]) [Ben Laurie and Google Security Team]
10952 * Change the Unix randomness entropy gathering to use poll() when
10953 possible instead of select(), since the latter has some
10954 undesirable limitations.
10956 *Darryl Miles via Richard Levitte and Bodo Moeller*
10958 * Disable rogue ciphersuites:
10960 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
10961 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
10962 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
10964 The latter two were purportedly from
10965 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
10968 Also deactivate the remaining ciphersuites from
10969 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
10970 unofficial, and the ID has long expired.
10974 * Fix RSA blinding Heisenbug (problems sometimes occurred on
10975 dual-core machines) and other potential thread-safety issues.
10979 ### Changes between 0.9.7i and 0.9.7j [04 May 2006]
10981 * Adapt fipsld and the build system to link against the validated FIPS
10982 module in FIPS mode.
10986 * Fixes for VC++ 2005 build under Windows.
10990 * Add new Windows build target VC-32-GMAKE for VC++. This uses GNU make
10991 from a Windows bash shell such as MSYS. It is autodetected from the
10992 "config" script when run from a VC++ environment. Modify standard VC++
10993 build to use fipscanister.o from the GNU make build.
10997 ### Changes between 0.9.7h and 0.9.7i [14 Oct 2005]
10999 * Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS.
11000 The value now differs depending on if you build for FIPS or not.
11001 BEWARE! A program linked with a shared FIPSed libcrypto can't be
11002 safely run with a non-FIPSed libcrypto, as it may crash because of
11003 the difference induced by this change.
11007 ### Changes between 0.9.7g and 0.9.7h [11 Oct 2005]
11009 * Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
11010 (part of SSL_OP_ALL). This option used to disable the
11011 countermeasure against man-in-the-middle protocol-version
11012 rollback in the SSL 2.0 server implementation, which is a bad
11013 idea. ([CVE-2005-2969])
11015 *Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
11016 for Information Security, National Institute of Advanced Industrial
11017 Science and Technology [AIST, Japan)]*
11019 * Minimal support for X9.31 signatures and PSS padding modes. This is
11020 mainly for FIPS compliance and not fully integrated at this stage.
11024 * For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform
11025 the exponentiation using a fixed-length exponent. (Otherwise,
11026 the information leaked through timing could expose the secret key
11027 after many signatures; cf. Bleichenbacher's attack on DSA with
11032 * Make a new fixed-window mod_exp implementation the default for
11033 RSA, DSA, and DH private-key operations so that the sequence of
11034 squares and multiplies and the memory access pattern are
11035 independent of the particular secret key. This will mitigate
11036 cache-timing and potential related attacks.
11038 BN_mod_exp_mont_consttime() is the new exponentiation implementation,
11039 and this is automatically used by BN_mod_exp_mont() if the new flag
11040 BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH
11041 will use this BN flag for private exponents unless the flag
11042 RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or
11043 DH_FLAG_NO_EXP_CONSTTIME, respectively, is set.
11045 *Matthew D Wood (Intel Corp), with some changes by Bodo Moeller*
11047 * Change the client implementation for SSLv23_method() and
11048 SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0
11049 Client Hello message format if the SSL_OP_NO_SSLv2 option is set.
11050 (Previously, the SSL 2.0 backwards compatible Client Hello
11051 message format would be used even with SSL_OP_NO_SSLv2.)
11055 * Add support for smime-type MIME parameter in S/MIME messages which some
11060 * New function BN_MONT_CTX_set_locked() to set montgomery parameters in
11061 a threadsafe manner. Modify rsa code to use new function and add calls
11062 to dsa and dh code (which had race conditions before).
11066 * Include the fixed error library code in the C error file definitions
11067 instead of fixing them up at runtime. This keeps the error code
11068 structures constant.
11072 ### Changes between 0.9.7f and 0.9.7g [11 Apr 2005]
11074 [NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after
11077 * Fixes for newer kerberos headers. NB: the casts are needed because
11078 the 'length' field is signed on one version and unsigned on another
11079 with no (?) obvious way to tell the difference, without these VC++
11080 complains. Also the "definition" of FAR (blank) is no longer included
11081 nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up
11082 some needed definitions.
11086 * Undo Cygwin change.
11090 * Added support for proxy certificates according to RFC 3820.
11091 Because they may be a security thread to unaware applications,
11092 they must be explicitly allowed in run-time. See
11093 docs/HOWTO/proxy_certificates.txt for further information.
11097 ### Changes between 0.9.7e and 0.9.7f [22 Mar 2005]
11099 * Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating
11100 server and client random values. Previously
11101 (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in
11102 less random data when sizeof(time_t) > 4 (some 64 bit platforms).
11104 This change has negligible security impact because:
11106 1. Server and client random values still have 24 bytes of pseudo random
11109 2. Server and client random values are sent in the clear in the initial
11112 3. The master secret is derived using the premaster secret (48 bytes in
11113 size for static RSA ciphersuites) as well as client server and random
11116 The OpenSSL team would like to thank the UK NISCC for bringing this issue
11119 *Stephen Henson, reported by UK NISCC*
11121 * Use Windows randomness collection on Cygwin.
11125 * Fix hang in EGD/PRNGD query when communication socket is closed
11126 prematurely by EGD/PRNGD.
11128 *Darren Tucker <dtucker@zip.com.au> via Lutz Jänicke, resolves #1014*
11130 * Prompt for pass phrases when appropriate for PKCS12 input format.
11134 * Back-port of selected performance improvements from development
11135 branch, as well as improved support for PowerPC platforms.
11139 * Add lots of checks for memory allocation failure, error codes to indicate
11140 failure and freeing up memory if a failure occurs.
11142 *Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson*
11144 * Add new -passin argument to dgst.
11148 * Perform some character comparisons of different types in X509_NAME_cmp:
11149 this is needed for some certificates that re-encode DNs into UTF8Strings
11150 (in violation of RFC3280) and can't or won't issue name rollover
11155 * Make an explicit check during certificate validation to see that
11156 the CA setting in each certificate on the chain is correct. As a
11157 side effect always do the following basic checks on extensions,
11158 not just when there's an associated purpose to the check:
11160 - if there is an unhandled critical extension (unless the user
11161 has chosen to ignore this fault)
11162 - if the path length has been exceeded (if one is set at all)
11163 - that certain extensions fit the associated purpose (if one has
11168 ### Changes between 0.9.7d and 0.9.7e [25 Oct 2004]
11170 * Avoid a race condition when CRLs are checked in a multi threaded
11171 environment. This would happen due to the reordering of the revoked
11172 entries during signature checking and serial number lookup. Now the
11173 encoding is cached and the serial number sort performed under a lock.
11174 Add new STACK function sk_is_sorted().
11178 * Add Delta CRL to the extension code.
11182 * Various fixes to s3_pkt.c so alerts are sent properly.
11184 *David Holmes <d.holmes@f5.com>*
11186 * Reduce the chances of duplicate issuer name and serial numbers (in
11187 violation of RFC3280) using the OpenSSL certificate creation utilities.
11188 This is done by creating a random 64 bit value for the initial serial
11189 number when a serial number file is created or when a self signed
11190 certificate is created using 'openssl req -x509'. The initial serial
11191 number file is created using 'openssl x509 -next_serial' in CA.pl
11192 rather than being initialized to 1.
11196 ### Changes between 0.9.7c and 0.9.7d [17 Mar 2004]
11198 * Fix null-pointer assignment in do_change_cipher_spec() revealed
11199 by using the Codenomicon TLS Test Tool ([CVE-2004-0079])
11201 *Joe Orton, Steve Henson*
11203 * Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites
11206 *Joe Orton, Steve Henson*
11208 * Make it possible to have multiple active certificates with the same
11209 subject in the CA index file. This is done only if the keyword
11210 'unique_subject' is set to 'no' in the main CA section (default
11211 if 'CA_default') of the configuration file. The value is saved
11212 with the database itself in a separate index attribute file,
11213 named like the index file with '.attr' appended to the name.
11217 * X509 verify fixes. Disable broken certificate workarounds when
11218 X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if
11219 keyUsage extension present. Don't accept CRLs with unhandled critical
11220 extensions: since verify currently doesn't process CRL extensions this
11221 rejects a CRL with *any* critical extensions. Add new verify error codes
11226 * When creating an OCSP nonce use an OCTET STRING inside the extnValue.
11227 A clarification of RFC2560 will require the use of OCTET STRINGs and
11228 some implementations cannot handle the current raw format. Since OpenSSL
11229 copies and compares OCSP nonces as opaque blobs without any attempt at
11230 parsing them this should not create any compatibility issues.
11234 * New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when
11235 calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without
11236 this HMAC (and other) operations are several times slower than OpenSSL
11241 * Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex().
11243 *Peter Sylvester <Peter.Sylvester@EdelWeb.fr>*
11245 * Use the correct content when signing type "other".
11249 ### Changes between 0.9.7b and 0.9.7c [30 Sep 2003]
11251 * Fix various bugs revealed by running the NISCC test suite:
11253 Stop out of bounds reads in the ASN1 code when presented with
11254 invalid tags (CVE-2003-0543 and CVE-2003-0544).
11256 Free up ASN1_TYPE correctly if ANY type is invalid ([CVE-2003-0545]).
11258 If verify callback ignores invalid public key errors don't try to check
11259 certificate signature with the NULL public key.
11263 * New -ignore_err option in ocsp application to stop the server
11264 exiting on the first error in a request.
11268 * In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
11269 if the server requested one: as stated in TLS 1.0 and SSL 3.0
11274 * In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
11275 extra data after the compression methods not only for TLS 1.0
11276 but also for SSL 3.0 (as required by the specification).
11278 *Bodo Moeller; problem pointed out by Matthias Loepfe*
11280 * Change X509_certificate_type() to mark the key as exported/exportable
11281 when it's 512 *bits* long, not 512 bytes.
11285 * Change AES_cbc_encrypt() so it outputs exact multiple of
11286 blocks during encryption.
11290 * Various fixes to base64 BIO and non blocking I/O. On write
11291 flushes were not handled properly if the BIO retried. On read
11292 data was not being buffered properly and had various logic bugs.
11293 This also affects blocking I/O when the data being decoded is a
11298 * Various S/MIME bugfixes and compatibility changes:
11299 output correct application/pkcs7 MIME type if
11300 PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures.
11301 Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening
11302 of files as .eml work). Correctly handle very long lines in MIME
11307 ### Changes between 0.9.7a and 0.9.7b [10 Apr 2003]
11309 * Countermeasure against the Klima-Pokorny-Rosa extension of
11310 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
11311 a protocol version number mismatch like a decryption error
11312 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
11316 * Turn on RSA blinding by default in the default implementation
11317 to avoid a timing attack. Applications that don't want it can call
11318 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
11319 They would be ill-advised to do so in most cases.
11321 *Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller*
11323 * Change RSA blinding code so that it works when the PRNG is not
11324 seeded (in this case, the secret RSA exponent is abused as
11325 an unpredictable seed -- if it is not unpredictable, there
11326 is no point in blinding anyway). Make RSA blinding thread-safe
11327 by remembering the creator's thread ID in rsa->blinding and
11328 having all other threads use local one-time blinding factors
11329 (this requires more computation than sharing rsa->blinding, but
11330 avoids excessive locking; and if an RSA object is not shared
11331 between threads, blinding will still be very fast).
11335 * Fixed a typo bug that would cause ENGINE_set_default() to set an
11336 ENGINE as defaults for all supported algorithms irrespective of
11337 the 'flags' parameter. 'flags' is now honoured, so applications
11338 should make sure they are passing it correctly.
11342 * Target "mingw" now allows native Windows code to be generated in
11343 the Cygwin environment as well as with the MinGW compiler.
11347 ### Changes between 0.9.7 and 0.9.7a [19 Feb 2003]
11349 * In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
11350 via timing by performing a MAC computation even if incorrect
11351 block cipher padding has been found. This is a countermeasure
11352 against active attacks where the attacker has to distinguish
11353 between bad padding and a MAC verification error. ([CVE-2003-0078])
11355 *Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
11356 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
11357 Martin Vuagnoux (EPFL, Ilion)*
11359 * Make the no-err option work as intended. The intention with no-err
11360 is not to have the whole error stack handling routines removed from
11361 libcrypto, it's only intended to remove all the function name and
11362 reason texts, thereby removing some of the footprint that may not
11363 be interesting if those errors aren't displayed anyway.
11365 NOTE: it's still possible for any application or module to have its
11366 own set of error texts inserted. The routines are there, just not
11367 used by default when no-err is given.
11371 * Add support for FreeBSD on IA64.
11373 *dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454*
11375 * Adjust DES_cbc_cksum() so it returns the same value as the MIT
11376 Kerberos function mit_des_cbc_cksum(). Before this change,
11377 the value returned by DES_cbc_cksum() was like the one from
11378 mit_des_cbc_cksum(), except the bytes were swapped.
11380 *Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte*
11382 * Allow an application to disable the automatic SSL chain building.
11383 Before this a rather primitive chain build was always performed in
11384 ssl3_output_cert_chain(): an application had no way to send the
11385 correct chain if the automatic operation produced an incorrect result.
11387 Now the chain builder is disabled if either:
11389 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert().
11391 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set.
11393 The reasoning behind this is that an application would not want the
11394 auto chain building to take place if extra chain certificates are
11395 present and it might also want a means of sending no additional
11396 certificates (for example the chain has two certificates and the
11401 * Add the possibility to build without the ENGINE framework.
11403 *Steven Reddie <smr@essemer.com.au> via Richard Levitte*
11405 * Under Win32 gmtime() can return NULL: check return value in
11406 OPENSSL_gmtime(). Add error code for case where gmtime() fails.
11410 * DSA routines: under certain error conditions uninitialized BN objects
11411 could be freed. Solution: make sure initialization is performed early
11412 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>,
11413 Nils Larsch <nla@trustcenter.de> via PR#459)
11417 * Another fix for SSLv2 session ID handling: the session ID was incorrectly
11418 checked on reconnect on the client side, therefore session resumption
11419 could still fail with a "ssl session id is different" error. This
11420 behaviour is masked when SSL_OP_ALL is used due to
11421 SSL_OP_MICROSOFT_SESS_ID_BUG being set.
11422 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
11423 followup to PR #377.
11427 * IA-32 assembler support enhancements: unified ELF targets, support
11428 for SCO/Caldera platforms, fix for Cygwin shared build.
11432 * Add support for FreeBSD on sparc64. As a consequence, support for
11433 FreeBSD on non-x86 processors is separate from x86 processors on
11434 the config script, much like the NetBSD support.
11436 *Richard Levitte & Kris Kennaway <kris@obsecurity.org>*
11438 ### Changes between 0.9.6h and 0.9.7 [31 Dec 2002]
11440 [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after
11443 * Fix session ID handling in SSLv2 client code: the SERVER FINISHED
11444 code (06) was taken as the first octet of the session ID and the last
11445 octet was ignored consequently. As a result SSLv2 client side session
11446 caching could not have worked due to the session ID mismatch between
11448 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
11453 * Change the declaration of needed Kerberos libraries to use EX_LIBS
11454 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is
11459 * The hw_ncipher.c engine requires dynamic locks. Unfortunately, it
11460 seems that in spite of existing for more than a year, many application
11461 author have done nothing to provide the necessary callbacks, which
11462 means that this particular engine will not work properly anywhere.
11463 This is a very unfortunate situation which forces us, in the name
11464 of usability, to give the hw_ncipher.c a static lock, which is part
11466 NOTE: This is for the 0.9.7 series ONLY. This hack will never
11467 appear in 0.9.8 or later. We EXPECT application authors to have
11468 dealt properly with this when 0.9.8 is released (unless we actually
11469 make such changes in the libcrypto locking code that changes will
11470 have to be made anyway).
11474 * In asn1_d2i_read_bio() repeatedly call BIO_read() until all content
11475 octets have been read, EOF or an error occurs. Without this change
11476 some truncated ASN1 structures will not produce an error.
11480 * Disable Heimdal support, since it hasn't been fully implemented.
11481 Still give the possibility to force the use of Heimdal, but with
11482 warnings and a request that patches get sent to openssl-dev.
11486 * Add the VC-CE target, introduce the WINCE sysname, and add
11487 INSTALL.WCE and appropriate conditionals to make it build.
11489 *Steven Reddie <smr@essemer.com.au> via Richard Levitte*
11491 * Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and
11492 cygssl-x.y.z.dll, where x, y and z are the major, minor and
11493 edit numbers of the version.
11495 *Corinna Vinschen <vinschen@redhat.com> and Richard Levitte*
11497 * Introduce safe string copy and catenation functions
11498 (BUF_strlcpy() and BUF_strlcat()).
11500 *Ben Laurie (CHATS) and Richard Levitte*
11502 * Avoid using fixed-size buffers for one-line DNs.
11504 *Ben Laurie (CHATS)*
11506 * Add BUF_MEM_grow_clean() to avoid information leakage when
11507 resizing buffers containing secrets, and use where appropriate.
11509 *Ben Laurie (CHATS)*
11511 * Avoid using fixed size buffers for configuration file location.
11513 *Ben Laurie (CHATS)*
11515 * Avoid filename truncation for various CA files.
11517 *Ben Laurie (CHATS)*
11519 * Use sizeof in preference to magic numbers.
11521 *Ben Laurie (CHATS)*
11523 * Avoid filename truncation in cert requests.
11525 *Ben Laurie (CHATS)*
11527 * Add assertions to check for (supposedly impossible) buffer
11530 *Ben Laurie (CHATS)*
11532 * Don't cache truncated DNS entries in the local cache (this could
11533 potentially lead to a spoofing attack).
11535 *Ben Laurie (CHATS)*
11537 * Fix various buffers to be large enough for hex/decimal
11538 representations in a platform independent manner.
11540 *Ben Laurie (CHATS)*
11542 * Add CRYPTO_realloc_clean() to avoid information leakage when
11543 resizing buffers containing secrets, and use where appropriate.
11545 *Ben Laurie (CHATS)*
11547 * Add BIO_indent() to avoid much slightly worrying code to do
11550 *Ben Laurie (CHATS)*
11552 * Convert sprintf()/BIO_puts() to BIO_printf().
11554 *Ben Laurie (CHATS)*
11556 * buffer_gets() could terminate with the buffer only half
11559 *Ben Laurie (CHATS)*
11561 * Add assertions to prevent user-supplied crypto functions from
11562 overflowing internal buffers by having large block sizes, etc.
11564 *Ben Laurie (CHATS)*
11566 * New OPENSSL_assert() macro (similar to assert(), but enabled
11569 *Ben Laurie (CHATS)*
11571 * Eliminate unused copy of key in RC4.
11573 *Ben Laurie (CHATS)*
11575 * Eliminate unused and incorrectly sized buffers for IV in pem.h.
11577 *Ben Laurie (CHATS)*
11579 * Fix off-by-one error in EGD path.
11581 *Ben Laurie (CHATS)*
11583 * If RANDFILE path is too long, ignore instead of truncating.
11585 *Ben Laurie (CHATS)*
11587 * Eliminate unused and incorrectly sized X.509 structure
11590 *Ben Laurie (CHATS)*
11592 * Eliminate unused and dangerous function knumber().
11594 *Ben Laurie (CHATS)*
11596 * Eliminate unused and dangerous structure, KSSL_ERR.
11598 *Ben Laurie (CHATS)*
11600 * Protect against overlong session ID context length in an encoded
11601 session object. Since these are local, this does not appear to be
11604 *Ben Laurie (CHATS)*
11606 * Change from security patch (see 0.9.6e below) that did not affect
11607 the 0.9.6 release series:
11609 Remote buffer overflow in SSL3 protocol - an attacker could
11610 supply an oversized master key in Kerberos-enabled versions.
11613 *Ben Laurie (CHATS)*
11615 * Change the SSL kerb5 codes to match RFC 2712.
11619 * Make -nameopt work fully for req and add -reqopt switch.
11621 *Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson*
11623 * The "block size" for block ciphers in CFB and OFB mode should be 1.
11625 *Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>*
11627 * Make sure tests can be performed even if the corresponding algorithms
11628 have been removed entirely. This was also the last step to make
11629 OpenSSL compilable with DJGPP under all reasonable conditions.
11631 *Richard Levitte, Doug Kaufman <dkaufman@rahul.net>*
11633 * Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT
11634 to allow version independent disabling of normally unselected ciphers,
11635 which may be activated as a side-effect of selecting a single cipher.
11637 (E.g., cipher list string "RSA" enables ciphersuites that are left
11638 out of "ALL" because they do not provide symmetric encryption.
11639 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.)
11641 *Lutz Jaenicke, Bodo Moeller*
11643 * Add appropriate support for separate platform-dependent build
11644 directories. The recommended way to make a platform-dependent
11645 build directory is the following (tested on Linux), maybe with
11648 # Place yourself outside of the OpenSSL source tree. In
11649 # this example, the environment variable OPENSSL_SOURCE
11650 # is assumed to contain the absolute OpenSSL source directory.
11651 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`"
11652 cd objtree/"`uname -s`-`uname -r`-`uname -m`"
11653 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do
11654 mkdir -p `dirname $F`
11655 ln -s $OPENSSL_SOURCE/$F $F
11658 To be absolutely sure not to disturb the source tree, a "make clean"
11659 is a good thing. If it isn't successful, don't worry about it,
11660 it probably means the source directory is very clean.
11664 * Make sure any ENGINE control commands make local copies of string
11665 pointers passed to them whenever necessary. Otherwise it is possible
11666 the caller may have overwritten (or deallocated) the original string
11667 data when a later ENGINE operation tries to use the stored values.
11669 *Götz Babin-Ebell <babinebell@trustcenter.de>*
11671 * Improve diagnostics in file reading and command-line digests.
11673 *Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>*
11675 * Add AES modes CFB and OFB to the object database. Correct an
11676 error in AES-CFB decryption.
11680 * Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this
11681 allows existing EVP_CIPHER_CTX structures to be reused after
11682 calling `EVP_*Final()`. This behaviour is used by encryption
11683 BIOs and some applications. This has the side effect that
11684 applications must explicitly clean up cipher contexts with
11685 EVP_CIPHER_CTX_cleanup() or they will leak memory.
11689 * Check the values of dna and dnb in bn_mul_recursive before calling
11690 bn_mul_comba (a non zero value means the a or b arrays do not contain
11691 n2 elements) and fallback to bn_mul_normal if either is not zero.
11695 * Fix escaping of non-ASCII characters when using the -subj option
11696 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>)
11700 * Make object definitions compliant to LDAP (RFC2256): SN is the short
11701 form for "surname", serialNumber has no short form.
11702 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798;
11703 therefore remove "mail" short name for "internet 7".
11704 The OID for unique identifiers in X509 certificates is
11705 x500UniqueIdentifier, not uniqueIdentifier.
11706 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>)
11710 * Add an "init" command to the ENGINE config module and auto initialize
11711 ENGINEs. Without any "init" command the ENGINE will be initialized
11712 after all ctrl commands have been executed on it. If init=1 the
11713 ENGINE is initialized at that point (ctrls before that point are run
11714 on the uninitialized ENGINE and after on the initialized one). If
11715 init=0 then the ENGINE will not be initialized at all.
11719 * Fix the 'app_verify_callback' interface so that the user-defined
11720 argument is actually passed to the callback: In the
11721 SSL_CTX_set_cert_verify_callback() prototype, the callback
11722 declaration has been changed from
11725 int (*cb)(X509_STORE_CTX *,void *);
11726 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
11727 i=s->ctx->app_verify_callback(&ctx)
11728 has been changed into
11729 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
11731 To update applications using SSL_CTX_set_cert_verify_callback(),
11732 a dummy argument can be added to their callback functions.
11734 *D. K. Smetters <smetters@parc.xerox.com>*
11736 * Added the '4758cca' ENGINE to support IBM 4758 cards.
11738 *Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe*
11740 * Add and OPENSSL_LOAD_CONF define which will cause
11741 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
11742 This allows older applications to transparently support certain
11743 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
11744 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
11745 load the config file and OPENSSL_add_all_algorithms_conf() which will
11746 always load it have also been added.
11750 * Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
11751 Adjust NIDs and EVP layer.
11753 *Stephen Sprunk <stephen@sprunk.org> and Richard Levitte*
11755 * Config modules support in openssl utility.
11757 Most commands now load modules from the config file,
11758 though in a few (such as version) this isn't done
11759 because it couldn't be used for anything.
11761 In the case of ca and req the config file used is
11762 the same as the utility itself: that is the -config
11763 command line option can be used to specify an
11768 * Move default behaviour from OPENSSL_config(). If appname is NULL
11769 use "openssl_conf" if filename is NULL use default openssl config file.
11773 * Add an argument to OPENSSL_config() to allow the use of an alternative
11774 config section name. Add a new flag to tolerate a missing config file
11775 and move code to CONF_modules_load_file().
11779 * Support for crypto accelerator cards from Accelerated Encryption
11780 Processing, www.aep.ie. (Use engine 'aep')
11781 The support was copied from 0.9.6c [engine] and adapted/corrected
11782 to work with the new engine framework.
11784 *AEP Inc. and Richard Levitte*
11786 * Support for SureWare crypto accelerator cards from Baltimore
11787 Technologies. (Use engine 'sureware')
11788 The support was copied from 0.9.6c [engine] and adapted
11789 to work with the new engine framework.
11793 * Have the CHIL engine fork-safe (as defined by nCipher) and actually
11794 make the newer ENGINE framework commands for the CHIL engine work.
11796 *Toomas Kiisk <vix@cyber.ee> and Richard Levitte*
11798 * Make it possible to produce shared libraries on ReliantUNIX.
11800 *Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte*
11802 * Add the configuration target debug-linux-ppro.
11803 Make 'openssl rsa' use the general key loading routines
11804 implemented in `apps.c`, and make those routines able to
11805 handle the key format FORMAT_NETSCAPE and the variant
11808 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
11810 * Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
11812 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
11814 * Add -keyform to rsautl, and document -engine.
11816 *Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>*
11818 * Change BIO_new_file (crypto/bio/bss_file.c) to use new
11819 BIO_R_NO_SUCH_FILE error code rather than the generic
11820 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
11824 * Add new functions
11825 ERR_peek_last_error
11826 ERR_peek_last_error_line
11827 ERR_peek_last_error_line_data.
11828 These are similar to
11830 ERR_peek_error_line
11831 ERR_peek_error_line_data,
11832 but report on the latest error recorded rather than the first one
11833 still in the error queue.
11835 *Ben Laurie, Bodo Moeller*
11837 * default_algorithms option in ENGINE config module. This allows things
11839 default_algorithms = ALL
11840 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
11844 * Preliminary ENGINE config module.
11848 * New experimental application configuration code.
11852 * Change the AES code to follow the same name structure as all other
11853 symmetric ciphers, and behave the same way. Move everything to
11854 the directory crypto/aes, thereby obsoleting crypto/rijndael.
11856 *Stephen Sprunk <stephen@sprunk.org> and Richard Levitte*
11858 * SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
11860 *Ben Laurie and Theo de Raadt*
11862 * Add option to output public keys in req command.
11864 *Massimiliano Pala madwolf@openca.org*
11866 * Use wNAFs in EC_POINTs_mul() for improved efficiency
11867 (up to about 10% better than before for P-192 and P-224).
11871 * New functions/macros
11873 SSL_CTX_set_msg_callback(ctx, cb)
11874 SSL_CTX_set_msg_callback_arg(ctx, arg)
11875 SSL_set_msg_callback(ssl, cb)
11876 SSL_set_msg_callback_arg(ssl, arg)
11878 to request calling a callback function
11880 void cb(int write_p, int version, int content_type,
11881 const void *buf, size_t len, SSL *ssl, void *arg)
11883 whenever a protocol message has been completely received
11884 (write_p == 0) or sent (write_p == 1). Here 'version' is the
11885 protocol version according to which the SSL library interprets
11886 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
11887 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
11888 the content type as defined in the SSL 3.0/TLS 1.0 protocol
11889 specification (change_cipher_spec(20), alert(21), handshake(22)).
11890 'buf' and 'len' point to the actual message, 'ssl' to the
11891 SSL object, and 'arg' is the application-defined value set by
11892 SSL[_CTX]_set_msg_callback_arg().
11894 'openssl s_client' and 'openssl s_server' have new '-msg' options
11895 to enable a callback that displays all protocol messages.
11899 * Change the shared library support so shared libraries are built as
11900 soon as the corresponding static library is finished, and thereby get
11901 openssl and the test programs linked against the shared library.
11902 This still only happens when the keyword "shard" has been given to
11903 the configuration scripts.
11905 NOTE: shared library support is still an experimental thing, and
11906 backward binary compatibility is still not guaranteed.
11908 *"Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte*
11910 * Add support for Subject Information Access extension.
11912 *Peter Sylvester <Peter.Sylvester@EdelWeb.fr>*
11914 * Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
11915 additional bytes when new memory had to be allocated, not just
11916 when reusing an existing buffer.
11920 * New command line and configuration option 'utf8' for the req command.
11921 This allows field values to be specified as UTF8 strings.
11925 * Add -multi and -mr options to "openssl speed" - giving multiple parallel
11926 runs for the former and machine-readable output for the latter.
11930 * Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
11931 of the e-mail address in the DN (i.e., it will go into a certificate
11932 extension only). The new configuration file option 'email_in_dn = no'
11933 has the same effect.
11935 *Massimiliano Pala madwolf@openca.org*
11937 * Change all functions with names starting with `des_` to be starting
11938 with `DES_` instead. Add wrappers that are compatible with libdes,
11939 but are named `_ossl_old_des_*`. Finally, add macros that map the
11940 `des_*` symbols to the corresponding `_ossl_old_des_*` if libdes
11941 compatibility is desired. If OpenSSL 0.9.6c compatibility is
11942 desired, the `des_*` symbols will be mapped to `DES_*`, with one
11945 Since we provide two compatibility mappings, the user needs to
11946 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
11947 compatibility is desired. The default (i.e., when that macro
11948 isn't defined) is OpenSSL 0.9.6c compatibility.
11950 There are also macros that enable and disable the support of old
11951 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
11952 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
11953 are defined, the default will apply: to support the old des routines.
11955 In either case, one must include openssl/des.h to get the correct
11956 definitions. Do not try to just include openssl/des_old.h, that
11959 NOTE: This is a major break of an old API into a new one. Software
11960 authors are encouraged to switch to the `DES_` style functions. Some
11961 time in the future, des_old.h and the libdes compatibility functions
11962 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
11963 default), and then completely removed.
11967 * Test for certificates which contain unsupported critical extensions.
11968 If such a certificate is found during a verify operation it is
11969 rejected by default: this behaviour can be overridden by either
11970 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
11971 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
11972 X509_supported_extension() has also been added which returns 1 if a
11973 particular extension is supported.
11977 * Modify the behaviour of EVP cipher functions in similar way to digests
11978 to retain compatibility with existing code.
11982 * Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
11983 compatibility with existing code. In particular the 'ctx' parameter does
11984 not have to be to be initialized before the call to EVP_DigestInit() and
11985 it is tidied up after a call to EVP_DigestFinal(). New function
11986 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
11987 EVP_MD_CTX_copy() changed to not require the destination to be
11988 initialized valid and new function EVP_MD_CTX_copy_ex() added which
11989 requires the destination to be valid.
11991 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
11992 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
11996 * Change ssl3_get_message (ssl/s3_both.c) and the functions using it
11997 so that complete 'Handshake' protocol structures are kept in memory
11998 instead of overwriting 'msg_type' and 'length' with 'body' data.
12002 * Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
12004 *Massimo Santin via Richard Levitte*
12006 * Major restructuring to the underlying ENGINE code. This includes
12007 reduction of linker bloat, separation of pure "ENGINE" manipulation
12008 (initialisation, etc) from functionality dealing with implementations
12009 of specific crypto interfaces. This change also introduces integrated
12010 support for symmetric ciphers and digest implementations - so ENGINEs
12011 can now accelerate these by providing EVP_CIPHER and EVP_MD
12012 implementations of their own. This is detailed in
12013 [crypto/engine/README.md](crypto/engine/README.md)
12014 as it couldn't be adequately described here. However, there are a few
12015 API changes worth noting - some RSA, DSA, DH, and RAND functions that
12016 were changed in the original introduction of ENGINE code have now
12017 reverted back - the hooking from this code to ENGINE is now a good
12018 deal more passive and at run-time, operations deal directly with
12019 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
12020 dereferencing through an ENGINE pointer any more. Also, the ENGINE
12021 functions dealing with `BN_MOD_EXP[_CRT]` handlers have been removed -
12022 they were not being used by the framework as there is no concept of a
12023 BIGNUM_METHOD and they could not be generalised to the new
12024 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
12025 ENGINE_cpy() has been removed as it cannot be consistently defined in
12030 * Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
12034 * Change mkdef.pl to sort symbols that get the same entry number,
12035 and make sure the automatically generated functions `ERR_load_*`
12036 become part of libeay.num as well.
12040 * New function SSL_renegotiate_pending(). This returns true once
12041 renegotiation has been requested (either SSL_renegotiate() call
12042 or HelloRequest/ClientHello received from the peer) and becomes
12043 false once a handshake has been completed.
12044 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
12045 sends a HelloRequest, but does not ensure that a handshake takes
12046 place. SSL_renegotiate_pending() is useful for checking if the
12047 client has followed the request.)
12051 * New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
12052 By default, clients may request session resumption even during
12053 renegotiation (if session ID contexts permit); with this option,
12054 session resumption is possible only in the first handshake.
12056 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes
12057 more bits available for options that should not be part of
12058 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION).
12062 * Add some demos for certificate and certificate request creation.
12066 * Make maximum certificate chain size accepted from the peer application
12067 settable (`SSL*_get/set_max_cert_list()`), as proposed by
12068 "Douglas E. Engert" <deengert@anl.gov>.
12072 * Add support for shared libraries for Unixware-7
12073 (Boyd Lynn Gerber <gerberb@zenez.com>).
12077 * Add a "destroy" handler to ENGINEs that allows structural cleanup to
12078 be done prior to destruction. Use this to unload error strings from
12079 ENGINEs that load their own error strings. NB: This adds two new API
12080 functions to "get" and "set" this destroy handler in an ENGINE.
12084 * Alter all existing ENGINE implementations (except "openssl" and
12085 "openbsd") to dynamically instantiate their own error strings. This
12086 makes them more flexible to be built both as statically-linked ENGINEs
12087 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
12088 Also, add stub code to each that makes building them as self-contained
12089 shared-libraries easier (see [README-Engine.md](README-Engine.md)).
12093 * Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
12094 implementations into applications that are completely implemented in
12095 self-contained shared-libraries. The "dynamic" ENGINE exposes control
12096 commands that can be used to configure what shared-library to load and
12097 to control aspects of the way it is handled. Also, made an update to
12098 the [README-Engine.md](README-Engine.md) file
12099 that brings its information up-to-date and
12100 provides some information and instructions on the "dynamic" ENGINE
12101 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
12105 * Make it possible to unload ranges of ERR strings with a new
12106 "ERR_unload_strings" function.
12110 * Add a copy() function to EVP_MD.
12114 * Make EVP_MD routines take a context pointer instead of just the
12115 md_data void pointer.
12119 * Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
12120 that the digest can only process a single chunk of data
12121 (typically because it is provided by a piece of
12122 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
12123 is only going to provide a single chunk of data, and hence the
12124 framework needn't accumulate the data for oneshot drivers.
12128 * As with "ERR", make it possible to replace the underlying "ex_data"
12129 functions. This change also alters the storage and management of global
12130 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
12131 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
12132 index counters. The API functions that use this state have been changed
12133 to take a "class_index" rather than pointers to the class's local STACK
12134 and counter, and there is now an API function to dynamically create new
12135 classes. This centralisation allows us to (a) plug a lot of the
12136 thread-safety problems that existed, and (b) makes it possible to clean
12137 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
12138 such data would previously have always leaked in application code and
12139 workarounds were in place to make the memory debugging turn a blind eye
12140 to it. Application code that doesn't use this new function will still
12141 leak as before, but their memory debugging output will announce it now
12142 rather than letting it slide.
12144 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
12145 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
12146 has a return value to indicate success or failure.
12150 * Make it possible to replace the underlying "ERR" functions such that the
12151 global state (2 LHASH tables and 2 locks) is only used by the "default"
12152 implementation. This change also adds two functions to "get" and "set"
12153 the implementation prior to it being automatically set the first time
12154 any other ERR function takes place. Ie. an application can call "get",
12155 pass the return value to a module it has just loaded, and that module
12156 can call its own "set" function using that value. This means the
12157 module's "ERR" operations will use (and modify) the error state in the
12158 application and not in its own statically linked copy of OpenSSL code.
12162 * Give DH, DSA, and RSA types their own `*_up_ref()` function to increment
12163 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
12164 the operation, and provides a more encapsulated way for external code
12165 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
12166 to use these functions rather than manually incrementing the counts.
12168 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
12172 * Add EVP test program.
12176 * Add symmetric cipher support to ENGINE. Expect the API to change!
12180 * New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
12181 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
12182 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
12183 These allow a CRL to be built without having to access X509_CRL fields
12184 directly. Modify 'ca' application to use new functions.
12188 * Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
12189 bug workarounds. Rollback attack detection is a security feature.
12190 The problem will only arise on OpenSSL servers when TLSv1 is not
12191 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
12192 Software authors not wanting to support TLSv1 will have special reasons
12193 for their choice and can explicitly enable this option.
12195 *Bodo Moeller, Lutz Jaenicke*
12197 * Rationalise EVP so it can be extended: don't include a union of
12198 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX
12199 (similar to those existing for EVP_CIPHER_CTX).
12204 EVP_MD_CTX_init(&md); /* new function call */
12205 EVP_DigestInit(&md, EVP_sha1());
12206 EVP_DigestUpdate(&md, in, len);
12207 EVP_DigestFinal(&md, out, NULL);
12208 EVP_MD_CTX_cleanup(&md); /* new function call */
12212 * Make DES key schedule conform to the usual scheme, as well as
12213 correcting its structure. This means that calls to DES functions
12214 now have to pass a pointer to a des_key_schedule instead of a
12215 plain des_key_schedule (which was actually always a pointer
12218 des_key_schedule ks;
12220 des_set_key_checked(..., &ks);
12221 des_ncbc_encrypt(..., &ks, ...);
12223 (Note that a later change renames 'des_...' into 'DES_...'.)
12227 * Initial reduction of linker bloat: the use of some functions, such as
12228 PEM causes large amounts of unused functions to be linked in due to
12229 poor organisation. For example pem_all.c contains every PEM function
12230 which has a knock on effect of linking in large amounts of (unused)
12231 ASN1 code. Grouping together similar functions and splitting unrelated
12232 functions prevents this.
12236 * Cleanup of EVP macros.
12240 * Change historical references to `{NID,SN,LN}_des_ede` and ede3 to add the
12241 correct `_ecb suffix`.
12245 * Add initial OCSP responder support to ocsp application. The
12246 revocation information is handled using the text based index
12247 use by the ca application. The responder can either handle
12248 requests generated internally, supplied in files (for example
12249 via a CGI script) or using an internal minimal server.
12253 * Add configuration choices to get zlib compression for TLS.
12257 * Changes to Kerberos SSL for RFC 2712 compliance:
12258 1. Implemented real KerberosWrapper, instead of just using
12259 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
12260 2. Implemented optional authenticator field of KerberosWrapper.
12262 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
12263 and authenticator structs; see crypto/krb5/.
12265 Generalized Kerberos calls to support multiple Kerberos libraries.
12266 *Vern Staats <staatsvr@asc.hpc.mil>, Jeffrey Altman <jaltman@columbia.edu>
12267 via Richard Levitte*
12269 * Cause 'openssl speed' to use fully hard-coded DSA keys as it
12270 already does with RSA. testdsa.h now has 'priv_key/pub_key'
12271 values for each of the key sizes rather than having just
12272 parameters (and 'speed' generating keys each time).
12276 * Speed up EVP routines.
12279 pe 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
12280 s-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
12281 s-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
12282 s-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
12284 s-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
12285 s-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
12286 s-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
12289 s-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
12291 s-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
12295 * Added the OS2-EMX target.
12297 *"Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte*
12299 * Rewrite commands to use `NCONF` routines instead of the old `CONF`.
12300 New functions to support `NCONF` routines in extension code.
12301 New function `CONF_set_nconf()`
12302 to allow functions which take an `NCONF` to also handle the old `LHASH`
12303 structure: this means that the old `CONF` compatible routines can be
12304 retained (in particular w.rt. extensions) without having to duplicate the
12305 code. New function `X509V3_add_ext_nconf_sk()` to add extensions to a stack.
12309 * Enhance the general user interface with mechanisms for inner control
12310 and with possibilities to have yes/no kind of prompts.
12314 * Change all calls to low-level digest routines in the library and
12315 applications to use EVP. Add missing calls to HMAC_cleanup() and
12316 don't assume HMAC_CTX can be copied using memcpy().
12318 *Verdon Walker <VWalker@novell.com>, Steve Henson*
12320 * Add the possibility to control engines through control names but with
12321 arbitrary arguments instead of just a string.
12322 Change the key loaders to take a UI_METHOD instead of a callback
12323 function pointer. NOTE: this breaks binary compatibility with earlier
12324 versions of OpenSSL [engine].
12325 Adapt the nCipher code for these new conditions and add a card insertion
12330 * Enhance the general user interface with mechanisms to better support
12331 dialog box interfaces, application-defined prompts, the possibility
12332 to use defaults (for example default passwords from somewhere else)
12333 and interrupts/cancellations.
12337 * Tidy up PKCS#12 attribute handling. Add support for the CSP name
12338 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
12342 * Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
12343 tidy up some unnecessarily weird code in 'sk_new()').
12345 *Geoff, reported by Diego Tartara <dtartara@novamens.com>*
12347 * Change the key loading routines for ENGINEs to use the same kind
12348 callback (pem_password_cb) as all other routines that need this
12353 * Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
12354 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
12355 than this minimum value is recommended.
12359 * New random seeder for OpenVMS, using the system process statistics
12360 that are easily reachable.
12364 * Windows apparently can't transparently handle global
12365 variables defined in DLLs. Initialisations such as:
12367 const ASN1_ITEM *it = &ASN1_INTEGER_it;
12369 won't compile. This is used by the any applications that need to
12370 declare their own ASN1 modules. This was fixed by adding the option
12371 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
12372 needed for static libraries under Win32.
12376 * New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
12377 setting of purpose and trust fields. New X509_STORE trust and
12378 purpose functions and tidy up setting in other SSL functions.
12382 * Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
12383 structure. These are inherited by X509_STORE_CTX when it is
12384 initialised. This allows various defaults to be set in the
12385 X509_STORE structure (such as flags for CRL checking and custom
12386 purpose or trust settings) for functions which only use X509_STORE_CTX
12387 internally such as S/MIME.
12389 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
12390 trust settings if they are not set in X509_STORE. This allows X509_STORE
12391 purposes and trust (in S/MIME for example) to override any set by default.
12393 Add command line options for CRL checking to smime, s_client and s_server
12398 * Initial CRL based revocation checking. If the CRL checking flag(s)
12399 are set then the CRL is looked up in the X509_STORE structure and
12400 its validity and signature checked, then if the certificate is found
12401 in the CRL the verify fails with a revoked error.
12403 Various new CRL related callbacks added to X509_STORE_CTX structure.
12405 Command line options added to 'verify' application to support this.
12407 This needs some additional work, such as being able to handle multiple
12408 CRLs with different times, extension based lookup (rather than just
12409 by subject name) and ultimately more complete V2 CRL extension
12414 * Add a general user interface API (crypto/ui/). This is designed
12415 to replace things like des_read_password and friends (backward
12416 compatibility functions using this new API are provided).
12417 The purpose is to remove prompting functions from the DES code
12418 section as well as provide for prompting through dialog boxes in
12419 a window system and the like.
12423 * Add "ex_data" support to ENGINE so implementations can add state at a
12424 per-structure level rather than having to store it globally.
12428 * Make it possible for ENGINE structures to be copied when retrieved by
12429 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
12430 This causes the "original" ENGINE structure to act like a template,
12431 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
12432 operational state can be localised to each ENGINE structure, despite the
12433 fact they all share the same "methods". New ENGINE structures returned in
12434 this case have no functional references and the return value is the single
12435 structural reference. This matches the single structural reference returned
12436 by ENGINE_by_id() normally, when it is incremented on the pre-existing
12441 * Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
12442 needs to match any other type at all we need to manually clear the
12447 * Changes to the "openssl engine" utility to include;
12448 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
12449 about an ENGINE's available control commands.
12450 - executing control commands from command line arguments using the
12451 '-pre' and '-post' switches. '-post' is only used if '-t' is
12452 specified and the ENGINE is successfully initialised. The syntax for
12453 the individual commands are colon-separated, for example;
12454 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
12458 * New dynamic control command support for ENGINEs. ENGINEs can now
12459 declare their own commands (numbers), names (strings), descriptions,
12460 and input types for run-time discovery by calling applications. A
12461 subset of these commands are implicitly classed as "executable"
12462 depending on their input type, and only these can be invoked through
12463 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
12464 can be based on user input, config files, etc). The distinction is
12465 that "executable" commands cannot return anything other than a boolean
12466 result and can only support numeric or string input, whereas some
12467 discoverable commands may only be for direct use through
12468 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
12469 pointers, or other custom uses. The "executable" commands are to
12470 support parameterisations of ENGINE behaviour that can be
12471 unambiguously defined by ENGINEs and used consistently across any
12472 OpenSSL-based application. Commands have been added to all the
12473 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
12474 control over shared-library paths without source code alterations.
12478 * Changed all ENGINE implementations to dynamically allocate their
12479 ENGINEs rather than declaring them statically. Apart from this being
12480 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
12481 this also allows the implementations to compile without using the
12482 internal engine_int.h header.
12486 * Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
12487 'const' value. Any code that should be able to modify a RAND_METHOD
12488 should already have non-const pointers to it (ie. they should only
12489 modify their own ones).
12493 * Made a variety of little tweaks to the ENGINE code.
12494 - "atalla" and "ubsec" string definitions were moved from header files
12495 to C code. "nuron" string definitions were placed in variables
12496 rather than hard-coded - allowing parameterisation of these values
12497 later on via ctrl() commands.
12498 - Removed unused "#if 0"'d code.
12499 - Fixed engine list iteration code so it uses ENGINE_free() to release
12500 structural references.
12501 - Constified the RAND_METHOD element of ENGINE structures.
12502 - Constified various get/set functions as appropriate and added
12503 missing functions (including a catch-all ENGINE_cpy that duplicates
12504 all ENGINE values onto a new ENGINE except reference counts/state).
12505 - Removed NULL parameter checks in get/set functions. Setting a method
12506 or function to NULL is a way of cancelling out a previously set
12507 value. Passing a NULL ENGINE parameter is just plain stupid anyway
12508 and doesn't justify the extra error symbols and code.
12509 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
12510 flags from engine_int.h to engine.h.
12511 - Changed prototypes for ENGINE handler functions (init(), finish(),
12512 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
12516 * Implement binary inversion algorithm for BN_mod_inverse in addition
12517 to the algorithm using long division. The binary algorithm can be
12518 used only if the modulus is odd. On 32-bit systems, it is faster
12519 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
12520 roughly 5-15% for 256-bit moduli), so we use it only for moduli
12521 up to 450 bits. In 64-bit environments, the binary algorithm
12522 appears to be advantageous for much longer moduli; here we use it
12523 for moduli up to 2048 bits.
12527 * Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
12528 could not support the combine flag in choice fields.
12532 * Add a 'copy_extensions' option to the 'ca' utility. This copies
12533 extensions from a certificate request to the certificate.
12537 * Allow multiple 'certopt' and 'nameopt' options to be separated
12538 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
12539 file: this allows the display of the certificate about to be
12540 signed to be customised, to allow certain fields to be included
12541 or excluded and extension details. The old system didn't display
12542 multicharacter strings properly, omitted fields not in the policy
12543 and couldn't display additional details such as extensions.
12547 * Function EC_POINTs_mul for multiple scalar multiplication
12548 of an arbitrary number of elliptic curve points
12549 \sum scalars[i]*points[i],
12550 optionally including the generator defined for the EC_GROUP:
12551 scalar*generator + \sum scalars[i]*points[i].
12553 EC_POINT_mul is a simple wrapper function for the typical case
12554 that the point list has just one item (besides the optional
12559 * First EC_METHODs for curves over GF(p):
12561 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
12562 operations and provides various method functions that can also
12563 operate with faster implementations of modular arithmetic.
12565 EC_GFp_mont_method() reuses most functions that are part of
12566 EC_GFp_simple_method, but uses Montgomery arithmetic.
12568 *Bodo Moeller; point addition and point doubling
12569 implementation directly derived from source code provided by
12570 Lenka Fibikova <fibikova@exp-math.uni-essen.de>*
12572 * Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
12573 crypto/ec/ec_lib.c):
12575 Curves are EC_GROUP objects (with an optional group generator)
12576 based on EC_METHODs that are built into the library.
12578 Points are EC_POINT objects based on EC_GROUP objects.
12580 Most of the framework would be able to handle curves over arbitrary
12581 finite fields, but as there are no obvious types for fields other
12582 than GF(p), some functions are limited to that for now.
12586 * Add the -HTTP option to s_server. It is similar to -WWW, but requires
12587 that the file contains a complete HTTP response.
12591 * Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
12592 change the def and num file printf format specifier from "%-40sXXX"
12593 to "%-39s XXX". The latter will always guarantee a space after the
12594 field while the former will cause them to run together if the field
12595 is 40 of more characters long.
12599 * Constify the cipher and digest 'method' functions and structures
12600 and modify related functions to take constant EVP_MD and EVP_CIPHER
12605 * Hide BN_CTX structure details in bn_lcl.h instead of publishing them
12606 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
12610 * Modify `EVP_Digest*()` routines so they now return values. Although the
12611 internal software routines can never fail additional hardware versions
12616 * Clean up crypto/err/err.h and change some error codes to avoid conflicts:
12618 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
12619 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
12622 ERR_R_NESTED_ASN1_ERROR
12624 ERR_R_MISSING_ASN1_EOS
12625 were 4 .. 9, conflicting with
12626 ERR_LIB_RSA (= ERR_R_RSA_LIB)
12628 ERR_LIB_PEM (= ERR_R_PEM_LIB).
12629 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
12631 Add new error code 'ERR_R_INTERNAL_ERROR'.
12635 * Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
12640 * New option '-subj arg' for 'openssl req' and 'openssl ca'. This
12641 sets the subject name for a new request or supersedes the
12642 subject name in a given request. Formats that can be parsed are
12643 'CN=Some Name, OU=myOU, C=IT'
12645 'CN=Some Name/OU=myOU/C=IT'.
12647 Add options '-batch' and '-verbose' to 'openssl req'.
12649 *Massimiliano Pala <madwolf@hackmasters.net>*
12651 * Introduce the possibility to access global variables through
12652 functions on platform were that's the best way to handle exporting
12653 global variables in shared libraries. To enable this functionality,
12654 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
12655 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
12656 is normally done by Configure or something similar).
12658 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
12659 in the source file (foo.c) like this:
12661 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
12662 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
12664 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
12665 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
12667 OPENSSL_DECLARE_GLOBAL(int,foo);
12668 #define foo OPENSSL_GLOBAL_REF(foo)
12669 OPENSSL_DECLARE_GLOBAL(double,bar);
12670 #define bar OPENSSL_GLOBAL_REF(bar)
12672 The #defines are very important, and therefore so is including the
12673 header file everywhere where the defined globals are used.
12675 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
12676 of ASN.1 items, but that structure is a bit different.
12678 The largest change is in util/mkdef.pl which has been enhanced with
12679 better and easier to understand logic to choose which symbols should
12680 go into the Windows .def files as well as a number of fixes and code
12681 cleanup (among others, algorithm keywords are now sorted
12682 lexicographically to avoid constant rewrites).
12686 * In BN_div() keep a copy of the sign of 'num' before writing the
12687 result to 'rm' because if rm==num the value will be overwritten
12688 and produce the wrong result if 'num' is negative: this caused
12689 problems with BN_mod() and BN_nnmod().
12693 * Function OCSP_request_verify(). This checks the signature on an
12694 OCSP request and verifies the signer certificate. The signer
12695 certificate is just checked for a generic purpose and OCSP request
12700 * Add OCSP_check_validity() function to check the validity of OCSP
12701 responses. OCSP responses are prepared in real time and may only
12702 be a few seconds old. Simply checking that the current time lies
12703 between thisUpdate and nextUpdate max reject otherwise valid responses
12704 caused by either OCSP responder or client clock inaccuracy. Instead
12705 we allow thisUpdate and nextUpdate to fall within a certain period of
12706 the current time. The age of the response can also optionally be
12707 checked. Two new options -validity_period and -status_age added to
12712 * If signature or public key algorithm is unrecognized print out its
12713 OID rather that just UNKNOWN.
12717 * Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
12718 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
12719 ID to be generated from the issuer certificate alone which can then be
12720 passed to OCSP_id_issuer_cmp().
12724 * New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
12725 ASN1 modules to export functions returning ASN1_ITEM pointers
12726 instead of the ASN1_ITEM structures themselves. This adds several
12727 new macros which allow the underlying ASN1 function/structure to
12728 be accessed transparently. As a result code should not use ASN1_ITEM
12729 references directly (such as &X509_it) but instead use the relevant
12730 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
12731 use of the new ASN1 code on platforms where exporting structures
12732 is problematical (for example in shared libraries) but exporting
12733 functions returning pointers to structures is not.
12737 * Add support for overriding the generation of SSL/TLS session IDs.
12738 These callbacks can be registered either in an SSL_CTX or per SSL.
12739 The purpose of this is to allow applications to control, if they wish,
12740 the arbitrary values chosen for use as session IDs, particularly as it
12741 can be useful for session caching in multiple-server environments. A
12742 command-line switch for testing this (and any client code that wishes
12743 to use such a feature) has been added to "s_server".
12745 *Geoff Thorpe, Lutz Jaenicke*
12747 * Modify mkdef.pl to recognise and parse preprocessor conditionals
12748 of the form `#if defined(...) || defined(...) || ...` and
12749 `#if !defined(...) && !defined(...) && ...`. This also avoids
12750 the growing number of special cases it was previously handling.
12754 * Make all configuration macros available for application by making
12755 sure they are available in opensslconf.h, by giving them names starting
12756 with `OPENSSL_` to avoid conflicts with other packages and by making
12757 sure e_os2.h will cover all platform-specific cases together with
12759 Additionally, it is now possible to define configuration/platform-
12760 specific names (called "system identities"). In the C code, these
12761 are prefixed with `OPENSSL_SYSNAME_`. e_os2.h will create another
12762 macro with the name beginning with `OPENSSL_SYS_`, which is determined
12763 from `OPENSSL_SYSNAME_*` or compiler-specific macros depending on
12768 * New option -set_serial to 'req' and 'x509' this allows the serial
12769 number to use to be specified on the command line. Previously self
12770 signed certificates were hard coded with serial number 0 and the
12771 CA options of 'x509' had to use a serial number in a file which was
12776 * New options to 'ca' utility to support V2 CRL entry extensions.
12777 Currently CRL reason, invalidity date and hold instruction are
12778 supported. Add new CRL extensions to V3 code and some new objects.
12782 * New function EVP_CIPHER_CTX_set_padding() this is used to
12783 disable standard block padding (aka PKCS#5 padding) in the EVP
12784 API, which was previously mandatory. This means that the data is
12785 not padded in any way and so the total length much be a multiple
12786 of the block size, otherwise an error occurs.
12790 * Initial (incomplete) OCSP SSL support.
12794 * New function OCSP_parse_url(). This splits up a URL into its host,
12795 port and path components: primarily to parse OCSP URLs. New -url
12796 option to ocsp utility.
12800 * New nonce behavior. The return value of OCSP_check_nonce() now
12801 reflects the various checks performed. Applications can decide
12802 whether to tolerate certain situations such as an absent nonce
12803 in a response when one was present in a request: the ocsp application
12804 just prints out a warning. New function OCSP_add1_basic_nonce()
12805 this is to allow responders to include a nonce in a response even if
12806 the request is nonce-less.
12810 * Disable stdin buffering in `load_cert()` (`apps/apps.c`) so that no certs are
12811 skipped when using openssl x509 multiple times on a single input file,
12812 e.g. `(openssl x509 -out cert1; openssl x509 -out cert2) <certs`.
12816 * Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
12817 set string type: to handle setting ASN1_TIME structures. Fix ca
12818 utility to correctly initialize revocation date of CRLs.
12822 * New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
12823 the clients preferred ciphersuites and rather use its own preferences.
12824 Should help to work around M$ SGC (Server Gated Cryptography) bug in
12825 Internet Explorer by ensuring unchanged hash method during stepup.
12826 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
12830 * Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
12831 to aes and add a new 'exist' option to print out symbols that don't
12836 * Additional options to ocsp utility to allow flags to be set and
12837 additional certificates supplied.
12841 * Add the option -VAfile to 'openssl ocsp', so the user can give the
12842 OCSP client a number of certificate to only verify the response
12847 * Update Rijndael code to version 3.0 and change EVP AES ciphers to
12848 handle the new API. Currently only ECB, CBC modes supported. Add new
12851 Add TLS AES ciphersuites as described in RFC3268, "Advanced
12852 Encryption Standard (AES) Ciphersuites for Transport Layer
12853 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were
12854 not enabled by default and were not part of the "ALL" ciphersuite
12855 alias because they were not yet official; they could be
12856 explicitly requested by specifying the "AESdraft" ciphersuite
12857 group alias. In the final release of OpenSSL 0.9.7, the group
12858 alias is called "AES" and is part of "ALL".)
12860 *Ben Laurie, Steve Henson, Bodo Moeller*
12862 * New function OCSP_copy_nonce() to copy nonce value (if present) from
12863 request to response.
12867 * Functions for OCSP responders. OCSP_request_onereq_count(),
12868 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
12869 extract information from a certificate request. OCSP_response_create()
12870 creates a response and optionally adds a basic response structure.
12871 OCSP_basic_add1_status() adds a complete single response to a basic
12872 response and returns the OCSP_SINGLERESP structure just added (to allow
12873 extensions to be included for example). OCSP_basic_add1_cert() adds a
12874 certificate to a basic response and OCSP_basic_sign() signs a basic
12875 response with various flags. New helper functions ASN1_TIME_check()
12876 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
12877 (converts ASN1_TIME to GeneralizedTime).
12881 * Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
12882 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
12883 structure from a certificate. X509_pubkey_digest() digests the public_key
12884 contents: this is used in various key identifiers.
12888 * Make sk_sort() tolerate a NULL argument.
12890 *Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>*
12892 * New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
12893 passed by the function are trusted implicitly. If any of them signed the
12894 response then it is assumed to be valid and is not verified.
12898 * In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
12899 to data. This was previously part of the PKCS7 ASN1 code. This
12900 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
12901 *Steve Henson, reported by Kenneth R. Robinette
12902 <support@securenetterm.com>*
12904 * Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
12905 routines: without these tracing memory leaks is very painful.
12906 Fix leaks in PKCS12 and PKCS7 routines.
12910 * Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
12911 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
12912 effectively meant GeneralizedTime would never be used. Now it
12913 is initialised to -1 but X509_time_adj() now has to check the value
12914 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
12915 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
12916 *Steve Henson, reported by Kenneth R. Robinette
12917 <support@securenetterm.com>*
12919 * Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
12920 result in a zero length in the ASN1_INTEGER structure which was
12921 not consistent with the structure when d2i_ASN1_INTEGER() was used
12922 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
12923 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
12924 where it did not print out a minus for negative ASN1_INTEGER.
12928 * Add summary printout to ocsp utility. The various functions which
12929 convert status values to strings have been renamed to:
12930 OCSP_response_status_str(), OCSP_cert_status_str() and
12931 OCSP_crl_reason_str() and are no longer static. New options
12932 to verify nonce values and to disable verification. OCSP response
12933 printout format cleaned up.
12937 * Add additional OCSP certificate checks. These are those specified
12938 in RFC2560. This consists of two separate checks: the CA of the
12939 certificate being checked must either be the OCSP signer certificate
12940 or the issuer of the OCSP signer certificate. In the latter case the
12941 OCSP signer certificate must contain the OCSP signing extended key
12942 usage. This check is performed by attempting to match the OCSP
12943 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
12944 in the OCSP_CERTID structures of the response.
12948 * Initial OCSP certificate verification added to OCSP_basic_verify()
12949 and related routines. This uses the standard OpenSSL certificate
12950 verify routines to perform initial checks (just CA validity) and
12951 to obtain the certificate chain. Then additional checks will be
12952 performed on the chain. Currently the root CA is checked to see
12953 if it is explicitly trusted for OCSP signing. This is used to set
12954 a root CA as a global signing root: that is any certificate that
12955 chains to that CA is an acceptable OCSP signing certificate.
12959 * New '-extfile ...' option to 'openssl ca' for reading X.509v3
12960 extensions from a separate configuration file.
12961 As when reading extensions from the main configuration file,
12962 the '-extensions ...' option may be used for specifying the
12965 *Massimiliano Pala <madwolf@comune.modena.it>*
12967 * New OCSP utility. Allows OCSP requests to be generated or
12968 read. The request can be sent to a responder and the output
12969 parsed, outputted or printed in text form. Not complete yet:
12970 still needs to check the OCSP response validity.
12974 * New subcommands for 'openssl ca':
12975 `openssl ca -status <serial>` prints the status of the cert with
12976 the given serial number (according to the index file).
12977 `openssl ca -updatedb` updates the expiry status of certificates
12980 *Massimiliano Pala <madwolf@comune.modena.it>*
12982 * New '-newreq-nodes' command option to CA.pl. This is like
12983 '-newreq', but calls 'openssl req' with the '-nodes' option
12984 so that the resulting key is not encrypted.
12986 *Damien Miller <djm@mindrot.org>*
12988 * New configuration for the GNU Hurd.
12990 *Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte*
12992 * Initial code to implement OCSP basic response verify. This
12993 is currently incomplete. Currently just finds the signer's
12994 certificate and verifies the signature on the response.
12998 * New SSLeay_version code SSLEAY_DIR to determine the compiled-in
12999 value of OPENSSLDIR. This is available via the new '-d' option
13000 to 'openssl version', and is also included in 'openssl version -a'.
13004 * Allowing defining memory allocation callbacks that will be given
13005 file name and line number information in additional arguments
13006 (a `const char*` and an int). The basic functionality remains, as
13007 well as the original possibility to just replace malloc(),
13008 realloc() and free() by functions that do not know about these
13009 additional arguments. To register and find out the current
13010 settings for extended allocation functions, the following
13011 functions are provided:
13013 CRYPTO_set_mem_ex_functions
13014 CRYPTO_set_locked_mem_ex_functions
13015 CRYPTO_get_mem_ex_functions
13016 CRYPTO_get_locked_mem_ex_functions
13018 These work the same way as CRYPTO_set_mem_functions and friends.
13019 `CRYPTO_get_[locked_]mem_functions` now writes 0 where such an
13020 extended allocation function is enabled.
13021 Similarly, `CRYPTO_get_[locked_]mem_ex_functions` writes 0 where
13022 a conventional allocation function is enabled.
13024 *Richard Levitte, Bodo Moeller*
13026 * Finish off removing the remaining LHASH function pointer casts.
13027 There should no longer be any prototype-casting required when using
13028 the LHASH abstraction, and any casts that remain are "bugs". See
13029 the callback types and macros at the head of lhash.h for details
13030 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
13034 * Add automatic query of EGD sockets in RAND_poll() for the unix variant.
13035 If /dev/[u]random devices are not available or do not return enough
13036 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
13038 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
13039 /etc/entropy will be queried once each in this sequence, querying stops
13040 when enough entropy was collected without querying more sockets.
13044 * Change the Unix RAND_poll() variant to be able to poll several
13045 random devices, as specified by DEVRANDOM, until a sufficient amount
13046 of data has been collected. We spend at most 10 ms on each file
13047 (select timeout) and read in non-blocking mode. DEVRANDOM now
13048 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
13049 (previously it was just the string "/dev/urandom"), so on typical
13050 platforms the 10 ms delay will never occur.
13051 Also separate out the Unix variant to its own file, rand_unix.c.
13052 For VMS, there's a currently-empty rand_vms.c.
13056 * Move OCSP client related routines to ocsp_cl.c. These
13057 provide utility functions which an application needing
13058 to issue a request to an OCSP responder and analyse the
13059 response will typically need: as opposed to those which an
13060 OCSP responder itself would need which will be added later.
13062 OCSP_request_sign() signs an OCSP request with an API similar
13063 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
13064 response. OCSP_response_get1_basic() extracts basic response
13065 from response. OCSP_resp_find_status(): finds and extracts status
13066 information from an OCSP_CERTID structure (which will be created
13067 when the request structure is built). These are built from lower
13068 level functions which work on OCSP_SINGLERESP structures but
13069 won't normally be used unless the application wishes to examine
13070 extensions in the OCSP response for example.
13072 Replace nonce routines with a pair of functions.
13073 OCSP_request_add1_nonce() adds a nonce value and optionally
13074 generates a random value. OCSP_check_nonce() checks the
13075 validity of the nonce in an OCSP response.
13079 * Change function OCSP_request_add() to OCSP_request_add0_id().
13080 This doesn't copy the supplied OCSP_CERTID and avoids the
13081 need to free up the newly created id. Change return type
13082 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
13083 This can then be used to add extensions to the request.
13084 Deleted OCSP_request_new(), since most of its functionality
13085 is now in OCSP_REQUEST_new() (and the case insensitive name
13086 clash) apart from the ability to set the request name which
13087 will be added elsewhere.
13091 * Update OCSP API. Remove obsolete extensions argument from
13092 various functions. Extensions are now handled using the new
13093 OCSP extension code. New simple OCSP HTTP function which
13094 can be used to send requests and parse the response.
13098 * Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
13099 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
13100 uses the special reorder version of SET OF to sort the attributes
13101 and reorder them to match the encoded order. This resolves a long
13102 standing problem: a verify on a PKCS7 structure just after signing
13103 it used to fail because the attribute order did not match the
13104 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
13105 it uses the received order. This is necessary to tolerate some broken
13106 software that does not order SET OF. This is handled by encoding
13107 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
13108 to produce the required SET OF.
13112 * Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
13113 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
13114 files to get correct declarations of the ASN.1 item variables.
13118 * Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
13119 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
13120 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
13121 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
13122 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
13123 ASN1_ITEM and no wrapper functions.
13127 * New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
13128 replace the old function pointer based I/O routines. Change most of
13129 the `*_d2i_bio()` and `*_d2i_fp()` functions to use these.
13133 * Enhance mkdef.pl to be more accepting about spacing in C preprocessor
13134 lines, recognize more "algorithms" that can be deselected, and make
13135 it complain about algorithm deselection that isn't recognised.
13139 * New ASN1 functions to handle dup, sign, verify, digest, pack and
13140 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
13141 to use new functions. Add NO_ASN1_OLD which can be set to remove
13142 some old style ASN1 functions: this can be used to determine if old
13143 code will still work when these eventually go away.
13147 * New extension functions for OCSP structures, these follow the
13148 same conventions as certificates and CRLs.
13152 * New function X509V3_add1_i2d(). This automatically encodes and
13153 adds an extension. Its behaviour can be customised with various
13154 flags to append, replace or delete. Various wrappers added for
13155 certificates and CRLs.
13159 * Fix to avoid calling the underlying ASN1 print routine when
13160 an extension cannot be parsed. Correct a typo in the
13161 OCSP_SERVICELOC extension. Tidy up print OCSP format.
13165 * Make mkdef.pl parse some of the ASN1 macros and add appropriate
13166 entries for variables.
13170 * Add functionality to `apps/openssl.c` for detecting locking
13171 problems: As the program is single-threaded, all we have
13172 to do is register a locking callback using an array for
13173 storing which locks are currently held by the program.
13177 * Use a lock around the call to CRYPTO_get_ex_new_index() in
13178 SSL_get_ex_data_X509_STORE_idx(), which is used in
13179 ssl_verify_cert_chain() and thus can be called at any time
13180 during TLS/SSL handshakes so that thread-safety is essential.
13181 Unfortunately, the ex_data design is not at all suited
13182 for multi-threaded use, so it probably should be abolished.
13186 * Added Broadcom "ubsec" ENGINE to OpenSSL.
13188 *Broadcom, tweaked and integrated by Geoff Thorpe*
13190 * Move common extension printing code to new function
13191 X509V3_print_extensions(). Reorganise OCSP print routines and
13192 implement some needed OCSP ASN1 functions. Add OCSP extensions.
13196 * New function X509_signature_print() to remove duplication in some
13201 * Add a special meaning when SET OF and SEQUENCE OF flags are both
13202 set (this was treated exactly the same as SET OF previously). This
13203 is used to reorder the STACK representing the structure to match the
13204 encoding. This will be used to get round a problem where a PKCS7
13205 structure which was signed could not be verified because the STACK
13206 order did not reflect the encoded order.
13210 * Reimplement the OCSP ASN1 module using the new code.
13214 * Update the X509V3 code to permit the use of an ASN1_ITEM structure
13215 for its ASN1 operations. The old style function pointers still exist
13216 for now but they will eventually go away.
13220 * Merge in replacement ASN1 code from the ASN1 branch. This almost
13221 completely replaces the old ASN1 functionality with a table driven
13222 encoder and decoder which interprets an ASN1_ITEM structure describing
13223 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
13224 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
13225 has also been converted to the new form.
13229 * Change BN_mod_exp_recp so that negative moduli are tolerated
13230 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
13231 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
13232 for negative moduli.
13236 * Fix BN_uadd and BN_usub: Always return non-negative results instead
13237 of not touching the result's sign bit.
13241 * BN_div bugfix: If the result is 0, the sign (res->neg) must not be
13246 * Changed the LHASH code to use prototypes for callbacks, and created
13247 macros to declare and implement thin (optionally static) functions
13248 that provide type-safety and avoid function pointer casting for the
13249 type-specific callbacks.
13253 * Added Kerberos Cipher Suites to be used with TLS, as written in
13255 *Veers Staats <staatsvr@asc.hpc.mil>,
13256 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte*
13258 * Reformat the FAQ so the different questions and answers can be divided
13259 in sections depending on the subject.
13263 * Have the zlib compression code load ZLIB.DLL dynamically under
13268 * New function BN_mod_sqrt for computing square roots modulo a prime
13269 (using the probabilistic Tonelli-Shanks algorithm unless
13270 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
13271 be handled deterministically).
13273 *Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller*
13275 * Make BN_mod_inverse faster by explicitly handling small quotients
13276 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
13277 512 bits], about 30% for larger ones [1024 or 2048 bits].)
13281 * New function BN_kronecker.
13285 * Fix BN_gcd so that it works on negative inputs; the result is
13286 positive unless both parameters are zero.
13287 Previously something reasonably close to an infinite loop was
13288 possible because numbers could be growing instead of shrinking
13289 in the implementation of Euclid's algorithm.
13293 * Fix BN_is_word() and BN_is_one() macros to take into account the
13294 sign of the number in question.
13296 Fix BN_is_word(a,w) to work correctly for w == 0.
13298 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
13299 because its test if the absolute value of 'a' equals 'w'.
13300 Note that BN_abs_is_word does *not* handle w == 0 reliably;
13301 it exists mostly for use in the implementations of BN_is_zero(),
13302 BN_is_one(), and BN_is_word().
13306 * New function BN_swap.
13310 * Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
13311 the exponentiation functions are more likely to produce reasonable
13312 results on negative inputs.
13316 * Change BN_mod_mul so that the result is always non-negative.
13317 Previously, it could be negative if one of the factors was negative;
13318 I don't think anyone really wanted that behaviour.
13322 * Move `BN_mod_...` functions into new file `crypto/bn/bn_mod.c`
13323 (except for exponentiation, which stays in `crypto/bn/bn_exp.c`,
13324 and `BN_mod_mul_reciprocal`, which stays in `crypto/bn/bn_recp.c`)
13325 and add new functions:
13334 BN_mod_lshift1_quick
13336 BN_mod_lshift_quick
13338 These functions always generate non-negative results.
13340 `BN_nnmod` otherwise is `like BN_mod` (if `BN_mod` computes a remainder `r`
13341 such that `|m| < r < 0`, `BN_nnmod` will output `rem + |m|` instead).
13343 `BN_mod_XXX_quick(r, a, [b,] m)` generates the same result as
13344 `BN_mod_XXX(r, a, [b,] m, ctx)`, but requires that `a` [and `b`]
13345 be reduced modulo `m`.
13347 *Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller*
13350 The following entry accidentally appeared in the CHANGES file
13351 distributed with OpenSSL 0.9.7. The modifications described in
13352 it do *not* apply to OpenSSL 0.9.7.
13354 * Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
13355 was actually never needed) and in BN_mul(). The removal in BN_mul()
13356 required a small change in bn_mul_part_recursive() and the addition
13357 of the functions bn_cmp_part_words(), bn_sub_part_words() and
13358 bn_add_part_words(), which do the same thing as bn_cmp_words(),
13359 bn_sub_words() and bn_add_words() except they take arrays with
13365 * In 'openssl passwd', verify passwords read from the terminal
13366 unless the '-salt' option is used (which usually means that
13367 verification would just waste user's time since the resulting
13368 hash is going to be compared with some given password hash)
13369 or the new '-noverify' option is used.
13371 This is an incompatible change, but it does not affect
13372 non-interactive use of 'openssl passwd' (passwords on the command
13373 line, '-stdin' option, '-in ...' option) and thus should not
13374 cause any problems.
13378 * Remove all references to RSAref, since there's no more need for it.
13382 * Make DSO load along a path given through an environment variable
13383 (SHLIB_PATH) with shl_load().
13387 * Constify the ENGINE code as a result of BIGNUM constification.
13388 Also constify the RSA code and most things related to it. In a
13389 few places, most notable in the depth of the ASN.1 code, ugly
13390 casts back to non-const were required (to be solved at a later
13395 * Make it so the openssl application has all engines loaded by default.
13399 * Constify the BIGNUM routines a little more.
13403 * Add the following functions:
13405 ENGINE_load_cswift()
13407 ENGINE_load_atalla()
13408 ENGINE_load_nuron()
13409 ENGINE_load_builtin_engines()
13411 That way, an application can itself choose if external engines that
13412 are built-in in OpenSSL shall ever be used or not. The benefit is
13413 that applications won't have to be linked with libdl or other dso
13414 libraries unless it's really needed.
13416 Changed 'openssl engine' to load all engines on demand.
13417 Changed the engine header files to avoid the duplication of some
13418 declarations (they differed!).
13422 * 'openssl engine' can now list capabilities.
13426 * Better error reporting in 'openssl engine'.
13430 * Never call load_dh_param(NULL) in s_server.
13434 * Add engine application. It can currently list engines by name and
13435 identity, and test if they are actually available.
13439 * Improve RPM specification file by forcing symbolic linking and making
13440 sure the installed documentation is also owned by root.root.
13442 *Damien Miller <djm@mindrot.org>*
13444 * Give the OpenSSL applications more possibilities to make use of
13445 keys (public as well as private) handled by engines.
13449 * Add OCSP code that comes from CertCo.
13453 * Add VMS support for the Rijndael code.
13457 * Added untested support for Nuron crypto accelerator.
13461 * Add support for external cryptographic devices. This code was
13462 previously distributed separately as the "engine" branch.
13464 *Geoff Thorpe, Richard Levitte*
13466 * Rework the filename-translation in the DSO code. It is now possible to
13467 have far greater control over how a "name" is turned into a filename
13468 depending on the operating environment and any oddities about the
13469 different shared library filenames on each system.
13473 * Support threads on FreeBSD-elf in Configure.
13477 * Fix for SHA1 assembly problem with MASM: it produces
13478 warnings about corrupt line number information when assembling
13479 with debugging information. This is caused by the overlapping
13482 *Bernd Matthes <mainbug@celocom.de>, Steve Henson*
13485 NCONF_get_number() has no error checking at all. As a replacement,
13486 NCONF_get_number_e() is defined (`_e` for "error checking") and is
13487 promoted strongly. The old NCONF_get_number is kept around for
13488 binary backward compatibility.
13489 Make it possible for methods to load from something other than a BIO,
13490 by providing a function pointer that is given a name instead of a BIO.
13491 For example, this could be used to load configuration data from an
13496 * Fix for non blocking accept BIOs. Added new I/O special reason
13497 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
13498 with non blocking I/O was not possible because no retry code was
13499 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
13504 * Added the beginnings of Rijndael support.
13508 * Fix for bug in DirectoryString mask setting. Add support for
13509 X509_NAME_print_ex() in 'req' and X509_print_ex() function
13510 to allow certificate printing to more controllable, additional
13511 'certopt' option to 'x509' to allow new printing options to be
13516 * Clean old EAY MD5 hack from e_os.h.
13520 ### Changes between 0.9.6l and 0.9.6m [17 Mar 2004]
13522 * Fix null-pointer assignment in do_change_cipher_spec() revealed
13523 by using the Codenomicon TLS Test Tool ([CVE-2004-0079])
13525 *Joe Orton, Steve Henson*
13527 ### Changes between 0.9.6k and 0.9.6l [04 Nov 2003]
13529 * Fix additional bug revealed by the NISCC test suite:
13531 Stop bug triggering large recursion when presented with
13532 certain ASN.1 tags ([CVE-2003-0851])
13536 ### Changes between 0.9.6j and 0.9.6k [30 Sep 2003]
13538 * Fix various bugs revealed by running the NISCC test suite:
13540 Stop out of bounds reads in the ASN1 code when presented with
13541 invalid tags (CVE-2003-0543 and CVE-2003-0544).
13543 If verify callback ignores invalid public key errors don't try to check
13544 certificate signature with the NULL public key.
13548 * In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
13549 if the server requested one: as stated in TLS 1.0 and SSL 3.0
13554 * In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
13555 extra data after the compression methods not only for TLS 1.0
13556 but also for SSL 3.0 (as required by the specification).
13558 *Bodo Moeller; problem pointed out by Matthias Loepfe*
13560 * Change X509_certificate_type() to mark the key as exported/exportable
13561 when it's 512 *bits* long, not 512 bytes.
13565 ### Changes between 0.9.6i and 0.9.6j [10 Apr 2003]
13567 * Countermeasure against the Klima-Pokorny-Rosa extension of
13568 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
13569 a protocol version number mismatch like a decryption error
13570 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
13574 * Turn on RSA blinding by default in the default implementation
13575 to avoid a timing attack. Applications that don't want it can call
13576 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
13577 They would be ill-advised to do so in most cases.
13579 *Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller*
13581 * Change RSA blinding code so that it works when the PRNG is not
13582 seeded (in this case, the secret RSA exponent is abused as
13583 an unpredictable seed -- if it is not unpredictable, there
13584 is no point in blinding anyway). Make RSA blinding thread-safe
13585 by remembering the creator's thread ID in rsa->blinding and
13586 having all other threads use local one-time blinding factors
13587 (this requires more computation than sharing rsa->blinding, but
13588 avoids excessive locking; and if an RSA object is not shared
13589 between threads, blinding will still be very fast).
13593 ### Changes between 0.9.6h and 0.9.6i [19 Feb 2003]
13595 * In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
13596 via timing by performing a MAC computation even if incorrect
13597 block cipher padding has been found. This is a countermeasure
13598 against active attacks where the attacker has to distinguish
13599 between bad padding and a MAC verification error. ([CVE-2003-0078])
13601 *Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
13602 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
13603 Martin Vuagnoux (EPFL, Ilion)*
13605 ### Changes between 0.9.6g and 0.9.6h [5 Dec 2002]
13607 * New function OPENSSL_cleanse(), which is used to cleanse a section of
13608 memory from its contents. This is done with a counter that will
13609 place alternating values in each byte. This can be used to solve
13610 two issues: 1) the removal of calls to memset() by highly optimizing
13611 compilers, and 2) cleansing with other values than 0, since those can
13612 be read through on certain media, for example a swap space on disk.
13616 * Bugfix: client side session caching did not work with external caching,
13617 because the session->cipher setting was not restored when reloading
13618 from the external cache. This problem was masked, when
13619 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set.
13620 (Found by Steve Haslam <steve@araqnid.ddts.net>.)
13624 * Fix client_certificate (ssl/s2_clnt.c): The permissible total
13625 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33.
13627 *Zeev Lieber <zeev-l@yahoo.com>*
13629 * Undo an undocumented change introduced in 0.9.6e which caused
13630 repeated calls to OpenSSL_add_all_ciphers() and
13631 OpenSSL_add_all_digests() to be ignored, even after calling
13636 * Change the default configuration reader to deal with last line not
13637 being properly terminated.
13641 * Change X509_NAME_cmp() so it applies the special rules on handling
13642 DN values that are of type PrintableString, as well as RDNs of type
13643 emailAddress where the value has the type ia5String.
13645 *stefank@valicert.com via Richard Levitte*
13647 * Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half
13648 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently
13649 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be
13650 the bitwise-OR of the two for use by the majority of applications
13651 wanting this behaviour, and update the docs. The documented
13652 behaviour and actual behaviour were inconsistent and had been
13653 changing anyway, so this is more a bug-fix than a behavioural
13656 *Geoff Thorpe, diagnosed by Nadav Har'El*
13658 * Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c
13659 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes).
13663 * Fix initialization code race conditions in
13664 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(),
13665 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(),
13666 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(),
13667 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(),
13668 ssl2_get_cipher_by_char(),
13669 ssl3_get_cipher_by_char().
13671 *Patrick McCormick <patrick@tellme.com>, Bodo Moeller*
13673 * Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after
13674 the cached sessions are flushed, as the remove_cb() might use ex_data
13675 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com>
13676 (see [openssl.org #212]).
13678 *Geoff Thorpe, Lutz Jaenicke*
13680 * Fix typo in OBJ_txt2obj which incorrectly passed the content
13681 length, instead of the encoding length to d2i_ASN1_OBJECT.
13685 ### Changes between 0.9.6f and 0.9.6g [9 Aug 2002]
13687 * [In 0.9.6g-engine release:]
13688 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use `_stdcall`).
13690 *Lynn Gazis <lgazis@rainbow.com>*
13692 ### Changes between 0.9.6e and 0.9.6f [8 Aug 2002]
13694 * Fix ASN1 checks. Check for overflow by comparing with LONG_MAX
13695 and get fix the header length calculation.
13696 *Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>,
13697 Alon Kantor <alonk@checkpoint.com> (and others), Steve Henson*
13699 * Use proper error handling instead of 'assertions' in buffer
13700 overflow checks added in 0.9.6e. This prevents DoS (the
13701 assertions could call abort()).
13703 *Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller*
13705 ### Changes between 0.9.6d and 0.9.6e [30 Jul 2002]
13707 * Add various sanity checks to asn1_get_length() to reject
13708 the ASN1 length bytes if they exceed sizeof(long), will appear
13709 negative or the content length exceeds the length of the
13712 *Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>*
13714 * Fix cipher selection routines: ciphers without encryption had no flags
13715 for the cipher strength set and where therefore not handled correctly
13716 by the selection routines (PR #130).
13720 * Fix EVP_dsa_sha macro.
13725 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS
13726 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure
13727 that was added in OpenSSL 0.9.6d.
13729 As the countermeasure turned out to be incompatible with some
13730 broken SSL implementations, the new option is part of SSL_OP_ALL.
13731 SSL_OP_ALL is usually employed when compatibility with weird SSL
13732 implementations is desired (e.g. '-bugs' option to 's_client' and
13733 's_server'), so the new option is automatically set in many
13738 * Changes in security patch:
13740 Changes marked "(CHATS)" were sponsored by the Defense Advanced
13741 Research Projects Agency (DARPA) and Air Force Research Laboratory,
13742 Air Force Materiel Command, USAF, under agreement number
13745 * Add various sanity checks to asn1_get_length() to reject
13746 the ASN1 length bytes if they exceed sizeof(long), will appear
13747 negative or the content length exceeds the length of the
13748 supplied buffer. ([CVE-2002-0659])
13750 *Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>*
13752 * Assertions for various potential buffer overflows, not known to
13753 happen in practice.
13755 *Ben Laurie (CHATS)*
13757 * Various temporary buffers to hold ASCII versions of integers were
13758 too small for 64 bit platforms. ([CVE-2002-0655])
13759 *Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>*
13761 * Remote buffer overflow in SSL3 protocol - an attacker could
13762 supply an oversized session ID to a client. ([CVE-2002-0656])
13764 *Ben Laurie (CHATS)*
13766 * Remote buffer overflow in SSL2 protocol - an attacker could
13767 supply an oversized client master key. ([CVE-2002-0656])
13769 *Ben Laurie (CHATS)*
13771 ### Changes between 0.9.6c and 0.9.6d [9 May 2002]
13773 * Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
13774 encoded as NULL) with id-dsa-with-sha1.
13776 *Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller*
13778 * Check various `X509_...()` return values in `apps/req.c`.
13780 *Nils Larsch <nla@trustcenter.de>*
13782 * Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
13783 an end-of-file condition would erroneously be flagged, when the CRLF
13784 was just at the end of a processed block. The bug was discovered when
13785 processing data through a buffering memory BIO handing the data to a
13786 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
13787 <ptsekov@syntrex.com> and Nedelcho Stanev.
13791 * Implement a countermeasure against a vulnerability recently found
13792 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
13793 before application data chunks to avoid the use of known IVs
13794 with data potentially chosen by the attacker.
13798 * Fix length checks in ssl3_get_client_hello().
13802 * TLS/SSL library bugfix: use s->s3->in_read_app_data differently
13803 to prevent ssl3_read_internal() from incorrectly assuming that
13804 ssl3_read_bytes() found application data while handshake
13805 processing was enabled when in fact s->s3->in_read_app_data was
13806 merely automatically cleared during the initial handshake.
13808 *Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>*
13810 * Fix object definitions for Private and Enterprise: they were not
13811 recognized in their shortname (=lowercase) representation. Extend
13812 obj_dat.pl to issue an error when using undefined keywords instead
13813 of silently ignoring the problem (Svenning Sorensen
13814 <sss@sss.dnsalias.net>).
13818 * Fix DH_generate_parameters() so that it works for 'non-standard'
13819 generators, i.e. generators other than 2 and 5. (Previously, the
13820 code did not properly initialise the 'add' and 'rem' values to
13821 BN_generate_prime().)
13823 In the new general case, we do not insist that 'generator' is
13824 actually a primitive root: This requirement is rather pointless;
13825 a generator of the order-q subgroup is just as good, if not
13830 * Map new X509 verification errors to alerts. Discovered and submitted by
13831 Tom Wu <tom@arcot.com>.
13835 * Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
13836 returning non-zero before the data has been completely received
13837 when using non-blocking I/O.
13839 *Bodo Moeller; problem pointed out by John Hughes*
13841 * Some of the ciphers missed the strength entry (SSL_LOW etc).
13843 *Ben Laurie, Lutz Jaenicke*
13845 * Fix bug in SSL_clear(): bad sessions were not removed (found by
13846 Yoram Zahavi <YoramZ@gilian.com>).
13850 * Add information about CygWin 1.3 and on, and preserve proper
13851 configuration for the versions before that.
13853 *Corinna Vinschen <vinschen@redhat.com> and Richard Levitte*
13855 * Make removal from session cache (SSL_CTX_remove_session()) more robust:
13856 check whether we deal with a copy of a session and do not delete from
13857 the cache in this case. Problem reported by "Izhar Shoshani Levi"
13858 <izhar@checkpoint.com>.
13862 * Do not store session data into the internal session cache, if it
13863 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
13864 flag is set). Proposed by Aslam <aslam@funk.com>.
13868 * Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
13873 * [In 0.9.6d-engine release:]
13874 Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
13876 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
13878 * Add the configuration target linux-s390x.
13880 *Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte*
13882 * The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
13883 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
13884 variable as an indication that a ClientHello message has been
13885 received. As the flag value will be lost between multiple
13886 invocations of ssl3_accept when using non-blocking I/O, the
13887 function may not be aware that a handshake has actually taken
13888 place, thus preventing a new session from being added to the
13891 To avoid this problem, we now set s->new_session to 2 instead of
13892 using a local variable.
13894 *Lutz Jaenicke, Bodo Moeller*
13896 * Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
13897 if the SSL_R_LENGTH_MISMATCH error is detected.
13899 *Geoff Thorpe, Bodo Moeller*
13901 * New 'shared_ldflag' column in Configure platform table.
13905 * Fix EVP_CIPHER_mode macro.
13907 *"Dan S. Camper" <dan@bti.net>*
13909 * Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
13910 type, we must throw them away by setting rr->length to 0.
13912 *D P Chang <dpc@qualys.com>*
13914 ### Changes between 0.9.6b and 0.9.6c [21 dec 2001]
13916 * Fix BN_rand_range bug pointed out by Dominikus Scherkl
13917 <Dominikus.Scherkl@biodata.com>. (The previous implementation
13918 worked incorrectly for those cases where range = `10..._2` and
13919 `3*range` is two bits longer than range.)
13923 * Only add signing time to PKCS7 structures if it is not already
13928 * Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
13929 OBJ_ld_ce should be OBJ_id_ce.
13930 Also some ip-pda OIDs in crypto/objects/objects.txt were
13931 incorrect (cf. RFC 3039).
13933 *Matt Cooper, Frederic Giudicelli, Bodo Moeller*
13935 * Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
13936 returns early because it has nothing to do.
13938 *Andy Schneider <andy.schneider@bjss.co.uk>*
13940 * [In 0.9.6c-engine release:]
13941 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
13943 *Andy Schneider <andy.schneider@bjss.co.uk>*
13945 * [In 0.9.6c-engine release:]
13946 Add support for Cryptographic Appliance's keyserver technology.
13947 (Use engine 'keyclient')
13949 *Cryptographic Appliances and Geoff Thorpe*
13951 * Add a configuration entry for OS/390 Unix. The C compiler 'c89'
13952 is called via tools/c89.sh because arguments have to be
13953 rearranged (all '-L' options must appear before the first object
13956 *Richard Shapiro <rshapiro@abinitio.com>*
13958 * [In 0.9.6c-engine release:]
13959 Add support for Broadcom crypto accelerator cards, backported
13962 *Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox*
13964 * [In 0.9.6c-engine release:]
13965 Add support for SureWare crypto accelerator cards from
13966 Baltimore Technologies. (Use engine 'sureware')
13968 *Baltimore Technologies and Mark Cox*
13970 * [In 0.9.6c-engine release:]
13971 Add support for crypto accelerator cards from Accelerated
13972 Encryption Processing, www.aep.ie. (Use engine 'aep')
13974 *AEP Inc. and Mark Cox*
13976 * Add a configuration entry for gcc on UnixWare.
13978 *Gary Benson <gbenson@redhat.com>*
13980 * Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
13981 messages are stored in a single piece (fixed-length part and
13982 variable-length part combined) and fix various bugs found on the way.
13986 * Disable caching in BIO_gethostbyname(), directly use gethostbyname()
13987 instead. BIO_gethostbyname() does not know what timeouts are
13988 appropriate, so entries would stay in cache even when they have
13990 *Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>*
13992 * Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
13993 faced with a pathologically small ClientHello fragment that does
13994 not contain client_version: Instead of aborting with an error,
13995 simply choose the highest available protocol version (i.e.,
13996 TLS 1.0 unless it is disabled). In practice, ClientHello
13997 messages are never sent like this, but this change gives us
13998 strictly correct behaviour at least for TLS.
14002 * Fix SSL handshake functions and SSL_clear() such that SSL_clear()
14003 never resets s->method to s->ctx->method when called from within
14004 one of the SSL handshake functions.
14006 *Bodo Moeller; problem pointed out by Niko Baric*
14008 * In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
14009 (sent using the client's version number) if client_version is
14010 smaller than the protocol version in use. Also change
14011 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
14012 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
14013 the client will at least see that alert.
14017 * Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
14022 * Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
14023 client receives HelloRequest while in a handshake.
14025 *Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>*
14027 * Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
14028 should end in 'break', not 'goto end' which circumvents various
14029 cleanups done in state SSL_ST_OK. But session related stuff
14030 must be disabled for SSL_ST_OK in the case that we just sent a
14033 Also avoid some overhead by not calling ssl_init_wbio_buffer()
14034 before just sending a HelloRequest.
14036 *Bodo Moeller, Eric Rescorla <ekr@rtfm.com>*
14038 * Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
14039 reveal whether illegal block cipher padding was found or a MAC
14040 verification error occurred. (Neither SSLerr() codes nor alerts
14041 are directly visible to potential attackers, but the information
14042 may leak via logfiles.)
14044 Similar changes are not required for the SSL 2.0 implementation
14045 because the number of padding bytes is sent in clear for SSL 2.0,
14046 and the extra bytes are just ignored. However ssl/s2_pkt.c
14047 failed to verify that the purported number of padding bytes is in
14052 * Add OpenUNIX-8 support including shared libraries
14053 (Boyd Lynn Gerber <gerberb@zenez.com>).
14057 * Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
14058 'wristwatch attack' using huge encoding parameters (cf.
14059 James H. Manger's CRYPTO 2001 paper). Note that the
14060 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
14061 encoding parameters and hence was not vulnerable.
14065 * BN_sqr() bug fix.
14067 *Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>*
14069 * Rabin-Miller test analyses assume uniformly distributed witnesses,
14070 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
14071 followed by modular reduction.
14073 *Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>*
14075 * Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
14076 equivalent based on BN_pseudo_rand() instead of BN_rand().
14080 * s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
14081 This function was broken, as the check for a new client hello message
14082 to handle SGC did not allow these large messages.
14083 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
14087 * Add alert descriptions for TLSv1 to `SSL_alert_desc_string[_long]()`.
14091 * Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
14092 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
14096 * Rework the configuration and shared library support for Tru64 Unix.
14097 The configuration part makes use of modern compiler features and
14098 still retains old compiler behavior for those that run older versions
14099 of the OS. The shared library support part includes a variant that
14100 uses the RPATH feature, and is available through the special
14101 configuration target "alpha-cc-rpath", which will never be selected
14104 *Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte*
14106 * In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
14107 with the same message size as in ssl3_get_certificate_request().
14108 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
14109 messages might inadvertently be reject as too long.
14111 *Petr Lampa <lampa@fee.vutbr.cz>*
14113 * Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
14117 * Modified SSL library such that the verify_callback that has been set
14118 specifically for an SSL object with SSL_set_verify() is actually being
14119 used. Before the change, a verify_callback set with this function was
14120 ignored and the verify_callback() set in the SSL_CTX at the time of
14121 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
14122 to allow the necessary settings.
14126 * Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
14127 explicitly to NULL, as at least on Solaris 8 this seems not always to be
14128 done automatically (in contradiction to the requirements of the C
14129 standard). This made problems when used from OpenSSH.
14133 * In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
14134 dh->length and always used
14136 BN_rand_range(priv_key, dh->p).
14138 BN_rand_range() is not necessary for Diffie-Hellman, and this
14139 specific range makes Diffie-Hellman unnecessarily inefficient if
14140 dh->length (recommended exponent length) is much smaller than the
14141 length of dh->p. We could use BN_rand_range() if the order of
14142 the subgroup was stored in the DH structure, but we only have
14147 BN_rand(priv_key, l, ...)
14149 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
14156 RSA_eay_public_encrypt
14157 RSA_eay_private_decrypt
14158 RSA_eay_private_encrypt (signing)
14159 RSA_eay_public_decrypt (signature verification)
14161 (default implementations for RSA_public_encrypt,
14162 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
14163 always reject numbers >= n.
14167 * In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
14168 to synchronize access to 'locking_thread'. This is necessary on
14169 systems where access to 'locking_thread' (an 'unsigned long'
14170 variable) is not atomic.
14174 * In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
14175 *before* setting the 'crypto_lock_rand' flag. The previous code had
14176 a race condition if 0 is a valid thread ID.
14178 *Travis Vitek <vitek@roguewave.com>*
14180 * Add support for shared libraries under Irix.
14182 *Albert Chin-A-Young <china@thewrittenword.com>*
14184 * Add configuration option to build on Linux on both big-endian and
14185 little-endian MIPS.
14187 *Ralf Baechle <ralf@uni-koblenz.de>*
14189 * Add the possibility to create shared libraries on HP-UX.
14193 ### Changes between 0.9.6a and 0.9.6b [9 Jul 2001]
14195 * Change ssleay_rand_bytes (crypto/rand/md_rand.c)
14196 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
14197 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
14198 PRNG state recovery was possible based on the output of
14199 one PRNG request appropriately sized to gain knowledge on
14200 'md' followed by enough consecutive 1-byte PRNG requests
14201 to traverse all of 'state'.
14203 1. When updating 'md_local' (the current thread's copy of 'md')
14204 during PRNG output generation, hash all of the previous
14205 'md_local' value, not just the half used for PRNG output.
14207 2. Make the number of bytes from 'state' included into the hash
14208 independent from the number of PRNG bytes requested.
14210 The first measure alone would be sufficient to avoid
14211 Markku-Juhani's attack. (Actually it had never occurred
14212 to me that the half of 'md_local' used for chaining was the
14213 half from which PRNG output bytes were taken -- I had always
14214 assumed that the secret half would be used.) The second
14215 measure makes sure that additional data from 'state' is never
14216 mixed into 'md_local' in small portions; this heuristically
14217 further strengthens the PRNG.
14221 * Fix crypto/bn/asm/mips3.s.
14225 * When only the key is given to "enc", the IV is undefined. Print out
14226 an error message in this case.
14230 * Handle special case when X509_NAME is empty in X509 printing routines.
14234 * In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
14235 positive and less than q.
14239 * Don't change `*pointer` in CRYPTO_add_lock() is add_lock_callback is
14240 used: it isn't thread safe and the add_lock_callback should handle
14243 *Paul Rose <Paul.Rose@bridge.com>*
14245 * Verify that incoming data obeys the block size in
14246 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
14252 *Ulf Möller, Bodo Möller*
14254 * The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
14255 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
14256 when fixing the server behaviour for backwards-compatible 'client
14257 hello' messages. (Note that the attack is impractical against
14258 SSL 3.0 and TLS 1.0 anyway because length and version checking
14259 means that the probability of guessing a valid ciphertext is
14260 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
14263 Before 0.9.5, the countermeasure (hide the error by generating a
14264 random 'decryption result') did not work properly because
14265 ERR_clear_error() was missing, meaning that SSL_get_error() would
14266 detect the supposedly ignored error.
14268 Both problems are now fixed.
14272 * In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
14273 (previously it was 1024).
14277 * Fix for compatibility mode trust settings: ignore trust settings
14278 unless some valid trust or reject settings are present.
14282 * Fix for blowfish EVP: its a variable length cipher.
14286 * Fix various bugs related to DSA S/MIME verification. Handle missing
14287 parameters in DSA public key structures and return an error in the
14288 DSA routines if parameters are absent.
14292 * In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
14293 in the current directory if neither $RANDFILE nor $HOME was set.
14294 RAND_file_name() in 0.9.6a returned NULL in this case. This has
14295 caused some confusion to Windows users who haven't defined $HOME.
14296 Thus RAND_file_name() is changed again: e_os.h can define a
14297 DEFAULT_HOME, which will be used if $HOME is not set.
14298 For Windows, we use "C:"; on other platforms, we still require
14299 environment variables.
14301 * Move 'if (!initialized) RAND_poll()' into regions protected by
14302 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
14303 having multiple threads call RAND_poll() concurrently.
14307 * In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
14308 combination of a flag and a thread ID variable.
14309 Otherwise while one thread is in ssleay_rand_bytes (which sets the
14310 flag), *other* threads can enter ssleay_add_bytes without obeying
14311 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
14312 that they do not hold after the first thread unsets add_do_not_lock).
14316 * Change bctest again: '-x' expressions are not available in all
14317 versions of 'test'.
14321 ### Changes between 0.9.6 and 0.9.6a [5 Apr 2001]
14323 * Fix a couple of memory leaks in PKCS7_dataDecode()
14325 *Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>*
14327 * Change Configure and Makefiles to provide EXE_EXT, which will contain
14328 the default extension for executables, if any. Also, make the perl
14329 scripts that use symlink() to test if it really exists and use "cp"
14330 if it doesn't. All this made OpenSSL compilable and installable in
14335 * Fix for asn1_GetSequence() for indefinite length constructed data.
14336 If SEQUENCE is length is indefinite just set c->slen to the total
14337 amount of data available.
14339 *Steve Henson, reported by shige@FreeBSD.org*
14341 *This change does not apply to 0.9.7.*
14343 * Change bctest to avoid here-documents inside command substitution
14344 (workaround for FreeBSD /bin/sh bug).
14345 For compatibility with Ultrix, avoid shell functions (introduced
14346 in the bctest version that searches along $PATH).
14350 * Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
14351 with des_encrypt() defined on some operating systems, like Solaris
14356 * Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
14357 On the Importance of Eliminating Errors in Cryptographic
14358 Computations, J. Cryptology 14 (2001) 2, 101-119,
14359 <http://theory.stanford.edu/~dabo/papers/faults.ps.gz>).
14363 * MIPS assembler BIGNUM division bug fix.
14367 * Disabled incorrect Alpha assembler code.
14371 * Fix PKCS#7 decode routines so they correctly update the length
14372 after reading an EOC for the EXPLICIT tag.
14376 *This change does not apply to 0.9.7.*
14378 * Fix bug in PKCS#12 key generation routines. This was triggered
14379 if a 3DES key was generated with a 0 initial byte. Include
14380 PKCS12_BROKEN_KEYGEN compilation option to retain the old
14381 (but broken) behaviour.
14385 * Enhance bctest to search for a working bc along $PATH and print
14388 *Tim Rice <tim@multitalents.net> via Richard Levitte*
14390 * Fix memory leaks in err.c: free err_data string if necessary;
14391 don't write to the wrong index in ERR_set_error_data.
14395 * Implement ssl23_peek (analogous to ssl23_read), which previously
14400 * Replace rdtsc with `_emit` statements for VC++ version 5.
14402 *Jeremy Cooper <jeremy@baymoo.org>*
14404 * Make it possible to reuse SSLv2 sessions.
14408 * In copy_email() check for >= 0 as a return value for
14409 X509_NAME_get_index_by_NID() since 0 is a valid index.
14411 *Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>*
14413 * Avoid coredump with unsupported or invalid public keys by checking if
14414 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
14415 PKCS7_verify() fails with non detached data.
14419 * Don't use getenv in library functions when run as setuid/setgid.
14420 New function OPENSSL_issetugid().
14424 * Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
14425 due to incorrect handling of multi-threading:
14427 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
14429 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
14431 3. Count how many times MemCheck_off() has been called so that
14432 nested use can be treated correctly. This also avoids
14433 inband-signalling in the previous code (which relied on the
14434 assumption that thread ID 0 is impossible).
14438 * Add "-rand" option also to s_client and s_server.
14442 * Fix CPU detection on Irix 6.x.
14443 *Kurt Hockenbury <khockenb@stevens-tech.edu> and
14444 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>*
14446 * Fix X509_NAME bug which produced incorrect encoding if X509_NAME
14451 *This change does not apply to 0.9.7.*
14453 * Use the cached encoding of an X509_NAME structure rather than
14454 copying it. This is apparently the reason for the libsafe "errors"
14455 but the code is actually correct.
14459 * Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
14460 Bleichenbacher's DSA attack.
14461 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
14462 to be set and top=0 forces the highest bit to be set; top=-1 is new
14463 and leaves the highest bit random.
14465 *Ulf Moeller, Bodo Moeller*
14467 * In the `NCONF_...`-based implementations for `CONF_...` queries
14468 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
14469 a temporary CONF structure with the data component set to NULL
14470 (which gives segmentation faults in lh_retrieve).
14471 Instead, use NULL for the CONF pointer in CONF_get_string and
14472 CONF_get_number (which may use environment variables) and directly
14473 return NULL from CONF_get_section.
14477 * Fix potential buffer overrun for EBCDIC.
14481 * Tolerate nonRepudiation as being valid for S/MIME signing and certSign
14482 keyUsage if basicConstraints absent for a CA.
14486 * Make SMIME_write_PKCS7() write mail header values with a format that
14487 is more generally accepted (no spaces before the semicolon), since
14488 some programs can't parse those values properly otherwise. Also make
14489 sure BIO's that break lines after each write do not create invalid
14494 * Make the CRL encoding routines work with empty SEQUENCE OF. The
14495 macros previously used would not encode an empty SEQUENCE OF
14496 and break the signature.
14500 *This change does not apply to 0.9.7.*
14502 * Zero the premaster secret after deriving the master secret in
14507 * Add some EVP_add_digest_alias registrations (as found in
14508 OpenSSL_add_all_digests()) to SSL_library_init()
14509 aka OpenSSL_add_ssl_algorithms(). This provides improved
14510 compatibility with peers using X.509 certificates
14511 with unconventional AlgorithmIdentifier OIDs.
14515 * Fix for Irix with NO_ASM.
14517 *"Bruce W. Forsberg" <bruce.forsberg@baesystems.com>*
14519 * ./config script fixes.
14521 *Ulf Moeller, Richard Levitte*
14523 * Fix 'openssl passwd -1'.
14527 * Change PKCS12_key_gen_asc() so it can cope with non null
14528 terminated strings whose length is passed in the passlen
14529 parameter, for example from PEM callbacks. This was done
14530 by adding an extra length parameter to asc2uni().
14532 *Steve Henson, reported by <oddissey@samsung.co.kr>*
14534 * Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
14535 call failed, free the DSA structure.
14539 * Fix to uni2asc() to cope with zero length Unicode strings.
14540 These are present in some PKCS#12 files.
14544 * Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
14545 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
14546 when writing a 32767 byte record.
14548 *Bodo Moeller; problem reported by Eric Day <eday@concentric.net>*
14550 * In `RSA_eay_public_{en,ed}crypt` and RSA_eay_mod_exp (rsa_eay.c),
14551 obtain lock CRYPTO_LOCK_RSA before setting `rsa->_method_mod_{n,p,q}`.
14553 (RSA objects have a reference count access to which is protected
14554 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
14555 so they are meant to be shared between threads.)
14556 *Bodo Moeller, Geoff Thorpe; original patch submitted by
14557 "Reddie, Steven" <Steven.Reddie@ca.com>*
14559 * Fix a deadlock in CRYPTO_mem_leaks().
14563 * Use better test patterns in bntest.
14567 * rand_win.c fix for Borland C.
14571 * BN_rshift bugfix for n == 0.
14575 * Add a 'bctest' script that checks for some known 'bc' bugs
14576 so that 'make test' does not abort just because 'bc' is broken.
14580 * Store verify_result within SSL_SESSION also for client side to
14581 avoid potential security hole. (Re-used sessions on the client side
14582 always resulted in verify_result==X509_V_OK, not using the original
14583 result of the server certificate verification.)
14587 * Fix ssl3_pending: If the record in s->s3->rrec is not of type
14588 SSL3_RT_APPLICATION_DATA, return 0.
14589 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
14594 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
14595 releases, have been re-implemented by renaming the previous
14596 implementations of ssl2_read and ssl3_read to ssl2_read_internal
14597 and ssl3_read_internal, respectively, and adding 'peek' parameters
14598 to them. The new ssl[23]_{read,peek} functions are calls to
14599 ssl[23]_read_internal with the 'peek' flag set appropriately.
14600 A 'peek' parameter has also been added to ssl3_read_bytes, which
14601 does the actual work for ssl3_read_internal.
14605 * Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
14606 the method-specific "init()" handler. Also clean up ex_data after
14607 calling the method-specific "finish()" handler. Previously, this was
14608 happening the other way round.
14612 * Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
14613 The previous value, 12, was not always sufficient for BN_mod_exp().
14617 * Make sure that shared libraries get the internal name engine with
14618 the full version number and not just 0. This should mark the
14619 shared libraries as not backward compatible. Of course, this should
14620 be changed again when we can guarantee backward binary compatibility.
14624 * Fix typo in get_cert_by_subject() in by_dir.c
14626 *Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>*
14628 * Rework the system to generate shared libraries:
14630 - Make note of the expected extension for the shared libraries and
14631 if there is a need for symbolic links from for example libcrypto.so.0
14632 to libcrypto.so.0.9.7. There is extended info in Configure for
14635 - Make as few rebuilds of the shared libraries as possible.
14637 - Still avoid linking the OpenSSL programs with the shared libraries.
14639 - When installing, install the shared libraries separately from the
14644 * Fix SSL_CTX_set_read_ahead macro to actually use its argument.
14646 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
14647 and not in SSL_clear because the latter is also used by the
14648 accept/connect functions; previously, the settings made by
14649 SSL_set_read_ahead would be lost during the handshake.
14651 *Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>*
14653 * Correct util/mkdef.pl to be selective about disabled algorithms.
14654 Previously, it would create entries for disabled algorithms no
14659 * Added several new manual pages for SSL_* function.
14663 ### Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
14665 * In ssl23_get_client_hello, generate an error message when faced
14666 with an initial SSL 3.0/TLS record that is too small to contain the
14667 first two bytes of the ClientHello message, i.e. client_version.
14668 (Note that this is a pathologic case that probably has never happened
14669 in real life.) The previous approach was to use the version number
14670 from the record header as a substitute; but our protocol choice
14671 should not depend on that one because it is not authenticated
14672 by the Finished messages.
14676 * More robust randomness gathering functions for Windows.
14678 *Jeffrey Altman <jaltman@columbia.edu>*
14680 * For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
14681 not set then we don't setup the error code for issuer check errors
14682 to avoid possibly overwriting other errors which the callback does
14683 handle. If an application does set the flag then we assume it knows
14684 what it is doing and can handle the new informational codes
14689 * Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
14690 a general "ANY" type, as such it should be able to decode anything
14691 including tagged types. However it didn't check the class so it would
14692 wrongly interpret tagged types in the same way as their universal
14693 counterpart and unknown types were just rejected. Changed so that the
14694 tagged and unknown types are handled in the same way as a SEQUENCE:
14695 that is the encoding is stored intact. There is also a new type
14696 "V_ASN1_OTHER" which is used when the class is not universal, in this
14697 case we have no idea what the actual type is so we just lump them all
14702 * On VMS, stdout may very well lead to a file that is written to
14703 in a record-oriented fashion. That means that every write() will
14704 write a separate record, which will be read separately by the
14705 programs trying to read from it. This can be very confusing.
14707 The solution is to put a BIO filter in the way that will buffer
14708 text until a linefeed is reached, and then write everything a
14709 line at a time, so every record written will be an actual line,
14710 not chunks of lines and not (usually doesn't happen, but I've
14711 seen it once) several lines in one record. BIO_f_linebuffer() is
14714 Currently, it's a VMS-only method, because that's where it has
14715 been tested well enough.
14719 * Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
14720 it can return incorrect results.
14721 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
14722 but it was in 0.9.6-beta[12].)
14726 * Disable the check for content being present when verifying detached
14727 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
14728 include zero length content when signing messages.
14732 * New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
14733 BIO_ctrl (for BIO pairs).
14737 * Add DSO method for VMS.
14741 * Bug fix: Montgomery multiplication could produce results with the
14746 * Add RPM specification openssl.spec and modify it to build three
14747 packages. The default package contains applications, application
14748 documentation and run-time libraries. The devel package contains
14749 include files, static libraries and function documentation. The
14750 doc package contains the contents of the doc directory. The original
14751 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
14755 * Add a large number of documentation files for many SSL routines.
14757 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>*
14759 * Add a configuration entry for Sony News 4.
14761 *NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>*
14763 * Don't set the two most significant bits to one when generating a
14764 random number < q in the DSA library.
14768 * New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
14769 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
14770 the underlying transport is blocking) if a handshake took place.
14771 (The default behaviour is needed by applications such as s_client
14772 and s_server that use select() to determine when to use SSL_read;
14773 but for applications that know in advance when to expect data, it
14774 just makes things more complicated.)
14778 * Add RAND_egd_bytes(), which gives control over the number of bytes read
14783 * Add a few more EBCDIC conditionals that make `req` and `x509`
14784 work better on such systems.
14786 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
14788 * Add two demo programs for PKCS12_parse() and PKCS12_create().
14789 Update PKCS12_parse() so it copies the friendlyName and the
14790 keyid to the certificates aux info.
14794 * Fix bug in PKCS7_verify() which caused an infinite loop
14795 if there was more than one signature.
14797 *Sven Uszpelkat <su@celocom.de>*
14799 * Major change in util/mkdef.pl to include extra information
14800 about each symbol, as well as presenting variables as well
14801 as functions. This change means that there's n more need
14802 to rebuild the .num files when some algorithms are excluded.
14806 * Allow the verify time to be set by an application,
14807 rather than always using the current time.
14811 * Phase 2 verify code reorganisation. The certificate
14812 verify code now looks up an issuer certificate by a
14813 number of criteria: subject name, authority key id
14814 and key usage. It also verifies self signed certificates
14815 by the same criteria. The main comparison function is
14816 X509_check_issued() which performs these checks.
14818 Lot of changes were necessary in order to support this
14819 without completely rewriting the lookup code.
14821 Authority and subject key identifier are now cached.
14823 The LHASH 'certs' is X509_STORE has now been replaced
14824 by a STACK_OF(X509_OBJECT). This is mainly because an
14825 LHASH can't store or retrieve multiple objects with
14826 the same hash value.
14828 As a result various functions (which were all internal
14829 use only) have changed to handle the new X509_STORE
14830 structure. This will break anything that messed round
14831 with X509_STORE internally.
14833 The functions X509_STORE_add_cert() now checks for an
14834 exact match, rather than just subject name.
14836 The X509_STORE API doesn't directly support the retrieval
14837 of multiple certificates matching a given criteria, however
14838 this can be worked round by performing a lookup first
14839 (which will fill the cache with candidate certificates)
14840 and then examining the cache for matches. This is probably
14841 the best we can do without throwing out X509_LOOKUP
14842 entirely (maybe later...).
14844 The X509_VERIFY_CTX structure has been enhanced considerably.
14846 All certificate lookup operations now go via a get_issuer()
14847 callback. Although this currently uses an X509_STORE it
14848 can be replaced by custom lookups. This is a simple way
14849 to bypass the X509_STORE hackery necessary to make this
14850 work and makes it possible to use more efficient techniques
14851 in future. A very simple version which uses a simple
14852 STACK for its trusted certificate store is also provided
14853 using X509_STORE_CTX_trusted_stack().
14855 The verify_cb() and verify() callbacks now have equivalents
14856 in the X509_STORE_CTX structure.
14858 X509_STORE_CTX also has a 'flags' field which can be used
14859 to customise the verify behaviour.
14863 * Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
14864 excludes S/MIME capabilities.
14868 * When a certificate request is read in keep a copy of the
14869 original encoding of the signed data and use it when outputting
14870 again. Signatures then use the original encoding rather than
14871 a decoded, encoded version which may cause problems if the
14872 request is improperly encoded.
14876 * For consistency with other BIO_puts implementations, call
14877 buffer_write(b, ...) directly in buffer_puts instead of calling
14880 In BIO_puts, increment b->num_write as in BIO_write.
14882 *Peter.Sylvester@EdelWeb.fr*
14884 * Fix BN_mul_word for the case where the word is 0. (We have to use
14885 BN_zero, we may not return a BIGNUM with an array consisting of
14886 words set to zero.)
14890 * Avoid calling abort() from within the library when problems are
14891 detected, except if preprocessor symbols have been defined
14892 (such as REF_CHECK, BN_DEBUG etc.).
14896 * New openssl application 'rsautl'. This utility can be
14897 used for low-level RSA operations. DER public key
14898 BIO/fp routines also added.
14902 * New Configure entry and patches for compiling on QNX 4.
14904 *Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>*
14906 * A demo state-machine implementation was sponsored by
14907 Nuron (<http://www.nuron.com/>) and is now available in
14908 demos/state_machine.
14912 * New options added to the 'dgst' utility for signature
14913 generation and verification.
14917 * Unrecognized PKCS#7 content types are now handled via a
14918 catch all ASN1_TYPE structure. This allows unsupported
14919 types to be stored as a "blob" and an application can
14920 encode and decode it manually.
14924 * Fix various signed/unsigned issues to make a_strex.c
14925 compile under VC++.
14927 *Oscar Jacobsson <oscar.jacobsson@celocom.com>*
14929 * ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
14930 length if passed a buffer. ASN1_INTEGER_to_BN failed
14931 if passed a NULL BN and its argument was negative.
14933 *Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>*
14935 * Modification to PKCS#7 encoding routines to output definite
14936 length encoding. Since currently the whole structures are in
14937 memory there's not real point in using indefinite length
14938 constructed encoding. However if OpenSSL is compiled with
14939 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
14943 * Added BIO_vprintf() and BIO_vsnprintf().
14947 * Added more prefixes to parse for in the strings written
14948 through a logging bio, to cover all the levels that are available
14949 through syslog. The prefixes are now:
14951 PANIC, EMERG, EMR => LOG_EMERG
14952 ALERT, ALR => LOG_ALERT
14953 CRIT, CRI => LOG_CRIT
14954 ERROR, ERR => LOG_ERR
14955 WARNING, WARN, WAR => LOG_WARNING
14956 NOTICE, NOTE, NOT => LOG_NOTICE
14957 INFO, INF => LOG_INFO
14958 DEBUG, DBG => LOG_DEBUG
14960 and as before, if none of those prefixes are present at the
14961 beginning of the string, LOG_ERR is chosen.
14963 On Win32, the `LOG_*` levels are mapped according to this:
14965 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
14966 LOG_WARNING => EVENTLOG_WARNING_TYPE
14967 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
14971 * Made it possible to reconfigure with just the configuration
14972 argument "reconf" or "reconfigure". The command line arguments
14973 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
14974 and are retrieved from there when reconfiguring.
14980 *Assar Westerlund <assar@sics.se>, Richard Levitte*
14982 * Add the arguments -CAfile and -CApath to the pkcs12 utility.
14986 * The obj_dat.pl script was messing up the sorting of object
14987 names. The reason was that it compared the quoted version
14988 of strings as a result "OCSP" > "OCSP Signing" because
14989 " > SPACE. Changed script to store unquoted versions of
14990 names and add quotes on output. It was also omitting some
14991 names from the lookup table if they were given a default
14992 value (that is if SN is missing it is given the same
14993 value as LN and vice versa), these are now added on the
14994 grounds that if an object has a name we should be able to
14995 look it up. Finally added warning output when duplicate
14996 short or long names are found.
15000 * Changes needed for Tandem NSK.
15002 *Scott Uroff <scott@xypro.com>*
15004 * Fix SSL 2.0 rollback checking: Due to an off-by-one error in
15005 RSA_padding_check_SSLv23(), special padding was never detected
15006 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
15007 version rollback attacks was not effective.
15009 In s23_clnt.c, don't use special rollback-attack detection padding
15010 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
15011 client; similarly, in s23_srvr.c, don't do the rollback check if
15012 SSL 2.0 is the only protocol enabled in the server.
15016 * Make it possible to get hexdumps of unprintable data with 'openssl
15017 asn1parse'. By implication, the functions ASN1_parse_dump() and
15018 BIO_dump_indent() are added.
15022 * New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
15023 these print out strings and name structures based on various
15024 flags including RFC2253 support and proper handling of
15025 multibyte characters. Added options to the 'x509' utility
15026 to allow the various flags to be set.
15030 * Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
15031 Also change the functions X509_cmp_current_time() and
15032 X509_gmtime_adj() work with an ASN1_TIME structure,
15033 this will enable certificates using GeneralizedTime in validity
15034 dates to be checked.
15038 * Make the NEG_PUBKEY_BUG code (which tolerates invalid
15039 negative public key encodings) on by default,
15040 NO_NEG_PUBKEY_BUG can be set to disable it.
15044 * New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
15045 content octets. An i2c_ASN1_OBJECT is unnecessary because
15046 the encoding can be trivially obtained from the structure.
15050 * crypto/err.c locking bugfix: Use write locks (`CRYPTO_w_[un]lock`),
15051 not read locks (`CRYPTO_r_[un]lock`).
15055 * A first attempt at creating official support for shared
15056 libraries through configuration. I've kept it so the
15057 default is static libraries only, and the OpenSSL programs
15058 are always statically linked for now, but there are
15059 preparations for dynamic linking in place.
15060 This has been tested on Linux and Tru64.
15064 * Randomness polling function for Win9x, as described in:
15065 Peter Gutmann, Software Generation of Practically Strong
15070 * Fix so PRNG is seeded in req if using an already existing
15075 * New options to smime application. -inform and -outform
15076 allow alternative formats for the S/MIME message including
15077 PEM and DER. The -content option allows the content to be
15078 specified separately. This should allow things like Netscape
15079 form signing output easier to verify.
15083 * Fix the ASN1 encoding of tags using the 'long form'.
15087 * New ASN1 functions, `i2c_*` and `c2i_*` for INTEGER and BIT
15088 STRING types. These convert content octets to and from the
15089 underlying type. The actual tag and length octets are
15090 already assumed to have been read in and checked. These
15091 are needed because all other string types have virtually
15092 identical handling apart from the tag. By having versions
15093 of the ASN1 functions that just operate on content octets
15094 IMPLICIT tagging can be handled properly. It also allows
15095 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
15096 and ASN1_INTEGER are identical apart from the tag.
15100 * Change the handling of OID objects as follows:
15102 - New object identifiers are inserted in objects.txt, following
15103 the syntax given in [crypto/objects/README.md](crypto/objects/README.md).
15104 - objects.pl is used to process obj_mac.num and create a new
15106 - obj_dat.pl is used to create a new obj_dat.h, using the data in
15109 This is currently kind of a hack, and the perl code in objects.pl
15110 isn't very elegant, but it works as I intended. The simplest way
15111 to check that it worked correctly is to look in obj_dat.h and
15112 check the array nid_objs and make sure the objects haven't moved
15113 around (this is important!). Additions are OK, as well as
15114 consistent name changes.
15118 * Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
15122 * Addition of the command line parameter '-rand file' to 'openssl req'.
15123 The given file adds to whatever has already been seeded into the
15124 random pool through the RANDFILE configuration file option or
15125 environment variable, or the default random state file.
15129 * mkstack.pl now sorts each macro group into lexical order.
15130 Previously the output order depended on the order the files
15131 appeared in the directory, resulting in needless rewriting
15136 * Patches to make OpenSSL compile under Win32 again. Mostly
15137 work arounds for the VC++ problem that it treats func() as
15138 func(void). Also stripped out the parts of mkdef.pl that
15139 added extra typesafe functions: these no longer exist.
15143 * Reorganisation of the stack code. The macros are now all
15144 collected in safestack.h . Each macro is defined in terms of
15145 a "stack macro" of the form `SKM_<name>(type, a, b)`. The
15146 DEBUG_SAFESTACK is now handled in terms of function casts,
15147 this has the advantage of retaining type safety without the
15148 use of additional functions. If DEBUG_SAFESTACK is not defined
15149 then the non typesafe macros are used instead. Also modified the
15150 mkstack.pl script to handle the new form. Needs testing to see
15151 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
15152 the default if no major problems. Similar behaviour for ASN1_SET_OF
15153 and PKCS12_STACK_OF.
15157 * When some versions of IIS use the 'NET' form of private key the
15158 key derivation algorithm is different. Normally MD5(password) is
15159 used as a 128 bit RC4 key. In the modified case
15160 MD5(MD5(password) + "SGCKEYSALT") is used instead. Added some
15161 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
15162 as the old Netscape_RSA functions except they have an additional
15163 'sgckey' parameter which uses the modified algorithm. Also added
15164 an -sgckey command line option to the rsa utility. Thanks to
15165 Adrian Peck <bertie@ncipher.com> for posting details of the modified
15166 algorithm to openssl-dev.
15170 * The evp_local.h macros were using 'c.##kname' which resulted in
15171 invalid expansion on some systems (SCO 5.0.5 for example).
15172 Corrected to 'c.kname'.
15174 *Phillip Porch <root@theporch.com>*
15176 * New X509_get1_email() and X509_REQ_get1_email() functions that return
15177 a STACK of email addresses from a certificate or request, these look
15178 in the subject name and the subject alternative name extensions and
15179 omit any duplicate addresses.
15183 * Re-implement BN_mod_exp2_mont using independent (and larger) windows.
15184 This makes DSA verification about 2 % faster.
15188 * Increase maximum window size in `BN_mod_exp_...` to 6 bits instead of 5
15189 (meaning that now 2^5 values will be precomputed, which is only 4 KB
15190 plus overhead for 1024 bit moduli).
15191 This makes exponentiations about 0.5 % faster for 1024 bit
15192 exponents (as measured by "openssl speed rsa2048").
15196 * Rename memory handling macros to avoid conflicts with other
15198 Malloc => OPENSSL_malloc
15199 Malloc_locked => OPENSSL_malloc_locked
15200 Realloc => OPENSSL_realloc
15201 Free => OPENSSL_free
15205 * New function BN_mod_exp_mont_word for small bases (roughly 15%
15206 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
15210 * CygWin32 support.
15212 *John Jarvie <jjarvie@newsguy.com>*
15214 * The type-safe stack code has been rejigged. It is now only compiled
15215 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
15216 by default all type-specific stack functions are "#define"d back to
15217 standard stack functions. This results in more streamlined output
15218 but retains the type-safety checking possibilities of the original
15223 * The STACK code has been cleaned up, and certain type declarations
15224 that didn't make a lot of sense have been brought in line. This has
15225 also involved a cleanup of sorts in safestack.h to more correctly
15226 map type-safe stack functions onto their plain stack counterparts.
15227 This work has also resulted in a variety of "const"ifications of
15228 lots of the code, especially `_cmp` operations which should normally
15229 be prototyped with "const" parameters anyway.
15233 * When generating bytes for the first time in md_rand.c, 'stir the pool'
15234 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
15235 (The PRNG state consists of two parts, the large pool 'state' and 'md',
15236 where all of 'md' is used each time the PRNG is used, but 'state'
15237 is used only indexed by a cyclic counter. As entropy may not be
15238 well distributed from the beginning, 'md' is important as a
15239 chaining variable. However, the output function chains only half
15240 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
15241 all of 'md', and seeding with STATE_SIZE dummy bytes will result
15242 in all of 'state' being rewritten, with the new values depending
15243 on virtually all of 'md'. This overcomes the 80 bit limitation.)
15247 * In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
15248 the handshake is continued after ssl_verify_cert_chain();
15249 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
15250 can lead to 'unexplainable' connection aborts later.
15252 *Bodo Moeller; problem tracked down by Lutz Jaenicke*
15254 * Major EVP API cipher revision.
15255 Add hooks for extra EVP features. This allows various cipher
15256 parameters to be set in the EVP interface. Support added for variable
15257 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
15258 setting of RC2 and RC5 parameters.
15260 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
15263 Remove lots of duplicated code from the EVP library. For example *every*
15264 cipher init() function handles the 'iv' in the same way according to the
15265 cipher mode. They also all do nothing if the 'key' parameter is NULL and
15266 for CFB and OFB modes they zero ctx->num.
15268 New functionality allows removal of S/MIME code RC2 hack.
15270 Most of the routines have the same form and so can be declared in terms
15273 By shifting this to the top level EVP_CipherInit() it can be removed from
15274 all individual ciphers. If the cipher wants to handle IVs or keys
15275 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
15278 Change lots of functions like EVP_EncryptUpdate() to now return a
15279 value: although software versions of the algorithms cannot fail
15280 any installed hardware versions can.
15284 * Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
15285 this option is set, tolerate broken clients that send the negotiated
15286 protocol version number instead of the requested protocol version
15291 * Call dh_tmp_cb (set by `..._TMP_DH_CB`) with correct 'is_export' flag;
15292 i.e. non-zero for export ciphersuites, zero otherwise.
15293 Previous versions had this flag inverted, inconsistent with
15294 rsa_tmp_cb (..._TMP_RSA_CB).
15296 *Bodo Moeller; problem reported by Amit Chopra*
15298 * Add missing DSA library text string. Work around for some IIS
15299 key files with invalid SEQUENCE encoding.
15303 * Add a document (doc/standards.txt) that list all kinds of standards
15304 and so on that are implemented in OpenSSL.
15308 * Enhance c_rehash script. Old version would mishandle certificates
15309 with the same subject name hash and wouldn't handle CRLs at all.
15310 Added -fingerprint option to crl utility, to support new c_rehash
15315 * Eliminate non-ANSI declarations in crypto.h and stack.h.
15319 * Fix for SSL server purpose checking. Server checking was
15320 rejecting certificates which had extended key usage present
15321 but no ssl client purpose.
15323 *Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>*
15325 * Make PKCS#12 code work with no password. The PKCS#12 spec
15326 is a little unclear about how a blank password is handled.
15327 Since the password in encoded as a BMPString with terminating
15328 double NULL a zero length password would end up as just the
15329 double NULL. However no password at all is different and is
15330 handled differently in the PKCS#12 key generation code. NS
15331 treats a blank password as zero length. MSIE treats it as no
15332 password on export: but it will try both on import. We now do
15333 the same: PKCS12_parse() tries zero length and no password if
15334 the password is set to "" or NULL (NULL is now a valid password:
15335 it wasn't before) as does the pkcs12 application.
15339 * Bugfixes in `apps/x509.c`: Avoid a memory leak; and don't use
15340 perror when PEM_read_bio_X509_REQ fails, the error message must
15341 be obtained from the error queue.
15345 * Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
15346 it in ERR_remove_state if appropriate, and change ERR_get_state
15347 accordingly to avoid race conditions (this is necessary because
15348 thread_hash is no longer constant once set).
15352 * Bugfix for linux-elf makefile.one.
15356 * RSA_get_default_method() will now cause a default
15357 RSA_METHOD to be chosen if one doesn't exist already.
15358 Previously this was only set during a call to RSA_new()
15359 or RSA_new_method(NULL) meaning it was possible for
15360 RSA_get_default_method() to return NULL.
15364 * Added native name translation to the existing DSO code
15365 that will convert (if the flag to do so is set) filenames
15366 that are sufficiently small and have no path information
15367 into a canonical native form. Eg. "blah" converted to
15368 "libblah.so" or "blah.dll" etc.
15372 * New function ERR_error_string_n(e, buf, len) which is like
15373 ERR_error_string(e, buf), but writes at most 'len' bytes
15374 including the 0 terminator. For ERR_error_string_n, 'buf'
15377 *Damien Miller <djm@mindrot.org>, Bodo Moeller*
15379 * CONF library reworked to become more general. A new CONF
15380 configuration file reader "class" is implemented as well as a
15381 new functions (`NCONF_*`, for "New CONF") to handle it. The now
15382 old `CONF_*` functions are still there, but are reimplemented to
15383 work in terms of the new functions. Also, a set of functions
15384 to handle the internal storage of the configuration data is
15385 provided to make it easier to write new configuration file
15386 reader "classes" (I can definitely see something reading a
15387 configuration file in XML format, for example), called `_CONF_*`,
15388 or "the configuration storage API"...
15390 The new configuration file reading functions are:
15392 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
15393 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
15395 NCONF_default, NCONF_WIN32
15397 NCONF_dump_fp, NCONF_dump_bio
15399 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
15400 NCONF_new creates a new CONF object. This works in the same way
15401 as other interfaces in OpenSSL, like the BIO interface.
15402 `NCONF_dump_*` dump the internal storage of the configuration file,
15403 which is useful for debugging. All other functions take the same
15404 arguments as the old `CONF_*` functions with the exception of the
15405 first that must be a `CONF *` instead of a `LHASH *`.
15407 To make it easier to use the new classes with the old `CONF_*` functions,
15408 the function CONF_set_default_method is provided.
15412 * Add '-tls1' option to 'openssl ciphers', which was already
15413 mentioned in the documentation but had not been implemented.
15414 (This option is not yet really useful because even the additional
15415 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
15419 * Initial DSO code added into libcrypto for letting OpenSSL (and
15420 OpenSSL-based applications) load shared libraries and bind to
15421 them in a portable way.
15423 *Geoff Thorpe, with contributions from Richard Levitte*
15425 ### Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
15427 * Make sure _lrotl and _lrotr are only used with MSVC.
15429 * Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
15430 (the default implementation of RAND_status).
15432 * Rename openssl x509 option '-crlext', which was added in 0.9.5,
15433 to '-clrext' (= clear extensions), as intended and documented.
15434 *Bodo Moeller; inconsistency pointed out by Michael Attili
15435 <attili@amaxo.com>*
15437 * Fix for HMAC. It wasn't zeroing the rest of the block if the key length
15438 was larger than the MD block size.
15440 *Steve Henson, pointed out by Yost William <YostW@tce.com>*
15442 * Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
15443 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
15444 using the passed key: if the passed key was a private key the result
15445 of X509_print(), for example, would be to print out all the private key
15450 * des_quad_cksum() byte order bug fix.
15451 *Ulf Möller, using the problem description in krb4-0.9.7, where
15452 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>*
15454 * Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
15457 *Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>*
15459 * For easily testing in shell scripts whether some command
15460 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
15461 returns with exit code 0 iff no command of the given name is available.
15462 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
15463 the output goes to stdout and nothing is printed to stderr.
15464 Additional arguments are always ignored.
15466 Since for each cipher there is a command of the same name,
15467 the 'no-cipher' compilation switches can be tested this way.
15469 ('openssl no-XXX' is not able to detect pseudo-commands such
15470 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
15474 * Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
15478 * For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
15479 is set; it will be thrown away anyway because each handshake creates
15481 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
15482 to parameters -- in previous versions (since OpenSSL 0.9.3) the
15483 'default key' from SSL_CTX_set_tmp_dh would always be lost, meaning
15484 you effectively got SSL_OP_SINGLE_DH_USE when using this macro.
15488 * New s_client option -ign_eof: EOF at stdin is ignored, and
15489 'Q' and 'R' lose their special meanings (quit/renegotiate).
15490 This is part of what -quiet does; unlike -quiet, -ign_eof
15491 does not suppress any output.
15495 * Add compatibility options to the purpose and trust code. The
15496 purpose X509_PURPOSE_ANY is "any purpose" which automatically
15497 accepts a certificate or CA, this was the previous behaviour,
15498 with all the associated security issues.
15500 X509_TRUST_COMPAT is the old trust behaviour: only and
15501 automatically trust self signed roots in certificate store. A
15502 new trust setting X509_TRUST_DEFAULT is used to specify that
15503 a purpose has no associated trust setting and it should instead
15504 use the value in the default purpose.
15508 * Fix the PKCS#8 DSA private key code so it decodes keys again
15509 and fix a memory leak.
15513 * In util/mkerr.pl (which implements 'make errors'), preserve
15514 reason strings from the previous version of the .c file, as
15515 the default to have only downcase letters (and digits) in
15516 automatically generated reasons codes is not always appropriate.
15520 * In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
15521 using strerror. Previously, ERR_reason_error_string() returned
15522 library names as reason strings for SYSerr; but SYSerr is a special
15523 case where small numbers are errno values, not library numbers.
15527 * Add '-dsaparam' option to 'openssl dhparam' application. This
15528 converts DSA parameters into DH parameters. (When creating parameters,
15529 DSA_generate_parameters is used.)
15533 * Include 'length' (recommended exponent length) in C code generated
15534 by 'openssl dhparam -C'.
15538 * The second argument to set_label in perlasm was already being used
15539 so couldn't be used as a "file scope" flag. Moved to third argument
15544 * In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
15545 instead of RAND_bytes for encryption IVs and salts.
15549 * Include RAND_status() into RAND_METHOD instead of implementing
15550 it only for md_rand.c Otherwise replacing the PRNG by calling
15551 RAND_set_rand_method would be impossible.
15555 * Don't let DSA_generate_key() enter an infinite loop if the random
15556 number generation fails.
15560 * New 'rand' application for creating pseudo-random output.
15564 * Added configuration support for Linux/IA64
15566 *Rolf Haberrecker <rolf@suse.de>*
15568 * Assembler module support for Mingw32.
15572 * Shared library support for HPUX (in shlib/).
15574 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous*
15576 * Shared library support for Solaris gcc.
15578 *Lutz Behnke <behnke@trustcenter.de>*
15580 ### Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
15582 * PKCS7_encrypt() was adding text MIME headers twice because they
15583 were added manually and by SMIME_crlf_copy().
15587 * In bntest.c don't call BN_rand with zero bits argument.
15589 *Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>*
15591 * BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
15592 case was implemented. This caused BN_div_recp() to fail occasionally.
15596 * Add an optional second argument to the set_label() in the perl
15597 assembly language builder. If this argument exists and is set
15598 to 1 it signals that the assembler should use a symbol whose
15599 scope is the entire file, not just the current function. This
15600 is needed with MASM which uses the format label:: for this scope.
15602 *Steve Henson, pointed out by Peter Runestig <peter@runestig.com>*
15604 * Change the ASN1 types so they are typedefs by default. Before
15605 almost all types were #define'd to ASN1_STRING which was causing
15606 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
15611 * Change names of new functions to the new get1/get0 naming
15612 convention: After 'get1', the caller owns a reference count
15613 and has to call `..._free`; 'get0' returns a pointer to some
15614 data structure without incrementing reference counters.
15615 (Some of the existing 'get' functions increment a reference
15616 counter, some don't.)
15617 Similarly, 'set1' and 'add1' functions increase reference
15618 counters or duplicate objects.
15622 * Allow for the possibility of temp RSA key generation failure:
15623 the code used to assume it always worked and crashed on failure.
15627 * Fix potential buffer overrun problem in BIO_printf().
15628 *Ulf Möller, using public domain code by Patrick Powell; problem
15629 pointed out by David Sacerdote <das33@cornell.edu>*
15631 * Support EGD <http://www.lothar.com/tech/crypto/>. New functions
15632 RAND_egd() and RAND_status(). In the command line application,
15633 the EGD socket can be specified like a seed file using RANDFILE
15638 * Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
15639 Some CAs (e.g. Verisign) distribute certificates in this form.
15643 * Remove the SSL_ALLOW_ADH compile option and set the default cipher
15644 list to exclude them. This means that no special compilation option
15645 is needed to use anonymous DH: it just needs to be included in the
15650 * Change the EVP_MD_CTX_type macro so its meaning consistent with
15651 EVP_MD_type. The old functionality is available in a new macro called
15652 EVP_MD_md(). Change code that uses it and update docs.
15656 * `..._ctrl` functions now have corresponding `..._callback_ctrl` functions
15657 where the `void *` argument is replaced by a function pointer argument.
15658 Previously `void *` was abused to point to functions, which works on
15659 many platforms, but is not correct. As these functions are usually
15660 called by macros defined in OpenSSL header files, most source code
15661 should work without changes.
15665 * `<openssl/opensslconf.h>` (which is created by Configure) now contains
15666 sections with information on -D... compiler switches used for
15667 compiling the library so that applications can see them. To enable
15668 one of these sections, a pre-processor symbol `OPENSSL_..._DEFINES`
15669 must be defined. E.g.,
15670 #define OPENSSL_ALGORITHM_DEFINES
15671 #include <openssl/opensslconf.h>
15672 defines all pertinent `NO_<algo>` symbols, such as NO_IDEA, NO_RSA, etc.
15674 *Richard Levitte, Ulf and Bodo Möller*
15676 * Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
15681 * Change the 'other' type in certificate aux info to a STACK_OF
15682 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
15683 the required ASN1 format: arbitrary types determined by an OID.
15687 * Add some PEM_write_X509_REQ_NEW() functions and a command line
15688 argument to 'req'. This is not because the function is newer or
15689 better than others it just uses the work 'NEW' in the certificate
15690 request header lines. Some software needs this.
15694 * Reorganise password command line arguments: now passwords can be
15695 obtained from various sources. Delete the PEM_cb function and make
15696 it the default behaviour: i.e. if the callback is NULL and the
15697 usrdata argument is not NULL interpret it as a null terminated pass
15698 phrase. If usrdata and the callback are NULL then the pass phrase
15699 is prompted for as usual.
15703 * Add support for the Compaq Atalla crypto accelerator. If it is installed,
15704 the support is automatically enabled. The resulting binaries will
15705 autodetect the card and use it if present.
15707 *Ben Laurie and Compaq Inc.*
15709 * Work around for Netscape hang bug. This sends certificate request
15710 and server done in one record. Since this is perfectly legal in the
15711 SSL/TLS protocol it isn't a "bug" option and is on by default. See
15712 the bugs/SSLv3 entry for more info.
15716 * HP-UX tune-up: new unified configs, HP C compiler bug workaround.
15720 * Add -rand argument to smime and pkcs12 applications and read/write
15725 * New 'passwd' tool for crypt(3) and apr1 password hashes.
15729 * Add command line password options to the remaining applications.
15733 * Bug fix for BN_div_recp() for numerators with an even number of
15738 * More tests in bntest.c, and changed test_bn output.
15742 * ./config recognizes MacOS X now.
15746 * Bug fix for BN_div() when the first words of num and divisor are
15747 equal (it gave wrong results if `(rem=(n1-q*d0)&BN_MASK2) < d0)`.
15751 * Add support for various broken PKCS#8 formats, and command line
15752 options to produce them.
15756 * New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
15757 get temporary BIGNUMs from a BN_CTX.
15761 * Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
15766 * Change the `SSLeay_add_all_*()` functions to `OpenSSL_add_all_*()` and
15767 include a #define from the old name to the new. The original intent
15768 was that statically linked binaries could for example just call
15769 SSLeay_add_all_ciphers() to just add ciphers to the table and not
15770 link with digests. This never worked because SSLeay_add_all_digests()
15771 and SSLeay_add_all_ciphers() were in the same source file so calling
15772 one would link with the other. They are now in separate source files.
15776 * Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
15780 * Use a less unusual form of the Miller-Rabin primality test (it used
15781 a binary algorithm for exponentiation integrated into the Miller-Rabin
15782 loop, our standard modexp algorithms are faster).
15786 * Support for the EBCDIC character set completed.
15788 *Martin Kraemer <Martin.Kraemer@Mch.SNI.De>*
15790 * Source code cleanups: use const where appropriate, eliminate casts,
15791 use `void *` instead of `char *` in lhash.
15795 * Bugfix: ssl3_send_server_key_exchange was not restartable
15796 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
15797 this the server could overwrite ephemeral keys that the client
15802 * Turn DSA_is_prime into a macro that calls BN_is_prime,
15803 using 50 iterations of the Rabin-Miller test.
15805 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
15806 iterations of the Rabin-Miller test as required by the appendix
15807 to FIPS PUB 186[-1]) instead of DSA_is_prime.
15808 As BN_is_prime_fasttest includes trial division, DSA parameter
15809 generation becomes much faster.
15811 This implies a change for the callback functions in DSA_is_prime
15812 and DSA_generate_parameters: The callback function is called once
15813 for each positive witness in the Rabin-Miller test, not just
15814 occasionally in the inner loop; and the parameters to the
15815 callback function now provide an iteration count for the outer
15816 loop rather than for the current invocation of the inner loop.
15817 DSA_generate_parameters additionally can call the callback
15818 function with an 'iteration count' of -1, meaning that a
15819 candidate has passed the trial division test (when q is generated
15820 from an application-provided seed, trial division is skipped).
15824 * New function BN_is_prime_fasttest that optionally does trial
15825 division before starting the Rabin-Miller test and has
15826 an additional BN_CTX * argument (whereas BN_is_prime always
15827 has to allocate at least one BN_CTX).
15828 'callback(1, -1, cb_arg)' is called when a number has passed the
15829 trial division stage.
15833 * Fix for bug in CRL encoding. The validity dates weren't being handled
15838 * New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
15842 * New function BN_pseudo_rand().
15846 * Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
15847 bignum version of BN_from_montgomery() with the working code from
15848 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
15853 * Avoid a race condition in s2_clnt.c (function get_server_hello) that
15854 made it impossible to use the same SSL_SESSION data structure in
15855 SSL2 clients in multiple threads.
15859 * The return value of RAND_load_file() no longer counts bytes obtained
15860 by stat(). RAND_load_file(..., -1) is new and uses the complete file
15861 to seed the PRNG (previously an explicit byte count was required).
15863 *Ulf Möller, Bodo Möller*
15865 * Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
15866 used `char *` instead of `void *` and had casts all over the place.
15870 * Make BN_generate_prime() return NULL on error if ret!=NULL.
15874 * Retain source code compatibility for BN_prime_checks macro:
15875 BN_is_prime(..., BN_prime_checks, ...) now uses
15876 BN_prime_checks_for_size to determine the appropriate number of
15877 Rabin-Miller iterations.
15881 * Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
15882 DH_CHECK_P_NOT_SAFE_PRIME.
15883 (Check if this is true? OpenPGP calls them "strong".)
15887 * Merge the functionality of "dh" and "gendh" programs into a new program
15888 "dhparam". The old programs are retained for now but will handle DH keys
15889 (instead of parameters) in future.
15893 * Make the ciphers, s_server and s_client programs check the return values
15894 when a new cipher list is set.
15898 * Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
15899 ciphers. Before when the 56bit ciphers were enabled the sorting was
15902 The syntax for the cipher sorting has been extended to support sorting by
15903 cipher-strength (using the strength_bits hard coded in the tables).
15904 The new command is `@STRENGTH` (see also `doc/apps/ciphers.pod`).
15906 Fix a bug in the cipher-command parser: when supplying a cipher command
15907 string with an "undefined" symbol (neither command nor alphanumeric
15908 *A-Za-z0-9*, ssl_set_cipher_list used to hang in an endless loop. Now
15909 an error is flagged.
15911 Due to the strength-sorting extension, the code of the
15912 ssl_create_cipher_list() function was completely rearranged. I hope that
15913 the readability was also increased :-)
15915 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>*
15917 * Minor change to 'x509' utility. The -CAcreateserial option now uses 1
15918 for the first serial number and places 2 in the serial number file. This
15919 avoids problems when the root CA is created with serial number zero and
15920 the first user certificate has the same issuer name and serial number
15925 * Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
15926 the new code. Add documentation for this stuff.
15930 * Changes to X509_ATTRIBUTE utilities. These have been renamed from
15931 `X509_*()` to `X509at_*()` on the grounds that they don't handle X509
15932 structures and behave in an analogous way to the X509v3 functions:
15933 they shouldn't be called directly but wrapper functions should be used
15936 So we also now have some wrapper functions that call the X509at functions
15937 when passed certificate requests. (TO DO: similar things can be done with
15938 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
15939 things. Some of these need some d2i or i2d and print functionality
15940 because they handle more complex structures.)
15944 * Add missing #ifndefs that caused missing symbols when building libssl
15945 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
15946 NO_RSA in `ssl/s2*.c`.
15948 *Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller*
15950 * Precautions against using the PRNG uninitialized: RAND_bytes() now
15951 has a return value which indicates the quality of the random data
15952 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
15953 error queue. New function RAND_pseudo_bytes() generates output that is
15954 guaranteed to be unique but not unpredictable. RAND_add is like
15955 RAND_seed, but takes an extra argument for an entropy estimate
15956 (RAND_seed always assumes full entropy).
15960 * Do more iterations of Rabin-Miller probable prime test (specifically,
15961 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
15962 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
15963 in crypto/bn/bn_prime.c for the complete table). This guarantees a
15964 false-positive rate of at most 2^-80 for random input.
15968 * Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
15972 * New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
15973 in the 0.9.5 release), this returns the chain
15974 from an X509_CTX structure with a dup of the stack and all
15975 the X509 reference counts upped: so the stack will exist
15976 after X509_CTX_cleanup() has been called. Modify pkcs12.c
15979 Also make SSL_SESSION_print() print out the verify return
15984 * Add manpage for the pkcs12 command. Also change the default
15985 behaviour so MAC iteration counts are used unless the new
15986 -nomaciter option is used. This improves file security and
15987 only older versions of MSIE (4.0 for example) need it.
15991 * Honor the no-xxx Configure options when creating .DEF files.
15995 * Add PKCS#10 attributes to field table: challengePassword,
15996 unstructuredName and unstructuredAddress. These are taken from
15997 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
15998 international characters are used.
16000 More changes to X509_ATTRIBUTE code: allow the setting of types
16001 based on strings. Remove the 'loc' parameter when adding
16002 attributes because these will be a SET OF encoding which is sorted
16007 * Initial changes to the 'req' utility to allow request generation
16008 automation. This will allow an application to just generate a template
16009 file containing all the field values and have req construct the
16012 Initial support for X509_ATTRIBUTE handling. Stacks of these are
16013 used all over the place including certificate requests and PKCS#7
16014 structures. They are currently handled manually where necessary with
16015 some primitive wrappers for PKCS#7. The new functions behave in a
16016 manner analogous to the X509 extension functions: they allow
16017 attributes to be looked up by NID and added.
16019 Later something similar to the X509V3 code would be desirable to
16020 automatically handle the encoding, decoding and printing of the
16021 more complex types. The string types like challengePassword can
16022 be handled by the string table functions.
16024 Also modified the multi byte string table handling. Now there is
16025 a 'global mask' which masks out certain types. The table itself
16026 can use the flag STABLE_NO_MASK to ignore the mask setting: this
16027 is useful when for example there is only one permissible type
16028 (as in countryName) and using the mask might result in no valid
16033 * Clean up 'Finished' handling, and add functions SSL_get_finished and
16034 SSL_get_peer_finished to allow applications to obtain the latest
16035 Finished messages sent to the peer or expected from the peer,
16036 respectively. (SSL_get_peer_finished is usually the Finished message
16037 actually received from the peer, otherwise the protocol will be aborted.)
16039 As the Finished message are message digests of the complete handshake
16040 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
16041 be used for external authentication procedures when the authentication
16042 provided by SSL/TLS is not desired or is not enough.
16046 * Enhanced support for Alpha Linux is added. Now ./config checks if
16047 the host supports BWX extension and if Compaq C is present on the
16048 $PATH. Just exploiting of the BWX extension results in 20-30%
16049 performance kick for some algorithms, e.g. DES and RC4 to mention
16050 a couple. Compaq C in turn generates ~20% faster code for MD5 and
16055 * Add support for MS "fast SGC". This is arguably a violation of the
16056 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
16057 weak crypto and after checking the certificate is SGC a second one
16058 with strong crypto. MS SGC stops the first handshake after receiving
16059 the server certificate message and sends a second client hello. Since
16060 a server will typically do all the time consuming operations before
16061 expecting any further messages from the client (server key exchange
16062 is the most expensive) there is little difference between the two.
16064 To get OpenSSL to support MS SGC we have to permit a second client
16065 hello message after we have sent server done. In addition we have to
16066 reset the MAC if we do get this second client hello.
16070 * Add a function 'd2i_AutoPrivateKey()' this will automatically decide
16071 if a DER encoded private key is RSA or DSA traditional format. Changed
16072 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
16073 format DER encoded private key. Newer code should use PKCS#8 format which
16074 has the key type encoded in the ASN1 structure. Added DER private key
16075 support to pkcs8 application.
16079 * SSL 3/TLS 1 servers now don't request certificates when an anonymous
16080 ciphersuites has been selected (as required by the SSL 3/TLS 1
16081 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
16082 is set, we interpret this as a request to violate the specification
16083 (the worst that can happen is a handshake failure, and 'correct'
16084 behaviour would result in a handshake failure anyway).
16088 * In SSL_CTX_add_session, take into account that there might be multiple
16089 SSL_SESSION structures with the same session ID (e.g. when two threads
16090 concurrently obtain them from an external cache).
16091 The internal cache can handle only one SSL_SESSION with a given ID,
16092 so if there's a conflict, we now throw out the old one to achieve
16097 * Add OIDs for idea and blowfish in CBC mode. This will allow both
16098 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
16099 some routines that use cipher OIDs: some ciphers do not have OIDs
16100 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
16105 * Simplify the trust setting structure and code. Now we just have
16106 two sequences of OIDs for trusted and rejected settings. These will
16107 typically have values the same as the extended key usage extension
16108 and any application specific purposes.
16110 The trust checking code now has a default behaviour: it will just
16111 check for an object with the same NID as the passed id. Functions can
16112 be provided to override either the default behaviour or the behaviour
16113 for a given id. SSL client, server and email already have functions
16114 in place for compatibility: they check the NID and also return "trusted"
16115 if the certificate is self signed.
16119 * Add d2i,i2d bio/fp functions for PrivateKey: these convert the
16120 traditional format into an EVP_PKEY structure.
16124 * Add a password callback function PEM_cb() which either prompts for
16125 a password if usr_data is NULL or otherwise assumes it is a null
16126 terminated password. Allow passwords to be passed on command line
16127 environment or config files in a few more utilities.
16131 * Add a bunch of DER and PEM functions to handle PKCS#8 format private
16132 keys. Add some short names for PKCS#8 PBE algorithms and allow them
16133 to be specified on the command line for the pkcs8 and pkcs12 utilities.
16134 Update documentation.
16138 * Support for ASN1 "NULL" type. This could be handled before by using
16139 ASN1_TYPE but there wasn't any function that would try to read a NULL
16140 and produce an error if it couldn't. For compatibility we also have
16141 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
16142 don't allocate anything because they don't need to.
16146 * Initial support for MacOS is now provided. Examine INSTALL.MacOS
16149 *Andy Polyakov, Roy Woods <roy@centicsystems.ca>*
16151 * Rebuild of the memory allocation routines used by OpenSSL code and
16152 possibly others as well. The purpose is to make an interface that
16153 provide hooks so anyone can build a separate set of allocation and
16154 deallocation routines to be used by OpenSSL, for example memory
16155 pool implementations, or something else, which was previously hard
16156 since Malloc(), Realloc() and Free() were defined as macros having
16157 the values malloc, realloc and free, respectively (except for Win32
16158 compilations). The same is provided for memory debugging code.
16159 OpenSSL already comes with functionality to find memory leaks, but
16160 this gives people a chance to debug other memory problems.
16162 With these changes, a new set of functions and macros have appeared:
16164 CRYPTO_set_mem_debug_functions() [F]
16165 CRYPTO_get_mem_debug_functions() [F]
16166 CRYPTO_dbg_set_options() [F]
16167 CRYPTO_dbg_get_options() [F]
16168 CRYPTO_malloc_debug_init() [M]
16170 The memory debug functions are NULL by default, unless the library
16171 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
16172 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
16173 gives the standard debugging functions that come with OpenSSL) or
16174 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
16175 provided by the library user) must be used. When the standard
16176 debugging functions are used, CRYPTO_dbg_set_options can be used to
16177 request additional information:
16178 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
16179 the CRYPTO_MDEBUG_xxx macro when compiling the library.
16181 Also, things like CRYPTO_set_mem_functions will always give the
16182 expected result (the new set of functions is used for allocation
16183 and deallocation) at all times, regardless of platform and compiler
16186 To finish it up, some functions that were never use in any other
16187 way than through macros have a new API and new semantic:
16189 CRYPTO_dbg_malloc()
16190 CRYPTO_dbg_realloc()
16193 All macros of value have retained their old syntax.
16195 *Richard Levitte and Bodo Moeller*
16197 * Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
16198 ordering of SMIMECapabilities wasn't in "strength order" and there
16199 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
16204 * Some ASN1 types with illegal zero length encoding (INTEGER,
16205 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
16207 *Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson*
16209 * Merge in my S/MIME library for OpenSSL. This provides a simple
16210 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
16211 functionality to handle multipart/signed properly) and a utility
16212 called 'smime' to call all this stuff. This is based on code I
16213 originally wrote for Celo who have kindly allowed it to be
16214 included in OpenSSL.
16218 * Add variants des_set_key_checked and des_set_key_unchecked of
16219 des_set_key (aka des_key_sched). Global variable des_check_key
16220 decides which of these is called by des_set_key; this way
16221 des_check_key behaves as it always did, but applications and
16222 the library itself, which was buggy for des_check_key == 1,
16223 have a cleaner way to pick the version they need.
16227 * New function PKCS12_newpass() which changes the password of a
16232 * Modify X509_TRUST and X509_PURPOSE so it also uses a static and
16233 dynamic mix. In both cases the ids can be used as an index into the
16234 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
16235 functions so they accept a list of the field values and the
16236 application doesn't need to directly manipulate the X509_TRUST
16241 * Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
16246 * Modify the way the V3 extension code looks up extensions. This now
16247 works in a similar way to the object code: we have some "standard"
16248 extensions in a static table which is searched with OBJ_bsearch()
16249 and the application can add dynamic ones if needed. The file
16250 crypto/x509v3/ext_dat.h now has the info: this file needs to be
16251 updated whenever a new extension is added to the core code and kept
16252 in ext_nid order. There is a simple program 'tabtest.c' which checks
16253 this. New extensions are not added too often so this file can readily
16254 be maintained manually.
16256 There are two big advantages in doing things this way. The extensions
16257 can be looked up immediately and no longer need to be "added" using
16258 X509V3_add_standard_extensions(): this function now does nothing.
16259 Side note: I get *lots* of email saying the extension code doesn't
16260 work because people forget to call this function.
16261 Also no dynamic allocation is done unless new extensions are added:
16262 so if we don't add custom extensions there is no need to call
16263 X509V3_EXT_cleanup().
16267 * Modify enc utility's salting as follows: make salting the default. Add a
16268 magic header, so unsalted files fail gracefully instead of just decrypting
16269 to garbage. This is because not salting is a big security hole, so people
16270 should be discouraged from doing it.
16274 * Fixes and enhancements to the 'x509' utility. It allowed a message
16275 digest to be passed on the command line but it only used this
16276 parameter when signing a certificate. Modified so all relevant
16277 operations are affected by the digest parameter including the
16278 -fingerprint and -x509toreq options. Also -x509toreq choked if a
16279 DSA key was used because it didn't fix the digest.
16283 * Initial certificate chain verify code. Currently tests the untrusted
16284 certificates for consistency with the verify purpose (which is set
16285 when the X509_STORE_CTX structure is set up) and checks the pathlength.
16287 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
16288 this is because it will reject chains with invalid extensions whereas
16289 every previous version of OpenSSL and SSLeay made no checks at all.
16291 Trust code: checks the root CA for the relevant trust settings. Trust
16292 settings have an initial value consistent with the verify purpose: e.g.
16293 if the verify purpose is for SSL client use it expects the CA to be
16294 trusted for SSL client use. However the default value can be changed to
16295 permit custom trust settings: one example of this would be to only trust
16296 certificates from a specific "secure" set of CAs.
16298 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
16299 which should be used for version portability: especially since the
16300 verify structure is likely to change more often now.
16302 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
16303 to set them. If not set then assume SSL clients will verify SSL servers
16306 Two new options to the verify program: -untrusted allows a set of
16307 untrusted certificates to be passed in and -purpose which sets the
16308 intended purpose of the certificate. If a purpose is set then the
16309 new chain verify code is used to check extension consistency.
16313 * Support for the authority information access extension.
16317 * Modify RSA and DSA PEM read routines to transparently handle
16318 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
16319 public keys in a format compatible with certificate
16320 SubjectPublicKeyInfo structures. Unfortunately there were already
16321 functions called *_PublicKey_* which used various odd formats so
16322 these are retained for compatibility: however the DSA variants were
16323 never in a public release so they have been deleted. Changed dsa/rsa
16324 utilities to handle the new format: note no releases ever handled public
16325 keys so we should be OK.
16327 The primary motivation for this change is to avoid the same fiasco
16328 that dogs private keys: there are several incompatible private key
16329 formats some of which are standard and some OpenSSL specific and
16330 require various evil hacks to allow partial transparent handling and
16331 even then it doesn't work with DER formats. Given the option anything
16332 other than PKCS#8 should be dumped: but the other formats have to
16333 stay in the name of compatibility.
16335 With public keys and the benefit of hindsight one standard format
16336 is used which works with EVP_PKEY, RSA or DSA structures: though
16337 it clearly returns an error if you try to read the wrong kind of key.
16339 Added a -pubkey option to the 'x509' utility to output the public key.
16340 Also rename the `EVP_PKEY_get_*()` to `EVP_PKEY_rget_*()`
16341 (renamed to `EVP_PKEY_get1_*()` in the OpenSSL 0.9.5 release) and add
16342 `EVP_PKEY_rset_*()` functions (renamed to `EVP_PKEY_set1_*()`)
16343 that do the same as the `EVP_PKEY_assign_*()` except they up the
16344 reference count of the added key (they don't "swallow" the
16349 * Fixes to crypto/x509/by_file.c the code to read in certificates and
16350 CRLs would fail if the file contained no certificates or no CRLs:
16351 added a new function to read in both types and return the number
16352 read: this means that if none are read it will be an error. The
16353 DER versions of the certificate and CRL reader would always fail
16354 because it isn't possible to mix certificates and CRLs in DER format
16355 without choking one or the other routine. Changed this to just read
16356 a certificate: this is the best we can do. Also modified the code
16357 in `apps/verify.c` to take notice of return codes: it was previously
16358 attempting to read in certificates from NULL pointers and ignoring
16359 any errors: this is one reason why the cert and CRL reader seemed
16360 to work. It doesn't check return codes from the default certificate
16361 routines: these may well fail if the certificates aren't installed.
16365 * Code to support otherName option in GeneralName.
16369 * First update to verify code. Change the verify utility
16370 so it warns if it is passed a self signed certificate:
16371 for consistency with the normal behaviour. X509_verify
16372 has been modified to it will now verify a self signed
16373 certificate if *exactly* the same certificate appears
16374 in the store: it was previously impossible to trust a
16375 single self signed certificate. This means that:
16376 openssl verify ss.pem
16377 now gives a warning about a self signed certificate but
16378 openssl verify -CAfile ss.pem ss.pem
16383 * For servers, store verify_result in SSL_SESSION data structure
16384 (and add it to external session representation).
16385 This is needed when client certificate verifications fails,
16386 but an application-provided verification callback (set by
16387 SSL_CTX_set_cert_verify_callback) allows accepting the session
16388 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
16389 but returns 1): When the session is reused, we have to set
16390 ssl->verify_result to the appropriate error code to avoid
16393 *Bodo Moeller, problem pointed out by Lutz Jaenicke*
16395 * Fix a bug in the new PKCS#7 code: it didn't consider the
16396 case in PKCS7_dataInit() where the signed PKCS7 structure
16397 didn't contain any existing data because it was being created.
16399 *Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson*
16401 * Add a salt to the key derivation routines in enc.c. This
16402 forms the first 8 bytes of the encrypted file. Also add a
16403 -S option to allow a salt to be input on the command line.
16407 * New function X509_cmp(). Oddly enough there wasn't a function
16408 to compare two certificates. We do this by working out the SHA1
16409 hash and comparing that. X509_cmp() will be needed by the trust
16414 * SSL_get1_session() is like SSL_get_session(), but increments
16415 the reference count in the SSL_SESSION returned.
16417 *Geoff Thorpe <geoff@eu.c2.net>*
16419 * Fix for 'req': it was adding a null to request attributes.
16420 Also change the X509_LOOKUP and X509_INFO code to handle
16421 certificate auxiliary information.
16425 * Add support for 40 and 64 bit RC2 and RC4 algorithms: document
16430 * Add the possibility to add extra information to the memory leak
16431 detecting output, to form tracebacks, showing from where each
16432 allocation was originated: CRYPTO_push_info("constant string") adds
16433 the string plus current file name and line number to a per-thread
16434 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
16435 is like calling CYRPTO_pop_info() until the stack is empty.
16436 Also updated memory leak detection code to be multi-thread-safe.
16440 * Add options -text and -noout to pkcs7 utility and delete the
16441 encryption options which never did anything. Update docs.
16445 * Add options to some of the utilities to allow the pass phrase
16446 to be included on either the command line (not recommended on
16447 OSes like Unix) or read from the environment. Update the
16448 manpages and fix a few bugs.
16452 * Add a few manpages for some of the openssl commands.
16456 * Fix the -revoke option in ca. It was freeing up memory twice,
16457 leaking and not finding already revoked certificates.
16461 * Extensive changes to support certificate auxiliary information.
16462 This involves the use of X509_CERT_AUX structure and X509_AUX
16463 functions. An X509_AUX function such as PEM_read_X509_AUX()
16464 can still read in a certificate file in the usual way but it
16465 will also read in any additional "auxiliary information". By
16466 doing things this way a fair degree of compatibility can be
16467 retained: existing certificates can have this information added
16468 using the new 'x509' options.
16470 Current auxiliary information includes an "alias" and some trust
16471 settings. The trust settings will ultimately be used in enhanced
16472 certificate chain verification routines: currently a certificate
16473 can only be trusted if it is self signed and then it is trusted
16478 * Fix assembler for Alpha (tested only on DEC OSF not Linux or `*BSD`).
16479 The problem was that one of the replacement routines had not been working
16480 since SSLeay releases. For now the offending routine has been replaced
16481 with non-optimised assembler. Even so, this now gives around 95%
16482 performance improvement for 1024 bit RSA signs.
16486 * Hack to fix PKCS#7 decryption when used with some unorthodox RC2
16487 handling. Most clients have the effective key size in bits equal to
16488 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
16489 A few however don't do this and instead use the size of the decrypted key
16490 to determine the RC2 key length and the AlgorithmIdentifier to determine
16491 the effective key length. In this case the effective key length can still
16492 be 40 bits but the key length can be 168 bits for example. This is fixed
16493 by manually forcing an RC2 key into the EVP_PKEY structure because the
16494 EVP code can't currently handle unusual RC2 key sizes: it always assumes
16495 the key length and effective key length are equal.
16499 * Add a bunch of functions that should simplify the creation of
16500 X509_NAME structures. Now you should be able to do:
16501 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
16502 and have it automatically work out the correct field type and fill in
16503 the structures. The more adventurous can try:
16504 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
16505 and it will (hopefully) work out the correct multibyte encoding.
16509 * Change the 'req' utility to use the new field handling and multibyte
16510 copy routines. Before the DN field creation was handled in an ad hoc
16511 way in req, ca, and x509 which was rather broken and didn't support
16512 BMPStrings or UTF8Strings. Since some software doesn't implement
16513 BMPStrings or UTF8Strings yet, they can be enabled using the config file
16514 using the dirstring_type option. See the new comment in the default
16515 openssl.cnf for more info.
16519 * Make crypto/rand/md_rand.c more robust:
16520 - Assure unique random numbers after fork().
16521 - Make sure that concurrent threads access the global counter and
16522 md serializably so that we never lose entropy in them
16523 or use exactly the same state in multiple threads.
16524 Access to the large state is not always serializable because
16525 the additional locking could be a performance killer, and
16526 md should be large enough anyway.
16530 * New file `apps/app_rand.c` with commonly needed functionality
16531 for handling the random seed file.
16533 Use the random seed file in some applications that previously did not:
16535 dsaparam -genkey (which also ignored its '-rand' option),
16538 x509 (when signing).
16539 Except on systems with /dev/urandom, it is crucial to have a random
16540 seed file at least for key creation, DSA signing, and for DH exchanges;
16541 for RSA signatures we could do without one.
16543 gendh and gendsa (unlike genrsa) used to read only the first byte
16544 of each file listed in the '-rand' option. The function as previously
16545 found in genrsa is now in app_rand.c and is used by all programs
16546 that support '-rand'.
16550 * In RAND_write_file, use mode 0600 for creating files;
16551 don't just chmod when it may be too late.
16555 * Report an error from X509_STORE_load_locations
16556 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
16560 * New function ASN1_mbstring_copy() this copies a string in either
16561 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
16562 into an ASN1_STRING type. A mask of permissible types is passed
16563 and it chooses the "minimal" type to use or an error if not type
16568 * Add function equivalents to the various macros in asn1.h. The old
16569 macros are retained with an `M_` prefix. Code inside the library can
16570 use the `M_` macros. External code (including the openssl utility)
16571 should *NOT* in order to be "shared library friendly".
16575 * Add various functions that can check a certificate's extensions
16576 to see if it usable for various purposes such as SSL client,
16577 server or S/MIME and CAs of these types. This is currently
16578 VERY EXPERIMENTAL but will ultimately be used for certificate chain
16579 verification. Also added a -purpose flag to x509 utility to
16580 print out all the purposes.
16584 * Add a CRYPTO_EX_DATA to X509 certificate structure and associated
16589 * New `X509V3_{X509,CRL,REVOKED}_get_d2i()` functions. These will search
16590 for, obtain and decode and extension and obtain its critical flag.
16591 This allows all the necessary extension code to be handled in a
16592 single function call.
16596 * RC4 tune-up featuring 30-40% performance improvement on most RISC
16597 platforms. See crypto/rc4/rc4_enc.c for further details.
16601 * New -noout option to asn1parse. This causes no output to be produced
16602 its main use is when combined with -strparse and -out to extract data
16603 from a file (which may not be in ASN.1 format).
16607 * Fix for pkcs12 program. It was hashing an invalid certificate pointer
16608 when producing the local key id.
16610 *Richard Levitte <levitte@stacken.kth.se>*
16612 * New option -dhparam in s_server. This allows a DH parameter file to be
16613 stated explicitly. If it is not stated then it tries the first server
16614 certificate file. The previous behaviour hard coded the filename
16619 * Add -pubin and -pubout options to the rsa and dsa commands. These allow
16620 a public key to be input or output. For example:
16621 openssl rsa -in key.pem -pubout -out pubkey.pem
16622 Also added necessary DSA public key functions to handle this.
16626 * Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
16627 in the message. This was handled by allowing
16628 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
16630 *Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>*
16632 * Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
16633 to the end of the strings whereas this didn't. This would cause problems
16634 if strings read with d2i_ASN1_bytes() were later modified.
16636 *Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>*
16638 * Fix for base64 decode bug. When a base64 bio reads only one line of
16639 data and it contains EOF it will end up returning an error. This is
16640 caused by input 46 bytes long. The cause is due to the way base64
16641 BIOs find the start of base64 encoded data. They do this by trying a
16642 trial decode on each line until they find one that works. When they
16643 do a flag is set and it starts again knowing it can pass all the
16644 data directly through the decoder. Unfortunately it doesn't reset
16645 the context it uses. This means that if EOF is reached an attempt
16646 is made to pass two EOFs through the context and this causes the
16647 resulting error. This can also cause other problems as well. As is
16648 usual with these problems it takes *ages* to find and the fix is
16649 trivial: move one line.
16651 *Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer)*
16653 * Ugly workaround to get s_client and s_server working under Windows. The
16654 old code wouldn't work because it needed to select() on sockets and the
16655 tty (for keypresses and to see if data could be written). Win32 only
16656 supports select() on sockets so we select() with a 1s timeout on the
16657 sockets and then see if any characters are waiting to be read, if none
16658 are present then we retry, we also assume we can always write data to
16659 the tty. This isn't nice because the code then blocks until we've
16660 received a complete line of data and it is effectively polling the
16661 keyboard at 1s intervals: however it's quite a bit better than not
16662 working at all :-) A dedicated Windows application might handle this
16663 with an event loop for example.
16667 * Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
16668 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
16669 will be called when RSA_sign() and RSA_verify() are used. This is useful
16670 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
16671 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
16672 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
16673 This necessitated the support of an extra signature type NID_md5_sha1
16674 for SSL signatures and modifications to the SSL library to use it instead
16675 of calling RSA_public_decrypt() and RSA_private_encrypt().
16679 * Add new -verify -CAfile and -CApath options to the crl program, these
16680 will lookup a CRL issuers certificate and verify the signature in a
16681 similar way to the verify program. Tidy up the crl program so it
16682 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
16683 less strict. It will now permit CRL extensions even if it is not
16684 a V2 CRL: this will allow it to tolerate some broken CRLs.
16688 * Initialize all non-automatic variables each time one of the openssl
16689 sub-programs is started (this is necessary as they may be started
16690 multiple times from the "OpenSSL>" prompt).
16692 *Lennart Bang, Bodo Moeller*
16694 * Preliminary compilation option RSA_NULL which disables RSA crypto without
16695 removing all other RSA functionality (this is what NO_RSA does). This
16696 is so (for example) those in the US can disable those operations covered
16697 by the RSA patent while allowing storage and parsing of RSA keys and RSA
16702 * Non-copying interface to BIO pairs.
16703 (still largely untested)
16707 * New function ASN1_tag2str() to convert an ASN1 tag to a descriptive
16708 ASCII string. This was handled independently in various places before.
16712 * New functions UTF8_getc() and UTF8_putc() that parse and generate
16713 UTF8 strings a character at a time.
16717 * Use client_version from client hello to select the protocol
16718 (s23_srvr.c) and for RSA client key exchange verification
16719 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
16723 * Add various utility functions to handle SPKACs, these were previously
16724 handled by poking round in the structure internals. Added new function
16725 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
16726 print, verify and generate SPKACs. Based on an original idea from
16727 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
16731 * RIPEMD160 is operational on all platforms and is back in 'make test'.
16735 * Allow the config file extension section to be overwritten on the
16736 command line. Based on an original idea from Massimiliano Pala
16737 <madwolf@comune.modena.it>. The new option is called -extensions
16738 and can be applied to ca, req and x509. Also -reqexts to override
16739 the request extensions in req and -crlexts to override the crl extensions
16744 * Add new feature to the SPKAC handling in ca. Now you can include
16745 the same field multiple times by preceding it by "XXXX." for example:
16748 this is the same syntax as used in the req config file.
16752 * Allow certificate extensions to be added to certificate requests. These
16753 are specified in a 'req_extensions' option of the req section of the
16754 config file. They can be printed out with the -text option to req but
16755 are otherwise ignored at present.
16759 * Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
16760 data read consists of only the final block it would not decrypted because
16761 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
16762 A misplaced 'break' also meant the decrypted final block might not be
16763 copied until the next read.
16767 * Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
16768 a few extra parameters to the DH structure: these will be useful if
16769 for example we want the value of 'q' or implement X9.42 DH.
16773 * Initial support for DSA_METHOD. This is based on the RSA_METHOD and
16774 provides hooks that allow the default DSA functions or functions on a
16775 "per key" basis to be replaced. This allows hardware acceleration and
16776 hardware key storage to be handled without major modification to the
16777 library. Also added low-level modexp hooks and CRYPTO_EX structure and
16778 associated functions.
16782 * Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
16783 as "read only": it can't be written to and the buffer it points to will
16784 not be freed. Reading from a read only BIO is much more efficient than
16785 a normal memory BIO. This was added because there are several times when
16786 an area of memory needs to be read from a BIO. The previous method was
16787 to create a memory BIO and write the data to it, this results in two
16788 copies of the data and an O(n^2) reading algorithm. There is a new
16789 function BIO_new_mem_buf() which creates a read only memory BIO from
16790 an area of memory. Also modified the PKCS#7 routines to use read only
16795 * Bugfix: ssl23_get_client_hello did not work properly when called in
16796 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
16797 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
16798 but a retry condition occurred while trying to read the rest.
16802 * The PKCS7_ENC_CONTENT_new() function was setting the content type as
16803 NID_pkcs7_encrypted by default: this was wrong since this should almost
16804 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
16805 the encrypted data type: this is a more sensible place to put it and it
16806 allows the PKCS#12 code to be tidied up that duplicated this
16811 * Changed obj_dat.pl script so it takes its input and output files on
16812 the command line. This should avoid shell escape redirection problems
16817 * Initial support for certificate extension requests, these are included
16818 in things like Xenroll certificate requests. Included functions to allow
16819 extensions to be obtained and added.
16823 * -crlf option to s_client and s_server for sending newlines as
16824 CRLF (as required by many protocols).
16828 ### Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
16830 * Install libRSAglue.a when OpenSSL is built with RSAref.
16832 *Ralf S. Engelschall*
16834 * A few more `#ifndef NO_FP_API / #endif` pairs for consistency.
16836 *Andrija Antonijevic <TheAntony2@bigfoot.com>*
16838 * Fix -startdate and -enddate (which was missing) arguments to 'ca'
16843 * New function DSA_dup_DH, which duplicates DSA parameters/keys as
16844 DH parameters/keys (q is lost during that conversion, but the resulting
16845 DH parameters contain its length).
16847 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
16848 much faster than DH_generate_parameters (which creates parameters
16849 where `p = 2*q + 1`), and also the smaller q makes DH computations
16850 much more efficient (160-bit exponentiation instead of 1024-bit
16851 exponentiation); so this provides a convenient way to support DHE
16852 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
16853 utter importance to use
16854 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
16856 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
16857 when such DH parameters are used, because otherwise small subgroup
16858 attacks may become possible!
16862 * Avoid memory leak in i2d_DHparams.
16866 * Allow the -k option to be used more than once in the enc program:
16867 this allows the same encrypted message to be read by multiple recipients.
16871 * New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
16872 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
16873 it will always use the numerical form of the OID, even if it has a short
16878 * Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
16879 method only got called if p,q,dmp1,dmq1,iqmp components were present,
16880 otherwise bn_mod_exp was called. In the case of hardware keys for example
16881 no private key components need be present and it might store extra data
16882 in the RSA structure, which cannot be accessed from bn_mod_exp.
16883 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
16884 private key operations.
16888 * Added support for SPARC Linux.
16892 * pem_password_cb function type incompatibly changed from
16893 typedef int pem_password_cb(char *buf, int size, int rwflag);
16895 ....(char *buf, int size, int rwflag, void *userdata);
16896 so that applications can pass data to their callbacks:
16897 The `PEM[_ASN1]_{read,write}...` functions and macros now take an
16898 additional void * argument, which is just handed through whenever
16899 the password callback is called.
16901 *Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller*
16903 New function SSL_CTX_set_default_passwd_cb_userdata.
16905 Compatibility note: As many C implementations push function arguments
16906 onto the stack in reverse order, the new library version is likely to
16907 interoperate with programs that have been compiled with the old
16908 pem_password_cb definition (PEM_whatever takes some data that
16909 happens to be on the stack as its last argument, and the callback
16910 just ignores this garbage); but there is no guarantee whatsoever that
16913 * The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
16914 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
16915 problems not only on Windows, but also on some Unix platforms.
16916 To avoid problematic command lines, these definitions are now in an
16917 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
16918 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
16922 * MIPS III/IV assembler module is reimplemented.
16926 * More DES library cleanups: remove references to srand/rand and
16927 delete an unused file.
16931 * Add support for the free Netwide assembler (NASM) under Win32,
16932 since not many people have MASM (ml) and it can be hard to obtain.
16933 This is currently experimental but it seems to work OK and pass all
16934 the tests. Check out INSTALL.W32 for info.
16938 * Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
16939 without temporary keys kept an extra copy of the server key,
16940 and connections with temporary keys did not free everything in case
16945 * New function RSA_check_key and new openssl rsa option -check
16946 for verifying the consistency of RSA keys.
16948 *Ulf Moeller, Bodo Moeller*
16950 * Various changes to make Win32 compile work:
16951 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
16952 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
16953 comparison" warnings.
16954 3. Add `sk_<TYPE>_sort` to DEF file generator and do make update.
16958 * Add a debugging option to PKCS#5 v2 key generation function: when
16959 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
16960 derived keys are printed to stderr.
16964 * Copy the flags in ASN1_STRING_dup().
16966 *Roman E. Pavlov <pre@mo.msk.ru>*
16968 * The x509 application mishandled signing requests containing DSA
16969 keys when the signing key was also DSA and the parameters didn't match.
16971 It was supposed to omit the parameters when they matched the signing key:
16972 the verifying software was then supposed to automatically use the CA's
16973 parameters if they were absent from the end user certificate.
16975 Omitting parameters is no longer recommended. The test was also
16976 the wrong way round! This was probably due to unusual behaviour in
16977 EVP_cmp_parameters() which returns 1 if the parameters match.
16978 This meant that parameters were omitted when they *didn't* match and
16979 the certificate was useless. Certificates signed with 'ca' didn't have
16982 *Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>*
16984 * Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
16985 The interface is as follows:
16986 Applications can use
16987 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
16988 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
16989 "off" is now the default.
16990 The library internally uses
16991 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
16992 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
16993 to disable memory-checking temporarily.
16995 Some inconsistent states that previously were possible (and were
16996 even the default) are now avoided.
16998 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
16999 with each memory chunk allocated; this is occasionally more helpful
17000 than just having a counter.
17002 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
17004 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
17009 * Introduce "mode" for SSL structures (with defaults in SSL_CTX),
17010 which largely parallels "options", but is for changing API behaviour,
17011 whereas "options" are about protocol behaviour.
17012 Initial "mode" flags are:
17014 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
17015 a single record has been written.
17016 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
17017 retries use the same buffer location.
17018 (But all of the contents must be
17023 * Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
17026 * Fix problems with no-hmac etc.
17028 *Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>*
17030 * New functions RSA_get_default_method(), RSA_set_method() and
17031 RSA_get_method(). These allows replacement of RSA_METHODs without having
17032 to mess around with the internals of an RSA structure.
17036 * Fix memory leaks in DSA_do_sign and DSA_is_prime.
17037 Also really enable memory leak checks in openssl.c and in some
17040 *Chad C. Mulligan, Bodo Moeller*
17042 * Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
17043 up the length of negative integers. This has now been simplified to just
17044 store the length when it is first determined and use it later, rather
17045 than trying to keep track of where data is copied and updating it to
17047 *Steve Henson, reported by Brien Wheeler <bwheeler@authentica-security.com>*
17049 * Add a new function PKCS7_signatureVerify. This allows the verification
17050 of a PKCS#7 signature but with the signing certificate passed to the
17051 function itself. This contrasts with PKCS7_dataVerify which assumes the
17052 certificate is present in the PKCS#7 structure. This isn't always the
17053 case: certificates can be omitted from a PKCS#7 structure and be
17054 distributed by "out of band" means (such as a certificate database).
17058 * Complete the `PEM_*` macros with DECLARE_PEM versions to replace the
17059 function prototypes in pem.h, also change util/mkdef.pl to add the
17060 necessary function names.
17064 * mk1mf.pl (used by Windows builds) did not properly read the
17065 options set by Configure in the top level Makefile, and Configure
17066 was not even able to write more than one option correctly.
17067 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
17071 * New functions CONF_load_bio() and CONF_load_fp() to allow a config
17072 file to be loaded from a BIO or FILE pointer. The BIO version will
17073 for example allow memory BIOs to contain config info.
17077 * New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
17078 Whoever hopes to achieve shared-library compatibility across versions
17079 must use this, not the compile-time macro.
17080 (Exercise 0.9.4: Which is the minimum library version required by
17082 Note: All this applies only to multi-threaded programs, others don't
17087 * Add missing case to s3_clnt.c state machine -- one of the new SSL tests
17088 through a BIO pair triggered the default case, i.e.
17089 SSLerr(...,SSL_R_UNKNOWN_STATE).
17093 * New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
17094 can use the SSL library even if none of the specific BIOs is
17099 * Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
17100 for the encoded length.
17102 *Jeon KyoungHo <khjeon@sds.samsung.co.kr>*
17104 * Add initial documentation of the X509V3 functions.
17108 * Add a new pair of functions PEM_write_PKCS8PrivateKey() and
17109 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
17110 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
17111 secure PKCS#8 private key format with a high iteration count.
17115 * Fix determination of Perl interpreter: A perl or perl5
17116 *directory* in $PATH was also accepted as the interpreter.
17118 *Ralf S. Engelschall*
17120 * Fix demos/sign/sign.c: well there wasn't anything strictly speaking
17121 wrong with it but it was very old and did things like calling
17122 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
17123 unusual formatting.
17127 * Fix demos/selfsign.c: it used obsolete and deleted functions, changed
17128 to use the new extension code.
17132 * Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
17133 with macros. This should make it easier to change their form, add extra
17134 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
17139 * Add to configuration table a new entry that can specify an alternative
17140 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
17141 according to Mark Crispin <MRC@Panda.COM>.
17145 * DES CBC did not update the IV. Weird.
17149 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
17150 Changing the behaviour of the former might break existing programs --
17151 where IV updating is needed, des_ncbc_encrypt can be used.
17154 * When bntest is run from "make test" it drives bc to check its
17155 calculations, as well as internally checking them. If an internal check
17156 fails, it needs to cause bc to give a non-zero result or make test carries
17157 on without noticing the failure. Fixed.
17161 * DES library cleanups.
17165 * Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
17166 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
17167 ciphers. NOTE: although the key derivation function has been verified
17168 against some published test vectors it has not been extensively tested
17169 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
17174 * Instead of "mkdir -p", which is not fully portable, use new
17175 Perl script "util/mkdir-p.pl".
17179 * Rewrite the way password based encryption (PBE) is handled. It used to
17180 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
17181 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
17182 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
17183 the 'parameter' field of the AlgorithmIdentifier is passed to the
17184 underlying key generation function so it must do its own ASN1 parsing.
17185 This has also changed the EVP_PBE_CipherInit() function which now has a
17186 'parameter' argument instead of literal salt and iteration count values
17187 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
17191 * Support for PKCS#5 v1.5 compatible password based encryption algorithms
17192 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
17193 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
17194 KEY" because this clashed with PKCS#8 unencrypted string. Since this
17195 value was just used as a "magic string" and not used directly its
17196 value doesn't matter.
17200 * Introduce some semblance of const correctness to BN. Shame C doesn't
17205 * "linux-sparc64" configuration (ultrapenguin).
17207 *Ray Miller <ray.miller@oucs.ox.ac.uk>*
17208 "linux-sparc" configuration.
17210 *Christian Forster <fo@hawo.stw.uni-erlangen.de>*
17212 * config now generates no-xxx options for missing ciphers.
17216 * Support the EBCDIC character set (work in progress).
17217 File ebcdic.c not yet included because it has a different license.
17219 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
17221 * Support BS2000/OSD-POSIX.
17223 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
17225 * Make callbacks for key generation use `void *` instead of `char *`.
17229 * Make S/MIME samples compile (not yet tested).
17233 * Additional typesafe stacks.
17237 * New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
17241 ### Changes between 0.9.3 and 0.9.3a [29 May 1999]
17243 * New configuration variant "sco5-gcc".
17245 * Updated some demos.
17247 *Sean O Riordain, Wade Scholine*
17249 * Add missing BIO_free at exit of pkcs12 application.
17253 * Fix memory leak in conf.c.
17257 * Updates for Win32 to assembler version of MD5.
17261 * Set #! path to perl in `apps/der_chop` to where we found it
17262 instead of using a fixed path.
17266 * SHA library changes for irix64-mips4-cc.
17270 * Improvements for VMS support.
17274 ### Changes between 0.9.2b and 0.9.3 [24 May 1999]
17276 * Bignum library bug fix. IRIX 6 passes "make test" now!
17277 This also avoids the problems with SC4.2 and unpatched SC5.
17279 *Andy Polyakov <appro@fy.chalmers.se>*
17281 * New functions sk_num, sk_value and sk_set to replace the previous macros.
17282 These are required because of the typesafe stack would otherwise break
17283 existing code. If old code used a structure member which used to be STACK
17284 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
17285 sk_num or sk_value it would produce an error because the num, data members
17286 are not present in STACK_OF. Now it just produces a warning. sk_set
17287 replaces the old method of assigning a value to sk_value
17288 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
17289 that does this will no longer work (and should use sk_set instead) but
17290 this could be regarded as a "questionable" behaviour anyway.
17294 * Fix most of the other PKCS#7 bugs. The "experimental" code can now
17295 correctly handle encrypted S/MIME data.
17299 * Change type of various DES function arguments from des_cblock
17300 (which means, in function argument declarations, pointer to char)
17301 to des_cblock * (meaning pointer to array with 8 char elements),
17302 which allows the compiler to do more typechecking; it was like
17303 that back in SSLeay, but with lots of ugly casts.
17305 Introduce new type const_des_cblock.
17309 * Reorganise the PKCS#7 library and get rid of some of the more obvious
17310 problems: find RecipientInfo structure that matches recipient certificate
17311 and initialise the ASN1 structures properly based on passed cipher.
17315 * Belatedly make the BN tests actually check the results.
17319 * Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
17320 to and from BNs: it was completely broken. New compilation option
17321 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
17322 key elements as negative integers.
17326 * Reorganize and speed up MD5.
17328 *Andy Polyakov <appro@fy.chalmers.se>*
17332 *Richard Levitte <richard@levitte.org>*
17334 * New option -out to asn1parse to allow the parsed structure to be
17335 output to a file. This is most useful when combined with the -strparse
17336 option to examine the output of things like OCTET STRINGS.
17340 * Make SSL library a little more fool-proof by not requiring any longer
17341 that `SSL_set_{accept,connect}_state` be called before
17342 `SSL_{accept,connect}` may be used (`SSL_set_..._state` is omitted
17343 in many applications because usually everything *appeared* to work as
17344 intended anyway -- now it really works as intended).
17348 * Move openssl.cnf out of lib/.
17352 * Fix various things to let OpenSSL even pass "egcc -pipe -O2 -Wall
17353 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
17354 -Wmissing-declarations -Wnested-externs -Winline" with EGCS 1.1.2+
17356 *Ralf S. Engelschall*
17358 * Various fixes to the EVP and PKCS#7 code. It may now be able to
17359 handle PKCS#7 enveloped data properly.
17361 *Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve*
17363 * Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
17364 copying pointers. The cert_st handling is changed by this in
17365 various ways (and thus what used to be known as ctx->default_cert
17366 is now called ctx->cert, since we don't resort to `s->ctx->[default_]cert`
17367 any longer when s->cert does not give us what we need).
17368 ssl_cert_instantiate becomes obsolete by this change.
17369 As soon as we've got the new code right (possibly it already is?),
17370 we have solved a couple of bugs of the earlier code where s->cert
17371 was used as if it could not have been shared with other SSL structures.
17373 Note that using the SSL API in certain dirty ways now will result
17374 in different behaviour than observed with earlier library versions:
17375 Changing settings for an `SSL_CTX *ctx` after having done s = SSL_new(ctx)
17376 does not influence s as it used to.
17378 In order to clean up things more thoroughly, inside SSL_SESSION
17379 we don't use CERT any longer, but a new structure SESS_CERT
17380 that holds per-session data (if available); currently, this is
17381 the peer's certificate chain and, for clients, the server's certificate
17382 and temporary key. CERT holds only those values that can have
17383 meaningful defaults in an SSL_CTX.
17387 * New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
17388 from the internal representation. Various PKCS#7 fixes: remove some
17389 evil casts and set the enc_dig_alg field properly based on the signing
17394 * Allow PKCS#12 password to be set from the command line or the
17395 environment. Let 'ca' get its config file name from the environment
17396 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
17401 * Allow certificate policies extension to use an IA5STRING for the
17402 organization field. This is contrary to the PKIX definition but
17403 VeriSign uses it and IE5 only recognises this form. Document 'x509'
17408 * Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
17409 without disallowing inline assembler and the like for non-pedantic builds.
17413 * Support Borland C++ builder.
17415 *Janez Jere <jj@void.si>, modified by Ulf Möller*
17421 * SHA-1 cleanups and performance enhancements.
17423 *Andy Polyakov <appro@fy.chalmers.se>*
17425 * Sparc v8plus assembler for the bignum library.
17427 *Andy Polyakov <appro@fy.chalmers.se>*
17429 * Accept any -xxx and +xxx compiler options in Configure.
17433 * Update HPUX configuration.
17437 * Add missing `sk_<type>_unshift()` function to safestack.h
17439 *Ralf S. Engelschall*
17441 * New function SSL_CTX_use_certificate_chain_file that sets the
17442 "extra_cert"s in addition to the certificate. (This makes sense
17443 only for "PEM" format files, as chains as a whole are not
17448 * Support verify_depth from the SSL API.
17449 x509_vfy.c had what can be considered an off-by-one-error:
17450 Its depth (which was not part of the external interface)
17451 was actually counting the number of certificates in a chain;
17452 now it really counts the depth.
17456 * Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
17457 instead of X509err, which often resulted in confusing error
17458 messages since the error codes are not globally unique
17459 (e.g. an alleged error in ssl3_accept when a certificate
17460 didn't match the private key).
17462 * New function SSL_CTX_set_session_id_context that allows to set a default
17463 value (so that you don't need SSL_set_session_id_context for each
17464 connection using the SSL_CTX).
17468 * OAEP decoding bug fix.
17472 * Support INSTALL_PREFIX for package builders, as proposed by
17477 * New Configure options "threads" and "no-threads". For systems
17478 where the proper compiler options are known (currently Solaris
17479 and Linux), "threads" is the default.
17483 * New script util/mklink.pl as a faster substitute for util/mklink.sh.
17487 * Install various scripts to $(OPENSSLDIR)/misc, not to
17488 $(INSTALLTOP)/bin -- they shouldn't clutter directories
17489 such as /usr/local/bin.
17493 * "make linux-shared" to build shared libraries.
17495 *Niels Poppe <niels@netbox.org>*
17497 * New Configure option `no-<cipher>` (rsa, idea, rc5, ...).
17501 * Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
17502 extension adding in x509 utility.
17506 * Remove NOPROTO sections and error code comments.
17510 * Partial rewrite of the DEF file generator to now parse the ANSI
17515 * New Configure options --prefix=DIR and --openssldir=DIR.
17519 * Complete rewrite of the error code script(s). It is all now handled
17520 by one script at the top level which handles error code gathering,
17521 header rewriting and C source file generation. It should be much better
17522 than the old method: it now uses a modified version of Ulf's parser to
17523 read the ANSI prototypes in all header files (thus the old K&R definitions
17524 aren't needed for error creation any more) and do a better job of
17525 translating function codes into names. The old 'ASN1 error code embedded
17526 in a comment' is no longer necessary and it doesn't use .err files which
17527 have now been deleted. Also the error code call doesn't have to appear all
17528 on one line (which resulted in some large lines...).
17532 * Change #include filenames from `<foo.h>` to `<openssl/foo.h>`.
17536 * Change behaviour of ssl2_read when facing length-0 packets: Don't return
17537 0 (which usually indicates a closed connection), but continue reading.
17541 * Fix some race conditions.
17545 * Add support for CRL distribution points extension. Add Certificate
17546 Policies and CRL distribution points documentation.
17550 * Move the autogenerated header file parts to crypto/opensslconf.h.
17554 * Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
17555 8 of keying material. Merlin has also confirmed interop with this fix
17556 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
17558 *Merlin Hughes <merlin@baltimore.ie>*
17560 * Fix lots of warnings.
17562 *Richard Levitte <levitte@stacken.kth.se>*
17564 * In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
17565 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
17567 *Richard Levitte <levitte@stacken.kth.se>*
17569 * Fix problems with sizeof(long) == 8.
17571 *Andy Polyakov <appro@fy.chalmers.se>*
17573 * Change functions to ANSI C.
17577 * Fix typos in error codes.
17579 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller*
17581 * Remove defunct assembler files from Configure.
17585 * SPARC v8 assembler BIGNUM implementation.
17587 *Andy Polyakov <appro@fy.chalmers.se>*
17589 * Support for Certificate Policies extension: both print and set.
17590 Various additions to support the r2i method this uses.
17594 * A lot of constification, and fix a bug in X509_NAME_oneline() that could
17595 return a const string when you are expecting an allocated buffer.
17599 * Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
17600 types DirectoryString and DisplayText.
17604 * Add code to allow r2i extensions to access the configuration database,
17605 add an LHASH database driver and add several ctx helper functions.
17609 * Fix an evil bug in bn_expand2() which caused various BN functions to
17610 fail when they extended the size of a BIGNUM.
17614 * Various utility functions to handle SXNet extension. Modify mkdef.pl to
17615 support typesafe stack.
17619 * Fix typo in SSL_[gs]et_options().
17621 *Nils Frostberg <nils@medcom.se>*
17623 * Delete various functions and files that belonged to the (now obsolete)
17624 old X509V3 handling code.
17628 * New Configure option "rsaref".
17632 * Don't auto-generate pem.h.
17636 * Introduce type-safe ASN.1 SETs.
17640 * Convert various additional casted stacks to type-safe STACK_OF() variants.
17642 *Ben Laurie, Ralf S. Engelschall, Steve Henson*
17644 * Introduce type-safe STACKs. This will almost certainly break lots of code
17645 that links with OpenSSL (well at least cause lots of warnings), but fear
17646 not: the conversion is trivial, and it eliminates loads of evil casts. A
17647 few STACKed things have been converted already. Feel free to convert more.
17648 In the fullness of time, I'll do away with the STACK type altogether.
17652 * Add `openssl ca -revoke <certfile>` facility which revokes a certificate
17653 specified in `<certfile>` by updating the entry in the index.txt file.
17654 This way one no longer has to edit the index.txt file manually for
17655 revoking a certificate. The -revoke option does the gory details now.
17657 *Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall*
17659 * Fix `openssl crl -noout -text` combination where `-noout` killed the
17660 `-text` option at all and this way the `-noout -text` combination was
17661 inconsistent in `openssl crl` with the friends in `openssl x509|rsa|dsa`.
17663 *Ralf S. Engelschall*
17665 * Make sure a corresponding plain text error message exists for the
17666 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
17667 verify callback function determined that a certificate was revoked.
17669 *Ralf S. Engelschall*
17671 * Bugfix: In test/testenc, don't test `openssl <cipher>` for
17672 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
17673 all available ciphers including rc5, which was forgotten until now.
17674 In order to let the testing shell script know which algorithms
17675 are available, a new (up to now undocumented) command
17676 `openssl list-cipher-commands` is used.
17680 * Bugfix: s_client occasionally would sleep in select() when
17681 it should have checked SSL_pending() first.
17685 * New functions DSA_do_sign and DSA_do_verify to provide access to
17686 the raw DSA values prior to ASN.1 encoding.
17690 * Tweaks to Configure
17692 *Niels Poppe <niels@netbox.org>*
17694 * Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
17699 * New variables $(RANLIB) and $(PERL) in the Makefiles.
17703 * New config option to avoid instructions that are illegal on the 80386.
17704 The default code is faster, but requires at least a 486.
17708 * Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
17709 SSL2_SERVER_VERSION (not used at all) macros, which are now the
17710 same as SSL2_VERSION anyway.
17714 * New "-showcerts" option for s_client.
17718 * Still more PKCS#12 integration. Add pkcs12 application to openssl
17719 application. Various cleanups and fixes.
17723 * More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
17724 modify error routines to work internally. Add error codes and PBE init
17725 to library startup routines.
17729 * Further PKCS#12 integration. Added password based encryption, PKCS#8 and
17730 packing functions to asn1 and evp. Changed function names and error
17731 codes along the way.
17735 * PKCS12 integration: and so it begins... First of several patches to
17736 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
17737 objects to objects.h
17741 * Add a new 'indent' option to some X509V3 extension code. Initial ASN1
17742 and display support for Thawte strong extranet extension.
17746 * Add LinuxPPC support.
17748 *Jeff Dubrule <igor@pobox.org>*
17750 * Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
17751 bn_div_words in alpha.s.
17753 *Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie*
17755 * Make sure the RSA OAEP test is skipped under -DRSAref because
17756 OAEP isn't supported when OpenSSL is built with RSAref.
17758 *Ulf Moeller <ulf@fitug.de>*
17760 * Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
17761 so they no longer are missing under -DNOPROTO.
17763 *Soren S. Jorvang <soren@t.dk>*
17765 ### Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
17767 * Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
17768 doesn't work when the session is reused. Coming soon!
17772 * Fix a security hole, that allows sessions to be reused in the wrong
17773 context thus bypassing client cert protection! All software that uses
17774 client certs and session caches in multiple contexts NEEDS PATCHING to
17775 allow session reuse! A fuller solution is in the works.
17777 *Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)*
17779 * Some more source tree cleanups (removed obsolete files
17780 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
17781 permission on "config" script to be executable) and a fix for the INSTALL
17784 *Ulf Moeller <ulf@fitug.de>*
17786 * Remove some legacy and erroneous uses of malloc, free instead of
17789 *Lennart Bang <lob@netstream.se>, with minor changes by Steve*
17791 * Make rsa_oaep_test return non-zero on error.
17793 *Ulf Moeller <ulf@fitug.de>*
17795 * Add support for native Solaris shared libraries. Configure
17796 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
17797 if someone would make that last step automatic.
17799 *Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>*
17801 * ctx_size was not built with the right compiler during "make links". Fixed.
17805 * Change the meaning of 'ALL' in the cipher list. It now means "everything
17806 except NULL ciphers". This means the default cipher list will no longer
17807 enable NULL ciphers. They need to be specifically enabled e.g. with
17808 the string "DEFAULT:eNULL".
17812 * Fix to RSA private encryption routines: if p < q then it would
17813 occasionally produce an invalid result. This will only happen with
17814 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
17818 * Be less restrictive and allow also `perl util/perlpath.pl
17819 /path/to/bin/perl` in addition to `perl util/perlpath.pl /path/to/bin`,
17820 because this way one can also use an interpreter named `perl5` (which is
17821 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
17822 installed as `perl`).
17824 *Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
17826 * Let util/clean-depend.pl work also with older Perl 5.00x versions.
17828 *Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
17830 * Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
17831 advapi32.lib to Win32 build and change the pem test comparison
17832 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
17833 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
17834 and crypto/des/ede_cbcm_enc.c.
17838 * DES quad checksum was broken on big-endian architectures. Fixed.
17842 * Comment out two functions in bio.h that aren't implemented. Fix up the
17843 Win32 test batch file so it (might) work again. The Win32 test batch file
17844 is horrible: I feel ill....
17848 * Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
17849 in e_os.h. Audit of header files to check ANSI and non ANSI
17850 sections: 10 functions were absent from non ANSI section and not exported
17851 from Windows DLLs. Fixed up libeay.num for new functions.
17855 * Make `openssl version` output lines consistent.
17857 *Ralf S. Engelschall*
17859 * Fix Win32 symbol export lists for BIO functions: Added
17860 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
17861 to ms/libeay{16,32}.def.
17863 *Ralf S. Engelschall*
17865 * Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
17866 fine under Unix and passes some trivial tests I've now added. But the
17867 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
17868 added to make sure no one expects that this stuff really works in the
17869 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
17870 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
17873 *Ralf S. Engelschall*
17875 * Fix the generation of two part addresses in perl.
17877 *Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie*
17879 * Add config entry for Linux on MIPS.
17881 *John Tobey <jtobey@channel1.com>*
17883 * Make links whenever Configure is run, unless we are on Windoze.
17887 * Permit extensions to be added to CRLs using crl_section in openssl.cnf.
17888 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
17893 * Add a useful kludge to allow package maintainers to specify compiler and
17894 other platforms details on the command line without having to patch the
17895 Configure script every time: One now can use
17896 `perl Configure <id>:<details>`,
17897 i.e. platform ids are allowed to have details appended
17898 to them (separated by colons). This is treated as there would be a static
17899 pre-configured entry in Configure's %table under key `<id>` with value
17900 `<details>` and `perl Configure <id>` is called. So, when you want to
17901 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
17902 assembler stuff you can use `perl Configure "FreeBSD-elf:pgcc:-O6:::"`
17903 now, which overrides the FreeBSD-elf entry on-the-fly.
17905 *Ralf S. Engelschall*
17907 * Disable new TLS1 ciphersuites by default: they aren't official yet.
17911 * Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
17912 on the `perl Configure ...` command line. This way one can compile
17913 OpenSSL libraries with Position Independent Code (PIC) which is needed
17914 for linking it into DSOs.
17916 *Ralf S. Engelschall*
17918 * Remarkably, export ciphers were totally broken and no-one had noticed!
17923 * Cleaned up the LICENSE document: The official contact for any license
17924 questions now is the OpenSSL core team under openssl-core@openssl.org.
17925 And add a paragraph about the dual-license situation to make sure people
17926 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
17927 to the OpenSSL toolkit.
17929 *Ralf S. Engelschall*
17931 * General source tree makefile cleanups: Made `making xxx in yyy...`
17932 display consistent in the source tree and replaced `/bin/rm` by `rm`.
17933 Additionally cleaned up the `make links` target: Remove unnecessary
17934 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
17935 to speed processing and no longer clutter the display with confusing
17936 stuff. Instead only the actually done links are displayed.
17938 *Ralf S. Engelschall*
17940 * Permit null encryption ciphersuites, used for authentication only. It used
17941 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
17942 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
17947 * Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
17948 signed attributes when verifying signatures (this would break them),
17949 the detached data encoding was wrong and public keys obtained using
17950 X509_get_pubkey() weren't freed.
17954 * Add text documentation for the BUFFER functions. Also added a work around
17955 to a Win95 console bug. This was triggered by the password read stuff: the
17956 last character typed gets carried over to the next fread(). If you were
17957 generating a new cert request using 'req' for example then the last
17958 character of the passphrase would be CR which would then enter the first
17963 * Added the new 'Includes OpenSSL Cryptography Software' button as
17964 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
17965 button and can be used by applications based on OpenSSL to show the
17966 relationship to the OpenSSL project.
17968 *Ralf S. Engelschall*
17970 * Remove confusing variables in function signatures in files
17971 ssl/ssl_lib.c and ssl/ssl.h.
17973 *Lennart Bong <lob@kulthea.stacken.kth.se>*
17975 * Don't install bss_file.c under PREFIX/include/
17977 *Lennart Bong <lob@kulthea.stacken.kth.se>*
17979 * Get the Win32 compile working again. Modify mkdef.pl so it can handle
17980 functions that return function pointers and has support for NT specific
17981 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
17982 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
17983 unsigned to signed types: this was killing the Win32 compile.
17987 * Add new certificate file to stack functions,
17988 SSL_add_dir_cert_subjects_to_stack() and
17989 SSL_add_file_cert_subjects_to_stack(). These largely supplant
17990 SSL_load_client_CA_file(), and can be used to add multiple certs easily
17991 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
17992 This means that Apache-SSL and similar packages don't have to mess around
17993 to add as many CAs as they want to the preferred list.
17997 * Experiment with doxygen documentation. Currently only partially applied to
17999 See <http://www.stack.nl/~dimitri/doxygen/index.html>, and run doxygen with
18000 openssl.doxy as the configuration file.
18004 * Get rid of remaining C++-style comments which strict C compilers hate.
18006 *Ralf S. Engelschall, pointed out by Carlos Amengual*
18008 * Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
18009 compiled in by default: it has problems with large keys.
18013 * Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
18014 DH private keys and/or callback functions which directly correspond to
18015 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
18016 is needed for applications which have to configure certificates on a
18017 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
18019 For the RSA certificate situation is makes no difference, but
18020 for the DSA certificate situation this fixes the "no shared cipher"
18021 problem where the OpenSSL cipher selection procedure failed because the
18022 temporary keys were not overtaken from the context and the API provided
18023 no way to reconfigure them.
18024 The new functions now let applications reconfigure the stuff and they
18025 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
18026 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
18027 non-public-API function ssl_cert_instantiate() is used as a helper
18028 function and also to reduce code redundancy inside ssl_rsa.c.
18030 *Ralf S. Engelschall*
18032 * Move s_server -dcert and -dkey options out of the undocumented feature
18033 area because they are useful for the DSA situation and should be
18034 recognized by the users.
18036 *Ralf S. Engelschall*
18038 * Fix the cipher decision scheme for export ciphers: the export bits are
18039 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
18040 SSL_EXP_MASK. So, the original variable has to be used instead of the
18041 already masked variable.
18043 *Richard Levitte <levitte@stacken.kth.se>*
18045 * Fix `port` variable from `int` to `unsigned int` in crypto/bio/b_sock.c
18047 *Richard Levitte <levitte@stacken.kth.se>*
18049 * Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
18050 from `int` to `unsigned int` because it is a length and initialized by
18051 EVP_DigestFinal() which expects an `unsigned int *`.
18053 *Richard Levitte <levitte@stacken.kth.se>*
18055 * Don't hard-code path to Perl interpreter on shebang line of Configure
18056 script. Instead use the usual Shell->Perl transition trick.
18058 *Ralf S. Engelschall*
18060 * Make `openssl x509 -noout -modulus`' functional also for DSA certificates
18061 (in addition to RSA certificates) to match the behaviour of `openssl dsa
18062 -noout -modulus` as it's already the case for `openssl rsa -noout
18063 -modulus`. For RSA the -modulus is the real "modulus" while for DSA
18064 currently the public key is printed (a decision which was already done by
18065 `openssl dsa -modulus` in the past) which serves a similar purpose.
18066 Additionally the NO_RSA no longer completely removes the whole -modulus
18067 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
18070 *Ralf S. Engelschall*
18072 * Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
18073 BIO. See the source (crypto/evp/bio_ok.c) for more info.
18075 *Arne Ansper <arne@ats.cyber.ee>*
18077 * Dump the old yucky req code that tried (and failed) to allow raw OIDs
18078 to be added. Now both 'req' and 'ca' can use new objects defined in the
18083 * Add cool BIO that does syslog (or event log on NT).
18085 *Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie*
18087 * Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
18088 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
18089 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
18090 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
18094 * Add preliminary config info for new extension code.
18098 * Make RSA_NO_PADDING really use no padding.
18100 *Ulf Moeller <ulf@fitug.de>*
18102 * Generate errors when private/public key check is done.
18106 * Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
18107 for some CRL extensions and new objects added.
18111 * Really fix the ASN1 IMPLICIT bug this time... Partial support for private
18112 key usage extension and fuller support for authority key id.
18116 * Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
18117 padding method for RSA, which is recommended for new applications in PKCS
18118 #1 v2.0 (RFC 2437, October 1998).
18119 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
18120 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
18121 against Bleichbacher's attack on RSA.
18122 *Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
18125 * Updates to the new SSL compression code
18127 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
18129 * Fix so that the version number in the master secret, when passed
18130 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
18131 (because the server will not accept higher), that the version number
18132 is 0x03,0x01, not 0x03,0x00
18134 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
18136 * Run extensive memory leak checks on SSL commands. Fixed *lots* of memory
18137 leaks in `ssl/` relating to new `X509_get_pubkey()` behaviour. Also fixes
18138 in `apps/` and an unrelated leak in `crypto/dsa/dsa_vrf.c`.
18142 * Support for RAW extensions where an arbitrary extension can be
18143 created by including its DER encoding. See `apps/openssl.cnf` for
18148 * Make sure latest Perl versions don't interpret some generated C array
18149 code as Perl array code in the crypto/err/err_genc.pl script.
18151 *Lars Weber <3weber@informatik.uni-hamburg.de>*
18153 * Modify ms/do_ms.bat to not generate assembly language makefiles since
18154 not many people have the assembler. Various Win32 compilation fixes and
18155 update to the INSTALL.W32 file with (hopefully) more accurate Win32
18156 build instructions.
18160 * Modify configure script 'Configure' to automatically create crypto/date.h
18161 file under Win32 and also build pem.h from pem.org. New script
18162 util/mkfiles.pl to create the MINFO file on environments that can't do a
18163 'make files': perl util/mkfiles.pl >MINFO should work.
18167 * Major rework of DES function declarations, in the pursuit of correctness
18168 and purity. As a result, many evil casts evaporated, and some weirdness,
18169 too. You may find this causes warnings in your code. Zapping your evil
18170 casts will probably fix them. Mostly.
18174 * Fix for a typo in asn1.h. Bug fix to object creation script
18175 obj_dat.pl. It considered a zero in an object definition to mean
18176 "end of object": none of the objects in objects.h have any zeros
18177 so it wasn't spotted.
18179 *Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>*
18181 * Add support for Triple DES Cipher Block Chaining with Output Feedback
18182 Masking (CBCM). In the absence of test vectors, the best I have been able
18183 to do is check that the decrypt undoes the encrypt, so far. Send me test
18184 vectors if you have them.
18188 * Correct calculation of key length for export ciphers (too much space was
18189 allocated for null ciphers). This has not been tested!
18193 * Modifications to the mkdef.pl for Win32 DEF file creation. The usage
18194 message is now correct (it understands "crypto" and "ssl" on its
18195 command line). There is also now an "update" option. This will update
18196 the util/ssleay.num and util/libeay.num files with any new functions.
18198 perl util/mkdef.pl crypto ssl update
18199 it will update them.
18203 * Overhauled the Perl interface:
18204 - ported BN stuff to OpenSSL's different BN library
18205 - made the perl/ source tree CVS-aware
18206 - renamed the package from SSLeay to OpenSSL (the files still contain
18207 their history because I've copied them in the repository)
18208 - removed obsolete files (the test scripts will be replaced
18209 by better Test::Harness variants in the future)
18211 *Ralf S. Engelschall*
18213 * First cut for a very conservative source tree cleanup:
18214 1. merge various obsolete readme texts into doc/ssleay.txt
18215 where we collect the old documents and readme texts.
18216 2. remove the first part of files where I'm already sure that we no
18217 longer need them because of three reasons: either they are just temporary
18218 files which were left by Eric or they are preserved original files where
18219 I've verified that the diff is also available in the CVS via "cvs diff
18220 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
18221 the crypto/md/ stuff).
18223 *Ralf S. Engelschall*
18225 * More extension code. Incomplete support for subject and issuer alt
18226 name, issuer and authority key id. Change the i2v function parameters
18227 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
18228 what that's for :-) Fix to ASN1 macro which messed up
18229 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
18233 * Preliminary support for ENUMERATED type. This is largely copied from the
18238 * Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
18240 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
18242 * Make sure `make rehash` target really finds the `openssl` program.
18244 *Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
18246 * Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
18247 like to hear about it if this slows down other processors.
18251 * Add CygWin32 platform information to Configure script.
18253 *Alan Batie <batie@aahz.jf.intel.com>*
18255 * Fixed ms/32all.bat script: `no_asm` -> `no-asm`
18257 *Rainer W. Gerling <gerling@mpg-gv.mpg.de>*
18259 * New program nseq to manipulate netscape certificate sequences
18263 * Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
18268 * Fixes to BN code. Previously the default was to define BN_RECURSION
18269 but the BN code had some problems that would cause failures when
18270 doing certificate verification and some other functions.
18272 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
18274 * Add ASN1 and PEM code to support netscape certificate sequences.
18278 * Add ASN1 and PEM code to support netscape certificate sequences.
18282 * Add several PKIX and private extended key usage OIDs.
18286 * Modify the 'ca' program to handle the new extension code. Modify
18287 openssl.cnf for new extension format, add comments.
18291 * More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
18292 and add a sample to openssl.cnf so req -x509 now adds appropriate
18297 * Continued X509 V3 changes. Add to other makefiles, integrate with the
18298 error code, add initial support to X509_print() and x509 application.
18302 * Takes a deep breath and start adding X509 V3 extension support code. Add
18303 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
18304 stuff is currently isolated and isn't even compiled yet.
18308 * Continuing patches for GeneralizedTime. Fix up certificate and CRL
18309 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
18310 Removed the versions check from X509 routines when loading extensions:
18311 this allows certain broken certificates that don't set the version
18312 properly to be processed.
18316 * Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
18317 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
18318 can still be regenerated with "make depend".
18322 * Spelling mistake in C version of CAST-128.
18324 *Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>*
18326 * Changes to the error generation code. The perl script err-code.pl
18327 now reads in the old error codes and retains the old numbers, only
18328 adding new ones if necessary. It also only changes the .err files if new
18329 codes are added. The makefiles have been modified to only insert errors
18330 when needed (to avoid needlessly modifying header files). This is done
18331 by only inserting errors if the .err file is newer than the auto generated
18332 C file. To rebuild all the error codes from scratch (the old behaviour)
18333 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
18334 or delete all the .err files.
18338 * CAST-128 was incorrectly implemented for short keys. The C version has
18339 been fixed, but is untested. The assembler versions are also fixed, but
18340 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
18341 to regenerate it if needed.
18342 *Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
18343 Hagino <itojun@kame.net>*
18345 * File was opened incorrectly in randfile.c.
18347 *Ulf Möller <ulf@fitug.de>*
18349 * Beginning of support for GeneralizedTime. d2i, i2d, check and print
18350 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
18351 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
18352 al: it's just almost always a UTCTime. Note this patch adds new error
18353 codes so do a "make errors" if there are problems.
18357 * Correct Linux 1 recognition in config.
18359 *Ulf Möller <ulf@fitug.de>*
18361 * Remove pointless MD5 hash when using DSA keys in ca.
18363 *Anonymous <nobody@replay.com>*
18365 * Generate an error if given an empty string as a cert directory. Also
18366 generate an error if handed NULL (previously returned 0 to indicate an
18367 error, but didn't set one).
18369 *Ben Laurie, reported by Anonymous <nobody@replay.com>*
18371 * Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
18375 * Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
18376 parameters. This was causing a warning which killed off the Win32 compile.
18380 * Remove C++ style comments from crypto/bn/bn_local.h.
18382 *Neil Costigan <neil.costigan@celocom.com>*
18384 * The function OBJ_txt2nid was broken. It was supposed to return a nid
18385 based on a text string, looking up short and long names and finally
18386 "dot" format. The "dot" format stuff didn't work. Added new function
18387 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
18388 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
18389 OID is not part of the table.
18393 * Add prototypes to X509 lookup/verify methods, fixing a bug in
18394 X509_LOOKUP_by_alias().
18398 * Sort openssl functions by name.
18402 * Get the `gendsa` command working and add it to the `list` command. Remove
18403 encryption from sample DSA keys (in case anyone is interested the password
18408 * Make *all* `*_free` functions accept a NULL pointer.
18410 *Frans Heymans <fheymans@isaserver.be>*
18412 * If a DH key is generated in s3_srvr.c, don't blow it by trying to use
18415 *Anonymous <nobody@replay.com>*
18417 * s_server should send the CAfile as acceptable CAs, not its own cert.
18419 *Bodo Moeller <3moeller@informatik.uni-hamburg.de>*
18421 * Don't blow it for numeric `-newkey` arguments to `apps/req`.
18423 *Bodo Moeller <3moeller@informatik.uni-hamburg.de>*
18425 * Temp key "for export" tests were wrong in s3_srvr.c.
18427 *Anonymous <nobody@replay.com>*
18429 * Add prototype for temp key callback functions
18430 SSL_CTX_set_tmp_{rsa,dh}_callback().
18434 * Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
18435 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
18439 * X509_name_add_entry() freed the wrong thing after an error.
18441 *Arne Ansper <arne@ats.cyber.ee>*
18443 * rsa_eay.c would attempt to free a NULL context.
18445 *Arne Ansper <arne@ats.cyber.ee>*
18447 * BIO_s_socket() had a broken should_retry() on Windoze.
18449 *Arne Ansper <arne@ats.cyber.ee>*
18451 * BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
18453 *Arne Ansper <arne@ats.cyber.ee>*
18455 * Make sure the already existing X509_STORE->depth variable is initialized
18456 in X509_STORE_new(), but document the fact that this variable is still
18457 unused in the certificate verification process.
18459 *Ralf S. Engelschall*
18461 * Fix the various library and `apps/` files to free up pkeys obtained from
18462 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
18466 * Fix reference counting in X509_PUBKEY_get(). This makes
18467 demos/maurice/example2.c work, amongst others, probably.
18469 *Steve Henson and Ben Laurie*
18471 * First cut of a cleanup for `apps/`. First the `ssleay` program is now named
18472 `openssl` and second, the shortcut symlinks for the `openssl <command>`
18473 are no longer created. This way we have a single and consistent command
18474 line interface `openssl <command>`, similar to `cvs <command>`.
18476 *Ralf S. Engelschall, Paul Sutton and Ben Laurie*
18478 * ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
18479 BIT STRING wrapper always have zero unused bits.
18483 * Add CA.pl, perl version of CA.sh, add extended key usage OID.
18487 * Make the top-level INSTALL documentation easier to understand.
18491 * Makefiles updated to exit if an error occurs in a sub-directory
18492 make (including if user presses ^C) [Paul Sutton]
18494 * Make Montgomery context stuff explicit in RSA data structure.
18498 * Fix build order of pem and err to allow for generated pem.h.
18502 * Fix renumbering bug in X509_NAME_delete_entry().
18506 * Enhanced the err-ins.pl script so it makes the error library number
18507 global and can add a library name. This is needed for external ASN1 and
18508 other error libraries.
18512 * Fixed sk_insert which never worked properly.
18516 * Fix ASN1 macros so they can handle indefinite length constructed
18517 EXPLICIT tags. Some non standard certificates use these: they can now
18522 * Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
18523 into a single doc/ssleay.txt bundle. This way the information is still
18524 preserved but no longer messes up this directory. Now it's new room for
18525 the new set of documentation files.
18527 *Ralf S. Engelschall*
18529 * SETs were incorrectly DER encoded. This was a major pain, because they
18530 shared code with SEQUENCEs, which aren't coded the same. This means that
18531 almost everything to do with SETs or SEQUENCEs has either changed name or
18532 number of arguments.
18534 *Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>*
18536 * Fix test data to work with the above.
18540 * Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
18541 was already fixed by Eric for 0.9.1 it seems.
18543 *Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>*
18545 * Autodetect FreeBSD3.
18549 * Fix various bugs in Configure. This affects the following platforms:
18553 unixware-2.0-pentium
18558 * Eliminate generated files from CVS. Reorder tests to regenerate files
18559 before they are needed.
18563 * Generate Makefile.ssl from Makefile.org (to keep CVS happy).
18567 ### Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
18569 * Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
18570 changed SSLeay to OpenSSL in version strings.
18572 *Ralf S. Engelschall*
18574 * Some fixups to the top-level documents.
18578 * Fixed the nasty bug where rsaref.h was not found under compile-time
18579 because the symlink to include/ was missing.
18581 *Ralf S. Engelschall*
18583 * Incorporated the popular no-RSA/DSA-only patches
18584 which allow to compile a RSA-free SSLeay.
18586 *Andrew Cooke / Interrader Ldt., Ralf S. Engelschall*
18588 * Fixed nasty rehash problem under `make -f Makefile.ssl links`
18589 when "ssleay" is still not found.
18591 *Ralf S. Engelschall*
18593 * Added more platforms to Configure: Cray T3E, HPUX 11,
18595 *Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>*
18597 * Updated the README file.
18599 *Ralf S. Engelschall*
18601 * Added various .cvsignore files in the CVS repository subdirs
18602 to make a "cvs update" really silent.
18604 *Ralf S. Engelschall*
18606 * Recompiled the error-definition header files and added
18607 missing symbols to the Win32 linker tables.
18609 *Ralf S. Engelschall*
18611 * Cleaned up the top-level documents;
18612 o new files: CHANGES and LICENSE
18613 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
18614 o merged COPYRIGHT into LICENSE
18615 o removed obsolete TODO file
18616 o renamed MICROSOFT to INSTALL.W32
18618 *Ralf S. Engelschall*
18620 * Removed dummy files from the 0.9.1b source tree:
18621 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
18622 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
18623 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
18624 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
18625 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
18627 *Ralf S. Engelschall*
18629 * Added various platform portability fixes.
18633 * The Genesis of the OpenSSL rpject:
18634 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
18635 Young and Tim J. Hudson created while they were working for C2Net until
18638 *The OpenSSL Project*
18640 ### Changes between 0.9.0b and 0.9.1b [not released]
18642 * Updated a few CA certificates under certs/
18646 * Changed some BIGNUM api stuff.
18650 * Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
18651 DGUX x86, Linux Alpha, etc.
18655 * New COMP library [crypto/comp/] for SSL Record Layer Compression:
18656 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
18661 * Add -strparse option to asn1pars program which parses nested
18664 *Dr Stephen Henson <shenson@bigfoot.com>*
18666 * Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
18670 * DSA fix for "ca" program.
18674 * Added "-genkey" option to "dsaparam" program.
18678 * Added RIPE MD160 (rmd160) message digest.
18682 * Added -a (all) option to "ssleay version" command.
18686 * Added PLATFORM define which is the id given to Configure.
18690 * Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
18694 * Extended the ASN.1 parser routines.
18698 * Extended BIO routines to support REUSEADDR, seek, tell, etc.
18702 * Added a BN_CTX to the BN library.
18706 * Fixed the weak key values in DES library
18710 * Changed API in EVP library for cipher aliases.
18714 * Added support for RC2/64bit cipher.
18718 * Converted the lhash library to the crypto/mem.c functions.
18722 * Added more recognized ASN.1 object ids.
18726 * Added more RSA padding checks for SSL/TLS.
18730 * Added BIO proxy/filter functionality.
18734 * Added extra_certs to SSL_CTX which can be used
18735 send extra CA certificates to the client in the CA cert chain sending
18736 process. It can be configured with SSL_CTX_add_extra_chain_cert().
18740 * Now Fortezza is denied in the authentication phase because
18741 this is key exchange mechanism is not supported by SSLeay at all.
18745 * Additional PKCS1 checks.
18749 * Support the string "TLSv1" for all TLS v1 ciphers.
18753 * Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
18754 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
18758 * Fixed a few memory leaks.
18762 * Fixed various code and comment typos.
18766 * A minor bug in ssl/s3_clnt.c where there would always be 4 0
18767 bytes sent in the client random.
18769 *Edward Bishop <ebishop@spyglass.com>*
18773 [CVE-2020-1971]: https://www.openssl.org/news/vulnerabilities.html#CVE-2020-1971
18774 [CVE-2020-1967]: https://www.openssl.org/news/vulnerabilities.html#CVE-2020-1967
18775 [CVE-2019-1563]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1563
18776 [CVE-2019-1559]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1559
18777 [CVE-2019-1552]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1552
18778 [CVE-2019-1551]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1551
18779 [CVE-2019-1549]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1549
18780 [CVE-2019-1547]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1547
18781 [CVE-2019-1543]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1543
18782 [CVE-2018-5407]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-5407
18783 [CVE-2018-0739]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0739
18784 [CVE-2018-0737]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0737
18785 [CVE-2018-0735]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0735
18786 [CVE-2018-0734]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0734
18787 [CVE-2018-0733]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0733
18788 [CVE-2018-0732]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0732
18789 [CVE-2017-3738]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3738
18790 [CVE-2017-3737]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3737
18791 [CVE-2017-3736]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3736
18792 [CVE-2017-3735]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3735
18793 [CVE-2017-3733]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3733
18794 [CVE-2017-3732]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3732
18795 [CVE-2017-3731]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3731
18796 [CVE-2017-3730]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3730
18797 [CVE-2016-7055]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7055
18798 [CVE-2016-7054]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7054
18799 [CVE-2016-7053]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7053
18800 [CVE-2016-7052]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7052
18801 [CVE-2016-6309]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6309
18802 [CVE-2016-6308]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6308
18803 [CVE-2016-6307]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6307
18804 [CVE-2016-6306]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6306
18805 [CVE-2016-6305]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6305
18806 [CVE-2016-6304]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6304
18807 [CVE-2016-6303]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6303
18808 [CVE-2016-6302]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6302
18809 [CVE-2016-2183]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2183
18810 [CVE-2016-2182]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2182
18811 [CVE-2016-2181]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2181
18812 [CVE-2016-2180]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2180
18813 [CVE-2016-2179]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2179
18814 [CVE-2016-2178]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2178
18815 [CVE-2016-2177]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2177
18816 [CVE-2016-2176]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2176
18817 [CVE-2016-2109]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2109
18818 [CVE-2016-2107]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2107
18819 [CVE-2016-2106]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2106
18820 [CVE-2016-2105]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2105
18821 [CVE-2016-0800]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0800
18822 [CVE-2016-0799]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0799
18823 [CVE-2016-0798]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0798
18824 [CVE-2016-0797]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0797
18825 [CVE-2016-0705]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0705
18826 [CVE-2016-0702]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0702
18827 [CVE-2016-0701]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0701
18828 [CVE-2015-3197]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3197
18829 [CVE-2015-3196]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3196
18830 [CVE-2015-3195]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3195
18831 [CVE-2015-3194]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3194
18832 [CVE-2015-3193]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3193
18833 [CVE-2015-1793]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1793
18834 [CVE-2015-1792]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1792
18835 [CVE-2015-1791]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1791
18836 [CVE-2015-1790]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1790
18837 [CVE-2015-1789]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1789
18838 [CVE-2015-1788]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1788
18839 [CVE-2015-1787]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1787
18840 [CVE-2015-0293]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0293
18841 [CVE-2015-0291]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0291
18842 [CVE-2015-0290]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0290
18843 [CVE-2015-0289]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0289
18844 [CVE-2015-0288]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0288
18845 [CVE-2015-0287]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0287
18846 [CVE-2015-0286]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0286
18847 [CVE-2015-0285]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0285
18848 [CVE-2015-0209]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0209
18849 [CVE-2015-0208]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0208
18850 [CVE-2015-0207]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0207
18851 [CVE-2015-0206]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0206
18852 [CVE-2015-0205]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0205
18853 [CVE-2015-0204]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0204
18854 [CVE-2014-8275]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-8275
18855 [CVE-2014-5139]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-5139
18856 [CVE-2014-3572]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3572
18857 [CVE-2014-3571]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3571
18858 [CVE-2014-3570]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3570
18859 [CVE-2014-3569]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3569
18860 [CVE-2014-3568]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3568
18861 [CVE-2014-3567]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3567
18862 [CVE-2014-3566]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3566
18863 [CVE-2014-3513]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3513
18864 [CVE-2014-3512]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3512
18865 [CVE-2014-3511]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3511
18866 [CVE-2014-3510]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3510
18867 [CVE-2014-3509]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3509
18868 [CVE-2014-3508]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3508
18869 [CVE-2014-3507]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3507
18870 [CVE-2014-3506]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3506
18871 [CVE-2014-3505]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3505
18872 [CVE-2014-3470]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3470
18873 [CVE-2014-0224]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0224
18874 [CVE-2014-0221]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0221
18875 [CVE-2014-0195]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0195
18876 [CVE-2014-0160]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0160
18877 [CVE-2014-0076]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0076
18878 [CVE-2013-6450]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-6450
18879 [CVE-2013-4353]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-4353
18880 [CVE-2013-0169]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-0169
18881 [CVE-2013-0166]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-0166
18882 [CVE-2012-2686]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2686
18883 [CVE-2012-2333]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2333
18884 [CVE-2012-2110]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2110
18885 [CVE-2012-0884]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0884
18886 [CVE-2012-0050]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0050
18887 [CVE-2012-0027]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0027
18888 [CVE-2011-4619]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4619
18889 [CVE-2011-4577]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4577
18890 [CVE-2011-4576]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4576
18891 [CVE-2011-4109]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4109
18892 [CVE-2011-4108]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4108
18893 [CVE-2011-3210]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-3210
18894 [CVE-2011-3207]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-3207
18895 [CVE-2011-0014]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-0014
18896 [CVE-2010-4252]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-4252
18897 [CVE-2010-4180]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-4180
18898 [CVE-2010-3864]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-3864
18899 [CVE-2010-1633]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-1633
18900 [CVE-2010-0740]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-0740
18901 [CVE-2010-0433]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-0433
18902 [CVE-2009-4355]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-4355
18903 [CVE-2009-3555]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-3555
18904 [CVE-2009-3245]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-3245
18905 [CVE-2009-1386]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1386
18906 [CVE-2009-1379]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1379
18907 [CVE-2009-1378]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1378
18908 [CVE-2009-1377]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1377
18909 [CVE-2009-0789]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0789
18910 [CVE-2009-0591]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0591
18911 [CVE-2009-0590]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0590
18912 [CVE-2008-5077]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-5077
18913 [CVE-2008-1678]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-1678
18914 [CVE-2008-1672]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-1672
18915 [CVE-2008-0891]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-0891
18916 [CVE-2007-5135]: https://www.openssl.org/news/vulnerabilities.html#CVE-2007-5135
18917 [CVE-2007-4995]: https://www.openssl.org/news/vulnerabilities.html#CVE-2007-4995
18918 [CVE-2006-4343]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-4343
18919 [CVE-2006-4339]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-4339
18920 [CVE-2006-3738]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-3738
18921 [CVE-2006-2940]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-2940
18922 [CVE-2006-2937]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-2937
18923 [CVE-2005-2969]: https://www.openssl.org/news/vulnerabilities.html#CVE-2005-2969
18924 [CVE-2004-0112]: https://www.openssl.org/news/vulnerabilities.html#CVE-2004-0112
18925 [CVE-2004-0079]: https://www.openssl.org/news/vulnerabilities.html#CVE-2004-0079
18926 [CVE-2003-0851]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0851
18927 [CVE-2003-0545]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0545
18928 [CVE-2003-0544]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0544
18929 [CVE-2003-0543]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0543
18930 [CVE-2003-0078]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0078
18931 [CVE-2002-0659]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0659
18932 [CVE-2002-0657]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0657
18933 [CVE-2002-0656]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0656
18934 [CVE-2002-0655]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0655